WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? ENTERPRISE SURVEY LESSONS FROM THE WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? ENTERPRISE SURVEY LESSONS FROM THE ii WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? Contributors What’s Holding Back the Private Sector in MENA? Lessons from the Enterprise Survey has been jointly produced by the European Bank for Reconstruction and Development (EBRD), the European Investment Bank (EIB), and the World Bank (WB). The team was led by, in alphabeti- cal order, Pedro de Lima (EIB), Debora Revoltella (EIB), Jorge Luis Rodriguez Meza (WB), and Helena Schweiger (EBRD). Chapter contributors are as follows: Chapter 1: Tim Bending (EIB) and Pedro de Lima (EIB). Chapter 2: Mohammad Amin (WB, chapter leader), Tim Bending (EIB), David C. Francis (WB), Asif M. Islam (WB), Arvind Jain (WB), and Jorge Rodriguez Meza (WB). Chapter 3: Frank Betz (EIB, chapter leader), Ralph de Haas (EBRD), Farshad R. Ravasan (Paris School of Economics and Université Paris 1 Panthéon-Sorbonne), Asif M. Islam (WB), and Jorge Rodriguez Meza (WB). Chapter 4: Mohammad Amin (WB, chapter leader), Frank Betz (EIB), David C. Francis (WB), Asif M. Islam (WB), Valeria Perotti (WB), Jorge Rodriguez Meza (WB), and Christoph Weiss (EIB). Chapter 5: Helena Schweiger (EBRD, chapter leader), David C. Francis (WB), Alexander Stepanov (EBRD), and Sarah Stölting (EIB). Economy fiches: Frank Betz (EIB), David C. Francis (WB), Arvind Jain (WB), Helena Schweiger (EBRD), and Sarah Stölting (EIB). LESSONS FROM THE ENTERPRISE SURVEY iii Acknowledgments The authors are grateful to the peer reviewers Mark Allen (Fellow at CASE – Center for Social and Economic Research), Ahmed Galal (Managing Director of the Economic Research Forum and Chairman of the Board of Forum Euroméditerranéen des Instituts des Sciences Économiques) and Jan Svejnar (James T. Shotwell Professor of Global Political Economy at Columbia University). Caroline Freund (Peterson Institute for International Economics) provided comments to the Concept Note. The report benefited from comments and feedback from Husam Abudagga (WB), Simon Bell (WB), Fabrizio Coricelli (Paris School of Economics), Shanta Devarajan (WB), Rand Fakhoury (EBRD), Carolin Geginat (WB), Sergei Guriev (Sciences Po), Fadel Jaoui (EBRD), Augusto Lopez-Claros (WB), Hanan Morsy (EBRD), Peter Mousley (WB), Tarek Osman (EBRD), Alexander Plekhanov (EBRD), and Aminur Rahman (WB). Tim Bending (EIB) and Romesh Vaitilingam provided editorial support. Most of the material in this report is based on the EBRD-EIB-WB MENA Enterprise Survey (MENA ES), which was funded by the World Bank, European Investment Bank, and the EBRD (Southern and Eastern Mediterranean cooperation funds account). This funding is gratefully acknowledged. iv Table of Contents Foreword . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi Executive summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Lessons from the MENA Enterprise Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Firm productivity and the business environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Access to finance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Jobs and skills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Competitiveness: trade, innovation, and management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Chapter 1: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 The MENA Enterprise Survey . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 The MENA ES region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 The focus of this report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Appendix A1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Chapter 2: Firm productivity and the business environment . . . . . . . . . . . . . . . . . . . 15 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Firm productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Box 2.1: Estimating total factor productivity with survey data . . . . . . . . . . . . . . . . . . . . . . . 16 The business environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 What are the main obstacles perceived by firms? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Political instability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Corruption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Unreliable electricity supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 The business environment experiences of large and small firms . . . . . . . . . . . . . . . . . . . . . . . . 27 Box 2.2: Political instability and electricity supply . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Policy conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Appendix A2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 Chapter 3: Access to finance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 The context: financial sectors in the MENA region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Firms in the MENA region are not typically credit-constrained, but many are disconnected . . . . . . . . . . 39 Box 3.1: A measure of credit constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 What explains the “disconnect” between firms and the banking sector and what are the consequences? . . 43 Box 3.2: The case of collateral practices for employment growth . . . . . . . . . . . . . . . . . . . . . . . . 47 Banking sector competition and firm access to credit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Policy conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Appendix A3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 v Chapter 4: Jobs and skills in the formal private sector . . . . . . . . . . . . . . . . . . . . . . . 59 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Employment in the formal private sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 The formal private sector’s contribution to women’s employment . . . . . . . . . . . . . . . . . . . . . . . . 63 The formal private sector’s contribution to youth employment . . . . . . . . . . . . . . . . . . . . . . . . . 67 Employment dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 Box 4.1: Comparing ES transition matrix data with census findings from Tunisia that include information on rates of firm exit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 Skills, training, and employment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 The wage bill per worker in the formal private sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Policy conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74 Appendix A4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Trade participation and competitiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 Increasing firm productivity through innovation and better management . . . . . . . . . . . . . . . . . . . . 92 Box 5.1: Estimating the impact of innovation on labor productivity . . . . . . . . . . . . . . . . . . . . . . . 93 Box 5.2: Types of firm-level innovation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 Box 5.3: Management practices in the MENA region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 Policy conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Appendix A5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112 Enterprise Survey–Economy Fiches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115 Djibouti . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 Arab Republic of Egypt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 Jordan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 Lebanon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122 Morocco . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124 Tunisia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126 West Bank and Gaza . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128 Republic of Yemen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130 Glossary of terms and acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 All ES economies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 vi WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Foreword I n the Middle East and North Africa (MENA), the development of a vibrant private sector is essential to fire up the engines of economic growth and that, in turn, is necessary to meet the needs and aspirations of the people in the region. The formal private sector represents a relatively small part of these economies; nonetheless it has the potential to become a powerful driver of job creation and rising living standards in the region. Creating an environment that is con- ducive to private sector development depends on a detailed understanding of the key determinants of firm performance and the major challenges that firms face, and also the role of government in providing the right business environment. This is why three of the leading international institutions active in the MENA region, the European Bank for Reconstruction and Development (EBRD), the European Investment Bank (EIB), and the World Bank Group (WBG) have joined forces to produce this report. It presents the results of the MENA Enterprise Surveys (MENA ES) conducted during 2013-2014 in eight economies: Djibouti, the Arab Republic of Egypt, Jordan, Lebanon, Morocco, Tunisia, the West Bank and Gaza, and the Republic of Yemen. By analysing detailed information on more than 6,000 private firms in the manufacturing and services sectors, the report provides fine-grained insights into their performance and the business environment in which they operate. Firms in MENA face many distorting incentives and barriers to competition. On the financing side, many appear to be not so much constrained by financing conditions as completely disconnected from the financial sector, thereby forgoing opportunities for growth. Workforce skills are another constraint, with a need for the re-orientation of education and training, so that workers have greater workplace skill and are prepared for the modern work environment. Enhancing the productivity of firms in the region also requires greater openness to international trade, something which will support innovation by facilitating the acquisition of knowledge about new products and processes. In 2015, the EIB, the EBRD and the World Bank Group together provided more than USD 7 .7 billion in financing for development in MENA. Looking forward, we remain committed to supporting private sector development in the region, each institution according to its strategy and remit, and in partnership with local authorities and stakeholders. Hans Peter Lankes Debora Revoltella Kaushik Basu MD, Acting Chief Economist Director Economics Department Sr. VP and Chief Economist EBRD EIB WB 1 Executive summary O ver the last few years, the Middle East and Enhancing the prospects for more inclusive North Africa (MENA) region has witnessed growth—with accessible opportunities for sustain- unprecedented transformation. In the Arab able employment, particularly for young people Uprisings, thousands of young people took to the and women—is vital to raise living standards, to streets to voice their frustration with the lack of underpin stability, and to offer an alternative to economic and social opportunities. These events economic migration out of the region. There is reflected demands for improvements in living con- an overwhelming consensus among economists ditions, infrastructure, job quality, education, and that the development of a vibrant private sector healthcare services, as well as better governance. is essential for delivering that growth. Creating an environment that is conducive to private sector The Arab Uprisings were a response to the failure development depends on a detailed understanding of the region’s economic models to satisfy people’s of the key drivers of private firms’ performance and needs and expectations. These models typically fea- the major challenges of the business environment in tured strong protectionism, lack of integration into which they operate. international markets, misguided state intervention, and inadequate support for a business environment that fosters innovation, entrepreneurship, and good management practices. 2 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Lessons from the MENA Enterprise Firm productivity and the business Survey environment This report is an assessment of the constraints on private sector development, which has been jointly conducted While the formal private sector represents a small by the three leading international institutions active in part of the MENA ES economies, it has the potential the MENA region. The report presents the results of to become the driver of a more sustainable model of the MENA Enterprise Survey (MENA ES) conducted growth in 2013 and 2014 in eight middle-income economies in Firms in the MENA ES have comparatively higher labor the region: Djibouti, the Arab Republic of Egypt, Jordan, productivity than their middle-income peer economies Lebanon, Morocco, Tunisia, the West Bank and Gaza, and outside the region; yet following global financial turmoil the Republic of Yemen. Implemented and co-financed by and the Arab Uprisings, their labor productivity has been the European Bank for Reconstruction and Development declining. Moreover, higher labor productivity belies lower (EBRD), the European Investment Bank (EIB), and the total factor productivity (TFP), in part due to relatively high World Bank Group (WBG), the MENA ES provides data use of capital. on a representative sample of the formal private sector. Large firms, which provide the majority of formal private Covering more than 6,000 private firms in the manufactur- jobs in the MENA ES, tend to be more efficient, but their ing and services sectors, the MENA ES includes data on activities are skewed toward more capital-intensive pro- the experiences of firms with a broad range of dimensions duction. In general private sector firms are typically small, of the business environment, including access to finance, old, and faced with limited growth opportunities. corruption, infrastructure, crime, and competition. The surveys also provide information on firm characteristics On the positive side, economic fundamentals seem to be and the cost of labor and other inputs; workforce com- at work in the formal private sector. For example, it is the position and women’s participation in the labor market; more productive firms that are the most likely to grow. trade, innovation, and management practices. After taking size into account, the most productive firms also have higher wage bills and greater access to finance. This unique set of information is an extremely valuable Encouragingly, these positive relationships reveal that in complement to the macroeconomic data most commonly certain areas at least, market forces are working as might used by researchers. Firm-level data permit fine-grained be expected. Policies should allow these forces to operate analysis of the drivers of firm performance, disaggregat- more efficiently. ing effects by key firm characteristics, such as their size, their sector, their inputs and output, and their involvement in innovation and international trade. The data also provide Addressing key constraints in the business a window on how managers and CEOs themselves environment is vital to help the private sector grow perceive the challenges and opportunities that they face. Addressing some of the key concerns of firms about the While the region is far from homogeneous, with manag- environment in which they operate is a way to unlock their ers reporting widely different experiences, analysis of the transformative potential. In the MENA ES region, four data helps to provide a basis for sound policies for private particular areas of concern stand out: political instability, sector development. corruption, unreliable electricity supply, and inadequate access to finance. Executive summary 3 Political instability is the leading concern for firms in technologies, linked with a lack of incentives for invest- most of the region, and it has a negative impact on ment in critical infrastructure, while creating room for sales and productivity growth vested interests. As part of the reform program in recent years, various international institutions, including the IMF Reflecting the effects of the Arab Uprisings, unresolved and the World Bank, have called for a comprehensive social tensions, and conflicts in the wider region, politi- reform of subsidies to open the way to a more efficient cal instability stands out as the greatest concern of firm energy sector. managers and CEOs in Egypt, Lebanon, Tunisia, the West Bank and Gaza, and the Republic of Yemen. In most of these economies, political instability seems to have nega- Inefficiencies in the business environment are tively affected firm and productivity growth. felt disproportionately by small and medium-sized enterprises High perceived levels of corruption are associated While several elements of the business environment— with lower growth of sales and employment, as well as notably political instability, unreliable electricity supply, lower labor productivity and inadequate access to finance—are widely reported as constraints on firms, inefficiencies stemming from these Corruption stands out as a key concern of firm managers factors have a more negative impact on smaller firms. and CEOs. High perceived corruption is associated with SMEs are more likely than large firms to report these lower sales, employment growth, and labor productiv- three elements as major obstacles to their operations, ity. There is also evidence that corruption deters firms’ though they are no more likely to report corruption as a interactions with public authorities, preventing them from major obstacle. making full use of available opportunities. In addition, concerns about corruption seem to go beyond petty cor- ruption, possibly indicating deeper problems in the econo- Access to finance mies concerned, such as state capture by interest groups or elites, corruption at high levels, or even under-reporting Financial and banking sectors in the region are for fear of potentially adverse consequences. relatively large, but credit is mostly channeled to a small number of large firms An unreliable electricity supply is a serious obstacle The financial sector of the MENA ES economies is for firms in several economies dominated by a relatively large banking sector, with loans- Unreliable electricity supply remains a significant problem to-GDP ratios above the standards in peer economies. for firms in Egypt, Lebanon, the West Bank and Gaza, and Bank lending is highly concentrated, however, with credit the Republic of Yemen despite efforts by some govern- targeting only a limited number of large companies, leav- ments to tackle this problem. An irregular power supply ing the bulk of firms with little or no access to credit. accounts for a significant loss of sales for many firms, and is associated with lower productivity levels. MENA ES firms finance their operations and investments in a similar way to firms in peer The relevance of poor electricity access as a constraint on economies firm growth should be read in the context of the overall institutional framework that characterizes the energy There is considerable variation in the use of internal funds sector in the region. Many economies have used energy to finance operations and investments across the region. subsidies as a safety net when their systems of social The use of bank credit and credit from suppliers and welfare have proved inadequate or ineffective. But this customers is in line with peer economies. Equity finance is costly and, by distorting prices, there have been sys- plays a negligible role in the region, while other sources tematic incentives to move toward more capital-intensive of finance, including microfinance, are only significant in Tunisia and the West Bank and Gaza. 4 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY A large share of firms are not credit-constrained Regardless of age, firms are less likely to disconnect from the banking sector and more likely to create new The MENA ES economies have a smaller share of credit- jobs if banks accept movable assets as collateral. Since constrained firms than other regions of the world. But a large share of firms’ assets consists of machinery and this is not driven by successful loan applications; instead, equipment, banks’ willingness to accept movable assets many firms report that they have enough capital and thus as collateral can be considered a business-friendly collat- do not need a loan. eral standard. This suggests a potential link between the adoption of business-friendly collateral standards and the There is a notable disconnect between firms and banks potential for job creation. in the region A significant share of firms that are not credit-constrained Jobs and skills have disconnected from the banking sector altogether. Compared with firms that have encountered difficulties Compared with other regions, formal private sector obtaining credit, the disconnected firms are more likely to employment is concentrated in manufacturing and be small, less likely to have audited financial reports, and exporting firms; but employment of women is low; and less likely to use the banking system even for payments. youth employment is strongest in young innovative firms Disconnected firms resemble credit-constrained firms, as they both have low propensity to invest and are less likely The structure of employment in the region’s formal private to plan for expansion, even when capacity constraints are sector is in many ways similar to comparable economies binding. The difference is that disconnected firms seem elsewhere, although the manufacturing sector and ex- content with their current situation and do not complain porting firms play a comparatively larger role in providing about access to finance. employment, with the retail sector lagging behind. The business cycle alone cannot account for this pattern The employment of women in a typical firm is much as a downturn may prompt firms to seek loans for pur- lower than elsewhere in the world, and the same is true poses of liquidity management. It seems that many of the for women as top managers and firm owners. Within the disconnected firms have adapted production strategies to region, the share of women’s employment is higher in an environment in which they do not consider banks as labor-intensive sectors and among exporting firms. Youth a financing option, albeit at the cost of reduced growth employment is higher among firms that are young and prospects. fast-growing, and which tend to innovate. Collateral standards affect firms’ propensity to Firm dynamics are generally weak, but high labor disconnect from the banking sector and ultimately their productivity firms are still more likely to grow fast growth prospects Overall, firm dynamics are weak in the region: compara- In the MENA ES economies, more than four in five loans tively few firms move between size categories, whether require collateral with an average value of just over twice expanding or downsizing. In a difficult period for the pri- the loan amount, slightly above that in peer economies. vate sector in the MENA region, medium-sized firms have The higher the relative collateral requirements, the more been more likely to become small firms and less likely likely young firms are to disconnect from the banking sec- to grow over a three-year period, compared with other tor. Older firms, on average, have more assets that they regions. Fast-growing firms over the period 2009-2012, can use to secure loans and are relatively less affected by however, had higher levels of initial labor productivity, an collateral standards. But they also create jobs at a slower indication of reallocation of resources toward the more rate than young firms and, as such, collateral practices productive firms and a signal of potentially positive private may constrain employment growth. sector developments. Executive summary 5 Skills shortages affect the fastest-growing firms The region’s exporters are numerous but small, Across the region, firms that have grown the fastest are with labor productivity gains concentrated in large more likely to perceive the lack of an adequately educated “superstar” exporters workforce as a major constraint. Unlike other firms, fast- Trade per se is not the problem underlying relatively weak growing firms are also more likely to invest in the formal competitiveness: firms in the MENA ES economies are training of employees, suggesting that the supply of more likely to export, to import, or to do both than their relevant knowledge and skills is a severe constraint for counterparts elsewhere, but these firms are also more the most promising, high-growth firms in the region. likely to be SMEs. Furthermore, the average size and productivity differentials between exporting and non- exporting firms are smaller than in other regions. Indeed, More productive firms pay higher wages, but larger the region’s exporter size and productivity premia are firms do not achieved almost entirely by a small number of superstar The MENA ES results confirm the expectation that exporters. The inability or unwillingness of small exporters more productive firms pay higher wages. This suggests to scale up their operations may indicate barriers to mar- that labor markets are, to some extent, able to facilitate ket entry or distortions, such as subsidized energy costs. the reallocation of labor resources to the firms with the most potential to grow and provide rewarding jobs. Nonetheless, such high-productivity, high-paying private Access to foreign technology and supply chains can sector jobs remain scarce, which is likely to encourage raise the productivity of importing firms jobseekers to pursue public sector jobs instead. In terms of productivity gains from trade, the winners in the region are importers. This could be due to the ac- In most economies, larger firms pay higher wages, but cess to foreign technology and supply chains from which that standard result does not hold in the MENA ES region. they benefit. This is despite the fact that importers face It seems that larger firms, which are more productive considerable obstacles in terms of relatively high tariffs, mostly due to inefficiently high capital intensity, focus non-tariff restrictions on trade, and the time it takes for on stronger capital remuneration rather than labor remu- imports to clear customs. neration. This gives an indication that distorting incentives, which are at the base of the decision to favor more capital- intensive production, might also affect the quality and Nearly a third of firms in the region engage in basic remuneration of jobs. forms of innovation Firms in the region engage in both technological and non- technological innovation, introducing new products, new Competitiveness: trade, innovation, processes, and new organizational or marketing methods and management at a similar rate. Much of this innovation activity involves adapting existing products to local conditions or upgrading The growth of the region’s small yet productive private machinery and equipment, practices that are typical of sector may be constrained by wider considerations of firms in developing economies. competitiveness The MENA ES economies generally perform worse on Innovation by firms is associated with certain various global competitiveness rankings than their peer supporting conditions: human capital, access to economies in other regions. The apparent inability of the knowledge, and access to finance region’s small yet productive firms to scale up their opera- tions may indicate distortions and uncertainties underlying Firm-specific human capital—obtained through formal the competitiveness of these economies. training or by giving employees time to develop new approaches and ideas—is associated with innovation, as is access to knowledge and information and communica- tions technology facilitated by firms. Two-way traders 6 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY (firms that both import and export), in particular, are more and international trade. The more productive firms in the likely to license foreign technology and more likely to region are able to grow faster and pay higher wages to introduce technological innovations. Firms with access attract workers. This suggests an encouraging potential to credit are more likely to introduce new products and for MENA economies to reallocate resources to the most processes. promising firms. In this way, it is possible to see the potential of the private Innovation is positively linked to increases in labor sector in the region to grow and meet the aspirations of productivity the growing workforce for rewarding employment. Indeed, Labor productivity gains from innovation are in line with it is through more widespread employment creation that those found in developed economies, but lower than private sector growth can principally be expected to con- those observed in developing economies. This may be tribute to a more inclusive growth model in the region. explained by the general lack of competition in many MENA ES economies compared with other developing At the same time, it is essential to understand that firms economies. Returns to innovation vary by sector, with operate under conditions that are often very difficult. high-tech manufacturers benefitting most from product Distortive incentives push large firms toward inefficient innovation and low-tech firms benefitting more from non- more capital-intensive production models; SMEs face technological innovations. limited growth opportunities and are more negatively affected by the business environment. Almost all firms in the region are severely affected by issues of political in- Poorly managed firms benefit more from improving stability, corruption, and unreliable electricity supply. Firm their management practices than from innovation innovation and growth are also constrained by barriers to The quality of management practices is positively corre- trade and a scarcity of appropriately trained workers. In lated with GDP per capita but not significantly associated many places, there is a striking disconnect between firms with firm-level labor productivity, except for firms that and formal financing channels, with the result that firms score below the median for their management practices. are not seeking external finance, inevitably reducing their While better-managed firms are more likely to benefit growth potential. from innovation, poorly managed firms are more likely to benefit from improving their management practices. Strategies to support firms in enhancing their productivity—as well as the process of resource real- location toward more productive firms—should be a high In economies with lower energy subsidies, better priority for public authorities in the region. The report sug- management practices are associated with lower gests some key areas for policy attention. These include energy intensity and higher labor productivity looking at the complex system of distortive incentives, Where energy subsidies are high, better management is privileges, and barriers to competition, as well as their associated with the opposite effect: higher energy inten- intended and unintended consequences. sity and lower labor productivity. Policies to improve the business environment Conclusions Achieving political stability is obviously a critical issue. The formal private sector in the MENA ES economies Across many of the economies, tackling corruption and an is relatively small, but its size belies its significance for unreliable electricity supply are also likely to be important economic development. The labor productivity of formal priorities. Corruption may be deterring many firms from private firms in the region is higher than that of their strategies that require engagement with public authori- counterparts in comparable regions of the world; yet TFP ties, limiting their opportunities. Dealing with the reliability lags behind. Many firms are successful in enhancing their of electricity may also depend on a policy approach that productivity though significant engagement in innovation addresses corruption and vested interests. Executive summary 7 More generally, the region is known for a large number of are likely to be particularly positive for the employment of distorting incentives, which form the basis of the current young people. They will also boost aggregate productivity system of transfers. Unintended consequences are often growth and raise living standards through better-paid jobs. addressed by adopting new and potentially distorting incentives. A serious reassessment of distorting incen- A re-orientation of the region’s education systems tives, transfers, privileges, and barriers to competition is toward learning skills that are relevant for private sector of central importance. employment—with greater status given to vocational training—will facilitate the growth of high quality employ- ment. Fast-growing and more productive firms are already Policies to enhance firms’ access to finance providing more training to their employees as well as While disconnecting from the financial sector is a choice better-paid jobs. More appropriate education and training that many firms make, the fact that this has an impact on of young people before they join the labor market would their growth potential reveals the need for policy action. help to address skill shortages in these firms. Several issues may need to be addressed to facilitate firms’ access to finance, to encourage them to connect with the formal financial sector, and to seize opportunities Policies to promote trade, competition, and innovation for growth that rely on external financing. Enhancing the productivity of firms in the region requires greater openness to international trade. In particular, this Capacity building for banks to strengthen their credit means more effective customs and trade regulations— risk assessment would help those interested in lending both in terms of imports and exports—and reducing entry to SMEs, without putting financial stability at risk. This costs for all firms. Importing should not be viewed solely should be accompanied by reforms to establish modern through the lens of trade deficits and foreign exchange secured transactions laws and an efficient collateral reg- reserves; imports allow firms to source component parts istry; to introduce credit guarantee schemes to alleviate of better quality or lower cost than those available in the collateral constraints; and to build capacity for SMEs to domestic market. They also facilitate the acquisition of improve their transparency and reduce the information knowledge about new products and processes. asymmetries. Other essential measures include promoting greater competition by reducing restrictions on firm entry and Policies for better education, employment and skills exit, and on foreign investment. Measures that give in- With regard to employment in the formal private sector, cumbent firms undue advantage—for example, privileged there is considerable scope for improvements, particularly access to markets, licensing, and contracts—should be in relation to women’s employment. Supporting the expan- eliminated, along with regulations protecting state-owned sion of labor-intensive and exporting sectors may help to or politically connected firms. Improving access to finance provide more jobs for women, but opportunities are also and improving the skills of the workforce will also support required in capital-intensive sectors. Measures that sup- the ability of firms to innovate and grow. port the emergence and growth of young innovative firms 8 1. Introduction  O ver the last few years, the Middle East and North Africa (MENA) region has witnessed unprecedented social and economic transfor- mation. In the Arab Uprisings, thousands of young certainly require significant growth of the private sector.1 The Arab Uprisings reflected the failure of the re- people took to the streets to voice their frustration gion’s economic models to satisfy people’s needs with the lack of economic and social opportunities. and expectations. These models typically featured These events reflected the demand for improve- strong protectionism, lack of integration in interna- ments in living conditions, infrastructure, job quality, tional markets, misguided state intervention, and lack of support for a business environment that education and healthcare services, as well as better fosters innovation, entrepreneurship, and good governance. management practices. In contrast, there is cur- rently an overwhelming analytical consensus that The Arab Uprisings took place against the back- the development of a vibrant private sector is crucial ground of a rapidly expanding workforce and for creating more opportunities, more sustainable rising unemployment—particularly among young employment, and thus more inclusive growth in the people—to some of the highest levels in the world. region.2 Indeed, a World Bank study conducted in the early 2000s concluded that the region would need to In light of this, sound policies for private sector create about 6 million new jobs each year to absorb development need to be based on a thorough as- new labor market entrants and bring down unem- sessment of the state of the private sector in the ployment, especially among young people. This will region and the challenges it faces. As a diagnostic Chapter 1: Introduction 9 tool, this report presents and discusses firm-level data The MENA ES provides a representative sample of the collected by the MENA Enterprise Survey (ES) in eight non-agricultural, formal private sector (figure 1.1). As economies—Djibouti, the Arab Republic of Egypt, Jordan, shown in table 1.1, the survey respondents comprised Lebanon, Morocco, Tunisia, the West Bank and Gaza, and 6,083 formal (registered) firms in the private sector across the Republic of Yemen—which are collectively referred to the eight economies. Table A1.1 in the Appendix provides in the report as the MENA ES region. a breakdown of the type of firms that were interviewed, along with the geographic regions of the surveys. Firm-level survey data are a crucial complement to macro- economic data. They make it possible to analyze firm-level To be included in the survey, firms needed to have at least productivity and performance—as well as their variation five employees and to operate in the manufacturing or across different types of firms—to understand what services sectors. “Services” include retail, wholesale, drives firm performance and the specific challenges faced hospitality, repairs, construction, information and com- by the private sector. munication technology (ICT) and transport. Not included in the survey are agriculture, fishing and extractive indus- The collection of these data took place in 2013 and 2014, tries, as well as utilities and some services sectors, such amid considerable social and economic upheaval. In particu- as financial services, education and healthcare.3 Also not lar, entrepreneurs and firms across the MENA region faced included are firms with 100 percent state ownership. increased uncertainty with negative implications for their business and investment decisions. In addition to domestic Table 1.1: MENA Enterprise Survey characteristics developments, international economic conditions were also Economies Djibouti; Egypt, Arab Rep.; Jordan; Lebanon; unfavorable, particularly as economic activity in Europe— covered* Morocco; Tunisia; West Bank and Gaza; and Yemen, Rep. one of the region’s major trading partners—was weak. Sample 6,083 private firms, which are: • registered, This difficult environment is inevitably reflected in the • with five or more employees, and • in the manufacturing or services sectors. snapshot that the data provide on firm performance, the Sampling Random, stratified by: business environment, and firms’ expectations of the • regional location within each economy, future. Nonetheless, the survey also reveals much about • firm size, and • sector of activity. the objective status and performance of firms in the Reference period Fiscal year 2012 region, as well as the structural challenges that they face. * The MENA ES rollout included the economies of Djibouti, the Arab Republic Both factors—the short-run events and the long-standing of Egypt, Jordan, Lebanon, Morocco, Tunisia, the West Bank and Gaza, and environment in which they emerged—are important for the Republic of Yemen. Initial plans to include Algeria, Libya, and the Syrian Arab Republic were suspended due to the security situation and additional understanding the potential foundations of prosperity in budget constraints. the region. The survey used random sampling, stratified by firm The MENA Enterprise Survey size, sector of activity, and regional location within each The Enterprise Survey provide a rich source of information economy. Stratification ensures that there are enough about firms and the business environment in which they observations for robust analysis within each stratum. The operate. Topics include firm characteristics, annual sales, survey design, comprehensive sample frames, and sam- costs of labor and other inputs, performance measures, pling weights used in the MENA ES together ensure that access to finance, workforce composition, women’s the surveys are statistically representative of the private participation in the labor market, and many aspects of the sector in each economy. business environment. Survey data are not only useful for corroborating findings based on macroeconomic data but also for exploring heterogeneity at the firm level and The MENA ES region examining how firms experience laws and regulations. Given the differences among MENA ES economies, it is useful to benchmark results against other regions covered by the surveys, as well as distinguishing 10 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 1.1: Number, size and sector of firms surveyed Lebanon Morocco Tunisia West Bank and Gaza Jordan Eqypt, Arab Rep. Firm size Firm sector Number of surveyed firms < 500 Small Manufacturing 500–2,000 Medium Retail > 2,000 Large Other services Yemen, Rep. Djibouti Source: Enterprise Surveys. between lower-middle-income and upper-middle-income Table 1.2: Selected indicators for the MENA ES economies.4 Using the World Bank Group classifications, economies Jordan, Lebanon, and Tunisia are upper-middle-income Human GDP per economies, while Djibouti, Egypt, Morocco, the West Population Development capita, 2012 Bank and Gaza, and the Republic of Yemen are lower- 2012 Index rank (constant Economy (millions) (HDIR), 2013 USD 2005) middle-income economies. Djibouti 0.9 170 1,144 Egypt, Arab Rep. 81.0 110 1,560 It should be stressed that while this report makes many Jordan 6.3 77 2,839 references to the MENA ES private sector as a whole, Lebanon 4.4 65 7,245 the region is far from homogeneous, with entrepreneurs Morocco 33.0 129 2,462 across the region facing wide differences. As table 1.2 Tunisia 11.0 90 3,921 shows, MENA ES economies have very different popula- West Bank and Gaza 4.0 107 1,564 tion sizes, ranging from just under one million (Djibouti) to Yemen, Rep. 24.0 154 729 81 million (Egypt). The MENA ES economies also have dif- Lower-middle-income 60.0 1,509 ferent levels of development: at the extremes, Lebanon’s Upper-middle-income 55.0 5,291 GDP per capita in 2012 was 10 times that of the Republic Source: UN National Accounts Main Aggregated Database (2005 USD). of Yemen. Using a broader measure of well-being, such HDIR: low value indicates better performance. as the Human Development Index (which takes account of life expectancy, levels of schooling, and income), also reveals wide disparities, with Djibouti ranked 170th out of 187 economies in 2013 and the upper-middle-income economies of Jordan, Lebanon, and Tunisia nearer the middle of the rankings. Chapter 1: Introduction 11 As is to be expected for middle-income economies, all 13 percent of GDP and 11 percent of employment, and of the MENA ES economies have undergone consider- most segments of the services sector, which provides able structural transformation,5 with industry on average 35 percent of employment on average.6 The MENA accounting for 30 percent of GDP and services for 61 ES excludes informal economic activity. The size of the percent (table 1.3). Agriculture plays a lesser role in terms informal economy in the MENA ES economies has been of value-added (although providing 14 percent of GDP estimated as equivalent to 33 percent of GDP on aver- in Egypt and Morocco), but it remains very important in age (for economies where estimates are available, table terms of employment, particularly in Egypt, Morocco, and 1.3), but this activity is only partially included in the GDP the West Bank and Gaza (table 1.4). measure. The informal sector is estimated to provide up to 67 percent of non-agricultural employment in Morocco The sectors analyzed in the MENA ES include the full and 49 percent overall (table 1.4). manufacturing sector, which provides an average of Table 1.3: GDP by sector and the informal economy in the MENA ES economies GDP by sector (%)* Informal economy Industry (% of GDP), Agriculture Of which manuf.: Services 2004/2005** Djibouti 3.5 16.9 2.5 79.3 n.a. Egypt, Arab Rep. 14.5 39.2 15.8 46.3 35 Jordan 3.1 30.1 18.8 66.8 20 Lebanon 6.1 20.5 8.5 73.4 37 Morocco 14.4 30.3 15.9 55.3 37 Tunisia 9.2 31.1 17.0 59.7 38 West Bank and Gaza 5.3 25.1 16.2 69.6 n.a. Yemen, Rep. 10.1 49.2 7.8 40.6 27 MENA ES 8.3 30.3 12.8 61.4 33 Source: * World Development Indicators. Data are for 2012, except for Djibouti (2007) and the Republic of Yemen (2006). ** Schneider (2007). Informal sector activity is only partly estimated in official figures for GDP. Note: n.a.—not available. Table 1.4: Employment by sector and the informal economy in the MENA ES economies Employment by sector (formal and informal sectors, %)* Employment in the informal economy (% non-agricultural Services Public administration, employment), Agriculture Manufacturing (ISIC groups F-I) defense & education Other sectors 2000-2004** Djibouti n.a. n.a. n.a. n.a. n.a. n.a. Egypt, Arab Rep. 27 11 32 18 12 46 Jordan 2 10 33 36 18 n.a. Lebanon n.a. n.a. n.a. n.a. n.a. 52 Morocco 38 11 32 8 11 67 Tunisia 16 19 35 19 11 35 West Bank and Gaza 38 12 40 25 13 43 Yemen, Rep. 16 7 39 22 8 51 MENA ES 20 11 35 21 12 49 Source: * ILO, KILM. ** Charmes (2012). Note: n.a.—not available. Agriculture refers to ISIC Rev. 3.1 group A; manufacturing to Group D; services includes groups F-I (construction; wholesale and retail trade; repair of motor vehicles, motorcycles and personal and household goods; hotels and restaurants; transport, storage and communications); public administration, defense and education refers to groups L and M; other sectors refers to groups B, C, E, J, K, N-X (including mining, utilities, finance, real estate and health). Data are for 2012, except the Republic of Yemen (2010). 12 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY The MENA ES does not cover publicly provided services competition, all of which make finance more difficult to and purely state-owned firms, which are important provid- access for many firms in the region. ers of employment in the region. Jobs in public admin- istration, defense and education comprise 21 percent Chapter 4 examines the contribution of different segments of employment on average, but this may underestimate of the private sector to employment, with a particular the total number of employees in the public sector, as it focus on youth employment, women’s employment, and excludes for example, state-owned enterprises. the role of women in management. The chapter highlights the lack of dynamism of medium-sized firms and explores The universe of sectors included in the surveys com- the relationship between employment growth and factors prises a relatively restricted segment of the MENA ES such as access to finance and labor productivity. It also economies. But the segment it covers is important for considers the extent to which firms face constraints in development: it is the growth of both employment and terms of access to adequately skilled workers and the value-added in the formal private sector that provides the impact of skills on productivity and wages. best opportunities for improving the long-term prospects and prosperity for the growing workforces of the econo- Chapter 5 explores the broad issue of firm competitive- mies in the region. ness, and specifically the effect on firm performance of participation in international trade, innovative activity, and management practices. The chapter investigates the The focus of this report extent to which firms in the region have been able to take The following four chapters focus on key issues for un- advantage of opportunities for output and productivity derstanding the growing role played by the formal private growth through participation in trade. It investigates the sector in the MENA ES region, the constraints faced by state of innovation and its relationship to productivity these firms, and the opportunities for promoting faster growth. It also discusses the effect of management prac- private sector growth and job creation. tices on firm productivity and the efficient use of energy resources. Chapter 2 examines firm productivity and its relationship to the general business environment. Political instability is a particularly acute concern for firms in the economies Endnotes most directly affected by the Arab Uprisings, and this seems to have affected their economic performance. The 1 World Bank (2004). chapter also discusses corruption and unreliable electric- 2 Kabbani and Kothari (2005, p.16). ity supply, which are two further core concerns of firms 3 More information on the ES methodology along across the region. with inclusion/exclusion criteria can be found in www.enterprisesurveys.org/methodology. Chapter 3 explores a critical issue in the business environ- 4 These regions are Latin America and Caribbean (LAC), Eastern Europe and Central Asia (ECA), Sub-Saharan Africa ment: the extent to which firms experience difficulties in (AFR), South Asia (SAR) and East Asia and Pacific (EAP). getting access to finance, and whether some may even 5 Atiyas and others (2015). choose to opt out of the formal financial system. The chap- 6 The MENA ES excludes certain segments of the services ter argues that by disconnecting from financial services, category, such as public administration, healthcare and firms forgo growth opportunities. More financial sector education. The services covered by the ES are represented flexibility and competition would help firms to re-engage. by ISIC Rev. 3.1 groups F–I (including construction, retail, The chapter highlights the specific issues of collateral vehicle repair, hotels and restaurants, transport and requirements, branch density, and lack of banking sector communications) as well as K.72 (computer and related activities). Chapter 1: Introduction 13 Appendix A1 Table A1.1: Enterprise Survey in the MENA region: Number of firms interviewed (n= 6,083) and levels of stratification by economy Stratification level Firm size (no. employees) Size of the Small Medium Large economy Economy (5-19) (20-99) (100+) Total Sector of activity Locations Large Egypt, Arab 1,273 1,029 595 2,897 Food (257), Textiles (224), Garments (206), Leather (111), Cairo (794), Giza (476), Upper Egypt Rep. Printing & Publishing (58), Chemicals (173), Rubber (355), Kafr-El-Sheikh/Menoufiya/ & Plastic (121), Non-Metallic Mineral Products (245), Beheira (226), Alexandria (192), Sharqia Fabricated Metal Products (91), Furniture (142), Wood (187), Qualyubia (144), Gharbiya (132), Products (78), Other Manufacturing (316), Construction Port Said/Suez/Ismalia (124), Damietta (134), Services of Motor Vehicles (49), Wholesale (122), (117), Dakahliya (114), Red Sea/ Retail (147), Hotels & Restaurants (163), Transport, Matrouh/Wadi Al Jadid/Sinai (36) Storage & Communications (256), IT (4) Medium Tunisia 199 237 156 592 Food (83), Garments (84), Other Manufacturing (163), South Coast/West (148), Northeast Retail (34), Other Services (228) (141), Tunis (135), Sfax (126), Interior (42) Jordan 266 181 126 573 Food (88), Garments (66), Other Manufacturing (181), Amman (274), Zarqa (99), Irbid (97), Retail (106), Other Services (132) Aqaba (52), Balqa (51) Lebanon 264 207 90 561 Food (89), Other Manufacturing (150), Retail & Mount Lebanon (139), Beirut (123), Wholesale (231), Other Services (91) South Lebanon (98), Bekaa Valley (85), North Lebanon (77), Nabatieh (39) Morocco 141 153 113 407 Food (49), Garments (38), Other Manufacturing (100), Grand Casablanca (107), South (100), Retail (36), Other Services (184) Central (80), Rabat/Sale/Zemmour/Zaer (71), North (49) Yemen, Rep. 211 102 40 353 Manufacturing (117), Retail (126), Other Services (110) Rest of the country (204), Amanat Al-Asemah (149) Small West Bank 292 119 23 434 Manufacturing (158), Retail (112), Other Services (164) West Bank (295), Gaza (139) and Gaza Very Small Djibouti 169 79 18 266 Manufacturing (62), Retail (59), Other Services (145) Djibouti City (266) Source: Enterprise Surveys. Note: The Enterprise Survey sample designs are based on stratified random sampling. More information can be found in http://www.enterprisesurveys.org/methodology. For sector of activity and locations, the number in the parentheses indicates the number of interviewed firms in that particular strata level. For comparisons with other geographic regions, the number of firms is as follows: ECA (8,730), LAC (12,046), EAP (9,026), SAR (13,381), and AFR (16,968). 14 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY References Atiyas, Izak, Ahmed Galal, and Hoda Selim. 2015. “Structural Transformation and Industrial Policy: Volume 1: A Comparative Analysis of Egypt, Morocco, Tunisia and Turkey, ” Luxembourg: European Investment Bank. Charmes, Jacques. 2012. “The Informal Economy Worldwide: ” The Journal of Applied Trends and Characteristics, Economic Research 6(2): 103-132. Kabbani, Nader, and Ekta Kothari. 2005. “Youth Employment in the MENA Region: A Situational Assessment, ” Social Protection Discussion Paper No. 534, Washington, DC: World Bank. Schneider, Friedrich. 2007. “Shadow Economies and Corruption All Over the World: New Estimates for 145 Countries, ” Economics, No. 2007-9. World Bank. 2004. “Unlocking the Employment Potential in the Middle East and North Africa: Toward a New Social ” Washington, DC: World Bank. Contract, 15 2. Firm productivity and the business environment Introduction of labor productivity, but that labor productivity has been declining over time. Furthermore, high labor Firms’ productivity—their effectiveness in produc- ing output from inputs—is the basis for their ability productivity has been achieved through inefficiently to survive and compete in national and international high capital intensity, resulting in lower total factor markets. Rising productivity in the private sector is productivity (TFP). Large firms are generally more key for economic growth, and it is a good indicator productive, but tend to be more capital-intensive of a well-functioning private economy. In the ab- and to focus on capital remuneration. sence of market frictions, resources are reallocated toward more productive firms, thereby reinforcing The chapter also presents evidence on the impact of the process of growth and opening opportunities for the business environment on firm performance and more productive jobs. growth. Firms perceive political instability, unreliable electricity, corruption, and inadequate access to This chapter analyzes firm productivity and the busi- finance as key constraints. Small and medium-sized ness environment in the MENA ES region. It shows enterprises (SMEs) tend to experience a more chal- that firms are relatively more productive in terms lenging operating environment than larger firms. 16 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Firm productivity Box 2.1: Estimating total factor productivity with survey data Labor productivity is somewhat higher than in The use of micro or firm-level data to estimate total factor peer economies of the MENA ES region, but productivity (TFP)—the portion of output not explained by the total factor productivity lags behind amount of inputs utilized—has enabled analysts to explore how the efficiency of production varies with heterogeneous Figure 2.1 shows the distribution of firm-level firm characteristics. Most analytical work begins with a Cobb- labor productivity and TFP in the formal private Douglas production function in the form: yi = ai kiβk liβl miβm where sector for each of the MENA ES economies in firm-level output yi is a function of inputs of capital (ki), labor (li), and other inputs such as materials (mi); firms’ efficiency of comparison with the median productivity level production is measured by the term ai which is the portion of for economies at a similar income level outside output that cannot be directly attributed to the utilized inputs. the region—its “peer economies. ”1 If the dis- Analytically straightforward, estimation can be troublesome. tribution of either performance measure in an Often only monetary (as opposed to physical) output and inputs economy is similar to that in peer economies, are observed, and the resulting productivity measures thus in- roughly half of firms will fall below the compa- corporate market dynamics through clearing prices; such reve- nue-based TFP is often referred to as TFPR.a In addition, it has rable median and roughly half will be above this been widely noted that even within narrowly defined industries level. Likewise, if a relatively higher proportion results exhibit large and persistent differences across firms.b of firms are above the income-group median, Empirically, TFPR is generally estimated by regressions in the this indicates generally higher levels of firm form of: Yi=βk Ki + βl Li + βmMi + εi , where capital letters indicate performance, with the converse being true if natural logarithm of monetary inputs and outputs. εi is the natu- more firms fall below the median. ral logarithm of firm-specific productivity. Capital, Ki , is proxied by the replacement value of machinery and equipment. Labor, In most MENA ES economies, firms have labor Li , inputs are represented by total wage bill, while materials, Mi , are measured as the cost of raw materials and intermedi- productivity levels that are somewhat above the ate goods used in production. TFPR is thus only meaningful comparable income-level median—that is, more for manufacturing firms. It should be noted that since data are than half of firms report higher revenues per cross-sectional (and not time-series), corrections for the endo- worker compared with peer economies. Jordan geneity of inputs (that is when firms have knowledge of their and the Republic of Yemen, where a majority productivity and set their capital and labor inputs simultane- falls below the median, are the only exceptions.2 ously) is not possible. Since the above specification assumes a common production This higher labor productivity could result from technology, TFPR was estimated separately for each industry— greater efficiency, superior technology, and/ grouped by two-digit ISIC codes, s —and pooling economies by income level—grouped by the World Bank classifications, w. or the intensive use of complementary inputs, To allow for an average economy-level effect, a dummy variable such as capital or material intermediates. The for each economy c is included.c The final estimation is then latter explanation seems to be confirmed by Yisw = βksw Kisw + βlsw Lisw + βmsw Misw + ∑βcc + εisw. The firm-level the fact that TFP lags behind peer economies TFPR is the sum of the economy-industry-level effect and firm- in most MENA ES economies (figure 2.1),3 the specific productivity: TFPRi = εisw + βc. only exceptions being Jordan and Morocco. TFP For an economy-level measure of productivity, the firm-level measures the efficiency of use of all factors of TFPR is aggregated by taking into account each firm’s share TFPRi • ( i ) , where three differ- s production including not only labor but also capi- in the economy: TFPRc = ∑i=1 Nc ∑si tal and intermediates (see box 1.2 for details on ent measures of shares (s) were used: (i) sample weights, ωi , TFP computation). The results in figure 2.1 thus giving each firm a weight equal to the share of firms it repre- suggest that in most MENA ES economies, sents in the economy; (ii) sales share, yiωi ; and (iii) employment higher levels of labor productivity are achieved share, eiωi with ei being the number of permanent employees. at the expense of an over-reliance on capital and a Foster and others (2008). intermediates—and not underlying technologi- b See, e.g., Syverson (2011). cal superiority—with a resulting lag in TFP. c Halvorsen and Palmquist (1980). Chapter 2: Firm productivity and the business environment 17 In all economies but Jordan, Morocco, and the Republic of due to inefficiencies or imperfectly competitive markets, Yemen, higher-than-median labor productivity goes hand in or through incentives favoring greater factor intensity than hand with lower-than-median TFP . This is an indication that would otherwise be optimal. while labor is used somewhat efficiently, when all factors are taken into consideration, firms are actually less produc- Figure 2.2 shows the median factor shares of three main tive. In Jordan, firms tend to be more inefficient in using inputs used by manufacturers—their labor, intermediate labor, as reflected by below-median labor productivity, inputs, and capital costs respectively. The capital factor but above-median TFP . The Republic of Yemen stands out: share is above the respective peers in Egypt, Jordan, firms are relatively inefficient and characterized by low la- Lebanon, and Tunisia, revealing higher capital intensity. bor productivity and TFP . By contrast, in Morocco, relatively The labor factor share is above the respective peers only high labor productivity is also associated with relatively in Jordan and Lebanon. high TFP , indicating a comparatively efficient system. Egypt and Tunisia’s story is consistent with a pattern of relative investment in energy-intensive (subsidized) and MENA ES manufacturers tend to have lower labor capital-intensive industries—for example, in metal and ce- intensity and higher capital and intermediates intensity ment production.4 Moreover, the subsidization of energy Factor shares have long been used to study the impor- inputs (and the subsequent favoring of capital-intensive tance of each type of input in the production process. production) renders labor relatively more expensive. This Each ratio—expressed as a proportion of total annual limits the potential of job expansion through greater labor revenues—shows the relative intensity of those input intensity. Furthermore, if labor is relatively more abundant costs to revenue output, and is thus itself a simple mea- relative to private sector demand, wages will slump. sure of productivity. If a firm’s ability to command greater While this will increase employment, it will be at the cost revenue is high relative to inputs, it is generally regarded of lower wages rather than a result of more labor-intensive as more productive, a sign of underlying efficiency; if, production techniques (chapter 4). however, factor shares are high relative to revenues (as well as to each other and vis-à-vis comparators), they may reveal lower underlying productivity—a disproportionate Larger firms have higher levels of productivity in expense on inputs. The latter scenario can be due to in- manufacturing but not in services ferior technology and/or comparatively expensive costs of In the MENA ES economies, there is no significant asso- production—as would be the case if input costs were high ciation between firm size and productivity in the services Figure 2.1: Distribution of firm-level labor productivity and total factor productivity LABOR PRODUCTIVITY TOTAL FACTOR PRODUCTIVITY (REVENUE) Upper-middle-income economies Lebanon Lebanon 69% Lebanon 41% Tunisia Tunisia 53% Tunisia 42% Jordan Jordan 40% Jordan 57% Lower-middle-income economies Morocco Morocco 74% Morocco 58% Djibouti Djibouti 71% Data unavailable Data unavailable West Bank and Gaza West Bank and Gaza 60% West Bank and Gaza 46% Egypt, Arab Rep. Egypt, Arab Rep. 50% Egypt, Arab Rep. 37% Yemen, Rep. Yemen, Rep. 31% Yemen, Rep. 48% Percent of firms above income group median Percent of firms below income group median Income group median Source: Authors’ calculations based on Enterprise Surveys. Economies are ordered within income group by labor productivity distribution. 18 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 2.2: Median factor shares 50 40 Percent 30 20 10 0 Morocco West Bank Yemen, Egypt, Lower- Jordan Lebanon Tunisia Upper- and Gaza Rep. Arab Rep. middle- middle- income income Labor Intermediates Capital Source: Enterprise Surveys. Note: Within each income group, economies are ordered by the total sum of factor shares, low to high from left to right. sector; this is not the case in the manufacturing sector. capital across the region (figure 2.3C). In both income When only labor is considered as a factor of production, groups, this allocation increases with firm size. manufacturers in the MENA ES economies show a sig- nificant and positive relationship between the number of The analysis above suggests that the higher productivity workers they employ and their revenues per worker—that of larger firms overlies their higher capital intensity. If is, labor productivity (table A2.1, column 2). But when oth- this is efficiency-enhancing in terms of physical produc- er factors of production are taken into account—namely tion (and not just in commanding greater revenue), the the costs of capital and intermediate inputs—the addition substitution of labor intensity with capital inputs would of more workers reduces labor productivity. expand overall productivity. But the MENA ES economies’ relatively poor TFP compared with peers suggests that This finding may imply that the positive association this relative intensity may be less than optimal, possibly between firm size and labor productivity is due to the a consequence of distortive incentives pushing toward extensive use of capital by large firms, not necessarily due capital intensity. to the number of employees in the firm. This again points to a strong bias towards capital and intermediates relative to labor: that is, firms with comparably more employees Labor productivity is declining maintain higher labor productivity precisely through their Despite comparatively higher labor productivity, revenues intensive use of other inputs of production. per worker are contracting over time in all MENA ES economies. This may be partly explained by the wide- spread social and political upheaval. The surveys make use Relative to revenue, larger firms spend more on capital of recall on sales and employment data from fiscal years than on labor inputs 2009 and 2012, allowing for indicators of performance be- In the MENA ES economies, larger firms allocate relatively fore and during the upheaval.5 Compared with their peer fewer resources to labor costs. While this pattern is con- economies, the MENA ES economies tend to lag behind sistent with lower-middle-income economies elsewhere, on average in sales, employment, and labor productivity it is not consistent with other upper-middle-income growth rates (figure 2.4). In fact, the annual rate of growth economies, in which there is no change in the labor-to- of labor productivity for every economy in the MENA revenue ratio as firms grow (figure 2.3A). More striking ES region over the period 2009–2012 is negative (figure is the relatively large amount of resources allocated to 2.4). This is the result of steady and positive employment Chapter 2: Firm productivity and the business environment 19 growth, except in Egypt and the Republic of Yemen, ac- Figure 2.3a: Labor cost factor share by size companied by weaker, and sometimes negative, sales 60 growth. 50 In response to the Arab Uprisings, several governments in 40 the region responded by rapidly increasing public spend- ing on food and energy subsidies; between 2009 and Percent 30 2012, subsidy expenses in real terms more than tripled 20 in Jordan, more than doubled in Tunisia, and increased by over 40 percent in Lebanon. The increase in Egypt was 10 only 4 percent, but it constituted a 1.6 billion expansion 0 Sm Med Lg Sm Med Lg Sm Med Lg Sm Med Lg in public spending in 2012 U.S. dollars.6 While making Lower-middle- Lower-middle- Upper-middle- Upper-middle- debt levels somewhat untenable, this additional public income income income income MENA ES MENA ES spending may induce further misallocations in the private sector, biasing firms toward capital and energy intensity and against further employment generation.7 Figure 2.3b: Intermediate input cost factor share by size Governments in Egypt, Jordan, Tunisia, and Morocco have 60 announced and begun energy subsidy reforms—allowing 50 gasoline and other fuel prices to rise as well as electricity tariffs.8 Though initial efforts have proceeded—in 2015, 40 the Egyptian government cut subsidies by nearly a third Percent 30 compared with the previous year—these reforms face 20 persistent political resistance.9 The starkest example is in the Republic of Yemen, where protests erupted after 10 the Saleh government cut energy subsidies in 2014, and 0 Sm Med Lg Sm Med Lg Sm Med Lg Sm Med Lg these reforms have been withdrawn further following the Lower-middle- Lower-middle- Upper-middle- Upper-middle- conflict.10 income income income income MENA ES MENA ES The business environment Figure 2.3c: Capital replacement cost factor share by The business environment includes regulatory compli- size ance, access to finance, infrastructure, and several other 60 contextual elements that affect the day-to-day experiences of firms. Productivity is as much dependent on internal 50 factors, such as technology, research and development 40 (R&D), management practices, and human capital as it is on the external factors of the business environment.11 But Percent 30 external factors can affect “within” aggregate productivity 20 growth by forcing individual firms to become more effi- cient; and they can affect “between” aggregate productiv- 10 ity by allowing more efficient firms to grow faster than 0 Sm Med Lg Sm Med Lg Sm Med Lg Sm Med Lg less efficient ones or by replacing less efficient firms with Lower-middle- Lower-middle- Upper-middle- Upper-middle- income income income income newer more efficient entrants. MENA ES MENA ES Many studies have established the effect of differ- Source: Enterprise Surveys, all shares shown on the same scale for comparison. ent dimensions of the business environment on firm performance, particularly in developing economies.12 20 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 2.4: Labor productivity growth has been negative in all MENA ES economies 10 5 Percent 0 -5 -10 -15 Djibouti Egypt Morocco West Bank And Gaza Yemen Lower- middle-income 10 Real annual sales growth (%) Annual employment growth (%) 5 Annual labor productivity growth (%) Percent 0 -5 -10 -15 Jordan Lebanon Tunisia Upper- middle-income Source: Enterprise Surveys. The business environment can affect firm productivity Corruption Perceptions Index in 2013, while the Republic directly—for example, through the reliability of electricity of Yemen ranked 167th out of 177 economies worldwide.14 supply—or indirectly by affecting decisions on the alloca- tion of resources.13 For example, corruption or burden- The World Economic Forum’s Global Competitiveness some regulation can create incentives for the reallocation Index, which covers a broader range of issues from of labor or capital resources from productive tasks to less infrastructure to financial markets and innovation-related than optimal uses, leading to lower aggregate productivity issues, reveals a similar picture of heterogeneity across and output. the MENA ES economies. While Jordan, Morocco, and Tunisia rank in the middle of the range, Egypt, Lebanon, Several aggregate measures of the business environment and the Republic of Yemen rank much lower, with the in the MENA ES economies point to substantial differ- Republic of Yemen at 145 out of 148 economies. ences among them. The World Bank’s Doing Business Index measures the overall regulatory environment by Similarly, according to the six World Governance Indicators considering the cost and complexity across 10 common for 2013, on average, Jordan, Tunisia, and Morocco tend business transactions for a medium-sized limited liability to rank just below the middle of the range among 210 company. According to this measure, in 2013, Tunisia was economies and the remaining MENA ES economies rank the 50th business-friendly economy in the world, while much lower. For example, the Rule of Law index ranks Djibouti was 170th. Tunisia, together with Jordan, also Jordan at 79th, Tunisia at 103rd, and Morocco at 111th; ranked relatively high in the Transparency International’s Egypt, Lebanon, and Yemen are ranked much lower at Chapter 2: Firm productivity and the business environment 21 business environment measures and benchmarking them Table 2.1: Selected business environment indicators for against other regions of the world. the MENA ES economies Doing Corruption Global Business Perception Competitiveness Economy rank, 2013 rank, 2013 rank, 2013-14 What are the main obstacles Djibouti 171 94 – perceived by firms? Egypt, Arab Rep. 109 114 118 Jordan 106 66 68 Political instability, corruption, and electricity are most Lebanon 115 127 103 commonly identified as “top obstacles” Morocco 97 91 77 Managers and CEOs who took part in the MENA ES were Tunisia 50 77 83 asked to select the “top obstacle” from a list of 15 potential West Bank and Gaza 135 N/A – obstacles. As figure 2.5 shows, political instability is the Yemen, Rep. 118 167 145 most commonly chosen top obstacle in five of the eight Sources: World Bank Group, Doing Business Index 2013; Transparency International, Corruption Perceptions Index 2013; World Economic Forum, Global economies. In the three economies that experienced a Competitiveness Index 2013-2014 edition. change of regime in the Arab Uprisings—Egypt, Tunisia, Note: Larger numbers represent worse performance. and the Republic of Yemen—one out of two firms cite po- litical instability as the top obstacle. Similarly, in Lebanon, with a history of political struggle compounded by the 140th, 158th, and 185th respectively. Overall, these ag- effects of the conflict in neighboring Syria, this percentage gregate measures indicate that even in the more prosper- nears 60 percent. Likewise, in the West Bank and Gaza— ous economies of the region, there is ample room for which was entering a period of heightened tension with improvement. Israel at the time of the survey—political instability is also the top obstacle for the private sector. In Jordan, political The World Bank’s Enterprise Survey also provide a valuable instability is still among the top three cited obstacles, window into an economy’s business environment, rooted primarily due to the spillovers from regional instability. in the day-to-day experiences of firms. The evaluation can be made from either a perception-based view of the In five economies, electricity is among the top three cited obstacles faced by the firm or by looking at factually-based obstacles. In comparatively stable Djibouti, nearly half of Figure 2.5: Political instability is most commonly chosen as the top obstacle in the MENA ES economies 60 50 Percent of firms 40 30 20 10 0 Egypt, Tunisia Yemen, Rep. Lebanon West Bank Jordan Djibouti Morocco Arab Rep. and Gaza Access to finance Electricity Informality Tax rates Corruption Inadequately educated workforce Political instability Source: Enterprise Surveys. Note: For each economy the three obstacles most frequently chosen as the top obstacle by firms are shown. 22 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY firms consider electricity to be their top obstacle. Indeed, concerns of firms about electricity, corruption, and access electricity seems to be a particular problem for firms to finance, all of which contribute to and are fed by the in three of the lower-middle-income economies in the overarching political instability. group—Egypt, the West Bank and Gaza, and the Republic of Yemen—as well as in one upper-middle-income economy, Lebanon. Corruption is among the three most Experience-based indicators of the business frequently cited top obstacles in four economies, which environment reveal specific areas of concern is largely consistent with the rankings of Transparency The MENA ES data also contain measures of firms’ actual International. Access to finance is ranked among the day-to-day experience dealing with specific elements of top three obstacles in three economies of the region; in the business environment. These include indicators of Jordan, it is the top obstacle. regulatory costs, such as the time that senior manage- ment spends in dealing with government regulations— In addition to the top obstacle ranking, respondents were the “time tax”; indicators of administrative efficiency, for given the opportunity to evaluate individual elements of example, the number of meetings held with tax officials the business environment to determine whether each and the waiting times to obtain licenses and permits; indi- element was a major or very severe concern to the cators of the exposure to crime and bribery; and indicators operations of the firm. Since this evaluation was done of the quality of infrastructure and market conditions, such independently of the other elements of the business envi- as shipment losses or power outages (table 2.2). ronment, it can be used to benchmark the extent to which any given obstacle is perceived as severe compared with For many of these indicators, the overall average for other economies.15 Figure 2.6 shows that political instabil- MENA ES economies is comparable to results elsewhere, ity and corruption stand out: they are considered severe though there are a few areas of concern. For example, by a much larger share of firms than in all ES economies. the time tax for Tunisia is exceptionally high, the highest Electricity and access to finance are also above the av- of any economy with ES data. Respondents there also erage of all economies with ES data, but the difference report three-month delays, on average, in getting an is not as large. The future growth of the formal private electricity connection. In Lebanon and Tunisia, obtaining sector requires reforms aimed at addressing the specific an operating license can take over 40 days; in Egypt, this waiting time is substantially longer with nearly a third of applicants reporting that their request was still in process. Figure 2.6: Political instability, corruption, and unreliable electricity supply are considered severe obstacles more In Lebanon, Morocco, and the West Bank and Gaza, frequently in the MENA ES region obtaining an import license may take up to a month, well above the time in the other economies. While overall, 70 the MENA ES economies do not show particularly poor 60 business environments, these specific deficiencies may 50 still be binding and can provide a starting point for policy Percent of firms reforms. 40 30 20 Political instability 10 Between 2010 and 2013, diverging growth patterns 0 Political Corruption Electricity Access reflected different levels of political stability instability to finance One useful way of viewing the private sector in different MENA ES All ES economies economies in the region is to look at relative trends fol- Source: Enterprise Surveys. lowing the period of upheaval around the Arab Uprisings and the onset of the Syrian civil war. While in the lead-up to 2010, all economies in the region showed positive Chapter 2: Firm productivity and the business environment 23 Table 2.2: MENA ES business environment averages mask individual areas of concern time spent in dealing Senior management Percentage of firms clear direct exports robbery, vandalism, Losses due to theft, paying for security Products exported water connection operating license with government through customs Average days to Days to obtain a directly lost due and arson (% of to breakage or Days to obtain Days to obtain Days to obtain import license an electrical spoilage (%) regulations connection sales) Djibouti 5.3 7.7 8.8 10.4 34.1 16.1 49.8 0.5 0.1 Egypt, Arab Rep. 3.1 19.8 138.9 7.4 75.7 20.5 20.4 0.6 0.5 Jordan 5.3 2.1 1.4 4.6 13.1 21.0 12.4 0.3 0.4 Lebanon 4.1 28.0 50.0 4.9 56.0 40.2 21.8 0.2 0.2 Morocco 4.6 30.6 24.1 3.5 13.8 49.8 39.5 0.3 0.6 Tunisia 46.5 12.9 39.2 3.0 89.3 17.2 68.7 1.0 0.2 West Bank and Gaza 4.4 35.4 11.5 2.5 42.5 13.4 35.2 1.9 4.2 Yemen, Rep. 1.9 11.6 7.0 11.2 25.6 35.9 27.1 0.6 1.6 MENA ES 9.4 18.5 35.1 5.9 43.8 26.8 34.4 0.7 1.0 Lower-middle-income 10.3 17.3 24.3 9.2 25.5 24.2 57.1 1.3 1.3 Upper-middle-income 10.6 22.7 35.7 7.3 23.7 29.8 55.5 0.7 0.8 All ES economies 9.8 18.4 30.1 7.9 30.0 27.7 57.5 1.0 1.0 Source: Enterprise Surveys. Figure 2.7: In economies with higher political instability, growth stagnated between 2010 and 2013 Panel A: Arab Uprisings economies Panel B: Continuing political Panel C: Politically stable instability 140 130 120 Percent 110 100 90 80 2008 2009 2010 2011 2012 2013 2008 2009 2010 2011 2012 2013 2008 2009 2010 2011 2012 2013 Egypt, Arab. Rep. Lebanon Djibouti Tunisia West Bank and Gaza Jordan Yemen, Rep. Morocco Source: WDI, authors’ calculations. Note: Figures are indexed to GDP per capita levels in 2008, which is set to 100. growth, they differed sharply as political events unfolded. of the government of Mohamed Morsi in 2012, GDP per In the three Arab Uprisings economies that underwent capita growth stagnated. In Tunisia, which experienced a regime change—Egypt, Tunisia, and the Republic of a relatively smoother political transition, growth initially Yemen—a distinct pattern is clear (figure 2.7 , panel A). In dropped, though it recovered after 2011. the Republic of Yemen after 2010, GDP per capita dropped precipitously amid tension leading up to the civil conflict. In the two other economies where political instability was In Egypt, which after the Arab Uprisings saw the removal most often ranked as the top obstacle—Lebanon and the 24 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY West Bank and Gaza—growth also seems to have been by the surveys), was less dramatically affected in Egypt, affected by geo-political events. As figure 2.7 shows Lebanon and Tunisia, although these trends resulted in ex- (panel B), growth flattened in Lebanon after 2010, a period plosions in public debt.16 In the politically stable Djibouti, that includes the civil war in neighboring Syria. While GDP Jordan, and Morocco, the shares of these sectors in GDP per capita in the West Bank and Gaza has grown consider- growth have changed comparatively little. ably relative to 2008, this was punctuated by periods of conflict, including in 2008–2009. Djibouti, Jordan, and On this basis, it can be suggested that private sector activ- Morocco (figure 2.7 , panel C) can be considered relatively ity has tended to be disproportionately affected by political stable. In Djibouti and Morocco, growth seems to have instability in the region, while other sectors, many associ- been little affected by instability, either domestic or in the ated with the public sector, were bolstered by high—and wider region. Growth in Jordan has been relatively flat probably unsustainable—levels of public spending and since 2009, which may partly reflect the economy’s expo- incurred deficits. sure to events in neighboring Syria. While the causal effect of this pattern is hard to discern—whether low growth has resulted in instability or the other way around—the Political instability is associated with negative sales association is clear. and labor productivity growth Between 2009 and 2012, the typical firm in the Republic of Yemen, Tunisia, and Egypt saw revenues collapse by a The formal private sector is disproportionately affected rate of -11, -7, and -6 percent per year respectively (figure by political instability 2.4). In contrast, firms in Lebanon and Jordan saw their In all of the economies severely affected by political revenues remain virtually flat (at a rate of -1 percent per instability, the formal private sector’s contribution to GDP year) over the same period. Only in Djibouti, Morocco and growth—as represented by the manufacturing and servic- the West Bank and Gaza was annual sales growth posi- es categories covered by the MENA ES—seems to have tive, though these rates lagged behind comparable rates fallen considerably, comparing periods before and after in other upper-middle-income and lower-middle-income 2010 (figure 2.8). In contrast, over the same period, the economies. contribution to growth of other sectors, including public administration, defense, health, education, the financial Poor sales growth performance in the Arab Uprisings econ- sector, and extractive industries (all sectors not covered omies was accompanied by a contraction in employment: Figure 2.8: The role of the private sector in real growth in value-added GDP 25 Percent growth in value-added GDP 20 15 10 5 0 -5 -10 -15 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 ‘08-10 ‘10-12 Egypt, Arab Tunisia Yemen, Rep. Jordan Lebanon West Bank and Djibouti Morocco Rep. Gaza Proxy for formal private sector - MENA ES comparable Other sectors Source: UN National Accounts Main Aggregated Database (2005 USD). Note: 2008–2010 is relative to overall value-added in 2008; 2010–2012 is relative to overall value-added in 2010. The values for the private sector are proxied by manufactur- ing (ISIC Rev. 3.1 Group D) and selected services (Groups F–I). “Other sectors” include Groups A–C, E and J–P. Chapter 2: Firm productivity and the business environment 25 the average firm shed jobs in Egypt and the Republic of Table 2.3: Perceptions of corruption score much higher Yemen, and kept its employment level virtually the same than factual indicators of exposure to bribery in Tunisia (figure 2.4). In contrast, in Jordan and Lebanon, firms added jobs, showing positive employment growth, Identifying corruption but sales did not keep apace, resulting in a contraction of Bribery Bribery as a major sales per worker (labor productivity). In Djibouti, Morocco, depth (% of incidence (% obstacle (% Economy transactions) of firms) of firms) and the West Bank and Gaza, firms both added jobs and Djibouti 8 11 39 increased their sales on average, indicating a potential Egypt, Arab Rep. 16 17 59 driving force for current and future growth. Jordan 10 13 21 Lebanon 14 19 61 The relatively poor growth performance of firms in econo- Morocco 29 37 53 mies suffering from greater political instability—coupled Tunisia 9 10 36 with the large number of firms that find political instability West Bank and Gaza 5 7 49 a key constraint on their performance—make a strong Yemen, Rep. 61 64 97 case for social, political and economic reforms to provide MENA ES 19 22 52 greater political stability in the region. Lower-middle-income 16 21 38 Upper-middle-income 9 12 33 Corruption Source: Enterprise Surveys. Note: “Major obstacle” refers to a rating by respondents as “major or “very severe”. Bribery depth refers to the frequency with which firms are confronted with bribe requests. Bribery incidence shows the average share of firms Perceptions of corruption as an obstacle may be driven exposed to at least one bribe. by factors beyond the scope of individual firms’ activity Corruption can result in a misallocation of resources, both through the allocation of resources to bribery and The share of firms recognizing corruption as a serious through the distortions in decision making that it creates. impediment is above 50 percent in Egypt, Lebanon, Corruption is the second most frequently rated major Morocco, and the Republic of Yemen. The higher incidence obstacle in the survey, after political instability. In addition, of corruption in the perception indicator compared with the survey collected information on the actual experience the transaction-based bribery indicators seems to indicate of firms dealing with petty corruption when engaging in that firms may be perceiving corruption in elements of the six different transactions, including applications for utili- business environment that are not related to their day-to- ties (water and electricity), imports, operating licenses, day operations. Some of these elements could include construction permits, and when paying taxes.17 corruption at high political levels and/or state capture by particular interest groups or elites. Furthermore, respon- As table 2.3 shows, the average share of firms exposed dents may be reticent and not report an interaction where to at least one bribe in the MENA ES economies (bribery a bribe was requested.19 Each of these could be a possible incidence) is considerably lower than the percentage of explanation for higher perceptions of corruption that are firms that consider corruption as a major obstacle to their not reflected in the experience-based information in the operations. On average, the frequency with which firms in MENA ES. the MENA ES region are confronted with bribe requests (bribery depth) is somewhat greater than the average for Corruption perceptions may deter firms from lower-middle-income and upper-middle-income econo- interactions with public authorities mies. But there is considerable variation across econo- mies, with Morocco and the Republic of Yemen standing In the MENA ES economies, firms engage in transactions out as having the highest values for bribery incidence and with public officials at a considerably lower rate than in depth.18 other regions (table 2.4). Excluding visits by tax officials— a transaction that is rarely voluntary—only a third of firms in the MENA ES economies engage in a public transac- tion, which is well below the average for peer economies. 26 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY High perceived corruption is associated with lower Table 2.4: Firms in the MENA ES economies engage in public transactions less frequently sales and employment growth and lower labor productivity Engaging in transaction The difference between perception-based and transac- Engaging in excluding visits by Percent of firms transaction* tax officials tion-based measures of corruption also matters for the Djibouti 72 53 relationship between corruption and firm performance, Egypt, Arab Rep. 78 16 even after taking several firm characteristics into account. Jordan 75 47 The survey results suggest that bribery incidence and Lebanon 50 24 depth are not related to firm performance, whereas firms Morocco 44 35 that perceive corruption as a severe obstacle tend to Tunisia 48 31 experience lower growth rates of sales and employment, West Bank and Gaza 69 43 and a lower level of labor productivity (table A.2.1).20 Yemen, Rep. 89 26 MENA ES 66 34 Together, these results suggest that while petty corrup- Lower-middle-income 80 54 tion may not limit firms’ performance, more widespread Upper-middle-income 72 41 corruption is problematic. Firms that see corruption as an Source: Enterprise Surveys. important constraint perform more poorly. Add to this the Note: *Transactions include applications for: an import license, an operating fact that firms in the MENA ES economies are less likely to license, water connection, electrical connection, a construction permit, or visits by tax officials. engage in transactions with public officials, and the case for reforms that go beyond petty corruption is strengthened. Assuming this is partly driven by the demand for transac- tions by firms, this may be an indication of the effects Unreliable electricity supply of economic uncertainty and the investment environment The quality of electricity provision varies greatly among on firms’ willingness to undertake activities that require the MENA ES economies applications for licenses and permits. Firms’ expecta- Electricity is the third most frequently cited major obstacle tions of bribe requests and poor service may also be a in the MENA ES economies. This measure, based on the significant factor deterring such interactions with public perceptions of managers and CEOs, can be compared to administrators. a number of experience-based measures of power supply Table 2.5: Electricity provision in the MENA ES economies Number of electrical Average total time of Value lost due to outages in a typical power outages per electrical outages Firms owning or sharing Electricity from month month (hours) (% of sales) a generator (%) generator (%) Djibouti 1.6 2.3 2.8 69.1 13.3 Egypt, Arab Rep. 16.3 28.8 5.6 5.9 1.0 Jordan 0.2 0.7 0.2 8.1 2.0 Lebanon 50.5 211.0 5.7 84.6 40.1 Morocco 0.6 1.0 0.2 11.2 2.3 Tunisia 0.3 4.1 0.2 4.3 1.8 West Bank and Gaza 8.7 66.5 6.4 21.4 6.3 Yemen, Rep. 38.8 158.4 16.1 80.5 38.5 MENA ES 14.6 59.1 4.7 35.6 13.2 Lower-middle-income 6.7 32.3 3.3 35.4 9.3 Upper-middle-income 2.1 7.6 1.1 25.8 3.4 Source: Enterprise Surveys. Chapter 2: Firm productivity and the business environment 27 quality in the survey. These include the number of power cited as a major constraint, firms are also heavily reliant outages in a typical month, the total duration of power on generators: 7 in 10 firms own or share a generator, outages in a typical month, and the total losses due to and firms using those generators draw over a fifth of their power outages as a percentage of the firm’s annual sales electricity from those sources. (table 2.5). On each of these indicators, the MENA ES economies perform worse than peer economies with available data. For example, for a typical firm in the MENA Poor quality electricity provision is associated with ES economies, losses due to power outages equal 5 lower labor productivity percent of annual sales, while the corresponding figures The observed high losses due to power outages suggest for peer economies are 3.3 and 1.1 percent. that improvements in the quality of power supply could result in a substantial increase in firms’ output and pro- This picture is somewhat misleading, however, as the ductivity. Indeed, it turns out that there is a significant and economies in the survey should really be split into two negative relationship between poorer supply of electricity groups in relation to power supply. In Egypt,21 Lebanon, and labor productivity (table A2.3).22 the Republic of Yemen, and the West Bank and Gaza, the quality of power supply as measured by the three objec- The relevance of electricity access as a constraint for tive indicators is much worse than in Djibouti, Jordan, firms’ growth in the region should be read in the context of Morocco, and Tunisia. On all three power supply indica- the overall institutional framework characterizing the local tors, the first group of economies performs significantly energy sector. Economies in the MENA ES have tradition- worse than peer economies, while the second group ally used energy subsidies as a safety net, in the context performs better than peer economies within and outside of ineffective systems of social welfare. This generated the region. high associated costs and inefficiencies. By distorting prices, there has been a systematic lack of incentives for The poor quality of power supply in the first group of investment in critical infrastructure, while creating room economies can be attributed to a number of factors. for vested interests. The distorted prices have also led to These include the rapid expansion of demand for electric- inefficiently high usage of electricity. ity, distorting energy subsidies that lead to inefficiently high use of electricity, inefficiencies resulting from state As part of the reform program in recent years, various control of the power supply, and a lack of adequate invest- international institutions, including the IMF and the World ment in the power sector (see box 2.2). In the case of Bank, have been vocal in calling for a comprehensive Egypt, however, there is evidence that the situation has reform of subsidies, to open the way to a more efficient improved since the time of the survey, with considerable energy sector. investment in bolstering electricity supply. The need for policy measures to improve the quality of The business environment power supply in some of the MENA ES economies is experiences of large and small evident. In the meantime, use of generators has helped to firms reduce the impact of the failure in electricity provision. For As discussed at the beginning of the chapter, large firms example, while power cuts in Lebanon last on average 7 in the MENA ES are generally more productive. Chapter 4 times as long as those in Egypt, firms in both economies also shows that relative to SMEs, large firms are the major lose an equivalent percentage of sales to these outages. employers in the private sector but they are also com- This may be largely explained by the fact that 85 percent of paratively static. It has been shown elsewhere that this firms in Lebanon own generators, which together provide dynamic may be due to the privileged positions enjoyed by 40 percent of the supply, while in Egypt, only 6 percent large firms, both directly and indirectly. If this holds true, it of firms own or share generators, which produce only 1 may also be the case that SMEs experience poorer condi- percent of the supply. In Djibouti, where reported disrup- tions in the business environment more broadly. tions due to outages are low but electricity is frequently 28 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Box 2.2: Political instability and electricity supply While political and civil conflict can have a pervasive surged past capacity, due to a growing population and impact on economic activity and the private sector, one energy-intensive investments. Near the time of the sur- specific and tangible consequence can be the deteriora- vey, the World Bank estimated that demand was grow- tion of electricity supply. The conflict in the Republic of ing at 6 percent per annum, overwhelming capacity and Yemen has had stark effects on the electricity supply: resulting in recurring outages.h New investments in both entire cities have been without power for months at a traditional and alternative energy sources have been de- time, exacerbated by bombing campaigns damaging ex- veloped, with several sources due to come online in the isting electricity networks.a According to one estimate next few years.i These include a gas-powered Helwan in 2012, 90 percent of firms reported that the conflict South Power Plant, which will produce 1950 MW. had resulted in power-related losses to their business, Even with expanded capacity through further invest- a figure that has certainly not improved in the middle of ment, chronic under-provision may present further upheaval.b challenges for securing a well-integrated electricity Such conflict can have persistent and lasting effects. system. In the West Bank and Gaza, there have been Lebanon’s 1975–1990 civil war (as well as its later war recent efforts to develop and support local electricity with Israel) seriously damaged the economy’s power production, which has been lacking: nearly 90 percent infrastructure: even today Lebanese consumers often of the economy’s electricity supply is imported. As of face outages lasting up to 12 hours.c As of December 2014, with support from both the World Bank and the 2012, total electricity production in Lebanon stood at EIB, the Palestinian Authority started the Electric Utility 1500 MW while the demand exceeded 2400 MW at Management Project, to improve and streamline elec- peak times.d The state electricity company, Electricite tricity distribution into four substations. While this prom- du Liban (EDL), accounts for about 75 percent of power ises improved capacity and lower costs, one challenge generation. The company is beset with inadequate ca- going forward will be the integration of non-payers into pacity, inefficient production and distribution, subscriber this network. Currently, nearly 60 percent of the cost of delinquency, and corruption. Half of EDL ’s existing ca- electricity provision is lost due to non-payment, up from pacity was installed in the 1970s and 1980s, making it 37 percent in 2013.j extremely inefficient and unreliable. According to a gov- ernment study, EDL ’s cost of production was 22.7 cents a Al-Harazi (2015). per KWh, one of the highest in the world.e EDL is highly subsidized as well. At the end of 2014, the total accu- b Stone and others (2012). mulated deficit of EDL stood at 27 billion U.S. dollars or c http://www.businessweek.com/ap/financialnews/ about 40 percent of the total Lebanese public debt and D9H029MG0.htm 55 percent of the economy’s GNP . The annual payout by d http://www.georgessassine.com/ the state to cover EDL ’s losses stood at US$2.1 billion in lebanon-electricity-regulation/ 2014.f Recent influxes of refugees from the civil war in e http://www.al-monitor.com/pulse/originals/2015/01/ neighboring Syria threatens to put further stress on the lebanon-electricity-supply-debt-disaster.html# limited capacity of the Lebanese electricity supply. One f Ibid. recent estimate put the cost of providing electricity to g World Bank (2013a). refugees at US$393 million in 2014.g h World Bank (2013b). Further investment in electricity capacity may be re- i Ibid. quired for several economies with politically uncertain j World Bank (2014c). environments. In Egypt, for example, demand has Chapter 2: Firm productivity and the business environment 29 In fact, SMEs do report different experiences and percep- This broad undercurrent is impossible to separate from tions in their day-to-day operations. SMEs are more likely various aspects of the business environment. Corruption to indicate that political instability is a major obstacle. They stands out as a key concern of managers and CEOs. High are also more likely to experience longer periods without perceived corruption is associated with lower sales and power and less likely to use a generator to offset those employment growth, as well as lower labor productivity. disruptions. It is thus not surprising that SMEs also more There is also evidence that it may deter interactions with frequently report unreliable electricity supply as a major public authorities, preventing firms from making full use obstacle in their daily work. Similar results hold for access of the opportunities available to them. to finance, which is explored more fully in chapter 3: SMEs are more likely to be credit-constrained and to report ac- Firms’ experiences with petty corruption affecting day-to- cess to finance as a major obstacle. For the four most day operations do not seem to account for the severity of frequently cited top obstacles by MENA ES firms, only in corruption perceptions, suggesting the influence of wider the area of corruption are there no significant differences problems of corruption and state capture in the societies between SMEs and large firms (table A2.4). concerned. Hence, policies aimed at reducing corruption in the region must look beyond petty corruption and at the broader institutional environment that governs public- Policy conclusions private interactions. Such a general perception of corrup- A supportive business environment is a critical factor tion as a constraint and an unwillingness to engage the underpinning the ability of firms to survive, invest, create state can have wide impacts. Indeed, it was the harass- jobs, and innovate, which in turn raises productivity and ment and attempted bribe extraction from a street vendor, competitiveness. Overall, the level of productivity of firms and his subsequent self-immolation, that led to the start in the region is not too different from firms in economies of Tunisia’s uprising. with similar income levels—labor productivity is some- what higher in the region, but TFP lags behind, possibly Electricity is frequently cited as a major constraint—notably due to inefficiently high capital intensity, particularly in in Egypt, Lebanon, the West Bank and Gaza, and the larger firms. Yet labor productivity is declining as revenues Republic of Yemen. Each of these economies has been are also falling, notably in politically unstable economies. characterized by political instability as well as difficulties in the provision of electricity, accounting for a significant Understanding the factors that may be impeding the direct loss of sales and associated with lower sales and growth of private firms, as well as addressing these labor productivity growth at the firm level. constraints through policy reforms, remains a top priority for policy makers. The MENA ES data point to political This relationship may be self-reinforcing: the inadequate instability as the most commonly cited impediment to provision of services such as electricity supply may feed private sector development, reflecting the impact of the broad discontent, just as political upheaval may allow Arab Uprisings and their aftermath, as well as unresolved infrastructure to deteriorate through lack of investment social tensions and conflicts in the region. or to be destroyed by violent conflict. Reform agendas to improve energy-sector efficiency and investment, includ- In Egypt, Lebanon, Tunisia, the West Bank and Gaza, the ing through the streamlining or removal of distorting sub- Republic of Yemen, and, to some extent, Jordan, political sidies, should be seen through where they have begun instability seems to have negatively affected firms’ sales, and taken up again where they have stalled. As subsidies employment, and labor productivity growth. The impact can lead to a sub-optimal use of resources, such reforms of political instability goes beyond the obvious disruptive may also lead to increased TFP . impact of political turmoil and armed conflict. It needs to be seen as creating a general environment of uncertainty Another key business environment constraint in the with regard to economic policy and the regulatory environ- region, not discussed in this chapter, is access to finance. ment that may reach across national boundaries. This is the most frequently named top obstacle by firms in Jordan, and also features prominently in the results for two other economies. Indeed, access to finance might 30 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY appear as more of a concern, were it not for the influence 12 See Kinda and others (2015), Xu (2011), Ayyagari and of factors such as political instability that are likely to deter others (2015). investment by firms, reducing their demand for capital, 13 Restuccia and Rogerson (2008). and encouraging a level of disconnectedness from the 14 The Corruption Perceptions Index scores the degree of formal financial sector. These issues are discussed in the public sector corruption in an economy based on a series of broad perception questions. next chapter. 15 For each element of the business environment respondents are asked to assess its degree of difficulty Finally, addressing constraints related to the business on scale from 0 to 4, with 0 indicating no obstacle and 4 environment might also support competition and overall very severe obstacle. The graph provides the percentage efficiency in the economy. As indicated above, large firms of firms that chose major or very severe, 3 or 4, for each are more productive but inefficiently capital-intensive. At obstacle. the same time, SMEs are disproportionately affected by 16 Calculations based on IMF Government Finance Statistics inefficiencies in the business environment. Ameliorating show changes of near 400 percent increases in debt (negative gross operating balance) of Tunisia and Egypt, in these constraints, carefully assessing distorting incen- real terms. In Lebanon, public deficits grew by 30 percent. tives, removing privileges and more generally enhancing See also Mottaghi (2014). competition, can be effective policies toward a more 17 Two indicators are derived to measure the degree of a inclusive growth. firm’s exposure to corruption. Bribery depth captures the percentage of these six transactions in which a gift or informal payment was requested. Bribery incidence is the percentage of firms experiencing at least one bribery Endnotes request in any of the six transactions. 1 Note that total factor productivity estimate is available only 18 It should also be noted that the bribery depth and for manufacturing firms. incidence indicators are based on questions that are only asked to firms that engage in at least one of the six 2 The comparison income group includes either 36 upper- transactions. Therefore, results across economies may middle-income or 38 lower-middle-income economies not be fully comparable if there are systematic differences (according to the World Bank income classification, as in the way firms engage in these transactions across of 2012) for which Enterprise Survey data are available, economies. excluding MENA economies. Survey years for the comparators can run from 2009 to 2014. 19 Kraay and Murrell (2013). 3 The pattern shown in figure 2.1 remains when labor 20 A firm-level regression model including all the usual firm productivity and TFP are restricted to the same sub- characteristics is used to assess the relationship between sample of only manufacturing firms with TFP estimates several measures of firm performance (sales growth, available. employment growth, and labor productivity levels) with the three measures of corruption (whether corruption is 4 Schiffbauer and others (2015). perceived as a severe obstacle, and the bribery depth 5 While this is the case, several caveats should be noted. and incidence indicators). As noted above, the bribery First, the use of recall data is subject to potential bias, indicators are only available for firms that engage in at as is the fact that the ES necessarily can only gather least one of the transactions. To make sure that the result information on surviving firms. Yet while the nature of on the perception indicator is not driven by the larger such upheaval in these economies is not necessarily easily sample, an alternative specification was used by excluding confined to this period, the two periods do offer useful firms for which the bribery indicators are not available. indicators of firm performance. Results for the perception-based indicator still hold 6 Based on authors’ calculations from IMF Government although at a lower significance level. Finance Statistics; see also Mottaghi (2014). 21 Egypt experienced a major deterioration in electricity 7 Mottaghi (2014). supply reliability in 2012, the reference year for the survey. The situation has since improved. 8 Sdralevich and others (2014). 22 But there is no significant relationship for manufacturing 9 Coats (2015). between total factor productivity and poor quality of power 10 Devarajan and others (2014). supply as defined in table A2.3. 11 Syverson (2011). Chapter 2: Firm productivity and the business environment 31 Appendix A2 Table A2.1: Firm performance and firm size Services Manufacturing Labor productivity Labor productivity TFP (1) (2) (3) (4) Log of size 0.01 0.12** -0.58*** 0.18*** (0.085) (0.057) (0.066) (0.067) Log of cost of capital 0.10*** (0.040) Log of cost of intermediate goods 0.50*** (0.053) Foreign ownership (Y/N) 0.11 -0.08 0.01 0.27 (0.242) (0.170) (0.144) (0.220) Exports 10% or more of sales (Y/N) 0.22 -0.16 0.08 -0.01 (0.168) (0.166) (0.105) (0.247) Firm is part of a larger firm (Y/N) 0.09 -0.19 -0.18 -0.14 (0.221) (0.255) (0.258) (0.316) Constant 9.26*** 8.97*** 4.111*** 1.67*** (0.580) (0.305) (0.479) (0.384) Number of observations 2,201 2,218 2,218 2,218 R-squared 0.208 0.181 0.600 0.088 Source: Enterprise Surveys. Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix). Linearized Taylor standard errors that account for survey stratification are indicated in parentheses. ***, **, and * denote statistical significance at the 1, 5, and 10 percent levels respectively. Economy and locality fixed effects not shown. Columns 2, 3, and 4 are run over the sub- sample of manufacturing firms for which TFPR is available.   32 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A2.2: Association of perceptions of the severity of Table A2.3: Deficiencies in the provision of electricity and corruption with performance measures labor productivity Log labor Log labor productivity Real annual Annual productivity (sales per worker, USD) sales growth employment (sales per (1) (2) (%) growth (%) worker, USD) Number of electrical outages in -0.01* Dependent variable (1) (2) (3) a typical month (0.003) Corruption as major/ -3.18** -2.25*** -0.14* severe obstacle (Y/N) Duration of electrical outages -0.02** (1.239) (0.851) (0.085) (hours) (0.009) Log of size 1.55*** -0.040 Log of size -0.04 -0.03 (0.500) (0.051) (0.050) (0.051) Log of size, 3 FY ago -4.00*** Young firms (5 years or less) -0.03 -0.05 (0.534) (Y/N) (0.116) (0.116) Young firms (5 years or 3.21 5.29*** -0.03 less) (Y/N) Firm is part of a larger firm (Y/N) 0.09 0.07 (2.621) (1.838) (0.119) (0.164) (0.188) Firm is part of a larger -3.31 2.19* 0.07 firm (Y/N) Manager has university 0.49*** 0.49*** (3.556) (1.184) (0.189) education (Y/N) (0.100) (0.104) Manager has university 0.33 1.86 0.50*** education (Y/N) Manager experience in sector 0.004 0.003 (1.802) (1.210) (0.105) (years) (0.004) (0.004) Manager experience in -0.18*** -0.11*** 0.00 sector (years) Exports 10% or more of sales 0.14 0.10 (0.057) (0.036) (0.004) (Y/N) (0.109) (0.114) Exports 10% or more of 0.67 1.46 0.09 sales (Y/N) Foreign ownership (Y/N) 0.02 0.02 (1.397) (1.061) (0.113) (0.161) (0.163) Foreign ownership (Y/N) -0.78 2.36* 0.03 Retail firms (Y/N) 0.28** 0.29** (2.695) (1.408) (0.161) (0.120) (0.119) Retail firms (Y/N) -2.52 -2.34* 0.32** Other services firms (Y/N) -0.01 -0.03 (1.719) (1.229) (0.124) (0.111) (0.111) Other services firms (Y/N) -2.20 -0.30 -0.004 Constant 9.36*** 9.29*** (2.018) (0.978) (0.113) (0.258) (0.306) Constant -8.60* 13.62*** 9.33*** Number of observations 4,912 4,890 (4.815) (2.476) (0.311) R-squared 0.220 0.212 Number of observations 4,019 4,848 4,908 Source: Enterprise Surveys. R-squared 0.128 0.191 0.216 Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix Source: Enterprise Surveys. command). Linearized Taylor standard errors that account for survey stratification are Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix indicated in parentheses. ***, ** and * denote statistical significance at the 1, 5 and command). Linearized Taylor standard errors that account for survey stratification are 10 percent levels respectively. All regressions include economy fixed effects. indicated in parentheses. ***, ** and * denote statistical significance at the 1, 5 and   10 percent levels respectively. All regressions include economy fixed effects. Chapter 2: Firm productivity and the business environment 33 Table A2.4: The experiences of SMEs and large firms with the business environment Political Electricity Corruption Access to finance instability OLS Probit OLS Probit OLS Probit Probit Probit Probit (1)a (2)a (3) a (4) (5) (6) (7)a (8)a (9) Typical Uses power Proportion of Electricity Credit- Political power generator electricity from major Bribery Bribery constrained Finance major instability major outage (hrs) (Y/N) generator (%) obstacle (Y/N) depth incidence - (Y/N) obstacle (Y/N) obstacle (Y/N) SME (Y/N) (<100 0.42** -0.73*** -3.61** 0.26* 0.27 0.01 0.65*** 0.30** 0.26** employees) (0.192) (0.139) (1.590) (0.155) (3.766) (0.155) (0.165) (0.125) (0.111) Foreign -0.16 0.24 2.58 0.01 1.47 0.13 -0.07 -0.26* 0.05 ownership (Y/N) (0.219) (0.163) (2.458) (0.177) (3.876) (0.221) (0.168) (0.150) (0.127) Exports 10% or -0.56** 0.38*** 3.29** -0.04 4.96 0.19 -0.12 0.12 0.26** more of sales (0.285) (0.118) (1.503) (0.165) (4.801) (0.184) (0.137) (0.135) (0.108) (Y/N) Firm is part of a 0.11 0.16 0.68 0.28*** 1.43 0.04 0.13 -0.14 -0.14 larger firm (Y/N) (0.311) (0.142) (2.354) (0.103) (4.823) (0.171) (0.105) (0.146) (0.170) Constant 1.26*** -1.10*** 9.67*** -0.31* 13.80** -1.01*** -0.95*** -0.58*** 0.70*** (0.342) (0.194) (2.059) (0.185) (5.723) (0.237) (0.202) (0.175) (0.166) Number of 5,690 5,903 5,765 5,894 4,258 4,258 5,565 5,865 5,850 observations R-squared 0.215 0.470 0.272 Source: Enterprise Surveys. Note: Using Stata’s svy prefix command. Linearized Taylor standard errors that account for survey stratification are indicated in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. All regressions include economy, sector, and location fixed effects. Bribery depth is the number of transactions that were subject to a bribe request. Bribery incidence is a dummy variable if a firm was subjected to such a request in any transaction. a. Indicates that the log of size is also statistically significant with an opposite sign from SME dummy variable. 34 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY References Al-Harazi, Ebrahim Mohammad Yahya. 2015. “The Restuccia, Diego and Richard Rogerson. 2008. “Policy unheard voices of exhausted Yemenis” . Blog, distortions and aggregate productivity with accessed at: http://blogs.worldbank.org/arabvoices/ heterogeneous establishments. ” Review of Economic unheard-voices-exhausted-yemenis Dynamics 11(4), 707–20. Ayaggari, Meghana, Asli Demirguc-Kunt, and Vojislav Schiffbauer, Marc, Abdoulaye Sy, Sahar Hussain, Hania Maksimovic. 2015. “Are larger firms born or made: Sahnoun, and Philip Keefer. 2015. Jobs or Privileges: Evidence from developing countries. ” World Bank Unleashing the Employment Potential of the Middle Policy Research Working Paper No.7406. World Bank: East and North Africa. Washington, DC: World Bank. Washington DC. Sdralevich, Carlo, Randa Sab, Younes Zouhar, and Giorgia Coats, Christopher. 2015. “Energy subsidy cuts finally Albertin. 2014. Subsidy Reform in the Middle East and providing relief for Egypt”. Forbes. Online. North Africa: Recent Progress and Challenges Ahead. Available at: http://www.forbes.com/sites/ Washington, DC: International Monetary Fund. christophercoats/2015/03/30/energy-subsidy-cuts-finally- Syverson, Chad. 2011. “What determines productivity?” providing-relief-for-egypt/#6f3818c53b97 Journal of Economic Literature 49(2): 326-65. Devarajan, Shanta, Lili Mottaghi, Farrukh Iqbal, Gabriela Stone, Andew, Lina Badaway, and Nabila Assaf. 2012. “The Mundaca, Maria Laursen, Maria Vagliasindi, Simon plight of Yemeni private enterprises since the 2011 Commander, and Isabelle Chaal-Dabi. 2014. MENA crisis: a rapid assessment”. MENA Knowledge and economic monitor: corrosive subsidies. Middle Learning, Quick Note Series No. 72, World Bank, East and North Africa (MENA) Economic Monitor. Washington, DC. Washington, DC : World Bank Group. Xu, Lixin Colin. 2011. “The effect of business environment on Foster, Lucia, John Haltiwanger, and Chad Syverson. 2008. development: Surveying new firm level evidence. ” The “Reallocation, firm turnover and efficiency: Selection World Bank Research Observer 26(2): 310-40. on productivity or profitability?” American Economic World Bank. 2013a. Lebanon: Economic and Social Impact Review, 98(1): 394-425. Assessment of the Syrian Conflict. Washington, DC: Halvorsen, R., and R. Palmquist, 1980. “The interpretation World Bank Group. of dummy variables in semilogarithmic equations. ” World Bank. 2013b. “World Bank Supports Boosting Much American Economic Review 70(3): 474-475. ” Needed Electricity Generation Capacity across Egypt. Kinda, Tidiane, Patrick Plane, and Marie-Ange Véganzonès- World Bank, Washington, DC. Available at: http://www. Varoudakis. 2015. “Does investment climate matter worldbank.org/en/news/press-release/2013/06/27/ for firms technical efficiency: An application to Middle world-bank-supports-boosting-much-needed-electricity- Eastern and North African manufacturing. ” Journal of generation-capacity-across-egypt International Development 27(7): 1267-93. World Bank. 2014c. West Bank and Gaza - Assessment and Kraay, Aart, and Peter Murrell. 2013. “Misunderestimating action plan to improve payment for electricity services ” World Bank Policy Research Working Paper corruption. in the Palestinian territories: study on electricity sector No. 6488. contribution to net lending. Washington, DC; World Bank Mottaghi, Lili. 2014. Predictions, Perceptions and Economic Group. https://hubs.worldbank.org/docs/ImageBank/ Reality: Challenges of seven Middle East and North Pages/DocProfile.aspx?nodeid=23040668 Africa countries described in 14 charts. MENA Quarterly Economic Brief No. 3. 35 3. Access to finance Introduction A well-functioning financial sector can facilitate the exchange of goods and services, the diversification if this comes at the cost of losing possible growth opportunities. This “disconnect” between firms and banks goes so far that in some economies, even the use of checking and savings accounts by firms is of risk, the mobilization of savings, and the identifi- low. Instead, firms rely to a large extent on internal cation of good business opportunities—all of which financing. encourage investment and entrepreneurship.1 These functions enable rapid accumulation of physical and On the supply side, the financial sector is dominated human capital, boost technological advances, and by banks. Banks in the MENA ES region seem to have thus promote faster growth and higher levels of adopted a cautious approach, based on traditional employment.2 lending technologies and conservative practices. Thus, despite comparatively high volumes of private This chapter explains the relationship between the credit, only a small segment of the private sector is financial sector and the formal non-financial private financed by the formal financial sector. Credit is highly sector in the MENA ES economies. A few fairly concentrated, favoring a small number of large clients. consistent patterns emerge. On the borrower side, a large proportion of firms exclude themselves from This chapter first provides some context for the formal financial markets. More importantly, the evi- survey results by examining financial sector charac- dence is highly suggestive that firms have adjusted teristics, drawing on other data sources, including production strategies and expectations to the reality relevant Doing Business indicators. It then turns of limited involvement with the financial sector, even to the question of whether firms in the region are 36 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY credit-constrained, presenting an evidence-based indicator Table 3.1: Banking sector characteristics of credit constraint. In light of the finding that a substantial proportion of firms seem to be disconnected from the Credit to Credit to private formal financial sector, and are therefore likely to forgo Deposits Loans to government sector growth opportunities, the third section draws on data from Economy (% of GDP) deposits (% of GDP) (% GDP) the surveys to examine some supply-side factors that may Djibouti 71 38 4 28 have contributed to this situation. The last section outlines Egypt, Arab 60 48 35 29 Rep. policy implications. Jordan 94 76 41 70 Lebanon 228 38 72 84 Morocco 89 81 17 71 The context: financial sectors in Tunisia 55 128 5 69 the MENA region West Bank and 64 43 12 24 Gaza The formal financial sector is dominated by a banking Yemen, Rep. 21 20 13 5 sector that is typically large compared with peer MENA ES 85 59 25 48 economies Lower-middle- 35 102 7 31 income The banking sector dominates the formal financing Upper-middle- channels available in the MENA ES region. Bank de- income 49 100 8 47 posits account for 85 percent of GDP in the MENA ES High-income: 78 82 15 68 economies, compared with only 49 percent for the aver- non-OECD age upper-middle-income economy (see table 3.1). The High-income: 99 120 17 122 OECD region’s banking sectors are therefore large in relation to Sources: World Bank Global Financial Development Database, Palestine peer economies in other regions. The size of the banking Monetary Authority, reference year 2012. sectors reflects the capacity of the banking sector to at- tract relatively large amounts of deposits. The supply of deposits is supported by remittances and capital inflows.3 Compared with banks, the role of institutional investors In 2012, the MENA ES economies attracted remittances and equity markets is limited. With the exception of worth 9.6 percent of GDP , compared with an average of Morocco, the mutual fund industry is small compared 3.5 percent for upper-middle-income economies. with peer economies. The size of the insurance industry is also limited. While equity markets display comparatively Lebanon serves as the most striking example. The high levels of market capitalization, they are dominated by economy benefits from a large and loyal diaspora, which financial and infrastructure firms. According to the World contributes remittances equivalent to around 16 to 20 per- Bank,7 the market capitalization of the industry (excluding cent of Lebanon’s GDP . Due in large part to the diaspora, infrastructure firms) and non-financial services sectors in bank deposits have been growing steadily over the years the wider MENA region represents less than 12 percent despite episodes of high political instability.4 The inflows of GDP , which suggests that equity markets play a limited have been supported by the ability to hold deposits in for- role in funding the real economy. eign currency and the unrestricted convertibility between local and foreign currency deposits.5 The leasing industry is similarly small by international standards.8 Leasing firms retain ownership of the leased Another example is Morocco, where the size of the bank- asset, which should facilitate repossession in case the ing sector may be attributed to successful financial sector lessee defaults. Thus, leasing can be an attractive alterna- reforms, notably between 1986 and 1996. The reforms led tive to bank finance in an environment characterized by to the elimination of credit controls, deregulation of inter- weak creditor rights. Among the MENA ES economies, est rates, improved prudential regulation and supervision, leasing is most prevalent in Tunisia, followed by Jordan, and the first steps toward the liberalization of international Morocco, and Egypt. Most leasing firms are banks or capital flows.6 bank-related institutions, reflecting their easy access to Chapter 3: Access to finance 37 deposit funding. Factoring plays only a minor role in the governments increased by 6 percentage points between MENA ES economies. 2010 and 2013.12 In contrast, the average level of credit to governments in lower-middle-income and upper-middle- income economies in the rest of the world did not Ratios of loans to deposits are low in many MENA ES increase. economies and they are often associated with high levels of credit to governments Tunisia is the only MENA ES economy with a loan-to- At 59 percent, the region’s loan-to-deposit ratio is well deposit ratio exceeding 100 percent. Relative to Jordan below the average of all income brackets. This means that and Morocco, which have similar levels of private credit, comparatively few of the deposits received by banks are the Tunisian deposit base is relatively small. Banks there- translated into lending to the non-financial private sector. fore have to rely on wholesale (and cross-border) fund- The low ratios reflect both a large supply of deposits and ing.13 Tunisia is the only economy in the region where plentiful opportunities to hold local government debt.9 banks experienced significant withdrawals of deposits during the Arab Uprisings; it also suffered from a high ratio The MENA ES economies receive substantial remittance of non-performing loans, 13 percent in 2011.14 inflows. Under a floating exchange rate, such capital inflows would put upward pressure on the exchange rate. But all MENA ES central banks that issue their own Credit to the private sector is relatively high in the legal tender pursue an exchange rate arrangement that is region’s upper-middle-income economies, but lending pegged in some way. To resist appreciation, the central is concentrated bank buys foreign currency and thereby creates liquidity in Despite the low loan-to-deposit ratios, private credit the domestic currency. As a result, capital inflows lead to to GDP for the MENA ES is well above the average for the creation of local currency bank deposits. peer economies. Private credit is especially high in the upper-middle-income economies—Jordan, Lebanon, and Large local currency deposits in the MENA ES economies Tunisia—and lower-middle-income Morocco. In the other are also the result of banks’ policies to hold large volumes lower-middle-income economies—Djibouti, Egypt, and of local public debt, which is widely available in the region. the West Bank and Gaza—private credit to GDP is in line Household savings are mostly held in the form of bank with peer economies in other regions. Only the Republic deposits rather than direct holdings of government debt. of Yemen is lagging behind. Monetary financing of public debt also increases bank deposits as the government spends the borrowed money While high volumes of private credit are desirable, they do to pay employees and suppliers. not necessarily translate into financial access for a broad cross-section of firms. Figure 3.1 shows that credit con- But large-scale lending to governments also has a cycli- centration ratios in non-Gulf Cooperation Council MENA, cal component that is closely associated with the Arab which is the aggregate corresponding most closely to the Uprisings. Egypt is the most striking example. Following MENA ES economies, are among the highest in the world. the protests of 2011, bank claims on the public sector Within the region, Egypt has the highest credit concentra- increased from 27 percent of GDP in 2010 to over 50 tion ratio. In 2010, the top 20 exposures accounted for percent in 2015. This can be attributed to both deteriorat- more than half of total loans in the economy, implying that ing fiscal balances and capital flight.10 As foreign investors credit is absorbed primarily by large corporate clients.15 withdrew, the domestic banking system stepped in. With local treasury bill rates approaching 16 percent in 2012, A similar divergence between depth and access can be bank claims on the private sector decreased. Because the observed on the deposit side. The share of the population government was able to offer more attractive risk-adjusted that saves in formal financial institutions is much lower returns, parts of the private sector were crowded out.11 than in economies with similar deposit volumes, suggest- ing a lopsided distribution of wealth. It is the strength Similar patterns, albeit less pronounced, prevail in the of surveys such as the MENA ES that they can give a other MENA ES economies, where, on average, credit to 38 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 3.1: Top 20 loan exposures as a percentage of total bank equity, by world region 250 200 150 Percent 100 50 0 Non-GCC Europe and Middle East GCC Central Western Asia Australia Latin North MENA Central Asia and North Europe Europe America America Africa Source: World Bank (2011). Note: Data are regional averages computed between 2005 and 2010. detailed representation of financial access that is not The index thus assesses the quality of the secured unduly affected by the largest players. transaction framework. The MENA ES economies have a particularly poor record on legal rights, suggesting that collateral regimes in the MENA ES economies have seri- The institutional financial infrastructure does not ous deficiencies across the board, a result highlighted in facilitate expansion of credit to small and medium- other studies.16 sized enterprises Financial intermediation in the MENA ES economies takes The depth of credit information index measures rules and place against an unfavorable institutional background. practices affecting the coverage, scope, and accessibility Table 3.2 presents institutional quality as represented of credit information available through either a private by the getting credit dimension of Doing Business. This credit bureau or a public credit registry. The index provides set of indicators is based on a case study that seeks to a measure of the extent to which these institutions help to represent the institutions faced by a domestically owned mitigate the informational asymmetries that impede lend- limited liability company that has up to 50 employees and ing to SMEs. In terms of the depth of credit information, operates in the largest business city. With an average rank the economies of the region fall into two groups. Djibouti, of 135, the region scores worse than economies in any Jordan, and the Republic of Yemen receive a score of 0, income bracket. Jordan and the Republic of Yemen both while the other economies obtain scores between 5 and rank 185 out of the 185 economies examined. 8, indicating advanced credit information systems. The getting credit ranking has two components: a legal The last two columns of table 3.2 present data on the rights index; and a depth of credit information index. The coverage of credit information systems, which do not af- strength of legal rights index measures the degree to fect the index score. In the region, public credit registries which collateral and bankruptcy laws protect the rights have on average better coverage than private credit bu- of borrowers and lenders, thereby facilitating lending. reaus. The only economies with functioning private credit bureaus are Egypt and Morocco. Chapter 3: Access to finance 39 Table 3.2: Doing Business getting credit indicators Strength of legal rights Depth of credit Public credit registry Private credit bureau Economy Getting credit rank index (0-12) information index (0-8) coverage (% of adults) coverage (% of adults) Djibouti 181 1 0 0 0 Egypt, Arab Rep. 79 2 8 7 21 Jordan 185 0 0 2 0 Lebanon 109 2 6 24 0 Morocco 109 2 6 0 23 Tunisia 126 2 5 29 0 West Bank and Gaza 109 0 8 23 0 Yemen, Rep. 185 0 0 1 0 MENA ES 135 1 4 11 6 Lower-middle-income 90 5 4 8 15 Upper-middle-income 82 5 5 20 33 High income: non-OECD 91 4 5 16 37 High income: OECD 54 6 6 12 67 Source: Doing Business report, 2016. Note: GCR: low value better performance. SLRI and DCII: high value, better performance. Firms in the MENA region are not Unsurprisingly, firms in the MENA ES region are more typically credit-constrained, but likely to use external finance from banks wherever finan- many are disconnected cial deepening is greater, as measured by private credit to GDP . The share of bank finance in Lebanon (20 percent), The composition of firm finance in the region is similar Morocco (21 percent), and Tunisia (16 percent) is well to peer economies, but with a slightly larger role for above that of their peer economies’ average of 12 percent internal funds and great variation in the use of bank for lower-middle-income economies and 14 percent for and supplier credit upper-middle-income economies with ES data. Jordan is the only economy where high levels of financial deepening To examine whether firms are credit-constrained, it is first are not associated with a strong use of bank financing by useful to examine the types of finance that they use. The the average firm. In Egypt, the West Bank and Gaza, and MENA ES data provide detailed information on firms’ use the Republic of Yemen, banks play a negligible role for firm of the different sources of funds for both their working financing, with Jordan and Djibouti an intermediate case. capital and their purchases of fixed assets. For each firm, information is available on the relative use of internal The use of credit from input suppliers and customers in funds, bank finance, credit from suppliers or customers, the MENA ES economies is broadly comparable to peer equity finance, and other sources of finance, including economies, accounting for, on average, 8 percent of firm informal sources and non-deposit-taking institutions. financing in the region. The use of input supplier credit does not seem to be associated with the level of income Figure 3.2 presents the composition of firm financing. of the economy. Supplier credit is most widely used in With 77 percent of working capital and investment fi- Tunisia and the West Bank and Gaza, whereas firms in nanced internally, firms in the MENA ES region rely more Djibouti and Lebanon rarely resort to this source of on internal funds than the average lower-middle-income financing. and upper-middle-income economy. 40 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 3.2: External sources of finance are similar to comparable economies elsewhere 100 80 60 Percent 40 20 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, MENA ES Lower- Upper- Arab Rep. and Gaza Rep. middle- middle- income income Internal Banks Suppliers Others Equity Source: Enterprise Surveys. The use of equity finance is negligible throughout the re- do not contain detailed information on loan outcomes, gion, reaching a maximum of only 2 percent in the case of the figure can only provide boundaries for not credit- Tunisia, which confirms the limited role of equity markets constrained firms in other regions.19 Regardless, the share for funding the real economy. Other sources of financing, of not credit-constrained firms in the MENA ES region ex- which include non-deposit-taking financial institutions, ceeds the upper bound for all other world regions except microfinance operators, and Islamic finance, are not for ECA, where the upper bound estimate matches the prevalent either.17 These sources of finance matter most MENA ES average. in Tunisia and the West Bank and Gaza. Djibouti and Morocco have the highest share of not Although the discussion on sources of finance used by credit-constrained firms (87 percent) in the region while firms elucidates important features of the relationship the Republic of Yemen has the lowest share of not credit- between the private sector and the financial sector, it constrained firms (51 percent), followed by Jordan (64 does not measure credit constraints. Combining informa- percent), as shown in figure 3.4. tion on loan applications and their outcomes with data on the sources of finance for both working capital and the purchase of fixed assets yields a measure of the Credit-constrained firms have weaker performance on prevalence of credit-constraints faced by firms in the fis- average cal year 2012. The credit-constraint measure splits firms Fully and partially credit-constrained firms (FCC and PCC) into three categories—fully credit-constrained, partially in the MENA ES region are associated with lower employ- credit-constrained, and not credit-constrained (see box ment growth, lower levels of capacity utilization, and 3.1 for details). Fully and partially constrained firms are lower levels of labor productivity as measured as sales considered to be credit-constrained in this report. per employee (table A3.1).20 The negative relationship between performance measures The MENA ES economies are characterized by an and credit constraints can be interpreted in a number unusually high share of firms that are not credit- of ways. It is possible that firms face credit constraints constrained because they were evaluated by financial intermediaries Figure 3.3 shows that on average 73 percent of firms to lack creditworthiness, because they proposed proj- in the MENA ES are not credit-constrained.18 Because ects that were not financially viable, or simply because previous Enterprise Survey implemented in other regions they did not have good accounting records. All of these Chapter 3: Access to finance 41 Box 3.1: A measure of credit constraints Figure B3.1 shows how external and bank finance us- (FCC), partially credit-constrained (PCC), and not credit- age and applications are used to compute the credit constrained (NCC) firms. Credit-constrained firms are constraint indicator. Based on this indicator, three cat- defined as those that are fully (FCC) or partially con- egories of firms are defined: fully credit-constrained strained (PCC). Figure B3.1: Correspondence between credit-constrained classification and ES questions Did the firm have any source of external finance? Yes No Did the firm apply for a loan or line of credit? Did the firm apply for a loan or line of credit? No Yes No Yes Why not? Why not? Has enough Terms and Approved Approved Rejected Has enough Terms and Rejected capital conditions in full in part capital conditions Not Credit Constrained (NCC) Partially Credit Constrained (PCC) Fully Credit Constrained (FCC) Source: Methodology based on Kuntchev et al. 2014. Fully credit-constrained firms (FCC) are those that find external source of financing and applied for a loan that it challenging to obtain credit. These are firms that have was partially approved or rejected. no source of external financing and typically fall into two Not credit-constrained firms (NCC) are those that do categories: those that applied for a loan and were re- not seem to have any difficulties accessing credit or do jected; and those that were discouraged from applying not need credit. Firms under this category encompass either because of unfavorable terms and conditions or those that did not apply for a loan as they have sufficient because they did not think the application would be ap- capital either on their own or from other sources; and proved. The terms and conditions that discourage firms firms that applied for a loan and the application was ap- include complex application procedures, unfavorable proved in full. interest rates, high collateral requirements, and insuf- ficient size of loan and maturity. There are limitations to the credit constraint indicator. The indicator does not incorporate any information on Partially credit-constrained firms (PCC) are those that creditworthiness of the firm, and therefore among the have been somewhat successful in obtaining external credit-constrained firms there may be some that were financing. PCC firms include those that have external rationed for good reasons, such as insufficiently produc- financing but were discouraged from applying for a tive projects or a bad repayment history. loan from a financial institution; and firms that have an 42 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 3.3: A high percentage of firms are not credit-constrained 100 80 60 Percent 40 20 0 MENA ES Africa East Asia Europe and Latin America South Asia Lower- Upper- and Pacific Central Asia and Caribbean middle- income middle- income Not credit constrained (lower bound) Not credit constrained (upper bound) MENA ES average Source: Enterprise Surveys. Figure 3.4: The percentage of not credit-constrained firms varies considerably across MENA ES economies 100 80 Percent 60 40 20 0 Djibouti Morocco West Bank Lebanon Egypt, Tunisia Jordan Yemen, MENA ES and Gaza Arab Rep. Rep. Not credit constrained Partially credit-constrained Fully credit-constrained Source: Enterprise Surveys. factors could be correlated with weak firm performance. Many firms in the region are disconnected from the But lack of access to credit may also be the cause of banking sector low performance as firms are unable to expand due to Why do the data show such high levels of not credit- limited finance. The negative association between credit constrained firms in the MENA ES region? A closer constraints and performance measures implies that the examination offers important insights. Firms are not credit- evidence does not contradict the possibility that credit constrained for one of two reasons: either they have their is being properly allocated and that financial markets are loan application approved; or they see themselves as working appropriately even if only a limited cross-section having sufficient amounts of capital and therefore see no of the private sector benefits. need to engage financial intermediaries. In the MENA ES Chapter 3: Access to finance 43 Figure 3.5: Firms’ credit relationship with the financial sector 100 80 Percent of firms 60 40 20 0 Djibouti West Bank Egypt, Morocco Jordan Yemen, Lebanon Tunisia and Gaza Arab Rep. Rep. Disconnected (sufficient capital—no loan needed) Discouraged Connected Source: Enterprise Surveys. economies, the latter type accounts for the vast majority What explains the “disconnect” of unconstrained firms. An important question is whether between firms and the banking these latter firms are in fact losing growth opportunities sector and what are the because of their stance. consequences? Figure 3.5 shows these results by decomposing the Firm-bank disconnectedness reflects a number population of firms into three categories: connected, of different factors and may lead to lost growth disconnected, and discouraged. Connected firms are opportunities those that applied for loans regardless of whether their Firms in economies where there are lower levels of credit application was approved or rejected. They are “con- to the private sector relative to GDP—such as Djibouti, nected” in the sense that they see financial markets as Egypt, and the West Bank and Gaza—tend to have a an option. Disconnected firms are those that did not apply higher percentage of firms disconnected from the finan- for any loan as they had sufficient capital. Discouraged cial sector. It may be that the prevailing banking systems firms are those that did not apply for any loans due to have led firms to adjust their expectations and produc- terms and conditions. Given these definitions, it fol- tion strategies to an environment in which they do not lows that all disconnected firms are unconstrained (not consider banks as an option. It is plausible that some of credit-constrained), but not all unconstrained firms are these firms would engage with the formal financial sector disconnected. by applying for loans if the banking system were more attuned to their needs. The share of firms that are disconnected, explicitly stating that they do not need a loan, is highest in Djibouti, the In some regards, disconnected firms resemble credit- West Bank and Gaza, and Egypt. These figures largely constrained firms more closely than firms with a successful drive the share of unconstrained firms in these econo- loan application. Both disconnected and credit-constrained mies. At the other end of the spectrum are Tunisia and firms are significantly less likely to invest and less likely Lebanon, suggesting that firms in these economies do to have expansion plans. The major difference is that dis- generally see bank finance as an option. In Morocco, a connected firms are content with their situation whereas particularly low share of discouraged firms mirrors the credit-constrained firms are not (table A3.2). Indeed, the high prevalence of bank financing of firms. propensity to view access to finance as a major constraint 44 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY is much lower for disconnected firms than for credit- Figure 3.6: Firms’ disconnect from the banking sector constrained firms and firms that obtained a loan.21 concerns both credit and the use of payment services 100 West Bank For manufacturing firms, it is possible to examine how and Gaza the propensity to invest changes with capacity utilization. Yemen, Egypt, 90 Arab Rep. Disconnected firms in percent Disconnected firms with above median capacity utilization Rep. of unconstrained firms have a propensity to invest that is 22 percentage points Djibouti 80 lower than firms that obtained a loan. The corresponding Jordan Morocco difference for firms with below median capacity utilization 70 is zero. Thus, disconnected firms are less likely to invest, 60 especially when they are doing well, and they may well be Lebanon forgoing growth opportunities (table A3.3). 50 Tunisia It is possible that firms may also disconnect because of 40 40 60 80 100 limited growth opportunities. Firms that had no intentions Percent of firms with a checking or savings account of investing during the fiscal year 2012, the reference Source: Enterprise Surveys. period of the survey, may have had no need to apply for a loan. This could be a likely scenario given the political situation in some economies of the region. But the high Figure 3.7: Checking or saving accounts are more prevalence of disconnected firms across the region makes prevalent in economies where a larger share of firms were registered when starting operations it difficult to claim that this reflects just idiosyncratic varia- tion in project timing. 100 Morocco Tunisia 90 Lebanon Percent of firms with a checking Similarly, it is unlikely that the macroeconomic environ- Djibouti ment is fully responsible for the larger share of discon- or savings account 80 Jordan West Bank nected firms. It could be argued that lack of demand for and Gaza loans is a consequence of the downturn that most of the 70 MENA ES economies experienced following the events of 2011. While a downturn may explain the lack of demand 60 Yemen, Egypt, for investment finance, it does not necessarily explain the Rep. Arab Rep. 50 lack of demand for working capital. In fact, the demand for working capital may increase to bridge temporary liquidity 40 60 70 80 90 100 problems. Furthermore, there is considerable variation in Percent of firms registered when starting operations the proportion of disconnected firms across the MENA ES region even though there is little variation in the mac- Source: Enterprise Surveys. roeconomic environment, which was consistently difficult in most economies. even use banks for cash-flow management and payment services supports the notion that these firms are indeed Disconnected firms are also less likely to use banks for opting out of the banking system. cash-flow management and payment services. It turns out that the share of firms with a bank account is lowest in Firms that were not registered when starting operations the Republic of Yemen, where only 48 percent of firms in are less likely to have a checking or savings account (fig- the formal sector have a bank account, followed by Egypt ure 3.7). The share of firms that were not registered when and the West Bank and Gaza. These economies also have starting operations is likely to be higher in economies with the highest share of disconnected firms as a proportion of a larger informal sector. It is therefore likely that the pro- not credit-constrained firms, which exceeds 90 percent in pensity of firms to disconnect from the banking system all three economies (figure 3.6). The fact that a substantial also depends on the costs and benefits of participating in share of the private sector in these economies does not Chapter 3: Access to finance 45 the formal economy. This association is consistent with than those that do not. This is to be expected given that anecdotal evidence from Egypt, according to which the audited financial reports reduce informational asymme- Egyptians themselves characterize their economy as a tries, or alternatively signal better-managed firms.22 Both cash economy, and in line with the strong role typically relationships vary with the depth of the banking sector. It ascribed to Egypt’s informal sector. is only in economies with deep banking sectors—Jordan, Lebanon, Morocco, and Tunisia—that the relationship between access to credit and both firm size and audited Loan rejection rates are very low, while firms financial reports applies. For economies in the MENA ES connected to the banking sector tend to be large and with lower levels of financial deepening—Djibouti, Egypt, more likely to have audited financial reports and the Republic of Yemen—these relationships are not One salient result emerging from the MENA ES data is statistically significant. the small share of rejected loan applications. Thus, most of the firms that decide to apply for a loan are successful. The absence of an association between firm size and ac- As figure 3.8 shows, the rate of rejection of loan applica- cess to credit in economies lacking depth in the financial tions per firm varies from zero percent in Djibouti to three sector is probably due to a very small overall share of firms percent in Tunisia. This seems to indicate that the private with a bank loan or line of credit. The lack of significance of sector in the MENA ES economies is divided into two financial reports may be the result of banks attaching little sets of firms. On the one hand, there is a large set of importance to screening borrowers in economies lacking disconnected firms that have adjusted to operate without financial depth. financing options from financial markets; on the other hand, there is a smaller set of firms—with the exception of Tunisia—that is linked to financial markets and is able to The availability and type of collateral can play an raise funds through credit from financial organizations. In important role in facilitating access to credit between these two sets are the discouraged firms. One important aspect of the financial sector that may influence the connectivity with the private sector is the Firms in the MENA ES region that have a loan or line of use of collateral. Collateral can facilitate lending when credit differ significantly from those that do not (table banks face a risky operating environment dominated by A3.4). SMEs are less likely than large firms to have a opaque firms—that is, firms for which information is dif- loan or a line of credit. Firms that have audited financial ficult to obtain and costly to process. Collateral serves to reports are also more likely to have a loan or a line of credit reduce the risk faced by lenders as losses are recoverable Figure 3.8: Loan applications are rarely rejected 40 35 30 Percent of firms 25 20 15 10 5 0 Yemen, Egypt, West Bank Djibouti Jordan Morocco Lebanon Tunisia Rep. Arab Rep. and Gaza Loan fully or partially approved Loan rejected Source: Enterprise Surveys. 46 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 3.9: Collateral requirements in the MENA ES economies are comparable to peer economies 300 250 200 Percent 150 100 50 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, MENA ES Lower- Upper- Arab Rep. and Gaza Rep. average middle- middle- income income Collateral ratio Collateral incidence Movable collateral incidence Source: Enterprise Surveys. Note: The collateral ratio is the average ratio of the value of the collateral to the value of the loan at disbursement. Collateral incidence is the share of outstanding loans that are collateralized (with any form of underlying asset, namely real estate, land, movable assets, etc.). Finally, movable collateral incidence is the share of collateralized loans where either machinery and equipment or receivables were pledged as collateral. through collateral in cases of default. Collateral also The MENA ES economies have both collateral ratios (the increases the incentives for borrowers to repay given the value of collateral to the value of the loan) and collateral consequences of losing the collateral in case of default. incidence (the share of collateralized loans) above the av- It further mitigates informational asymmetries, as infor- erages for lower-middle-income and upper-middle-income mation on the quality of the collateral can substitute for economies. Higher collateral ratios are often required by borrower information. Consequently, it has been shown banks to compensate for costly and long processes to that loans secured by collateral tend to have much more foreclose collateral, while a high overall collateral incidence favorable terms—higher loan volumes, longer repayment reflects systems based on relatively prudent and conser- periods and lower interest rates—than unsecured loans.23 vative lending practices. Collateral ratios in particular differ widely across economies with the average collateral ratio Collateralized lending also has drawbacks as collateral in Egypt more than twice the level observed in Jordan. requirements can affect the allocation of credit. The avail- ability of assets that can be pledged can become a binding As movable assets represent a substantial share of firms’ constraint on access to credit when loans need to be assets, collateral practices allowing the posting of machin- collateralized. Secured lending also favors investment in ery, equipment, or receivables to secure a loan can be assets that can be pledged as collateral, and thus tilts considered business-friendly. The high regional average is production toward capital-intensive strategies. As the vast driven by West Bank and Gaza where weak land property majority of firms’ assets are movable, a collateral regime rights prevent the use of real estate assets as collateral. that allows for movable assets tends to facilitate financial In fact, a large share of land in the West Bank is simply access.24 Movable assets, such as machinery, equipment, not registered. Without West Bank and Gaza the regional or receivables, account for 78 percent of the capital stock average is much closer to the average for lower-middle- of firms in developing economies.25 But banks have income economies. At 24 percent, Jordan has the second shown reluctance to accept movable assets as collateral highest share of loans secured by movable collateral. In and prefer land or real estate instead. Several metrics of contrast, Lebanon and the Republic of Yemen have only 2 collateral use in the MENA ES economies are presented and respectively 1 percent of loans secured by movable in figure 3.9. collateral. Chapter 3: Access to finance 47 It should be noted, however, that in the MENA region The analysis shows that young and old firms respond dif- movable assets are often used as secondary collateral, in ferently to the collateral standards prevalent in the area addition to real estate.27 Owing to uncertain foreclosure where they operate. Young firms are less likely to discon- outcomes, banks may ask to complement real estate col- nect from the banking system when they are located in an lateral with more liquid assets. In this case, the relevant area where the value of required collateral is low relative measure to assess the tightness of the collateral require- to the volume of the loan.28 This may be a reflection of ment is the overall collateral ratio. the fact that young firms more frequently experience lack of assets to be pledged as collateral as a binding constraint. Older firms, which over time have been able Firms are more likely to disconnect from the banking to accumulate assets, are in a better position to pledge system when faced with stringent collateral practices them as collateral. The collateral regime affects firms’ propensity to discon- nect from the banking system. Table A3.5 presents the Firms located in areas where banks accept movable as- results of an analysis that explains the propensity of a sets as collateral are less likely to disconnect from the firm to disconnect with the prevailing collateral standards banking system. This applies to both young and old firms. required by banks located in the area where the firm Again, this result holds after accounting for other potential operates. The approach addresses potential reverse cau- determinants of being disconnected (table A3.6).29 sality (from poor firm quality to stringent collateral require- ments) by obtaining an estimate of collateral requirements When firms differ in their ability to meet collateral require- cleansed of client firm characteristics. For a more detailed ments such requirements can affect the allocation of description of the methodology, see box 3.2. credit. Box 3.2 goes one step further and links collateral Box 3.2: The case of collateral practices for employment growtha Empirical evidence has highlighted the central role of are certainly demanding, collateral can facilitate lending young firms for job creation.b There is some debate when information asymmetries are salient and therefore about whether employment growth is driven by market banks face high credit risk. But collateralized lending can entry itself or the expansion of existing firms. Earlier also bring about problems if only a small fraction of firms’ work emphasizes the importance of the fast expansion assets can be pledged as collateral. As machinery, equip- of firms in early stages of their life cycle in the United ment, and tradables account for most of firms’ assets, States compared with slow expansion in Mexico and no banking practices allowing movable property as collateral expansion in India.c This suggests that insufficient job might help. But financial institutions may be reluctant to creation could partly be explained by firms’ limited ability accept movable property as collateral if they lack the to expand in early stages of their life cycle. creditor protection that comes with a modern secured transaction regime that encompasses movable property. What can unlock firms’ ability to expand? The MENA ES provides a unique source of information The availability and cost of external finance is one to investigate the extent to which financing constraints of the factors that affect the ability of a business to generated through the collateral channel restricts firms’ expand.d Furthermore firms in different stages of their ability to expand and create new jobs. Simply document- lifecycle face different external financing environments.e ing the association between collateral posted by the Empirical evidence indicates that due to their opacity and firm, access to finance, and employment growth is not the limited availability of assets that can be pledged as enough. The central methodological problem that the collateral, young firms face a larger wedge between the research design needs to address is reverse causality. cost of internal and external finance that makes external Do stringent collateral requirement lead firms to grow finance less attractive.f slower or do banks require more collateral from slow- growing firms? Both channels are plausible and both In the MENA ES economies, banks rely extensively on imply a negative association between collateral require- collateralized lending. About 83 percent of loans require ments, access to finance, and employment growth. some type of collateral with an average value exceed- ing twice the loan amount. While these requirements (continued on next page) 48 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY To address the reverse causality problem, the analysis collateral value to the value of the loan, higher values needs to be based on a measure of collateral require- imply lower collateral ratios. The movable collateral index ments that is not affected by the characteristics of the measures the weighted share of branches of banks will- specific firm. In practice this measure is derived through ing to lend against movable collateral in the area and var- a two-stage procedure. The first stage recovers each ies between zero and one. Thus, if banks that are more bank’s collateral policies. In the second stage, the es- likely to accept movable collateral have a larger share of timated collateral policies are aggregated into collateral branches close to the firm, this will be represented by indices, reflecting market conditions applied by banks in a higher score of the corresponding movable collateral the area where the firm is located. environment index. The MENA ES provides information on the identity of The collateral index is then used to explain firms’ em- the bank that granted the last loan or line of credit to ployment growth. Table A3.7 shows that firms create the firm. This information is used to identify borrower- more jobs when they are young—under 5 years old. The lender linkages. Single banks’ collateral policies are then results also show that these young firms have higher defined as the average conditional collateral requirement employment growth if they are located in areas where for all clients of that specific bank and can be recovered banks with less stringent collateral policies have a stron- through a regression of the collateral requirement on ger presence. Table A3.8 presents results on movable firm characteristics and a bank-specific parameter.g The collateral. The regressions indicate that firms’ ability to bank-specific parameter represents the collateral policy, expand diminishes if they are located in areas with a while the firm-level explanatory variables account for stronger presence of banks less likely to accept movable firm features that may affect the collateral requirement. assets as collateral. This result applies both to young and Data on the location of bank branches and the firms are old firms. The analysis thus provides evidence that collat- then used to obtain a representation of the collateral re- eral practices, by influencing firms’ financial choices and quirements prevalent in the specific market where the options, influence employment creation. firm is located. This idea is implemented by averaging the estimated collateral policies of all banks that have branches in a circle within a radius of 10km centered a Based on Betz and Ravasan (2015). on the sample firm. This index is branch-weighted, thus b http://www.oecd.org/sti/Flyer_DynEmp.pdf, Haltiwanger banks that have more branches in the circle receive and others (2013), Schiffbauer and others (2015), Anyadike- greater weight in the index. Danes and others (2013), Ayyagari and others (2011), Birch In practice two collateral indices are constructed to rep- (1979, 1981, and 1987). resent different aspects of the collateral environment. c Hsieh and Klenow (2012). The first index tracks the ratio of collateral to loan value d See Binks and Ennew (1996a) and Oliveira and Fortunato (the collateral ratio index), whereas the second index (2006) for empirical evidence, and Clementi and measures the share of collateralized loans where either Hopenhayn (2006) for a theoretical exposition. machinery and equipment or receivables were pledged e This literature is known as financial growth cycle paradigm. as collateral (the movable collateral index). The collat- f Schiantarelli (1996), Hubbard (1998). eral ratio index is given by the negative of the average g Technically the bank-specific parameter is a fixed effect. collateral ratio applied by branches of banks located in the area close to the firm. As it is the negative of the requirements to economic performance. It turns out that ability of firms to access credit.30 Two features of the bank- firms located in areas where stringent collateral practices ing sector are explored: the density of bank branches; and are dominant have lower employment growth on average. banks’ net interest margin, considered as a measure of profitability of banks traditional intermediation activities. While the analysis is far from exhaustive, both features Banking sector competition and are relevant, as they help to shed some light on the rela- firm access to credit tionship between banking sector competition and firms’ The section examines the relationship between some access to finance. specific characteristics of the banking system and the Chapter 3: Access to finance 49 A denser network of bank branches is associated with rates and also credit rationing.38 High margins can also be greater access to credit due to high fixed costs as a side-effect of a small financial system. Running a bank involves fixed costs that arise, Branches serve an important role in relationships between for example, from the necessity to develop and sustain borrowers and lenders. These relationships are important a branch network or IT infrastructure. If these fixed costs to facilitate better access to credit. But banks consider are borne by a small number of clients bank lending will several factors when deciding whether to increase the be more expensive. number of branches. At one extreme, they have the option of branchless banking, which in recent years has High interest margins can also be driven by the macroeco- received a lot of attention from both market participants nomic environment; inflation can affect margins if changes and international financial institutions.31 Branchless bank- in monetary policy affect lending and deposit rates at ing is attractive given that branches are expensive and different speeds. In addition, the creditworthiness of bor- require a minimum level of economic activity close to the rowers varies over the business cycle and can likewise location to be viable. affect lending rates. Finally, monopoly rents can lead to high interest margins in the absence of competitive forces In the MENA ES economies,32 however, a denser network to drive down the margins. of bank branches is associated with greater access to credit by firms. Firms are more likely to have a loan or In the present context, it seems likely that elevated inter- line of credit outstanding if they are located in areas with est margins result from lack of competition among banks higher branch density (table A3.9).33 A concern with this in the MENA ES region. The institutional and macroeco- finding is that it may be that branches choose to locate nomic factors do not vary within economies and therefore in areas of high population density—and therefore high they cannot explain the observed variation of interest economic activity—where firms are more likely to de- margins within an economy. Most banks operate in one mand credit. But the positive association between branch economy and thus rely on the local market to cover their density and access to credit holds after accounting for the fixed costs. Furthermore, monetary policy is set at the effects of population density. national level. High bank profit margins may deter access to credit Previous studies indicate that banking markets in the MENA region are less competitive than in other regions The MENA ES data show that profit margins may be of the world.39 Lack of competition in the banking sector negatively associated with access to credit. Firms located is attributed to a poor credit information environment and in regions where banks earn higher net interest margins lack of market contestability. Additional findings from the are less likely to have a bank loan than firms in regions MENA ES support this explanation: using the return on av- where banks earn lower margins.34 This finding holds after erage assets as an alternative measure of profit margins accounting for several other factors that could also explain provides consistent results (table A3.9). Firms located in the result, such as firm size, age, sector of activity, owner areas where banks with high returns on assets have a and manager characteristics, and level of engagement of strong presence are less likely to have a bank loan or line the firm in trade and with the real economy (table A3.9). of credit. This result is consistent with the literature that finds high interest margins to be impediments to financial access.35 Policy conclusions The literature provides several potential explanations for high interest margins: information asymmetries between This chapter highlights that in most MENA ES economies, lenders and borrowers, high fixed costs for banks, a substantial share of the private sector does not use banks macroeconomic factors,36 and monopoly rents from lack but chooses to remain disconnected from the financial of competition in the banking sector.37 Information asym- sector. This may be seriously undermining the potential metries make it difficult for a bank to assess borrowers’ for growth of the private sector. The chapter also provides creditworthiness effectively, leading to higher lending evidence that financial exclusion carries costs in terms of 50 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY forgone employment growth. Such costs are particularly Credit guarantee schemes can be an alternative mecha- high in societies plagued by persistent underemployment. nism to alleviate collateral constraints.42 But the ability of While such financial exclusion may be caused by both guarantee schemes to foster financial inclusion hinges demand and supply factors, it clearly represents a sub- critically on operational design. In particular, incentives optimal outcome. The chapter also highlights potential between lender, borrower, and guarantor need to be pathways to re-connect firms with the financial system. aligned.43 In principle, collateral and guarantees can be used on the same loan. Putting up collateral reduces the More bank competition and lower government funding borrower’s incentives to default. If, however, guarantees needs are likely to have a positive effect on access to simply provide back-up protection for collateralized loans, finance. The first section shows that the MENA ES region they no longer contribute to financial inclusion. It is there- stands out for the high level of credit to governments and fore crucial that contractual mechanisms governing the state-owned enterprises. Governments can offer more level of collateralization prevent this scenario. A modern attractive risk-adjusted returns than private sector borrow- secured transactions framework is likely to increase the ers, crowding out the marginal private sector borrowers. appeal of bank finance. The second section shows how Following the popular protests of 2011, governments have a rigid collateral regime can induce firms to disconnect increased spending to maintain economic activity as well from the banking system. The MENA ES economies have as social cohesion. In Egypt, for example, claims on the for many years scored poorly on the legal rights index public sector increased from 27 percent of GDP in 2010 of Doing Business, and earlier work by the World Bank44 to above 50 percent in 2015. The expansionary policies highlights the benefits of a modern secured transactions have strained fiscal buffers, leaving few alternatives to law and an efficient collateral registry. While it is under- fiscal consolidation, which is likely to undo some of the standable that policy makers have prioritized other issues, crowding-out observed in recent years. there should now be scope to tackle secured transaction reform, at least in those jurisdictions that experience a Programs aimed at strengthening banks’ capacity to as- return to political stability. sess credit risk should accompany a shift in the regulatory stance toward increased competition. Improvements in The chapter also shows that access to finance suffers in financial access should not come at the expense of finan- regions where banks with high interest margins have a cial stability. The institutional framework therefore needs stronger presence. This is consistent with earlier work to be adapted so that competition does not lead to irre- finding that competition between banks is weaker than in sponsible lending practices.40 Capacity-building measures other regions.45 Banks’ market power has been attributed could help banks interested in entering the SME segment to a lack of market contestability: indeed, the region has to avoid pitfalls. Such programs may also lower potential the highest share of rejected applications for bank licenses resistance to reform from incumbents, as they will be in a among emerging economies. better position to cope with the challenges that increased competition entails. Increased competition could provide incentives for banks to seek out new market segments such as SME lending.46 Governments and donors can support capacity-building SME lending may not be attractive for banks focusing measures that increase banks’ screening capacity and the on corporates as long as the appropriate organizational supply of bankable firms. Such measures should aim to structure is not in place. Competition, however, could pro- make SMEs less opaque and thus reduce the information mote organizational and procedural change, and thereby asymmetries that plague lending to them. In practice, this facilitate access to finance. Thus, bank regulators may may involve helping entrepreneurs develop a business want to take account of the competitive landscape when plan or define an organizational structure.41 A limitation of evaluating applications for a banking license. such programs is that they are typically bound to be small relative to the size of the economy. The association between the share of firms with a check- ing or savings account and the share of firms that were registered when they started their operations suggests Chapter 3: Access to finance 51 that the banking sector disconnect is also associated with 21 These findings hold after accounting for various firm the perceived costs and benefits of formalization. Informal characteristics such as firm size, age, sector of activity, exporting status, as well as owner and manager firms may economize on taxes, but informality also implies characteristics. opportunity costs in terms of forgone growth. Addressing 22 This pattern holds after accounting for other factors informality, however, is beyond the scope of this report. such as firm sector of activity, firm age, and attributes of ownership. 23 World Bank (2006). Endnotes 24 Alvarez de la Campa (2011), Love and others (2015). 25 World Bank (2006). 1 Creane and others (2004). 26 Collateral can be enforced in zone A of the West Bank 2 An extensive empirical literature provides country-level, only, which substantially constrains lending. industry-level and firm-level evidence to show that 27 Alvarez de la Campa (2011). financial development is conducive to economic growth, see Levine (1997), Levine (1998) and Levine and others 28 These results hold after accounting for other potential (2000). determinants of firms being disconnected. 3 World Bank (2011). 29 This finding is likely to understate the full benefits of adopting a modern secured transaction framework as they 4 IMF (2014, 2015a), http://blog.blominvestbank.com/wp- are based on the observed behavioral variation under the content/uploads/2014/10/The-banking-sector-in-Lebanon. existing regimes. It is likely that structural changes to a pdf. secured transaction regime would bring additional benefits 5 ESCWA (2014). not accounted for in this analysis. 6 Yu and others (2014). 30 Similar to the approach outlined in box 2.2 the 7 World Bank (2011). methodology uses data on the location of firms and 8 World Bank (2011). bank branches to construct branch weighted measures of banking system properties at the subnational level 9 Gray and others (2014). and relate them to credit outcomes at the level of the 10 Herrera and others (2013), Gray and others (2014). firm. This within-economy design has the advantage of 11 IMF (2015b). removing potential confounding factors at the economy 12 Owing to lack of data, the average does not take into level. Confounders are variables that are correlated with account Lebanon and the West Bank and Gaza. the explanatory variable of interest and the outcome. Economies differ in so many dimensions that it is 13 Gray and others (2014). impossible to statistically remove all possible confounding 14 AMF and EBRD (2015), Gray and others (2014), IMF (2013). factors. The within-economy approach circumvents this 15 World Bank (2011). problem by effectively comparing only entities within the 16 World Bank (2011). same economy. 17 With the exception of the Republic of Yemen, Islamic 31 EIB (2014a). banks have only a small presence in the MENA ES 32 Owing to lack of data on the location of bank branches, economies. Therefore the survey makes no effort to the analysis does not take into account Djibouti and the distinguish Islamic banks from banks and non-bank Republic of Yemen. financial institutions. 33 This finding stands after accounting for several factors 18 The results closely resemble those that would be obtained such as firm size, age, sector, and owner characteristics. from a simpler measure that classifies as credit- 34 Bank profit margins are proxied by the net interest margin constrained those firms that had their loan application defined as interest income minus interest expense rejected or were discouraged from applying in the first expressed as a percentage of interest earning assets. place; see, for example, Popov and Udell (2012). 35 Brock and Rojas-Suarez (2000), Beck and Hesse (2009). 19 While the MENA ES allows for the measurement of the 36 Beck and Hesse (2009). degree of credit constraint by loan outcomes, for older surveys this information is not available. Thus the outcome 37 Anzoategui and others (2010). of the loan is approximated by whether or not the firm has 38 Stiglitz and Weiss (1981). a loan outstanding. 39 Anzoategui and others (2010). 20 These findings are significant after accounting for firm 40 Economic theory makes conflicting predictions on the size, age and sector of activity. relationship between bank competition and financial 52 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY stability. According to the “competition-fragility” view, 43 EIB (2014a). increased competition erodes the charter value of banks 44 World Bank (2011). (Keeley, 1990). Intense completion between banks 45 Anzoategui and others (2010). leads to excessive risk taking. The quality of the loan portfolio deteriorates with the quality of the marginal 46 According to standard economic theory, market power borrower, increasing financial stability. According to the leads to reduced supply at higher cost. In the presence “competition-stability” view, market power leads to of asymmetric information, on the other hand, banks high interest rates, which triggers excessive risk taking with market power have greater incentives to establish on the side of borrowers (Boyd and de Nicoló, 2005). relationships with young or distressed firms by shifting The evidence in Beck and others (2013) suggests that interest payments into the future (Petersen and Rajan, the trade-off between competition and financial stability 1995). In a cross country setting, Beck and others (2004) depends on country-specific characteristics. find that firms in countries with low levels of economic and institutional development perceive access to 41 McKenzie (2015). finance as a greater obstacle when banking markets are 42 See, for example, EIB (2014b). concentrated. Chapter 3: Access to finance 53 Appendix A3 Table A3.1: Credit constraints and firm performance Table A3.2: Characteristics of disconnected firms Dependent variable: Probit (marginal effects) Probit (marginal effects) credit constrained (1) (2) (3) (FCC, PCC - Y/N) (1) (2) (3) Investment— Plans to Access to Annual employment -0.46*** purchased increase size finance: major growth (%) (0.104) fixed assets of establish- or severe Dependent variable (Y/N) ment (Y/N) obstacle (Y/N) Capacity utilization (%) -0.21** Disconnected -0.16*** -0.10** -0.12*** (0.091) (no need for a loan due to Log of sales per worker -0.03*** sufficient funds - Y/N) (0.041) (0.042) (0.044) (USD) (0.012) Credit constrained (FCC, -0.19*** -0.17*** 0.15*** PCC) (Y/N) Log of size, 2010 -0.09*** (0.041) (0.063) (0.054) (0.016) Wald test: disconnected = credit constrained 1.17 2.63 43.25*** Log of size -0.09*** -0.08*** P-value 0.280 0.105 0.000 (0.020) (0.014) Number of observations 5,403 5,316 5,394 Young firms: 0-5 years 0.05 0.07 0.05 (Y/N) Source: Enterprise Surveys. (0.050) (0.065) (0.038) Note: Marginal effects from probit regression using survey-weighted observations Firm is part of a larger 0.12*** -0.02 0.13*** (Stata’s svy prefix). Other control variables included but not reported include size, age, firm (Y/N) manager education, manager experience in the sector, exporting status, gender of the (0.039) (0.074) (0.038) owner, foreign ownership, multi-establishment firm and legal status. ***, ** and * Manager has university -0.04 -0.04 -0.02 denote statistical significance at the 1, 5 and 10 percent levels respectively. education (Y/N) (0.036) (0.057) (0.035) Manager experience in -0.00 0.00 -0.00 sector (years) Table A3.3: Investment and capacity utilization (0.002) (0.002) (0.002) Exports 10% or more of -0.01 0.06 -0.03 (1) sales (Y/N) Investment—purchased fixed (0.044) (0.063) (0.042) Dependent variable assets (Y/N) Foreign ownership (Y/N) 0.01 -0.06 0.02 Disconnected (no need for a loan due to -0.02 (0.061) (0.069) (0.058) sufficient funds—Y/N) (0.958) Number of observations 4,715 2,760 4,772 Above median capacity utilization (Y/N) 0.43 Source: Enterprise Surveys. Note: Marginal effects from probit regression using survey-weighted observations (0.180) (Stata’s svy prefix). Standard errors are reported in parentheses below the Disconnected * above median capacity -0.64* coefficient. The dependent variable is the credit-constraint indicator described in box utilization 3.1. All specifications consider a firms as credit constrained if it is either partially or (0.080) fully credit constrained and include both economy and sector fixed effects. Capacity Marginal effects of interaction utilization is defined only for manufacturing firms. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Disconnected | above median capacity 0.00 utilization = 0 (0.095) Disconnected | above median capacity -0.22 utilization = 1 (0.094) P-value of the difference 0.087* Number of observations 2,202 Source: Enterprise Surveys. Note: Coefficient estimates and marginal effects from Probit regression using survey- weighted observations (Stata’s svy prefix). The marginal effects show the difference in the probability to invest relative to firms that obtained a loan condition on the state of capacity utilization. Capacity utilization is defined only for manufacturing firms. Control variables included but not reported include size, age, manager education, manager experience in the sector, exporting status, gender of the owner, foreign ownership, multi-establishment firm and legal status. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. 54 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A3.4: Probability of having a loan or line of credit Table A3.5: Collateralized lending and the banking system Probit (marginal effects) disconnect (1) (2) (3) Probit (marginal effects) Jordan, (1) (2) (3) Dependent variable: Firm Djibouti, Egypt, Lebanon, Dependent variable: Collateral Single firms has a loan or line of credit West Bank and Morocco, Disconnect (no need for environment or HQ of multi- from a bank (Y/N) All MENA ES Gaza, Yemen Tunisia a loan due to sufficient based on loans establishment Young firms: 0-5 years -0.08*** -0.06** -0.09* funds—Y/N) Full sample only after 2005 firms (Y/N) Collateral Environment 0.00 (0.031) (0.026) (0.053) Index (higher Small and medium firms -0.10*** -0.06 -0.16*** (0.003) values means less (less than 100 full time collateralization of loans) employees) (Y/N) (0.037) (0.050) (0.047) Collateral Environment -0.01** Female principal owner 0.04 -0.02 0.09** Index * young firms (Y/N) (0.005) (0.029) (0.031) (0.043) (younger than five) Foreign ownership (Y/N) -0.05 -0.00 -0.10* Collateral Environment 0.00 0.00 Index 2005 (based only on (0.036) (0.043) (0.055) (0.003) (0.003) loans after 2005) External auditor reviewed 0.07*** -0.00 0.16*** Collateral Environment -0.01* -0.01* financial statements (Y/N) Index 2005 * young firms (0.025) (0.024) (0.042) (0.005) (0.005) (younger than five) Shareholding company 0.07** 0.06 0.07 (Y/N) Young firms (younger than -0.05 -0.04 -0.05 (0.032) (0.046) (0.047) five) (Y/N) (0.046) (0.046) (0.046) Manager has university 0.05** 0.05* 0.04 education (Y/N) Number of observations 4,855 4,855 4,054 (0.025) (0.028) (0.039) Source: Enterprise Surveys. Manager experience in 0.00 0.00 0.00 Note: Marginal effects from probit regression using survey-weighted observations sector (years) (Stata’s svy prefix). Standard errors are reported in parentheses below the (0.001) (0.001) (0.002) coefficient. The collateral ratio index is a branch-weighted average of the collateral Exports 10% or more of -0.00 -0.02 0.01 policies of banks that have branches in a circle with radius 10km centered on the sales (Y/N) sample firm. The MENA ES has information on the identity of the bank that granted (0.027) (0.029) (0.039) the last loan or line of credit. It is therefore possible to estimate banks’ collateral Firm is part of a larger 0.07** 0.10** 0.03 policies as bank-specific effects in a fixed effect regression of the collateral ratio on firm (Y:1 N:0) firm characteristics (not shown). Other control variables included but not reported (0.034) (0.042) (0.048) include size, manager education, exporting status, gender of the manager, foreign Number of observations 5,486 3,597 1,889 ownership, multi-establishment firms, having a website, having audited financial reports. Firms and banks from Djibouti and the Republic of Yemen are not part of Source: Enterprise Surveys. the sample. For more details on the methodology see box 3.2. ***, ** and * denote Note: Marginal effects from probit regression using survey-weighted observations statistical significance at the 1, 5 and 10 percent levels respectively. (Stata’s svy prefix). Standard errors are reported in parentheses below the coefficient. All regressions include economy and sector fixed effects. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 3: Access to finance 55 Table A3.6: Movable collateral and the banking system Table A3.7: Collateralized lending and employment growth disconnect (1) (2) (3) Probit (marginal effects) Collateral Single firms (1) (2) (3) environment or HQ of multi- Dependent variable: based on loans establishment Dependent variable: Collateral Single firms employment growth Full sample only after 2005 firms disconnect (no need for environment or HQ of multi- a loan due to sufficient based on loans establishment Collateral Environment 0.00 funds—Y/N) Full sample only after 2005 firms Index (higher (0.002) values mean less Movable Collateral -0.96** collateralization of loans) Environment Index (0.455) Collateral Environment 0.01** (higher values means greater acceptance of Index * young firms (0.005) movable collateral for (younger than five) loans) Collateral Environment 0.00 0.00 Movable Collateral -1.02* -1.11** Index 2005 (based only (0.002) (0.002) Environment Index 2005 on loans after 2005) (0.525) (0.528) (based on loans after Collateral Environment 0.01** 0.01** 2005) Index 2005 * young firms (0.005) (0.005) Young firms (younger -0.04 -0.04 -0.04 (younger than five) than five) (Y/N) Young firms (younger 0.13** 0.13** 0.13** (0.045) (0.045) (0.045) than five) (Y/N) Number of observations 4,855 4,855 4,625 (0.053) (0.053) (0.053) Source: Enterprise Surveys. Number of observations 4,256 4,256 4,054 Note: Marginal effects from probit regression using survey-weighted observations Source: Enterprise Surveys. (Stata’s svy prefix). Standard errors are reported in parentheses below the coefficient. Note: OLS using survey-weighted observations (Stata’s svy prefix). Standard errors The movable collateral index is a branch-weighted average of the collateral policies are reported in parentheses below the coefficient. The collateral ratio index is a of banks that have branches in a circle with radius 10km centered on the sample branch-weighted average of the collateral policies of banks that have branches in a firm. The MENA ES has information on the identity of the bank that granted the last circle with radius 10km centered on the sample firm. The MENA ES has information loan or line of credit. It is therefore possible to estimate banks’ collateral policies as on the identity of the bank that granted the last loan or line of credit. It is therefore bank-specific effects in a fixed effect regression of an indicator for movable collateral possible to estimate banks’ collateral policies as bank-specific fixed effects in a on firm characteristics (not shown). Other control variables included but not reported regression of collateral ratio on firm characteristics (not shown). Other control include size, manager education, exporting status, gender of the manager, foreign variables included but not reported include initial size (log), manager education, ownership, multi-establishment firms, having a website, having audited financial exporting status, gender of the manager, foreign ownership, multi-establishment reports as well as economy and sector fixed effects. Firms and banks from Djibouti firms, having a website, having audited financial reports and economy and sector and the Republic of Yemen are not part of the sample. For more details on the fixed effects. Firms and banks from Djibouti and the Republic of Yemen are not part of methodology see box 3.2. ***, ** and * denote statistical significance at the 1, 5 and the sample. For more details on the methodology see box 3.2. ***, ** and * denote 10 percent levels respectively. statistical significance at the 1, 5 and 10 percent levels respectively. 56 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A3.8: Movable collateral and employment growth Table A3.9: Probability of firms having a loan and (1) (2) (3) characteristics of the banking system Collateral Single firms Dependent variable: firm has Probit (marginal effects) environment or HQ of multi- a loan or line of credit from a bank (Y/N) (1) (2) (3) Dependent variable: based on loans establishment employment growth Full sample after 2005 firms Log of bank branches per firm 0.09*** Movable Collateral 0.66** (0.003) Environment Index (higher (0.312) Net interest margin, 2nd tercile -0.03 values mean greater acceptance of movable (0.378) collateral for loans) Net interest margin, 3rd tercile -0.09*** Movable Collateral 0.77** 0.83** Environment Index 2005 (0.002) (based on loans after 2005) (0.362) (0.362) Return on assets, 2nd tercile -0.03 Young firms (younger than 0.13** 0.13** 0.14** (0.274) five) (Y/N) (0.054) (0.054) (0.054) Return on assets, 3rd tercile -0.06* Number of observations 4,855 4,855 4,625 (0.084) Source: Enterprise Surveys. Young firms: 0-5 years (Y/N) -0.08** -0.08** -0.07* Note: OLS using survey-weighted observations (Stata’s svy prefix). Standard errors are reported in parentheses below the coefficient. The movable collateral index is a (0.030) (0.042) (0.058) branch-weighted average of the collateral policies of banks that have branches in a Small and medium firms (less circle with radius 10km centered on the sample firm. The MENA ES has information -0.13*** -0.13*** -0.12*** than 100 full time employees) on the identity of the bank that granted the last loan or line of credit. It is therefore (Y/N) (0.001) (0.000) (0.001) possible to estimate banks’ collateral policies as bank-specific effects in a fixed effect regression of an indicator for movable collateral on firm characteristics (not Female principal owner (Y/N) 0.04 0.05 0.05 shown). Other control variables included but not reported include initial size (log), (0.178) (0.164) (0.127) manager education, exporting status, gender of the manager, foreign ownership, multi-establishment firms, having a website, having audited financial reports. Firms Foreign ownership (Y/N) -0.07* -0.06 -0.07 and banks from Djibouti and the Republic of Yemen are not part of the sample. (0.090) (0.119) (0.107) For more details on the methodology see box 3.2. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Financial statement reviewed 0.10*** 0.11*** 0.11*** by external auditor (Y/N) (0.000) (0.000) (0.000) Shareholding firm (Y/N) 0.06* 0.06* 0.06* (0.095) (0.088) (0.087) Manager education: university 0.03 0.03 0.04 (Y/N) (0.310) (0.307) (0.164) Years of experience of the top 0.00 0.00 0.00 manager working in the firm’s sector (0.228) (0.296) (0.275) Exporter (Y/N) -0.01 -0.01 0.00 (0.706) (0.765) (0.928) Firm is part of a larger firm (Y/N) 0.04 0.05 0.05 (0.218) (0.201) (0.132) Log of population density -0.01 0.00 0.00 (0.228) (0.894) (0.800) Number of observations 5,155 5,155 5,155 Source: Enterprise Surveys, Bankscope. Note: Marginal effects from Probit regression using survey-weighted observations (Stata’s svy prefix). Branch density is given by the log of bank branches at the locality level divided by the number of sample firms in that locality. Net interest margin and return on assets are branch weighted averages at the locality level. The resulting distributions exhibit bunching at the country level. To generate sufficient within- country variation they are then split into terciles. Firms and banks from Djibouti and the Republic of Yemen are not part of the sample. Bank balance sheet data comes from Bureau van Dijk’s Bankscope. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 3: Access to finance 57 References Boyd, John, and Gianni de Nicoló. 2005. “The Theory of ” Journal of Finance 60 (3): Bank Risk Taking Revisited. Alvarez de la Campa, Alejandro. 2011. “Increasing Access to 1329-1343. Credit through Reforming Secured Transactions in the MENA Region. ” Policy Research Working Paper 5313, Brock, Philip, and Liliana Rojas Suarez. 2000. “Understanding World Bank, Washington, DC. the behavior of bank spreads in Latin America. ” Journal of Development Economics 63: 113-34. Anyadike-Danes, Michael, Mark Hart, and Jun Du. 2013. “Firm dynamics and job creation in the UK.” ERC White Paper Clementi, Gian Luca, and Hugo A. Hopenhayn. 2006. “A 6. ” The theory of financing constraints and firm dynamics. Quarterly Journal of Economics 121(1): 229-65. Anzoategui, Diego, Maria Soledad Martínez Peria, and Roberto Rocha. 2011. “Bank Competition in the Middle East and Creane, Susan, Rishi Goyal, Ahmed Mushfiq Mobarak, North Africa. ” Policy Research Working Paper 5363, and Randa Sab. 2004. “Evaluating Financial Sector World Bank, Washington, DC. Development in the Middle East and North Africa: New Methodology and Some New Results. ” Topics in Middle Arab Monetary Fund and European Bank for Reconstruction Eastern and North African Economies, Proceedings of and Development. 2015. Joint IFI Needs Assessment the Middle East Economic Association, Vol. 6. on Local Capital Market Development: Egypt—Jordan – Morocco—Tunisia. Abu Dhabi: Arab Monetary Fund Economic and Social Commission for Western Asia. 2014. and London: European Bank for Reconstruction and “Remittances to Lebanon: Economic Impact and Development. ” Beirut: Economic and Social the Role of Banks. Commission for Western Asia. Ayyagari, Meghana, Asli Demirgüç-Kunt, and Vojislav Maksimovic. 2011. “Small vs. young firms across the European Investment Bank. 2014a. Digital Financial Services in world: contribution to employment, job creation, and Africa: Beyond the Kenyan Success Story. Luxembourg: growth. ” World Bank Policy Research Working Paper European Investment Bank and UN Capital 5631. Development Fund. Beck, Thorsten, Olivier De Jonghe, and Glenn Schepens. European Investment Bank. 2014b. Credit Guarantee Schemes 2013. “Bank Competition and Stability: Cross-country for SME lending in Central, Eastern and South-Eastern heterogeneity. ” Journal of Financial Intermediation 22(2): Europe. Luxembourg: European Investment Bank. 2018-244. Gray, Simon, Philippe Karam, and Rima Turk. 2014. “Are banks Beck, Thorsten, Asli Demirgüç-Kunt, and Vojislav Maksimovic. really lazy? Evidence from Middle East and North 2004. “Bank Competition and Access to Finance: Africa.” IMF Working Paper WP/14/86. International Evidence. ” Journal of Money, Credit, and Haltiwanger, John, Ron S. Jarmin, and Javier Miranda. 2013. Banking 36(3): 627-648. “Who creates jobs? Small versus large versus young. ” Beck, Thorsten, Ross Levine, and Norman Loayza. 2000. Review of Economics and Statistics 95(2): 347-61. “Finance and the sources of growth. ” Journal of Hsieh, Chang-Tai, and Peter J. Klenow. 2014. “The Life Cycle Financial Economics 58: 261-300. ” The Quarterly Journal of of Plants in India and Mexico. Beck, Thorsten, and Heiko Hesse. 2009. “Why are interest Economics 129(3): 1035-84. spreads so high in Uganda. ” Journal of Development Herrera, Santiago, Christophe Hurlin, and Chahir Zaki. 2013. Economics 88(2): 192-204. “Why don’t banks lend to Egypt’s private sector?” Betz, Frank, and Farshad Ravasan. 2015. “Collateral regimes Economic Modelling 33 (1): 347-56. and missing job creation in the MENA region.” Hubbard, Glenn R. 1998. “Capital-market imperfections and Unpublished manuscript. investment.” Journal of Economic Literature 36 (1): Birch, David GW. 1979. “The job generation process. ” MIT 193-225. Program on Neighborhood and Regional Change. International Monetary Fund. 2013. Regional Economic ———.1981. ‘‘Who Creates Jobs?’’ Public Interest 65(1), 3–14. Outlook Update, Middle East and Central Asia, World Economic and Financial Surveys, Washington D.C. ———. 1987 . “Job creation in America: How our smallest companies put the most people to work. ” University International Monetary Fund. 2014. Lebanon 2014 Article of Illinois at Urbana-Champaign’s Academy for IV Consultation—Staff Report. Washington, DC: Entrepreneurial Leadership Historical Research International Monetary Fund. Reference in Entrepreneurship. International Monetary Fund. 2015a. Lebanon 2015 Article Binks, Martin R., and Christine T. Ennew. 1996. “Growing firms IV Consultation—Staff Report. Washington, DC: and the credit constraint.” Small Business Economics International Monetary Fund. 8(1): 17-25. 58 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY International Monetary Fund. 2015b. Arab Republic of Egypt Perez-Quiros, Gabriel, and Allan Timmermann. 2000. “Firm 2014 Article IV Consultation—Staff Report. Washington, ” The Journal size and cyclical variations in stock returns. DC: International Monetary Fund. of Finance 55(3): 1229-62. Keeley, Michael. 1990. “Deposit insurance, risk, and market Petersen, Mitchell and Raghuram Rajan. 1995. “The Effect of power in banking. ” American Economic Review 80: Credit Market Competition on Lending Relationships.” 1183-1200. Quarterly Journal of Economics 110(2): 407-443. Kuntchev, Veselin, Rita Ramalho, Jorge Rodríguez-Meza, and Popov, Alexander, and George Udell. 2012. “Cross-border Judy S. Yang. 2014. “What Have We Learned from the ” Journal banking, credit access, and the financial crisis. Enterprise Survey Regarding Access to Finance by of International Economics 87(1): 147-161 SMEs?” Policy Research Working Paper 6670, updated Schiantarelli, Fabio. 1996. “Financial constraints and May 2014, World Bank, Washington, DC. investment: methodological issues and international Levine, Ross. 1997. “Financial development and economic evidence. ” Oxford Review of Economic Policy 12(2): growth: views and agenda. ” Journal of Economic 70-89. Literature 35(2): 688–726. Schiffbauer, Marc, Abdoulaye Sy, Sahar Hussain, Hania Levine, Ross. 1998. “The legal environment, banks and Sahnoun, and Philip Keefer. 2014. Jobs or Privileges: long-run economic growth. ” Journal of Money Credit Unleashing the employment potential of the Middle and Banking 30 (3): 596-613. East and North Africa. World Bank Publications. Levine, Ross, Norman Loayza, and Thorsten Beck. 2000. Stiglitz, Joseph, and Andrew Weiss. 1981. “Credit Rationing “Financial intermediation and growth: causality and ” American in Markets with Imperfect Information. causes.” Journal of Monetary Economics 46(1): 31–77 . Economic Review 71(3): 393-410. Love, Inessa, Maria Soledad Martinez Peria, and Sandeep World Bank. 2006. Reforming Collateral Laws to Expand Singh. 2015. “Collateral Registries for Movable Assets: Access to Finance. Washington, DC: World Bank. Does Their Introduction Spur Firms’ Access to Bank World Bank. 2011. Financial Access and Stability—A Financing?” Journal of Financial Services Research, Roadmap for the Middle East and North Africa. MENA forthcoming DOI 10.1007/s10693-015-0213-2. Development Report. Washington, DC: World Bank. Oliveira, Blandina, and Adelino Fortunato. 2006. “Firm growth World Bank. 2016. Doing Business 2016: Measuring ” Small and liquidity constraints: A dynamic analysis. Regulatory Quality and Efficiency. Washington, DC: Business Economics 27(2): 139-56. World Bank. McKenzie, David J. 2015. Identifying and spurring high-growth Yu, Jung-Suk, M. Kabir Hassan, Abudllah Mamun, and Abul entrepreneurship: experimental evidence from a Hassan. 2014. “Financial Sectors Reform and Economic business plan competition. Policy Research working Growth in Morocco: An Empirical Analysis.” Journal of paper; no. WPS 7391; Impact Evaluation series. Emerging Market Finance 13(1): 69–102. Washington, D.C.: World Bank Group. 59 4. Jobs and skills in the formal private sector Introduction Recent developments, punctuated by downturns in growth following the Arab Uprisings, have made the The importance of jobs in the MENA region can situation more tenuous. Several governments in the hardly be exaggerated. The region has been suffer- region responded to this uncertainty by ramping up ing from structural unemployment for years, with public expenditure, particularly on food and energy an unemployment rate averaging over 12 percent subsidies, resulting in government fiscal deficits of in the 1990s and 2000s, substantially higher than approximately 10 percent of GDP in Egypt, Jordan, elsewhere in the world.1 While the economic per- Lebanon, Tunisia, and the Republic of Yemen.4 Given formance of the region over the last two decades the tight fiscal and budgetary situation, it is highly has been reasonably good, it has failed to keep unlikely that the public sector—long a desired source pace with large increases in population and demand of employment—alone will be able to create enough for jobs. A World Bank study from the early 2000s jobs in the coming years. The only solution to high estimated that close to 6 million new jobs each unemployment rates lies with the development of a year would be required to absorb new labor market dynamic and competitive formal private sector. entrants.2 But the MENA region was able to add only 3.2 million jobs per year during the 2000s.3 Aside from overall job creation, employment op- portunities for young people and women in the MENA region are important, not only for economic 60 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY reasons, but also for social and political ones. Women’s Despite its importance, there is little systematic research participation in the labor market in the region is one of on the role of the formal private sector in providing em- the lowest in the world; youth unemployment is one of ployment in the MENA region. Lack of data is one reason. the highest.5 The youth unemployment rate neared 30 This chapter uses the MENA ES data to shed light on percent in the region in 2013, more than twice as high key issues, such as the share of jobs provided by differ- as the global average.6 Failure to provide jobs for millions ent types of firms, women and youth employment, firm of people can lead to social unrest and political turmoil, dynamism, and the relationship between employment, as was evident during the Arab Uprisings. Along with the skills, and wages. demand for more political inclusion, young people in par- ticular took to the streets out of frustration with the lack of opportunities to put their skills and talents to productive Employment in the formal private use.7 Creating these jobs remains a key challenge. sector The formal private sector constitutes only a fraction of Larger firms provide the majority of formal private total employment in the MENA ES region, which is known sector jobs in the MENA ES region for a high level of public sector employment and a large The MENA ES data provide a unique source of information informal sector.8 Precise estimates of the importance of on employment provided by different types of firms, com- the formal private sector for employment are difficult to bined with evidence on firm productivity. This will help obtain, but labor force and household surveys suggest policy makers to identify appropriate policies and actions that the share of private formal employment ranges from for fostering job growth. A pattern that has been widely around 10 percent in Morocco and Egypt to 15 percent observed—particularly in developing economies—is that in Tunisia and 25 percent in Jordan. Public sector em- private sector jobs tend to be clustered in either a vast ployment also ranges from just under 10 percent in the abundance of smaller firms or a handful of substantially Republic of Yemen to more than 30 percent in Jordan.9 larger ones.11 In the MENA region, previous analysis has At the same time, informality accounts for roughly 50 found that most jobs are in large firms.12 In all MENA ES percent of non-agricultural sector employment. Given the economies except the West Bank and Gaza, the largest limits of public sector employment creation and the typi- share of private sector jobs is indeed in large firms (fig- cally low productivity and wages of the informal sector,10 ure 4.1).13 greater attention must be devoted to the role of the formal private sector in creating productive employment. Figure 4.1: The proportion of employment by firm size 100 80 60 Percent 40 20 0 West Bank Yemen, Djibouti Egypt, Morocco Lower- Lebanon Jordan Tunisia Upper- and Gaza Rep. Arab Rep. middle- middle- income income Small (5-19 employees) Medium (20-99 employees) Large (100+ employees) Source: Enterprise Surveys. Chapter 4: Jobs and skills in the formal private sector 61 The relatively small share of employment in small and of employment growth; but if firms do not grow in size medium-sized enterprises (SMEs) is notable, particularly over time, the large presence of small firms may be given a strong policy focus on those firms as sources of indicative of market distortions that hamper competition gainful employment in the private sector. This share is and obstruct the incentives or opportunities for firms to not explained by a relative lack of SMEs: in the MENA grow.14 It is often argued that younger firms tend to be ES economies, 96 percent of establishments have more dynamic, learn faster from mistakes, provide better fewer than 100 employees. Rather, firms in the MENA quality jobs, and generate higher employment growth ES economies tend to be smaller (with a handful of large than their older counterparts.15 Conversely, older firms firms being exceptionally large). Figure 4.2 shows the may tend to have better political connections and enjoy proportion of firms in ES economies below the median protection from competitive forces, undermining eco- size of the same income group. In all economies except nomic dynamism.16 One study concludes that the latter Morocco and Tunisia, the majority of MENA ES firms are forces are more predominant in the MENA region.17 smaller than the comparative median size (15 employees in lower-middle-income economies and 13 employees in The dominance of older firms is borne out by the distribu- upper-middle-income economies). tion of jobs between young and old firms in the MENA ES economies. About three quarters of jobs are provided by Put simply, firms in the MENA ES economies are smaller firms that are more than 10 years old. The contribution of on average. Morocco and Tunisia have the smallest young firms to private sector employment stands out as employment shares in SMEs; they are also the only particularly high in Djibouti, Egypt, and the West Bank and two MENA ES economies with average firm size higher Gaza.18 In contrast, in Lebanon, Tunisia, and the Republic than in their peer economies. This distribution may have of Yemen, older firms are the source of more jobs. implications for overall productivity: although larger firms in the MENA ES economies tend to be more productive Firms in the MENA ES also tend to be older on average (as explained in chapter 2), they are rare. Less productive (table 4.1), which may be indicative of high barriers to SMEs are abundant. entry for new firms.19 A continual and efficient entry of new firms would necessarily lower the average age of firms, but this seems not to be the case in the MENA ES Firms older than 10 years provide three quarters of jobs economies. If small firms are in the early stages of their lifecycle, they may represent dynamic sectors and new sources Figure 4.2: The proportion of firms below income-group median size 100 80 Percent of firms 60 40 20 0 West Bank Yemen, Rep. Egypt, Arab Djibouti Morocco Jordan Lebanon Tunisia and Gaza Rep. Lower-middle-income Upper-middle-income Source: Enterprise Surveys. 62 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 4.3: The proportion of young firms in total formal private sector employment is lower than elsewhere in the world 50 40 Percent of firms 30 20 10 0 Yemen, Rep. Morocco Egypt, West Bank Djibouti Lower- Tunisia Lebanon Jordan Upper- Arab Rep. and Gaza middle- middle- income income Five years old or younger More than 5 but 10 years or less in age Source: Enterprise Surveys. As discussed in chapter 2, the business environment var- Table 4.1: Average firm age ies substantially across these economies. This variation is Mean Median also found in the Doing Business sub-indicator measur- Djibouti 16 12 ing the ease of starting a business.20 Several MENA ES Egypt, Arab Rep. 14 12 economies maintain burdensome regulations for busi- Morocco 20 16 ness start-ups: Djibouti ranks last of 185 economies, the West Bank and Gaza 16 12 West Bank and Gaza 179th, and the Republic of Yemen Yemen, Rep. 24 21 and Lebanon rank 110th and 114th respectively. While Lower-middle-income 16 13 Egypt ranks 26th globally, followed by Tunisia, which is Jordan 16 13 50th, recent upheaval in both economies risks discourag- Lebanon 22 20 ing new entrants, which will limit competitive pressure on Tunisia 20 17 incumbent firms. At the same time, MENA ES economies Upper-middle-income 16 13 maintain remarkably high shares of employment in micro- Source: Enterprise Surveys. sized firms21 (which are not covered by the MENA ES) as well as pervasive informality.22 If productive firms are unable to grow over their lifecycle, the incentives for new other parts of the world (22 percent in all other economies firms to enter the market will be undermined. with ES data). There is substantial variation, however. Tunisia stands out Exporters account for a higher proportion of formal with exporters providing close to 61 percent of formal pri- jobs in the region than elsewhere in the world vate sector jobs (the result of an explicit policy of focusing One additional source of competitive market forces can on the export sector), followed by Jordan and Lebanon (47 come from abroad to the extent that economies engage and 32 percent). At the other extreme, only 15 percent of in foreign trade. As detailed in chapter 5, several firms in jobs in the Republic of Yemen are provided by exporting the MENA ES economies are internationally engaged; but firms. Not surprisingly, exporting firms contribute more a very large share of these traders tend to be SMEs, pos- to jobs in the relatively rich economies (Jordan, Lebanon, sibly due to market distortions. Similarly, the distribution and Tunisia) than elsewhere.23 From a policy perspective, of jobs shows that on average 30 percent of employment this international exposure may result in global factors in the formal private sector in the MENA ES economies influencing domestic employment. The task of policy then occurs in exporting firms (figure 4.4), more so than in Chapter 4: Jobs and skills in the formal private sector 63 Figure 4.4: The proportion of jobs provided by exporters is higher than in the rest of the world 80 70 Percent of total employees 60 50 40 30 20 10 0 Yemen, Rep. Morocco Djibouti Egypt, West Bank Lower- Lebanon Jordan Tunisia Upper- Arab Rep. and Gaza middle- middle- income income Source: Enterprise Surveys. is to maximize the gains offered by positive global shocks including upper-middle-income economies (37 percent) and guard against the negative ones. and lower-middle-income economies (29 percent). In the MENA ES region, the average percentage of The formal private sector’s women employed in the formal private sector as a whole contribution to women’s is even lower than the proportion of women in the labor employment force (18 percent compared with 24 percent, as shown in According to the United Nations Development Program’s figure 4.5).27 Because labor force data also include the un- Gender Inequality Index, in 2014, the MENA region was employed and sectors not covered by the MENA ES (such the second most unequal region for women, preceded as agriculture, government, the informal sector, and the only by Sub-Saharan Africa.24 These results are in part financial and social services sectors), the lower proportion driven by women’s low participation in the labor market: of women employed in firms in the MENA ES economies the region’s women tend to be comparatively well may be due to different factors: either unemployment educated, showing important advances in investment in is higher among women, or women tend to work more human capital, but their labor market participation remains in sectors not covered by the Enterprise Survey.28 Both low.25 Increasing women’s employment in the MENA ES factors seem to be at play and are suggestive of a gap in economies is important not only for purely economic women’s employment in the formal private sector.29 reasons, increasing the productive capacity of the region, but also for society’s well-being and stability. Women’s employment is higher in labor-intensive sectors and exporting firms Women’s employment is low compared with other Previous evidence suggests that women are more likely to regions be employed in sectors that are relatively labor-intensive Labor force participation rates for women in the MENA as well as in retail.30 In the MENA ES economies, labor- region are lower than the average for low- and middle- intensive manufacturing sectors—such as the production income economies, as previous reports have shown exten- of garments, footwear, leather, and furniture—have the sively.26 In the average firm in the MENA ES economies, highest average share of women workers (21 percent), women constitute 17 percent of the workforce (full-time followed by retail (20 percent), and other services (17 permanent workers). This is significantly lower than what percent). In other manufacturing sectors, which are is found in the rest of the world with ES data (34 percent), less labor-intensive, only 13 percent of employees are 64 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 4.5: The percentage of women employed in the formal private sector is smaller than the percentage of women in the total labor force 35 30 25 Percent 20 15 10 5 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, Rep. MENA ES Arab Rep. and Gaza Of total permanent workers in the formal private sector Of labor force aged 15-64 Source: Enterprise Surveys; ILOSTAT. women.31 There are important differences, however: business cities also tend to have a higher percentage of Djibouti, the West Bank and Gaza, and the Republic of women. These results indicate that other factors might Yemen stand out with low shares of women employed by explain the differences in women’s employment across labor-intensive manufacturers (figure 4.6). and within economies, factors probably associated with cultural norms and differential enforcement of customs Differences also emerge after accounting for basic firm and laws.32 characteristics (table A4.1): firms in the formal private sector in Djibouti, Lebanon, Morocco, and Tunisia tend Earlier studies generally support a positive effect of global- to employ significantly more women than firms in Egypt, ization on women’s employment.33 One reason could be Jordan, the West Bank and Gaza, or the Republic of that women tend to be concentrated in labor-intensive ex- Yemen. All else equal, firms located in capitals or main porting sectors that expand following trade liberalization. Figure 4.6: Labor-intensive manufacturing and retail have the highest percentage of women employees 60 50 Percent of women employees 40 30 20 10 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, Rep. MENA ES Arab Rep. and Gaza High labor intensity manufacturing Other manufacturing Retail Other services Source: Enterprise Surveys. Note: High labor intensity manufacturing includes manufacturing of wearing apparel, leather, and furniture. Chapter 4: Jobs and skills in the formal private sector 65 Another possibility is that by increasing competition, eight MENA ES economies have more than 10 such legal international trade makes it more expensive for employ- differences. ers to discriminate against women employees. MENA ES results confirm that the share of women employees The MENA ES data show real implications in terms of is 4 percentage points higher for firms that export,34 even women’s participation in ownership and top manage- after discounting other potential explanations such as the ment. Women own on average less than 8 percent of sector of activity and labor intensity (first column, table firms in the MENA ES economies, significantly lower A4.1). The larger percentage of women employed by than 16 percent in upper-middle-income economies and firms in manufacturing sectors with high labor intensity 13 percent in lower-middle-income economies. Similarly, compared with sectors with lower labor intensity is also only about 5 percent of firms in the MENA ES economies confirmed when accounting for basic firm characteristics have a woman top manager, compared with 19 percent (first column, table A4.1). in both lower-middle-income and upper-middle-income economies (figure 4.7). Women’s participation in top management and firm There is substantial variation across MENA ES economies ownership is low by international standards in the level of women’s participation in ownership and top Looking at women’s participation in entrepreneurship, management (table A4.1, columns 2 and 3). Even Tunisia MENA has the highest gender gap in the world: 12 per- and Lebanon—where women’s ownership is higher than cent of adult women are entrepreneurs compared with in peer economies—lag behind in terms of women in top 31 percent of adult men.35 MENA also has many legal management. Looking across the MENA ES economies, restrictions on women’s employment and entrepreneur- Egypt, Jordan, the West Bank and Gaza, and the Republic ship. The Women, Business, and the Law (WBL) 2016 of Yemen perform worse than any of the other economies. report measures legal gender differences in the areas of accessing institutions, using property, getting a job, A significantly larger percentage of women is employed by building credit, and going to court; it also measures legal firms with a woman top manager or by firms with one or incentives for women’s work and legislation on violence more women owners (figure 4.8). This is consistent with against women. According to the report, MENA hosts 18 previous literature indicating that women in top leader- of the 30 economies around the world that have 10 or ship positions can increase hiring of women, reduce sex more legal differences favoring men over women.36 All the segregation, and improve retention rates among women Figure 4.7: Women’s participation in top management positions is low Percent of firms with a woman as top manager 20 15 10 5 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, Lower- Upper- MENA ES All ES Arab Rep. and Gaza Rep. middle- middle- economies income income Source: Enterprise Surveys. 66 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 4.8: Firms managed by women have a higher proportion of women employees 60 Percent of full-time female workers 50 40 30 20 10 0 West Bank Yemen, Egypt, Djibouti Morocco Lower- Lebanon Tunisia Jordan Upper- and Gaza Rep. Arab Rep. middle- middle- income income Male top manager Female top manager Source: Enterprise Surveys. staff.37 Strikingly, women’s participation in ownership is the The business environment is not worse for women top only factor that helps to explain the probability of a firm managers and owners than for their male counterparts having a woman as the top manager. In addition, in firms Twenty-two objective measures and 17 subjective mea- where the top manager has fewer years of experience, the sures were used to detect potential differences in the same manager is more likely to be a woman (table A4.1, business environment faced by firms with women top column 3). managers compared with firms with men top managers.40 Only two indicators point to a worse business environ- ment for women: the percentage of firms that spent on Firm performance is not related to the gender of top security; and security costs as a percentage of annual managers, owners, or employees sales. In contrast, firms with women top managers enjoy A number of studies have shown that firms managed and a significantly better business environment according to owned by women tend to lag behind their male coun- indicators related to interactions with the government terparts in terms of productivity, growth, and firm size.38 (meetings with or inspections by tax officials, time to This could be due to gender discrimination in obtaining obtain licenses). The picture does not change much when finance or dealing with government, and prevailing laws looking at firms with at least one woman among the own- that tend to favor men over women. MENA ES results ers compared with firms with all male owners. provide no evidence of worse performance among firms managed or owned by women. Labor productivity and MENA ES data therefore contribute to the debate on the TFP levels, as well as growth rates of sales and employ- region’s low participation of women in the labor market by ment, are not associated with the top manager’s gender, ruling out the influence of firm performance or aspects of the proportion of women employed, or the presence of the business environment measured by the survey. The at least one woman owner.39 On the other hand, firms majority of such aspects are not affected by legal discrimi- that have at least one woman owner are more likely to nation, as they refer to power outages, custom clearance invest and innovate (columns 4 through 8 in table A4.1). waiting times, or bribes, for example. The results further Overall, performance does not help to explain the gender corroborate the idea that the legal and social framework gap in entrepreneurship and management rates. The next could instead play an important role in women’s participa- question is therefore whether women experience a more tion in the private sector.41 Furthermore, legal obstacles hostile business environment compared with men, limit- to starting a business may be such that only women who ing the ability of women-owned and women-managed can navigate this environment are ultimately able to run a businesses to survive. Chapter 4: Jobs and skills in the formal private sector 67 business, and those women encounter fewer difficulties The formal private sector’s in certain areas of the business environment. contribution to youth employment Youth employment is greater in young, dynamic, and The role of the legal framework for women’s innovative firms employment and entrepreneurship Youth employment is another major labor market chal- The WBL report shows that across the world, a higher lenge for the MENA region. On average across the MENA number of legal gender differences is associated with ES economies, labor force data show that young people more negative social and economic outcomes for women, between the ages of 15 and 29 represent 47 percent of such as a lower proportion of girls enrolled in secondary the working-age population and 40 percent of the labor school compared with boys, a lower employment rate for force.44 Compared with a total unemployment rate of 13 women, and a more pronounced wage gap between men percent in the region, unemployment among the young is and women.42 The same report and other previous work more than double at 30 percent.45 using ES data show that more legal gender differences are also associated with a lower percentage of firms with Previous research shows that unemployment in the a woman top manager and a lower percentage of women MENA region is mostly due to difficulties in entering the in a firm’s workforce.43 This is also true in the MENA ES labor market, since the majority of the unemployed are economies, as figure 4.9 shows. These results, combined first-time jobseekers.46 Hence, policies aimed at improv- with the fact that the business environment—as mea- ing labor market flexibility for new entrants, facilitating sured by the survey indicators—does not seem to be a information on entry-level jobs, and improving the linkages constraining factor for women’s entrepreneurship, sug- between the private sector and education institutions gest that eliminating gender discrimination would lead to could be key avenues for addressing the issue of youth better integration of women in the economy and therefore unemployment in the region. contribute to the development of the private sector in the MENA ES economies. The average share of workers under 30 in the formal pri- vate sector is 43 percent across the MENA ES economies. While there is no evidence of a systematic difference in Figure 4.9: More gender legal differences are associated youth employment across sectors (table A4.2, column with a lower percentage of women working in firms in the region 1), figure 4.10 shows that within manufacturing firms, a much smaller percentage of young people is employed 35 in non-production jobs (29 percent) compared with 30 Tunisia production jobs (45 percent). Since non-production jobs Percent of women in the workforce Djibouti Morocco in manufacturing firms typically require higher skills than 25 Lebanon production jobs,47 this evidence potentially points toward 20 a problem of skills mismatch for qualified young workers in the MENA ES economies. 15 Egypt, 10 Arab Rep. Further indication of the skills mismatch problem for young Jordan West Bank workers comes from firms’ propensity to provide training 5 and Gaza Yemen, Rep. to their workers and the severity of inadequate worker 0 10 15 20 25 30 education as an obstacle. In the MENA ES economies, Total number of gender legal differences firms with larger shares of young workers are more likely to provide training to their workers (table A4.3, column 1). Source: Enterprise Surveys and Women, Business and the Law 2016. This points to a skills mismatch problem with young work- ers since the need for training may arise because workers do not have the necessary skills for their job. 68 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 4.10: Fewer young people are employed in non-production manufacturing jobs than in other manufacturing or services jobs 80 70 Percent of workers <30 years old 60 50 40 30 20 10 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, Rep. MENA ES Arab Rep. and Gaza Manufacturing (production jobs) Manufacturing (non-production jobs) Retail Other services Source: Enterprise Surveys. In fact, the higher the share of university-educated em- Employment dynamics ployees the higher the probability of providing training Understanding the dynamics of employment or net job (tables A4.3 column 2). In addition, firms that use propor- additions—jobs created minus jobs terminated—can pro- tionately more young workers are significantly more likely vide useful insights on policy measures aimed at increas- to report skills shortages as a very severe constraint (table ing job creation in the MENA region. Dynamic analysis A4.4). Thus, a closer alignment of education curricula with using MENA ES data needs to be interpreted with caution: the requirements of industries is likely to improve job the data provide information only on growth rates for sur- prospects for the young in the region. viving firms, not for firms that exit the market. They also exclude very recent entrants and micro firms, which may The MENA ES results also indicate that firms that are affect the observed short-run growth rate of employment younger or larger tend to employ proportionately more and any conclusions about the impact of policy measures workers under 30 (table A4.2, column 1). This result, com- or economic shocks. Nonetheless, understanding growth bined with the evidence that younger firms in the MENA among surviving firms remains a useful starting point for ES economies create more jobs documented below and analyzing long-run employment growth, the size distribu- in previous work,48 suggests that encouraging firm entry tion of existing firms, and the impact of the entry and exit would help boost youth employment in the formal private patterns on surviving firms.49 sector. The survey results also indicate that firms with proportion- Young firms grow faster, but the average number of net ately more young employees are significantly more likely jobs created is similar for young and old firms in the to increase employment, to invest in fixed assets and to MENA ES economies innovate (table A4.2, columns 2–4). Although these results Consistent with the broader literature,50 firm-level growth cannot be interpreted as evidence of a causal relationship, rates of employment in the MENA ES economies be- they seem to indicate the presence of a “virtuous circle” tween 2009 and 2012 is much higher among relatively of young and innovative firms hiring younger employees younger firms. For example, for a typical firm that has and creating more jobs. been operating for five years or fewer, employment grows on average by 9.4 percent per annum compared with only 1.7 percent for a typical firm older than five years.51 The total number of new jobs does not vary much between Chapter 4: Jobs and skills in the formal private sector 69 4 percent became large (table 4.2).53 In the Republic of Table 4.2: Firm transitions across size categories Yemen, almost a third of firms that were medium-sized between 2009 and 2012 in 2009 became small in 2012, quite possibly due to the Average for the full sample conflict. These findings stand in contrast with ES data Status in 2012 on other lower-middle-income and upper-middle-income Small Medium Large Status in 2009 (5-19) (20-99) (100+) economies.54 In lower-middle-income economies, only 6 Small (5-19 employees) 93% 7% 0% percent of medium-sized firms became small after three Medium (20-99 employees) 14% 82% 4% years, while 4 percent became large. In upper-middle- Large (100+ employees) 0% 9% 91% income economies, 6 percent of medium-sized firms Source: Enterprise Surveys. became small, while 7 percent became large. This indicates that the period 2009–2012 may have been young and old firms. The average number of net jobs particularly difficult for medium-sized firms in the MENA added between 2009 and 2012 by firms under five years ES region in the context of challenging economic and old is not significantly different from the result for older political circumstances. Despite this, labor productivity firms (3.1 and 4.6 permanent employees respectively). in 2009 seems to be positively associated with a higher probability of becoming a medium-sized or large firm in 2012 (table A4.5).55 This suggests that productive firms Few firms expand or downsize over time were able to grow or maintain firm size despite political An extensive literature in labor economics suggests instability, which may have constrained greater growth. that the growth of firms over time reflects an important process of learning and selection, with some firms exiting Moreover, as stated in chapter 2, SMEs are at a disad- and others growing, thereby improving aggregate firm vantage, since they are more negatively affected by the productivity. The data show that firm dynamics in the inefficiencies of the business environment. Measures to MENA ES economies are weak. Relatively few firms address these inefficiencies might also serve as drivers moved from one size category (small, medium, or large) for more dynamic growth of such firms. to another between 2009 and 2012. This finding is illustrated in table 4.2, which summarizes Between 2009 and 2012, growth was faster for more the percentage of firms that move from one size category productive firms and slower for credit-constrained to another. Of the firms that were small in 2009, 93 per- firms cent were still small in 2012. Only 7 percent grew beyond In the MENA ES region, the employment growth rate 19 employees in 2012. Similarly, 82 percent of medium- between 2009 and 2012 is strongly associated with the la- sized firms and 91 percent of large firms remained in the bor productivity in 2009 (table A4.6), indicating that highly same size category. These findings are consistent with productive firms are able to generate new jobs at a faster the idea that distorted competition and privileged access rate than less productive firms, leading to the mostly posi- to the government by some firms—known to be widely tive employment growth rates presented in chapter 2. prevalent in the region—have blunted the dynamic forces that force firms to learn and grow over time.52 Although Another important factor for firm performance and firm dy- the employment transition matrix using ES data only con- namics is access to finance. Using the definition of credit siders surviving firms, the findings are in line with findings constraint introduced in chapter 3, the results in table A4.6 for Tunisia based on census data that also take account of show that the growth rate of employment in firms in firm exit (see box 4.1). the MENA ES economies is significantly lower for firms that are credit-constrained compared with those that are not. The employment growth rate is also lower for firms Medium-sized firms struggle to grow that report that corruption is a major constraint on their Across the region, nearly 14 percent of firms that were operations. In addition to and in line with the economic medium-sized in 2009 became small in 2012, while only literature discussing which firms create more jobs,56 table 70 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Box 4.1: Comparing ES transition matrix data with census findings from Tunisia that include information on rates of firm exit The ES data only consider firms that exist in 2012, and slightly different than the estimates reported in table 4.2 exclude firms that exited the market between 2009 and but this does not affect the results qualitatively. 2012. To help gauge the extent to which ES results may be biased by this fact, it is useful to compare the ES Table B4.1.2: Reweighted employment transition matrix findings with recent findings for Tunisia that are based for Tunisian firms between 2007 and 2011 based on on census data and that also take account of firm exit. census data but excluding firm exit and 1-person firms Status in 2011 Table B4.1.1 reproduces the employment transition ma- Status in 2007 Micro (2-9) SME (10-99) Large (100+) trix for Tunisian firms using census data over the period 2007-2011 and shows that the probability of exit is sub- Micro (2-9) 96% 4% 0% stantially larger for smaller firms: while only 6 percent of SME (10-99) 19% 76% 5% SMEs and large firms exited the market over this period, Large (100+) 4% 18% 78 % 9 percent of micro firms (2 to 9 employees) and 22 per- Source: Calculations based on Schiffbauer and others (2015). cent of one-person firms ceased to exist.a The estimates are very much in line with the ES data for Table B4.1.1: Employment transition matrix for Tunisian Tunisia reported in table B4.1.3. This lends support to the firms between 2007 and 2011 using census data finding that medium-sized firms in MENA ES are more Status in 2011 likely to become small than grow to large size, in contrast Status in Micro with other regions of the world, despite the lack of data 2007 Exited 1-person (2-9) SME (10-99) Large (100+) on firm exit. 1-person 22% 76% 2% 0% 0% Micro (2-9) 9% 21% 67% 3% 0% Table B4.1.3: Employment transition matrix for Tunisian SME 6% 11% 16% 63% 4% firms between 2009 and 2012 using ES data (10-99) Status in 2012 Large 6% 11% 3% 15% 65% Status in 2009 Small (5-19) Medium (20-99) Large (100+) (100+) Small (5-19) 94% 7% 0% Source: Schiffbauer and others (2015). Medium (20-99) 11% 85% 4% Large (100+) 0% 9% 91% To make it comparable to the employment transition matrix for MENA ES, table B4.1.2 reweights the Tunisian Source: Enterprise Surveys. census data to omit firms that exited the market and one-person firms that are not captured in MENA ES data. The firm size categories and the time period are a Schiffbauer and others (2015). A4.6 shows that younger and smaller firms have higher mismatch between the aspirations of graduates and the employment growth rates than older and larger firms. supply of rewarding jobs, it has also been argued that the region’s education systems fail to provide private sector employers with employees with the relevant skills. Skills, training, and employment Despite massive improvement in enrollment rates in Surprisingly, the share of firms in the formal private sector secondary and tertiary education, the quality of education that consider an inadequately educated workforce as a in the MENA region remains poor, particularly in providing major or very severe obstacle in the MENA ES econo- skills that are relevant for private sector employment.57 A mies is relatively low.58 Only in Morocco, Tunisia, and the major problem in education systems seems to be a focus Republic of Yemen is this share above the average levels in on competitive examinations as a screening mechanism lower-middle-income and upper-middle-income economies mainly aimed at securing access to public sector employ- (figure 4.11). Skills as an obstacle to firm growth are likely ment. Technical and vocational education and training, to have a cyclical component. During the period under which may be more suitable for private sector jobs, study, the MENA ES economies experienced growth are associated with lower status. While there is a great rates barely above population growth, making skills a less Chapter 4: Jobs and skills in the formal private sector 71 Figure 4.11: The proportion of firms reporting an inadequately educated workforce as a major or very severe constraint 35 30 25 Percent of firms 20 15 10 5 0 West Bank Egypt, Djibouti Yemen, Morocco Lower- Jordan Lebanon Tunisia Upper- and Gaza Arab Rep. Rep. middle- middle- income income SSource: Enterprise Surveys. pressing issue. On the other hand, skill shortages may be- three years.59 In other words, skill shortages seem to be come more salient once these economies start to recover. a particular concern for those firms that may have the highest growth potential. Firms that view an inadequately educated workforce as a very severe obstacle also tend to Skills-related constraints are seen as more severe by employ a higher share of university-educated employees firms that have grown quickly (see figure 4.12B).60 Figure 4.12A shows that firms that report an inadequately educated workforce as a very severe obstacle to their This could be interpreted in at least two different ways. operations tend to have grown faster in the preceding First, it could be that the inadequacy of the workforce is Figure 4.12 :Skill shortages are a particular concern for firms that grow rapidly and that rely more on university-educated employees Panel A: Employment growth and an inadequately educated Panel B: The share of university-educated employees and workforce as an obstacle to the enterprise inadequately educated workforce as an obstacle to the enterprise 8 50 Percent of university educated employees 7 40 6 Employment growth (%) 5 30 4 3 20 2 10 1 0 0 No Minor Moderate Major Very No Minor Moderate Major Very severe severe Source: Enterprise Surveys. 72 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY a problem for firms requiring higher levels of skills, there- The wage bill per worker in the fore indicating a scarcity of workers with tertiary-level formal private sector skills. Second, firms may have to resort to hiring more In addition to the number of jobs, the quality of jobs—in tertiary graduates to address the lack of skills in workers terms of wage rates—is also important, particularly with lower levels of education, reflecting a problem in the for the MENA ES region where the private sector has education system. failed to provide high-paying jobs to attract talent. Many in the region, especially young people, prefer to remain Training provision is low in MENA ES economies unemployed while seeking high-paying jobs in the public sector rather than taking up low-paying jobs offered by The education systems in the MENA ES economies have the private sector.62 This creates greater pressure on the failed to provide the necessary skills required by the government to provide more public sector jobs, adds to private sector. Training by the private sector could fill the unemployment, and dries up the flow of talent to the gap left by the education system. Across the MENA ES private sector. economies, however, the intensity of training provided by firms is low. A higher proportion of firms provide training in One narrative that has emerged in the MENA region as Morocco, Tunisia, Lebanon, and Djibouti (ranging from 22 a whole is that inflexible wages, including formal and de to 29 percent), but none of these economies exceeds the facto wage floors, may limit employment mobility and average shares of firms providing training in lower-middle- exacerbate skill mismatches. One report finds that a income and upper-middle-income economies at around “measurable share” of firms in Jordan and Egypt pay their 38 percent (figure 4.13). This is consistent with previous workers less than the mandated minimum wage.63 The findings that although training plays a prominent role in same report notes that minimum wage rules and collective active labor market programs in the region, it tends to wage agreements at the sector level—which establish a be class-based rather than on-the-job, and supply-driven negotiated wage minimum often linked to education level rather than coordinated with the private sector, thus and seniority—are often shirked.64 As these rules are often diverging from international best practices.61 tied to education level, private sector employers “do not absorb an ever growing graduate population at the wages foreseen for graduates” .65 Figure 4.13: Percent of firms offering formal training 45 40 35 30 Percent of firms 25 20 15 10 5 0 Egypt, West Bank Yemen, Rep. Djibouti Morocco Lower- Jordan Lebanon Tunisia Upper- Arab Rep. and Gaza middle- middle- income income Source: Enterprise Surveys. Chapter 4: Jobs and skills in the formal private sector 73 A lens to evaluate these trends is provided by the total firms do pay higher wages, then encouraging a business wage bill per worker. This is given by the total remunera- environment that allows firms to scale up their operations tion cost including wages, taxes, and social security pay- will lead to higher living standards for workers as well as ments divided by the number of employees at the firm a more equitable distribution of income between owners level. To account for local cost adjustments, it is defined of capital and labor. in terms of U.S. dollars adjusted for purchase power par- ity (PPP). While remuneration may reflect higher wages, However, the MENA ES economies seem to defy this it also includes taxes and social security contributions, trend. Larger firms in the MENA ES economies do not which can vary substantially between firms and across dedicate a greater share of their revenues toward their economies. wage bill; in fact, all else equal, larger firms tend to spend significantly less (table A4.7). This is consistent with the findings from chapter 2, which showed that larger firms More productive firms have higher wage bills per are actually less labor-intensive (measured by the wage worker bill cost) relative to smaller ones. Ideally, competitive forces should drive wages higher for more productive workers; but labor market imperfections One possible explanation is that larger firms tend to trans- suggest that ties between wages and worker productivity fer a larger share of returns to remunerate capital rather are not always watertight. In the MENA ES economies, than labor. Small firms may also adopt fewer labor-saving more productive firms—on a sales per employee basis66 technologies, and so are more reliant on labor relative —do have significantly larger wage bills per worker, in to their revenues, resulting in their higher average wage line with previous research (table A4.7), and this holds bill. Similarly, large firms may be able to leverage their in both lower-middle-income and upper-middle-income market position or privileged status to drive down wages economies (table A4.8).67 or other remuneration costs, including labor-related taxes. They may also be in a position to pay less given their This dynamic may indicate that more productive firms in comparably low labor demand (relative to other inputs) in the MENA ES region also dedicate more of that revenue an environment of high unemployment. per worker toward total remuneration (and this relationship is higher in lower-middle-income economies, table A4.8). While this may be considered as a sign of sound labor Higher wage bills are associated with university markets on the surface, it is important to take account of education in upper-middle-income MENA ES the limited size of the formal private sector and conse- economies quently the possible scarcity of those fairly remunerated A higher share of employees with tertiary education is or high-productivity private sector jobs. Consequently, it also related directly to higher wage bills per worker in is likely that many new entrants to the job market seek the upper-middle-income MENA ES economies: Tunisia, and are trained for public sector jobs—and not jobs in the Lebanon, and Jordan (table A4.8). While this may be an private sector—due in part to a relative scarcity of fairly indication of firms’ ability to recruit and pay skilled work- remunerated private sector jobs.68 ers, it is also likely to be a consequence of education-tied wage levels in these economies and possibly driven by public sector policy. In contrast, the percentage of em- Relative to revenue, larger firms spend less on ployees with a university degree is not tied to the average remuneration wage bill in lower-middle-income MENA ES economies, A well-established finding in the literature is that large which is a possible indication of distortions in the labor firms tend to pay their employees more.69 This so-called market, low quality higher education, or skills mismatch. “wage-size effect” has been linked to management qual- ity, the capacity of larger firms to attract and recruit better Similarly, exporting firms that provide a large share of total employees, and issues of scale for larger firms that make jobs have a much higher wage bill per worker than non- it harder to monitor and evaluate employees.70 This rela- exporting firms (figure 4.14). Lastly, the median wage bill tionship can have important policy implications. If large per worker for firms more than 5 years old is higher than 74 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 4.14: The wage bill per worker is higher for exporting firms in most MENA ES economies than for non-exporting firms $25,000 Wage bill per worker (USD) $20,000 $15,000 $10,000 $5,000 $0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, Rep. Arab Rep. and Gaza Non-exporter Exporter Source: Enterprise Surveys. in young firms in all the MENA ES economies with the in the formal training of employees, they are also more exception of Morocco. On average across all the MENA likely to complain about the adequacy of workforce educa- ES economies, it equals US$10,888 for old firms and a tion levels. Skill shortages are striking in the context of much lower US$8,832 for the young firms. the high share of university educated young people in the region. There seems to be evidence of a mismatch be- tween the skills learned in the formal education process Policy conclusions and those required by the business community, indicating Recent political upheaval as well as pressure on public the need for more effective training. budgets will limit public sector employment as a source of jobs in the MENA ES region. This means that the formal Policies should not constrain firm growth or discourage private sector will need to play an increasingly important new firm entry. In some MENA ES economies, burden- role in providing critically needed jobs. some regulations for start-up businesses may prevent new and dynamic firms from entering the market. Well- Large firms provide the majority of jobs in the formal heeled firms can take advantage of a lack of competitive private sector employment, compared with formal SMEs. forces to extract rents and reduce overall efficiency. Other They are also more productive, though their activities are forces that hamper competition (such as privileged access skewed toward inefficiently high capital intensity, with to markets, licensing and contracts) would have similar associated lower remuneration of labor. At the same effects. time, SMEs in the region typically fail to grow. Given that distorting incentives may favor capital at the cost of labor While lowering the barriers to entry for new (possibly and that the SMEs seem to be more penalized by the more efficient and competitive firms) is one avenue for business environment, carefully assessing current poli- employment growth, ensuring that future job creation is cies, removing privileges, and more generally supporting inclusive of women and young people is another. Inclusive competition may have implications for inclusive growth. growth is important not just for economic or egalitarian reasons, but also for ensuring greater political stability and Fast-growing firms are also those that have higher pro- for coping with cross-border migration and the refugee ductivity, possibly indicating a partial reallocation of jobs crisis currently affecting the region. toward more productive firms. Fostering such firms can encourage the development of the private sector as a The MENA ES economies are characterized by lower whole. While fast-growing firms are more likely to invest rates of women’s employment, management, and private Chapter 4: Jobs and skills in the formal private sector 75 sector ownership compared with the rest of the world. 13 Distributions for MENA ES and comparators are based The benefits of job growth will be limited if women are on the coverage of the ES and therefore they exclude the micro sector—less than 5 employees—and the informal or prevented from being employees or employers, either unregistered sector. through restrictions on jobs they can do or on their access 14 Hsieh and Klenow (2012). to real assets. Similarly, women’s employment is higher 15 See for example, Pages and others (2009) and Haltiwanger in labor-intensive sectors and among exporting firms. An and others (2014). expansion of labor-intensive and exporting sectors may 16 See Rijkers and others (2014) or Diwan and others (2015). help to provide more jobs for women, but more opportuni- 17 Schiffbauer and others (2015). ties are also needed in capital-intensive sectors to reduce 18 The methodology of the ES can introduce a downward sector segregation and women’s greater vulnerability to bias to the contribution of young firms as samples are external shocks to the economy. drawn from sampling frames that typically are several years old. In the MENA ES project, however, most Likewise, young jobseekers and newly employed workers sampling frames were current and whenever older frames in the region must be in a position to be well integrated were used—the oldest dating to 2012—the frame was into the private sector. Young, fast-growing, and in- updated with current listings of firms operating in the economy. novative firms tend to employ a greater share of young 19 One caveat of these results is that firms are randomly workers. Ensuring the entry and growth of such firms drawn from a sample frame that often necessarily omits will likely have knock-on effects on youth employment. A the youngest start-ups, having an upward age bias. re-orientation of education and training systems toward 20 Figures are based on Doing Business 2013. learning skills that are relevant for private sector employ- 21 Micro firms are defined as those registered firms with ment, with greater status given to vocational training, will less than 5 employees and informal firms are unregistered be likely to facilitate growth of high quality employment in firms. the region. Similarly, creating conditions that allow larger 22 Schiffbauer and others (2015), World Bank (2014). firms to provide greater remuneration to employees—or 23 Chapter 5 discusses various issues related to exporting allowing better-remunerated small firms to add jobs—will firms in more detail. attract talented workers into the private sector. 24 The UNDP’s Gender Inequality Index includes measures of reproductive health, empowerment and labor market participation. A low value indicates low inequality between women and men. The scores for the different regions in Endnotes 2014 were: 0.317 for ECA, 0.331 for EAP , 0.416 for LAC, 0.539 for SAR, 0.546 for MENA, and 0.578 for AFR (UNDP 1 See, for example, ILO (2013), World Bank (2013a), 2014). ILO-KILM Database, via World Bank (2013a), World Bank 25 World Bank (2013b). (2011). 26 See World Bank (2013b), Verme (2014). 2 World Bank (2004). 27 To compare ES micro data to aggregate indicators from 3 World Bank (2011). other sources, such as labor force statistics, the average 4 Schiffbauer and others (2015). percentage of women employed in the formal private 5 See, for example, World Bank (2011, 2013a). sector as a whole was calculated as the weighted sum 6 ILO (2014). of all female employees across all firms, divided by the total number of employees in all firms. In all other 7 See, for example, World Bank (2013a). instances in this section, the percentage of women in the 8 See, for example, World Bank (2013a) and Devarajan and workforce refers to the firm-level average. Labor force others (2014). data from ILOSTAT: 2013 for West Bank and Gaza; 2012 for 9 World Bank (2013a). Morocco, Tunisia; 2010 for Egypt; 2007 for Lebanon; 2004 10 See, for example, Devarajan and others (2014). for Jordan; 1999 for the Republic of Yemen. Data are not available for Djibouti. 11 Ayyagari and others (2014) and Aga and others (2015)). 28 The methodology is also very different between the ES 12 Schiffbauer and others (2015) and Rijkers and others (2014) data and labor force data. The latter are typically based on note, for example, that outside of micro firms, which are surveys of the population. not included in the MENA ES data, large firms are the second largest source of private sector employment. 76 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY 29 The World Bank (2013b) report indicated that in many labor laws, lack of skilled workers, corruption, courts, MENA economies, unemployment rates among women access to finance, tax rates, tax administration, customs aged 15–24 were around 50 percent compared with 10 and trade regulations, competition from informal sector percent to 20 percent for men of the same ages. At the firms, access to land, crime and security, obtaining same time, the report highlights that women in MENA licenses and permits, and regulatory policy uncertainty. economies tend to consider employment in the public 41 World Bank (2013b). sector preferable to a job in the private sector (p. 20). 42 World Bank (2015). 30 See Joekes (1995), Bardasi and others (2011), Amin and 43 Amin and Islam (2015). Islam (2014). 44 Labor force data from ILOSTAT: 2013 for West Bank & 31 Manufacturing sectors are classified as follows, based Gaza; 2012 for Morocco, Tunisia; 2010 for Egypt; 2007 for on Xu (2003). High labor intensity: wearing apparel, Lebanon; 2004 for Jordan; 1999 for the Republic of Yemen leather, furniture; moderate labor intensity: wood (no data available for Djibouti). Population data: authors’ products, publishing, printing; low labor intensity: food, calculation from UN Population Division World Population tobacco, textiles, paper and paper products, rubber and Prospects: The 2012 Revision (no data available for West plastics, machinery and equipment, electrical machinery Bank and Gaza). and apparatus, motor vehicles, transport equipment, other manufacturing; very low labor intensity: coke, 45 Data from World Development Indicators (2013). Youth refined petroleum products and nuclear fuel, chemicals category refers to labor force participants between 15 and and chemical products, basic metals, fabricated metal 24 years old. products, other non-metallic mineral products. 46 Kabbani and Kothari (2005). 32 For example, Amin and Islam (2015, 2016) show provision 47 Berndt, Morrison and Rosenblum (1992), Davis and of paternity leave and presence of laws prohibiting Haltiwanger (1991). discrimination against women in hiring practices as 48 Schiffbauer and others (2015). defined by World Bank’s Women, Business and Law (WBL) 49 See, for example, the large literature on the size data boost women’s employment prospects. distribution of surviving firms and its economic 33 See Amin and others (2015) and papers cited therein. implications following the seminal work of Gibrat (1931). 34 That is, firms that export more than 10 percent of their 50 See, for example, Haltiwanger and others (2013) and sales compared with firms that export less than 10 Ayyagari and others (2011, 2014). percent or do not export at all. 51 This results holds after accounting for firm characteristics. 35 Result from an OECD (2014) report using data from the 52 See Rijkers and others (2014) and Schiffbauer and others Global Entrepreneurship Monitor (GEM). (2015). 36 World Bank (2015). 53 The finding that medium-sized firms were more likely 37 See Carrington and Troske (1995) and (1998); Huffman and to reduce their size and become small firms (than others (2010); Giuliano and others (2006); Kurtulus and increase size and become large firms) between 2009 Tomaskovic-Devey (2012). and 2012 also holds in a regression that controls for firm 38 See, for example, Brush (1992) and Sabarwal and Terrell characteristics—including age, economy, sector and (2008). locality. 39 Estimates of total factor productivity are available for 54 The time period for the transition matrices for the MENA manufacturers only. Chapter 2 provides details on how ES economies and other ES economies are different. these estimates are obtained. Hence, some caution is needed in comparing these. 40 The 22 objective measures cover areas including the 55 The results in Table A4.5 are qualitatively similar when quality of power supply, water shortages, waiting time total factor productivity in 2012 is used in the analysis to obtain various licenses and permits, customs delays instead of labor productivity in 2009. In other words, firms (in exporting and importing goods), bribes paid or asked that were medium-sized in 2009 and are less productive for in dealing with government officials, inspections are more likely to become small in 2012, while medium- and meetings with tax officials, time spent by senior sized firms that have higher total factor productivity are management of the firm in dealing with business more likely to maintain or expand their size. regulations (time tax), crime and security losses (incidence 56 See, for example, Schiffbauer and others (2015, pp.25-26) and cost). The 17 subjective measures include the firm’s and the references therein. perception on the amount of bribes paid to public officials 57 World Bank (2013a). by other firms like itself to get things done, and whether or not the following is a major obstacle for the firm’s 58 Firms in MENA have to report the severity of the obstacle operations—electricity, transport, telecommunications, on a scale ranging from 0 to 4 where “no obstacle” is coded as 0 and “very severe obstacle” as 4. Chapter 4: Jobs and skills in the formal private sector 77 59 The positive relationship between employment growth and reporting inadequately educated workforce as an obstacle holds when using the scale 0-4 as presented in figure 4.13A and when controlling for firm characteristics and economy fixed effects. 60 The relationships in figure 4.12 hold after accounting for firm and economy characteristics. 61 World Bank (2013a). 62 See, for example, Devarajan and others (2014) and World Bank (2013a). 63 World Bank (2013a). 64 Ibid. Musette and Mohamed-Meziani (2011). 65 World Bank (2013a). 66 Estimations are only provided for labor productivity as total wage bill cost is the main factor input for TFPR. 67 See, for example, Haltiwanger and others (1999), Haltiwanger and others (2007) and Dunne and others (2004). 68 World Bank (2013a). 69 These findings are expansive and build on the seminal work of Brown and Medoff (1989), using data from the U.S. 70 For a detailed discussion see Idson and Oi (1999). 78 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Appendix A4 Table A4.1: Percentage of women workers, probability of a firm having a woman owner or top manager, and key performance indicators (1) (2) (3) (4) (5) (6) (7) (8) Female Female Labor Real annual Annual Purchase of full-time participation in Female top productivity sales growth employment fixed assets in Innovator Dependent variable workers (%) ownership (Y/N) manager (Y/N) (log) (%) growth (%) last FY (Y/N) (Y/N) Size (log) 0.80 0.07* -0.08 -0.03 1.90*** 3.02*** 0.32*** 0.19*** (0.605) (0.038) (0.052) (0.040) (0.560) (0.398) (0.040) (0.040) Age (log) -0.91 0.10* 0.07 0.01 -1.85* -4.02*** -0.17*** 0.00 (0.803) (0.059) (0.079) (0.061) (1.103) (0.857) (0.056) (0.056) High labor intensity 8.40*** 0.06 0.03 -0.76*** -4.06 -3.34 -0.55** 0.11 manufacturing (Y/N) (2.624) (0.163) (0.179) (0.193) (2.937) (2.266) (0.213) (0.151) Moderate labor intensity 1.27 -0.09 0.20 -0.68*** 1.24 0.49 0.01 0.30 manufacturing (Y/N) (3.289) (0.239) (0.416) (0.227) (2.898) (2.282) (0.240) (0.259) Very low labor intensity -7.01*** -0.14 -0.19 -0.02 -0.75 1.07 -0.18 0.06 manufacturing (Y/N) (2.083) (0.151) (0.266) (0.166) (3.025) (1.480) (0.164) (0.147) Retail (Y/N) 5.00** 0.00 -0.03 0.16 -4.54* 0.06 -0.12 -0.27* (2.313) (0.136) (0.193) (0.163) (2.401) (1.509) (0.143) (0.147) Other services (Y/N) 1.34 0.01 -0.19 -0.26* -3.66 0.50 -0.16 -0.17 (1.949) (0.113) (0.164) (0.157) (3.113) (1.127) (0.144) (0.114) Exporter (Y/N) 3.84** 0.09 -0.03 0.18* 0.39 -1.07 -0.09 0.15 (1.492) (0.117) (0.173) (0.103) (1.569) (1.200) (0.115) (0.118) Capital/main business 4.32*** 0.08 0.29* 0.43*** -1.01 1.20 0.16 -0.21** city (Y/N) (1.619) (0.098) (0.163) (0.132) (2.164) (1.040) (0.125) (0.101) Manager experience -0.05 0.00 -0.03*** 0.00 -0.17** -0.12** 0.00 0.00 (years) (0.053) (0.004) (0.008) (0.005) (0.080) (0.045) (0.004) (0.005) Djibouti (Y/N) 10.54*** 0.17 0.22 0.07 12.81*** 6.84*** 0.77*** 0.91*** (2.808) (0.153) (0.202) (0.184) (4.008) (1.555) (0.170) (0.157) Jordan (Y/N) -4.69** -0.03 -0.93*** 0.14 7.30*** 6.13*** 0.34* 0.25 (1.875) (0.161) (0.285) (0.147) (1.688) (1.432) (0.189) (0.169) Lebanon (Y/N) 11.06*** 0.77*** -0.79*** 0.90*** 9.02*** 5.58*** 1.08*** 0.82*** (2.238) (0.146) (0.276) (0.154) (2.928) (1.554) (0.160) (0.155) Morocco (Y/N) 14.09*** 0.45*** -0.38* 0.49*** 10.25*** 6.69*** 0.87*** 0.65*** (1.907) (0.118) (0.201) (0.163) (2.074) (1.177) (0.143) (0.139) Tunisia (Y/N) 17.72*** 0.96*** -0.22 0.81*** 1.19 2.82** 0.99*** 0.50*** (2.056) (0.123) (0.212) (0.123) (2.194) (1.273) (0.146) (0.132) West Bank And Gaza -3.22 -0.04 -0.88** -0.09 14.25*** 10.47*** 0.90*** 0.40*** (Y/N) (2.177) (0.184) (0.412) (0.139) (2.866) (1.703) (0.158) (0.154) Yemen, Rep. (Y/N) -4.75** -0.46** -0.47 -0.86*** -0.52 -1.02 0.89*** 1.09*** (2.183) (0.191) (0.316) (0.288) (7.148) (1.786) (0.289) (0.124) Female participation in 0.03** 0.01*** 0.00 0.00 -0.02 0.00** 0.00** ownership (Y/N) (0.015) (0.002) (0.001) (0.015) (0.011) (0.001) (0.001) Female top manager 0.14*** -0.00* 0.01 -0.01 0.00 0.00 (Y/N) (0.036) (0.002) (0.025) (0.017) (0.002) (0.002) Female full time workers 0.00 0.04 0.00 0.00 0.00 (%) (0.003) (0.034) (0.018) (0.002) (0.002) Constant 7.28** -1.54*** -1.41*** 10.88*** -2.24 3.02 -1.60*** -1.39*** (2.834) (0.197) (0.244) (0.186) (3.739) (2.141) (0.190) (0.180) Observations 5,077 5,625 5,624 4,553 3,697 4,476 5,048 5,034 Source: Enterprise Surveys. Note: Standard errors in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Innovator means the firm has introduced a new or significantly improved product, service, or process. Manufacturing sectors are classified as follows, based on Xu (2003): High labor intensity: wearing apparel, leather, furniture; moderate labor intensity: wood products, publishing, printing; low labor intensity: food, tobacco, textiles, paper and paper products, rubber and plastics, machinery and equip- ment, electrical machinery and apparatus, motor vehicles, transport equipment, other manufacturing; very low labor intensity: coke, refined petroleum products and nuclear fuel, chemicals and chemical products, basic metals, fabricated metal products, other non-metallic mineral products. All regressions control for a dummy variable indicating whether at least 10 percent of the firm is owned by foreign agents and economy fixed effects. Ordinary least squares regression coefficients reported for columns 1, 4, 5, 6; probit regression coefficients reported for columns 2, 3, 7 and 8. Chapter 4: Jobs and skills in the formal private sector 79 Table A4.2: Percentage of workers under 30 and key performance indicators (1) (2) (3) (4) Percentage of workers Annual employment Purchase of fixed assets Dependent variable under 30 growth (%) in last FY (Y/N) Innovator (Y/N) Size (log) 3.39*** 2.84*** 0.33*** 0.18*** (0.819) (0.455) (0.045) (0.048) Age (log) -8.52*** -2.84*** -0.18*** 0.01 (1.231) (0.896) (0.063) (0.067) High labor intensity manufacturing (Y/N) -0.56 -3.10 -0.50** 0.12 (2.989) (2.276) (0.228) (0.167) Moderate labor intensity manufacturing (Y/N) -1.87 2.95 0.08 0.21 (3.234) (2.788) (0.278) (0.331) Very low labor intensity manufacturing (Y/N) -0.22 1.83 0.10 0.32* (3.641) (1.597) (0.199) (0.194) Retail (Y/N) -1.44 0.73 -0.03 -0.34** (2.937) (1.613) (0.159) (0.163) Other services (Y/N) -3.27 1.42 -0.04 -0.16 (2.652) (1.193) (0.155) (0.135) Percentage of under 30 0.06*** 0.01*** 0.01** (0.02) (0.002) (0.002) Constant 52.81*** -2.16 -1.92*** -1.55*** (4.147) (2.532) (0.246) (0.239) Observations 4,149 3,689 4,135 4,115 Source: Enterprise Surveys. Note: Standard errors in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Innovator means the firm has introduced a new or significantly improved product, service, or process. Manufacturing sectors are classified as follows, based on Xu (2003): High labor intensity: wearing apparel, leather, furniture; moderate labor intensity: wood products, publishing, printing; low labor intensity: food, tobacco, textiles, paper and paper products, rubber and plastics, machinery and equipment, electrical machinery and apparatus, motor vehicles, transport equipment, other manufacturing; very low labor intensity: coke, refined petroleum products and nuclear fuel, chemi- cals and chemical products, basic metals, fabricated metal products, other non-metallic mineral products. All regressions control for dummy variables indicating if there is at least one woman among the owners, if the top manager of the firm is a woman, if at least 10 percent of the firm is owned by foreign agents, if at least 10 percent of annual sales of the firm are made abroad, and they control for the years of experience the top manager of the firm has working in the industry, and economy fixed effects. Ordinary least squares regression coefficients reported for columns 1 and 2; probit regression coefficients reported for columns 3 and 4. Table A4.3: Probability of offering training Table A4.4: Probability of reporting skill shortages as a Formal training (Y/N) constraint Dependent variable (1) (2) Inadequately educated workforce a very Dependent variable severe constraint (Y/N) Proportion of workers 0.66*** 0.61** younger than 30 Proportion of workers 0.67** (0.235) (0.241) younger than 30 (0.292) Share of university 0.79*** educated employees Constant -2.84*** (0.215) (0.383) Constant -2.21*** -2.30*** Number of observations 4,386 (0.264) (0.272) Source: Enterprise Surveys. Number of observations 4,461 4,331 Note: Simple OLS estimations using survey-weighted observations (using Stata’s svy Source: Enterprise Surveys. prefix). Standard errors in parentheses. ***, ** and * denote statistical significance Note: Simple probit estimations using survey-weighted observations (using at the 1, 5 and 10 percent levels respectively. Variables omitted from the table: Stata’s svy prefix). Standard errors in parentheses. ***, ** and * denote statistical Foreign ownership, exports, young firms, firm size, manager university education, significance at the 1, 5 and 10 percent levels respectively. Variables omitted from manager experience, sector, locality, and economy fixed effects. the table: Foreign ownership, exports, young firms, firm size, manager university education, manager experience, sector, locality, and economy fixed effects. 80 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A4.5: More productive firms are more likely to Table A4.7: The wage-size effect in the MENA ES region expand in size Log (Average wage bill, PPP-adjusted) (1) (2) (3) Dependent variable (1) (2) Small firm in Medium firm in Large firm in Size (log) -0.06** -0.09*** Dependent variable 2012 (Y/N) 2012 (Y/N) 2012 (Y/N) (0.031) (0.033) Log of labor productivity -0.16*** 0.09** 0.19*** (PPP) in 2009 Age (log) 0.07** (0.041) (0.035) (0.049) (0.035) Small (5-19 employees) 2.64*** -2.40*** -1.66*** in 2009 (Y/N) Labor productivity (2012 USD) 0.39*** (0.119) (0.105) (0.238) (0.033) Large (+100 employees) -3.07*** -2.29*** 3.13*** in 2009 (Y/N) Manager has university education -0.03 (0.364) (0.137) (0.156) (Y/N) (0.074) Young firms (0-10 years) -0.08 0.09 -0.08 (Y/N) Percentage of workers with 0.14 (0.119) (0.11) (0.162) university degree (0.133) Constant 0.87* -0.37 -3.49*** Formal training (Y/N) 0.13 (0.477) (0.413) (0.586) (0.080) Number of observations 4,365 4,365 4,365 Constant 9.74*** 5.41*** Source: Enterprise Surveys. Note: PPP—purchasing power parity. The regressions include controls for economy, (0.232) (0.440) sector and locality fixed effects. Standard errors in parentheses. ***, ** and * Observations 5,348 4,668 denote statistical significance at the 1, 5 and 10 percent levels respectively. Probit regression coefficients are reported. R-squared 0.166 0.376 Source: Enterprise Surveys. Note: OLS regressions. Standard errors in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Economy and 2-digit sector fixed effects not shown. Table A4.6: The rate of growth of employment is lower for firms that have lower initial labor productivity level and for credit-constrained firms Annual employment growth (%) Credit-constrained (partially and fully) (Y/N) -4.06*** (1.399) Log of labor productivity (PPP) winsorized, 1.26*** 3 FY ago (0.382) Corruption: major constraint (Y/N) -1.99** (0.997) Small firms (based on size 3 FY ago) (Y/N) 6.81*** (1.419) Large firms (based on size 3 FY ago) (Y/N) 1.71 (1.273) Young firms (0-10 years) (Y/N) 3.52*** (1.309) Constant -18.19*** (5.075) Sample size 3,911 R-squared 0.171 Source: Enterprise Surveys. Note: OLS regressions. Standard errors in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. The regressions include controls for economy, 2-digit sector and locality fixed effects. Chapter 4: Jobs and skills in the formal private sector 81 Table A4.8: The wage-size effect in the MENA ES region Lower-middle-income Upper-middle-income Dependent variable: Log (Average wage bill, PPP-adjusted) (1) (2) (3) (4) Size (log) -0.10** -0.12*** -0.03 -0.06 (0.047) (0.046) (0.034) (0.042) Age (log) 0.08 0.06 (0.051) (0.036) Labor productivity (2012 USD) 0.43*** 0.29*** (0.042) (0.048) Manager has university education (Y/N) -0.06 0.04 (0.114) (0.080) Percentage of workers with university degree 0.00 0.38** (0.171) (0.167) Formal training (Y/N) 0.17 0.09 (0.112) (0.110) Constant 9.85*** 5.07*** 9.57*** 6.25*** (0.271) (0.574) (0.228) (0.520) Observations 3,782 3,207 1,566 1,461 R-squared 0.152 0.387 0.039 0.198 Source: Enterprise Surveys. Note: OLS regressions. Standard errors in parentheses. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Economy and 2-digit sector fixed effects not shown. 82 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY References Davis, Steve J., and John Haltiwanger. 1991. “Wage dispersion between and within U.S. manufacturing plants, Aga, Gemechu, David. C. Francis, and Jorge Rodriguez Meza. 1963-1986, ” Brookings Papers on Economic Activity: 2015. “SMEs, age, and jobs: A review of the literature, Microeconomics, (1991): 115-200. ” World Bank Policy Research metrics, and evidence. Working Paper No. 7493, World Bank, Washington, DC. Devarajan, Shanta, and Lili Mottaghi. 2014 “Predictions, Perceptions and Economic Reality—Challenges of Amin, Mohammad and Asif Islam. 2014. “Are there more seven Middle East and North Africa countries described female managers in the retail sector? Evidence from in 14 Charts” Middle East and North Africa Quarterly ” Journal of Applied survey data in developing countries. Economic Brief, (July), World Bank, Washington, DC. Economics, 17(2): 213-28. Diwan, Ishac, Phillip Keefer, and Marck Schiffbauer. 2015. Amin, Mohammad and Asif Islam. 2015. “Does mandating “Pyramid capitalism: Political connections, regulation, nondiscrimination in hiring practices influence women’s and firm productivity in Egypt.” World Bank Policy employment? Evidence using firm-level data. ” Feminist Research Working Paper No. 7354. Washington DC: Economics, 21(4): 28-60. World Bank. Amin, Mohammad, Asif Islam, and Khrystyna Kushnir. 2015. Dunne, Timothy, Lucia Foster, John Haltiwanger, and Kenneth “Do exports hurt or help women employment in R. Troske. 2004. “Wage and productivity dispersion ” developing economies? Evidence using firm-level data. in United States manufacturing: The role of computer Mimeograph. Washington DC: World Bank. investment. ” Journal of Labor Economics 22(2): Ayyagari, Meghana, Asli Demirgüç-Kunt, and Vojislav 397-430. Maksimovic. 2011. “Small vs. young firms across the Gibrat, Robert. 1931. Les inégalités économiques; applications: world: Contribution to employment, job creation, and aux inégalités des richesses, à la concentration des growth. ” World Bank Policy Research Working Paper No. entreprises, aux populations des villes, aux statistiques 5631, World Bank, Washington, DC. des familles, etc., d’une loi nouvelle, la loi de l’effet Ayyagari, Meghana, Asli Demirgüç-Kunt, and Vojislav proportionnel. Paris: Librairie du Recueil Sirey. Maksimovic. 2014. “Who creates jobs in developing Giuliano, Laura, Jonathan Leonard, and David I. Levine. 2006. countries?” Small Business Economics 43(1): 75-99. “Do race, gender, and age differences affect manager- Bardasi, Elena, Shwetlana Sabarwal, and Katherine Terrell. employee relations? An analysis of quits, dismissals, 2011. “How do female entrepreneurs perform? Evidence and promotions at a large retail firm.” Institute for ” Small Business from three developing regions. Research on Labor and Employment Working Paper No. Economics 37(4):417-41. 151-07. University of California, Berkeley, CA. Berndt, Ernst R., Catherine J. Morrison, and Larry S. Haltiwanger, John, Julia Lane, and James Spletzer. 2007. Rosenblum. 1992. “High-tech capital formation and “Wages, productivity, and the dynamic interaction of labor composition in US manufacturing industries: an businesses and workers. ” Labour Economics 14(3): ” NBER Working Paper No. 4010. exploratory analysis. 575-602. Brown, Charles, and James Medoff. 1989. “The employer ———. 1999. “Productivity differences across employers: size-wage effect.” Journal of Political Economy 97(5): ” The roles of employer size, age, and human capital. 1027-59. American Economic Review, Papers and Proceedings Brush, Candida G. 1992. “Research on women business 89(2): 94–98. owners: Past trends, a new perspective and future Haltiwanger, John, Ron Jarmin, and Javier Miranda. 2013. ” Entrepreneurship Theory and Practice 16(4): directions. “Who creates jobs? Small vs. large vs. young.” The 5-30. Review of Economics and Statistics 95(2): 347-361. Carrington, William J., and Kenneth R. Troske. 1995. “Gender Haltiwanger, John, Stefano Scarpetta, and Helena Schweiger. ” Journal of Human segregation in small firms. 2014. “Cross country differences in job reallocation: Resources 30(3): 505–33. ” Labour The role of industry, firm size and regulations. ———. 1998. “Sex segregation in U.S. manufacturing.” Economics 26(1): 11-25. Industrial and Labor Relations Review 51(3): 445–64. Huffman, Matt L., Philip N. Cohen, and Jessica Pearlman. Crivelli, Ernesto, Davide Furceri and Joël Toujas-Bernaté. 2010. “Engendering change: Organizational dynamics 2012. “Can policies affect employment intensity of and workplace gender desegregation, 1975–2005. ” growth? A cross-country analysis.” IMF Working Paper Administrative Science Quarterly 55(2):255–77 . No. WP/12/218, International Monetary Fund (IMF), ILO. 2013. Global Employment Trends 2013: Recovering from Washington, DC. a Second Jobs Dip, International Labour Organization (ILO), Geneva. Chapter 4: Jobs and skills in the formal private sector 83 ILO. 2014. Global Employment Trends 2014: Risk of a jobless Schiffbauer, Marc, Abdoulaye Sy, Sahar Hussain, Hania recovery? International Labour Organization (ILO), Sahnoun, and Philip Keefer. 2015. Jobs or Privileges: Geneva. Unleashing the Employment Potential of the Middle Joekes, Susan. 1995. “Trade-related employment for women East and North Africa. Washington, DC: World Bank. in industry and services in developing countries. ” Verme, Paolo. 2014. “Economic development and female labor Occasional Paper 5, Geneva: United Nations Research participation in the Middle East and North Africa: A test Institute for Social Development. of the U-shape hypothesis, ” Policy Research Working Kabbani, Nader, and Ekta Kothari. 2005. “Youth employment Paper 6927 , World Bank, Washington, DC. in the MENA region: A situational assessment. ” Xu, Bin. 2003. “Capital abundance and developing country World Bank Social Protection Discussion Paper 534. production patterns.” University of Florida, Department Washington DC: World Bank. of Economics. http://www.ceibs.edu/faculty/xubin/ Kurtulus, Fidan Ana, and Donald Tomaskovic-Devey. 2012. Puzzle.pdf “Do female top managers help women to advance? World Bank. 2004. Unlocking the Employment Potential in A panel study using EEO-1 records.” The ANNALS of the Middle East and North Africa: Toward a New Social the American Academy of Political and Social Science Contract. Washington DC: World Bank. 639(1): 173-197 . World Bank. 2011. Investing for Growth and Jobs. OECD. 2014. Women in Business 2014: Accelerating Washington DC: World Bank. Entrepreneurship in the Middle East and North Africa World Bank. 2013a. Jobs for Shared Prosperity: Time for Action Region, Paris: OECD Publishing. in the Middle East and North Africa. Washington DC: Idson, Todd L., and Walter Y. Oi. 1999. “Workers are more World Bank. ” American Economic Review productive in large firms. World Bank. 2013b. Opening Doors: Gender Equality and 89(2): 104-108. Development in the Middle East and North Africa. Pagés, Carmen, Gaëlle Pierre, and Stefano Scarpetta. 2009. Washington DC: World Bank. Job creation in Latin America and the Caribbean: Recent World Bank. 2014. More jobs, better jobs: a priority for Egypt. trends and policy challenges. World Bank, Washington Washington, DC: World Bank. DC. World Bank. 2015. Women, Business and the Law 2016: Rijkers, Bob, Caroline Freund, and Antonio Nucifora. 2014. “All Getting to Equal. Washington DC: World Bank. ” Policy Research in the family: State capture in Tunisia. Working Paper Series 6810. Washington, DC: World Bank. Sabarwal, Shwetlana, and Katherine Terrell. 2008. “Does gender matter for firm performance? Evidence from Eastern Europe and Central Asia.” Policy Research Working Paper Series No. 4705, World Bank, Washington DC. 84 5. Competitiveness in the MENA region: Trade, innovation, and management practices Introduction advantages—competing on cost, quality, delivery, and flexibility.3 Competitiveness is much talked about, but com- plex to define. There is neither a shared definition On average, the MENA ES economies are middle- of competitiveness nor a consensus on how to income, though their performance in recent years measure it consistently across economies and has been disappointing. In the World Economic over time—unsurprisingly, as it is firms rather than Forum Global Competitiveness Report 2015-2016, economies that compete in the global market.1 At the highest ranked developing economy in the the level of a firm, competitiveness can be thought MENA ES region was Jordan, in 64th place (out of as the ability to sustain market position by sup- of 140 economies). Moreover, economies in the plying quality products on time—at competitive region have on average regressed by five places in prices2—and the ability to adapt quickly to changes the rankings since 2012-2013. The average value of in the external environment. It requires continuous the global competitiveness index in the MENA ES increases in productivity, shifting from comparative region was below that of their middle-income peer advantages, such as low cost labor, to competitive economies.4 Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 85 This chapter sheds light on the position of firms in the and benefit from exporting.8 In contrast, the presence of MENA ES economies in terms of labor productivity factors that affect entry costs for selected firms only—such and competitiveness.5 Perhaps surprisingly, the survey as subsidies, access to cheaper inputs, regulatory capture, results reveal that the labor productivity of firms in the or preferential access to foreign markets—may distort region compares favorably with that in economies with which firms benefit from exporting. comparable incomes.6 The proportion of firms with labor productivity above the median labor productivity in peer Likewise, the learning-by-exporting mechanism argues economies is higher than 50 percent in most MENA ES that exporters gain knowledge from exposure to foreign economies. Yet despite somewhat higher labor productiv- markets and practices, allowing them to grow and ity levels, firms in the MENA ES economies remain small: increase their efficiency. Evidence of the significance everywhere except Morocco, a majority of firms employ of this mechanism for the greater size and productivity fewer workers than the typical firm in similar economies. of exporters is mixed. Such forces may be increasingly The fact that these firms are unable (or unwilling) to scale important, however, with the presence of vertically inte- up their operations may indicate distortions and uncertain- grated production, where firms export as part of a “global ties underlying the competitiveness of these economies. value chain” (GVC) and may gain knowledge from parent companies, partners, and competitors, or through reacting A wide variety of factors have been suggested as driv- to the demands of foreign markets.9 Studies have indeed ers of productivity and competitiveness. This chapter confirmed the existence of similar size and productivity considers two broad areas: entrance and exposure to premia for importers: firms that import their inputs are on international markets through trade; and firms’ innovation average larger and more productive than firms that do not and management practices. These factors are interlinked. use foreign inputs. Innovation and management quality affect how inputs are employed and influence competitiveness. It is often The presence of barriers to trade, either through non-tariff only competitive firms that are able to be involved in a or tariff measures, is expected to reduce market competi- globalized system of production, allowing them to make tion and therefore average productivity in the market.10 the most of trading across borders. Under the right conditions, trade—whether exporting, importing, or both—presents an opportunity for firms to capitalize on and often improve their competitive position. Trade participation and But when those conditions are distorted and resources competitiveness are allocated inefficiently, many productive firms might Exposure to international trade has long been viewed as a not be able to access foreign markets and reap the scale driver of competition both within and across economies. and efficiency benefits from trade. An extensive and diverse literature has found the exis- tence of positive exporter size and productivity premia: Indeed, empirical work shows that the MENA region may firms that export are on average larger and more produc- be failing to realize such gains fully. Given its capacity and tive than their non-exporting competitors.7 The two main proximity to Europe, the region’s exports are estimated to mechanisms underlying this relationship are self-selection be roughly only a third of their potential level.11 The litera- into the export market and “learning-by-exporting.” ture also suggests that the profile of the region’s traders is characterized by a large number of firms engaging in low- The self-selection mechanism implies that firms must incur level trade, with a few solitary “superstars” facing few sunk costs to enter the export market, which only a select competitors.12 This section assesses whether these sug- few—presumably larger and more productive firms—find gestions are supported by the MENA ES data. It focuses advantageous to bear. Lowering these barriers to entry, on the size and labor productivity premia of exporting and for example, through decreased regulatory time and pro- importing firms, and on certain constraints faced by both cedures as well as transport costs, may ensure that this types of firms in the business environment. selection process works more efficiently: while the least productive firms, faced with expanded competition from home and abroad, will exit the market, more firms can enter 86 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Exporting firms in the region are numerous but small Figure 5.2: MENA ES exporters have lower size and One in four manufacturers in the MENA ES region directly productivity premia exports goods abroad, a proportion appreciably higher Panel A: Size premium of exporters vs. non-exporters than averages for lower-middle-income and upper-middle- income economies (14 and 18 percent respectively).13 500 This proportion varies considerably across the region. In 450 Lebanon, Tunisia, and the West Bank and Gaza, exporters 400 account for approximately 40 percent of all manufacturing 350 firms, but this share is as low as 8 percent in Egypt and 5 300 Percent percent in the Republic of Yemen. Although exporters are 250 200 numerous in the MENA ES economies, they tend to be 150 small firms. Nearly 80 percent of exporting manufacturers 100 in the region employ fewer than 100 full-time employees, 50 compared with 60 and 74 percent in lower-middle-income 0 and upper-middle-income economies respectively (figure MENA LAC ECA AFR EAP SAR Lower- Upper- ES middle- middle- 5.1). income income Figure 5.1: More of MENA ES exporters are small and Panel B: Labor productivity premium of exporters vs. medium-sized enterprises non-exporters 50 100 90 40 Percent of exporters that are SMEs 80 70 30 60 Percent 50 20 40 30 10 20 10 0 0 MENA AFR ECA LAC EAP SAR Lower- Upper- MENA AFR ECA SAR LAC EAP Lower- Upper- ES middle- middle- ES middle- middle- income income income income Source: Enterprise Surveys. Source: Enterprise Surveys. Note: The figure for MENA ES is based on coefficients from table A5.1, columns 1 and 2. Comparable figures are based on identically specified regressions for regions (income groups) as indicated. Bars with patterned fill indicate that coefficients are not significant at a 10 percent level. Exporter size and productivity premia are low compared with other regions in both lower-middle-income and upper-middle-income Reflecting the prevalence of small exporters, the so-called economies are on average 28 percent more productive exporter size premium (figure 5.2, panel A)—the average than non-exporters, while MENA ES and Sub-Saharan size differential between exporting and non-exporting Africa (AFR) are the only regions where on average firms—is considerably smaller in the MENA ES region (71 exporters are not significantly more productive than non- percent more permanent full-time employees on average) exporters (panel B). than it is in all other regions in the world or in comparable income groups (see table A5.1). This low size premium is mirrored by a low labor productivity premium. Exporters Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 87 Figure 5.3: The smaller size premium in the MENA ES region is dampened by big and small player exporters 3,400 Size premium of exporting groups 3,200 vs. non-exporters 1,200 1,000 Percent 800 600 400 200 0 -200 MENA ES AFR LAC ECA EAP SAR Lower- Upper- middle-income middle-income Superstar Big player Small player Source: Enterprise Surveys. Note: The figure for MENA ES is based on coefficients from table A5.1, column 3. Comparable figures are based on identically specified regressions for regions (income groups) as indicated. A few “superstar” exporters account for nearly all of learning-by-exporting, some firms might be willing to ac- the exporter size and productivity premia in the region; cept entering the export market at a short-term cost for a the numerous small player exporters experience no long-term gain. such premia The relative abundance of SME exporters in the MENA A striking picture emerges by differentiating exporting ES economies coupled with all but the top-tier, superstar firms by their export sales volume into “superstar” export- exporters, operating without an apparent ability or need to ers (the top 5 percent of firms), big player exporters (firms scale up their operations or improve their labor productiv- between the 50th and 94th percentile), and small player ity may be linked to the subsidization and the selective exporters (firms below the 50th percentile).14 Figure 5.3 lowering of export costs offered primarily to SMEs by shows the size premia for all three groups. In line with export promotion agencies.17 Such strategies that focus findings from a World Bank Group report,15 there is a wide on SME-based exporting may draw firms into foreign gap between the superstar exporters and other exporting markets through subsidized cost reductions, rather than firms (and compared with non-exporters). Furthermore, the underlying efficiency of those firms. Indeed, one re- the size premium for small player exporters in the MENA port argues that it is important to understand the reason ES region is very marginally negative. why these exporting SMEs remain comparatively small. If the reason is their lower productivity, policies focusing Looking at labor productivity, superstar exporters in the on helping them to export may be misguided. If they are MENA ES region generate revenues per worker that are prevented from growing by distortions, the focus should 4.5 times higher than non-exporters (and more than 3.5 be on policies that help eliminate such constraints.18 times as big player exporters). Small player exporters are actually less productive than firms that do not export at This relative abundance of low-volume exporters is also all (figure 5.4). In other words, these firms generate less consistent with potentially overvalued exchange rates, revenue per worker than their non-exporting peers. One which may dampen exports. Pegged exchange rates— reason for this negative productivity premium is that small such as those in Lebanon, Morocco, and Jordan—as well player exporters are significantly less capital-intensive as “crawl-like” ones in Egypt and Tunisia may limit export than other manufacturers, thus relying on more labor volume and hurt exporters’ international competitiveness relative to their revenues.16 Another possible explanation if they keep tradable goods more expensive abroad.19 If is that in expectation of increased productivity thanks to some exporting firms—particularly smaller ones—are 88 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 5.4: Small player exporters are less productive than non-exporters 1,400 Labor productivity premium of exporting groups vs. non-exporters 1,300 500 400 Percent 300 200 100 0 -100 MENA ES LAC AFR ECA SAR EAP Lower- Upper- middle-income middle-income Superstar Big player Small player Source: Enterprise Surveys. Note: The figure for MENA ES is based on coefficients from table A5.1, column 4. Comparable figures are based on identically specified regressions for regions (income groups) as indicated. disadvantaged in international markets by overvalued private sector in reviewing policies and identifying priori- exchange rates rather than their underlying productive ”20 ties have been largely absent. capacity, they may similarly lack incentives to scale up their operations. Table 5.1 provides some further context: superstar export- ers begin with remarkably more employees at start-up The much higher superstar exporter premia may also (on average 111) and begin exporting much earlier in their be explained by the presence of policies favoring large lifecycle, on average after only three years of operation.21 exporters and privileging relative capital intensity—for In other words, these top-tier firms start larger and are in a example, through lines of credit as well as land and position to enter international markets sooner, reinforcing energy subsidies—and in lieu of other subsidies such as evidence that it is a firm’s initial position in the market that those for R&D. One World Bank Group report addresses allows it to retain its size as a dominant exporter.22 this issue more directly, noting, “Discretion and lack of transparency in the allocation of subsidies or credit lines Superstar exporters in the MENA ES economies have on fuel the impression that less deserving firms are often the average seen a three-fold increase in their size over their beneficiaries. Successful exporters, large firms, or mul- lifecycle; the same factor for big players is less than 2.5 tinationals receive subsidies, protection, and privileges times. In contrast, small players grow from a starting size they do not need. Institutional processes that involve the of nearly 20 employees to just over 30, even after being Table 5.1: Superstar exporters start larger, while small player exporters are far less trade-intensive and take longer to begin exporting Age Employees Percentage of Exported directly When firm began Foreign firms in high- Exporter type (% sales) exporting As of 2012 At start-up As of 2012 ownership tech sectors Superstars 85 3 20 111 340 29 14 Big players 64 4 21 39 94 16 3 Small players 41 7 19 19 31 12 1 Source: Enterprise Surveys. Note: Indicators show values after controlling for industry and economy fixed effects. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 89 in operation for nearly 20 years, indicating a comparatively Manufacturers in the region are heavily import-reliant flat growth trajectory, despite being exporters. Moreover, Export activity is only one part of the story: manufacturing superstars are more likely to be foreign-owned than other firms frequently realize productivity and size gains from exporters: 29 percent of superstar exporters are at least importing their inputs as well. Increasingly, there has been 10 percent foreign-owned, compared with only 16 and 12 a focus on the role of these imports and firms’ position in percent for big and small player exporters respectively. international trade flows.23 Analysis of trade in the MENA The large initial size of superstar exporters could also region has noted that while trade levels are possibly below be explained by the strong presence of firms that use their potential, they are not particularly low; in fact, these technology intensively in this category: 14 percent of levels seem to be bolstered by imports to the MENA ES superstar firms are active in high-tech sectors. economies, which import goods and services at an aver- age of 57 percent of GDP . 24 When barriers to entry to exporting are low, they allow for the efficient entry of new and productive exporters into The MENA ES data show that manufacturers are particu- the market, as well as the exit of less competitive firms. larly reliant on imports, with 63 percent importing material Table 5.2 shows several proxy measures for the cost of inputs, trailing only manufacturers in the Latin America and firms to export. The table shows that, on average, the time the Caribbean (LAC) region (figure 5.5). Moreover, firms in and cost to export is lower in the MENA ES economies the MENA ES region use foreign inputs more intensively: than in peer economies. The exceptions are Lebanon, the 46 percent of manufacturers’ inputs are of foreign origin, West Bank and Gaza, and the Republic of Yemen, where above the average in peer economies, possibly indicating exporting is more timely and costly. Likewise, there are that firms are unable to find inputs of sufficient quality on often indirect costs to trading, for example, the quality of the domestic market. This pattern holds despite relatively domestic infrastructure. One proxy for this is the percent- high restrictions on imports (see below). This may be due age of products lost due to breakage or spoilage, which is to a combination of the lack of domestic alternatives as high in Djibouti, Lebanon, the West Bank and Gaza, and well as policies overvaluing currencies, for example, due the Republic of Yemen. Moreover, in large economies to pegged rates to hard currencies, such as the dollar peg such as Egypt, internal distance from borders can add in Lebanon or the peg to a euro-dollar basket in Morocco.25 further time and cost. Table 5.2: Costs of exporting in the MENA ES region are comparable to peer economies De facto time to clear Cost to export Percentage of products lost De jure time to export (days) customs (days) (USD per container) due to breakage/spoilage Djibouti 20 10 886 1.6 Egypt, Arab Rep. 12 7 625 0.8 Jordan 13 5 825 0.8 Lebanon 22 5 1,080 1.2 Morocco 11 3 577 1.0 Tunisia 13 3 773 0.6 West Bank and Gaza 23 3 1,685 4.1 Yemen, Rep. 29 11 995 2.4 MENA ES 18 6 931 1.6 Lower-middle-income 26 9 1,665 1.2 Upper-middle-income 21 7 1,445 0.8 Source: Enterprise Surveys, Doing Business database for 2013. 90 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 5.5: The import reliance of manufacturers 70 60 50 40 Percent 30 20 10 0 MENA ES LAC AFR ECA EAP SAR Lower- Upper- middle-income middle-income Percent of manufacturers importing inputs Percent of inputs of foreign origin Source: Enterprise Surveys. Importer size and productivity premia are high Figure 5.6: Manufacturers that import inputs are more compared with other regions productive than those that do not import Several works have examined the size and productivity 80 Labor productivity premium of importers vs. non-importers premia related to importing intermediate inputs.26 Indeed, MENA ES manufacturers that import inputs experience 70 significant and comparatively large premia over non- 60 importers in terms of both size and labor productivity. 50 Percent Firms that import their inputs are on average 55 percent 40 larger in terms of the number of employees, compared 30 with manufacturers that do not import (see table A5.2). 20 Only in the South Asia region (SAR) is this size premium even greater. In addition, importing firms in the MENA ES 10 region are nearly 75 percent more productive than non- 0 MENA ECA SAR AFR LAC EAP Lower- Upper- importers, a premium that is also considerably larger than ES middle- middle- income income in peer economies (figure 5.6). Source: Enterprise Surveys. Note: The figure for MENA ES is based on coefficients from table A5.2, column The importer size premium is driven by two-way 2. Comparable figures are based on identically specified regressions for regions (income groups) as indicated. traders, but the importer productivity premium is independent of export activity Manufacturing firms that directly import inputs may export percent foreign-owned, compared with less than 10 per- their final output as well. Comparing two-way traders with cent of exporters only, importers only, or non-traders. This firms that only export, only import, or do not trade, it is result holds even when superstar exporters are excluded clear that the size premium for manufacturing firms in the (table A5.3, column 2). MENA ES region is driven by two-way traders. As in other regions, importing inputs alone has little association with Importers have a labor productivity premium whether or larger size. Unsurprisingly, it is the larger firms that tend not they also export. Access to foreign inputs is strongly to be engaged in both importing and exporting, possibly associated with higher labor productivity—revenue per within GVCs, and almost a quarter of them are at least 10 worker (figure 5.7). For two-way traders, this association Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 91 Figure 5.7: Importing inputs drives productivity premium in the region 160 Labor productivity premium of traders vs. non-traders 140 120 100 Percent 80 60 40 20 0 MENA ES AFR EAP ECA LAC SAR Lower- Upper- middle-income middle-income Two-way trader Export only Import only SSource: Enterprise Surveys. Note: The figure for MENA ES is based on coefficients from table A5.3, column 3. Comparable figures are based on identically specified regressions for regions (income groups) as indicated. is driven largely by superstar exporters. Once these are tend to be larger and more productive than those that do excluded, the association with higher labor productiv- not, the region maintains substantial restrictions on trade ity is larger for firms that only import their inputs, again from abroad through higher tariffs and non-tariff restric- confirming that large and small player exporters in the tions.27 Tariff rates vary substantially within the region region seem to be unable to reap the efficiency gains that (table 5.3), as do the average usage of foreign inputs and emerge from exporting (table A5.3, column 4). the time to clear customs. Average tariff rates are highest in Djibouti and Tunisia, economies where manufacturers use foreign inputs at comparatively high rates (63 and 55 The business environment is not conducive to importing percent respectively), though in Tunisia the offshore sec- While manufacturers in the MENA ES economies are tor’s low-tariff access to inputs and well-documented tariff comparatively import-reliant, and while those that import evasion have played a role.28 Moreover, waiting times at Table 5.3: Restrictions on imports from abroad vary substantially Average manufacturing tariff rate De facto time (2008–12) Percent of inputs to clear imports Cost to that are of foreign De jure time to through customs import (USD per Intermediates Raw materials origin import (days) (days) container) Djibouti 3.6 3.0 63.3 18 5.2 911 Egypt, Arab Rep. 4.5 2.4 28.8 15 9.2 755 Jordan 1.9 7.6 42.3 15 5.3 1,335 Lebanon n.a. n.a. 51.6 30 9.7 1,365 Morocco 11.6 19.9 47.7 15 7.6 950 Tunisia 11.5 15.4 55.3 17 7.4 858 West Bank and Gaza n.a. n.a. 56.6 38 17.0 1,295 Yemen, Rep. 3.2 6.1 26.5 25 8.0 1,623 Lower-middle-income 4.0 5.8 37.0 33 13.1 669 Upper-middle-income 4.2 6.4 34.9 21 9.3 762 Source: Authors’ calculations based on UNCTAD Trade Analysis Information System (TRAINS); Enterprise Surveys, Doing Business database for 2013. Note: n.a.—not available. 92 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY customs for manufacturers importing inputs directly are links acquisition of knowledge, innovation, and labor roughly on par with peer economies.29 In addition, while productivity (see box 5.1 for more details).35 costs to import are also comparable, they are generally more expensive than those to export shown in table 5.2. Two in every five firms in the region innovate, but Given this combination of factors, it is somewhat product innovation is dominated by the adoption surprising that manufacturers in the MENA ES are so of existing technologies import-reliant. This pattern is consistent with a pattern Innovation is often associated with groundbreaking of “under-export/over-import” previously noted in the technology: the type that advances the global production region.30 Furthermore, this import reliance may translate frontier, typically in high-tech sectors. Innovation is also into higher input costs for the MENA ES region’s manu- a much broader concept, which includes the introduction facturing, eroding the gains from more sales per worker of new products and processes (technological innovation) (labor productivity). This can be a constraint on the growth as well as new organizational and marketing methods of efficient firms, and may result in low value-added or (non-technological innovation)—see box 5.2 for examples. what has been called “just-in-time production” rather than Moreover, most new products (as well as processes) are high value-added production.31 based on the adoption of existing technologies developed elsewhere, possibly with some adaptation to suit the needs of the local market. They are still considered to Increasing firm productivity be an innovation, though, as long as they are, at the very through innovation and better least, new to the firm itself. management Many firms in the MENA ES region compete in the Comparable Enterprise Survey data on innovation are international market but do not appear to achieve the available only for the Eastern Europe and Central Asia maximum benefits from doing so. This may reflect an in- (ECA) and MENA ES regions. These data show that in ability to improve their productivity continuously. One way both regions, firms engage in technological and non- to improve productivity is through innovation. A positive technological innovation at similar rates; on average, correlation between the introduction of a new or signifi- nearly 40 percent of firms engaged in at least one type cantly improved product (“product innovation”) and firms’ of innovation. In neither region are many of the new or performance has been established for European firms, improved products truly new to the global market (figure but evidence for developing economies has been mixed.32 5.8). The adoption (and adaptation) of existing products Similar studies do not exist for MENA economies. and processes is particularly important for emerging mar- kets and developing economies—including those in the Firms can also increase their productivity through other MENA ES region—where firms have considerable room means, such as making better use of excess capacity for improvement relative to the technological frontier. (provided there is any) or by improving management or business practices. Studies show that there is a strong R&D and other forms of knowledge acquisition are correlation between the quality of management practices dominated by high-tech sectors, but two-way trading and firms’ performance, and this also applies to developing seems to favor knowledge acquisition in lower-tech economies.33 Furthermore, lack of management skills has sectors as well been shown to be one explanation for the low productivity of state-owned firms or politically connected firms in the Firms can use a range of different approaches to acquir- absence of regulations that target their competitors.34 ing knowledge. They can create (“make”) it themselves through in-house spending on R&D.36 Firms can also To account for factors that may affect both firms’ pro- “buy” this knowledge by contracting R&D with other ductivity and the decision to innovate, this chapter uses companies and institutions or by purchasing or licensing a modified version of a well-known model devised by patented technologies, non-patented inventions, and Crépon, Duguet, and Mairesse (the “CDM model”) that know-how. Acquisition of knowledge does not always lead Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 93 Box 5.1: Estimating the impact of innovation on labor productivity The impact of innovation on productivity is estimated here treats innovation as an outcome of firms’ invest- using a modified version of a well-known three stage ment in the acquisition of knowledge, either created by model by Crépon, Duguet, and Mairesse (the “CDM the firm (R&D) or obtained from external sources. That model”).a The original model links productivity to firms’ is, it explains the decision to acquire knowledge; the de- innovation activities and, in turn, treats innovation as an cision to introduce a new product or process; and the outcome of firms’ investment in R&D. The model used firm’s labor productivity (figure B5.1). Figure B5.1: Version of the CDM model used in the chapter XCONTROL = Size; Age; Foreign ownership; State ownership; Trading status; Skilled workforce; Sector- and economy-specific effects XK = Manager’s education— XI = Use of foreign technology; XP = Location type; Fuel intensity; university Develop new ideas; Capital per worker; Capacity Formal training; Main utilization; Management market-local; ICT usage practices Acquisition of knowledge Innovation Productivity XKI = Access to finance; Manager’s sector experience XKP = Sole proprietorship Source: Authors’ representation of the model. Note: Based on Crépon and others (1998). ICT = information and communication technology. Variables in italics are available for manufacturing firms only. All stages are estimated simultaneously using an as- The second stage of the model determines the prob- ymptotic least squares estimator (ALS). The recursive ability of a firm implementing innovation, taking into ac- model accounts for the simultaneity and unobserved count its decision to acquire knowledge. The latent vari- variable problems arising from estimating the effect of able Knowledgei* derived from the first stage is used to the acquisition of knowledge and innovation activities, explain the impact that the acquisition of knowledge has which are likely to influence each other, on productivity.b on innovative activities: The model does not allow establishing causal relation- ships because the system does not permit the identifi- (2) Innovationi = 1[Innovationi* > 0] where cation of true instruments. Instead, the model imposes γ1 Knowledgei* + Xi,I γ2 + Xi,KIγ3 + Innovationi* =  exclusion restrictions grounded in economic theory and Xi,CONTROLγ4 + εi2 previous empirical work. In this equation, coefficient γ1 denotes the impact of The first stage estimates the innovation input equation: the acquisition of knowledge on the probability of a firm (1) Knowledgei = 1[Knowledgei* > 0] where introducing an innovation. Innovationi refers to the oc- currence of the various types of innovation. The prob- Knowledgei* = Xi,K β1 + Xi,KI β2 + Xi,KP β3 + Xi,CONTROL β4 + εi1 ability of observing such an innovation is explained by This represents the probability of the spending on the Xi,KI , Xi,CONTROL and Xi,I , which include variables listed in acquisition of knowledge (including R&D) by firm i, figure B5.1. where Knowledgei takes the value of 1 whenever the la- The final stage of the model relates the firm’s innova- tent value of spending on the acquisition of knowledge tive activities—or more precisely, the latent variable reported by the firm, Knowledgei*, is larger than zero. Xi,K, Xi,KI, Xi,KP and Xi,CONTROL include variables listed in fig- ure B5.1. (continued on next page) 94 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY (continued from previous page) Box 5.2: Types of firm-level innovationa that determines whether or not to Productivity-enhancing innovations are not limited to new prod- innovate—to labor productivity (mea- ucts. Significant improvements in technical specifications, com- sured as revenue per employee, con- ponents and materials, incorporated software, user-friendliness, verted into U.S. dollars, in log terms): and other functional characteristics of existing goods and services (3) Productivityi = ξInnovationi* + Xi,P δ1 count too. They can also entail new or significantly improved pro- + Xi,KP δ2 + Xi,CONTROLδ3 + ε i3 duction or delivery methods, such as the automation of work that used to be done manually or the introduction of new software to The coefficient ξ reflects the impact manage inventories. of innovation on labor productivity. In Moreover, innovations do not necessarily need to involve new addition to Xi,CONTROL and Xi,KP , the aug- technologies: they may also be in the form of organizational or mented production function includes marketing improvements. Examples of organizational innovation variables in vector Xi,P (see figure include introduction of a supply chain management system or B.5.1). For manufacturing firms, Xi,P decentralization of decision making, giving employees greater au- also includes their fuel intensity, capi- tonomy. Marketing innovations could be aimed at better address- tal per worker, and capacity utilization. ing customers’ needs, opening up new markets, or repositioning a firm’s product in the market. Examples include the introduction of a new flavor for a food product to target a new group of cus- a See Crépon and others (1998). tomers or the introduction of variable pricing based on demand. b The model also addresses issues relating to measurement errors in a Based on OECD, European Commission and Eurostat (2005). innovation surveys. Figure 5.8: Product innovation at the global technological frontier and the adoption of existing technologies 25 20 Percent of firms 15 10 5 0 Djibouti Egypt, Jordan Lebanon Morocco Tunisia West Bank Yemen, MENA ES ECA Arab Rep. and Gaza Rep. New to international market New to country, local market or firm Source: Enterprise Surveys. Note: Self-reported innovation and degree of novelty. Comparable ES data on innovation are available only for ECA and MENA ES. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 95 to successful innovation; conversely, innovation may not of employees does not seem to play an important role always require the acquisition of knowledge. in that process (tables A5.4 and A5.5).40 This may reflect both the general quality of education in the MENA ES The percentage of firms that engage in R&D is similar in region as well as a mismatch between the skills provided the MENA ES and ECA economies, but firms in the MENA by formal education and those demanded by the private ES region are less likely to engage in acquiring knowledge sector. Formal training helps workers learn the skills that more broadly. The MENA ES region compares favorably they need for their particular tasks as well as new produc- with the ECA region in higher-tech manufacturing sectors, tion techniques. such as pharmaceuticals, and medium-low-tech sectors, such as basic metals, but lags statistically significantly The formal level of education of managers, however, mat- behind in low-tech sectors, such as food products or ters for the decision to acquire knowledge: firms in which textiles (figure 5.9).37 Differences between different managers have a university degree are much more likely types of sectors are particularly large in Jordan, where to do so either through R&D or from external sources. almost a quarter of higher-tech firms engage in acquiring Such managers may be more familiar with the external knowledge, but less than 5 percent do so in other manu- knowledge already available, more open to investing in facturing sectors. This could be related to their exposure R&D, or more supportive of implementing various ways of to the international market: almost a quarter of higher-tech acquiring knowledge in their workplace (tables A5.4 and Jordanian firms are exporters, compared with less than 13 A5.5, column 1). percent of firms in other manufacturing sectors. Second, in the MENA ES region, access to knowledge In contrast, in Morocco and Tunisia, the gap between and information plays a crucial role in the ability of firms higher-tech manufacturing and lower-tech manufacturing to innovate (tables A5.4 and A5.5). Most firms do not and services is much lower. Both economies are charac- introduce innovation new to the technological frontier terized by greater integration into GVCs than their regional and often rely on existing knowledge of what their peers peers. In general, GVCs are considered to be crucial for are doing. The results show that two-way trader status knowledge transfer to local firms.38 Tunisia, for example, is positively and significantly associated with innovation has opted for an economic model oriented toward exports directly and indirectly, and it is a possible channel for the and industrialization supported by a pro-active policy of labor productivity premium shown above. Two-way trad- public investment in physical and human capital, and of ers are more likely to license foreign technology as well as attracting foreign direct investment (FDI). In Morocco, the introduce technological innovations. Similarly, manufactur- clothing industry, for example, has become a key supplier ers with at least 10 percent foreign ownership are more for fast fashion supply chains, as have automobile parts likely to acquire knowledge, introduce new products, and manufacturers and the aeronautical industry.39 implement technological innovations. There are several reasons why foreign ownership and two- Innovation benefits from firm-specific human capital: way trading—where, for example, firms are involved in access to knowledge through foreign ownership, two- GVCs—may be particularly important sources of informa- way trading, and ICT as well as access to finance tion for innovation. First, to satisfy a GVC’s product quality The analysis shows that there are a number of firm and process efficiency requirements, managers may need characteristics that are important determinants of firm to adapt their production methods or acquire technology innovation. First, a suitably skilled workforce (including via licensing arrangements. Second, to ensure smooth strong management skills) is a key prerequisite for suc- delivery to foreign clients, improved delivery methods cessful innovation. In the MENA ES region, firms that may be required. Third, by importing intermediate goods, provide formal training to their employees or give them firms may also import state-of-the-art technology that has time to develop new approaches and ideas are more likely not previously been available in the domestic market. to introduce new products, processes, organizational or This may require further training of workers, enhancing marketing methods, while the formal level of education their technical skills—which may, in turn, enable firms to introduce their own new products.41 96 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 5.9 :The proportion of firms that acquire knowledge A: Percent of firms, high- and medium-high-tech industries Djibouti Egypt, Arab Rep. Jordan Lebanon Morocco Tunisia West Bank and Gaza Yemen, Rep. MENA ES ECA 0 5 10 15 20 25 30 35 B: Percent of firms, medium-low-tech industries Djibouti Egypt, Arab Rep. Jordan Lebanon Morocco Tunisia West Bank and Gaza Yemen, Rep. MENA ES ECA 0 5 10 15 20 25 30 35 40 C: Percent of firms, low-tech industries Djibouti Egypt, Arab Rep. Jordan Lebanon Morocco Tunisia West Bank and Gaza Yemen, Rep. MENA ES EAC 0 5 10 15 20 Make only Make and buy Buy only Source: Enterprise Surveys. Note: Based on International Standard Industrial Classification (ISIC), Rev 3.1. Higher-tech manufacturing sectors include pharmaceuticals (24), machinery and equipment (29), electrical and optical equipment (30–33), and transport equipment (34–35, excluding 35.1). Low-tech manufacturing sectors include food products, beverages and tobacco (15– 16), textiles (17–18), leather (19), wood (20), paper, publishing and printing (21–22), and other manufacturing (36–37). Data represent cross-economy averages. Comparable ES data on innovation are available only for ECA and MENA ES. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 97 Furthermore, firms that use email to communicate with Figure 5.10: Association between innovation and labor their clients or suppliers are also significantly more likely productivity to introduce both technological and non-technological innovations. This may attest to the importance of both 30 Percent increase in labor productivity modern organizational practices and supporting ICT infra- 25 structure in facilitating innovation. 20 Finally, the results suggest that firms in the MENA ES 15 economies—as in many other economies—are much 10 more likely to introduce new products, processes, or both if they have access to finance in the form of a line of credit 5 or a loan. Introduction of non-technological innovation is 0 less affected by access to finance and foreign technologies Full private sector Manufacturing sector, 20+ employees (tables A5.4 and A5.5). Adapting external technologies, Product Process products, and processes to local circumstances can be Technological Non-technological costly, and firms may need sufficient financial resources to do so. While banks might not be willing or able to fund Source: Enterprise Surveys database and authors’ calculations. Note: This figure is based on coefficients from tables A5.6 and A5.7, columns innovative firms at the technological frontier, they may 1–4. For a detailed description, see box 5.1. fund firms that innovate by imitation, which is arguably less risky. They can also stimulate innovation by provid- result may be related to limited competition, as well as the ing firms with working capital or short-term loans, which presence of politically connected firms in several MENA can free up internal resources that the firms can use to ES economies and the regulations protecting them,45 finance innovation.42 which prevent innovative firms without political connec- tions from obtaining a larger market share and higher Firm innovation is associated with higher labor labor productivity. productivity, but less than in other developing economies Non-technological innovations, which are probably less risky and costly than technological innovations, are also Figure 5.10 shows that all types of innovation are associ- significantly associated with higher labor productivity (21 ated with higher labor productivity in both the full private percent higher than in the private sector overall). Given sector and in particular in manufacturing firms with more that this is comparable to or higher than productivity yields than 20 employees (tables A5.6 and A5.7). This correlation associated with technological innovation, it is perhaps is highest for product innovation, which is associated with surprising that only 29 percent of firms in the MENA ES labor productivity that is 28 percent higher than that of economies engage in either. This could be due to a lack of firms that do not introduce new or significantly improved information on new organizational and marketing meth- products. It is lower for process innovation, which is ods, skepticism about their effectiveness, or resistance to associated with labor productivity that is 22 percent change within organizations.46 higher compared with firms not undertaking this type of innovation. The correlations are up to 62 percent lower for manufacturing firms. The somewhat lower returns to High-tech firms benefit most from product innovation, process innovation may be due to the fact that firms in while low-tech firms benefit most from non- the MENA ES region are more likely to introduce new technological innovation processes than new products,43 and hence the benefits of There are also differences in returns to innovation within engaging in process innovation are lower. manufacturing (figure 5.11). In sectors with high- and medium-tech intensity, introducing a new product is as- These returns are in line with those found for developed sociated with labor productivity levels that are almost 20 economies, but lower than those observed in developing percent higher compared with firms that did not introduce economies, especially for the manufacturing sector.44 This 98 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Figure 5.11: Association between innovation and labor Figure 5.12: Distribution of the quality of management productivity by technological intensity practices compared with income-group median 25 Upper-middle-income economies Percent increase in labor productivity Tunisia Tunisia 64% 20 Lebanon Lebanon 52% Jordan Jordan 46% 15 Lower-middle-income economies 10 West Bank and Gaza West Bank and Gaza 57% Egypt, Arab Rep. Egypt, Arab Rep. 56% 5 Yemen, Rep. Yemen, Rep. 46% 0 Morocco Morocco 43% High- and Low-tech medium-tech Percent of firms above income group median Product Process Percent of firms below income group median Technological Non-technological Income group median Source: Enterprise Surveys. Source: Enterprise Surveys. Note: This figure is based on coefficients from table A5.8, columns 1–4. For Note: Djibouti is excluded, as the quality of management practices is available a detailed description, refer to box 5.1. Bars with patterned fill indicate that for only 12 firms. Comparable data are available for a comparison income coefficients are not significant at the 10 percent level. group only in the ECA. For Egypt, Morocco, the West Bank and Gaza, and the Republic of Yemen, the comparison income group is the lower-middle-income group. For Jordan, Lebanon, and Tunisia, the comparison group is the upper- middle-income group. a new product (table A5.8). In manufacturing sectors with low-tech intensity, firms benefit more from introducing non-technological innovations; the latter are associated management practices relating to operations, monitoring, with 15 percent higher labor productivity levels.47 targets, and incentives. They range from dealing with machinery breakdowns to factors determining the remu- This variation in estimated returns to innovation can be neration of workers. On the basis of firms’ answers, the explained by differences in the probability of different quality of their management practices can be assessed types of innovations and the level of competitive pres- and given a rating (see box 5.3 for details). sures faced. In several MENA ES economies, more than one-fifth of low-tech firms are two-way traders and There are firms with good and bad management practices compete primarily in the international market.48 They face in all MENA ES economies (figure 5.12). The share of great pressure to deliver the required products quickly manufacturing firms with good management practices and efficiently. As a group, low-tech firms are less likely to in Tunisia, Lebanon, the West Bank and Gaza, and Egypt introduce new organizational or marketing methods, but is higher than in their peer economies.51 Jordan, the those that do so successfully may manage to capture a Republic of Yemen and Morocco, on the other hand, stand larger market share as a result, thereby increasing their out with a share of firms with bad management practices revenue per worker. Some innovations by firms in low- above their peer economies. With some exceptions, large tech manufacturing sectors may be due to European firms manufacturing firms are on average better managed than moving production to Tunisia and Morocco from China in their medium-sized counterparts. the period up to late 2014, as a result of rising wage costs and the increasing cost of fossil fuels during that period.49 The quality of management practices in the MENA ES economies is positively correlated with economic devel- opment (measured as GDP per capita, figure B5.3). It is Poorly managed firms benefit more from improving not significantly associated with firm-level labor productiv- their management practices than from innovation ity, either on its own or in combination with different types The MENA ES included a subset of questions on of innovation (table A5.7). This is in contrast with results management practices.50 These questions look at core found elsewhere, including in the ECA region.52 Among Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 99 Box 5.3: Management practices in the MENA region The MENA ES includes a section on management prac- There is a positive correlation between the average qual- tices in the areas of operations, monitoring, targets, and ity of management practices and log per capita GDP (see incentives. The operations question focuses on how the figure B5.3). firm handles a process-related problem, such as machin- ery breaking down. The monitoring question covers the Figure B5.3: There is a positive correlation between the collection of information on production indicators. The average quality of management practices and log per questions on targets focus on the timescale for produc- capita GDP tion targets, as well as their difficulty and employees’ awareness of them. Lastly, the incentives questions cov- 0.4 Tunisia er criteria governing promotion, practices for addressing poor performance by employees, and the basis on which Lebanon 0.2 Quality of management practices the achievement of production targets are rewarded. These questions were answered by all manufacturing 0.0 Jordan firms with at least 20 employees. The median number Yemen, (z-score) Rep. West Bank of completed interviews with sufficiently high response -0.2 and Gaza rates was just below 115 per economy, with totals rang- ing from 12 in Djibouti to 1,130 in Egypt.a -0.4 Egypt, Arab Rep. The scores for individual management practices (in other words, for individual questions) were converted into z- -0.6 Morocco scores by normalizing each practice so that the mean was 0 and the standard deviation was 1. To avoid putting -0.8 6.5 7.0 7.5 8.0 8.5 9.0 too much emphasis on targets or incentives, unweight- GDP per capita (constant 2005 USD), log ed averages were first calculated using the z-scores of individual areas of the four management practices. An Source: Enterprise Surveys. Note: Djibouti is excluded, as the measure of quality of management practices unweighted average was then taken across the z-scores is available for only 12 firms. for the four practices. Lastly, a z-score of the measure obtained was calculated. This means that the average management score across all firms in all economies in the sample is equal to zero. The management practices a The questions on management practices came at the of individual firms deviating either left or right from zero, end of a long face-to-face interview. This resulted in an with those to the left denoting bad practices and those unusually large number of people responding “don’t know” to the right indicating good practices. or refusing to answer. poorly managed firms, however, those that are somewhat In economies with fewer energy subsidies, better better managed tend to have higher labor productivity, managed firms use energy resources more efficiently while the association with innovation is not significant. In The MENA ES data also show that energy intensity, as contrast, for well-managed firms, management practices measured by fuel intensity, is negatively correlated with are not correlated with higher labor productivity, but in- labor productivity (tables A5.6 and A5.7). Theoretically, novations are (table A5.9). These results suggest that better management practices may either decrease usage poorly managed firms might achieve higher returns from of energy through more efficient production techniques or improving management practices than from being innova- increase it through higher capital utilization. Empirical evi- tive. Well-managed firms, on the other hand, might ben- dence shows that in the United Kingdom, better-managed efit more from engaging in innovation than from further firms use energy more efficiently.53 Similar analysis ap- improving their management practices. plied to the MENA ES region does not reveal the same relationship (table A5.10, column 1). This may be due to a remarkable difference in the level of subsidization of energy consumption: the average of energy subsidies (the sum of subsidies for petroleum products, natural gas, and 100 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Policy conclusions Figure 5.13: Petroleum products, natural gas and electricity subsidies in the MENA ES region are much MENA ES economies generally perform worse on various higher than in the United Kingdom competitiveness rankings compared with their middle- 5 income peer economies in other regions, even though the labor productivity of private sector firms is similar in both 4 groups. Percent of GDP 3 Trade is not the issue per se: firms in the MENA ES re- gion are more likely to export, import, or both than their 2 counterparts elsewhere; but those firms are also more likely to be SMEs. The differences lie in the productivity 1 premium: superstar exporters have similar productivity margins as elsewhere, but the bulk of exporters lag be- 0 MENA ES ECA United Kingdom hind. In other words, many exporters may find them- Petroleum products Electricity selves constrained or unwilling to expand, or they have an Natural gas Coal incentive to continue exporting despite being inefficient. Source: Clements and others (2013) and authors’ calculations. The winners in terms of productivity gains, however, are importers, which is perhaps due to the access they get to foreign technology and supply chains. This is despite the coal) in seven MENA ES economies54 in 2011 constituted obstacles that importers face in terms of higher tariffs, 5 percent of GDP , compared with 0.6 percent of GDP in non-tariff restrictions on trade from abroad, and the time the United Kingdom (figure 5.13). it takes for imports to clear customs. In the less-subsidized group of MENA ES economies—all Trade, access to information, and access to knowledge but Egypt and the Republic of Yemen—higher-quality more broadly—through two-way trading, foreign owner- management practices are associated with a lower level ship, firm-specific human capital, and ICT—are also of fuel spending per dollar of total revenue (table A5.10, important determinants of innovation in the MENA ES column 3).55 The estimate suggests that improving the region. The percentage of firms that engage in any type management quality from the 25th to the 75th percentile of innovation is comparable with the ECA region, but is associated with a 32 percent decrease in firm’s fuel labor productivity gains from innovation are smaller than intensity. More subsidized MENA ES economies do not those observed in other developing economies. Only well- follow this pattern (table A5.10, column 2) and, therefore, managed firms see productivity gains from innovation; do not benefit in a similar way from improvements in poorly managed firms would benefit more from improving management practices. their management practices. These results provide evidence of an indirect relationship Taken together, these findings suggest several measures between management practices and labor productivity in that policy makers in the MENA ES economies should the MENA ES economies: better management practices implement to reduce the differences in productivity gains. are associated with lower energy intensity and lower energy intensity is associated with higher productivity. First, firms would benefit from greater openness to in- This is true only in economies with a relatively low level of ternational trade and in particular more effective customs energy subsidies. If anything, more subsidized economies and trade regulations, both when exporting and importing. do not benefit from better management practices and, as The aim should be reducing entry costs for all firms; giving a consequence, they lack one of the ways to improve their preference to certain groups of firms—including SMEs— productivity. may result in less efficient and dynamic firms entering the export market. Moreover, while trade costs in the MENA ES economies seem to be comparable with trade costs Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 101 elsewhere, additional factors such as internal transport university levels. Governments could encourage firms costs are important for well-functioning export sectors. to provide training to their employees through dedicated training programs or training centers. Moreover, there is Second, importing should not be viewed solely through a need for more intensive training programs, particularly the lens of trade deficits and foreign exchange reserves. aimed at improving the management of SMEs. Despite the obstacles that importers face in terms of higher tariffs, non-tariff restrictions on trade from abroad Finally, there is an issue that is not discussed directly in and time to clear customs, firms in the MENA ES region the chapter due to data availability, but is related to many are import-reliant. Imports allow companies to source of its findings. Restrictions on firm entry and exit as well component parts of a better quality or at a lower cost as restrictions that give undue advantage to incumbent than those available in the domestic market, as well as to firms, particularly state-owned or politically connected acquire knowledge about new products and processes. firms (such as privileged access to subsidized energy and Time- and cost-efficient access to high-quality inputs, state procurement contracts or state-supported non-tariff either domestic or foreign, can thus be a means to en- barriers to trade), should be removed. courage more high value-added production. There is now a wealth of evidence showing that such Third, FDI-specific restrictions that hinder foreign invest- restrictions suppress productivity, aggregate growth, and ment should be removed. Manufacturers with at least employment growth. There are several reasons for this. 10 percent foreign ownership are more likely to acquire Unconnected firms might shrink due to fewer profitable knowledge, introduce new products, and implement investment opportunities or stop growing to stay small technological innovations. Yet despite this, the World enough to operate under the radar of their connected Bank’s Investing Across Borders reports that relative to larger competitors; they might also be forced to exit the other regions, the MENA economies are fairly restrictive market. Furthermore, undue advantages for incumbent on foreign equity ownership in many sectors, with the firms might discourage new and potentially more produc- exception of Tunisia, and it takes twice as long to start a tive and innovative firms from entering. Such distortions foreign firm as it does to start a domestic firm. have further knock-on effects: they may provide incentives for less efficient firms to enter export markets and gain Fourth, the governments should facilitate improvements or retain their market share, and prevent some more ef- in the skills of the workforce. Better communication and ficient ones from exporting or growing. cooperation between the private sector and universities would be beneficial and should be encouraged, with adequate funding provided at secondary, vocational, and 102 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY that fall below the median in terms of export value, by Endnotes economy. Jaud and Freund (2015) define superstars as the top 1 percent; since that report works from administrative 1 See Altomonte and Békés (2016). data and not a sample, a more conservative definition is 2 See Altenburg and others (1998). used here to ensure sufficient coverage. Their approach 3 See Porter (2000) and UNCTAD (2005). necessarily includes all firms at the frontier and so the observed effect they find is higher than presented here. 4 See World Economic Forum Global Competitiveness Report 2015–2016. 15 Jaud and Freund (2015). 5 The labor productivity results discussed in the chapter 16 Even after taking relative capital intensity into account, this disappear when total factor productivity is used pattern remains. instead: while the coefficients mostly keep their signs, 17 Jaud and Freund (2015). Examples include Jordan the significance disappears. This could be due to the Enterprise Development Corporation (JEDCO), whose assumptions used in TFP estimation, relatively smaller export promotion program has a strong focus on SMEs, sample size (not all manufacturing firms reported the and the Investment Development Authority of Lebanon capital measures), or higher-than-optimal capital intensity, (IDAL) which focuses its export promotion on the resulting from energy subsidies. Further, it may take agro-industry (agricultural and agro-industrial products) and longer for trade and innovation to be reflected in total therefore largely aims at relatively small firms. factor productivity improvements than in labor productivity 18 Ibid, p. 51. improvements—it may take more time for firms to adjust 19 International Monetary Fund (2014). capital and other non-labor inputs. Data availability does not allow us to determine the actual cause with certainty. 20 World Bank (2009), p. 151. 6 The comparison income group includes either upper- 21 Significant at a 10 percent level. middle-income or lower-middle-income economies 22 The ES data do not include productivity levels at the time (according to the World Bank income classification, as the firm started operating. Moreover, it should be noted of 2012) for which Enterprise Survey data are available, that by structure, only incumbent firms are considered, excluding MENA economies. and so entry and exit effects are not considered. 7 The firm-level literature on the profile of exporters is 23 See Seker (2012); Amador and di Mauro (2015). expansive, following the path of early works by Bernard 24 See Behar and Freund (2011). Figures from WDI, Imports and Jensen (1995, 1999); Bernard and others (2003, as a percentage of GDP . For Yemen and Djibouti the most 2006, 2007). For recent surveys of the literature, see recent year available is used: 2006 and 2007 respectively. Tybout (2003); Wagner (2007 , 2012); and Greenaway and 25 Jaud and Freund (2015). Kneller (2007). It should be noted that productivity is most often based on revenue rather than quantity output. This 26 See Amiti and Konings (2007); Seker (2012); Amin and distinction is important as firms may have higher revenue- Islam (2014). based productivity not only based on their productive 27 Jaud and Freund (2015) directly attribute unrealized growth efficiency but also through commanding higher prices for to these policies, “Closing MENA markets to competition the goods they sell or lower prices for their inputs. See with high tariffs and restrictive non-tariff measures (NTMs) Foster and others (2008). has not helped domestic exporters grow. ” (p. XV). 8 See Melitz (2003) and Bernard and others (2007). 28 World Bank (2014). According to the ES data, 96 percent 9 Bernard and others (2007). these so-called offshore firms import inputs, compared with 70 percent of comparators. Offshore firms use an 10 Ibid. average of 75 percent foreign inputs, compared with 50 11 See Behar and Freund (2011). percent for other Tunisian firms in the ES. 12 Jaud and Freund (2015, p. 57) find that while there are 29 Note that West Bank and Gaza do not control their borders dominant superstar traders, there are few near-level and customs themselves. trading firms. As they succulently characterize this 30 Jaud and Freund (2015). situation, “…in MENA the largest exporter is alone at the top—Zidane without a team. ” 31 Ibid. They note: “In addition, even if individual firms are able to source high-quality inputs from abroad, transport 13 Throughout the chapter, exporters are defined as firms costs and the increasing prevalence of “just-in-time” exporting more than 10 percent of their sales directly. production imply that a lack of high-quality locally available 14 “Superstar exporters” here are defined as the top 5 inputs is likely to hinder the ability of even the most percent of firms by their export sales value. “Big player” talented firms to succeed. ” p. 35. exporters are those accounting for between the 50th 32 See Mohnen and Hall (2013) for an overview. and 94th percentile. “Small player” exporters are those Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 103 33 In a management field experiment looking at large Indian 44 See Mohnen and Hall (2013) for an overview. Raffo and textile firms, Bloom and others (2013a) find that improved others (2008) found that a rise in product innovation management practices resulted in a 17 percent increase increased labor productivity of manufacturing firms by in productivity in the first year through improvements in 7.8, 24.6 and 36.8 percent in France, Brazil and Mexico the quality of products, increased efficiency and reduced respectively. inventories. For micro and small enterprises, McKenzie 45 See, for example, Rijkers and others (2014), Diwan and and Woodruff (2015) showed that micro and small firms others (2013). with better business practices in marketing, stock-keeping, 46 See, for example, Atkin and others (2015). record-keeping and financial planning have higher labor productivity, survival rates and faster sales growth. 47 Significant at a 1 percent level. 34 See Brown and others (2006); Estrin and others (2009); 48 International market is the main market for 37 .1 percent Bloom and Van Reenen (2010); Bloom and others (2012, of firms in Tunisia, 34.3 percent of firms in West Bank 2013); McKenzie and Woodruff (2015); and Rijkers and and Gaza, 32 percent of firms in Morocco, and 21.1 others (2014). percent of firms in Lebanon. In the remaining economies, comparable figures are below 8 percent. 35 See Crépon and others (1998). 49 Examples include lingerie manufacturer La Perla moving 36 R&D is the creative work undertaken on a systematic production from China to Tunisia and Turkey and ready-to- basis to increase a firm’s stock of knowledge. wear group Etam moving production to Morocco, Tunisia 37 The differences are significant at 10 percent level. The and Turkey (see Wendlandt, 2012). shares of higher- and medium-low-tech firms in ECA and 50 See Bloom and others (2013b). MENA ES are similar: 20.1 and 22.7 percent respectively. The definition of manufacturing sectors according to 51 Comparable ES data on management practices are technological intensity can be found at http://www.oecd. available only for ECA and MENA ES. org/sti/ind/48350231.pdf 52 See EBRD (2014), Chapter 2, Bloom and Van Reenen 38 See Saliola and Zanfei (2009). (2010), Bloom and others (2012), Bloom and others (2013a), Bartz and others (2016). 39 See AfD, OECD and UNDP (2014). 53 See Bloom and others (2010). 40 Stone and Tarek Badawy (2011) find a similar result using a sample of seven MENA economies (Egypt, Lebanon, 54 Data on energy subsidies for West Bank and Gaza are not Libya, Morocco, Saudi Arabia, Syria and the Republic of available. Yemen). 55 The share of energy subsidies in GDP is below the 41 See EBRD (2014), Box 3.2. relevant regional average in less subsidized economies and above it in more subsidized economies. 42 Ibid, Chapter 4 and Bircan and De Haas (2015). 43 Differences are significant at 10 percent level. 104 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Appendix A5 Table A5.1: Exporter size and labor productivity premia (1) (2) (3) (4) Dependent variable Log (PFTE) Log (LP) Log (PFTE) Log (LP) Direct exporter only (at least 10 percent of 0.54*** 0.09 sales) (Y/N) (0.105) (0.139) Superstar exporters (top 5th percentile by 2.33*** 1.71*** export value) (Y/N) (0.350) (0.332) Big player exporters (50th to 94th percentile 0.96*** 0.71*** by export value) (Y/N) (0.128) (0.143) Small player exporters (below 50th -0.08 -0.60*** percentile by export value) (Y/N) (0.125) (0.138) At least 10 percent foreign ownership (Y/N) 0.43** 0.11 0.35** 0.07 (0.185) (0.164) (0.159) (0.139) Log (LP) 0.00 -0.07** (0.033) (0.032) Log (PFTE) 0.00 -0.10** (0.048) (0.049) Constant 2.87*** 9.64*** 3.55*** 9.95*** (0.351) (0.191) (0.345) (0.193) Observations 3,011 3,011 3,011 3,011 R-squared 0.26 0.227 0.329 0.289 Source: Enterprise Surveys. Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix). Linearized Taylor standard errors clustered on strata are indicated in parentheses. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per permanent full-time employee, in 2012 USD. Variables omitted from the table: economy and sector fixed effects. Column 2 corresponds to marginal effects as presented in figure 5.4. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Table A5.2: Importer size and labor productivity premia (1) (2) Dependent variable Log (PFTE) Log (LP) Import inputs (at least 10 0.44*** 0.55*** percent foreign origin) (Y/N) (0.122) (0.132) At least 10 percent foreign 0.50*** 0.02 ownership (Y/N) (0.182) (0.149) Log (LP) -0.01 (0.036) Log (PFTE) -0.01 (0.050) Constant 2.89*** 9.52*** (0.373) (0.195) Observations 2,842 2,842 R-squared 0.262 0.277 Source: Enterprise Surveys. Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix). Linearized Taylor standard errors clustered on strata are indicated in parentheses. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Variables omitted from the table: economy and sector fixed effects. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 105 Table A5.3: Size and labor productivity premia by trader type (1) (2) (3) (4) Dependent variable Log (PFTE) Log (PFTE) Log (LP) Log (LP) Two-way trading firm (Y/N) 0.86*** 0.63*** 0.55*** 0.33* (0.147) (0.136) (0.167) (0.185) Direct exporter only (at least 10 percent 0.29 0.52** 0.17 0.37 of sales) (Y/N) (0.229) (0.249) (0.266) (0.249) Import inputs only (at least 10 percent 0.35** 0.05 0.58*** 0.44*** foreign origin) (Y/N) (0.138) (0.150) (0.151) (0.167) At least 10 percent foreign ownership 0.40** 0.05 0.02 -0.07 (Y/N) (0.183) (0.136) (0.150) (0.148) Log (LP) -0.01 -0.03 (0.036) (0.045) Log (PFTE) -0.01 -0.07 (0.052) (0.093) Log (Age) 0.05 -0.05 (0.046) (0.060) Log (Number of employees at start-up) 0.57*** 0.08 (0.044) (0.105) Log (Capital per employee) 0.00 0.24*** (0.027) (0.043) Constant 2.93*** 1.71*** 9.50*** 7.74*** (0.369) (0.425) (0.232) (0.475) Observations 2,842 2,145 2,828 2,145 R-squared 0.286 0.57 0.275 0.372 Source: Enterprise Surveys. Note: Simple OLS using survey-weighted observations (using Stata’s svy prefix). Linearized Taylor standard errors clustered on strata are indicated in parentheses. Two-way trading firm is a firm that exports at least 10 percent of revenue and imports at least 10 percent of inputs. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Columns 2 and 4 exclude superstar exporters. Variables omitted from the table: economy and sector fixed effects. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. 106 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A5.4: CDM, 1st and 2nd stages, full private sector Stage 1 Stage 2: Innovation (1) (2) (3) (4) (5) Spending on knowledge acquisition Non-technological Dependent variable (Y/N) Product (Y/N) Process (Y/N) Technological (Y/N) (Y/N) Spending on knowledge acquisition 0.19* -0.13 -0.00 0.09 (Y/N) (0.110) (0.134) (0.106) (0.110) Log (Age) -0.18* 0.09* 0.03 0.07 0.06 (0.100) (0.052) (0.064) (0.050) (0.050) Log (PFTE) 0.52*** -0.16** 0.10 -0.03 0.04 (0.068) (0.070) (0.086) (0.068) (0.069) At least 10 percent foreign 0.47** 0.11 0.07 0.16 0.12 ownership (Y/N) (0.234) (0.142) (0.168) (0.138) (0.138) At least 25 percent state ownership 0.76 -0.44 -0.59 -0.54 -0.51 (Y/N) (0.957) (0.473) (0.603) (0.462) (0.437) Direct exporter (at least 10 percent 0.14 0.05 0.13 0.15 0.07 of sales) (Y/N) (0.178) (0.100) (0.116) (0.096) (0.097) Percent PFTE with university degree 0.01*** -0.01** -0.00 -0.00* -0.00 (0.003) (0.002) (0.003) (0.002) (0.002) Percent PFTE with secondary 0.00 -0.00*** -0.00 -0.00 -0.00 education only (0.003) (0.001) (0.002) (0.001) (0.001) Years of manager's experience in -0.00 0.01** 0.01 0.01 -0.01** the sector (0.007) (0.003) (0.004) (0.003) (0.003) Line of credit or loan from a 0.28 0.35*** 0.51*** 0.53*** 0.20** financial institution (Y/N) (0.183) (0.094) (0.114) (0.093) (0.093) Foreign technology license (Y/N) 0.39*** 0.60*** 0.64*** 0.21* (0.111) (0.115) (0.109) (0.112) Employees receive time to develop 1.15*** 1.51*** 1.39*** 1.60*** new ideas (Y/N) (0.085) (0.089) (0.083) (0.083) Employees receive formal training 0.67*** 0.38*** 0.56*** 0.71*** (Y/N) (0.094) (0.100) (0.091) (0.090) Main market: local (Y/N) -0.23*** -0.24*** -0.20*** -0.16* (0.083) (0.091) (0.077) (0.081) Email usage (Y/N) 0.60*** 0.42*** 0.43*** 0.67*** (0.096) (0.105) (0.088) (0.095) Sole proprietorship (Y/N) 0.03 (0.180) Manager has a university degree 0.83*** (Y/N) (0.185) Source: Enterprise Surveys. Note: This table reports regression coefficients for the first and second stage of the model described in box 5.1. The results are estimated using asymptotic least squares (ALS). Standard errors are reported in parentheses below the coefficient. PFTE = permanent full-time employees. Variables omitted from the table: Percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 107 Table A5.5: CDM, 1st and 2nd stages, manufacturing firms with 20 or more employees only Stage 1 Stage 2: Innovation (1) (2) (3) (4) (5) Spending on knowledge acquisition Non-technological Dependent variable (Y/N) Product (Y/N) Process (Y/N) Technological (Y/N) (Y/N) Spending on knowledge acquisition -0.06 -0.26 -0.30 0.18 (Y/N) (0.184) (0.207) (0.206) (0.182) Log (Age) 0.03 0.17* 0.08 0.11 0.05 (0.163) (0.086) (0.100) (0.098) (0.083) Log (PFTE) 0.32*** -0.14 0.05 -0.01 0.06 (0.119) (0.092) (0.105) (0.106) (0.089) At least 10 percent foreign 0.97** 0.49* 0.35 0.72** 0.04 ownership (Y/N) (0.391) (0.269) (0.306) (0.303) (0.263) At least 25 percent state ownership 0.80 -0.02 -1.46 -0.27 -0.77 (Y/N) (0.995) (0.625) (1.119) (0.712) (0.644) Direct exporter only (at least 10 0.65 0.24 0.33 0.41 -0.15 percent of sales) (Y/N) (0.694) (0.338) (0.398) (0.399) (0.322) Import inputs only (at least 10 1.96*** 0.18 0.71 0.67 -0.37 percent foreign origin) (Y/N) (0.516) (0.407) (0.467) (0.465) (0.398) Two-way trading firm (Y/N) 1.73*** 0.51 0.85* 1.07** -0.08 (0.579) (0.392) (0.454) (0.456) (0.380) Percent PFTE with university degree 0.01 -0.00 -0.01* -0.00 0.00 (0.007) (0.004) (0.005) (0.005) (0.004) Percent PFTE with secondary -0.00 -0.01*** -0.01*** -0.01** -0.00 education only (0.004) (0.003) (0.003) (0.003) (0.002) Years of manager's experience in 0.01 0.01 0.01 0.01* -0.00 the sector (Y/N) (0.012) (0.006) (0.007) (0.007) (0.006) Line of credit or loan from a 0.79*** 0.43** 0.70*** 0.78*** 0.13 financial institution (Y/N) (0.300) (0.213) (0.236) (0.238) (0.208) Foreign technology license (Y/N) 0.45*** 0.76*** 0.57*** 0.28 (0.168) (0.177) (0.170) (0.175) Employees receive time to develop 0.90*** 1.56*** 1.30*** 1.59*** new ideas (Y/N) (0.138) (0.144) (0.140) (0.141) Employees receive formal training 0.92*** 0.45*** 0.80*** 0.60*** (Y/N) (0.148) (0.157) (0.148) (0.149) Main market: local (Y/N) 0.02 0.15 0.04 -0.07 (0.152) (0.168) (0.147) (0.159) Email usage (Y/N) 0.46*** 0.28 0.22 0.39** (0.161) (0.183) (0.150) (0.166) Sole proprietorship (Y/N) 0.16 (0.336) Manager has a university degree 0.93*** (0.312) Source: Enterprise Surveys. Note: This table reports regression coefficients for the first and second stage of the model described in box 5.1. The results are estimated using asymptotic least squares (ALS) on a sample of manufacturing firms with at least 20 employees. Standard errors are reported in parentheses below the coefficient. Two-way trading firm is a firm that exports at least 10 percent of revenue and imports at least 10 percent of inputs. PFTE = permanent full-time employees. Variables omitted from the table: Percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. 108 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A5.6: CDM, 3rd stage, full private sector Stage 3 (1) (2) (3) (4) Dependent variable: Log (LP) Product Process Technological Non-technological Innovation (Y/N) 0.25*** 0.20*** 0.22*** 0.19*** (0.030) (0.026) (0.026) (0.024) Capital or main business city (Y/N) 0.19*** 0.20*** 0.19*** 0.19*** (0.044) (0.044) (0.044) (0.044) Log (Age) -0.04* -0.03 -0.04* -0.02 (0.023) (0.022) (0.022) (0.022) Log (PFTE) -0.00 -0.01 0.00 -0.02 (0.017) (0.017) (0.017) (0.018) At least 10 percent foreign ownership 0.05 0.09 0.06 0.07 (Y/N) (0.074) (0.071) (0.071) (0.071) At least 25 percent state ownership (Y/N) 0.44* 0.50** 0.48** 0.43* (0.250) (0.243) (0.242) (0.242) Direct exporter (at least 10 percent of 0.15*** 0.16*** 0.14*** 0.17*** sales) (Y/N) (0.055) (0.053) (0.054) (0.054) Percent PFTE with university degree 0.01*** 0.01*** 0.01*** 0.01*** (0.001) (0.001) (0.001) (0.001) Percent PFTE with secondary education 0.00*** 0.00*** 0.00*** 0.00*** only (0.001) (0.001) (0.001) (0.001) Sole proprietorship (Y/N) -0.33*** -0.37*** -0.34*** -0.35*** (0.041) (0.042) (0.041) (0.041) Source: Enterprise Surveys. Note: This table reports regression coefficients for the third stage of the model described in box 5.1. The results are estimated using asymptotic least squares (ALS). Standard errors are reported in parentheses below the coefficient. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Variables omitted from the table: percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 109 Table A5.7: CDM, 3rd stage, manufacturing firms with 20 or more employees only Stage 3 (1) (2) (3) (4) Dependent variable: Log (LP) Product Process Technological Non-technological Innovation (Y/N) 0.14*** 0.08** 0.10** 0.13*** (0.048) (0.039) (0.041) (0.041) Management practices -0.04 -0.04 -0.04 -0.04 (0.035) (0.035) (0.035) (0.035) Log (Capital per employee) 0.27*** 0.27*** 0.27*** 0.27*** (0.019) (0.019) (0.019) (0.019) Capacity utilization 0.00* 0.00* 0.00* 0.00* (0.001) (0.001) (0.001) (0.001) Capital or main business city (Y/N) 0.03 0.03 0.03 0.03 (0.088) (0.088) (0.088) (0.088) Log (Age) -0.04 -0.02 -0.02 -0.01 (0.041) (0.039) (0.039) (0.039) Log (PFTE) 0.01 -0.00 0.00 -0.03 (0.033) (0.033) (0.032) (0.035) At least 10 percent foreign ownership -0.01 0.04 0.00 0.01 (Y/N) (0.100) (0.097) (0.099) (0.098) At least 25 percent state ownership (Y/N) 0.15 0.28 0.19 0.24 (0.413) (0.421) (0.412) (0.415) Direct exporter only (at least 10 percent 0.11 0.14 0.13 0.15 of sales) (Y/N) (0.160) (0.157) (0.158) (0.158) Import inputs only (at least 10 percent 0.14 0.14 0.15* 0.15* foreign origin) (Y/N) (0.088) (0.087) (0.086) (0.087) Two-way trading firm (Y/N) 0.23** 0.27** 0.25** 0.26** (0.113) (0.109) (0.112) (0.109) Percent PFTE with university degree 0.01*** 0.01*** 0.01*** 0.01*** (0.002) (0.002) (0.002) (0.002) Percent PFTE with secondary education 0.00*** 0.00** 0.00** 0.00** only (0.001) (0.001) (0.001) (0.001) Fuel intensity (fuel cost as a fraction of -0.47*** -0.47*** -0.47*** -0.47*** sales) (0.087) (0.087) (0.087) (0.087) Sole proprietorship (Y/N) -0.09 -0.11 -0.10 -0.11 (0.078) (0.079) (0.078) (0.079) Source: Enterprise Surveys. Note: This table reports regression coefficients for the third stage of the model described in box 5.1 for the sample of manufacturing firms with at least 20 employees. The results are estimated using asymptotic least squares (ALS). Standard errors are reported in parentheses below the coefficient. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Two-way trading firm is a firm that exports at least 10 percent of revenue and imports at least 10 percent of inputs. Variables omitted from the table: percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. 110 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Table A5.8: CDM, 3rd stage, manufacturing firms with 20 or more employees only, by technology intensity Stage 3 (1) (2) (3) (4) Dependent variable: Log (LP) Product Process Technological Non-technological High- and medium-technology intensity Innovation (Y/N) 0.18** 0.06 0.08 0.08 (0.076) (0.062) (0.064) (0.054) Low-technology intensity Innovation (Y/N) 0.08 0.06 0.08 0.14*** (0.060) (0.046) (0.052) (0.055) Source: Enterprise Surveys. Note: This table reports regression coefficients for the third stage of the model described in box 5.1 for the sample of manufacturing firms with at least 20 employees by technology intensity. The results are estimated using asymptotic least squares (ALS). Standard errors are reported in parentheses below the coefficient. PFTE—permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Variables omitted from the table in addition to those shown in table A5.7: percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Table A5.9: CDM, 3rd stage, manufacturing firms with 20 or more employees only, by management quality above or below median Stage 3 (1) (2) (3) (4) Dependent variable: Log (LP) Product Process Technological Non-technological Firms with management quality above median Innovation (Y/N) 0.12* 0.11** 0.09* 0.13** (0.063) (0.052) (0.052) (0.051) Management practices -0.01 -0.01 -0.01 -0.02 (0.081) (0.081) (0.081) (0.081) Firms with management quality below median Innovation (Y/N) 0.13** 0.06 0.09 0.08 (0.065) (0.052) (0.060) (0.060) Management practices 0.14* 0.16** 0.15** 0.15** (0.074) (0.073) (0.073) (0.073) Source: Enterprise Surveys. Note: This table reports regression coefficients for the third stage of the model described in box 5.1 for the sample of manufacturing firms with at least 20 employees where the quality of management practice is above or below the MENA ES weighted median. The results are estimated using asymptotic least squares (ALS). Standard errors are reported in parentheses below the coefficient. PFTE = permanent full-time employees. LP = labor productivity. Labor productivity is measured as total revenue per PFTE, in 2012 USD. Variables omitted from the table in addition to those shown in table A5.7: percent PFTE with university degree (don’t know), percent PFTE with secondary education (don’t know), sector and economy fixed effects, and the intercept. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 111 Table A5.10: Management practices and fuel intensity (1) (2) (3) Dependent variable: Fuel intensity (fuel cost as a percent of revenue) All economies More subsidized Less subsidized Management practices -0.41 0.39 -0.87** (0.304) (0.439) (0.373) Log (Age) -0.13 -0.15 -0.53* (0.241) (0.432) (0.316) Log (Sales) -0.86*** -1.33*** -0.57 (0.313) (0.473) (0.402) Log (PFTE) 0.45 1.40* 0.11 (0.331) (0.725) (0.393) Log (Capital stock) 0.61*** 0.34 0.72*** (0.192) (0.273) (0.215) Percent PFTE with university degree -0.01 -0.02 0.00 (0.014) (0.018) (0.020) Constant 4.68 10.34** 2.82 (2.942) (4.302) (4.105) Observations 2,498 1,542 956 R-squared 0.217 0.204 0.276 Source: Enterprise Surveys. Note: This table reports regression coefficients for the sample of manufacturing firms with at least 20 employees using OLS regression on survey-weighted observations (using Stata’s svy prefix). Standard errors are reported in parentheses below the coefficient. PFTE = permanent full-time employees. Variables omitted from the table: percent PFTE with university degree (don’t know), sector and economy fixed effects. ***, ** and * denote statistical significance at the 1, 5 and 10 percent levels respectively. 112 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY References ———. 1999. “Exceptional Exporter Performance: Cause, Effect or Both?” Journal of International Economics AfD (African Development Bank), OECD (Organization for 47(1): 1–25. Economic Co-operation and Development), and UNDP (United Nations Development Programme). 2014. Bernard, Andrew B., J. Eaton, J. Bradford Jensen, and S. African Economic Outlook 2014: Global Value Chains Kortum. 2003. “Plants and Productivity in International and Africa’s Industrialization. Paris. ” American Economic Review 93 (4): 1268–90. Trade. Altenburg, Tilman, Wolfgang Hillebrand, and Jörg Meyer- Bernard, Andrew B., J. Bradford Jensen, and Peter K. Schott. Stamer. 1998. “Building Systemic Competitiveness: ” Journal of 2006. “Trade Costs, Firms and Productivity. Concept and Case Studies from Mexico, Brazil, Monetary Economics 53 (2006): 917–37 . Paraguay, Korea and Thailand.” German Development Bernard, Andrew B., J. Bradford Jensen, S. Redding, and Peter Institute, Reports and Working Papers 3/1998, German K. Schott. 2007 ” Journal of . “Firms in International Trade. Development Institute, Berlin. Economic Perspectives 21 (3): 105–30. Altomonte, Carlo, and Gabor Békés, eds. 2016. Measuring Bircan, Çagatay, and Ralph De Haas. 2015. “The Limits competitiveness in Europe: Resource allocation, of Lending: Banks and Technology Adoption across granularity and trade. Bruegel Blueprint Series, Volume Russia.” EBRD Working Paper No. 178, European Bank XVIV. http://bruegel.org/wp-content/uploads/2016/01/ for Reconstruction and Development, London. Blueprint-XXIV.pdf Bloom, Nicholas, and John Van Reenen. 2010. “Why Do Amador, João, and Filippo di Mauro, eds. 2015. The Age of Management Practices Differ across Firms and Global Value Chains: Maps and Policy Issues. VoxEU Countries?” Journal of Economic Perspectives 24 (1): eBook. http://www.voxeu.org/sites/default/files/GVCs- 203–24. ebook.pdf Bloom, Nicholas, Christos Genakos, Ralf Martin, and Raffaella Amin, Mohammad, and Asif Islam. 2015. “Use of Imported Sadun. 2010. “Modern Management: Good for the Inputs and the Cost of Importing: Evidence from Environment or Just Hot Air?” The Economic Journal Developing Countries.” Applied Economics Letters 22 120 (544): 551–72. (6): 488–92. Bloom, Nicholas, Helena Schweiger, and John Van Reenen. Amiti, M., and J. Konings. 2007. “Trade Liberalization, 2012. “The Land that Lean Manufacturing Forgot? Intermediate Inputs, and Productivity: Evidence from ” Management Practices in Transition Countries. Indonesia. ” American Economic Review 97 (5):1611–38. Economics of Transition 20 (4): 593–635. Arnold, Jens, Giuseppe Nicoletti, and Stefano Scarpetta. 2008. Bloom, Nicholas, Benn Eifert, A. Mahajan, David McKenzie, “Regulation, Allocative Efficiency and Productivity in and J. Roberts. 2013a. “Does Management Matter? OECD Countries: Industry and Firm-level Evidence. ” Evidence from India.” Quarterly Journal of Economics OECD Economics Department Working Paper 128(1), 1–51. No. 616, Organization for Economic Co-operation and Bloom, Nicholas, Erik Brynjolfsson, Lucia Foster, Ron Jarmin, Development, Paris. Itay Saporta-Eksten, and John Van Reenen. 2013b. Atkin, David, Azam Chaudhry, Shamyla Chaudry, Amit K. “Management in America. ” Center for Economic Khandelwal, and Eric Verhoogen. 2015. “Organizational Studies Paper No. CES-WP-13-01, U.S. Census Bureau, Barriers to Technology Adoption: Evidence from Washington, DC. Soccer-ball Producers in Pakistan.” NBER Working Paper Brown, J. David, John S. Earle, and Almos Telegdy. 2006. No. 21417 , National Bureau for Economic Research, “The productivity effects of privatization: Longitudinal Cambridge, MA. estimates from Hungary, Romania, Russia, and Bartz, Wiebke, Pierre Mohnen, and Helena Schweiger. 2016. Ukraine.” Journal of Political Economy 114(1): 61-99. “The Role of Innovation and Management Practices in Clements, Benedict, David Coady, Stefania Fabrizio, Sanjeev Determining Firm Productivity: Evidence from Transition Gupta, Trevor Alleyne, and Carlo Sdralevich, eds. 2013. Economies. ” EBRD Working Paper No. 188, European Energy Subsidy Reform: Lessons and Implications. Bank for Reconstruction and Development, London. Washington, DC: International Monetary Fund. Behar, Alberto, and Caroline Freund. 2011. “The Trade Crépon, Bruno, Emmanuel Duguet, and Jacques Mairesse. Performance of the Middle East and North Africa.” 1998. “Research, Innovation and Productivity: An Middle East and North Africa Working Paper Series No. ” Economics of Econometric Analysis at the Firm Level. 53, World Bank, Washington, DC. Innovation and New Technology 7 (2): 115–58. Bernard, Andrew B., and J. Bradford Jensen. 1995. “Exporters, Jobs, and Wages in U.S. Manufacturing: 1976–1987.” Brooking Papers on Economic Activity: Microeconomics: 67–119. Chapter 5: Competitiveness in the MENA region: Trade, innovation, and management practices 113 Diwan, Ishac, Philip Keefer, and Marc Schiffbauer. 2015. Rijkers, Bob, Caroline Freund, and Antonio Nucifora. 2014. “All “Pyramid Capitalism: Cronyism, Regulation, and ” Policy Research in the Family: State Capture in Tunisia. Firm Productivity in Egypt. ” CID Working Paper No. Working Paper 6810, World Bank, Washington, DC. 291, Center for International Development, Harvard Saliola, Federica, and Antonello Zanfei. 2009. “Multinational University. Firms, Global Value Chains and the Organization of EBRD (European Bank for Reconstruction and Development). Knowledge Transfer. ” Research Policy 38(2), 369–81. 2014. Transition Report 2014: Innovation in Transition. Schiffbauer, Marc, Abdoulaye Sy, Sahar Hussain, Hania London: EBRD. Sahnoun, and Philip Keefer. 2015. Jobs or Privileges: Estrin, Saul, Jan Hanousek, Evžen Kocenda, and Jan Svejnar. Unleashing the Employment Potential of the Middle 2009. “The effects of privatization and ownership in East and North Africa. World Bank, Washington, DC. transition economies.” Journal of Economic Literature Seker, Murat. 2012. “Importing, Exporting, and Innovation 47(3): 699-728. in Developing Countries.” Review of International Foster, Lucia, John C. Haltiwanger, and Chad Syverson. 2008. Economics 20 (2): 299–314. “Reallocation, Firm Turnover and Efficiency: Selection Stone, Andrew, and Lina Tarek Badawy. 2011. SME Innovators on Productivity or Profitability?” American Economic and Gazelles in MENA–Educate, Train, Certify, Compete! Review 98 (1): 394–425. MENA Knowledge and Learning Quick Notes Series No. Greenaway, David, and Richard Kneller. 2007. “Firm 64502. Washington, DC: World Bank. ” heterogeneity, exporting and foreign direct investment. Tybout, J. 2003. “Plant and Firm Level Evidence on ‘New’ Economic Journal 117(517): F134-F161. Trade Theories.” In Handbook of International Trade, eds. Hsieh, Chang-Tai, and Peter J. Klenow. 2009. “Misallocation E. K. Choi and J. Harrigan, 388–435. Oxford: Blackwell. and Manufacturing TFP in China and India.” Quarterly UNCTAD (United Nations Conference on Trade and Journal of Economics 124 (4): 1403–48. Development). 2005. Improving the Competitiveness International Monetary Fund. 2014. Toward New Horizons: of SMEs through Enhancing Productive Capacity. Arab Economic Transformation Amid Political Transitions. Proceedings of Four Expert Meetings. New York and IMF , Washington, DC. Geneva: United Nations. http://unctad.org/en/Docs/ Jaud, Mélise, and Caroline Freund. 2015. Champions Wanted. iteteb20051_en.pdf Promoting Exports in the Middle East and North Africa. UIS (UNESCO Institute for Statistics). 2015. “Summary Report Directions in Development. Washington, DC: World of the 2013 UIS Innovation Data Collection.” Information Bank. Paper No. 24, UNESCO Institute for Statistics. McKenzie, David, and Christopher M. Woodruff. 2015. Wagner, Joachim. 2007. “Exports and Productivity–A Survey of “Business Practices in Small Firms in Developing the Evidence from Firm-level Data.” The World Economy Countries. ” Policy Research Working Paper 7405, World 30 (1): 60–82. Bank, Washington, DC. ———. 2012. “International Trade and Firm Performance: Melitz, Marc J. 2003. “The Impact of Trade on Intra-Industry A Survey of Empirical Studies since 2006. ” Review of Reallocations and Aggregate Industry Productivity.” World Economics 148 (2): 235–67 . Econometrica 71 (6): 1695–1725. Wendlandt, Astrid. 2012. “Out of Love with China, Fashion Mohnen, Pierre, and Bronwyn H. Hall. 2013. “Innovation and Factory Moves. ” Reuters, March 22. http://www. Productivity: An Update. ” Eurasian Business Review 3 reuters.com/article/2012/03/22/us-china-fashion- (1): 47–65. idUSBRE82L0SC20120322 (accessed October 18, OECD (Organization for Economic Co-operation and 2015). Development), European Commission and Eurostat. World Bank. 2009. From Privilege to Competition: Unlocking 2005. Oslo Manual: Guidelines for Collecting and Private-Led Growth in the Middle East and North Africa. Interpreting Innovation Data. Third edition. OECD, World Bank, Washington, DC. European Commission, and Eurostat. ———. 2014. The Unfinished Revolution: Bringing Opportunity, Porter, Michael E. 2000. The Current Competitiveness Good Jobs, and Greater Wealth to All Tunisians. World Index: Measuring the Microeconomic Foundations of Bank, Washington, DC. Prosperity. Global Competitiveness Report 2000. New World Economic Forum. 2013. The Global Competitiveness York/Oxford: Oxford University Press. Report 2013–2014: Full Data Edition. Geneva: World Raffo, Julio, Stephane Lhuillery, and Luis Miotti. 2008. Economic Forum. “Northern and southern innovativity: a comparison across European and Latin American countries. ” The European Journal of Development Research 20(2): 219-239. 114 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Enterprise Survey–Economy Fiches 116 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Djibouti Figure 1: Ranking of the top business environment obstacles for firms in Djibouti 50 Electricity Corruption Tax rates Inadequately educated workforce Informality Customs and trade regulations Transportation Access to finance Licenses and permits Labor regulations 40 Percent of firms 30 20 10 0 Djibouti MENA ES Firms in Djibouti are heavily dependent on generators Figure 2: Firms owning or Figure 3: Electrical for electricity sharing a generator outages in a typical month Nearly half of firms in Djibouti choose electricity as their top obstacle in the business environment (figure 1). Firms cope 80 20 with an unreliable electricity supply by using power from Percent of firms 60 15 generators, represented by the 69 percent of firms in Djibouti Number who own or share a generator, much higher than the MENA 40 10 ES average of 36 percent (figure 2). Probably due to the high prevalence of generators, firms report fewer power outages 20 5 in a typical month: on average just under two, compared with a MENA ES average of almost 15 per month (figure 3). In 0 0 Djibouti MENA ES Djibouti MENA ES addition to an unreliable supply of electricity in the business environment, firms face issues with corruption as well as tax rates: respectively, 13 percent and 12 percent of firms indicate that these are top obstacles. Within the MENA ES region, Djibouti has the largest percentage of firms reporting Figure 4: Firms not that they do not need a loan needing a loan Djibouti’s financial sector has grown dramatically since the early 2000s, and today it is quite robust when compared with its peers. The increase in the number of banks in operation, the introduction 80 of Islamic financial instruments, and the opening of accounts for small savers have increased bank 60 Percent of firms deposits. Almost 92 percent of firms in Djibouti have a checking or savings account, well above the MENA ES average of 80 percent. In terms of access to credit, about 12 percent of working 40 capital needs are financed by banks. This is higher than the other MENA ES lower-middle-income economies, with the exception of Morocco. Djibouti also stands out in that 75 percent of firms 20 indicate that they do not need a loan (figure 4). This is the highest percentage in the region. Indeed, only 2 percent of firms rank access to finance as their top business environment obstacle. 0 Djibouti MENA ES Enterprise Survey–Economy Fiches 117 The majority of jobs in Djibouti’s private sector are in services Figure 5: The proportion of Djibouti’s economy differs from its MENA ES peers, as the majority of the private sec- jobs in the services sector tor is composed of the services sector. With an economy dominated by its deep-water port, 82 percent of jobs in Djibouti’s formal private sector covered by the survey are in 100 transport and related services sector. This is much higher than the average of 40 per- 80 cent across all MENA ES economies (figure 5). With an estimated unemployment rate 60 Percent of over 50 percent, job creation remains a challenging national priority. Among MENA ES economies, Djibouti has the highest share of firms (14 percent) indicating that labor 40 regulations are a major or very severe obstacle to the operations of their establishment. 20 0 Djibouti MENA ES Compared with the MENA ES region, firms in Djibouti are more reliant on foreign Figure 6: The proportion inputs of inputs that are of foreign Manufacturing firms in Djibouti are relatively more reliant on inputs of foreign origin, which is a origin result of the country’s lack of natural resources and harsh climate. On average, 63 percent of 80 manufacturing inputs are of foreign origin, well above the average for all MENA ES economies (46 percent, figure 6). This is despite the fact that its import tariff rates are among the highest in 60 the region. In terms of innovation across all business sectors, almost a third of firms in Djibouti Percent introduce new processes, higher than elsewhere in the MENA ES region. The majority of process 40 innovations occur through upgrading existing machinery and equipment, as well as software. 20 0 Djibouti MENA ES Djibouti has the highest proportion of Figure 7: Firms with a woman top manager Figure 8: Permanent full- firms with women in top management time employees that are positions in the MENA ES region women When compared with the rest of the world, 15 30 the MENA ES region lags behind in terms of Percent of firms women’s participation in the workforce, firm 10 20 ownership and top management positions. Percent Within this group, Djibouti stands out in 5 terms of having a relatively large percentage 10 of firms with a woman top manager: 14 per- 0 Djibouti MENA ES cent (figure 7), which was much higher than Female top manager 0 the MENA ES region average of 5 percent. Djibouti MENA ES Majority female ownership Djibouti also has the highest percentage of firms with majority female ownership: 7 percent, which is almost twice the regional average (4 percent). The proportion of permanent full-time employees that are women is also higher than the MENA ES average (figure 8). The relatively strong participation of women in the local workforce and firm management may be partly the result of the preponderance of the services sector in Djibouti’s economy, since services firms are typically more open to women. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 118 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Arab Republic of Egypt Figure 1: Ranking of the top business environment obstacles for firms in Egypt 50 Political instability Access to finance Electricity Corruption Licenses and permits Informality Crime Tax rates Labor regulations Inadequately educated workforce 40 Percent of firms 30 20 10 0 Egypt, Arab Rep. MENA ES Political instability is the top obstacle reported by Egyptian firms Figure 2: Sales and Nearly half of Egyptian firms choose political instability as their top obstacle, which was higher employment growth than the MENA ES average (figure 1). The uncertain business environment that followed the 0 2011 uprising and developments in the summer of 2013 was reflected in firms’ economic performance: between 2009 and 2012, the typical firm in Egypt saw revenues decline by 6.4 -2 percent per year and employment by more than one percent per year (figure 2). Access to Percent -4 finance is named as the top obstacle by one in every ten firms—not surprising, given that fewer than 60 percent of firms have a checking or savings account and only 6 percent of them -6 have a bank loan or a line of credit. Electricity issues emerge in third place, linked to a major -8 deterioration in electricity supply reliability in 2012, the reference year for the survey. Although Real annual Annual named as the top obstacle by only 6 percent of firms, corruption is widespread: 17 percent of sales employment growth growth firms report being exposed to at least one bribe request. Access to finance remains a key issue for Egyptian Figure 3: Degree of Figure 4: Firms with firms connection to the banking a checking or savings Banks account for only 2 percent of firm finance in Egypt, sector account well below the MENA ES average of 12 percent. The low 100 80 prevalence of bank finance is mirrored by a high share of 80 Percent of firms disconnected firms—those that did not apply for a loan be- Percent of firms 60 60 cause they have sufficient capital (figure 3). The fact that 40 40 40 percent of formal private sector firms do not have a checking 20 or savings account (figure 4) and therefore do not use the 0 20 Egypt, MENA ES financial system even for payment services suggests that Arab Rep. 0 the disconnect is structural. Anecdotal evidence suggests Egypt, MENA Disconnected that Egyptians themselves characterize their economy as a Discouraged Arab Rep. ES cash economy. This is in line with the strong role typically Connected ascribed to Egypt’s informal economy—estimates from the Egyptian Center for Economic Studies suggest that it con- stitutes around 40 percent of GDP and 66 percent of total non-agricultural private sector employment. Enterprise Survey–Economy Fiches 119 Egyptian manufacturers have high capital intensity and the use of Figure 5: Median factor shares capital seems inefficient Egyptian firms have labor productivity levels on par with firms in lower-middle- 50 Percent of firms 40 income economies. Where they lag behind is in total factor productivity (TFP), 30 which measures the efficiency of use of not only labor, but also capital and 20 intermediate inputs. When comparing the median factor shares of the three 10 main inputs used by manufacturers—their labor, intermediate inputs, and capi- 0 tal costs—Egyptian manufacturers are more capital-intensive than the average Egypt, MENA Lower- Arab Rep. ES middle- manufacturer in MENA ES as well as in their peer economies (figure 5). Among income the MENA ES economies, only Tunisian manufacturers are more capital- Labor Capital intensive. This can partly be explained by the presence of energy subsidies, Intermediates which distort production structures by promoting energy- and capital-intensive industries. Compared with larger firms, SMEs in Egypt are less likely to provide Figure 6: Firms offering formal training training to their employees Egypt is suffering from a mismatch between labor supply and demand, par- 50 ticularly in the area of technical and vocational skills. Post-secondary vocational 40 Percent of firms education and training are often perceived as low status and low quality, without 30 systematic engagement of employers in developing the programs and curricula. 20 Moreover, only 5 percent of Egyptian firms offer formal training, far lower than the MENA ES average of 17 percent. The difference is driven primarily by the 10 low percentage of SMEs providing formal training for their employees—only 2 0 and 6 percent of them do so, compared with 12 and 23 percent in the MENA All Small Medium Large ES region on average respectively (figure 6). Lack of skilled workers affects Egypt, Arab Rep. MENA ES fast-growing firms in particular, and as such, has important implications for aggregate growth and productivity. Due to the large domestic market, fewer firms are Figure 7: Firms by trade Figure 8: Firms engaged engaged in international trade status in at least one type of Given the large size of its domestic market, it is not surprising innovation that Egypt has one of the highest proportions of non-trading firms in the MENA ES region. Almost half of all manufacturing 100 40 Percent of firms 80 firms do not engage in either export or import activities (fig- 60 Percent of firms 30 ure 7). Moreover, only a quarter of firms in Egypt are engaged 40 20 in at least one type of innovation, compared with more than 0 20 two-thirds in the MENA ES region (figure 8). This may be due Egypt, MENA Arab Rep. ES 10 to the fact that the Egyptian market is vast and underserved, which means that firms do not need to compete for custom- Non-trader 0 Import only Egypt, MENA ES ers and hence do not feel the pressure to innovate. Moreover, Export only Arab Rep. only 3 percent of firms engage in knowledge acquisition, Two-way traders either through R&D or other sources. Compared with other MENA ES economies, this proportion is particularly low in high- and medium-high-tech manufacturing sectors. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 120 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Jordan Figure 1: Ranking of the top business environment obstacles for firms in Jordan 35 Access to finance Tax rates Political instability Labor regulations Inadequately educated workforce Access to land Corruption Informality Customs and trade regulations Licenses and permits 30 25 Percent of firms 20 15 10 5 0 Jordan MENA ES Access to finance is the top obstacle reported by Jordanian firms Figure 2: Sales and Almost a third of all Jordanian firms report access to finance as the top obstacle to their opera- employment growth tions (figure 1), the highest proportion among the MENA ES economies. Cyclical factors might partly explain this result. In 2012, the reference period of the survey, Jordan experienced several 5 adverse shocks. Reductions in gas supply from Egypt forced Jordan to resort to more expensive 4 3 fuel imports, putting pressure on the current account and reserves as well as the budget. Public Percent 2 debt increased from 71 percent of GDP in 2011 to 82 percent in 2012, potentially crowding out 1 the private sector. These adverse shocks also decreased firms’ propensity to invest and hence 0 reduced their demand for credit. Tax rates are the top obstacle for nearly a quarter of all firms, -1 possibly linked to an increase in the time it takes to prepare, file, and pay taxes. Political instabil- -2 Real annual Annual ity is in third place. Jordan faces security challenges mostly as a result of spillovers of regional sales employment turmoil. These problems notwithstanding, firms in Jordan experienced a relatively small drop in growth growth sales and robust growth in employment between 2009 and 2012 (figure 2). Jordanian firms are among the most credit-constrained in the MENA ES region Figure 3: Degree of credit MENA ES data indicate that problems of access to finance seem to go beyond cyclical consid- constraint erations and their potential impact on demand for and the supply of credit. While Jordan has comparatively deep financial and banking sectors, with private sector credit to GDP accounting 100 for about 70 percent of GDP from a peak of around 90 percent of GDP in 2007 , bank finance 80 Percent of firms accounts for only 10 percent of SME financing in Jordan. The banking sector’s exposure to 60 40 the government and public sector entities increased since 2010. Data indicate that loans to 20 SMEs account for about 10 percent of total loans, which could explain the divergence between measures of financial depth and financial access. Only 64 percent of firms—second lowest after 0 Jordan MENA ES the Republic of Yemen—are not credit-constrained, compared with 73 percent in the MENA ES Fully credit-constrained region (figure 3). Moreover, more than a third of Jordanian firms report being discouraged from Partially credit-constrained applying for a loan due to terms and conditions. Jordan also ranks last in terms of the Doing Not credit-constrained Business measure for ease of getting credit (185 out of 185, tying with the Republic of Yemen). Enterprise Survey–Economy Fiches 121 Women’s employment in Jordan is below the MENA Figure 4: Permanent full- Figure 5: Firms offering ES average time employees that are formal training The proportion of women among the full-time permanent women employees in the MENA ES region is very low by international 20 20 standards, and Jordan compares relatively poorly with other economies in the region. Only 8 percent of the workforce in a 15 15 Percent of firms typical Jordanian firm is composed of women, compared with Percent an average of 17 percent for MENA ES economies (figure 4). 10 10 Jordan also stands out among the MENA ES economies as having the lowest percentage of firms that provide training 5 5 to their employees—only 3 percent of Jordanian firms do so, compared with the MENA ES average of 17 percent (figure 5). 0 0 Jordan MENA ES Jordan MENA ES Jordanian manufacturing firms are competitive by Figure 6: Firms by trading Figure 7: Days to clear regional standards imports through customs status At 68th place, Jordan was the highest ranked MENA ES econ- omy in the World Economic Forum Global Competitiveness 100 10 Percent of firms 80 Report 2013–2014. Jordan’s manufacturing firms are relatively 8 60 well integrated into international trade, with 26 percent of 40 6 them both importing and exporting, compared with the aver- 20 Days ages of 20 percent in the region (figure 6) and 13 percent 0 4 Jordan MENA ES in upper-middle-income economies. The firms benefit from Non-Trader 2 relatively low manufacturing tariff rates on both intermediates Import only and raw materials. In addition, the reported number of days Export only 0 Jordan MENA ES to clear imports through customs is also among the lowest in Two-way traders the MENA ES region (figure 7). Among the MENA ES economies, the proportion of Figure 8: Firms engaged in Figure 9: Firms engaged firms engaged in at least one type of innovation is the knowledge acquisition in at least one type of lowest in Jordan innovation About a fifth of Jordanian firms are engaged in at least one 25 60 type of innovation (the lowest proportion in the MENA ES Percent of firms Percent of firms 20 50 region) and less than 5 percent of them acquire knowledge 40 15 by engaging in R&D and purchasing or licensing patented 30 10 20 technologies, non-patented inventions, and know-how. There 5 10 are, however, large differences across sectors. In higher-tech 0 0 industries, almost a quarter of firms acquire knowledge Jordan MENA ES Jordan MENA ES (figure 8) and more than half introduce new products, pro- High- and medium-high-tech High- and medium-high-tech cesses, and organizational or marketing methods (figure 9), Medium-low-tech Medium-low-tech Low-tech Low-tech on par with the MENA ES average. In other sectors, less than 5 percent of firms acquire knowledge, and the proportion of firms engaged in at least one type of innovation also lags behind the MENA ES average. These discrepancies could be driven by differences in trade integration: among firms in high- and medium-high-tech industries, more than 60 percent are exporters and more than 90 percent import their inputs. In the medium-low and low-tech industries, roughly 40 percent of the firms are exporters and about half import their inputs. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 122 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Lebanon Figure 1: Ranking of the top business environment obstacle for firms in Lebanon 60 Political instability Electricity Corruption Access to finance Tax rates Customs and trade regulations Informality Access to land Courts Tax administration 50 Percent of firms 40 30 20 10 0 Lebanon MENA ES Political instability is the top obstacle reported by Lebanese firms Figure 2: Political instability and sales Lebanese firms perceive political instability as the most important obstacle growth (figure 1). This probably refers to negative spillovers from the conflict in Syria, as well as more generally to the country’s confessional governance and the 6 Sales growth (percent) consequent inertia in structural reforms and weakening of institutions. The 4 country has not had a President since May 2014, and Parliament has voted 2 twice to extend its own term. The four-year term scheduled to end in 2013 0 is now foreseen to end in 2017 . In this difficult political and economic envi- -2 ronment, the performance of firms has come under pressure. In a question -4 -6 that considers obstacles independently from each other, political instability is Major/ Lesser identified as a major or severe obstacle by 91 percent of firms in Lebanon. severe obstacle obstacle Those firms performed worse in terms of sales growth over the survey refer- Lebanon MENA ES ence period 2009 to 2012 than firms that identify political instability as a lesser obstacle (figure 2). Electricity remains a key issue for Lebanese firms Figure 3: Quality of electricity supply For 11 percent of Lebanese firms, electricity is the most important obstacle (figure 1). Political divisions have forestalled reform of the energy sector, pre- 100 venting much needed investments in generating capacity and transmission. 80 Number/percent Moreover, tariffs have not been adjusted since the 1990s, implying substantial 60 fiscal transfers to the state-owned Electricité du Liban (EdL). As a result, firms 40 suffer from frequent power outages. Firms experience on average 51 power outages per month, far exceeding the MENA ES average (figure 3). The poor 20 quality of electricity supply forces firms to rely on expensive electricity from 0 Power outages Percent of firms generators. Not surprisingly, they are much more prevalent in Lebanon—where per month owning a generator 85 percent of firms own or share one—than in the other MENA ES economies. Lebanon MENA ES Enterprise Survey–Economy Fiches 123 Bank finance plays an important role for financing working capital and fixed capital Figure 4: Firm finance Lebanon has one of the highest levels of financial depth among the MENA ES economies, coming from banks reflecting persistent, large-scale deposit inflows that result from its traditional role as a financial hub for the region and a large and loyal diaspora. Overall, financial intermedia- 25 tion seems to be working well in Lebanon. Banks account for 21 percent of firm financing, 20 Percent of finance exceeding the MENA ES average by a wide margin (figure 4). There is a mixed picture for the collateral framework. On the one hand, Lebanese banks are more willing to lend unse- 15 cured than banks in an average MENA ES economy; on the other hand, banks rarely lend 10 against movable collateral. Only 4 percent of loans are secured by machinery and equip- ment or receivables, compared with a MENA ES average of 14 percent. A reform of the 5 secured transactions framework could further improve access to finance for Lebanese firms. 0 Lebanon MENA ES Workforce skills do not seem to be a major constraint for Lebanese firms Figure 5: Firms providing Less than 1 percent of firms in Lebanon consider workforce skills as the most important formal training obstacle, while 15 percent see it as a serious impediment to operations. This relatively good outcome may reflect the fact that Lebanon has one of the highest tertiary school 30 enrollment ratios in the region. Moreover, it is one of the MENA ES economies with 25 the highest training intensity. About 27 percent of firms offer formal training, compared Percent of firms 20 with a MENA ES average of 17 percent (figure 5). Moreover, Lebanon has the second 15 highest share of firms with women’s ownership in the MENA ES region at 43 percent, outperformed only by Tunisia (50 percent). This compares with a regional average of 25 10 percent. When considering the percentage of firms with a woman top manager, Lebanon 5 (4 percent) lags well behind Tunisia (8 percent) and below the regional average (5 percent). 0 Lebanon MENA ES Lebanese firms are among those most likely to engage in at least one Figure 6: Different types of innovation type of innovation across the MENA ES economies Lebanon has the highest proportion of firms engaged in innovation in the Product MENA ES region, with half of them introducing at least one type of innovation. Lebanese firms are more likely to introduce new products than firms in any Process other MENA ES economy (figure 6). They also exceed the MENA ES average for Organization the proportion of firms engaged in marketing and organizational innovations. In terms of involvement in international markets, Lebanon’s firms are outperform- Marketing ing most economies in the region. Only 20 percent of manufacturing firms do 0 5 10 15 20 25 30 35 not engage in any trade activities, compared with 33 percent in the MENA ES Percent of firms region on average. Lebanon has a strikingly high share of domestically owned Lebanon MENA ES exporters (95 percent compared with a regional average of 85 percent). This could be explained by the traditionally very high political and security uncer- tainty in the country, which leads domestic firms to seek stable markets for their products and foreign investors to stay away. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 124 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Morocco Figure 1: Ranking of the top business environment obstacle for firms in Morocco 35 Corruption Inadequately educated workforce Informality Access to finance Tax rates Political instability Customs and trade regulations Courts Tax administration Transportation 30 25 Percent of firms 20 15 10 5 0 Morocco MENA ES Corruption is the top obstacle reported by Moroccan firms Figure 2: Bribery depth and incidence Morocco is one of the few economies in the MENA ES region where political instability does not rank highly as a top obstacle. Instead, Moroccan firms 40 perceive corruption as the most important impediment to the business environ- Percent of firms 30 ment (figure 1): 21 percent of firms identify corruption as the top obstacle, compared with the MENA ES average of only 8 percent. Indeed, Morocco 20 has one of the highest reported bribery depths in the MENA ES region, at 30 percent (compared with a MENA ES average of 21 percent). Bribery depth 10 reflects the percentage of transactions where a firm is asked or expected 0 to pay a bribe when soliciting public services, permits, or licenses. Bribery Bribery depth Bribery incidence incidence—the percentage of firms experiencing at least one bribe payment re- Morocco MENA ES quest—is, at 37 percent, above the MENA ES average of 24 percent. Morocco also compares poorly with other lower-middle-income economies, where the averages for bribery depth and incidence are 16 and 21 percent respectively. An inadequately educated workforce ranks second as top obstacle in Morocco, and practices of competitors in the informal sector emerge in third place. Indeed, 47 percent of firms in Morocco report that they are competing against unregistered or informal firms, which is significantly higher than the regional average of 16 percent and trailing only the Republic of Yemen. Morocco lacks an adequately educated workforce Figure 3: Firms offering Of surveyed firms in Morocco, 13 percent identify an inadequately educated workforce as the formal training top business obstacle. Morocco has one of the lowest tertiary school enrollments in the region, with only the Republic of Yemen and Djibouti performing worse. In Morocco, gross enrolment 30 at the tertiary level is only 16 percent of the total tertiary age population, which compares 25 Percent of firms poorly to 30 percent in the MENA ES region as a whole. Moreover, the quality of education lags 20 behind and often does not correspond to the business needs of the private sector. At the same 15 time, Morocco is one of the MENA ES economies where the intensity of training provided by firms is one of the highest, with 26 percent of firms offering formal training compared with a 10 regional average of 17 percent (figure 3). This formal training provision remains well below the 5 lower-middle-income average of 37 percent. 0 Morocco MENA ES Enterprise Survey–Economy Fiches 125 Financial intermediation in Morocco compares well Figure 4: Use of banks as Figure 5: Not credit- with other economies in the MENA ES region financing source constrained firms Morocco has one of the highest levels of financial depth 25 100 among MENA ES economies, despite being a lower-middle- income economy, and is one of only two economies in the 20 80 Percent of finance Percent of firms region that have fully functioning credit bureaus. Overall, fi- 15 60 nancial intermediation seems to be working well in Morocco. Twenty-one percent of working capital and investment is 10 40 financed through banks (figure 4), the highest proportion 5 20 among MENA ES economies and by far exceeding the lower- middle-income average of 12 percent. The high prevalence of 0 0 Morocco MENA ES Morocco MENA ES bank finance is mirrored by the highest share of not credit- constrained firms (those that either did not need a loan or whose loan was approved in full): 87 percent, compared with the average of 73 percent in the MENA ES region (figure 5). Moreover, a low share of firms are discouraged from applying for a loan due to unfavorable terms and conditions such as complex application procedures, unfavorable interest rates, high collateral requirements, or insufficient size of loan and maturity. In fact, the share of discouraged firms in Morocco is the lowest among all MENA ES economies, as only 10 percent indicate being discouraged from applying for a loan while this proportion ranges from 13 percent in Djibouti to 49 percent in the Republic of Yemen. Morocco also has one of the lowest collateral ratios (the ratio of the value of collateral to the value of the loan) in the MENA ES region at 166 percent. The higher regional average of 208 percent is driven by the very high collateral ratios of the Republic of Yemen (281 percent), Egypt (272 percent) and Tunisia (252 percent). Moroccan firms engage more frequently in marketing Figure 6: Foreign Figure 7: Firms engaged than in other types of innovation ownership of exporting in R&D (inhouse or Morocco has one of the highest shares of foreign-owned manufacturing firms contracted) manufacturing exporters in the MENA ES region (27 percent 30 12 in Morocco compared with 15 percent in the MENA ES region on average, figure 6). This can, at least partly, be explained by 25 10 Percent of firms Percent of firms the country’s political stability, its capacity to attract foreign 20 8 investors, and its proximity to Europe. In terms of innovative 15 6 activities, Moroccan firms engage most frequently, at 28 10 4 percent each, in process and marketing innovation, which is well ahead of the regional averages of 19 and 20 percent re- 5 2 spectively. Moreover, a higher proportion of firms in Morocco 0 0 Morocco MENA ES Morocco MENA ES report engaging in R&D or buying external knowledge (10 percent) than in the MENA ES region on average (7 percent) (figure 7). This could be explained by greater integration of Moroccan firms into GVCs than their regional peers (with the exception of Tunisia) as well as the higher share of foreign ownership. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 126 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Tunisia Figure 1: Ranking of the top business environment obstacle for firms in Tunisia 50 Political instability Informality Access to finance Inadequately educated workforce Corruption Customs and trade regulations Tax rates Tax administration Labor regulations Transportation 40 Percent of firms 30 20 10 0 Tunisia MENA ES Political instability is the top obstacle reported by Tunisian firms Figure 2: Political instability and sales Tunisian firms perceive political instability as the most important obstacle to growth business activity, with half of all firms identifying this issue as their top obstacle (figure 1). Many firms suffered from the uncertain business environment that 2 Sales growth (percent) followed the Jasmine revolution in 2011, notably the uncertainty about policy 0 directions. Economic performance was at a low and most firms saw their sales -2 contract significantly in this difficult environment. Considered independently -4 of the other obstacles, political instability is identified as a major or severe -6 obstacle by 60 percent of firms in Tunisia. Those firms saw their sales decline -8 most dramatically over the survey reference period 2009 to 2012, by 9 percent, -10 Major/severe Lesser obstacle compared with a decline in sales of 3 percent for the firms that identify politi- obstacle cal instability as a lesser obstacle (figure 2). Informality ranks second as top Tunisia MENA ES obstacle in Tunisia, where 45 percent of firms report competing against un- registered or informal firms. Access to finance emerges in third place, despite Tunisian firms relying more heavily on external financing than firms in any other MENA ES economy, with only 59 percent of working capital and investment financed through internal sources. Tunisian manufacturers have high capital intensity and the use of Figure 3: Median factor shares capital seems inefficient Manufacturing firms in Tunisia are significantly more capital-intensive than firms 50 in upper-middle-income economies on average (figure 3). When comparing 40 Percent 30 the median factor shares of three main inputs used by manufacturers, that 20 is, their labor, intermediate inputs, and capital costs, Tunisian manufacturers 10 stand out as the most capital-intensive in the MENA ES region. This can partly 0 be explained by the presence of energy subsidies, which distort production Tunisia MENA ES Upper- middle- structures by promoting energy and capital-intensive industries. Indeed, while income Tunisian manufacturers have labor productivity levels comparable to those Labor Capital of manufacturers in upper-middle-income economies, their TFP lags behind, Intermediates indicating that capital is used inefficiently. Enterprise Survey–Economy Fiches 127 Tunisian firms have a lower degree of financial Figure 4: Financial sector Figure 5: Collateral disconnect, but high collateralization of loans is disconnect characteristics hampering access to finance 60 300 Despite ranking third, access to finance is identified as top 50 obstacle by only 10 percent of Tunisian firms. This compares 200 Percent of firms Percent favorably with the averages of both the MENA ES region 40 (11 percent) as well as upper-middle-income economies 30 100 (16 percent). Tunisian firms report a relatively low degree of 20 disconnect of the private sector from financial markets, with 0 10 Collateral Collateral 37 percent of firms disconnected in Tunisia compared with 58 ratio incidence percent of firms in the MENA ES region on average (figure 4). 0 Tunisia MENA ES Tunisia Disconnected firms are those that did not apply for any loan MENA ES in the survey reference year and explicitly state that they did not need a loan thanks to sufficient capital. Tunisian financial institutions rely heavily on the use of collateral as guarantees for loans. Both the collateral ratio (the ratio of the value of collateral to the value of the loan) and the collateral incidence (the share of loans that are collateralized) are high, the former being above any other MENA ES upper-middle-income economy at about 252 percent, and the latter (at 87 percent) being above the MENA ES average of 83 percent (figure 5). These two measures of collateral requirements also compare poorly with upper-middle-income averages (190 percent for the collateral ratio and 75 percent for the collateral incidence). Tunisian firms are competitive by Figure 6: Firms by trade Figure 7: Different types of innovation regional standards status Tunisia has the highest proportion of two-way traders—firms that both export and import— 100 Percent of firms Product 80 in the MENA ES region, with 35 percent of 60 firms exporting 10 percent or more of their Process 40 sales directly and importing 10 percent or 20 Organization more of intermediate inputs (figure 6). This 0 Tunisia MENA ES can partly be explained by the importance of Marketing Non-trader the offshore industry in Tunisia, which com- Import only 0 5 10 15 20 25 30 prises fully export-oriented firms that benefit Export only Percent of firms from tax exemptions, duty free access to Two-way traders Tunisia MENA ES inputs and equipment, and streamlined cus- toms procedures. Given this special status, these firms tend to be well integrated into GVCs. Moreover, a higher percentage of Tunisian firms are engaged in innovation than in the MENA ES region on average (figure 7). The proportion of firms undertaking process innovation is particularly high at almost a quarter of all firms—this may be related to the knowledge transfer from their GVC partners. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 128 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY West Bank and Gaza Figure 1: Ranking of top business environment obstacles for firms in the West Bank and Gaza 35 Political instability Electricity Informality Tax rates Access to finance Customs and trade regulations Corruption Licenses and permits Access to land Transportation 30 25 Percent of firms 20 15 10 5 0 West Bank and Gaza MENA ES Political instability is the top obstacle reported by firms in the West Bank and Gaza Figure 2: Sales and Roughly one in three firms in the West Bank and Gaza report political instability as the top employment growth obstacle in the business environment, in line with the average in the MENA ES economies (figure 1). Electricity and practices of the informal sector are the second and third ranked top 8 obstacle. Despite persistent instability, firms in the West Bank and Gaza experienced robust 6 growth rates in the period 2009–2012, in terms of both sales revenues, which increased at Percent nearly 6 percent per year, and employment, with an annual growth rate of nearly 8 percent 4 (figure 2). Although not ranked as the top obstacle, corruption is considered a major or very 2 severe obstacle to their operations by half of all firms. In addition, over half of all firms consider both access to finance and electricity as major/very severe obstacles to their current opera- 0 Real annual Annual tions. The current and future economic outlook, however, is much more uncertain as overall sales employment growth growth donor aid and disbursements have decreased, Israeli-Palestinian peace talks remain stalled, and fiscal pressures on the Palestinian Authority continue to grow. Given continuing political instability in the West Bank and Gaza and the uncertain economic outlook, policies are needed to promote private sector growth. The unreliable provision of electricity is particularly acute in Gaza Figure 3: Losses due to Figure 4: Electrical electrical outages outages in a typical month Firms in the West Bank and Gaza report losses due to power outages of above 6 percent of annual sales, larger 8 30 than losses reported by any other MENA ES economy’s firms Percent of sales 6 (figure 3). The supply of electricity is particularly unreliable 20 Number in Gaza, where losses due to power outages average over 4 22 percent of annual sales and firms experience nearly 29 10 2 outages per month, compared with reported losses of just above 1 percent and almost two power outages each month 0 0 in the West Bank (figure 4). The blockade of the Gaza strip, West Bank MENA ES West Gaza and Gaza Bank political infighting, perpetual fuel shortages, a crumbling infrastructure, and perpetual conflict and insecurity all result in the very unreliable supply of electricity in Gaza. Enterprise Survey–Economy Fiches 129 Many firms in the West Bank and Gaza are disconnected from financial services Figure 5: Firms with a The majority of firms’ working capital is financed by internal funds and supplier credit. Banks bank loan or line of credit account for only 3 percent of working capital financing in the West Bank and Gaza, which is well below the MENA ES average of 10 percent. Almost three-quarters of firms did not apply for a 30 loan as they have sufficient capital and are thus classified as disconnected from the financial Percent of firms sector, the second highest share of firms in the MENA ES. The fact that almost 30 percent of 20 formal private sector firms do not have a checking or savings account and therefore do not use the financial system even for payment services suggests that the disconnect is structural. 10 Indeed, only 6 percent of firms indicate having a loan or line of credit (figure 5). Despite the low prevalence of business loans, the West Bank and Gaza does stand out in terms of client- 0 friendly collateral practices. The share of movable collateral, such as machinery and equipment West Bank MENA ES and Gaza or receivables, is the highest among the MENA ES economies. At the same time, the average collateral ratio is the second lowest among the MENA ES economies. Women’s participation in the private sector lags behind other MENA ES economies Figure 6: Permanent full-time workers that are The West Bank and Gaza has some of the lowest rates of women’s participation both in women the workforce and in firm ownership or management among the MENA ES economies. Of permanent full-time workers, only 6 percent are women, lower than the MENA ES average 20 of 17 percent (figure 6). In addition, only 13 percent of firms have women’s participation in 15 ownership and 1 percent of firms have a woman top manager; the comparable averages for Percent the MENA ES region are 25 percent and 6 percent. Commonly cited reasons for this lack 10 of women’s participation include a dearth of opportunities as well as social, cultural, and 5 institutional norms. Due to persistent conflict and instability, additional concerns of personal safety and mobility restrictions further inhibit women’s participation in the formal private sector. 0 West Bank MENA ES and Gaza Firms in the West Bank and Gaza spend less on Figure 7: Degree of foreign engagement in Figure 8: Different types of knowledge R&D manufacturing acquisition In the West Bank and Gaza, 60 12 exporters account for ap- Percent of firms 9 proximately 40 percent of all 40 Percent 6 manufacturers and more than half of those firms’ inputs are 20 3 of foreign origin (figure 7). 0 0 Knowledge Engaged in R&D Nonetheless, importers face Direct exporters Inputs of foreign acquisition (either in-house by far the longest customs (% of firms) origin (% of inputs) (either R&D or external) or contracted) waiting times in the MENA West Bank and Gaza MENA ES West Bank and Gaza MENA ES ES region: 17 days. In ad- dition, compared with the MENA ES region as a whole, a slightly lower percentage of firms in the West Bank and Gaza are spending on R&D or the acquisition of external knowledge (figure 8). Almost a third of higher technology manufacturing firms do so, on par with Tunisia and Djibouti. The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 130 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Republic of Yemen Figure 1: Ranking of top business environment obstacles for firms in the Republic of Yemen 50 Political instability Electricity Corruption Crime Access to finance Access to land Labor regulations Informality Courts Inadequately educated workforce 40 Percent of firms 30 20 10 0 Yemen, Rep. MENA ES Political instability is the top obstacle reported by firms in the Republic of Yemen Figure 2: Sales and The ES fieldwork took place between March 2013 and July 2014, during a period of instability employment growth in the Republic of Yemen, which deteriorated into civil war in early 2015. Unsurprisingly, nearly half of all firms identify political instability as their top obstacle in the business environment 0 (figure 1). Nearly a quarter of firms indicate electricity as their top obstacle. Although not ranked -3 as the top obstacle, corruption is considered a major or very severe obstacle by 97 percent of Percent firms; among all economies with ES data, this is the highest percentage. In addition, over 60 -6 percent of all firms consider crime as a major/very severe obstacle to their current operations; -9 and 17 percent of firms experience losses due to theft and vandalism, the highest percentage among MENA ES economies. Not surprisingly, following this deterioration of the business en- -12 Real annual Annual vironment, private sector activity over the period contracted. A typical firm, between 2009 and sales employment 2012, saw sales revenues strongly decline by nearly 11 percent per year and an employment growth growth contraction of 5 percent per year (figure 2). Electricity remains a key issue for firms in the Republic of Yemen Figure 3: Electrical outages and losses Figure 4: Firms owning or sharing a generator After political instability, electricity is the second most-often cited top obstacle to 40 80 firms in the Republic of Yemen. Private sector 30 Percent of firms firms experience nearly 40 power outages in 60 Percent 20 a typical month and lose over 16 percent of 40 their annual sales as a result of these power 10 outages (figure 3). Closely linked to this, 0 20 Number of Losses due to the private sector reports heavy reliance on electrical outages electrical outages 0 private generators: eight in 10 firms in the in a typical month (% of annual sales) Yemen, MENA Republic of Yemen own or share a genera- Rep. ES Yemen, Rep. MENA ES tor (figure 4), and overall, 39 percent of the private sector’s power provision comes from these generators. Enterprise Survey–Economy Fiches 131 Firms in the Republic of Yemen remain largely Figure 5: Degree of credit Figure 6: Firms with disconnected from the financial sector constraint a checking or savings The Republic of Yemen has the highest share of credit- account constrained firms—those that were rejected (or partially 100 100 approved) on loan applications and/or were discouraged from Percent of firms 80 applying due to unfavorable terms and conditions—among 80 Percent of firms 60 MENA ES economies (figure 5). This is driven by a high 40 60 share of firms that are discouraged from applying for loans. 20 40 Moreover, only 1 percent of financing is sourced from banks, 0 Yemen, MENA the lowest proportion among all MENA ES economies. The Rep. ES 20 fact that over 50 percent of formal private sector firms do not Not credit-constrained 0 have a checking or savings account (figure 6) and therefore Partially credit-constrained Yemen, MENA Rep. ES do not use the financial system even for payment services Fully credit-constrained suggests that the disconnect is structural. Republic of Yemen manufacturers are the least Figure 7: Degree of Figure 8: Days to clear integrated into global markets foreign engagement in direct exports through Well behind the MENA ES average, only 37 percent of the manufacturing customs Republic of Yemen manufacturers import at least a tenth of 80 15 their material inputs or supplies from abroad (figure 7). In Percent of firms 60 contrast, this rate is on average over 60 percent in the MENA ES region. Manufacturers are even less integrated on the 40 10 Number exporting side. Only 5 percent of the economy’s manufactur- 20 ers export at least 10 percent of their sales abroad, a fifth 0 5 Importing Directly of the MENA ES average. Not surprisingly, the Republic of inputs and exporting Yemen has the lowest proportion of two-way trading manu- supplies 0 facturing firms (those that import and export), indicating that Yemen, Rep. Yemen, MENA Rep. ES this sector is quite removed from GVCs. In addition, Republic MENA ES of Yemen firms face longer waiting times to clear customs when directly exporting, compared with firms across the MENA ES region (figure 8). Innovation rates in the Republic of Yemen are comparable with Figure 9: Firms engaged in innovation MENA ES averages More than 40 percent of firms in the Republic of Yemen engage in at least one At least one type type of innovation (figure 9). These are introductions of new or significantly im- of innovation proved products or processes (technological innovations) or new or significantly improved organizational or marketing methods (non-technological innovations). Technological Most of the innovations are new to the firm rather than new to the Republic innovation of Yemen or to international markets. In the Republic of Yemen, firms primarily Non-technological introduce new marketing and organizational methods rather than new products innovation and processes; but firms also report technical innovation at rates slightly above 0 15 30 45 the MENA ES average. Yemen, Rep. MENA ES The Economy Fiches summarize the economy-specific findings of the report “What’s Holding Back the Private Sector in MENA?” Note that annualized sales and employment growth statistics are calculated using the reference years 2009 and 2012; these reference years are used due to when the Enterprise Survey was administered. The findings, interpretations, and conclusions expressed in this fiche are entirely those of the authors. They do not necessarily represent the views of the European Bank for Reconstruction and Development/European Investment Bank/World Bank and its affiliated organizations, or those of their Executive Directors or the governments they represent. 132 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Glossary of terms and acronyms AFR Sub-Saharan Africa. For a full list see pages 135–137 Big player exporters Firms between the 50th and 94th percentile by their export sales volume Business environment The various domains that affect the day-to-day experiences of firms. Examples include accessing finance, meeting regulatory requirements, infrastructure, corruption, etc. Collateral incidence The share of outstanding loans that are collateralized Collateral ratio The average ratio of the value of the collateral to the value of the loan Collateral ratio index The inverse of the collateral ratio, calculated from bank-specific information, reflecting the prevalent requirements applied by the bank to its client. It is presented as a measure reflecting the environment in the area where the bank is located, by providing a weighted average based on the relevance of the branches of banks located in a circle with radius of 10km centered on the specific firm. Competitiveness At the firm level, competitiveness can be thought of as the ability to sustain market position by supplying quality products on time—at competitive prices—and the ability to adapt quickly to changes in the external environment. It requires continuous increases in productivity, by shifting from comparative advantages, such as low cost labor, to competitive advantages—competing on cost and quality, delivery, and flexibility. Connected firms Firms that applied for a loan, regardless of whether their application was approved or rejected Credit-constrained firms Firms that had a loan application rejected or were discouraged from applying in the first place. They can be fully or partially credit-constrained. DEC Development Economics Vice Presidency, the research arm of the World Bank Group Disconnected firms Firms that did not apply for any loan, as they had sufficient capital Discouraged firms Firms that did not apply for any loan, due to terms and conditions EAP East Asia Pacific. For a full list see pages 135–137 EBRD European Bank for Reconstruction and Development ECA Eastern Europe and Central Asia. For a full list see pages 135–137 EIB European Investment Bank Enterprise Survey A survey that asks firms in the formal private sector questions about the business environment and their economic inputs and output ES Enterprise Survey Exporter productivity Average labor productivity differential between exporting and non-exporting firms premium Exporters Firms that export at least 10 percent of their sales Factor share The ratio of total input costs to overall revenues. Factor shares are included for total employment costs (including wages, bonuses, and social security payments), the total cost of raw materials and intermediate inputs, and the replacement cost of machinery, vehicles, and equipment (capital). FDI Foreign direct investment Firms Firms are the respondents to the Enterprise Survey. A firm is a business in the private sector that meets the eligibility criteria for the survey. Formal private sector Firms registered with a government authority and the firm has at least 1 percent of private ownership Formal training Training that has a structured, defined curriculum offered to employees; this type of training does not include employee orientation Fully credit-constrained firms Firms that have no source of external finance and typically applied for a loan and were rejected or did not apply for a loan (FCC) due to unfavorable terms and conditions GVC Global value chain Higher-tech manufacturing High- and medium-high-tech intensity manufacturing (ISIC Rev. 3.1) sectors include chemicals (24), machinery and equipment (29), electrical and optical equipment (30-33), and transport equipment (34-35, excluding 35.1). See http://www.oecd.org/sti/ind/48350231.pdf Glossary of Terms and Acronyms 133 IBRD International Bank for Reconstruction and Development ICT Information and communication technology ILO International Labour Organization IMF International Monetary Fund Importer productivity Average labor productivity differential between importing and non-importing firms premium Importer size premium Average size differential between importing and non-importing firms, where size is measured as number of permanent full-time employees Importers Firms that import at least 10 percent of their inputs Informal firms Unregistered firms Innovation Introduction of new or significantly improved products and processes, as well as new or significantly improved organizational and marketing methods ISIC International Standard Industrial Classification (UN) Jobless growth When the broader economy is growing yet new job creation is very limited Knowledge acquisition Includes contracting R&D with other firms and institutions, or by purchasing or licensing patented technologies, non-patented inventions, and know-how Labor productivity (LP) Total annual sales divided by the number of full-time permanent employees (expressed in 2012 USD) Labor productivity growth Growth in labor productivity between 2009 and 2012, annualized and expressed in constant 2012 USD. The growth measure is calculated by dividing the difference in labor productivity in the two periods by their average. LAC Latin America and the Caribbean. For a full list see pages 135–137 Large firm A firm with at least 100 full-time employees Lower-middle-income Using ES data as a benchmark, the average across lower-middle-income economies where ES data are available. For a full list see pages 135–137 Low-tech manufacturing Low-technology intensity manufacturing (ISIC Rev. 3.1) sectors include food products, beverages and tobacco (15-16), textiles (17-18), leather (19), wood (20), paper, publishing, and printing (21-22), and other manufacturing (36-37). See http://www.oecd.org/sti/ind/48350231.pdf Major obstacle Firms are asked to rate an individual business environment obstacle on a 5 point scale. If the firm chooses a 4 or a 5, then that obstacle is a “major obstacle” for the firm. Management practices Core management practices relating to operations, monitoring, targets, and incentives Marketing innovation Introduction of new or significantly improved marketing methods Medium-sized firm A firm with 20–99 full-time employees Medium-low-tech Medium-low-technology intensity manufacturing (ISIC Rev. 3.1) sectors include building and repairing of ships and boats manufacturing (35.1), rubber and plastics products (25), coke, refined petroleum products, and nuclear fuel (23), other non-metallic mineral products (26), and basic metals and fabricated metal products (27–28). See http://www.oecd.org/sti/ind/48350231.pdf MENA Middle East and North Africa. For a full list see pages 135–137 MENA ES The eight economies in the MENA region that are the focus of this report: Djibouti; Egypt, Arab Rep.; Jordan; Lebanon; Morocco; Tunisia; the West Bank and Gaza; and the Republic of Yemen Movable collateral Collateral based on machinery, equipment, or receivables as underlying assets Movable collateral incidence The share of collateralized loans where either machinery and equipment or receivables were pledged as collateral Not credit-constrained firms Firms that have no difficulties in accessing credit or do not need credit. This category includes firms that did not apply for a loan as they have enough capital (on their own or from other sources) and firms that applied for a loan and the loan application was approved in full. Non-technological innovation Introduction of new or significantly improved organizational or marketing methods OECD Organisation for Economic Co-operation and Development Organizational innovation Introduction of new or significantly improved organizational methods Other services Services firms, excluding retail firms 134 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Partially credit-constrained Firms that have external financing but were discouraged in applying for a new loan due to terms and conditions. Also firms (PCC) included are firms that have external financing and applied for a new loan that was only partially approved. PFTE Permanent full-time employee Process innovation Introduction of new or significantly improved processes Product innovation Introduction of new or significantly improved products R&D Research and development SAR South Asia. For a full list see pages 135–137 Sector The business activity of a firm. The ES classifies firms as manufacturing, retail, or other services. Small firm A firm with fewer than 20 full-time employees Small player exporters Firms below the 50th percentile by their export sales volume SMEs Small and medium-sized enterprises Superstar exporters Top 5 percent of firms by their export sales volume Technological innovation Introduction of new or significantly improved products or processes Top manager The most senior-level manager of the firm, who is making the key decisions on a day-to-day basis Top obstacle The obstacle from a list of 15 possible business environment obstacles, which firms choose as the biggest obstacle to their establishment Transition matrix from census The transition matrix estimates the probability of firms moving from one size category to the other. Based on census data, data it accounts for both entry and exit of firms over the period. Transition matrix from ES The transition matrix estimates the probability of firms moving from one size category to the other. The survey includes data firms existing in 2012 and excludes firms that exited the market in 2009-2012. The transition matrix is thus biased, as it does not account for the exit of firms. Two-way traders Firms that are both importers and exporters UN United Nations UNDP United Nations Development Programme Upper-middle-income Using ES data as a benchmark, the average across upper-middle-income economies where ES data are available. For a full list see pages 135–137 USD United States dollars Wage-size effect A finding in the literature that larger firms tend to pay their employees more compared with smaller firms WBL Women, Business and the Law Report WDI World Development Indicators Young firm A firm that is 5 years old or younger Youth employment Workers below 30 years of age Enterprise Survey–All ES Economy Data 135 All ES economies Country name Survey year Fiscal year Income Region Afghanistan 2014 2012/2013 Low income SAR Albania 2013 2011 Upper-middle-income ECA Angola 2010 2009 Upper-middle-income AFR Antigua and Barbuda 2010 2009 High income: non-OECD High income: non-OECD Argentina 2010 2009 Upper-middle-income LAC Armenia 2013 2011 Lower-middle-income ECA Azerbaijan 2013 2011 Upper-middle-income ECA Bahamas, The 2010 2009 High income: non-OECD High income: non-OECD Bangladesh 2013 2012 Low income SAR Barbados 2010 2009 High income: non-OECD High income: non-OECD Belarus 2013 2011 Upper-middle-income ECA Belize 2010 2009 Upper-middle-income LAC Benin 2009 2008 Low income AFR Bhutan 2009 2008 Lower-middle-income SAR Bolivia 2010 2009 Lower-middle-income LAC Bosnia and Herzegovina 2013 2011 Upper-middle-income ECA Botswana 2010 2009 Upper-middle-income AFR Brazil 2009 2007 Upper-middle-income LAC Bulgaria 2013 2011 Upper-middle-income ECA Burkina Faso 2009 2008 Low income AFR Burundi 2014 2013 Low income AFR Cabo Verde 2009 2008 Lower-middle-income AFR Cambodia 2013 2012 Low income EAP Cameroon 2009 2008 Lower-middle-income AFR Central African Republic 2011 2010 Low income AFR Chad 2009 2008 Low income AFR Chile 2010 2009 High income: OECD High income: OECD China 2012 2011 Upper-middle-income EAP Colombia 2010 2009 Upper-middle-income LAC Congo, Democratic Rep. 2013 2012 Low income AFR Congo, Rep. 2009 2007 Lower-middle-income AFR Costa Rica 2010 2009 Upper-middle-income LAC Côte d’Ivoire 2009 2007 Lower-middle-income AFR Croatia 2013 2011 High income: non-OECD High income: non-OECD Czech Republic 2013 2011 High income: OECD High income: OECD Djibouti 2013 2012 Lower-middle-income MENA Dominica 2010 2009 Upper-middle-income LAC Dominican Republic 2010 2009 Upper-middle-income LAC Ecuador 2010 2009 Upper-middle-income LAC Egypt, Arab Rep. 2013 2012 Lower-middle-income MENA El Salvador 2010 2009 Lower-middle-income LAC Eritrea 2009 2008 Low income AFR Estonia 2013 2011 High income: OECD High income: OECD Ethiopia 2011 2011 Low income AFR continued on next page 136 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Country name Survey year Fiscal year Income Region Fiji 2009 2008 Upper-middle-income EAP Gabon 2009 2007 Upper-middle-income AFR Georgia 2013 2011 Lower-middle-income ECA Ghana 2013 2012 Lower-middle-income AFR Grenada 2010 2009 Upper-middle-income LAC Guatemala 2010 2009 Lower-middle-income LAC Guyana 2010 2009 Lower-middle-income LAC Honduras 2010 2009 Lower-middle-income LAC Hungary 2013 2011 Upper-middle-income ECA India 2014 2012/2013 Lower-middle-income SAR Indonesia 2009 2008 Lower-middle-income EAP Israel 2013 2012 High income: OECD High income: OECD Jamaica 2010 2009 Upper-middle-income LAC Jordan 2013 2012 Upper-middle-income MENA Kazakhstan 2013 2011 Upper-middle-income ECA Kenya 2013 2012 Low income AFR Kosovo 2013 2011 Lower-middle-income ECA Kyrgyz Republic 2013 2011 Lower-middle-income ECA Lao PDR 2012 2011 Lower-middle-income EAP Latvia 2013 2011 High income: non-OECD High income: non-OECD Lebanon 2013 2012 Upper-middle-income MENA Lesotho 2009 2007 Lower-middle-income AFR Liberia 2009 2007 Low income AFR Lithuania 2013 2011 High income: non-OECD High income: non-OECD Macedonia, FYR 2013 2011 Upper-middle-income ECA Madagascar 2013 2012 Low income AFR Malawi 2014 2013 Low income AFR Mali 2010 2009 Low income AFR Mauritania 2014 2013 Lower-middle-income AFR Mauritius 2009 2007 Upper-middle-income AFR Mexico 2010 2009 Upper-middle-income LAC Micronesia, Fed. Sts. 2009 2008 Lower-middle-income EAP Moldova 2013 2011 Lower-middle-income ECA Mongolia 2013 2011 Lower-middle-income EAP Montenegro 2013 2011 Upper-middle-income ECA Morocco 2013 2012 Lower-middle-income MENA Myanmar 2014 2012 Low income EAP Namibia 2014 2013 Upper-middle-income AFR Nepal 2013 2012 Low income SAR Nicaragua 2010 2009 Lower-middle-income LAC Niger 2009 2008 Low income AFR Nigeria 2014 2013 Lower-middle-income AFR Pakistan 2013 2011/2012 Lower-middle-income SAR Panama 2010 2009 Upper-middle-income LAC Paraguay 2010 2009 Lower-middle-income LAC Peru 2010 2009 Upper-middle-income LAC Philippines 2009 2008 Lower-middle-income EAP continued on next page Enterprise Survey–All ES Economy Data 137 Country name Survey year Fiscal year Income Region Poland 2013 2011 High income: OECD High income: OECD Romania 2013 2011 Upper-middle-income ECA Russian Federation 2011 2010 High income: non-OECD High income: non-OECD Rwanda 2011 2010 Low income AFR Samoa 2009 2008 Lower-middle-income EAP Senegal 2014 2013 Lower-middle-income AFR Serbia 2013 2011 Upper-middle-income ECA Sierra Leone 2009 2007 Low income AFR Slovak Republic 2013 2011 High income: OECD High income: OECD Slovenia 2013 2011 High income: OECD High income: OECD South Sudan 2014 2013 Lower-middle-income AFR Sri Lanka 2011 2010 Lower-middle-income SAR St. Kitts and Nevis 2010 2009 High income: non-OECD High income: non-OECD St. Lucia 2010 2009 Upper-middle-income LAC St. Vincent and the Grenadines 2010 2009 Upper-middle-income LAC Sudan 2014 2013 Lower-middle-income AFR Suriname 2010 2009 Upper-middle-income LAC Sweden 2014 2013 High income: OECD High income: OECD Tajikistan 2013 2011 Low income ECA Tanzania 2013 2011/2012 Low income AFR Timor-Leste 2009 2008 Lower-middle-income EAP Togo 2009 2008 Low income AFR Tonga 2009 2008 Upper-middle-income EAP Trinidad and Tobago 2010 2009 High income: non-OECD High income: non-OECD Tunisia 2013 2012 Upper-middle-income MENA Turkey 2013 2011 Upper-middle-income ECA Uganda 2013 2012 Low income AFR Ukraine 2013 2011 Lower-middle-income ECA Uruguay 2010 2009 High income: non-OECD High income: non-OECD Uzbekistan 2013 2011 Lower-middle-income ECA Vanuatu 2009 2008 Lower-middle-income EAP Venezuela RB 2010 2009 Upper-middle-income LAC Vietnam 2009 2008 Lower-middle-income EAP West Bank and Gaza 2013 2012 Lower-middle-income MENA Yemen, Rep. 2013 2012 Lower-middle-income MENA Zambia 2013 2012 Lower-middle-income AFR Zimbabwe 2011 2010 Low income AFR Sources: Enterprise Surveys, income/region categories are from the World Bank lending group definitions for 2012. 138 WHAT’S HOLDING BACK THE PRIVATE SECTOR IN MENA? LESSONS FROM THE ENTERPRISE SURVEY Djibouti, Arab Republic of Egypt, Jordan, Lebanon, Morocco, Tunisia, West Bank and Gaza, Republic of Yemen