41704 noTE no. 29 ­ SEpT. 2007GRIDLINES Sharing knowledge, experiences, and innovations in public-private partnerships in infrastructure What it takes to lower regulatory risk in infrastructure industries An assessment and benchmarking of Brazilian regulators Paulo Correa R egulatory governance--how regulators Good decisions are more likely if regulatory manage concession contracts, or other design is sound. But even with sound design, public-private contractual arrangements poor decisions can be made. Actual regulatory and sector laws--can affect the private sector's decisions--how regulators manage concession perception of regulatory risk and thus the avail- contracts and sector laws in practice--therefore ability of private capital for infrastructure projects. become a critical part of regulatory risk. Four key elements of the regulatory governance structure can reduce the risk of regulatory failure: No regulation is risk free. But several conditions political and financial autonomy, decision-making appear to reduce the likelihood of regulatory fail- structures that reduce regulatory discretion, ure: sufficient political and financial autonomy; access to effective enforcement and other regu- structures for decision making that constrain latory tools, and efficient rules of accountability. regulatory discretion; adequate access to regula- This note presents an analytical framework based tory means, including legal provisions for effective on those four elements and applies it in assessing enforcement of decisions; and efficient rules of regulatory governance in Brazil. accountability. Together, these conditions form the structure of regulatory governance. Since the early 1990s governments seeking to expand private investment in infrastructure have delegated regulatory authority to independent A framework for assessing entities. Through this delegation of power, govern- regulatory governance ments sought to assure private investors that they would not intervene arbitrarily and expropriate Regulatory governance defines the conditions economic rents once investments were sunk. under which regulators enforce concession contracts and sector laws. While there is no univer- Brazil was among them. As investment in sal framework for assessing regulatory governance, infrastructure projects with private participa- some common attributes are increasingly cited. tion climbed to US$125 billion in the late 1990s--more than in China, Indonesia, and the Autonomy Philippines combined--Brazil created 5 federal Regulators need autonomy from both the and 22 state regulatory agencies for infrastructure government and the regulated industry. Political industries and reformed most sector laws. autonomy insulates regulators from short-term political interference. Two policies favor politi- Yet delegating regulatory powers, including cal autonomy: providing regulators tenure and enforcement, may not be enough to minimize staggered terms (not coinciding with the term of the risk of expropriation, or regulatory risk. As the executive) and ensuring their financial inde- a result of unforeseen contingencies and insuffi- pendence (Smith 1997a). Insulation from private cient consensus on detailed regulation, concession PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY contracts and sector laws eventually become more Paulo Correa is a senior economist in the Finance and flexible, and some degree of regulatory discretion Private Sector Unit of the World Bank's Europe and Central and unpredictability is inevitable. Asia Region. Helping to eliminate poverty and achieve sustainable development through public-private partnerships in infrastructure PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY sector interests is also important, as decisions by Efficient rules of accountability provide for eval- captured regulators tend not to be sustainable. uation of regulators' performance and periodic review of the strategy for delegation of regulatory Decision-making structures powers. These create a mechanism for improving Procedural requirements limit the choices of regu- regulation without arbitrary intervention in regu- lators, protecting private investors from abuses latory decisions. But excessive interference from of discretionary power (Smith 1997c). They other organizations may reduce the predictability may also help induce systematic deliberations, of regulatory decisions and raise transaction costs strengthening coherence and predictability in the (such as when the court reverses regulatory deci- enforcement of laws and contracts (Stern and sions on the basis of content rather than a failure Holder 1999; Noll 2001). Compliance with due of due process). process reduces the risk that a regulator's decisions will be reversed in court (Berg 2001). Transpar- ency (disclosing relevant information, announcing The case of Brazil meeting schedules in advance, ensuring that stake- holders have equitable access to decision makers) The framework, based on these four key elements, also increases the legitimacy of regulatory deci- was applied to analyze regulatory governance in sions (Smith 1997c). Brazilian infrastructure industries. The analysis identified key parameters to characterize regula- Decision-making rules are key in assessing the tory institutions and quantitative indicators to decision-making process (Smith 1997a). Also assess their governance. The assessment was based important is the extent to which precedent binds on a survey of 21 of the 27 infrastructure regula- future decisions (Berg 1998). Requiring written tory agencies (5 federal, 22 state) in Brazil.1 explanations of regulatory decisions can help lay the foundation for consistent implementation of Autonomy the law (Guasch and Spiller 1999). Most of the agencies reported that directors' positions were tenured. But in a surprisingly large Decision tools number of agencies (7 of 21) directors can be Regulators lack information on cost and demand, dismissed for vague reasons (for example, "threat- and regulated firms have little incentive to reveal ening the agency's integrity"). Directors' terms do No regulation this information. This asymmetry of information not coincide with the executive's except in three creates opportunities for strategic behavior by state agencies. Seventeen agencies declared that is risk free. both the regulator and the regulated firms (Baron they had the financial autonomy to control their But some key 1989). Informational problems are aggravated by budgets. Only two reported that the government evolving market structures and rapid technologi- was their sole source of revenue. conditions can cal advances. reduce risk. Although all the agencies have legal autonomy, To reduce the risk of ineffective regulation or 13 reported that the executive (ministry or gover- mismanagement of contracts, access to informa- nor) had made "highly" or "very highly" effective tion and resources to process it are essential (Noll attempts to intervene. Eight agencies (two federal 2001; Smith 1997b). Good management of the and six state) reported direct interventions, though regulatory process requires, at a minimum, the legal defending the interests of the government (as means to collect information, budgetary resources shareholder or consumer) was never the primary for managing and processing information, quali- objective. fied staff, and standard regulatory tools. Decision-making structures Accountability The main voting rule used by the regulatory agen- Who regulates the regulator? Oversight mecha- cies to make decisions is simple majority. Only nisms are important tools of accountability. two agencies rely mostly on qualified majority. Regulatory agencies should be subject to legisla- Two others use both rules, depending on the type tive oversight and to requirements for periodic of decision. Directors have veto power over the reporting. Like any public entity, regulators also board's decisions in only two agencies. Most agen- should be monitored by the public prosecutor's cies allocate cases among their directors on the office and by the appropriate audit office. Another basis of their areas of specialization or the decision important tool of accountability is judicial appeal, of the director general. In almost all agencies any usually available when regulatory decisions are director can obstruct the normal development of administrative. a case after it has been assigned. What it takes to lower regulatory risk in infrastructure industries Almost all agencies must document their deci- have affected decisions by at least once changing sions in writing and back them with technical the original proposal (15 agencies), delaying a deci- analysis. Interestingly, among the 18 agencies sion (11), or blocking the original proposal (4). formally required to document their decisions, only 7 are also required to cite previous regula- Ranking the regulators tory decisions. Only 6 agencies make decisions To rank the Brazilian regulators, results for 83 without prior communication and discussion attributes of regulatory governance were combined among board members. And only 3 (of 18) agen- in a summary index (RGI-83). One striking result Formal cies reported that directors are forbidden by law to is the degree of homogeneity: two-thirds of the attributes do hold informal meetings with stakeholders. Seven- sample are within one standard deviation of the teen agencies are required by law to allow public mean. Based on formal attributes, those ranking not always participation, with 15 reporting that such partici- highest are the three federal agencies--ANATEL translate pation has led to changes in decisions. (Federal Telecommunications Agency), ANEEL (Federal Electricity Agency), and ANP (Federal into actual Decision tools Petroleum Agency). But when only questions practice Most regulators reported having appropriate legal about actual practices are used (in the RGI-28), means to enforce regulations and collect informa- the rankings change (figure 1). tion. Responses on staffing, however, revealed that only about 20 percent of the agencies' personnel Another exercise disaggregated the summary index had been admitted through public examinations, (RGI-83) by the four components of regulatory reflecting a lack of professional or specialized governance, by the age of the regulatory agencies, staff.2 Of the 21 agencies surveyed, 12 considered and by their jurisdiction. Results show that older salaries for top technical and managerial positions agencies have better decision tools and account- in regulatory agencies to be at least 25 percent ability mechanisms on average. But no statistically lower than the salary of the attorney general or significant difference was found between old and state finance secretary (used as a benchmark). In young agencies in autonomy and decision-making 18 agencies technical staff had been in their jobs structures. One possible explanation for these for more than two years on average, indicating opposite results: while autonomy and decision- relatively low turnover. making structures are defined when regulatory powers are initially delegated, decision tools and Technical qualifications were limited. Among the accountability mechanisms may evolve as a result 14 agencies providing data on this issue, 9 had of learning and adaptation. employees with master's degrees (representing 4 percent of the technical staff on average), and only 6 had employees with PhDs (2 percent). Conclusion Most agencies reported having access to standard The assessment of regulatory governance in Brazil regulatory tools. More sophisticated tools for led to four broad findings worth highlighting: economic regulation (such as models for bench- marking firm performance) were less available. · Formal attributes of governance do not always translate into effective attributes in practice, Accountability particularly in the case of autonomy. Decisions can be appealed first to the agency-- · Federal regulatory agencies generally score then, if not resolved, to the court. Five agencies higher than state agencies on formal attributes, reported having had a case reach the Supreme especially human resources. Court. Given the short life of most of the agencies, · Older agencies have better decision tools and this number may not be negligible. accountability (but not better decision-making rules) than younger agencies, suggesting a signif- Legislatures exert some control over 17 of the 21 icant learning curve. agencies surveyed. The most effective mechanism · Weakness in the decision-making process under- was requesting explanations, followed by summon- mines the consistency of regulatory decisions ing the directors and holding public hearings. For over time and the fairness of the regulatory all three mechanisms state agencies indicated process. higher levels of control than federal agencies. The findings suggest that well-crafted government Of the 21 agencies surveyed, 15 hold public hear- policy might help regulators fill the holes of the ings. These agencies reported that the hearings regulatory process. And the (quasi) public good FIguRe 1 Rankings on actual practice differ from those on formal attributes Scores of selected Brazilian infrastructure regulators on regulatory governance indexes (0­1) RGI-83 equal weight RGI-28 mean 0.8 RGI-28 One standard deviation 0.7 of RGI-28 mean 0.6 0.5 0.4 0.3 0.2 0.1 0 A (MA) TEL AQ AN ANEEL ARPE ANP AGR (RN) ARCE ANA TESP ANA ARSAL CSPE AGER ASEP ANTT ANT AGERGS AGERBA AGEP AAGIS AGEEL AR ARSEP ARSEP Source: Correa and others 2006. Note: RGI-83 is a summary index measuring 83 attributes of regulatory governance; RGI-28 measures a subset reflecting actual prac- tice. For the full names of the agencies, see Correa and others (2006). nature of effective regulatory enforcement suggests Framework Report on Infrastructure in Mexico." World Bank, a rationale for multilateral institutions becoming Washington, DC. more systematically involved in this effort. Smith, Warrick. 1997a. "Utility Regulators--Decisionmaking Structures, Resources, and Start-Up Strategy." In The Private Sector in Infrastructure: Strategy, Regulation, and Risk. Washington, DC: World References Bank and International Forum for Utility Regulation, 29­32. Baron, David. 1989. "Design of Regulatory Mechanisms and ------. 1997b. "Utility Regulators--Roles and Responsibilities." Institutions." In Handbook of Industrial Organization, vol. 2, ed. In The Private Sector in Infrastructure: Strategy, Regulation, and Risk. Richard Schmalensee and Robert Willig. Amsterdam: North- Washington, DC: World Bank and International Forum for Utility Holland. Regulation, 25­28. Berg, Sanford V. 1998. "Lessons in Electricity Market Reform: ------. 1997c. "Utility Regulators--The Independence Debate." Regulatory Processes and Performance." Electricity Journal 11 (5): In The Private Sector in Infrastructure: Strategy, Regulation, and Risk. 13­20. Washington, DC: World Bank and International Forum for Utility ------. 2001. "Infrastructure Regulation: Risk, Return, and Regulation, 21­24. Performance." Global Utilities 1 (May): 3­10. Stern, Jon, and Stuart Holder. 1999. "Criteria for Assessing the Brown, Ashley C., Jon Stern, and Bernard W. Tenenbaum. 2006. Performance of Regulatory Systems." Utilities Policy 8 (1): 33­50. Handbook for Evaluating Infrastructure Regulatory Systems. Washington, Notes DC: World Bank. This note is based on a World Bank­PPIAF report (Correa and others Correa, Paulo, Carlos Pereira, Bernardo Mueller, and Marcus 2006), also forthcoming in a shorter version in the Quarterly Review Melo. 2006. Regulatory Governance in Infrastructure of Economics and Finance. The report was prepared while the author GRIDLINES Industries: Assessment and Measurement of Brazilian was in the Finance, Infrastructure, and Private Sector Unit of the Regulators. Trends and Policy Options 3. Washington, World Bank's Latin America and the Caribbean Region. The author DC: World Bank and PPIAF. http://www.ppiaf.org/ gratefully acknowledges comments by Eric Groom and Clemencia Gridlines share emerging knowledge sections/Trendsandpolicy.htm. Torres de Mästle and editorial assistance by Jasna Vukoje. on public-private partnership and give an overview of a wide selection of projects from Guasch, J. Luis, and Pablo T. Spiller. 1999. 1. The survey questionnaire included 106 questions divided among various regions of the world. Past notes can be Managing the Regulatory Process: Design, four sections: autonomy, decision making, decision tools, and found at www.ppiaf.org/gridlines. Gridlines are a Concepts, Issues and the Latin American and accountability. High-level officials in each regulatory agency, most Caribbean Story. Washington, DC: World often its president or directors, responded to the questionnaire. The publication of PPIAF (Public-Private Infrastructure PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY Bank. survey was conducted in April­June 2005. Advisory Facility), a multidonor technical assistance facility. Through technical assistance and knowledge Noll, Roger G. 2001. "Progress in 2. Those not hired through public exams either hold temporary dissemination PPIAF supports the efforts of policy Telecommunications Reform in Mexico." contracts or belong to another government body (in several cases Background paper prepared for "Country the privatized state enterprise). makers, nongovernmental organizations, research institutions, and others in designing and implementing strategies to tap the full potential of private involvement in c/o The World Bank, 1818 H St., N.W., Washington, DC 20433, USA infrastructure. The views are those of the authors and do PhoNe (+1) 202 458 5588 FAX (+1) 202 522 7466 not necessarily reflect the views or the policy of PPIAF, PUBLIC-PRIVATE INFRASTRUCTURE ADVISORY FACILITY geNeRAl eMAIl ppiaf@ppiaf.org weB www.ppiaf.org the World Bank, or any other affiliated organization.