WPS5612 Policy Research Working Paper 5612 A Review of Credit Guarantee Schemes in the Middle East and North Africa Region Youssef Saadani Zsofia Arvai Roberto Rocha The World Bank Middle East and North Africa Region Financial and Private Sector Development Unit March 2011 Policy Research Working Paper 5612 Abstract Many countries in the MENA region have established is in line with the international average, although there partial credit guarantee schemes to facilitate SME access are wide differences across countries, and some schemes to finance. These schemes can play an important role, seem too small to make any significant impact. Most especially in a period where MENA governments are importantly, the number of guarantees looks generally making efforts to improve the effectiveness of credit small while their average value looks large. This suggests registries and bureaus and strengthen creditor rights. that guarantee schemes are not yet reaching the smaller This paper reviews the design of partial credit guarantee firms. Guarantee schemes in MENA look financially schemes in MENA, and assesses their preliminary sound and most schemes have room to grow. However, outcomes. The paper is based on a survey conducted in this growth should be accompanied by an improvement 10 MENA countries in early 2010. The authors find of some key design and management features, as well as that the average size of guarantee schemes in MENA the introduction of systematic impact evaluation reviews. (measured by the total value of outstanding guarantees) This paper is a product of the Financial and Private Sector Development Unit, Middle East and North Africa Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank. org. The authors may be contacted at rrocha@worldbank.org, zarvai@worldbank.org, and ysaadanihassani@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team A Review of Credit Guarantee Schemes in the Middle East and North Africa Region Youssef Saadani, Zsofia Arvai and Roberto Rocha1 October 2010 The World Bank 1 The authors have benefitted from the comments provided by the participants of the MENA's financial sector flagship workshop organized jointly with the Arab Monetary Fund in Abu Dhabi in June 2010. The authors are particularly grateful to Khater Abi Habib (Kafalat Lebanon), Jean-Louis Leloir (OSEO France), John Khoury (EPGF Palestine), and Nagla Bahr (CGC Egypt) for their detailed comments and suggestions. The paper has also benefitted from comments from Juan Carlos Mendoza, Michael Fuchs, Leora Klapper, Kevin Carey and Aquiles Almansi. Table of Contents 1 Introduction ............................................................................................................................ 4 2 A Brief Survey of the Literature ............................................................................................ 5 3 The MENA PCG Survey and the Review Methodology ....................................................... 6 3.1 Basic Description of the MENA Survey .................................................................... 6 3.2 The Review Methodology .......................................................................................... 7 4 Reviewing the Design of Partial Credit Guarantee Schemes in MENA ................................ 8 4.1 Main Objectives of the Scheme.................................................................................... 8 4.2 Eligibility Criteria......................................................................................................... 9 4.3 Coverage Ratios ......................................................................................................... 12 4.4 Fees ............................................................................................................................. 14 4.5 Payment Rules ............................................................................................................ 17 4.6 Collateral and Down-payment Rules ......................................................................... 19 4.7 Operational Mechanisms ............................................................................................ 20 4.8 Risk Management and Regulation ............................................................................. 22 4.9 Capacity Building to Participating Institutions .......................................................... 23 4.10 The Use of Counter-Guarantees ................................................................................. 24 5 Preliminary Assessment of Outcomes.................................................................................. 24 5.1 Outreach ..................................................................................................................... 24 5.2 Additionality............................................................................................................... 28 5.3 Financial Sustainability .............................................................................................. 30 6 Summary of Findings and Policy Implications .................................................................... 33 References ................................................................................................................................ 35 Annex 1. Mission Statements of Benchmark PCGs ................................................................ 39 Annex 2. Mission Statements of MENA PCGs ....................................................................... 40 Annex 3. Alternative Methods to Assess Additionality........................................................... 40 Tables Table 1: MENA Partial Credit Guarantee Schemes included in the Survey .............................7 Table 2: Basic Design Components and Outcomes Assessed ...................................................8 Table 3: Eligibility Criteria in Benchmark Countries ..............................................................10 Table 4: Eligibility Criteria Adopted by MENA PCGs ...........................................................11 Table 5: Coverage Ratios in Selected Benchmark Countries ..................................................13 Table 6 : Coverage ratios of MENA PCGs ..............................................................................14 Table 7 : Coverage ratios in MENA and Benchmark countries ..............................................14 Table 8: Fees in Selected Benchmark Countries .....................................................................15 Table 9: Fees Charged by MENA PCGs .................................................................................16 Table 10: Efficiency of the Judicial Process in MENA and Benchmark Countries ................18 Table 11: Payment Rules in MENA ........................................................................................19 2 Table 12: Collateral and Down-payment Rules in MENA ......................................................20 Table 13: Operational Mechanisms adopted in the Benchmark Countries ..............................21 Table 14: Operational Mechanisms in MENA ........................................................................22 Table 15: Outreach of Guarantee Schemes in Benchmark Countries ......................................25 Table 16: Outreach of MENA PCGs .......................................................................................26 Table 17: Alternative Methods to Measure Additionality .......................................................29 Table 18: Findings of Selected Country Studies......................................................................30 Table 19: Building blocks for a comprehensive review ..........................................................34 Figures: Figure 1: Maximum Size of Eligible loans: MENA and Benchmark countries ......................12 Figure 2: Standardized Fees .....................................................................................................17 Figure 3: Outstanding guarantees ............................................................................................27 Figure 4: Outstanding guarantees ............................................................................................27 Figure 6: Average size of guarantees .......................................................................................28 Figure 7: Equity multiplier.......................................................................................................32 Figure 8: Net loss ratios ...........................................................................................................32 3 1 Introduction Expanding SME access to finance has proved a challenge in many developing countries, especially in the Middle East and North Africa (MENA) region. Research shows that SMEs contribute to a large share of employment and GDP in developing economies.2 Despite their importance, SMEs are significantly more financially constrained than large firms, especially in developing countries. This problem seems severe in MENA countries, where about 33% of SMEs report difficulties in getting finance, compared to 25% on average in other emerging countries3. The lack of SME access to finance is to a large extent the consequence of weaknesses in the enabling environment for finance (e.g. weak credit reporting systems, collateral regimes) that result in informational asymmetries and high risks to creditors4. Deficiencies in the enabling environment have motivated government interventions designed to expand SME finance. Government interventions may be justified when it takes time to build an effective enabling environment, or where some groups remain difficult to reach, even when efficient financial infrastructure and regulations are in place. Traditionally, such policy interventions have included partial credit guarantee schemes, direct lending facilities, and lending by state-owned financial institutions. Partial Credit Guarantee Schemes (PCGs) are operated by a large number of countries and are considered one of the most market-friendly types of interventions. In developed countries such schemes have been operational for over four decades while their use in developing countries is more recent. PCGs facilitate access to finance by creditworthy firms when such access is constrained by insufficient credit information and collateral. As a risk- sharing mechanism, PCGs reduce the risks and potential losses of creditors, inducing lending to riskier types of borrowers. Arguably, PCGs generate fewer market distortions compared to other policy interventions, such as directed lending programs or state banks, because they usually entail less interference in credit allocation and use private banks as the main vehicles for loan origination. Many countries have also used PCGs as a countercyclical policy tool. Korea is one of the most notable examples of a country that have used credit guarantees during crises to alleviate the adverse effects on SMEs5. As another example, in the current global crisis the European Union has allowed PCGs in member states to increase the coverage ratio to 90 percent for distressed SMEs until end-2010, and allowed the possibility for subsidized guarantee premiums. In addition, some guarantee schemes introduced simplified and faster approval processes (e.g. Portugal, Romania, Greece) or raised the maximum guaranteed loan amount (e.g. Germany). Many countries in the MENA region have established PCGs to facilitate SME access to finance. These schemes can play an important role, especially in a period where MENA governments are making efforts to improve the effectiveness of credit registries and bureaus 2 Ayyagari et al., (2003) . 3 See www.entreprisesurveys.org 4 Beck, Demirguc-Kunt and Peria (2009) show that differences in the quality of the legal framework explain the differences in SME lending between developed and developing countries. Rocha, Farazi, Khouri, and Pearce (2010) provide similar evidence for MENA. A review of SME finance is provided in World Bank (2008) and IFC (2010) 5 IFC (2010). 4 and strengthen creditor rights. There is some evidence that credit guarantee schemes have contributed to more SME lending in the region ­ the MENA countries that have larger and more established PCGs have larger shares of SME lending and this result seems to hold when controlling for other factors.6 The central policy question, however, is whether these schemes are cost-effective, i.e., whether they are able to target financially constrained SMEs, reach a significant number of these firms, and remain financially sustainable. The objective of this paper is to review the design of PCGs in MENA and assess their preliminary outcomes. A survey was conducted in 10 MENA countries in early 2010 to gather the information needed for the assessment. In each country, the largest credit guarantee scheme was surveyed. The survey covered the main rules of the scheme, the management of the scheme, and the key outcome indicators. The survey results allow for a review of these schemes based on comparisons with other mature schemes outside the MENA region. The paper is structured as follows. The next section provides a brief survey of the literature on PCGs. The third section describes the MENA survey and the methodology adopted for reviewing PCGs. The fourth section reviews the rules of PCGs in MENA, while the fifth section provides a preliminary analysis of their outcomes. Finally, the sixth section concludes and identifies the main elements of the agenda for improving the effectiveness of PCGs in MENA. 2 A Brief Survey of the Literature There is a growing body of literature on partial credit guarantee schemes, reflecting the increasing interest on this type of policy intervention to support SME access to finance. This literature can be classified into three broad areas. The first consists of cross-country surveys describing the main features of guarantee schemes (e.g. Beck and al. (2008), Bennett and al. (2005)). The second consists of individual country studies, including efforts to assess additionality (e.g. Ridding (2007), Cowan and al. (2009)). Finally, a third category focuses on best practices and design issues, drawing on the international experience (e.g. Deelen and Molenaar (2004), Green (2003)). The World Bank conducted the first large scale cross-country survey of PCGs in 2008 (Beck, Klapper, and Mendoza, 2008). The objective of this survey was to provide an overview of the key features of guarantee schemes around the world, such as eligibility criteria, coverage ratios, fees, and selected indicators of operational and financial performance. The sample comprised 76 guarantee schemes operating in 46 developed and developing countries (However, Egypt was the only MENA country included). The survey shows that there are large differences in the organizational features and rules of guarantee schemes around the world. Interestingly, these differences are not systematically related to financial and economic development. One of the many interesting findings of the survey is that few guarantee schemes around the world use risk-based pricing or risk-management mechanisms. The authors call for further empirical research on specific schemes to better understand which features work best in practice. They also stress the importance of doing proper cost-benefit analysis to assess whether guarantee schemes are cost-effective. 6 Rocha, Farazi, Khouri, and Pearce (2010). 5 Some country studies have concluded that PCGs have contributed positively to SME access to finance. Although measuring the impact of PCGs accurately remains technically challenging (Section 5), some recent studies have concluded that PCGs have been able to extend finance to firms that otherwise would have remained constrained. For example, in Canada Ridding (2007) estimates that 75% of guarantees are used by firms that would not have been able to obtain a loan otherwise. In Chile, Larrain and Quiroz (2006) find that the guarantee scheme increases the probability of small firms to get a loan by 14%. At the same time, PCG schemes may add limited value and prove costly when they are not well designed. As noted by Honohan (2008), loose eligibility criteria, low fees, and overly generous coverage ratios may result in the provision of guarantees to enterprises that would have obtained credit anyway. They may also result in financial imbalances requiring recurrent government contributions. Along these lines, Bechri et al. (2001) studied the case of the Tunisian scheme FOPROPI, which became unsustainable and finally collapsed in 1997 as a result of major institutional failures. Guarantee schemes around the world vary on fundamental design features, but there is a growing effort to identify good practices. The failure of several guarantee schems in the 1980s led to an intensive debate on their role (Levistky, 1997). As noted by Green (2003) the weaknesses of early guarantee schemes can be avoided through proper design and institutional arrangement. Some recent studies provide guidelines and discuss operational parameters of guarantee funds, based on international experience. Deelen and Moleenar (2004) published a practical manual for guarantee funds managers. Along these lines, the European Commission established an expert group on guarantee schemes to identify and disseminate best practices (European Commission, 2006). This literature converges on broad principles, including the need to build attractiveness while ensuring additionality through well designed eligibility criteria, proper coverage ratio and fees, sound risk management, and efficient operational procedures. 3 The MENA PCG Survey and the Review Methodology 3.1 Basic Description of the MENA Survey This paper is based on a survey of MENA PCGs conducted in the first quarter of 2010. The questionnaire prepared for the survey covered the institutional set up, the main operational rules, and the main performance indicators. The questionnaire was partly based on Beck and al. (2008) to ensure comparability with other guarantee schemes around the world. The survey was initiated in February 2010 and completed in April 2010. The authors met with several managers of surveyed schemes to present the objectives of the survey and discuss technical issues. There was also a follow-up effort to ensure the timely completion of the survey and check the accuracy of the data. The survey covered the largest credit guarantee schemes in 10 MENA countries. As shown in Table 1, the oldest guarantee fund in MENA was established in Morocco in 1949, while the youngest one in Syria starts operation in 2010. The average equity is US$50 million, ranging from US$10 million in Syria, to US$75 million in Morocco. Half of these guarantee schemes are majority state-owned (Morocco, Tunisia, Jordan, Syria, Saudi Arabia, UAE), while the others are majority owned by banks (Lebanon, Egypt, Iraq) or donors (Palestine). 6 3.2 The Review Methodology The outcomes of a guarantee scheme can be assessed along three main dimensions: outreach, additionality, and financial sustainability. Outreach refers to the scale of the guarantee scheme, as measured by the number of guarantees issued to eligible SMEs and the amount of outstanding guarantees. The greater the outreach, the stronger is the impact of the scheme on the SME sector. However, the impact of the guarantee scheme will also depend on whether guarantees are extended to firms that are credit constrained, and not to firms that would be able to obtain a loan anyway. This is why additionality is another key outcome that is taken into account. Furthermore, reaching firms that are credit constrained involves risk- taking and financial losses. Even if the objective of a guarantee scheme is not to make a profit, the scheme should still be financially sustainable through sound rules, effective risk management, and regular funding. Table 1: MENA Partial Credit Guarantee Schemes included in the Survey Equity Shareholders (%) Starting Name date Government Banks Other (US$ million) Egypt Credit Guarantee Company 1991 52 - 90 10 Iraq Iraqi Company For Bank Guarantees 2007 12 100 Jordan Jordanian Loan Guarantee Corp 1994 60 14 20 Lebanon Kafalat 1999 50 37.5 62.5 - Morocco Caisse Centrale de Garantie 1949 75 100 - Palestine European-Palestinian Credit Guarantee 2005 40 100 - Fund Saudi Saudi Industrial Development Fund 2005 57 (funds50 donated 50 - Arabia by donors) Syria Loan Guarantee Institution of Syria 2010 10 94 - 6 Arabia Tunisia Sotugar 2003 48 100 - UAE Khalifa Fund 2010 NA 90 10 N/A Designing a guarantee scheme may entail trade-offs among the main objectives. The design of a guarantee scheme must strike a balance between the objectives of outreach, additionality, and financial sustainability. For example, targeting riskier types of borrowers through strict eligibility criteria may have a positive impact on additionality, but may also reduce outreach and lead to larger losses. Similarly, very high fees improve additionality by discouraging banks to use the guarantee for good borrowers, but may reduce outreach, and may generate adverse selection effects. The optimal balance between these three objectives will depend to a good extent on country conditions. For example, in countries with more serious shortcomings in financial infrastructure and limited SME financing, high outreach and high additionality may be achieved simultaneously, while more advanced countries may only increase outreach at the expense of additionality. The design of guarantee schemes in MENA was reviewed against general guiding principles and international practice. There is no unique recipe or one-size-fits-all formula 7 for designing effective guarantee schemes. Our review is based on general guiding principles derived from general insurance principles, a thorough literature review, and international practice7. For the international benchmarking, we selected a number of credit guarantee schemes in developing and developed countries that are reasonably well-established, including Canada's SLFP, Chile's FOGAPE, Colombia's Fondo Nacional de Garantías, France's OSEO, Hungary's Garantiqa, India's CGTMSE, Korea's KODIT, the Netherland's BMKB, Romania's National Credit Guarantee Fund for SMEs, Taiwan's SMEG, and the US SBA. Table 2 summarizes the design components and the outcomes that are assessed. Table 2: Basic Design Components and Outcomes Assessed MAIN OBJECTIVES OF THE SCHEME Main objectives Mission statement of the PCG scheme RULES OF THE SCHEME Eligibility criteria Characteristics of eligible firms (size, sectors, age) and eligible financing Coverage ratio Percentage of risk taken by the guarantee fund Fees Price of the guarantee Payment rules Triggers related to the payment of the guarantee Collateral and down payment Collateral and down payment required when using the guarantee MANAGEMENT OF THE SCHEME Operational mechanism Individual, portfolio or hybrid approach Credit risk management Credit risk management tools (credit scoring and rating, credit registry) Assistance to participating institutions designed to increase their lending Capacity building and risk management capacity OUTCOMES OF THE SCHEME Outreach Number of eligible firms that are covered by the scheme Additionality Capacity to target firms that are effectively credit constrained Capacity to contain losses and maintain an adequate level of equity given Financial sustainability the expected liabilities 4 Reviewing the Design of Partial Credit Guarantee Schemes in MENA 4.1 Main Objectives of the Scheme MENA PCGs generally have broader objectives than those in the benchmark countries. As shown in Annex 1, the mission statements of Guarantee Schemes in benchmark countries emphasize access to finance for SMEs that lack adequate collateral (Annex 1). By contrast, Annex 2 shows that MENA Guarantee Schemes have broader developmental objectives, such as supporting export capacity (Jordan, Morocco), fostering entrepreneurial spirit (UAE), improving the financial sector's skill base (Syria, Iraq), facilitating investment in innovation (Morocco), and supporting national industrialization programs (Saudi Arabia). These broad 7 Beck and al. 2008; Honohan 2008 ; Green 2003; European Commission 2006; Deelen and Molenaar 2004. 8 objectives suggest that MENA schemes interpret the additionality objective more liberally than schemes in other countries. 4.2 Eligibility Criteria General guiding principles and international experience Eligibility criteria should target financially constrained SMEs while providing for some flexibility. Targeting is important to ensure additionality, although overly restrictive eligibility criteria should be avoided because there is uncertainty in practice about the firms that are credit constrained and the type of financing that is lacking. Very low ceilings, excessive restrictions on the types of loans or eligible sectors may exclude firms that are credit constrained and generate threshold effects (excluding many firms just above the threshold, even if credit constrained). The relevance of eligibility criteria can be strengthened through market surveys that identify SME financing gaps. Most guarantee schemes in the benchmark group target SMEs in a broad sense and generally do not restrict sectors or types of loans (Table 3). All the countries in the benchmark group allow start-ups to apply for guarantees (though there is no uniform definition of start-ups across countries other than in the EU). It is also noticeable that these schemes do not impose restrictions on sectors (except for a general restriction on agriculture in the case of Canada), or type of loan (again, except for Canada, which does not guarantee working capital loans). The main differences seem to lie in the limits imposed on firm and loan size. Korea does not impose any limits on firm size, while France and the Netherlands target SMEs following the EU's definition (maximum turnover of 50 million euros and 250 employees). The other countries impose much lower limits on firm size, especially regarding turnover. However, the limits imposed on loan size are probably the binding ones, and here the ranking changes significantly, especially when the limit is defined in relation to per capita income. As shown in Table 3, the Asian schemes look more generous in this case, while the Canadian, Dutch and US schemes look restrictive by comparison. Reviewing Eligibility Criteria in MENA Eligibility criteria differ significantly across MENA guarantee schemes. All schemes cover start-ups except for the Palestine, but there are significant differences regarding firm size (Table 4). Some schemes seem generous regarding firm size ­ Morocco and Tunisia do not set any ceilings, while Jordan and Syria set their ceilings at the high EU level (250 employees). By contrast, Egypt, Lebanon and the Palestine restrict the use of guarantees to smaller firms (respectively 50, 40 and 20 employees. The employee limit for the Palestine scheme is especially low by international comparison. There are significant differences regarding the maximum size of loans. The guarantee schemes in Morocco and Tunisia cover loans up to US$ 2 million, or the equivalent of 600 times GDP per capita. These are high ratios by international standards as shown in Figure 1. The ratios in Egypt, Jordan and Syria are lower (150 times GDP per capita), but still high by international standards. By contrast, eligible loans in Lebanon, the Palestine, and Saudi Arabia are smaller and more comparable to other PCGs outside MENA (50-60 times GDP per capita). 9 Table 3: Eligibility Criteria in Benchmark Countries Loan size Working Start-ups Firm size limit limit Sectors capital (US$ million) All (except Canada Yes Sales: US$5 million 0.5 No agriculture) Chile Yes Sales: US$3 million 0.45 All Yes All (except Colombia Yes Assets: US$7.3 million 0.97 Yes agriculture) All (except for Sales: 50 million euros most France Yes 3.51,2 Yes Employees: 250 agriculture firms) Sales: 50 million euros or Balance sheet total: 43 million 12.5 million Hungary Yes All Yes euros euro1, 3 Employees: 250 India Yes Assets: US$1 million 0.2 All Yes Korea Yes All 3 All Yes Sales: US$1.6 million Employees: 50 Malaysia Yes 3 All Yes Manufacturing: US$7 million Employees: 150 Sales: 50 million euros Netherlands Yes 1.8 All Yes Employees: 250 Sales: 50 million euros Romania 3.2 All Yes Employees: 250 Services: US$3 million and 100 Taiwan, China Yes employees; 31 All Yes Manufacturing: 200 employees US Yes Sales: US$7 million 2 All Yes 1) Exposure limit 2) 800/2008 EU regulation for state aid applies 3) 800/2008 EU regulation for state aid applies to loans counter-guaranteed by the state, and is usually binding at a loan amount well below the exposure limit. 10 Table 4: Eligibility Criteria Adopted by MENA PCGs Max Loan Max Loan Short-term Start-ups Size Maturity Sectors Working Firm size (US$ Million) (years) capital Max 50 Egypt Yes 7 All Yes employees 0.35 Max 50 Iraq Yes 0.25 5 All employees Max 250 Jordan Yes 0.6 8 All Yes employees Agriculture, Max 40 Industry, Lebanon Yes 0.4 7 Yes employees Tourism, High Tech, Crafts Morocco Yes All 1.5* 12 All Yes Max 20 Palestine No 0.1 5 All Yes employees Saudi Max sales US$ All, except Arabia Yes 0.4 7 N/A 5.Million trading Max 250 Syria N/A 0.4 7 All No employees Manufacturing, Tunisia Yes All 2.5 15 No some services UAE Yes All 1.3 7 All Yes *Exposure limit on each transaction There are also significant differences regarding eligible sectors. Morocco, Egypt, Jordan, Palestine, and Syria allow the use of the guarantee for all sectors. A second group of countries, composed of Tunisia, Lebanon and Saudi Arabia, excludes trading and some services. However, there is some uniformity regarding maximum loan maturity, with most schemes setting the maximum maturity at 7-8 years, except for Morocco and Tunisia, which guarantee loans up to 12-15 years, and the Palestine, which imposes a very short maximum maturity. It is also noticeable that some schemes do not guarantee working capital loans. There is scope for revising eligibility criteria in some MENA schemes. In some cases, eligibility criteria could be tightened to enhance additionality, while in others they look overly restrictive and could be relaxed in order to extend finance to small but promising firms. For example, in Morocco and Tunisia, there is no ceiling on firm size and the maximum loan ceiling is well above the international average (Figure 1). This may encourage banks to use the guarantee for large firms and loans, weakening the additionality of those schemes. The definition of SMEs in Jordan and Syria are very similar to the definition used in the EU (maximum 250 employees), which might not be relevant given their economic structures. On the other extreme, in the Palestine, the maximum size of firms (20 employees) seems overly restrictive. This limit can bias against labor-intensive sectors, such as small manufacturing firms, and firms having a higher share of formal employees (compared to firms with large share of informal employees). It can also generate threshold effects, excluding firms just above the threshold, even if credit constrained. 11 It is also surprising that some types of financing are restricted in some countries, such as the restriction on start-up loans in the Palestine. Similarly, guarantees cannot be used for working capital loans in Tunisia and Syria. Therefore, more flexibility might be needed in some schemes in MENA to allow a broader range of firms facing credit constraints to use the guarantee. At the same time, the rationale for guaranteeing loans with very long maturities (12-15 years) is not clear as most investment projects implemented by SMEs do not have such long durations. Figure 1: Maximum Size of Eligible loans: MENA and Benchmark countries (Scaled by GDP per capita, 2009) 700 625 600 570 500 426 428 400 300 200 200 176 187 145 153 157 100 83 54 66 37 43 50 50 5 13 0 4.3 Coverage Ratios General guiding principles and international experience Coverage ratios should preserve incentives for effective loan origination and monitoring while providing sufficient protection against the risk of default. The coverage ratio needs to provide sufficient protection against credit risk, while also preserving incentives for banks to screen and monitor borrowers. Beck et al (2008) show that the median coverage ratio in a large sample of PCGs is 80%. The Chilean experience with bidding procedures shows that banks demand a coverage ratio of about 70% to extend long term loans to riskier types of borrowers (Benavente, 2006). The bidding procedure adopted in Chile provides an interesting market test of the levels of coverage that make the scheme attractive to lenders.8 In our comparator group (Table 5), the coverage ratio ranges from 30% to 100%, with a median value of about 75%. 8 Banks bid for a given amount of guarantees indicating the coverage ratios they are willing to accept for a given level of fees. Banks requesting the lowest coverage ratio are those who win the auction. 12 Several PCGs provide higher coverage ratios to riskier types of borrowers. Banks will require higher coverage to extend loans to riskier borrowers. Many PCGs extend such higher coverage while also charging a higher fee. As shown in Table 5, in France and the Netherlands, the coverage ratio is higher for innovative firms and start-up loans. In Korea, risky firms with low credit scores get higher coverage. In Chile, the maximum coverage ratio for small firms is 80%, compared to 50% for medium firms. Setting a higher coverage ratio for riskier types of borrowers is a way to enhance additionality while providing some flexibility (less risky borrowers can use the benefit from the guarantee but with a lower coverage ratio, and paying a lower fee). Table 5: Coverage Ratios in Selected Benchmark Countries Coverage ratio Link to Risk Exposure Min Median Max Canada 85% 85% 85% No scalability 80% Small firms (Max sales US$ 750,000; Loan US$ 100,000); Chile 50% 65% 80% 50% Medium firms (Max sales US$ 3 million; Loan US$400,000) Colombia 40% 60% 80& According to type of loan/firm France 40% 55% 70% 40%-50% in general, 60% Innovation, 70% start-ups Max 80% in general, Hungary n/a n/a 90% Max 60% on agricultural loans Max 90% firms affected by the crisis (until end-2010) 75% in general India 75% 80% 85% 85% on loans to micro firms <= US$ 10,000 Depending on firms credit score: Eligible firms with the lowest Korea 50% 70% 90% credit score: 90%, Firms with the highest credit score: 50% Malaysia 30% 65% 100% According to type of loan/firm Netherlands 50% 65% 80% 50% in general, 60% innovative businesses, 80% start-ups Romania n/a n/a 80% According to type of loan/firm Taiwan 50% 65% 80% According to type of loan/firm 75 % on loans >US$ 150,000 USA 75% 80% 85% 85 % on loans<= US$ 150,000 Reviewing Coverage Ratios in MENA Coverage ratios in MENA are generally in line with international practice, but some schemes seem to offer high coverage. As shown in Tables 6 and 7, the minimum, median, and maximum coverage ratios in MENA are similar to those in the benchmark group. The average minimum ratio in MENA is just slightly higher than the equivalent average in the benchmark group, the average median is very similar, and the average maximum is actually lower (Table 7). However, there are some differences across counties. Some schemes seem 13 to have high minimum ratios (Iraq, Jordan, Lebanon, UAE), and some of these have high maximum ratios as well (Lebanon, UAE). There is scope for calibrating coverage ratios in some of these cases. Most importantly, some schemes in MENA do not link coverage ratios to the borrowers' risk profile. Morocco, Tunisia, Egypt and Saudi Arabia offer higher coverage ratios for riskier types of borrowers. However, in Syria, Jordan, Iraq, UAE and the Palestine, the coverage ratio is flat and not linked to the risk exposure. These schemes could consider introducing variable coverage ratios, in line with international practice. Table 6 : Coverage Ratios of MENA PCGs Coverage ratio Link to Risk Exposure Min Median Max Egypt 50% 60% 70% Medium firms 50% ( >10 employees); Small firms 75% (< 10 employees), Iraq 75% 75% 75% No scalability Jordan 70% 70% 70% No scalability Small-sized loans (< US$ 200,000): 75%, Medium-sized loans Lebanon 75% 82.5 90% (< US$ 400,000): 85%;, Innovative loans: 90% Working capital 50% , Fixed assets 60%, Start-ups 80% (70% Morocco 50% 65% 80% for loans > US$125,000) Palestine 60% 60% 60% No scalability Saudi Arabia 50% 62.5% 75% General: 50%; Start-ups 75%, Syria 50% 50% 50% No scalability General: 60%; Prioritized firms 75% (Development zones, start- Tunisia 60% 67.5% 75% ups) UAE 90% 90% 90% No scalability Table 7 : Average Coverage ratios in MENA and Benchmark countries Average Min Average Median Average Max Benchmark countries 54% 69% 84% MENA 63% 68% 74% 4.4 Fees General guiding principles and international experience Fees should be related to the risk exposure and contribute to the financial sustainability of the guarantee scheme. Linking the price of the guarantee to the risk exposure is a basic insurance principle that should generally be adopted by guarantee schemes. Moreover, fees are not only a critical source of revenue (and therefore financial sustainability) for guarantee 14 schemes; they also play an important role in building additionality. When fees are sufficiently high, banks are discouraged to use the guarantee for good clients who can obtain loans without additional guarantees. In our benchmark group, the level of fees ranges from 0.8% to 2.3% p.a., with an average fee of 1.5% p.a (Table 8). Note that these are basic standardized rates expressed as a percentage of the guarantee that are comparable across countries9. Although Beck et al (2008) report that only 21% of guarantee schemes around the world utilize risk-based fees, most of the schemes in our benchmark group link fees to the risk exposure. For example, in the Netherlands, higher fees are charged on guarantees to riskier types of firms, such as start- ups or innovative firms. In Korea, Malaysia and Taiwan, fees vary according to the credit rating of the borrower. In Hungary, fees are determined based on the credit rating of the borrower and the risk rating of the loan in the case of loans over approx. 350,000 with government counterguarantee, and all loans without government counterguarantees. The lower the credit score, the higher the fee. In Chile, the level of fees varies across banks according to the quality of their portfolio as measured by the default rate. Table 8: Fees in Selected Benchmark Countries Fees Basic Link to Risk Official standardized Definition rate (% p.a.) 2% of the loan amount + 1.25% Canada 2.3% No scalability p.a. calculated on the loan balance Higher fees for banks with higher default Chile 1% to 2% p.a. 1.5% rates Fees are link to the product and coverage Colombia 0.95% - 3.85% p.a. ratio Fees are linked to the coverage ratio: 0.6% to 0.9% p.a. of the loan France 1.3% 0.6% (40% coverage ratio), 0.9% (70% value coverage ratio) For loans over 350,000 euros, fees vary Hungary 1% - 5% p.a. of guarantee amount 2% according to firms' credit ratings Fees are lower for loans up to US$ India 1.5% upfront + 0.75% p.a. 1.5% 10,000 (1.25% per annum) Higher fees for low credit rating along Korea 0.5 % to 3% p.a. 1.2% with higher coverage ratio Malaysia 0.5% to 3.6% p.a. 1.5% Higher fees for low credit rating Netherlands 2% to 3.6% one-off 1.7% Fees are linked to the coverage ratio Romania 1.5% per annum 1.5% Fees are linked to the coverage ratio Taiwan, China 0.75% to 1.5% per annum 0.8% Fees are linked to risk profile 2%-3.5% of the loan amount + United States annual rate of 0.55% of the 1.9% Higher fees for larger loan amounts outstanding guarantee balance Note: see footnote (7) 9 To ensure comparability across guarantees schemes, we converted flat rates into per annum rates, assuming a loan maturity is 4 years. The "standardized fee rate" is expressed as a percentage of the guarantee amount. When several fee rates exist, we take the fee of the most important guarantee product (the "basic rate"). 15 Reviewing Guarantee Fees in MENA Some MENA schemes do not seem to price their guarantees adequately. The average fee charged by MENA schemes is 1.5% p.a., similar to the average fee in the benchmark group (Table 9). However, some MENA schemes seem to underprice the guarantee (Figure 2), which may undermine financial sustainability and weaken additionality. Moreover, most MENA schemes do not link the price of the guarantee to the risk exposure, excepting for Morocco. These countries may consider linking more closely the fee to the coverage ratio and other aspects of the risk exposure. Table 9: Fees Charged by MENA PCGs Fees Official Standardized Scalability Definition (% p.a.)* Egypt 2% per annum 2% Lower fees for health care Iraq 2% per annum 2% No Jordan 1%-1.5% N/A N/A Lebanon 2.5% per annum 2.5% No 2% flat in general, Morocco 2% of the loan value (flat) 1% 0.5% on working capital 1.5% for start-ups US$ 125,000 1% of the original loan amount Palestine 1. 5% annual commission on No the outstanding guarantee 2.1% Saudi Arabia N/A N/A No Syria N/A N/A No 1% flat short-term loan (standardized Tunisia 0.6% per annum 0.6% 1.2%) UAE N/A N/A No Note: see footnote (7) 16 Figure 2: Standardized Fees (Percentage per annum, standardized basic rate) 3 2.5 2.5 2.4 2.3 2.1 2 2 2 2 2 1.7 1.5 1.5 1.5 1.3 1.2 1 1 0.8 0.6 0.5 0 Note: see footnote (8) 4.5 Payment Rules General guiding principles and international experience The payment of claims should be quick and predictable in order to build the credibility of the guarantee scheme, while encouraging loan collection. The capacity to pay promptly the claims is a key factor to induce banks to use the guarantee. However, the challenge is to design a payment rule which is reliable while providing incentives for loan recovery. There are four types of payment rules that can be considered: (i) a single payment after default is validated; (ii) a single payment after legal actions are initiated; (iii) partial payment at the time of default, followed by the remaining payment when judicial procedures are exhausted; and (iv) single payment when judicial procedures are exhausted. Beck et al (2008) show that in 66% of guarantee schemes around the world, banks are responsible for the recovery of defaulting loans. Moreover, in 34% of the schemes payouts are made after the borrower defaults. In 42% of the schemes, payout takes place when the bank initiates legal actions. In only 14% of the schemes payment is held until the bank writes off the loan. The choice of a payment rule should take into account the efficiency of the judicial system. In countries with efficient judicial systems, the payment of claims can be made when all judicial procedures are exhausted. In France, Canada, and the US, claims are paid on the basis of realized losses, once all judicial procedures are completed. However, in countries where the judicial system is less efficient, paying claims at the end of the judicial process may result in long waiting periods and losses to lenders, and hinder the attractiveness of the guarantee scheme. Table 10 provides an illustration of the differences in the efficiency of loan collection among MENA countries and the benchmark group. The numbers in the table apply to the judicial system, and do not necessarily represent recovery rates and times for the respective PCG portfolios. 17 Table 10: Efficiency of the Judicial Process in MENA and Benchmark Countries (Doing Business 2010) Country Recovery rate Time (years) MENA (cents on the dollar) 27.7 3.4 Egypt 16.8 4.2 Iraq NA NA Jordan 27.3 4.3 Lebanon 19 4 Morocco 35.1 1.8 Saudi Arabia 37.5 1.5 Syria 29.5 4.1 Tunisia 52.3 1.3 UAE 10.2 5.1 Benchmark Countries 70.4 1.6 Canada 88.7 0.8 Chile 21.3 4.5 Colombia 35.3 1.7 France 44.7 1.9 Hungary 38.4 2.0 India 15 7 Korea 80.5 1.5 Malaysia 38.6 2.3 Netherlands 82.7 1.1 Romania 28.5 3.3 Taiwan, China 80.9 1.9 United States 76.7 1.5 Reviewing Payment Rules in MENA MENA countries are exploring different ways to reconcile payment efficiency and loan collection. Most MENA guarantee schemes have rules that allow payment of claims before legal procedures are exhausted (Table 11). This is probably the right approach, given the relatively low efficiency of judicial procedures in MENA ­ as shown in Table 11, the average recovery rate in MENA is 28% compared to 70% in the benchmark group, and the time needed to complete the process is 3.4 years, compared to 1.6 years in the benchmark group. In order to induce banks to collect defaulting loans, some guarantee schemes in MENA are testing different incentive structures. Morocco and Tunisia provide an advance payment of 50% once the claim is presented, followed by the balance once legal procedures are exhausted. Lebanon's Kafalat makes the payment 90 days after the claim is validated, but recovers itself the collateral. In Syria, the payment is deposited in an escrow account at the bank until the legal procedures are exhausted. These models have not been sufficiently tested yet, and it is too early to assess their effectiveness. The hybrid payment rules offer a potential solution to the challenge of building credibility while promoting loan collection by the banks, but these systems have not 18 been sufficiently tested. The hybrid rules will need to be further assessed using cost-benefit analysis and feedback from lenders. Hybrid systems should in any case include a maximum period (as in the case of Morocco) to avoid too much uncertainty for guarantee users. Table 11: Payment Rules in MENA Iraq Single payment 30 days after the bank initiates legal procedures. Jordan After the bank initiates legal procedures to recover the debt. Single payment once the claim is presented and validated. Lebanon Payment 90 days after 3 unpaid installments. Single payment once the claim is presented and validated. The guarantor is mainly responsible for recovering the collateral Morocco Advance payment of a 50% once the claim is presented and validated, followed by the balance once legal procedures have been exhausted (maximum 3 years) Palestine Single payment after six months from default date Saudi Arabia After a fixed number of days following default. Single payment once the claim is presented and validated Syria Money is deposited in an escrow account at the bank until the legal procedures are exhausted. Tunisia Advance payment of a 50% once the claim is presented and validated, followed by the balance once legal procedures have been exhausted UAE Single payment once the claim is presented and validated. The guarantor is mainly responsible for recovering the collateral 4.6 Collateral and Down-payment Rules General guiding principles and international experience Guarantee schemes should be allowed to require collateral and down payments subject to reasonable limits. One of the main roles of guarantee schemes is precisely to compensate for the lack of collateral hindering SME access to finance. However, the complete absence of collateral may generate adverse selection and moral hazard effects and ultimately result in large losses for the scheme. To mitigate this risk, the scheme should be allowed to require whatever collateral is available up to reasonable limits. For example, in France and in Canada, the schemes are allowed to require personal guarantees but these guarantees are capped respectively at 50% and 25% of the loan value. Reviewing Collateral Rules in MENA Most MENA guarantee schemes allow banks to take collateral but do not impose ceilings (Table 12). This may contradict the objectives of PCGs, although there is little information on additional collateral provided by guarantee users in MENA. This information should be collected and disclosed. Some enterprise surveys conducted by the World Bank indicate that banks in MENA tend to require high levels of collateral even when using 19 guarantees, which can defeat the purpose of a guarantee scheme 10. Some countries set ceilings for collateral, such as Morocco (maximum collateral of 100%), and more recently Lebanon (50%). However, most other schemes do not impose ceilings. Many MENA schemes also impose minimum down-payment rules. Requiring minimum contributions from borrowers can be an effective way to reduce adverse selection and moral hazard, especially for riskier types of loans, such as start-ups or long term investments. This feature is used in many MENA countries. With a significant down payment the loan amount represents a smaller percentage of the value of collaterized assets, thus improving the (theoretical) recovery expectancy. Moreover, the down-payment rules seem reasonable, capped at about 20%-30% of the project value. Table 12: Collateral and Down-payment Rules in MENA Down-payment Collateral Egypt Medium firms: 20% Allowed, no ceiling Jordan 30% SME loan, 30% industrial loan, Allowed, no ceiling Leasing 50% Lebanon Kafalat plus: 20% of total investment, Allowed, Ceiling of 50% of the loan amount 30% if start ups (Kafalat Basic) Morocco Start-up loans: 10%-20% depending on Allowed, Ceiling of 100% of the loan amount the loan amount Palestine No Allowed, no ceiling (in practice, the majority of loans are not secured against collateral) Tunisia 30% of the cost of the investment Allowed, no ceiling Saudi Arabia No Allowed, no ceiling Syria N/A N/A 4.7 Operational Mechanisms General guiding principles and international experience Guarantees can be delivered though the individual, portfolio or hybrid approaches. Under the individual approach, every loan application is assessed and approved by the guarantee scheme. The portfolio approach is more flexible and allows banks to extend guarantees without consulting the guarantee scheme. Each bank receives a guarantee allocation which can be used for eligible firms. The hybrid approach mixes elements of individual and portfolio approaches: certified lenders can extend guarantees without referring to the guarantee scheme up to a limit; above a certain threshold, the guarantee scheme adopts an individual approach and appraises the loan application before extending the guarantee. Beck et al (2008) report that 72% of schemes surveyed use the individual approach, 14% the portfolio approach, and 9% the hybrid approach. The schemes in our benchmark group adopt either the portfolio approach (Canada, Netherlands, UK, and Chile) or the hybrid approach (France, USA, Taiwan, Hungary, and Korea) as shown in Table 13. 10 Lebanon Investment Climate Assessment, World Bank, 2009 20 Each approach has its advantages and limits. The main advantage of the individual approach is its potential to better control credit risk and ensure financial sustainability. In the case of banking systems with less experience with SME lending, the individual approach has another important value added, namely it allows the provision of information and technical support by the scheme to the bank through exchanges during the decision making process. By contrast, the portfolio approach involves higher risks for guarantee schemes, but reduces substantially operational and transaction costs. The hybrid approach aims to combine the advantages of the two approaches, while overcoming their limitations. Table 13: Operational Mechanisms Adopted in the Benchmark Countries Countries Operational Mechanism Canada Portfolio Portfolio: FOGAPE auctions available guarantee amounts, with the lenders bidding Chile on the coverage ratio. Colombia Hybrid Hybrid : individual in general, delegation of guarantee decision to banks for France loans