Mozambique 51326 Investment Climate Assessment - 2009 Su tai i 9 n de i 9 G owth 'ntIWORLD lANK MOZAMBIQUE 145 51326-MZ Mozambique Investment Climate Assessment October 2009 World Bank Regional Program for Enterprise Development (RPED) Africa Finance and Private Sector (AFTFP) REPUBLIC OF MOZAMBIQUE--FISCAL YEAR (January 31--December 31) CURRENCY EQUIVALENTS (Exchange Rate Effective as of 09/14/2009) Currency Unit = Mozambique Metical US$ 1.00 = 27.18 ABBREVIATIONS AND ACRONYMS ANE Administração Nacional de Estradas (National Road Administration) ATM Automated Teller Machine BCI Banco Comercial e de Investimentos (Commercial and Investment Bank) BIM Banco Internacional de Moçambique (International Bank of Mozambique) CEM Country Economic Memorandum CFM Portos e Caminhos de Ferro de Moçambique, E.P.(Mozambique Railroad) CNELEC National Council of Electricity COMTRADE United Nations Commodity Trade Statistics Database CPI Corruption Perception Index DNEAP National Directory of Studies and Policy Analysis EdM Electricity of Mozambique GCI Global Competitiveness Index GDP Gross Domestic Product GSP Generalized System of Preferences HCB Cahora Bassa Hydroelectric HIV/AIDS Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome ICA Investment Climate Assessment IFC International Finance Corporation IGEPE Institute for the Management of State Holdings IMF International Monetary Fund INCM Instituto Nacional das Comunicações de Moçambique (National Institute of Communications of Mozambique) ISO International Standards Organization MSE Micro and Small Enterprises MW Megawatt ODA Official Development Assistance OLS Ordinary Least Squares PARPA Poverty Reduction Support Strategy POS Point of Sale RF Road Fund SADC Southern Africa Development Community SME Small and Medium Enterprise TdM Telecomunicações de Moçambique (Telecommunications of Mozambique) TFP Total Factor Productivity TVET Technical and vocational and education and training USAID United States Agency for International Development WDR World Development Report WTO World Trade Organization ii TABLE OF CONTENTS ACKNOWLEDGMENTS ........................................................................................................................ VII EXECUTIVE SUMMARY .....................................................................................................................VIII Main Findings....................................................................................................................................................x Moving Forward.............................................................................................................................................xiv Summary of Selected Policy Recommendations ...........................................................................................xvii 1. RECENT ECONOMIC PERFORMANCE ...................................................................................... 1 Recovery and Growth........................................................................................................................................1 Export Performance...........................................................................................................................................4 Trade .................................................................................................................................................................7 Moving Forward................................................................................................................................................8 2. BUSINESS CLIMATE...................................................................................................................... 10 Introduction .....................................................................................................................................................10 The Main Obstacles according to Manager's Perceptions ...............................................................................13 Indirect Costs...................................................................................................................................................17 Crime...............................................................................................................................................................18 Tax Rates.........................................................................................................................................................19 Corruption .......................................................................................................................................................21 Petty Corruption ..............................................................................................................................................23 Grand Corruption.............................................................................................................................................24 Conclusions .....................................................................................................................................................25 3. INVESTMENT CLIMATE, PRODUCTIVITY, AND EXPORTS............................................... 27 Impact of Investment Climate Variables on Firm Productivity .......................................................................27 Mozambique Only Regressions ..................................................................................................................27 Pooled Sample (Mozambique and Comparator Countries)........................................................................28 Summary of Results and Comparison of Mozambique Versus Pooled Sample...........................................28 Exports and Exporter Premium .......................................................................................................................29 Labor Productivity...........................................................................................................................................30 Conclusions .....................................................................................................................................................34 4. ACCESS AND COST OF FINANCE .............................................................................................. 36 The Mozambican Financial System.................................................................................................................36 Cross-Country Comparisons............................................................................................................................38 Access to Finance ............................................................................................................................................41 Characteristics of Credit Constrained Firms....................................................................................................43 Why Has Access to Finance Decreased? .........................................................................................................45 Conclusions .....................................................................................................................................................46 Banking Sector Competitiveness ................................................................................................................47 High Collateralized Lending ......................................................................................................................47 Lack of Audited Accounts ...........................................................................................................................47 5. INFORMALITY IN MOZAMBIQUE ............................................................................................ 49 Assessing the Impacts of Regulation, Governance, and Structural Factors on Informality .............................49 Firm-Level Data: The Enterprise Survey Results ............................................................................................54 Microenterprises and Informality ....................................................................................................................55 The Investment Climate for Microenterprises .................................................................................................57 Microenterprises and Financing ................................................................................................................61 Conclusions .....................................................................................................................................................63 6. INFRASTRUCTURE........................................................................................................................ 65 Electricity ........................................................................................................................................................68 Telecommunications........................................................................................................................................74 Transport and Logistics ...................................................................................................................................77 Hardware of Logistics--Transportation Infrastructure .............................................................................78 Software of Logistics--Customs and Trade Regulation .............................................................................80 iii Conclusions .....................................................................................................................................................83 7. LABOR MARKETS AND HUMAN CAPITAL............................................................................. 87 Labor Market Characteristics ..........................................................................................................................87 Cross Country Comparisons ............................................................................................................................88 Labor Regulation.............................................................................................................................................90 Decomposition of Recent Employment Growth..............................................................................................93 A Look at Technology and Human Capital in the Mozambican Firm .............................................................96 Conclusions .....................................................................................................................................................99 8. REFERENCES ................................................................................................................................ 101 9. TECHNICAL ANNEXES............................................................................................................... 105 Annex I: Sampling Methodology ....................................................................................................... 105 Annex II: Informality Correlations.................................................................................................... 108 Annex III: Modeling Demand for External Finance and Credit Constraints .................................... 110 Results ...........................................................................................................................................................113 Annex IV: Productivity and Exports in Mozambican Firms.............................................................. 120 Methodology .................................................................................................................................................120 Investment Climate and Total Factor Productivity (TFP)..............................................................................120 Results ...........................................................................................................................................................123 Summary of Results and Comparison Mozambique Versus Pooled Sample .................................................127 Exports and Exporter Premium .....................................................................................................................127 Results ...........................................................................................................................................................128 Figures and Table for Labor Productivity and Employment Analysis...........................................................133 Annex V: Gender and Entrepreneurship in Mozambique.................................................................. 136 Gender Differences in Firm Productivity ......................................................................................................136 Gender Differences in Business Climate Constraints ....................................................................................137 Access to Finance ..........................................................................................................................................139 Crime, Theft, and Disorder............................................................................................................................141 Gender Differences in Labor Market Participation........................................................................................142 Conclusions ...................................................................................................................................................144 LIST OF BOXES Box 1. What Is an Investment Climate Assessment and How Does it Differ from the Doing Business Reports?............................................................................................................................................... i Box 1.1. Government Initiatives to Improve the Business Environment........................................................ 1 Box 2.1. The Enterprise Survey in Mozambique ............................................................................................ 1 Box 3.1. Olley and Pakes Decomposition....................................................................................................... 1 Box 5.1. Different Views on the Costs and Benefits of Informality ............................................................... 1 Box 5.2. Tax Reform for Micro and Small Firms........................................................................................... 1 Box 7.1. Labor Reforms in Mozambique ....................................................................................................... 1 LIST OF FIGURES Figure 1. Cross-Country Firm Performance Relative to Mozambique .......................................................... ix Figure 2. Main Obstacles ............................................................................................................................... xi Figure 1.1. War, Recovery and Growth .......................................................................................................... 1 Figure 1.2. Sectoral Output Share and Real Growth....................................................................................... 2 Figure 1.3. Contribution to GDP Growth by Economic Sector ...................................................................... 2 Figure 1.4. Growth Decomposition................................................................................................................. 3 Figure 1.5. Contributions to GDP Growth ...................................................................................................... 3 Figure 1.6. Labor Productivity (Normalized to Agriculture) .......................................................................... 4 Figure 1.7. Export Performance, 1994­2008 .................................................................................................. 6 Figure 1.8. Mozambique Main Exports, 2006 ................................................................................................ 6 Figure 1.9. Concentration of Exports.............................................................................................................. 7 iv Figure 2.1. Main Obstacles ........................................................................................................................... 14 Figure 2.2. Breakdown of Taxes, International Comparison ........................................................................ 20 Figure 2.3. Perceptions on Tax Compliance ................................................................................................. 21 Figure 2.4. Corruption Perceptions Index ..................................................................................................... 22 Figure 2.5. Control of Corruption Index ....................................................................................................... 22 Figure 2.6. Percentage of Firms Requested to Pay Bribes when Applying for Public Licenses/Services .... 23 Figure 2.7. Corruption in Tax Inspections--Cross-Country Comparison .................................................... 24 Figure 2.8. Favoritism in Decisions by Government Officials ..................................................................... 24 Figure 2.9. Informal Payments in Public Procurement--Cross-Country Comparison ................................. 25 Figure 3.1. Potential Gains from Improvements in IC Indices ..................................................................... 29 Figure 3.2. Labor Productivity (value added per worker, US$).................................................................... 31 Figure 3.3. Labor Productivity Decomposition--Olley and Pakes............................................................... 33 Figure 4.1. Percentage of Firms with Overdraft Facility/Lines of Credit or Loans (2003-08) ..................... 38 Figure 4.2. Sources of Working Capital Financing (2003-08)...................................................................... 39 Figure 4.3. Sources of Investment Financing (2003-08)............................................................................... 40 Figure 4.4. Collateral Requirements--Incidence and Burden (2003-08)...................................................... 40 Figure 5.1. Predicted Informality in Mozambique as Percentage of Actual Indicator .................................. 53 Figure 5.2. Predicted/Actual Differences in Informality for Mozambique and Chile................................... 53 Figure 5.3. Contribution to Differences in Predicted Informality, Mozambique and Chile.......................... 54 Figure 5.4. Percentage of Top Managers with Education beyond Secondary............................................... 56 Figure 5.5. Indirect Costs of Bribes by Firm Size (as percentage of sales)................................................... 56 Figure 5.6. Major or Very Severe Obstacles for Operations of Microenterprises......................................... 58 Figure 5.7. Microenterprise Financial Characteristics .................................................................................. 62 Figure 5.8. Reason for Not Applying for Line Credit or Loan ..................................................................... 63 Figure 6.1. Infrastructure Costs as Share of Total Sales ............................................................................... 66 Figure 6.2. Bilateral and Multilateral ODA .................................................................................................. 66 Figure 6.3. ODA Transfers to Mozambique by Sector ................................................................................. 67 Figure 6.4. Private Investment in Infrastructure, 1997­2006........................................................................ 68 Figure 6.5. Evolution in the Coverage of Electricity Services...................................................................... 70 Figure 6.6. Electricity of Mozambique--Production and Productivity ........................................................ 70 Figure 6.7. Electricity as a Business Constraint............................................................................................ 71 Figure 6.8. Power Outages--Frequency and Losses..................................................................................... 73 Figure 6.9. Generator Ownership.................................................................................................................. 74 Figure 6.10. (a) Average Time to Connect to Electricity, and (b) Electricity Services as Percentage of Total Expenditure...................................................................................................................... 74 Figure 6.11. Telecommunications--(a) Fixed lines per 100 Inhabitants; (b) Mobile Subscribers per 100 Inhabitants; (c) Internet Users per 100 Inhabitants ................................................................... 75 Figure 6.12. Telecommunications as Business Constraint............................................................................ 76 Figure 6.13. E-mail Usage and Web Site Ownership ................................................................................... 77 Figure 6.14. Average Days of Waiting to Connect a Telephone .................................................................. 77 Figure 6.15. Transportation as Business Constraint...................................................................................... 79 Figure 6.16. Cost Associated with Transportation: Product Losses, Inadequate Roads, and Direct Costs... 80 Figure 6.17. Trade across Borders Index and Logistics Performance Index................................................. 81 Figure 6.18. Customs and Trade Regulation as Business Constraint............................................................ 82 Figure 6.19. Share of Total Sales.................................................................................................................. 83 Figure 7.1. Average Educational Attainment of Typical Production Worker (by firm size) ........................ 89 Figure 7.2. Median Monthly Compensation of Production Workers (by firm size, in US$) ..................... 89 Figure 7.3. Unit Labor Costs......................................................................................................................... 90 Figure 7.4. Labor Regulation Rankings ........................................................................................................ 91 Figure 7.5. Unionization Rate (by firm size, %) ........................................................................................... 92 Figure 7.6. Employment Change between 2003 and 2006 ........................................................................... 94 Figure 7.7. Continuing and Entering Firms by Regions ............................................................................... 95 Figure 7.8. Percentage of Firms that Offer Formal Training Programs ........................................................ 97 Figure 7.9. Percentage of Full-Time Workers that Received Training ......................................................... 97 Figure 9.1. Labor Productivity Decomposition--Olley and Pakes (by firm size) ...................................... 133 Figure 9.2. Histograms of Log TFP in Manufacturing (2006).................................................................... 134 v Figure 9.3. Male- and Female-Owned Firms Rating Investment Climate as Major or Very Severe Constraint ................................................................................................................................ 138 Figure 9.4. Access to Finance Is Perceived to Be a Serious Constraint...................................................... 139 Figure 9.5. Crime, Theft, and Disorder as a Constraint to Business ........................................................... 141 LIST OF TABLES Table 1. Most Often Cited Obstacles, 2003­08 ............................................................................................ xii Table 1.1. Discoveries between 1994 and 2006.............................................................................................. 5 Table 2.1. Mozambique Doing Business Indicators, 2009 and 2010............................................................ 12 Table 2.2. Firms Reporting Major or Very Severe Business Constraints (percentage) ................................ 13 Table 2.4. Most Often Cited Constraints for Business across African Comparators .................................... 16 Table 2.5. Most Often Cited Constraints for Business across Asian Comparators ....................................... 17 Table 2.6. Indirect Costs as a Percentage of Annual Sales (weighted averages) .......................................... 17 Table 2.7 Indirect Costs as Percentage of Annual Sales (cross country comparisons) ................................. 18 Table 4.1. Characteristics of Loans and Credit Lines in Mozambique ......................................................... 41 Table 4.2. Use of Bank Financing in 2003 and 2008, by Firm Size ............................................................. 42 Table 4.3 Current Financial Market Participation Characteristics by Firm Size........................................... 42 Table 4.4. Investment Financing................................................................................................................... 43 Table 5.2 Firm-Level Correlates of Perceptions on Tax Evasion ................................................................. 55 Table 5.3 Location of Formal and Informal Microenterprises ...................................................................... 57 Table 5.5. Sources of Working Capital Funding........................................................................................... 61 Table 6.1. Power Outage Indicators, 2003 and 2008 .................................................................................... 71 Table 7.1. Size and Sector Distribution of Sampled Firms ........................................................................... 87 Table 7.2. Percentage of Permanent Workforce Paid the Minimum Legal Wage ........................................ 92 Table 7.3. Full-Time Employment Growth................................................................................................... 93 Table 7.4. Technology and Human Capital .................................................................................................. 98 Table 7.5. Technology and Human Capital and Firm Characteristics (Manufacturing Firms Only) ........... 99 Table 9.1. Sample Distribution by Region, Sector, and Firm Size ............................................................. 106 Table 9.2. External Finance Demand and Credit Constraint Definitions.................................................... 110 Table 9.6. Credit Demand and Constraints by Size, Legal Structure, Location and Sector........................ 116 Table 9.9. Firm Performance and Investment Climate Variables--Mozambique Only ............................. 123 Table 9.10. Firm Performance and IC Variables--Mozambique and Comparator Countries .................... 125 Table 9.11. Productivity and Wage Exporter Premium in Mozambique and Comparator Countries ......... 128 Table 9.12. Summary Statistics for Investment Climate Variables ............................................................ 129 Table 9.14. Partial Correlations between Performance, Investment Climate and Control Variables.......... 132 Table 9.15. Partial Correlations (Multiple Linear Regression) ................................................................... 135 Table 9.16. Size of Female-Owned Firms .................................................................................................. 136 Table 9.17. Performance in Male- and Female-Owned Firms .................................................................... 136 Table 9.18. Female-Owned Firms Are More Likely to Be Located in the Retail Sector............................ 139 Table 9.19. Most Financing Comes from Internal Funds or Retained Earnings ......................................... 140 Table 9.20 Security Costs and Losses Due to Theft ................................................................................... 141 Table 9.21. Female Workers in Manufacturing Firms ................................................................................ 142 Table 9.22. Cross-Sector Comparisons: Number of Female Employees and Average Monthly Compensation ........................................................................................................................... 142 Table 9.23. Full-Time Permanent Workforce Paid Minimum Wage and Expenditures on HIV/AIDS Programs and Activities ............................................................................................................ 143 Table 9.24. Educational Attainment in Firms with Larger Versus Smaller Female Workforce ................. 143 Table 9.25. Absenteeism Due to Sickness and HIV/AIDS ......................................................................... 144 vi ACKNOWLEDGMENTS This report was prepared by a team composed of José Guilherme Reis, Mazen Bouri, Samuel Munzele Maimbo and Cristian Quijada. The team had inputs from Professor John Rand (University of Copenhagen and Ministry of Planning and Development, government of Mozambique, chapter on Finance), Fabiano Bastos (World Bank, chapters on Labor and Productivity), Barbara Cunha (World Bank, chapter on Infrastructure), Michaela Weber and Sina Grasmann (World Bank, appendix on Gender), Sidney Nakahodo (World Bank, chapter on Informality) and Magueye Dia (World Bank, chapter on the Business Environment). The productivity analysis was prepared by Veronica Alaimo and Cristian Quijada. The report was prepared under the guidance of Michael Baxter and Gerardo Corrochano and benefited from comments made by Susan Hume, Gregor Binkert, Young Kim, Antonio Nucifora, Wendy Hughes, Jose Chembeze, and José Luis Guasch, and from peer reviewers Mary Hallward-Driemeier, Giuseppe Iarossi, and Paulo Correa. The team acknowledges important contributions received from officials of the government of Mozambique. We would like to thank especially His Excellency Antonio Cruz, National Director at the Ministry of Planning and Development, for his insights and very helpful comments on a preliminary draft of the report. We would also like to thank His Excellency Antonio Fernando, Dr. Domingas Muchine, and her team at the Ministry of Industry and Commerce for useful comments and insights, as well as Mr. Adbul Carimo at the Legal Reform Unit. Preparation of the report benefited from an intense exchange of ideas with development partners and private sector representatives based in Mozambique, especially with participants of the Private Sector Working Group. The dissemination of the study was done through three workshops in three cities: Maputo, Beira, and Nampula. Overall the workshops were quite successful and more than 100 participants, from Government (national and local), private sector, academia and donors attended the workshops. The debates centered on mechanisms to improve access to finance small and medium firms and policies to address the business climate and the impact of informality. In the workshops outside Maputo, discussion also focused on the effectiveness and problems of implementation of the one stop shop experience, as well as on the impact on businesses of the new labor legislation. Participants tended to agree that while reforms in Mozambique are on track, implementation has been slow and should be accelerated. We would like to thank all the participants, especially the Minister of Industry and Commerce, Mr. Antonio Fernando, who participated in the workshop in Maputo. The findings and views expressed here are exclusively those of the World Bank and do not necessarily represent the views of the government of Mozambique. vii EXECUTIVE SUMMARY Mozambique's recent history is a rare example of a successful post-conflict recovery and economic takeoff. Emerging from decades of economic stagnation and decline, a consequence first of a failed socialist economic experience and then of a vicious civil war that only ended in 1992, the country has achieved a commendable degree of political stability. This has been accompanied by prudent and stable economic policy continuity, as well as coordinated and ever more efficient use of substantial international aid. These factors have contributed to sustained economic growth that averaged 7.8 percent between 1992 and 2006. Furthermore, this growth has been "pro-poor": increasing output has been accompanied by real and significant decreases in poverty levels, with the poverty headcount index declining from 69 percent in 1997 to 54 percent in 2003. In spite of these achievements, sustained and broad-based growth is still a challenge for Mozambique. Some of the key drivers of shared economic growth present in the recent past are bound to fade over time--the catch-up effect associated with the end of the war as well as the current volumes of development partners' aid--and need to be replaced by long-run productivity growth. Moreover, growth in the past decade has been concentrated at one end of the productive spectrum: foreign-owned, capital-intensive, and export-oriented "mega-projects". These projects are dependent on world commodity prices, have had limited impact on employment creation and productivity spillovers, and until recently, have also brought limited fiscal benefits. At the other end of the productive spectrum are the vast majority of firms, primarily small and medium enterprises (SME), which sell mostly to the local market, face severe resource constraints, and contribute modestly to economic growth and exports. Sustained and broad-based growth in output and employment requires a diversification of exports and production and the creation of a better business environment for greater participation of this part of the Mozambican private sector in the country's economic activity. Achieving sustained, broad-based growth in Mozambique requires a strategy that maximizes Mozambique's export potential through productivity improvements across firms in the country. The Mozambican economy is small, relatively open and integrated economy, and is strategically located as a conduit for regional trade. Therefore, a strategy that focuses on maximizing Mozambique's capacity for export can help to bring about higher growth in the private sector.1 However, as will be shown in this report, Mozambican firms do not have the productivity levels that allow them to compete in global markets (see Figure 1). Improvements in key aspects of the investment climate in the country can help firms raise their competitiveness and dynamism, allowing them to link to global markets. The government of Mozambique has already recognized this and has placed the private sector as the main engine for investment, growth, and employment, as shown by the second Poverty Reduction Support Strategy (PARPA II) covering 2006­09. 1 See the forthcoming Mozambique CEM (World Bank 2009) for a more detailed discussion. viii Figure 1. Cross-Country Firm Performance Relative to Mozambique S W V A P W TFP W / M A I M S A Z Source: Author's calculations based on Enterprise Surveys 2003-08 Note: Results normalized so that Mozambique performance equals 100. Manufacturing firms only, micro firms excluded. Results are controlled by sector, firm size and other variables (see Table 9.10). Based on the Enterprise Survey2 results for Mozambique, this report assesses the main obstacles to achieving an investment climate that supports private sector growth and provides policy options for improving the business environment and increasing competitiveness with the goal of achieving sustained and broad-based growth. The focus is on microeconomic constraints and reforms where, according to a recent World Bank report, most of the challenges for sustainable growth are concentrated.3 Therefore, the analysis presented in this report should be of interest to policy makers, academics, non-governmental organizations and representatives of the private sector involved in the policy dialogue in the country. 2 http://www.enterprisesurveys.org 3 ""The growth of GDP may be measured up in the macroeconomic treetops, but all the action is in the microeconomic undergrowth, where new limbs sprout, and dead wood is cleared away." Most growth- oriented policies and reforms are designed to foster this microeconomics of creation and destruction, and, crucially, to protect people who are adversely affected by these dynamics." Commission on Growth and Development, 2008. ix Box 0. What Is an Investment Climate Assessment and How Does it Differ from the Doing Business Reports? The main objective of the Investment Climate Assessment (ICA) is to develop a better understanding of the constraints to investment and of the key elements that affect sustained productivity growth in a given country. The ICA uses as its primary data source an Enterprise Survey which, for Mozambique, was undertaken in the first half of 2008 and covered 599 micro, small, medium, and large firms (see Box 3 for further details on sample distribution). The ICA can add to the policy debate by improving the quantitative basis for discussion and, where possible, estimating effects of the constraints on firm growth and productivity. In particular, the proposed ICA seeks to: (i) measure, in a standardized way, the investment climate conditions in Mozambique; (ii) develop a set of indicators at the regional level that allow for regional comparison within the country; (iii) provide comparisons of conditions with those prevailing in other countries and regions; and (iv) identify the features of the investment climate that matter most for productivity and, therefore, GDP growth. In the simplest terms, the investment climate establishes the rules of the game and the environment within which all firms must operate. It includes such items as: (i) macroeconomic policies regarding interest and exchange rates, inflation, and both the tax and trade regimes; (ii) the quantity and quality of infrastructure services; (iii) quality of governance, including (a) rules and regulations regarding entry, exit, and access to land and utilities; (b) administrative regulations over operations including labor policy, operational permits and inspections (health, fire, social security, and so forth), which, if excessive, can raise costs and push firms into the shadow economy, or, if politically motivated, can favor one enterprise at the expense of its competitors; (c) corruption, bribery, and other forms of official harassment; and (d) crime and violence; (iv) access to and cost of finance; (v) access to new technologies and protection of intellectual property rights; (vi) the level and quality of workers education and the opportunities and incentives for training; and (vii) general levels of competition. The Doing Business indicators deal with regulations that represent only one dimension of the overall investment climate in which firms operate. There are significant methodological differences: while ICA data is based on firm-level surveys, Doing Business collects information from expert informants in each country, mostly lawyers and accountants. Although the latter can provide reliable information, it is not based on a representative sample and some data points are generated by just one or two firms. On the other hand, the strengths of the Doing Business methodology include its cost-effectiveness, simplicity, and the quick and direct link to reforms. Despite differences in approach and the depth and range of the investment climate data, both the ICA and Doing Business provide complementary information and analysis. Consolidating the findings, the Bank Group works collaboratively to help clients develop a more friendly business environment in the country. Main Findings Improvements to the business environment and increased access to finance are the most critical aspects to firm growth in Mozambique identified in this report. Despite recent progress, the business environment for the Mozambican enterprise sector is still in many ways problematic. Based on econometric evidence as well as on business perceptions and quantitative data, this study indicates that while all aspects of the investment climate are important, reform priorities should focus on increasing access to finance and improving the business environment as discussed below. Mozambican businesses identified competition from the informal sector as the top constraint to their operations. While unfair competition is well known as one of the undesirable consequences from informality, this result in Mozambique is somewhat x surprising, as concerns with access to finance and infrastructure aspects normally top the list of most important obstacles for businesses in Africa. Our interpretation for this result is that informal competition is partly a consequence of challenges in the regulatory framework for business operation and the quality of governance in Mozambique. This conclusion is supported by results from cross-country analysis of determinants of informality, as well as by survey results showing that crime, corruption, and tax rates were also cited among the top obstacles by firms. Indeed, there is now considerable evidence showing that the varying incidence of informality found across countries with similar levels of per capita income is significantly linked not only to the nature of business regulations but also to the quality of governance. The analysis presented in this report suggests that the scope for action is wide: it should include policies that encourage entry in the formal sector, including taxation and labor issues, as well as policies aimed at improving governance, including the strengthening of institutions and rule of law. Furthermore, as informality tends to prevail among smaller firms, policies targeted to improve the business environment for microenterprises and for SME are particularly important.4 Figure 2. Main Obstacles Practices of informal competitors 49% Access to finance 42% Tax rates 36% Crime, theft, and disorder 31% Transportation 27% Electricity 25% Corruption 15% % of firms that report obstacle among top three constraints Source: Enterprise Survey 2008 It is also useful to compare perceptions from this survey with the figures from a previous survey, carried out in 2003. The figures in Table 1 show that there were important changes in the ranking of obstacles: Electricity dropped from 2nd to 6th position, and labor regulation from 11th to 15th. On the other hand, crime, tax rates, and transport moved up in the rankings, being perceived as more important constraints in this last survey. 4 In Latin America, a region where informality is a ubiquitous phenomenon, Chile stands out as an exception where a policy mix encompassing human capital investments and sensible tax and pension reforms have led to a significant growth record, a reduction in poverty, high trust in institutions, and relatively low levels of informality. xi Table 1. Most Often Cited Obstacles, 2003­08 Ranking Ranking Major or Severe Obstacles to Business in 2008 in 2003 Practices of informal competition 1 5 Access to finance 2 1 Crime 3 8 Tax rates 4 7 Corruption 5 3 Electricity 6 2 Transport 7 14 Tax administration 8 10 Workforce education 9 12 Licensing & permits 10 13 Customs & trade regulations 11 9 Access to land 12 15 Telecommunications 13 16 Political instability 14 Labor regulation 15 11 Courts 16 Source: Enterprise Surveys 2003 and 2008 Access to finance was the second most often cited obstacle to business operations by Mozambican firms and is confirmed as a key obstacle through empirical analysis linking the access to finance indicators to firm-level productivity in Mozambique.5 The empirical analysis of this Investment Climate Assessment (ICA), taking into consideration demand factors, found a highly credit constrained environment: more than two-thirds of the sample was credit constrained. The percentage increased to 82 percent when the sample was limited to credit-demanding firms that (i) had applied for and been denied credit or (ii) had not applied for credit due to reasons such as "application procedures too complex," "collateral requirements unattainable," or "possible loan size and maturity insufficient." Survey results also showed that lending is highly collateralized and that financing for investment is largely based on retained earnings. Furthermore, the percentage of firms accessing lines of credit is not only low at 13 percent, but also has declined from previous surveys in 1997 (35 percent), 2002 (29 percent), and in 2006 (26 percent)--especially for SMEs. There are several possible explanations for the reduction as something more than an artifact of different representative samples. First, this survey may denote a better, albeit lower, credit quality portfolio, because the privatizations in the financial and industrial sectors may have resulted in more selective lending. Thus, while the volume of loans may have increased, it was to a smaller number of better-managed, privately owned firms. Second, the reduction could point to the recent regulation Aviso 5/2005.6 The regulation resulted in a 5 While the cost of finance is affected by both macroeconomic and microeconomic factors, the analysis presented in this report focuses primarily on the micro-level constraints to finance. Macroeconomic aspects affecting interest rates are dealt with in other Bank reports, such as the recently concluded 2009 joint IMF- World Bank Financial Sector Assessment. 6 The regulation requires banks to set aside a 50 percent loan-loss provision against bad debts for loans denominated in foreign currency that are issued to nonexporters. Aviso 5/2005 is intended to discourage xii decrease of aggregate lending in foreign currency from 60 percent in early 2005 to 36 percent in late 2006. Several firms now had to borrow in meticais at a higher interest rate (around 25 percent) rather than at an interest rate of about 10 percent. According to United States Agency for International Development (USAID) (2007), this interest rate gap was greater than exchange rate risk alone would suggest and Aviso 5/2005 might have reduced credit demand for certain types of firms. Third, but likely to a lesser degree, the improving regulatory, accounting, and provisioning regulations and the entry of international banks into the sector with home country regulators requiring higher standards of credit assessment may have contributed to a reduction in firms with access to easy credit. Finally, there is evidence that overdraft use across firms increased compared to the Enterprise Survey results of 2003. This may suggest that there has been some degree of substitution in how firms finance their operations. Reconstruction of roads and ports has been significant but physical infrastructure including the provision of reliable energy remains a key constraint for businesses. Adequate infrastructure facilitates access to inputs and foreign markets, increases the productivity of existing firms, and gives incentives for the creation of new businesses. After more than 15 years of reconstruction, Mozambique's infrastructure is beginning to approach the level of its regional competitors. Reflecting this improvement, entrepreneurs in Mozambique today are less worried with energy provision, for example, than they were in 2003. Yet firms still face significant losses due to outages and losses of merchandise while in transit. Combining direct and indirect costs (sales losses due to energy outages and breakage during transportation and inventory), the median firm in Mozambique spends on average 22 percent of its sales on infrastructure. Moreover, electricity provision outside of Maputo is still inadequate both in terms of access and quality, especially in peri-urban and rural areas. Results from the econometric analysis confirm the regional disparities in firm performance, as well as infrastructure's current negative impact on firm productivity in Mozambique. A large fraction of firms considers the road transportation system inadequate for their production activities. In addition to high tariffs (direct cost), firms face significant indirect costs imposed by production losses (theft and breakage) during transportation. Finally, despite reform, customs' weak performance still represents an important obstacle for firms trying to access international markets. Another challenge for Mozambique is weak governance structures--institutions, the rule of law and security. Crime and corruption are perceived to be a serious constraint to growth, particularly by exporters and foreign firms. As mentioned earlier, crime was the third most often cited constraint to doing business in Mozambique, a significant increase when compared to the 2003 survey, when it was ranked eighth. One firm out of three rated it as a severe or major constraint for growth. The incidence of crime adds significantly to the costs of doing business in Mozambique, both directly through the loss of plant equipment and product to vandalism or theft and indirectly through the cost of employing security agencies to protect firms' property. Thirty-six percent of firms in Mozambique experienced losses as a result of theft, vandalism, or arson. These losses end up amounting to 2 percent of annual sales. Such costs are considerably higher than costs experienced in most comparator countries, including South Africa (0.61 percent). banks from lending in foreign currency to borrowers who do not have an income stream in that currency, to avoid a potential systemic risk of the Asian financial crisis type. xiii As a consequence of the incidence and concerns over crime, up to 34 percent of firms in Mozambique report the use of security services (equipment, personnel or professional security services). This additional cost of having to pay for security services, on average, amounts to 1.5 percent of sales each year. Altogether, these costs exact a significant toll on businesses activities and have a direct impact on reducing domestic investment as well as in dissuading foreign investment. Finally, there is a complementary agenda (at the macro level) linked to human capital and technology absorption. The country's endowments of human capital are comparable to other countries in the region and lack of skilled workers is not perceived as a major obstacle by firms. However, the empirical analysis done with the survey results shows, not surprisingly, a strong association between firm-level productivity and human capital variables. Moreover, in order to tackle the informality issue, at least in some segments of the informal economy, policies to directly address productivity are required. The fact that informal firms exhibit lower productivity than formal firms, even controlling for observable attributes, is not convincing evidence that there would be a significant growth dividend from reducing informality. Assuming that firms have some knowledge about their current productivity and future growth prospects, such productivity differentials between the formal and informal sectors could be explained by self-selection among firms.7 In other words, policies to address informality may need to go beyond trying to make firms formal and should include specific actions to help lift productivity levels across firms, including informal ones. Moving Forward Creating a long-term reform agenda with a broad-based consensus among the various segments of society can be an important step forward. Substantial reforms are often politically costly in the near term, while the benefits may not be visible for years to come. In this regard, the recommendations made in this report must be seen as but a small part of a broader economic reform process, including regional integration. This Mozambique ICA identifies potential areas of reform that can accelerate the private sector's contribution to higher growth and a reduction in poverty. In order to discuss policy recommendations, the report does not restrict itself to the results of the survey. The survey is essentially a diagnostic tool that helps identify key bottlenecks for firms' growth but it was not designed to devise policy recommendations. Several sources of information are used to complement the analysis, especially other documents prepared by the World Bank. The current crisis poses new risks and challenges for Mozambique. The real sector contagion is already a major concern for policy makers around the world, as the increasingly unfavorable external environment slows economic activity worldwide, leading to lower overall foreign direct investment and trade flows (due to a flight to quality by international investors and to lower export demand), as well as a decrease in remittances. These emerging concerns put a premium on improving the business environment, as moving ahead with key reforms that can reduce the cost of operating 7 If this were the real explanation, then removing barriers to formality would then lead to more formal firms but no change in firm productivity. xiv businesses will be vital to stimulate trade and investments, two of the most important factors to restore growth. Mozambique should continue reforming its regulations for private business activity--particularly for SMEs. The Strategy and Action Plan to Improve the Business Environment, along with the SME strategy approved by the Cabinet, are encouraging indications of a commitment to undertake needed reforms. Additionally, the government has undertaken recent measures related to business start-up and licensing (elimination of minimum capital requirement, streamlining of license requirements for SMEs) as well as the establishment of commercial courts and the simplification of tax and customs requirements for small investors. Lastly, the government has recently introduced a Simplified Tax for Small Contributors, which aims to reduce tax compliance costs by streamlining procedures. Cumulatively, these measures could have a positive impact on the business environment. In addition, there are plans to pass the bankruptcy law and to further strengthen the one-stop shops for business licensing delivery. However, even with the progress that has been made, there is still a need to deepen and broaden the reforms so they involve all affected ministries and also to tackle some fundamental issues that impact the business environment such as labor regulations and land and property registration. Trade facilitation remains a major bottleneck and there is a need to streamline the regulations related to import and export clearances. Furthermore, in order to have an enduring impact, the reforms passed through legislation should be complemented with capacity building of public sector agencies in charge of implementing these reforms. There is also a need for raising awareness and improving consultative processes through strengthened public-private dialogue with business representative organizations. As mentioned earlier, the government's passage of the Business Environment Strategy is a positive sign of commitment and the reform effort should continue to be championed at the highest political level. This can be accomplished by empowering the interministerial committee in charge of implementing the Strategy and Action Plan to Improve the Business Environment so that the targeted reforms are realized as planned across the different ministries. On corruption and governance issues, the country would benefit from modernizing conflict of interest regulations and adopting codes of conduct and instituting ethics committees at different levels of government. A periodic updating of the country's legal and regulatory framework is necessary to ensure greater access to finance for the private sector. Rules on transparency, capital, leveraging, accounting, and auditing must be regularly modernized. Based on the findings of this assessment, the government should consider: (i) revising the recent regulation Aviso 5/2005 (this should help to bring down interest rates for a group of firms unnecessarily affected by this legislation); (ii) exploring an alternative mechanism for collateral-based security, including the use of credit risk guarantee schemes; (iii) enhancing efforts to improve the accounting and auditing standards in the country, given that firms that are audited are less likely to be credit constrained, and that improved standards are likely to yield positive results (assuming there is no reverse causality); and (iv) improving competition among banks by taking steps to promote greater transparency of bank fees and charges, increase financial literacy and encourage the unbundling of financial services. As with all legal and regulatory reforms, the above recommendations should involve working on the statutory framework; professional education and training, including strengthening the capacity of public sector agencies responsible for the xv implementation of reforms and establishing/strengthening mechanisms for the monitoring and enforcement of the applicable standards. Furthermore, regular and closer consultations between the public and private sectors would go a long way to ensure that reforms passed continue to reflect the main constraints faced by entrepreneurs. In infrastructure, the most pressing needs in terms of policy action are concentrated in energy and transport. In the energy sector, Mozambique should continue investing in capacity building and the gradual transition to a fully functional independent regulator. Meanwhile, the Ministry of Energy should periodically review the performance targets for the government-owned electricity utility, Electricity of Mozambique (EdM), and intensify monitoring. The design of additional incentive mechanisms, such as fines for unwarned power outages and an excessive number of complaints, is also an effective tool to enhance service quality. Mozambique's government is currently investing heavily in the extension of the electricity grid to peri-urban areas. Nevertheless, the problem of attracting private investment to construct independent grids in low-density rural areas remains unsolved. Increasing transparency and commitment to contracts will reduce the risks for private investors and attract the necessary resources to the sector. In regards to transportation, the "soft" side seems to deserve more attention. A strategic reduction in the number of procedures and inspections, and the intensification of the ongoing process of automation of border activities can improve the exchange of information with neighboring countries and the monitoring of activities. This would help to shorten times spent at customs without compromising the need for controls and inspections. Finally, on the hardware logistics side, ensuring adequate maintenance and rehabilitation of the transport network and improving coordination between rail and ports can significantly reduce travel costs and turnaround times. Furthermore, investments in network expansion (roads and railroads) should focus on improving the precarious link between the north and the south of the country. Finally, there is a clear need for the government to support efforts to increase firm productivity, especially among SMEs, so that Mozambican firms can link to and compete more effectively in global markets. Important aspects to focus on, that are suggested by the analysis presented here, are employee training and the need for companies to focus on process and quality improvement in their operations. For professional skills, employee training can be promoted by partnerships between higher education and the private sector through on-the-job training. In order to address specific short-term skill shortages, import of skilled labor may still be the most readily available solution. However, as noted in Box 7.1, initial evidence from implementation of the new labor law points to concerns over quotas and qualifications vetting and their impact on businesses' ability to hire foreign expertise. For technical skills more broadly, the strengthening of the technical and vocational education and training (TVET) system can provide workers with an opportunity to upgrade their skills. The current model of delivery needs to be modified from supply to demand-driven to ensure that the system is producing the skills needed to meet the demand of the economy. Current government efforts that are being supported by the World Bank are beginning to address these concerns. Finally, the great improvements achieved in expanding primary education need now to be focused on improving the access to and the quality of post-primary education. As noted by a recent World Bank report (World Bank 2008b), further investments are needed to expand school networks and reduce the costs of education. xvi Summary of Selected Policy Recommendations Investment climate topic Diagnostic Policy recommendations Financial system Analysis of Enterprise Survey data found a highly credit constrained · Revise the recent regulation Aviso 5/2005 (this should environment in Mozambique: more than two-thirds of the sample was help to bring down interest rates for a group of firms credit constrained. Moreover, the percentage of firms accessing lines of unnecessarily affected by this legislation). credit has declined in the past five years. · Explore alternative mechanism for collateral-based Limited competition has resulted in a nondynamic financial system security including the use of credit risk guarantee schemes. with low levels of financial intermediation, high profit margins (helped · Government should consider enhancing its efforts to by excessively high fees on banking services), and a large interest rate improve the accounting and auditing standards in the spread between deposits and loans (to some extent driven by high country as they are likely to yield positive results. reserve requirements). · Improve competition among banks by taking steps to promote greater transparency of bank fees and charges to Most lending in Mozambique is collateralized (around 90 percent of increase financial literacy and to encourage the unbundling loans) and a significant percentage of firms applied for and were denied of financial services. credit or did not apply for credit due to unattainable collateral requirements. Firms without audited accounts are less likely to receive financing. Accounting and auditing standards are weak and, as a consequence, banks look more at collateral potential than corporate financial statements when reviewing prospective borrowers. Business regulation The Enterprise Survey shows that informality, crime and theft, tax · Accelerate implementation of the Strategy and Action Plan rates, and corruption are major constraints to business according to to Improve the Business Environment and implementation firms interviewed. of the SME strategy. International comparisons continue to show that Mozambique's · Adopt codes of conduct and instituting ethics committees business environment remains constrained. Among others, performance at different levels of government. in the Doing Business rankings for Mozambique is particularly poor on · Streamline the business licensing regime. Employing Workers, Dealing with Construction Permits, Registering · Update and modernization of the system of property Property, and Trading across Borders. registry. High rates of informality may be a consequence of institutional · Support the creation of commercial courts. constraints, including lengthy and cumbersome registration procedures · Pass a new bankruptcy law. and difficult interactions with government agencies. · Improve government coordination in implementation of Investment climate topic Diagnostic Policy recommendations and difficult interactions with government agencies. business environment reforms. Infrastructure Electricity Electricity Although EdM experienced improvements in the expansion of services, · Continue investing in the capacity building and the gradual labor productivity, and financial management, the company still transition to a fully functional independent regulator. struggles with high distributional loss and unwarned interruptions. · Periodically review the performance targets for EdM and intensify monitoring. The design of additional incentive Electricity provision is still restricted in access and quality outside mechanisms such as fines for unwarned power outages and Maputo, especially in peri-urban and rural areas. excessive number of complaints is also a helpful tool to enhance service quality. Transportation & Logistics · Reevaluate some aspects of the concession contracts and Despite the recent improvements, transportation still imposes high bidding process. Instead of promoting competition over direct and indirect costs on firms. tariffs, the government should fix tariffs and award Customs requirements are commonly perceived as one of the most contracts to firms proposing the most efficient subsidy relevant in determining border crossing time. Mozambique's border schemes. In addition, increasing transparency and also experiences a shortage of skilled labor for operational, legal, and commitment to contracts will reduce the risks for private supervising functions. investors and attract the necessary resources to the sector. Transportation & Logistics · Ensure adequate maintenance and rehabilitation of transport network, improve the coordination between rail and ports, and invest on network expansion (roads and railroads) to improve the strategic link between the northern and southern regions of the country. · Reduce the number of procedures and inspections, and intensify the ongoing process of automation of border activities to improve the exchange of information and to facilitate the monitoring of activities in a shorter amount of time. xviii Recent Economic Performance Recovery and Growth By most accounts, Mozambique's recent history has been an example of a successful post-conflict recovery and economic takeoff. Emerging from decades of economic stagnation and decline, a consequence first of a failed socialist economic experience and then of a vicious civil war that only ended in 1992, the country has been able to achieve an unusual degree of political stability for the African context (Figure 1.1). This has been accompanied by prudent and stable economic policy continuity, as well as coordinated and ever more efficient use of substantial international aid. These factors have contributed to sustained economic growth that averaged 7.8 percent between 1992 and 2006. Furthermore, this growth has been "pro-poor". Increasing output has been accompanied by real and significant decreases in poverty levels with the poverty headcount index declining from 69 percent in 1997 to 54 percent in 2003. Figure 0.1. War, Recovery and Growth Mozambique Real GDP 1970­2004 (billions of metical at constant 1980 prices) 300.0 Real GDP 1986: start of Economic Rehabilitation Programme 250.0 1981: intensification of civil war 200.0 1992: End of civil war 150.0 1975: Independence from Portugal 100.0 1970 1974 1978 1982 1986 1990 1994 1998 2002 Source: Arndt, Jones, and Tarp (2006) Prudent fiscal policies combined with tight monetary policies have been key to stabilizing the economy and thus to laying out the macro fundamentals for growth recovery. The government of Mozambique has been consistently pursuing sound fiscal, monetary, and external policies, demonstrating clear understanding of the benefits associated with a solid underlying macroeconomic framework. Although reform efforts began as early as 1987, more sustained improvements were not possible until peace was achieved in 1992, ending more than a decade of civil 1 war. The initial period of peace allowed for agricultural output to grow as refugees returned to their villages, devastated infrastructure was rebuilt, agricultural extension programs were revived and the full impact of earlier price and product movement liberalization took effect.8 With continuing political stability and deepening of economic reforms, new investments were attracted at first in agriculture, where export-oriented production of cash crops began to expand. Since 1998, foreign direct investment in large- scale, capital-intensive extraction projects--the so called megaprojects--have led to a significant increase in output and exports and have put the country on the international map of mining and resource-related companies. The megaprojects have also helped the construction sector to experience significant growth rates during the same period. The structure of output growth has changed. While agriculture has seen above- average growth rates for the region since the early 1990s (5.4 percent annual average growth rate 1996­2003), its share in total output has decreased, going from 34.4 percent in 1996 to 27.3 percent in 2003. Services' share of total Gross Domestic Product (GDP) has increased slightly while industry has experienced significant growth in total output participation, going from 16 percent in 1996 to 26.9 percent in 2003. Specifically, gas and electricity and the construction sectors have all experienced double-digit growth rates (1993­2006) related to the megaprojects, albeit all starting from low levels of contribution to GDP (see Figures 1.2­1.4). Figure 0.2. Sectoral Output Share and Figure 0.3. Contribution to GDP Real Growth Growth by Economic Sector 20 0.60 60 0.50 50 15 Real Gorwth % Sectoral Share 0.40 40 0.30 30 10 0.20 20 5 0.10 10 0.00 0 0 gas, water Agriculture Mining Manufacturing Construction Electricity, Services 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 -5 Year Agriculture Mining Manufacturing Sector Electricity, gas, water Construction Services 1993 Share 2006 Share Average annual real growth (93-06) GDP Growth Source: World Bank (2009, forthcoming) Source: World Bank (2009, forthcoming) 8 Coughlin (2006). 2 Figure 0.4. Growth Decomposition Figure 0.5. Contributions to GDP Growth Mozambique Growth Decomposition Average Contributions to Growth (1993 - 2006) 14% 12% 10% 8% 32% 6% 4% 45% 2% 0% 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 -2% -4% 23% -6% Capital Contribution Labor Contribution (quality adjusted) TFP GDP Growth Capital Labor (quality-adjusted) TFP Source: World Bank (2009, forthcoming) Source: World Bank (2009, forthcoming) Growth-accounting exercises for Mozambique have shed light on the sources of the overall economic growth the country has experienced. While factor accumulation has been substantial, increases in total factor productivity (TFP) have also significantly contributed to growth.9 This aggregate productivity growth mostly reflects the structural changes in output that the economy has experienced since the early 1990s. A shift from lower-productivity agricultural activities towards more productive activities in the industry and services sector explains much of the increase in overall productivity.10 This shift in sources of income also explains a large proportion of the reduction in poverty levels.11 However, these gains in productivity from structural changes may have a limit. Some estimates point to the fact that differences in labor productivity between the agricultural sector and services, while still large, have been declining.12 The productivity gap in relation to industry, however, has been widening. While not the focus of this report, the agricultural sector still represents an important opportunity for increased productivity (Figure 1.6), as evidenced in the gap between current and potential yields for some of its crops presented in the most recent World Development Report (World Bank 2007b).13 9 World Bank (2009, forthcoming), Chapter 1. 10 World Bank (2009, forthcoming), Chapter 1. 11 Clement and Peiris (2008). 12 World Bank (2009, forthcoming), Chapter 1. 13 World Bank (2007b), p. 67. 3 Figure 0.6. Labor Productivity (Normalized to Agriculture) 30 25 1996 2003 20 Percent 15 10 5 0 Industry Services (Private) Services (Public) Sector Source: World Bank (2009, forthcoming) Therefore, while growth in the past years has been robust, certain features of this growth give concern over the sustainability of Mozambique's recent achievements in terms of poverty reduction. The concentration of growth in the services and industry sectors is problematic for two reasons. First, growth in the services sector has meant the creation of mostly informal jobs, productivity gains have also been limited, and more recently even decreasing. Second, growth in the industrial sector has come mostly, as mentioned previously, from the capital intensive megaprojects. While the increases in output and exports have been impressive, reliance on manufacturing growth in only these types of projects presents very limited opportunities for job creation, for economic linkages with the wider economy, and, until recently, for revenue creation.14 Export Performance Between 1994 and 2006, Mozambique was able to introduce 18 new products to its export basket.15 These discoveries, as Table 1.1 shows, were not concentrated on specific categories, instead they were comprised of both agricultural commodities and manufactures. 14 The government has been making efforts at trying to streamline taxation policies and fiscal incentives in order to raise revenues. This has included new rules for future foreign direct investment. For a more comprehensive review of these issues, see Clement and Peiris (2008). For discussion of the tax regime's current bias in favor of capital rather than labor deployment, see World Bank (2009, forthcoming). 15 We adapted the methodology described in Klinger and Lederman (2004) to identify discoveries in Mozambique. "Discovery" is the authors' definition of episodes when countries begin exporting a new product. Disaggregating the United Nations Commodity Trade Statistics Database (COMTRADE) data on exports at the four-digit level, we computed the number of products whose exports in 1994, 1995, or 1996 had been less than US$10,000 and 10 years later accounted for more than US$1,000,000 (filter 1). In order to avoid "false" discoveries, we used a second filter to identify exports that satisfied filter 1 more than once over the period in consideration. For instance, exports of "grain sorghum" did not show up in 1995 and 1996, but totaled US$1,486,491 and US$1,311,476 in 2005 and 2006, respectively. As such, the product discovery was computed only once. 4 Table 0.1. Discoveries between 1994 and 2006 Code Product description H0-2925 Carboxyimide-function and imine-function compounds H0-8534 Electronic printed circuits H0-8410 Hydraulic turbines, water wheels, regulators thereof H0-4803 Paper, household, sanitary, width > 36 centimeters H0-7301 Sheet piling, welded angles, sections of iron or steel H0-5211 Woven fabric < 85% cotton with manmade fiber > 200g/m2 H0-1502 Bovine, sheep, and goat fats, raw or rendered H0-5810 Embroidery in the piece, in strips or in motifs H0-1007 Grain sorghum H0-5303 Jute, other bast fiber, raw or processed, not spun H0-1703 Molasses from the extraction or refining of sugar H0-8602 Rail locomotives, diesel, steam, locomotive tenders H0-7601 Unwrought aluminum H0-6901 Bricks, blocks, and ceramic goods of siliceous earths H0-4014 Hygienic or pharmaceutical articles of rubber H0-2519 Natural magnesium carbonate, magnesium oxide H0-0802 Nuts except coconut, brazil and cashew, fresh or dried Source: Authors' calculations Despite these new product introductions, Mozambique has experienced an overall concentration of its exports. Growth of exports in the last 10 years has come almost exclusively from three megaprojects: the Mozal aluminum smelter, hydroelectricity from the Cahora Bassa dam, and natural gas exported to South Africa.16 All other exports (prawns, cotton, cashews, forestry, tea, and so forth) have grown only slowly, with the exception of tobacco, sugar, and recently, tourism. Figures 1.7 and 1.8 below help to illustrate the megaprojects' increased importance and preponderance in Mozambique's export output. Another measure of this greater concentration is given by calculations of the Herfindahl-Hirschmann index17 which, as shown in Figure 1.9, also confirms a steady increase in export concentration. 16 The Cahora Bassa dam was built in the early 1970s by the Portuguese, but production of electricity was interrupted during the war. After rehabilitation, exports started again in 1998. The first aluminum exports took place in 2000, and natural gas was first exported in 2004 (World Bank 2009, forthcoming). 17 The Herfindahl-Hirschmann is defined as the sum of the squares of the shares of each individual product in total exports, weighted by export share. The index is normalized here to a 0-1 scale--the closer to 1, the higher the concentration. 5 Figure 0.7. Export Performance, 1994­2008 3,000 Total merchandise 2,500 Fuels and mining products US$ in millions Agricultural products 2,000 1,500 1,000 500 - 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 Year Source: World Trade Organization Figure 0.8. Mozambique Main Exports, 2006 1,600 1,404 1,400 1,200 US$ millions 1,000 800 600 400 178 110 110 200 86 71 46 37 36 5 1 - Frozen shrimp Aluminium Tobacco Cotton Maize Gas Sugar Lobster Wood products Cashew nuts Electricity Exports Source: World Bank (2009, forthcoming) 6 Figure 0.9. Concentration of Exports 0.12 Herfindhal Index (normalized) 0.10 0.08 0.06 0.04 0.02 0.00 2000 2001 2002 2003 2004 2005 2006 Year Source: Authors' calculations Trade Mozambique is today one of the most open economies in sub-Saharan Africa, thanks to the process of trade and gradual capital account liberalization initiated by the country in the 1990s (Clement and Peiris, 2008). Similarly, Mozambique is relatively well positioned in regards to access to markets. The country is a beneficiary of the Generalized System of Preferences (GSP) with Australia, Canada, Japan, New Zealand, and the United States, as well as the European Union's Everything But Arms amendment to its GSP. It is also a signatory of the African, Caribbean, and Pacific States­European Commission Partnership (Cotonou) Agreement. In addition, Mozambique's exports are also eligible for preferential access to U.S. markets under the African Growth and Opportunity Act and to India under the Global System of Trade Preferences. At the regional level, Mozambique is a member of the Southern Africa Development Community (SADC).18 The country has held World Trade Organization membership since 1995. Maximizing the benefits from and increasing its openness to trade will be essential for Mozambique to grow at the rates required to bring down absolute poverty and create the number and types of jobs needed. Given the relative small domestic market (in size and in purchasing power), the country's growth strategy needs to rely to a large extent on increasing exports. This represents Mozambique's best long-term potential to sustain a high rate of economic growth. Increasing exports will require increasing selected imports, especially of intermediate and capital goods. In this context, a decisive step is the current regional trade integration within SADC, which envisions a customs union by 2010 and a common market by 2015. However, as noted in the Country Economic Memorandum (CEM), additional markets outside of SADC are increasingly 18 SADC membership comprises Angola, Botswana, the Democratic Republic of Congo, Lesotho, Madagascar, Malawi, Mauritius, Mozambique, Namibia, South Africa, Swaziland, Tanzania, Zambia, and Zimbabwe. As a first step towards regional trade integration, 85 percent of products traded within this zone have enjoyed a zero import tariff since January 2008. 7 important, particularly when considering export diversification into products other than aluminum, gas and electricity, which account for the bulk of current exports to EU and SADC countries, respectively. Moving Forward The agenda to increase competitiveness--working on the investment climate of the country--is, therefore, essential to reap the benefits of increased integration and maximize Mozambique's export potential. The Country Economic Memorandum (World Bank 2009, forthcoming), several value chain studies, and a previous Enterprise Survey (2003) all have already pointed to key constraints to private sector growth: access to credit, barriers to firm entry, the uncertainty of laws and regulations, and access to and quality of infrastructure. Indeed, Mozambique's business investment conditions remain restrictive, and this is illustrated by its ranking in the Doing Business Indicators (2010), where the country's ranking was 135, out of a total of 183 economies. By seeking to deepen the integration of the Mozambican economy into world markets and by addressing competitiveness bottlenecks through additional structural reforms and targeted policy interventions, it is possible for established firms and potential entrants to compete in larger markets. This competition in larger and more sophisticated markets enables the economy as a whole to tap into technology, incentives, and selection mechanisms that are found to command systematic productivity gains over time. Such a strategy contributes directly to the strengthening of domestic markets to the extent that local factors of production participate in the wealth-generating process. The analysis presented in this report highlights competitiveness promotion and trade integration as vehicles for higher productivity growth and wealth generation. The current crisis makes this focus even more important as reductions in the cost of operating businesses will be essential to stimulate trade and investment again. The government has a crucial role to play in this process, by diligently pursuing reforms and policies today. In the next chapters we present our assessment of the investment climate in Mozambique based on the results of the survey conducted in 2007­08. 8 Box 0.0. Government Initiatives to Improve the Business Environment The government of Mozambique's second Poverty Reduction Support Strategy (PARPA II), covering 2006­09, identified the private sector as the main engine for investment, growth, and employment. Realizing the vision set out in the PARPA II requires improving the business environment, stimulating exports, promoting growth in sectors where Mozambique has a comparative advantage, and supporting SMEs, which generate the greatest number of jobs. According to the Ministry of Planning and Development, in the process of evaluating the PARPA II, a study will be undertaken to analyze the development and growth of the private sector in terms of employment creation. In response to stronger advocacy by the private sector and international partners' advice, the government has embarked on a five-year strategy to improve the business environment. The Business Environment Strategy contains the following four pillars: (i) Legal Reform, which aims to improve the business environment through reforms covering business registration, licensing, inspection, and closure; labor law; competition law; import-export facilitation; property registration; and simplification of tax administration; (ii) Fiscal and Financial Sector Reform, which aims to improve the fiscal environment and promote access to finance for Small and Medium Enterprises (SME); (iii) Infrastructure, which aims to reduce costs and increase access for private sector to basic infrastructure services, such as electricity, and telecom; and (iv) Governance and Implementation Mechanisms, which covers the capacity building needed to implement the reforms and establishes a mechanism for overseeing and monitoring the strategy implementation To further complement the strategy, the Council of Ministers in June 2008 adopted an action plan that provides a road map of timelines, reform actions, indicators of achievement, and the responsible agencies under each of the main pillars of the Business Environment Strategy. The Ministry of Industry and Commerce has also recently developed a strategy for SME development that was adopted by the Council of Ministers in 2007. The SME strategy is supposed to complement the overall Business Environment Strategy and focuses on specific areas of support needed for SMEs, through promoting access to finance, improving SME management and technical capacity, and ensuring that reforms are supportive of SME development. The government has also been working on sector-specific strategies in agriculture and tourism, as well as an overall approach to dealing with trade integration. Source: World Bank (2009) 9 Business Climate Introduction This chapter puts Mozambique's business climate into perspective by summarizing the micro data and benchmarking Mozambique's results with respect to other countries. The disaggregated nature of the Enterprise Survey's data allows for the possibility of exploring differences within countries, across firm sizes, across regions, and among industries, among other categories. In addition, the standard design of the questionnaire that is applied for these surveys allows for comparisons of results across countries.19 While much of the data presented in this chapter is qualitative, representing manager's perceptions of various constraints related to the business climate in which they operate, we also make use of quantitative results (hard data) to augment the findings. Firms' perceptions provide an initial weighing of the importance of different obstacles but need to be interpreted with caution. Perceptions may be biased by recent events, displaying, for example, temporary optimism or pessimism associated with recent economic performance, and they may also reflect inherent cultural traits and specific socioeconomic backgrounds. Whenever possible (and especially for international comparisons) we use the ranking of obstacles to draw conclusions from the data, instead of relying solely on percentage of respondents. Moreover, the Enterprise Survey data is complemented with data from a variety of other sources, including Doing Business indicators, World Economic Forum rankings, Transparency International, and the World Bank's World Governance Indicators. These additional sources of information can provide a richer and more nuanced view of the investment climate in Mozambique. 19 We use as comparator countries in this study Angola, Malawi, South Africa, Zambia, Indonesia, Thailand, and Vietnam. 10 Box 0.0. The Enterprise Survey in Mozambique The Mozambique Enterprise Survey sample included 599 firms from a variety of industries, regions, and sizes. The objective of the survey was to collect both subjective indicators of firm perceptions about key obstacles related to the business environment and objective data for econometric analyses of firm productivity determinants. The survey for Mozambique was fielded in the first half of 2008. The basic breakdown of the sample is shown below. The analysis and description of the data is based on weighted averages, but we avoid drawing conclusions from very small samples. Detailed information about the sample and sampling methodology employed for this survey is found in Annex I. Detailed information about World Bank's Enterprise Surveys in general can be found at http://www.enterprisesurveys.org. Nampula Regional distribution Medium F irm s ize 10% (20-99) (by employees ) 24% B eira 9% Large Matola 00+ (1 ) 10% Small (5- 5% 19) 51 % Micro (5 Maputo 71% or les s ) 20% D is tribut ion Manufa- by s ect or cturing 60% Cons tru- ction & T rans po- S ervices rt 39% 1% Overall, rankings and indicators from different sources consistently show that Mozambique has a weak investment climate. For example, according to the Global Competitiveness Index (GCI) (2009­10),20 Mozambique ranks 129 out of 133 countries; this is the lowest ranking among its comparator countries. The Ease of Doing Business 20 The Global Competitiveness Index (GCI) summarizes 12 pillars of competitiveness--Institutions, Infrastructure, Macroeconomic Stability, Health and Primary Education, Higher Education and Training, Goods Market Efficiency, Labor Market Efficiency, Financial Market Sophistication, Technological Readiness, Market Size, Business Sophistication, and Innovation. The GCI rankings are calculated from both publicly available data and the Executive Opinion Survey, an annual poll the World Economic Forum conducts in partnership with leading research institutes and business organizations in countries the report covers. 11 rankings21 put forth by the World Bank places Mozambique in the position 135 out of 183 economies. Despite gaining five positions in the overall ranking, within SADC countries Mozambique ranks 14th out of 15 countries (see Table 2.1). Among comparator countries, only Angola was ranked lower than Mozambique according to the Doing Business Report in 2010. Performance for Mozambique is particularly poor on the following indicators: Dealing with Construction Permits (159), Employing Workers (156), Registering Property (151),22 Starting a Business (144), Trading across Borders (136), Closing a Business (136), Enforcing Contracts (129), and Getting Credit (127). On the other hand, it ranks relatively well in Protecting Investors (41). Table 0.1. Mozambique Doing Business Indicators, 2009 and 2010 Best Best Mozambique Performance Country in Rank Indicator 2010 2009 in SADC SADC in SADC Ease of doing business (overall rank) 135 140 17 Mauritius 14 Time to start a business (days) 26 26 6 Mauritius 5 Time to deal with construction permits (days) 381 381 93 Swaziland 12 Rigidity of employment index (0-100) 40 49 10 Swaziland 10 Firing costs (weeks of salary) 134 134 4 Mauritius 12 Time to register property (days) 42 42 16 Botswana 8 Tax payments per year (number) 37 37 7 Mauritius 6 Time to pay taxes (hours per year) 230 230 76 Seychelles 10 Documents to import (number) 10 10 5 Seychelles 14 Documents to export (number) 7 8 4 Madagascar 8 Time to enforce contract (days) 730 730 270 Namibia 9 Time to close a business-bankruptcy (years) 5 5 1.5 Namibia 11 Getting credit ­ strength of credit information index (0-10) 4 4 6 South Africa 5 Protecting investors- strength of investor protection index (0-10) 6 6 8 South Africa 4 Source: Doing Business, 2010 and 2009 reports 21 The Doing Business rankings summarize a wide range of characteristics, such as starting a business, dealing with licenses, and closing a business. As with the GCI, economies are ranked from best to worst. A low number on the ease of doing business index connotes a regulatory environment conducive to operating a business. This index averages the country's percentile rankings on 10 topics made up of a variety of indicators, giving equal weight to each topic. 22 Registration of property is of particular importance for industries with potential for growth in Mozambique. The most recent CEM for Mozambique highlights that the process of land acquisition is both risky and lengthy, and that this is particularly relevant for new investments in the agribusiness and tourism sectors. As emphasized by the report, the laws themselves are generally adequate, but implementation remains a challenge. 12 In the remainder of this chapter, we report and discuss the main results of the survey in terms of firm's perceptions of main obstacles to their operations. Some of these main constraints are elaborated further in this chapter, while issues of informality (specifically relating to microenterprises), access to finance, infrastructure, and human capital are discussed in subsequent individual chapters. The Main Obstacles according to Manager's Perceptions Surveyed firms were asked to report how serious of an obstacle different components of the investment climate are to their operation or development. They reported their perceptions using a five-point scale ranging from "no obstacle" to "very severe obstacle." Table 2.2 presents the proportion of firms that rated the different areas of the business environment in consideration as a major or very severe obstacle to their business operation. The two constraints cited most often in Mozambique are practices of informal competitors and access to finance. Crime, theft and disorder, tax rates, corruption, electricity, and transport represent the other most commonly cited constraints. When firms were asked to prioritize among the obstacles presented, the results were very similar. Figure 2.1 shows that informal competition and access to finance remain the main constraints, with only a slight rearrangement in the order of the next five obstacles. In other words, tax rates, crime, electricity, and transportation still represent the main concerns according to firms, after informality, and finance. Corruption is the only issue that seems to lose some urgency as a binding constraint, especially in relation to infrastructure concerns. Table 0.2. Firms Reporting Major or Very Severe Business Constraints (percentage) Constraint Total Manufacturing Services Other Micro Small Medium Large Practices of informal 51.09 51.12 51.28 42.86 46.67 55.70 46.48 43.33 competition Access to finance 45.74 50.56 38.46 42.86 25.83 53.42 46.48 43.33 Crime 32.22 32.96 30.34 57.14 21.67 34.53 35.21 36.67 Tax rates 31.05 27.65 35.47 57.14 19.17 35.18 33.10 26.67 Corruption 26.38 23.74 30.77 14.29 20.00 28.99 27.46 20.00 Electricity 21.87 24.58 17.09 42.86 17.50 22.48 22.54 30.00 Transport 21.04 22.63 17.95 42.86 16.67 24.43 12.68 43.33 Tax administration 15.53 10.06 23.08 42.86 7.50 16.61 16.90 30.00 Workforce 15.19 15.92 13.25 42.86 4.17 15.64 19.72 33.33 education Licensing & permits 14.69 12.85 17.95 0 10.83 18.57 10.56 10.00 Customs & trade 14.52 8.66 23.50 14.29 15.00 13.36 12.68 33.33 regulations Access to land 13.02 12.29 13.68 28.57 16.67 12.7 8.45 23.33 Telecommunications 6.34 6.15 5.98 28.57 4.17 7.17 3.52 20.00 Political instability 6.18 6.42 5.98 0 5.83 9.45 0.70 0 Labor regulation 5.51 5.31 5.98 0 4.17 3.91 8.45 13.33 Courts 4.67 3.07 6.84 14.29 2.50 4.23 6.34 10.00 13 Table 2.2 continued Outside Constraint Domestic Foreign Nonexporter Exporter Maputo Maputo Practices of informal 51.52 49.04 51.83 34.62 54.82 41.95 competition Access to finance 45.25 48.08 46.25 34.62 47.06 42.53 Crime 31.31 36.54 31.59 46.15 34.82 25.86 Tax rates 30.1 35.58 31.76 15.38 34.59 22.41 Corruption 25.66 29.81 26.00 34.62 30.35 16.67 Electricity 21.01 25.96 21.99 19.23 20.94 24.14 Transport 19.39 28.85 20.94 23.08 20.47 22.41 Tax administration 14.75 19.23 15.18 23.08 17.41 10.92 Workforce 13.54 23.08 14.66 26.92 18.35 7.47 education Licensing & permits 13.33 21.15 14.14 26.92 17.88 6.90 Customs & trade 11.52 28.85 13.61 34.62 17.88 6.32 regulations Access to land 12.32 16.35 12.91 15.38 12.00 15.52 Telecommunications 5.05 12.50 6.28 7.69 5.88 7.47 Political instability 6.67 3.85 6.46 0 7.53 2.87 Labor regulation 4.44 10.58 5.06 15.38 6.12 4.02 Courts 3.23 11.54 4.01 19.23 4.47 5.17 Source: Enterprise Survey 2008 Note: Response rates for these perception questions are close to 100 percent. Figure 0.1. Main Obstacles P ractices of Informal Competitors 49% Acces s to F inance 42% T ax Rates 36% Crime, T heft, andDis order 31% T rans portation 27% Electricity 25% Corruption 15% Inadequately EducatedWorkforce 14% Acces s to Land 13% Cus toms andT rade Regulations 12% Licens ing andP ermits 12% T ax Adminis tration 7% Labor Regulations 4% P olitical Ins tability 4% Courts 2% % of firms report obstacle among top three constraints Source: Enterprise Survey 2008 A look across the spectrum of firms gives more insights about how firms perceive the business environment in Mozambique. Constraints ranked as major or very severe are very similar for most subcategories of firms. Exceptions worth noting are exporters 14 and firms in the industry categorized as "others," which includes firms in the construction and transportation sector. Among exporters, more firms perceive crime (46 percent) as a major or very severe obstacle for business than any other obstacle; practices of informal competition and access to finance follow, and are perceived as equally problematic as corruption and, not surprisingly for exporters, as customs and trade regulations (35 percent of firms rate them as major or very severe obstacles for business). Among firms in the construction and transportation industries, crime and tax rates draw the highest response (57 percent) among firms as major or very severe business constraints, while informality and access to finance share the same percentage of respondents as infrastructure issues (electricity and transport), tax administration, and workforce education (43 percent). Differences within subcategories are taken up in the following sections of this chapter. A comparison with 2003 shows that while concerns with corruption and especially with energy have declined among firms, issues like informal competition, crime, and tax rates are perceived as more important constraints today. It is useful to compare the perceptions presented in Table 2.2 with the figures from a previous survey, carried out in 2003. Table 2.3 shows that there were important changes in the ranking of obstacles.23 Perhaps reflecting some of the advances that have been made in the provision of electricity across the country, concerns with energy dropped from 2nd to the 6th most often cited constraint. Concerns with labor regulation also showed a decline, going from 11th to 15th in the rankings. On the other hand, crime, tax rates, and transport moved up in the rankings, being perceived as more important constraints in this last survey. Access to finance is a top constraint in most African countries, according to business perceptions. Another interesting exercise is the comparison of the main constraints identified by firms in Mozambique with those in other countries. Among African countries, the constraints most often cited by firms as major or severe obstacles to business are fairly similar (Table 2.4), but some differences are worth noting. Informal competition, while the most often cited concern in Mozambique and Vietnam and also a major concern for Zambian firms, ranks lower in the list of concerns for all other comparator countries (Table 2.5).24 Workforce education, and labor regulation issues are the most prominent issues raised by South African firms, while firms in most other countries, including Mozambique, do not perceive these two issues as severe constraints to business.25 23 Care must be exercised when comparing results from 2003 and 2008, since some of the issues included in the survey in 2003 were not included in the newest survey. For example, concerns with macroeconomic stability ranked as the fourth most often cited major or severe constraint in 2003, while not being part of the survey in 2008. Moreover, the two surveys followed different processes for sample selection. The 2003 survey included only manufacturing firms, while the 2008 survey included firms in both the manufacturing and services sectors. 24 Concerns with practices from informal competition have shown up more often as a significant obstacle to business in the most recent round of Enterprise Surveys. This is not confined to African countries, as it has also been noted in the most recent surveys in Latin America. 25 Firms in Malawi also perceive workforce education as a significant issue, ranking it the fourth most often cited concern. 15 Table 0.3. Most Often Cited Obstacles, 2003­08 Major or Severe Obstacles to Ranking Ranking Business in 2008 in 2003 Practices of informal competition 1 5 Access to finance 2 1 Crime 3 8 Tax rates 4 7 Corruption 5 3 Electricity 6 2 Transport 7 14 Tax administration 8 10 Workforce education 9 12 Licensing & permits 10 13 Customs & trade regulations 11 9 Access to land 12 15 Telecommunications 13 16 Political instability 14 Labor regulation 15 11 Courts 16 Source: Enterprise Surveys 2003 and 2008 Table 0.4. Most Often Cited Constraints for Business across African Comparators Mozambique (2008) Angola (2006) Malawi (2005) Informal competition (51%) Access to finance (53%) Access to finance (76%) Access to finance (46%) Electricity (46%) Electricity (60%) Crime (32%) Corruption (34%) Tax rates (56%) Tax rates (31%) Crime (33%) Workforce education (50%) Corruption (26%) Licensing & permits (31%) Crime (47%) Electricity (22%) Transport (29%) Corruption (47%) Transport (21%) Informal competition (24%) Tax administration (43%) South Africa (2003) Zambia (2007) Workforce education (35%) Access to finance (28%) Labor regulation (33%) Informal competition (26%) Crime (29%) Tax rates (25%) Access to finance (23%) Access to land (16%) Tax rates (19%) Corruption (15%) Customs & trade regulations (17%) Electricity (14%) Corruption (16%) Crime (13%) Source: Enterprise Surveys Note: Number in parenthesis next to each constraint represents percentage of firms reporting the issue as a major or very severe business constraint. 16 Table 0.5. Most Often Cited Constraints for Business across Asian Comparators Indonesia (2003) Thailand (2004) Vietnam (2005) Corruption (41%) Workforce education (30%) Informal competition (20%) Access to finance (31%) Electricity (26%) Access to finance (16%) Tax rates (29%) Tax rates (24%) Corruption (15%) Labor regulation (26%) Customs & trade regulations (24%) Transport (12%) Tax administration (23%) Access to finance (23%) Electricity (9%) Electricity (22%) Tax administration (22%) Crime (6%) Crime (22%) Informal competition (20%) Telecommunications (6%) Source: Enterprise Surveys Note: Number in parenthesis next to each constraint represents percentage of firms reporting the issue as a major or very severe business constraint. Indirect Costs Quantitative data collected by the survey reinforce some of the qualitative data represented in firms' perceptions. Namely, the relative lower burden of electricity- related issues on firms' operating costs mirrors the drop in the level of concern with energy issues detected between the 2003 and 2008 surveys. On the other hand, crime is imposing relatively larger cost burdens on Mozambican firms, and this has also been reflected with an increased number of firms that perceive it as a serious obstacle to business. Firm's perceptions about the business environment can be contrasted with some of the real costs reported by companies. Table 2.6 reports the estimated impact of some of these constraints on the indirect costs of firms in Mozambique. Mozambican companies report indirect costs of 7 percent of their annual sales. Indirect costs include costs imposed by crime-related issues (criminal acts and security services) and losses associated with infrastructure (transit and power outages). From an international point of view, Table 2.7 reveals that Mozambique compares well when contrasted with levels of indirect cost reported in Angola and Malawi, and is on par with results for Zambia. South African and Vietnamese firms, on the other hand, report the lowest levels of costs or losses due to crime and poor infrastructure. Table 0.6. Indirect Costs as a Percentage of Annual Sales (weighted averages) Crime & Security Infrastructure % Sales % Sales % Sales % Losses-- % Losses-- % Sales Lost Lost Due to Spent in Lost while Theft in Breakage/Spoilage Due to Crime Power Security in Transit Transit in Transit Outages 1.49 2.02 1.50 1.98 0.80 1.18 Source: Enterprise Survey 2008 17 Table 0.7 Indirect Costs as Percentage of Annual Sales (cross country comparisons) Indirect Costs South as % of Angola Malawi Vietnam Zambia Mozambique Africa Annual Sales % sales spent 1.69 2.10 0.89 0.62 1.45 1.49 on security % sales lost 0.59 2.20 0.61 0.04 1.16 2.02 due to crime % sales lost due to power 3.94 22.24 0.92 1.49 2.68 1.50 outages % sales lost 1.92 3.34 0.80 n.a. 1.65 1.98 while in transit Total 8.14 29.88 3.22 2.15 6.94 6.99 Source: Enterprise Survey 2008 n.a. Not available. Figure does not include percentage of sales lost while in transit. Crime Crime was the third most often cited constraint to doing business in Mozambique, a significant increase when compared to the 2003 survey, when it was ranked 8th. One firm out of three rated it as a severe or major constraint for growth. Crime is also often cited as a constraint to business in other African comparator countries, ranking 5th in Malawi, 3rd in South Africa, and 4th in Angola. On the other hand firms in Zambia and the other Asian comparator countries (Indonesia, Thailand, and Vietnam) do not cite crime as an obstacle to business as often as the former countries. Within Mozambique, concerns about crime vary across firms. Maputo-based firms voiced more concerns about security than firms located outside of Maputo. More exporting firms seem to view crime as a serious obstacle to their operations than nonexporters. Manufacturing and retail firms share a similar level of concern about criminality (33 and 30.3 percent respectively), whereas a significantly larger proportion of firms in the construction and transport sectors view crime as a constraint (57 percent). Incidence of crime adds significantly to the costs of doing business in Mozambique, both directly through the loss of plant equipment and product to vandalism or theft, and indirectly through the cost of employing security agencies to protect firms' property. Objective data included in the Enterprise Survey can complement perception- based data. Indeed, 36 percent of firms in Mozambique experienced losses as a result of theft, vandalism, or arson, while in Angola, a similar concern about crime reflects a much lower actual incidence than in Mozambique: only 18 percent of firms reported losses stemming from criminal acts. South Africa, on the other hand, has a higher incidence of crime: 52 percent of manufacturing firms experienced losses due to it. Mozambican firms reported having suffered losses amounting to 2 percent of annual sales due to arson, theft, or vandalism (Table 2.7). Such costs are considerably higher than costs experienced in most comparator countries, including South Africa (0.61 percent). As a consequence of the incidence and concerns over crime, up to 34 percent of firms in Mozambique report the use of security services (equipment, personnel or professional security services). This 18 additional cost, on average, amounts to 1.5 percent of sales each year. This cost is similar to figures in Angola and Zambia (1.69 and 1.45 percent, respectively), lower than in Malawi (2.1 percent), but much higher than in South Africa or Vietnam (0.89 and 0.62 percent respectively). Tax Rates Across firms, the financial burden of taxation seems to constrain more small and medium enterprises (SME) and firms based in Maputo. Businesses complain worldwide about tax rates. According to the World Development Report 2005, managers ranked tax rates among the top five obstacles in all upper middle-income countries. The same rank was obtained in over four out of five countries in Sub-Saharan Africa where World Bank Enterprise Surveys had been completed at the time (World Bank, 2004). In Mozambique, tax rates ranked as the fourth most often cited obstacle, with 30 percent of firms perceiving it as a major or very severe constraint to their business activity. Enterprise Survey results from other comparator countries also placed tax rates among the top five obstacles. Exceptions to these perceptions where results in Angola and Vietnam, where firms were much less likely to see tax rates as a significant constraint. Nonexporting firms were more likely to complain about tax levels than exporting ones. Indicators of overall fiscal pressure in Mozambique presented in the Doing Business report of 2009­10 suggest that the country compares favorably to some comparator countries, notably Angola, Vietnam, Thailand, and Indonesia, while South Africa, Malawi, and especially Zambia, have lower overall tax burdens on businesses. Figure 2.2 presents the full breakdown of taxes that an average business pays. It shows that, although the burden of total tax in Mozambique is similar to South Africa, the compositions are different. Mozambique extracts higher taxes on profits and for labor contributions while other taxes account for a larger percentage of total tax rates in South Africa. Even though tax rates in Mozambique, when compared to other countries in the region and outside, are not particularly high, the system of tax and tariffs could be made significantly more efficient. Analysis presented in the most recent CEM for Mozambique indicates that the tax and tariff system is "characterized by many different exemptions whose cumulative effects are unclear. And small entrepreneurs normally do not benefit from such exemptions since there is a fixed cost involved in applying for them, which they cannot afford" (World Bank, 2009, forthcoming). The report suggests that tariff rates could be reduced while maintaining revenue levels unchanged if tariff exemptions were eliminated. Thus, not only could this reform make the country more attractive for investors, but it would likely make the tax treatment of different firm sizes more equitable. 19 Figure 0.2. Breakdown of Taxes, International Comparison O L P W Source: Doing Business indicators 2009­10 One expected consequence of heavy profit tax regimes is tax evasion. As noted in Chapter 5 of this report (Table 5.2), Enterprise Survey results for Mozambique show a strong and significant correlation between firms that see tax rates as a constraint on business and their perceptions of tax evasion in their own business sectors. Figure 2.3 provides more disaggregated information on firms' perceptions of tax compliance in their own sectors for Mozambique. Tax compliance is seen as more prevalent among firms in the construction and transport sector, when compared to either manufacturing or services firms. Differences in perceptions are not statistically significant for firms in Maputo when compared to firms outside Maputo, or between exporters versus nonexporters. On the other hand, the differences in perceptions between foreign-owned and locally-owned firms are statistically significant. Differences in perceptions among firms of different size are only statistically significant between large and small firms and between small and medium firms. 20 Figure 0.3. Perceptions on Tax Compliance What percentage of total annual sales would you estimate a typical establishment in your sector of activity reports for tax purposes? 78.5 70.0 55.5 56.7 52.8 46.8 47.0 44.1 42.1 41.6 43.7 39.5 37.1 Services Manufacturing Transport Exporter Non-exporter Maputo Outside Foreign Micro Large Small Medium Constr. & Domestic Maputo Sector Location Nationality Export Size Source: Enterprise Survey 2008 Corruption Corruption is still perceived to be a serious constraint to business operation in Mozambique. Table 2.2 indicates that corruption ranked as the fifth most often cited obstacle, a slight improvement over the 2003 survey, when it was ranked the third most severe obstacle. Across different categories of firms, complaints about corruption arise more often among SMEs, firms located in Maputo, exporting firms, and among firms operating in the services sector. Across countries, corruption consistently ranked among the top five most often cited constraints, with the exception of South Africa, where it ranked 7th, and Thailand, where it ranked 8th. Other cross-country rankings provide further evidence of the importance of the issue. The Corruption Perception Index (CPI) compiled yearly by Transparency International supports the view that corruption is widely perceived as a serious constraint in Mozambique.26 Figure 2.4 reveals that perceptions about corruption in Mozambique are similar to those in most comparator countries, with the exception of better perceptions in South Africa and Thailand, and more negative perceptions in Angola. Kaufman, Kraay and Mastruzzi's (2008) control of corruption index confirms Transparency International's overall ratings, showing again significantly better performance for South Africa and to a lesser extent Thailand, while 26 The Transparency International CPI measures the perceived levels of public sector corruption in a given country and is a composite index, drawing on different expert and business surveys. The 2008 CPI scores 180 countries (the same number as the 2007 CPI) on a scale from zero (highly corrupt) to ten (highly clean). Mozambique's 2008 ranking actually decreased from 2007, when it ranked 111th (2.8 index). 21 Mozambique ranks at a similar level with Zambia and Angola presents the lowest score (Figure 2.5). Figure 0.4. Corruption Perceptions Index 6 4.9 5 4 3.5 Index 2.8 2.8 2.7 2.6 2.6 3 1.9 2 1 0 e a a nd m i la a aw qu bi si na ric go la ne m bi al ai Af et Za An am do M Th Vi h In ut oz So M Country Source: Transparency International, 2008 Figure 0.5. Control of Corruption Index 0.4 0.2 0 Mozambique Thailand Zambia Indonesia Vietnam Malawi South Angola Africa -0.2 Index -0.4 -0.6 -0.8 -1 -1.2 Country Source: Kaufman, Kraay and Mastruzzi (2008) Note: All scores lie between -2.5 and 2.5; higher scores correspond to better outcomes. International comparisons put Mozambique, while not among the very bottom, certainly in a position where much improvement is needed. Before confronting these findings with more objective indicators, it is important to note that corruption can take various forms. It can consist of irregular payments to get things done with respect to customs, taxes, licenses, regulations, and services, usually taking place at the implementation end of politics and policies, where "public officials meet the public."27 This petty corruption, while usually involving small amounts of money per transaction, if 27 "Corruption: Definitions and Forms," Anti-Corruption Resource Centre (http://www.u4.no). 22 pervasive can add up to significant costs to society. Grand corruption, on the other hand, usually involves more senior officials and larger sums of money. It is characterized by, among other things, attempts to secure public contracts, ensure favoritism of government officials, or influence government policy making. Both forms of corruption are explored separately next. Petty Corruption Figure 2.6 shows that corruption to obtain infrastructure services varies given the type of service being requested. Informal payments to secure electrical or water connections are reported by 15 percent of firms surveyed, almost twice as many as firms that reported having to make informal payments to secure a telephone connection (8.3 percent). Construction permit issuance had the lowest reported incidence of bribes: only 4.7 percent of firms reported requests for such informal payments. On the other hand, 13.3 percent and 9.5 percent of businesses surveyed that needed import or operating licenses, respectively, reported being asked for a bribe when applying for the license. Figure 0.6. Percentage of Firms Requested to Pay Bribes when Applying for Public Licenses/Services % of firms that responded "Yes" 15.3% 15.5% 13.3% 9.5% 8.3% 4.7% Telephone Electrical Water Construction Import license Operating connection connection connection related permit license Source: Enterprise Survey 2008 Interactions with tax inspectors present another opportunity for informal payments to take place. About 11 percent of firms reported informal payments to tax inspectors in Mozambique. This is similar to the level of incidence reported in Indonesia, lower than levels in Angola (16.6 percent) and Malawi (15.3 percent), but higher than levels in Zambia (5.4 percent) and especially South Africa (0.6 percent) (Figure 2.7). 23 Figure 0.7. Corruption in Tax Inspections--Cross-Country Comparison In any visits, inspections or meetings w ith tax officials, w as a gift or informal payment expected/requested? 48.4% % of firms that responded "Yes" 16.6% 15.3% 11.2% 10.7% 5.4% 0.6% Vietnam Angola Malaw i Indonesia Mozambique Zambia South Africa Source: Enterprise Surveys 2003-08 Grand Corruption Figure 2.8 shows businesses perceptions of favoritism when government officials decide upon policies and contracts. The data collected by the World Economic Forum illustrates that perceptions of African countries regarding favoritism are negative overall. Among comparators countries, Mozambique shows a comparable level of favoritism. It fares slightly better than Angola and Zambia. Nonetheless, perceptions seem clear when having to decide on policies or contracts, businesses believe that government officials favor well-connected firms and individuals. Figure 0.8. Favoritism in Decisions by Government Officials When deciding upon policies and contracts, government officials (1 = usually favor w ell-connected firms and individuals, 7 = are neutral) 3.1 2.9 2.7 2.4 2.2 South Africa Malawi Mozambique Angola Zambia Source: World Economic Forum (2007) 24 Firms were asked if and how much they would typically have to pay to secure a government contract. The answer provided by Mozambican firms is high.28 Among comparator countries only Angolan firms reported higher levels of informal payments to secure public contracts. On average firms in Mozambique are required to pay approximately 3 percent of the value of a government contract as informal payments (see Figure 2.9). This amount, while similar to that reported by Zambian firms, is much higher than values reported in Malawi and South Africa. Figure 0.9. Informal Payments in Public Procurement--Cross-Country Comparison When establishments like this one do business w ith the government, w hat % of contract value w ould typically be paid in informal payments/gifts to secure the contract? 4.7 3.0 2.9 1.6 1.0 0.8 0.1 Angola Mozambique Zambia Malaw i Vietnam Indonesia South Africa Source: Enterprise Survey 2003-08 Conclusions While Mozambique's performance in many of the governance indicators reviewed in this chapter does not compare poorly with many of its regional comparators, certainly the general assessment is that much can be done to better utilize current resources while also improving the incentives for investment in more productive activities. Weak governance can affect the competitiveness and efficiency with which a country uses its resources. More directly, it increases firm's costs by creating additional "taxes" on businesses, such as informal payments needed to "get things done," costs related to regulatory and legal uncertainty, and additional security measures needed to prevent theft and to repair damages. More indirectly, it provides incentives for rent- seeking and nonproductive activities, diverting efforts and talent from more productive 28 Results for Mozambique need to be taken with caution since the response rate for this question was particularly low. 25 activities. These costs have a direct impact on reducing domestic investment as well as in dissuading foreign investment. More evidence on these impacts is presented in the next chapter. 26 Investment Climate, Productivity, and Exports This chapter takes advantage of the additional information gathered by the survey to present a detailed assessment of firms' performance in Mozambique. This information, largely quantitative, allows for a more rigorous analysis of how different dimensions of the investment climate in Mozambique are correlated with firm performance. Drawing on the data for enterprises in Mozambique as well as for comparator countries, we first examine correlations between investment climate variables, firm-specific characteristics, and firm productivity. We then explore export performance and firm characteristics. The chapter concludes with a look at labor productivity differences across countries, followed by a decomposition of the recent changes in labor productivity observed in Mozambique, using the methodology from Olley and Pakes (1996). Impact of Investment Climate Variables on Firm Productivity In this section we examine the correlations between Investment Climate Assessment (ICA) variables, firm-specific characteristics, and firm performance (measured as sales and value added per worker, TFP, and wages paid to workers), using data for enterprises in Mozambique and comparator countries. This pooling of data allows for more efficient estimates of the correlations between investment climate variables and firm performance, as well as providing information for comparisons of associations across countries. However, since our data are cross-sectional, inferences about causality are not possible. Nonetheless, significance in the correlations allow us to determine which indicators have a significant association with enterprise performance, and identify those whose association is minimal. Mozambique Only Regressions The indices for infrastructure, finance and technology, and human capital are highly correlated to improved firm performance, and their coefficients are also statistically significant. Table 9.9 in Annex IV presents the full results for the regressions between the investment climate indices and firm productivity for the sample of Mozambican enterprises only. Firms that have better access to finance and that show greater efforts at technology upgrading and worker training also show significantly higher levels of productivity as measured by both sales and value added per worker indicators. Poor infrastructure, on the other hand, is correlated with lower firm productivity as measured by both value added per worker and TFP. The variables for governance, corruption, and informality29 do not show a significant correlation to firm performance, with the exception of informality and value added per worker. Finally, regional disparities are significant: firms in Maputo perform better, particularly with regards to value added per worker, TFP, and wages paid to its workers. 29 The governance index tries to capture crime and security issues related to firm performance. A separate binary variable that captures the extent to which firms pay bribes to access government services and/or public contracts tries to measure the degree of corruption. The informality index summarizes firm's perceptions about the extent of informality in their sectors. For a more detailed explanation of the investment climate indices used in the analysis, see Annex IV. 27 Pooled Sample (Mozambique and Comparator Countries) Across the pool of countries, governance issues matter for firm productivity. Table 9.10 presents results for the relationship between the investment climate indices and firm performance for the pooled sample. The first column shows that expenses associated with security and crime--our Governance index--do take a toll on firms' productivity, measured as sales per worker, across the sample of countries pooled. Perception of poorer firm compliance in reporting total sales for tax purposes--the measure of informality used in this analysis--is more prevalent among more productive firms. As in the regression results for Mozambique only, infrastructure and finance have similar correlations to firm productivity for this larger sample, although the magnitudes of the correlations are lower when other countries are taken into account. When it comes to value added per worker and TFP, technology appears as a significant investment climate variable associated with enhanced firm performance, while Access to Finance and Poor Infrastructure lose statistical significance. Informality and the Governance indices are still significant and their correlations to TFP are larger than to productivity as measured by sales per worker. Using labor productivity for the restricted TFP sample, we observe that only the informality index is correlated with higher firm productivity. The last column shows the relationship between wages and the investment climate indices. None of the investment climate variables seem to be significantly correlated to wages paid by firms in the pooled sample. It is worth noting that foreign ownership is highly correlated to improved firm performance under all five specifications (sales and value added per worker, TFP, sales per worker for TFP sample, and wages). Finally, Mozambique's performance gap in relation to the other countries compared in this analysis is evident from the country dummy coefficients: significant differences in performance are captured in relation to most countries and measures of productivity, as well as wages paid to workers. Summary of Results and Comparison of Mozambique Versus Pooled Sample Similar results, larger correlations. Comparing the results from Tables 9.9 and 9.10, we observe that while poor infrastructure and access to finance are all correlated to firm performance in Mozambique and in the comparator countries, the extent of this association seems to be greater for Mozambican firms. Therefore, for example, the ability to access finance separates to a larger extent more productive firms from less productive ones in Mozambique than in the other countries included in the sample. The technology index also shows a greater divide between more productive firms and less productive ones in Mozambique than in the other countries, where the index also loses statistical significance when productivity is measured as sales per worker. On the other hand, issues of crime and security captured by the Governance index, show a greater correlation with firm performance for the pooled sample of countries than in Mozambique. A final exercise was conducted to assess potential impact of improvements in investment climate variables. Figure 3.1 below illustrates the relative magnitude of 28 potential productivity gains that can be achieved if improvements in the main investment climate indices addressed in this report were to be made.30 Figure 0.1. Potential Gains from Improvements in IC Indices 120 Labor productivity gain (%) 100 80 60 40 20 0 Governance Technology Infrastructure Bribes Access to finance Indices Source: Author's calculations based on Enterprise Survey 2008 results Note: Blue bars are statistically significant results (at the 10 percent level or better), while bars in gray represent results that are not statistically significant. Exports and Exporter Premium As exports become more important for Mozambique, it is worthwhile estimating if firms that export perform better than firms that do not. The literature on this relationship between exports and firm growth has firmly established that the two are highly correlated, while the direction of causality between them remains less certain.31 Some studies point to evidence that exporting exposes firms to knowledge and competition that pressures them to become more productive. Larger foreign demand for products, particularly important for smaller economies, may also allow firms to exploit economies of scale, increasing their overall productivity. Other studies, however, point to evidence that in fact firms self-select into exporting. In other words, more productive firms are better able to meet the demands and costs of more sophisticated, foreign markets.32 Due to lack of adequate data, the exercise that is attempted here does not deal 30 "Improvements" are the impact on valued added per worker of an increase (or decrease, depending on the sign of the regression coefficient) of one standard deviation in the investment climate indices for the Mozambique only regression. Large gains are reported, especially for infrastructure and technology, due to large coefficients and/or large standard deviations. Results therefore should be interpreted more on a relative basis. 31 For a brief overview of the literature on this topic, see Carlos Casacuberta et al. in Fajnzylber et al. (2009), Chapter 7. 32 According to Casacuberta et al. (in Fajnzylber et al., 2009), the evidence for the self-selection hypotheses seems more robust. Costs in this case are related to costs incurred to reach foreign markets, such as transportation, distribution, marketing, and even production costs associated with changes needed to meet higher foreign standards. 29 with this issue of causality, focusing instead on estimating if this exporter premium also exists in Mozambique and how it compares to other countries. Results by and large confirm the idea that exporters perform "better."33 In fact, exporters in Mozambique have a labor productivity that is 135 and 145 percent higher than nonexporters, when measured by sales per worker and value added per worker, respectively. If measured using TFP, Mozambican exporters are 85 percent more productive than non exporters, and wages paid are also 67 percent higher for exporters. This difference in performance between exporters and nonexporters seems to be larger in Mozambique than in other comparator countries: the coefficients for both sets of pooled countries are smaller than for the regression with Mozambique alone. If we take the hypothesis that more productive firms self-select into exporting, these results may suggest that costs taken broadly to mean barriers to reaching foreign markets, may be much higher in Mozambique than in other countries. Only highly productive enterprises may be able to overcome these barriers. If we take the opposite hypothesis, of learning by exporting, these results show the potential benefits that an export-driven strategy can bring to the Mozambican economy. Regardless of the causality direction, these benefits extend beyond the reduction in external macroeconomic vulnerability: exporters are more productive and they pay higher wages. Thus, there are clear potential benefits of an export strategy for the country: export growth can be expected to increase the average performance of the manufacturing sector. Labor Productivity Mozambique's average labor productivity (value added per worker) is one of the lowest among the group of countries compared (only Vietnam is lower).34 South African firms have the highest productivity not only among their African comparators but also among Asian comparator countries (Figure 3.2). Labor productivity varies considerably among Mozambican firms and medium firms are more than three times more productive than small ones, and almost 50 percent more productive than large firms.35 33 See Table 9.11 in Annex IV for full results. 34 Outliers for Mozambique (5 observations) and South Africa (3 observations) were dropped from the analysis. 35 When other controls are added, this difference between medium and large firms does not persist. 30 Figure 0.2. Labor Productivity (Value Added per Worker, US$) 40,000 35,000 Small Medium 30,000 Large 25,000 US$ Overall 20,000 15,000 10,000 5,000 - a nd a e a m i aw ol qu ic bi na la fr ng m al bi ai A et Za A M am Th h Vi ut oz So M Country Source: Author's calculations based on Enterprise Surveys 2003-08 Increases in labor productivity in the manufacturing and services sector were significant. Data from the survey allows a look at changes in productivity levels among Mozambican firms between 2003 and 2006.36 Labor productivity, measured as sales per full-time employee, improved in the manufacturing and services sectors by approximately 20 and 17 percent, respectively. The construction and transportation sector, in contrast, underwent a reduction of approximately 7 percent in labor productivity. A decomposition exercise suggests that the labor productivity increase reflects improvements in allocational efficiency, as employment has become more concentrated in more productive firms. Following the methodology developed by Olley and Pakes (1996), it is possible to inquire further into the nature of these productivity changes. Results from this analysis are shown in Figure 3.3 below, which illustrates differences in labor productivity levels across the three sectors analyzed, their evolution over time, and changes in their composition meant to capture allocational efficiency shifts.37 The decomposition exercise indeed suggests that the labor 36 The 2008 Enterprise Survey includes firm information on sales, input costs, and employment for 2003 and 2006 that allow for the estimation of productivity. 37 The concept of allocation efficiency used here is specific to the Olley and Pakes decomposition and it captures whether the most productive firms are the ones with greater employment share in each industry. In a dynamic sense, increases in allocational efficiency imply that the most productive firms are the ones gaining more employment share in the industry. This is an identity decomposition exercise and it does not involve statistical estimation, as described in Box 3.1. Additionally, sample adjustments were carried out to control for lack of information on 2003 seasonal workers and its potential effect on initial labor productivity levels. Both could introduce bias in the employment share shifts or labor productivity growth. At the same time, it is important to acknowledge that particularly small sample sizes across sector and firm size may cause some results to be particularly nonrepresentative for out-of-sample comparisons. 31 productivity increase observed in manufacturing and services between 2003 and 2006 do reflect improvements in allocational efficiency within-industry; that is, employment has become more concentrated in more productive firms in these two sectors. In the case of construction and transportation, despite a reduction in overall labor productivity, it is interesting to note that improvements in allocational efficiency have taken place. In a context where employment growth is stifled (see Chapter 0 on Labor Markets and Human Capital), the positive growth of labor productivity in the manufacturing and services sectors may, more than anything else, reflect more lackluster employment generation. On the other hand, the increase in the allocational efficiency term for all three sectors suggests a more optimistic view in the sense that, despite unsatisfactory pace of formal job creation in Mozambique, employment is increasingly concentrated in more productive firms. Box 0.0. Olley and Pakes Decomposition The Olley and Pakes (1996) decomposition of productivity is useful to evaluate the allocation efficiency of firm by sector/region/size during a given period of time. It allows determining if an increase in productivity is due to a reallocation towards more productive firms or to an increase in average productivity growth. The decomposition is given by the expression: N kt ~ Pkt = Pkt + ~kjt Pkjt s t =1 N kt ~ ~ where Pkt is the average productivity of industry k and s kjt Pkjt measures the covariance t =1 between productivity and output shares. If the covariance is positive, then the larger it is, the higher the share of sales that goes to more productive firms (that is, allocation efficiency) and the higher is the industry productivity. If the covariance is negative, it cannot be interpreted as allocation inefficiency, since the more negative it is, the higher is the output share that goes to less productive firms, reducing the industry's productivity. 32 Figure 0.3. Labor Productivity Decomposition--Olley and Pakes 1.5 2003 2006 1.5 1 1 .5 .5 0 0 Construction & TransportManufacturing Services Construction & TransportManufacturing Services Unweighted Mean Alloc Efficiency Unweighted Mean Alloc Efficiency Source: Enterprise Survey 2008 Note: The data in both graphs is normalized by 2003 industry-specific productivity levels such that, for each sector, adding up the stacked bars on the left-hand side graph yields 1 by construction. Adding up the stacked bars on the right-hand side graph indicates either increase or decrease in labor productivity levels has taken place in each industry. Labor productivity gains were primarily concentrated in SMEs, suggesting that allocational efficiency dynamics were superior among those firms. It is possible to refine the above results further by applying the decomposition across different firm sizes. This analysis is presented in Figure 9.1,38 where it is observed that labor productivity gains were primarily concentrated in SMEs, with large firms displaying negative productivity growth. This result, in and of itself, is not necessarily surprising: as shown in Chapter 7, employment growth has been mostly limited to large companies. Nonetheless, the decomposition suggests that allocational efficiency dynamics were superior among SMEs. This means that labor input has been allocated to the more productive firms within that industry-size category. While it is important to exercise caution when drawing conclusions from this decomposition exercise,39 one reasonable way to read the evidence here presented is that there is no clear "efficiency-based" story explaining the differential employment growth rates between large companies and SMEs (see Chapter 0). In other 38 Annex IV: Productivity and Exports in Mozambican Firms. 39 For instance, the extensive literature on firm learning dynamics could easily justify a situation whereby allocational efficiency reduces at the onset of labor inflows only to pick up strongly later as firms age and become more productive via learning-by-doing. 33 words, the bias in job growth toward large companies and the lagging performance of SMEs are not clearly efficiency-enhancing phenomena. Due to data limitations, it is not possible to perform the same exercise above using TFP as the criterion of performance. Nonetheless, to reinforce the message about SMEs not being less productive than large companies in general, Figure 9.240 provides distributions of log TFP across firm size for the manufacturing sector.41 Despite greater dispersion, SMEs do reveal potential for superior productivity. The stylized facts here presented help to substantiate the key concern with SME development in Mozambique. Although the topic has been in the policy agenda for a while, improving its political traction continues to be necessary, especially when policy tradeoffs against continuing focus on large firms are involved. The evidence brought by the survey should help to inform such debate. Conclusions The analysis conducted in this chapter confirms the importance of the investment climate for productivity in Mozambique. The correlation varies according to the investment climate factor. The indices for finance and technology and human capital are highly correlated to improved firm performance. There is substantial scope for improvement in infrastructure, especially in the energy and transport sectors, while governance factors are less significant for firm performance, at least in the dimensions captured by the variables analyzed here: crime and security issues and corruption.42 Regional disparities are pronounced: firms in Maputo do better in all three aspects of firm performance (sales per worker, TFP, and wages per worker), although a few of the correlations are not statistically significant.43 Most of these issues seem to affect Mozambique to a larger extent than comparator countries. Results from pooled regressions confirm the importance of investment climate variables to firm performance across countries, although the correlations between them are not as high as for Mozambique only. The Enterprise Survey also allowed for a look at the evolution of labor productivity in Mozambique between 2003 and 2006. While improvements were seen in the manufacturing and services sectors, the decomposition exercise presented in this chapter indicates that the observed labor productivity increase reflected improvements in 40 Annex IV: Productivity and Exports in Mozambican Firms. 41 TFP numbers were calculated using sector-level input shares extracted from the data (cost of materials and cost of labor divided by sales) and assuming constant returns to scale to obtain the capital share. Lack of data for 2003 prevented us from repeating the same exercise done for labor productivity. 42 It is important to note that while the econometric results presented in this chapter did not point to a significant correlation between crime and firm performance, it is clear from the qualitative results presented in Chapter 2 that crime is perceived as an important constraint to business growth. Moreover concern among firms about this issue has clearly increased since 2003. 43 A control variable that designated whether or not the firm was located in an export-processing or other industrial zone was also included in the regressions. The variable itself was not significantly correlated to any of the measures of firm performance; furthermore, it did not alter significantly any of the other correlation coefficients. 34 allocational efficiency within-industry. While this improvement represents an overall positive trend for the economy, average firm productivity will also need to be raised if the country is to fully benefit from the export strategy being emphasized here. This is made evident by the econometric results presented in this chapter: significant gaps exist between the performance of Mozambican firms and that of competitor countries. Addressing the investment climate challenges identified in this report will, therefore, be crucial to increase average firm productivity, while promoting sustainable growth and improving Mozambique's overall competitiveness. This will create the conditions for the country's improved integration into the global economy, taking full advantage of its trade agreements such as SADC. 35 Access and Cost of Finance Lack of finance ranks as the second most important constraint for businesses in Mozambique. Although there are 13 commercial banks licensed to operate in Mozambique, most are concentrated in Maputo and provide services primarily to enterprises that are larger, audited, and able to provide collateral. Although 78 percent of firms surveyed had an effective demand for credit, only 13 percent of firms had obtained credit from the banking sector. While there are a small but increasing number of nonbank financial institutions, their contribution to increasing access to finance remains limited. Many SMEs, especially outside the main urban areas, still lack adequate supply of formal financial services. Limited access to finance is a serious constraint to the growth and development of the private sector in Mozambique. Finance can reduce the barriers to entry for entrepreneurs, allowing the economy at large to benefit in terms of employment, price, and quality of services. Supportive policy for financial sector development is increasingly widely acknowledged to be a key component of national development policy. Careful long-run comparative analysis of growth rates across different countries has produced convincing evidence that having a deeper financial system contributes to growth. Countries with deep financial systems also seem to have a lower incidence of poverty than others at the same level of national income. Access to financial services in Mozambique is improving, but more needs to be done to support the growth of the private sector. During the last few years, numerous government initiatives aimed at increasing the outreach of financial services and improving the performance of financial institutions have been introduced, including a financial sector technical assistance project with multiple development partners. However, the survey results contained in this report suggest that more work--in addition to the ongoing reforms to strengthen the financial sector infrastructure--needs to be undertaken to improve access to finance. This chapter will take a closer look at the survey findings to better understand the most critical financing constraints and the appropriate policy responses. The analysis focuses on the micro-level constraints to finance. It does not deal with the issues at the macro level that affect interest rates. A brief overview of the financial system in Mozambique is presented first, followed by a cross-country comparison of basic features of the financial systems for countries of interest. A summary of survey results and the implications of these results are outlined and discussed next. The chapter concludes with some policy recommendations and discussion of implementation challenges that lie ahead. The Mozambican Financial System Understanding the current financial system in Mozambique is important when examining and trying to recognize key financing constraints facing private firms. Commercial banks (primarily foreign-owned) dominate the financial sector. The three 36 largest banks alone account for over 80 percent of total bank assets. In 2007, the largest banks--the Banco Internacional de Moçambique (BIM), Banco Comercial e de Investimentos (BCI), Standard Bank, and Banco Austral (now Barclays Bank)--held altogether 87.9 percent of assets, 84.8 percent of credit, 91.4 percent of deposits, and 75.7 percent of the banking system's own funds. The geographical distribution of branches has not experienced substantial changes. Most of the credit institutions are still concentrated in Maputo city, and of the 128 existing districts in the country only 28 districts have bank branches, culminating in a poor distribution of banking services in rural areas. This concentration is also observed in the automated teller machine (ATM) and point of sale (POS) networks. Close to 60 percent of the ATMs and about 70 percent of the POSs currently installed in the country are in Maputo. The Mozambican financial market is characterized by being tight and weak. Several papers have recently tried to summarize the main problems and challenges facing the financial system in Mozambique.44 Specifically, some point to the fact that limited competition has resulted in a nondynamic financial system with low levels of financial intermediation, high profit margins (helped by excessively high fees on banking services), and a large interest rate spread between deposits and loans (to some extent driven by high reserve requirements). Excessive domestic borrowing by the government until recently (and noting that interest earned on T-bills is tax free) also put an upward pressure on interest rates and crowded out available funds to the private sector. While tight credit may reflect weak entrepreneurship and poor quality of investment projects, a careful analysis by United States Agency for International Development (USAID) (2007) concludes that interest rates on metical loans appear to be higher than economic fundamentals merit. Beyond the main commercial banks, several institutions provide finance to SMEs. These include two leasing companies and 68 microfinance institutions (three banks, one micro bank, six credit unions, and 58 microfinance lenders). However, these institutions represent a relatively minor addition to overall credit availability (Byiers et al., 2008) and average loan sizes are considerably smaller than in the commercial bank sector and they often require higher interest rates. Moreover, USAID (2007) provides anecdotal evidence for the limited coverage of informal (local moneylenders) and semi-informal (trade credits) loan markets in Mozambique. Trade credits have played an important role in financing SMEs in South East Asia, but in Mozambique previous evidence concludes that this arrangement only takes place within ethnic networks. This suggests that microfinance institutions or the informal sector do not constitute an easier and more accessible source of funding than the formal banking system. All of the above suggest that the credit market may not be functioning well in Mozambique, with inherent negative implications for future economic private sector activity and growth. Identifying which firms are credit constrained and establishing the magnitude of the financing shortage have important policy implications. It is well known 44 See USAID (2007) for an excellent overview. 37 that the wedge between costs of internal and external financing influence the effects of fiscal and monetary policy differently for constrained and nonconstrained firms. For example, Hubbard (1998) shows that contractionary monetary policy increases the cost of external financing more for credit constrained firms than nonconstrained ones, both in terms of higher interest rates and as a larger spread between internal and external funds. Investment, therefore, tends to be lower among credit constrained firms than in nonconstrained firms. Moreover, identifying characteristics of credit constrained firms can help policy makers target firms that are most likely to face binding financial constraints. Before a more detailed description of the survey results for Mozambique is introduced, some cross-country comparisons of basic features of access to finance are presented first. Cross-Country Comparisons Access to finance in Mozambique is low when compared to other African countries. A comparison of basic features of firm financing is presented in order to place Mozambique in context with some of the practices and characteristics of firms in other countries. A basic measure of access to finance is compared first. Survey results for Mozambique shows that while only 13 percent of firms indicated having a line of credit or loan from a financial institution, this number improves to 21 percent when also considering overdraft facilities. In comparison to other countries, Figure 4.1 shows that Mozambican firms have better access to credit than firms in Angola. However, firms in Zambia and especially in South Africa seem to have much easier access to lines of credit. Figure 0.1. Percentage of Firms with Overdraft Facility/Lines of Credit or Loans (2003-08) 80 67.9 70 60 50 Percent 38.6 40 30 21.2 20 10 6.3 0 Angola Mozambique Zambia South Africa Country Source: Enterprise Surveys 2003-08 Mozambican firms' financing of their working capital requirement relies to a large extent on internal funds or retained earnings. This is similar to most other African comparators; in contrast, internal financing/retained earnings accounts for much less of working capital financing in the Asian comparator countries, where bank financing plays a much larger role, especially in Thailand. In addition, Figure 4.2 shows that bank 38 financing in Mozambique is significantly less important for financing working capital than in South Africa or Malawi. Zambia and Angola, on the other hand, present similar levels of bank financing to those in Mozambique. Figure 0.2. Sources of Working Capital Financing (2003-08) Share of working capital financed (%) 100 90 80 70 60 50 40 30 20 10 0 a i a m e nd a sia aw ri c l bi qu go na ila m al e Af bi An et on a M Za am h Th Vi d ut In oz So M Country Internal funds/retained earnings Bank financing Trade credit Other Source: Enterprise Surveys 2003-08 Purchases of fixed assets in Mozambique are dependent on internal funds or retained earnings to an even greater extent than working capital. Zambia and Angola exhibit a similar structure of financing. Firms in South Africa and Malawi, while also using retained earnings for the majority of its new investment funding, show a much greater proportion of funding that comes from bank sources. Asian comparator countries, namely Thailand and Vietnam, stand out again for their much larger share of investment funding derived from banking sources (Figure 4.3). Finally, a large share of Mozambican firms report having to post collateral in order to receive loans or lines of credit. Levels in South Africa and Malawi are significantly lower. However, the average value of the collateral required is lower in Mozambique (as a percentage of the loan being requested) when compared to Zambia and South Africa, while essentially the same as values in Malawi. 39 Figure 0.3. Sources of Investment Financing (2003-08) Share of fixed assets investment financed (%) 100 90 80 70 60 50 40 30 20 10 0 e a a nd m i la a aw qu bi si na ric go la ne m bi al ai Af et An Za am do M Th Vi h In ut oz So M Country Internal funds/retained earnings Bank financing Trade credit Other Source: Enterprise Surveys 2003-08 Figure 0.4. Collateral Requirements--Incidence and Burden (2003-08) 200 148 153 154 150 101 101 Percent 89 91 91 100 74 61 50 0 South Africa Malaw i Vietnam Zambia Mozambique Country % of firms reporting having to post collateral Average value of collateral as % of loan Source: Enterprise Surveys 2003-08 Mozambique shows a clear deficit in terms of firms that have access to lines of credit or financing more broadly when compared to other countries in the region. While the sources of funding are more or less typical for the region, a greater diversification away from internal funds/retained earnings is desirable, following other countries in the region and elsewhere that show a much greater role for formal financial institutions. Lastly, comparisons with other countries show that the percentage of firms having to post collateral is high in Mozambique. These basic comparisons show that 40 Mozambique has room for improvement in integrating firms to the financial system. A more disaggregated description of the survey results is presented next. Access to Finance Access to finance has declined. Table 4.1 gives an overview of the current loan situation for firms interviewed in the survey. As previously mentioned, only 13 percent of firms have a line of credit or a loan from a formal financial institution. Compared to previous industrial surveys in Mozambique, this number is significantly lower than in 1997 (35 percent--Biggs, Nasir, and Fisman, 1999), in 2002 (29 percent--Enterprise Survey 2003), and in 2006 (26 percent--National Directory of Studies and Policy Analysis [DNEAP], 2006). It is especially a decline in the number of SMEs having bank loans that drives the decrease. A possible explanation for this decrease is suggested by the figures for overdraft utilization. Table 4.2 shows that, while the percentage of firms having bank loans has decreased, the use of overdraft facilities has increased across all firm sizes. This may suggest that there has been some degree of substitution in how firms finance their operations.45 Table 0.1. Characteristics of Loans and Credit Lines in Mozambique Full Sample Excluding Microenterprises Obs. Percent Obs. Percent Has a line of credit or loan (79) 13.3 (60) 12.8 Details on most recent line of credit or loan Loan Line of Credit Loan Line of Credit Loan or line of credit (48) (31) (32) (28) 60.8 39.2 53.3 46.7 Collateral required (45) (27) (29) (24) 93.8 87.1 90.6 85.7 Average annual interest rate 21.4 22.5 22.4 21.6 Details on creditor Loan Line of Credit Loan Line of Credit Private commercial bank 70.2 77.4 87.1 78.6 State-owned bank/government agency 0.0 6.5 0 7.1 Nonbank financial institution 29.8 12.9 12.9 10.7 Other 0.0 3.2 0 3.6 Source: Enterprise Survey 2008 Note: Loans are on average larger in size and have longer duration as compared to a line of credit. Obs. = observations. 45 Care must be exercised with this comparison since the categories used for firm size in 2003 were slightly different than in 2008. Notwithstanding this fact, the substitution explanation may still be plausible. 41 Table 0.2. Use of Bank Financing in 2003 and 2008, by Firm Size Size Percentage of Firms that Percentage of Firms that Use/have Overdraft Use/have Loans Facilities 2003 2008 2003 2008 Micro 0 8 17 15 Small 6 8 18 9 Medium 10 21 36 17 Large 22 41 41 41 Very large 41 - 59 - Source: Enterprise Surveys 2003, 2008 -- Not available Most lending is collateralized (around 90 percent of loans) and the average annual interest rate of around 22 percent seems independent of the type of loan considered. Loans are generally characterized by being larger and having a longer duration (term) as compared to a line of credit. As expected, most bank credit is extended by private commercial banks. A firm-size analysis of the data further suggests that larger firms have fewer difficulties in accessing credit (even after controlling for several different proxies for access to collateral and loan transaction costs). Micro firms on average obtain formal credit as often as SMEs. However, micro firms are more likely to need collateral when applying for credit. Moreover, loans obtained by micro enterprises are often relatively small and they are often provided by nonbank financial/microfinance institutions (61 percent). Finally, keeping external audited accounting books seems an especially efficient tool to reduce credit constraints (Table 4.3). Table 0.3 Current Financial Market Participation Characteristics by Firm Size Micro Small Medium Large Have loan or line of credit 15.1 8.5 17.0 40.7 Percent with no collateral 0.0 11.5 12.5 9.1 Percent of loans provided by nonbank financial institutions 61.1 15.4 13.0 0.0 Average annual interest rate 22.3 24.0 21.0 17.5 Have current account or savings account 68.1 69.9 87.9 85.2 Have an overdraft facility 8.4 8.2 21.3 40.7 Have annual financial statements checked and certified by 27.7 28.4 60.3 74.1 external auditor Source: Enterprise Survey 2008 Note: All numbers in percentage 42 Use of other potential collateral substitutes increases with firm size. In terms of other financial services that can be seen as potential collateral substitutes (for details, see Bigsten et al., 2003), the number of firms with (i) audited accounting books, (ii) checking/saving accounts, and (iii) an overdraft facility seems to increase with firm size (there seems to be evidence of a threshold between SMEs). However, our data does not suggest that Mozambican firms use overdrafts more than loans. Firms, regardless of size, rely primarily on internal funds/retained earnings for their financing. Table 4.4 provides figures related to firm investment. Results show that (i) the probability of carrying out new investments increases with firm size, and that (ii) financing is done primarily using internal funds/retained earnings, even for large firms. Interestingly, only one firm in the sample obtained investment financing through informal sources (a moneylender or from family or friends), countering the view that informal loans play a major part in micro and small enterprise (MSE) financing. This aspect of the Mozambican credit market is notably different from that in South East Asia, where informal sources of lending play a more important role in SME financing. Rand (2007), for example, reports that in the case of Vietnam almost one-third of credit to SMEs comes from informal sources. Table 0.4. Investment Financing Micro Small Medium Large Total Purchased fixed assets in 2006 26.1 31.4 39.7 66.7 33.9 Percentage financed by internal funds/retained earnings Micro Small Medium Large Total Fixed assets 89.7 90.8 86.4 74.7 87.8 Working capital 85.4 81.3 74.2 73.9 80.1 Source: Enterprise Survey 2008 Note: All numbers in percentages Characteristics of Credit Constrained Firms An econometric analysis suggests a highly credit constrained environment in Mozambique, with more than two-thirds of the formal firms being credit constrained.46 Recognizing that credit constraint studies might be subject to selection bias since not all enterprises have credit demand. The survey analysis (i) classified firms with demand for external finance, and (ii) established the characteristics of credit constrained Mozambican firms, conditional on credit demand. The study found a highly credit constrained environment. More than two-thirds of the sample was credit constrained. The percentage increased to 82 percent when the sample was of credit demanding firms that had (i) applied and has been denied credit or (ii) did not apply for credit due to reasons such as "application procedures to complex," "collateral requirements unattainable," or "possible loan size and maturity insufficient." 46 See Annex III: Modeling Demand for External Finance and Credit Constraints. 43 High interest rates keep firms from applying for loans. Analysis of the survey results for firms that had applied but not succeeded in getting a loan is particularly revealing. Almost half of the firms give "interest rates not favorable" as the main reason for not applying for loans: these are mainly non-micro firms who state this as a problem. As in Bigsten et al. (2003), the survey does not include firms answering "interest rates too high" or "did not believe it would be approved" as being part of the constrained firms as these answers may reflect that investment projects are simply not competitive at going interest rates. Collateral requirements seem to be mainly responsible for rejection of credit. Some of the firms applying and being denied credit may not have had a sustainable business plan and could, therefore, not be labeled as credit rationed. The questionnaire enabled us to provide reasons for why the project proposal was rejected. For between 48.5 and 55 percent of the cases, the rejection of an application for credit was based on collateral issues. In the remainder of cases, the reasons ranged from "insufficient profitability (of the project proposal and historically)," "problems with credit history," or "other objections." Increased auditing could help improve access to finance through lessening of collateral requirements. It is notable that firms without audited accounts are only 9­14 percent less likely to receive financing. Ordinarily, in a country with an active accounting and audit profession, the percentage would have been substantially much higher. In Mozambique, this is not the case. Accounting and audit standards in the country are absent or weak at best. In addition, the country is currently suffering from a serious shortage of qualified accountants and trained accounting technicians. SMEs are especially affected. Consequently, banks do not place any reliance on corporate financial statements. Lenders, in general, manage credit risks through the use of collateral. Moreover, personal relations and familiarity with the performance of a business entity are considered more important for business decisions than access to reliable financial information. With the exception of a few large entities with foreign participation, companies in Mozambique are still largely family-owned and the business community generally sees limited interest in external financial reporting. In fact, many businesses are reluctant to provide transparent financial information, mostly for fear of tax consequences. The absence of reliable financial statements in the private sector then reduces the potential for cash flow­based lending. Banks and financial institutions are compelled to rely on collateral even when reliable financial statements might have otherwise facilitated cash flow­based lending. Other noteworthy results of the analysis are the following:47 · Largely, firms face financing constraints regardless of size. Only in two of the cases modeled does the study find a well-determined negative association for supporting the perception that credit constraints are more binding for smaller firms. 47 See Annex III: Modeling Demand for External Finance and Credit Constraints for complete results. 44 · Firms that are audited are less likely to be credit constrained. More specifically, the study finds that those enterprises with external audits are 4 to 19 percent less likely to be credit constrained than firms not keeping formal accounts. · Firms with a higher share of foreign ownership are less likely to be credit constrained. · Credit demand for firms in the food processing sector seems particularly sensitive to the cost of credit, whereas enterprises in retail are less sensitive. However, firms in retail, on average, do seem to have a relatively lower demand for external funds. Firms in the category "other manufacturing" are the most financially constrained of the categories considered. Why Has Access to Finance Decreased? The decline in the proportion of firms that have a line of credit or a loan from a formal financial institution in the survey is surprising. Explaining the reduction by other than being an artifact of different representative samples could point to a number of possibilities: i. First, it maybe that this survey merely denotes a better, albeit lower, credit quality portfolio because the privatizations in the financial sector and the private sector may have resulted in more selective lending. Thus, while the volume of loans may have increased, it was directed to a smaller number of better-managed, privately owned firms. By the end of 2006, the number of companies controlled by Mozambique's Institute for the Management of State Holdings (IGEPE) had fallen from the 279 it took over in 2002 to 155 companies, which are heavily concentrated in the real sector (investments in the financial sector are now limited to a small shareholding in BIM and majority ownership of EMOSE, Mozambique's largest insurance company). Of the 124 companies that left IGEPE's portfolio, some 33 were privatized, with 76 percent of sales made to private investors and the remaining 24 percent to companies' managers and employees. The balance of the reduction in the number of companies (91 firms) was achieved by the formal closure and liquidation of unviable companies, many of which had been inactive for more than a decade. In addition to selling and closing companies, IGEPE has also focused on restructuring larger state-owned enterprises (such as Telecomunicações de Moçambique (TdM) and Petromoc) in order to create viable businesses. ii. Second, the reduction could point to the recent regulation Aviso 5/2005. The regulation resulted in a decrease of aggregate lending in foreign currency from 60 percent in early 2005 to 36 percent in late 2006. Several firms now had to borrow in meticais at a higher interest rate (around 25 percent) rather than at an interest rate of about 10 percent. According to USAID (2007), this interest rate gap was greater than exchange rate risk alone would suggest and the Aviso 5/2005 might have reduced credit demand for certain types of firms.48 A recent competitiveness review of the banking sector argues that banks are dissatisfied with current regulations 48 For additional details on the Aviso 5/2005, see USAID (2007), Section 8. 45 requiring an automatic provision of 50 percent against loans denominated in foreign currency, except where the borrower is an exporter (additional credit provisions are required depending on the performance of the loan). Banks point out that: · Approximately 55 percent of bank deposits are in foreign currency. Forcing banks to lend in metical, thus, exposes them to risks from metical inflation and currency mismatches; · The provision requirement substantially increases the cost of foreign currency denominated loans, driving up borrower costs; · New entrant banks see the provision requirement as a significant barrier to competition with longer established banks because the latter have had more time to build metical deposit bases. A lack of metical liquidity was described as the main barrier to expanding lending by new entrant, SME-focused banks. iii. Third, but likely to a lesser degree, the improving regulatory, accounting, and provisioning regulations, coupled with the entry of international banks into the sector with home country regulators requiring higher standards of credit assessment, may have contributed to a reduction in firms with access to easy credit. iv. Finally, the evidence presented in this chapter shows that while the percentage of firms having bank loans has decreased, the use of overdraft facilities has increased across all firm sizes. As mentioned earlier in this chapter, this may suggest that there has been some degree of substitution in how firms finance their operations. Conclusions A deliberate government strategy to improve and expand financial services in Mozambique is beginning to yield tentatively positive results. In 2007, the central bank announced a package of initiatives targeted to reduce the operational costs of providing banking services in rural areas through infrastructure improvements, and by relaxing reserve requirements in rural branches so as to include cash in vault. Progress was also made in improving the institutional lending environment. In October 2007, Parliament approved a new bankruptcy and insolvency law for banks, which is expected to help accelerate dispute settlement and improve contract enforcement. In terms of infrastructure, Parliament approved a new national payment system law consistent with the SADC payment system initiative, and the central bank has launched a survey of the banking sector to analyze the rationale for the fees and charges structure. In addition to growing bank interest in retail banking, these incentives contributed to the growth in financial institutions and total bank branches. During 2007, the central bank approved nine new license applications (one micro bank and seven credit unions), and total bank branches grew to 246 (compared to 228 in 2006). Bank branches are now present in 32 districts around the country (28 districts in 2006). The number of savings accounts also increased to an estimated 128,629 (88,972 in 2006), the majority of which are held with microfinance institutions (53,421 in 2007). With an estimated population of 21 million, however, this still represents a small albeit growing percentage of the population with access to formal financial services. 46 Banking Sector Competitiveness Despite the positive steps mentioned above already taken by the government, important and fundamental work remains to be done. The challenges the government faces in increasing access to finance, such as poor infrastructure and human capital, are structural and will take time to address. Improving the credit registry, property registry, accounting and audit standards, and the judicial system are all essential recommendations for improving access to finance. On many of these activities, the government is already in the process of taking action. However, these efforts require time to generate tangible improvements. In the meantime, it is essential that the government work on increasing the competitiveness of the banking sector by supporting the increase in the number of financial institutions and products and by improving the quality of the users of financial services. Key recommended policy actions to improve competition among banks include: (i) taking steps to promote greater transparency of bank fees and charges; (ii) increasing financial literacy; and (iii) encouraging the unbundling of financial services. Publishing fees enables the public to choose between banks and financial services. Increased financial literacy empowers individuals to better manage their earnings, and hence, to better manage life events like education, illness, and retirement. In addition, financial literacy, together with the unbundling of financial services and products enables clients to move between financial institutions, increasing pressure on banks to behave more competitively, to improve the quality of the services that they offer, and to be more responsive to client's needs. High Collateralized Lending To address the significant percentage of firms that are not able to secure lending due to unattainable collateral requirements, the government may wish to consider exploring alternative mechanisms for collateral-based security. One such option is supporting the establishment of credit risk guarantee facilities. Similar risk-sharing instruments can facilitate greater lending to SMEs by reducing the risk that commercial banks face when lending to SMEs. Admittedly, there is always the risk that the partial credit guarantees may not be utilized as anticipated due to lack of interest from private banks. Designing the scheme and confirming firm demand with commercial banks is, therefore, critical to its success. If well designed, there is a potentially important catalytic role for a partial credit guarantee scheme in increasing access to finance. Lack of Audited Accounts The absence of reliable financial statements in the private sector reduces the potential for cash flow­based lending, pushing banks and financial institutions to rely on collateral when deciding loan approvals. Therefore, it is paramount that the government accelerate its efforts to strengthen the private sector's accounting, financial reporting, and auditing practices, as a means to increase confidence in the financial statements, and in turn, the level of reliance on these statements by banks in their lending decisions. 47 Financial sector development is a strategic driver of growth and employment in Mozambique. This survey found that firms view access to and cost of finance as key constraints to doing business. Finance in Mozambique has been strengthened by a wave of reforms since the 1990s. However, despite recent improvements, more needs to be done. The challenge is to deepen the Mozambican financial system and increase its intermediation capacity, particularly for SMEs. The government has a major policy role in making finance work for Mozambique as a convener, facilitator, and enabler in promoting competition, ensuring robust regulation, and overcoming coordination failures in providing financial system infrastructure. 48 Informality in Mozambique Interviewed managers in Mozambique rated competition from informal firms as the top bottleneck to their business. When compared to other countries, informal competition does not appear to be as problematic for Malawi, South Africa, Indonesia, or Thailand, where this issue does not come up as one of the most often cited constraints faced by businesses. However, informal competition does seem to be a significant constraint for many firms in Vietnam and Zambia, where it ranks as the most often cited and second most often cited obstacle to business, respectively. Moreover, businesses tend to be particularly concerned with the unfair competition associated with informality: recent World Bank surveys of firms around the world show that firms tend to rank competition from informal firms as one of their top three obstacles to doing business. This can be seen in 14 Latin American countries, where 38.7 percent of the manufacturing firms ranked informality ahead of issues such as tax rates and access to finance (Gonzalez and Lamanna, 2007). This chapter explores in more depth different aspects of informality in Mozambique. It includes a more detailed description of the Enterprise Surveys results for firms as a whole as well as an evaluation of the characteristics of microenterprises, the impediments to their growth, and some of the possible reasons behind their informality. These are essential issues that should inform the design of appropriate policies to enable microenterprises to grow in Mozambique. As investment climate surveys are a limited instrument to explore causes and consequences of informality, the chapter begins with a cross-country analysis of determinants of informality, trying to assess the relative importance of factors like business regulation compared to more structural ones (like educational levels of the workforce) to explain the high levels of informality in Mozambique. Assessing the Impacts of Regulation, Governance, and Structural Factors on Informality There is considerable evidence linking the incidence of informality with the institutional characteristics associated with the regulatory environment, the prevalence of the rule of law, and the quality of government in general. In an attempt to include both the exclusion and exit perspectives of informality (see Box 5.1), Loayza (2009) developed a model that estimates how institutional and structural factors contribute to explaining differences in informality levels across countries. In order to have a general assessment of the impact of regulatory factors on informality in Mozambique, the model is applied to Mozambique using Chile, a fast-growing developing economy where levels of informality are relatively low, as a comparator country. According to the cross-country model, informality decreases when law and order, business regulatory freedom, or schooling achievement rise, as well as when the share of agriculture and the demographic pressures from youth and rural populations decline. The main features of the model are reproduced in Table 5.1, where four proposed determinants of informality (explanatory variables) representing 49 institutional and structural factors are presented.49 Three different measures of informality are regressed against these explanatory variables. It is important to highlight the statistical and economic significance of the four coefficients for each explanatory variable and their expected signs, as well as the large share of cross-country informality variation that is explained by the model. The model predicts the various measures of informality in Mozambique quite well. The coefficients for each explanatory variable derived from the cross-country regressions can be used to estimate the level of informality in Mozambique. Comparing these predictions with the actual informality values for Mozambique, Figure 5.1 shows that the model predicts the various measures of informality quite well, explaining between 98 percent and 107 percent of the actual indicators.50 Differences in levels of informality between Mozambique and Chile can also be compared. The predicted and actual differences between the two countries for each one of the informality measures are calculated. Figure 5.2 shows that the model gives fairly accurate predictions for the differences in informality levels using both the Heritage Foundation Index and Schneider's Shadow Economy Index. When informality is measured by the percentage of labor force without pension coverage, the model slightly overpredicts the level of informality. This model can then be used to assess the relative importance of each one of the independent variables in explaining the differences in predicted informality between Mozambique and Chile. By multiplying the estimated coefficients by the differences across countries in the averages of the corresponding variables, it is possible to decompose cross-country informality differences in fractions attributable to institutional and structural factors. 49 These variables are: (i) an index on the prevalence of law and order--obtained from The International Country Risk Guide (ICRG)--to proxy for the quality of governance and institutions; (ii) an index of business regulatory freedom--taken from the Fraser Foundation's Economic Freedom of the World Report--to proxy for the quantity of regulations imposed on businesses; (iii) the average years of secondary schooling of the adult population--taken from Barro and Lee (2001)--to control for the human capital of the labor force; and (iv) an index of sociodemographic factors--constructed from the World Bank's World Development Indicators database, as well as International Labor Organization and UN data--which includes the share of agriculture in GDP, the share of youth in the population, and the share of rural population. 50 Actually, the model predicts quite well the levels of informality in most countries; there is only a small number of countries for which the residual is relatively large and significantly different from zero. This seems to suggest that there is low variance in the coefficients across countries. 50 Box 0.0. Different Views on the Costs and Benefits of Informality Interviewed managers in Mozambique rated competition from informal firms as the top constraint to their business. Concerns with practices from informal competition have shown up more often as a significant obstacle to business in the most recent round of Enterprise Surveys. This is not confined to African countries, as it has also been noted in the most recent surveys in Latin America. Unfair competition is one of the channels through which informality can have negative effects on firm productivity, curtailing the prospects of economic growth. The rationale is that firms that are insufficiently competitive and that in other circumstances would be forced to withdraw from the market can remain active in the informal sector, slowing down the flow of human and capital resources to more dynamic and productive sectors, triggering a low productivity trap. From a theoretical point of view, there are several other potential mechanisms by which informality may impact growth, including providing incentives for firms to remain at a suboptimal size, lowering incentives to innovate, and the general undermining of the rule of law and governance that results from tax evasion and noncompliance with safety, health, and work regulations. The economic literature has most often associated informality with negative economic outcomes. Moreover, the prevailing view has been that firms and workers are driven into informality by stifling regulations and burdensome taxations. As argued by Perry et al. (2007), however, there is considerable evidence suggesting that the sector is fairly heterogeneous, with workers and firms that have been excluded from the formal economy coexisting with others that have opted out on the basis of implicit cost benefit analysis. To a certain extent, firms can choose their "degree" of formality, weighing the cost reductions associated with tax and social security evasion against the risk of being caught by regulators and the possible negative effects that informality may have on their ability to widen their markets and hire higher-productivity workers. Firms may, thus, opportunistically choose to evade a fraction of their taxes and social security contributions. In this context, policy interventions aimed at reducing informality are likely to require both "sticks" and "carrots." By increasing the costs of informality (for example, increasing enforcement of legislation) and/or the benefits of formalization, policies can induce firms to move towards formalization. However, the impacts on firms will likely depend on the "depth" of informality: firms that are at the fringe may be more prone to react favorably to "carrots," while firms that are "deep into informality" may require more profound changes on both legislation and enforcement mechanisms. In this sense, these actions may need to be complemented by reforms aimed at improving the quality of governance and institutions in general. Whereas this is certainly a complex and long-term agenda, it is a critical one for maximizing the impact of improved enforcement on firms' levels of regulatory compliance. The means by which informality is reduced is important and likely to determine the corresponding impact on productivity and growth. For example, increasing enforcement without addressing the underlying causes of informality could lead to unemployment and reduced growth. Finally, there is growing evidence that for many individuals there are not just monetary gains from being self-employed but also nonmonetary gains that may need to be considered when formulating policies aimed at reducing informality (Oviedo 2008). Evidence of this is reported in Perry et al. (2007) for Argentina, Colombia, Bolivia, and the Dominican Republic, where a large proportion of self- employed informal workers cite as the main reasons for being informal not only the higher earnings that self-employment brings but also the greater flexibility (especially for women). In addition, some individuals in Latin America report that the social protection benefits offered by formal job contracts do not provide good enough services to be worth the contributions, which are therefore seen as useless taxation. This is particularly relevant for people who live in remote areas, where infrastructure and public service provision is poor. In developing countries, informal protection networks also substitute for the state in many cases, with lower transaction costs and more efficient monitoring, so that again people see little benefit in having access to formal services. For instance, Morduch (1999) reports results from numerous studies that find informal insurance has a significant role in reducing the impact of income shocks. Finally, some informal salaried jobs might offer training opportunities, which are valuable especially for young, uneducated workers, for whom training could increase their chances of landing a formal job later in life. 51 Table 0.1 Cross-Country Determinants of Informality Informality measures Schneider Heritage Non- Shadow Foundation contributor to Economy Index Informal Market Pension Scheme (% of GDP) Index (% of labor (1-5: higher, more) force) [1] [2] [3] Law and order -3.2360** -0.0969* -2.9764* (index from ICRG, range 0-6: higher, -2.57 -1.76 -1.67 better; country average) Business Regulatory Freedom --2.0074* -0.5333*** -5.8675** (index from Economic Freedom of the -1.80 -9.95 -2.28 World by The Fraser Institute, range 0- 10: higher, less regulated; country average) Average Years of Secondary Schooling -1.9684* -0.1152** -5.8114*** (from Barro and Lee (2001); country -1.70 -2.00 -3.27 average) Sociodemographic Factors 3.8438** 0.5027*** 21.6130*** (simple average of share of youth 2.00 4.99 7.31 (aged 10-24) population, share of rural population, and share of agriculture in GDP; country average) Constant 60.3429*** 6.6326*** 113.3110*** 10.48 31.72 11.40 No. of observations 84 86 70 R-squared 0.57 0.89 0.88 Source: Loayza (2009) Notes: t-statistics are presented below the corresponding coefficients. See Loayza (2009) for further information on countries included in each regression, definitions and sources of variables used to compute country averages of informality measures. *, **, and *** denote significance at the 10 percent, 5 percent, and 1 percent levels, respectively. 52 Figure 0.1. Predicted Informality in Mozambique as Percentage of Actual Indicator 120 107 107 98 100 80 Percent 60 40 20 0 Schneider Shadow Heritage Informality No Pension Economy Index Coverage Source: Authors' calculations, based on Loayza (2009) Figure 0.2. Predicted/Actual Differences in Informality for Mozambique and Chile 120 110 99 100 100 80 Percent 60 40 20 0 Schneider Heritage No Pension Shadow Economy Informality Index Coverage Source: Authors' calculations, based on Loayza (2009) Both institutional and structural factors play a significant role in explaining the differences in informality levels between Mozambique and Chile. Figure 5.3 shows that both structural factors, especially sociodemographic ones, and institutional factors, particularly the regulatory burden indicator, do play a significant role in explaining the differences in informality levels between the two countries. Business regulation explains between 25 and 56 percent of the differences in informality between Mozambique and Chile, while the proxy for law and order explain much less, between 7 and 31 percent, depending on the index of informality used. The proportion of the difference explained by sociodemographic factors also varies according to the different indexes, explaining as much as 51 percent of the differences when the no pension coverage index is used. These results hint at the fact that reevaluating the country's regulatory framework and how it affects businesses operations will help to combat informality in Mozambique, but that 53 changes in more structural factors, such as rural/urban and demographic compositions, will tend to reduce informality in the long run. Figure 0.3. Contribution to Differences in Predicted Informality, Mozambique and Chile 60 56 51 50 40 31 Percent 28 30 30 26 25 20 15 15 9 10 7 7 0 Law and order Business Average years of Sociodemographic regulation freedom schooling factors Schneider Shadow Economy Heritage Informality Index No Pension Coverage Source: Authors' calculations, based on Loayza (2009) Firm-Level Data: The Enterprise Survey Results Perception of informality as a major obstacle to firms' growth is widespread in Mozambique. More than 50 percent of firms interviewed by the Enterprise Survey identified informal competition as a major constraint to business. Understanding the types of firms that perceive informal sector practices as detrimental to their businesses might help identify the reasons for their concerns. According to the survey, size does not seem to be an important factor since there is no strong pattern emerging among different categories of firms. More small firms seem to perceive informal competition as a major or severe constraint (55.7 percent of them), but the percentage of large, medium, and micro firms that had the same perception was also fairly high (43.3, 46.5, and 46.7 percent, respectively). More importantly, informal competition was the most often cited or among the most often cited constraints for businesses across all sectors, regions, exporters, and foreign ownership status. Specifically, approximately 55 percent of firms in Maputo voiced concerns about competition from the informal sector as opposed to 42 percent of firms operating outside the capital city. A greater percentage of manufacturing and retail firms perceive competition from informal firms as an important constraint (51 percent of firms in each sector), while only 43 percent of firms located in other sectors had the same perception. This finding is not unrelated to the fact that retail firms or those operating in light manufacturing sectors such as food processing are more susceptible to informal competitors than firms in other sectors. Finally, as expected, more nonexporters rank competition from the informal sector as a severe or major constraint for business than exporters (52.0 to 34.6 percent), confirming that firms with export operations face 54 less competition from informal firms, which are less likely to have access to international markets. Measured alternatively by perceived tax evasion, informality is closely correlated with business environment indicators. An exercise was carried out to complement the qualitative data resulting from firm's perceptions on informality. The Enterprise Survey asks firms' perceptions on the degree of tax evasion when reporting company's results51 in their own sectors. A simple Ordinary Least Squares (OLS) regression was run to compute correlations between this perception variable and a few selected firm-level correlates.52 Table 5.2 illustrates that, perhaps not surprisingly, firms that indicate tax rates as either a major or a very severe obstacle to business are also more likely to perceive greater tax evasion in their own sectors. On the other hand, firms that report spending more time dealing with government regulations are less likely to perceive pervasive amounts of sales underreporting in their industries. The same trend is seen for firms that generally have a positive impression about the rule of law in Mozambique, for foreign-owned firms, and for firms that were formal at the time of start up. Finally, while firms that report having been asked for informal payments when requesting public services (our indicator for incidence of corruption) do perceive greater tax evasion in their sectors, the correlation lacks statistical significance. Table 0.2 Firm-Level Correlates of Perceptions on Tax Evasion Unreported Sales Indicator Corruption indicator 5.588 Bureaucracy indicator 1.271*** Tax rate as constraint 11.232** Rule of law indicator 3.482** Foreign owned 13.595*** Formality status at time of start up 9.711* Source: Authors' calculations using Enterprise Survey 2008 * significant at 10%; ** 5%; *** 1% Microenterprises and Informality The Enterprise Survey included a separate survey of 120 microenterprises, defined as those firms with less than five employees. Microenterprises represent 20 percent of all firms in the sample, comprising 5 percent of the manufacturing firms and 44 percent of the establishments in the services sector in Mozambique. In general, microenterprises face the same obstacles as other firms, but some factors pose greater challenges for them. A major disadvantage is labor qualification: only 14 51 The exact question that firms are asked to respond to is, what percentage of total annual sales would you estimate a typical establishment in your sector of activity reports for tax purposes? In the analysis conducted here we subtracted these responses from 100 percent to create our perception variable of tax evasion. Strictly speaking this is not a direct measure of tax evasion (and thus informality), but it is a measure of how pervasive of an issue firms may think this is. 52 See Annex II: Informality Correlation, for details on firm level correlates used in this analysis. Firm size and age, labor productivity, sector and location are used as control variables. 55 percent of the top managers of small and microenterprises have studied beyond secondary education, compared to the median of 30 percent for all firms (Figure 5.4). Microenterprises also endure a heavier burden when confronting corruption. Indirect costs due to bribery payments correspond to 3.5 percent of microenterprise sales, a much higher rate compared to the burden faced by other firms (Figure 5.5). Figure 0.4. Percentage of Top Managers with Education beyond Secondary 90 80 70 60 Percent 50 40 30 20 10 0 Micro Small Medium Large Firm size Source: Enterprise Survey 2008 Figure 0.5. Indirect Costs of Bribes by Firm Size (as percentage of sales) 4 3.5 3 2.5 Percent 2 1.5 1 0.5 0 Micro Small Medium Large Firm size Source: Enterprise Survey 2008 Defining formal microenterprises as those firms registered with government authorities for tax purposes, only 26 percent of the micro firms in the sample are informal. Therefore, comparisons with formal microenterprises and conclusions drawn from this data need to be taken with caution. With this caveat in mind, we examine the characteristics of the group of firms that chose to register and formalize versus those that did not, in order to identify key factors governing such choices. Economic activity of 56 microenterprises is highly concentrated in two cities, Maputo and Beira, where 80 percent of them are located (Table 5.3). Incidence of informality differs depending on the urban area: almost all microenterprises in Beira are formal, whereas more than 30 percent of those ones located in Maputo do not have tax registration. The majority of microenterprises surveyed operate in the services sector (86 percent) and only 14 percent in the manufacturing sector. More than half of the microenterprises (51 percent) have sole proprietorship. Table 0.3 Location of Formal and Informal Microenterprises Informal Formal Total Beira 1 23 24 Maputo 22 49 71 Matola 7 8 15 Nampula 1 9 10 Total 31 89 120 Source: Enterprise Survey 2008 The Investment Climate for Microenterprises Formal and informal microenterprises deal with similar obstacles, but perceptions differ for important issues such as tax rates and tax administration. Formal microenterprises point out that the practice of informal firms has been a very severe or major obstacle to their operations. According to the Enterprise Survey, 45 percent of formal microenterprises identified competition with informal establishments as a major or severe barrier to their business activities (Figure 5.6). These figures are reinforced by an additional question where micro entrepreneurs were asked to prioritize among issues, identifying the three most serious obstacles to operations. Informal competition is identified as the most serious obstacle. Data also shows that almost 50 percent of both formal and informal microenterprises stress the difficulty of accessing credit as a major constraint to operations. Not surprisingly, formal firms cite tax rates and tax administration more often than informal firms as a major or very severe obstacle to business. 57 Figure 0.6. Major or Very Severe Obstacles for Operations of Microenterprises 60 50 Informal Formal 40 Percent 30 20 10 0 Tax rates Electricity Political instability Courts Customs & trade regulation Corruption Access to land Business licensing Labor regulation Transport Access to finance Crime Tax administration Practices from informal competition Inadequately educated labor Obstacles Source: Enterprise Survey 2008 Institutional constraints may be more important than level of taxes in determining formality among microenterprises. The Enterprise Survey provides additional information that enables a better understanding of the barriers microenterprises face when registering for tax purposes. Table 5.4 shows that access to information on what is needed to register represents the biggest obstacle for registering a business for both formal and informal microenterprises. Informal microenterprises also emphasize the time needed to complete registration procedures and the additional administrative burdens associated with inspections and meetings with government officials as major difficulties for formalization. The Enterprise Survey shows that more than half (52 percent) of informal microenterprises do not view the financial burden of taxes as a big impediment for registration. As such, addressing institutional constraints, including improving the efficiency and transparency of registration procedures and the interaction with government agencies, may play a more important role in improving formalization rates of microenterprises than the level of taxation. 58 Table 0.4 Main Obstacles for Registration--Formal and Informal Microenterprises (percent) Obstacles to Major or Very Minor or No Obstacle Moderate Obstacle Registration for Tax Severe Obstacle Purposes Informal Formal Informal Formal Informal Formal Difficulty to get information on what 19 18 13 6 68 76 you need to do to register Time to complete registration 32 37 3 15 65 48 procedures Financial cost of completing 32 33 16 24 52 44 registration procedures Minimum capital requirements for 19 33 23 12 58 55 registered enterprises Financial burden of 52 36 6 19 42 45 taxes Administrative burden to comply with tax 29 30 13 16 58 54 laws Other administrative burdens (such as inspections and 26 25 10 15 65 61 meetings with government officials) Strict labor market 23 20 19 4 58 75 rules Bribes and informal 39 28 10 7 52 65 gifts Source: Enterprise Survey 2008 59 Box 0.0. Tax Reform for Micro and Small Firms The Government of Mozambique has taken an important step in simplifying tax compliance for micro and small businesses. The tax authority has recently introduced a Simplified Tax for Small Contributors (ISPC, in its Portuguese acronym), which aims to reduce tax compliance costs by streamlining procedures. The government expects that the ISPC will contribute to the improvement of the business environment by reducing the average number of procedures and time necessary to prepare tax returns. The government also anticipates that a lower tax rate (companies with sales volume of less than 2.5 million meticais per year will pay a tax of 3 percent of sales volume) will increase the number of taxpayers, including businesses in the informal sector. Similar programs in other countries have led to significant positive outcomes. One of the most important microeconomic reforms implemented in Brazil in the last 15 years was the introduction of a simplified tax regime for micro and small enterprises (MSEs), the SIMPLES. The overhaul of the tax regime was aimed at making the tax burden on MSEs more manageable and at encouraging small business participation and the creation of formal jobs. Brazil's complex tax structure, with multiple layers and types of taxes and different tax bases, made tax compliance difficult, particularly for MSEs. High levels of taxation added to the problem, making hiring workers particularly costly. The SIMPLES was the first attempt to standardize the tax treatment given to MSEs in Brazil. The system established procedures for tax incidence and tax revenue collection by the federal government, and together with the Complementary Law No.123 of 2006, it unified six federal taxes with two subnational ones into a single tax rate based solely on firm's annual gross revenue. The flat tax introduced by the SIMPLES helped to reduce the tax burden on MSEs. Impact evaluation studies have shown that the SIMPLES reform has been successful in reducing informality among MSEs, both in terms of formal employment and firm registration. Moreover, the higher levels of registration among MSEs driven by SIMPLES have also been found to have had a positive impact on investment, levels of capital intensity, and productivity. For the government, the introduction of the simplified system of taxes for MSEs has also brought benefits. Enterprise Survey results for Brazil (2003) also suggest that the SIMPLES reform has helped to meet the goal of improving the tax system administration by reducing the average number of federal fiscal inspections and the average duration of each inspection. In addition, the implementation of SIMPLES has meant a decrease in both the number of tax forms and the length of the unified declaration, making the cost of controlling, overseeing, and storing information on about two million MSEs significantly lower. Finally, fears of significant revenue losses have not materialized. After the implementation of SIMPLES in 1996­97, an initial decrease in tax revenues in fiscal years 1997 and 1998 was quickly overcome and a positive trend in tax revenue resumed thereafter. Therefore, evidence on the impact of microeconomic reforms such as the SIMPLES in Brazil, that aim to simplify tax compliance and reduce tax burden on MSEs, has been so far encouraging. For Mozambique's recent reforms, it will be important to conduct impact evaluation studies to assess whether the goals of the reform have been achieved. Source: For ISPC, comments received from the Mozambican Ministry of Planning and Development. For SIMPLES: Fan et al. (2008). While labor regulation is perceived as a severe obstacle to formalization but not to growth, tax rates seem to impact formalization less but represent a major obstacle to operations. Three-fourths of formal establishments and 58 percent of informal ones 60 see strict labor market rules as a very severe or major obstacle for registration of microenterprises. At the same time, labor regulation was listed as one of the least- important hurdles for operations by both formal and informal microenterprises. This suggests that while deregulating the labor market might have a positive effect on formality rates, firms do not perceive a particularly important impact on performance. Conversely, the financial burden of taxes does not seem to be one of the main constraints to registration, relative to other issues. On the other hand, as expected, formal establishments seem to be highly affected by taxation in the operation of their businesses: tax rates were the second most often cited obstacle to business and, when asked to prioritize, almost 40 percent of formal micro entrepreneurs stressed taxes as one of the most serious obstacles for their operations. Microenterprises and Financing The great majority of microenterprises uses internal funds and/or retained earnings to finance capital investments and working capital. Microenterprises are also less likely to access other financial instruments. Credit from suppliers or anticipated advances from customers represent the only other significant sources of funding for working capital requirements: roughly one-third of formal and informal establishments seem to depend on it for partial funding. Loans from banks account for only 9 and 7 percent of funding for informal and formal microenterprises, respectively (Table 5.5). The financial characteristics of microenterprises reflect the way their growth is funded. Figure 5.7 shows that most formal and informal establishments hold checking or savings accounts, while only very few of them report having lines of credit or loans. Not surprisingly, 37 percent of formal microenterprises, compared to none of the informal ones, hold audited accounts. Table 0.5. Sources of Working Capital Funding 0% Between 0 and 50% Between 50 and 100% Informal Formal Informal Formal Informal Formal Internal funds 3 3 3 8 94 89 Loan from banks 90 92 3 3 6 4 Borrow from nonbank financial 100 98 0 2 0 0 institutions Purchases on credit from suppliers and 71 61 29 36 0 3 advances from customers Others 100 99 0 1 0 0 Source: Enterprise Survey 2008 Note: Proportion of financing is shown as percentage of working capital (current assets). 61 Figure 0.7. Microenterprise Financial Characteristics 80 70 60 Informal 50 Formal Percent 40 30 20 10 0 Checking or savings Audited accounts Current line of credit account or loan Characteristics Source: Enterprise Survey 2008 Microenterprises in Mozambique identified "lack of need" as the main reason for not applying for credit lines or loans. As previously mentioned, most microenterprises use internal funds to finance their growth. But does that result from external constraints or from voluntary decisions by microenterprises? The latter could be explained by the presence of "temporary" micro entrepreneurs, typically female and unemployed workers who started their business because they could not find a job and have no interest in expanding their business.53 For informal microenterprises, the complexity of having a loan approved is also more important than the existence of high interest rates in discouraging potential borrowers.54 More puzzling, however, is the presence of higher levels of "disinterest" for borrowing in the formal sector. Figure 5.8 shows that more than 50 percent of formal microenterprises report not having a need to apply for loans. One possible explanation is lack of an "indebtedness culture" due to previous downturns in economic cycles. 53 See Perry et al. (2007), p. 150. 54 According to Bila and Rand (2008), the high interest rate charged by microfinance institutions is a consequence of the labor intensive nature of these activities and the high cost of services delivered. Therefore, according to the authors, one way to reduce the high interest rate and increase the supply of microfinance credit is to promote initiatives that will cut the cost of these services. 62 Figure 0.8. Reason for Not Applying for Line Credit or Loan No need for loan Interest rates not favorable Complex application Collateral unattainable Did not think it w ould be approved Size of loan and maturity insufficient Formal Informal Other 0 10 20 30 40 50 60 Percent Source: Enterprise Survey 2008 Conclusions Interviewed managers in Mozambique rated competition from informal firms as the top bottleneck to their business. While analysis based on hard data did not confirm a significant impact of informality on productivity (see Chapter 3 of this report), the simple fact that it is perceived as the main constraint by most firms regardless of economic sector, firm size, region, and exporter status suggests the topic deserves to be followed by policy makers in the country. Using both firm-level data from the Enterprise Survey and cross-country evidence, this chapter tried to provide some evidence on the relationship of informality with the regulatory environment. Perceptions of informality are correlated with business environment indicators and, more importantly, cross-country evidence suggests that business regulation plays a significant role in explaining the differences of informality levels in Mozambique when compared to countries with a lower level of informality, such as Chile. However, the model suggests that structural factors also play an important role. While addressing the main flaws in the business environment will certainly contribute to reduce informality and likely have a positive impact on growth, other long-run policies are required. Informality among firms is a complex phenomenon, involving both exit and exclusion. It has negative consequences in terms of fiscal revenues, productivity, and access to markets. On the other hand, informal firms do generate minimum levels of income to a significant number of workers. In this context, policy interventions aimed at reducing evasion--or informality among formal firms--are likely to require both "sticks" and "carrots." 63 Microenterprises are more likely to be informal than other establishments because of their low productivity. Because of their small size, microenterprises may find that using informal arrangements for accessing inputs--for example, obtaining credit from family and friends rather than from formal financial institutions, or counting on trust rather than on formal contract enforcement mechanisms--is more efficient than having to pay the costs associated with regulatory compliance.55 On the other hand, wages and labor skills in informal microenterprises tend to be lower than in the formal sector.56 In particular, among informal microenterprises, the incidence of low productivity in many cases is a cause of informality and not a consequence of it. For example, micro entrepreneurs with very low levels of human and physical capital are likely to operate very small and low-productivity businesses with limited growth potential. Interventions should prioritize productivity improvements. In the case of entrepreneurs located farther away from the formal/informal frontier and with limited growth potential, incorporating them into the formal economy may require other types of interventions in addition to reducing registration and regulatory compliance costs, whose effects are likely to materialize only in the medium to long term. Those measures may include changes in educational policies and improvements in the investment climate aimed at raising productivity in the formal sector, and reducing the attractiveness of operating informal businesses. 55 Empirical evidence shows that higher levels of regulatory burden are correlated with lower growth and higher informality rates. However, the quality of regulation matters and can mitigate its harmful effects on growth. See Loayza, Oviedo, and Serven (2005). 56 Fajnzylber, Maloney, and Rojas (2006) argue that firms operating in the formal sector employ more paid workers, are more capital intensive, and exhibit higher levels of total factor productivity. 64 Infrastructure Infrastructure plays a critical role in fostering growth and productivity and reducing poverty and inequality. Adequate infrastructure facilitates access to input and output markets. It increases the productivity of existing firms and it gives incentives to the creation of new businesses. Good infrastructure facilitates the access to diversified labor markets, to better health and to knowledge, strengthening the overall level of human capital in the economy. Finally, infrastructure allows firms to grow and reach an efficient scale of production, gaining competitiveness internally and internationally. Mozambique's infrastructure performance is strongly associated with the low productivity levels in the country. The results of the econometric analysis developed in the productivity chapter (Chapter 3) show that even after controlling for endogeneity, the quality of infrastructure provision significantly determines a firm's productivity. The link between infrastructure performance and productivity growth has been previously documented in Clément et al. (2008). In a benchmark analysis against economies that achieved sustainable growth, Mozambique stands out by its accentuated gap in infrastructure provision. The authors conclude that improving infrastructure is a necessary step for Mozambique in order to reach a sustainable growth path. Since the end of civil war in 1992, Mozambique's infrastructure has been reconstructed and now finally surpasses prewar levels of service. Mozambique's infrastructure was devastated during the prolonged war. Most bridges and roads were mined and railroad and power transmission were heavily damaged. Land communications between the south and the center of the country were completely destroyed. After more than 15 years of reconstruction Mozambique's infrastructure approaches the level of its regional competitors. Despite the recent improvements, infrastructure still represents a significant fraction of costs. Infrastructure costs have an important impact on private returns and investment decisions. Combining direct and indirect costs (such as sales losses due to power outages, delays on customs, and days of inventory) the median firm in Mozambique spends on average 22 percent of its sales in infrastructure (Figure 6.1). When compared to most of its regional peers, total costs are low but the direct component of costs is the second highest ranked level of expenditure. Mozambique's infrastructure reconstruction was largely financed by Official Development Assistance (ODA). Mozambique is, by far, the country in our regional sample that received the largest amount of ODA transfers. Combining multilateral and bilateral assistance, more than US$3.5 billion has been committed to the country since 1992 (Figure 6.2). This is more than twice the amount committed to the next largest recipient, Zambia. The transportation sector received almost 52 percent of the total assistance, used in the reconstruction of the road and railroad systems (Figure 6.3). 65 Figure 0.1. Infrastructure Costs as Share of Total Sales 50 45 40 35 30 Percent 25 20 15 10 5 0 Mozambique Angola Malaw i South Africa Zambia Vietnam Country Direct costs - % sales Indirect costs - % sales Source: Enterprise Survey 2008 Figure 0.2. Bilateral and Multilateral ODA 700 600 . US$ million (const. 2006) 500 400 300 200 100 0 1992/93 1994/95 1996/97 1998/99 2000/01 2002/03 2004/05 Year Angola Malaw i Mozambique South Africa Zambia Source: OECD statistics 66 Figure 0.3. ODA Transfers to Mozambique by Sector 500 450 . 400 US$ million (const. 2006) 350 300 250 200 150 100 50 0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 Year Water & sanitation Energy Transport Telecom Source: OECD statistics Mozambique also experienced growing participation of private investment in infrastructure in the last decade. Private participation in infrastructure in the last 10 years adds up to US$2.1 billion and it was primarily concentrated in energy (55 percent) and transportation projects (36 percent). Private investment is small when compared to middle-income countries such as South Africa (US$17.1 billion), but it is much larger than the investment in the other low-income countries in our sample (Figure 6.4). 67 Figure 0.4. Private Investment in Infrastructure, 1997­2006 a. Investment per Year b. Investment per Year per 100 Inhabitants 1710.1 40 35.7 1,800 1,600 35 Avg US $ Commited per 1,400 il. Avg US $ m Commited 30 year/100 hab 1,200 per year 25 1,000 20 800 15 10.2 600 398.8 10 6.7 400 3.4 4.6 217.5 80.3 5 0.7 200 58.3 9.6 0 - Ä Ä Ã Ã Ä Ä h h s s h h Count r y Co unt r y Source: PPI Database (World Bank) Source: PPI Database (World Bank) In this chapter, we discuss the result of the infrastructure reconstruction in electricity, telecommunications, and transport and logistics from the perspective of Mozambican firms. The infrastructure reconstruction process consisted of a series of sector reforms that legally restructured the provision of services and liberalized markets to private investment and competition. This chapter analyzes the results of the ICA survey, highlighting the outcomes of the reforms in the perception of Mozambican enterprises. The first section presents the results for the electricity sector. The following section analyzes telecommunication provision, while the third section provides an overview of transportation and logistics in Mozambique. Finally, conclusions and policy recommendations are discussed in the last section of this chapter. Electricity Since the end of the civil war in 1992, the Mozambican government started a period of reconstruction and comprehensive reforms in the electricity sector. The introduction of a new electricity law in 1997 and the consolidation of a National Energy Strategy in 2000 provided the legal framework for the restructuring and opening of the energy sector to private participation. With the support of the World Bank, Mozambique's government designed an Energy Reform and Access Program to implement the National Energy Strategy. The main objectives of the program were to accelerate the expansion and to improve the quality of modern energy services with particular emphasis on the provision to rural and peri-urban localities. The restructuring of Electricity of Mozambique (EdM) is one of the main steps in the reform.57 EdM is a vertically integrated, government-owned electricity utility. It is a quasi-monopolist in transmission and distribution activities. It has an installed capacity of 57 Details about the steps of the reform are described in the Project Appraisal Document (The World Bank, 2003) and in the Proposed Restructured Document (The World Bank, 2007) for the Energy Reform and Access Program. 68 140 megawatt (MW) hydropower (of which 86 MW is available) and 109 MW in thermal power stations (of which 82 MW is available). EdM buys most of its power supply (300 MW) from Hidroelectrica Cahora Bassa (HCB); the remaining production of HCB is exported to South Africa. EdM's reform started in 1995 when it gained the legal status of a state-owned enterprise. In order to improve EdM's weak financial and operational performance, the Energy Reform and Access Program promoted a deep restructuring process that included participation of the private sector, changes in management, training of personnel, and the setting of a performance target indicator strategy. The program is currently implementing the division of the utility into three separate business units-- generation, transmission, and distribution--with the goal of increasing transparency and efficiency and helping EdM meet its performance targets. Reforms also include activities that support the creation of an Independent Regulatory Board. One of the main challenges in Mozambique's electricity reform is the scarcity of skilled professionals in the sector, in particular those with experience in strategy planning, tariff determination, monitoring and other regulatory activities. Part of Energy Reform and Access Program is dedicated to capacity building, training, and technical assistance in the design of sector strategy, including the transformation of the National Council of Electricity (CNELEC) into a strong independent regulatory body. CNELEC has recently begun its work, but it currently holds only advisory responsibilities. Most of the strategic decisions for the sector are set by the Ministry of Energy. Reforms in the electricity sector have started to generate some positive results. It is too early to evaluate the full impact of the Electricity Reform, but it is already possible to observe some changes. Figures 6.5 and 6.6 show a steady expansion of services since 2002, when coverage increased on average 9 percent per year while consumption per capita increased an average of 3.9 percent per year. Coverage in the southern region, in particular Maputo, is still significant higher than in the other regions. So far, the reforms have been unable to boost provision in the peri-urban areas and close the gap between regions. EdM's performance has also started to show improvements: the number of connections per employees more than doubled in six years. However, energy losses are still high and account for more than 20 percent of the volume produced, they are significantly above the 10 percent target set by the Electricity Reform (World Bank, 2005). 69 Figure 0.5. Evolution in the Coverage of Electricity Services W z N C S N Source: Statistical Report EdM 2006 Figure 0.6. Electricity of Mozambique--Production and Productivity z C W Source: Statistical Report EdM 2006 Electricity performance also improved in the perception of firms according to the results of the 2008 Enterprise Survey. Electricity was perceived as a major constraint for business performance in 2003 by more than 60 percent of firms (Figure 6.7). It was reported as the second main obstacle to business, after cost and access to financial services. Currently, only 24.7 percent of the firms surveyed consider electricity among the three most important obstacles to business, ranking it as the sixth most important obstacle. When compared to other low-income countries in the region Mozambique's electricity sector relatively well perceived, but despite the improvements it does not reach 70 the performance of middle-income countries such as South Africa. Perceptions about electricity are quite unequal across firms; large firms outside Maputo seem particularly constrained by the quality and reliability of electricity services. Figure 0.7. Electricity as a Business Constraint W W N M M C M A M S A Z V C N M M W W Micro (les s than 5 S mall (5-19 Medium (20-99 Large ( 100 employees ) emplyees ) employees ) employees and N M M more) F C M O N M M S Source: Enterprise Survey 2008 The changes in firms' perceptions reflect the reduction of indirect costs associated with the use electricity services. Almost all indicators in our analysis improved significantly in the last five years. The frequency of electricity outages decreased significantly, while the median loss from power outages experienced a modest fall (Table 6.1). These changes justify the lower ownership of generators. The apparently puzzling increase in the percentage of self-generated power reflects a change in the composition of firms. As external electricity becomes more reliable, firms with a low dependency on self-generating power choose to become fully dependent on external sources. Table 0.1. Power Outage Indicators, 2003 and 2008 Percent of Firms Median percent of Median Number Median Total with Own Self-generated Power Outages/month Production Losses Generator Power due to Power Outages (%) 2003 23.0 8 5 2.0 2008 13.2 10 3 1.9 Source: Enterprise Survey 2008 71 Despite progress, costs associated with inadequate power supply are still relatively high. Inadequate provision, in particular unexpected interruptions and oscillations, affects production and may cause serious damage to machinery and equipment. The percentage of Mozambican manufacturing firms that experienced power outages is low compared with the sample average. Outages reached 52 percent of firms (63 percent of large firms) in Mozambique against an average of 74 percent in the comparison group of countries. However, the incidence of power outages is relatively high: the median manufacturing firm in Mozambique experiences an average of 3 outages per month, an incidence lower only than Angola (6 outages per month) and Malawi (4.2 outages per month). Moreover, the costs associated with inadequate provision in Mozambique are still significantly higher than in middle-income comparison countries. Losses due to inadequate power are also relatively high in Mozambique. The median manufacturing firm losses are the second highest in the sample, reaching 1.9 percent of the firms' total sales. Losses due to inadequate provision are also highly unequal across firms (Figure 6.8). Firms in Maputo have a clear advantage with respect to electricity provision. Average losses in other regions are twice as high as in the capital, reaching 8 percent of total sales. Service firms are more susceptible to power outages, but the poor quality of electricity services is much less costly for this group of firms. Exporting firms are the least susceptible to inadequate power supply. This group includes companies such as MOZAL (producer and exporter of aluminum), a mega-project that imports most of its electricity consumption directly from South Africa. Firms frequently respond to the uncertainty in electricity provision by investing in self-generation systems. Ownership of generators is particularly high among medium and large firms in Mozambique (compared to other countries in the region) (Figure 6.9). The investment in generators has a clear precautionary character and it is more frequent in areas and sectors where losses due to inadequate power are more severe. Ownership of self-generators is 80 percent higher outside Maputo than in the capital and it is concentrated in sectors like chemicals and machinery. Despite helping firms to balance their energy losses, generating energy is not a core business for most firms and it represents an inefficient allocation of resources. For this reason, firms use only a small fraction of consumption from generators: the median firm draws 10 percent of its total power consumption from its own self-generators. The overall cost of electricity is relatively low in Mozambique in comparison to other countries. Connecting to electricity in Mozambique takes on average 32 days, 43 percent less than in the other low-income countries in the region, but still more than the average for middle-income countries (17 days). Once connected, firms face very low costs of electricity services. The median firm in Mozambique spends 1.5 percent of its total sales on electricity, less than in any other country in the sample. Electricity provision in Mozambique has come a long way since the beginning of reforms. The results of the enterprise survey reflect this progress, but also expose some weaknesses. Mozambique's productive capacity seems adequate for current consumption and the cost of services is definitely low for regional standards. Our analysis points out the relative weakness of electricity provision outside Maputo, and in particular in rural and peri-urban areas. Although these areas were originally the main focus of the Energy 72 Reform and Access Program, the inability to sustain private investment has delayed the expansion of electricity services in these areas. Figure 0.8. Power Outages--Frequency and Losses à à i M A M S A Z V M A S A Z V C C à à i M O M O R R O M à à i M S M L M S M L F F O M Source: Enterprise Survey 2008 73 Figure 0.9. Generator Ownership 90 80 70 60 50 40 30 20 10 0 M A M S A Z V C o unt r y S M L Source: Enterprise Survey 2008 Figure 0.10. (a) Average Time to Connect to Electricity, and (b) Electricity Services as Percentage of Total Expenditure T C 120 4.5 4.0 100 3.5 80 3.0 à 2.5 60 à 2.0 i 40 W 1.5 20 1.0 0.5 0 0.0 M A M S A Z V M A M S A Z V C C N N Source: Enterprise Survey 2008 Telecommunications Mozambique was the first country in the region to reform its telecommunications landscape, starting immediately after the end of the civil war. The process of reforms started in 1992 with the transformation of TdM, the single provider of services at the time, into an independent state-owned enterprise. In the same year Instituto Nacional das Comunicações de Moçambique (INCM) was established as an autonomous regulator. At the end of 2001, the government started the process of liberalization of the sector by opening the mobile market to competition. The year 2003 was marked by the separation of TdM's mobile subsidiary mCel, a necessary condition for the entrance of the private mobile competitor Vodacom. The consolidation of the reforms arrived with the partial 74 privatization of TdM in 2004. Private investment permitted expansion and improvement in the quality of services, while competition assured a decrease in tariffs.58 Telecommunication reform has proven to be a great success in expanding access, in particular access to mobile services. As of 2000, there were 85,000 land lines and 51,000 mobile phones. Telephone charges were high and the quality of mobile phones was poor. By 2007, there were 3,060,000 mobile phone subscribers, many of them outside of Maputo and the quality of service had improved substantially.59 In fixed lines, however, competition provoked a strategic scaling down of services by TdM to 63,000 lines. Despite big improvements, the provision of communication services is still quite unequal across the country. In fact, 65 percent of fixed lines are concentrated in Maputo and only 1 percent of total lines serve rural areas (where around 70 percent of the population lives). Figure 6.11 compares the expansion of communication services with services in other countries in the region. Mozambique performs poorly in the relative number of mainlines--only 0.3 fixed lines per 100 individuals. The country's performance is similar to its regional peers in access to mobile phones (11.2 subscribers/100 individuals) and Internet usage (0.9 users/100 individuals). However, Mozambique's statistics for all three services (mobile, fixed line, and Internet) are still significantly below those for middle-income countries. Figure 0.11. Telecommunications--(a) Fixed lines per 100 Inhabitants; (b) Mobile Subscribers per 100 Inhabitants; (c) Internet Users per 100 Inhabitants 80 25 70 M L 20 60 50 15 40 E 10 E 30 20 5 10 0 0 M A M S A V Z S S M A M S A V Z S S C C 20 18 I 16 14 12 10 E 8 6 4 2 0 M A M S A V Z S S C Source: Enterprise Survey 2008 58 For example, the price of a three minutes call from Mozambique to the United States dropped from US$6.00 in 2001 to less than US$2.00 in 2004. This price reduction (67 percent) was larger than the average for the Sub-Saharan region (50 percent). Source: ICT at Glance Tables, The World Bank. 59 According to the Country Economic Memorandum for Mozambique, World Bank (2005a). 75 Telecommunication is perceived by firms as the least problematic among infrastructure services (Figure 6.12). In the previous ICA (2003), firms already perceived telecommunication provision as a minor obstacle to business performance. The low percentage of firms reporting telecommunication as a major problem decreased further in 2008. This improvement in firms' perception is consistent with the transformations that have occurred in the sector. Firms' perceptions are very similar in Maputo and in the other regions, but they vary significantly with firm size. Large firms are considerably more likely to perceive telecommunication as a major obstacle. Mozambique's telecommunication sector performs well in comparison to other countries. The percentage of firms reporting telecommunication services in Mozambique as a major problem is equivalent to the levels in the middle-income countries in our sample. Figure 0.12. Telecommunications as Business Constraint W W N M M M S N M M M S C C M O W W M M S A V Z M S M L C F N M M M S N M M M S Source: Enterprise Survey 2008 Access to telephone services is relatively easy for Mozambican firms (Figure 6.14). The average waiting time for a firm to connect a phone in Mozambique, 25 days, is far below the regional average (44 days). Mozambique's performance still lags in comparison to the middle income average (11 days), but larger firms are able to connect to services almost immediately. The overall cost of service is relatively low for connected firms. The median Mozambican firm spends 0.06 percent of total sales on communication services. 76 Figure 0.13. E-mail Usage and Web Site Ownership W W M A M S A Z V M A M S A Z V C C W W A S M L A S M L F F Source: Enterprise Survey 2008 Figure 0.14. Average Days of Waiting to Connect a Telephone S M L M A M S A Z V F C Source: Enterprise Survey 2008 Mozambique's pioneering telecommunication reforms delivered great results in access, quality, and cost of services, but sector performance remains below the regional average. Telecommunications was reported to be the least significant infrastructure problem. Nevertheless, the degree of uptake of Internet services by businesses and the availability of any kind of telecommunications service are still considerably low. Part of Mozambique's relatively weak performance is a reflection of the high levels of poverty and low levels of literacy (more than 50 percent of population is illiterate). Yet much of the responsibility lies with the small and geographically concentrated supply. Transport and Logistics Mozambique's ability to efficiently connect firms, suppliers, and consumers is a key factor for its future economic development. Mozambique's internal market is too small 77 in size and in purchasing power to sustain increases in production levels. International trade is, therefore, necessary to promote growth at high enough rates to bring down absolute poverty and to create the number and types of jobs needed. The implementation of an efficient logistics infrastructure is a fundamental step to guarantee a Mozambican firm's international competitiveness. Strategically located, Mozambique has the potential to be a major outlet for Southern Africa. The Maputo port is the closest route for the largest mining and manufacturing region in South Africa. Beira is the closest port to Zambia, Zimbabwe, and Lubumbashi (in the Democratic Republic of Congo), while Nacala port was built to give Malawi a shorter route to the ocean. The high costs and low transport reliability derived from a variety of logistical problems60 prevent Mozambique from fully exploring its potential for exporting services. Logistics include two main categories: hardware of logistics and software of logistics. Hardware of logistics refers to the quantity and quality of physical infrastructure needed to move goods effectively: roads, ports, airports and railroads. Software of logistics refers to the services and processes needed to move and trade goods effectively. It includes customs, licenses and fees, trade regulations, inspections and certificates, and packaging, among others. We analyze the perception of Mozambican enterprises and identify their main concerns with respect to each of these categories. Hardware of Logistics--Transportation Infrastructure Mozambique's transportation system is undergoing a major restructuring process. The system was severely damaged during the civil war. Government and massive foreign efforts allowed the country to rebuild and begin to expand the system. The transport sector in Mozambique is presently comprised of 26,235 km of classified roads, of which only 5,005 km (19.8 percent) are paved; and about 130,400 vehicles; 3,000 km of railway lines, of which about half are in operation; five international airports (Maputo, Beira, Nampula, Pemba, and Vilankulos) and 14 secondary airports; three major ports (Maputo, Beira, and Nacala), two secondary ports (Pemba and Quelimane) and 13 tertiary ports scattered along a 2,700 km of coastline (World Bank, 2008c). The sector experienced a series of structural reforms. Some positive results can already be observed, but the sector needs more improvements. In April 1996, the government of Mozambique approved a transportation policy reform that aimed at boosting the activities in the different components of the sector (maritime, land, and air transportation). The actions included legal and regulatory changes that promoted private participation, including the concession of railways, ports, and airports. They supported the creation of independent planning, finance, and regulatory bodies, strengthening the institutional capacity in the sector. By improving management efficiency and safety, the government expected to reduce transportation costs, thereby gaining competitiveness. 60 Examples of such problems include limited communication infrastructure between production sites and ports, in particular roads and railroads linking Nacala to Malawi; reduced shipping capacity, especially in Beira port; poor integration of trucking services; corruption; and high scanning fees. 78 The reforms succeeded in reconnecting the main strategic routes, but they still lag behind in reducing costs and expanding access to the large majority of population living in rural areas. Mozambican firms still perceive transportation as an obstacle to investment (Figure 6.15). Despite the reforms in the sector, the absolute perception of firms about the sector changed very little with respect to the previous ICA survey in 2003. Currently, almost one quarter of the enterprises consider transportation a major obstacle to investment against 27 percent in 2003. In relative terms, transportation increased in importance as an obstacle to business, moving from the 14th position in 2003 to the 7th in 2008. Larger firms and firms outside Maputo are the most constrained by transportation problems. Mozambique's performance is quite weak when compared with its regional peers. Malawi is the only country in which a larger percentage of firms consider transportation as an important constraint. Figure 0.15. Transportation as Business Constraint 70 70 60 60 50 50 40 40 30 W W 30 20 20 10 10 0 0 No/Minor Moderate Major/S evere All Maputo Other regions C R N M M M S 80 80 70 70 60 60 50 50 40 40 W 30 W 30 20 20 10 10 0 0 M M S A Z V S mall Medium L arge C F N M M M S N M M M S Source: Enterprise Survey 2008 The real costs of transportation justify the perception of firms. Transportation remains a significant factor in the high cost of doing business in Mozambique (Figure 6.16). Firms experience very high indirect costs from inadequate transportation. Sixteen percent of firms reported production losses from theft or breakage during transportation. The percentage is even higher for large firms--at least 33 percent experienced losses. Almost one half of firms report that the quality of roads is not suitable for their activities and this share reaches 75 percent among large firms. Direct expenditures on transportation services are relatively low for most firms. Firms outside Maputo, however, experience larger direct costs: 7.3 percent of their total sales are dedicated to transportation related activities. 79 Figure 0.16. Cost Associated with Transportation: Product Losses, Inadequate Roads, and Direct Costs 35 8 30 7 25 6 20 5 4 W 15 W3 10 2 5 1 0 All S mall Medium Large 0 All Maputo Other regions F R D 70 80 60 70 60 50 50 40 40 W 30 W 30 20 20 10 10 0 0 All Maputo Other regions E xporter All S mall Medium L arge R E F Source: Enterprise Survey 2008 Mozambican firms face a constant tradeoff between direct and indirect costs of transportation. The Mozambican ports and railways services were, in the past, the preferred transport modes for the import and export of goods in the region. In general, the cost of railway transport is one-third of the cost of formal road transport over similar distances. However, due to problems with security of cargo, loading, and delivery delays, these sectors are losing competitiveness. Despite high tariffs, road transportation and foreign ports are becoming the preferred option among enterprises. Software of Logistics--Customs and Trade Regulation Sixteen years after the civil war ended Mozambique has successfully completed an ambitious program of "first generation" reforms, including trade and customs reforms. Mozambique's trade policy has improved and is among the most progressive in the region. Mozambique reduced drastically the number and level of tariffs. Import- weighted tariff averages have declined from 17.4 percent from 1995­99, to just 9 percent in 2005 (World Bank 2009, forthcoming). The government also designed a new customs code and invested heavily in improving capacity, management, and transparency in customs operations.61 Even though changes are still underway, the moderate improvement facilitated the rapid expansion of exports (an average of 10 percent per year since 1995). As a result, Mozambique's exports-to-GDP ratio has doubled. Mozambique 61 De Wulf and Sokol (2004) provide a good description of Mozambique's Custom Reform. 80 has, thus, become one of the few countries in Africa to experience an increase in its share in world exports. International comparisons expose Mozambique's still fragile logistics. The trade across borders index presented in Figure 6.17 combines procedures, regulations, and costs of trade reported in the Doing Business database. Mozambique's performance is slightly superior to its closest neighbors (South Africa, Malawi, Angola, and Zambia), but well below Vietnam. Mozambique is by far the worse country with respect to logistic performance. Figure 0.17. Trade across Borders Index and Logistics Performance Index T A B I L P I / / M A M S A Z V M A M S A Z V S L I A C C R Source: Doing Business 2010; Logistics Performance Index 2007 Surprisingly, Mozambique's firms do not perceive customs and trade regulation as a major business constraint (Figure 6.18). Currently, only 11 percent of Mozambican firms consider customs and trade regulation as a major problem. This indicator is significantly lower than the 46 percent reported in 2003, far below the average in the comparison group. One possible explanation is the fact that only a small fraction of firms in our sample exports directly. If we restrict the analysis to firms that export, the fraction of firms that considers trade and customs as a major or very severe obstacle increases to 37 percent. The average perception varies drastically when firm size is taken into account. Large firms are usually more dependent on external markets. Dissatisfaction among this group is much higher: 32 percent of these enterprises report customs and regulations as a major bottleneck. High transportation costs and inefficient customs result in a low share of exports. Despite the reforms, Mozambican customs are still relatively slow and inefficient. An average firm might take up to 20 days to export a product. This is the second longest clearance period among the countries analyzed--only Angolan customs take longer. Firms also face delays importing machines and materials. An imported input takes up to 18 days to clear customs in Maputo and up to 30 days in other regions, by far the longest in the region. Easy access to input and output markets is a fundamental determinant of international competitiveness. High transportation costs, losses, and delays explain why Mozambique still has one of the lowest shares of sales export in the comparison group (Figure 6.19). Mozambican firms export only two percent of total sales, one-fifth of the average share of exports in the region. 81 Figure 0.18. Customs and Trade Regulation as Business Constraint 80 90 70 80 70 60 60 50 50 40 40 W W 30 30 20 20 10 10 0 0 N M M M S A M O C R N M M M S 90 90 80 80 70 70 60 60 50 50 40 40 W W 30 30 20 20 10 10 0 0 M M S A Z V S M L C F N M M M S N M M M S Source: Enterprise Survey 2008 Transport and trade reforms are still insufficient to bring Mozambique to its full potential. Reforms generated an expansion of exports and economic growth, but results were concentrated in a few projects, and the majority of firms face low quality and high cost of transportation. Mozambique is relatively well positioned in regards to access to markets, and has a potential comparative advantage in a variety of products. Nevertheless, logistic costs remain prohibitively high for most firms, rendering products that could be competitive less so. 82 Figure 0.19. Share of Total Sales 5 30 5 4 25 4 20 3 3 15 W2 W 2 10 1 1 5 0 A M O M S M L 0 M A M S A Z V R F C 35 45 40 30 35 25 30 20 25 15 20 15 10 10 5 5 0 0 M A M S A Z V A M O S M L C R F A L A L A L A L Source: Enterprise Survey 2008 Conclusions Mozambique still lags behind middle-income economies in terms of infrastructure; nevertheless, significant steps have been made in the right direction. Institutional and structural reforms helped the post-war reconstruction and expansion of infrastructure. They successfully increased private participation, investments, and efficiency, increasing access and reliability of services. Reforms generated positive outcomes, in particular those in electricity and telecommunication, yet much still needs to be done in order to guarantee current and future adequate provision of services. Electricity was definitely the sector that improved the most in the perception of firms, but it still faces many challenges. Access has been improving and capacity seems adequate for current demand, but Mozambique faces the challenge of expanding services at the same pace of the future expansion in demand (estimated around seven percent a year for the next years). Mozambique's government is currently addressing this problem. It plans to construct one thermal power plant, to expand the Cahora Bassa plant, and to expand the transmission system to connect the new sources of power. Increasing transparency and clearly assigning responsibilities will help mitigate the risks for private participators and speed up the implementation of these and other future projects. Although EdM experienced improvements in the expansion of services, labor productivity, and financial management, the company still struggles with very large losses in the distribution of electricity and unwarned interruptions. The creation of 83 an independent regulatory agency with monitoring and policy designing responsibilities is a successful strategy for many developing countries. Nevertheless, the premature implementation and transfer of responsibilities to the independent board might produce more harm than good. Mozambique should continue investing in capacity building and the gradual transition to a fully functional independent regulator. Meanwhile, the Ministry of Energy should periodically review the performance targets for EdM (such as maximum of losses acceptable in the system) and intensify monitoring. The design of additional incentive mechanisms such as fines for unwarned power outages and excessive number of complaints is also a helpful tool to enhance service quality. Electricity provision is still restricted in access and quality outside Maputo, especially in peri-urban and rural. Rural provision was one of the main focuses of the Energy Reform and Access Program, but the inability to sustain private investment has delayed the expansion of electricity service. Mozambique's government is currently investing heavily in the extension of the electricity grid to peri-urban areas. Nevertheless, the problem of attracting private investment to construct independent grids in low-density rural areas remains unresolved. Based on the experience of the pilot project62 the government should reevaluate some aspects of the concession contracts and bidding process. Instead of promoting competition over tariffs, the government should fix tariffs and award contracts to firms proposing the most efficient subsidy schemes. This model should reduce the risks of ex-post tariff adjustments that are poorly perceived by the local communities. In addition, increasing transparency and commitment to contracts will reduce the risks for private investors and attract the necessary resources to the sector. Telecommunication is the least problematic sector in the perception of firms, but Mozambique performance is still below regional standards. Despite the increase in private participation and competition, until recently investments in telecommunications were below the levels necessary to match performance standards in the region. Access to phone services, in particular to fixed lines, is very low and concentrated. As broadband networks are not widespread, inadequate fixed-line infrastructure hampers internet usage. Prepaid services are an efficient way of improving access with low billing and administrative costs. TdM, mCel and Vodacom are developing plans, massively expanding prepaid services (fixed and mobile). In addition, in order to intensify competition and attract resources into the sector, INCM is planning to expand wireless broadband networks and to release a third license for mobile operation.63 If successful, these measures will help Mozambique meet its demand and reach international standards. The results from the enterprises survey reveal transportation and logistics as the main infrastructure obstacles to business. Mozambique's railroads and ports are losing competitiveness and actual transportation volume due to breakage and theft losses, delays, and unnecessary 62 OB Approaches n.003 (World Bank, 2005b) provides a good description of the implementation of the pilot project. 63 Paul Budde Communication Pty Ltd. (2007). 84 charges. Mozambique's port and railway systems are managed by Portos e Caminhos de Ferro de Moçambique, E.P. (CFM). The government recently started a series of structural reforms to improve operations in the sector.64 The actions include the expansion of private participation in the management of ports and railway. The government started a process of restructuring of the CFM with the separation of strategy, regulatory, monitoring, and day-by-day operational function. Reforms also include training, specialization, and rationalization of the staff team.65 The full implementation of the described changes should enhance CFM performance. The agency should then focus on executing adequate maintenance and network rehabilitation, intensifying monitoring activities, and improving coordination between rail and port networks to reduce turnaround times. Finally, investments in network expansion should focus on improving the precarious link between the north and the south of the country. A large fraction of firms consider road transportation system inadequate for their production activities. Mozambique put great effort in reconstructing its road system, reconnecting the main corridors, and creating separated regulatory (Administração Nacional de Estradas [ANE]) and financial (Road Fund, [RF]) boards. The attention should now be directed to the quality and cost of road transportation. National road taxes are higher than those applied in the neighboring countries, which constitutes a barrier to potential incoming transporters. Road conditions also have a big impact on transportation costs: while east-west transport linkages for trucking are fairly well developed, roads are of variable quality and bridges are missing from major river crossings in the north-south meridian. Currently, it is more expensive to transport cargo within the country than shipping it to a different continent.66 Finally, poor integration of trucking services with the rest of the region decreases the attractiveness of Mozambican ports, leading to a suboptimal scale of operations.67 One of the main problems faced by the sector is the lack of technically qualified staff. This problem is reflected in the ability to manage financially and operationally an efficient maintenance program. The small supply of qualified professionals also affects the monitoring and accountability of operations. Mozambique should intensify capacity building and personnel restructuring in both ANE and RF. Despite the reform, customs' weak performance still represents an important obstacle in the access to international markets. According to the World Bank's Doing Business Report, Mozambique ranks in the lower third worldwide for imported and exported cargo procedures. Although there exist a variety of procedures associated with exit/entry for goods crossing a country's border, customs requirements are commonly perceived as one of the most relevant in determining border crossing time. In order to increase the speed and effectiveness of customs, the government should review the legal framework governing the timely submission of documents, payment of duty, and requirements for the timely release of goods. A strategic reduction in the number of 64 In particular port efficiency and shipping lines frequency. 65 World Bank (2007a) 66 Ocean shipping from northern Mozambique to southern Mozambique costs just as much as from Asia to South Africa. (Minor 2007, p. 16, cited in World Bank 2009 (forthcoming), v2, p. 127) 67 Maputo port operates at less than 30 percent of its capacity while Beira port has operated under capacity at less than 40 percent for the last 5 years (World Bank, 2009, forthcoming). 85 procedures and inspections, and the intensification of the ongoing process of automation and coordination of border activities,68 can improve the exchange of information with neighboring countries and the monitoring of activities in a shorter amount of time. Despite the progress made on customs reform,69 Mozambique's borders still experience capacity shortages in operational and supervising functions. A continued capacity- building program will not only enhance performance, but help to undermine corruption and irregularities in customs services. Infrastructure performance, while improving, remains a concern for a large group of firms in Mozambique. Many different issues affect production costs (either directly, indirectly through production losses and spoilage of merchandise, or implicitly through losses of time and opportunities). Isolated, each of the elements might carry a small weight, but the combination of all of them results in a substantial negative impact on productivity and access to markets. Therefore, infrastructure performance still represents an important obstacle to competitiveness and growth, and it should continue to be a priority. Improving the overall competitiveness of Mozambican enterprises requires combined actions. On the hardware logistics side, investments should be focused on improving deficient infrastructure. The objective should be to implement a multimodal framework, integrating ports, maritime and coastal shipping, railways, road transport, terminals, warehouses, and distribution centers. Another key infrastructure undertaking is linking the main transport corridors to rural areas with feeder roads. On the software side, the priorities should be integrating border institutions, building capacity, simplifying and streamlining customs procedures, improving technology utilization, and curbing corruption. 68 Many countries in the regions adopted a "One Stop Border Post" as way of harmonizing customs procedures and speeding up the clearance process. Such reforms have a very positive impact on trade flows and could be used to improve the capacity usage in the Maputo port. 69 De Wulf and Sokol (2004). 86 Labor Markets and Human Capital This chapter uses firm-level data provided by the surveyed firms to describe both employment generation in the Mozambican labor market and human capital levels and technology adoption among firms in Mozambique.70 The chapter starts with a broad characterization of Mozambican firms in relation to a variety of labor issues and their perceptions of labor skills and regulations as obstacles to business growth. Data collected from the Enterprise Survey allows a novel perspective on the growth dynamics of nonagricultural formal employment over the recent years in Mozambique. There is scattered evidence suggesting that the dynamics have not been satisfactory and the survey, confirming such evidence, allows a deeper examination of the issue. The chapter concludes with a look at human capital accumulation and technology adoption and a discussion of key policy implications. Labor Market Characteristics Mozambique's labor force comprises approximately 10 million workers. The overwhelming majority of them are concentrated in the informal agricultural sector (estimates range between 80 percent and 90 percent of total employment). Services, manufacturing, and public administration are all important employers in urban areas. Unemployment figures seemed to have increased since 1997, and according to one labor survey, levels of unemployment were as high as 18 percent in 2004.71 Significant employment generation in the nonagricultural formal sector, therefore, is crucial for improving inclusiveness and diversification of economic growth. Given data limitations, the labor and employment growth analysis presented in this chapter focuses on formal nonagricultural firms. The size and economic sector of sampled firms are presented in Table 7.1. Table 0.1. Size and Sector Distribution of Sampled Firms Industry Size Construction & Transport Manufacturing Services Total Micro 0 17 103 120 <5 (0) (2.84) (17.2) (20.03) Small 1 219 87 307 5­19 (0.17) (36.56) (14.52) (51.25) Medium 1 103 38 142 20­99 (0.17) (17.2) (6.34) (23.71) Large 5 19 6 30 > 100 (0.83) (3.17) (1) (5.01) Total 7 358 234 599 (1.17) (59.77) (39.07) (100) Source: Enterprise Survey 2008 Note: Numbers in parenthesis denote percentages over the total number of firms. 70 This chapter is partly based on the background paper "Employment Growth and the Lagging Performance of SMEs," prepared by Fabiano Bastos. 71 Employment and unemployment figures and estimates are presented in Chapter 1 of the 2009 Mozambique CEM (World Bank, 2009, forthcoming). 87 Cross Country Comparisons Workforce education does not top the list of concerns for Mozambican firms.72 In fact it ranks 9th as a constraint often cited by businesses. This perception is shared among some of Mozambique's comparators: it ranks outside the top seven most often cited constraints in Angola, Zambia, Indonesia, and Vietnam. On the other hand, firms in South Africa, Malawi, and Thailand considered workforce education to be a more important obstacle, ranking as the most often cited obstacle in South Africa and Thailand, and the 4th most often cited constraint in Malawi. The perception that workforce education is not a constraint to business in Mozambique is not reinforced by the quantitative data collected by the survey. As was shown in Chapter 0, the investment climate index that captured technology and human capital was significantly correlated to higher productivity. This index included the percentage of firms that offer formal training to their employees. As we will see later in this chapter, this percentage is low in Mozambique, particularly among SMEs. On the other hand, the data also shows that among the firms that do provide some formal training to their workers, most of them offer it to a fairly large percentage of their staff. Therefore, some firms are offering training to a high percentage of workers, but the number of firms doing so is relatively low. Another indicator that points to the need to improve educational levels in the Mozambican labor force is educational attainment. The average educational level of a typical production worker in Mozambique is considerably lower than in countries like South Africa and Zambia and only marginally higher than the education level of a production worker in Angola (Figure 7.1). This educational attainment changes little across firm size in all four countries for which comparable data is available, and for Mozambique, it reflects a production worker that on average has completed elementary education and has some lower but no higher secondary education. 72 Due to data availability issues and the fact that certain questions of concern were not comparable across surveys, the results presented in this chapter only include those variables and countries for which we could ensure comparability. 88 Figure 0.1. Average Educational Attainment of Typical Production Worker (by firm size) 3.5 3.0 Overall Small Medium Large Education Category 2.5 2.0 1.5 1.0 0.5 0.0 Angola Mozambique South Africa Zambia Country Source: Author's calculations based on Enterprise Surveys 2003-08 Note: Education categories: 1 = 0­6 years of education; 2 = 6­9 years of education; 3 = 10­12 years of education; 4 = more than 13 years of education. Perhaps reflecting lower levels of educational attainment, median monthly compensation in Mozambique is lower than most comparators. Angola's compensation numbers are almost three times that of Mozambique, even though average educational attainment figures are comparable for the two countries. Differences in compensation again change little across firm size in Mozambique (Figure 7.2). Figure 0.2. Median Monthly Compensation of Production Workers (by firm size, in US$) 900 800 Small Medium Large Overall 700 600 500 US$ 400 300 200 100 - Angola Malaw i Mozambique South Zambia Africa Country Source: Enterprise Surveys 2003-08 89 While wages may be lower in Mozambique, the competitiveness of its workers seems to be hindered by lower productivity levels. Unit labor costs, measured as the ratio of labor costs per worker to value added per worker, provide evidence supporting this finding. Unit labor costs in Mozambique, while lower than Malawi and South Africa, and marginally lower than Angola and Vietnam, are higher than in Zambia and Thailand (Figure 7.3). Accordingly, unit labor costs for small Mozambican firms are marginally higher than medium and large firms, given lower productivity levels among workers in these firms. Figure 0.3. Unit Labor Costs 1.4 1.2 Small Medium Large Overall 1.0 Ratio 0.8 0.6 0.4 0.2 0.0 nd a i a a m aw e ol ic bi qu na la ng fr m al ai bi A et Za M A Th am th Vi u oz So M Country Source: Author's calculations based on Enterprise Surveys 2003-08 Labor Regulation Labor regulation was one of the constraints least cited by firms in Mozambique, with only 5.5 percent of firms reporting as a major or severe obstacle for business. It ranked 15th out of 16 issues surveyed. While in the 2003 survey, this constraint was cited more often, ranking 11th out of 17 obstacles. Most other comparator countries also ranked labor regulation low among their most severe obstacles, with the only exceptions being South Africa, where it was the second most often cited constraint for business, and Indonesia, where the issue was the 4th most often cited obstacle. The result for Mozambique is in sharp contrast with the Doing Business indicators for the country. The Doing Business report collects detailed information on how labor regulations affect hiring, firing, and rigidity of employment. Based upon these regulations, the report calculates objective measures that assess how strict labor regulations are in the country. According to the 2010 Doing Business report, Mozambique ranks 156 (out of 183) in this area (Figure 7.4), one of the worst rankings for the country (overall, Mozambique's ranking is 135). One possible explanation for this discrepancy is the surprisingly high numbers for another constraint, practices from the informal sector, which ranked as the top constraint for business in Mozambique. This constraint may have biased results by 90 capturing all the attention among the other potential obstacles that may explain the exclusion of firms from the formal sector, including labor legislation (Box 7.1). Figure 0.4. Labor Regulation Rankings 200 178 180 156 160 149 140 Ranking 116 120 102 103 100 92 80 60 52 40 20 0 Country Source: Doing Business Report 2010 Box 0.0. Labor Reforms in Mozambique The 2007 labor legislation introduced major changes in labor regulation and represented an important step towards a more efficient use of labor inputs, especially in terms of flexibility to hiring workers. Historically, Mozambique's labor legislation has been highly restrictive and has hampered employment in the formal economy. After a consultative process involving government, unions, and private sector employers, a new labor law was enacted in November 2007. The new law (i) introduced more flexible rules on hiring and firing workers by reducing severance pay and termination notice periods (from 90 to 30 days); (ii) relaxed restrictions for fixed-term employment (the maximum duration of term contracts was extended to up to two years renewable twice); (iii) streamlined recruitment of foreign workers (contingent on quota requirements); and (iv) laid down new conflict resolution procedures envisioning a greater role for arbitrage committees. The new law represents major progress in comparison to the preceding legislation. However implementation and enforcement of different aspects of the new law will need to be closely followed to insure that the goals set forth for the new law are being met. This is especially relevant for the hiring of foreign labor, where initial evidence points to concerns over quotas and qualifications vetting and their impact on businesses' ability to hire foreign expertise. Source: Mozambique World Bank 2009 (forthcoming) Unionization rates in Mozambique are the second highest among the African cohort compared in this report, and only South Africa has higher unionization rates. Mozambique's unionization levels are slightly higher than in Zambia, while considerably higher than in Malawi and Angola. Not surprisingly, unionization rates among 91 Mozambican firms are highest for larger firms, where about 41 percent of workers are unionized, compared to around 10 percent in small firms (Figure 7.5). Figure 0.5. Unionization Rate (by firm size, %) 80 70 Small Medium 60 Large Overall Percent 50 40 30 20 10 0 a i a a aw e ol ic bi qu ng fr m al bi A Za M A am th u oz So M Country Source: Enterprise Survey 2003-08 There is some concern that the current minimum wage is very high in relation to the average wage. According to the Confederation of Economic Associations of Mozambique, "real rises in the minimum wage are encouraging lower levels of employment at the entrée level of employment, more underreporting wages, and strengthens the incentives for firms and jobs to remain in the informal economy." The ICA does not allow us to evaluate the impact of the minimum wage on productivity and employment, but it does provide an overall picture of the percentage of workers that are paid the minimum wage across firm size and regions (Table 7.2). On average, almost 50 percent of the workers surveyed are paid the minimum wage. This number increases to 75 percent for micro firms, while the corresponding figure for SMEs is around 46 percent. In poorer regions like Matola and Nampula, the proportion of workers receiving the minimum wage reaches 60 and 64 percent, respectively. Table 0.2. Percentage of Permanent Workforce Paid the Minimum Legal Wage Beira Maputo Matola Nampula Average Micro 75 53 82 85 75 Small 41 26 58 52 43 Size Medium 44 33 43 68 43 Large 26 35 27 55 34 Average by city 47 36 64 60 49 Source: Enterprise Survey 2008 Note: Weighted characteristics by firm size, sector, exporter, and foreign-owned 92 Decomposition of Recent Employment Growth Data from the 2008 Enterprise Survey indicates that the average growth rate of full- time employment between 2003 and 2006 was 44 percent, with employment growth rates of large firms dominating those of SMEs.73 This figure should be interpreted in the context of a very small base, representing a sample from the nonagricultural formal sector that, as discussed in the introduction, accounts for only a fraction of total employment. Across firm sizes, the most significant result revealed by the data is that (despite starting from a very small base) employment growth rates of large firms dominate those of SMEs over the period 2003­06 (Table 7.3). Indeed, by decomposing the actual levels of employment across size classes over the period, the majority of employment generated comes from large firms, including some medium-size firms (in 2003) that grew to become large firms in 2006. Overall, the data shows a limited role played by SMEs when it comes to formal-sector employment generation. Table 0.3. Full-Time Employment Growth Firm Size Growth Rate % Workers in Workers in Continuing Entering 2003 2006 Firms Firms Micro 63 408 666 190 68 Small 35 2,064 2,803 493 246 Medium 36 4,630 6,323 992 701 Large 60 2,922 4,685 772 991 Total 44 10,024 14,477 2,447 2,006 Source: author's calculations based on Enterprise Survey 2008 Note: Full-time employment in 2003 is used to determine firm size. For entering firms (that is, no employment figure for 2003) we use employment in 2006 to determine firm size. The data suggests a healthy relative participation of new firms in employment generation since 2003, as approximately 45 percent of employment growth was due to entering firms (Figure 7.6). The fact that start-up firms played a significant role when compared to existing firms is an indication that continuing to build a business environment supportive of entrepreneurial behavior has the potential to bring on board new resources that are currently in the sideline. Assessing employment changes across firm size and sectors reveal that the manufacturing sector accounted for most of the employment growth over the period. This represents a change from past years when manufacturing was a source of job destruction in Mozambique. Employment generation in the manufacturing sector has also been predominantly concentrated in larger firms, reflecting a business environment that is especially unfriendly to micro enterprises and SMEs.74 Additionally, expensive 73 Although representing one point in time in the survey, each interviewed firm is asked for the number of full-time permanent employees at the end of 2003 and 2006. Information on seasonal, temporary, and part- time workers is available only for 2006. For this reason, analysis of employment growth is restricted to full- time workers only. 74 As mentioned before, many of the 2003 medium-size firms that are driving employment growth performance in that size category became large in 2006. The appendix reproduces this graph adopting full- 93 active policies of incentive schemes focus on large companies. As an illustration, less than one percent of firms in the sample reported being targeted by incentives schemes, and all of them were large manufacturing firms. While extending such benefits to SMEs may be impractical, phasing out some of the existing ones can generate fiscal space for improving the overall business environment in ways particularly beneficial to SMEs (lowering statutory value added taxes against elimination of targeted tax breaks is one example). Figure 0.6. Employment Change between 2003 and 2006 Employment Change Between 2003 and 2006 1,500 1,000 500 0 Micro Small MediumLarge Micro Small Medium Large Micro Small MediumLarge Construction & Transport Manufacturing Services Continuing firms Entering firms Source: Author's calculations based on Enterprise Survey 2008 Note: Full-time permanent workers considered Both construction & transport and service sectors show a more restrained performance of employment generation. Low employment figures among small and micro firms in the service sector indicate an even more severe case of bias towards large companies; within services, the majority of firms sampled are in retail trade, where informality is also dominant and rampant. Additionally, it would be expected to see the construction sector in a growing economy showing up as a strong labor absorber, but in terms of formal jobs this is not happening. Anecdotal evidence suggests that informality in the construction sector is indeed rampant. Formal employment performance is also unevenly distributed among Mozambique's regions. As shown in Figure 7.7, Maputo has concentrated most of the employment gains. The danger of disruptive migration (should formal job creation pick time employment in 2006 as the criterion for firm size classification, which makes the bias towards larger firms even more clear. 94 up only in Maputo) and the uneven observed performance across regions suggest the importance of a regional perspective when discussing policy measures for job growth. Figure 0.7. Continuing and Entering Firms by Regions Region A Region B 0 200400600800 Micro Medium Small Large MicroSmall Medium Large Micro Medium Small Large Micro Medium Small Large MicroSmall Medium Large Medium MicroSmall Large Manufacturing Construction & Transport Services Manufacturing Construction & Transport Services Region C Region D 0 200400600800 Micro Medium Small Large MicroSmall Medium Large Micro Medium Small Large Micro Medium Small Large MicroSmall Medium Large Medium MicroSmall Large Manufacturing Construction & Transport Services Manufacturing Construction & Transport Services Continuing firms Entering firms Graphs by Region Source: Author's calculations based on Enterprise Survey 2008 Note: Region A= Maputo, Region B=Matola, Region C=Beira, Region D=Nampula. Correlates for firm-level employment dynamics show a strong association between firm size and employment performance. The data also allows the identification of correlates for firm-level employment dynamics. Uncovering causality strictly from the data though (that is, determinants of employment growth) is not a simple exercise given endogeneity issues involving potential explanatory variables.75 However, analyzing the association between employment performance and different variables can be a useful tool for understanding labor market dynamics. Table 9.15 reports partial correlation coefficients between two measures of employment performance and different potential correlates.76 Even controlling for a number of other factors, the association between firm size and employment performance persists. This is further confirmation of the lagging performance of SMEs when it comes to job creation. The evidence also suggests that export behavior as a lever for employment creation is absent, consistent with the lack of 75 Possible routes to address the issue involve instrumental variables in a regression context or the use of multivariate (simultaneous equation) approaches. 76 Each partial correlation coefficient is a linear association measure that holds all other correlates fixed. See Annex IV: Productivity and Exports in Mozambican Firms for Table 9.15 with results. 95 diversification in Mozambican exports and the need for implementing competitiveness promotion reforms that effectively target SMEs. A Look at Technology and Human Capital in the Mozambican Firm In addition to the general characteristics of the labor market and employment growth dynamics, the results presented in the econometric analysis section show that technological capabilities and human capital are also important aspects of firm characteristics that need to be examined in greater detail, given their significant correlation to improved firm performance. The Enterprise Survey provides information that allows us to explore in greater depth these two dimensions of the investment climate. We first present basic descriptive statistics for the variables that can best reflect technology adoption and human capital in the Mozambican firm. We follow this presentation with simple differential probit regressions that explore how firms' characteristics affect the probability that they will have greater technological capabilities and human capital stock. Only 20 percent of Mozambican firms, on average, offer formal training programs to their full-time employees (Figure 7.8).77 While varying by firm size, this average is lower than Malawi's, where 50 percent of firms offered formal training to its employees, and much lower than South Africa, Thailand, and Vietnam, where more than 60 percent of firms, on average, offer such training (Figure 7.9). Furthermore, training seems to be particularly low among small firms (12 percent). However, training programs seem to benefit a large percentage of Mozambican workers in the companies that do offer these programs.78 77 Formal training is defined as beyond "on the job" training. 78 Due to data availability and the fact that formulation of certain questions was not comparable across surveys, the results presented here are only for those variables and countries where we could ensure comparability. 96 Figure 0.8. Percentage of Firms that Offer Formal Training Programs 100 Small Medium Large Overall 90 80 70 Percent 60 50 40 30 20 10 0 e a a nd m i la a aw qu bi si na ric go la ne m bi al ai Af et An Za am do M Th Vi h In ut oz So M Country Source: Enterprise Surveys 2003-08 Figure 0.9. Percentage of Full-Time Workers that Received Training 100 Small Medium Large Overall 80 Percent 60 40 20 0 Angola Mozambique Zambia Country Source: Enterprise Surveys 2006-08 Overall, International Standards Organization (ISO) certification and the use of the Internet for business communication are particularly low among Mozambican firms when compared to other countries in the sample. Table 7.4 presents the variables that best reflect different dimensions of technological capabilities and human capital. The figures show that quality certification and the use of Information Technology in the relation with clients is low in the country. Use of foreign licenses, on the other hand, is comparably higher in Mozambique than in all other African comparators. In terms of human capital stock, Mozambique shows considerable deficiencies in both aspects of the human capital dimension--training and managers education level. While 70 percent of Mozambican managers in the manufacturing sector have a secondary or higher level of educational achievement, the figure for most other countries is well above 90 percent. Likewise, firms that offer training represent only 20 percent of the overall sample of 97 manufacturing firms, while in other countries such as South Africa and Malawi this number is above the 50 percent mark. Table 0.4. Technology and Human Capital Technological Capabilities (%) Human Capital (%) Firms that Use Managers Managers Firms Email and/or Own Firms with Firms with with More with Web Page to ISO Certifi- that Offer Secon- than Foreign Communicate with cation Training dary Secondary Licenses Clients and Education Education Suppliers Mozambique Small 8 26 11 12 41 17 Medium 23 41 41 29 33 55 Large 21 58 74 63 16 84 Overall 13 33 24 20 37 33 Angola 4 6 27 20 28 52 Overall Means Malawi 18 15 68 51 25 73 South Africa 43 23 99 64 8 91 Zambia 16 23 54 24 12 88 Thailand 45 -- 59 76 14 82 Vietnam 38 -- 66 60 15 0.81 Source: Enterprise Surveys 2003-08 -- Not available. Firms that export are associated with an increase of 46 percentage points in the likelihood of offering formal training programs to employees. Foreign ownership is also associated with an increase in the likelihood that a firm will acquire foreign licenses or use the Internet to communicate with clients and suppliers. We next explore how firm characteristics such as exporting experience, foreign ownership, age, and manager education are associated with the likelihood that firms will engage in more technologically sophisticated business practices, such as acquiring ISO certification and foreign licenses and using the Internet to communicate with clients and suppliers. We also model the likelihood that firms will invest more on developing human capital among their personnel by offering formal training programs. The results from these differential probit regressions are shown in Table 7.5 below and show that exporters are associated with an increased likelihood of providing training, while foreign-owned firms are more likely to have foreign licenses and use the Internet for business communication. Also worthy of notice is the association between the manager's education level and an increased likelihood for three of the four variables: offering formal training programs, acquiring ISO certification, and using the Internet for business communication. Firm size is also found to be positively associated with an increased likelihood of Internet usage for business communications. 98 Table 0.5. Technology and Human Capital and Firm Characteristics (Manufacturing Firms Only) Offers Has ISO Has Use Email and/or Formal Certifications Foreign Own Web Page to Training Licenses Communicate with Clients and Suppliers Direct exporter (10%+) 0.467*** 0.186 0.032 0.216 [0.143] [0.117] [0.131] [0.147] Foreign-owned (10%+) 0.006 0.052 0.196** 0.328*** [0.056] [0.046] [0.082] [0.080] Ln (firm's age) 0.003 0 0.025 0.03 [0.025] [0.015] [0.031] [0.027] Manager with secondary education 0.089 0.235*** 0.044 0.002 [0.064] [0.072] [0.067] [0.066] Managers with more than secondary 0.206*** 0.263*** 0.068 0.130* [0.075] [0.085] [0.079] [0.077] Medium 0.073 0.036 0.083 0.135** [0.054] [0.037] [0.065] [0.060] Large 0.165 -0.025 0.19 0.576*** [0.134] [0.052] [0.152] [0.153] Observations 340 340 340 340 Source: Authors' calculations using Enterprise Survey 2008 Note: Each coefficient reports the marginal effect (for continuous variables) or discrete change (for binomial variables) of the independent variable on the innovation variables of interest. In addition to manager's education and firm size, all regressions also included a control variable for geographical location. Small firm is the omitted category for firm size. Robust standard errors are provided in brackets. * significant at 10%; ** 5%; *** 1%. Conclusions Employment generation in Mozambique is a cause for concern. While large firms have shown expansion in their labor force, employment growth among SMEs is below its potential. Government policies that can facilitate employment creation among SMEs will be an important part of the strategy needed to reduce unemployment rates and diversify Mozambique's economic base. On the brighter side, despite this less than stellar employment creation performance, evidence presented elsewhere in this report suggests that SME employment is increasingly concentrated in more productive firms, which has helped to increase the overall allocational efficiency of the economy. But, as shown in the productivity chapter of this report (Chapter 3), current productivity levels among Mozambican firms are below many of its regional competitors. These results are supported by the evidence presented in this chapter: critical determinants of firm productivity such as technological capabilities and human capital accumulation among Mozambican firms are below that of most of its neighbors. More worryingly perhaps is the fact that for some of these human capital variables (workforce education, for 99 example) firm perceptions, as evidenced by the responses to the Enterprise Survey, seem to be misaligned with the competitive reality that they need to be a part of. Therefore, the recommendation that is borne out of this analysis and is reinforced throughout this report is the need for the government to support efforts to increase firm productivity, especially among SMEs, so that Mozambican firms can link to and compete more effectively in global markets. Important aspects to focus on and that are suggested by the analysis presented here are employee training and the need for companies to focus on process and quality improvement in their operations (as evidenced by the significance of the ISO certification variable). For professional skills, employee training can be promoted by partnerships between higher education and the private sector through on-the-job training. In order to address specific short-term skill shortages, import of skilled labor may still be the most readily available solution. However, as noted in Box 7.1, immigration rules for qualified specialists who can provide managerial and technical expertise to Mozambican enterprises are still very restrictive, particularly for companies providing professional services. According to Bank's recent analytical work in the country (World Bank 2009 forthcoming), the new labor law has not changed this situation significantly and in fact, there are many complaints that it has become even more difficult to hire highly skilled foreigners. In addition, the report highlights the need to create professional standards that are in line with the realities of the country's SMEs. The report recommends that for professional services, professional associations in Mozambique should create, with government support, a framework for regulation and appropriate standards, seeking also cooperation with regional organizations in developing these standards for services and service professionals which reflect more closely the needs of small enterprises, and not just those of large firms and more affluent consumers. For technical skills, more broadly the strengthening of the technical and vocational education and training (TVET) system can provide workers with an opportunity to upgrade their skills. For the latter, it is important to modify the current model of delivery from supply to demand-driven to ensure that the system is producing the skills needed to meet the demand of the economy. The World Bank has been involved in supporting the government's efforts in this area through a project to facilitate the transition of the existing TVET system to a demand-led training system and provide beneficiaries with more market relevant skills and improved economic opportunities (World Bank 2006). Finally, while great strides have been achieved in expanding primary education, access to and the quality of post-primary education remain serious issues. As noted by a recent World Bank report (World Bank 2008b), further investments are needed to expand school networks and reduce the costs of education. 100 References Anti-Corruption Resource Centre (http://www.u4.no). Arndt, C., S. Jones, and F. Tarp. 2006. "Aid and Development: The Mozambican Case." Discussion Paper No. 27E. Ministry of Planning and Development, Maputo, May. Barro, R. J., and J.-W. Lee. 2001. "International data on educational attainment: updates and implications". Oxford Economic Papers. 53: 541-563. Biggs, T., J. Nasir, and R. Fisman. 1999. "Structure and Performance of Manufacturing in Mozambique." RPED Paper No. 107. World Bank, Washington, DC. Bigsten, A., P Collier, S. Dercon, M. Fafchamps, B. Gauthier, J.W. Gunning, A. Oduro, R. Oostendorf, C. Patillo, M. Söderbom, F. Teal, and A. Zeufack. 2003. "Credit Constraints in Manufacturing Enterprises in Africa." Journal of African Economies 12: 104­125. ------. 2004. "Do African Manufacturing Firm Learn from Exporting?" Journal of Development Studies 40(3): 115­141. Bigsten, A., and M. Soderbom. 2006. "What Have We Learned from a Decade of Manufacturing Enterprise Surveys in Africa?" World Bank Researcher Observer. 21 (2): 241-266.. Bila, S. and Rand, J. (2008) " Are Microfinance Interest Rates too High in Mozambique?" available at www.mpd.gov.mz. Byiers, B., J. Rand, F. Tarp, and J. Bentzen. 2008. "Credit Demand in Mozambican Manufacturing." Working Paper, Department of Economics, University of Copenhagen. Available at: http://www.econ.ku.dk/rand/images/FINAL3_MZCREDIT.pdf. Clement, J.A.P., and S.J. Peiris. 2008. Post-Stabilization Economics in Sub-Saharan Africa: Lessons from Mozambique. Washington, DC: International Monetary Fund. Cockburn, M., and C. Low. 2005. "Output-Based Aid in Mozambique." GPOBA, OBA Approaches, Note No. 3. World Bank, Washington, DC, January. Cohen, W.M., and F. Malerba. 2001. "Is the Tendency to Variation a Chief Cause of Progress?" Industrial and Corporate Change 10(3): 587­608. Coleman, S. 2007. "The Role of Human and Financial Capital in the Profitability and Growth of Women-Owned Small Firms." Journal of Small Business Management 45(3): 303­319. Commission on Growth and Development. 2008. The Growth Report: Strategies for Sustained Growth and Inclusive Development. Washington, DC: World Bank, on behalf of the Commission on Growth and Development. Coughlin, P. 2006. "Agricultural Intensification in Mozambique: Infrastructure, Policy and Institutional Framework." EconPolicy Research Group, Maputo, September. Dehesa, G., and D.J. Snower. 1997. Unemployment Policy: Government Options for the Labor Market. Cambridge, UK: Cambridge University Press. de Paula, A. and J.A. Scheinkman. 2007. "The Informal Sector." NBER Working Paper 13486. National Bureau of Economic Research, Cambridge, MA. De Wulf, L., and J.B. Sokol. 2004. Customs Modernization Initiatives: Case Studies. Trade and Development Series. Washington, DC: World Bank. DNEAP (National Directory of Studies and Policy Analysis). 2006. "Enterprise Development in Mozambique: Results Based on Manufacturing Surveys Conducted in 101 2002 and 2006." DNEAP discussion paper 33E-2006, Maputo. Available at http://www.mpd.gov.mz/gest/publicat.htm. Escribano, A., and J.L. Guasch. 2005. "Assessing the Impact of the Investment Climate on Productivity Using Firm-Level Data: Methodology and the Cases of Guatemala, Honduras, and Nicaragua." Policy Research Working Paper 3621. World Bank, Washington, DC. Etude Economique Conseil. 2008. "Enterprise Survey (Productivity and Enterprise Survey) in Mozambique: Sample Survey Design". Quebec, Canada. Fajnzylber, P., J.L. Guasch, and J.H. Lopez. 2009. Does the Investment Climate Matter?: Microeconomic Foundations of Growth in Latin America. Washington, DC: World Bank. Fajnzylber, P., W. Maloney, and G.M. Rojas. 2006. "Microenterprise Dynamics in Developing Countries: How Similar Are They to Those in the Industrialized World? Evidence from Mexico." World Bank Economic Review 20(3): 389­420. Fan, Q., J.G. Reis, et al. 2008. The Investment Climate in Brazil, India, and South Africa: A Comparison of Approaches for Sustaining Economic Growth in Emerging Economies. WBI learning resources series. Washington, DC: World Bank. Gonzalez, A.S., and F. Lamanna. 2007. Who Fears Competition from Informal Firms? Evidence from Latin America. Washington, DC: World Bank. Government of Mozambique. 2006. Poverty Reduction Support Strategy (PARPA II). Gregg, P., and A. Manning. 1997. "Labour Market Regulation and Unemployment." In Unemployment Policy: Government Options for the Labor Market. Cambridge, UK: Cambridge University Press. Gwartney, James, Robert Lawson, Russell Sobel and Peter Leeson, 2007. Economic Freedom of the World: 2007 Annual Report, The Fraser Institute. www.freetheworld.com. Hausmann, R., J. Hwang, and D. Rodrik. 2005 "What You Export Matters." NBER Working Paper 11905. National Bureau of Economic Research, Cambridge, MA, December. Hubbard, R.G. 1998. "Capital-Market Imperfections and Investment." Journal of Economic Literature 36: 193­225. International Finance Corporation (IFC). 2008. Gender: Where We Work. Available at: www.ifc.org/ifcext/sustainability.nsf/Content/Gender_WhereWeWork. IFC, Washington, DC. International Labour Organization (ILO), Bureau of Statistics, LABORSTA Internet. www.laborsta.ilo.org. International Monetary Fund (IMF). "International Financial Statistics." Online database. Kaufmann D., A. Kraay, and M. Mastruzzi. 2008. Governance Matters VII: Governance Indicators for 1996­2007. Washington, DC: World Bank. Klinger, B., and D. Lederman. 2004. "Discovery and Development: An Empirical Exploration of `New' Products." Policy Research Working Paper 3450. World Bank, Washington, DC, November. Loayza, N. 2009. "Informality in Latin America and the Caribbean" Mimeo, World Bank, Washington, DC. Loayza, N.V., A.M. Oviedo, and L. Serven. 2005. "The Impact of Regulation on Growth and Informality--Cross-Country Evidence." Policy Research Working Paper 3623. World Bank, Washington, DC. 102 Maloney, W.F. 2004. "Informality Revisited." World Development 32(7): 1159­1178. Morduch, J. 1999. "Between the Market and State: Can Informal Insurance Patch the Safety Net?" World Bank Research Observer 14: 187­207. Olley, G.S., and A. Pakes. 1996. "The Dynamics of Productivity in the Telecommunications Equipment Industry." Econometrica 64(6): 1263­1298. Oviedo, A.M. 2008. "Economic Informality: Causes, Costs, and Policies. A Critical Survey of the International Literature." Background paper for Turkey Country Economic Memorandum Economic Growth and Informality. World Bank, Washington, DC. Paul Budde Communication Pty Ltd. 2007. "Mozambique--Telecoms Market Overview & Statistics." Bucketty, Australia, June 2007. Perry, G.E., W.F. Maloney, O.S. Arias, P. Fajnzylber, A.D. Mason, and J. Saavedra- Chanduvi. 2007. Informality: Exit and Exclusion. Washington, DC: World Bank. Schwab, K., & Sala-i-Martin, X. (2009). The global competitiveness report 2009-2010. Geneva: World Economic Forum. The PRS Group, International Country Risk Guide (ICRG). www.icrgonline.com. Rand, J. 2007. "Credit Constraints and Determinants of the Cost of Capital in Vietnamese Manufacturing." Small Business Economics 29: 1­13. Rosa, P., S. Carter, and D. Hamilton. 1996. "Gender as a Determinant of Small Business Performance: Insights from a British Study." Small Business Economics 8(6): 463­478. Schneider, F. 2007. "Shadow Economies and Corruption all over the World: New Estimates for 145 Countries." Economics: The Open-Access, Open-Assessment E- Journal, Vol. 1, 2007­9. http://www.economics-ejournal.org/economics/journalarticles/ 2007-9. USAID (United States Agency for International Development). 2007. "Financial Sector Constraints on Private Sector Development in Mozambique." USAID, Washington, DC. Available at: http://www.tipmoz.com/page.php?cat1=117&cat2=262&cat3=557. Van Biesebroeck, J. 2005. "Exporting raises productivity in sub-Saharan African manufacturing firms". Journal of International Economics. 67 (2): 373-391. World Bank. Data Development Platform. Washington, DC: World Bank Group. ------. ICT at a glance tables. Washington, DC: World Bank Group. ------. Mozambique ­ Energy Reform and Access Program. Washington, DC: World Bank Group. ------.Various years. Doing Business Indicators. Online at: http://www.doingbusiness.org/. ------. Various years. Doing Business. Report. Washington, DC: World Bank. ------. 2004. World Development Report 2005: A Better Investment Climate for Everyone. Washington, DC: World Bank. ------. 2005a. "Mozambique Country Economic Memorandum." World Bank, Washington, DC. ------. 2005b. OBApproaches Note Number 03. World Bank, Washington, DC. ------. 2006. Mozambique ­ Technical and Vocational Education and Training Project ­ Project Appraisal Document. World Bank, Washington, DC. 103 ------. 2007a. "Mozambique--Brief on Railways and Ports." Unpublished. World Bank, Washington, DC. ------. 2007b. World Development Report 2008: Agriculture for Development. Washington, DC: World Bank. ------. 2008a. Africa Development Indicators 2007. Washington DC: World Bank. ------. 2008b. Beating the Odds: Sustaining Inclusion in Mozambique's Growing Economy. Washington, DC: World Bank. ------. 2008c. "Mozambique Transport Cost Study." Unpublished. World Bank, Washington, DC. ------. 2009. Mozambique ­ Competitiveness and Private Sector Development Project. World Bank, Washington, DC. ------. 2009, forthcoming. "Reshaping Growth and Creating Jobs through Trade and Regional Integration." Mozambique Country Economic Memorandum. World Bank, Washington, DC. ------. Forthcoming. "Zambia Investment Climate Assessment." World Bank, Washington, DC. World Bank and International Monetary Fund. 2007. Global Monitoring Report 2007. Washington, DC: World Bank and IMF. World Bank and United Nations Conference on Trade and Development (UNCTAD). World Integrated Trade Solution (WITS). COMTRADE database, maintained by the United Nation Statistical Division (UNSD). Available at: http://wits.worldbank.org. World Economic Forum. 2007. The Africa competitiveness report 2007. Geneva, Switzerland: World Economic Forum. World Trade Organization Statistical Database. Available at: http://stat.wto.org/. 104 Technical Annexes ANNEX I: SAMPLING METHODOLOGY The Mozambique Enterprise Survey included 599 firms located in Maputo, Matola, Beira, and Nampula. The survey followed the sampling methodology for the global roll- out of Enterprise Surveys.79 The sample consists of a range of micro, small, medium, and large firms in the different sectors identified in Table 9.1. The survey is based on the standard Enterprise Survey questionnaire to ensure international comparability. The data was collected by face-to-face interviews of company management, performed by trained enumerators during the period 2007­08. Data collection was overseen by the Enterprise Analysis Unit of the World Bank's Investment Climate Department. The survey targeted establishments in the following industries (according to Industrial Classification of All Economic Activities [ISIC], revision 3.1): all manufacturing sectors (Group D), construction (Group F), retail and wholesale services (subgroups 52 and 51 of Group G), hotels and restaurants (Group H), transport, storage, and communications (Group I), and computer and related activities (subgroup 72 of Group K). For establishments with five or more full-time permanent paid employees, this universe was stratified according to the following categories of industry: 1. Manufacturing: Food and Beverages (Group D, subgroup 15); 2. Manufacturing: Garment (Group D, subgroup 18); 3. Manufacturing: Furniture (Group D, subgroup 36); 4. Retail Trade: (Group G, subgroup 52); 5. Rest of the universe, including: - Construction (Group F); - Wholesale trade (Group G, subgroup 51); - Hotels, bars and restaurants (Group H); - Transportation, storage, and communications (Group I); - Computer related activities (Group K, subgroup 72); - Manufacturing: Other Manufacturing (Group D excluding subgroups 15, 18, and 36). 79 This section is based on information provided by Etude Economique Conseil, the company responsible for implementing the survey in Mozambique, in the document titled "Enterprise Survey (Productivity and Enterprise Survey) in Mozambique: Sample Survey Design." 105 Table 0.1. Sample Distribution by Region, Sector, and Firm Size Non- Metal & Chemicals & Other Construction Food Garments Electronics metallic Services Total Machinery Pharmaceuticals Manufacturing & Transport Products Maputo Micro (<=4) 5 4 0 0 0 0 2 60 0 71 Small (5­19) 25 32 50 0 1 6 50 72 0 236 Medium (20­99) 31 6 11 2 2 6 11 29 0 98 Large (>=100) 5 2 0 2 0 0 1 7 3 20 Subtotal 66 44 61 4 3 12 64 168 3 425 Matola Micro (<=4) 2 0 0 0 0 0 0 13 0 15 Small (5­19) 3 1 4 1 0 3 3 4 0 19 Medium (20­99) 6 2 7 1 0 2 5 2 0 25 Large (>=100) 0 0 1 1 0 1 0 0 0 3 Subtotal 11 3 12 3 0 6 8 19 0 62 Beira Micro (<=4) 0 0 0 0 0 0 0 24 0 24 Small (5­19) 4 4 4 0 0 0 4 1 0 17 Medium (20­99) 2 1 1 0 0 0 2 5 1 12 Large (>=100) 1 0 0 0 0 0 0 0 1 2 Subtotal 7 5 5 0 0 0 6 30 2 55 Nampula Micro (<=4) 0 1 1 0 0 0 2 6 0 10 Small (5­19) 5 3 4 0 0 1 10 9 1 33 Medium (20­99) 3 0 1 0 0 1 1 2 0 8 Large (>=100) 4 1 0 0 0 0 0 0 1 6 Subtotal 12 5 6 0 0 2 13 17 2 57 Total 96 57 84 7 3 20 91 234 7 599 Source: Enterprise Survey 2008 106 A satisfactory list of establishments with five or more full-time paid permanent employees was obtained from the following sources: - Instituto Nacional de Estatistica (INE) , CEMPRE_2002_2003; - Centro de Promoção de Investimentos, Investimentos por sector e nacionalidade; - Instituto de Promoção de Exportações (IPEX), IPEX_Directory 2006_2007; - KPMG, Top 100 - Centro de Promoção de Investimentos--Nampula, 2000­05 - Ministério da Industria e Comércio--Nacala, Lista de empresas exportadoras e seus produtos a exportar - Bases de dados de empresas--zona norte - Ministério dos Transportes e Comunicações--Empresas de transporte de carga The survey also sampled a selection of micro establishments (establishments with less than five full-time permanent paid employees) from the targeted universe, without stratification by industry. During the survey period, the master list was updated as new information regarding establishments that had closed or were out-of- scope was gathered. The final population size in all strata and locations is estimated at 15,546, of which 1,163 are non-micro establishments. The sample was selected at random from the master list by a computer program. For many reasons including the small size of establishments, their expected high rate of turnovers, the high level of "informality" of establishments in many activities and consequently, the difficulty to obtain trustworthy information from official sources, EEC Canada selected an aerial sampling approach to estimate the population of micro establishments and select the sample in this stratum for all regions of the survey. 107 ANNEX II: INFORMALITY CORRELATIONS A simple OLS regression was run to compute correlations between firm's perception of tax evasion in their own sectors and firm level correlates. The results of the regression and the explanation of the different indicators used are presented below. Unreported Sales Corruption Dummy 5.588 [4.737] Bureaucracy Indicator 1.271*** [0.262] Tax Inspections 0.057 [0.036] Tax rate as Constraint 11.232** [4.562] Rule of Law Indicator 3.482** [1.612] Bank Loans 0.672 [6.243] Firm's Age 0.394** [0.181] Labor Productivity 0.000*** [0.000] Exporter 7.073 [9.934] Foreign-owned 13.595*** [5.127] Formality Status at time of start up 9.711* [4.960] Small (5­19 employees) 7.141 [11.572] Medium (20­99 employees) 2.658 [12.094] Large (100+ employees) 0.437 [15.050] Services 0.185 [11.939] Others 0 [0.000] Matola 2.249 [5.797] Beira 39.394*** [5.010] Nampula 22.859*** [5.528] Constant 67.184*** [12.363] Observations 374 Source: Authors' calculations using Enterprise Survey 2008 Notes: * significant at 10%; ** 5%; *** 1% Robust standard errors in brackets Omitted categories: Maputo, manufacturing industry, micro firms. 108 Unreported Sales: The exact question that firms are asked to respond to is the following: What percentage of total annual sales would you estimate a typical establishment in your sector of activity reports for tax purposes? Responses were coded as variable j8 and we used (100-j8) as the variable for unreported sales. Corruption Indicator: Dummy variable that answers the following question: Has establishment been asked for informal payments when requesting public services? Bureaucracy Indicator: What percentage of total senior management's time was spent in dealing with requirements imposed by government regulations? Tax Inspections: Over the last 12 months, how many times was this establishment visited by, inspected by or required to meet with tax officials? Tax Rates as Constraint: Dummy variable equals 1, if firm perceived tax rates as a major or very severe obstacle to business. Rule of Law Indicator: First component of Principal Component Analysis for two variables, both asked the firm, "To what extent do you agree with the statement:" - j1a: that government officials' interpretation of the laws and regulations affecting this establishment are consistent and predictable. - H7a: the court system is fair, impartial and uncorrupted. Bank Loans: Does your establishment currently have a line of credit or loan from a financial institution? Firm's age: in years. Labor productivity: value added per worker. 109 ANNEX III: MODELING DEMAND FOR EXTERNAL FINANCE AND CREDIT CONSTRAINTS In this section we employ a direct approach to analyzing credit rationing using different definitions of "credit constraints" in line with Byiers et al. (2008). Moreover, we recognize that credit constraint studies might be subject to selection bias since not all enterprises have credit demand. We, therefore, follow Bigsten et al. (2003) by: (i) first classifying firms with demand for external finance; and (ii) conditional on credit demand, we establish the characteristics of credit constrained Mozambican firms. The underlying credit constraints model in focus is given by equation (1): ci = 1[ xi + i > 0] (1) where ci is a binary variable taking the value one if the firm is credit constrained, and zero otherwise. xi is a vector of constraint proxies, and i is a random error term for firm i. Estimating equation 1 requires that we define what is meant by credit constrained. We define three different credit constraint indicator variables. First, we use the simplest indicator of firms being constrained in their access to credit by defining a firm as credit constrained if it has applied for credit in 2006 and has been denied credit. Table 9.2 shows that only 17.5 percent of firms applied for credit (15.8 percent excluding micro establishments) and 6.7 percent of the full sample is credit constrained according to this proxy measure (38.5 percent of those applying). Table 0.2. External Finance Demand and Credit Constraint Definitions Full Sample Non-micro 1 2 3 4 With Credit With Credit All Demand All Demand Applied for credit 0.175 0.226 0.158 0.196 Constrained def. 1 0.067 0.087 0.071 0.087 Constrained def. 2 0.243 0.313 0.248 0.307 Constrained def. 3 0.668 0.861 0.720 0.892 Have credit demand 0.776 1.000 0.808 1.000 Constrained def. 1a 0.037 0.048 0.034 0.042 Constrained def. 2a 0.212 0.274 0.212 0.262 Constrained def. 3a 0.637 0.822 0.684 0.847 Observations 593 460 468 378 Source: Enterprise Survey 2008 Note: Credit demand Definition 1 used (including firms not applying for loans due to high interest rates as part of the credit demanding group). However, not all firms have credit demand. The survey instrument has information explaining why firms did not apply for credit. One reason is that the firm has "no need for a loan--establishment has sufficient capital." Table 9.2 shows that between 77.6 and 80.8 percent of the enterprises, depending on the sample considered, have a demand for external financing, and columns 2 and 4 in Table 9.2 show the summary statistics including only firms with credit demand. 110 The second measure of firm credit rationing labels a firm constrained in their access to credit if (i) it applied and has been denied credit, or (ii) did not apply for credit due to reasons such as "application procedures to complex," "collateral requirements unattainable," or "possible loan size and maturity insufficient." Table 9.3 shows reasons for not applying. Panel A shows the aggregate result for the full sample and non-micro enterprises only, whereas Panel B document shows results by firm size. Almost half of the firms give "interest rates not favorable" as the main reason for not applying for loans, and mainly non-micro firms state this as a problem. As in Bigsten et al. (2003), we do not include those firms answering "interest rates to high" or "did not believe it would be approved" as being part of the constrained firms in Definition 2 as these answers may reflect that investment projects simply are not competitive at going interest rates. For comparison, we include these firms as constrained in our third credit constraint definition.80 Table 9.2 shows that using Definition 3 around 67 percent of firms can be labeled credit constrained and considering only firms with credit demand this number increases to 86 percent. These numbers correspond to the high estimates obtained in Byiers et al. (2008) using a closely related credit constraint definition. Some of the firms applying and being denied credit may not have a sustainable business plan and should, therefore, not be labeled as credit rationed. The questionnaire enables us to provide reasons for why the project proposal was rejected. Table 9.4 shows loan rejection reasons cited by the lender. Between 48.5 and 55.0 percent of the cases, the rejection was based on collateral issues. However, in the remainder, the reasons were "Insufficient profitability (of the project proposal and historically)," "Problems with credit history," or "Other objections." As a robustness check, we exclude these as credit constrained firms from the definitions described above (resulting in constraint variable 1a to 3a documented in Table 9.2). However, this does not change any of the overall conclusions found in the subsequent analysis and in the following we use Definitions 2a to 3a as our key constraint measures. Finally, as noted above and documented in Table 9.2, some firms not applying for credit are simply not demanding external finance. A potential bias may therefore be present when estimating equation (1) directly. We, therefore, carry out a joint analysis of firm credit demand and credit constraints using a sample selection model. Hence, we record if enterprises had a credit demand in 2006 and conditional on this demand we estimate the probability of being credit constrained. Table 0.3. Main Reason for Not Applying for Credit PANEL A Full Sample Non-micro Obs. Percent Obs. Percent 80 In Bigsten et al. (2003), firms not applying for credit due to "too high interest rates" are grouped in the "unconstrained, no credit demand" category (credit demand definition 2 in this paper). In this section, we include these firms in the group demanding credit (definition 1), given the nature of the current Mozambican financial market summarized in the introduction and described in Byiers et al. (2008). 111 No need for loan (133) 27.1 (90) 22.8 Application procedures to complex (60) 12.3 (46) 11.7 Interest rates not favorable (213) 43.6 (192) 48.7 Collateral requirements are unattainable (28) 5.7 (22) 5.6 Size of loan/maturity insufficient (16) 3.3 (15) 3.8 Did not think it would be approved (17) 3.5 (15) 3.8 Other (22) 4.5 (14) 3.6 PANEL B Micro Small Medium Large No need for loan 47.2 19.0 31.7 17.7 Application procedures to complex 15.7 15.2 5.0 0.0 Interest rates not favorable 18.0 48.3 50.8 52.9 Collateral requirements are unattainable 6.7 5.7 5.0 5.9 Size of loan/maturity insufficient 1.1 4.9 0.8 5.9 Did not think it would be approved 2.3 4.6 2.5 0.0 Other 9.0 2.3 4.2 17.7 Observations (89) (263) (120) (17) Source: Enterprise Survey 2008 Note: Numbers in percentage (observations in parenthesis) Table 0.4. Loan Rejection Reasons Full Sample Non-micro Obs. Percent Obs. Percent Collateral or cosigners unacceptable (22) 55,0 (16) 48.5 Insufficient profitability (9) 22,5 (9) 27.3 Problems with credit history/report (4) 10,0 (4) 12.1 Other objections (5) 12,5 (4) 12.1 112 Source: Enterprise Survey 2008 Note: Numbers in percentage (observations in parenthesis) Following Bigsten et al. (2003), a model of credit demand should be able to capture differences between the capital returns and the cost of capital. Ideally, we would proxy capital returns by the gross profit share of total assets (profit rate). However, due to missing data we only include location (region) and sector dummies (sector) as controls for shocks to the conditions faced by firms in the different areas of the economy. To model the cost of capital side of the equation we need to include variables capturing: (i) the opportunity cost of capital; (ii) collateral requirements; and (iii) loan transaction costs. First, for opportunity costs, we use several different proxies for the availability of alternative sources of funds. In the absence of an informal credit market (for the sample considered), we include an indicator variable for having access to an overdraft facility (overdraft) or being a subsidiary of another firm (branch). Moreover, several papers (see Bigsten and Soderbom 2006 for an overview) have found trade credit as an important source of especially working capital financing and van Biesebroeck (2005) documents that exporters use trade credit more frequently. We, therefore, include an indicator variable for export participation (export) and a trade credit proxy for the proportion of working capital financed via credit from suppliers and advances from customers (trade credit). Second, we include a variety of controls for access to collateral and loan transaction costs. Relevant proxies for these controls include assets information and a full debt history for the firm. Due to missing data for assets, we do not fully cover this aspect of access to collateral, and we rely on a simple indicator variable for currently having debt (debt) obtained before 2006 as the measure of the firm's debt history. However, we use a wide range of variables representing various aspects of transaction costs: (i) legal status dummies (legal); (ii) ownership structure (share of foreign ownership--foreign) dummies; (iii) firm age (age); (iv) and indicator variable representing whether the firm keeps external audited accounts (audit); and (v) owner-specific characteristics (gender [equal to one if any firm owners are female, zero otherwise], ethnic [equal to one if the largest shareholder is non-African, zero otherwise] and education [equal to one if owner has no education or below secondary schooling, zero otherwise]). Finally, we include firm size (number of employees: size) allowing us to analyze whether there exists a firm size bias in Mozambican credit allocation even when controlling for different aspects of firm heterogeneity. Results Table 9.5 shows the summary statistics by firm size of the variables described above. Without commenting in great detail on these summary statistics, the following is worth noting with respect to credit demand: (i) Some 28 percent of micro and small firms have external audits as compared to 60 and 74 percent for medium and large firms, respectively; (ii) very few firms obtained loans before 2006 (7.1 percent of sample considered), combined with the above evidence this suggests that more firms have formal loans today than compared to just a couple of years ago; (iii) having an overdraft facility 113 increases with firm size; (iv) the probability of exporting increases with firm size and almost none of the micro and small enterprises engage in export activities; and (v) finally, around 15 percent of firms' working capital is financed using trade credits. This is clearly different from the conclusion reached in USAID (2007) suggesting almost no use of trade credits as part of working capital financing in Mozambique. Table 9.6 provides an overview of credit demand and constraints by firm size, legal structure, location, and sector. It should be noted that definitions "Demand 1" and "Credit Constrained 3" correspond to each other as they consider enterprises that do not apply for credit due to high interest rates as having demand for external funds. Demand 2 and Credit Constrained 2 consider these firms as not demanding credit. Moreover, summary statistics for the two credit constraint definitions are reported conditional on having credit demand.81 According to the first credit demand definition, 78 percent have demand for external financing. Micro firms have on average less credit demand than their small, medium, or large counterparts. This picture, however, changes when turning to the demand definition following Bigsten et al. (2003) where "non-loan applicants due to high interest rates" are not considered in the credit demanding group. Only 42 percent of the firms considered have credit demand and it is especially small, medium, and large firms that stay out of the credit market due to high interest rates. Turning to credit demand patterns by legal structure, location, and sector, we observe the following: (i) private held limited liability companies have lower demand for external financing than the other three groups considered; (ii) firms in Beira have the lowest credit demand according to Definition 1, but the highest demand using Definition 2 which suggests that the high cost of financing influence firm credit demand relatively less in Beira than in the other regions considered; and finally (iii) credit demand for firms in food processing seem particularly sensitive to the cost of credit (going from 85 percent demanding credit according to Definition 1 to merely 33 percent using Definition 2), whereas, enterprises in retail are less sensitive. However, firms in retail do on average seem to have a relatively lower demand for external funds. Table 0.5. Credit Demand Determinants: Summary Statistics by Firm Size Total Micro Small Medium Large Size 24.273 2.218 8.712 40.553 212.815 Age 13.560 8.311 13.235 17.759 18.444 Gender 0.255 0.294 0.212 0.305 0.296 Ethnic 0.268 0.076 0.219 0.468 0.630 Manager Education 0.290 0.471 0.330 0.106 0.000 Audit 0.379 0.277 0.284 0.603 0.741 Debt 0.071 0.050 0.052 0.085 0.296 Foreign 0.170 0.076 0.134 0.262 0.519 81 Note that the difference between the summary statistics reported in Table 9.2 and in Table 9.6 for constraint definition 2 is due to the fact that Table 9.2 conditions on credit demand Definition 1, whereas Table 9.6 uses credit demand Definition 2. 114 Overdraft 0.128 0.084 0.082 0.213 0.407 Branch 0.093 0.034 0.046 0.206 0.296 Export 0.046 0.008 0.013 0.092 0.333 Trade credit 0.147 0.078 0.148 0.196 0.183 Legal 1 (Note 1) 0.132 0.134 0.059 0.255 0.296 Legal 2 0.636 0.731 0.739 0.404 0.259 Legal 3 0.103 0.025 0.111 0.121 0.259 Legal 4 0.130 0.109 0.092 0.220 0.185 Region 1 (Note 2) 0.710 0.597 0.778 0.681 0.593 Region 2 0.105 0.126 0.062 0.177 0.111 Region 3 0.093 0.202 0.056 0.085 0.074 Region 4 0.093 0.076 0.105 0.057 0.222 Sector 1 (Note 3) 0.160 0.059 0.118 0.298 0.370 Sector 2 0.094 0.042 0.131 0.057 0.111 Sector 3 0.126 0.034 0.183 0.106 0.000 Sector 4 0.322 0.723 0.242 0.184 0.185 Sector 5 0.297 0.143 0.327 0.355 0.333 Observations 593 119 306 141 27 Source: Enterprise Survey 2008 Note 1: Legal 1 = Private held limited liability company. Legal 2 = Sole Proprietorship. Legal 3 = Partnership. Legal 4 = Limited Partnership. Note 2: Region/City 1 = Region A/Maputo. Region/City 2 = Region B/Matola. Region/City 3 = Region C/Beira. Region/City 4 = Region D/Nampula. Note 3: Sector 1 = Food. Sector 2 = Garments. Sector 3 = Other Manufacturing. Sector 4 = Retail. Sector 5 = Other Services incl. Construction. 115 Table 0.6. Credit Demand and Constraints by Size, Legal Structure, Location and Sector Credit Credit Demand 1 Demand 2 Constrained 2 Constrained 3 Full sample 0.776 0.417 0.510 0.822 Micro (1-4 employees) 0.647 0.513 0.443 0.688 Small (5-19 employees) 0.837 0.422 0.636 0.887 Medium (20-99 employees) 0.730 0.298 0.310 0.796 Large (100 employees and above) 0.889 0.556 0.267 0.667 Private held limited liability company 0.718 0.321 0.440 0.750 Sole Proprietorship 0.790 0.422 0.535 0.846 Partnership 0.754 0.459 0.571 0.848 Limited Partnership 0.779 0.455 0.400 0.750 Maputo 0.786 0.397 0.521 0.849 Matola 0.774 0.468 0.586 0.750 Beira 0.618 0.473 0.346 0.647 Nampula 0.855 0.455 0.520 0.830 Food 0.853 0.326 0.484 0.827 Garments 0.839 0.446 0.520 0.766 Other manufacturing 0.853 0.413 0.677 0.938 Retail 0.654 0.429 0.427 0.776 Other services incl. Construction 0.813 0.443 0.538 0.825 Source: Enterprise Survey 2008 Note: "Demand 1" and "Credit Constrained 3" include enterprises that do not apply for credit due to high interest rates as demanding credit/being constrained. "Demand 2" and "Credit constrained 2" do not include these firms as part of the credit demanding group. The credit constraint measures in Table 9.7 (columns 3 and 4) are presented conditional on having credit demand. Using constraint measure 2, some 51 percent of those demanding credit are constrained in the financial market. This number corresponds to 21 percent of the total sample. The equivalent figures for credit constraint 3 are 82 and 64 percent. It is noteworthy that small enterprises are on average more constrained with over 60 percent of those demanding credit being credit rationed. Firms in Beira are relatively less credit constrained as compared to firms in Maputo, Matola, and Nampula. Finally, firms in the category "Other manufacturing" are the most financially constrained of the categories considered. Switching from simple cross tabulations to multivariate analysis, Table 9.7 presents probit results of the demand for credit equation. Columns 1 and 2 report the results of the full sample (unweighted) and the non-micro firm sample (applying appropriate weights) using credit demand Definition 1 where firms not applying for credit due to "too high interest rates" are grouped are included in the group demanding. Columns 3 and 4 follow the Bigsten et al. (2003) approach by including these firms in the "unconstrained, no credit demand" category (credit demand Definition 2). 116 Table 0.7. Credit Demand Determinants Credit Demand Def. 1 Credit Demand Def. 2 (1) (2) (3) (4) Size 0.000 0.001*** 0.001 0.002*** (1.27) (3.32) (1.28) (2.95) Age 0.001 0.000 0.001 0.002 (0.47) (0.11) (0.81) (0.75) Gender 0.061 0.076* 0.095* 0.106 (1.58) (1.88) (1.90) (1.60) Ethnic 0.108** 0.082* 0.092* 0.039 (2.41) (1.80) (1.73) (0.59) Manager Education 0.068* 0.043 0.118** 0.023 (1.68) (0.88) (2.28) (0.33) Audit 0.066* 0.126*** 0.129*** 0.225*** (1.73) (3.16) (2.75) (3.72) Debt 0.138** 0.083 0.322*** 0.306*** (2.06) (1.14) (3.50) (2.85) Overdraft 0.049 0.027 0.170** 0.144* (0.95) (0.52) (2.41) (1.70) Branch 0.041 0.049 0.022 0.029 (0.64) (0.76) (0.28) (0.32) Export 0.028 0.035 0.035 0.025 (0.31) (0.42) (0.32) (0.21) Trade credit 0.221** 0.350*** 0.234* 0.187 (2.33) (3.47) (1.92) (1.20) Legal structure dummies Yes Yes Yes Yes Regional dummies Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Weights used No Yes No Yes Observations 593 468 593 468 Pseudo R-squared 0.10 0.12 0.08 0.11 Source: Enterprise Survey 2008 Note: Probit, marginal effects. Unweighted and weighted estimates and robust standard errors. t-statistics in parenthesis. *, **, *** indicates significance at a 10 percent, 5 percent and 1 percent level, respectively. Base: Private held, limited liability company, Maputo, Retail. There is an indication of larger firms having higher demand for external funds although this is only well-determined in the non-micro weighted estimations. Second, owner-specific characteristics seem to matter for credit demand. Having female shareholders, an African as major shareholder, and less-educated managers are positively associated with larger credit demand. Third, somewhat surprisingly, firms having external audits have less demand for outside credit. Since these audits may reduce transaction costs of obtaining credit, one would expect a positive association between audits and credit demand. Fourth, firms with debt are more likely to have a demand for external finance. This is in line with the results in Bigsten et al. (2003) and suggests that firm debt/credit history to a larger degree serves as an indicator reducing firm transaction costs rather than a signal of increased default risk. The same argument seems to apply for having access to overdrafts (only well-determined using credit demand Definition 2). 117 Finally, as expected, firms with access to trade credit have a lower demand or external finance since they have access to alternative external funds. When modeling the second step of the sample selection model, we assume that banks only have limited firm-specific information available when deciding which firms should be granted credit. In line with Bigsten et al. (2003), we model the credit constraint equation using the following variables: firm size, firm age, legal structure, industry, location, and a set of firm-level characteristics often available within the credit system (debt, overdraft, audit, foreign, branch, export). Table 9.8 presents the results of the credit constraint equation, unadjusted (columns 1, 2, 5, and 6) and adjusted (columns 3, 4, 7, and 8) for sample selection. In addition, the results are presented for two different credit constraint definitions; Definition 2 (columns 1 to 4) and Definition 3 (columns 5 to 8). The number of uncensored observations using credit constrained Definition 2 is lower than for Definition 3 since firms not applying for loans due to "interest rates to high" are categorized as not demanding credit (columns 3 and 4). Several findings stand out. First, all coefficients on firm size are negative in line with the results in Byiers et al. (2008). However, only in two cases (columns 6 and 8) do we find a well-determined negative association for supporting the perception that credit constraints are more binding for smaller firms. Second, audited firms are less likely to be credit constrained. More specifically we find (using columns credit constraint Definition 2) that enterprises with external audits are 4 to 19 percent less likely to be credit constrained than firms not keeping formal accounts. Third, considering constraint Definition 3, firms with a higher share of foreign ownership are less likely to be credit constrained, which supports the findings in Byiers et al. (2008). Moreover, having an overdraft facility is negatively associated with being credit constrained, supporting the above creditworthiness interpretation. 118 Table 0.8. Credit Constraint Determinants (1)(a) (2)(b) (3) (4) (5) (6) (7) (8) Size 0.003 0.001 0.004 0.002 0.001 0.002* 0.000 0.002* (1.16) (0.69) (1.18) (0.60) (0.55) (1.85) (0.26) (1.81) Age 0.015** 0.013 0.018* 0.017 0.001 0.001 0.001 0.003 (2.51) (1.59) (1.94) (1.26) (0.22) (0.15) (0.27) (0.52) Audit 0.481*** 0.778*** 0.488** 0.845** 0.135 0.317** 0.058 0.212 (3.50) (4.46) (2.01) (2.23) (1.10) (2.06) (0.42) (0.98) Debt 0.115 0.378 0.695 0.776 0.238 0.134 0.361 0.390 (0.41) (1.17) (1.34) (1.06) (1.03) (0.47) (1.56) (1.13) Foreign 0.007 0.007 0.085 0.014 0.457*** 0.463** 0.266* 0.292 (0.04) (0.03) (0.28) (0.04) (2.89) (2.42) (1.69) (1.35) Overdraft 0.060 0.237 0.372 0.618 0.425** 0.618*** 0.531*** 0.869** * (0.29) (0.93) (1.08) (1.22) (2.51) (2.99) (2.84) (2.61) Branch 0.171 0.060 0.198 0.038 0.250 0.161 0.268 0.146 (0.67) (0.21) (0.56) (0.08) (1.24) (0.67) (1.26) (0.54) Export 0.757 0.514 1.108 0.996* 0.262 0.104 0.167 0.017 (1.61) (1.09) (1.50) (1.68) (0.90) (0.33) (0.55) (0.05) Legal structure Yes Yes Yes Yes Yes Yes Yes Yes dummies Regional dummies Yes Yes Yes Yes Yes Yes Yes Yes Sector dummies Yes Yes Yes Yes Yes Yes Yes Yes Weights used No Yes No Yes No Yes No Yes Sample selection No No Yes Yes No No Yes Yes Model Obs. 593 468 593 468 593 468 593 468 Uncensored .. .. 247 186 .. .. 460 378 Pseudo R-squared 0.07 0.11 0.07 0.11 0.09 0.28 0.09 0.28 LR-test for indep. .. .. 0.80 0.87 .. .. 0.01 0.27 eq. Source: Mozambique Enterprise Survey 2008 Note: (a) Credit Constraint Def. 2. (b) Credit Constraint Def. 3. Probit with and without sample selection. Unweighted and weighted estimates and robust standard errors. t-statistics in parenthesis. *, **, *** indicates significance at a 10 percent, 5 percent and 1 percent level, respectively. All regressions included a constant. Base: Private held, limited liability company, Maputo, Retail. .. Negligible. 119 ANNEX IV: PRODUCTIVITY AND EXPORTS IN MOZAMBICAN FIRMS In this section, we present the calculations and methodology used to examine the correlations between ICA variables, firm-specific characteristics, and firm productivity. Figures for the labor productivity analysis and a table with results for firm- level employment analysis are also included at the end of the section. Since our data are cross-sectional, inferences about causality are not possible. However, significance in the correlations allows us to determine which indicators have a significant impact on enterprise performance and identify those whose impact is minimal. The data generated by the Enterprise Surveys allows for exploring differences both within and across countries. As presented in the main part of the document, the analysis conducted for this ICA includes estimations of firm performance in Mozambique alone and estimations where this data is pooled with other data from countries considered to be important comparators for Mozambique. This pooling allows for more efficient estimates of the correlations between investment climate variables and firm performance, as well as providing information for comparisons of impacts across countries. Methodology The basic idea is that productivity is a function of firms' characteristics (age, size, industry, and location, for example) and investment climate attributes (regulations, access to finance, infrastructure, technology, and so forth). The basic specifications explored in this analysis are the following: Investment climate and labor productivity: log(Yi / Li ) = j 1 j IC ij + j 1 j C ij + v1i (1) where Y is output as measured by the firm sales or value added (when material inputs are subtracted from sales), and L is labor (so that the dependent variable can be interpreted as labor productivity). IC refers to investment climate variables, C to other firm control variables, and v is a random error term. The set of control variables Ci includes sector, region, and country dummies (for pooled data calculations), aimed at capturing unobservable firm characteristics not directly related to the investment climate, but that can be expected to affect firm performance. In addition, control variables are included for: (i) log of firm age; (ii) firm size dummies; (iii) a dummy for partially or entirely foreign-owned firms; and (iv) an exporter dummy. Investment Climate and Total Factor Productivity (TFP)82 Yi = M i + Pi K i L1- i (2a) Pi = P ( IC pi , C pi ) exp(v 2i ) (2b) 82 This derivation follows Fajnzylber et al. (2009), Chapter 2. 120 where K is capital, M is intermediate inputs, P is TFP, and u is a random error term. Rearranging (2a) and taking logs yields: log(Yi / Li - M i / Li ) = log(VAi / Li ) = log( Pi ) + log( K i / Li ) . (3) Substituting (2b) into (3), and assuming a linear specification for log(P(.)) such as log( Pi ) = j 2 j IC ij + j 2 j C ij + u i , (4) then (3) can be rewritten as: log(VAi / Li ) = j 2 j ICij + j 2 j C ij + log( K i / Li ) + v 2i . (5) It follows from (4) and (5) that in this second specification, the estimated parameters of the investment climate attributes can somewhat be interpreted as affecting productivity. To explore the impact of investment climate variables on wages, a variation of equation (1) is estimated, where the dependent variable has been substituted by the average wage per worker paid by the firm: log(Wagei / Li ) = j 3 j IC ij + j 3 j C ij + v3i . (6) One of the challenges in estimating (1), (5), and (6) is that it assumes homogeneity in the production function across sectors, countries, regions, and firms, which is unrealistic. This is mitigated to some extent by including sector, country, and regional dummies in the different econometric specifications to capture as much heterogeneity as possible. An additional challenge, however, is the difficulty in establishing causality from investment climate attributes to firm productivity and growth or from exporter status to firm productivity. Does a better business environment make firms more productive, or do more productive firms tend to locate where the investment climate is more conducive to growth? The argument is similar to the one for exporter premium and productivity. Are productive firms more likely to export? Or do firms become more productive by exporting? Escribano and Guasch (2005) addressed this issue by using a statistical method that takes into account the endogeneity between firm performance and investment climate attributes at the firm level.83 This part of their approach is followed in this 83 The statistical method is known as an instrumental variables approach where the potentially endogenous investment climate variables--at the firm level--are instrumented using the same variables aggregated by industry and region. Following the regional study for Latin America and Caribbean, we impose a minimum size of 10 observations to compute the region-industry-size average. If every group had at least 10 observations, the estimates corresponding to the group sample average and weighted average should be the same because the weights are identical for firms within a region-industry-size group. Where we do not have a minimum of 10 observations, we follow the same procedure for calculating group means for the purpose of constructing instrumental variables. We use the average of the corresponding region-industry regardless of firm size; if we are still short of 10 observations, we use industry or region averages. In the worst case scenario, we treat the observation as missing. In the cases where weighted averages have to be computed at 121 analysis, while we depart from Escribano and Guasch in the treatment of the investment climate variables. Whereas, Escribano and Guasch use a "general to particular" approach to include a large set of characteristics as explanatory variables and then reduce the number based on the significance level of each variable, the approach used for this analysis was to construct indexes that summarize investment climate attributes into a few variables. This follows the methodology used in other ICAs,84 where the first factor of principal component analysis is used to construct indexes for five main IC areas-- governance, informality, infrastructure, finance, and human capital and technology. We follow this approach for our sample of Mozambique and its comparators countries. The governance index tries to capture crime and security issues related to firm performance. This index is constructed using the following variables: the percentage of annual sales spent in security expenses and the percentage of annual sales lost due to crime. A separate binary variable that captures the extent to which firms pay bribes to access government services and/or public contracts tries to measure the degree of corruption.85 The informality index summarizes firm's perceptions about the extent of informality in their sectors. We expect the governance and bribe variables to be negatively related to firm performance because enterprises are diverting resources from productive uses--for example, to fight crime and violence. The infrastructure index measures the proportion of sales lost due to power outages and the proportion of shipment value lost in domestic transport. It is also expected to be negatively associated with firms' performance. The finance index captures whether firms have access to different financial services, such as overdraft facilities or lines of credit and whether they use banks to finance their working capital. We expect a positive relationship to firm performance. Finally, the technology index used for the pooled sample includes indicators for training offered to workers, and for Internet use to communicate with clients and suppliers. The technology and human capital index used for the Mozambique only sample includes the indicator for training offered to workers, but does not include the Internet use indicator, while supplementing the index with indicators for ISO certification and manager's education level. We expect these factors to be positively correlated with enhanced firm performance. Table 9.12 presents summary statistics for the investment climate variables and Table 9.13 presents statistics for the control variables used in the regressions. As can be seen from Table 9.12, the issue of low response was not a significant factor for the majority of the variables. For Mozambique, Bribe and ISO Certification variables present the highest non-response rates, but even here these rates do not exceed 20 percent and the higher levels of aggregation than region-industry-size, we redefine the weights to take into account the different selection probability associated with the different groups. 84 See among others, the ICA reports for El Salvador (2005), for Panama (2007), and for Central America (forthcoming). 85 More specifically, bribe payments are measured as a variable that equals one if: (i) when the interviewer asks "When establishments like this one do business with the government, what percent of the contract value would be typically paid in additional or informal payments or gifts to secure the contract?", the firm's manager answers with a percentage greater than zero; or (ii) if the firm answers affirmatively to any of the questions: "Was an informal gift or payment expected or requested to obtain the [service]?" in reference to a telephone, water, or electricity connection; an import, or operating license, or a construction permit. Bribe equals zero if the manager answers "zero" to the first question (procurement), and "no" to all other questions (public services). 122 averages for the variables change little with substitution. Missing observations in each country are substituted with the mean for the corresponding region, industry, and firm size category. We establish a threshold of 10 observations as a minimum for the mean to be representative of the population. If the region-industry-size category has less than 10 firms (observations) in it, we use the region-industry mean, provided that the minimum number of observations is reached. If not, we use the industry, or region mean, depending on the number of observations in each. If none of these categories contains at least 10 observations, we do not substitute the missing value. Finally, pair-wise correlations are presented in Table 9.14 for dependent variables and investment climate indicators and control variables. Correlations are generally low, well below 0.5, and many of the variables present correlation coefficients that are significant at the 10 percent level, especially in the pooled sample. Results Table 9.9 presents the results for the regressions between the investment climate indices and firm performance for the pool of Mozambican enterprises only. The indices for Finance and Technology and Human Capital are highly correlated to improved firm performance, and their coefficients are also statistically significant. Poor infrastructure is correlated with lower firm productivity, as measured by both value added per worker and TFP. The variables for governance, corruption (Bribes), and informality do not show a significant correlation to firm performance, with the exception of informality and value added per worker, where a positive correlation is detected.86 This shows that firms that have a better perception of tax compliance in their own sectors are also more productive. On the other hand, informality as measured by another (perception) variable87 showed a statistically significant and economically meaningful negative correlation to firm productivity measured as TFP. Finally, regional disparities in firm performance are significant: firms in Maputo perform better, particularly with regards to value added per worker, TFP and wages paid to its workers. Table 0.9. Firm Performance and Investment Climate Variables--Mozambique Only Value Added Sales per Worker TFP Wages per Worker (in Logs) (One Step) (in Logs) (in Logs) 86 Results from the econometric analysis were not able to fully capture the impact that informality may have on firm's productivity--possibly because many of the standard drivers of productivity and growth are also likely to affect informality. For example, low levels of human capital or governance leading to both lower productivity and higher informality. For a thorough discussion about these difficulties, see Perry et al. (2007). 87 This variable (coded as e11) is available in the Mozambican survey but not across most of the surveys. It answers the following question: Does this establishment compete against unregistered or informal firms? To ensure better comparability across countries, the variable used for both sets of regressions (Mozambique and Mozambique + comparators, variable coded as j8) answers the following question: What percentage of total annual sales would you estimate a typical establishment in your sector of activity reports for tax purposes? This question is part of all surveys used for comparison in this productivity analysis. 123 Value Added Sales per Worker TFP Wages per Worker (in Logs) (One Step) (in Logs) (in Logs) Poor Governance 0.003 0.252 -0.139 0.083 (security & crime) [0.233] [0.239] [0.110] [0.149] Bribes 0.571 0.871 0.8 0.041 [1.081] [1.027] [1.113] [0.572] Informality 0.011 0.026* 0.001 0.012 [0.020] [0.014] [0.007] [0.011] Poor Infrastructure 0.857 1.107*** 0.530** 0.423 [0.594] [0.395] [0.247] [0.329] Access to Finance 1.287*** 0.624** 0.852*** 0.713*** [0.345] [0.300] [0.253] [0.219] Technology & Human Capital 0.898*** 1.004** 0.472 0.354** [0.271] [0.474] [0.415] [0.152] Direct exporter (10%+ of sales) 0.921 0.304 0.598 0.58 [0.901] [1.016] [0.653] [0.567] Foreign owned (10%+) 0.107 0.502 0.282 0.063 [0.483] [0.511] [0.330] [0.235] Firm's age (in logs) 0.046 0.114 0.042 0.057 [0.154] [0.116] [0.109] [0.104] Capital per worker (in logs) 0.004 [0.073] Matola 0.756* 0.795** 0.593** 0.366* [0.386] [0.358] [0.273] [0.216] Beira 0.8 0.005 0.862** 0.104 [0.700] [0.615] [0.419] [0.392] Nampula 1.269 1.883*** 1.411*** 0.941** [0.987] [0.459] [0.370] [0.443] Services 0.111 0 0 0.265 [0.438] [0.000] [0.000] [0.253] Small (5-19 employees) 0.954 [0.674] Medium (20-99 employees) 0.778** 0.529 0.853 0.419* [0.392] [0.483] [0.558] [0.241] Large (100+ empl) 0.794 1.036 0.053 [0.723] [0.721] [0.531] Constant 10.060*** 8.706*** 8.046*** 4.829*** [0.552] [0.604] [0.832] [0.300] Observations 377 269 217 377 Source: Authors' calculations using Enterprise Survey 2008 Note: Dependent variables are expressed in logs, so that the reported coefficients can be interpreted as relative changes in respectively sales per worker, TFP and wages. Omitted categories: Maputo, Manufacturing, small firm size (for labor productivity and wages regressions), large firm size (for TFP). * significant at 10%; ** 5%; *** 1% Robust standard errors in brackets, data clustered by sample stratum. For the pool of countries, Table 9.10 presents the relationship between the investment climate indices and firm productivity as measured by sales and value 124 added per worker, TFP88 and wages. The first column shows that expenses with security and crime do take a toll on firms' productivity, measured as sales per worker, across the sample of countries pooled. Perception of poorer firm compliance in reporting total sales for tax purposes, the measure of informality used in this analysis, is more prevalent among more productive firms. Infrastructure and finance have similar correlations to firm productivity (sales per worker) for the larger sample, although the magnitudes of the correlations are lower when other countries are taken into account. When it comes to value added per worker and TFP, technology appears as a significant investment climate variable associated with enhanced firm performance, while access to finance and poor infrastructure lose significance. Informality and the governance indices are still significant and their correlations to TFP are larger than to productivity as measured by sales per worker. Using labor productivity for the restricted TFP sample, we observe that only the informality index is correlated with higher firm productivity. The last column shows the relationship between wages and the investment climate indices. None of the investment climate variables seem to be significantly correlated to wages paid by firms in the pooled sample. Note that foreign ownership is highly correlated to firm performance under all of the four specifications. Finally, Mozambique's performance gap in relation to the other countries compared in this analysis is evident from the country dummy coefficients: significant differences in performance are captured in relation to most countries and measures of productivity, as well as wages paid to workers. Table 0.10. Firm Performance and IC Variables--Mozambique and Comparator Countries Sales per Sales per Value Added TFP (One Worker (in Wages (in Worker (in per Worker Step) Logs) for TFP Logs) Logs) (in Logs) Sample (1) (2) (3) (4) (5) Poor Governance 0.236* 0.23 0.314* 0.23 0.178 (security & crime) [0.132] [0.153] [0.179] [0.199] [0.112] Bribes 0.036 0.221 0.184 0.197 0.217 [0.249] [0.229] [0.216] [0.240] [0.190] Informality 0.007** 0.007** 0.014*** 0.011*** 0.002 [0.003] [0.003] [0.004] [0.004] [0.003] Poor Infrastructure 0.162** 0.126 0.069 0.08 0.065 [0.081] [0.080] [0.071] [0.079] [0.061] Access to Finance 0.335** 0.028 0.07 0.076 0.031 [0.170] [0.174] [0.155] [0.172] [0.133] Better Technology 0.178 0.228* 0.311** 0.185 0.04 [0.144] [0.136] [0.133] [0.148] [0.114] Direct exports 10%+ of sales 0.056 0.104 0.05 0.118 0.089 [0.093] [0.089] [0.083] [0.092] [0.071] Foreign owned 10%+ 0.500*** 0.554*** 0.275*** 0.596*** 0.307*** [0.085] [0.087] [0.083] [0.092] [0.065] Firm age (in logs) 0.035 0.007 0.019 0.008 0 88 In the case of TFP, the dependent variable is value added per worker controlled for capital per worker. Since many firms do not report their stock of capital, we also estimate the model for sales per worker restricting the sample to the TFP one. 125 [0.047] [0.046] [0.042] [0.047] [0.036] Capital per worker (in logs) 0.306*** [0.018] Manager - secondary education 0.149 [0.095] Manager - more secondary 0.338*** [0.097] Textiles 0.564*** 0.420*** 0.503*** 0.830*** 0.049 [0.124] [0.119] [0.115] [0.128] [0.094] Garments 0.931*** 0.448*** 0.249** 0.987*** 0.064 [0.117] [0.109] [0.107] [0.118] [0.089] Metals & machinery 0.123 0.158 0.074 0.123 0.239*** [0.113] [0.107] [0.099] [0.110] [0.087] Chemicals 0.372*** 0.482*** 0.326*** 0.375*** 0.249** [0.138] [0.128] [0.122] [0.135] [0.106] Electronics 0.663*** 0.185 0.013 0.442* 0.121 [0.250] [0.246] [0.243] [0.269] [0.188] Nonmetallic products 0.109 0.158 0.027 0.1 0.172* [0.130] [0.122] [0.113] [0.125] [0.101] Other manufacturing 0.257*** 0.082 0.121 0.294*** 0.076 [0.088] [0.083] [0.080] [0.088] [0.068] Services 0.958*** 0.876 0 0 0.381*** [0.144] [0.889] [0.000] [0.000] [0.111] Medium (20-99 empl) 0.072 0.085 0.163 0.188* 0.122 [0.110] [0.107] [0.100] [0.110] [0.085] Large (100+ empl) 0.173 0.122 0.101 0.205 0.330** [0.184] [0.172] [0.168] [0.186] [0.140] Angola 0.701*** 0.880*** 0.696*** 0.825*** 0.973*** [0.169] [0.169] [0.168] [0.187] [0.130] Indonesia 0.101 0.620* 0.839** 1.169*** 0.06 [0.325] [0.351] [0.331] [0.366] [0.250] Malawi 0.727* 0.791** 0.726** 1.302*** 0.051 [0.377] [0.380] [0.343] [0.383] [0.288] South Africa 1.956*** 2.098*** 1.572*** 2.467*** 2.182*** [0.271] [0.264] [0.252] [0.290] [0.209] Vietnam 0.27 0.281 0.082 0.976** 0.478 [0.396] [0.411] [0.386] [0.426] [0.306] Zambia 0.654*** 0.909*** 0.410** 0.937*** 0.265** [0.132] [0.177] [0.177] [0.198] [0.105] Constant 9.387*** 8.165*** 6.592*** 8.875*** 3.965*** [0.392] [0.442] [0.461] [0.475] [0.314] Observations 2,461 2,154 1,895 1,895 2,466 Source: Authors' calculations using Enterprise Surveys 2003­08 Note: Dependent variables are expressed in logs, so that the reported coefficients can be interpreted as relative changes in respectively sales per worker, TFP, and wages. Micro firms and Thailand excluded from the analysis. Omitted categories: food industry, small firms, and Mozambique. Robust standard errors in brackets. * significant at 10%; ** 5%; *** 1%. 126 Summary of Results and Comparison Mozambique Versus Pooled Sample Comparing the results from Tables 9.9 and 9.10, we observe that while poor infrastructure and access to finance are all correlated to firm performance in Mozambique and in the comparator countries, the extent of this association seems to be greater for Mozambican firms. Therefore, for example, the ability to access finance separates to a larger extent more productive firms from less productive ones in Mozambique than in the other countries included in the sample. The technology index also shows a greater divide between more productive firms and less productive ones in Mozambique than in the other countries, where the index also loses statistical significance when productivity is measured as sales per worker. On the other hand, issues of crime and security, captured by the governance index, show a greater correlation with firm performance for the pooled sample of countries than in Mozambique. Exports and Exporter Premium As exports become more important for Mozambique, it is worthwhile estimating if firms that export perform better than firms that do not. The literature on this relationship between exports and firm growth has firmly established that the two are highly correlated, while the direction of causality between them remains less certain.89 Some studies point to evidence that exporting exposes firms to knowledge and competition that pressures them to become more productive. Larger foreign demand for products, particularly important for smaller economies, may also allow firms to exploit economies of scale, increasing their overall productivity. Other studies, however, point to evidence that in fact firms self-select into exporting; in other words, more productive firms are better able to meet the demands and costs of more sophisticated, foreign markets.90 Due to lack of adequate data, the exercise that is attempted here does not deal with this issue of causality, focusing instead on estimating if this exporter premium also exists in Mozambique and how it compares to other countries. Specifically, we estimate productivity and wage exporter premiums for Mozambique and its comparator countries, for the subsample of African countries, and for Mozambique alone. To do so, we run simple OLS regressions of sales and value added per worker, TFP, and average wages on variables capturing firms' age, size, foreign ownership, log of capital per worker (for TFP calculations), and country-region and industry dummies. The basic specifications employed follow closely the ones already used, with the difference that the IC variables are dropped: log(Yi / Li ) = 4 Expi + j 4 j C ij + v 4i log(VAi / Li ) = 5 Expi + j 5 j C ij + log( K i / Li ) + v5i 89 For a brief overview of the literature on this topic, see Carlos Casacuberta et al in Fajnzylber et al (2009), Chapter 7. 90 According to Casacuberta et al. (in Fajnzylber et al., 2009) the evidence for the self-selection hypotheses seems more robust. Costs in this case are related to costs incurred to reach foreign markets, such as transportation, distribution, marketing, and even production costs associated with changes needed to meet higher foreign standards. 127 log(Wagei / Li ) = 6 Expi + j 6 j C ij + v6i Results The estimated coefficients on the exporter dummy and their standard errors in each case are reported in Table 9.11. Table 0.11. Productivity and Wage Exporter Premium in Mozambique and Comparator Countries Sales per Value Added TFP (One Step) Wages (in Logs) Worker (in per Worker Logs) (in Logs) Mozambique and all 0.229*** 0.217*** 0.139*** 0.117*** comparator countries [0.045] [0.042] [0.037] [0.034] Mozambique and African 0.345*** 0.407*** 0.245*** 0.236*** comparator countries [0.090] [0.086] [0.070] [0.069] Mozambique 1.346*** 1.445*** 0.850** 0.671** [0.446] [0.453] [0.379] [0.340] Source: Authors' calculations using Enterprise Surveys 2003-08 Note: Each coefficient corresponds to the exporter dummy in a separate OLS regression. All regressions include as control variables firm-, region-, and industry-level characteristics (see text for more details). Standard errors are provided in brackets. * Significant at 10%; ** 5%; *** 1%. Results by and large confirm the idea that exporters perform "better." In fact, exporters in Mozambique have a labor productivity that is 135 percent higher than nonexporters when measured as sales per worker and 145 percent higher when measured as value added per worker. If measured using TFP, Mozambican exporters are 85 percent more productive than nonexporters and wages paid are also 67 percent higher for exporters. This difference in performance between exporters and nonexporters seems to be larger in Mozambique than in other comparator countries: the coefficients for both sets of pooled countries are smaller than for the regression with Mozambique alone. If we take the hypothesis that more productive firms self-select into exporting, these results may suggest that costs, taken broadly to mean barriers to reaching foreign markets, may be much higher in Mozambique than in the other countries. Only highly productive enterprises may be able to overcome these barriers. If we take the opposite hypothesis, of learning by exporting, these results show the potential benefits that an export-driven strategy can bring to the Mozambican economy. 128 Table 0.12. Summary Statistics for Investment Climate Variables % Sales Lost to Power % Consignment Lost in % Sales Spent in Security % Sales Lost Due to Crime Bribe Informality Outages Transit No With No With No With No With No With No With Country Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Angola Mean 1.72 1.72 0.68 0.68 0.44 0.44 52.00 51.75 3.90 3.92 1.10 1.10 Observations 537 540 540 540 535 540 528 540 519 540 540 540 Indonesia Mean 1.79 1.79 0.49 0.49 0.24 0.24 73.14 73.14 2.02 2.02 2.27 2.27 Observations 713 713 708 713 713 713 713 713 713 713 713 713 Malawi Mean 2.11 2.14 2.20 2.31 0.24 0.24 69.72 69.83 22.25 22.62 3.19 3.19 Observations 154 159 97 153 153 160 132 160 114 152 160 160 Mozambique Mean 1.14 1.14 2.25 2.25 0.19 0.19 45.05 45.06 1.59 1.60 1.38 1.38 Observations 599 599 598 599 495 598 597 599 592 599 599 599 South Africa Mean 0.89 0.89 0.61 0.62 0.02 0.02 90.78 90.78 0.92 0.93 0.83 0.83 Observations 603 603 595 603 529 601 568 601 303 567 603 603 Thailand Mean 0.58 0.58 0.26 0.25 1.51 1.51 0.53 0.53 Observations 1377 1385 1367 1385 1367 1385 1385 1385 Vietnam Mean 0.62 0.63 0.04 0.04 0.54 0.54 93.64 93.65 1.49 1.51 0.37 0.37 Observations 1137 1150 1136 1150 1135 1150 1114 1150 975 1150 1150 1150 Zambia Mean 1.31 1.31 1.19 1.19 0.16 0.16 44.13 44.78 2.84 2.84 2.02 2.02 Observations 603 603 603 603 519 601 484 601 585 602 603 603 Total Mean 1.06 1.06 0.67 0.68 0.30 0.30 70.83 70.44 2.40 2.46 1.12 1.12 Observations 5723 5752 5644 5746 4079 4363 4136 4364 5168 5708 5753 5753 129 Table 9.12 continued Line of Credit or Loan Managers with More than Use of Internet to % Financing from Banks from Financial Institution Offers Training ISO Certification Secondary Education Communicate with Clients No With No With No With No With No With No With Country Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Substitution Angola Mean 1.37 1.36 0.06 0.06 0.20 0.21 0.05 0.06 0.51 0.51 0.29 0.29 Observations 535 540 540 540 329 527 423 534 540 540 540 540 Indonesia Mean 15.67 15.67 1.75 1.75 0.24 0.24 0.22 0.22 0.67 0.67 0.50 0.51 Observations 713 713 713 713 596 709 696 713 713 713 711 713 Malawi Mean 17.37 17.37 1.36 1.36 0.52 0.52 0.18 0.19 0.73 0.73 0.68 0.68 Observations 160 160 160 160 155 160 148 160 160 160 159 160 Mozambique Mean 3.14 3.14 0.21 0.21 0.18 0.18 0.20 0.22 0.33 0.33 0.26 0.26 Observations 599 599 599 599 599 599 479 599 599 599 599 599 South Africa Mean 16.56 16.56 0.68 0.68 0.64 0.64 0.42 0.42 0.92 0.92 0.99 0.99 Observations 603 603 603 603 603 603 599 603 603 603 603 603 Thailand Mean 46.27 46.27 1.15 1.15 0.76 0.76 0.45 0.44 0.82 0.82 0.59 0.59 Observations 1385 1385 1385 1385 1385 1385 1183 1385 1385 1385 1385 1385 Vietnam Mean 27.65 27.65 1.60 1.60 0.60 0.60 0.38 0.40 0.81 0.81 0.66 0.66 Observations 1150 1150 1150 1150 1114 1150 925 1150 1150 1150 1150 1150 Zambia Mean 4.03 4.03 0.39 0.39 0.19 0.19 0.14 0.14 0.77 0.77 0.46 0.46 Observations 603 603 603 603 603 603 478 601 603 603 603 603 Total Mean 21.72 21.70 0.99 0.99 0.49 0.48 0.30 0.30 0.72 0.72 0.56 0.56 Observations 5748 5753 5753 5753 5384 5736 4931 5745 5753 5753 5750 5753 Source: Enterprise Survey 2008 130 Table 0.13. Summary Statistics for Control Variables Country Foreign-owned Exporter Log of Firm Age Angola Mean 0.14 0.01 1.87 Observations 540 540 540 Indonesia Mean 0.16 0.35 2.98 Observations 713 713 711 Malawi Mean 0.26 0.18 2.60 Observations 160 160 156 Mozambique Mean 0.17 0.04 2.26 Observations 599 599 597 South Africa Mean 0.19 0.38 3.08 Observations 603 603 602 Thailand Mean 0.26 0.49 2.81 Observations 1385 1385 1385 Vietnam Mean 0.12 0.34 2.33 Observations 1150 1150 1147 Zambia Mean 0.21 0.08 2.12 Observations 603 603 602 Total Mean 0.19 0.29 2.54 Observations 5,753 5,753 5,740 Source: Enterprise Survey 2008 131 Table 0.14. Partial Correlations between Performance, Investment Climate and Control Variables Informality Technology governance Infrastruct Added per Log Value Log Wage Log Sales Access to Exporter Foreign- Log Age Worker Worker Finance owned Bribes Better Poor Poor per ure Mozambique Log sales per worker 1 Log value added per worker 0.9698* 1 Log wage 0.7127* 0.7085* 1 Poor governance 0.0417 0.0291 0.0776* 1 Bribes 0.0407 0.1196* 0.0665 0.0403 1 Informality 0.0015 0.0699 0.0511 0.0251 0.0804* 1 Poor infrastructure 0.0292 0.0872* 0.0388 0.2101* 0.0003 0.0163 1 Access to finance 0.1516* 0.1301* 0.1293* 0.0007 0.0283 0.0849* 0.0228 1 Better technology 0.1772* 0.1670* 0.1490* 0.0786* 0.0172 0.0129 0.0009 0.2671* 1 Foreign-owned 0.2102* 0.1667* 0.2021* 0.0352 0.0518 0.1062* 0.0740* 0.1646* 0.2622* 1 Exporter 0.1534* 0.2334* 0.1412* 0.1330* 0.0631 0.0364 0.0157 0.1197* 0.3752* 0.1835* 1 Log age 0.1028* 0.1425* 0.0269 0.0112 0.0617 0.0404 0.0161 0.0153 0.0740* 0.0204 0.0401 1 Full Sample Log sales per worker 1 Log value added per worker 0.9074* 1 Log wage 0.6508* 0.7346* 1 Poor governance 0.1075* 0.0419* 0.0633* 1 Bribes 0.1030* 0.1552* 0.1527* 0.0173 1 Informality 0.0867* 0.0205 0.0616* 0.0907* 0.0171 1 Poor infrastructure 0.0308* 0.0086 0.0211 0.1522* 0.0302* 0.0608* 1 Access to finance 0.008 0.0746* 0.1439* 0.0995* 0.1350* 0.3761* 0.0386* 1 Better technology 0.2948* 0.2941* 0.2360* 0.0626* 0.0238 0.2606* 0.0482* 0.2464* 1 Foreign-owned 0.2113* 0.2369* 0.1497* 0.0162 0.0405* 0.0093 0.0122 0.0151 0.1829* 1 Exporter 0.1683* 0.1681* 0.1170* 0.0621* 0.0115 0.2082* 0.0544* 0.2993* 0.3666* 0.2335* 1 Log age 0.1416* 0.1731* 0.1573* 0.0494* 0.1289* 0.1997* 0.0352* 0.2385* 0.2703* 0.0002 0.1784* 1 Source: Enterprise Survey 2008 132 Figures and Table for Labor Productivity and Employment Analysis Figure 0.1. Labor Productivity Decomposition--Olley and Pakes (by firm size) Small Small 2003 2006 1.5 1 1 .8 .6 .5 .4 .2 0 0 Construction & Transport Manufacturing Services Construction & Transport Manufacturing Services Unweighted Mean Alloc Efficiency Unweighted Mean Alloc Efficiency Medium Medium 2003 2 1.5 1 2006 1 .8 .6 .5 .4 .2 0 0 Manufacturing Services Manufacturing Services Unweighted Mean Alloc Efficiency Unweighted Mean Alloc Efficiency Large Large 2003 2006 1.5 1 1 .2 .4 .6 .8 .5 0 0 -.5 Construction & Transport Manufacturing Services Construction & Transport Manufacturing Services Unweighted Mean Alloc Efficiency Unweighted Mean Alloc Efficiency Source: Enterprise Survey 2008 133 Figure 0.2. Histograms of Log Total Factor Productivity (TFP) in Manufacturing (2006) Manufacturing - Small Firms 0 .1.2 .3.4.5 Density 0 2 4 6 8 10 lntfp1 Manufacturing - Medium Firms 0 .1 .2 .3 .4 Density 0 2 4 6 8 10 lntfp1 Manufacturing - Large Firms 0 .2 .4 .6 .8 Density 0 2 4 6 8 10 lntfp1 Source: Enterprise Survey 2008 134 Table 0.15. Partial Correlations (Multiple Linear Regression) Employment Performance Measure 2003­2006 Binary Employment Symmetric Employment Growth (1 for Growth Rate Positive Growth) Partial p- Partial p- Variables Correlation value Correlation value Industry Construction and transport (dummy) (omitted) .195 .244 .028 .876 Manufacturing .081 .626 .077 .668 Services Region Maputo (omitted) (dummy) Matola .060 .404 .010 .867 Beira .029 .662 .023 .715 Nampula .025 .728 .136*** .092 Size Large (omitted) (dummy) Medium .104 .371 .209*** 0.07 Micro .510* .000 .417* .003 Small .040 .738 .278** .028 Age Log of age .135* .000 .145* .000 Legal status Private held, limited company (dummy) (omitted) Sole proprietorship .173* .003 .128* .010 Partnership .165** .048 .122** .086 Limited partnership .170** .028 .057 .350 Export zone No (omitted) (dummy) Yes .068 .232 .022 .575 Part of larger firm No (omitted) (dummy) Yes .272* .000 .157** .026 Export directly No (omitted) (dummy) Yes .017 .870 .028 .785 Minimum wage Percentage of full-time workers paid minimum wage .001** .016 .0006 .174 Time tax Percentage of management time spent on government regulations .006** .043 .003 .444 Purchase fixed No (omitted) assets in 2006 Yes .124* .005 .083** .034 (dummy) Court is fair Strongly disagree (dummy) Otherwise (omitted) .081*** .059 .093*** .015 Number observations 506 513 R2 0.21 0.15 F test p-value .000 .000 Source: Author's calculations based on Enterprise Survey 2008 Note: * Significant at less than 1%; ** significant at less than 5%; *** significant at less than 10%; constant term was estimated but it is not reported; symmetric growth rate is a growth rate that takes as base the average of employment in the period. 135 ANNEX V: GENDER AND ENTREPRENEURSHIP IN MOZAMBIQUE The 2007 Global Monitoring Report (World Bank 2007) stresses the importance of gender equality and women's empowerment to promote economic development and attain the Millennium Development Goals. Higher numbers of entrepreneurs and firms increase diversity, and accordingly, economic performance in a region or country (Cohen and Malerba, 2001). Therefore, greater female participation in entrepreneurship and the workforce has the potential to increase numbers and diversity of firms. However, research has shown that women entrepreneurs face a multitude of institutional, sociocultural, and technical constraints when trying to start or grow their own enterprises. Female-owned businesses in developing countries have been found to underperform in comparison to their male counterparts in a multitude of studies. However, data from the 2007 Enterprise Survey in Mozambique suggests that female-owned firms are as productive as their male counterparts. Nevertheless, only 22 percent (86 firms) of Mozambican sole ownership firms surveyed are owned by women. Most female-owned firms were small in terms of the number of employees. About 75 percent of female-owned firms had between 5­19 employees. The same is the case for male-owned Mozambican firms. Whereas, there were no single-ownership, female-owned large firms (100 employees and more) there were at least seven such firms owned by men. Table 0.16. Size of Female-Owned Firms Male-owned Firms Female-owned Firms Observations Percent Observations Percent Microenterprises (less than 5 employees) 72 22.3 27 30.3 Small (5­19 employees) 186 57.6 44 49.4 Medium (20­99 employees) 57 17.6 17 19.1 Large (100 employees and more) 8 2.5 1 1.1 Total 323 100.0 89 100.0 Source: Enterprise Survey 2008 Gender Differences in Firm Productivity The data from the Enterprise Survey suggests that there is no significant difference in performance between male- and female-owned firms.91 In neither of the three different measures of firm productivity used--number of sales per worker (labor productivity), the levels of sales per fixed assets (capital productivity) and profit rates-- are female-owned firms found to underperform in relation to male-owned ones.92 Table 0.17. Performance in Male- and Female-Owned Firms Labor Capital Profit Rates Productivity Productivity 91 In this section the analysis examines a subset of firms, namely firms that are owned by a single owner. Because currently there is no question in the Enterprise Surveys that asks about the gender of the top manager, the sample is restricted by assuming that if the sole owner is a woman, there is a high likelihood that she is also the top manager. 92 Note that this conclusion applies to firms in the Enterprise Survey only and may be the result of a small sample size and large standard errors. 136 Productivity Productivity (Log (Log Sales (Log Sales/Fixed Profits/Fixed per Worker) Assets) Assets) Female owned 0.249 0.527 0.527 (0.252) (0.472) (0.472) Services sector 0.291 (0.479) Garments sector 0.302 0.199 0.291 (0.258) (0.463) (0.479) Food sector 0.379 0.277 0.199 (0.254) (0.381) (0.463) Capital intensity 0.034 (0.050) Secondary 0.577** 0.278 0.277 (0.205) (0.670) (0.381) Wage rate 1.14** (0.114) reg2 0.067 0.663 0.278 (0.357) (0.731) (0.670) reg3 0.336 0.631 0.663 (0.396) (0.564) (0.731) reg4 0.233 0.527 0.631 (0.307) (0.472) (0.564) Constant 0.428 1.758** 1.758** (1.157) (0.302) (0.302) Observations 191 191 191 R-squared 0.44 0.03 0.03 Source: Enterprise Survey 2008 Note: Standard errors in parentheses * significant at 5%; ** significant at 1% Using the same technique, analysis for Tanzania yielded similar results: no difference in performance was found between male- and female-owned firms. This was not the case, however, in Zambia and South Africa, where results from the Enterprise Survey data indicated that in Zambia female-owned firms underperform in terms of both labor productivity and profit rates, while in South Africa female-owned firms underperform in terms of labor productivity.93 Gender Differences in Business Climate Constraints If Mozambican firms owned by women seem to have similar levels of performance to those owned by men, one possible explanation for the overwhelming prevalence of male-owned firms is that women face more severe constraints than men when trying 93 Zambia Investment Climate Assessment (World Bank, forthcoming). 137 to start a business or operate one.94 We examine this proposition by analyzing results from the Enterprise Survey. Specifically, we would want to see a gender differential on the percentages of owners reporting various aspects of the investment climate as major or severe constraints to their operations. Namely, perceptions about business constraints would reflect actual differences in constraints faced by female versus male firm owners. The data, however, suggest that female-owned firms do not seem to face larger business climate constraints than male-owned firms (see Figure 9.3). In none of the top five business climate constraints--competition from the informal sector, access to finance, crime, theft and disorder, tax rates and corruption--is there a significant difference in terms of the percentage of female- versus male-owned firms that rate these as serious constraints to the operations of their business. Figure 0.3. Male- and Female-Owned Firms Rating Investment Climate as Major or Very Severe Constraint Informal Sector Access to Finance Crime, Theft and Disorder Tax rates Corruption Electricity Customs and Trade Regulations Tax administration Business Licensing and Permits Workforce Education Access to Land Political instability Labor Regulations Courts 0 10 20 30 40 50 60 Serious Constraint (% of firms) Female Male Source: Enterprise Survey 2008 Thus, it seems that from the perception results of the Enterprise Survey, challenges facing women's businesses, such as financing, tax rates, corruption, and competition from the informal sector, are common to most SMEs. However, it is important to highlight some aspects of the investment climate that are known to constraint female entrepreneurs to a larger degree than male entrepreneurs in most developing countries. The fact that the data gathered in the Enterprise Surveys did not point to a difference in constraints faced by women and men when doing business does not mean that restrictions to women's participation are not at work. Evidence from other countries suggests that women's participation as entrepreneurs may be hindered through a variety of mechanisms. We focus here on access to finance and issues of crime and security. 94 Legal aspects, such as the Family Law that required the husband's permission to take on a job, were reformed in 2005. Women now also can buy, own, and manage property. 138 Access to Finance Access to finance is necessary for entrepreneurial activity and stood out as the second biggest concern for both female- and male-owned firms. Out of all sole- ownership, female-owned firms, 37 percent considered access to finance to present a serious constraint to the operations of their business. In comparison, in Mauritius only 13 percent of firms considered access to finance a serious constraint, but in Kenya the number reached 55 percent (see Figure 9.4). Figure 0.4. Access to Finance Is Perceived to Be a Serious Constraint 60 Serious Constraint (% of firms) 50 40 30 20 10 0 Kenya Tanzania Botswana Mozambique South Africa Zambia Namibia Mauritius Access to Finance Female Male Source: Enterprise Surveys, World Bank Many studies have suggested limited access to finance to be a major constraint to female-owned businesses across many countries. Coleman (2007) finds that among small business owners in the United States, female-owned firms were significantly less likely to have loans from banks or lines of credit. In the United Kingdom, Rosa, Carter, and Hamilton (1996) find that the average female-owned firm uses a third less external finance than the average male-owned firm. Specific barriers for female entrepreneurs to obtain the capital available in the market may limit the opportunities for female-owned firms to expand and use optimal quantities of capital and labor. Female entrepreneurs are therefore more likely to start smaller businesses in services and the informal sector, which on average will be less productive. In Mozambique, 43 percent of female-owned businesses are located in the retail sector, which has low startup costs. They are much less likely to be located in the manufacturing sector, which requires a larger starting capital (see Table 9.18). Table 0.18. Female-Owned Firms Are More Likely to Be Located in the Retail Sector Female Male Sector Frequency Percent Frequency Percent Other manufacturing 5 5.81 71 23.36 139 Food 18 20.93 46 15.13 Garments 11 12.79 38 12.5 Chemicals .. .. 2 0.66 Plastics and rubber .. .. 1 0.33 Nonmetallic mineral products 2 2.33 3 0.99 Fabricated metal products 7 8.14 53 17.43 Machinery and equipment .. .. 2 0.66 Other services 3 3.49 9 2.96 Retail 37 43.02 79 25.99 Hotels and restaurants 3 3.49 .. .. Total 86 100 304 100 Source: Enterprise Survey 2008 .. Negligible High collateral requirements in Mozambique make it difficult for firms to access capital. Possibly for this reason, for both current and fixed assets, the Enterprise Survey data indicates there to be a near-total lack of financing from financial institutions: among both male- and female-owned firms over 80 percent of financing comes from internal funds or retained earnings (see Table 9.19). Table 0.19. Most Financing Comes from Internal Funds or Retained Earnings Proportion of Financing from Different Sources Current Assets (%) Purchases of Fixed Assets (%) Financing source Female Male Female Male Internal funds/retained earnings 81.0 82.7 94.2 88.1 Borrowed from financial institutions 5.4 2.5 0.8 5.6 Trade credit 11.7 14.2 5.0 6.4 Source: Enterprise Survey 2008 140 With access to financing being the key issue for women in a collateral-based banking system, improving access to microcredit and other forms of financing could contribute significantly to increasing the number of women entrepreneurs. In line with this during the year 2008, the International Finance Corporation's (IFC) gender program and Small and Medium Sized Enterprise Development Initiative Maputo anticipated undertaking a diagnostic study on access to finance for women entrepreneurs in Mozambique. This study was intended to be undertaken in cooperation with BCI Fomento Bank in Mozambique and aimed at increasing the World Bank's lending to female entrepreneurs (IFC, 2008). The upcoming Mozambique Competitiveness Project will include a component on matching grants for entrepreneurs to upgrade their business operations and marketing efforts as well as an access to finance component with a guarantee scheme and capacity building for accounting and auditing standards for SMEs. The project will, furthermore, monitor the participation and performance of female- owned and female-managed enterprises in all project components. Crime, Theft, and Disorder Security services to avoid crime and theft, and losses due to theft, increase firm operating costs. The impact of crime, theft, and disorder was the concern that the third largest number of female-owned firms, 30 percent, reported to represent a serious constraint to the operation of their business. In comparison, more South African and Kenyan firms, but less Zambian and Mauritian firms, reported this to represent a serious constraint (see Figure 9.5). Figure 0.5. Crime, Theft, and Disorder as a Constraint to Business 60 Serious Constraint (% of firms) 50 40 30 20 10 0 South Africa Kenya Namibia Botswana Mozambique Tanzania Zambia Mauritius Crime, Theft and Disorder Female Male Source: Enterprise Surveys (various years) Security costs as a percentage of total sales were high for both male- and female- owned firms (Table 9.20). Also, over 30 percent of firms experienced losses due to theft, robbery, vandalism, or arson. Table 0.20 Security Costs and Losses Due to Theft Male-owned Cost Female-owned Firms Firms 141 Firms Employ security services (% of total) 24.70 25.70 Security costs (% of sales) 3.31 3.82 Losses due to theft (% of total) 31.46 33.75 Source: Enterprise Survey 2008 Gender Differences in Labor Market Participation This section uses firm-level data provided by the surveyed manufacturing firms to compare male and female labor market participation across different industries. Overall, the percentage of total employees that are female amongst all the firms surveyed, at 15 percent, is very low. In only 9 percent of firms observed, more than 50 percent of the total workforce is female. There are even less female production workers, 12 percent, but at 37 percent, comparatively more female nonproduction workers (Table 9.21). Table 0.21. Female Workers in Manufacturing Firms Number of Firms Mean Female employees as a percentage of total employees 341 15.1 (%) Female production workers as a percentage of total production workers 341 12.1 (%) Female nonproduction workers as a percentage of total nonproduction workers 256 36.6 (%) Source: Enterprise Survey 2008 Note: This data is for manufacturing firms only because this question was not asked of nonmanufacturing firms. Breaking the manufacturing firm data down into different sectors suggests that among the top four sectors surveyed, the number of female employees as a percentage of total employees varies significantly between sectors, ranging from 5 percent in the fabricated metal products sector to 34 percent in the garment sector. Average growth in firm employment size between 2003 and 2006 ranged from one additional worker in the garment and other manufacturing sectors to six additional workers in the food sector. The garment sector, which employs the most female workers, has the lowest average monthly compensation per employee (Table 9.22). Similarly, average monthly compensation in firms with more than 30 percent female employees, at 2,115 metical is significantly lower than average monthly compensation, 3,180 metical, in those firms where less than 30 percent of employees are female. Table 0.22. Cross-Sector Comparisons: Number of Female Employees and Average Monthly Compensation Female Employees Average Monthly Number of as a Percentage of Compensation per Firms Total Employees Employee (Metical) Food 89 18.6 2371 Garments 52 34.2 1839 Fabricated metal products 79 5.2 2081 142 Other manufacturing 87 9.9 1988 Source: Enterprise Survey 2008 Also, the percentage of the total permanent workforce that is paid the minimum wage is the lowest in the garments sector. However, expenditure on programs and activities to prevent Human Immunodeficiency Virus/Acquired Immunodeficiency Syndrome (HIV/AIDS) among employees was the second highest in the garments sector (Table 9.23). Table 0.23. Full-Time Permanent Workforce Paid Minimum Wage and Expenditures on HIV/AIDS Programs and Activities Number of % Paid Expenditures on Firms Minimum Wage HIV/AIDS Food 96 49.2 1994.3 Garments 57 44.8 564.9 Fabricated metal products 80 55.8 313.8 Other manufacturing 91 52.9 441.8 Source: Enterprise Survey 2008 This lower compensation in firms, in which more than 30 percent of the workforce is female compared to that in firms where less than 30 percent of the workforce is female, may partly be due to lower levels of average educational attainment of a typical production worker in the firms with higher female worker concentration (Table 9.24). However, maybe because of initially lower levels of education of the female workforce,95 we notice that the percentage of firms that run formal training programs for permanent, full-time employees is, at 27.6 percent, higher in those firms where 30 percent or more of the workforce is female than in those firms where this is not the case, where only 18.4 percent of firms run training programs. Table 0.24. Educational Attainment in Firms with Larger Versus Smaller Female Workforce Firms with Female Workforce Firms with Female Workforce under 30% above 30% Number of Percentage Number of Percentage Firms Firms 0­3 years of education 15 5.30 5 8.62 4­6 years of education 105 37.10 24 41.38 7­9 years of education 117 41.34 23 39.66 10­12 years of education 42 14.84 6 10.34 13 years and above of education 4 1.41 Total 283 100.00 58 100.00 Source: Enterprise Survey 2008 The analysis also shows that absenteeism due to sickness, both because of worker sickness or because the worker needs to care for family members or friends, in both 95 Mozambique's economic growth was held back by low female education levels (for example gross female tertiary education enrolment was only 1 percent in 2005, see World Bank Data Development Platform). The increase of educated women in the labor force explains 11 percent of the GDP growth, compared to 17 percent attributed to the increased supply of educated men (World Bank, 2008b, p. 49). 143 categories of firms is very high. Although absenteeism for HIV/AIDS96 is lower than for general sickness, it is interesting to note that absenteeism due to taking care of family or friends because of sickness or HIV/AIDS is more of an issue in those firms where more than 30 percent of the workforce is female--40 percent of firms with a female workforce of 30 percent and higher compared to 28 percent of firms (Table 9.25). This suggests that part of the lower formal female labor participation rate may be due to women staying at home to take care of sick family members or friends. Table 0.25. Absenteeism Due to Sickness and HIV/AIDS Under 30% of 30% or More of Workforce Workforce Female Female (283 Firms) (58 Firms) Absenteeism due to worker sickness 36.7 37.9 Absenteeism due to worker's family or friends sickness 22.6 29.3 Absenteeism among workers due to HIV/AIDS 3.5 3.4 Absenteeism among workers due to need to take care of family members or friends with HIV/AIDS 4.9 10.3 Source: Enterprise Survey 2008 Conclusions The Enterprise Survey did not point to specific differences in performance or in business environment constraints faced by female- versus male-owned firms in Mozambique. Lower monthly compensation per employee and a higher percentage of firms reporting high absenteeism due to caretaking of family or friends are observed in firms with higher female participation. Also, there are significantly less female-owned firms, and fewer women participate in the labor markets. Nonetheless, the results may suggest interesting questions that, while beyond the scope of this ICA, merit future inquiries. Among them is the fact that female-owned firms, while seemingly as productive as male-owned firms, represent a lower number of firms and are concentrated on the retail sector. What exactly drives these outcomes and the consequences of such concentration in the retail sector may be interesting topics for future research. 96 The prevalence rate for Mozambique 2004 was 16.4 percent (World Bank, 2008b, p. 101). 144