53380 Understanding Sector Performance: The Case of Utilities in Latin America and the Caribbean June 29, 2009 Sustainable Development Department Economics Unit Latin America and the Caribbean Region Document of the World Bank i ii T ABLE OF C ONTENTS ACKNOWLEDGEMENTS ......................................................................................................................... III ABBREVIATIONS AND ACRONYMS..................................................................................................... IV EXECUTIVE SUMMARY ............................................................................................................................ V 1. INTRODUCTION .................................................................................................................................. 1 1.1. ANALYTICAL FRAMEW ORK AND SCOPE ........................................................... 4 2. BENCHMARKING LAC'S UTILITY PERFORMANCE ............................................................... 8 2.1. ELECTRICITY DISTRIBUTION .................................................................................. 11 2.1.1. ELECTRICITY DISTRIBUTION BENCHMARKING RESULTS ...................... 12 2.2. W ATER AND SANITATION SECTOR ..................................................................... 15 2.2.1. W ATER AND SANITATION BENCHMARKING RESULTS.............................. 16 2.3. FIXED TELECOMMUNICATIONS SECTOR ........................................................ 18 2.3.1. FIXED TELECOMMUNICATION BENCHMARKING RESULTS ................... 18 2.4. PUBLIC VS PRIVATE BENCHMARKING - ELECTRICITY DISTRIBU TION ............................................................................................................................................. 19 2.5. FINAL REMARKS ......................................................................................................... 21 3. UNDERSTANDING THE IMPACT OF PRIVATE SECTOR PARTICIPATION ON PERFORMANCE OF UTILITIES ................................................................................................... 23 3.1. THE IMPACT OF PSP IN ELECTRICITY DISTRIBUTION .............................. 26 3.2. THE IMPACT OF PSP ON W ATER AND SEW ERAGE....................................... 33 3.3. THE IMPACT OF PSP ON FIXED-LINE TELECOMMUNICATIONS ............ 40 3.4. THE IMPACT OF CONTRACTS DESIGN .............................................................. 47 3.5. MAIN CONCLUSIONS ................................................................................................. 49 4. REGULATORY INSTITUTIONAL DESIGN AND SECTOR PERFORMANCE ..................... 50 4.1. BENCHMARKING OF REGU LATORY INSTITUTIONAL DESIGN .............. 51 4.2. RESULTS OF BENCHMARKING OF AGENCIES AT R EGIONAL LEVEL . 54 4.3. RESULTS OF REGULATOR Y GOVERNANCE BENCHMARKING AT AGENCY LEVEL ........................................................................................................... 55 4.4. RESULTS ON DIFFERENT DIMENSIONS ............................................................. 58 4.5. REGULATORY GOVERNANC E AND SECTOR PERFORMANCE ................ 61 4.6. RESULTS .......................................................................................................................... 61 4.7. CONCLUSIONS .............................................................................................................. 64 5. CORPORATE GOVERNANCE FOR STATE-OWNED ENTERPRISES .................................. 66 5.1. METHODOLOGY/FRAMEW OR K OF ANALYSIS .............................................. 68 5.2. RESULTS OF CORPORATE GOVERNANCE BENCHMARKING .................. 70 i 5.3. CORPORATE GOVERNANCE AND PERFORMANCE ...................................... 79 5.4. CONCLUSIONS .............................................................................................................. 82 6. OTHER DETERMINANTS FOR SECTOR PERFORMANCE.................................................... 84 7. CONCLUSIONS ................................................................................................................................... 89 ANNEX 1: EMPIRICAL APPROACH:................................................................................................... 95 ANNEX 2: DATA SETS .............................................................................................................................. 97 ANNEX 3: BENCHMARKIN G ANALYSIS .........................................................................................107 ANNEX 4: DETAILED RESULTS OF THE EMPIRIC AL ANALYSIS ........................................123 ANNEX 5: REGULATORY GOVERNANCE DIMENSIONS ..................................................... - 135 - ANNEX 6: REGULATORY GOVERNANCE AND PERFORMANCE .................................... - 139 - ANNEX 7: CORPORATE GOVERNANCE AND PERFOR MANCE.............................................142 REFERENCES ............................................................................................................................................145 ii ACKNOWLEDGEMENTS Understanding Sector Performance: The Case of Utilities in Latin America and the Caribbean is the product of a team effort by the Economics Unit at the Sustainable Development Department in the Latin America and Caribbean Region of the World Bank, co-led by Luis A. Andrés (LCSSD) and J. Luis Guasch (LCSSD). The team gratefully acknowledges the support of Diana Cubas, Barbara Cunha, Jose Guillermo Diaz, Georgeta Dragoiu, Raquel Fernandez, Julio Gonzalez, Alejandro Guerrero, and Maria Claudia Pachon. The team would also thank Sebastian Lopez Azumendi who coauthored the background papers for Chapters 4 and 5. The team received valuable feedback through a rich consultation and peer review process. The team values the ongoing support and technical inputs from the regional Chief Economist, Augusto de la Torre, and his team, including Tito Cordella, Pablo Fajnzylber, and William Maloney. The team also gratefully acknowledges the early inputs from the Sustainable Development Department team of Latin America and the Caribbean (LCSSD), which contributed its ideas and suggestions via several meetings and workshops, and subsequently through comments on earlier versions of the chapters and at a seminar on the main findings and messages of the report. The team appreciated early inputs from Daniel Benitez, Philippe Benoit, Susan Bogach, Juan Miguel Cayo, Makhtar Diop, Joshua Gallo, Manuel Mariño, Martin Rossi, Tomas Serebrisky, Tova Solo, Maria Angelica Sotomayor, and Carlos Velez. Insightful and constructive comments were received as well from peer reviewers, Marianna Fay, Antonio Estache, Maximo Torero, and Maria Vagliasindi. The team also gratefully acknowledges the guidance provided by Jordan Schwartz. The team is grateful for the financial support received from: the LAC Chief Economist Office for the overall preparation of this report, ESMAP for funding the data collection of the performance data in the electricity distribution sector, PPIAF for its support to the background material for Chapter 3, and DFID for its support to the background material on Corporate Governance of State-Owned Enterprises. The findings, interpretations, and conclusions expressed in this document are those of the authors and do not necessarily reflect the views of the Executive Directors of The World Bank, the governments they represent, or the counterparts consulted or engaged with during the informality study process. Any factual errors are, as well, the responsibility of the team. iii ABBREVIATIONS AND ACRONYMS CAPEX Capital Expenditures ERGI Electricity Regulatory Governance Index GDP Gross Domestic Product GWh Giga Watt Hour IEA International Energy Agency ICT Information and Communication technology IRA Independent Regulatory Agency ITU International Telecommunication Union kWh kilo Watt hour LAC Latin American and the Caribbean M3 Cubic meters MDG Millennium Development Goals MW MegaWatt MWh Mega Watt Hours NEC Chile's National Energy Commission OECD Organization for Economic Co-operation and Development OLADE Organización Latinoamericana de Energía OPEX Operational Expenditures PERs Public Expenditure Reviews PPI Private Participation in Infrastructure PPP Private Public Partnership PPIAF Public-Private Infrastructure Advisory Facility PSP Private Sector Participation SIN National Interconnected System (Sistema Nacional Interconectado) SOE State-owned enterprise TOTEX Total Expenditures Vice President: Pamela Cox Regional Chief Economist: Augusto de la Torre Sector Director: Laura Tuck Project Co-Team Leaders: Luis A. Andres and J. Luis Guasch iv EXECUTIVE SUMMARY This report provides innovative approaches to better understand infrastructure sector performance by focusing on the links between key indicators for private and public utilities, and changes in ownership, regulatory agency governance, and corporate governance, among other dimensions. By linking inputs and outputs over the last 15 years, the analysis proposes key determinants that have impacted sector performance in infrastructure sectors in Latin America and the Caribbean (LAC). It is about understanding how and to what extent the effect of such elements result in significant changes in the performance of infrastructure service provision. The lack of adequate infrastructure is hampering the region's ability to grow, compete, and reduce poverty 1. By understanding the various interventions and conditions that explain LAC sector performance, this report proposes a framework of analysis that addresses elements key to the design of mechanisms that contribute to minimizing the region's infrastructure gap. For a region as diverse as LAC, there is no one model that fits all solution for improving service provision; what this report does show however, is that improving performance requires a comprehensive approach that integrates mechanisms to address the different components of sector performance. This report focuses on the distribution segment of basic infrastructure services. It covers electricity distribution, water distribution and sewerage, and fixed telecommunications. It uses previously unavailable data on performance of utility companies. Furthermore, data was collected through surveys sent to regulatory agencies and state-owned enterprises (SOEs) throughout the region. The entire analysis undertaken for this report is based on a dataset specifically constructed for this purpose. For most of the analysis, the data collected is original and has not been used previously 2. The wealth of information pulled together through this exercise lends itself to further far-reaching analysis. A simple, yet comprehensive, web- based analysis tool allows the reader to conduct ad hoc queries and download information used for regression analyses. By making this information available to a broader audience, this report hopes that such benchmarking efforts provide a regional and utility-level frame of reference for good and/or poor sector performance in LAC. The methodology we propose allows for analyzing trends from cross sector comparisons, for features common to all three sectors. It develops a useful tool to understand the determinants of sector performance. Particularly, we emphasize the importance of not only increasing regional infrastructure stocks, but also striving for improvements in sector performance. This report answer key questions on utility performance, and poses further challenges for the region. More specifically, the report does: i) Depict sector performance with a broad set of indicators that describes the current situation as well its evolution during the last 15 years; ii) Propose analytical frameworks for themes less developed in the literature such as regulatory governance and corporate governance for SOEs; iii) Benchmark the institutional designs of the regulatory agencies in the region for the water and electricity sectors; and iv) Analyze the relationship between sector performance and regulation, private sector participation, and corporate governance. The report, however, does not: i) Presume to describe all the possible factors and resulting conditions that may impact sector performance; ii) Focus on the external environment, 1 Fay and Morrison 2007. 2 Exception to this is the ITU database for telecommunications and the contract design database (Guasch 2004). v which cannot be changed/impacted by the change in sector performance; and iii) Analyze other factors that may belong to the sector or utility environments but, for different reasons may not be standardized. Benchmarking assessments have long been used to compare performance amongst utility companies. For the LAC region, previous analysis has been reduced to using small samples of companies, in limited countries and limited indicators, within data constraints. This initiative is more ambitious. By collecting information for over 250 electricity distribution companies, and more than 1,700 water and sanitation companies, this study is able to more comprehensively analyze sector performance, through service provider performance. The report focuses on the relationship between sector performance and the following determinants: ownership structure, regulatory agencies, and corporate governance. It also includes other related conditions such as contract design and market structure. The databases are rich, not only in the number and types of utilities surveyed, but also in the diversity and comprehensiveness of the collected indicators. For electricity and water alone, over 20 indicators for each sector, were collected. The benchmarking assessments are based on the following performance indicators: output, coverage, labor productivity, inputs, operating performance, service quality, and prices show the details that have shaped the electricity, the water and sanitation, and the fixed telecommunication sectors during the last decade. Furthermore, detailed surveys were filled by a large group of regulatory agencies and SOEs in the region. Each questionnaire, of over 100 questions each, requested information of key aspects that were then used to assess the current situation and find the implications of key determinants on sector performance. The analysis performed in this report emphasizes the effects of key determinants in sector performance. As a result, this report offers three key messages: M ESSAGE 1: Sector performance for electricity distribution, water and sanitation, and fixed telecommunications significantly improved in LAC but there is still much room for improvement. Sector performance for electricity distribution, water and sanitation, and fixed telecommunications significantly improved in LAC. During the last 15 years, the region has witnessed significant improvements especially in coverage, service quality, and labor productivity in all sectors. Between 1990 and 2005, LAC's coverage increased to 95 percent for electricity distribution, 97 percent for the water utilities within our sample 3, and 62 percent in fixed telecommunications. Regional coverage is close to 92 percent in electricity, 80 percent in water, and 62 percent in fixed telecommunications for those households with access to these services. A similar pattern of improvement is visible for all three sectors for labor productivity. For electricity distribution, labor productivity doubled since 1995 and for water, labor productivity almost doubled from 252 connections per employee to 425 in 2006. When measuring labor productivity for telecommunications, the sector has experienced a seven-fold increase between 1995 and 2007. 3 Our database includes 59 percent of the water connections in LAC. vi Another indicator that reflects a leap forward for LAC in utility performance is quality of service. For electricity distribution, the quality indicators exhibit significant improvements as measured by the frequency and duration of interruptions per connection showing a 42 and 40 percent reduction accordingly. The water sector also experienced a significant 8 percent increase in the continuity of service during this time period. Last but not least, the telecommunications sector reports gradual but significant increments in the percent of digital main lines, and telephone faults cleared by the next working day. For example, digital main lines have increased from 63 percent in 1995 to 100 percent in 2007. Comparably, the number of telephone faults per 100 main fixed lines per year dropped from 23 to 8 by the end of 2007. Such improvements in quality have also been accompanied by a reduction in the waiting list for main fixed lines, which by 2007 averaged to zero. Sector performance in LAC is also a story of diversity in the types of service providers; the specific environment and conditions in which they operate, and the impact on performance. Ranking companies and comparing top-middle-bottom performers for subsets of providers enables the benchmarking of companies not only at the country level, but also at the utility level. The wealth of data used to benchmark utility performance in LAC concludes throughout the past 15 years, the region has hosted a wide range of good and poor performers. For example, in water and sanitation, the top ten percent performers have 100 percent water and sanitation coverage on average. In contrast, the bottom ten percent water utilities average 66 percent coverage, while sewerage utilities average a low 15 percent coverage. Another example shows that electricity distribution utilities in the top ten percentile were ten times more productive in 2005 and sold six times the amount of energy (per connection) compared to utilities in the bottom ten percent. Finally, on average private utilities outperformed public utilities--but more importantly, there are good public and private utilities and underperforming private and public utilities. For several indicators, the average of the top 10 percent public utilities performed better than the average private utilities and in other cases the bottom 10 percent of the private utilities performed worse than the average public utilities. In the case of distribution losses, it is noteworthy that the public utilities in the bottom 10 percent perform better than the average private utilities. Likewise the private utilities forming the top decile experience more distribution losses than the average public utilities By identifying key indicators where poor performers lag and by serving as a tool for utilities to compare their performance and strive to achieve top performance, the benchmarking exercise and data allow utilities to target improvements towards those areas in which they lag the most. For even a few top performers, improvement can be achieved through a careful analysis of selected indicators. LAC is performing well in contrast to other comparable regions; even so millions of people still lack access to basic services. In 2007, the (weighted) average for phone penetration (Mobile and fixed-line telephone subscribers per 100 people) worldwide was 70.8; for MIC it was 64.3, for EAP it was 66.7, while for LAC it was 85.1 4. In 2004, LAC's household water coverage was 80 percent; the worldwide coverage was only 54 percent. More so, the coverage for East and South Asia was 70 and 20 percent, respectively; and Africa's water coverage was at a low of 26 percent. 5 Despite the fact that electricity coverage in LAC increased from 84.7 to 94.6 percent 6 in 2005 in our sample, there are still many people, almost all poor and in rural areas, without electricity. There are 29 million additional households that do not have a water connection. These 4 Calculations by authors using data from the ITU dataset. 5 JMP data. 6 These regional estimates correspond to the weighted average across the 250 utilities in the sample that represents 89 percent of the total number of electricity connections. vii figures present a strong need to expand electrification and water and sanitation services in rural areas in LAC countries since these areas lag behind. The differences in performance amongst utilities within countries, and within providers with similar characteristics, pose a number of questions about key determinants of sector performance. Has PSP in service provision changed the dynamics of the sector? Does the type of regulation and more importantly the way it is govern, affect utility performance? Do corporate govern ance frameworks that provide SOEs similar incentives to privately held providers positively affect performance? Previous research endeavors have few answers to these important questions. Differences in ownership, regulatory governance, and corporate governance of SOEs explain some of the dispersion on utilities' performance. M ESSAGE 2: Results show that both the government (as a regulator and a service provider) and the private sector (as a service provider) can play an active role in enhancing sector performance. a. When carefully designed and implemented, Private Sector Participation in service provision has a positive effect on sector performance. When analyzing PSP in service provision, this report presents a comprehensive and systemic assessment of the impact of PSP in LAC to date. The report considers what happened before, during, and after the change in ownership in three sectors--electricity, water, and telecommunications--by focusing on a range of performance variables. It is necessary to look at all these three periods, because often the most dramatic effects of PSP are found in the transition period, when the enterprise is overhauled as part of the transaction process. The report also focuses on changes and rates of change in the three different periods, rather than on absolute numbers, because in many cases, the performance variables exhibit natural changes over time (with or without private sector participation). Hence, the analysis controls for such naturally occurring rates of change. The changes associated with PSP have had a significant positive effect on labor productivity, efficiency, and quality of the service. In addition, for telecommunications, PSP had significant effects on output and coverage. After controlling for firm specific time trends, there do not appear to be significant impacts on output and coverage, but prices tended to increase somewhat although the picture is highly variable across sectors. For electricity, labor productivity ended up being twice as high for private utilities than that of public utilities. Distribution losses improved in private utilities 12 percent, while public utilities saw their performance deteriorate by 5 percent. For continuity of service, both groups started at around 24 interruptions per year. The private utilities reduced this to around 12 compared with a reduction to around 19 interruptions for public utilities. Similarly, public utilities saw the average duration of their outages increase by almost 50 percent compared with a reduction of almost 30 percent from the private utilities, from a similar starting value. It is worth noting the differences between publicly and privately operated distribution utilities which occurred primarily in regard to labor productivity, distribution losses, quality of service, and tariffs. The average of the top 10 percent of performers in the public utility group outperformed the average private utility, and the average private utility outperformed the bottom ten percent of the private utility group. In the case of distribution losses, it is noteworthy that the public utilities in the bottom 10 percent perform better than the average private utilities. Likewise, viii the private utilities forming the top decile experience more distribution losses than the average public utilities. By introducing a number of PSP contracts and process variables, we analyze how distinct PSP design variables can have a different impact on performance outcomes. Depending on the priorities of a country when considering the change in ownership, certain PSP contract characteristics might be more important than others: i) contract characteristics matter: the way of PSP is undertaken can generate significant performance differences; ii) each contract characteristic affects each performance variable differently. In other words, a certain contract characteristic could have a positive influence on one performance variable while having a negative or insignificant impact on another; and iii) some contract variables have bigger impacts than others. b. An Independent Regulatory Agency (IRA) designed to be transparent, accountable , and free of political interference contributes positively to sector performance. We find that the existence of a regulatory agency has significant impact on sector performance. Under the presence of a regulatory agency, utilities increased labor productivity from 18.2 to 19.4 percent. Similarly, utilities reported 18.9 percent less average interruption duration and 17.3 percent less frequency of interruptions. Furthermore, operational expenses, residential tariffs and the cost recovery ratio had a positive change. The experience of the regulatory agency also contributes to improvements in performance. For instance, after controlling for changes in ownership, utilities resulted with 1.4 additional increments per year in labor productivity. Similarly, for distribution losses there was a 1.8 percent reduction per year. Finally, quality improved about 9 percent annually. The principal component analysis reveals that different elements of the regulatory governance design influence performance indicators differently. Changes in the formal component of regulatory governance positively affect labor productivity, frequency of interruptions, and residential tariffs. Changes related to the formal autonomy and the attributions of the agency in terms of tariff setting are associated with higher labor productivity levels and reductions in the average duration of service interruptions. Thus, improvements in sector performance come from introducing in an IRA with characteristics that promote transparency, autonomy, independence, and accountability. The results are consistent with the literature on the impact of PSP and show the relevance of the existence of a regulatory agency and its governance, defined as the agency's institutional design and structure that allows it to carry its functions as an independent regulator. Furthermore, our results indicate a significant improvement in utility performance through the involvement of a regulatory agency even in the case of SOEs. c. A strong accountability mechanism that prevents discriminatory management is fundamental for improving SOEs performance Corporate Governance arrangements in SOEs in water and electricity present a wide spectrum of designs. While private enterprises are characterized by the adoption of standard corporate strategies, SOEs standards vary depending on countries' institutional systems and the characteristics of the service. Thus, the variety of arrangements calls for a careful systematization of governance practices and the identification of successful experiences. SOEs are part of the public sector and factors of good and bad performance are directly or indirectly related to countries'/provinces' overall governance. ix A best practice corporate governance design for SOEs7 that perform well includes: an independent performance-driven Board of directors, a professional staff, transparency and clear disclosure policies, and a clear mechanism to evaluate performance. A corporate structure that prevents political intervention, rewards performance, and is subject to public scrutiny serves as a benchmark for design comparison. Generally, SOEs are subject to influences of different authorities, particularly during their planning process. Rather than focusing on profit maximization, SOEs emphasize social goals and human capital improvement. Thus, manpower is a critical factor of state enterprises' performance. Moreover, in several cases the company's bureaucracy has built a reputation for good performance that has prevented political interference. Good management of SOEs presents government bureaucrats with different challenges. First and foremost, state enterprises face conflicting goals that affect the establishment of a business strategy. Several departments usually compete to have their agenda prioritized, often at the expense of the company's service. Most importantly, interference in the companies' business adopt informal, ad-hoc, approaches, that prevent the company from making explicit the costs and prevents management from identifying ways to improve efficiency and performance. Because low revenues can be compensated by government subsidies, efforts to make the company sustainable fall to second place. Third, poor accountability systems (either at the regulatory or management levels) prevent the development of an ownership structure that triggers efficient behavior from senior management. It is in this context that accountability emerges as the main governance aspect of SOEs. In the cases of companies with high levels of corruption and inefficiency, accountability systems should prevent discretional management (both from management and political authorities) and create the incentives for good performance. Regulation and performance-based management could be considered complementary ways of achieving these goals. Furthermore, a good checks and balances, such as parliamentary oversight and state auditing, should be built into the governance design. Good corporate governance is associated with high levels of performance. As expected, performance orientation and professional management characteristics seem to be the highest contributors for performance; however, all the other dimensions are associated with some of the performance indicators. M ESSAGE 3: Improving sector performance requires a holistic and case-based approach When considered together or in various interactions, the aforementioned issues they may all influence the performance indicators this study uses, thus affecting utility performance. Either through direct links, for example subsidy mechanisms that result in non-cost recovery tariffs and restrict the firm's financial ability to expand coverage and provide adequate service quality; or through more indirect links, like improving social accountability by introducing a mechanism that can hold service providers more directly accountable to their users for the outcomes of their work; the issues reviewed in this report interact to explain the type of incentive framework utilities use to make management and operation decisions. Our objective was not to fully explain sector performance but to evaluate the main drives while recognizing and acknowledging those other issues that might influence utility behavior and the type of incentives they have to perform efficiently. 7 With a corporatized framework. x Our analysis is based on a number of key dimensions; however there are certainly other elements that can influence and explain sector performance. While the purpose of this report is to focus on particular utility level variables as determinants of sector performance, the report briefly summarizes a number of additional factors and the interaction of some of these factors, as they may impact sector performance. On one hand, academics and researchers have modeled and empirically tested the influence of such issues as corruption, market structure, economies of scope and density, renegotiation, and reputation. On the other hand, some have proposed that other issues like subsidy mechanisms, lack of cost recovery, the political economy of the different sectors and social accountability also play a role in sector performance. Although widely discussed, few econometric studies exist and most analyses rely on comprehensive analytical case studies. In this report, sector performance is defined by a number of variables linked and related to each other. By proposing a new framework of analysis and building a comprehensive data set, this report builds a foundation for innovative research that can explain links and variables for which very little empirical analysis exists but for which much theoretical models and case based evidence exist. The report suggests that by identifying the differences in performance amongst utilities, policy decision makers and utility managers can find ways to improve service provision. The heterogeneity amongst utilities warrants a holistic approach to solving any and all shortcomings in performance. When designing any type of solution to improve sector performance, key determinants like ownership structure, regulatory governance and corporate governance, among some, need to be addressed strategically, and not in isolation. Improving service provision is not easy task, and it requires a well designed and comprehensive strategy. M OVING F ORWARD Improving sector performance goes beyond conducting a comprehensive assessment of a key determinant and proposing specific designs that address issues related to that determinant; it entails an approach that integrates policies that address a wide range of issues, some of which are introduced in detail in this report. The LAC region can afford universal coverage of water, sanitation and electricity if appropriate technologies and standards are used. Scarce resources imply that investments need to focus on bottlenecks in existing systems and not on overall expansion 8. By acknowledging and determining the differences amongst service providers and the environments in which they operate, policy makers can design comprehensive solutions to complex problems in infrastructure service provision. Utility sector performance is a complex undertaking that encompasses a variety of dimensions. Impacts on each of these dimensions are not necessarily straightforward, with differences determined by sector, and internal and external environments. Policy makers considering future sector reforms should first prioritize their performance objectives. Once the objectives are identified, the detailed results presented by the analysis can be mined to determine the circumstances in which those objectives can be achieved. For instance, if a utility prioritizes quality and efficiency over retaining employees, private sector participation would be an attractive option. Similarly, if reducing distributional losses is a key objective, in a SOE, then a sound design of its corporate governance with well designed performance orientation rules can be considered. 8 Fay and Morrison (2007) xi The results presented in this report are instructive to policy makers in terms of highlighting pitfalls in sector reform programs. Poor design and faulty implementation explain many of the shortcomings in reform processes. Identifying the potential for these in advance can assist policy makers in the design of proactive counter measures. Consider the case of an electricity distribution policy maker who has prioritized improving quality and reducing distributional losses--and hence decided to move ahead with PSP. By drawing lessons from the experience detailed in this analysis, the policy maker could design a public relations campaign emphasizing expected benefits and cautioning consumers of potential price increases and reductions in sector employment. As a whole, this report can help policy makers make informed decisions and well designed change strategies, allowing them to maximize both technical and political objectives. By securing an environment that maximizes the benefits of reform and promotes a broad consensus, reform programs in the infrastructure sectors can be successfully implemented. In moving forward, the lessons from the past need to be accounted for and corrected. The ultimate objective is to secure improved sector performance and long-term efficiency, reduce poverty through better concession design and regulation, and foster compliance with the terms agreed to by both the government and the operator. To establish such an environment, concession laws and contracts should (i) focus on securing long-term sector efficiency and proper risk assignments and mitigation, as well as discourage opportunistic bidding and renegotiation; (ii) be embedded in regulations that foster transparency and predictability, support incentives for efficient behavior, impede opportunistic renegotiation and force contract compliance; (iii) address social concerns and focus on poverty; and (iv) promote accountability as the main governance aspect of SOEs. Governments remain at the heart of infrastructure service delivery. SOEs that have a corporate governance structure that reduces political interference, rewards performance, and opens decisions to public scrutiny perform better than those that have a structure that allows politics to influence decision making. Furthermore, even under the presence of PSP, there may be a need for public involvement. Governments need to regulate infrastructure provision as well as contribute a good share of the investment. They must leverage their resources to attract complementary financing. Moreover, they are responsible for setting distributional objectives and ensuring that resources and policies are available to increase access for the poor. To make new reforms sustainable, not only the technical and financial aspects need to be addressed, but also the social aspects most responsible for the backlash. Better communication is critical to create popular support. It is essential to promote the program's infrastructure improvements, advertise the initiative, explain the impact of not improving (but rather maintaining) the status quo, and realistically argue the program's cost-benefit tradeoff. The communication strategy must not only justify the programs, but also periodically inform on the progress of the program, as well as of any changes or problems. The reforms must not only be successful, but that success must be communicated. Communication also serves as a safeguard against corruption at all the levels and as a tool to obtain greater popular support. Greater fairness and support to those adversely affected in the design of the transaction is needed. This can be achieved through the incorporation of social policies, such as social tariffs and financial assistance to those adversely affected by the programs such as those losing their jobs. Programs or policies should be implemented to support users and workers. Sector performance should play a major role in defining the proper sectoral reforms. The newer modalities of PSP--beyond strict privatization-- and proper corporate governance design for SOEs offer significant potential for sector performance improvement. In particular, chances of success will be highly enhanced for programs that comply with the above-listed elements. Improvements in infrastructure for growth and poverty cannot be delayed. There are significant xii threats and opportunities. Most countries, including those in LAC, are at a crossroads on how to improve sector performance. Success may require some form of private sector involvement and financing. If obstacles such as poor perception of PSP are not removed, the significant gains and the very necessary modernization of the sector might fail, and the private financing will prove costly if not difficult. Conversely, opportunity exists to refine the model, attacking the problems and deficiencies of the past, through second-generation reforms that are constructive and broadly participatory. New reform processes that incorporate lessons learned with a clear participation of all the stakeholders and a protagonist role of the public sector are crucial. xiii 1. INTRODUCTION 1. This report conducts a micro-level analysis of various determinants of infrastructure sector performance that have impacts on development at large. Analyzing infrastructure sector performance is about measuring, understanding, and improving conditions at the micro-level, in order to understand how utilities, and regulatory agents contribute to the broader development agenda. Ultimately, sector performance is about the delivery of efficient, affordable, and sustainable infrastructure services. By correlating inputs and outcomes over the last 15 years, this report aims to understand the various determinants tha t have impacted sector performance in infrastructure sectors in Latin America and the Caribbean (LAC). It is about understanding how, and to what extent the effect of several potential elements (including private sector participation, regulation, corporate governance) has resulted in significant changes in the performance of infrastructure services. 2. A large body of empirical literature documents the impacts of infrastructure on poverty reduction and on growth. Evidence shows that infrastructure development promotes economic growth and poverty reduction. Moreover by facilitating access of basic services for the poor, infrastructure fosters development along all levels of the results chain. There are different players involved at each level of sector performance: consumers, communities, service providers, regulators, investors, governments, and nongovernmental organizations. A holistic understanding of infrastructure sector performance creates and strengthens a positive dynamic among key stakeholders. 3. During the 1990's most LAC countries implemented substantial reforms in the infrastructure sector to increase private sector participation, economic regulation, and when possible, promote competition as the main instruments to improve the quality, accessibility, and efficiency of services. While some reforms successfully achieved these objectives, overall the reforms encountered difficulties and currently most of the countries in the region are facing new challenges. By the late 1990´s and early 2000´s, the region faced a series of financial and economic crises, corporate scandals, and some market failures in LAC and around the world. These challenges led to a significant drop of private investment, an increase in political opposition, and some dissatisfaction with privatization and liberalization policies. Ultimately this resulted in difficulties in ensuring access to affordable services for the poor. 4. LAC's infrastructure history leaves no room for complacency, including the sectors in which LAC has performed relatively well: 112 million Latin Americans lack access to household water connections, and 47 million have no access to electricity. While time trends point to improved coverage and performance in LAC, they also shed light on a gap of infrastructure services for many people. According to recent figures, LAC increased coverage of potable water, from 68 percent of the population in 1990 to roughly 80 percent to date. 9 However there are significant disparities between countries, extending from 71 percent water coverage in Haiti to 98 percent in Uruguay (2002). Similarly, while electricity coverage in LAC increased from 82 to 92 percent by 2007, there are many households, most rural, who have been left behind in these steps of progress. About 220 million people live in poverty in Latin America and there is still a long road before reaching the MDGs in a sustainable way. 9 WHO-UNICEF Joint Monitoring Programme (JMP) for water supply and sanitation. 1 5. An integral component of the findings presented in this report is the data collected for each chapter. While the conclusions of the research are entirely and comprehensively derived from this data, the wealth of information produced lends itself to further analysis. The data is easily accessible for repurposing so that the reader is able to conduct ad hoc queries and regression analyses. The benchmarking efforts provide a regional and utility-level frame of reference for good and/or poor sector performance in LAC. The following databases feed into the analysis: a. LAC Electricity Distribution Database: This database contains detailed annual information on 250 public and private utilities of 26 countries that cover 89 percent of the connections in the region. It contains data for more than 20 variables indicating output, input, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990 but the main focus is the period of 1995- 2005. The data is now publicly available. 10 b. LAC Water and Sanitation Database: This database contains detailed annual information on 1700+ public and private utilities of 16 countries that cover 59 percent of the water connections in the region. Similarly to the previous database, it contains data for more than 20 variables indicating output, input, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990 but the main focus is the period of 1995-2006 11. c. ITU World Telecommunication/ICT Indicators Database: This database contains annual time series from 1975-2007 for around 100 sets of telecommunication statistics covering telephone network size and dimension, mobile services, quality of service, traffic, staff, tariffs, revenue, and investment 12. d. The report also draws from a comprehensive database on the impact of private sector participation in LAC. The data covers what happened before, during, and after private sector participation in three sectors--electricity distribution, water and sewerage, and telecommunications--by focusing on a range of performance variables. e. Additional data explores the governance of independent regulatory agencies (IRAs) in the water and electricity distribution sectors of LAC and the link between the governance of IRAs and the performance of both sectors. The analysis of the data first tackles the institutional design of regulatory agencies. The comparison of the different governance modes of IRAs is carried out through different measures of autonomy, transparency, accountability, and tools. The second part of the analysis establishes the methodology and results of the correlation between institutional design and sector performance. f. Data on Corporate Governance of State-Owned Enterprises (SOEs) was collected through surveys sent to different utilities of the region in both the electricity distribution and water sectors. Final respondents were 45 SOEs. The initiative included both public companies with full state ownership and companies where 10 The complete database can be accessed in the external homepage at: http://info.worldbank.org/etools/lacelectricity/home.htm. From there you can compare and download data at utility and country level. 11 This database is not publicly available yet but soon it will be posted on the external site. 12 http://www.itu.int/ITU-D/ict/publications/world/world.html 2 despite there is private investment state ownership is at least 51 percent of total shares (only a few in this category). 6. Understanding the various interventions and conditions that explain LAC sector performance is an indispensible milestone in minimizing the region's infrastructure gap. This report focuses on the distribution segment of basic infrastructure services: It covers electricity distribution, water and sanitation, and fixed telecommunications. The pros of this approach are that some of the features are significantly comparable across the sectors and, hence, we can learn from their comparison. Note however that, although we have advanced in collecting data for the electricity distribution and telecommunications sector, the information available on water needs additional efforts in order to have comparable coverage to the data available for these sectors. 7. In this endeavor, this study aims to answer the following questions: a. What are the main performance trends in the region and how heterogeneous are they? This section will present the results of a benchmarking exercise for each of the utility sectors. The conclusions from this section set out to guide future analytical work in this area of research. b. Difference in performance between state-owned and private enterprises. In this section, we will analyze the determinants of performance (productivity, quality of service, prices, coverage, etc.). Notwithstanding the existing studies, there are still important questions to be addressed in this regard. For instance, what correlations can we make between performance and regulation, and between performance and specific characteristics of market reforms (such as the introduction of wholesale markets, third party access, etc.)? What impact did private sector participation have on the performance of utilities? Does regulatory quality matter? Does competition (when possible) matter? What can be done to increase the efficiency of SOEs? What are the conditions for success? Are firms recovering cost? c. Institutional design of regulatory agencies as a tool for sector performance. This section will focus in more detail on regulatory governance. A number of possible actions and models have been analyzed in relation to their impact on sector performance. However, several important questions remain unanswered: To what extent does regulatory quality matter? Does regulation have any effect on sector performance? Is the independent regulatory agency model still valid for the region? Are there better alternatives? Who are the leaders in the region? How to contrast what is happening in terms of the procedures aimed at improving the governance of regulatory agencies (formal regulation) and the implementation of some of its components (informal regulation)? d. Management mechanisms to create incentives for improved performance. A number of possible actions affecting budget allocations, compensation, and managerial interventions can be provided so as to create incentives for improved performance. Likewise, identified performance indicators can be highly publicized so as to create a context for changes. This section tackles the following questions: What have the boards and the managers of the most competitive and efficient utilities done and are currently doing to improve their governance? What have the governments that own them done and are currently 3 doing? What expectations do they have and what results have they already achieved? What is and has been their rationale and reasons for focusing their efforts in this area? What are and have been the measures they are taking and/or are planning to take? How are and have they been organized to perform this task? What are the main legal difficulties and other obstacles they face in this work? How important is it to enjoy a good reputation and solid social support in carrying over these types of reforms? In which circumstances does social support make reform easier? How does operating in regions characterized by challenging social difficulties affect the chances of introducing reforms? What are the main lessons so far? Is there any difference across sectors? 1.1. A NALYTICAL FRAMEWORK AND SCOPE 8. During the 1990´s most LAC countries implemented substantial reforms in the infrastructure sector to increase private sector participation, economic regulation, and when possible, promote competition as the main instruments to improve the quality, reliability, and efficiency of services, as well as improve the government's fiscal position. 9. This report first delves into the various dimensions of sector performance by describing the main elements that characterize it. The report defines sector performance as the delivery of a reliable affordable service that complies with certain quality standards. Although this definition may be questionable, the set of indicators to be analyzed will provide an overall assessment of the utilities and a different selection of indicators as a whole will not significantly change the key messages of this analysis. In this regard, there are some intermediate outcomes that will also be analyzed. For instance, distributional losses and labor productivity, as a proxy of efficiency of the utilities, may be highly correlated with the quality of the service provided. Figure 1. Analytical Framework 10. Figure 1 depicts the overall framework of this analysis: As we will see in Chapter 2 the region, as a whole, observed significant improvements during the last 15 years; however, at utility level the results are far from homogeneous. A second stage of this work will aim to understand 4 the potential drivers for these differences. Furthermore, the main changes in policies will be analyzed as potential hypotheses for determinants of changes in performance. The report focuses on the relationship between sector performance and the following determinants: private sector participation, regulatory agencies, and corporate governance. It also includes other related aspects such as contract design and market structure, and in particular, the competition in the market, specifically for telecommunications. 11. This report argues that these determinants significantly changed the landscape of the sectors we are studying. However, we acknowledge that there are other elements that may potentially impact sector performance. Figure 2 depicts some of these elements. Even when exploring the specified determinants of sector performance, the report does not assume to describe all the possible links or spheres of influence between each variable and sector performance. More specifically, the following figure represents different levels or environments of impact related to sector performance. When assessing sector performance at the utility-level, the utility faces several constraints in influencing or impacting the condition or actors of its surrounding environments. For example, while a change in utility performance may impact aspects of its sectoral environment, such as regulation and market structure, it will never be ab le to influence its external environment such as the macro economy and geography. Therefore, the actions of the utility will at most influence its own sector performance and its sector environment. However, when assessing the impact that the external environment may have at the sector and utility level, there are no such constraints. The factors in the external environment can certainly impact the elements in the sectoral and utility environments. For example, the state of the macro - economy will surely impact the conditions at both the sectoral and utility level. Figure 2. Scope of the Report 12. This report proposes to firstly explain the dynamics and interactions between the utility and the respective sector. It tackles the issues of sector environment as it may determine the incentives for the utility to better perform. While the report may refer to the components and impact (on sector performance) of the external environment, it does not assume to explain these elements as they relate to sector performance. The main objective of this report is to provide a factual description of the changes and policies that can be empirically tested and analyzed. Since many of the processes and policies in the external environment cannot be measur ed or standardized, it would only marginally (if at all) benefit our understanding of utility performance. For example, while the report is aware that macroeconomic circumstances (external environment) 5 may trickle down and affect the overall sector and utility, it does not promise to expound on this relationship. From the utility and the sector perspective, the external environment is a given and we consider it very unlikely that the sector or the utility may influence it, at least in the short run. Therefore, we restricted the scope of some of the potential policies that could be developed within the sectors. While there are many other policies that could be significantly relevant in a specific context, to draw systematic conclusions about them raises tremendous challenges in terms of data availability and methodology. For this reason, this report focuses only on those policies that can be systematically and empirically analyzed. 13. More specifically, the previous Figure clarifies what this report intends and does not intend to do. The report does: a. Depict sector performance with a broad set of indicators that describes the current situation as well its evolution during the last 15 years; b. Propose analytical frameworks for themes less developed in the literature such as regulatory governance and corporate governance for state-owned enterprises; c. Benchmark the institutional designs of the regulatory agencies in the region for the water and electricity distribution sectors; and d. Analyze the relationship between sector performance and regulation, private sector participation, and corporate governance. .... However, the report does not: a. Presume to describe all the possible factors and resulting conditions that may impact sector performance; b. Focus on the external environment (Figure 2), which cannot be changed/impacted by the change in sector performance; and c. Analyze other factors that may belong to the sector or utility environments but, for different reasons may not be systematized. 14. The report is organized into the six following chapters that serve to document the changes that have occurred in the last 15 years and their respective impacts. Each core chapter is dedicated to understanding the particular characteristics of each of the determinants within a specific time and political context. 15. The Chapter 2 of this report outlines the changes that have occurred in the Electricity distribution, Water and Sanitation, and Fixed Telecommunications sectors in Latin America during the last 15 years. These changes are captured through benchmarking assessments based on the results of performance indicators such as output, coverage, labor productivity, inputs, operating performance, service quality and prices. This chapter tells multiple stories of the substantial improvement in these sectors and fills in the knowledge gaps that exist regarding the status of the sectors by benchmarking utility performance at the regional, country, and utility level. 16. The Chapter 3 synthesizes the impact that Private Sector Participation has had on various infrastructure sectors, including electricity distribution, water and sewerage, and fixed -line 6 telecommunications. In an attempt to understand the true impacts and determinants of private sector participation in LAC, this chapter directs our attention to what happened before, during, and after private sector participation in these three sectors by focusing on a range of performance variables. Looking at changes before, during, and after private sector participation, with and without private sector participation scenarios are then examined. This chapter will also identify whether privatization characteristics like the sale method (for example, auction), investor nationality, and award criterion affect the performance variables discussed in previous sections. 17. The Chapter 4 explores the institutional design of the regulatory agencies in the region and the link between regulatory governance and sector performance. The first part of the chapter is dedicated to the evaluation and benchmarking of the governance of regulatory agencies in the electricity sector in the region. This chapter draws heavily on previous work in which an index of regulatory governance was developed in order to rank all the agencies in the LAC countries. The index is an aggregate number of the evaluation of four key governance characteristics: autonomy, transparency, accountability, and regulatory tools, including not only formal aspects of regulation but also indicators related to actual implementation. The second part builds upon the benchmarking analysis and questions whether there is a correlation between regulatory governance and sector performance. The results suggest that the mere existence of a regulatory agency, regardless of the utilities' ownership, has a significant impact on performance. 18. The Chapter 5 of this report is an assessment of the Governance of SOEs in Infrastructure. The results of this chapter are based on the cases of 45 state owned companies in the water and electricity distribution sector of LAC. It proposes an analytical framework for analyzing corporate governance of these utilities and benchmarks their institutional internal design. Finally, the chapter will evaluate the contribution that these dimensions have on sector performance. 19. Chapters 6 and 7 will be a summary of other potential determinants for sector performance and a compilation of the main results, and will serve to channel the region's attention to an array of possibilities for moving forward. 7 2. BENCHMARKING LAC'S UTILITY PERFORMANCE 20. This chapter outlines the changes that have occurred in the Electricity distribution, Water and Sanitation, and Fixed Telecommunications sectors in LAC. The first part analyzes the changes that have shaped the performance of these sectors. This analysis is derived from previous benchmarking initiatives for the electricity distribution sector (Andres et al., 2008b) as well as a database for water sector and a well know database on the telecommunication sector (ITU, 2009). The chapter documents these changes and accounts for the current performance (according to our data) of the respective sectors at the regional and utility level. 21. The findings reported in this chapter were captured through benchmarking assessments based on the following performance indicators: output, coverage, labor productivity, inputs, operating performance, service quality, and prices. Considering the changes that have shaped the electricity distribution, the water and sanitation, and the fixed telecommunication sectors during the last decade, such benchmarking efforts provide a regional and utility-level frame of reference for good and/or poor sector performance in LAC. In sum, this chapter documents the substantial improvement in the electricity distribution and water and sanitation. 22. There is a sharp divide between rural and urban coverage within countries. For water, electricity, roads and telecommunications, coverage rates in rural areas tend to be much lower. While more than 90 percent of the urban population of most countries in the region have access to safe water, rural access in Brazil (58 percent) and Chile (59 percent) is worse than in several much poorer African nations such as Burundi (78 percent) and Zimbabwe (74 percent) (Fay and Morrison, 2006). Given that poverty rates are usually much higher in the countryside, lower rural access rates explain much, though by no means all of the great disparity in coverage between the rich and poor in Latin America. While this Chapter does not make a distinction between urban and rural electricity and water and sanitation, it nonetheless acknowledges this discrepancy and intends to provoke further work in order to bridge this gap. W HY B ENCHMARK THE I NFRASTRUCTURE S ECTORS? 23. Benchmarking is a means of providing countries and utilities with a point of reference regarding their performance. Electricity lights homes and powers industries, but in many LAC countries, service quality remains unreliable -- even for those who can afford to pay high prices. Further service expansion to the people in the region who live without basic infrastructure services, and improving the quality and reliability of service delivery are urgent socio-economic priorities. More so, the lack of good infrastructure services cost Latin American businesses dearly. Against this backdrop, the benchmarking initiatives outlined in this chapter, serve to provide regional and utility level direction and a framework of comparison for identifying where LAC utilities stands in relation to the others, detecting their strengths and weaknesses, and setting goals for improvement. 8 24. The purpose of benchmarking infrastructure utilities is to provide a detailed description of the sectors in LAC and to identify and rank the best performers in the region. For instance, a number of empirical studies have used benchmarking methods within the electricity supply industry. These studies have traditionally focused on generation or on vertically integrated utilities; however, perhaps due to regulators' demand, the interest in benchmarking the natural monopoly segments (i.e., transmission and distribution) has recently increased. Surveys of the benchmarking literature (Jamasb and Pollitt, 2001; Mota, 2004) have concluded that, due to issues of data standardization and currency conversion, international benchmarking has not been widely used. When international efficiency comparisons have been used, they have traditionally focused on developed countries. 25. An analytical framework was designed to produce a comprehensive description of the sectors as well as a mechanism for ranking countries and utilities for best performance. By serving as a mirror of good performance, this chapter allows for a comparative analysis and the ranking of utilities and countries according to the following indicators used to measure performance. The data collected tells various stories about the distribution sector based on accomplishments in output, coverage, inputs, labor productivity, operating performance, the quality of service and prices. The following sections are dedicated to analyzing these results. This chapter is designed to be solely factual, aimed at describing sectoral performance at the regional and utility levels and does not assume, at this stage, an analytical or explanatory role. Additionally, this benchmarking exercise will contribute towards a more consistent benchmarking analysis in the distribution segments and serves as a path-breaker for other regional benchmarking initiatives. 26. This benchmarking exercise, covers the following databases (see Annex 2 for details): a. LAC Electricity Distribution Database: This database contains detailed annual information on 250 public and private utilities of 26 countries that cover 89 percent of the connections in the region. It contains data for more than 20 variables indicating output, input, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990 but the main focus is the period of 1995- 2005. The data is now publicly available. 13 b. LAC Water and Sanitation Database: This database contains detailed annual information on 1700+ public and private utilities of 16 countries that cover 59 percent of the water connections in the region. Similarly to the previous database, it contains data for more than 20 variables indicating output, input, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990 but the main focus is the period of 1995-2006 14. c. ITU World Telecommunication/ICT Indicators Database. This database contains annual time series from 1975-2007 for around 100 sets of telecommunication statistics covering telephone network size and dimension, mobile services, quality of service, traffic, staff, tariffs, revenue, and investment 15. 27. Table 2.1 presents the definitions of the variables used in the present analysis. 13 The complete database can be accessed in the external homepage at: http://info.worldbank.org/etools/lacelectricity/home.htm. From there you can compare and download data at utility and country level. 14 This database is not publicly available yet but soon it will be posted on the external site. 15 http://www.itu.int/ITU-D/ict/publications/world/world.html 9 Table 2.1. Variable Definitions Electricity Distribution Fixed Telecommunications Water Distribution Output Total number of subscribers and residential Total number of active connections. December of Total number of water subscribers and residential subscribers, December of each year each year water subscribers Total energy sold per year (in MWh) Total number of local minutes per year Total number of residential sewerage subscribers Energy sold per connection Total minutes per active connection and residential sewerage subscribers Total water production per year Total water sold per year Labor Number of employees Number of employees Number of employees Labor Number of subscribers per employee Number of active connections per employee Number of water connections per employee Productivity Total energy sold each year per employee Local minutes per employee Water sold per employee Efficiency Energy lost in the distribution (due to technical Percentage of incomplete calls Percentage of total water produced that was not losses and illegal connections) charged to the consumers Quality Average duration of interruptions per consumer Percentage of incomplete calls [faults] Average number of hours per day with water (hours/year) Percentage of digital connections in the network service Average frequency of interruptions per consumer Percentage of the samples that passed a potability (number/year) test Coverage Number residential subscribers per 100 Number of active connections per 100 Number of residential water subscribers per 100 households inhabitants households Number of residential sewerage subscribers per 100 households Prices Average tariff for 1 MWh for a residential Average cost for a three-minute, nonpeak local Average price per cubic meter of supplied water (in service in dollars (it includes fixed and variable call (dollars) dollars) costs), December of each year Average monthly cost for residential service Average price per cubic meter of collected waste Average tariff for 1 MWh for a industrial service (dollars) (in dollars) in dollars (it includes fixed and variable costs), December of each year Expenses Operation expenses per year per connection Operation expenses per year per water connection (dollars) (dollars) Operation expenses per year per MWh sold Operation expenses per year per cubic meter sold (dollars) (dollars) Source: Authors' elaboration. 10 28. The following sections describe the benchmarking analyses for the three sectors evaluated. For simplicity, we present in this Chapter the summary of the results. For detailed information, the results for each of the indicators can be found in Annex 3. 2.1. ELECTRICITY DISTRIBUTION 29. Since the late 1980s, a wave of reform has transformed the institutional framework, organization, and operational environment of the infrastructure industries, particularly those in the electricity sectors in most developed and developing countries. Although the structure of the power sectors and the approaches to reform vary across countries, their main objectives are to improve the efficiency of the sector as well as to increase the coverage and quality of service. Separation of roles, unbundling, competition and private participation were used as key instruments to increase efficiency, improve the government's fiscal position and increase access to electricity service for the poor. In many countries in the region the combination of private participation, competition and better regulation was effective in improving productive efficiency and quality of service. 30. The last decade has witnessed significant progress in the power sector of LAC. While there are differences between countries, overall supply has increased substantially and with it access to electricity. The best electricity distribution performer is Uruguay with 97 percent coverage followed by Costa Rica, Brazil, Argentina, Chile, and Mexico with more than 95 percent coverage. However, equally important is to consider the overall improvement in coverage as reflected in the growth rate of countries such as Peru, Paraguay, Honduras, and El Salvador achieving an average growth of 20 percentage points in the last 10 years. 31. Electricity distribution is at the forefront of infrastructure improvement in LAC with 95 percent coverage16 and a 10 percentage point increase by 2005. Since 1995, most countries in the region have made considerable progress in expanding access to electricity and improving the quality of their service. In the period covered in this report, private sector participation increased from 11 percent to 60 percent of electricity connections and labor productivity doubled since 1995. In addition, the results of this chapter exhibit improvements in the frequency and duration of interruptions per connection showing a 42 and 40 percent reduction accordingly. While there are no clear trends in operational expenditures, these values have grown between 41 and 44 percent in the last decade. Furthermore, there are no considerable changes in distributional losses and tariffs have grown steadily with a cumulative increase of 70 and 91 percent for residential and industrial users, respectively. 32. Despite the fact that electricity coverage in LAC increased from 85 to 95 percent in 2005, evidence suggests that the poor and rural areas were not the main beneficiaries of the improvements in productive efficiency and coverage. According to the LCR Energy Strategy (World Bank, 2007a), in many countries, industrial consumers and high income residential consumers were the main beneficiaries of competition and rebalancing of tariffs, which reduced substantial cross-subsidies of the pre-reform period. However, it is also true that privatization and cost-covering tariffs ensured the financial feasibility of efficient electricity providers, which were able to expand access and improve the quality of service to a large number of consumers in urban and peri-urban areas, including poor people. 16 These coverage figures correspond to the weighted average of the 250 utilities in the LAC Benchmarking database. The total regional electricity coverage is estimated in 92 percent by 2007. 11 2.1.1. E LECTRICITY DISTRIBUTION B ENCHMARKING R ESULTS R EGIONAL B ENCHMARKING A SSESSMENT 33. During the last 10 years, electrification in the LAC increased significantly at an annual growth rate of 1.1 percent, from 84.7 percent coverage in 1995 to 94.6 percent (within the sample covered) in 2005. The growing trend in electricity coverage reflects a high demand for access to the electricity network by a growing number of residential, non-residential, and rural users. 34. As demand for electricity increased, so has private participation in electricity distribution throughout the LAC region. Private participation has grown substantially since 1990, and especially between 1995-1998. While in 1990 there was little significant participation of the private sector in electricity distribution, by 1995, 11.1 percent of electricity connections in the region were served by the private sector. By the end of our period of analysis, 60 percent of electrical connections were supplied by private utilities. Based on the data from the Private Participation in Infrastructure (PPI) Project Database, during the last 15 years, US$ 102.6 billion was invested in 384 private electricity projects in LAC. Most LAC countries have introduced private participation in electricity distribution as part of broader reforms attempting to establish a more competitive market structure. However, in the last four years private participation has remained mostly stagnant with low levels of investments. It is worth considering this phenomenon when analyzing the regional performance of the electricity distribution in the following sections. 35. Despite the fact that electricity coverage in LAC increased from 84.7 to 94.6 percent 17 in 2005 there are still many people, almost all poor and in rural areas, without electricity. There is still a strong need to expand electrification in rural areas in LAC countries since these areas lag behind. For example, large increases in electricity coverage in Argentina are related to the normalization of illegal connections in urban slums rather than the expansion of electricity service to rural areas. Private investors were effective in connecting consumers in urban and rural areas near the power grid but are reluctant to extend access to rural areas where electricity service is not financially viable. In Bolivia and Nicaragua, countries that privatized distribution, only 30 percent of rural population has access to electricity. Further increases of coverage in rural areas usually require substantial investment subsidies and strong government support. The government of Chile, a leader in reform and privatization, provided investment subsidies of about US$1,500/household to increase electricity coverage in rural areas from 62 to 92 percent in 1995- 2005. 36. The energy sold per connection per year, as an output measure, exhibits an increasing trend until 2000 with a total increase of 0.29 MWh sold per connection, after which ther e is a sudden drop in sales that continues to decrease until the end of 2005, with a total reduction in MWh sold per connection of 2.9 percent. During the last 10 years, the average energy sold per connection is 5.5 MWh. Albeit a 45 percent increase in the number of connections from 1995- 2005, the total amount of energy sold per connection has declined. When considering the 17 These regional estimates correspond to the weighted average across the 250 utilities in the sample that represents 90 percent of the total number of electricity connections. 12 evolution of energy sold, the fluctuating values of the energy sold per connection may be attributed to the increase in residential and industrial tariffs and thus a decrease in demand. 18 37. When assessing regional distribution losses, there is no apparent trend but rather sporadic increases and decreases throughout the 10 year period. The lowest distributional loss was observed in 2001, with a 0.9 percentage point decrease over a 14.5 percent distributional loss in 1995. Since 2001, the region has experienced a one percentage point increase, resulting in a 14.7 percent distributional loss in 2005. 38. A look at the quality of electricity distribution in LAC allows one to qualify the region as improving in the delivery of its services. In the last ten years, the frequency of interruptions in the region has decreased by almost half, with a 42.4 percent drop in the frequency of the interruptions and 40.2 percent decrease in the duration of the interruptions per connection per year. When measuring the quality of service, there has been a steady decline in the number of interruptions per connection. While the average number of interruptions per connection was 20.5 times in 1995, this dropped to 11.8 times in 2005, a reduction of 5.4 percent per year, totaling a 42.4 percent reduction in ten years. A second indicator used to measure quality of the service is the average number of hours the customer did not have service. The last decade presents a generally downward trend with a 40.2 percent decrease in the duration time per connection. The indicator presents a remarkable increase in 2002 in the duration of interruptions. As explained in the next chapter, Brazil and Paraguay are the main contributors for the 1996 increase while the peak in 2002 is explained by the hurricanes that affected the quality of service in Mexico. These two indicators successfully encapsulate two root causes of interruptions: the reduction in the number of outages per connection shows managerial improvement, while the duration of the interruption serves as a proxy for natural events or disasters that affect service. 39. The decrease in the number of employees through the past ten years is inversely related to the rise in private participation. The 23.2 percent reduction in employees is visible in the trend between 1995 and 2000, when PSP reached its peak. For the last five years the database suggests that no significant changes in the regional level of the labor force have occurred, consistent with decreased private participation levels. 40. Amongst the measures used for estimating labor productivity is the number of residential connections per employee. During the 1995-2005 timeframe, this value has doubled from 384 residential connections in 1995 to 701 in 2005. The natural growth trend in population (approximately 1.1 percent per year) accounts for the suggestive "natural" growth in the number of connections contributing at most for one fifth of the improvement in labor productivity. A second contributing factor is the substantial improvement in electricity coverage. The final concurrent factor that drives this change is the reduction of the labor force in the sector. Based on this analysis, there was a 23.2 cumulative percent reduction in the number of employees for the period analyzed. 41. In the same vein, an analysis of the regional labor productivity tells a story of results that doubled throughout the decade. Labor productivity, measured as the energy sold per employee increased gradually from 2.2 TWh sold per employee in 1995 with a peak of 4.1 TWh in 2005, totaling a 85.1 percent growth for the last decade. 18 This reduction in energy sold per connection could also be related to the increase in residential access to poor families which brings down the average intensity of electricity usage. 13 42. Average end-user tariffs for electricity (dollars/MWh) supplied to residential connections show an overall increase with the exception of 1999, with a 12 percent decrease, mainly caused by the crisis in Brazil. By the end of 2005, the average residential tariff was $104 per MWh, a 70.3 percent accumulative increase over 1995's $61.33 average residential tariff. Following the same pattern, the average industrial tariff increased by 90.8 percent since 1995. While the weighted average in 1995 was $44.28, in 2005 the weighed industrial tariff reached $84.48. The figure shows a steady increasing trend with the exception of the period between 1997 and 1999 where there was a slight decrease in prices. 19 43. With respect to input indicators, the region has witnessed fluctuating values of operational expenditures (OPEX) with more prominent changes towards the end of the decade. Operation expenditures per connection have increased 40.8 percent throughout the decade. Despite the irregular activity between 1995-2005 with unexpected changes in expenditures between 2000 and 2003, the regional average for OPEX was $128 with an average 3.5 percent increase per year. The results for total expenditures (TOTEX) per connection express the overall direction of operational and capital expenditures for LAC in the last decade. Defined as the total operation and capital expenditures, TOTEX exhibits a steady increase with the exception of a drop between 1998-1999 and 2001-2003. By the end of 2005 total expenditures reached $173.7 per connection, from $ 99, a two-fold increase since 1995. The results for OPEX per MWh energy sold show a similar tendency to that of OPEX per connection. OPEX per energy sold demonstrates a 44 percent increase throughout the last 10 years with an annual growth rate of 3.7 percent. With respect to the regional average of $26.6 per connection, OPEX reached $33.28 per connection by 2005. U TILITY -L EVEL B ENCHMARKING A SSESSMENT - E LECTRICITY D ISTRIBUTION 44. Three main messages characterize LAC's electricity distribution performance: First, there are significant discrepancies amongst utility performance. Second, there has been an overall improvement of the underperforming utilities during the last ten years. Third, there are cases with significant deterioration in distribution performance as reflected by indicators such as the average tariffs and distributional losses. In order to assess the performance of LAC's electricity distributors at the utility level, the 250 utilities studied were ranked according to the top ten percent, middle 80 percent, or bottom ten percent of distribution performance. The best performing utilities are listed in the top or bottom ten percent depending on the indicator being measured. Amongst the characteristics of the top performing utilities are utilities with 100 percent electrification, an average of 897.1 residential connections or 6,402 MWh of energy sold per employee, 6.5 percent distributional losses, and residential prices in the range of $591 per MWh consumed. 45. In summary, for the time period of 1995-2005, the lower performing utilities have doubled their electricity coverage and labor productivity, curtailed the frequency of interruptions per connection by 73 percent and the duration of interruptions by 55.9 percent, and decreased their total expenditures per connection by 26 percent. As attested by the aforementioned results, 19 Tariffs in absolute terms are not, per se an efficiency measure of utilities since retail tariffs are related to generation costs. Ideally, it would be necessary to measure the tariff gap or the Value Added of Distribution (VAD) to isolate the cost for the distribution segment from the rest of the value chain. Cost estimates are extremely complicate to collect. We attempted to collect this data with our OPEX indicators. However, the coverage of these series are no so high. 14 significant progress has been made by the majority of the utilities in all categories throughout the last decade. 46. While assessing the performance of the distribution utilities presented in this study, we have encountered significant discrepancies amongst utilities. For instance, in 2005, utilities in the top ten percentile were ten times more productive and sold six times the amount of energy (per connection) of utilities in the bottom ten percent. The best performing utilities had less than four interruptions per year and less than 4 hours of duration of the utilities represented in the bottom ten percent. In the same vein, the bottom decile utilities had one fifth of the distributional losses that characterized the underperforming utilities. 47. A second but equally important message is the overall improvement of the underperforming utilities during the last ten years. As attested by the time trends, the utilities in the bottom ten percent improved significantly in coverage from an initial 40 percent electrification in 1995 to 61 percent electrification in 2005. Similar improvements were observed in the frequency and duration of interruptions especially by the underperforming utilities. The evolution of the level of labor productivity illustrates the progress of the poorer performers with a three-fold improvement in the last ten years. 48. Third, there were cases with significant deterioration in performance reflected in indicators such as distributional losses. While the residential tariff increased from $44.4 in 1995 to $114.4 in 2005 for the middle 80 percent, the top ten percent increased their residential ta riffs by 36.7 percent compared to the initial $127 per MWh sold in 1995. With respect to distributional losses, while the middle 80 percent did not exhibit a significant change during the decade, the underperforming ten percent showed a 27 percent increase in distribution losses. 49. Chronicling the story of the best performing distribution utilities during the last decade is a story of universal electrification and significant improvements of the utilities represented in the middle 80 percent. Electrification increased by almost 15 percentage points for the middle 80 percent and 20 percentage points for the bottom ten percent attaining 88 percent and 61 percent coverage in 2005 respectively. 50. Finally, the chapter identifies the utilities that set the standard of good performance for each indicator. Although there are some variations within and between countries, in general, several companies in Brazil lead with best performance in terms of labor productivity, distributional losses, OPEX, and coverage. In addition, Costa Rica benchmarks good performance in coverage, OPEX, and tariffs. Finally, several utilities in Chile produced leaders for indicators measuring labor productivity and technical efficiency. 2.2. W ATER AND SANITATION SECTOR 51. Approximately 220 million people live in poverty in the LAC region; out of these, 112 million people in urban and rural areas lack access to a water connection. These figures attest to the region's challenge in meeting the MDGs in a sustainable way and the need for timely and efficient interventions in the sector. This benchmarking initiative proposes to fill in knowledge gaps for the identification of the best performers of the region and thus improve water and sanitation interventions and sector performance. 15 52. In order to conduct a benchmarking analysis, the study collected detailed information of 16 countries and 1,700 water and sanitation utilities in the region (see Annex 2 for details). An analytical framework was designed to produce a comprehensive description of the sector as well as a mechanism for ranking countries and utilities for best performance. The data collected for this benchmarking project is representative of 59 percent of the water and sanitation connections in the region from 1995 to 2006. By serving as a mirror of good performance, the report allows for a comparative analysis and the ranking of utilities and countries according to the indicators used to measure performance. Through in-house and field data collection, consultants compiled data to tell various stories about the distribution sector based on accomplishments in output, coverage, input, labor productivity, operating performance, the quality of service and prices. 2.2.1. W ATER AND SANITATION B ENCHMARKING R ESULTS R EGIONAL-L EVEL B ENCHMARKING A SSESSMENT 53. The main finding of this chapter is one of overall improvement across the region during the studied time period with significant changes in the following areas: a 4 percent increase in water coverage reaching 97 percent in 2006 within the coverage area of the utilities in the database20; labor productivity that almost doubled from 252 connections per employee to 425 in 2006; slight tariff increases for both water and sanitation--the water tariff increased by 27 percent and the sewerage average tariff increased by 35 percent; and a significant 8 percent increase in the continuity of service. While these figures are solely a glimpse of what characterizes the water and sanitation sector in LAC, it provides important insights regarding the sector's strengths and weaknesses. 54. It is important to note that the following data and results are only representative of our data sample. Therefore the results must be interpreted within the specified data sample: 1,700 utilities serving 65 million households--or 59 percent of the households with a water connection in the region. More so, this chapter is designed to be solely factual, aimed at describing water and sanitation performance at the regional and utility level and does not assume, at this stage, to take on an analytical role. 55. Water and sanitation coverage for the utilities benchmarked in LAC increased from 93 percent to 97 percent between 1999 and 2006. The 98 percent coverage represents approximately half of the households in the utility's respective area of service. This 4 percentage point improvement is consistent with the overall improvement in coverage for all the water and sanitation operators in the region which resulted in 81 percent in 2006. Sewerage coverage results in a significant 12 percentage point increase from 72 percent in 1999 to 84 percent in 2006. When measuring the percentage point increase in coverage, it is important to consider that such changes may depend on other factors such as demographics. While this chapter considers possible 20 Note that these figures correspond to the coverage area for the concessionaires. As we mentioned earlier, there are 146 million inhabitants in LAC without adequate access to water supply that would be equivalent of access to potable water in 2004 (source: WHO-UNICEF, 2008). The main difference between these estimates is that the lack of service estimate is calculated using census and household surveys; hence, these sources include rural population and population in areas not covered by the concessionaires in our sample. A second consideration is that the evolution in coverage presented here corresponds to the results based on our database that includes 50 percent of the water connections in LAC. Extrapolating these figures to the reminder connections not covered by the sample could be biased since the other utilities that we were not able to include in our sample may present lower level of water coverage. 16 determinants for such observed changes, explaining the possible link between the results and determinants is beyond the scope of this chapter. U TILITY -L EVEL B ENCHMARKING A SSESSMENT - W ATER AND S ANITATION 56. As we presented for electricity utilities, we rank water and sanitation utilities according to the top ten percent, bottom ten percent, and the simple average of the rest of the 80 percent performers. The utilities are evaluated based on their achievements in coverage, labor productivity, output, input, operating performance, service quality, and prices. Best performing utilities are listed in the top or bottom ten percent depending on the variable at hand. While the best performer for water coverage is listed in the top ten percent, the best performer for non- revenue water would form part of the bottom ten percent. In addition, it is worth noting that for certain indicators such as operation and capital expenditures, ranking in the top or bottom ten percent is not necessarily a benchmark of good performance. 57. Benchmarking the water and sanitation utilities in our sample, allows us to relate some key messages in the region. The best performing utilities are listed in the top or bottom ten percent depending on the variable being measured. The best performers in our database are characterized by 100 percent water and sanitation coverage, an average of 581 m 3 water sold per connection a year, an average of 541 residential connections per employee, 15 percent losses in non-revenue water, water residential prices in the range of $0.11 per m3 water a year, and sewerage residential prices averaging $0.07/m 3 water a year. It is important to note that these numbers, while serving as benchmarks, also highlight the heterogeneity of water and sanitation utility performance not only between countries but also within countries. 58. When considering the results for coverage and output for our sample, we find substantive discrepancies between the top ten and bottom ten performers. On average, the top ten percent performers have 100 percent water and sanitation coverage. In contrast, the bottom ten percent water utilities average 66 percent coverage, while sewerage utilities average a low 15 percent coverage. Between 2000 and 2006, the top ten performers maintained an average of 100 percent coverage in water and sanitation and there has been a small improvement for the middle 80 percent performers. It is worth nothing however that both the bottom ten percent water and sanitation utilities exhibit a 23 percent increase in coverage between 2000 and 2006. 59. When assessing the efficiency of the water and sanitation sector in LAC, this study focuses on the results of the following indicators: labor productivity, non-revenue water, collection ratio, and water connections that are micro metered. Within this respective sample set, the best performers have: 541 connections per employee, 15 percent non-revenue water, a collection ratio and micro-metered water of 100 percent. When compared to the poorest performers in our sample, the top decile performers are 5 time more productive, incur ¼ of the non-revenue water losses of the bottom ten performers, collects 50 percent more revenues per total water billed, and has 5 times more micro-metered households that the bottom ten percent. 60. Continuity of service and potability are two main indicators that determine the quality of service for the water and sanitation sector. The best utility performers provide 24 hours of water service a day--approximately 1.5 hours more than the middle 80 percent by the end of 2006. However, there is a substantial gap between the 100 percent average of the top ten percent performers and the 8 hour per day average of the bottom decile during the 1997-2006 time period. 17 61. In an attempt to benchmark the water and sanitation performance of the utilities presented in this study, we have encountered significant heterogeneity amongst utilities. For example, when measuring continuity of service, there is a substantial gap between the 100 percent average of the top ten percent performers and the 8 hour per day average of the bottom decile during the 1997- 2006 time period. Similarly, when measuring efficiency, the top decile performers are 5 times more productive that the bottom ten percent and incur ¼ of the non-revenue water losses of the bottom ten performers. 62. A second concluding remark is that our results illustrate an overall improvement of the underperforming water and sanitation utilities between 1995 and 2006. For example, there is a 23 percent increase in water and sewerage coverage by the bottom decile between 2000 and 2006. Similar progress can be seen in the overall quality of service, especially for the middle 80 percent. 63. The third important message drawn from this benchmarking analysis is that there is still much room for improvement. On average, the 1,700 water and sanitation utilities in our sample continue to face challenges such as high non-revenue water levels, low collection rations-- averaging 50 percent amongst the region's poorest performers, insufficient tariffs, amongst other factors. 2.3. F IXED T ELECOMMUNICATION S S ECTOR 64. During the 1980s and the 1990s, the state owned the fixed telecommunications company, which operated in a monopolistic market. After Chile's experience in the 1980s, most of the countries privatized their telecom companies. The new owners generally had to comply with requirements such as network expansion and quality standards. In exchange, they were granted a monopoly period, after which new firms could enter the market. 65. In most countries, liberalization of the long-distance market took place within a few years after privatization (Andres et al., 2008c). Hence, there is a possibility that the impacts of privatization perceived were actually instead caused by liberalization. Even though the indicators used above refer to local telephone service, liberalization of the long-distance market could be an indicator that liberalization of the local market was to come. 66. Nowadays, close to 85 percent of the fixed lines are operated by private companies. Just countries like Colombia, Uruguay, and Paraguay still possess their main telecommunication operator. 2.3.1. F IXED T ELECOMMUNICATION B ENCHMARKING RESULTS 67. During the 1980s and the 1990s, the government owned the fixed teleco mmunication companies, which operated in a monopolistic market. After Chile's experience in the 1980s, most of the countries privatized their telecom companies. The new owners generally had to comply with requirements such as network expansion and quality standards. In exchange, they were granted a monopoly period, after which new firms could enter the market. The following section describes the main trends in the LAC telecommunications sector between 1995 and 2007. These results represent all the countries in the region and allow for a regional benchmark of good 18 telecommunication performance according to the following indicators: output, coverage, labor productivity, investments, and the quality of service and prices. 68. By 2007 the LAC region invested an average of $12 billion in telecommunication services--a 4 percentage point increase since the $8 billion invested in 1995. Out of the total $12 billion invested in 2007, $2.3 billion was allocated for fixed telephone services. By the end of our analysis in 2007, the region's coverage for households with a fixed telephone line reached 62 percent--a 100 percent increase between 1995 and 2007. Similarly, the main (fixed) telephone lines per 100 habitants increased from 9 percent to 20 percent, at an annual percent change of 102 during our studied time period. When considering the change in subscribers per 100 inhabitants, our analysis takes into consideration both fixed and mobile lines. While the total change in subscribers per 100 inhabitants for both fixed and mobile increased from 10 subscribers in 1995 to 83 in 2007, there is a significant divergence when considering the change in subscribers for fixed and mobile lines separately. The subscribers per 100 inhabitants for fixed lines increased one fold from 10 to 20 subscribers. However, mobile lines surged from 0.7 subscribers in 1995 to 64 subscribers per 100 inhabitants in 2007--an impressive 8253 percent increase over 12 years. 69. By 2007, the region serves a total of 464 million (fixed and mobile line) telepho ne subscribers, representing a significant 921 percent increase between 1995 and 2007. The main fixed lines in operation at the regional level increased by 150 percent between 1995 and 2007, out of which 73 percent were residential main lines. While the trend for residential main lines has increased steadily between 1995 and 2006, it experienced a sharp 2 percentage point decrease between 2006 and 2007. 70. Labor productivity, measured as the number of fixed and mobile connections per employee, has increased by 1,072 percentage points at annual change of 611.1 percent, with the exception of an abrupt decrease in the last two years. 71. When measuring quality, the telecommunication sector in LAC has taken an important leap forward with significant gradual increments in the percent of digital main lines, and telephone faults cleared by the next working day. For example, digital main lines have increased from 63 percent in 1995 to 100 percent in 2007. Comparably, the number of telephone faults per 100 main fixed lines per year dropped from 23 to 8 by the end of 2007. Such improvements in quality have also been accompanied by a reduction in the waiting list for main fixed lines, which by 2007 averaged to zero. 2.4. P UBLIC VS PRIVATE B ENCHMARKING - E LECTRICITY D ISTRIBUTION 72. The data allows for different desegregations such as by country, size, ownership, and structure, among other characteristics. Although all these exercises permit to understand different questions, there exists a wide range of possible scenarios. The data is publicly available 21 and allows the users to identify and produce their own benchmarking exercises. All the Figures for this exercise are available in Annex 3.4. 21 For instance, the electricity data can be accessed at http://info.worldbank.org/etools/lacelectricity/home.htm. 19 73. Among these scenarios we selected, as an example, the comparison between public utilities and those with private sector participation. The goal is to contrast the differences between these two groups of utilities. While the previous sections indicate major improvements in coverage and quality at the regional and utility-levels, this section provides insight on utility performance based on the means of ownership. The following results are based on the simple average across the utilities in the Electricity distribution Benchmarking Database. We will present the information in two ways: i) the utilities will firstly be presented in the following three categories: public utilities throughout the period of 1995-2005, utilities that privatized before 1995 and remained private throughout 2005, and utilities that privatized after 1995 and remained private throughout 2005. In order to most accurately assess and compare the performance of public and private distribution utilities, we considered the initial conditions in 1995 as well as the overall trend of the last ten years; and ii) we present the variance of change and improvement of the studied indicators. As we did in the previous sections, we report the average top ten, bottom ten, and middle eighty percent public and private utilities. 74. The main findings of this section attest to the considerable improvement in the performance of the electricity distribution sector. The following results show the public and/or private utilities that benchmark good performance for each respective indicator. When comparing the performance of private and public utilities, the main differences in performance are marked by: labor productivity, distribution losses, quality of service, and tariffs. In contrast, other indicators such as coverage and operation expenditures exhibit similar trends and/or do not present significant changes between the groups. On average, private utilities performed better than public utilities with clear differences after the change in ownership. Significant improvements in labor productivity are a distinguishing factor when assessing the performance of the sector. When measuring the number of connections per employee in 1995, the labor productivity of post-1995 privatized utilities was only 10.7 percent greater that that of public utilities. Yet by the end of the decade, the labor productivity of post 1995 privatizations increased three-fold and doubled the amount of public utilities. Another indicator exhibiting significant improvement after the change in ownership is that of distribution losses. In 1995, public and post-1995 utilities experienced on average 17.9 and 15.3 distributional losses. However whereas private utilities by 2005 reduced distribution losses by 12.6 percent, public utilities resulted with a 4.9 percent increase. More remarkable are the cases in which public utilities and post 1995 utilities experienced similar initial conditions in 1995, yet after the change in ownership diverged in their performance. One such instance is noted when assessing the quality of service. In 1995, public utilities experienced on average a frequency of 22 interruptions per connection, 5 interruptions less than that of private utilities. However by the end of the decade, public utilities reduced the average frequency of interruptions by 4, a modest improvement considering that private utilities cut their average frequency of interruptions by half. Moreover, this contradistinction is more evident when comparing the average duration time of private and public utilities. Whereas public and private utilities were separated by one hour duration in 1995, by the end of 2005, public utilities exhibit a 48.8 percent increase in the duration per connection, while private utilities improved the quality of service by reducing the duration per connection by 28.2 percent. There are good public and private utilities and underperforming private and public utilities. For several indicators the top 10 percent public utilities performed better than 20 the average private utilities and in other cases the bottom 10 percent of the private utilities performed worse than the average public utilities. In the case of distribution losses, it is noteworthy that the public utilities in the bottom 10 percent perform better than the average private utilities. Likewise the private utilities forming the top decile experience more distribution losses than the average public utilities. 2.5. F INAL R EMARKS 75. This chapter has not only weaved together results about the various trends and messages that characterize the electricity, water, and fixed communication sectors in LAC, but also accounted for the heterogeneity inherent in the elements that construe these respective sectors. By collecting information for over 250 public and private electricity distribution companies, 1700+ water and sanitation companies, and 40+ telecommunications companies in 32 countries, this study has comprehensively analyzed sector performance, through service provider performance. The database is rich, not only in the number and types of utilities surveyed, but also in the diversity and representativeness of the collected indicators. As a result, our conclusions are equally diverse and conditioned by the unique characteristics of each sector and service provider. LAC T AKES A L EAP F ORWARD IN S ECTOR P ERFORMANCE 76. Sector performance for electricity distribution, water and sanitation, and fixed telecommunications significantly improved in LAC. During the last 15 years, the region has witnessed significant improvements especially in coverage, service quality, and labor productivity in all sectors. Between 1990 and 2005, LAC's coverage in our sample increased to 95 percent for electricity distribution, 97 percent for the water utilities, and 62 percent in fixed telecommunications. A similar pattern of improvement is visible for all three sectors for labor productivity. For electricity distribution, labor productivity doubled since 1995 and for water, labor productivity almost doubled from 252 connections per employee to 425 in 2006. Wh en measuring labor productivity for telecommunications, the sector has experienced a seven-fold increase between 1995 and 2007. Another indicator that reflects a leap forward for LAC in utility performance is quality of service. For electricity distribution, the quality indicators exhibit significant improvements as measured by the frequency and duration of interruptions per connection showing a 42 and 40 percent reduction accordingly. The water sector also experienced a significant 8 percent increase in the continuity of service during this time period. Last but not least, the telecommunications sector reports gradual but significant increments in the percent of digital main lines, and telephone faults cleared by the next working day. For example, digital main lines have increased from 63 percent in 1995 to 100 percent in 2007. Comparably, the number of telephone faults per 100 main fixed lines per year dropped from 23 to 8 by the end of 2007. Such improvements in quality have also been accompanied by a reduction in the waiting list for main fixed lines, which by 2007 averaged to zero. LAC S TILL F ACES C HALLENGES : 77. While LAC has taken significant steps of progress in improving sector performance, the region still faces difficulties particularly in expanding services in rural areas, minimizing distributional losses, and in increasing cost-recovery levels. Despite the fact that electricity 21 coverage in LAC increased from 10 percentage points to 92 percent in 2005 there are still many people, almost all poor and in rural areas, without electricity. In the same vein, there are 29 million additional households, mostly in rural areas that are do not have a water connection. These figures present a strong need to expand electrification and water and sanitation services in rural areas in LAC countries since these areas lag behind. When assessing regional distribution losses, there is no apparent trend but rather sporadic increases and decreases throughout the last 15 year period. For electricity, the lowest distributional loss was observed in 2001, with a 0.9 percentage point decrease over a 14.5 percent distributional loss in 1995. Since 2001, the region has experienced a one percentage point increase, resulting in a 14.7 percent distributional loss in 2005. According to the water benchmarking results, non-revenue water has slightly increased during the last 10 years. While small, a one percentage point increase, from 38 to 39 percent, remains an obstacle for the region's water utilities. Last but not least, cost-recovery continues to hamper many utility performers in LAC. As outlined in the previous chapter, improving cost - recovery invokes an integrated approach often conditioned by tariff setting and OPEX and CAPEX. Considering the diversity of our sample set, this approach will depend on the specific conditions and environment of each sector and utility provider. 78. While these figures are solely a glimpse of what characterizes the electricity, water and sanitation, and telecommunications sectors in LAC, it provides important insights regarding the Region's strengths and weaknesses. It was part of the scope of this work to provoke further research that would build upon this knowledge and delve into the specific case studies represented by this data. LAC P RESENTS A W IDE SPECTRUM OF G OOD AND P OOR U TILITY P ERFORMERS 79. To better understand the dynamics in the sectors, we ranked companies and compared top, middle, and bottom performers for subsets of providers. We were able to benchmark companies not only at the country level, but also at the utility level. The wealth of data used to benchmark utility performance in LAC concludes throughout the past 15 years, the region has hosted a wide range of good and poor performers. For example, in water and sanitation, the top ten percent performers have 100 percent water and sanitation coverage on average. In contrast, the bottom ten percent water utilities average 66 percent coverage, while sewerage utilities average a low 15 percent coverage. Another example shows that electricity distribution utilities in the top ten percentile were ten times more productive in 2005 and sold six times the amount of energy (per connection) compared to utilities in the bottom ten percent. Finally, when comparing private and public utilities, our analysis shows that on average private utilities outperformed public utilities--but more importantly, there are good public and private utilities and underperforming private and public utilities. 80. Better understanding the progress, challenges, and diversity of LAC's infrastructure performers is in and of itself is an accomplished objective of this report. The utility performance data presented in this report is intended not only to answer basic questions about LAC's utility performance but to pose further analytical questions on how to move forward from the status quo. 22 3. UNDERSTANDING THE IMPACT OF PRIVATE SECTOR PARTICIPATION ON PERFORMANCE OF UTILITIES 81. Considering that infrastructure plays an essential role in fostering growth and reducing poverty and inequality, this chapter focuses on the private players that have significantly shaped the region's infrastructure trends. The 1990s were characterized by a massive policy redirection toward private participation in infrastructure (PPI). 22 The introduction of PPI was an attempt to compensate for the short-comings of state-operated utilities and thus improve the coverage and quality of the infrastructure sectors. In LAC, between 1995 and 1998, private participation went from roughly US$17 billion to a peak of more than US$70 billion, and then dropped back to US$20 billion by 2002 (World Bank 2007b). 82. Over the last decade, LAC governments have sought to involve the private sector in the provision of infrastructure services as a new source of managerial expertise and financing. Up until the 1980s, infrastructure services in the region and the rest of the world were exclusively operated and financed by public sector entities. This began to change in the 1990s, as a growing number of countries turned to a new approach for the infrastructure sectors. This phenomenon was based on the coincidence of two distinct but complementary trends. On the one hand, governments began to see the private sector as an attractive and manageable solution to the problems posed by infrastructure services. On the other hand, the private sector began to see the commercial attraction of investing in emerging economies. 83. As a result, private capital flows to infrastructure projects in developing countries grew six fold during the mid-1990s, but they declined sharply thereafter. From a baseline of US$20 billion in 1990s, investments swelled to a peak of US$131 billion in 1997. The increase was primarily driven by the rapid adoption of the new model in Latin America and East Asia. The countries of Eastern Europe and Central Asia were partly responsible for the increase, as the transition economies launched mass privatization programs. From 1997 until recently, private capital flows have been in marked decline. Triggered by the financial crises--and resulting currency devaluations--in East Asia and Latin America, this fall coincided with various corporate crises. Some of the major global energy and telecommunications companies were investing in emerging economies, which saw their average share prices fall by 90 percent and 70 percent, respectively. Private investment has fallen from $71bn in 1998 to $16bn in 2003, and the average contract only attracts two bidders. 84. This report acknowledges that there is an uphill battle in rebuilding public and business confidence in private-public partnerships in LAC. At the same time, it intends to demystify the notions surrounding Private Sector Participation (PSP) by establishing an empirical understanding of the main differences between private and public utilities when measuring performance are marked by: labor productivity, distribution losses, quality of service, and tariffs. In contrast, other indicators such as coverage and operation expenditures exhibit similar trends and/or do not present significant changes between the groups. 85. On average, private utilities performed better than public utilities with clear differences after the change in ownership. As it was described in Section 2.4, significant 22 The four main types of PPI are: i) management and lease contracts; ii) concessions; iii) greenfield projects; and iv) divestitures. In this chapter, PPI and privatization are used interchangeably to cover all four types. 23 improvements in labor productivity are a distinguishing factor when assessing the performance of the sector. When measuring the number of connections per employee in 1995, the labor productivity of post-1995 privatized utilities was only 10.7 percent greater than that of public utilities. Yet by the end of the decade, the labor productivity of post 1995 privatizations increased three-fold and doubled the amount of public utilities. Another indicator exhibiting significant improvement after the change in ownership is that of distribution losses. In 1995, public and post- 1995 utilities experienced on average 17.9 and 15.3 distributional losses. However whereas private utilities by 2005 reduced distribution losses by 12.6 percent, public utilities resulted with a 4.9 percent increase. 86. More remarkable are the cases in which public utilities and post 1995 utilities experienced similar initial conditions in 1995, yet after the change in ownership diverged in their performance. One such instance is noted when assessing the quality of service. In 1995, public utilities experienced on average a frequency of 22 interruptions per connection, 5 interruptions less than that of private utilities. However by the end of the decade, public utilities reduced the average frequency of interruptions by 4 (interruptions), a modest improvement considering that private utilities cut their average frequency on interruptions by half. Moreover, this contradistinction is more evident when comparing the average duration time of private and public utilities. Whereas public and private utilities were separated by one hour duration in 1995, by the end of 2005, public utilities exhibit a 48.8 percent increase in the average duration per connection, while private utilities improved the quality of service by reducing the average duration per connection by 28.2 percent. 87. There are good public and private utilities and underperforming private and public utilities. For several indicators the top 10 percent public utilities performed better than the average private utilities and in other cases the bottom 10 percent of the private utilities performed worse than the average public utilities. In the case of distribution losses, it is noteworthy that the public utilities in the bottom 10 percent perform better than the average private utilities. Likewise the private utilities forming the top decile experience more distribution losses than the average public utilities. 88. However, since the beginning of this decade, PSP has become a topic of contention amongst LAC governments and the region's ability to attract investors has diminished. In November 2000, 36 percent of Argentines believed that infrastructure services should come back under government control; five years later, 78 percent did (El Cronista April 18, 2005 ). This reflects a general trend in Latin America: with the exception of Panama, where about 40 percent of the population was expressing discontent with PSP in 1998. Today, the average is closer to 75 percent (see figure 2.17). Public resistance has become a major constraint on PPI in some countries, both politically and operationally. Currently, the average number of bidders for power distribution privatizations in Latin America fell from more than four in 1998 to less than two in 2000 and 2001 (Harris, 2003). 89. There is a remarkable contrast between generally positive evaluations of PSP and the extreme public disaffection of it. Martimort and Straub (2005) concluded that either important failures have gone unreported (although clearly not unnoticed by those who suffered) or there has been a major problem with perceptions (and therefore a massive communication failure). While estimates of impact on service coverage and quality, and redistribution are generally positive, it is possible that some negative aspects were under-reported. Some of these may include quality of the service that may have deteriorated or at least failed to improve as much as expected, re-distributional impact of price increases may not have been sufficiently mitigated by subsidies, and the record on job losses is clearly negative although the argument is 24 that losses tended to be reversed in the medium term (Fay and Morrison, 2006). Perceptions may be the main driver for this disaffection: More precisely, negative public perception of privatization may be due to the downturn in the economic cycle (Boix, 2005), perception may have suffered from a gap between actual and expected performance, and the perceived transparency of the privatization process is likely to be crucial in shaping public perceptions, among other issues (Fay and Morrison, 2006). 90. Perhaps the gravest misconception during the peak of PPI was that governments could now pass on responsibility for infrastructure financing and management to the private sector. While PPI held promises of a new flow of finance and technological innovations, it was not intended to substitute or play the role of the public sector, but rather complement it. As heavily emphasized in Fay and Morrison (2006), governments remain at the heart of infrastructure service delivery. Governments should continue to regulate and oversee infrastructure provision and pay for a large share of investments. 91. The challenge that currently faces the LAC region is the low level of public and private infrastructure investment. Low levels of infrastructure investment are a concern because of the widely documented link between infrastructure and growth, productivity, and poverty reduction (see Briceño-Garmendia, Estache, and Shafik 2004; Calderón and Servén 2004a; Fay and Morrison 2006). Public investment in infrastructure dropped from 3 percent of GDP in 1980 to less than 1 percent in 2001 in LAC (De Ferranti, Perry, Ferreira, and Walton 2004). In order to revive both public and private investment in the region, it is important to understand their distinct yet complementary roles and the true impacts and determinants of PSP in LAC. More so, if Latin American governments and private actors are to increase infrastructure investments in feasible ways, it is critical that they learn from past experiences, in order to make better investment and maintenance choices. 92. This chapter contributes to that aim by presenting a comprehensive and systemic analysis on the impact of PSP in LAC to date. This chapter 23 looks at what happened before, during, and after PSP in three sectors--electricity distribution, water and sewerage, and telecommunications--by focusing on a range of performance variables. It is necessary to look at all three periods, because often the most dramatic effects of PSP are found in the transition period, when the enterprise is overhauled as part of the transaction process. These constitute a one-time adjustment, however, and present a pace of improvement that is not necessarily sustained in the long run. The chapter focuses on changes and rates of changes in the three different periods, rather than on absolute numbers, because in many cases, the performance variables exhibit natural changes over time (with or without PSP). Hence, the analysis controls for such naturally occurring rates of changes. 93. The main results of this analysis, accounting for the counterfactual, are that the changes associated with PSP had a significant positive effect on labor productivity, efficiency, and quality. In addition, for telecommunications, it had significant effects on output and coverage. There were not conclusive results with respect to prices, although care should be exercised in any price impact analysis, because, most prices were highly distorted --most did not represent cost recovery--before the PSP programs. After the transition period to PSP, however, the improvements are not so striking. The main points to note are the consistent improvements in efficiency and quality, and reductions in the workforce. There do not appear to be significant 23 This chapter draws heavily on the work of Andres, L., J.L. Guasch, V. Foster, and T. Haven (2008), The Impact of Private Sector Participation in Infrastructure: Lights, Shadows, and the Road Ahead. 25 impacts on output and coverage, but prices tended to increase somewhat although the picture is highly variable across sectors. Table 3.1. Private Participation in Electricity Distribution, Telecommunications, and Water and Sewerage Source: Andres et al. (2008c). Note: Up and down arrows indicate that a positive or negative change occurred in addition to the natural change that would be expected in the absence of privatization. An equal sign indicates that the trend perceived during the previous period was sustained but not substantially exceeded or diminished. A question mark indicates that insufficient observations were available to reach a conclusion. The arrow size represents the size of the change. 94. The differences between publicly and privately operated distribution utilities showed up primarily with regard to labor productivity, distribution losses, quality of service, and tariffs. In contrast, other indicators such as coverage and operation expenditures exhibit similar trends or do not present significant changes between the groups. As seen in the previous chapter, there is significant variation in performance within both groups. The top 10 percent of performers in the public utility group outperformed the average private utility, and the average private utility outperformed the bottom ten percent of the private utility group. The analysis also addresses the determinants of performance. 95. The following sections of this chapter present a detailed analysis of the impact of privatization on the electricity distribution, telecommunications, and water sectors, respectively. 3.1. T HE I MPACT OF PSP IN E LECTRICITY D ISTRIBUTION 96. In the 1990s, the poor performance of the public model in electricity distribution beckoned for a reform in the sector. The reform was based on the introduction of market 26 principles, aimed to solve the main problems that besieged the public sector model: improve the quality, reliability, and efficiency of electricity services; improve the government's fiscal position; and increase affordable access to energy services for the poor. To achieve these objectives, a market-oriented reform promoted the following: i) the separation of roles of policymaking, regulation, and service provider, limiting the role of the state to policymaking and regulation, and relying on the private sector to be the main investor and provider of electricity service; and ii) the introduction of competition wherever possible and of economic regulation in the natural monopolies to improve economic efficiency. This market model would improve the government's fiscal position and ensure the financial sustainability of the sector by promoting the participation of private investment and the establishment of competitive prices for generation and cost-covering tariffs for transmission and distribution. It would be sustainable from a social and political point of view by improving access to energy services by the poor, based on a scheme of efficient subsidies. 97. This chapter summarizes the elements of the electricity distribution sector reform in the LAC Region and evaluates its impact. It draws on the work conducted by Andres et al. (2006) in which built an original data set and used documentation from 116 electricity distribution companies in the region for the years before and after their privatization. This study used two complementary methodologies to learn about the effects of changes in ownership: i) a means and medians analysis and ii) an econometric analysis. This section synthesizes the results from these two methodologies, presenting summaries of the impact, geared toward policymakers in the following areas: outputs and coverage, employment, labor productivity, efficiency, and prices. As mentioned previously, the period under analysis is separated into three parts: pre-privatization (pre-transition), a three-year transition period, and post-privatization (post-transition) 24. This allows for the study of short- versus long-term effects. 98. An in-depth analysis of the impact that privatization had on the electricity distribution sector indicates that the change in ownership did not change the growth trend for number of connections, energy sold, and coverage. Employment fell during both periods, but primarily during the transition. The labor productivity growth accelerated during the transition, followed by a deceleration during the post-transition period. Distributional losses and quality improved during both periods. Average prices in real local currency increased somewhat over both periods, although results for dollar price changes were less robust given Brazil's currency devaluation in 1999. 99. Output and Coverage: The number of connections, energy sold each year, and coverage levels increased across all three periods--pre-transition, transition, and post-transition--but effects were driven by trends. The trend in energy sold declined slightly after privatization. 100. Energy Sold: Two measures are used to estimate output: the MWhs of energy sold each year and the total number of connections at the end of each year. The amount of energy sold increased over all three periods: pre-transition, transition, and post-transition (see Figure 3.1). These increases were found to be statistically significant by both the means and median analysis (table A4.1) and the econometric analysis (table A4.3). According to the econometric analysis, the average amount of energy sold increased by 22.3 percent during the transition; the average 25 amount sold after the transition was 18.4 percent higher than transition levels. These output 24 We defined the transition period, as the time starting two years before the privatization or concession was awarded -- an approximation of when the reform was announced--and ending one year after the awarding; The pre-transition or pre-privatization period refers to the three years before the transition period and the post-transition or post-privatization period refers to the four years after the transition. 25 For the rest of the chapter, we refer to average for a given variable as the simple average within the country. 27 indicators seem to exhibit a natural rate of growth that must be controlled for to isolate the impacts of privatization. The econometric results show that there was a slight improvement in the growth trend during the transition. After the transition (during the post-privatization phase), however, the growth trend in the number of MWhs sold seems to have slowed slightly. 26 Figure 3.1: Electricity Distribution: Energy Sold and Number of Connections Electricity - Distribution Electricity - Distribution Avg number of MWHs Avg Number of Connections 160 100 120 140 160 140 120 100 80 80 60 60 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia El Salvador Guatemala Nicaragua Colombia El Salvador Guatemala Nicaragua Panama Peru ALL Panama Peru ALL Source: Andres et al (2008). Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 101. Number of Connections: The number of connections for electricity distribution increased significantly during the three periods. 27 According to the econometric analysis, the average level of connection numbers was 16.2 percent higher during the transition than in the previous period. The average level after the transition was 19.2 percent higher than during the transition (table A4.3). These increases were found to be statistically significant by both the means and median analysis (table A4.1) and the econometric analysis (table A4.3). An examination of Figure 3.1, however, shows that the increases largely followed a trend. The cross-country differences in the evolution of connection numbers potentially could be explained by differences in initial coverage conditions or differences in contract and regulator characteristics. 102. Coverage: There were improvements in electricity distribution coverage across all three periods: the average increase during the transition was 5.4 percent, and the average increase after that (with respect to transition levels) was 8 percent. Like the output increases, the coverage increases were statistically significant. But after controlling for time trends or when looking at changes in growth patterns, the impacts of privatization become negligible or difficu lt to discern. Actual differences in coverage across countries can be seen in figure 3.2. Brazil overtook Argentina to have the highest coverage level--more than 95 percent--during the post-transition period, and Guatemala experienced the largest jump between the before transition and after transition periods. 26 Possible reasons for this include the following: i) An overall decrease in the average consumption per household, perhaps because of the increase in prices (as it will be seen later in this chapter); ii) a change in the composition of the average household. Of those households that did receive electricity connections after a concession was awarded, it is very likely that they were mostly low-income families, with a smaller average consumption of energy; and iii) a reduction in distributional technical and commercial losses (as it will be seen later in this chapter). The data series was built using the total energy supplied to the distributional network, hence a reduction in losses could lead to a drop in MWh. 27 These increases were found to be statistically significant by both the means and median analysis (table A4.1) and the econometric analysis (table A4.3). 28 Figure 3.2: Before and After Comparison of Electricity Distribution: Coverage Levels Electricity - Distribution Coverage: Residential Connnection per 100 HHs 100 80 60 40 20 0 Argentina Bolivia Brazil Colombia El Salvador Guatemala Nicaragua Panama Peru Before Transition After Transition Source: Andres et al (2008). 103. Employment levels dropped substantially during the transition, not controlling for time trends. They also fell after the transition, but to a lesser extent. Most of the SOEs were characterized by having excess personnel. Hence, as expected, significant reductions in the number of employees clearly were observed across the three periods (Figure 3.3).28 The literature found that, in some cases, the government reduced the number of employees before privatization to increase the value of the firms (Chong and López-de-Silanes 2003). Consistent with the literature, this analysis found that labor force reductions during the transition were substantially larger than those after. Specifically, the econometric analysis found a 26.4 percent drop in the number of employees during the transition; after the transition, there was an additional drop of 17.6 percent. 29 Employment levels dropped substantially during the transition; they also fell after the transition, but to a lesser extent. Table A4.3 shows the changes in employment levels found by the econometric analysis. 104. With respect to distributional losses, the situation during public ownership was heterogeneous. Some countries had increasing distributional losses, but others had decreasing losses. After the transition, however, almost all the countries reduced their average distributional losses. The reason for the upturn in losses partway through the post-transition period in some countries is unclear (Figure 3.3). 105. The transition period saw an average drop in distributional losses of 3.1 percent, according to the econometric analysis. In contrast, distributional losses plunged 13.2 percent during the post-transition period (with respect to the transition period). When looking at the means and medians analysis, results tell a slightly different story. The mean for the transition period was 11.5 percent lower than the mean during the pre-transition period; the mean during the post-transition period was about 10 percent lower than during the transition period. When considering changes in the median, the results are more similar to the econometric analysis. The distributional loss median was 6 percent lower during the transition period and 11 percent lower during the post-transition period with respect to the previous period (see Table A4.1). In this case, it makes more sense to analyze changes in loss levels, rather than trends, because a natural trend is not expected. 28 Statistically significant drops were found by both the means and median analysis (Tables A4.1 and A4.2) and the econometric analysis (table A4.3). 29 The means and medians analysis found complementary results: The mean number of employees during the transition was 38 percent lower than before the transition, and the mean number of employees after the transition was 14 percent lower than during the transition (see table A4.1). 29 Figure 3.3: Electricity Distribution: Employment Electricity - Distribution Electricity - Distribution Avg Number of Employees Distributional Losses 250 120 200 100 150 80 100 60 50 40 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia El Salvador Guatemala Nicaragua Colombia El Salvador Guatemala Nicaragua Panama Peru ALL Panama Peru ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 106. The mixed results are likely the result of a conflation of the two types of distributional losses: technical and commercial. To curb technical losses, new investments and upgrades are required that take time to implement. Hence, they would be expected to occur following the transition period. Commercial losses, on the other hand, can often be reduced quickly by shutting off the connections of nonpaying customers. Thus, drops in distributional losses during the transitional period could be attributed to commercial losses. 107. Labor productivity: Connections per employee and energy sold per employee showed large gains in levels during both the transition and post-transition periods. With respect to the connections and energy per employee, the results are a composition of the previous comparisons. These results are driven by the positive trend in the output measures and by the reduction in the number of employees. Although the greatest gains came during the transition period, levels of both connections per employee and energy per employee showed significant improvements during the transition and post-transition periods relative to the previous period (Figure 3.4). 30 According to the econometric analysis, connections per employee were 55.6 percent higher during the transition and another 44.5 percent higher after the transition. Equivalent numbers for energy sold per employee are 60.6 percent and 41.3 percent. 108. Given the underlying data--connections and energy sold, which follow natural trends versus employment, which does not--it is argued that it is more appropriate to analyze labor productivity after controlling for trends. As was the case for the output and labor indicators, controlling for trends dramatically reduces the privatization impacts. With the effect of time trends removed, connections per employee and energy per employee increased by 5 percent and 9 percent, respectively, during the transition. Levels after the transition decreased slightly (-3.6 percent for connections per employee and -7.7 percent for energy per employee, with respect to transition levels). The econometric growth rate analysis produced similar results: the average annual growth rate for both connections per employee and MWh per employee increased during the transition and decreased after the transition. 30 The level increases were found to be statistically significant by both the means and median analysis (Tables A4.1 and A4.2) and the econometric analysis (Table A4.3). 30 Figure 3.4: Electricity Distribution: Labor Productivity Electricity - Distribution Electricity - Distribution Labor Productivity: Connections per employee Labor Productivity: MWHs per employee 300 100 150 200 250 300 250 200 150 100 50 50 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia El Salvador Guatemala Nicaragua Colombia El Salvador Guatemala Nicaragua Panama Peru ALL Panama Peru ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 109. Connections per employee and energy sold per employee showed large gains in levels during both the transition and post-transition periods. When looking at growth rates, however, a temporary growth acceleration occurred during the transition followed by a deceleration after the transition. Distributional losses declined in both periods. 110. Average prices in real local currency increased somewhat during transition and post- transition. Dollar prices appear to have fallen, but after excluding Brazil (which experienced a currency devaluation in 1999), dollar prices seem to have increased slightly. Average residential electricity prices in U.S. dollars and in real local currency are analyzed. The results seem somewhat peculiar--the tariffs in real local currency show a clearly increasing trend, but prices in dollars seem to be decreasing in the same period. The econometric analysis showed statistically significant increases in real local currency prices of 11.1 percent during the transition and 7.4 percent after the transition (with respect to the transition level). In dollars, there was no significant change during the transition period and a -2.8 percent drop during the post-transition period. 111. A plausible explanation for this is, in part, the 1999 currency devaluation in Brazil. To test this explanation, the analysis was repeated with Brazil excluded from the sample. With Brazil excluded, both series show increasing prices, but at a much lower rate. As a result of the smaller sample size and relatively small price changes, no significant differences were found between consecutive periods in the means and medians analysis. According to the same analysis, there were small but significant price increases in both local currency and dollars when comparing the pre-transition and post-transition periods. 112. Quality of the Service: There is a relatively small amount of quality data from the pre- transition period, but available data do indicate that both (i) the average duration of interruptions per consumer and (ii) the average frequency of interruptions per consumer fell during both the transition and post-transition periods. Combining these two indicators yields an overall quality measure that shows improvement in both periods. The quality of electricity distribution is measured by the frequency and duration of service interruptions per consumer. In general, these measures were defined at the time of reform, along with the creation of regulatory agencies, making it difficult to build long time series. Only Argentina and Brazil had some information for the years before the transition. Despite the lack of historical data, quality improvements on average have been substantial. Argentina stands out as having been particularly successf ul in reducing the average duration and frequency of interruptions per consumer, both in relative and 31 absolute terms. Bolivia, on the other hand, has experienced some quality deteriorations since the privatization transition, even though absolute quality levels are second only to Argentina (Figures 3.5). In countries where quantitative quality data before privatization are not available, strong anecdotal evidence suggests that quality was poor. Figure 3.5: Electricity Distribution: Quality Electricity - Distribution Electricity - Distribution Quality: Avg Duration of interruptions per consumer Quality: Avg Frequency of interruptions per consumer 200 250 200 150 150 100 100 50 50 0 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia El Salvador Guatemala Nicaragua Colombia El Salvador Guatemala Nicaragua Panama Peru ALL Panama Peru ALL Source: Authors' calculations. Note: The average line for all countries appears more erratic because of the relative scarcity of data. The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 113. Both of the analysis methodologies found improvements in average frequency and duration of interruptions. According to the econometric analysis, the duration of interruptions fell 13.4 percent during the transition and an additional 29.1 percent after the transition. Similarly, the frequency of interruptions fell 10.1 percent during the transition and an additional 26.5 percent after it. 31 The means and medians analysis found similar quality improvements, although the frequency of interruptions results was not statistically significant for the post-transition period. 32 114. Available data suggest that both (i) the average duration of interruptions per consumer and (ii) the average frequency of interruptions per consumer fell during both the transition and post-transition periods. Combining these two indicators yields an overall quality measure that shows improvement in both periods. ... I N S UMMARY ... 115. The main results of this section are that the change in ownership did not change the growth trend for number of connections, energy sold, and coverage. 33 Employment fell during both periods, but primarily during the transition. The labor productivity growth accelerated during the transition, followed by a deceleration during the post-transition period. Distributional losses and quality improved during both periods. Average prices in real local currency increased somewhat over both periods, although results for dollar price changes were less robust given Brazil's currency devaluation in 1999. 31 These drops in interruptions were all statistically significant. 32 The means and medians analysis found a 23 percent drop in the duration of interruptions between the pre-transition and transition periods and a 25 percent drop between the transition and post-transition periods. Both of these drops were significant. The frequency of interruptions fell 26 percent between the pre-transition and post-transition periods and no statistically significant change occurred between the transition and post-transition periods (Table A4.1). 33 The results for the output, coverage, and labor productivity indicators are reported after controlling for time trends. If time trends were not controlled for, each of these indicators would show significant increases. A natural increase is expected for each of these variables, regardless of whether ownership is public or private. 32 3.2. T HE I MPACT OF PSP ON W ATER AND S EWERAGE 116. Growing dissatisfaction with the performance of the national water monopolies, combined with wider political pressure for devolution across all areas of government, created the conditions for a move toward decentralized control in the 1980s and 1990s. Chile was the first to attempt to modernize its water sector with new legislation passed as early as 1988. By 1991, both Argentina and Mexico were beginning to conduct a series of experiments with PSP. In a second wave, Peru, Colombia, and Bolivia enacted ambitious new legislation in the mid-1990s. During the second half of the decade, reform began to take root in Brazil and Central America. By the end of the 1990s, few countries remained that had not either completed reforms, had major reforms in process, or were actively considering reforms. In general, the water sector reforms were composed of three components: decentralization, regulation, and PSP. 117. As part of the reform process, many countries created national regulatory agencies for water, similar to the Water Services Regulation Authority (Ofwat) model developed in the United Kingdom. The responsibilities of these agencies typically included the determination of tariffs, approval of investment plans, oversight on quality of service, and consumer protection. In some cases (for example, Peru), the agencies did not have final authority to determine tariffs. In the larger federal countries (Argentina, Brazil, and Mexico), regulatory functions were often organized at the state or provincial level. The regulatory agencies were seen as a precursor to private participation in the sector, although the ultimate scope of private participation was modest relative to initial expectations. 118. Historically, the water and sewerage sectors have not been well analyzed in Latin America. In contrast to electricity distribution and telecommunications, firms tend to be based at the local or regional government level, making the private participation process slower and more fragmented. Despite the slow process, currently, at least 11 percent of the water in Latin American households is supplied by private firms. For the analysis in this section, data were collected for 49 firms with a change in ownership in the last 15 years. Similar to the electricity distribution and telecommunications sections, two complementary methodologies were used to learn about the effects of changes in ownership: a means and medians analysis and an econometric analysis. 119. The following section is a brief summary of the impact that privatization has had on the water and sewerage sector. Output and coverage measures improved, but the improvements were consistent with the existing trend. Meanwhile, the number of employees dropped substantially during the last years under public management. These changes significantly increased labor productivity, especially during the transition period, but when looking at growth rates, labor productivity rates accelerated during the transition and decelerated in the post-transition period. Efficiency--measured by distributional losses--improved mainly after the transition. Price increases were seen in both water and sewerage, although the increases for sewerage were generally not robust because of a small sample size. Two measures were used for quality: the continuity of the water service and the number of water samples that passed a potability test. Both measures improved in both periods, but potability improvements occurred mainly during the transition. 33 120. Output and Coverage: The number of water and sewage connections increased during the transition and post-transition periods, but these improvements were consistent with existing trends. Similar results were found for both water and sewerage coverage. Water production increased somewhat in both periods, but after controlling for trends, a small growth deceleration occurred in the post-transition period. Two variables are used to measure output in the water and sewerage sector: the number of residential connections (for both water and sewerage) and the amount of water produced (in cubic meters) each year. The number of connections for both water and sewerage increased substantially during both the transition and post-transition periods (Figure 3.6). In fact, the econometric analysis found significant increases of between 15 and 20 percent for each period (see Annex 4). The means and medians analysis found similar results, which can be found in Annex 4. A closer look at the results shows the increases can be accounted for by the existence of a trend. After controlling for firm-specific time trends, the econometric analysis found no significant changes in the number of water or sewerage connections. When considering growth rates, the econometric analysis found no significant changes during the transition, while the average annual growth rate fell by 1 percent after the transition for both water and sewerage. 34 121. The second output indicator is the number of cubic meters of water produced per year (Figure 3.6). The econometric analysis found that water production increased by 4.1 percent during the transition and an additional 1.5 percent after the transition. However, taking trends into account--by controlling for firm-specific time trends or looking at changes in growth rates-- erases those gains. In fact, the econometric analysis found no significant change in water production during the transition and a small drop after the transition. 35 As will be seen later, a possible justification for this deceleration is the improvement in efficiency caused by the reduction of distributional losses. 122. Coverage in both water and sewerage improved during the transition and post-transition periods (figure 3.6). According to the econometric analysis, these improvements were statistically significant and ranged from 2.5 percent to 6.7 percent. The means and medians analysis found similar increases of between 6.9 and 11.1 percent (table A3.10). These improvements apparently were driven by trends, however, and they likely would have occurred in the absence of privatization. After controlling for firm-specific time trends, the econometric analysis found no significant changes. And looking at growth rates yielded no significant changes during the transition period, combined with a small drop in the average annual growth rate of 0.4 percentage points for water and 0.8 percentage points for sewerage after the transition. Not surprisingly, these results are quantitatively similar to those found for the number of connections above. 34 When actual (as opposed to normalized) water connection numbers are considered, Argentina and Chile stand out as having the largest water distribution companies. For sewerage, Argentina, Chile, and Colombia have companies of roughly the same size. In contrast to the results found in other sectors, water and sewerage companies in Brazil and Mexico fall at the small end of the spectrum (Andres et al., 2008c). 35 The only significant result of the means and medians analysis was a drop of roughly 3 percent in the mean amount of water produced between the transition and post-transition periods. 34 Figure 3.6: Water and Sewerage: Water - Distribution Water - Distribution 140 Avg Number of Total Connections (water) Avg Number of Total Connections (sewer) 200 120 150 100 100 80 60 50 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Water - Distribution Water - Distribution Coverage: Water Production of Water (m3/yr) 130 120 120 110 110 100 100 90 90 80 70 80 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Water - Distribution Coverage: Sewer 160 140 120 100 80 -5 0 5 t Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 123. When actual (not normalized) water coverage levels are considered, levels for most countries are relatively high--more than 90 percent (Figure 3.7). Mexico stands out as an exception with less than 80 percent coverage. For sewerage, actual coverage levels are lower -- closer to 60 percent for some countries. Chile is the outlier with close to 100 percent sewerage coverage. 35 Figure 3.7: Before and After Comparison of Water and Sewerage: Coverage Levels W ater - Distribution Coverage: Water 100 80 60 40 20 0 Argentina Bolivia Brazil Chile Colombia Mexico Before Transition After Transition Source: Andres et al (2008). Note: The y axis is the number of connections per 100 inhabitants. 124. The number of employees declined during transition and post-transition, not accounting for time trends. All of the analyses found significant drops in employment during both of these periods, although the drop during the transition seems to have been the greatest (Figure 3.8). Specifically, the means and medians analysis found a 26.3 percent drop during the transition and a 11.7 percent drop after the transition. The econometric analysis found a 16.5 percent drop during the transition and a 17.6 percent drop after the transition. 36 Figure 3.8: Water and Sewerage: Employment and Labor Productivity Water - Distribution Water - Distribution Number of Employees Connections of drinkable water per employee 140 250 120 200 100 150 80 100 60 50 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 125. Given that most SOEs had excess numbers of personnel, the drops seen during the transition period should not be surprising. Many governments opted to trim the labor force before the ownership change in an attempt to increase the value of the firm. Argentina stands out as having by far the most employees as well as experiencing the largest absolute reduction in employee numbers between the pre-transition and post-transition periods. 126. Labor productivity--measured by the number of water connections per employee-- clearly increased greatly during both the transition and post-transition periods (Figure 3.8). This 36 While a natural trend in employment is not expected, the numbers after controlling for trends are reported in Annex 3. 36 was a result of changes in the underlying indicators: the number of connections increased while the number of employees fell. The econometric analysis found that water connections per employee increased 30.7 percent during the transition and another 42.5 percent after the transition. The means and medians analysis found similar large jumps. 127. Controlling for trends tells a somewhat different story. According to the econometric growth rate analysis, the average annual growth rate of connections per employee increased by 4.7 percentage points during the transition. This was followed by a drop of 3.7 percentage points after the transition (relative to the transition levels). In other words, there was a temporary acceleration in labor productivity growth (largely because of employment changes) during the transition, and then the annual growth rate returned to roughly 1 percentage point above the pre- transition level. The means and medians analysis identified similar changes: a 11.6 percentage point increase during the transition followed by a 9.6 percentage point decrease after the transition. There was no statistically significant difference between the pre-transition and post- transition growth rates in the means and medians analysis. 128. Distributional losses clearly fell substantially during both the transition and post- transition periods (Figure 3.9). Indeed, the econometric analysis found a 3.8 percent drop in the percentage of water lost during the transition period followed by a 14.4 percent drop during the post-transition period. The means and medians analysis found results of a slightly larger magnitude (8.1 percent and 18.3 percent, respectively). Trends are not controlled for because a natural trend is not expected, and figure 3.9 does not signal a trend in the pre-privatization period. Figure 3.9: Water and Sewerage: Distributional Losses Water - Distribution % of lost water 120 110 100 90 80 70 -5 0 5 t Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 129. Water prices in dollars showed little change during the transition (due to Brazil's devaluation) and rose after the transition. Water prices in real local currency increased fairly substantially in both the transition and post-transition periods. Because of the small sample size, not much can be said about sewerage prices; however, a significant sewerage price increase in real local currency occurred during the post-transition period. 130. Four measures of prices were analyzed: (i) water prices in dollars, (ii) water prices in real local currency, (iii) sewerage prices in dollars, and (iv) sewerage prices in real local currency. Water prices seem to have increased in both periods in dollars and in real local currency (Figure 3.10). Brazil's currency devaluation in 1999 accounted for the main difference between the two types of currencies. As a result of the devaluation, Brazil's water prices in dollars fell, while they 37 mainly rose in real local currency. Given that the Brazil devaluation skewed the dollar price's results so that they appeared artificially low, it is preferable to look at the changes in real local currency. 131. According to the econometric analysis, water prices in dollars did not change significantly during the transition, but increased by 10.2 percent after the transition. In contrast, water prices showed statistically significant increases in real local currency of 15.7 percent during the transition and 23.7 percent after transition. In the means and medians analysis, there were no significant changes between adjacent periods in dollars, but there was a statistically significant increase between the pre- and post-transition periods. In real local currency, the means and medians analysis found significant price increases in each period. When Brazil was excluded from the sample, the means and medians analysis found statistically significant increases of 32.6 percent during the transition and 16.9 percent after the transition. Figure 3.10: Water and Sewerage: Water Prices Water - Distribution Water - Distribution Average price per m3 in dollars [water] Average price per m3 in real local currency [water] 200 160 140 150 120 100 100 80 50 60 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Water - Distribution Water - Distribution Average price per m3 in dollars [sewer] Average price per m3 in real local currency [sewer] 300 200 250 150 200 150 100 100 50 50 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 132. Sewerage prices seem to have behaved in a similar fashion as water prices (Figure 3.10). Because of the small number of observations, that results were not statistically significant for the most part. In fact, according to both the econometric and means and medians analyses, the only significant change was an increase in real local currency prices after the transition period (this increase was 24.9 percent in the econometric analysis). 133. Water prices in dollars showed little change during the transition (thanks to Brazil's devaluation) and rose after the transition. Water prices in real local currency increased fairly 38 substantially in both periods. Not much can be said about sewerage prices because of the small sample size. However, a significant sewerage price increase in real local currency in the post- transition period was found. 134. Improvements in service continuity appear to have occurred during both the transition and post-transition periods, although no improvements occurred during the pre-transition period (Figure 3.11). The means and medians analysis found that on average continuity improved by 27.8 percent during the transition period and 14.8 percent after the transition Presumably because of a relatively small sample size, the econometric analysis found a statistically significant improvement (of 7.7 percent) only in the post-transition period. Figure 3.11: Water and Sewerage: Service Continuity and Quality Water - Distribution Water - Distribution Continuity of the service (hs per day) % of pass the potability test 120 140 100 120 80 100 60 80 60 40 -5 0 5 -5 0 5 t t Argentina Bolivia Brazil Chile Argentina Bolivia Brazil Chile Colombia Mexico Trinidad ALL Colombia Mexico Trinidad ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 135. Despite a relatively small number of observations, it seems evident that water potability improved (Figure 3.11). Most of the changes occurred during the transition: according to the econometric analysis, potability improved by 6.1 percent during the transition and 1.2 percent in the post-transition period. Given that potability numbers were already close to 100 percent for many countries (with the exception of Colombia), it is not surprising that the improvements seen in the post-transition period were quite modest. ... I N S UMMARY ... 136. In the Water and Sewerage Sector, output and coverage measures improved, but the improvements were consistent with the existing trend. Meanwhile, the number of employees dropped substantially during the last years under public management. These changes significantly increased labor productivity, especially during the transition period, but when looking at growth rates, labor productivity rates accelerated during the transition and decelerated in the post- transition period. Efficiency--measured by distributional losses--improved mainly after the transition. Price increases were seen in both water and sewerage, although the increases for sewerage were generally not robust because of a small sample size. Two measures were used for quality: the continuity of the water service and the number of water samples that passed a potability test. Both measures improved in both periods, but potability improvements occurred mainly during the transition. 39 3.3. T HE I MPACT OF PSP ON F IXED - LINE T ELECOMMUNICATIONS 137. During the 1980s and the 1990s, the state owned the fixed telecommunications company, which operated in a monopolistic market. After Chile's experience in the 1980s, most LAC countries privatized their telecom companies. 37 The new owners generally had to comply with requirements such as network expansion and quality standards. In exchange, they were granted a monopoly period, after which new firms could enter the market. In most countries, liberalization of the long-distance market took place within a few years after privatization. Therefore this analysis takes into account the possibility that the impacts of privatization perceived were actually instead caused by liberalization. 138. This section analyzes a data set 38 built by ITU (2008), 39 that covers 16 fixed telecommunication companies that were privatized in LAC and contains information for the years before and after their privatization. Similar to the previous sectors, two complementary methodologies were used to learn about the effects of changes in ownership: a means and medians analysis and an econometric analysis. In addition, the period under analysis is separated into three parts: pre-privatization (pre-transition), a three-year transition period, and post- privatization (post-transition). 139. Number of Connections: Two variables are used to measure output in the fixed telecommunications sector: the number of connections and the number of local minutes consumed each year. As seen in Figure 3.12, the number of connections increased during all three periods for almost all countries. Both the means and medians analysis and the econometric analysis confirmed that there were statistically significant increases in the number of connections between the pre-transition, transition, and post-transition periods (see Annex 4, tables A4.4 and A4.6). In fact, the econometric analysis found a 29 percent increase in the number of connections during the transition period and an additional 64 percent increase during the post-transition period. Figure 3.12: Telecommunications: Number of Connections and Number of Minutes Fixed Telecommunications Fixed Telecommunications Avg Number of Connections Avg Number of Minutes 300 200 250 n_connections_100 150 output_100 200 100 150 50 100 50 0 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 37 Andres et al (2008c). 38 Currently, only six countries remain with public telecommunications firms: Colombia, Costa Rica, Ecuador, Honduras, Paraguay, and Uruguay. 39 See description in Annex 2. 40 140. Turning to growth trends, Figure 3.12 indicates that growth in the number of connections accelerated, possibly temporarily, in the first few years of private ownership. The means and medians analysis found that average annual growth in the number of connections increased from 6.9 percent in the pre-transition period to 12.7 percent during the transition period, before falling back to 7.2 percent in the post-transition period. Similarly, the econometric analysis found that the average annual growth rate increased 2.7 percentage points during the transition, while there was no statistically significant change from that level after the transition. 40 After controlling for trends it seems that an increase occurred during the transition, but growth rates returned to normal levels after the transition. 141. One possible explanation for the surge in the number of connections during and shortly after the transition is that newly privatized companies took action to meet pent-up demand. According to the ITU, waiting lists for connections in the year before the reform numbered 780,000 in Argentina, 308,247 in Peru, and 175,000 in El Salvador. These numbers accounted for 26 percent, 46 percent, and 54 percent of the connections in operation at the time in Argentina, Peru, and El Salvador, respectively. Another contributing factor was the spread of mobile telecommunications, especially during the second half of the 1990s, which likely reduced the demand for new fixed connections. 142. As Ros (1999) pointed out, private ownership in fixed telecommunications could shift priorities away from network expansion. This shift occurs because, in a private company, shareholders may be reluctant to increase the network unless it is profitable or made mandatory in the contract. While this may be true, the analysis in this book finds that privatization led to greater network expansion. 143. The second output indicator is the number of minutes consumed per year. Figure 3.12 shows that, with the exception of Argentina, the average number of minutes consumed was generally increasing and growth was particularly strong after the transition. These results are not surprising given the increasing number of connections discussed above. The means and medians and econometric analyses generally confirm what can be seen in the figure, although results are not always robust because of the relatively small number of observations. For instance, the econometric analysis found statistically significant increases of 8.2 percent and 37.6 percent during the transition and post-transition periods, respectively. 41 144. When time trends are taken into account in the econometric analysis, there is no significant change during the transition period, whereas the post-transition period shows an increase of 14.2 percent over transition levels. The growth regressions, on the oth er hand, find statistically significant increases in the growth rates of 6.9 percentage points during the transition period and 5.3 percentage points during the post-transition period. Hence, the preponderance of evidence suggests that the number of minutes of fixed telecom services increased in both the transition and post-transition periods after controlling for the trend. 145. Consistent with the output measures, coverage (or teledensity, defined as the number of connections per 100 inhabitants) increased substantially during the periods under study (Figure 40 Results from the econometric analysis that controls for firm-specific time trends tell a somewhat different story . The number of connections fell by 4.9 percent during the transition, but then increased by 12 percent after the transition (with respect to transition levels). This model specification is less useful in this particular case, however, given the fluctuating nature of the underlying data. 41 The means and medians analysis did not find a statistically significant difference between the pre-transition and transition periods. Based on two observations, the analysis found that the average number of minutes was 40 percent higher during the post-transition period than during the transition period (see Table A4.4). 41 3.13). In fact, the econometric analysis found an increase of 18.3 percent during the transition period and an additional increase of 52.3 percent during the post-transition period. Similarly, the means and medians analysis found substantial, statistically significant increases. Figure 3.13: Telecommunications: Coverage and Employment Fixed Telecommunications Fixed Telecommunications Coverage Number of Employees 150 250 200 cn_100hb_100 n_wrks_100 100 150 100 50 50 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownersh ip. The y axis is normalized at 100 when time=0. 146. Looking at trends and growth rates indicates that coverage grew more rapidly during the transition period. The econometric analysis found that the annual growth rate increased by 3.7 percentage points during the transition period and registered no additional changes after the transition. The means and medians analysis found that the average annual growth rate increased by 6.1 percentage points during the transition period, but then fell by 5.9 percentage points (relative to transition rates) during the post-transition period. 42 147. Figure 3.14 compares actual coverage levels across countries. While considerable heterogeneity exists, most countries in the sample have coverage levels of between 10 and 20 connections per 100 inhabitants. 148. The number of connections increased during both periods, but after controlling for trends, only the transition period showed abnormally high growth rates. Again, after controlling for trends, the number of minutes increased in both periods, whereas the increases in coverage occurred mainly in the transition period. 149. The number of employees declined during the transition and post-transition periods, not accounting for time trends. On average, the number of employees in fixed telecommunications companies has been declining steadily since before the transition period. However, this average decline masks considerable differences across firms and countries (Figure 3.13). The econometric analysis found that employment declined by 9.2 percent during the transition period and a further 23.2 percent after the transition period. 43 A natural trend in employment is not expected, but employment growth rates became increasingly negative during the transition and post-transition periods. The econometric analysis found that, during the transition, the annual growth rate of 42 The econometric analysis that controlled for firm-specific time trends found that coverage fell by 6.3 percent during the transition and then increased by 9.5 percent during the post-transition period. This model specification may be less applicable, however, given the shape of the underlying data, (that is, the time trend analysis becomes less accu rate when there is more than one shift in the presumed trend). 43 The means and medians analysis found that employment fell 14.5 percent during the transition and 18.2 percent more after the transition. All of these changes were statistically significant. 42 employment was 4.1 percentage points lower than during the previous period; annual growth fell an additional 2.6 percentage points after the transition. Figure 3.14: Before and After Comparison of Telecommunications: Coverage Levels Fixed Telecommunications Coverage 25 20 15 10 5 0 Argentina Bolivia Brazil Chile El SalvadorGuatemala Guyana Jamaica Mexico Nicaragua Panama Peru Trinidad Venezuela Before Transition After Transition Source: Andres et al. (2008c). Note: The y axis is the number of connections per 100 inhabitants. 150. One reason for the fall in employment during that transition period is that governments decided to trim the labor force reform before the ownership change with the intention of increasing the value of the firm and bringing employment to a more sustainable, equilibrium level. As mentioned in the electricity distribution section, investors proved indifferent to these policies and, in the end, the value of the firm remained at the same level or was even reduced when the government applied layoff programs in advance. One explanation is selection issues that provide incentives for good employees to leave while bad employees remain in the company (Chong and Lopez-de-Silanes 2003a). 151. Two indicators were used to measure labor productivity: connections per employee and minutes per employee. As a consequence of the increase in the output measures and the general negative trend in the number of employees, labor productivity improved substantially, especially after the transition (Figures 3.14). Almost all of the countries in the data set at least doubled labor productivity in less than five years after the reform. The only exception was Panama, which already had a relatively high teledensity (that is, the number of active connections per 100 inhabitants). At the time of the reform, Panama's teledensity was 13 percent; neighboring Nicaragua, El Salvador, and Guatemala had teledensities of 3 percent, 6 percent, and 4 percent, respectively. 44 152. According to the econometric analysis, the number of connections per employee increased 35.1 percent during the transition (compared with the pre-transition period) and a whopping 106.9 percent after the transition. The results of the means and medians analysis were even greater: the increase during the transition was 65.6 percent, and the increase after the transition was 117.9 percent (see table A4.4). All changes were statistically significant. 153. Fewer data were available for minutes per employee, but the econometric analysis still found impressive statistically significant improvements: 32 percent during the transition and an 44 Panama was a special case in that it actually had more connections in 1998 than in 2003. In 1998, 419,000 subscribers had fixed connections; at the end of 2003, only 380,000 had fixed connections. Not surprisingly, mobile telecommunications proliferated during the same years. In fact, mobile subscribers surpassed fixed-line subscribers, jumping from 49,000 in 1998 to 834,000 in 2003 (Ente Regulador de los Servicios Públicos, 2004). 43 additional 92.9 percent after the transition. Again, the means and medians analysis found even larger increases: 43.2 percent during the transition and 117.2 percent after the transition. Figure 3.14: Telecommunications: Labor Productivity Fixed Telecommunications Fixed Telecommunications Labor Productivity: Connections per Employee Labor Productivity: Minutes per Employee 400 600 300 output_worker_100 conn_worker_100 400 200 200 100 0 0 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 154. As was the case for the output indicators, controlling for trends dramatically reduces the impact of privatization on labor productivity (the results can be seen in Annex 4). Yet, it is more appropriate to look at the changes in trends given the underlying indicators: in the previous sections, it was argued that the output indicators follow natural trends, but the number of employees does not. One way to examine trend changes is through growth rates. In this case, the annual growth rate of number of connections per employee increased by 7 percentage points during the transition period and 3.3 percentage points after the transition. The annual growth rate of minutes per employee increased by 8.5 percentage points during the transition period, but did not register any additional statistically significant changes during the post-transition period. 155. Actual (that is, not normalized) labor productivity measures show a surprising amount of variance. Brazil is by far the most productive with more than 1,000 connection s per employee during the post-transition period. The next-closest country, Bolivia, had less than one-half that number. The number of minutes per employee in Brazil vastly exceeds that of other countries. 156. The percentage of incomplete calls was chosen as the most feasible measure of fixed telecommunications efficiency. While considerable heterogeneity exists across countries, figure 3.15 shows a substantial drop in the average percentage of incomplete calls. Despite a relatively small number of observations, the econometric analysis confirmed that there was indeed a statistically significant drop of 29.7 percent in the post-transition period. Neither the econometric results from the transition period nor the results of the means and medians analysis were statistically significant. 157. The network digitization percentage was selected as a proxy for quality in fixed telecommunications. Network digitization increased during the transition and post-transition periods, with the largest increase coming during the transition, not controlling for time trends. The econometric analysis found increases of 36.3 percent during the transition and 58.1 percent after the transition. Similarly, the means and medians analysis found increases of 75.4 percent and 69.5 percent in the two periods, respectively. 44 Figure 3.15: Telecommunications: Percentage of Incomplete Calls Fixed Telecommunications Fixed Telecommunications % of Incompleted Calls Quality Index 120 300 250 100 quality_index_100 dis_loss_100 200 80 150 60 100 40 50 20 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least 6 months of public ownership. The y axis is normalized at 100 when time=0. 158. A natural trend is not assumed, but it is still interesting to control for trends and examine growth rate changes. The econometric analysis found that after controlling for firm-specific time trends, there was a statistically significant increase of 4.9 percent during the transition period; there was no significant change after the transition. On the other hand, the econometric growth analysis found a significant drop in the average annual growth rate of 5.6 percentage points after the transition but no significant change during the transition. 159. To provide a somewhat more robust measure of quality, a quality index was created that combines the percentage of completed calls and the share of the network that was digitized. The quality index steadily improved across all periods (Figure 3.15) and actual quality levels after the transition were generally comparable across countries (Figure 3.15). One exception was República Bolivariana de Venezuela, which experienced large gains but fell well short of actual levels in other countries. Network digitization increased during both periods, with the largest increase coming during the transition. 160. Three measures of fixed telecommunications prices were analyzed in both dollars and real local currency: (i) the average price of a three-minute local call, (ii) the average monthly charge for residential service, and (iii) the average charge for the installation of a residential line. The average price of a three-minute local call was mainly increasing during public ownership. One exception was Chile, which experienced a tremendous fall in prices leading up to the ownership change. On average, however, prices increased during the first part of the transition, reaching a high point during the last year of public ownership. Prices then began to fall, but not as rapidly as the increases of previous years (Figure 3.16). Trends in U.S. dollars and real local currency followed roughly similar patterns, although the 1999 devaluation in Brazil introduced some variation. 161. The econometric analysis found that average prices in both dollars and real local currency for a three-minute call increased by roughly 45 percent. There were no significant changes during the post-transition period, and the means and medians analysis did not find any statistically significant changes during either period. 162. Monthly charges for residential service increased significantly during and after the transition, both in dollars and in real local currency. The changes were largest during the transition: prices in dollars grew 75.9 percent and prices in real local currency grew 62.6 percent. After the transition, both dollar and real local currency prices were roughly 22 percent higher than 45 transition levels. The means and medians analysis also found significant jumps (see Table A4.4). Judging from Figure 3.16 and the econometric trend analysis, it appears that residential monthly charges experienced an abnormal jump during the transition before returning to a slower rate of growth similar to the pre-transition period. Figure 3.16: Telecommunications: Three-Minute Call Prices Fixed Telecommunications Fixed Telecommunications Price of a 3-minutes Call (in Dollars) Price of a 3-minutes Call (in Real Local Currency) 400 250 200 300 cost_3m_d_100 cost_3m_c_100 150 200 100 100 50 0 0 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Fixed Telecommunications Fixed Telecommunications Monthly Charge for a Residential Service (in Real Local Currency) Monthly Charge for a Residential Service (in Real Local Currency) 1000 1000 800 800 mon_chg_c_100 mon_chg_c_100 600 600 400 400 200 200 0 0 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Fixed Telecommunications Fixed Telecommunications Price of the Instalation of a Residential Line (in Dollars) Price of the Instalation of a Residential Line (in Real Local Currency) 1000 600 800 conex_cost_c_100 conex_cost_100 400 600 400 200 200 0 0 -5 0 5 -5 0 5 t t Argen Bol Brazil Chile Guyana Argen Bol Brazil Chile Guyana Salv Guate Jama Mex Nica Salv Guate Jama Mex Nica Pana Peru Trin Venez ALL Pana Peru Trin Venez ALL Source: Authors' calculations. Note: The x axis is time; t=0 is the last year with at least six months of public ownership. The y axis is normalized at 100 when time=0. 163. The analysis of average installation charges for a residential line produced somewhat mixed results, although the preponderance of evidence suggests that prices declined during the transition and post-transition periods. Figure 3.14 shows a big drop in installation charges during the transition and more modest falls after that. The means and medians analysis found a large statistically significant drop during the transition period, but the drop during the post-transition period was not significant. The econometric analysis found the reverse: the drop during the 46 transition was not significant, whereas the drop during the post-transition period was significant and roughly 25 percent in both dollars and real local currency. There were no significant changes in the growth rate. 164. Prices for a three-minute local call increased during the transition, but did not significantly change after that. Residential monthly service charges increased during both periods, with the greatest increase coming during the transition. Residential line installation charges seem to have decreased during both periods. These results hold for prices in both dollars and real local currency. ... I N SUMMARY ... 165. The change in ownership generally increased output and coverage, even after controlling for firm-specific time trends in the fixed-telecommunications sector. Employment fell and labor productivity increased during the transition and post-transition periods, while efficiency (percentage of incomplete calls) improved during the post-transition period. Prices showed mixed results: the price of a local call increased during the transition; residential monthly charges increased in both periods; and installation charges decreased in both periods. Quality--as measured by network digitization--generally improved. 3.4. T HE I MPACT OF C ONTRACTS D ESIGN 166. This Section deepens the previous analysis by introducing a number of privatization contract and process variables. The variables come from a separate World Bank data set containing the characteristics of nearly 1,000 infrastructure transactions in LAC between 1989 and 2002 (see Annex 2). This data set was merged with the data sets containing utilities performance information. Merging the two databases makes it possible to identify whether privatization characteristics like the sale method (for example, auction), investor nationality, and award criterion affect the performance variables discussed in previous sections. 167. The main aim of this section is not to advocate a certain type of contract design. Rather it is to emphasize that privatization is not simply a yes-no decision. Indeed, different privatization design variables can influence differently each performance outcome. The results in this section show that, depending on the priorities of a country when considering the change in ownership, certain privatization contract characteristics might be more important than others. 168. There are many reasons to suspect that characteristics of the privatization process and regulatory environment would affect firm performance both during and after the transition to private ownership. First, large unexplained differences in performance across firms were found earlier. For instance, large drops in employment occurred on average during both the transition and post-transition periods in the electricity distribution sector. However, some firms experienced much larger drops than others. These large performance differences suggest that differences in privatization procedures or the regulatory environment may have played a significant role. 169. The three sectors were pooled to maximize the amount of variation in the data set. 45 For more details on the data and methodology, see Annex 1. 46 Similar to the preceding sections, 45 The models were run for each sector separately (these tables are available upon request). Results were qualitatively similar to the ones presented in this section. See Andres et al. (2008c) for more details. 47 results from two time periods are analyzed: i) changes between the period before the transition to private ownership and the transition period; and ii) changes between the transition and post- transition periods. Overall changes are not reported in this section. Rather, the changes shown are relative to the base case for each variable. For instance, when it is reported that the number of connections decreased by 5.8 to 6.8 percent when an auction process was used, this change is relative to cases in which auctions were not used--no auction being the base case. Table 3.2. Base Case for Regulatory and Contract Variables Category Base Case Variables Sale method No auction Auction Investor nationality Local only Foreign only; foreign and local Award criteria Other criteria Highest price; best investment plan Tariff regulation Other regulation Rate of return; price cap Source: Andres et al (2008). 170. The following summary of the econometric results further illustrates these points: Sale method: Auctions tended to decrease employee numbers and increase quality by fairly large amounts. Auctions also resulted in price increases during the transition and price decreases after the transition, as well as distributional loss reductions after the transition. Investor nationality: The presence of only foreign investors caused output to fall somewhat during both periods, coverage to fall during the transition, the number of employees to fall substantially during the transition, and distributional losses to fall after the transition. Average dollar prices seem to have increased during both periods, while prices in real local currency first decreased then increased. When both foreign and local investors were involved, employment fell during both periods, distributional losses fell after the transition, prices in dollars first fell then rose, and quality improved. Award criterion: When concessions were awarded according to the best investment plan, employment fell substantially during both periods, prices in dollars appear to have risen after the transition, and prices in real local currency appear to have fallen during the transition. When concessions were awarded based on the highest price, the number of connections fell slightly during the transition, coverage first fell slightly then increased, the number of employees fell substantially, and prices in real local currency fell moderately during both periods. Tariff regulation: Price-cap tariff regulation caused output and quality to increase slightly and number of employees and labor productivity to decrease slightly, all during the transition. Distributional losses increased after the transition, and prices in real local currency increased during both periods. Rate-of-return regulation caused the number of connections to increase moderately, coverage to increase slightly during the transition, 46 The econometric analysis included several different regression specifications using different combinations of independent variables. In other words, for each performance variable, the impact of each contract variable was tested while controlling for different combinations of other contract variables. Controlling for other contract variables addresses collinearity issues, while tending to reduce the number of statistically significant results. Multiple regression specifications also can produce a range of results. For this reason, the following sections mention either a range of impacts or mixed results. Andres et al. (2008c) reports the minimum and maximum percentage changes in each performance variable disaggregated by the contract variables. 48 and employment to drop dramatically during both periods. Distributional losses fell after the transition, and prices in dollars first increased then decreased. 171. The findings of the section can be summarized in three main points. First, contract characteristics matter: the way privatizations are undertaken can generate significant performanc e differences. Second, each contract characteristic affects each performance variable differently. In other words, a certain contract characteristic could have a positive influence on one performance variable while having a negative or insignificant impact on another. Third, some contract variables have bigger impacts than others. 3.5. M AIN C ONCLUSIONS 172. The main finding of this chapter on the impact of private participation in the electricity distribution, telecommunication, and water and sanitation sectors in LAC is that overall significant improvements in sector performance were associated with PSP. The highlights are the consistent improvements in efficiency and quality, and reductions in the workforce. There do not appear to be significant impacts on output and coverage. Prices tended to increase somewhat, although the picture is highly variable across sectors. 173. The differences between publicly and privately operated distribution utilities showed up primarily with regard to labor productivity, distribution losses, quality of service, and tariffs. In contrast, other indicators such as coverage and operation expenditures exhibited similar trends or did not present significant changes between the groups. Nevertheless, the distribution of performance overlaps with the top decile of performers in the public utility group outperforming the average private utility, and the bottom decile of performers in the private utility group outperformed by the average public utility. 174. In the case of labor productivity and distribution losses, both groups of utilities displayed similar starting values. Following privatization, the performance of the privatized group improved substantially. For example, labor productivity ended up being twice as high as that of the public utilities. In the case of distribution losses, private utilities improved their performance by 12 percent, while public utilities saw their performance deteriorate by 5 percent. With regard to continuity of service, both groups started at around 24 interruptions per year. The private utilities reduced this to around 12 compared with a reduction to around 19 for the public utilities. Similarly, public utilities saw the average duration of their outages increase by almost 50 percent, compared with a reduction of almost 30 percent from the private utilities, from a similar starting value. 175. PSP should not be considered a homogenous event, but rather one whose results differ enormously with respect to the way it is designed. Key dimensions of the design include sale method, award criteria, nationality of the firm, and details of the subsequent regulatory framework, including degree of autonomy of any regulatory body and principles used to determine tariff. According to economic theory, each of these aspects can sign ificantly affect the incentives faced by the private party and, hence, could be expected to influence the different aspects of enterprise behavior reviewed above. By pooling all the cases available across sectors, and adding a new set of variables to capture the transactional and regulatory environment, it was possible to measure the impact of each of these factors. 49 4. REGULATORY INSTITUTIONAL DESIGN AND SECTOR PERFORMANCE 176. This chapter explores the governance of independent regulatory agencies (IRAs) in the water and electricity distribution sectors of LAC and the link between the governance of IRAs and the performance of both sectors. The first part of the chapter analyzes the institutional design of regulatory agencies. The comparison of the different governance modes of IRAs is carried out through different measures of autonomy, transparency, accountability, and tools. Measures of agencies' governance are the result of both formal and informal practices of IRAs. The second part establishes the methodology and results of the correlation between institutional design and sector performance. 177. This analysis focuses on the institutional design of IRAs. Thus, the reader should be aware that when we measure, for instance, autonomy, we are not automatically attributing it a degree of independence. Rather, we attempt to determine the inputs that we consider to be important in order to achieve good regulatory outcomes. By institutional design, we are referring to the inputs or characteristics of IRAs that would allow them to be more autonomous and accountable. Nevertheless, its existence does not guarantee either effective autonomy or accountability. 178. The second phase of this work will involve the use of qualitative comparative politics techniques to address issues of causality, sequencing, and complex interaction effects that better explain IRAs in policy-making. The rationale behind this methodological approach is to capture different aspects of the governance of IRAs that could be assessed against sector performance. 179. A review of the literature on the governance of IRAs in LAC does not align in scope or provide a sound basis for comparison to our approach. The majority of the research has been conducted by international donors (especially the World Bank) with the policy goal of establishing comparisons among countries of the region in terms of formal attributes of IRAs. Analysis of causality, addressing explanatory factors of agencies' performance is limited if not nonexistent. 180. LAC shows the highest diffusion of IRAs among developing countries (Sosay et al, 2005). Created within the context of wide privatization programs, IRAs were the chosen institutional arrangement to insulate decision-making in various economic sectors, such as the infrastructure sector, from political intervention (Thatcher, 2005). This was particularly the case for the electricity sector where, after the unbundling of the industry, regulatory agencies were assigned the task of enforcing concession contracts and protecting consumers. Beginning with Argentina's National Electricity Regulatory Agency in 1993 and ending in 2001 with Barbados's Fair Trading Commission, today 70 percent of the countries in the region have a separate entity, with different degrees of independence, to regulate electricity markets (Andres et al 2007). 181. While there is a growing consensus that institutions matter for growth and development (Aron, 2000; Rodrik, 2004), this chapter emphasizes the positive externalities associated with the presence and good governance of an independent regulatory agency. 50 4.1. B ENCHMARKING OF R EGULATORY INSTITUTIONAL D ESIGN 182. The different studies that have assessed regulatory agencies in the infrastructure sector have considered the US model of the independent commissions as their benchmark of comparison and analysis. An institutional design model that emphasizes agencies that make decisions independently from the Executive branch, are subject to the accountability of the Parliament, and have budgeting autonomy, has emerged as the paradigm of an infrastructure regulator. The literature has dealt in both ways with the design of regulatory agencies: by focusing only on independence and by considering other variables of agencies' functioning, namely accountability and transparency. The first attempts to evaluate infrastructure regulatory agencies made use of frameworks to assess the independence of Central Banks (Stern and Cubbin, 2005; Oliverira, Machado, Novaes, and Ferreira, 2005). This explains the original emphasis on agencies' independence and the reduced significance given to other aspects of their functioning such as accountability and transparency. The evolution of the subject and the initial stages of agencies' functioning changed the original approach and introduced more comprehensive strategies to assessment. A different approach (OECD, 1997) involves the consideration of mechanisms to achieve high quality regulation such as cost-benefit analysis of regulations and administrative simplification. 183. Johannsen (2003) measures the formal independence of energy regulators in eight European countries. Gilardi (2002) develops an independence index, covering regulators from five sectors in seven European countries. He also proposes three ways of evaluating independent regulators. Gutierrez (2003) develops a regulatory framework to assess the evolution of regulatory governance in the telecommunications sector during the period 1980­2001 in 25 LAC countries. Stern and Holder (1999) develop a framework to assess the governance of economic regulators in several sectors in six developing Asian economies. In addition they attempted to measure informal regulation. Three comprehensive approaches to assessing the governance of regulatory agencies have been those developed by Correa et al. (2006), Brown et al. (2006), and Andres et al. (2008). Correa et al. provide a detailed analysis of Brazilian regulatory agencies. Brown et al. (2006) develop a framework to assess the effectiveness of a regulatory system. Finally, Andres et al. (2008) develop a framework for LAC that will be discussed more in detail in this Chapter. 184. This chapter defines regulatory governance as the agency's institutional design and structure that allows it to carry its functions as an independent regulator. Based on selected literature on the subject, this chapter defines and assesses regulatory agencies governance according to four main characteristics: i) autonomy from political authorities and autonomy of their management and regulatory competencies; ii) transparency before institutional and non- institutional stakeholders; iii) accountability to the three branches of government (Executive, Legislative, and Judiciary); and iv) tools and capacities for the conduct of the regulatory policy and the improvement of its institutional development. 185. We measure the governance of IRAs through a main aggregated index and other indexes covering different aspects of governance. 47 Indexes were built with data from a survey completed 47 In other words, our measurement of agencies' governance is not an indicator of the effectiveness of the use of their regulatory instruments (such as the methodology to calculate tariff readjustment) or the quality of stakeholders' involvement in public consultations. It is aimed at capturing the institutional conditions necessary to achieve good regulation regardless of their scope and impact on the sector's performance (Correa et al, 2006). 51 by nineteen countries 48 for the electricity distribution and water and sanitation sectors of the region. Responses from the survey covered 43 electricity and 28 water regulatory agencies, whose coverage in terms of consumers exceeds 90 percent of the region. With the exception of Chile and Colombia, the rest of the LAC countries, have introduced regulatory agencies where the agency has both regulatory and oversight responsibilities with different degrees of independence from the government. 49 186. We conceive regulatory agencies as both public bodies that are part of the public administration--and as such in charge of the delivery of public services--and as instruments to implement regulatory policies. This approach to assessing regulatory agencies' governance led us to consider not only existing research on infrastructure agencies' designs (documented in the literature review), but also notions and tools of public sector governance applied to decentralized structures of government. 187. Figure 4.1 represents the selected framework to assess the governance of independent regulatory agencies in the infrastructure sectors of LAC. The regulatory governance of independent agencies is defined and assessed according to four variables of their design and functioning. Each of the variables, with the exception of accountability, is composed of several elements. We consider only an institutional perspective of accountability as defined by the relationships of the agency with the Executive, Legislative, and Judiciary branches of government. Autonomy is divided into political, managerial, and regulatory autonomy; transparency, into social and institutional transparency; and tools are divided into regulatory and institutional tools. 188. Variables for agencies' governance reflect not only formal aspects (procedures and tools established in the agency's statute or laws) but also the practices that derive from their implementation (informal regulation). Indicators for the informal elements of autonomy, accountability, and transparency represent the operationalization of some aspects of the se variables. The variable tools is excluded from this analysis as the mere existence of these instruments implies their actual implementation. 189. The first variable of agencies' regulatory governance is autonomy. We define autonomy as the procedures, mechanisms, and instruments aimed at guaranteeing the independence of the agency from political authorities (political autonomy), the autonomous management of its resources (managerial autonomy), and the regulation of the sector (regulatory autonomy). Political autonomy represents the level of independence of the agency from government authorities and is measured by indicators that reflect the autonomy of the agency's decision- making. Managerial autonomy involves the freedom of the agency to determine the administration of its resources and is measured by indicators that reflect the powers of the agency to determine its organizational structure and the use of its budget. Regulatory autonomy is defined by the extension of the agency's regulatory powers in the electricity sector and is represented by indicators that capture agencies' responsibilities in electricity regulation. 190. The second aspect of agency's governance is accountability, which we define as the procedures, mechanisms, and instruments aimed at guaranteeing an adequate level of control of 48 Trinidad and Tobago, Peru, Mexico, El Salvador, Colombia, Brazil, Bolivia, Nicaragua, Costa Rica, Panama, Guatemala, Ecuador, the Dominican Republic, Argentina, Jamaica, Honduras, Chile, and Uruguay. 49 In the case of Chile and Colombia, they have split regulatory responsibilities in two agencies, one in charge of the main regulatory functions (National Energy Commission) and one in charge of enforcement of the regulatory framework, particularly in terms of the imposition of sanctions and the observance of service quality standards (Superintendencia). 52 the agency's budget and performance by political authorities, namely the Parliament. Despite the successful use of mechanisms to assess the performance of agencies by governments, we prioritize the accountability of the agency before the Parliament. This decision was made because of the following two reasons: First, the fact that the institutional design model we follow is that of a US independent commission, where agencies are subject to parliamentary oversight. Second, the history of political interference of LAC line ministries in utilities underscores the importance of including other political stakeholders, such as the Parliament, in the regulatory process. We consider an institutional perspective of accountability only as defined by the relationships of the agency with the three branches of government (Executive, Legislative, and the Judiciary) and do not further disaggregate the variable. Figure 4.1 The Assessment Framework AUTONOMY: TRANSPARENCY: Political Social Regulatory Institutional Managerial Formal/informal Formal/informal SELECTED VARIABLES AND DIMENSIONS OF ELECTRICITY AGENCIES GOVERNANCE Formal/informal No distinction is made TOOLS: ACCOUNTABILITY Regulatory (No distinction is made) Institutional 191. The third variable is transparency. We define transparency as the procedures, mechanisms, and instruments aimed at guaranteeing the disclosure and publication of relevant regulatory and institutional information, the participation of stakeholders in the agency's regulatory decisions and decision-making, and the application of rules aimed at governing the integrity and behavior of agency officials. We cover two dimensions of transparency: social transparency and institutional transparency. Social transparency is composed of indicators related to the involvement of non-institutional actors in the agency's policy-making, including their access to the agency's information. Institutional transparency is composed of indicators related to the transparent management of the agency that are not directly linked to stakeholder involvement, and includes issues such as the publication of the agency's annual report, the use of norms of ethics, and the existence of public exams for hiring employees. 192. The fourth variable is tools, which we define as the instruments and mechanisms that contribute to the strengthening of different aspects of an agency's functioning and the quality of 53 its regulations. We include not only regulatory tools (e.g. mechanisms for tariff revision, regulatory accountability, instruments for monitoring technical standards), but also those instruments aimed at improving the institutional quality of the agency, or institutional tools (e.g. audits of agencies' accounts, electronic files for consumer complaints, performance-based payments for employees, regulatory quality standards). This is the only variable whose analysis does not consider its formal and informal aspects; the sole existence of agencies' tools implies their actual implementation. 4.2. R ESULTS OF B ENCHMARKING OF A GENCIES AT R EGIONAL L EVEL 193. The LAC region presents a wide spectrum of institutional design in its regulatory agencies. A regional analysis of regulatory governance indicates the prevalence of autonomy over the rest of the variables, with tools as the index's component with the lowest score. While there are degrees of variation, the majority of independent regulators in the electricity sector have a board of directors appointed by the President with the authorization of the Congress, a separate status from the line ministry, and separate budgeting (although there are different levels of autonomy in the management of funds). The lowest levels of autonomy can be found in agencies in charge of both regulation and sector planning, where the government, through the line minister and other ministers, is part of the agency's decision-making process. 194. The top ranking of the autonomy variable and the lower scores given to transparency and institutional and regulatory tools might be explained by the lack of progress in improving the institutional quality of the agencies (represented in the Infrastructure Regulatory Governance Index by several components of the transparency and tools variables). With some exceptions, the process that started with the initial creation of regulatory agencies in the LAC region has not been furthered nor improved. For instance, few agencies publicize their job posts or have developed public exams for hiring employees. On the tools side, the utilization of regulatory quality standards (such as the use of cost-benefit analysis to assess the impact of regulations) or performance-based payments for employees are practices that have been rarely implemented. 195. Figure 4.1 presents the distribution of the aggregated index for each of the sectors. Agencies in the electricity sector show an overall better performance than those in the water sector. This is evident not only in the general indexes but also in the rest of the specif ic measures. Figure 4.1: Regulatory Governance Electricity Regulatory Governance Water Regulatory Governance 1 1 .8 .8 Index (0-1) .6 Index (0-1) .6 .4 .2 .4 0 CR - ERSAPS BZ - ADASA HN - ERSAPS BZ - ARSAM AR - ENRESS BZ - ARSAE T1 T2 BZ - ARSAL PA - ASEP PE- SUNASS BZ - AGENERSA BZ - AGERSA BZ - AMAE AR-ERAS CO - CRA BZ - ARCE BZ - AGERGS AR - ERSACT BZ - AGER PY - ERSSAN BA - FTC BZ - ATR BZ - AGESC TT - RIC AR - ERSPyOC BZ - AGR AR - ERSAC .2 0 T1 BOL T2 COL MEX JAM BRA ESV GUA T&T ARG PER BAR URU PAN RD NIC CR ECU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. 54 Box 4.1: Multi-agency regulatory schemes Agencies included in the index are those that exhibit a similar design to that of a formal independent regulator. Although several agencies embody the institutional patterns of a formal independent regulatory agency (IRA), the Region's most salient characteristic is a Board composed of independent members. In other words, the members appointed to the Board should not be ministers, state secretaries, or o ther officials whose autonomy could be compromised by holding a policy-formulation position. Chile's National Energy Commission (NEC) does not follow this criterion. More precisely, NEC's Board is composed of the Ministers of Mining, Finance, Defense, Planning, and the Secretary General of the Presidency. This circumstance makes Chile a standalone case incomparable to the rest of the IRAs, regardless of its regulatory effectiveness in relation to other regulatory agencies in the region. Moreover, tariffs in Chile are not determined by the NEC but by the Minister of Finance, the only authority t hat approves electricity tariffs in the country. In Chile, regulatory competencies are complemented by the Superintendencia de Electricidad y Combustibles, which is responsible for the enforcement of regulations as well as quality and technical standards. Despite the fact that the Dominican Republic and Colombia also show a similar institutional design (National Energy Commissions, integrated by public officials, and Superintendencias, composed of independent members), these cases were included in the analysis for several reasons. In the case of Colombia, the Board of the National Energy and Gas Regulatory Commission is also composed by five independent experts that balance the influence of public sector officials such as the Ministers of Mining and Energy, Finance, and the National Director for Planning. Moreover, the country's position in the score is the result of the combination of the complementary roles of the Regulatory Commission (in charge of the main economic regulation responsibilities) and the Superintendencia de Servicios Publicos Domiciliarios (responsible for enforcing standards and regulations). In the case of the Dominican Republic, only the Superintendencia de Electricidad was included because it is the only electricity regulator; the National Commission of Energy with a policy-formulation responsibility. 4.3. R ESULTS OF R EGULATORY G OVERNANCE B ENCHMARKING AT A GENCY L EVEL 196. The challenges of benchmarking the regulatory agencies in LAC called upon three tiers of analysis. Agencies were discretionally grouped based on their performance in several indicators. Tier 1 encompasses agencies that have desirable conditions to develop good regulatory governance. The responses of agencies in this tier are similar to the highest value for each of the questions. Tier 2, countries that are between T1 and T2 in the different graphs, encompasses agencies that only meet the minimum conditions that we consider necessary to implement the independent regulator model. Agencies in Tier 2 have fewer responsibilities than those in T1 and lowers levels of autonomy from the line minister. They also have fewer sophisticated mechanisms for publishing their decisions and policies. Tier 3, countries below T2 in the different graphs, includes agencies that would not meet what we define as the minimum conditions to implement our benchmark model of regulatory governance. 197. Consistent with the regional analysis, autonomy is the variable with the highest score for Tier 2 and Tier 3 countries, with a slight difference towards accountability in the case of countries 55 above T1. Bolivia's Superintendencia de Electricidad 50, Nicaragua's Comision Nacional de Energia, and the Dominican Republic's Superintendencia de Electricidad have the highest score. Figure 4.2: Regulatory Governance Aspects - Electricity Autonomy Accountability 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BOL T1 T2 COL JAM ESV MEX BRA GUA ARG T&T NIC RD PER CR BAR PAN URU HON ECU T1 BOL T2 COL MEX JAM GUA BRA ESV T&T ARG BAR PER URU PAN RD NIC CR ECU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Transparency Tools/Capacity 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 T1 BOL T2 COL JAM MEX GUA BRA ESV ARG T&T PER URU NIC CR BAR PAN RD ECU HON T1 BOL T2 COL JAM ESV MEX BRA GUA T&T ARG PER BAR URU PAN CR RD NIC ECU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. 198. Accountability is the second highest variable. Trinidad and Tobago's Regulated Industries Commission is the agency with the highest score. The main difference between the best and poorest performers in accountability is explained by greater obligations to the Executive of the latter. Countries at the top of the aggregated measure of regulatory governance, with the exception of Bolivia and Peru, have a more balanced distribution of obligations between the Executive and the Parliament and are not fully accountable to the Executive. In contrast, countries at the bottom of this distribution are heavily dependent on the Executive, to which they are, in most of the cases, fully accountable. 199. Transparency is the third variable in order of prevalence. Trinidad and Tobago's Regulated Industries Commission is the agency with the highest score and Honduras' Comisión Nacional de Energía is the agency with the lowest performance. Differences are not significant between best and worst performers in transparency, with the exception of Ecuador among the latter. Both best and poorest performers have collective decision-making structures, mechanisms to allow the participation of their stakeholders in their rule-making processes, adequate 50 The government of Evo Morales has recently announced the elimination of Superintendencias as sector regulators in Bolivia and the creation of Autoridades de Fiscalizacion y Control Social. Article 138 of supreme decree 29894, published February 7, 2009, states that with the exception of the hydrocarbons regulator, all regulators that form part of the sector regulatory system or the renewable natural resources regulatory system will disappear within 60 days from the date of the decree's publication, and their functions will be taken by the corresponding ministries or a new regulatory authority. Although their institutional design is still not clear, their levels of autonomy as IRAs will be reduced as the new law makes them directly accountable to the line Minister. 56 mechanisms to report their activities to the required institutions and to publish their annual reports. The only aspect in which poorest performing countries show lower scores is in public consultations. 200. When considering the results for the region, tools is the variable where countries, regardless of ranking, have their lowest scores. This variable is not only a measure of tools related to the application of the agencies' regulatory policies such as benchmarking or the methodology for tariff revision, but also of instruments aimed at improving institutional and managerial quality (e.g., the publication of the agency's annual report or the use of performance-based payments). Guatemala's Comisión Nacional de Energía Eléctrica is the agency with the highest ranking for tools and both Honduras' Comisión Nacional de Energía and Mexico's Comisión Nacional Reguladora de Energía have the lowest scores for this variable. The main factors that explain the differences between best and worst performers in terms of the tools variable are: i) the use of benchmarking; ii) the extent and number of regulatory instruments; iii) the publication of the agency's annual report; iv) the registration of users' claims; v) the utilization of regulatory quality standards; and vi) the existence of a structure of posts and salaries. Figure 4.3: Regulatory Governance Aspects - Water Water - Autonomy Water - Accountability 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BZ - ADASA CR - ERSAPS HN - ERSAPS BZ - ARSAM BZ - ARSAE CR - ERSAPS BZ - ADASA BZ - ARSAM HN - ERSAPS AR - ENRESS BZ - ARSAE AR - ENRESS T1 T2 T1 T2 BZ - ARSAL BZ - ARSAL PA - ASEP PA - ASEP BZ - AGENERSA PE- SUNASS BZ - AGERSA BZ - AMAE AR-ERAS BZ - AGENERSA BZ - AGERSA PE- SUNASS BZ - AMAE AR-ERAS BZ - ARCE CO - CRA BZ - AGERGS CO - CRA BZ - ARCE BZ - AGERGS AR - ERSACT AR - ERSACT BZ - AGER BZ - AGESC PY - ERSSAN BZ - AGER PY - ERSSAN BA - FTC BZ - ATR BA - FTC BZ - ATR BZ - AGESC TT - RIC AR - ERSPyOC TT - RIC AR - ERSPyOC BZ - AGR BZ - AGR AR - ERSAC AR - ERSAC Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Water - Transparency Water - Tools/Capacity 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BZ - ADASA CR - ERSAPS HN - ERSAPS BZ - ARSAM AR - ENRESS BZ - ARSAE HN - ERSAPS CR - ERSAPS BZ - ARSAM AR - ENRESS BZ - ADASA BZ - ARSAE T1 T2 T1 T2 BZ - ARSAL BZ - ARSAL PA - ASEP PA - ASEP PE- SUNASS BZ - AGERSA BZ - AMAE AR-ERAS AR-ERAS BZ - AGENERSA BZ - AGENERSA PE- SUNASS BZ - AGERSA BZ - AMAE CO - CRA BZ - ARCE BZ - AGERGS BZ - AGERGS CO - CRA BZ - ARCE AR - ERSACT AR - ERSACT BZ - AGER PY - ERSSAN BZ - AGESC BZ - AGER BZ - AGESC PY - ERSSAN BZ - ATR BA - FTC BA - FTC BZ - ATR TT - RIC AR - ERSPyOC TT - RIC AR - ERSPyOC BZ - AGR BZ - AGR AR - ERSAC AR - ERSAC Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. 201. Differences between IRAs in water and electricity are more notorious in informal transparency, formal accountability, tools, regulatory autonomy, social transparency, regulatory tools, and institutional tools. Although it would be expected to have better indicators in electricity than in water, it could also be expected to have better results in aspects where the water sector is considered to be stronger, such as social public involvement in rule-making. Nevertheless, our measure of social transparency shows one of the largest differences between governance in electricity (where countries above T1 and on average are higher than in the water sector) and in the water sector. Similar results are seen in informal transparency. 57 4.4. R ESULTS ON D IFFERENT D IMENSIONS 202. This section of the paper disaggregates the variables: Autonomy is broken down into political, managerial, and regulatory autonomy; transparency into social and institutional transparency; and tools into regulatory and institutional tools. Accountability considers only an institutional perspective regarding the relationships between the agency and the other branches of government (Executive, Legislative, and Judicial) and no further division is made of its different indicators. 203. Political autonomy: This dimension of an agency's autonomy reflects the level of independence of the agency from political authorities. Its focus is on the independence of the agency's decision-making from authorities in charge of policy formulation, namely the line minister. It includes issues such as the mechanism to select agencies' directors, the renewability of directors' mandates, the number of directors that have not completed their terms, the reasons directors leave their positions, the interference of the minister in the agency's decisions, and the composition of the agency's budget. 204. This variable shows the largest number of countries among Tier 3 agencies (See Annex 5). It could be understood from this finding that the independence of agencies from political authorities is the most significant deficiency of agencies in terms of their autonomy. Only Brazil is among Tier1 countries. Tier 3 countries represent a wide variety of agencies. The scores of best performers in Political Autonomy significantly differ from countries at the bottom of that index. The agencies of Brazil, the Dominican Republic, and Bolivia are designed with a separate status from the line ministry, the separation of roles between the agency and the government authorities, and a budget composed exclusively of a regulation tax charged to electricity distribution companies. Directors leave mostly due to retirement, voluntary leave, or the completion of their appointments, and the line minister, according to the opinions of the agency, has a low level of influence on the agency's affairs. In contrast, those agencies in the bottom of this ranking present separate entities but with no autonomy from the line minister. Moreover, the sector ministry is part of both agencies, and chairs their Boards; their budgets are composed exclusively of government funds without any type of income from companies (regulation tax). 205. Managerial Autonomy: Managerial autonomy involves the freedom of the agency to determine the use of its budget and the organization of its resources. It includes aspects such as the ability of the agency to determine its organizational structure, the freedom to make its own decisions on personnel, the financial autonomy to determine its own expenses, and the type of legal regime that applies to its employees (private law, civil service law, or both). It also includes other aspects related to tools that contribute to improving its management, such as the existence of its own structure of posts and salaries and of performance-based payments for its employees. Jamaica, Guatemala, Brazil, Argentina, Trinidad and Tobago, Peru, and Barbados are among those countries with desirable conditions to manage its resources. These agencies show the existence of adequate mechanisms and procedures to guarantee an autonomous administration of the agency by its authorities. On the contrary, Colombia and Honduras have less managerial freedom and space to decide its organizational structure and the use of its resources is limited. Results in this section are not an indication of the effectiveness of the agency's management, but of powers aimed at allowing the agency an autonomous administration. Countries at the top of the distribution have full powers in all the aspects mentioned in the first paragraph. Brazil is among the leading countries in managerial autonomy. 58 206. Regulatory Autonomy: This dimension of agency's autonomy is represented by the extension of the agency's regulatory powers. It includes characteristics such as the institution responsible for the regulation of the sector (the agency, Parliament, the Executive or some combination among them); the type of the agency's powers (consultative, oversight, pricing, and rule-making); the agency's responsibilities regarding particular issues (tariffs, service quality, consumer complaints, companies' investment plans, wholesale market, anti-competitive behavior, technical standards); and the agency's powers to enforce its decisions. The majority of countries of the region are in Tier 1, with only four countries below T2. Countries with desirable conditions in regulatory autonomy have full responsibilities for areas such as tariffs, service quality, standards, and investments, as well as the power to implement sanctions and regulations. By contrast, countries that do not meet the minimum requirements in terms of the extension of their regulatory prerogatives have little responsibility for specific regulatory issues and no power to enforce regulations. 207. The changes experienced by regulatory agencies in political vs. regulatory autonomy explain the importance of linking political independence to the expansion of the agencies' regulatory powers. In other words, an agency can have the highest level of independence from political authorities, but no relevant power in the regulation of the sector, making independence an abstract characteristic of the agency's functioning with no real impact on regulation. The same conclusion was observed in an assessment of European electricity regulators, where it was found that even if regulatory agencies shared the same regulatory objectives, there were significant variations in the means the regulators had to pursue those objectives (Johannsen, 2003). 208. Social Transparency: The social aspects of transparency are related to the involvement of stakeholders in the agency's decision and rule-making processes and their access to the agency's information. Social transparency includes issues such as the participation of stakeholders in the agency's rule-making process, the publication by the agency of its decisions, the organization by the agency of public consultations, the existence of advisory committees in the agency's structure, the existence of a website, and the registration of users' claims. Agencies' positions in social transparency are represented in Annex 4. This standard of governance is headed by Trinidad and Tobago and followed by Colombia, the Dominican Republic, Peru, El Salvador, and Bolivia. Differences between countries at the top and bottom of Social Transparency center on three main aspects. The first aspect is the participation of the stakeholders in the agency's rule- making process. While public consultations or public hearings are aimed at allowing the involvement of stakeholders in the agency's main decisions, the rule-making process is the mechanism through which the regulatees are invited to contribute with their opinions in the elaboration of the agency's regulations. Contrary to countries at the top, countries at the bottom of the distribution either lack provisions to involve stakeholders in the rule-making process or, even though these provisions exist, stakeholders do not actually get involved in that process. The second aspect is the existence of advisory committees integrated by different stakeholders in the structures of best performing agencies. These committees are supposed to play an important role in the agency's decision-making by representing and promoting different interest groups (mainly consumers). The third and last aspect is the registration of users' claims. Best performing agencies register consumer claims through both paper-based and electronic mechanisms, allowing a faster resolution of the user's case and easier access to those files by the regulatees (both at the agency and through the website). 209. Institutional Transparency: This dimension is composed of indicators related to the transparent management of the agency that are not directly linked to the involvement of the sector's stakeholders. It includes aspects such as the nature of the agency's decision-making (collective or individual), the existence of quarantine rules for directors, the agency's reporting 59 instruments (annual report and public hearing before the Congress), the publication of the agency's institutional strategy and annual report, the publication of the agency's audit accounts and of its career posts, the existence of norms of ethics, the record of the Board's meetings, and the use of public exams to hire employees. Several factors cause agencies to be positioned at the top of the index. The first factor is related to the existence of collective decision -making via a Board of Directors. As opposed to a single decision-making structure, a Board composed of directors with varied technical backgrounds allows for more comprehensive and diverse debates on regulatory issues than a decision made by a single policy maker. The second factor is related to the publication of information such as job vacancies, an annual report, an institutional strategy, and audited accounts. Finally, a record of the Board's meetings and the existence of quarantine rules for directors that leave the agency also contribute to the high performance of countries at the top of the index. Agencies with good performance in institutional transparency tend to possess characteristics related to administrative modernization. For instance, the publication of the organization's institutional strategy, annual report, and job vacancies are indicators of agen cies concerned not only with sector-based policies related to transparency (such as the conducting of public hearings) but also with mechanisms and procedures aimed at making them more effective as administrative bodies. 210. Accountability was not disaggregated into different aspects. Its indicators represent different institutional elements (e.g. reporting obligations to the Executive and the Parliament and ability to appeal its decisions before the Executive and the Judiciary) of the agency's relationships with the Executive, the Legislative, and the Judiciary. Hence, we only considered the institutional aspect of agencies' accountability design. 211. Regulatory Tools: Benchmarking, mainly to determine tariffs, is used in 78 percent of the region, with a smaller percentage of countries having the full complement of tools listed in the survey. A significant number of countries are in Tier 1, reflecting the importance given by LAC agencies to the development of several tools to implement their regulatory decisions. Peru, Guatemala, El Salvador, and Brazil achieve full score leading those countries grouped in Tier 1. Leading countries in this dimension make use not only of benchmarking but also of tools to conduct regulatory policies such as a database for regulatory accountability, methodology for tariff revision, methodology for annual tariff readjustment, instruments for monitoring quality and technical standards, methodology for monitoring technical standards, methodology for defining interconnection tariffs, and five-year revisions of these tools. Moreover, most of the countries in the region have developed specific legislation to regulate consumers' rights. 212. Institutional Tools: The region shows a better performance in regulatory than in institutional tools. There are large disparities between countries at the top and the bottom of this measure. The former agencies have certain regulatory quality standards tools (cost-benefit analysis, regulatory impact analysis, and administrative simplification), the full use of performance-based payments for their employees, the publication of both annual reports and institutional strategies and, with the exception of Peru, a structure of posts and salaries. By contrast, worst performing countries lack regulatory quality standards, do not use incentives for their employees, and have not developed institutional strategies. In addition, the registration of consumer complaints is facilitated through paper-based mechanisms, not using electronic devices to perform that task. 60 4.5. R EGULATORY G OVERNANCE AND S ECTOR P ERFORMANCE 213. This section combines data on infrastructure agencies' governance with data collected at the company level and assesses the impact of regulatory agencies on utility performance in the electricity distribution and water and sanitation sectors in the region. This work fills a gap in the literature on the subject since previous attempts to interrelate the notions of agency governance and utility performance have focused on limited factors, affecting the scope and explanatory potential of the research. Previous research on governance has only focused on the existence of an agency, a legal framework, or particular aspects of its governance--mainly its autonomy, emphasizing formal attributes. In terms of performance, only electricity generation per capita was used as an indicator related to governance (Stern and Cubbin, 2005). Estache and Rossi have recently studied the relationship between the establishment of an agency and the efficiency of the utilities as well as the welfare of the consumers (Estache and Rossi, 2008). 214. In this report, we assess the relationship between two different pieces of literature. The first is related to the impact of PSP on sector performance. The second is the literature related to measuring the governance of regulatory agencies. There is little knowledge on the relationship between these two. 215. A few papers have focused on the relationship between regulatory characteristics and performance. Sirtaine et al. (2004) define a Regulatory Quality Index, considering three key aspects of regulatory quality: legal solidity, financial strength, and decision-making autonomy. Despite their small sample sizes, three out of the four models show that the regulatory quality variables are significant in overall terms, and that alone, they are capable of explaining 20-25 percent of the internal rate of return of private investment in infrastructure projects in LAC. More recently, Estache and Rossi (2007) explored the causal relation between the establishment of a regulatory agency and the performance of the electricity distribution sector. They analyze a unique dataset comprising firm-level information on a representative sample of 220 electric utilities from 51 developing and transition countries for the years 1985 to 2005. Their results indicate that regulatory agencies are associated with more efficient firms and with higher consumer welfare. 216. We have used unique databases. Annex 2 provides the description of these sources. More precisely, this section merges the performance data described in Chapter 2 with the regulatory governance analyzed in the previous section. Each country or state was matched with its own regulatory agency, with the exception of Colombia, for which we assigned only one score since there are two different agencies with regulatory functions. 4.6. R ESULTS 51 217. This section describes the results with different specifications. 51 All the specifications were run using a semi-logarithmic functional form of these models for each of the indicators. First, we included a dummy for the existence of a regulatory agency as well as its interactions with the ownership dummies. After this, we included a quadratic form of experience of the regulatory agency. Following this, we introduced the IRGI in the specifications as well as its interactions with ownership. Finally, we decomposed the regulatory index through a Principal Component approach and obtained three principal components that are introduced in the models. 61 E XISTENCE OF A R EGULATORY A GENCY 218. We defined a dummy with a value equal to one starting in the year when the regulatory agency was established 52. We ran two different specifications. First, we ran the ownership dummies and the one for the existence of a regulatory agency (see Table A5.1). These specifications allowed for the identification of the impact of ownership when we controlled by the existence of a regulatory agency and the effect of the existence of regulation when we controlled by ownership. In a second set of specifications, we interacted the ownership dummies with the one for existence. This allowed us to identify some complementarities between both phenomena (see Table A5.2). 219. Most of the results regarding the change in ownership from the previous chapter hold when we control by the existence of a regulatory agency; however, their magnitude is slightly reduced. For instance, the effect on labor productivity is reduced by one fourth. Similar to the quality of the service, the result during the transition becomes non-significant. On the contrary, the results for the post transition remain significant with a 10 and 17 percent reduction of the impact of the change in ownership, with respect to the results previously described when we did not account for the existence of an agency. 220. With respect to the existence of a regulatory agency when we controlled by change in ownership we found a significant and desirable impact in most of the indicators. For instance, under the presence of a regulatory agency, utilities resulted with 19.4 and 18.2 p ercent higher labor productivity. Similarly, utilities reported 18.9 percent less average duration and 17.3 percent less frequency of interruptions. With respect to operational expenditures, utilities regulated by an agency resulted between 27.4 and 32.1 percent less expenditures. Residential tariffs reported a 13.5 percent increase under the presence of a regulatory agency while industrial ones presented a 4.6 percent reduction. In addition the cost recovery ratio resulted significantly higher with 13.3 percent. E XPERIENCE OF THE R EGULATORY A GENCY 221. We defined an experience variable as the years since the establishment of the regulatory agency. We argue that agencies can learn by doing in order to obtain the desired outcomes. We assumed a quadratic form for this experience factor. As expected, these results are correlated with the ones with the existence of a regulatory agency. However, these estimations support the hypothesis of gradual improvements of utilities' performance under the presence of regulat ory agencies. 222. As stated previously, most of the results on the change in ownership from the previous section hold when we control by the experience of a regulatory agency; however, we also observed reductions in the magnitude of their effect when we introduced experience variables in the model. For instance, after controlling for the change in ownership, utilities resulted with 1.4 additional increments per year in labor productivity. Similarly, distributional losses and average consumption per connection reported a 1.8 percent reduction per year. Both quality indicators resulted with an annual improvement of 9.0 percent. Operational expenditures presented between 52 Note that there are some differences between when the agency was created (in general by law) with respect to the year when it was established. The governance data reported both dates. Despite this discrepancy, we selected the year when it was established, and ran similar specifications with the year of creation, and we obtained similar results. 62 1.6 and 5.5 percent per year. Residential tariffs reported an increase of 2.6 percent per year while industrial ones reported a 1.3 percent annual reduction. Consequently, the cost recovery ratio resulted with significant annual improvements. A GGREGATED M EASURE OF R EGULATORY G OVERNANCE 223. The models also include different measures of regulatory governance developed in the previous sections. The IRGI was defined as an index between zero and one. The average of this index was 0.483 with a standard deviation of 0.343. The purpose of these models is to test not only the existence of regulatory agencies but also the governance of these agencies. As seen before, the sole existence of a regulatory agency has a significant impact on performance. However, we propose to test if there are additional effects that protrude under the presence of an agency with good regulatory governance. 53 224. As in the previous sections, most of the results on change in ownership from the pr evious description hold when we control by the regulatory governance of a regulatory agency; however, we also observed some reduction in the magnitude of their effect when we introduce the IRGI in the model. A standard deviation in the IRGI is associated with a 8.7 and 9.1 percent additional increase in labor productivity, between a 7.5 and 8.2 reduction in duration and frequency of interruptions. Furthermore, operational expenditures resulted in more than a 10 percent reduction while a 5.7 percent increase was observed in residential tariffs. Consequently there was also an improvement in the cost recovery ratio. P RINCIPAL C OMPONENTS OF THE GOVERNANCE OF R EGULATORY A GENCIES 225. Although the previous section illustrated the impacts of total governance on performance, it is interesting to disentangle the different aspects of governance. As described before, the IRGI was defined as a combination of seven different indicators (See Section 4.1). Although each of them had a particular scope and interpretation it is likely that some of them behave similarly. Hence, we applied a Principal Component Analysis (PCA) in order to comprise the eight indicators in the relevant components, thus minimizing the loss of information. 54 Since patterns in data can be hard to determine in data of high dimension, PCA may contribute in analyzing data. Furthermore, an additional advantage of PCA is that once you have found these patterns in the data, you may compress the information by reducing the dimensions, without 55 much loss of information. 226. As we did for the IRGI, the results may be better interpreted when we compute the impact on performance given an increase of one standard deviation for each factor. 56 Factor 1 53 For this end, this section reports the results with an increase of one standard deviation in governance. Our data is cross section; hence, the underlying assumption is that once the agency was created it followed a similar institutional design and, therefore, its governance is assumed constant. 54 The PCA develops a composite index by defining a real valued function over the relevant variables objectively. The principle of this method lies in the fact that when different characteristics are observed about a set of events, the characteristic with higher variation explains a higher proportion of the variation in the dependent variable compared to a variable displaying less variation. Therefore, the issue is one of finding weights to be assigned to each of the concerned variables determined by the principle that the objective is to maximize the variation in the linear composite of these variables. In other words, this approach allows for identifying patterns in data, and expressing the data in such a way as to highlight their similarities and differences. 55 See Andres et al (2008c) for more details on factor scores and their eigenvalues. 56 Standard deviations resulted in 1.51, 1.41, and 1.28 for each of the three principal components, respectively. 63 reflects informal governance aspects in a regulatory agency, as it is correlated with informal autonomy, informal transparency, informal accountability, and tools and capacities. Factor 2 reflects formal aspects of regulatory governance and is highly correlated with formal transparency and formal accountability. Factor 3 reflects formal aspects of autonomy and the formal power of the agency to determine tariff's structure and level. This factor is highly correlated with the Tariff Regulatory and the Formal Autonomy indexes. 227. Most of the coefficients for the three principal components resulted significant and with the expected signs in most of the cases; however, it seems that each of them has a distinct effect on each of the performance indicators. For instance, a standard deviation in the formal component has a higher effect on improving labor productivity by 15.9 percent and reducing frequency of interruptions and the residential tariffs by 13.8 and 19.0 percent, respectively. A standard deviation improvement in the third component that is related to formal autonomy and the attributions of the agency in terms of setting tariffs is associated with higher labor productivity by 11.4 percent and a 17.2 percent reduction in the average duration of interruptions. Furthermore, it produced a reduction in operation expenditure between 42.8 and 49.3 percent with consequent improvements in the cost recovery ratio. Finally, the first component resulted in less influence given that only three out of eleven coefficients resulted significant. 4.7. C ONCLUSIONS 228. Regulatory agencies in the LAC region were originally created to isolate regulatory decisions from political intervention and this has been reflected in their governance design. Around 75 percent of the agencies in the region have final decision responsibilities in the determination of tariff structure and levels. 229. Nevertheless, the region has encountered difficulties in the implementation of the safeguards to guarantee the autonomous management of agencies. Factor 2 of the principal component analysis (informal aspects of agencies' governance) and our measure of political autonomy show the largest number of agencies among Tier 3 countries. In the former, which accounts for 14 percent of the variance in governance variables and reflects informal autonomy, transparency, accountability, and tools, almost 40 percent of the agencies do not meet minimum governance conditions. In the latter, almost 70 percent of them do not meet the minimum governance requirements to guarantee the insulation of the agency from political influence. Moreover, the informal accountability, which assesses the degree of agency's accountability to the Executive, shows a large number of agencies below T2. 230. Regulatory agencies of the region do not perform well when measuring their institutional, non-regulatory, mechanisms aimed at improving its transparency and overall institutional quality. For instance, the use of regulatory quality standards such as administrative simplification or the use of cost-benefit analysis in the assessment of regulations has not, with exceptions, been reflected in their governance. Moreover, 30 percent of agencies do not publish their job vacancies and almost 50 percent do not use public examinations to hire employees. 231. The implementation of the independent agency model depends on its particular context. For example, in some LAC countries, independent agencies became an appendix to the line Minister, serving only for technical aspects of regulation. Yet, this observation does not aim to overshadow important achievements that many LAC countries have reached in their efforts to 64 strengthen independent agencies. The results of this chapter suggest that the mere existence of a regulatory agency, independent of the utilities' ownership has a significant impact on performance. Furthermore, after controlling for the existence of a regulatory agency, the ownership dummies are still significant and with the expected signs. This chapter also proposed an experience measure in order to identify the gradual impact of the regulatory agency on utility performance. In addition, this chapter explores two different measures of governance; we used the IRGI, an aggregated measure of regulatory governance and then we decomposed the regulatory governance indexes into three main principal components related to informal and formal aspects of the agencies' governance, also considering the regulation of tariffs by agencies as an independent variable of their governance. The results suggest that governance matters and has significant impacts on performance when we simulated a standard deviation in each of these indexes. 232. Regulatory governance matters for sector performance. We have shown that the existence of a regulatory agency matters, that the experience of the regulatory agency matters and that its governance matters as well. The results are consistent with the literature on the impact of PSP and show the relevance of the existence of a regulatory agency and its governance, defined as the agency's institutional design and structure that allows it to carry its functions as an independent regulator. Our results indicate a significant improvement in utility performance through the involvement of a regulatory agency even in the case of SOEs. 65 5. CORPORATE GOVERNANCE FOR STATE-OWNED ENTERPRISES 233. After the wave of privatizations that characterized the end of the 1980s and the 1990s, the governance of state-owned enterprises (SOEs) in LAC has recaptured the interest of the development community. A one model fits all approach is no longer used by governments and international donors to address the management framework of SOEs. Public enterprises faced different problems related to deficiencies in service provision and financial shortcomings unique to the environment in which they operated. Addressing issues such as performance-based management, the role of incentives, the professionalization of senior management, and open policies regarding transparency of their information systems required a more pragmatic, case specific approach to reform. 234. Led by the work of the Organization for Economic Cooperation and Development (OECD) on Corporate Governance and supported by concepts and tools of the New Public Management theories, state enterprises were now viewed as corporations driven by incentives that reward efficiency and transparency. The notion of Corporate Governance as applied to public enterprises tries to reflect as close as possible the incentives that exist in a private enterprise. In the particular case of SOEs, corporate governance is used to refer to the organization of decision- making in a public corporation. 235. The OECD Guidelines of Corporate Governance in SOEs (OECD, 2005) emphasize the importance of a legal framework that clearly establishes the different roles of the State, as owner, regulator, and policy formulator. The institutional setting for SOEs should ensure a fair level- playing field vis-à-vis private enterprises in order to avoid distortions and inefficiency. They also stress the importance of an explicit legal mandate that regulates the provision of public service obligations, its sources of funding and scope. The guidelines also recommend the develo pment of an ownership policy that defines the overall objectives of state ownership, the state's role in the corporate governance of SOEs, and how it will implement its ownership policy. It also recommends clear and equitable rules for all shareholders, particularly the small investors. Finally, it emphasizes the need of a Board of Directors composed of officials with high qualifications, reasonable levels of autonomy, and effective mechanisms of accountability. 236. Two main approaches can be observed in existing literature on the subject. The first approach emphasizes improved corporate governance in SOEs as a prerequisite to PSP. This approach assumes that the resemblance to a private enterprise with higher levels of autonomy in the management of funds which is subject to corporate law, and eventually, listed in the stock markets aligns internal incentives and, consequently, improves performance, clearing the way to privatization. Critics of this view emphasize the approach's focus on one of the several ways of organizing state corporations. The second approach adopts a more comprehensive, less dogmatic, view of the governance of SOEs. First, it considers the improvement of the governance of SOEs as an end in itself and not as a strategy to privatization. Second, it presents SOEs with different strategies to improve performance, including, and not limited to, PSP. 237. From Whincop (2005)'s perspective government corporations face three main problems. The first problem is related to the alignment of the interests of the government corporations' managers with those of its ultimate owners, the citizens (agency costs of management). The constituency to whom the government corporations are ultimately accountable ­the people­ 66 stands in a dual relation to the government corporation. On the one hand, they are the government corporations' residual claimants, as shareholders in a business corporation. On the other hand, they are also frequently the principal recipients of the goods and services the government corporation provides. This dualistic relation between the government corporation and the public makes it difficult to concretize the meaning of acting in the best interests of the public. The second problem is associated with the alignment of the interests of that wielding delegated governance power over managers with those of its ultimate owners (agency costs of governance). As actors in the political process, questions arise regarding the extent to which these persons are inclined to use those government powers for political advantage. The third issue is the reduction of social costs associated with anti-competitive behavior by the government corporation (anti- competitive behavior costs) 238. Whincop explores how the governance of government corporations can be evaluated in terms of three objectives --reduction of management costs, anti-competitive behavior costs, and costs of governance)--. He evaluates that from a constituency perspective. He examines the major active players whose interests may be affected by the governance of the government corporation and their relation to the ultimate principal, the public at large. Principal players are the managers, the empowered political agents, and a group of active stakeholders including customers and employees. 239. Vagliasindi (2008; 2009) develops a detailed review of substantial research related to theoretical models of Board effectiveness and ownerships structures. Although applied to the private sector, the literature stresses the importance of independent directors. In the case of SOEs, even more than in private enterprises, the appointment of directors with technical expertise and a reasonable level of independence acquire central relevance. Vagliasindi also emphasizes the importance of external governance for the management of SOEs such as the role of the government agency in charge of ownership decisions and the relevance of regulation. 240. Schwartz (2006) brings light to the discussions about the organizational model in state water utilities. He distinguishes two main organizational approaches, the Bureaucratic Model and the New Public Management Model, and applies them to public water utilities in Mexico. He defines the Bureaucratic model as one based on the preeminence of the law and rules, composed of civil servants with stability and civil service careers in public administration, and organized under the principles of hierarchy and levels. The New Public Management framework proposes higher levels of decentralization and autonomy to government entities, the use of performance- based instruments such as performance-based payments, and accountability focused on results. The author challenges conventional wisdom about the effectiveness of New Public Management institutions to state enterprises, finding that well performing public utilities tend to display a stronger adherence to the Weberian ideal-type than poorly functioning public service providers. He concludes by asserting that rather than opposite strategies, they are better viewed as complementary, focusing on the one hand on reducing patronage and depoliticizing the management of the utility (Bureaucratic model), whilst at the same time emphasizing the levels of service that must be delivered by the utility (New Public Management model). 241. Whether we consider corporate governance as a mean to privatization or as an end in itself, both approaches lack empirical evidence about the impact of governance on performance. For instance, there is no assessment about the contributions of corporatization to access to finance or productivity, or the role of shares in not-for-profit enterprises. There is, though, some evidence about performance contracts, although section 5.4 will address that in more detail. 67 5.1. M ETHODOLOGY /F RAMEWORK OF ANALYSIS 242. This chapter focuses on the governance of SOEs in the water and electricity distribution sectors of LAC. Our definition of governance is similar to that of Chapter 4. In other words, our focus is on governance design rather than effectiveness. The data collected for this report reflects the different corporate arrangements that shape 45 state run companies in the region. We included both public companies with full state ownership and companies where, despite private investment, state ownership is at least 51 percent of the total shares (only a few are in this category). The methodology we use studies governance of SOEs through six indexes. The Corporate Governance Index (CGI), the main index, results from aggregating the other five: legal soundness, board competitiveness, professional management, performance-oriented, and transparency and disclosure. 57 Figure 5.1. Corporate Governance for SOEs Framework 7. FINANCIAL MAN AGEMEN 1. T OWNERSHIP 6. LABOR CORPORATE 2. LEG AL/ INSTITUTION GOVERNANCE AL FRAMEWORK OF A SOE 3 5. BOARD PERFORMAN CE ORIENT ATIO N 4 TRANSPARE NCY/ DISCLOSURE Source: Andres et al (2009) 243. The data was collected through a survey sent to SOEs in the region. The questions and value assignment mechanism was designed considering a corporatized public enterprise (with similar access to finance and auditing requirements as private enterprises) as the benchmark. The benchmark was adjusted to allow sector specificities such as the mechanisms to appoint the Board of Directors, economic regulation, and performance-based orientation. As a novelty approach, we also included the study of the selection, appointment, salary, and educational levels of the staff. Previous approaches only emphasized the role of the Board and its relationship with the shareholder/s. For infrastructure provision state enterprises, the role of the staff is a vital aspect of good management. Because most of these enterprises are not profit-oriented and do not focus on revenues as a parameter for good performance and because a good bureaucracy is a good filter to political intervention; we believe that a an index that reflects the professionalism (given by 57 Indexes are composed of different variables representing various aspects of the management of SOEs. Questions were valued between 0 (worst) and 1 (best). Information was collected through surveys sent to 110 different utilities of the region in both the electricity and water sectors. In the aggregated index, we introduced a sixth index related to the listing of the company in the stock exchange. 68 educational levels, hiring criteria, and rewards) of the staff might give us a good proxy for the performance of the enterprise. 244. Table 5.1 describes the different components of this framework of analysis and the weight criteria for each of the questions. Table 5.1. Corporate Governance for SOEs ­ Analytical Framework Ownership/ Management/ Transparency Performance Board/CEO Legal framework Staff Disclosure Orientation Ownership Appointment's Educational Website's contents, participation Assessment of the structure, tax process (authority, levels, training, of civil society in decision-making, performance of the regime, criteria), origin and criteria to hire annual performance report, company's and its corporatization, background of the company's auditing of company's accounting, decision-making regulatory bodies directors, employees, financial disclosure standards, authorities, criteria, COMPONENTS and functions, deliberative or mechanism to involvement of consumers and tools and restructuring, executive roles, reward civil society representatives in the mechanisms, procurement, salary levels, scope employees, company's decision-making, evaluation public listing. of responsibilities, salary levels. criteria to appoint the company's authorities, systems assessment of top authorities, criteria and to reward performance. mechanisms to hire the company's employees. employees. Focus on a Emphasis on a Our benchmark We emphasize a decision-making Model of an SOE company that has Board of Directors is a company process where civil society has a with a focus on a corporate and CEO appointed that hires its say in the company's decisions performance-based structure, subject under meritocratic employees (accountability effect) and with a management. Our to the same criteria, with a through an strong focus towards the benchmark conditions to the reasonable level of external publication of institutional and compensates the BENCHMARK private sector, independence, and competition, performance information. We also lack of incentives and the whose that rewards prioritize the involvement of provided by the possibility of performance is employees' private auditors and the profitability of a accessing private assessed regularly. performance, publication of financial private company and public and whose information through best with a framework financing. salary levels are international practices. We also where the close to private give importance to the ways the performance of sector company hire its employees (open public companies is standards. process). properly assessed. Source: Andres et al (2009). 245. As done in the previous Chapter, three tiers were created to distinguish between groups of countries. Tier 1 encompasses enterprises that have the desirable conditions to develop good corporate governance. Utilities' responses in this tier are close to the highest value for each of the questions. It reflects corporate governance design characterized by high standards. Tier 2, utilities that are between T1 and T2 in the different graphs, encompass SOEs that only meet the minimum conditions considered necessary to implement a corporate governance program. It reflects an institutional design we believe needs to be at least in place to guarantee acceptable levels of governance. Utilities in Tier 2 have weaker institutional design and less sophisticated mechanisms than those in T1. Tier 3, SOEs below T2 in the different graphs, includes enterprises that do not meet the minimum conditions to implement our benchmark model of corporate governance. 246. The following section presents the graphs for the aggregated index and for each of the dimensions described earlier. The scores are aggregated for each enterprise to the country level. 69 Although this approach simplifies the results presentation; averages can hide significant heterogeneity in the governance structures of utilities within a country. For instance, the cases of Brazil and Colombia where there are best performing utilities such as SABESP and EPM and worst performing service providers (at the bottom quartile of this distribution). 5.2. R ESULTS OF C ORPORATE G OVERNANCE B ENCHMARKING A GGREGATED C ORPORATE G OVERNANCE 247. The aggregated measure of Corporate Governance ranks the companies in the region taking into account information from all five components in the framework: legal soundness, board competitiveness, professional management, performance-orientation, and transparency and disclosure. SOEs in Brazil, Peru, and Colombia show the best performance among state companies in the region. Although, none of them are above T1 levels, defined as that of desirable corporate governance conditions. The majority of companies are located within T2 and T3 levels, representing companies with the minimum and below minimum corporate governance conditions. Corporate Governance 1 .8 Index (0-1) .6 .4 .2 0 PER PAR JAM COL MEX ARG TT ECU DR URU CR T1 T2 BZL BOL Source: LAC SOEs Governance Database, The World Bank, 2009. C OMPONENT 1: O WNERSHIP AND L EGAL F RAMEWORK 248. As previously mentioned, we privilege a legal framework in which companies are corporatized and subject to similar standards as other private companies. We also give priority to companies whose policies are established and monitored by a specialized govern ment agency. The index gives higher scores to companies regulated by independent commissions or agencies and subject to the same tax obligations as any other private enterprise. The public listing of companies have a privileged score, since we assume that a company subject to the standards of the Stock Commission has better corporate governance. 70 Legal Soundness 1 .8 Index (0-1) .6 .4 .2 0 PER PAR JAM COL MEX ARG TT ECU DR CR URU T1 T2 BZL BOL Source: LAC SOEs Governance Database, The World Bank, 2009. 249. Corporatization: The majority of the companies in this sample have been corporatized, adopting different corporate modalities. The most common is to subject SOEs to the same legal framework as a limited liability enterprise, what in LAC countries is called Sociedades Anonimas, Capital Variable, etc. SABESP (Brazil) is the only company of our sample that is publicly listed, and, hence, subject to more quality controls by authorities and investors. Corporatized enterprises are also subject to corporate law, with an institutional design closer to a private company than a non-incorporated enterprise. Around 70 percent of SOEs can bankrupt in case of insolvency, have a Board of Directors, and ownership is organized under a shares' structure. The pursuit of benefits is only required in 35 percent of the cases. Moreover, and with the exception of one company (SABESP), the rest of the companies are not listed in stock markets. 250. The landscape of companies with shares is diverse. There are cases such as Aguas de Rio Negro S.A. (Argentina) where despite companies being organized as private enterprises under shares, shares have not been implemented. Others have distributed profits but at very low levels. In some cases, shares have been used to reimburse users for the money spent in the extension of the network (Peru). There are also companies that despite not being integrated by shares, they have achieved significant profits. This is the case of enterprises such as Empresas Publicas de Medellin which transferred around two hundred thousand dollars to the Municipality of M edellin, the company's shareholder. 251. Ownership structure: Almost half of our sample of SOEs has some sort of PSP. Nevertheless, percentages, with some exceptions, are very small. SABESP and Aguas de Saltillo are clear exceptions where private investors account for 49.7 and 49.0 percent of the shares, respectively. Some alternative mechanisms for PSP include share ownership by employees, trade associations58, citizens, and users; although, they usually do not account for more than 10 percent of total shares. In Argentina, for example, employees of the enterprise represented by the unions are the largest private shareholders. 252. Authorities that exercise power over the company: Ownership rights are usually exercised by the sector or line minister. In some cases, ministers of finance and auditing bodies also possess ownership rights. In those cases where SOEs are subsidiaries of larger state enterprises, ownership rights are exercised by a holding company. 253. Policy formulation authorities: For the sample of SOEs specialized agencies for policy formulation are rare. Only 23 percent of companies in the LAC region have an agency 58 This is the case of associations such as the National Association of Coffee Producers of Colombia in the enterprises Centrales Electricas Norte de Santander S.A. E.S.P. and the Association of Manufacturers of Pichincha (Camara de Industriales de Pichincha) in the enterprise Electrica de Quito S.A. 71 specifically in charge of SOEs policies. 59 The rest of the SOEs show a wide range of policy formulation authorities. The sector ministry or some ministerial agency constitutes the most frequent policy formulation authority. 254. Regulatory role: Economic regulation is a critical aspect for sustainable management of SOEs, particularly the relationship between tariffs and the quality standards of service provision. Only a very specific division of roles between the State as policy formulator, provider, and regulator can provide a framework to enforce economic sustainability and quality of service from SOEs. In this survey, 72 percent of the SOEs claimed that the regulator has final decision power in the sector in specific aspects such as: tariffs, quality standards, and service expansion. The survey suggest that the involvement of the government is heavier when it comes to critical issues such as tariff levels and expansion of service, but that it is lighter when it is related to more technical, less controversial, aspects of the service such as technical standards, and service quality. The distribution of competencies between regulatory agencies and the line ministry shows that critical decisions are taken by the latter. 255. Tax regime: Ideally, it would be expected that SOEs are subject to the same tax obligation of private enterprises. The data suggests that more than half of the SOEs in the sample have some exemption or discount to their tax obligations. Only 43 percent of the sample declared not having any type of fiscal privilege. Exemptions and discounts usually come from differential treatment of income taxes and value added taxes. The data shows that even though state companies are formally not exempted from income tax, in practice, they do not pay income taxes because they either do not generate revenues or capitalize revenues as reserves. 256. The Legal Soundness index provides a benchmark of SOEs based on their legal framework. Priority was given to a legal structure that levels the playing field for SOEs vis-a-vis private enterprises. Results were despair. Companies well known for their sector performance, such as Agua y Drenaje de Monterrey, are ranked low in this index; while others known for their operational gaps, are ranked high in the index. Overall, companies under a limited liability framework and subject to similar rules as those of other private enterprises score high. On the contrary, those with a legal typology of government department or of a private enterprise, but subject to several public rules, score low. 257. This section identifies patterns in the legal design of SOEs and the regulatory environment in which they operate. The majority of SOEs in LAC has been corporatized and has adopted the legal typology of a private enterprise. Moreover, several of them are integrated by shares and have varying degrees of PSP. From our pool of SOEs, SABESP and Agua de Saltillo are the companies with the higher levels of PSP which have implemented a share structure that provides benefits to shareholders. Another pattern of the sample of SOEs is the existence of a wide spectrum of separate commissions or agencies as an instrument to regulate and fiscalize the functioning of SOEs. Their influence, although we do not focus on effectiveness or results, seems to be higher in issues such as the regulation of quality standards. Line ministries seem to be the most influential actor in regulation. 59 We can mention the cases of Aguas Rionegrinas S.A. whose ownership policies are determined by the Secretary of Management and Control of SOEs of the government of Rio Negro (Secretaria de Estado de Control de Gestion de Empresas Publicas) and the cases of some SOEs in Paraguay, subject to the Ove rsight Council of SOEs (Consejo Supervisor de las Empresas del Estado). In the case of Paraguay, the Oversight Council of SOEs is the entity also in charge of signing and enforcing performance contracts with the state enterprises. 72 C OMPONENT 2: B OARD AND CEO 258. This section focuses on the composition, qualifications, and performance of the Board of Directors of SOEs. It prioritizes a Board of Directors where political discretion is low, where members to the Board are selected on pre-defined criteria (particularly related to merits and experience), and whose performance is assessed based on different governance arrangements.. The more the emphasis on transparency and accountability of the decision-making authorities of a SOE, the higher the possibilities of improving performance. Board Competitiveness CEO Competitiveness 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 PER PAR JAM COL MEX ARG TT ECU URU DR CR T1 BOL T2 BZL PER PAR JAM COL MEX ARG TT CR DR ECU URU T1 T2 BZL BOL Source: LAC SOEs Governance Database, The World Bank, 2009. Source: LAC SOEs Governance Database, The World Bank, 2009. 259. The low Board Competitiveness results indicate the prevalence of political authorities in the appointment of Board of Directors, low selection rates of Directors coming from within the SOE ranks, or from private independent experts and lack of Board selection criteria. In only 36 percent of the cases, the law establishes the need to select Directors upon certain criteria. Among those which have an established procedure, sector experience and a university degree seem to be the most common requirements. Only in 2 percent of the cases, political independence is a pre- condition to Board eligibility. 260. The appointment of directors constitutes an interesting example of the differences that exist between SOEs and private sector enterprises and where certain caveats need to be taken into account. In the case of a profit-oriented private enterprise, shareholders are interested in appointing a President/CEO and Executive Directors with the skills to improve financial performance. Hence, the selection process, whether conducted through the Human Resources Department or based on the sole decision of shareholders, emphasizes the candidates' ability to increase the company's revenue. For SOEs the selection criteria should focus on reducing political discretion in the appointment of decision-making authorities and creating the incentives for good performance beyond the pursue of financial gains. Very few companies have developed specific criteria, beyond legal impediments, to select independent qualified Directors to the Board. 60 60 FONFAE, in Peru, developed a Guideline that regulates the appointment, payments, and obligations of directors to state companies fully owned by the State or where the State has some participation. This directive asserts that only directors with a university degree and with 5 years of professional experience can be appointed to the Boards. They also need to be exempted from ethical and legal impediments. They are not employees of the enterprise and are hired under a professional services contract (locacion de servicios). The regulation also establishes their obligations and responsibilities. Empresas Publicas de Medellin, in Colombia, has also developed a Corporate Governance Code in which it addresses, among other issues, the criteria to appoint directors to the Board. In addition to a university degree and related professional experience, the directive requires that five out of the nine members of the Board to be independent. EPM is one of the few state enterprises that require independence as a criterion for appointment. 73 261. Of critical importance for SOE management is performance evaluation. Although responsibility for the achievement of certain criteria can adopt different ways, the Board of Directors and the company's CEO are ultimately responsible for the conduct of the business. And it is in this aspect when the differences between a private company and a SOE become more evident. The profit mazimization is the main criteria to reward or dismiss directors in a private enterprise. All the company's policies are aligned around this objective and its organizational structure and strategies also reflect this orientation. In some state enterprises, the dispersion and opposed interests of stakeholders prevent the formulation of consistent strategies and policies. Hence, the assessment of the performance of the companies' authorities becomes a challenging exercise. 262. The survey attempts to capture the way directors are evaluated. A significant number of SOEs answered positively to the question of whether their directors were evaluated. However, answers given to the rest of the questions, which go into the details of evaluation, remained unclear. When asked about the methodology/criteria to carry these assessments, only 17 percent of the SOEs responded by identifying specific criteria. The majority expressed that although Directors are indeed assessed, there is no specific criteria for that purpose, confirming the existence of ad-hoc, more informal, mechanisms of evaluation. Moreover, when asked about the instruments used to undertake the evaluation, very few identified a particular mechanism against which performance is evaluated. Reflecting practices of private enterprises, Directors are assessed at the end of the fiscal year by the shareholders. In some cases, the head of the Executive by a decree approves, after the approval of the accounting and financing reports, the performance of Directors. Strikingly, those companies that declared having a specific criteria to set objectives, responded that they do not have a particular mechanism (especially written) to facilitate the evaluation of Directors. C OMPONENT 3: M ANAGEMENT/S TAFF 263. This index measures the composition and characteristics of the enterprise's staff by levels of education, type of training, legal status, salary and benefit levels, hiring and incentives. Employees are a central part of SOEs of the infrastructure sector. They may become a filter to political decisions as a professional and well organized bureaucracy can oppose measures that hinder their career prospects. Professional Management 1 .8 Index (0-1) .6 .4 .2 0 PER PAR JAM COL MEX ARG TT ECU DR URU CR T1 T2 BZL BOL Source: LAC SOEs Governance Database, The World Bank, 2009. 74 264. Education levels of SOE staff: The bulk of employees in SOEs are those dedicated to operational work 61. Thirty-seven percent are skilled workers and 31 percent are non-skilled workers. Twenty-four percent are non-operational, administrative, workers. A small percentage, around 15 percent of employees in SOEs, has a university degree. The average age in our sample of SOEs is 44 years old. The sample shows diverse educational backgrounds both in the members of the Board and the rest of the staff. Members of the Board show a reasonable academic background. In 70 percent of cases, all the members of the Board have a university degree and in 30 percent of the cases some members have a university degree. When asked about graduate studies, 15 percent of the companies said that all the members of the Board had postgraduate studies, 55 percent of the SOEs said that some of the members of the Board have a graduate degree, while in 30 percent of the cases none of the members of the Board had a graduate degree. 265. Educational levels are higher if we look at the CEOs and managers. In 56 percent of the cases, CEOs had postgraduate studies, in 38 percent of the cases CEOs only had undergraduate formation, and in only 6 percent of the cases CEOs did not have a university degree. When it comes to managers of the enterprise, in 78 percent of the cases all managers had a university degree and 22 percent said that some of their managers had a university degree. With respect to the graduate background of managers, 12 percent of the companies said that all their managers had a university degree and that in 58 percent of the cases some of the managers had a university degree. In 30 percent of the cases, none of the managers of the companies had graduate studies. A common assumption regarding the management of SOEs is the rigidity of labor schemes that prevent the restructuring of the labor force. According this data, in 62 percent of the cases employees are hired under private law and in the remaining 32 percent they are subject to civil service rules. The majority of the labor force is hired under a permanent basis. 62 According to the data, 84 percent of the employees were hired under a regime that gives dif ferent levels of stability. Training at SOEs takes place at the different segments of the enterprise's staff, but it's more frequent in managers and employees. Members of the Board rarely take training, and managers are the ones more benefited from capacity building. 266. Staff Selection Process: A crucial aspect related to the proficiency of the human resources of state companies is the mechanism to select employees. Political discretion and the undue influence of trade unions were frequently mentioned in the past as drivers of overstaffing and low capacity. The majority of the responses indicated the use of external competition as the primary way of selecting personnel. This is more evident when it comes to more qualified positions up to the managerial level. In the case of unskilled workers, 33 percent of the staff is selected discretionally. The rest of the mechanisms include internal competition and other combinations of internal competition with external selection. A similar situation can be seen in the case of non-operational workers where 25 percent of the workers are selected both discretionally and by the sector unions. In the case of managers, 50 percent of companies indicated that their managers were selected discretionally. These numbers are not necessarily an indication of political intervention or the undue influence of other stakeholders. It may actually reflect the need of professionals that deserves the trust of the CEO and/or the Board. 61 The survey defines operational workers in the following way: i) Operational qualified workers: employees (permanent and non-permanent) that, without a university degree, perform tasks that require a special knowledge and practice. We only include those positions that are operational, in other words those employees that work directly in the operation of the service, ii) Operational non-qualified workers: employees (permanent and non-permanent) that, without a university degree, perform tasks that do not require a special knowledge and practices. We only include those positions that are operational. 62 Those that are hired under a special regime such as those under Labor Agreement or Convenio Colectivo de Trabajo. 75 267. Performance Evaluation: In addition to open and meritocratic-based selection processes, staff of an SOE would benefit from a system of incentives that rewards good performance. The questionnaire asked about the criteria to reward performance and the ways performance is rewarded. Criteria includes years in the company, performance, and the discretional determination of rewards for employees. Options for rewards include promotion, salary increase, and bonuses. The majority of SOEs reward their staff through a combination of two main criteria: years in the company and performance. A significant number of companies use only discretionality or a combination of discretionality and performance/years in the company to rewards their employees. Very few companies pay employees a bonus after certain revenue targets are achieved. 268. Performance-based Incentives: The survey also enquired about the existence of performance-based incentive payments. Incentive payments in the public sector have been considered a way of motivating the civil service and of increasing efficiency and effectiveness. Although there is no empirical evidence on the consequences of this type of reforms in the public administration, its use in SOEs still receives both voices of doubt and support. In the sample, only 20 percent of companies have some type of performance-based payments. 269. Salary levels: Salary levels of employees are, on average, higher than the salary levels of the members of the Board. Board members that receive salary levels similar to the private sector or higher than those in the public administration constitute 30 percent of the sample. Board members with salaries similar to the public sector are in 34 percent of the companies. Among employees, 84 percent perceive salaries similar to private sector levels or in between the public and the private sectors, with 16 percent that perceive public sector salaries. Salary benefits follow the same trend. 90 percent of SOEs pay their employees benefits that are similar or higher than the private sector, or between private and public sector levels. C OMPONENT 4: T RANSPARENCY AND D ISCLOSURE 270. The transparency index measures the existence of mechanisms that allow transparent disclosure of the company's financial and non-financial information, the involvement of civil society in decision-making, and the independent auditing of SOEs' accounts. The three Tiers analysis indicates that the majority of SOEs show minimum conditions to achieve the open disclosure of their performance and accounts. In the sample, no SOEs fulfill the desirable criteria. Transparency and Disclosure 1 .8 Index (0-1) .6 .4 .2 0 PER PAR JAM COL MEX ARG TT URU CR ECU DR T1 BOL T2 BZL Source: LAC SOEs Governance Database, The World Bank, 2009. 76 271. Quality of companies' websites. The majority of the companies, with one exception, have a website. When it comes to the contents of companies' websites, four main aspects are emphasized: annual report, financial accounts, corporate structure (chart), and mechanisms to receive consumers' claims and suggestions. On the contrary, little importance is given to issues such as performance statistics (coverage, quality of service, costs, etc), vacancies, the names and backgrounds of Directors to the Board, procurement processes (stages, prices, etc), and educational content. 63 272. Involvement of consumers and the society in the formulation of the companies' policies. Civil society participation can be an important factor to reduce political discretion in the management of the company. We do not focus on civil society involvement in the Board, although it is a critical way of achieving transparency; rather we give attention to the existence of mechanisms through which some decisions are subject to the scrutiny of society. Among those companies that have certain mechanisms to involve civil society in 90 percent of the cases participation is not mandatory. In other words, the company is not obliged to request the views of users or other stakeholders on different aspects related to the delivery of services. Both mandatory and non-mandatory mechanisms include consultations on issues such as tariff increases and infrastructure works for contracts over a specific threshold. 273. Publication of Annual reports: Annual reports serve as accountability mechanisms since companies must describe their achievements. In the sample, the majority of the SOEs publish an annual report of their performance. The question does not go into the details of its components and its accuracy, but a closer look at some of them allows for comparisons. The analysis finds a wide range of reports from complete, detailed reports, to the simple enumeration of works developed during the fiscal year. 274. Auditing of financial accounts. Although traditionally subject to public sector scrutiny, a significant number of SOEs are also audited by private auditors. In our sample of SOEs, the majority of the enterprises are audited by both government audit agencies and private auditors. Only 5 percent of SOEs are audited exclusively by the government and 30 percent are audited only by private auditors. Moreover, 40 percent use international accounting standards to report financial information. The majority of SOEs also publish their audited accounts. 80 percent of the companies that publish their audited accounts use the website and other means such as newspapers and printed publications. Only 10 companies out of our total sample do not publish their audited accounts. 275. Integration of the Board. In this dimension, those companies whose Board has members from the civil society and the consumers receive a higher score. The data shows that Board members with this origin are very few; only 7 percent of the Boards have a member withdrawn from these sectors. Moreover, in a very small percentage of the cases Board members are either appointed with the intervention of the Parliament or by the private sector. 63 ElectroSureste (Peru), SABESP, and EPM have developed well designed websites with useful information for consumers, investors, and the general public. ElectroSureste offers an e-procurement system which allows participants to read biddings guidelines, deadlines, and results. It also publishes the projected time, responsible authorities, and purpose of the different claims users can pursue. Moreover, it provides consumers with a virtual office to attend their questions and concerns. 77 C OMPONENT 5: P ERFORMANCE O RIENTATION 276. This component addresses the existence of mechanisms to evaluate the performance of SOEs. We intend to identify those arrangements that allow for a performance-based management of the enterprise. In other words, a management that is oriented towards the fulfillment, and achievement, of objectives and goals. We believe this is one, among many, of the ways to increase state companies' accountability, particularly because of its orientation towards results. A performance oriented management, if properly implemented, would facilitate the identification of objectives and, consequently the efficiency of the company. This is particularly the case for SOEs, where because of the lack of private investors, incentives for performance are difficult to create. Performance Based 1 .8 Index (0-1) .6 .4 .2 0 PAR PER JAM COL MEX ARG TT URU DR ECU CR T1 T2 BZL BOL Source: LAC SOEs Governance Database, The World Bank, 2009. 277. We structure the analysis of performance orientation in SOEs through three dimensions: i) The process of setting objectives; ii) The instruments used to set objectives and its enforcement; and iii) The authority that conducts these assessments. 278. Objective setting: Answers from SOEs were not sufficiently clear about the ways performance objectives are established. The majority of responses focused on the instruments through which the evaluation takes place. A few, though, were explicit about targets and the process of identification and establishment. 64 64 State companies in the DR are under the authority of the DR Corporation of Electricity Companies, a holding responsible for the ownership of public companies in the electricity sector. ELECTROSUR, one of DR state corporations, agrees different objectives depending on the government unit. For instance, it discusses ob jectives related to coverage and quality of service with the government, efficiency and revenues issues with the holding company, and within the company issues related to work-related accidents, environment protection, etc. A different approach to the setting of objectives is the case of Colombia, where the control agency (Superintendencia de Servicios Publicos Domiciliarios) requires utilities to prepare different plans based on pre-selected criteria and indicators. The evaluation of financial and non-financial performance of SOEs takes place through an independent audit by a private firm. The assessment focuses on two aspects: corporate and social. The first evaluation is related to financial indicators and the second is related to administrative and techn ical parameters, and also to quality standards. Another set of companies coordinate policy goals and objectives through Performance Agreements. Some companies in Paraguay and Brazil sign a Performance Contract with government authorities through which they set objective and monitoring strategies. In Paraguay, the ANDE signs a performance agreement with the line minister and the ownership unit (Consejo Supervisor de Empresas del Estado). The agreement is enforced by the ownership unit through periodical reports stating the level of achievement of targets. Grupo CEEE and CAESB in Brazil also sign a performance contract with policy formulation authorities. Other state utilities established different objective that are linked to Development Plans. For instance, SOEs in Costa Rica set, together with the sector minister, development goals which are monitored in the context of the National Evaluation System. Finally, some utilities use scorecard methodologies. These are the cases of ANDE in Paraguay, and both ERSSA and CentroSur in Peru. 78 279. Instruments: The Strategic Plan or Business Plan seems to be the most common mechanism used by SOEs to set objectives, and the annual report the way through which the company informs about the fulfillment of these achievements. Some companies also use public hearings as a way for the members of the Board to explain the results of the enterprise. It is not clear from the responses, what constitutes a performance agreement and what a business strategy. Three companies specifically recognized the use of a performance contract to guide the strategic direction of the enterprise. Other mechanisms that complement business plans are the balance score card and evaluation systems linked to national/local development strategies. 280. Evaluation Authorities: The assessment of the performance of SOEs is divided among several authorities. The line ministry, the regulator, and auditing agencies seem to be the principal centers of accountability for state enterprises. In some cases, the company is self-assessed through its Board of Directors. Although less common, some companies are subject to the control of a specific agency such as the SOEs Oversight Council of Paraguay and the Solidarity Fund of Ecuador. The Parliament has little saying in the accountability of SOEs. A greater involvement of the Congress in the discussion of management issues related to SOEs' performance could constitute a balance to political discretion of the Executive. 281. Assessment of Board of Directors and Staff: Another aspect considered in our measure of Performance-Orientation is the assessment of the performance of both the members of the Board and the CEO or the Executive Director. As mentioned earlier, SOEs have weak mechanisms to evaluate Board member performance. Executive directors seem to be subject to higher levels of scrutiny than the members of the Board. This is not surprising, and is consistent with the idea of the CEO responsible for the management of the enterprise. Thirty percent of the Executive Directors are not assessed on particular criteria. Arrangements to evaluate the performance of the CEO go from less formal, ad-hoc, mechanisms to more detailed structures of assessment. In the majority of the cases the CEO's performance is approved by the Board of the enterprise. For some cases, a specific criterion is established, but in other cases there are no agreed procedures to evaluate CEO performance. The most detailed mechanisms include memorandums of understanding between the government and the executive director or the assessment of his/her performance against the performance agreement or mechanism through which the company is evaluated (such as the balance score card). 5.3. C ORPORATE G OVERNANCE AND P ERFORMANCE 282. This section explores the relations between various dimensions of Corporate Governance and the operational performance of utilities in the water and electricity distribution sectors of LAC. We correlated the dimensions described in the previous section with the level and growth rates of the main performance indicators for utilities in the water and electricity sectors. We ran the governance indicators with the pool of utilities and analyzed the relationship between the governance indicators and each sector separately. Annex 7 presents the detailed results. 283. Legal Framework: his dimension focuses on the legal framework that governs the functioning of the enterprise. We privilege a legal arrangement in which companies are corporatized and subject to similar standards as other private companies. Results suggest that higher legal soundness is associated with utilities that exhibit low distributional losses and coverage; these utilities also show high labor productivity and tariffs. When we analyzed the correlation between these governance indicators and the growth rates of the set of performance 79 indicators, results suggest that our measure of legal framework is associated with a decrease in average quality of service and an increase in average tariffs. When we evaluated each of the sectors separately, we observed that in water utilities there are some differences in terms of labor productivity where higher soundness is associated with higher labor productivity; in electricity distribution we observe the opposite trend. The main other results hold. 284. Board of Directors and CEO: This section focuses on the composition, qualifications, and performance evaluation of the Board of Directors of SOEs. It prioritizes a Board of Directors where political discretion is low, where members to the Board are selected on pre-defined criteria, and whose performance is assessed based on different governance arrangements. The results suggest that the higher the scores in these dimensions, the lower the distributional losses and service coverage. Results also show that the higher the qualifications of the Board, the higher the level of average tariffs. Growth rates in these performance indicators seem not to be significantly affected by the Board and CEO competitiveness; however, when the sectors are analyzed separately, the change in performance in water seems to be more sensitive to these dimensions. Moreover, these dimensions are associated with higher continuity of the service. Our measure of CEO competitiveness is more related to positive changes in coverage and the reduction of average tariffs, while Board competitiveness is associated with positive changes in labor productivity and micro-metering. For the electricity sector, our results were not significantly different than zero. 285. Management/Staff: This index measures the composition and characteristics of the enterprise's staff. When we compare corporate governance indicators and operational performance in electricity and water, only labor productivity had a direct correlation with professional management. Nonetheless, when we disaggregate the results for the water sector, management is associated with higher levels of labor productivity and lower distributional losses. In addition, it is also related with positive significant changes in the continuity of the service, sewerage coverage, labor productivity, and micro-metering. 286. Transparency and Disclosure: The transparency index measures the existence of mechanisms that allows for a better publication of the company's financial and non -financial information, the involvement of civil society in decision-making, the disclosure of financial information, and the independent auditing of SOEs' accounts. Those utilities with higher transparency and disclosure standards are associated with higher levels of service coverage and lower average tariffs. When we analyze each sector separately, the data illustrate that electricity utilities have significant coverage increases and tariff reductions. The correlation results are stronger in the water sector where we find that transparency is related with higher levels of efficiency, lower non-revenue water, higher potability, metering, and coverage. 287. Performance Orientation: This component addresses the existence of mechanisms (both internal and external) to evaluate the performance of SOEs. We emphasize arrangements that allow for a performance-based management of the enterprise. As expected, this index is highly correlated with high levels of labor productivity and low distributional losses, as well as significant changes in coverage. Most of these results hold when we separately assess each sector. 288. Aggregated Corporate Governance: The aggregated measure of Corporate Governance presents the overall results for the region in terms of the ranking of companies according to the previous five components of the framework. We find that overall corporate governance is highly correlated with high levels of labor productivity and tariffs, as well as with low distributional losses. We also observed positive changes in coverage of the service. The correlation results are stronger in water utilities than in electricity providers. For water companies, the data highlight 80 that overall corporate governance is associated with low non revenue water, high quality standards, coverage, labor productivity as well as high average tariffs. When we analyze the relation between governance and changes in performance for this sector, we find significant contributions in the improvement of the continuity of the service, labor productivity, metering, and sewerage coverage, as well as a reduction in average tariffs. 289. Although the previous paragraph illustrates the impacts of aggregated corporate governance (as a simple average across the six dimensions evaluated) on performance, it is interesting to disentangle the different aspects of governance. Although each dimension had a particular scope and interpretation it is likely that some of them behave similarly. Hence, we applied a Principal Component approach in order to comprise the six indicators in the relevant components, thus minimizing the lost of information. The Principal Component Analysis (PCA) develops a composite index by defining a real valued function over the relevant variables objectively. The principle of this method lies in the fact that when different characteristics are observed about a set of events, the characteristic with higher variation explains a higher proportion of the variation in the dependent variable compared to a variable displaying less variation. Therefore, the issue is one of finding weights to be assigned to each of the concerned variables determined by the principle that the objective is to maximize the variation in the linear composite of these variables. In other words, this approach allows for identifying patterns in data, and expressing the data in such a way as to highlight their similarities and differences. Since patterns in data can be hard to find in data of high dimension, PCA may contribute in analyzing data. Furthermore, an additional advantage of PCA is that once you have found these patterns in the data, you may compress the data by reducing the dimensions, without much loss of 65 information. 290. Among them, the first principal component factor accounts for 36 percent of the variance of the seven indexes. The other two component factors account for 23 and 17 percent of the variance respectively. The three factors together account for 76 of the total variance. 66 Table A7.9 (In Annex 7) presents the factor loading: Factor 1 is associated with the Professional Management and the Performance Orientation dimensions. Factor 2 reflects the Board Competitiveness aspects. Finally, Factor 3 is related to Legal Soundness and Transparency and Disclosure. The correlation between the aggregated index as a weighted average of these factors and our previous aggregated corporate governance index is significant (0.87). Furthermore, the ranking of countries presents some changes in the relative position for each country; however, the story previously described still holds. 291. This assessment is the first evaluation of this kind and these results are significantly promising: corporate governance is associated with high standards of utilities' performance and growth rates. As expected, performance orientation and professional management characteristics seem to be the greatest contributors to performance; all the other dimensions are associated with some of the performance indicators. Results in the water sector were stronger, presumably because of the higher number of water utilities in our questionnaire. Further analysis should 65 We use PCA to jointly take into account the information provided by our six main governance indicators ratios (Table A7.8) and generate orthogonal indexes to measure corporate governance. Factor scores were then calculated for each of the utilities. As a first step, we determine how many factors we may use in our analysis. Table A7.8 reports the estimated factors and their eigenvalues. Only those factors accounting for greater than 10 percent of the variance (eigenvalues >1) are kept in the analysis. As a result, only the first three factors are finally retained. 66 These factors allow for computing the factor score coefficient matrix. To enhance these factors' interpretability, we use the varimax factor rotation method to minimize the number of variables that have high loadings on a factor. In other words, varimax rotation produces results which make it the most likely to identify each variable with a single factor. This approach greatly enhances our ability to make substantive interpretation of the main factors. Table A7.9 presents the factor loadings, where variables with large loadings (N>0.4) for a given factor are highlighted in bold. 81 include more disaggregated data and higher coverage of the sample. It would be also critical to explore political economy approaches that address issues of causality, sequencing, and complex interaction effects that contribute to the explanation of SOEs' governance. It would also be important to complement the previous analysis with detailed case studies for improving the knowledge of the internal mechanisms for performance. 5.4. C ONCLUSIONS 292. Governance arrangements in SOEs in water and electricity distribution present a wide spectrum of designs. While private enterprises are characterized by the adoption of standard corporate strategies, SOEs standards vary depending on countries' institutional systems and the characteristics of the service. Thus, the variety of arrangements calls for a careful systematization of governance practices and the identification of successful experiences. SOEs are part of the public sector and factors of good and bad performance are directly or indirectly related to countries'/provinces' overall governance. 293. This chapter emphasized the need for a corporate structure that prevents political intervention, rewards performance, and is subject to public scrutiny. Additionally, it focused on the qualification of the staff of the enterprises. Although we tried to capture as many variables from state enterprises as possible, the focus of this work was on design. In other words, it did not consider the actual effectiveness of governance procedures. 294. Like a private enterprise, the organizational structure and decision-making of a SOE reflects the interests and involvement of their shareholders, and hence, their strengths and weaknesses. Because these enterprises are part of the public administration and, thus, subject to its governance schemes and leadership they can benefit or not from the performance of its bureaucracy. Government corporations remain a complex and unique organizational mode, caught between the norms of public sector governance and corporate governance (Whincop, xxx). Hence, although mimicking private enterprise arrangements in SOEs might cause significant improvements in management, it can also contribute to the consolidation of corruption and the lack of accountability in those enterprises with little controls and vested interests from governing stakeholders. 295. Our focus on five components of their design, allowed us to identify the major pitfalls in issues related to their performance orientation and the selection and composition of the Board of Directors. Companies do plan their strategies, but it is not clear the way objectives are agreed nor their monitoring and enforcement. Generally, SOEs are subject to influences of different authorities, particularly during their planning process. 296. Rather than focusing on profit maximization, SOEs emphasize social goals and human capital improvement. Thus, manpower is a critical factor of state enterprises' performance. Moreover, in several cases the company's bureaucracy has built a prestige for good performance that has prevented the intromission of political interests. In our sample of SOEs, directors to the Board have, on average, high educational levels. Almost all CEOs of public companies have a university degree and in 56 percent of the cases they also have postgraduate studies. 297. The major difference, with the private sector staff selection process, comes from the way management staff selected. From Board of Directors to low levels of employees, a significant 82 percentage is hired either internally or discretionally, with low levels of competition. Even though internal hiring is also a common pattern for private enterprises, in state enterprises the space for collusion is bigger and, hence, measures need to be taken to avoid low levels of professionalism and political appointees. 298. Good management of SOEs presents government bureaucrats with different challenges. First and most important, state enterprises face conflicting goals that affect the establishment of a business strategy. Several departments usually compete for moving their agenda into the priorities of the company, affecting the prioritization of the service. Most importantly, intromissions in the companies' business adopt informal, ad-hoc, approaches, that prevent the company from making explicit the costs of them. The lack of profit-orientation prevents SOEs from identifying ways to improve efficiency and performance. Because low revenues can be compensated by government subsidies, efforts to make the company sustainable fall to second place. Third, poor accountability systems (being at the regulatory or management levels) prevent the development of an ownership structure that triggers efficient behavior from senior management. 299. Although it is too early to formulate policy recommendations, both the literature and the practices in the region help identify some potential actions. Considering public enterprises as private companies can in some cases lead to wrong diagnoses and, hence, reform plans. Some, if not the majority of SOEs in water and electricity distribution, are not profit driven, which makes the corporate incentives on which private enterprises are based questionable. As Whincop pointed out, it makes sense to design governance appropriate to the form rather than to emulate the incentive structure of other alternatives. This calls for the identification of governance schemes that focus on the factors that may trigger efficiency, reducing the space for corruption and capture by vested interests. 300. It is in this context that accountability emerges as the main governance aspect of SOEs. In the cases of companies with high levels of corruption and inefficiency, accountability systems should prevent discretional management (both from management and political authorities) and create the incentives for good performance. Regulation and performance-based management could be considered complementary ways of achieving these goals; although good care needs to be taken in creating checks and balances such as parliamentary oversight and state auditing. 301. A final observation is related to the importance of tailoring governance strategies to companies' realities. This chapter analyzed both cases of full and partial state ownership. Among those with partial state ownership, particularly those with significant PSP, a governance design reflecting the incentives of private enterprises seems more appropriate. For companies with significant gaps in both performance and management, transparent accountability mechanisms should be considered. A third group of companies, those with full state ownership, characterized by good sector performance and management need to strike a balance between private sector orientation and public accountability. Finally, governance design needs to take into consideration sector differences. Technology and sector dynamics also determine management. 302. Corporate governance is associated with high standards of utilities' performance. As expected, performance orientation and professional management characteristics seem to be the highest contributors for performance; however, all the other dimensions are associated with some of the performance indicators. Results in the water sector were stronger, presumably because of the higher number of water utilities in our questionnaire. These results are quite encouraging in order to better understand sector performance. 83 6. OTHER DETERMINANTS FOR SECTOR PERFORMANCE 303. While the purpose of this report is to focus on particular utility level variables as determinants of sector performance, this chapter briefly summarizes a number of additional factors and the interaction of some of these factors, as they may impact sector performance. This chapter reviews and summarizes the results of previous empirical analysis of issues that in different ways impact utilities' decision making process. These decisions then have an impact which can be measured through the indicators proposed in this study. On one hand, academics and researchers have modeled and empirically tested the impact of such issues as corruption, market structure, economies of scope and density, renegotiation, and reputation. On the other hand, some have proposed that other issues like subsidy mechanisms, lack of cost recovery, the political economy of the different sectors and social accountability also play a role in sector performance. Although widely discussed, few econometric studies exist and most analyses rely on comprehensive analytical case studies. O THER D ETERMINANTS: 304. Corruption: Corruption can have a destructive effect in sector performance. As previous research has suggested, it affects the pace and nature of PSP in infrastructure service provision affecting competitive bidding and resulting in unequal allocation of bids that can further result in monopoly rents instead of efficiency gains (Andres et. al 2008). Through multiple transmission mechanisms, corruption directly impacts sector performance (as defined by this report). Various studies have linked corruption not only with lower levels of investments but also with types of investments, suboptimal levels of quality, access, and prices 67. By using models that associate corruption with the fate of firms and their ability to devote managerial efforts to supervision and coordination of the use of productive factors, Bó and Rossi (2007) show that corruption diverts managerial effort from the productive process. Firms, hence, need to use a different sub-optimal combination of inputs to meet their service obligations. The model shows that more corrupt countries have less efficient (with lower labor productivity) firms. 305. By measuring the impact of corruption on performance and the interaction between reforms (introducing private participation and/or an independent regulatory agency) and corruption, Estache et al (2009) test the extent to which these reforms can reinforce or off -set the impacts of corruption and the extent to which corruption reinforces or offsets the impacts of the policy changes. Results show that in electricity distribution, corruption off-sets the impact of reforms. An increase in the corruption index results in a decrease in energy use. Furthermore, an increase in corruption in countries with SOEs is associated with lower residential prices and deterioration of access and quality. For water, the model is less conclusive possibly because of the poor quality of the data available. The negative interaction between corruption and the effects of introducing an independent regulatory agency means that the presence of these agencies offsets the effects of corruption in electricity and telecommunications. Clarke and Wu (2004) provide evidence for the effects of petty corruption at the utility level 68, and the impact on service 67 A comprehensive review of recent studies on corruption and infrastructure can be found in Infrastructure: A survey of recent and upcoming issues, by Antonio Estache for The World Bank (2006) 68 Clarke and Wu use a unique dataset for 21 Eastern European countries that includes information about bribes paid by and to utilities for service provision. 84 provision and sector performance. They show that corruption creates greater constraints on utility capacity and lowers competition among utilities. 306. Cost recovery: Cost Recovery is considered the most significant policy aspect when attempting to explain water policy performance. Among the seven policy aspects considered in Evaluating Water Institutions and Water Sector Performance (Saleth and Dinar, 1999), the level of cost recovery ranks as the most significant factor in explaining water policy performance. This study evaluates the overall performance of water institutions and their ultimate impact on water sector performance by expounding upon various inter-linkages between the two. For example, better water sector performance in Mexico and China, among others, demonstrates the value of water policy that is being conditioned by macro-economic policies. Traditionally, utilities have charged tariffs far from cost recovery levels. As mentioned earlier, this was one of the fundamental reasons for promoting PSP in fixed telecommunications, electricity distribution, and water distribution during the 1990's in LAC. Low tariffs affected service provision through lack of network expansion, low coverage rates, and low service quality. Since poor customers could (can) not afford service at higher prices, subsidy mechanisms are still part of the price structure in utilities like electricity and water. Some studies suggest a link between type of subsidy mechanism and sector performance, since it created incentives for a particular behavior from customers that hinders the utilities ability to maximize its profits and perform efficiently (Komives et al, 2005). 307. Civil Society Role: While substantial attention is placed on the financial and technical governance of utilities, the voice of users is often overlooked. The lack of a mechanism for incorporating users' priorities and preferences into the decision-making processes of the service provider may lead to service deterioration, and client estrangement. In Ways to improve water services by making utilities more accountable to their users Muller et al (2008) explore innovative approaches to public management in order to hold service providers more directly accountable to their users for the outcomes of their work. Accountability in this context is about directly channeling users to service providers. Another work that looks at the role of civil society in water provider accountability in 18 Asian cities is Water in Asian Cities: Utilities Performance and Civil Society Views (ADB, 2004). This report explores priority areas for both user and service providers, such as improving governance and reducing corruption, and suggests that this overlap of priorities may be a powerful determinant for improved sector performance. 308. Contract Arrangements: When considering how to provide effective and equitable public services, decisions about investment levels and contract arrangements, have proven to be significant determinants. Various pieces of research have assessed the challenges, opportunities, and options for public-private partnerships (PPPs) and their impact on sector performance. According to Ogunbiyi (2004), several schemes have had a negative impact on the poorest of the poor by restricting their access to clean supplies due to high tariffs. The same author further asserts that PPP schemes involving management contracts, where the combination of public finance and private management of technical and commercial operations has been applied, could be the best type of contractual arrangement for water supply and sanitation in Africa. In Senegal for example, the choice of an affermage contract, which was enhanced by the addition of strong financial incentives to reduce leakage and improve billing and collection efficiency, was innovative. It addressed the needs of the Government and kept the assets in their hands, and operations and maintenance functions were clearly defined. Furthermore, the nature of the contract fostered a partnership between the Government and the private operator. 309. PSP and Renegotiation: While renegotiation may be the inevitable consequence of contract incompleteness, and sometimes the solution to some of the inefficiencies caused by it, 85 several authors have pointed to the importance of its negative practical consequences. For example, Engel et al (1997), who study the effects of government guarantees and renegotiation in the efficiency of the PPP contracts, note that renegotiations increase the discretionality of the government, reduce the incentives for efficiency for the firms and encourage them to lowball their bids for the projects, especially if they have experience in lobbying. Furthermore, Guasch (2004) notes that the lowball strategies in the bidding process undermine the efficiency of the allocation, and as a consequence consumer welfare and sector performance. The most relevant research conducted to understand the relation between renegotiation and lowballing bidding strategies are Guasch et al (2001), Estache and Quesada (2001), and Guasch et al, who developed theoretical models with lowballing as an equilibrium strategy for rational bidders. In recent work, Guasch et al (2004) provide a quantitative measure of the lowballing effect of th e expectations of future renegotiation over the bidding strategies. They conclude that renegotiation expectations appear to significantly affect the competitive bidding of PPP infrastructure projects. Disaggregating by renegotiation requesting party, there's evidence in favor of a positive effect if the requesting party is the winning firm, and slightly less clear evidence if the requesting party is the government. 310. PSP and Reputation: Bajari et al (2007) develop a structural auction model, in which they use data on the projects characteristics, including the amount of ex post adjustments to the original construction budgets (the auctions are for the right to construct highways in California, and the bids are for the lower construction costs), as a measure of the expected extra revenues the firm may obtain after requesting a contract renegotiation to the government. Since the extra revenues affect less than proportionally the bids in the auction stage, the authors conclude the existence of sizeable transaction costs from the renegotiation process. In the same line, Andres et al (2009) consider the ex post outcome (the occurrence of renegotiation) to measure ex ante expectations. More precisely, they explicitly model the expectations of renegotiation using ex post occurrences of renegotiation for a given country. The framework allows them to use more information and eliminate the possible bias from the estimation. The results suggest that bidders (especially the ones with the highest valuations for the project) adjust upwards their investment offer when renegotiation is a plausible outcome after the concession is awarded. 311. Economies of Scope, Scale, and Density: Research to determine the optimal size of utilities focuses on estimating cost or production functions, where firms either minimize costs or maximize profits. Through the use of these types of models, a number of studies have been able to establish the optimal size of a given utility firm and determine the existence or non-existence of economies of scale and scope in different sectors 69. 312. Additional research has focused on trying to measure the existence of economies of density in water and electricity. In the case of electricity, by using frontier analysis estimation methods, a number of studies have performed extensive comparisons among utilities to determine what factors affect individual firm productivity and efficiency. These frontier analysis models consider structural variables to account for potential existence of economies of scale and density. In some cases, they have found that settlement density, urban vs. rural (Cullman 2008), and consumer structure, affect the performance productivity of utilities (Von Hirschhausen and Kappeler, 2004). By estimating cost functions, several studies have been able to show the 69 For water supply, Kim and Clarke (1988) study the effects of economies of scale and scope in a multiproduct utility, using a translog multiproduct joint cost function. For electricity, Hjalmarsson and Veiderpass (1992) use Data Envelope Analysis to examine productivity growth, and the effects of economies of density, in electricity retail distribution for Sweden. 86 existence of economies of density and economies of scale for small and medium sized electric utilities. Since smaller utilities do not operate at an optimal service level, costs can be reduced by merging and increasing their service area (Filipino 1998; Filipino and Wild 2001). Low population density service areas can become a barrier for utilities performance, since it makes it harder for them to exploit economies of scale in management and physical plant (Gomez-Ibanez 2007). For a set of Southern African countries, using a total factor productivity analysis, it was found that there is a basic correlation between performance and market and that there are clear advantages in terms of the existence of a private actor or not, the dependence on hydr o source, the degree of vertical integration or the existence of an independent regulator (Estache et al, 2007). 313. For water, economies of density can also account for differences in sector performance. Using both models that estimate costs and production functions, studies suggest that under certain conditions, economies or diseconomies of scale and density exist. For a multiproduct utility where residential and non-residential service is considered a different product, models show that for residential service there are diseconomies of scale, but for non-residential service there are economies of scale. Perhaps more interesting is that the economies of scale achieved in water treatment are mostly lost in the distribution of water and the utility on the whole experiences economies of scope associated with joint production of the two services, water and sanitation (Kim and Clark, 1988). In a study of four developing countries, the results show that there is economies of scope in areas where utilities provide both water and sewerage services providing evidence that an integrated utility provides both services at a lower cost than two separate utilities with one specialized in water production and the other in wastewater collection (Nauges and Van den Berg, 2008). Market structure (vertical integration) can thus affect sector performance, particularly at the utility level. 314. Competition (in the Telecommunication Sector): During the 1990's in the LAC region both privatization and introduction of competition in the sector was recommended. There is broad agreement, among academics and practitioners, that competition is the most effective method of promoting investments in the telecom sector. A monopoly provider, whether a SOE or a private operator, faces fewer incentives to improve service and lower prices than enterprises operating in a competitive environment (Wallsten, 2001). In most countries, liberalization of the long-distance market took place a few years after privatization (Andres et al, 2008). In order to ident ify the effects of competition on performance, the literature uses two proxies for competition: the actual long-distance liberalization and existence and coverage of cellular phone providers, as a threat of competition for the fixed market. Petrazzini and Clark (1996) find that service coverage is higher in competitive markets while Wallsten (2001) finds that competition is associated with increased mainline penetration, payphones, connectivity capacity, and lower prices for local calls. Finally, Andres et al (2008) find that the main driver for sector performance in these markets is PSP. When a control variable for PSP is included in the model, the result of introducing competition in the market was a reduction in end prices. ... I N S UMMARY ... 315. While this chapter has outlined a number of factors that may impact sector performance, this report attempts to understand sector performance through the performance of individual utilities in the LAC region. The aforementioned issues alone do not explain sector performance, however they may all impact the performance indicators this study uses; thus, affecting utility performance. Either through direct links, for example subsidy mechanisms that result in non-cost recovery tariffs and restrict the firm's financial ability to expand coverage and provide adequate 87 service quality; or through more indirect links, like improving social accountability by introducing a mechanism that can hold service providers more directly accountable to their users for the outcomes of their work; the issues reviewed in this chapter interact to explain the type of incentive framework utilities use to make management and operation decisions. Our objective was not to fully explain sector performance but to recognize and acknowledge that other issues might influence utility behavior and the type of incentives they have to perform efficiently. The review in this chapter is precisely an attempt to recognize that the broader definition of sector performance includes how the external environment (Figure 2) and the sectoral environment shapes utility performance. 88 7. CONCLUSIONS 316. The report defined sector performance as the delivery of a reliable affordable service that complies with certain quality standards. While some reforms successfully achieved these objectives, overall they encountered difficulties and today several countries in the region are facing new challenges. As a result, the region has witnessed significant improvements in sector performance; however, recent years reported a significant drop in public and private investment, and an increase in dissatisfaction with some of the policies implemented during the 90s. Ultimately, this resulted in the inability to secure access to affordable services for the poor. 317. Against this backdrop, this report analyzed a selected number of determinants that have proved to impact sector performance between 1990 and 2006. By analyzing the trends in sector performance and several of its determinants, this report (i) responds to the need for a better understanding of LAC's infrastructure sector performance, and (ii) provides the empirical knowledge and foundation necessary for meeting the infrastructure challenges that the LAC region currently faces. Understanding the various interventions and conditions that explain LAC's sector performance is an indispensible milestone in minimizing the region's infrastructure gap. 318. The results of this report can be summarized in three main messages: Sector performance for electricity distribution, water and sanitation, and fixed telecommunications significantly improved in LAC but there is still much room for improvement. This report delves into the various dimensions of sector performance by describing the main elements that characterize sector performance as the delivery of a reliable affordable service that complies with certain quality standards. Throughout the past 15 years, coverage, service quality, and labor productivity in all sectors studied show noteworthy improvements. Coverage for the utilities covered in our databases increased to 95 percent in electricity, 97 percent in water, and 62 percent in fixed telecommunications by 2005 year. The quality of service improved: in electricity frequency of interruptions fell by half, continuity of service in water increased 8 percent, and the number of telephone faults per year dropped from 23 to 8. Private sector participation had a positive effect on labor productivity, efficiency and quality. Furthermore, introducing Independent Regulatory Agencies in the electric ity and water sectors promoted gradual improvements in the utilities' performance. The differences in performance amongst utilities suggest no one model as the best fit because there are different approaches and correlated variables that contribute to good performance. The results are usually dependent on initial conditions, and the implementation mechanisms. Throughout the last decade, service provision improved in both private and public companies. Even though the average top private performer outperforms the top public utility, in a number of cases top public performers outperform average private utilities. Similarly, smaller companies outperform larger companies. Some countries are top performers in electricity, others in water. This report explains how part of the heterogeneity in performance amongst utility distribution companies is explained by the different regulatory governance 89 arrangements, the degree of private sector participation, and the governance design for SOEs. Results show that both the government (as a regulator and a service provider) and the private sector (as a service provider) can play an active role in enhancing sector performance. Introducing Private Sector Participation (PSP) alone was not the answer to better sector performance. While, the government continues to be at the heart of infrastructure service delivery, private sector participation was an important partner in improving sector performance. However, the manner in which PSP was developed determined the extent of its impact on performance. By promoting transparent and accountable regulatory governance design, the government can make positive contributions to sector performance. An independent regulatory agency free of political interference and accountable for its decision significantly improves utility performance, even for SOEs. Furthermore, SOEs that have a corporate governance structure that reduces political interference, rewards performance, and opens decisions to public scrutiny perform better than those that have a structure that allows politics to influence decision making. Improving sector performance requires a holistic and case-based approach Improving sector performance goes beyond conducting a comprehensive assessment of a key determinant and proposing specific designs that address issues related to that determinant; it entails an approach that integrates policies that address a wide range of issues, some of which are introduced in detail in this report. By acknowledging and determining the differences amongst service providers and the environments in which they operate, policy makers can design comprehensive solutions to complex problems in infrastructure service provision. i. 319. There are other determinants issues that can affect sectors' performance but their interactions have not yet been thoroughly evaluated. This report briefly summarized other potential contributors to sector performance, but it did not make an attempt to further analyze the direct impact of these factors on sector performance. These factors include corruption, market structure, potential for contract renegotiation and reputation, type of contract arrangements for service provision, existence of social accountability mechanisms. The main objective of this report is to provide a factual description of the changes and policies that can be empirically tested and analyzed. Therefore, we restricted the scope to some of the potential policies that could be developed within the sectors. LOOKING AHEAD 320. While this initiative describes and benchmarks the region's good and poor utility performers, it also calls upon further analytical work to explain: (i) how the various determinants discussed in this report interact and impact specific performance indicators, and (ii) why there are such discrepancies among country and utilities albeit similar characteristics and environments. An in-depth analysis of the facts presented in this report, would allow us to draw further conclusions regarding the trends and changes that characterize the region. 90 321. A thorough understanding of how and why regional, country, and utility performance improved or worsened will allow LAC countries to share experiences and learn from each other by assessing what has worked and what has not worked. By doing so, different stakeholders can work together to establish the strongest possible foundation for efficient and reliable sectors in the future. In addition, future analytical work can target potential audiences such as the private sector, utility managers, political decision makers, policy makers, and regulators, among others. Such analytical exercise provides potential users with the knowledge and tools to move ahead, and impetus for future reform. 322. To move ahead, it is equally important to maintain, update and improve the quality of the data used in this report, so that it remains an on-going resource for the Bank and the community at large. Efforts to continue data collection and analysis are crucial in order for the World Bank to provide a resource that remains useful for LAC and other regions. 323. Utility sector performance is a complex undertaking that encompasses a variety of dimensions. Impacts on each of these dimensions are not necessarily straightforward, with differences determined by sector, and internal and external environments. Policy makers considering future sector reforms should first prioritize their performance objectives. Once the objectives are identified, the detailed results presented by the analysis can be mined to determine the circumstances in which those objectives can be achieved. For instance, if a utility prioritizes quality and efficiency over retaining employees, private sector participation would be an attractive option. Similarly, if reducing distributional losses is a key objective, in a SOE, then a sound design of its corporate governance with well designed performance orientation rules can be considered. 324. The results presented in this report are instructive to policy makers in terms of highlighting pitfalls in sector reform programs. Poor design and faulty implementation explain many of the shortcomings in reform processes. Identifying the potential for these in advance can assist policy makers in the design of proactive counter measures. Consider the case of an electricity distribution policy maker who has prioritized improving quality and reducing distributional losses--and hence decided to move ahead with PSP. By drawing lessons from the experience detailed in this analysis, the policy maker could design a public relations campaign emphasizing expected benefits and cautioning consumers of potential price increases and reductions in sector employment. As a whole, this report can help policy makers make informed decisions and well designed change strategies, allowing them to maximize both technical and political objectives. 325. As mentioned, the program and reforms could have been implemented better. The overall results are quite positive, but the perception appears quite negative. Although it seems a paradox, valid reasons explain the divergence between perceptions and facts. To solve this paradox, it is important to understand the reasons that generate the disco ntent of the citizens and their particular point of view. To achieve greater benefits and higher popular approval, in some countries, the process of introducing PSP could have been better prepared and communicated. The context in which the programs of PSP were developed in the region was one of: excessive optimism a belief in quick positive profits, too many promises, a lack of realism, the poor handling of the expectations, and a constant breach in contractual agreements by both parties. Social distribution and lack of transparency throughout the process appear as common denominators in that context. 91 326. By securing an environment that maximizes the benefits of reform and promotes a broad consensus, reform programs in the infrastructure sectors can be successfully implemented. In moving forward, the lessons from the past need to be accounted for and corrected. The ultimate objective is to secure improved sector performance and long-term efficiency, reduce poverty through better concession design and regulation, and foster compliance with the terms agreed to by both the government and the operator. To establish such an environment, concession laws and contracts should (i) focus on securing long-term sector efficiency and proper risk assignments and mitigation, as well as discourage opportunistic bidding and renegotiation; (ii) be embedded in regulations that foster transparency and predictability, support incentives for efficient behavior, impede opportunistic renegotiation and force contract compliance; (iii) address social concerns and focus on poverty; and (iv) promote accountability as the main governance aspect of SOEs. 327. Governments remain at the heart of infrastructure service delivery. SOEs that have a corporate governance structure that reduces political interference, rewards performance, and opens decisions to public scrutiny perform better than those that have a structure that allows politics to influence decision making. Furthermore, even under the presence of PSP, there may be a need for public involvement. Governments need to regulate infrastructure provision as well as contribute a good share of the investment. They must leverage their resources to attract complementary financing. Moreover, they are responsible for setting distributional objectives and ensuring that resources and policies are available to increase access for the poor. 328. Infrastructure service provision requires good performing SOEs and private companies that can disseminate good practices and with the government finance capital investments. Raising PSP to its previous levels requires addressing past problems and building on the lessons of the past decades. Under the current environment where infrastructure competes along side other investments for financial resources, increasing transparency, and improving the risk profile for projects rise as necessary conditions for further development. To do this regulatory risk must fall and the framework for PPI needs better risk mitigation mechanisms. Overwhelmingly negative public perceptions of PPI in some countries, is serious constraint on further participation that needs to be addressed. This, in turn, requires greater transparency, improved transaction design and oversight to reduce renegotiations and poor performance ; and better management of those who stand to lose out. 329. To make new reforms sustainable, not only the technical and financial aspects need to be addressed, but also the social aspects most responsible for the backlash. Better communication is critical to create popular support. It is essential to promote the program's infrastructure improvements, advertise the initiative, explain the impact of not improving (but rather maintaining) the status quo, and realistically argue the program's cost-benefit tradeoff. The communication strategy must not only justify the programs, but also periodically inform on the progress of the program, as well as of any changes or problems. The reforms must not only be successful, but that success must be communicated. Communication also serves as a safeguard against corruption at all the levels and as a tool to obtain greater popular support. Greater fairness and support to those adversely affected in the design of the transaction is needed. This can be achieved through the incorporation of social policies, such as social tariffs and financial assistance to those adversely affected by the programs such as those losing their jobs. Programs or policies should be implemented to support users and workers. Affected communities must be active participants in a successful program, and these communities must be involved from the start. Initiatives should be launched and supported from the bottom up in areas and locations where the benefits and costs will be incurred. 92 330. There are critical elements to be improved in order to move forward and secure success and maximum benefits of sector programs. The experiences in the region show that the key elements of a successful program must include the following: Improved Institutionality. Projects generally should be selected by the sectoral ministry, as a consequence of the country's strategic planning program and objectives. An interministerial group should be led by the finance minister to evaluate and approve the projects (accompanied by the appropriate economic and financial analysis) identified by the sectors. Improved Contract and Concession Design. Concession contracts should be awarded competitively and designed to avoid ambiguities as much as possible-- rather than direct adjudication or bilateral negotiation--and only after contracts have been carefully designed and reviewed and the qualifications of bidders have been screened. Outcome targets (regulation by objectives or service levels) should be the norm in contracts rather than investment obligations (regulation by means). Contracts should clearly define the treatment of assets, evaluation of investments, outcome indicators, procedures and guidelines to adjust and review tariffs, and criteria and penalties for early termination of concessions and procedures for resolution of conflicts. The sanctity of the bid is essential. For PSP to be successful and achieve the desired objectives, contracts and regulations need to be designed and enforced appropriately. The key objective should be to ensure that the contracting parties comply with the agreed conditions. Stronger Regulatory Framework. An appropriate regulatory framework and agency should be in place, with sufficient autonomy and implementation capacity to ensure high-quality enforcement and deter political opportunism. In addition, the tradeoffs between types of regulation--price cap and rate of return--should be well understood, including their different allocations of risk and implications for renegotiation. Technical regulation should fit information requirements and existing risks, and regulation should be by objectives and not by means. Thus, performance objectives should be used instead of investment obligations. Regulatory Instruments. Proper regulatory accounting of all assets and liabilities should be in place to avoid any ambiguity about the valuation of assets and liabilities and about the regulatory treatment and allocation of cost, investments, asset base, revenues, transactions with related parties, management fees, and operational and financial variables. Cost and financial models of the regulated utility should be standard regulatory instruments to assess performance, with particular emphasis on the evaluation of the cost of capital. Extensive use of benchmarking should be common best practice of regulatory agencies and is critical to assess the efficiency of operations and to assist in the ordinary five-year tariff reviews. Better Corporate Governance. Accountability emerges as the main governance aspect of SOEs. In the cases of companies with high levels of corruption and inefficiency, accountability systems should prevent discretional management (both from management and political authorities) and create the incentives for good performance. Regulation and performance-based management could be considered complementary ways of achieving these goals; although good care needs to be taken in creating checks and balances such as parliamentary oversight and state auditing. 93 An important observation is related to the importance of tailoring governance strategies to companies' realities. Among those with partial state ownership, particularly those with significant private sector participation, a governance design reflecting the incentives of private enterprises seems more appropriate. For companies with significant gaps in both performance and management, transparent accountability mechanisms should be considered. A third group of companies, those with full state ownership, characterized by good sector performance and management need to strike a balance between private sector orientation and public accountability. Finally, governance design needs to take into consideration sector differences. Technology and sector dynamics also determine management. Addressing Social Issues. Social tariffs, such as support for those adversely affected, should be a standard component of all programs. In particular, adoption and use of social tariffs and programs to subsidize access for the poor should be a part of all the relevant projects. In addition, programs or policies should be implemented to support adversely affected workers. Involvement of the affected communities from the start, at least in a consultative process, should be an integral part of any reform. Initiatives should be launched and supported from the bottom up in areas and locations where the benefits and costs will be incurred. Transparency and Communications. Better communication is essential to create popular support for reform, promote the program's infrastructure improvements, advertise the initiative, explain the likely impact and the consequences of maintaining the status quo, and realistically argue the program's cost-benefit tradeoff. The communication must not only justify the programs, but also periodically inform on the program's progress, as well as any changes or problems. The reforms must be successful, and that success must be communicated. Greater transparency in the overall process, financing, and use of funds is critical to provide a safeguard against corruption at all the levels and to obtain greater popular support. Evaluation and Monitoring. It is essential to periodically evaluate the accomplishments to improve efficiency and achieve the expected results and broadly communicate advances and pitfalls. 331. Sector performance should play a major role in defining the proper sectoral reforms. The newer modalities of PSP--beyond strict privatization-- and proper corporate governance design for SOEs offer significant potential for sector performance improvement. In particular, chances of success will be highly enhanced for programs that comply with the above- listed elements. Improvements in infrastructure for growth and poverty cannot be delayed. There are significant threats and opportunities. Most countries, including those in LAC, are at a crossroads on how to improve sector performance. Success may require some form of private sector involvement and financing. If obstacles such as poor perception of PSP are not removed, the significant gains and the very necessary modernization of the sector might fail, and the private financing will prove costly if not difficult. Conversely, opportunity exists to refine the model, attacking the problems and deficiencies of the past, through second-generation reforms that are constructive and broadly participatory. New reform processes that incorporate lessons learned with a clear participation of all the stakeholders and a protagonist role of the public sector are crucial. 94 A NNEX 1: E MPIRICAL A PPROACH : We will follow a similar approach of to econometric method proposed by Andres et al. (2007). The main difference is that we will build dummies for each characteristic and we will interact them with the ownership dummies both methodologies. In order to identify the effects of the characteristics we will modify (1') and (2') as follows: ln yijt T DUM _ TRAN ijt * X ijt P DUM _ POSTijt * X ijt ij Dij ijt (1) ij ln yijt T DUM _ TRAN ijt * X ijt P DUM _ POSTijt * X ijt ij Dij ij t ij ijt (2) ij ij where 1 if s ijt 1 DUM _ TRAN ijt 0 otherwise and 1 if s ijt 2 DUM _ POSTijt 0 otherwise where s ijt is a time trend that has a value equals to zero for the last year when the company had a public owner. Now, , that was a scalar number in our previous specifications, became a vector T 1 N with the coefficients for each characteristic of the vector X ijt than is of the form 1, xijt ,..., xijt with N as the total number of characteristics evaluated. Note the specifications used by Andres et al. (2008) were a particular case when we use a vector X ijt equals to 1,0,...,0 . In this case, the first coefficient will identify the average effect of change in ownership during the transitional period on a given indicator. As soon as we use the specification proposed in this Annex, the first coefficient of the vector will became the average effect of change in ownership during the T transitional period on a given indicator for a firm without the characteristics evaluated in the other elements of the vector X ijt . Equivalently, the vector will contain the coefficients for the different characteristics of vector P X ijt , but for the post-transitional years. As suggested by Andres et al. (2008), there are some indicators that present time trends. Hence, is more relevant their analysis using firm-specific time trend as shown in equation (2). Again, this relies on the assumption that trends between the three periods of analysis are the same. In order to relax this assumption, we will run a second set of equations (1) but using the (log) annual growth in each indicator. In this case, it will identify average changes in growth between the periods. 95 Given the fact that we are using a semi-logarithmic functional form of these models for each of the indicators, when interpreting the coefficient estimates of the dummy, it should be remembered that the percentage impact in each indicator is given by e 1 (Halvorsen and Palmquist, 1980). In order to correct for potential nonspherical errors a Generalized Least Square (GLS) approach will be more adequate. But, the GLS estimation requires the knowledge of the unconditional variance matrix of ijt , , up to scale. Hence, we must be able to write 2 C , where C is a known GxG positive definite matrix. But, in our case, as this matrix is not known, our second set of estimators will be a Feasible GLS (FGLS) that replaces the unknown matrix with a consistent estimator. 96 A NNEX 2: D ATA SETS Seven data sets were used and merged to provide a comprehensive analysis in this report. The performance indicators data set developed for this report is unique because of the comprehensiveness of the indicators and sectoral coverage. The data also have a relatively long time span, starting in 1995 and continuing until 2005 to 2007, depending on the sector. Data was collected from a variety of sources and was cross-checked, when possible. A particular effort was made in corroborating the company data with several public sources and with data of the firms provided by different government offices. In addition, the research was particularly cautious about the consistency and comparability of the data. In order to ensure high data quality and consistency, appropriate calculations and approximations were made to construct missing data points. For example, through the method of interpolation, data were constructed for the earlier years of certain variables, such as number of connections, number of employees, and so on. However, interpolation and other means of constructing data was the exception, and when used, it based on already concrete data and time trends. Specific methodologies were designed according to the variables at hand to ensure their comparability and consistency across time and utilities. 70 The data sets are the following: 1. Performance Indicators Data Set The performance indicators data set developed for Andres et al (2009) is unique because of the comprehensiveness of the indicators and sectoral coverage. It covers 181 infrastructure firms in Latin America that changed from public to private ownership during the 1990s. Many studies look only at the financial performance of privatized companies, which is just part of the story; this analysis considers changes in output, labor, efficiency, labor productivity, quality, coverage, and prices. In terms of sectors, the analysis includes the often-neglected water and electricity distribution sectors, in addition to fixed telecommunications. The analysis focuses on these sectors because of data availability and because they present similar characteristics (in the sense that they all have monopolistic features and are networking markets, allowing for similar interpretations of such indicators as labor productivity, coverage, and distributional losses), a feature that allows for cross-sectoral comparison. For these reasons, other sectors, such as transport, mobile telecommunications, and generation and transmission of electricity, among others, were excluded from the analysis. The data also have a relatively long time span, starting five years before the change in ownership and continuing five years after the privatization. The time span allows for the separation of short- run or transitional effects from long-run results. How short- and long-run effects are separated is discussed in the following methodology sections. The database targeted utilities privatized mainly in the period from 1990 to 2003--the main privatization wave in the region. The database also 70 This is the case, for instance, of the variable that measures number of employees in the case of utilities that were formerly vertically integrated. We compare the total number of employees of the different vertically disintegrated units and we compare with the total number of employees previous the change. We assumed that this change was proportionally similar to all the new units and then we use the growth rates for the previous years. 97 includes a few companies privatized during the 1980s (in cases in which pre-privatization data were available). Data was gathered from a variety of sources and was cross-checked, when possible. This research required the construction of an unbalanced panel data set of key indicators for utilities in LAC. For this, official data reported to investors by the firms and statistical reports of the regulator agencies of each country was used. Information was requested from each of the companies, as well as from each regulatory office. Furthermore, additional sources were used, such as ITU (International Telecommunication Union) and OLADE (Organización Latinoamericana de Energía, Latin American Organization of Energy). A particular effort was made in corroborating the company data with several public sources and with data of the firms provided by different government offices. In addition, the research was particularly cautious about the consistency and comparability of the data across time and across countries. 71 The analysis focused on several indicators of outcomes, employment, labor productivity, efficiency, quality, coverage, and prices. Some of these variables have been used by other authors in other samples, such as Ros (1999), who used equivalent indicators for coverage, labor productivity, quality, and prices, but did so for the telecommunications sector. Ramamurti (1996) used analogous indicators in output, coverage, and labor productivity for the four Latin American telecommunications firms of his study. Saal and Parker (2001) used similar indicators for output, employment, quality, and prices, but did so for water and sewerage companies of England and Wales. The countries analyzed in electricity distribution were Argentina, Bolivia, Brazil, Chile, Colombia, El Salvador, Guatemala, Nicaragua, Panama, and Peru. The sample consists of unbalanced panel data that includes 116 firms and 1,103 firm-year observations. Each of the firms included in the sample contains at least one year of pre-privatization data. In fact, 98 of the 116 firms have information for at least the previous three years. For water and sewerage, the paper reviewed companies in Argentina, Bolivia, Brazil, Chile, Colombia, Mexico, and Trinidad and Tobago. The sample consists of unbalanced panel data that includes 49 firms and 515 firm-year observations. Each of the firms included in the sample contains at least one year of pre-privatization data, and 35 of the 49 firms have information for at least the previous two years. The countries studied for the telecommunication sector were Argentina, Bolivia, Brazil, Chile, El Salvador, Guatemala, Guyana, Jamaica, Mexico, Nicaragua, Panama, Peru, Trinidad and Tobago, and República Bolivariana de Venezuela. The sample consists of an unbalanced panel data that includes 16 firms and 267 firm-year observations. Each of the firms included in the sample contains at least four years of pre-privatization data, and 17 out of the 18 firms have information for at least the previous four years. Table 1.1 presents the definitions of the variables used in the present analysis. 71 As quality indexes vary across countries, the most similar indexes were collected to compare their evolution across time, rather than absolute quality levels. 98 2. LAC Electricity Distribution Benchmarking Database The LAC electricity distribution benchmarking database was built by the World Bank (World Bank, 2008) and contains annual information of 250 private and state-owned utilities using 26 variables indicating coverage, output, input, labor productivity, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990, but the main focus is the period from 1995 to 2005. Data availability and data sources vary by country, often times depending on their ownership and means of regulation. While the benchmarking study uses a homogenous set of variables to collect data and measure performance, each country represents a special case and therefore efforts were made to ensure consistency of the data across time and utility. This database is representative of 89 percent of the electrification in the region (see Table A2.1). Furthermore, we argue that there is no significant self-selection in this database due the high data coverage. More precisely, most of the countries in the region were covered with, at least 75 percent of the electricity connections in the country. The only countries not covered were Cuba, Haiti, Guyana, Trinidad and Tobago, and some other islands in the Caribbean. The primary means of conducting research was field data collection and in-house data collection. A standard template and set of variables was used by both field and in-house consultants. Field consultants collected data to complement the information in some of the countries. Because of limited information available on the Web for these countries, local consultants were the most resourceful. For these selected countries and utilities, a preliminary feasibility screening was conducted to determine which countries would be likely to provide information. While field workers had direct access to the respective utility and government, the process of data collection was often hindered by unexpected factors, such as political affairs, bureaucracy, un -systematized data, and confidentiality issues, among other elements. The main sources for the in-house data collection were the World Wide Web, information collected by World Bank staff for other projects, and the internal World Bank databases (Development Data Platform, Integrated Records and Information System [IRIS], and so on). The main sources of information on the Internet were the utilities' Web sites. For some countries, the following sources proved to be useful: regulators, ministries, partnerships, central banks, online financial journals, papers, loan reports, financial reports, annual reports, monthly bulletins, statistics offices, and contacts with the companies and regulators. In addition, the following associations and organizations provided valuable statistics for the region: ARIAE ( Asociación Iberoamericana de Entidades Reguladores de Energía), ECLAC (Economic Commission for Latin America and the Caribbean), IEA (International Energy Agency), and CIER (Comisión de Integración Energética Regional). Because regulators, international organizations, and commissions often cover the electricity distribution of the entire region, most of the infor mation provided was aggregated at the country level and not disaggregated by utility. One of the challenges of data collection was the inconsistency between the data provided by utilities or regulators in annual and financial reports. To best describe the efficiency of the distribution sector of LAC, indicators were selected to determine utility-level performance. The utility-level indicators reflect relevant and feasible measurements in depicting the distribution segment of the electricity sector. The u tility-level indicators were computed to measure such factors as technical efficiency, operating efficiency, cost-efficiency, quality of service, and so on. Technical efficiency is defined as the capacity of the utility to achieve maximum output from a given set of inputs. To compute the technical efficiency of a utility, output and input indicators reflecting operating- and cost-efficiency were aggregated. 99 Table A2.1: Electricity Coverage and Data Coverage (Base year = 2005) Electricity Coverage Population Households with LAC Electricity Bench- (Census Dara; several sources) (Source: WDI & ITU) Power Connexion marking Database Urban Rural Total Total Pop % Urban Total HH (own calculation) Residential CXs % Total CXs Argentina 70% 95.4% 38,747,148 9140.0% 10,530,123 10,045,737 9,252,165 92% Bolivia 85% 28% 64.4% 9,182,015 6420.0% 2,135,003 1,374,942 942,805 69% Brazil 96.5% 186,830,759 8420.0% 54,223,593 52,325,767 49,600,000 95% Chile 90% 98.6% 16,295,102 8760.0% 4,791,755 4,724,670 4,486,053 95% Colombia 93% 55% 86.1% 42,889,000 7360.0% 9,028,323 7,773,386 7,773,386 100% Costa Rica 100% 87% 98.5% 4,327,228 6170.0% 1,006,053 990,962 990,962 100% Cuba 11,259,905 7560.0% 3,188,425 - Dominican Rep. 40% 82.5% 9,469,601 6680.0% 2,704,434 2,231,158 844,613 38% Ecuador 96% 54% 90.3% 13,060,993 6360.0% 2,902,443 2,620,906 2,620,906 100% El Salvador 97% 72% 79.5% 6,668,356 5980.0% 1,531,173 1,217,283 1,191,459 98% Guatemala 78.6% 12,709,564 4720.0% 2,955,713 2,323,190 1,583,268 68% Guyana 739,472 2820.0% 198,842 - Haiti 45% 36.0% 9,296,291 4270.0% 2,067,902 744,445 - 0% Honduras 94% 45% 69.0% 6,834,110 4650.0% 1,558,640 1,075,462 809,843 75% Jamaica 92.0% 2,650,400 5270.0% 764,827 703,641 491,452 70% Mexico 100% 95% 96.0% 103,089,133 7630.0% 24,703,635 23,715,490 23,715,490 100% Nicaragua 90% 40% 69.3% 5,462,539 5590.0% 974,652 675,434 534,886 79% Panama 85.2% 3,231,502 7080.0% 787,808 671,213 606,127 90% Paraguay 85.8% 5,898,651 5850.0% 1,453,110 1,246,769 871,717 70% Peru 30% 72.3% 27,274,266 7110.0% 6,244,176 4,514,540 3,597,326 80% Trinidad and T. 1,323,722 1220.0% 351,709 - Uruguay 95.4% 3,305,723 9200.0% 1,322,289 1,261,464 1,091,523 87% Venezuela 98.6% 26,577,000 9230.0% 5,945,522 5,862,285 4,802,261 82% Others (*) 90% 6,303,557 83.6% 1,969,846 1,772,861 230,707 LAC 91.6% 553,426,037 77.1% 143,019,699 131,006,044 116,036,948 89% (*) Antigua and Barbuda, Aruba, The Bahamas, Barbados, Belize, Cayman Islands, Dominica, Grenada, Netherlands Antilles, Puerto Rico, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname, and Virgin Islands (U.S.). The following table is a statistics summary for the datasets used in this report. We have calculated the number of observations, mean, standard deviation, minimum, and maximum, for the main indicators. The statistics show the heterogeneity and comprehensiveness of the data. Table A2.2: LAC Electricity Distribution Benchmarking Database ­ Summary Statistics Electricity Obs Mean Std. Dev. Min Max Number of Utilities 250 Total Connections (millions) 247 432,697 1,308,686 289 17,900,000 Total Residential Connections 247 377,889 1,147,699 251 15,800,000 Total Energy Sold (GWh) 248 2,605.0 8,997.4 4.1 111,000.0 Employees 235 1,244 3,062 13 36,942 Distribution Losses (%) 221 16.1 8.6 1.6 49.8 Average Duration of Interruptions per subscriber 149 26.1 33.2 0.5 209.2 Frequency of Interruptions per subscriber 137 27.6 47.7 1.0 285.2 Coverage (%) 151 78.8 17.0 28.1 100.0 Residential Connection per worker 231 392.9 273.5 58.3 2,694.1 Average GWh sold per Worker 230 1.8 1.4 0.14 11.8 Note: Each observation in this table corresponds to the simple average (across all the years with available information) for each utility. The LAC Electricity benchmarking dataset includes information for 250 utilities in 26 countries. The size of the utilities varies between over 17 million connections and as little as 289 connections. The dataset includes the information for most of the largest companies in the region, and some of the smaller companies. Evidence of this also is the difference in the energy sold yearly by each company. The utility with the lowest total energy sold sells 4.1 GWh per year, while the utility with the largest total energy sold sells 111,000 GWh per year. Distributional losses range between 1.6 and almost 50 percent of the energy produced. In terms of quality, the 100 indicators that show the differences amongst the observations in the sample are average duration of interruptions per subscriber, and frequency of interruptions per subscriber. The minimum and maximum for these indicators are respectively 0.5 and 209 minutes and 1.0 and 285.2 times. Labor productivity varies 58.3 and 2,694 connections per employee, and 0.14 and 11.8 GWh sold per employee. Average number of employees varies between 13 and 36,942. 3. LAC Water and Sanitation Benchmarking Database The LAC Water and Sanitation benchmarking database was built by the World Bank (World Bank, 2009) and contains annual information for 1700 private and state-owned utilities using 34 variables indicating coverage, output, input, labor productivity, operating performance, quality and customer services, and prices. The time frame covers data as early as 1990, but the main focus is the period from 1995 to 2006. Data availability and data sources vary by country, often times depending on their ownership and means of regulation. While the benchmarking study uses a homogenous set of variables to collect data and measure performance, each country represents a special case and therefore efforts were made to ensure consistency of the data across time and utility. This database is representative of 59 percent of the water connections in the region (See Table A2.3). Furthermore, most of the main utilities in the region covering urban areas were included in this database. The only countries not covered were Cuba, Dominican Republic, Guatemala, Guyana, Haiti, Jamaica, Venezuela, and some other islands in the Caribbean. The primary means of conducting research was in-house and direct data collection. A standard template and set of variables was used to collect the information. Because of limited information available on the Web for these countries, where feasible the information was requested directly to regulatory and sectoral agencies. In some cases, the utilities provided the information directly by completing the template. The main sources for the in-house data collection were the World Wide Web, information collected by World Bank staff for other projects, and the internal World Bank databases (Development Data Platform, Integrated Records and Information System [IRIS], and so on). The main sources of information on the Internet were the utilities' Web sites. For some countries, the following sources proved to be useful: regulators, ministries, partnerships, journals, papers, loan reports, financial reports, annual reports, monthly bulletins, and statistics offices, contacts with the companies and regulators. In addition, the following associations and organizations provided valuable statistics for the region: ADERASA (Asociación de Entes Reguladores de Agua Potable y Saneamiento de las Américas) and IBNET (International Benchmarking Network for Water and Sanitation Utilities). The information collected is for specific utility companies. In some cases, the existing data was at the municipal level. For those cases, we considered that the data for the municipality was that of the utility operator 72. In cases where the data was at the Municipal level and we were able to establish that the same operator serviced several municipalities, the data was aggregated at the utility level. One of the challenges of data collection was the inconsistency between the data provided by utilities or regulators and the annual and financial reports. Considering this, appropriate calculations and approximations were made to construct missing data points. For example, through the method of interpolation, data were constructed for the earlier years of certain variables, such as number of connections, number of employees, and so on. Interpolation and other means of constructing data was the exception based on already 72 For Mexico, the data submitted by the Consejo Nacional de Agua was at the municipal level. According to their description, the data for each Municipality corresponds to the data of the utility operator in the municipality area. For the few private operators, in Mexico, we were able to get data directly from the operator. 101 concrete data and time trends. Specific methodologies were designed according to the variables at hand to ensure their comparability and consistency across time and utilities. To best describe the efficiency of the distribution sector of LAC, indicators were selected to determine utility-level performance. The utility-level indicators reflect relevant and feasible measurements in depicting the distribution segment of the water and sanitation sector. The utility- level indicators were computed to measure such factors as technical efficiency, operating efficiency, cost-efficiency, quality of service, and so on. Technical efficiency is defined as the capacity of the utility to achieve maximum output from a given set of inputs. To compute the technical efficiency of a utility, output and input indicators reflecting operating- and cost- efficiency were aggregated. Table A.2.3: Water Coverage and Data Coverage (Base year = 2004) Water Coverage Population Households with LAC Water (Source: JMP) (Source: WDI & ITU) Water Connexion Benchmarking Database Urban Rural Total Total Pop % Urban Total HH (own calculation) Residential CXs % Total CXs Argentina 83% 45% 79.6% 38,371,527 91.1% 10,419,503 8,297,383 4,669,379 56% Bolivia 90% 44% 73.3% 9,009,045 63.7% 2,086,000 1,529,272 1,227,044 80% Brazil 91% 17% 78.9% 184,317,696 83.6% 51,939,168 40,961,306 37,100,000 91% Chile 99% 38% 91.2% 16,123,815 87.3% 4,741,622 4,325,715 3,555,960 82% Colombia 96% 51% 84.0% 42,306,000 73.3% 8,733,700 7,334,998 4,344,921 59% Costa Rica 99% 81% 92.0% 4,253,037 61.2% 989,172 910,125 397,902 44% Cuba 82% 49% 73.9% 11,246,670 75.6% 3,181,522 2,352,672 - 0% Dominican Rep. 92% 62% 81.8% 9,324,633 65.9% 2,663,357 2,177,986 - 0% Ecuador 82% 45% 68.3% 12,917,362 62.9% 2,870,525 1,960,218 617,605 32% El Salvador 81% 38% 63.6% 6,576,008 59.5% 1,542,091 980,671 545,223 56% Guatemala 89% 65% 76.2% 12,396,581 46.8% 2,817,405 2,147,629 - 0% Guyana 66% 45% 50.9% 738,992 28.3% 197,004 100,351 - 0% Haiti 24% 3% 11.7% 9,149,270 41.3% 2,008,392 234,355 - 0% Honduras 91% 62% 75.4% 6,702,291 46.1% 1,532,907 1,155,248 301,916 26% Jamaica 92% 46% 70.2% 2,638,100 52.5% 750,222 526,350 - 0% Mexico 96% 72% 90.2% 102,049,758 76.0% 24,626,697 22,221,949 8,241,126 37% Nicaragua 84% 27% 58.7% 5,393,597 55.7% 964,143 566,205 566,205 100% Panama 96% 72% 88.8% 3,175,354 69.8% 777,133 689,721 409,673 59% Paraguay 82% 25% 58.0% 5,788,088 57.9% 1,422,496 824,766 237,847 29% Peru 82% 39% 69.5% 26,958,549 71.0% 6,068,751 4,220,125 2,354,301 56% Trinidad and T. 80% 67% 68.5% 1,319,139 11.9% 350,223 240,076 240,076 100% Uruguay 97% 84% 95.9% 3,301,732 91.9% 1,303,720 1,250,813 715,563 57% Venezuela 84% 61% 82.1% 26,127,000 91.8% 5,743,930 4,716,306 - 0% Others (*) n.a. n.a. n.a. 6,255,641 83.2% 1,955,934 - LAC 90% 42% 79.2% 546,439,884 77.1% 139,392,648 110,336,799 65,524,741 59% (*) Antigua and Barbuda, Aruba, The Bahamas, Barbados, Belize, Cayman Islands, Dominica, Grenada, Netherlands Antilles, Puerto Rico, St. Kitts and Nevis, St. Lucia, St. Vincent and the Grenadines, Suriname, and Virgin Islands (U.S.). The following table summarizes the statistics for the datasets used in this report. As we presented for the LAC Electricity Distribution Benchmarking Database, we have calculated the number of observations, mean, standard deviation, minimum, and maximum, for the main indicators. The LAC Water and Sanitation Benchmarking dataset includes information for 1,708 utilities in 16 countries. The size of the utilities varies between over 6 million connections and as little as 110 connections. Coverage in the service area of the utilities in the sample varies between less than 10 and 100 percent. The dataset includes the information for most of the largest companies in the region, and some of the smaller companies. Evidence of this also is the difference in the volume of water produced by each utility. The utility with the lowest total volume of water produced 20,000 cubic meters per year, while the utility with the largest total volume of water produced per year produces 2.6 billion cubic meters. For sewerage collection, the wastewater collection varies between 0 and 420 million cubic meters. In terms of efficiency of service provision, the utility with the lowest collection rates, collect 16.4 percent of what they bill yearly, 102 while the best collect 100 percent. Metered connections vary between those that have no meter s, and those that have all connection with meters. Labor productivity, measured as number of connections per employee ranges between 38 and over 1,700 connections per employee. Table A2.4: LAC Water and Sanitation Benchmarking Database ­ Summary Statistics Water Obs Mean Std. Dev Min Max Number of Utilities 1,708 Total Water Connections 927 75,109 329,216 110 6,843,391 Total Residential Water Connections 927 68,652 300,203 100 6,247,583 Total Sewerage Connections 612 67,769 271,620 10 5,271,316 Total Residential Sewerage Connections 612 61,638 246,481 10 4,783,496 3 Total Volume of Water Produced (millions m ) 1,200 24.5 122.0 0.00002 2,610.0 Total Volume of Water Sold (millions m 3) 859 19.2 86.6 0.00021 1,720.0 3 Total Volume of Wastewater collected (millions m ) 722 6.3 31.6 0.0 420.0 Number of Employees 938 258 950 1 18,291 Unaccounted for Water (%) 803 35.4 17.3 0.0 99.9 Collection Rate (%) 1,006 89.5 15.0 16.4 100.0 Continuity of service (hrs) 523 22.7 3.7 1.9 24.0 Potability (%) 621 95.2 9.5 16.1 100.0 Water Coverage (%) 1,214 92.5 13.1 2.4 100.0 Sewerage Coverage (%) 1,073 79.3 25.9 1.2 100.0 Number of customer complaints 778 8,362 42,622 1 1,017,398 Labor Productivity (Connections per employee) 869 262.5 162.2 37.8 1,787.9 Metered (%) 699 72.2 31.7 0.0 100.0 Note: Each observation in this table corresponds to the simple average (across all the years with available information) for each utility. 4. ITU World Telecommunication/ICT Indicators Database This database contains annually time series from 1975-2007 for around 100 sets of telecommunication statistics covering telephone network size and dimension, mobile services, quality of service, traffic, staff, tariffs, revenue, and investment. Data for over 200 economies are available. The data is collected through an annual questionnaire sent out by the Telecommunication Development Bureau (BDT) of the ITU. The questionnaire is sent to the government agency in charge of the telecommunications sector, usually a line ministry or the regulator. The ITU's Market Information and Statistics (STAT) unit verifies, harmonizes, carries out additional research, and collects missing information from government websites, and operator's annual reports, particularly for those countries that do not provide answers to the questionnaire. Market research data is used to cross-check the data and complement missing values. In some cases, estimates are made by the ITU staff. For telecom, the ITU data includes information of 32 countries in LAC for most indicators. The sample includes small and large countries, as seen through the minimum and maximum statistics for telecom penetration and coverage. Furthermore, full-time staff varies between 188 to 90,576 employees. Quality also varies amongst the countries in the sample, from countries with 3.4 to 133 faults per 100 main (fixed) lines per year. Also, percentage of telephone faults cleared by next working day varies from 20 to 95 percent. Table A2.5 gives the reader a better idea of the diversity of countries in the data. 103 Table A2.5: ITU Database ­ Summary Statistics (for LAC) Telecom Obs Mean Std. Dev Min Max Main (fixed) lines in operation (millions) 32 2.0 4.9 0.2 24.9 Main (fixed) telephone lines per 100 inhabitants 32 16.7 15.3 1.0 79.6 Mobile cellular telephone subscribers per 100 inhabitants 32 30.0 22.3 2.2 109.4 % Households with a main line 27 41.4 26.0 4.3 90.0 % Residential main lines 32 73.2 5.7 59.3 85.0 % Digital main lines 32 83.5 16.3 38.3 100.0 Number of local (fixed) telephone (billions of calls) 17 1.9 2.9 0.0005 9.5 Number of local (fixed) telephone (billions of minutes) 24 12.1 30.5 0.0009 133.0 % of telephone faults cleared by next working day 29 58.2 20.7 19.7 95.0 Faults per 100 main (fixed) lines per year 29 42.9 30.9 3.4 133.2 Residential telephone connection charge (USD) 32 8.02 5.21 1.11 24.50 Price of a 3-minute fixed telephone local call (off-peak USD) 29 0.06 0.04 0.00 0.19 Price of a 3-minute fixed telephone local call (peak- USD) 30 0.10 0.11 0.00 0.64 Staff (Full-time telecommunications) 32 10,960 20,037 188 90,576 Waiting list for main (fixed) lines 30 146,816 242,245 615 980,262 Note: Each observation in this table corresponds to the simple average (across all the years with available information) for each utility. 5. Contract and Regulatory Characteristics Data Set The performance indicators data set was matched to a novel data set built by the World Bank that describes the characteristics of nearly 1,000 infrastructure projects awarded in Latin American and Caribbean countries from 1989 to 2002 (see Guasch 2004). The data set provides details on the privatization process, including how many bidders participated, the contract process, 73 the award criterion, 74 and the type of concession. 75 The data set covers the regulatory framework, including how the legal framework was established, 76 how tariffs are regulated, 77 if there was a possibility of renegotiation of the contract, and if so, who might be the initiator of the renegotiation. 78 The data set captures additional privatization contract details, including information about termination clauses, the arbitration process, claim-solving institutions, universal service obligations, contract duration, contract renewal, government guarantees, government subsidies, frequency of tariff review, and how the exchange and commercial risk were borne. If the contract was renegotiated, the reason given and the renegotiation outcome are also known. Characteristics of the regulator--such as an index of the regulator's autonomy, its budget source, the duration of the regulatory board member mandate, and the year of the regulatory board's inceptions--are captured in the data set. For this report's analysis, not all of the aforementioned variables could be used because of data constraints. Only the variables that had sufficient variation across firms were employed, making it 73 Bid, direct adjudication, invitation, petition, or request. 74 Highest cannon, highest price, tariff, lowest government subsidy, investment plan, shorter duration of the concession, or multiple criteria. 75 Operation, BOT, BOO, privatization, and so on. 76 Law, decree, contract, or license. 77 Revenue cap, price cap, rate of return, or no regulation. 78 The government, the concessionaire, both, or nobody. 104 possible to measure the effect of different contract and regulatory characteristics on performance outcomes. Table A2.6. Contract and Regulatory Variables Variable Description Privatization Process Auction Dummy with value 1 if the concession was awarded through an auction process. Award: Highest Price Dummy with value 1 if the concession was awarded according to the highest price. Award: Best Dummy with value 1 if the concession was awarded according to the best Investment Plan investment plan. Regulatory Board Full Autonomy Dummy with value 1 if the regulatory board was fully autonomous. Partial Autonomy Dummy with value 1 if the regulatory board was partially autonomous. Duration Dummy with value 1 if the duration of appointments to the regulatory board was five or more years. Investors Investors: Foreign Dummy with value 1 if the investors were foreign. Investors: Mixed Dummy with value 1 if some the investors were foreign. Tariff Regulation Tariffs: Rate of Return Dummy with value 1 if the tariffs were regulated according to the rate of return. Tariffs: Price Cap Dummy with value 1 if the tariffs were regulated according to price cap. Source: Andres et al (2008c). This database contains annually time series from 1975-2007 for around 100 sets of telecommunication statistics covering telephone network size and dimension, mobile services, quality of service, traffic, staff, tariffs, revenue, and investment. 6. Regulatory Governance In order to assess the governance of electricity regulators in LAC, we designed a survey that was distributed to all electricity regulatory agencies in the region, including not only national but also provincial or state regulators (particularly in the cases of Argentina and Brazil). All LAC countries that are members of the World Bank Group and have an electricity or water regulatory agency were included. The database comprises data from 43 electricity and 28 water regulatory agencies, whose coverage in terms of consumers exceeds 90 percent of the region. Each country was represented by its own regulatory agency, with the exception of Colombia and Chile, for which we assigned unique values since they each have two different entities with regulatory functions. In both Colombia and Chile, for instance, regulatory responsibilities are shared between a National Energy Commission in charge of the main regulatory aspects (tariffs, approval of contracts) and an Oversight Electricity Agency (in the case of Chile, the Superintendencia de Electricidad y Combustibles and in the case of Colombia, the Superintendencia de Servicios Públicos) in charge of the sector's oversight (service quality, sanctions' enforcing, consumer complaints). Considering that both agencies perform different tasks that in other countries are undertaken by only one regulator, the database merged both administrative bodies and assigned a unique value for the country. For those institutional aspects that should be reflected in both agencies, such as the independence of their decision-making (e.g.. the appointment of directors) or the transparency of their management (e.g. account audits), the data assigned the country an average score calculated from both agencies' scores on the same question. For instance, if the 105 Comisión Nacional de Energía of Chile was assigned 0 for not auditing its accounts and the Superintendencia de Electricidad y Combustibles was assigned 1 for auditing its accounts, then Chile would obtain 0.5 for that question. In those aspects where the agencies had separate responsibilities (e.g. the regulation of tariffs by the Comisión Reguladora de la Energía of Colombia and the reception of consumers' claims by the Superintendencia de Servicios Públicos), the data assigned the country the score achieved by the agency with responsibility in that issue, regardless of the score obtained by the other agency for the same issue. The questionnaire is composed of 97 questions (for the full version of the survey, see Andres et al, 2007) reflecting the four variables of agencies' governance and both formal and informal aspects of their functioning. The data also included a general section aimed at capturing characteristics of electricity markets such as the methodology for tariff calculation, the degree of market liberalization, and social tariffs. 7. Corporate Governance of State Owned Enterprises This data was collected through surveys sent to 110 different utilities of the region in both the electricity distribution and water sectors. Final respondents were 45 SOEs. The initiative included both public companies with full state ownership and companies where despite there is private investment state ownership is at least 51 percent of total shares (only a few in this category). This database compresses detailed information on the governance of SOEs in infrastructure through six indexes. The Corporate Governance Index (CGI) is the main index and is the result of the aggregation of the other five. Other indexes include: the Legal Soundness Index, the Board Competitiveness Index, the Professional Management Index, the Performance-Oriented Index, and the Transparency and Disclosure Index. Indexes are composed of different variables representing various aspects of the management of SOEs. Questions were valued between 0 (worst) and 1 (best). In selecting the questions and in giving values the data uses as a main benchmark a public enterprise that is corporatized and subject to same conditions, in terms of access to finance and auditing, than any other private enterprise. The data adjusted the benchmark to sector specificities such as the mechanisms to appoint the Board of Directors, economic regulation, and performance-based orientation. Different from other approaches to the governance of SOEs, it also included the study of the selection, appointment, salary, and educational levels of the staff. Previous approaches have only emphasized the role of the Board and its relationship with the shareholder/s. The data considered that in the infrastructure sector, the role of the staff of a state enterprise is a vital aspect of good management. Because most of these enterprises are not profit- oriented, not allowing to focus on revenues as parameters of good performance, and also because a good bureaucracy is a good filter to political intervention, we believe that a an index that reflects the professionalism (given by educational levels, hiring criteria, and rewards) of the staff might give us a good proxy of the performance of the enterprise. 106 A NNEX 3: B ENCHMARKING A NALYSIS A3.1 E LECTRICITY D ISTRIBUTION A. R EGIONAL B ENCHMARKING A SSESSMENT Table A3.1. Regional Benchmarking ­ Electricity Distribution: Output, Coverage, and Labor Productivity Energy Sold per Connection per year Coverage - Regional Level - - Regional Level - 5.8 .95 5.7 5.6 Percentage MWh/yr .9 5.5 5.4 .85 5.3 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Residential Connections per Employee Energy Sold per Employee - Regional Level - - Regional Level - 700 4000 3500 600 3000 500 2500 400 2000 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Number of Employees Private Sector Participation - Regional Level - - Regional Level - 65000 .6 60000 Share of total connections .5 .4 55000 .3 50000 .2 45000 .1 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 107 Table A3.2. Regional Benchmarking ­ Electricity Distribution: Distributional Losses and Quality of the Service Distributional Losses - Regional Level - .15 .145 .14 .135 .13 1995 2000 2005 year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Frequency of Interruptions per Connection Duration of Interruptions per Connection - Regional Level - - Regional Level - 20 20 18 18 Number Hours 16 16 14 14 12 12 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Table A3.3. Regional Benchmarking ­ Electricity Distribution: Tariffs and Expenses Average Residential Tariffs (dollars/MWh) Average Industrial Tariffs (dollars/MWh) - Regional Level - - Regional Level - 110 90 100 80 90 70 Dollars Dollars 80 60 70 50 60 40 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. OPEX per Connection (in dollars) OPEX per Energy Sold (in dollars) - Regional Level - - Regional Level - 160 34 150 32 30 140 Dollars Dollars 28 130 26 120 24 110 1995 2000 2005 1995 2000 2005 year year Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 108 B. U TILITY -L EVEL B ENCHMARKING A SSESSMENT Table A3.4. Utility Level Benchmarking ­ Electricity Distribution: Coverage, Output, and Labor Productivity Electricity Coverage Energy Sold per Connection per year (MWh) 1 8 .8 6 4 .6 2 .4 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Residential Connections per Employee Energy Sold per employee (MWh) 1500 6000 1000 4000 500 2000 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Table A3.5. Utility Level Benchmarking ­ Electricity Distribution: Distributional Losses and Quality of the Service % Distributional Losses .4 .3 .2 .1 0 1995 2000 2005 year Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 109 Avg Frequency of Interruptions per Connection (#/yr) Avg Duration of Interruptions per Connection (#/yr) 150 150 100 100 50 50 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Table A3.6. Utility Level Benchmarking ­ Electricity Distribution: Tariffs and Expenses Average Residential Tariffs (dollars/MWh) Average Industrial Tariffs (dollars/MWh) 200 140 120 150 100 80 100 60 50 40 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. OPEX per Connection (in dollars) OPEX per MWh sold (in dollars) 800 300 600 200 400 100 200 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. TOTEX per Connection (in dollars) TOTEX per MWh sold (in dollars) 400 1000 300 200 500 100 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 110 A3.2 W ATER AND SANITATION SECTOR A. R EGIONAL-L EVEL B ENCHMARKING A SSESSMENT Table A3.7. Regional Benchmarking ­ Water and Sanitation: Coverage, Output, and Labor Productivity Coverage (in the sample) Coverage (standarized) - Regional Level - - Regional Level - 100 80 90 70 70 % 80 % 60 60 50 50 40 40 1995 2000 2005 1995 2000 2005 year year Water Coverage Sewerage Coverage Water Coverage Sewerage Coverage Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: Joint Monitoring Programme (JMP) data and LAC Water Benchmarking Database, 2009. Number of Connections (in this sample) Number of Connections (standarized) - Regional Level - - Regional Level - 100 120 10 20 30 40 50 60 70 80 Millions Millions 60 40 20 0 0 1995 2000 2005 year 1995 2000 2005 year Total Water Conn's Residential Water Conn's Total Sewerage Conn's Residential Sewerage Conn's Residential Water Conn's Residential Sewerage Conn's Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: Joint Monitoring Programme (JMP) data and LAC Water Benchmarking Database, The World Bank, 2009. Total Water (Cubic meters per year) Water Sold per Inhabitant (liters per day) - Regional Level - - Regional Level - 160 30 25 155 Billions of m^3 20 Liters per day 150 15 10 145 5 140 0 1995 2000 2005 135 year Total Water Produced Total Water Sold 1995 2000 2005 year Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. 111 Table A3.8. Regional Benchmarking ­ Water and Sanitation: Efficiency, Labor Productivity, and Quality of the Service Efficiency Indicators (within this sample) Total Water Connections per Employee - Regional Level - - Regional Level - 100 450 80 400 40 % 60 Ratio 350 20 300 0 1995 2000 2005 250 year Non-Revenue Water (%) Collection Ratio Micrometering 1995 2000 2005 year Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Continuity of the Service (hours per day) Potability (%) - Regional Level - - Regional Level - 100 24 98 23 Hours per day 94 % 96 22 21 92 20 90 1995 2000 2005 1995 2000 2005 year year Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Table A3.9. Regional Benchmarking ­ Water and Sanitation: Tariffs and Expenses Average Residential Tariff (US$/m^3) - Regional Level - .8 .6 US$/m^3 .4 .2 0 1995 2000 2005 year Water Avg Tariff Sewerage Avg Tariff Source: LAC Water Benchmarking Database, The World Bank, 2009. Expenditures per Cubic Meter Sold (US$/m^3) Expenditures per Connection (US$/year) - Regional Level - - Regional Level - 100 120 140 1 .8 US$/m^3 US$/m^3 .6 80 60 .4 40 .2 20 0 0 1995 2000 2005 1995 2000 2005 year year Operational Expenditures Total Expenditures Operational Expenditures Total Expenditures Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. 112 B. U TILITY -L EVEL B ENCHMARKING A SSESSMENT Table A3.10. Utility Level Benchmarking ­ Water and Sanitation: Coverage and Output Water Coverage (%) Sewerage Coverage (%) 100 100 80 90 60 80 40 70 20 60 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Water Sold per Connection (cubic meters per year) Water Sold per Habitant (liters per day) 400 600 500 300 400 200 300 200 100 100 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. 113 Table A3.11. Utility Level Benchmarking ­ Water and Sanitation: Labor Productivity, Efficiency, and Quality of the Service Labor Productivity Non-Revenue Water (%) 800 80 600 60 400 40 200 20 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Collection Ratio Water Connection Micrometered (%) 100 100 90 80 80 60 70 40 60 20 50 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Continuity of the Service (hours per day) Potability (%) 100 24 21 95 18 90 15 85 12 80 9 75 6 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. 114 Table A3.12. Utility Level Benchmarking ­ Water and Sanitation: Tariffs and Expenses Average Residential Water Tariff (US$/m^3) Average Industrial Sewerage Tariff (US$/m^3) 1 1.5 .8 1 .6 .4 .5 .2 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Operational Expenditures per Connection (US$/year) Operational Expenditures per Cubic Meter Sold (US$/m^3) 150 1 100 .5 50 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. Total Expenditures per Connection (US$/year) Total Expenditures per Cubic Meter Sold (US$/m^3) 150 1.5 100 1 50 .5 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Water Benchmarking Database, The World Bank, 2009. Source: LAC Water Benchmarking Database, The World Bank, 2009. 115 A3.3 F IXED T ELECOMMUNICATIONS S ECTOR Households with a Fixed Telephone (%) Subscribers per 100 Inhabitants - Regional Level - - Regional Level - 60 60 50 40 % 30 40 20 20 10 0 1995 1997 1999 2001 2003 2005 2007 0 year 1995 1997 1999 2001 2003 2005 2007 Fixed Lines Mobile Lines year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. Main (fixed) telephone lines in operation Number of Local (fixed) Telephone Minutes - Regional Level - - Regional Level - 350 100 300 80 Billions of minutes Millions of lines 250 60 200 40 150 20 100 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. Number of National (fixed) Long Distance Telephone Minutes Total Minutes per Operational Line (minutes per month) - Regional Level - - Regional Level - 400 80 360 70 Billions of minutes 320 Minutes 60 280 50 240 40 200 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. Staff (Total full-time telecommunications staff) Total Labor Productivity - Regional Level - - Regional Level - 1500 400 Fixed and Mobile cx per employee 1000 350 Thousands 300 500 250 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. 116 % digital main lines Faults per 100 main (fixed) lines per year - Regional Level - - Regional Level - 100 25 90 20 Number of faults 80 15 %70 10 60 5 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. % of telephone faults cleared by next working day Waiting list for main (fixed) lines - Regional Level - - Regional Level - 80 2 1.5 60 Millions of lines 40 % 1 20 .5 0 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. Price of 3-minute local call (off-peak - US$) Price of 3-minute local call (peak - US$) - Regional Level - - Regional Level - .08 .12 .1 .06 .08 US$ US$ .04 .06 .04 .02 .02 0 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. 117 Residential monthly telephone subscription (US$) Business telephone monthly subscription (US$) - Regional Level - - Regional Level - 12 18 10 15 12 8 US$ US$ 6 9 4 6 2 3 0 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. Business telephone connection charge (US$) Residential telephone connection charge (US$) - Regional Level - - Regional Level - 400 200 300 150 200 100 % % 100 50 0 0 1995 1997 1999 2001 2003 2005 2007 1995 1997 1999 2001 2003 2005 2007 year year Source: International Telecommunication Union, 2008. Source: International Telecommunication Union, 2008. A3.4 P UBLIC VS PRIVATE B ENCHMARKING A SSESSMENT Electricity Coverage Electricity Coverage .95 .9 .9 .85 .85 .8 .8 .75 .75 .7 .7 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 118 Energy Sold per Connection per year (MWh) Energy Sold per Connection per year (MWh) 4.6 4.6 4.4 4.4 4.2 4.2 4 4 3.8 3.8 3.6 3.6 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Residential Connections per Employee Residential Connections per Employee 700 600 600 500 500 400 400 300 300 200 200 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Energy Sold per employee (MWh) Energy Sold per employee (MWh) 3000 3000 2500 2500 2000 2000 1500 1500 1000 1000 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. OPEX per Connection (in dollars) OPEX per Connection (in dollars) 350 350 300 300 250 250 200 200 150 150 100 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 119 OPEX per MWh sold (in dollars) OPEX per MWh sold (in dollars) 100 100 80 80 60 60 40 40 20 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Average Residential Tariffs (dollars/MWh) Average Residential Tariffs (dollars/MWh) 110 120 100 100 90 80 80 70 60 60 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Average Industrial Tariffs (dollars/MWh) Average Industrial Tariffs (dollars/MWh) 110 90 100 80 90 80 70 70 60 60 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. % Distributional Losses % Distributional Losses .2 .2 .18 .18 .16 .16 .14 .14 .12 .12 .1 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 120 Avg Frequency of Interruptions per Connection (#/yr) Avg Frequency of Interruptions per Connection (#/yr) 25 25 20 20 15 15 10 10 5 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Avg Duration of Interruptions per Connection (#/yr) Avg Duration of Interruptions per Connection (#/yr) 40 40 35 30 30 25 20 20 15 10 1995 2000 2005 1995 2000 2005 year year Public Utilities Private Utilities Public Utilities Privatized after '95 Privatized before '95 Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. ... TOP T EN and Bottom Ten Percent Performers... Public - Energy Sold per Connection per year (MWh) Private - Energy Sold per Connection per year (MWh) 10 10 8 8 6 6 4 4 2 2 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 121 Public - Residential Connections per Employee Private - Residential Connections per Employee 1500 1500 1000 1000 500 500 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Public - Energy Sold per employee (MWh) Private - Energy Sold per employee (MWh) 6000 6000 4000 4000 2000 2000 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Public - % Distributional Losses Private - % Distributional Losses .5 .5 .4 .4 .3 .3 .2 .2 .1 .1 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Public - Avg Frequency of Interruptions per Connection (#/yr) Private - Avg Frequency of Interruptions per Connection (#/yr) 250 250 200 200 150 150 100 100 50 50 0 0 1995 2000 2005 1995 2000 2005 year year Bottom 10% Mean Top 10% Bottom 10% Mean Top 10% Source: LAC Electricity Benchmarking Database, The World Bank, 2007. Source: LAC Electricity Benchmarking Database, The World Bank, 2007. 122 A NNEX 4: D ETAILED R ESULTS OF THE E MPIRICAL A NALYSIS Table A4.1. Means and Medians Analysis in Levels­­Electricity Distribution T-stat (Z-stat) for difference Variable stats Mean Diff in Levels in means (medians) in Levels Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Residential mean 85.83 102.26 120.48 17.32 17.11 35.16 -16.209*** -17.493*** -16.809*** Connections p50 85.94 102.00 119.59 17.11 16.55 34.33 -7.843*** -7.306*** -7.459*** sd 9.20 2.53 10.04 9.68 8.76 16.94 N 82 116 74 82 74 71 MWH sold mean 82.29 102.67 119.22 20.82 15.60 36.74 -13.119*** -11.882*** -7.554*** per year p50 82.59 101.20 117.13 19.88 15.17 34.60 -7.399*** -6.945*** -6.128*** sd 14.11 6.44 21.12 14.28 17.77 25.69 N 81 116 74 81 74 69 Inputs Number of mean 162.71 100.65 86.59 -61.37 -14.27 -78.19 8.949*** 8.678*** 5.432*** Employees p50 147.46 100.00 86.17 -48.38 -14.76 -63.63 6.252*** 5.903*** 5.057*** sd 54.42 6.76 23.63 52.22 20.18 63.71 N 58 116 59 58 59 50 Efficiency Connections mean 60.24 103.33 147.42 45.38 40.83 88.62 -14.738*** -13.344*** -9.334*** per employee p50 59.90 100.00 135.26 44.65 32.10 88.86 -6.543*** -6.093*** -6.438*** sd 18.65 9.86 42.10 23.25 33.31 46.49 N 57 116 58 57 58 49 GWH per mean 58.56 103.97 145.09 47.50 37.64 86.27 -17.097*** -11.362*** -6.901*** employee p50 59.68 100.00 129.76 46.04 26.76 71.15 -6.567*** -6.093*** -6.182*** sd 18.58 11.98 53.86 20.98 41.54 53.15 N 57 116 58 57 58 49 Distributional mean 112.19 98.73 87.78 -12.92 -9.75 -25.14 3.658*** 4.657*** 3.515*** losses p50 104.37 100.00 85.34 -6.13 -11.06 -19.93 3.268*** 4.272*** 3.341*** sd 26.96 7.33 26.03 27.14 21.12 37.79 N 59 116 58 59 58 49 Quality Duration of mean 134.49 100.34 72.42 -30.61 -25.32 -41.34 3.250*** 2.687*** 3.782*** Interruptions p50 123.37 100.00 65.42 -24.11 -30.41 -34.37 3.477*** 3.143*** 4.019*** per year sd 67.57 20.00 42.58 57.28 41.80 75.35 per consumer N 37 116 39 37 39 24 Frequency of mean 132.59 98.63 82.71 -34.90 -13.65 -31.66 4.256*** 1.300 1.078 Interruptions p50 119.54 100.00 67.96 -21.20 -29.20 -32.86 3.809*** 3.571*** 4.326*** per year sd 57.83 13.77 93.00 49.88 79.05 119.29 per consumer N 37 116 39 37 39 24 Coverage Residential mean 94.93 101.17 110.66 6.93 8.67 16.46 -6.886*** -8.162*** -8.333*** Connections p50 95.35 100.00 108.92 5.60 7.62 14.16 -6.016*** -6.110*** -6.323*** per 100 HHs sd 7.91 2.22 10.09 8.42 8.26 15.09 N 70 116 63 70 63 56 Prices Avg Tariff per mean 106.24 98.48 94.87 -9.49 -2.88 -9.91 3.305*** 2.808*** 1.313* residential GWH p50 97.85 100.00 95.61 -0.09 -1.38 -16.37 2.437** 2.690*** 1.702* (in dollars) sd 23.68 7.52 24.63 23.85 18.73 26.18 N 69 116 73 69 73 55 Avg Tariff per mean 91.77 100.81 109.61 9.21 8.46 17.90 -5.164*** -5.143*** -5.067*** residential GWH p50 88.27 100.00 107.07 15.25 4.64 24.26 -4.774*** -4.181*** -4.643*** (in real local sd 12.83 4.97 18.59 14.81 14.27 25.81 currency) N 69 116 73 69 73 55 Source: Andres et al. (2008). Note: GWh = gigawatt hours; HH = household; MWh = megawatt hours. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 123 Table A4.2. Means and Medians Analysis in Growth­­Electricity Distribution T-stat (Z-stat) for difference Variable stats Avg. Annual Growth Annual Diff in growth in means (medians) in growth Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Residential mean 4.3% 5.5% 3.4% 1.3% -2.8% -0.8% -1.787** 3.590*** 1.976** Connections p50 4.4% 4.7% 3.2% 0.4% -1.7% -1.0% -1.456 5.116*** 2.366** sd 2.6% 5.5% 2.0% N 79 84 60 79 60 56 MWH sold mean 6.7% 6.7% 3.2% -0.5% -5.0% -3.2% 0.616 3.085*** 3.362*** per year p50 6.6% 5.9% 2.8% -0.7% -2.9% -2.7% 0.708 4.096*** 3.159*** sd 4.5% 8.7% 4.7% N 74 85 57 74 57 51 Inputs Number of mean -6.6% -9.9% -2.1% -3.2% 9.7% 2.1% 2.056* -5.398*** -1.519* Employees p50 -6.1% -9.0% -1.8% -3.8% 8.7% 4.0% 2.306** -4.505*** -1.776* sd 8.1% 10.0% 4.8% N 53 69 44 53 44 32 Efficiency Connections mean 13.4% 18.4% 5.5% 4.2% -16.4% -4.2% -1.813** 5.691*** 2.183** per employee p50 11.1% 14.0% 5.6% 4.5% -10.6% -3.5% 2.333** 4.975*** 2.300** sd 12.6% 16.8% 5.1% N 53 66 43 53 43 32 GWH per mean 15.1% 20.3% 5.5% 3.7% -19.9% -6.7% 1.426* 6.539*** 2.826*** employee p50 12.8% 15.0% 4.0% 3.0% -16.4% -6.3% -1.624 5.084*** 3.011*** sd 13.5% 16.9% 7.6% N 53 66 43 53 43 32 Distributional mean 0.6% -5.5% -1.3% -4.7% 6.4% -2.0% 3.301*** -3.474*** 0.960 losses p50 0.1% -4.9% -0.1% -4.5% 6.5% -1.5% 3.317*** -2.944*** 0.786 sd 7.8% 10.2% 9.6% N 57 73 46 57 46 36 Quality Duration of mean 4.1% -9.8% -3.8% -11.2% 3.4% -10.5% 1.788* 4.476*** 5.122*** Interruptions p50 -5.2% -12.9% -3.2% -7.0% 8.5% -5.1% 2.132** -0.749 0.711 per year sd 31.6% 25.7% 24.8% per consumer N 32 51 26 32 26 11 Frequency of mean 2.7% -10.6% -11.4% -11.1% -2.9% -17.8% 1.653* 0.378 3.093*** Interruptions p50 -5.0% -10.8% -6.6% -2.8% -2.4% -14.4% 1.664* -0.165 2.490** per year sd 29.0% 20.3% 20.5% per consumer N 32 51 26 32 26 11 Coverage Residential mean 2.0% 2.2% 1.9% 0.4% -1.0% -0.6% -0.903 1.702** 0.780 Connections p50 1.5% 1.9% 1.3% 0.4% -0.9% -0.3% -1.408 3.186*** 0.619 per 100 HHs sd 3.9% 3.0% 3.6% N 65 76 50 65 50 42 Prices Avg Tariff per mean 9.3% -3.3% 2.0% -15.2% 4.3% -11.4% 6.251*** -'1.821** 3.172*** residential GWH p50 9.7% -6.3% 0.1% -15.1% 1.3% -13.1% 5.329*** -1.442 2.785*** (in dollars) sd 16.0% 9.0% 14.1% N 59 86 57 59 57 35 Avg Tariff per mean 10.2% 2.0% 0.6% -7.8% 0.2% -12.3% 4.744*** -0.172 4.899*** residential GWH p50 5.9% 2.3% 1.8% -5.3% 0.9% -9.7% 4.454*** -0.734 4.063*** (in real local sd 12.6% 7.3% 7.9% currency) N 59 86 56 59 56 35 Source: Andres et al. (2008). Note: GWh = gigawatt hours; HH = household; MWh = megawatt hours. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 124 Table A4.3. Econometric Analysis­­Electricity Distribution (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Number of Energy Sold Number of Connections Energy per Distributional Duration of Frequency of Coverage Avg price per Avg price per Connections per year Employees per employee employee Losses interruptions interruptions MWH MWH (in real (in dollars) local currency) Model 1: Log levels without firm-specific time trend Transition 0.150*** 0.201*** -0.307*** 0.442*** 0.474*** -0.031** -0.144*** -0.107*** 0.053*** -0.013 0.105*** (t>=-1) (0.005) (0.007) (0.016) (0.019) (0.021) (0.013) (0.028) (0.025) (0.004) (0.018) (0.008) Post-transition 0.176*** 0.169*** -0.193*** 0.368*** 0.346*** -0.141*** -0.344*** -0.308*** 0.077*** -0.028*** 0.071*** (t >=2) (0.005) (0.007) (0.016) (0.019) (0.021) (0.013) (0.026) (0.022) (0.004) (0.010) (0.007) Observations 823 808 586 575 570 614 376 377 698 687 685 Model 2: Log levels with firm-specific time trend Transition -0.002 0.040*** -0.054*** 0.049*** 0.086*** 0.021 0.068** 0.076*** -0.007*** 0.078*** 0.034*** (t>=-1) (0.002) (0.005) (0.013) (0.012) (0.017) (0.013) (0.033) (0.029) (0.002) (0.012) (0.008) Post-transition 0.009*** -0.014*** 0.047*** -0.037*** -0.080*** -0.040*** -0.115*** -0.120*** 0.009*** 0.036*** 0.007 (t >=2) (0.002) (0.005) (0.013) (0.013) (0.017) (0.013) (0.031) (0.027) (0.002) (0.009) (0.007) Observations 823 808 586 575 570 614 376 377 698 687 685 Model 3: Growth Transition 0.001 -0.002 -0.050*** 0.048*** 0.046*** -0.042*** -0.063*** -0.050** -0.000 -0.117*** -0.082*** (t>=-1) (0.001) (0.003) (0.008) (0.008) (0.010) (0.010) (0.023) (0.024) (0.001) (0.011) (0.007) Post-transition -0.003*** -0.027*** 0.064*** -0.065*** -0.092*** 0.015 0.001 -0.048** -0.000 0.023*** 0.009 (t >=2) (0.001) (0.003) (0.008) (0.008) (0.010) (0.010) (0.021) (0.021) (0.000) (0.008) (0.006) Observations 803 783 566 557 554 592 339 341 669 633 631 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. MWh = megawatt hours. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 125 Table A4.4. Means and Medians Analysis in Levels­­Fixed Telecommunications T-stat (Z-stat) for difference Variable stats Mean Diff in Levels in means (medians) in levels Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Total number mean 78.98 115.39 181.31 36.41 65.70 102.77 -10.022*** -8.627*** -6.742*** of lines p50 76.93 112.16 178.47 33.90 67.92 93.40 -3.516*** -3.408*** -3.408*** sd 12.55 13.76 48.91 14.53 37.74 46.14 N 16 16 15 16 15 15 Total number mean 107.32 103.05 146.89 0.82 41.13 69.57 -0.049 -3.973* -19.420** of minutes p50 97.39 100.00 146.89 9.05 41.13 69.57 0.105 -1.342 -1.342 sd 41.60 5.04 8.32 40.84 3.00 24.76 N 6 16 2 6 2 2 Inputs Number of mean 117.88 100.72 82.02 -17.12 -18.37 -37.18 2.213** 2.671*** 2.675*** employee p50 111.71 100.28 81.31 -22.64 -20.05 -50.94 1.761* 2.166** 2.291** sd 30.44 7.88 29.61 29.96 25.70 52.09 N 15 16 14 15 14 14 Efficiency Total number mean 72.98 119.54 262.84 47.86 140.97 191.73 -4.972*** -5.262*** -4.957*** of lines per p50 70.13 110.66 217.38 38.93 102.05 154.59 -3.237*** -3.233*** -3.233*** employee sd 24.63 26.54 126.18 37.28 106.41 136.35 N 15 16 14 15 14 14 Total number mean 79.81 105.38 238.94 34.53 123.54 172.50 -2.879** -2.059 -1.486 of minutes per p50 76.03 100.00 238.94 44.60 123.54 172.50 -1.782* -1.342 -1.342 employee sd 22.83 12.63 135.73 29.38 117.59 118.47 N 6 16 2 6 2 2 Percentage of mean 580.77 141.09 101.20 -368.95 -93.78 -472.93 1.050 1.098 1.378 Incomplete Calls p50 111.56 100.00 74.51 -17.23 -27.47 -37.37 1.782* 2.201** 2.366** sd 1133.58 167.34 74.92 860.92 180.06 1055.53 N 6 16 7 6 7 6 Quality Percentage of mean 68.64 116.56 199.92 51.75 81.00 138.97 -4.407*** -2.964*** -2.339** Digitalized p50 70.82 107.27 136.01 41.82 29.26 78.72 -3.180*** -3.180*** -3.129*** Network sd 22.80 31.58 161.58 42.33 129.55 169.03 N 13 16 14 13 14 13 Coverage Number of Lines mean 83.65 113.47 167.28 29.82 53.25 84.53 -7.573*** -7.708*** -6.025*** per 100 HHs p50 80.18 109.18 169.15 28.25 56.28 68.99 -3.516*** -3.408*** -3.351*** sd 12.73 13.75 45.46 15.75 34.23 42.48 N 16 16 15 16 15 15 Prices Avg Price for a mean 144.83 100.45 99.89 -46.64 -1.03 -58.79 0.718 0.710 0.058 3-minute call p50 57.48 99.98 91.72 34.44 -11.25 1.74 -0.866 -0.178 1.255 (in dollars) sd 219.85 15.00 63.61 205.46 61.29 248.59 N 10 16 12 10 12 9 Avg monthly charge mean 55.46 101.25 143.43 39.02 41.60 105.49 -2.983*** -2.083** -1.295 for residential p50 41.00 100.00 120.51 53.32 15.16 43.43 -2.293** -2.073** -0.804 Service (in dollars) sd 36.35 19.28 124.99 41.36 115.87 151.92 N 10 16 13 10 13 9 Avg Charge for the mean 634.94 123.11 100.51 -502.46 -25.83 -256.72 1.814* 0.777 1.122 installation of a p50 95.78 101.06 77.29 11.18 -39.79 8.92 0.051 -0.314 1.376 residential line sd 887.73 40.50 108.31 875.99 72.80 808.89 (in dollars) N 10 16 10 10 10 6 Avg Price for a mean 84.40 100.65 97.58 12.63 -3.46 16.28 -0.711 -0.599 0.250 3-minute call p50 64.40 100.00 87.14 30.96 -14.01 25.78 -0.980 -1.120 1.478 (in real local sd 50.71 7.71 44.03 50.24 43.72 76.87 currency) N 8 16 10 8 10 8 Avg monthly charge mean 60.42 100.26 135.11 36.59 34.54 88.96 -2.782** -2.750** -1.654* for residential p50 49.78 100.00 115.76 49.77 16.83 79.48 -2.191** -2.310** -1.334 Service (in real sd 35.69 12.69 77.55 41.60 69.27 97.05 local currency) N 10 16 11 10 11 9 Avg Charge for the mean 842.23 122.99 132.07 -699.77 1.25 -252.68 1.915** 0.692 -0.028 installation of a p50 108.37 100.00 58.62 -6.06 -31.83 1.91 0.700 -0.105 0.420 residential line (in sd 1045.40 41.81 152.59 1033.62 126.57 894.37 real local currency) N 8 16 8 8 8 6 Source: Andres et al. (2008). Note: HH = household. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 126 Table A4.5. Means and Medians Analysis in Growth­­Fixed Telecommunications T-stat (Z-stat) for difference Variable stats Avg. Annual Growth Annual Diff in growth in means (medians) in growth Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Total number mean 6.9% 12.7% 7.2% 5.8% -6.5% 0.4% -2.546** 1.917** -0.152 of lines p50 7.2% 11.7% 6.6% 3.8% -12.0% -2.1% -2.223** 1.852* -0.157 sd 6.2% 6.3% 8.2% 9.1% 12.8% 10.7% N 16 16 14 16 14 14 Total number mean 4.1% 2.1% 3.8% -6.7% 3.2% -0.8% 1.158 - - of minutes p50 4.6% 1.7% 3.8% -4.1% 3.2% -0.8% 1.219 - - sd 1.9% 15.3% . 12.9% . . N 5 6 1 5 1 1 Inputs Number of mean -0.5% -3.1% -6.9% -2.6% -3.4% -6.5% 0.916 1.258 2.861*** employee p50 -0.8% -4.5% -7.7% -1.5% -1.3% -3.9% 0.909 0.785 2.291** sd 6.9% 9.8% 9.0% 11.1% 10.0% 8.4% N 15 15 14 15 14 14 Efficiency Total number mean 7.8% 17.6% 16.0% 9.8% -3.1% 8.0% -2.452** 0.610 -1.791** of lines per p50 6.6% 21.3% 15.7% 10.9% -9.9% 9.4% -2.101** 0.659 -1.726* employee sd 11.6% 15.3% 11.5% 15.5% 18.9% 16.7% N 15 15 14 15 14 14 Total number mean 5.2% 13.2% 28.6% 5.5% 11.9% 19.1% -3.000** - - of minutes per p50 9.5% 16.3% 28.6% 4.4% 11.9% 19.1% -2.023** - - employee sd 9.6% 11.7% . 4.1% . . N 5 6 1 5 1 1 Percentage of mean -1.5% -16.4% -14.3% -13.9% -0.2% -13.7% 1.293 0.046 2.145** Incomplete Calls p50 -1.5% -7.8% -9.3% -5.1% 0.0% -8.8% 1.363 0.000 2.201** sd 1.0% 23.4% 14.7% 26.4% 14.0% 15.6% N 6 8 7 6 7 6 Quality Percentage of mean 51.5% 17.1% 4.9% -33.1% -13.5% -50.1% 1.085 3.602*** 1.434* Digitalized p50 22.1% 14.2% 0.9% -4.4% -12.0% -11.9% 1.293 2.734*** 2.824*** Network sd 116.3% 15.9% 6.8% 110.1% 13.5% 121.1% N 13 14 13 13 13 12 Coverage Number of Lines mean 4.9% 11.0% 6.0% 6.1% -5.9% 1.2% -3.001*** 2.040** -0.438 per 100 HHs p50 4.4% 9.4% 4.9% 4.5% -8.0% -0.1% -2.637*** 1.852* -0.471 sd 5.9% 6.2% 7.8% 8.1% 10.8% 10.0% N 16 16 14 16 14 14 Prices Avg Price for a mean 46.7% -3.1% -5.7% -44.4% -2.3% -60.8% 1.981** 0.295 1.788* 3-minute call p50 40.9% -1.3% -0.4% -41.4% -7.9% -52.5% 1.820* 0.459 1.572 (in dollars) sd 69.0% 16.8% 12.4% 63.5% 25.1% 83.3% N 8 13 10 8 10 6 Avg monthly charge mean 42.8% 13.9% 5.2% -21.9% -10.5% -45.8% 1.088 0.830 1.785* for residential p50 15.7% 6.0% 0.0% -33.1% -3.3% -28.4% 1.007 0.978 1.272 Service (in dollars) sd 54.6% 31.0% 28.1% 60.4% 41.9% 67.9% N 9 14 11 9 11 7 Avg Charge for the mean -1.9% -14.7% -13.7% -9.6% -5.7% -32.6% 0.785 0.381 1.626 installation of a p50 -1.8% -2.3% -29.3% -5.2% -2.6% -18.2% 1.008 0.533 1.826* residential line sd 25.8% 38.7% 33.7% 36.5% 44.6% 40.1% (in dollars) N 9 14 9 9 9 4 Avg Price for a mean 35.7% -2.5% -0.6% -30.5% 2.7% -36.7% 1.696* -0.389 1.549* 3-minute call p50 44.3% 4.3% 0.6% -32.1% -5.2% -21.2% 1.352 0.178 1.153 (in real local sd 55.4% 19.1% 4.9% 47.6% 21.1% 58.0% currency) N 7 10 9 7 9 6 Avg monthly charge mean 35.6% 16.5% 7.1% -12.7% -9.4% -29.4% 0.721 0.959 1.426 for residential p50 -0.9% 15.6% 3.2% -32.9% -1.9% 0.6% 0.770 0.866 0.676 Service (in real sd 50.1% 32.1% 13.1% 52.9% 30.9% 54.6% local currency) N 9 12 10 9 10 7 Avg Charge for the mean -8.6% -16.1% -11.6% -4.7% -6.7% -19.1% 0.289 0.370 0.789 installation of a p50 -26.3% -20.0% -30.5% -35.0% -2.0% 1.4% 0.000 0.845 -0.365 residential line (in sd 32.3% 46.4% 40.4% 43.5% 48.0% 48.4% real local currency) N 7 10 7 7 7 4 Source: Andres et al. (2008). Note: HH = household. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 127 Table A4.6. Econometric Analysis­­Fixed Telecommunications (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Cost of 3 Monthly Connection Cost of 3 Monthly Connection Number of Number of Number of Connections Minutes per Incomplete Network minute local charge charge minute local charge charge connections minutes employees per worker worker calls digitization Coverage call (dollars) (dollars) (dollars) call (r.l.c.) (r.l.c.) (r.l.c.) Model 1: Log levels without firm-specific time trend Transition 0.253*** 0.079** -0.097*** 0.301*** 0.278*** -0.133 0.310*** 0.168*** 0.384*** 0.565*** 0.095 0.371*** 0.486*** -0.178 (t>=-1) (0.030) (0.035) (0.033) (0.054) (0.059) (0.083) (0.053) (0.025) (0.080) (0.118) (0.114) (0.081) (0.113) (0.171) Post-transition 0.494*** 0.319*** -0.264*** 0.727*** 0.657*** -0.353*** 0.458*** 0.421*** -0.014 0.209*** -0.310*** -0.090 0.197** -0.286* (t >=2) (0.028) (0.032) (0.033) (0.054) (0.084) (0.057) (0.046) (0.026) (0.053) (0.049) (0.108) (0.063) (0.086) (0.153) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 2: Log levels with firm-specific time trend Transition -0.050** 0.002 0.031 -0.101*** -0.010 0.142*** 0.048** -0.065*** 0.523*** 0.281*** 0.300*** 0.358*** 0.067 0.118 (t>=-1) (0.024) (0.038) (0.026) (0.038) (0.044) (0.042) (0.024) (0.019) (0.104) (0.100) (0.063) (0.082) (0.092) (0.154) Post-transition 0.113*** 0.133*** -0.069** 0.185*** 0.173*** 0.006 0.024 0.091*** 0.051 -0.067 0.222*** -0.168** -0.099 0.244** (t >=2) (0.025) (0.041) (0.027) (0.041) (0.060) (0.044) (0.026) (0.021) (0.091) (0.087) (0.082) (0.082) (0.080) (0.097) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 3: Growth Transition 0.027** 0.069*** -0.041*** 0.070*** 0.085** -0.062 -0.008 0.037*** -0.052 -0.101 -0.003 -0.056 -0.047 -0.140 (t>=-1) (0.011) (0.012) (0.015) (0.021) (0.042) (0.041) (0.026) (0.010) (0.077) (0.097) (0.048) (0.065) (0.067) (0.107) Post-transition -0.002 0.053* -0.026* 0.033* 0.083 -0.035 -0.056*** 0.001 0.019 -0.034 -0.019 -0.025 0.001 0.036 (t >=2) (0.010) (0.031) (0.015) (0.020) (0.052) (0.028) (0.022) (0.010) (0.048) (0.056) (0.056) (0.046) (0.059) (0.073) Observations 165 60 158 158 59 64 122 162 93 105 98 82 102 79 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. r.l.c. = real local currency. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 128 Table A4.7. Econometric Analysis­­Fixed Telecommunications, Liberalization (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Cost of 3 Monthly Connection Cost of 3 Monthly Connection Number of Number of Number of Connections Minutes per Incomplete Network minute local charge charge minute local charge charge connections minutes employees per worker worker calls digitization Coverage call (dollars) (dollars) (dollars) call (r.l.c.) (r.l.c.) (r.l.c.) Model 1: Log levels without firm-specific time trend Transition 0.232*** 0.064* -0.046 0.272*** 0.232*** -0.140* 0.307*** 0.166*** 0.422*** 0.558*** 0.033 0.359*** 0.398*** -0.107 (t>=-1) (0.027) (0.036) (0.030) (0.049) (0.050) (0.081) (0.057) (0.025) (0.088) (0.131) (0.073) (0.085) (0.112) (0.191) Post-transition 0.432*** 0.279*** -0.151*** 0.602*** 0.432*** -0.335*** 0.446*** 0.364*** 0.011 0.220*** -0.151* -0.162** 0.102 -0.131 (t >=2) (0.028) (0.043) (0.031) (0.051) (0.078) (0.076) (0.055) (0.025) (0.057) (0.058) (0.083) (0.073) (0.086) (0.163) Liberalization 0.275*** 0.065 -0.361*** 0.673*** 0.487*** -0.027 0.023 0.230*** -0.097 0.001 -0.491*** 0.150* 0.443*** -0.529** dummy (0.037) (0.046) (0.047) (0.083) (0.082) (0.088) (0.069) (0.035) (0.088) (0.144) (0.171) (0.091) (0.155) (0.221) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 2: Log levels with firm-specific time trend Transition -0.050** 0.001 0.026 -0.089** -0.006 0.133*** 0.044* -0.066*** 0.441*** 0.192 0.245*** 0.296*** -0.007 0.130 (t>=-1) (0.024) (0.043) (0.026) (0.038) (0.049) (0.043) (0.025) (0.020) (0.109) (0.136) (0.083) (0.082) (0.087) (0.165) Post-transition 0.116*** 0.127*** -0.066** 0.192*** 0.164*** 0.009 0.023 0.091*** -0.011 -0.111 0.197** -0.193** -0.135* 0.246** (t >=2) (0.025) (0.041) (0.027) (0.041) (0.060) (0.043) (0.026) (0.021) (0.091) (0.093) (0.081) (0.078) (0.076) (0.097) Liberalization 0.002 0.037 -0.046 0.117** 0.108 -0.041 -0.016 -0.007 -0.356*** -0.410*** -0.030 -0.240*** -0.500*** 0.035 dummy (0.032) (0.063) (0.042) (0.049) (0.090) (0.053) (0.028) (0.025) (0.116) (0.147) (0.092) (0.090) (0.136) (0.169) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 3: Growth Transition 0.028** 0.066*** -0.041*** 0.075*** 0.073* -0.059 0.006 0.036*** 0.006 0.072 -0.021 -0.038 -0.004 -0.253* (t>=-1) (0.011) (0.013) (0.015) (0.020) (0.040) (0.040) (0.028) (0.011) (0.077) (0.095) (0.066) (0.065) (0.047) (0.138) Post-transition 0.010 0.030 -0.027* 0.047** -0.006 -0.011 -0.046* 0.008 0.142*** 0.038 -0.022 0.012 0.053 0.003 (t >=2) (0.011) (0.041) (0.016) (0.021) (0.058) (0.033) (0.025) (0.010) (0.053) (0.059) (0.067) (0.053) (0.049) (0.085) Liberalization -0.053*** 0.037 0.007 -0.075** 0.183*** -0.037 -0.044 -0.027 -0.451*** -0.428*** 0.002 -0.161** -0.387*** 0.251* dummy (0.019) (0.039) (0.029) (0.034) (0.067) (0.039) (0.031) (0.017) (0.080) (0.111) (0.098) (0.070) (0.108) (0.132) Observations 165 60 158 158 59 64 122 162 93 105 98 82 102 79 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. The Liberalization dummy = 1 for those years that the long-distance telecommunications market was liberalized. r.l.c. = real local currency. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 129 Table A4.8. Econometric Analysis­­Fixed Telecommunications, Mobile Competition (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Cost of 3 Monthly Connection Cost of 3 Monthly Connection Number of Number of Number of Connections Minutes per Incomplete Network minute local charge charge minute local charge charge connections minutes employees per worker worker calls digitization Coverage call (dollars) (dollars) (dollars) call (r.l.c.) (r.l.c.) (r.l.c.) Model 1: Log levels without firm-specific time trend Transition 0.247*** 0.047 -0.059** 0.291*** 0.178*** -0.143* 0.313*** 0.171*** 0.432*** 0.506*** -0.030 0.311*** 0.365*** -0.165 (t>=-1) (0.027) (0.037) (0.027) (0.043) (0.050) (0.077) (0.053) (0.022) (0.079) (0.120) (0.021) (0.075) (0.102) (0.106) Post-transition 0.413*** 0.221*** -0.089*** 0.500*** 0.269*** -0.337*** 0.442*** 0.342*** 0.038 0.189*** 0.032 -0.221*** 0.003 0.031 (t >=2) (0.027) (0.050) (0.030) (0.046) (0.085) (0.089) (0.053) (0.025) (0.053) (0.046) (0.030) (0.067) (0.077) (0.110) Mobile subs. 0.013*** 0.005** -0.025*** 0.037*** 0.030*** -0.000 0.001 0.014*** -0.015*** 0.013 -0.151*** 0.017*** 0.042*** -0.132*** (0.002) (0.002) (0.001) (0.002) (0.004) (0.004) (0.003) (0.002) (0.006) (0.010) (0.017) (0.004) (0.009) (0.017) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 2: Log levels with firm-specific time trend Transition -0.064*** 0.019 0.008 -0.070* 0.029 0.111** 0.017 -0.068*** 0.166*** -0.056 0.327*** 0.201*** -0.043 0.349** (t>=-1) (0.025) (0.051) (0.025) (0.039) (0.063) (0.045) (0.022) (0.021) (0.063) (0.105) (0.073) (0.047) (0.044) (0.161) Post-transition 0.120*** 0.112*** -0.044* 0.176*** 0.061 0.022 0.042* 0.099*** 0.293*** 0.055 0.195** 0.083* -0.005 0.225** (t >=2) (0.025) (0.034) (0.026) (0.041) (0.048) (0.046) (0.023) (0.022) (0.060) (0.061) (0.088) (0.049) (0.041) (0.090) Mobile subs. -0.006* 0.010** -0.017*** 0.010** 0.032*** -0.004 -0.021*** -0.003 -0.117*** -0.148*** 0.039* -0.063*** -0.105*** 0.076*** (0.003) (0.005) (0.003) (0.005) (0.006) (0.005) (0.003) (0.003) (0.007) (0.015) (0.024) (0.005) (0.011) (0.025) Observations 168 71 161 162 69 70 131 165 104 114 107 91 110 87 Model 3: Growth Transition 0.023** 0.068*** -0.043*** 0.068*** 0.075* -0.062 0.006 0.035*** -0.005 -0.076 -0.031 -0.023 -0.043 -0.175* (t>=-1) (0.011) (0.014) (0.015) (0.021) (0.040) (0.042) (0.025) (0.011) (0.063) (0.090) (0.054) (0.059) (0.059) (0.093) Post-transition 0.011 0.068 -0.017 0.039* -0.004 -0.033 -0.030 0.004 0.117*** 0.051 -0.063 0.051 0.071 -0.039 (t >=2) (0.011) (0.053) (0.016) (0.022) (0.064) (0.040) (0.024) (0.011) (0.042) (0.062) (0.060) (0.047) (0.056) (0.076) Mobile subs. -0.002** -0.001 -0.002 -0.001 0.006* -0.000 -0.005*** -0.001 -0.026*** -0.032*** 0.018* -0.014*** -0.025*** 0.028*** (0.001) (0.002) (0.002) (0.002) (0.003) (0.002) (0.001) (0.001) (0.004) (0.008) (0.011) (0.004) (0.007) (0.011) Observations 165 60 158 158 59 64 122 162 93 105 98 82 102 79 Number of firms 16 11 16 16 11 8 14 16 12 13 13 11 13 11 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. Mobile subs. is an independent variable measuring millions of mobile subscribers. r.l.c. = real local currency. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 130 Table A4.9. Econometric Analysis­­Fixed Telecommunications, Instrumental Variables (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Cost of 3 Monthly Connection Cost of 3 Monthly Connection Number of Number of Number of Connections Minutes per Incomplete Network minute local charge charge minute local charge charge connections minutes employees per worker worker calls digitization Coverage call (dollars) (dollars) (dollars) call (r.l.c.) (r.l.c.) (r.l.c.) Model 1: Log levels without firm-specific time trend Transition 0.462*** 0.326*** -0.198*** 0.646*** 0.717*** -0.086 0.490*** 0.377*** 0.877*** 1.041*** -0.692** 0.754*** 0.910*** -1.060*** (t>=-1) (0.052) (0.109) (0.070) (0.111) (0.135) (0.079) (0.105) (0.046) (0.147) (0.221) (0.300) (0.136) (0.209) (0.355) Post-transition 0.436*** 0.364*** -0.222*** 0.674*** 0.724*** -0.262*** 0.363*** 0.371*** -0.069 0.331* -0.204 0.012 0.332** 0.035 (t >=2) (0.043) (0.097) (0.059) (0.094) (0.120) (0.060) (0.084) (0.039) (0.111) (0.174) (0.260) (0.097) (0.163) (0.283) Observations 121 54 114 115 52 42 107 120 79 90 93 71 90 77 Model 2: Log levels with firm-specific time trend Transition 0.003 0.229* 0.160* -0.126 0.204 0.109** 0.129 0.027 1.370*** 0.982*** 0.912*** 0.837*** 0.507 0.862** (t>=-1) (0.063) (0.134) (0.087) (0.103) (0.153) (0.042) (0.199) (0.060) (0.278) (0.350) (0.309) (0.213) (0.304) (0.375) Post-transition 0.115** 0.114 0.057 0.095 0.173 -0.018 0.014 0.108** 0.099 -0.147 0.593*** -0.022 -0.209 0.723** (t >=2) (0.046) (0.138) (0.064) (0.077) (0.151) (0.042) (0.150) (0.045) (0.226) (0.264) (0.220) (0.176) (0.213) (0.271) Observations 121 54 114 115 52 42 107 120 79 90 93 71 90 77 Model 3: Growth Transition 0.035 0.056 -0.024 0.062 0.084 -0.049 0.243* 0.050** -0.559*** -0.477*** -0.197 -0.470*** -0.313** -0.095 (t>=-1) (0.024) (0.141) (0.031) (0.038) (0.152) (0.046) (0.124) (0.022) (0.170) (0.173) (0.144) (0.151) (0.150) (0.202) Post-transition -0.028 -0.049 -0.054** 0.023 -0.037 -0.036 -0.146* -0.038** -0.147 -0.116 0.043 -0.088 -0.088 0.046 (t >=2) (0.019) (0.113) (0.025) (0.030) (0.123) (0.028) (0.087) (0.018) (0.107) (0.118) (0.111) (0.085) (0.103) (0.140) Observations 118 45 111 112 44 37 101 117 72 84 87 64 84 71 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. r.l.c. = real local currency. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 131 Table A4.10. Means and Medians Analysis in Levels­­Water and Sewerage T-stat (Z-stat) for difference Variable stats Mean Diff in Levels in means (medians) in levels Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Residential Water mean 85.85 103.15 119.74 16.20 16.31 29.43 -10.988*** -8.762*** -12.059*** Connections p50 87.37 102.61 117.09 15.18 13.88 28.10 -4.197*** -5.086*** -3.724*** sd 6.32 3.72 13.17 7.07 10.85 10.35 N 23 49 34 23 34 18 Residential Sewer mean 84.88 102.75 122.59 18.83 19.43 32.90 -7.932*** -8.950*** -9.735*** Connections p50 85.48 101.89 119.62 18.62 17.46 29.38 -3.883*** -4.937*** -3.408*** sd 11.21 5.02 15.08 10.62 12.28 13.09 N 20 49 32 20 32 15 Cubic Meter of mean 99.98 103.62 97.27 2.21 -2.91 -1.33 -0.745 1.416* 0.299 produced water p50 100.99 100.00 99.04 1.95 -0.72 3.15 -0.879 1.078 -0.973 sd 8.89 22.20 14.80 11.88 11.45 16.60 N 16 49 31 16 31 14 Inputs Number of mean 141.43 103.97 92.35 -37.20 -12.18 -57.36 3.961*** 3.668*** 4.766*** Employees p50 125.11 100.00 97.04 -21.34 -8.36 -52.01 3.527*** 3.339*** 3.237*** sd 49.22 14.22 23.85 38.72 17.26 46.62 N 17 49 27 17 27 15 Efficiency Water Connections mean 70.50 103.34 144.11 36.53 38.73 83.86 -9.979*** -4.201*** -5.177*** per employee p50 68.46 100.00 125.05 36.39 20.71 69.30 -3.621*** -4.532*** -3.408*** sd 18.93 12.65 59.84 15.09 48.79 62.73 N 17 49 28 17 28 15 Distributional mean 107.22 100.02 82.08 -8.70 -18.26 -23.18 2.577** 3.755*** 3.110*** losses p50 106.01 100.00 81.64 -8.33 -16.63 -20.12 2.327** 3.254*** 2.605*** sd 16.43 7.42 21.22 13.51 23.33 27.88 N 16 49 23 16 23 14 Quality Continuity mean 78.34 101.01 116.79 21.81 14.94 21.66 -1.781* -2.748*** -1.330 (hs per day) p50 97.11 100.00 104.35 2.48 2.17 4.05 -2.192** -2.774*** -1.971** sd 37.52 4.68 24.68 36.74 21.06 46.07 N 9 49 15 9 15 8 % of the samples mean 88.35 100.30 103.89 11.55 2.58 4.94 -1.250 -2.088** -1.682* that passed the p50 99.50 100.00 100.51 0.58 0.46 1.08 -1.630 -2.603*** -1.941* potability test sd 27.92 1.53 6.87 26.14 4.62 7.20 N 8 49 14 8 14 6 Coverage Residential Water mean 94.25 101.84 111.12 6.52 8.71 10.37 -4.498*** -4.379*** -4.478*** Connections p50 95.13 100.00 106.88 4.86 5.26 8.76 -4.107*** -4.584*** -3.823*** per 100 HHs sd 5.70 3.96 14.11 6.80 10.71 10.10 N 22 49 29 22 29 19 Residential Sewer mean 91.47 101.77 110.03 10.23 8.67 13.59 -4.539*** -3.981*** -5.277*** Connections p50 91.72 100.00 106.87 8.02 5.76 8.98 -3.479*** -3.920*** -3.180*** per 100 HHs sd 8.76 6.88 11.55 9.29 9.74 9.29 N 17 49 20 17 20 13 Prices Avg price per cub. mean 93.62 101.39 106.70 10.43 1.46 40.24 -0.635 -0.173 -2.261** meter of water p50 87.95 100.00 98.60 11.81 3.27 32.70 -1.274 -0.314 -2.240** (in dollars) sd 43.54 9.53 37.16 51.89 30.57 50.34 N 10 49 13 10 13 8 Avg price per cub. mean 84.00 103.53 130.09 25.70 17.68 57.87 -2.478** -2.903*** -4.150*** meter of water p50 82.76 100.00 121.21 22.22 19.65 44.80 -1.988** -0.411** -2.521** (in real local sd 23.18 11.71 32.81 32.80 21.96 39.44 currency) N 10 49 13 10 13 8 Avg price per cub. mean 114.61 100.53 107.79 -19.43 0.03 44.29 0.375 0.001 -0.835 meter of sewer p50 79.43 100.00 107.68 16.46 -12.60 44.29 0.000 0.365 -0.447 (in dollars) sd 89.74 6.94 32.73 89.77 35.56 75.05 N 3 49 4 3 4 2 Avg price per cub. mean 93.06 101.80 152.44 13.26 32.25 53.34 -0.512 -3.012** -37.266*** meter of sewer p50 74.75 100.00 135.93 30.91 33.12 53.34 -0.535 -1.826* -1.342 (in real local sd 45.93 10.88 51.26 44.86 21.42 2.02 currency) N 3 49 4 3 4 2 Source: Andres et al. (2008). Note: HH = household. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 132 Table A4.11. Means and Medians Analysis in Growth­­Water and Sewerage T-stat (Z-stat) for difference Variable stats Avg. Annual Growth Annual Diff in growth in means (medians) in growth Preprivat Transition Postprivat (2)-(1) (3)-(2) (3)-(1) (2)-(1) (3)-(2) (3)-(1) (1) (2) (3) (4) (5) (6) (7) (8) (9) Outputs Residential Water mean 4.4% 6.5% 4.7% 0.9% -1.9% 1.5% -1.095 1.649* -1.113 Connections p50 4.1% 5.2% 3.8% -0.1% -1.8% 1.2% -0.923 2.229** -0.943 sd 3.0% 4.4% 4.6% 3.5% 5.6% 3.2% N 17 43 24 17 24 6 Residential Sewer mean 3.8% 6.7% 7.4% 3.1% 1.5% 0.0% -1.222 -0.569 0.009 Connections p50 4.3% 5.5% 3.6% 2.1% -1.4% 0.1% -0.966 0.693 -0.135 sd 5.9% 6.8% 10.7% 9.8% 12.3% 3.2% N 15 40 23 15 23 5 Cubic Meter of mean 2.1% 7.5% 0.5% -0.9% -1.8% 1.6% 0.741 1.117 -0.718 produced water p50 1.6% 1.0% 0.9% 0.0% 0.0% 1.5% 0.000 0.817 -0.674 sd 4.6% 38.6% 5.0% 4.1% 7.3% 5.0% N 12 38 21 12 21 5 Inputs Number of mean -0.4% -10.0% -1.5% -9.6% 7.5% -1.0% 3.425*** -3.460*** 0.309 Employees p50 0.1% -8.3% -1.0% -9.8% 7.8% -1.4% 2.432*** -2.765*** 0.135 sd 4.2% 10.2% 7.2% 9.7% 9.2% 7.4% N 12 32 18 12 18 5 Efficiency Water Connections mean 5.5% 17.5% 7.3% 11.6% -9.6% 1.2% -3.068*** 2.939*** -0.348 per employee p50 4.9% 15.8% 4.5% 9.9% -7.8% 0.1% 2.551** 2.656 0.105 sd 5.4% 13.5% 10.1% 13.7% 14.3% 8.3% N 13 32 19 13 19 6 Distributional mean -3.1% -0.6% -5.5% 0.5% 0.5% 0.6% -0.297 -0.310 -0.363 losses p50 -2.6% -2.0% -5.1% -0.1% 0.3% 0.8% -0.267 -0.450 -0.843 sd 3.8% 21.5% 9.1% 5.3% 6.2% 4.0% N 11 26 17 11 17 6 Quality Continuity mean 0.0% 7.2% 4.6% 22.4% -0.1% 0.0% -1.000 0.057 - (hs per day) p50 0.0% 0.0% 0.9% 0.0% 0.0% 0.0% -1.000 0.075 - sd 0.0% 16.0% 8.7% 38.7% 6.0% . N 3 18 11 3 11 1 % of the samples mean 0.8% 5.2% 0.4% 18.6% -0.5% -1.0% -1.074 1.273 1.000 that passed the p50 0.6% 0.2% 0.0% 2.2% 0.0% -1.0% -0.928 1.315 1.000 potability test sd 1.0% 16.4% 0.7% 34.6% 1.2% 1.4% N 4 18 9 4 9 2 Coverage Residential Water mean 1.0% 4.1% 3.3% 1.1% -1.3% 0.4% -2.050** 0.914 -0.570 Connections p50 0.3% 2.8% 1.6% 0.2% -1.3% 0.1% -1.448 1.690* -0.944 per 100 HHs sd 1.7% 5.0% 4.4% 2.1% 6.1% 1.7% N 16 34 19 16 19 5 Residential Sewer mean 1.6% 8.0% 2.8% 2.9% -0.9% -1.6% -1.815 0.529 2.735** Connections p50 1.4% 2.9% 0.6% 0.1% -1.6% -0.9% -1.036 1.601 2.023** per 100 HHs sd 17.9% 17.9% 6.1% 6.0% 6.2% 1.3% N 14 25 14 14 14 5 Prices Avg price per cub. mean 12.2% 1.9% -3.4% -12.1% -7.2% -3.9% 2.493** 0.835 0.666 meter of water p50 10.9% -2.2% -1.1% -13.8% -3.3% -2.1% 1.820* 0.889 0.535 (in dollars) sd 10.4% 22.2% 20.0% 13.8% 26.0% 10.1% N 8 17 9 8 9 3 Avg price per cub. mean 10.1% 9.4% 4.5% -6.0% -8.9% -0.8% 2.078** 1.060 0.346 meter of water p50 10.1% 5.4% 2.6% -4.3% -6.5% -2.5% 1.540 1.007 0.000 (in real local sd 6.7% 18.4% 10.0% 8.1% 25.1% 4.0% currency) N 8 17 9 8 9 3 Avg price per cub. mean -0.6% -5.1% -7.9% 2.3% -6.4% -7.7% -0.298 0.799 - meter of sewer p50 -0.6% -8.7% -7.9% 2.3% -10.8% -7.7% -0.447 1.069 - (in dollars) sd 17.1% 16.1% 11.6% 10.8% 13.9% . N 2 5 3 2 3 1 Avg price per cub. mean -1.1% 7.0% 9.7% 5.0% -4.3% -15.1% 3.881* 0.302 - meter of sewer p50 -1.1% 1.4% 9.8% 5.0% -18.4% -15.1% -1.342 0.000 - (in real local sd 13.9% 13.5% 16.0% 1.8% 24.7% . currency) N 2 5 3 2 3 1 Source: Andres et al. (2008). Note: HH = household. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 133 Table A4.12. Econometric Analysis­­Water Distribution and Sewerage (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Number of Number of Cubic Meters Number of Water Distributional Continuity of Potability Water Sewerage Avg price per Avg price per Avg price per Avg price per Water Sewerage per year Employees Connec. per losses the Service Coverage Coverage M3 of water M3 of water M3 for M3 for Connect's Connect's Employee (in dollars) (in R.L.C.) sewerage (in sewerage (in dollars) R.L.C.) Model 1: Log levels without firm-specific time trend Transition 0.141*** 0.174*** 0.040*** -0.180*** 0.268*** -0.039** 0.038 0.059* 0.025*** 0.053*** 0.055 0.146*** -0.014 0.104 (t>=-1) (0.010) (0.016) (0.009) (0.030) (0.034) (0.017) (0.064) (0.034) (0.007) (0.009) (0.041) (0.026) (0.142) (0.083) Post-transition 0.139*** 0.173*** 0.015*** -0.194*** 0.354*** -0.155*** 0.074*** 0.012** 0.049*** 0.065*** 0.097** 0.213*** -0.096 0.222*** (t >=2) (0.008) (0.011) (0.006) (0.024) (0.027) (0.015) (0.015) (0.005) (0.005) (0.007) (0.038) (0.027) (0.110) (0.077) Observations 259 239 195 201 199 179 97 90 243 198 112 112 37 37 Model 2: Log levels with firm-specific time trend Transition 0.006 -0.006 -0.007 0.083*** -0.076*** -0.014 0.000 -0.002 -0.000 -0.005 0.003 -0.048 0.026 0.017 (t>=-1) (0.004) (0.009) (0.010) (0.026) (0.023) (0.012) (0.006) (0.005) (0.001) (0.006) (0.050) (0.034) (0.093) (0.082) Post-transition -0.002 -0.005 -0.013* 0.069*** -0.027 0.000 0.000 -0.002 -0.001 -0.008 -0.047 -0.024 0.013 0.045 (t >=2) (0.003) (0.005) (0.007) (0.017) (0.019) (0.001) (0.002) (0.009) (0.001) (0.005) (0.031) (0.020) (0.088) (0.078) Observations 259 239 195 201 199 179 97 90 243 198 112 112 37 37 Model 3: Growth Transition 0.001 0.006 -0.008 -0.048*** 0.047*** -0.000 0.002 0.009 0.001 0.003 -0.203*** -0.099*** -0.054 0.007 (t>=-1) (0.004) (0.006) (0.009) (0.018) (0.018) (0.012) (0.020) (0.013) (0.002) (0.004) (0.034) (0.027) (0.080) (0.059) Post-transition -0.010*** -0.011*** -0.025*** 0.048*** -0.037*** -0.012* -0.001 -0.005 -0.004*** -0.008** -0.018 -0.011 -0.005 0.006 (t >=2) (0.002) (0.002) (0.007) (0.012) (0.012) (0.007) (0.005) (0.005) (0.002) (0.004) (0.021) (0.019) (0.076) (0.065) Observations 235 216 172 176 178 160 81 77 217 180 101 101 31 31 Source: Andres et al. (2008). Note: Standard errors are in parentheses. The Transition and Post-transition variables are dummy independent variables in regressions where the dependent variable is given by the column heading (Number of Connections). Transition = 1 starting two years before the privatization or concession was awarded and continuing for all years after. Post-transition = 1 for all years after the transition period, that is, starting one year after the privatization was awarded. r.l.c. = real local currency. * significant at 10 percent; ** significant at 5 percent; *** significant at 1 percent. 134 A NNEX 5: R EGULATORY G OVERNANCE D IMENSIONS Figure A5.1 Electricity Regulatory Agencies Regulatory Autonomy Managerial Autonomy 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BOL T1 T2 COL JAM ESV MEX GUA BRA ARG T&T PAN NIC BAR RD PER CR ECU URU HON BOL T1 T2 COL JAM ESV MEX GUA BRA ARG T&T PER BAR RD PAN NIC ECU CR URU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Political Autonomy 1 .8 Index (0-1) .6 .4 .2 0 T1 BOL T2 COL JAM BRA ESV MEX GUA ARG T&T RD NIC PER CR URU PAN HON BAR ECU Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Institutional Transparency Social Transparency 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 T1 BOL T2 COL MEX JAM BRA ESV GUA T&T ARG PER BAR CR URU RD NIC PAN HON ECU T1 BOL T2 COL ESV MEX JAM GUA BRA T&T ARG PER URU PAN ECU HON BAR RD NIC CR Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. 135 Regulatory Tools Institutional Tools 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BOL T1 T2 COL JAM GUA BRA ESV MEX ARG T&T PER URU NIC CR RD BAR PAN ECU HON T1 T2 COL BOL JAM ESV MEX GUA BRA T&T ARG PER URU CR BAR PAN HON RD NIC ECU Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Formal Autonomy Informal Autonomy 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BOL T1 T2 COL MEX JAM GUA BRA ESV ARG T&T URU NIC RD CR PER BAR PAN HON ECU BOL T1 T2 COL JAM ESV MEX BRA GUA ARG T&T PER RD NIC PAN BAR ECU CR URU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Formal Accountability Informal Accountability 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 T1 T2 T1 T2 COL BOL BOL COL MEX JAM MEX JAM ESV BRA GUA GUA BRA ESV T&T ARG T&T ARG PER BAR PAN ECU HON URU URU BAR PER NIC CR RD RD NIC CR ECU HON PAN Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Formal Transparency Informal Transparency 1 1 .8 .8 Index (0-1) Index (0-1) .6 .6 .4 .4 .2 .2 0 0 BOL T1 T2 COL ESV MEX JAM GUA BRA T&T ARG PER RD BAR URU CR HON NIC PAN ECU T1 BOL T2 COL MEX JAM ESV BRA GUA T&T ARG PER URU BAR PAN CR NIC RD ECU HON Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. Source: LAC Electricity Regulatory Governance Database, The World Bank, 2008. 136 Index (0-1) Index (0-1) Index (0-1) .2 .4 .6 .8 .2 .4 .6 .8 .2 .4 .6 .8 0 1 0 1 0 1 CO - CRA BZ - ARCE PE- SUNASS BZ - ARCE TT - RIC BZ - AGERSA T1 BZ - AGR BZ - ATR BZ - AGERGS PE- SUNASS AR - ERSPyOC TT - RIC T1 CR - ERSAPS PE- SUNASS BZ - ARSAL BZ - AMAE AR-ERAS CO - CRA TT - RIC HN - ERSAPS BZ - AGENERSA PA - ASEP BZ - AGERSA BZ - AGER Index (0-1) BZ - ARSAL CR - ERSAPS BZ - ATR AR - ERSAC BZ - AGER BZ - AGERGS BZ - ADASA BZ - AGENERSA CR - ERSAPS PY - ERSSAN BZ - ARSAM BA - FTC BZ - ARCE BZ - ATR BZ - ADASA BZ - AGER BZ - ARSAL T2 .2 .4 .6 .8 BZ - AGESC BZ - ADASA BZ - AGERSA 0 1 T1 BZ - AMAE HN - ERSAPS AR - ENRESS BZ - AGR PY - ERSSAN BZ - ADASA BZ - AGENERSA Institutional Tools BZ - AGESC BZ - ARSAM T1 BA - FTC Regulatory Autonomy BA - FTC BZ - AMAE PE- SUNASS T2 Institutional Transparency T2 BZ - ARSAE CR - ERSAPS HN - ERSAPS Figure A5.2 Water Regulatory Agencies PA - ASEP AR-ERAS BZ - AGENERSA BZ - ARSAE Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. AR - ENRESS BZ - AGESC T2 BZ - AGR PY - ERSSAN AR - ERSAC BZ - ARSAL CO - CRA AR - ERSAC PA - ASEP BZ - AGERGS BZ - AGERGS BZ - ARSAE Index (0-1) AR - ENRESS Index (0-1) BZ - AGER BZ - ARSAM Index (0-1) AR - ERSPyOC AR - ERSPyOC PY - ERSSAN AR - ERSACT AR - ERSACT AR - ERSACT HN - ERSAPS AR-ERAS BZ - ARCE AR - ENRESS BZ - AGR 137 .2 .4 .6 .8 .2 .4 .6 .8 BZ - AGESC .2 .4 .6 .8 0 1 0 1 BZ - AMAE 0 1 BZ - AGERSA BZ - AGERSA BZ - AGERSA AR - ERSPyOC BZ - ARCE Political Autonomy BZ - AGERGS AR-ERAS TT - RIC TT - RIC BZ - AGENERSA T1 BZ - ARSAM HN - ERSAPS PE- SUNASS PE- SUNASS BA - FTC BZ - AGER T1 CO - CRA BZ - ARSAE AR - ERSACT Source: LAC Water Regulatory Governance Database, The World Bank, 2009. HN - ERSAPS BZ - ARSAL AR - ERSAC BA - FTC CR - ERSAPS BZ - AGENERSA PA - ASEP PE- SUNASS BZ - AGR BZ - ARCE BZ - ATR PA - ASEP CO - CRA BZ - AMAE CO - CRA CR - ERSAPS AR-ERAS TT - RIC AR-ERAS CO - CRA BZ - ATR HN - ERSAPS AR - ERSACT BZ - AGERSA BZ - AGESC BZ - ATR BZ - AGENERSA PY - ERSSAN BZ - AGR AR-ERAS AR - ERSAC BZ - AGERGS AR - ERSAC BZ - AMAE BZ - ARSAM T1 TT - RIC BZ - AGER BZ - ARSAM BA - FTC BZ - ADASA BZ - ARSAL BZ - ARSAL CR - ERSAPS PY - ERSSAN Regulatory Tools BZ - ARSAE T2 BZ - ATR Social Transparency Managerial Autonomy BZ - AGER BA - FTC BZ - ARSAE BZ - ARSAM AR - ENRESS BZ - AMAE AR - ERSPyOC PY - ERSSAN BZ - AGR Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. T2 AR - ERSAC BZ - AGERGS PA - ASEP BZ - AGESC BZ - ADASA BZ - ADASA AR - ERSPyOC AR - ERSPyOC AR - ENRESS AR - ERSACT AR - ENRESS BZ - ARCE PA - ASEP T2 AR - ERSACT BZ - ARSAE BZ - AGESC Index (0-1) Index (0-1) Index (0-1) .2 .4 .6 .8 .2 .4 .6 .8 .2 .4 .6 .8 0 1 0 1 0 1 PE- SUNASS TT - RIC BZ - AGER T1 BZ - AGERSA BZ - AGENERSA HN - ERSAPS CO - CRA BZ - AGESC CO - CRA BZ - ARCE BZ - ADASA BZ - AGERSA BZ - AGER CR - ERSAPS AR-ERAS BZ - AGR BZ - ARSAL BZ - ATR T1 AR - ENRESS BZ - AGR BZ - AGENERSA HN - ERSAPS BZ - AGERGS BZ - ARSAL T1 TT - RIC BA - FTC TT - RIC BZ - ARSAL PE- SUNASS BZ - ARCE BZ - ADASA BZ - ATR BA - FTC BZ - ARCE T2 BZ - AGERSA BZ - AGER CR - ERSAPS PE- SUNASS BZ - ARSAM BZ - ARSAM BZ - ARSAM CR - ERSAPS BZ - AMAE BZ - AGERGS BA - FTC PY - ERSSAN PY - ERSSAN BZ - AMAE BZ - ADASA PA - ASEP Informal Autonomy T2 HN - ERSAPS BZ - AMAE Informal Transparency Informal Accountability BZ - AGENERSA PA - ASEP CO - CRA PA - ASEP BZ - AGESC BZ - AGR BZ - AGESC AR-ERAS BZ - ARSAE Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. PY - ERSSAN BZ - ARSAE AR - ERSAC AR - ENRESS BZ - AGERGS BZ - ATR AR - ERSAC AR - ERSPyOC T2 BZ - ARSAE Index (0-1) AR - ERSACT Index (0-1) AR - ERSPyOC Index (0-1) AR - ERSPyOC AR - ERSAC AR-ERAS AR - ERSACT AR - ENRESS AR - ERSACT 138 .2 .4 .6 .8 .2 .4 .6 .8 .2 .4 .6 .8 0 1 0 1 0 1 TT - RIC TT - RIC TT - RIC BZ - ARCE BZ - AGR BZ - ADASA BZ - AGR BA - FTC T1 CO - CRA T1 PE- SUNASS PE- SUNASS CR - ERSAPS CR - ERSAPS T1 CO - CRA BZ - ARSAL CR - ERSAPS BZ - AGENERSA BZ - AMAE BZ - ADASA PY - ERSSAN AR - ERSAC BZ - ARSAL PA - ASEP PY - ERSSAN BZ - AGER BZ - ARSAL PA - ASEP BZ - ATR BZ - ADASA BZ - AGERSA BZ - AMAE T2 BZ - ATR BZ - AGERSA BZ - ATR BZ - AGENERSA BZ - AGENERSA PE- SUNASS BZ - AGER AR - ERSAC BZ - AMAE BZ - ARCE BZ - AGERGS BZ - AGERSA BA - FTC PY - ERSSAN BZ - AGER T2 BA - FTC BZ - AGERGS BZ - AGR Formal Autonomy T2 AR - ERSACT BZ - ARSAM Formal Transparency Formal Accountability PA - ASEP BZ - ARCE CO - CRA HN - ERSAPS HN - ERSAPS HN - ERSAPS BZ - AGESC BZ - ARSAM AR - ERSPyOC Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. Source: LAC Water Regulatory Governance Database, The World Bank, 2009. BZ - ARSAE BZ - AGESC AR - ERSACT AR - ENRESS AR-ERAS BZ - AGESC AR - ERSPyOC AR - ERSAC BZ - AGERGS AR-ERAS AR - ERSPyOC AR - ENRESS BZ - ARSAM BZ - ARSAE BZ - ARSAE AR - ERSACT AR - ENRESS AR-ERAS ANNEX 6: REGULATORY GOVERNANCE AND PERFORMANCE Table A6.1 Residential Energy Sold Distribut. Coverage Energy Sold Duration Frequency OPEX per OPEX per Avg Resid. Avg Indust. Cost Conn.per per Losses per Conn. of of Connection MWH sold Tariff Tariff Recovery Employee Employee interrupt's interrupt's (in dollars) (in dollars) (in dollars) (in dollars) Ratio (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Dummy Transition of 0.131*** 0.169*** 0.043*** -0.011*** 0.065*** -0.014 0.032 -0.314 -0.352 0.042** 0.064*** -0.005 PSP (0.012) (0.014) (0.013) (0.002) (0.003) (0.032) (0.037) (0.223) (0.224) (0.019) (0.023) (0.059) Dummy Post Transition 0.045*** 0.015 -0.131*** 0.003* 0.003 -0.295*** -0.348*** -0.142*** -0.089** -0.019** -0.031 0.192*** of PSP (0.008) (0.010) (0.012) (0.002) (0.005) (0.024) (0.023) (0.034) (0.036) (0.009) (0.021) (0.050) Existence of Regulatory 0.177*** 0.167*** -0.045*** 0.004* -0.031*** -0.210*** -0.190*** -0.387*** -0.320*** 0.145*** -0.047** 0.125*** Agency (0.010) (0.012) (0.009) (0.002) (0.005) (0.028) (0.029) (0.051) (0.056) (0.016) (0.021) (0.032) Utility FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Utility Specific Yes Yes No Yes No No No No No No No No Time trend Observations 2000 1981 2073 1323 2515 1056 947 864 873 1728 840 669 Number of utilities 199 198 190 144 213 144 132 131 131 175 90 103 Standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% 139 Table A6.2 Residential Energy Sold Distribut. Coverage Energy Sold Duration Frequency OPEX per OPEX per Avg Resid. Avg Indust. Cost Conn.per per Losses per Conn. of of Connection MWH sold Tariff Tariff Recovery Employee Employee interrupt's interrupt's (in dollars) (in dollars) (in dollars) (in dollars) Ratio (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Dummy Transition of 0.121*** 0.170*** 0.125*** -0.012*** 0.057*** -0.018 0.059 -0.278 -0.230 0.164*** 0.055 -0.053 PSP (0.014) (0.016) (0.018) (0.003) (0.008) (0.040) (0.046) (0.228) (0.229) (0.024) (0.039) (0.107) Dummy Post Transition 0.018 -0.020 -0.123*** 0.006 0.095*** -0.561*** -0.429*** -0.110 -0.116 -0.087*** 0.123 0.308*** of PSP (0.015) (0.025) (0.035) (0.004) (0.015) (0.074) (0.064) (0.102) (0.099) (0.018) (0.100) (0.116) Existence of Regulatory 0.162*** 0.175*** -0.008 0.002 -0.023*** -0.239*** -0.176*** -0.351*** -0.233*** 0.286*** -0.039 0.146*** Agency (0.014) (0.017) (0.010) (0.002) (0.006) (0.038) (0.044) (0.069) (0.078) (0.024) (0.026) (0.042) Transition * Existence 0.026 -0.012 -0.144*** 0.004 0.005 0.019 -0.061 -0.016 -0.150 -0.315*** 0.000 0.045 (0.018) (0.022) (0.022) (0.004) (0.011) (0.054) (0.059) (0.121) (0.129) (0.033) (0.045) (0.114) Post Trans. * Existence 0.032* 0.041 0.020 -0.005 -0.102*** 0.284*** 0.107 -0.071 0.006 0.138*** -0.158 -0.123 (0.017) (0.026) (0.037) (0.005) (0.016) (0.078) (0.069) (0.109) (0.108) (0.021) (0.102) (0.121) Utility FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Utility Specific Yes Yes No Yes No No No No No No No No Time trend Observations 2000 1981 2073 1323 2515 1056 947 864 873 1728 840 669 Number of utilities 199 198 190 144 213 144 132 131 131 175 90 103 Standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% Table A6.3 Residential Distribut. Coverage Energy Sold Duration Frequency OPEX per OPEX per Avg Resid. Avg Indust. Cost Conn.per Losses per Conn. of of Connection MWH sold Tariff Tariff Recovery Employee interrupt's interrupt's (in dollars) (in dollars) (in dollars) (in dollars) Ratio (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Dummy Transition of 0.175*** 0.030*** -0.013*** 0.062*** -0.022 0.044 -0.463** -0.451** 0.053*** 0.027 0.043 PSP (0.014) (0.011) (0.002) (0.005) (0.027) (0.033) (0.220) (0.228) (0.016) (0.019) (0.059) Dummy Post Transition 0.101*** -0.091*** 0.006*** 0.024*** -0.112*** -0.167*** -0.152*** -0.158*** -0.089*** -0.058*** 0.157*** of PSP (0.008) (0.012) (0.002) (0.006) (0.025) (0.024) (0.036) (0.046) (0.011) (0.021) (0.049) Duration of the Regulat. -0.014*** -0.018*** 0.004*** -0.018*** -0.094*** -0.094*** -0.057*** -0.016*** 0.026*** -0.013*** 0.040*** Agency (0.003) (0.002) (0.001) (0.001) (0.008) (0.007) (0.005) (0.004) (0.004) (0.003) (0.005) Duration of the Regulat. -0.000 0.001*** -0.001*** 0.001*** 0.004*** 0.004*** 0.003*** 0.001*** 0.001*** 0.002*** -0.001*** Agency (Sq.) (0.000) (0.000) (0.000) (0.000) (0.001) (0.001) (0.000) (0.000) (0.000) (0.000) (0.000) Utility FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Utility Specific Yes No Yes No No No No No No No No Time trend Observations 2000 2073 1323 2515 1056 947 864 873 1728 840 669 Number of utilities 199 190 144 213 144 132 131 131 175 90 103 Standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% 140 Table A6.4 Residential Energy Sold Distribut. Coverage Energy Sold Duration Frequency OPEX per OPEX per Avg Resid. Avg Indust. Cost Conn.per per Losses per Conn. of of Connection MWH sold Tariff Tariff Recovery Employee Employee interrupt's interrupt's (in dollars) (in dollars) (in dollars) (in dollars) Ratio (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Dummy Transition of 0.124*** 0.159*** 0.045*** -0.012*** 0.054*** -0.010 0.031 -0.269 -0.293 0.041** 0.070*** -0.006 PSP (0.012) (0.014) (0.013) (0.003) (0.005) (0.033) (0.038) (0.225) (0.227) (0.018) (0.022) (0.060) Dummy Post Transition 0.062*** 0.030*** -0.118*** 0.001 -0.007 -0.276*** -0.332*** -0.213*** -0.179*** 0.019 -0.027 0.194*** of PSP (0.009) (0.011) (0.012) (0.002) (0.005) (0.024) (0.023) (0.036) (0.044) (0.012) (0.021) (0.050) Regulatory Governace 0.236*** 0.226*** -0.077*** 0.005* -0.029*** -0.274*** -0.248*** -0.495*** -0.373*** 0.154*** -0.074*** 0.150*** Index (ERGI) (0.013) (0.016) (0.013) (0.003) (0.007) (0.036) (0.038) (0.069) (0.076) (0.021) (0.028) (0.042) Utility FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Utility Specific Yes Yes No Yes No No No No No No No No Time trend Observations 1859 1840 1983 1247 2337 1030 924 841 850 1655 831 660 Number of utilities 181 180 175 137 195 139 127 126 126 159 85 98 Standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% Table A6.5 Residential Distribut. Coverage Energy Sold Duration Frequency OPEX per OPEX per Avg Resid. Avg Indust. Cost Conn.per Losses per Conn. of of Connection MWH sold Tariff Tariff Recovery Employee interrupt's interrupt's (in dollars) (in dollars) (in dollars) (in dollars) Ratio (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) Dummy Transition of 0.122*** 0.027** -0.014*** 0.059*** -0.043 0.046 -0.730* -0.808** 0.147*** 0.087*** 0.068 PSP (0.012) (0.013) (0.003) (0.004) (0.037) (0.044) (0.397) (0.400) (0.019) (0.021) (0.062) Dummy Post Transition 0.084*** -0.124*** 0.002 -0.008 -0.358*** -0.366*** -0.193*** -0.137*** 0.049*** -0.016 0.176*** of PSP (0.008) (0.013) (0.002) (0.005) (0.025) (0.024) (0.039) (0.049) (0.013) (0.021) (0.053) PCA 1 - Informal 0.001 -0.027*** -0.001 -0.048*** 0.014 0.010 0.046 0.053 0.087*** 0.010 -0.003 (0.007) (0.007) (0.001) (0.004) (0.018) (0.018) (0.042) (0.050) (0.010) (0.021) (0.018) PCA 2 - Formal 0.107*** -0.006 0.004* 0.037*** -0.024 -0.103*** 0.050 0.092 -0.145*** -0.051*** -0.071* (0.008) (0.006) (0.002) (0.004) (0.026) (0.028) (0.084) (0.085) (0.014) (0.016) (0.037) PCA 3 - Formal Autonomy 0.085*** -0.069*** 0.012*** -0.009 -0.144*** -0.080 -0.405*** -0.339** -0.036* -0.030 0.266*** and Tariffs (0.015) (0.012) (0.004) (0.009) (0.053) (0.049) (0.111) (0.132) (0.020) (0.029) (0.068) Utility FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Utility Specific Yes No Yes No No No No No No No No Time trend Observations 1782 1917 1190 2253 974 882 800 809 1596 820 619 Number of utilities 175 169 131 189 134 123 121 121 153 84 93 Standard errors in parentheses * significant at 10%; ** significant at 5%; *** significant at 1% 141 A NNEX 7: C ORPORATE G OVERNANCE AND P ERFORMANCE Table A7.1: Correlation between CG indexes and performance ­ Water and Electricity distribution Sectors (in levels) Distributional Quality of Coverage Labor Residential Losses the Service Productivity Tariffs Legal Soundness -0.41 0.05 -0.26 0.29 0.39 CEO Competitiveness -0.39 0.08 -0.33 0.08 0.36 Board Competitiveness -0.22 -0.14 -0.12 0.10 0.14 Professional Management -0.24 0.13 -0.08 0.34 0.22 Transparency & Disclosure 0.14 -0.16 0.37 0.24 -0.31 Performance Orientation -0.25 0.28 -0.09 0.26 0.22 Corporate Governance -0.44 0.09 -0.20 0.40 0.37 Table A7.2: Correlation between CG indexes and performance ­ Water and Electricity distribution Sectors (in growth rates) Distributional Quality of Coverage Labor Residential Losses the Service Productivity Tariffs Legal Soundness 0.04 -0.31 0.14 -0.10 0.26 CEO Competitiveness 0.05 -0.10 0.35 0.01 0.06 Board Competitiveness -0.06 -0.10 -0.08 0.18 0.00 Professional Management 0.03 -0.11 0.07 0.12 0.01 Transparency & Disclosure -0.02 -0.04 0.15 0.10 -0.37 Performance Orientation 0.18 0.09 0.30 0.13 0.01 Corporate Governance 0.07 -0.20 0.31 0.12 0.02 Table A7.3: Correlation between CG indexes and performance ­ Electricity distribution Sector (in levels) Distributional Duration of Frequency of Labor Residential Industrial Coverage Losses Interruptions Interruptions Productivity Tariifs Tariffs Legal Soundness 0.02 0.39 0.32 -0.32 -0.41 0.42 0.42 CEO Competitiveness 0.17 0.28 0.41 -0.02 -0.51 -0.19 0.22 Board Competitiveness -0.01 0.47 0.44 -0.03 -0.23 0.09 0.50 Professional Management 0.08 0.21 0.10 0.05 -0.07 0.40 0.18 Transparency & Disclosure -0.19 -0.18 0.00 -0.07 0.20 0.09 -0.23 Performance Orientation 0.06 -0.15 -0.04 0.14 0.31 0.23 -0.26 Corporate Governance 0.06 0.37 0.44 -0.11 -0.30 0.38 0.31 Table A7.4: Correlation between CG indexes and performance ­ Electricity distribution Sector (in growth rates) Distributional Duration of Frequency of Labor Residential Industrial Coverage Losses Interruptions Interruptions Productivity Tariifs Tariffs Legal Soundness -0.10 0.36 0.30 0.19 -0.10 0.15 -0.01 CEO Competitiveness -0.01 0.09 0.01 0.02 -0.19 -0.08 -0.26 Board Competitiveness -0.09 0.10 0.05 0.00 0.07 0.00 0.02 Professional Management 0.24 0.09 0.13 -0.15 -0.31 0.02 -0.30 Transparency & Disclosure -0.03 -0.03 0.16 0.32 0.17 -0.28 -0.49 Performance Orientation 0.28 -0.20 -0.14 0.03 0.04 -0.34 -0.16 Corporate Governance 0.09 0.16 0.18 0.17 -0.11 -0.18 -0.40 142 Table A7.5: Correlation between Corporate Governance indexes and performance ­ Water Sectors (in levels) Non Revenue Continuity of Potability Water Sewerage Res. Water Res. Sewerage Labor Metering Water the Service Coverage Coverage Tariffs Tariffs Productivity Legal Soundness -0.33 0.34 -0.05 -0.08 0.09 0.29 -0.01 0.54 -0.48 CEO Competitiveness -0.02 -0.52 -0.12 -0.13 0.26 0.23 -0.23 0.07 -0.02 Board Competitiveness -0.23 -0.12 0.31 0.29 -0.04 0.01 0.12 0.15 0.03 Professional Management -0.27 -0.13 0.24 0.23 -0.07 0.31 0.11 0.53 -0.09 Transparency & Disclosure -0.29 0.09 0.31 0.39 0.17 -0.11 0.32 0.26 0.26 Performance Orientation -0.37 -0.23 0.62 0.35 0.18 0.17 0.03 0.46 0.21 Corporate Governance -0.42 -0.14 0.41 0.32 0.17 0.30 0.10 0.59 -0.04 Table A7.6: Correlation between Corporate Governance indexes and performance ­ Water Sectors (in growth rates) Non Revenue Continuity of Potability Water Sewerage Res. Water Res. Sewerage Labor Metering Water the Service Coverage Coverage Tariffs Tariffs Productivity Legal Soundness 0.11 -0.37 0.25 -0.24 0.13 -0.04 0.17 -0.05 -0.03 CEO Competitiveness -0.04 0.70 0.17 0.24 0.33 -0.52 -0.38 0.01 -0.17 Board Competitiveness -0.10 0.36 0.22 0.02 -0.21 -0.03 -0.32 0.32 0.28 Professional Management -0.21 0.27 0.16 -0.23 0.25 -0.23 -0.20 0.29 0.36 Transparency & Disclosure 0.09 0.32 -0.01 0.28 0.20 -0.13 -0.09 0.08 0.32 Performance Orientation -0.05 0.42 -0.73 -0.11 0.55 0.13 0.37 0.41 0.51 Corporate Governance -0.06 0.48 0.05 -0.04 0.39 -0.25 -0.13 0.30 0.41 143 Table A7.8: Principal Component Analysis - Eigenvalues of factors Component Eigenvalue Difference Proportion Cumulative Comp1 2.173 0.810 0.362 0.362 Comp2 1.362 0.357 0.227 0.589 Comp3 1.006 0.289 0.168 0.757 Comp4 0.717 0.271 0.119 0.876 Comp5 0.445 0.148 0.074 0.950 Comp6 0.297 . 0.050 1.000 Table A7.9: Principal Component Analysis - Factor loadings of indexes after varimax rotation Variable Comp1 Comp2 Comp3 Unexplained Performance Orientation 0.678 -0.327 -0.128 0.128 Legal Soundness 0.217 0.151 0.624 0.328 Transparency & Disclosure 0.277 0.223 -0.692 0.157 Board Competitiveness -0.067 0.859 -0.050 0.076 CEO Competitiveness 0.374 0.162 0.335 0.485 Professional Management 0.522 0.236 0.028 0.287 144 R EFERENCES Andres, L., J.G. Diaz, and J.L. Guasch (2009) An Empirical Study of the Effects of Renegotiations over the Auctioning of PPP Concessions. The World Bank. Andres, L., M. Diop, and J.L. Guasch. (2008a) Achievements and Challenges of Private Participation in Infrastructure in Latin America: Evaluation and Future Prospects. In Henry Davis, ed., Euromoney Infrastructure Financing, Oxford: Oxford University Press. Andres, L. and G. Dragoiu (2008b) Benchmarking Electricity distribution Report 1995-2005, The World Bank. Andres, L., V. Foster, and J.L. Guasch (2006). The Impact of Privatization on the Performance of the Infrastructure Sector: The Case of Electricity Distribution in Latin American Countries, Policy Research Working Paper 3936, The World Bank. Andres, L., J.L. Guasch, V. Foster, and T. Haven (2008c), The Impact of Private Sector Participation in Infrastructure: Lights, Shadows, and the Road Ahead. The World Bank, Washington, DC. Andres, L., J.L. Guasch, M. Diop, and S. Lopez Azumendi (2007a) Assessing the Governance of Electricity Regulatory Agencies in the Latin American and the Caribbean Region: a Benchmarking Analysis, Policy Research Working Paper 4380, The World Bank. Andres, L., J.L. Guasch, and S. Lopez Azumendi (2008d). Regulatory Governance and Sector Performance: Methodology and Evaluation for Electricity Distribution in Latin America. Policy Research Working Papers Series 4494. The World Bank. Aron, J. (2000). "Growth and Institutions: A review of the evidence." The World Bank Research Observer, vol. 15, no. 1, pp. 99-135. Asian Development Bank (2004). Water in Asian Cities: Utilities Performance and Civil Society Views. Water for All Series No. 10, Asian Development Bank, Manila. Bajari, P., S. Houghton, and S. Tadelis (2007). Bidding for Incomplete Contracts: An Empirical Analysis of Adaptation Costs. Working Paper, University of Minnesota. Boix, C. (2005) Privatization and Public Discontent in Latin America. Background paper commissioned for Infrastructure in Latin America and the Caribbean: Recent Developments and Key Challenges. Inter-American Development Bank, Washington, DC. Briceño-Garmendia, C., A. Estache, and N. Shafik (2004). Infrastructure Services In Developing Countries: Access, Quality, Costs and Policy Reform. World Bank Policy Research Working Paper No. 3468. The World Bank. Brown, A.C., J. Stern, B. Tenenbaum, and D. Gencer (2006). Handbook for Evaluating Infrastructure Regulatory Systems. The World Bank, Washington DC. 145 Calderón, C. and L. Servén (2004). The Effects of Infrastructure Development on Growth and Income Distribution. Policy Research Working Paper No. 3400. The World Bank. Chong, A., and F. López-de-Silanes. (2003). The Truth about Privatization in Latin America. Latin American Research Network, Research Network Working Paper No. R-486, October. Inter- American Development Bank, Washington, DC. Chong, A. and F. Lopez de Silanes (2005). Privatization in Latin America: Myths and Reality, Alberto Chong and Florencio Lopez de Silanes (eds). Inter-American Development Bank and Stanford University press. Correa, P., C. Pereira, B. Mueller, and M. Melo (2006). Regulatory Governance in Infrastructure Industries: Assessment and Measurement of Brazilian Regulators. The World Bank/Public- Private Infrastructure Advisory Facility, Washington DC. Cubbin, J. and J. Stern (2006). "The Impact of Regulatory Governance and Privatization on Electricity Industry Generation Capacity in Developing Countries." The World Bank Economic Review Advance Access published. Washington, D.C. Cullmann, A., H. Crespo, and M. Plagnet (2008). International Benchmarking in Electricity Distribution: A Comparison of French and German Utilities. Discussion Papers of DIW Berlin. German Institute for Economic Research. Dal Bo, E. and M.A. Rossi (2007). "Corruption and Inefficiency: Theory and Evidence from Electric Utilities." Journal of Public Economics. Elsevier, vol. 91(5-6), pages 939-962, June. De Ferranti, D., G.E. Perry, F.H.G. Ferreira, and M. Walton. (2004). Inequality in Latin America: Breaking with History? The World Bank, Washington, DC Djerrari, M. F. and J.G. Janssens (1998). Overview Paper on Institutional Options for Urban Water and Sanitation Services in Africa, WUP, Abidjan. Engel, E., R. Fischer, and A. Galetovic (1997). Infrastructure Franchising and Government Guarantees. In Dealing with Public Risk in Private Infrastructure, edited by T. Irwin, M. Klein, G. Perry and M. Thobani. World Bank, Washington, DC. Estache, A. (2006). Infrastructure: A Survey of Recent and Upcoming Issues, The World Bank. Estache A., A. Goicoechea, and L. Trujillo (2009). Utilities reforms and corruption in developing countries. Utilities Policy 17 191­202. Estache, A., B. Tovar, and L. Trujillo (2008). How Efficient are African Electricity Companies? Evidence from the Southern African Countries. Energy policy, vol. 36, No6, pp. 1969 -1979. Estache, A. and M.A. Rossi (2008). "Regulatory Agencies: Impact on Firm Performance and Social Welfare," Policy Research Working Paper Series 4509, The World Bank. 146 Estache A., B. Tovar, and L. Trujillo (2007). Are African Electricity Distribution Companies Efficient? Evidence of Southern African Countries. Discussion Paper Series, Department of Economics, City University London. Estache, A, M. A. Rossi, and C. A. Ruzzier (2004). The Case for International Coordination of Electricity Regulation: Evidence from the Measurement of Efficiency in South America. Washington, D.C., Journal of Regulatory Economics, 25/3, 271-295. Estache, A. and L. Quesada (2001). Concession Contract Renegotiations: Some Efficiency versus Equity Dilemmas. World Bank Policy Research Working Paper No. 2705. The World Bank. Fay, M., and M. Morrison (2006). Infrastructure in Latin America and the Caribbean: Recent Developments and Key Challenges. The World Bank, Washington, DC Filippini, M. (1998). Are Municipal Electricity Distribution Utilities Natural Monopolies? Annals of Public and Cooperative Economics, 69/2, 157-174. Filippini, M. and J. Wild (2001). Regional Differences in Electricity Distribution Costs and Their Consequences for Yardstick Regulation of Access Prices, Energy Economics, 23, 477-488. Gassner, K., A. Popov, and N. Pushak (2007). An Empirical Assessment of Private Sector Participation in Electricity and Water Distribution in Developing Countries. The World Bank, Washington. Gilardi, F. (2002). Policy Credibility and Delegation to Independent Regulatory Agencies: A Comparative Empirical Analysis, Journal of European Public Policy, 9(6): 873-893. Gómez-Ibáñez, A.J. (2007). Private Infrastructure in Developing Countries: Lessons from Recent Experience." Paper presented to the Commission on Growth and Development at the Workshop on Global Trends and Challenges September 28 and 29, 2007. Yale Center for the Study of Globalization New Haven, Connecticut . Guasch, J.L., J.J. Laffont, and S. Straub (2005). Government-led Renegotiations. World Bank Policy Research Working Paper No. 3749. The World Bank. Guasch, J.L. (2004). Granting and Renegotiating Infrastructure Concessions. Doing it Right. The World Bank, Washington, DC. Guasch, J.L., J.J. Laffont and S. Straub (2003). Renegotiation of Concession Contracts in Latin America. World Bank Policy Research Working Paper No. 3011. The World Bank. Guasch, J. L., A. Kartasheva, and L. Quesada (2000). Concession Contracts in Latin America and Caribbean Region: An Economic Analysis and Empirical Evidence. The World Bank. Gutiérrez, L.H. (2003). The Effect of Endogenous Regulation on Telecommunications Expansion and Efficiency in Latin America. Journal of Regulatory Economics 23 (3): 257­86. Hirschhausen, V. C. and A. Kappeler (2004). Productivity Analysis of German Electricity Distribution Utilities. German Institute of Economic Research DIW Berlin, Discussion Papers 418. 147 Hjalmarsson, L. and A. Veiderpass (1992). Productivity in Swedish Electricity Retail Distribution, Scand. J. of Economics 94, Supplement. 193-205 Jamasb, T., R. Mota, D. Newbery, and M. Pollitt (2005). Electricity Sector Reform in Developing Countries: A Survey of Empirical Evidence on Determinants and Performance. World Bank Policy Research Working Paper No. 3549. The World Bank. Johannsen, K. S. (2003). Regulatory Independence in Theory and Practice: A Survey of Independent Energy Regulators in Eight European Countries. Energy Research Programme and the Danish Research Training Council, February 2003. Kenny, C. and T. Soreide (2008). Grand Corruption in Utilities World Bank Policy Research Working Paper No. 4805. The World Bank. Kim H.Y. and M.R. Clark (1988). Economies of Scale and Scope in Water Supply. Regional Science and Urban Economics, 479-502. North Holland. Komives, K., V. Foster, J. Halpern, and Q. Wodon (2005). Water, Electricity and the Poor: Who Benefits from Utility Subsidies? Directions in Development, The World Bank. Martimort, D. and S. Straub (2005). The Political Economy of Private Participation, Social Discontent and Regulatory Governance. Background paper commissioned for Infrastructure in Latin America and the Caribbean: Recent Developments and Key Challenges. The World Ban. Mizutani, F. and K. Urakami (2005). Identifying Network Density and Scale Economies for Japanese Water Supply Organizations. Graduate School of Business Administration, Kobe University. Muller M., R. Simpson, and M. van Ginneken (2008). Ways to Improve Water Services by Making Utilities More Accountable to Their Users: A Review. Water Working Notes, The World Bank. Ogunbiyi, C. (2004). PPPs: Fad or Good for SADC? SADC PPP Pathway, SADC Banking Association PPP Capacity Building Programme, Newsletter No. 1, July. Oliveira, G., E. Machado, L. Novaes, L. Martins, G. Ferreira, and C. Beatriz (2005). Aspects of the Independence of Regulatory Agencies and Competition Advocacy. Getulio Vargas Foundation, Rio de Janeiro. OECD (1999). Regulatory Reform in Mexico: Government Capacity to Assure High Quality Regulation. Organisation for Economic Co-operation and Development, Paris. Nauges, C., and van den Berg, C. (2008). Economies of Density, Scale and Scope in the Water Supply and Sewerage Sector: A Study of Four Developing and Transition Economies. Journal of Regulatory Economics 34:144­163. Petrazzini, B.A. (1996). Competition in Telecoms--Implications for Universal Service and Employment. Industry and Energy Department, Note No. 96. The World Bank. 148 Ramamurti, R. (1996). Privatizing Monopolies: Lessons from the Telecommunications and Transport Sector in Latin America. Baltimore, MD: The Hopkins University Press. Rodrick, D. (2004). "Getting Institutions Right." CESifo DICE Report: A Journal for International Comparisons 2:10-15. Ros, A. (1999). Does Ownership or Competition Matter? The Effects of Telecommunications Reform on the Network Expansion and Efficiency. Journal of Regulatory Economics 15 (1): 65­92. Saal, D. S., and D. Parker (2001). Productivity and Price Performance in the Privatized Water and Sewerage Companies of England and Wales. Journal of Regulatory Economics 20 (1): 61­90. Saleth, R. M., and a. Dinar (1999). Evaluating Water Institutions and Water Sector Performance. World Bank Technical Paper No. 447. Schwartz, J. (2007). The Feasibility of Regional Cooperation in Regulation of the Electricity Sector of the Eastern Caribbean States, The World Bank. Sirtaine, S. M.E. Pinglo, J. L. Guasch, and V. Foster (2005). How Profitable Are Infrastructure Concessions in Latin America? Empirical Evidence and Regulatory Implications. Trends and Policy Options, No. 2 PPIAF, The World Bank. Sosay, G. and U. Zenginobuz (2005). Independent Regulatory Agencies in Emerging Economies. Munich Personal RePEc Archive. Paper No. 380. Stern, J. and S. Holder. (1999). Regulatory Governance: Criteria for Assessing the Performance of Regulatory Systems: An application to Infrastructure in Developing Countries of Asia. Regulation Initiative Discussion Paper Series Number 20. Stern, J. and J. Cubbin (2005). "Regulatory Effectiveness: The Impact of Regulation and Regulatory Governance Arrangements on Electricity Industry Outcomes". The World Bank Policy Research Working Paper 3536. Stern, J. and J. Cubbin (2006). "The Impact of Regulatory Governance and Privatization on Electricity Industry Generation Capacity in Developing Countries." The World Bank Economic Review Advance Access published. Washington, D.C.. Thatcher, M. (2006). After Delegation: the Evolution of European Networks of Regulatory Agencies. ECPR Regulation Conference. Bath University. Thatcher, M. (2007). Regulatory Agencies, the State and Markets: A Franco-British Comparison. European Union University. Working Paper RSCAS 2007/17. United Nations Economic and Social Council Economic Commission for Africa (2005). Public- Private Partnerships for Services Delivery: Water and Sanitation". Third meeting of the Committee on Human Development and Civil Society (4-6 May 2005) Addis Ababa, Ethiopia. 149 Vagliasindi, M. (2008) Governance Arrangements for State Owned Enterprises. European Bank for Reconstruction and Development (EBRD). Vagliasindi, M. (2009) The Links Between Internal and External Governance and the Performance of Infrastructure Service Providers. The World Bank. Wallsten, S. (2001). Reverse Auctions and Universal Telecommunications Service: Lessons from Global Experience. Technology Policy Institute; Stanford University; Economists Incorporated. Whincop, M. J. (2005). Corporate Governance in Government Corporations. Aldershot, Hants, England. Burlington, VT. Ashgate. WHO-UNICEF (2008). Joint Monitoring Programme (JMP) for Water Supply and Sanitation. http://www.wssinfo.org/. World Bank (2007a). LCR Energy Strategy. The World Bank. World Bank (2007b). Private Participation in Infrastructure Projects (PPP) Database. The World Bank. http://ppi.worldbank.org/. 150