Report No. 47348-LAC Developing New Structured Financial Products to Channel Savings Towards Small and Medium Enterprises (SMEs) Growth February 2, 2009 Finance and Private Sector Unit Poverty Reduction and Economic Management Unit Latin America and the Caribbean Region Document of the World Bank Table of Contents Chapter 1. The role of financial marketsinensuring growth sustainability inPeru................8 Chapter 2 . The state of SME financing inPeru...................................................................... 16 Chapter 3. SMEs and structured finance inperspective......................................................... 35 Chapter 4 . Structured Finance inPeru: Institutional. Legal. and Regulatory Obstacles ........42 List of Tables Table 2 Reasonsfor not applying for a lqan (%) ......................................................................... Table 1.Bank Branches inInternational Comparison.................................................................. 25 Table 3 Sources of funding (%) ................................................................................................... 26 32 Table 4 Sources of funding for investment infixed assets (%) ... ................................................... 32 Table 5: Source of loan and average yearly interest rate (%) ....................................................... 34 Table 6: Participant's background and role .................................................................................. 44 Table 7: Issuancecosts of Corporate Bonds inthe Peruvian Capital Market ............................... 49 Table 8: Legal obstacles to implementation of CMAC S M E loan securitization......................... 57 Table 9: Legal obstacles to implementation of S M E receivables securitization .......................... 59 List of Figures Figure2: Peru- Total Credit to the Private Sector .......................................................................... Figure 1: Domestic public bondsby region. 1990-2005 (averageoutstanding. % of GDP)..........8 9 Figure3: Peru- Domestic Credit to the Private Sector ................................................................. Figure4: Peru.the supply-side of Capital Markets..................................................................... 10 11 Figure5: Peru-the demand-sideof capital markets .................................................................... 12 Figure6: Perufinancial systemdevelopment................................................................................ 15 Figure 7: Days to Process an SME Loan ...................................................................................... 17 Figure8: Fee for SME Loan, Relative to GDP Per Capita........................................................... 17 Figure 9.MinimumSME LoanAmount, Relative to GDP Per Capita ........................................ 18 Figure 10:Number of Locations where Sh4E Loan Application canbe Submitted 18 Figure 11: Development of LoanPortfolios, Volume. 2002 to 2007 ........................................... ..................... Figure 12: Development of LoanPortfolios, Borrowers. 2002 to 2007....................................... 19 .............................. 20 Figure 14: Access to financial services......................................................................................... Figure 13: MarketShare andMarketSize inMicro Lending2001 and 2007 21 22 Figure 15: Medianline of credit or loan size (inUS$ 1.000)...................................................... 23 Figure 16: Average Maturity of Line of Credit or Loan (inmonths) 24 Figure 17: Percentageof lines of credit or loans with short vs 24 Figure 18: Percentageof Firmsthat Applied for Credit inthe Last Year..................................... .long-term ........................................... maturity .................... 25 27 Figure 20: Percentageof Loan Applications that Were Rejected................................................. Figure 19: Percentageof Firmsthat are Credit Constrained or Unconstrained............................ 28 Figure 21: Sources of Fundingfor Working Capital (%) ............................................................. 28 Figure22: Percentageof Firmsthat PurchasedFixedAssets duringthe Last Year..................... 29 Figure23: Sources of Fundingfor FixedAssets (9%) ................................................................... 29 30 Figure 25: Loan amount and collateral requirement..................................................................... Figure24: Start up funding........................................................................................................... 30 ii Figure 26: Loan maturity ............................................................................................................... 31 32 Figure 28: Access to finance ......................................................................................................... Figure 27: Percentage of Firms that Purchased FixedAssets Last Year ...................................... 33 Figure 29: Small firms relationship with nonbank financial institutions..................................... 33 Figure30: Average and Median Yearly Interest Rates (96).......................................................... 34 A 1 The Structure of the FIDC Zoomp.................................................................................... A 2 The functioning of TREFI................................................................................................... .. . . . 54 56 iii ACKNOWLEDGEMENTS This study was preparedby ateam ledby Rogelio Marchetti (LCSPF), comprising Juan Carlos Mendoza (LCSPF), Yira J. Mascaro (LCSPF), Thorsten Beck (DECRG), MiriamBruhn (DECRG),Ilias Skamnelos (LCSPF). A teamledby Gonzalo Tamayo andMonicaCorn from Macroconsult provided valuable assistance to the team at the initial stages of the study. The team would also like to thank Marcel0Giugale (LCSPR), Lily Chu (LCSPF), Carlos Silva-Jauregui (LCSPR), Oliver Fratzscher(CCGCM), Ellis Juan (IADB), Oscar Calvo-Gonzalez (LCSPE), Victor Hugo Orozco (LCSPF), RossanaPolastri (LCSPE), Patrick Conroy (FPDVP), Jorge Luis Archimbau (LCCPE), and Livia Benavides (LCSHE) for their helpfulcomments and suggestions. iv DEVELOPING STRUCTUREDFINANCIAL NEW PRODUCTSTO CHANNELSAVINGSTOWARDS SMALLAND MEDIUMENTERPRISES (SMES)GROWTH This study responds to a request from the Government of Peru to provide recommendations promoting the development o f new financial instruments that could channel resources from institutional investors, particularly pension funds, more effectively towards underserved segments of the economy whose ability to grow i s being constrained by lack of financing. In summary, the objective of the study is to contribute to the development of new structured financial products in Peru, in particular as means to address the problems of access tofinance faced by undersewed segments such as Small and Medium Enterprises (SMEs). V EXECUTIVE SUMMARY The demand for investment instruments inPeru has increased considerably inthe last decade. Domestically, reforms in the pension area have created large pools o f funds seeking new instruments, while the appetite o f international investors for Peruvian assets has grown, particularly for debt instruments. The Government's local currency public debt programs, the process of privatization of firms, and the rise in the number o f new companies, consequence o f the economic growth of Peru, have all contributed to the increase in the supply o f investable assets. However, Peru's domestic bond markets remain smaller than those o f most Latin American countries, but also i s one in which the private sector has played a more important role. The supply o f securities i s heavily concentrated in a few types of instruments and in few issuers. In addition, the demand for securities is heavily and increasingly dominated by a very limited number of institutional investors, particularly pension fund administrators (Administradoras de Fondos de Pension - AFP). Small and MediumEnterprises (SMEs) constitute 95% of all enterprises representing 62% o f Peru's labor force. Compared with other countries in the region, Peruvian banks fund a larger share o f SMEs working capital. Anecdotal evidence suggests that many banks employ effective risk assessment techniques and at the same time have developed cost effective products. This situation allows them to deal effectively with the opacity of SMEs, achieving profitability when providing short term working capital financing. However, banks in Latin America as well as in Peru have not yet found an effective way to accommodate the long term financing needs o f SMEs. Smaller firms inPeru are able to get larger short term loans or lines of credit than in other Latin American countries, but the average maturity o f these loans or lines o f credit i s shorter and the amount o f long-term financing smaller firms receive i s smaller when compared to similar companies inLatin America. Structured finance could offer a solution to the Peruvian SMEs dilemma - how to finance their long term capital needs - by creating a channel with institutional investors. Structured finance involves the pooling of assets and the subsequent sale to institutional investors o f claims (to cash flows) backedby these asset pools. This enables the conversion o f illiquid and relatively high-risk SME-related assets into tradable securities that have the creditworthiness required by such institutional investors as pension funds. Inother parts ofthe world andspecifically inLatin America several securitizationtechniques have been used for this purpose. For example, the pooling of bank loans to SMEs, invoices, leasing contracts, factoring and other schemes like multi-originator transactions, and reverse factoring have been explored and implemented. The World Bank, with FIRST Initiative funding, supported the study o f alternative ways to improve the provision of finance to SMEs in Peru. This resulted in two innovative structures: (i) a local currency multi-originator securitization of micro-credits by Cajas Municipales de Ahorro y Crgdito (CMAC) and; (ii) scheme for medium size enterprise debt issuance backed by a revolving factoring of receivables from small enterprises (TREFI model). The C M A C transaction i s expected to be in the market by October 2008. The development and launch vi process created significant know-how and several valuable lessons were learned. These are discussed inmore detail at the end of the paper. Government, regulators, and professional associations can catalyze the development of securitization by promoting the development of securitization markets. One way i s through providing public goods i.e. developing instruments, building up legal expertise as well as educating market participants and potential investors. The costs associated with these initiatives are usually borne by the initial issuers, but once the market has developed, new issuers will benefit substantially by issuingsecurities with similar characteristics. This document reviews how structured financial products can provide an alternative channel for SMEs to access the market. Additionally it provides suggestions regarding policy issues aimed at improving the market environment. The study concludes that: (i) one way to effectively address Peruvian SMEs credit constraints, particularly with respect to longer term financing, i s the use of structured products and (ii)although the Peruvian capital markets regulatory framework regarding securitization has become more flexible in recent years, and despite being one of the least restrictive inthe region, it still shows significant weaknesses. vii Chapter 1. THEROLEOFFINANCIALMARKETSINENSURING GROWTHSUSTAINABILITYINPERU 1.1 Latin American capital markets have evolved significantly during the past fifteen years. Although bond markets have grown, equity markets in general have languished. Inthe case o f Peru, the country has recently made numerous positive strides. However; there are many financial needs that remain unmet. Diversifyingsources of funding: the role of local capital markets 1.2 Funding is available in Latin American capital markets, but there is the need to - diversify sources. Following financial liberalization and other reforms during the Figure 1: Domestic public bonds by region, 1990- early 1990s, international investors' appetite 2005 (averageoutstanding, % of GDP) for Latin American risks increased the amount Public Sector Bondsto GDP (YO) of funding available to governments and firms 45.1 45.0 1 in the region. The funding materialized through fixed income instruments and to a lesser degree, equity. However, the Tequila and Russian crises highlighted the need for diversification of funding as international investors "flight to quality" during market disturbances as well as during the contraction 199M4 1995-99 2000-05 of business cycles led to liquidity drying up OECD rn East Asia 8 Pacific 0 Latin America 8 Caribbean rn Peru quickly. In addition, funding in foreign Sources: based on informationfrom World Federationof Exchanges, currency exposes borrowers to the risks Banks for International Settlements, Inter-American Development Bank, andThe WorldBank. associated with exchange rate disruptions. As a result, countries in the region have started to move towards more diversified local currency- based financial instruments. In the case o f Peru, the development o f domestic bond markets constitutes an essential component o f financial sector diversification and also a mechanism that contributes to macroeconomic stability. 1.3 Governments and the private sector expanded the pool of fixed income investable assets. The public sector has played an important role in developing domestic bond markets by creating a benchmark yield curve to allow for the pricing o f private-sector bonds. The World Bank has been particularly involved in Latin America in the development of such local currenc public bond markets through the provision of technical assistance to debt management offices . Y InLatin America, the average outstanding domestic public-sector bonds as a percentage of GDP rose to 22% during the period 2000-2005, after oscillating around 13%-15% during the previous ten years (See Figure 1). WorldBank,2007. 8 1.4 Furthermore, this process has been accompanied by several reforms that have been a fillip to the development of local capital markets. The improvement of market regulation and infrastructure; the privatization o f utilities and other public enterprises, expanded the pool of enterprises that could tap the capital markets. Pension systems reform created a natural constituency o f long-term investors. The enhancement of macroeconomic stability has improved the incentives for investment inlong-term, local currency securities. 1.5 Latin American equity markets have grown but are relatively smaller when compared with other regions. At the end of 2006, stock market capitalization inLatin America and the Caribbean represented 48.1% o f GDP compared to 140.4% inEast Asia and Pacific. The value traded in 2006 represented only 5.6% o f GDP in Latin America, compared to 88.8% in East Asia and Pacific. Other signs of underdevelopment in Latin American equity markets are the concentration of trading injust a few firms, the de-listings of firms and the migration of large companies to the stock marketsindeveloped countries. These trends have been more accentuated in Latin America than in East Asia, in part due to the financial and capital market reforms - introduced by Latin American countries during the 1990s - that helped governments and firms obtain financing in international capital markets. However, this was done at the expense of domestic market development. Depth of the Peruvian capital market 1.6 The economy has grown but markets remained shallow. The broadest measure of financial sector deepening generally used Figure 2: Peru- Total Credit to the Private Sector (total credit to the private sector as ." percentage of GDP) peaked in Peru in 60' A 1999 at 28% (see Figure. 2). From there on, credit contracted such that by 2006, 50- the credit to the private sector stood at 40- 17%2of GDP. In2006, both indicators o f credit reached their lowest point since 30- 1997 (domestic credit to the private sector 20- represented 18% o f GDP and domestic 10' credit provided by the banking sector represented 15% of GDP). Both were also 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 well below the average levels of credit in Source: Basedon informationfrom the World Bank. the region. 1.7 Duringthe 1999 - 2004 period, domestic credit to the private sector as a share of credit provided by banks remained relatively constant, at an average 71%. While the net credit portfolios continued to decrease as a share o f GDP,3 as financial institutions removed non- performing assets from their balance sheets, portfolio growth in all client segments has been The average credit to the private sector for Latin America and the Caribbean was 31% of GDP at the end of 2006 (World Bank Development Indicators). The initial contraction in credit was generated by unexpected shocks during 1997-1998 (the impact of El Niiio on fishing and agriculture, the deterioration in the terms of trade, and the international financial crises) which left the banking sector with an impaired portfolio. 9 strong during the last three years. As shown in Figure 3 total credit among supervised financial entities4i s growing rapidly innominal terms. 1.8 Causes behind the shallowness Figure 3: Peru- Domestic Credit to the Private Sector of the Peruvian financial sector. The purpose of this paper i s not to identify and propose solutions regarding the limited size of the financial sector in Peru. Nevertheless, this topic will be examined to help formulate an approach to the development o f structured financial products as part of a constellation of policy actions. Several hypotheses can be identified to explain these circumstances. 1.9 Informality exacerbates the problem of information asymmetry. ISource: Basedon information from the World Bank. As Polastri et.al. (2007) record, "about 60% of production is done informally, 40% of the labor force i s self-employed in informal micro-enterprises, and 80% o f the labor force does not contribute to a formal pension plan." Over 74% o f micro and small firms operate in the underground economy. Informality does not always hinder access to credit but it does prevent firms from enjoying many of the other benefits o f formality. These benefits include: enforceable and impersonal contracts, credible financial statements, access to capital markets and access to risk pooling mechanisms which ultimately does lead to lower financing costs (Levenson & Maloney 1998). 1.10 Non-bank financing is limited due to high transaction costs and regulatory constraints. Transactions costs associated with accessing other financing vehicles such as capital markets (see also chapter 4 "Receivables financial model") are exacerbated by the dual regulatory role of SBS (Insurance and Bank Superintendency) and CONASEV (Securities and Exchange Superintendency), which are in charge of regulating institutional investors and issuers, respectively. Regulatory constraints and the lack o f competition based on performance have prevented AFPs from migrating beyond blue chips. In fact, 75% o f the total number of corporate bonds, financial-sector bonds and securitizations traded in 2006 were rated AAA, and the remaining 25% had ratings that did not fall below A+. These constraints are discussed in more detail inthe next section. 1.11 Businesses owners in general are reluctant to approach the capital market. This i s based mainly on the fear of opening the business to third parties and the lack o f an entrepreneurial culture that accepts good corporate governance practices (accounting and control). Other related issues include the perception that the issuance process i s very slow and bureaucratic, and the fear o f being treated thereafter differently by the tax authorities resultingin adverse consequences. Supervised financial entities include banks, financial companies (empresas financieras), municipal and rural savings and loans institutions (cajas municipales and cajas rurales) and microfinance institutions (edpymes). 10 Recent evolutionof the Peruvian capital markets 1.12 The domestic bond market in Peru is smaller than the ones in most other Latin American countries. However, the private sector has played a more important role. Indeed, with average outstanding domestic public-sector bonds representing only 3.7% of GDP in 2000- 2005, and domestic private-sector bonds representing 4%, Peru lies at the bottom of the ranking o f Latin American countries in size o f its fixed income market (public plus private). On the other hand, Peru private-sector participation in domestic bond markets i s o f much greater relative importance than other countries in the region. The private sector bonds represented 33% of the total amount issued in 2002, and 62% in 2006.(See figure 4a and 4b). This has been in part driven by an increase in the issuance of bonds from the corporate utilities sector and by the regulatory changes that gave private pension fund managers (Administradoras de Fondos de Pensiones, AFPs) more flexibility regarding the composition of their investment portfolio. The domestic private sector accounted for more than 95% of the new bonds listed in 2006, the highest contribution of the domestic private sector observed inthe region. 1.13 The Peruvian bond - markets remains concentrated in a Figure 4: Peru the supply-side of CapitalMarkets a. Private-sector issuancesin2002103 and 2006107 by type of few players. Privatization and instrument (% of total issuance) pension fund reforms in the early 1990s led to an unprecedented rate --- o f expansion of capital markets in 2006-07 Peru but in a context o f high concentration of supply and demand o f securities. Privatized utility firms became an increasing and 200243 predominant group of borrowers and pension funds became an increasing and predominant group o f long-term 0 CorporateBonds E3 Leasing Bonds institutional investors (De la Torre 0 SecuritisationBonds 0 Short-termInstruments and Schmukler, 2006). d b. Private-sector issuancesin2002-2007 by economic sector (average % of total issuance) 1.14 The supply of securities is I Aaroindustrial I heavily concentrated in a few instruments issued by a small number of entities. The number of private-sector issuers has been declining steadily since 1998 (from 249 issuers in that year to 221 issuers by the end of 2006). Corporate bonds have taken an increasingly Predominant role in the 0% hydrocarbon market over the last few years. `L 12% While the number of corporate bonds issued represented 43% o f Sources: Basedon information from CONASEV and SBS. total issuances in 2002-03, their share increased to 63% in 2006-07. On the other hand, the growth of asset-backed securities, such as mortgage and securitization bonds, remains stagnant. 11 Moreover, the issuance of corporate bonds was highly concentrated among a few players, with only five companies being responsible for 77% o f the total corporate bond issuance in 2006. Utility companies and financial sector institutions accounted for more than 74% of total issuances between 2003 and 2007. 1.15 The equity markets in Peru have grown significantly but market capitalization is highly concentrated inthe shares of a few issuers. The capitalization of the stock market grew from virtually nothing at the end of 1980s to US$111 billion by September 2007, representing 1.1 times the GDP. The traded volume in Lima's stock exchange has steadily increased since 2003, reaching a historic record in 2007 (4.7% of GDP). Lima's stock market was the most profitable market in the world in 2006 and the second most profitable in the region in 2007. The capitalization o f Lima's stock market i s highly concentrated. The top 12 companies by market cap represent 74% o f the total capitalization o f the stock market and the market cap of the largest company (Southern) represents 28% o f the stock market. However, this concentration problem i s considered less significant than the one existing in Argentina, Mexico and Brazil. In 2006, there were 196 Peruvian companies with shares listed on the Lima Stock Exchange. This means that for every 10,000 small, medium and large sized enterprises, there were 53 listed domestic companies. This i s twice the ratio in Chile and more than twice the ratio in Mexico, Brazil and Argentina. 1.16 Institutional investors, specifically AFPs (Pension Fund Administrators), dominate the demand for securities. The portfolio securities held by institutional investors grew from 19% to 27% of GDP between 2004 and 2007, (reaching US$27 billion). AFPs have always been the main demand-side players in capital markets. The total amount of securities held by institutional investors grew from US$ 13 billion in 2004 to US$ 27 billion in 2007. During the same period the share o f AFPs in that total grew from 60% to 71%, whereas Figure5: Peru-the demand-sideof capitalmarkets banks' share has fallen, from 27% to 19%.5 Because there are only four AFPs and eleven commercial banks in Peru, Securitizations competition for securities i s very low. Moreover, 87% of the total securities portfolio held by institutional investors i s inthe hands of only six economic groups. These groups participate in three or more business segments (banks, AFPs, insurance companies and mutual funds) and have a presence in the stock Investment Fund brokerage and investment banking businesses. (See Figures 5 a,b and c) Insurance companies account for the remaining portion of the securities portfolio. Their share has remained fairly stable, at around 12% o f the total. 12 b. Investmentinsecurities in 2004-2007. bv tv :of c. Securities portfolio of AFPs in 2004 and 2007, by institutionalinvestor(% of GDPand billionU 1 " " type of issuer (% of total) 27% F 2007 2004 0% 20% 40% 60% 80% 100% 2w4 2w5 2WB 2w7 -Government UllFinancial Sector I I I A F P s (R) IIICommerclal Banks (R) 5!CorporateSector 0Mutual&Investmt.Funds OlnrurancaCompanies(R) +as%ol GDP(L) 0SecuritizationAgencies c;3Foreign Source: Based on information from SBS. 1.17 AFPs' portfolio composition is restricted by the investment regulatory regime. AFPs' investment preferences are heavily inclined toward equity and public debt. Holdings of public debt among AFPs tend to be high because (i) regulation imposes maximum limits on the amount o f equity, corporate bonds and foreign assets that can be held, and (ii) because there are limited investment grade opportunities inthe private sector. 1.18 The Peruvian banking system falls short of various levels predicted by international comparisons along several dimensions. The results show the actual levels o f several financial development indicators and the levels predicted when taking into account Peru's GDP per capita level, population size and density, and other country characteristiw6 (See Figure 6: Peru financial system development). 1.19 The evaluation compares the actual values o f various financial development indicators for Peru with predicted values, which come from broad cross-country, cross-year panel regressions. Rather than comparing Peru to specific countries, this exercise therefore compares Peru's financial system to a synthetic benchmark, constructed from a broad cross-section o f countries. 1.20 The size of the banking system and public bond market are below the predicted level. However, bank efficiency (as measured by the interest rate margin) and the liquidity o f the stock marketare below their respective predicted values. Overall, this benchmarking exercise suggests that Peru's financial system i s not realizing its potential. It i s important to note that this i s a For details, see Beck et al. (2008). 13 structural rather than cyclical phenomenon, as the results are the same across the past seven years. 1.21 Peru has made large improvements in credit information sharing over the past years. There are one public and two private credit bureaus providing information for all financial institutions, including non-bank financial institutions serving the lower end of the credit market. The SBS (Superintendencia de Banca, Seguros, y AFP) collects and provides both positive and negative information to the supervised institutions. The two private credit bureaus offer their clients additional information to the standardized SBS information, such as clients' history with non-financial firms, mergers, bankruptcies and so on. This relatively sound credit information sharing system - at least when measured by regional standards - i s also reflected in the Doing Business Indicators and rankings. Only a few other countries in the region rank similarly highinterms of the depth and efficiency of credit information sharing. 1.22 Despite the improvements in the credit information systems, coverage of the population by these registries has remained relatively low. Compared to countries with less efficient credit information but deeper financial systems, Peru's credit information sharing system has not achieved the same "outreach." Specifically, as of 2007, the public registries cover 21% of all individuals or firms, while the private registries covered 33%. This compares to 40% for Colombian private credit registries, 46% for Brazilian private credit registries and 61% for Mexican private credit registries. 1.23 International experience has shown that smaller enterprises are typically more adversely affected by deficiencies in the financial system and the underlying financial infrastructure. Financial deepening typically results inleveling the playing field between small and large enterprises and allows more small enterprises to gain access to formal financial service^.^ International comparisons have also shown that improvements in credit information sharing help smaller enterprises in particular to gain greater access to financial services (Love and Mylenko, 2003). 1.24 Demand exceeds supply of securities. The perception o f market participants i s that there i s an excess demand for securities resulting primarily from the combination of the two factors: 1) Rapid growth of private pension funds and 2) the reluctance o f companies to diversify their traditional sources of funding, intruding on the stock market. Market analysts observe that valuations o f certain financial instruments are difficult to justify. There i s thus the opportunity to redirect a significant amount o f funds to companies and projects that represent an alternative to the traditional funding mechanisms. See survey in Beck and Demirguc-Kunt(2006). 14 +Predicted Observed -sE (+ ,.) 60 0Predictedwlregionaleflects 50 40 30 30 20 20 10 10 0 2000 2001 2002 2003 2004 2005 2006 2000 2W1 2002 2003 2004 2005 2044 2000 2001 2002 2003 2004 2005 2044 E ~ ~ I T ~ ~ ~ Stock Markat 1urnovm Ratio Number of Lrsrad Firmi pfrQQ-23GBi "S of G t P :2oon.2oc5: / hiril ot Pacpin 0 Observed OObservad +Predicted +Predicted -0SPredisled E (r,-) 90 60 - wl regionaleflesla 80 70 60 50 40 30 20 10 0 2WO 2001 2002 2003 2004 2WS 2006 2000 200i 2002 2003 2004 2005 2006 2000 2001 2002 2003 2004 2005 2006 0 Observed +-SE(+,.) + lnciudes assets lo GDP of pensionfunds. insurance Predicted companies and mutual iunds - 0Pradioledw/ mgionsl dfeclr Note: The predictions are based on pooled, year-fined effects OLS regressions for the psrlod 2000-06with the logarithm of X as a dependent variable Far the predictionsthat take into account regional infiuences. separate regressions are used that include region-fixed effects The independent variables are the same for both estimationsand include the logarithmsof GDP per capita. population size and dansty, the value of fuel expolli to :u GDP, the povoity gap. tho productai GDP per capita and papulation size and a dummy that indicates whether the country IS an offshore financial center or not (see http l l w imf orgiensmawnplofcaloica asp) 2000 2001 2002 2003 2004 2005 2006 2000 2001 2002 2003 2004 2005 2006 Source. FSAP Benchmarking database.staff analysts 15 Chapter2. THESTATEOFSMEFINANCINGINPERU' 2.1 Peru seems to be at the forefront among Latin American countries in terms of the availability and use of short term financing by SMEs. A cross-country analysis shows that unlike other countries, small firms in Peru are more likely to obtain bank finance for working capital than in other countries in the region. However, they are less likely to get long-term financing. Demand for credit i s higher among small firms in Peru than in other countries in the region. Unlike other countries, Peruvian firms are more likely to receive a portion o f their start- up funding from a bank. Small firms in Peru obtain larger loans or lines of credit than in other countries. In spite of recent improvements by banks and other financial institutions aimed at catering specifically to small and micro-enterprises, the Peruvian financial system i s yet to realize its full potential. Our findings point to the lack o f long-term finance for fixed asset investment as a significant growth constraint for small and micro enterprises in Peru. Policies that can help alleviate this constraint and improve access to long-term financing will thus facilitate the growth of this important segment of the enterprise population and thus contribute overall to private sector and economic growth. Peru's financial system and the SME sector 2.2 Small enterprises constitute 95% of all enterprises and 62% of the total labor force inPeru. The total number of small enterprisesg, is estimated to be between 2.3 million and 3.2 million, more than 97% o f which are micro and 83% of which have fewer than 5 employees. While the estimate for small enterprises in urban areas varies from 1.1 to 1.35 million, the estimate for rural small enterprises varies between 1.2 and 1.9 million, so the exact number i s much more uncertain. Ingeneral, enterprises inrural areas are significantly smaller. 2.3 Around a quarter to a third of small enterprises are formal, while the remainder are informal, which makes estimating the exact number of enterprises difficult. Employees in small enterprises also dominate the labor force, with remuneration increasing with firm size. Data on the share o f small enterprises in total sales or value added i s not available. Also, 92% o f employees insmall enterprises work ininformal enterprises. 2.4 Interms of sector delineation, 60% of small enterprises are in mining, 19% in trade and 12% in services. 19% o f all small enterprises are located in Lima, with Cajamarca being the second largest province interms of small enterprises (8%). * Thorsten Beck and Miriam Bruhn, with research assistance from Carlos Espina, were significant contributors to this chapter. The two sources for statistics are: Direccidn Nacional de la Micro y Pequeiia Empresa (2007): Actualizacidn de Estadisticas de la Micro y Pequeiia Empresa and World Bank's Enterprise Survey 2006. Desdpite having slightly different definitions of micro, small and medium enterprises, both sources show similar trends on the issues discussed in this study. The first source defines Small and micro enterprises are defined according to the number of employees or sales turnover. Specifically, microenterprises are defined as enterprises with fewer than 10 employees, while small enterprises are those with 11 to 50 employees. In terms of sales volume, microenterprises are those with less than $170,000 in sales, while small enterprises are those with sales between $170,000 and $1.0 million. These thresholds are actually defined in terms of Unidades de Impuesto Tributario, which are updated every year. As of 2007, one UIT was 3.450 Soles. For World Bank's Enterprise Survey 2006 definitions, see footnote 11. 16 2.5 Peru's lending practices to SME compared with other countries inthe region shows a positive picture. Using data from a bank-level survey in 2004, we compared barriers o f physical access, eligibility and affordability (Beck, Demirguc-Kunt and Martinez Peria, 2008) see Figure 7-10)," Figure 7: Days to Processan SME Loan 45 ,i I Sample size: 55 countries Note: This figure shows the number of days it takes to process an SME loan application, according to the surveyed banks. The numberfor each country is the averageacross all banksinthe respectivecountry Source: Beck, Demirguc-KuntandMartinez Peria(2008). Figure 8: Fee for SME Loan, Relative to GDP Per Capita Sample size: 48 countries Note: This figure shows the fee for an SME loan, relative to GDP per capita, according to the surveyedbanks. The number for each country i s the average across all banks inthe respectivecountry. Source: Beck, Demirguc-Kuntand Martinez Peria (2008). loThe four banksincludedfrom Peruare: Bancode CrBdito,Interbank,BancoWiese Sudameris (now Scotiabank) andBBVA Banco Continental. 17 Figure9.MinimumSMELoanAmount. Relativeto GDPPer CaDita 10000 1000 100 10 1 0.1 0.01 0.001 0.0001 Sample size: 51 countries Note: This figure shows the minimum loan amount for SME loans, relative to GDP per capita, according to the surveyed banks. The number for each country is the average across all banks inthe respective country. Source: Beck, Demirguc-Kuntand Martinez Peria (2008). Figure 10: Numberof Locationswhere SMELoanApplication canbeSubmitted (l=headquarter, 2=branch, 3= branch-like office, 4=phone, 5=Internet) Y Sample size: 55 countries Note: This figure shows the locations where it i s possible to submit an SME loan application, according to the surveyed banks. 1 indicates that it i s possible only at headquarters, 2 at a branch, 3 at a branch-like office, 4 by phone and 5 by Internet. The number for each country i s the average across all banks inthe respective country. Source: Beck, Demirguc-Kuntand Martinez Peria (2008). 2.6 Peru compares very favorably, it has one of the lowest minimum balance and loan fees environments in relation to GDP per capita in our sample of 55 countries. Peruvian banks also seem to process loan applications rather quickly, in less than 4 days, similar to Brazil and faster 18 than in Colombia (8 days) or Mexico (10 days). However, Peru has not entirely embraced the opportunities offered by modern technology, such as phone and Internet banking - loan applicants still have to personally appear at a branch or branch-like office, unlike in Brazil, Chile, Colombia and Mexico, where banks use alternative channels to receive loan applications, such by phone or the Internet. 2.7 The micro enterprise segment of the financial system's aggregate loan portfolio has grown more than the other segments in the past six years. SBS data allow for a detailed analysis of lending trends across different size segments over the period March 2002 to December 2007. As of December 2007, total outstanding loans by supervised financial institutions to larger enterprises were US$ 3.6 billion (or 27% of total lending), while loans to medium-sized enterprises amounted to a total o f US$ 3.2 billion (23%). Lending to small and micro enterprises amounted US$ 4.0 billion (29%) and US$ 2.8 billion (21%), respectively (Figure 11). Figure 11:Development of Loan Portfolios, Volume, 2002 to 2007 4,000,000,000 3,500,000,000 3,000,000,000 2,500,000,000 2,000,000,000 1,500,000,000 Note: This graph shows the volume of loans by supervised financial institutions to large, medium size, small and micro enterprises over the period March 2002to December 2007. Source: Own calculations basedon data by SBS. 2.8 Behind the volume statistics, there were 3,800 large borrowers, 4,900 medium-sized borrowers, 21,200 small borrowers and 12,900 micro borrowers (Figure 12). We found that the micro segment has grown fastest in terms o f volume over the past six years, with annualized growth of 25.1%, followed by the small enterprise segment with 23.1%. In terms of number o f borrowers, however, the micro lending segment increased by 484 firms (3.9%) per year, behind the larger enterprise segment whose number grew in 800 firms (26.6%) and the number o f small enterprise segment increased by 2,355 firms (12.5%). 19 Figure 12: Development of Loan Portfolios, Borrowers, 2002 to 2007 25,000.00 In -t Large 20,000.00 15,000.00 10,000.00 5,000.00 Note: This graph shows the number of loans by supervisedfinancial institutions to large, medium size, small and micro enterprisesover the periodMarch2002 to December 2007. Source: Own calculations basedon databy SBS. 2.9 The average loan size o f micro loans, on the other hand has more than doubled, while the average loan size of large enterprises actually decreased. Without borrower-level data it i s impossible to differentiate between different hypotheses that can explain these trends. Absent a frequent turnover o f clients in the micro enterprise segment, however, this seems to suggest that existing micro and small borrowers increased their exposure to the financial system, with few new borrowers getting access. This would be consistent with the "graduation theory" micro clients graduating into larger loans, even if remaining micro 2.10 The banking system dominates the lending market for small and micro enterprises, but is losing ground to other regulated non-bank financial institutions. Its market share has been continuously declining over the past seven years. Specifically, while inJanuary 2001, banks still had 67% market share among micro lending, this share dropped to 51% by the end of 2007 (See Figure 13). 2.11 The banks leading in small enterprise financing are MiBanco (with 52% of the microcredit market), Banco de Trabajo (19%) and Banco de Crkdito (14%). Other banks are pushing into this segment as well, including foreign banks, such as Scotiabank, which ranked fourth in micro enterprise lending as of December 2007. However, there are several other institutions that provide credit services to small and micro enterprises, though with different degrees of efficiency and soundness. 20 Figure 13: Market Share and Mark1 Size inMicro Lending 2001and 2007 Jan 2001 .Total Market Size 1.9 Billion Soles Dec 2007. Total Market Size 7.2 Billion Soles u 67% 0% 0 Banks 0 Financieras 0 Cajas Municipales Banks 0 Financieras nCajas Municipales Q Caias Rurales Edpyrnes Caias RuraIes rn EdDvrneS Source: Own calculationsbasedondata by SBS. 2.12 The Cajas Municipales de ahorro y credito (CMACs)are the main group inthe non- bank financial sector, with rapid loan portfolio and borrower growth over the past years. Their market share in the microcredit market was 30% as of December 2007 and has almost doubled since 2001. While originally limited to the respective municipality, three o f these institutions have been allowed to enter the competitive Lima banking market. CMACs are owned and runby local governments, which might present governance challenges and the risk o f political capture and might impede the process of market-drivenconsolidation. 2.13 Inspite of their name, CajasRurales de ahorro y credito are private companies that lend mostly in urban areas, with a focus on consumer, commercial and mortgage lending, with little agricultural lending. Their share in microlending has remained constant at 6% over the past seven years. They are in stiff competition with the CMACs who likewise face serious governance and management challenges. 2.14 EDPYMEs (Entidades de desarrollo de la pequefia y microempresa) - created in the 1990s to facilitate the formalization of informal lenders - are the third most important group in providing small enterprise credit. Unlike the other institutions, they are not allowed to take deposits, and thus most of their funding comes from donors. Their share in microlending has increased from 5% in 2001 to 13% by the end of 2007. A regional comparisonof SMEs access to financial services 2.15 Regional comparisons show that small (less than 20 employees) and medium size firms (20 - 100 employees) in Peru have more access to financing than i s the case in other countries. However, they have less access to longer-term investment financing and therefore invest less in fixed assets than small firms in the comparator countries. Loans for working capital are widely available in Peru. However, interest rates prevent a larger fraction of small and medium size firms from applying for loans than inother countries. Loan maturities for small and medium size firms are shorter vs. other countries inthe region. 21 2.16 These findings are consistent with the accounts of local small firms and associations o f small firms. Section 3 o f this chapter draws more detailed parallels between local accounts and the statistics o f a more detailed survey of small enterprises inPeru." 2.17 Most mediumsize firms inPeru currently access financial services, but access is less widespread among small firms. Similarly to Argentina, Colombia, and Chile, almost 100% o f medium size and large firms in Peru have a checking or savings account, a number that drops to 90% among small firms (Figure 14a). The majority of Peruvian firms also have overdraft facilities and lines of credit/loans. Here again, the numbers are lower for small firms, 60% of them have an overdraft facility and 60% have a line o f credit or loan, as opposed to over 80% and over 70% for medium size and large firms, respectively. Two facts stand out in the cross- country comparison. First, medium size firms in Peru are more likely to have overdraft facilities or lines of creditnoans than in other countries. Second, although a slightly lower fraction of small firms in Peru has overdraft facilities, a higher fraction of small firms has loans than in most other countries (See Figures 14b and 14c). Figure 14: Access to financial services a- Percentageof Firmswith a Checkingor Savings Account 99 98 98 98 98 99100 I00100 100 90 80 70 60 50 40 30 20 10 0 Peru Argentina Chile Colombia Mexico 11 This section describes the use of financial services in Peru as compared to four other countries in the region, Argentina, Chile, Colombia, and Mexico. It also seeks to identify areas where access to finance is lacking and constraining firm growth. All statistics come from the World Bank's Enterprise Survey 2006, which includes 628 Peruvian firms. The section pays particular attention to distinguishing the financial situation of small, medium size and large firms. Firmsize definitions in the Enterprise Surveys are as follows: Small = up to 19 employees, Medium = 20 to 99 employees, Large = over 100 employees. Employees here are calculated as "permanent equivalent employees" according to the following formula: permanent employees + temporary employees*(average length of employment per year in months A2). The Peruvian firm size definition for small firms differs from the one in the Enterprise Surveys, but the Enterprise Survey sample is only representative within its firm size definition. The next section addresses the Peruviancontext and firm size definition in more depth. 22 b-Percentage of Firmswith Overdraft Facilities Ic Percentage of Firms with Line of Credit or Loans - 97 99 io0 90 87 80 70 60 50 40 30 20 i o 0 I Peru Argentina Chile Colombia Mexico Peru Argentina Chile Colombia Mexico 'aSmall El Medium large 1 laSmall 0 Medium Larqe 1 Source: Enterprise Surveys 2006 portfolios has been driven by larger ,,ooo 943 loans to existing borrowers rather than an increase in the borrower 7oo universe. Interestingly, the loan size numbers vary significantly between 4oo demand and supply side analysis, which i s partly due to (i) different firm size definitions and (ii)the survey nature o f the demand-side Peru Argentina Chile Colombia Mexico data. Medium size firms also receive Small 0 Medium 0 Large substantially larger loans or lines Of Source: Enterprise Surveys 2006 butlarge firms receive smaller loans. 2.19 The average maturity of loans or lines of credit in Peru is shorter than in other countries. Although the previous point illustrated that the loan or line of credit size i s greater for small and medimm size firms in Peru than in other countries, the opposite i s true for maturity. In Argentina, Chile, and Colombia, average loan or line o f credit maturity for SMEs i s at least six months longer than inPem13.Within Peru, small firms receive loans or lines o f credit for only 17 12 Almost all lines of credit or loans in Peru come from private commercial banks. In other countries, state-owned banks also play a role, particularly for small and medium size firms, but this is not the case inPeru. l3Note that one has to be careful when comparing loan characteristics across countries. This data i s only available for firms who have a loan and, as shown in Figure 9c, the fraction of firms with a loan varies greatly from country to country. This could imply that for a country like Mexico, where only a small fraction of firms has a loan, firms that have a loan are better firms since banks only lend to the best firms. We would thus be comparing loan maturity for good firms with loan maturity for a much larger pool of firms, including less sound firms. To address this concern, the analysis was also done controlling for firm's age, total annual sales, export share of sales, and a variable 23 months on average, while medium firms receive them for 21 months and large firms for 32 months (Figure 16). compared to firms of the same size 53 in other countries. Looking at the 50 ! fraction of loans that have a maturity o f up 1 year, 1-3 years, and over 3 40- year reveals that only 5% of loans to small firms and only 11% of loans to medium size firms in Peru have a maturity o f over 3 years. In other countries, up to 32% of loans to small Peru Argentina Chile Colombia Mexico firms and up to 53% of loans to medium size firms have a maturity of W Small 61 Medium 0 Large over 3 years (Figure 17). Source Enterpnse Surveys 2006 Figure 17: Percentage of lines of credit or loans with short vs. long-term maturity I 100 90 80 70 60 50 40 30 20 10 0 `USmall: Up to 1 year 0 Medium:Upto 1year Large: Up to 1 year Small: 1-3 years ~1 Medium: 1-3 years Large: 1;3 years Small: More than 3 years E3 Medium: More than 3 years Large: More than 3 years I Source: Enterprise Surveys2006. 2.21 The demand for credit is higher among small firms in Peru than in other countries. Duringthe last year, 48% of Peru's small firms applied for aloan, which is at least 18percentage points higher than in Argentina, Chile and Mexico, and 8% more than in Colombia. Medium size and large firms are even more likely to apply for a loan (58 and 63%, respectively), but these numbers are similar inthe other countries, except for Mexico (See Figure 18). indicating whether the firm owns the land it operates on. Controlling for these variables changes average loan maturities only little, suggesting that the selection of who receives loans and who doesn't does not affect the results here. 24 2.22 A possible explanation for why Figure 18: Percentageof Firms that Appliedfor Credit in more small firms in Peru applied for a , the LastYear loan than occurred in other countries 90 82 could be that banks are more accessible 80 n 70 in Peru than in other countries. For 60 example, Peru could have more bank 50 branches per capita or per square 40 kilometer than other countries. Table 1, 30 however, shows that i s not the case. 2a 10 0 2.23 In fact, Peru has the lowest Peru Argentina Chile Colombia Mexico number of branches per 1,000 sq km and a Small I(Medium 0 Large 1 per 100,000 peopie out o f all the I A Source: Enterprise Surveys 2006 countries in this comparison. It i s important to emphasize that the branch data i s only for commercial banks and outlets from other significant participants in the financial sector, such as Cajas Municipales, are not included in these statistics. Furthermore, it does not include corresponding banking agencies which have developed very quickly in the last two years. It could be that small firms in Peru apply and obtain more loans from non-bank financial institutions, because such institutions are more widespread inPeru than inother countries. Table 1.Bank BranchesinInternational Comparison Colombia Source: Beck, Demirguc-KuntandMartinezPeria(2007) 2.24 Finally, the fact that more small firms in Peru apply for loans could be explained by the presence of more favorable SME lending practices. Section 1 of this chapter shows that it takes fewer days to process an SME loan in Peru than in the other countries and that fees for SME loans are relatively low, making bank financing more accessible. Anecdotal evidence suggest that many banks in Peru apparently are very interested in catering the financing needs o f SMEs. Anecdotal evidence suggests that many banks employ effective risk assessment techniques and at the same time have developed cost effective products. This situation allows them to deal effectively with the opacity of SMEs, achieving profitability when providing short term working capital financing. l4 14This confirms the findings of A. de la Torre, M.Martinez Periaand S. Schmukler (2008)Bank Involvementwith SMEs:BeyondRelationshipLending" World Bank. 25 2.25 Unfavorable interest rates are one major reason why when compared to the experience of other countries; many small firms in Peru do not apply for loans. Table 2 shows that small firms in Peru are less likely than small firms in other countries to mention that they do not need a loan as the reason for not applying. Instead, 31% o f Peruvian small firms report unfavorable interest rates as the reason why they did not apply for a loan. Small Peruvian firms also mention that they did not apply for a loan because they did not believe their application would be approved (11.3% of small firms). This reason was not at all common in other countries. Unfavorable interest rates are also an important concern for medium size firms inPeru. For a comparison on interest rates, please see section 3.6 inthis chapter. Table 2. Reasoi for not ai Peru Mexico Smallfirms No need for loan 50.6 53.9 77.1 Applicationprocedures are complex 1.8 6.3 5.5 Interest rates not favorable 31.4 21.9 8.8 Collateral requirements are unattainable 1.3 2.9 4.4 Size of loan and maturity are insuficient 0.4 1.6 0.6 Didnot think it would be approved 11.3 1.9 2.9 Other 3.1 11.6 0.6 5.7 0.8 Mediumsize firms No need for loan 77.4 50.5 89.1 39.4 86.9 Application procedures are complex 0.9 1.6 5.9 1.3 1.5 Interest rates not favorable 15.2 25.5 0.8 5.3 4.9 Collateral requirements are unattainable 3.2 0.5 0.1 42.6 3.9 Size of loan and maturity are insuficient 0.2 1.7 0.0 0.5 0.2 Didnot think it would be approved 0.2 3.9 0.8 4.2 0.0 Other 3.O 16.3 3.4 6.5 2.6 Largefirms No need for loan 56.4 65.2 86.1 92.4 91.6 Application procedures are complex 0.0 4.7 1.o 0.0 0.7 Interest rates not favorable 0.4 10.7 0.3 3.0 3.5 Collateral requirements are unattainable 21.0 2.8 0.6 0.0 1.o Size of loan and maturity are insuficient 0.0 0.4 0.0 0.0 1.1 IDidnot think it would be approved 0.0 4.4 0.0 0.0 0.1 Other 22.2 11.9 12.1 4.6 2.0 Source: Enternrise Survevs 2006. Note: Some of the percentagesare based on very few firms giving this particular reasonfor not applying for a loan and should thus be interpreted with caution. For example, only 4 out of 107 medium size firms in Colombia mention the reason that "Collateral requirements are unattainable." However, the Enterprise Survey statistics need to be weighted according to the sampling weights to make them nationally representative. This increases the percentage from an un-weighted 3.7% to a weighted 42.6%. In the un- weighted data, 69.2% of medium size Colombian firms say that they had no need for a loan. For most of the other statistics in this table, un-weighted and weighted numbers are not very far apart. The other notable exception is large Peruvian firms that give the reasonthat "Collateral requirements are unattainable." The un-weightedpercentagehere i s 3% as opposedto 21% weighted. Inthe un- weighted data, 81.8% of large Peruvian firms say that they had no need for a loan. 2.26 A smaller fraction of small firms is credit-constrained in Peru than in other countries. Following Bingsten et al. (2000), the information on whether or not a firm applied for a loan successfully and if not why not can be used to classify firms into three categories: unconstrained with credit15, unconstrained without credit16, and ~onstrained.'~ With 43.5%, Peru l5Category includes firms that applied and received credit. 26 has by far the most unconstrained small firms with credit (in second place i s Colombia with 35.8%). The number of constrained small firms i s slightly lower in Peru than in other countries (12.1%). While in other countries medium size firms tend to be less constrained than small firms and more o f them tend to have credit, the situation for medium size firms in Peru i s similar to that of the small firms in these other countries. About 21% of large firms in Peru are credit constrained, at least three times more than inother countries (Figure 19). Figure 19: Percentage of Firms that are Credit Constrained or Unconstrained S M L S M L S M L S M L S M L Peru Argentina Chile Colombia Mexico 0 Small: Unconstrainedwicredit 0 Medium: Unconstrainedwkredit 0 Large:Unconstrained wicredit 0 Small. Unconstrainedw/o credit 0 Medium: Unconstrainedwio credit 0 Large:Unconstrained w/o credit Small: Constrained Medium: Constrained 0 Large:Constrained h3 Small: Not classifiable Medium: Not classifiable 0 Large:Not classifiable ource Enterpnse Surveys 2006 2.27 In Peru, fewer loan applications are rejected than in other countries, (Figure 20). For small firms in Peru and in other countries, the main reason for rejection of loan applications i s unacceptable collateral or cosigners. Insufficient profitability i s another important reason why small Peruvian and also medium size Peruvian firms have loan applications rejected, but this is rarely a concern for firms of any size inother countries. 2.28 Compared to other countries, Peruvian banks fund a larger share of working capital. About 29% of the funding for the working capital of Peruvian firms comes from banks, making it the second most important source after internal funds (51%). In other countries, all firms, but particularly small firms, rely more on internal funds or supplier credit to fund working capital, and banks play a smaller role. For small firms in other countries, banks are only the third or fourth most important source o f funding (Figure 21). l6 Category includes firms that did not apply because they did not need credit or because the interest rate was not favorable. l7Category includes firms that either applied and had at least one application rejected or that did not apply because the application procedures were too complex, collateral requirement were unattainable, size and maturity were insufficient, or did not think they would be approved. 27 Figure 20: Percentaee of Loan Amlications that Were Reiected - 35 30 30 25 25 24 20 15 10 5 0 Peru Argentina Chile Colombia Mexico lm Small EIl Medium [3Large 1 Source: Enterprise Surveys 2006. Figure 21: Sources of Funding for Working Capital (%) 100 90 80 70 60 50 40 30 20 10 0 S M L S M L S M L S M L S M L Peru Argentina Chile Colombia Mexrco - 0 Small Internal OSmall Banks Q Medium Internal 0 Medium Banks 0 Large Internal OLarge Banks ~ Small Suppliers or clients 0 Medium Suppliers or clients 0 Large Suppliers or clients OSmall Other cf Medium Other 0 Large Others 1 Note "Other" includes non-bank financial institutions,fmly/fnends, informalsources, and other Source EnterpnseSurveys 2006 2.29 A smaller fraction of small Peruvian firms purchased fixed assets during the last year than in other countries. Medium size and large firms are approximately as likely to purchase fixed assets as in other countries, but small firms in Peru are less likely to do so than firms inall other countries considered here, except for Mexico (Figure 22). 28 Figure 22: Percentage of Firmsthat Purchased Fixed Assets during the Last Year 100 94 90 88 86 85 80 70 60 50 40 30 20 10 0 Peru Argentina Chile Colombia Mexico Source: Enterprise Surveys 2006. 2.30 In Chile, Colombia and Mexico, small firms finance a larger fraction of investment in fixed assets with bank funding compared to their financing o f working capital. In Peru, the issuance of new equity and/or debt i s a source of financing for only 1% o f medium size firms and not at all for small and large firms. InArgentina, Chile, and Colombia, this source of financing i s slightly more wide spread, particularly for medium size and large size firms, providing up to 5% o f funds for investment infixed assets. Figure 23: Sources of Funding for Fixed Assets (%) S M L S M L S M L S M L S M L Peru Argentina Chile Colombia Memo Small Internal Medium Internal 0Large Internal OSmall Banks Medium Banks oLarge Banks 0 Small Suppliers or clients Medium Suppliers or clients 0Large Suppliers or clients Small New equityor debt LB Medium New equityor debt E3 Large New equityor debt OSmall Other Medium Other DLarge Other Note "Other" includes non-bank financial mstltutions, famlyifnends, informal sources, and other Source Enterprise Surveys 2006 2.3 1 Peruvian firms are more likely than firms in other countries to receive part of their start-up funding from a bank. Figure 24a shows that around 19% of small Peruvian firms receive some start-up funding from a bank. The number i s still higher than in all four of the other countries. The percentages for medium and large firms are anecdotal. The question referred to the time when the firm was a start up. At that time the firm was probable not a medium or large firm. The survey does not indicate whether the start-up funding obtained from banks goes towards working capital or fixed assets. 29 a. Percentage of Firms that Received Part of their b. Average Percentage of Start-up Funding Start-up Funding from a Bank I funding) Received from a Bank (for firms that received 26 24 40 38 30 20 10 .,n Peru Argentina Chile Colombia Mexico Peru Argentina Chile Colombia Mexico Source: EnterpriseSurveys 2006. SMEaccessto financialservicesinPeru1' 2.32 Overall, this section confirms the findings of the previous section that most credit i s short-term and for relatively small loan amounts, making long-term investment in large fixed assets difficult. a. Average loanamount (US$) b. Percentof loans requiring collateral 20,000 15,000 10,000 5,000 0 Source World Bank FirmProductivity Survey 2008 2.33 Loan amounts are quite small, especially for micro firms, making large investments in fixed assets difficult. The average loan amount is more than twice as big for small firms as for micro firms, presumably due to the fact that most loans require collateral (Figures 25a and 25b). Averaging across cities and firm sizes, loans used exclusively for investment infixed assets have a typical size of about US$lO,OOO. These small loan sizes are in contrast to the recent increases in average loan size reported in section 1of this chapter and to the higher average loan sizes cited in the Enterprise Survey. Working capital loans are even smaller (US$5,500) and insufficient for firms that want to make large investments inmachinery or other technology. This section makes use of the 2008 World Bank's FirmProductivity Survey for Peru (Encuesta de Productividad de Empresas), which includes 802 Peruvian firms under the size of 50 employees. Apart from the exclusive focus on small and micro firms, this survey also differs from the Enterprise Survey in that it covers informal firms. Another important feature of the Productivity Survey i s that it i s tailored more towards the local situation, asking about financing from different Peruvian lenders, such as Cajns Municipales and EDPYMEs. In addition this section draws on the conclusions of "Peru: trajectories towards formality" May 2008 unpublishedWorld Bank paper. 30 Box 1.Difficulty insecuringfinancing for fixed asset investments. For example, a local consortium of firms mentioned that they were inneed of a lab that would certify their products for export, but the cost of the lab was too high (US$850,000) to be financed with a bank loan. Similarly, a local association of firms mentioned that some firm owners had visited trade fairs abroad to source new machinery. However, realistically they can afford only 2"dgeneration machinery due to the lack of sufficient funding for state-of-the-art technology and were thus only "window shopping" for these latest machines. They mentioned that they had approached various leasing companies, but these companies would lend only to firms with US$1,500,000 in sales, and the firm seeking financing had sales ranging between US$500,000 and US$l,OOO,OOO. The opportunities given in these examples seem to be possibilities for an angel investment or venture capital investment however; these types of funding are still intheir early stages inPeru. 2.34 Average loan maturity ranges between one and one and a half years and less than 3% of firms secure loans with a maturity over 3 years. Loan maturity tends to be rather short for both micro and small firms located inLima, as well as in other cities (Figures 20a and 20b). This is another potential deterrent to long-term investment in fixed assets. A manager from the micro and small firm division of a major commercial bank explained that banks tend to give short maturity loans to micro and small firms since these firms often have lower survival rates than large firms and thus often disappear after a year. a. Average LoanMaturity (months) b. Percentage of Loans with Short vs. Long-termMaturity 100 17 I 90 80 70 60 Lima Other cities 0 Lima Other cities I Source: World Bank FirmProductivity Survey 2008. 2.35 CajasMunicipales are an important source of funding for micro firms incities other than Lima. Most firms finance start-up and working capital through internal or personal funds. Banks provide up to 15% of start-up funding and up to 6% o f working capital funding. Consistent with the supply data discussed in section 1, Cujas Municipales are the second-most important supplier o f bank credit with supply with 5.6% and 2.6% of start-up capital in cities other than Lima (Table 3). EDPYMEs and Microfinancierus are the third most important provider o f bank credit for start-up financing, but still more important than informal money lenders. 31 Table 3. Sourcesof funding(%) Lima Other cities Internal Personal savings 66.6 64.0 61.1 64.8 9.8 3.7 6.3 4.9 Family/friends 14.5 13.7 13.3 9.2 1.2 0.9 1.3 0.4 Bank 9.8 14.9 11.0 8.8 3.9 5.4 5.8 1.8 Caja Municipal 0.0 0.5 5.6 2.6 0.0 0.0 2.8 1.8 Edpyme or Microfinanciera 0.7 0.6 1.5 1.5 0.6 0.2 0.9 1.6 Money lender 1.3 1.2 0.4 0.7 1.3 0.0 0.7 0.0 Suppliers or clients 3.O 3.6 3.8 6.2 8.6 11.5 9.3 ~ Other 4.0 1.8 3.4 8.6 Total I 100 100 100 100 Observations I 253 109 363 66 2.36 Only about 30% of micro firms Figure27: Percentageof Firmsthat PurchasedFixed purchased fixed asset during the last year AssetsLastYear (Figure 27), but twice as many small firms 64 did so in Lima and 46% of small firms in other cities. 2.37 These purchases are also largely financed with internal or personal funds, but up to 15% of the financing comes from banks Lima Other cities and 8% comes from Cajas Municipales for micro firms in cities other than Lima (Table 4). These findings are consistent with the analysis o f supply data discussed above. Note: Investmentin fixed assets includesequipment, machinery, parts andtools. Source: World Bank FirmProductivitySurvey 2008 Table 4. Sourcesof fundingfor investmentinfiiedassets(%) Othe cities Micro Small Internal 57.0 66.8 Personal savings 13.2 10.6 Family/friends 3.5 0.0 Bank 15.2 10.5 13.0 13.5 Caja Municipal 0.7 0.0 8.3 3.1 Edpyme or Microfinanciera 0.0 0.2 0.5 1.6 Money lender 0.0 0.0 0.0 0.0 Suppliers or clients 2.3 0.7 1.4 2.3 Other I 2.9 I 0.0 3.1 2.1 ITotal I 100 100 100 100 Observations I 253 II 109 363 66 Source: World Bank Firm ProductivitySurvey 2008 32 2.38 Micro firms perceive access to finance to be a greater obstacle than do small firms and this obstacle for micro firms is even greater in cities other than Lima. About 38% of micro firms in Lima see both loan requirements and the cost o f finance as a major obstacle. These numbers are even higher in other cities (about 45%). For small firms, loan requirements and the cost o f finance constitute less of an obstacle, but they are still a major concern for at least 24% of small firms (Figures 28a and 28b). a - Percentage of Firms that See Loan Requirements b - Percentage of Firms that See Cost of Finance (e.g. collateral) as an Obstacle (e.g. interest rates) as an Obstacle 100 EO 60 40 20 0 Lima Other cities Lima Other cities 0 Micro No obstacle Small No obstacle 13Micro Minoror moderate0Small Minor or moderate rn Micro Major 0 Small Major Source World Bank FirmProductivity Survey 2008 2.39 Micro firms are about half as likely as small firms to have a bank account and they are also less likely to have a loan, particularly inLima. In all cities, about two thirds of small firms have an account at either a bank, Caja Municipal or Microfinanciera (Figure 29a). However, in Lima only one third of micro firms have such an account. In other cities, a slightly higher fraction o f micro firms have an account (40%), reflecting the presence and importance of Cajas Municipales in these cities. Cajas Municipales, as well as Microfinancieras and EDPYMEs also account for a significant fraction of loans in cities other than Lima. This contributes to the fact that, in cities other than Lima, over half of the firms received a loan during the last year (Figure 29b). In Lima, half o f the small firms received a loan and a comparatively smaller fraction (38%) o f micro firms received loans. About 45% o f all loans were used exclusively to fund investment in fixed assets, another 27% fund working capital only, and the remaining 28% were used for a combination of the two or for other purposes. a - Percentage of Firms with Account at a Bank, Caja b - Percentage of Firms that Received a Loan Municipalor Microfinanciera during Last Year 60 40 20 0 Lima Other cities Lima Other cities I Source: World Bank FirmProductivity Survey 2008 33 2.40 Interest rates for micro and small firms are high compared to interest rates for larger borrowers. The Enterprise Survey data revealed that many small Peruvian firms do not apply for a loan due to high interest rates. The Productivity Survey reveals that average interest rates are above 50% per year in Lima for both micro and small firms. In other cities average interest rates are around 40% per year. Median interest rates are lower - about 30% per year (Figure 30). Table 5 shows that average interest rates are higher at banks than at Cajas Municipales, Microfinancieras or EDPYMEs. Money lenders charge almost three times the interest rate of banks, but they are a much less common source of funding. The reported interest rates for banks match the small loan rates recorded by SBS. For comparison, average rates for commercial credit to firms with larger total loan positions in the system vary between 6.38% and micro and small firms pay to different financial institutions. 11.8%, depending on the length of the loan. Table 5 below19provides a comparison of the rates Table 5: Source of loan andaverage yearly interest rate (%) Lima Other cities Average yearly Micro Small Micro Small interest rate (%) Bank 84.4 88.9 44.1 67.5 48.2 Caja Municipal 4.2 0.0 36.6 10.0 36.1 Microfinanciera 1.o 0.0 6.9 7.5 37.0 EDPYME 2.1 1.9 7.4 7.5 25.8 Money lender 5.2 9.3 2.5 5.0 114.2 Other 3.1 0.0 2.5 2.5 44.5 Total 100 100 100 100 Observations 96 54 202 54 Source: World Bank FirmProductivitySurvey 2008 Figure 30: Average and MedianYearly Interest Rates (%) 60 50 40 30 20 10 0 Micro Lima Micro Othercities Small Lima Small Other cities Average Median Note: When monthly ratewas reported, interest rate was multiplied by 12 to get yearly rate Source: World Bank FirmProductivitySurvey 2008 lgThecategory other covers a wide variety of sources such as familylfriend loans with no interest rates as well as other non identified lenders. 34 Chapter3. SMEsANDSTRUCTUREDFINANCE IN PERSPECTIVE^' 3.1 This chapter reviews the international and Latin American experiences of SME involvement in structured finance. We found that Latin America presented interesting examples of innovative solutions that facilitated SME access to the capital markets. These solutions were implemented within market realities and legal structures very similar to the ones prevailing in Peru. This chapter summarizes many o f these examples to highlight the key issues relevant for development o f a structured finance approach to SME financing. SMEsaccessto financeinLatinAmerica 3.2 As shown in the previous chapter, SMEs have a significant role in Latin American economies. However, they frequently face difficulties in obtaining credit and especially long term financing from the formal financial sector, including commercial banks and nonbank financial institutions. Their financing of working capital and investment relies heavily on funds provided by owners, trade credit from suppliers, and credit from informal lenders. This type o f financing can be insufficient, constraining growth. Debt issued by SMEs would hardly achieve a credit rating that would appeal to institutional investors without credit enhancements (such as guarantees). In addition, most SMEs are not ready for the discipline, formality, governance, and disclosure required by capital markets. 3.3 Several Latin American countries have attempted to provide SMEs direct access to cap- ital markets as a way to enhance their financing opportunities. Brazil created the Novo Mercado (or small-cap market). The Chilean stock exchangebegantargeting SMEs in2001 by makingit easier for them to be publicly traded. Colombia lowered the minimumthreshold for listings to less than US$200,000 in assets. Mexico's 2005 securities law created a new legal form for enterprises that want to attract private equity, leading to a potential future listing. Its impact has yet to be measured. S M E access to equity markets has not fully materialized. After a promising start, stocks of Chilean SMEs were seldom traded, and the limited liquidity and thus exit opportunities led to losses and negative perceptions of the market. Listings in other markets have not yet gained momentum. 3.4 In summary, capital markets have provided limited funding to SMEs, despite significant attempts to facilitate their access. The fixed costs of equity and debt issues do not make it economically viable for them to raise capital directly through the capital markets. Due to their typical small issuance size institutional investors cannot justify the cost of analyzing them individually. However, the use of structured finance dimensions o f capital markets may facilitate indirect access by SMEs by lowering associated transaction costs. 3.5 There i s a nascent market for securitizing SME-related assets in Latin America. Securitizations of accounts receivable have become more common in various countries. In 2004- 2oThis chapter is based on Hela Chiekhrouhou, W. Britt Gwinner, John Poller, Emanuel Salinas, Sopphie Sirtaine and DimitriVittas; "Structured Finance in Latin America, ChannelingPensionFundsto Housing, Infrastructureand SmallBusiness." (2007)WorldBank. 35 05, there were more than 10 issues of structured bonds backed by accounts receivable in Brazil, and similar transactions have begun in Mexico, Peru, and Argentina. So far, other SME-related assets, such as loans and leasing contracts, have not been securitized in the region, though practitioners expect this soon. The amounts issued through securitizations ranged from US$4 million to more than US$200 million. The originators ranged from petrochemical and construction material firms to clothing and retail firms. Enhancing financing of SMEs trough securitization 3.6 Securitization involves structuring and packaging a pool o f cash flow-generating, underlying assets (receivables, mortgage loans, future revenues) into marketable securities, which are then sold to investors. In almost all cases, a special-purpose vehicle i s created to purchase the assets, which it does by issuing securities (such as bonds) in its own name. By this process, the risk associated with the initial owner and seller of the cash-generating assets i s removedfrom the risk inherent inthe underlying assets. Investors who purchase the bonds issued by the special-purpose vehicle rely on the credit quality of the special-purpose vehicle, which depends on the future cash flow o f the cash-generating assets-not on the credit quality o f the initial owner and seller o f the assets. 3.7 Securitizations can enhance financing to SMEs by creating a channel from institutional investors. Asset securitization has proven an efficient mechanism to convert illiquid and relatively high-risk SME-related assets into tradable securities that have the creditworthiness required by such institutional investors as pension funds. They are a complement to other financing mechanisms but have yet to fully develop and are not intended to replace financing through bank credit lines, ordinary bond issuance, and equity. They can be a significant alternative source o f financing for companies that can only take advantage of them. Securitizations can, however, be costly to structure and issue, and they impose highrequirements on issuers. Until now, the number and volume of SME-related securitizations in the region has been relatively modest compared with mortgage-backed securities. 3.8 The remaining part o f this section will be dedicated to review a few examples of securitizations to provide perspective on different alternatives that could be pursued in Peru. The following cases will be discussed: securitization of loans, invoice factoring, reverse factoring, securitization o f leasing contracts and multi-originators. 3.9 Securitizations of loans, known as collateralized loan obligations, are fairly common in many developed financial markets. They apply to SME loans when commercial banks have significant penetration in the SME segment, as in the United States, Spain, Italy, and Germany. In most Latin American countries, commercial bank portfolios of SME loans are growing and will soon reach critical mass. Many practitioners believe that once liquidity and capital structure o f banks are aligned the market will take off. Many o f the existing transactions in Latin America have been originated by non-bank financial institutions. The credit activities of nonbank financial institutions (finance companies, savings banks, credit unions, and to a lesser extent microfinance institutions [MFIs]) inthe SME market in Latin America have grown significantly, albeit from a small base. Structured transactions in the region have been originated by nonbank financial institutions. One example i s the securitization of loans to acquire machinery by Servicios Financieros Navistar in Mexico. A second transaction i s the 2005 securitization o f 36 bridge loans to construction companies originated by Hipotecaria Su Casita, a housing finance company inMexico. 21 3.10 In December 2004, Servicios Financieros Navistar (SFN), a nonbank captive financial institution (subsidiary of Navistar International Corp., a North American manufacturer of truck and buses) securitized a portfolio of 1,545 loans (most o f them to SMEs) for the local currency equivalent o f about US$50million. All loans inthe transaction were provided by SFN in sales of Navistar equipment to clients. These loans have an average original maturity of 40 months and carry a fixed interest rate. This transaction was static-that is, all loans transferred to the market existed at the beginning o f the transaction and they would not be substituted by new assets once they mature. 3.11 For smaller or less creditworthy commercial banks and nonbank financial institutions, the mainbenefits are enhanced financing opportunities, as is the case for leasingand factoring firms. Securitizations can help smaller (and routinely less creditworthy) financial intermediaries to obtain funding, reduce financing costs, and diversify funding sources through access to securities markets. For larger or highly creditworthy commercial banks, the most important benefit i s capital relief. Through securitizations, banks can transfer to the market the credit risk o f their SME loans, freeing up the capital set aside for them. This capital relief can enhance the capacity o f lenders to maintain or even increase the origination o f SME credit without requiring additional capital, enhancing their incentives for SME lending. Thus contributing to further expansion of the funding available to a broader range o f SMEs. 3.12 Invoice factoring, or the purchase of receivables, offers several benefits to SMEs. Factoring improves the liquidity of a company by substituting accounts receivables with cash, reducing the need for bank and supplier financing. Factors can often help recent startups and undercapitalizedcompanies, since their main focus i s the creditworthiness of the end-customers. Factoring companies tend to have fast and convenient service compared with that o f banks. On the downside, factoring i s typically a source of short-term working capital financing, not useful for capital expenditure. 3.13 The factoring process involves checking the quality of receivables, providing a cash advance, and keeping a reserve amount until the receivables are fully paid. The factoring firm will conduct a lien search to determine if any lender has taken a first lien position on the business's accounts receivable as collateral. The factoring company focuses primarily on assessing the creditworthiness o f the business customers, rather than that of the business itself Invoices are eligible for factoring only once the services (or products) have been provided and accepted by the customer. Typically, an advance i s made of 75-95 percent of the invoice value, and the difference, called the reserve, i s held back until the customer pays the invoices in full. A fee and an agreed upon interest rate i s deducted from the reserve. 21The transaction had a cross-border component (with U S $ l O O million rated AAA based on a guarantee from Financial Security Assurance) and a local component (with Mex$137 millionrated mxA). 37 3.14 Reverse factoring** i s used in countries with weak credit information infrastructure. In ordinary factoring, the factor (lender) generally purchases many accounts receivable from a limited number of sellers, which requires the factor to collect credit information and calculate the credit risk for a large number o f customers. Inreverse factoring, the factor purchases receivables payable b y only a few high-quality customers from many smaller suppliers. Thus the lender needs to collect credit information and calculate the credit risk for only a few large, transparent, internationally accredited firms. 3.15 Reverse factoring facilitates access to affordable finance to SMEs in emerging markets that supply goods and services to highly creditworthy, local, private companies and government and sub-sovereign entities. As such, this mechanism does not substitute for an existing factoring industry ina country, but it can complement it. 3.16 To make reverse factoring work, it i s necessary to develop incentives for large buyers to make it easier for their small providers to use factoring. This was identified as an issue in Chile, Peru, and Guatemala, because large clients typically like to use their negotiation advantage and keep the flexibility to request an extension from their small suppliers, while factored receivables have to be paid on time or a default i s triggered. Part of the solution may be to emphasize the benefit to large clients through more favorable supplier credit terms. 3.17 A well-known example of reverse factoring i s that implemented by Nafin, a Mexican development bank. A major feature of Nafin's program i s the development of a Web-based systemwhere participating large buyers report information on their accounts payable. 3.18 This system is a trustable registry of active invoices, minimizing the risk of fraud. Inthis system, the supplier can discount the invoice through several different financial institutions (including banks and factors) participating in the program, which compete on price. Because the source of the future payment o f the invoice i s a high-qualitybuyer, the credit risk assumed by financial institutions i s low, which drives down the interest rates charged on these type o f operations. The collection o f receivables i s also centralized in the system under Nafin's management, thereby reducingthe administrative costs to both suppliers and buyers. 3.19 Securitizations of leasing contracts use a similar framework to those of trade receivables. Instruments such as the Brazilian FIDCs provide enough flexibility to cater to different types of originators and assets. Leasing could be an alternative solution to provide financing for equipment needed by SMEs to grow their operations. The main benefit o f securitizations for leasing companies i s the greater potential to access funding through capital markets. Unlike banks, leasing companies do not benefit from low-cost funding through deposits. Their funding sources are mostly expensive bank credit lines and in some cases, financing from parent companies. 3.20 The relative homogeneity of leasing contracts makes them particularly suitable for securitization. Leasing contracts usually have similar maturities, have relatively standardized 22Klapper, Leora. 2004. "The Role of 'ReverseFactoring'in Supplier Financingof SmallandMedium-Sized Enterprises."WorldBank,Washington, DC. 38 terms and conditions, and have homogeneous cash flows and similar types of collateral. The relatively long maturities of leasing contracts match the (usually) long terms of securitization transactions. The prerequisite for the development of leasing securitization i s a well-developed primary leasing market. The volume of the leasing sector's operations in Latin America is still modest. As a result, the potential for leasing securitizations may be limited until leasing i s adequately developed as an SME funding tool. 3.2 1 Multi-originator securitization transactions. In some Latin American countries, the limited credit from financial institutions to SMEs has constrained the size of lenders' credit portfolios, limiting their access to financial markets through securitizations. One way to overcome the limited securitizable pool i s to pool several originators. Pooling portfolios from various originators into one securitization can create a large volume o f assets and achieve diversification. These transactions can help originators reach the critical mass required by the market and benefit from economies of scale. Multi-originator securitization transactions have been arranged in Colombia and Argentina through specialized conduits. But these transactions have been based on more standardized assets, such as mortgage loans. 3.22 Going forward, one o f the main requirements for multi-originator transactions with SME- related assets will be homogenizing the basic characteristics of these assets-as well as their origination and administration processes. This requirement, and the fact that originators are jointly affected by losses in the overall portfolio, can make multi-originator transactions more viable for companies that belong to the same industrial group or trade association. The Brazilian FIDCs (Fundo de Investimento em Dereitos Creditorios) has features that allow them to function as multi-originator conduits. See Chapter 3 appendix: An example of Multi-originator Government and market support for SME loan securitization 3.23 The development of mechanisms that create a channel between institutional investors and SME borrowers requires active involvement of government, originators and investors. The development of securitization markets - setting up instruments, building up legal expertise, edu- cating market participants and potential investors - can be costly. These costs are usually covered by the first issuers, but once the market i s developed, new issuers can benefit substantially by issuing securities with similar characteristics. The primary development o f the market i s thus akin to a public good that private issuers may not be willing to cover. Inthis con- text, the active participation of governments in enabling the development o f these markets i s important. For an example on how Government can reduce transaction costs see appendix to chapter 3- Government support through developing SME's securitization infrastructure. 3.24 At the borrower level Governments can support SME financing through the provision of partial direct guarantees on loans to SMEs, but this requires adequate guarantee design and pricing to align incentives and minimize fiscal cost. The main rationale for these guarantees i s that lenders may not lend to SMEs, given their perceived high credit risk, lack o f credit history, and the lack of lenders' normal screening and risk assessment processes. By sharing the credit risk with lenders through partial guarantees, governments aim at enhancing lender incentives to start or increase lending. Inturn, banks are expected to develop knowledge o f the market and tailored credit risk management processes that enable them to maintain SME lending even after the guarantee i s eventually phased out. 39 3.25 Authorities can facilitate the development of primary markets by ensuring the basic legal framework for credit operations. This i s especially so for such instruments as factoring and leasing, which in many countries do not have specific legal and regulatory recognition, reducing lender rights and increasing the cost of working out nonperforming accounts. The basic legal requirements for these instruments include the recognition o f the concept of true sale o f credit assets or the transfer of credit rights and the existence o f regulations for a concept o f special- purpose vehicles and the recognition of its bankruptcyremoteness. 3.26 Clear accounting and tax rules regarding securitizations must exist. If a securitization structure does retain some originator recourse (to replace deteriorating assets of the special- purpose vehicle), clarification i s needed on whether the securitized assets should be kept on the balance sheet of the originator. Other rules that need to be sufficiently transparent include tax rules for originators regarding capital gains on assets transferred to the market. Regulatory institutional capacity i s paramount for the functioning o f the market. The dynamic pace o f securitization markets, along with the relative complexity of the transactions, poses challenges for authorities to acquire the skills and resources to maintain security in the system without undulyhindering or delaying approval of viable transactions. 3.27 Experience inLatin America has yielded mixedresults,23but generally it seems important that lenders remain exposed to some credit risk to have an incentive to originate and manage guaranteed loans efficiently. A more detailed analysis o f Latin American guarantee programs i s needed, but so far there i s little evidence o f successful phase-outs of such guarantees or thorough impact analysis of additional lending to SMEs The development of adequate securities instruments facilitates market development, as in Brazil and Mexico. These instruments by themselves will not necessarily ignite a market in the absence o f adequate macroeconomic and primary market conditions. But inadequate securitization vehicles and securities instruments can hinder the development of the market even when all other necessary conditions are present. 3.28 A strong primary market (for securitizable assets) is essential for the development of these financing mechanisms. Securitizations help originators expand their operations by providing liquidity through the transfer o f existing assets to the market. The strength o f primary markets depends largely on the development o f adequate products and processes by lenders. These products must recognize the underlying characteristics of borrowers and be based on clear and consistent origination processes. While standardization of terms and conditions o f credit to SMEs-especially bank loans-may be difficult, lenders must strive to standardize processes and underwriting practices. To the extent that originators can demonstrate a consistent track record in origination, monitoring, and collection of their credits, access to markets will be facilitated. 3.29 Institutional investors should be allowed to invest in new types of securities as long as their creditworthiness i s within the limits o f their investment rules. Securitizations can give investors a new asset class that provides appealing yields with high levels o f creditworthiness. 23 See "Sistema de garantia de credit0 en America Latina: Orientaciones Operativas" by Juan J. Llisterri et al., IADB, 2006. 40 Within fund managers' mandate, institutional investors should develop the expertise to assess the risk and adequately price these instruments. Regulation should not impede investment in structured securities, especially when they possess a credit rating from an independent and recognized rating agency. Price vendors may be needed to help with marking to market when bonds are illiquid or nonstandard. 41 Chapter 4. STRUCTUREDFINANCE PERU: INSTITUTIONAL, IN LEGAL, REGULATORY AND OBSTACLES 4.1 Inthis chapter we analyze what is the potential role of structured finance in alleviating financing difficulties o f SMEs in Peru, what has been learned from recent experiences and what i s preventing further growth. The potential role of structured finance inPeru 4.2 The development of structured products provides an opportunity not only to expand the sources of funding to the private sector, but also to contribute to the development of capital markets in Peru. Because institutional investors in general, and AFPs in particular, dominate the demand side for corporate debt, the expansion o f local capital markets requires that domestic issuers develop securities that adapt to the interests o f these dominating investorsz4. Pension funds have a predictable funding flow and fairly predictable liabilities for long periods of time. As a consequence, they have a long-term planning horizon and look for assets that generate a compatible flow of real income. In terms o f maturity, size, and credit quality, pooled securitizations represent an attractive investment option for AFPs. 25 4.3 The Government has shown an interest infurther promoting the development of the capital market, focusing on addressing the unmet financial needs of the private sector, particularlyof smaller firms26.Inthe past five years broad based sector deepening has resumed slowly but steadily, with efforts focused on trying to expand both the demand and the supply of securities. On the supply side, CONASEVz7,the government agency that supervises enterprises and securities issuances, hasbeen working on simplifyingregistration for Tier-2 companies. 4.4 On the demand side, SBSz8, the agency that regulates and supervises institutional investors, has implemented various measures that seek to provide more flexibility for AFPs to choose the composition of their securities portfolios, including: (i) an increase in the maximum 24 Zervos (2004). 25 De la Torre & Schmukler (2004) page 19 26 During the last days of June 2008, the Peruvian Congress approved important amendments to the legal framework of the financial system and the capital market by eliminating legal obstacles and, as a consequence, promoting their development. As a result of the recently approved amendment to the Financial System Law, non-bank financial companies are allowed to execute thirty six additional operations, which before were only executed by banks, including factoring operations, structured financing operations and participation in securitization processes. Additionally, the number of operations allowed for municipal and rural savings and loan institutions (cajas municipales and cajas rurales) as well as microfinance institutions (edpymes) were increased with the mentioned amendment including factoring operations. This action that helps SME' access to finance has been complemented by the provision of collateral for SME credits (Le. Fondo de Cobertura Multiple, COBMYPE). Moreover, an amendment of the Securities Market Law incorporated significant aspects related with the regulation o f securities valuation companies in order to improve the mechanism for setting prices. 27 Comisidn Nacional Supervisora de Empresa y Valores, or National Supervisory Commission of Enterprises and Securities. 28 Superintendencia de Banca, Seguros y AFP, or Superintendence of Banking, Insurance and AFPs. 42 limits on foreign asset holdings (from 10% to 13.5% o f assets); (ii) introduction of a multi- the fund scheme following the example of Chile; (iii) regulations pertainingto the improvement new o f risk evaluation and information disclosure; and (iv) a moderation of the limits applicable to investment inprivate-sector infrastructure projects and structured securities. 4.5 Further steps under consideration include removing regulatory impediments to issuing securities, publishinga yield curve for private securities, unifyingthe tax treatment of securitized commission transactions, and clarifying the tax treatment of financial derivatives while aligning it with that for securitization transactions. 4.6 Structured finance products, and securitizations inparticular, are a promising set of financial instruments that can facilitate access to capital markets for the underserved sectors in Peru. As we have seen in the previous section, if successfully implemented, these instruments can allow the underserved sectors to deal with the main obstacles that hinder their participation in capital markets. First, the problem o f firm size can be solved by the pooling o f a large number o f firms in a structured security. Second, in terms of the legal and institutional requirements to exercise creditor rights, the bar i s set much lower for asset-backed securities than other securities, because the assets are backed by collateral and because foreclosing should be less complicated than going through bankruptcy proceedings, as would be required in the case o f unsecured corporate bonds. Third, securitization breeds securitization, because any given structured security will contribute to solving information asymmetries between institutional investors and SMEs, which in turn can lead to further intermediation with the SMEs by these or other financial sector players. The situation inLatin America i s described by29: 4.7 In Peru, the volume of securitizations doubled between 2004 and 2005, reaching US$l.1 billion (1.4% of GDP). Most o f these were absorbed by AFPs: the share o f securitizations in AFPs' securities portfolio more than tripled between 2004 and 2007, reaching 5.3% (US$ 1 billion) of total investment by AFPs. 4.8 The World Bank has played an important role in promoting innovative securitizations in Peru since 2005. The Bank, with support from the FIRST initiative financial sector development facility, has provided technical support for the design o f two innovative transactions that gave SMEs access to capital markets. Recent experiences and lessons learned 4.9 FIRST initiative3'. The project financed the study o f alternative ways to support the provision of finance to SMEs in Peru. This resulted in two innovating structures: (i) a local currency multi-originator securitization of micro-credits by Cajas Municipales de Ahorro y CrCdito (CMAC) and; (ii) Revolving factoring o f SME receivables (TREFImodel). a. Multi-originator securitization of SME loans 29 Cheikhrouhou et.al. (2007). 30 A partnership between the World Bank, the multidonor programFIRST Initiative, and the National Supervisory Commission for Companies and Securities (CONASEV) aims to bring to the Peruvian market two innovative structures that could create an interesting model for SME financing structures in the region. 43 4.10 The CMAC is a pilot program for the securitization of revolving local currency denominated S M E loans. The project includes capacity-building efforts, regulatory fine-tuning, and a tentative local currency bond issuance for the equivalent o f US$20 million, for which three CMACs (Huancayo, Piura and Sullana) have applied for a multi-originator approach. 4.11 The securitization process involves a number of steps.: (i)selection of the asset portfolio; (ii) design of the financial and legal structure; (iii) rating o f bond issues; (iv) credit issuance, authorization by the financial regulator and central bank; (v) public offer authorization by the securities market regulator; (vi) creation of the SPV; (vii) placement of the securities issued; (viii) administration of the SPV, control o f the process and originator accounting. 4.12 The timeframe for the securitization is October 2008. This largely depends on the market conditions at the time, especially coincidence/overlapping with other relevant bond issuances. A number of important steps are still pending, the operation's prospectus needs to be submitted to CONASEV and approval could be obtained by September/October 2008. For participants, see Table 6 . CMACs (Huancayo, Piura, The true originator of the transaction, whose incentive i s Owner/seller Sullana) to raise additional funding by leveraging the underlying cash-generating assets it owns. Arranger and BBVA- Banco Continental A financial institution marketing and distributing the issuanceagent bonds to be issued. Rating agency Apoyo & Asociados An institution whose mission i s to assess the Internacionales S.A.C.- Fitch creditworthiness of the underlying assets and the securitization structure; it will assign a credit rating to the bonds to be issued. Special purpose Securitization trust An independent legal entity specifically created to hold vehicle the underlying assetsuntilthe maturity of the financing. Credit enhancer Tbd by costhenefit analysis * An entity or process that increases the creditworthiness of the underlying assets and improves the marketability of the bonds to be issued. Investors Institucional investors (AFP, Purchasers and holders of the bonds issued by the special Fondos Mutuos, Compaiiias de purpose vehicle. Seguros, ONP) 4.13 CMACs were selected because their lending to SMEs is large enough for a partial portfolio sec~ritization.~~ initial analysis o f potential originators picked CMACs (Peru's The largest providers of microcredit) as they had been very successful in building a sizeable, well- managed and profitable portfolio of SME loans (worth an average of US$l,OOO each). This portfolio had a loan growth of 28.5% annually, totaling US$642 million. 31Thestudy also examined the demand side, where the market shows an excess demand for financial assets of less relative risk, conducive to the securitization of the CMAC SME portfolio. For example, in 2005 the private sector had issued debt securities for a total of US$1,474 million, while the portfolio of just AFPs had increased by US$1,616 million. Between 2004 and 2007, the total securities portfolio held by institutional investors increased from 19% to 27% of GDP (US$27 billion). 44 4.14 The securitization isolates the cash flow-generating assets from the credit standing of CMACs through a separate legal entity. This special purpose vehicle (SPV) will be a securitization trust. The cash-generating assets will be sold to the SPV and the investors who purchase the bonds issued by the SPV will rely on its credit quality, which depends on the future cash flow o f the cash-generating assets-not on the credit quality of the initial owner and seller of the assets (CMACs). However, loan recovery activities will remain under CMACs responsibility and the financial entities will have the obligation to renew the pool of assets due to: i)higher than expected non performing loans, and; (ii) the fact of the different maturit o f assets, which implies that CMACS will be required to sell SME loans on a continuous basis. 3 1 4.15 Pooling portfolios from various CMACs into one securitization creates a large base of assets achieving good diversification. The multi-originator securitization helps CMACs reach the critical mass required by the market and benefit from economies o f scale. Notably, such transactions have been arranged in Colombia and Argentina, but have been based on more standardized assets, such as mortgage loans. 4.16 To enhance the credit quality the underwriter will use internal and external techniques. Internal techniques may involve: (i) senior-subordinated structure, with investors holding the senior bonds having the first claim on payments generated by the pool of underlying assets; (ii) an overcollateralized structure, transferring a pool of assets with a face value larger than the face value o f the initial bond issuance. External credit enhancements may involve: (i) partial credit guarantees (provided by commercial security assurance companies, public entities such as development banks, and multilateral organizations) to either reduce the probability o f default or enhance recovery in case o f default33; (ii) seller guarantee or recourse to seller where the initial seller commits to transfer additional assets to the special purpose vehicle if there i s a shortfall in the assets that initially backed the issuance o f the bonds; (iii) hedging part or all o f those risks by purchasing products such as credit derivatives or put options. The CMAC operation may include a subordinated structure and some degree of overcollateralization (among the internal techniques), a partial credit guarantee by third parties and recourse to the issuer (among the external techniques). 4.17 Two credit rating agencies34will provide ratings of the transaction. These ratings will be specific to the actual portfolio to be securitized and the credit quality achieved through the credit enhancement techniques employed. The rating will take into account the likely performance o f the underlying assets (under best and worst case scenarios) and the likely effect on those assets o f changing market conditions. Lessons learned: 4.18 Government owned issuers may face budgetary and procurement constraints. CMACs are relatively small financial intermediaries in the Peruvian markets and given the fact that they are government owned entities, they faced central government budgetary and 32SMEloans usually have a 3 to 6 months term and the new securities will have a 3 year term period. 33 The former when the guarantee i s intended to cover liquidity problems, and the latter when the guarantee i s intended to cover shortfalls after the resources from the securitized assets have been depleted. 34Apoyo& Asociados Internacionales S.A.C. and Class Asociados. 45 procurement restrictions on salary levels on advisory services. The FIRST Initiative project helped to develop a special framework status for contracting structuring and arranging financial services regarding the issuance that could promote other new issuers. Process required continuous consulting and legal support to CMACs at the Ministry o f Finance and CONSUCODE (government procurement body). 4.19 The lack of skilled humancapital can be a limitation. The issuance process i s new for CMACs and skilled human capital i s scarce and has to be allocated into various tasks with issuance being one of them. The time required to collect, prepare, and process statistical information in order to provide estimates o f probability o f default among cohorts and get a risk rating has led to delays in implementation. Similar issues appeared in the preparation of origination manuals and additional operational and legal documentation. This requires early identification of human capital requirements. 4.20 Issuers' diverse priorities may delay the securitization process. Even though CMACS are small, some o f them are aggressively seeking regional and international diversification through acquisitions of smaller financial institutions or greenfield projects. The process has put strain inhuman resources and distracts some institutionsfrom the established schedule. 4.21 The commitment of participating issuers is critical. Given the fact that 3 CMACs are involved in the process, any delay by one of the institutions will affect the group. If the transaction i s not progressing the others will loose interest. Given the need to attain a critical mass to access the market and dilute transaction costs, handling this risk i s vital. CMAC Huancayo has been more proactive in the process and has provided a pilot case for the others involved. Efforts were also applied to include additional CMACs in the process to achieve a critical mass as a preventive measure. However, these measures were only feasible for the period before the initiation o f any regulatory procedures at the SBS, due to the legal provisions requiring the precise identification of the issuer. 6. Revolvingfactoring of SME receivables (TREFI model) 4.22 The TREFI model was evaluated as a `proof of concept'; (See the TREFI model Appendix for details on structure andprocess.) The first step was to compare the cost and risk o f providing credit by Peruvian financial institutions to large, medium, small and micro sized enterprises, with similar arrangements offered by corporate suppliers to SMEs across Peru.35The conclusion was that corporate suppliers have lower cost when compared to financial institution^.^^ However, corporate suppliers generally try to minimize the amount of credit provided to their SME clients, because (i) additional financing diminishes the cash position of the corporation, and; (ii)f a higher perceived risk o f extending credit to the SME and the o 35This analysis was based on (i) received from SBS, consisting of 4.5 million loan observations for a total of data 144,096 different companies, reported monthly, from the end of January 2001 until the end of May 2006, and; (ii) data obtained from M G Rocsa, a large supplier in Peru, with all credits provided in the period December 2002 to November 2006. 36 The underlying reasons for this result are: (i)Suppliers have better information on their small credits than financial institutions, as they perform more transactions and visit their clients more often; (ii)Suppliers generally have better industry knowledge than financiers; (iii)Suppliers have better tools to get their money in case of distress of a debtor than financiers, because suppliers can stop delivery, resulting in immediate ending of the debtor's business,and; (iv) Suppliers are better at remarketing returned or seized goods. 46 inability to efficiently obtain adequate credit information on its clients. The proof of concept then turned to examining a possible methodology that could erve to satisfactory resolve these two concerns. To do so the study looked at the TREFI model, from perspective of the corporate supplier to the SME. The study also looked at different approaches whereby financiers might be encouraged to participate inproviding financing to SMEs through such a platform. 4.23 The integration of appropriate risk management tools into the TREFI financing model could lower the perceived SME's risk. The model incorporates effective and easy-to- use tools for use by corporations in managing their risk exposure. The tools leverage the knowledge of SMEs available to potential corporations, without revealing confidential business information relating to the corporation interms o f who its sells to, what it sells, how muchand at what pricing. In addition, due to the knowledge collected from the corporations over time, better credit indicators can become available through the platform compared to alternatives that rely on accounting or registration methods. Notably, the credit indicators are used to calculate the finance cost o f a receivable, which substantially simplifies credit management, and, as a result, the risk perceptiono f SMEs and SME portfolios may decrease. 4.24 InPeru, several vehicles could be used to implement the TREFImodel: (i) Investment an Fund, public or private, with an SME receivable fund established to issue participation certificates and invests inTREFIfinance (similar to the Compass corporate receivable fund - see Box 2); (ii) an Asset Backed Commercial Paper program, possibly stimulated through partial guarantees; (iii) Securitized bonds backed by SME invoices through a Securitization fund. An initial analysis suggests that it would be best to start-up with an investment fund and after a while establish a commercial paper program. Lessons learned: 4.25 A way to improve SME risk assessment, risk management and to reduce transaction costs. The scheme i s easier to administer for government than an SME guarantee programs and enables an effective refinancing o f the SME credits through the capital markets. Nevertheless, SME receivable financing needs more resources dedicated to promotion and dissemination. This will allow local agents to gain greater familiarity with the products and learn more about their structure and benefits. 4.26 Despite the barriers both products are feasible under the current legal framework. The key success factors have been the ability of originators to obtain a critical mass of underlying assets. In the case o f CMACs this has been the grouping together o f loans, to form an SME loan's portfolio attractive in both amount and composition. Inthe case o f invoices for TREFI, it i s creating a group of assets large and diversified enough to reduce the costs o f issuance and the risk perception of investors. SME receivable financing necessitates more time dedicated to promotion and dissemination, allowing local agents to gain familiarity with the product and learn more about structure and benefits3' 37 In terms of COMPASS, the existing investment fund in corporate receivables in Peru, no main legal or regulatory restrictions inhibit its development, but it could benefit from a numberof changes in terms of costs. 47 Box 2: Reverse Factoring CompassFondo de Inversi6n paraPymes The initial market study by the FIRST Initiative project identified Reverse Factoring as another important vehicle for helping improve SME accessto finance in Peru. Incountries with weak credit information infrastructure, the case for several emerging market economies, factoring companies use a variant called "reverse factoring.", whereby the factor purchases receivables payable by only a few high- quality customers from many smaller suppliers.This implies that the lender needs to collect credit information and calculate the credit risk for only a few large, transparent, internationally accredited firms. The FIRST Initiative project studied the institutional environment for the operation of Reverse Factoring. No major legal, tax or institutional barriers that would impede this product's development were identified, although some aspects could be improved to facilitate operations. Nevertheless, reverse factoring has its limits, since corporations could elect to pay their SME suppliers more promptly. Infact, no reverse factoring experiences exist indeveloped countries since, ultimately, it i s a mechanism for markets in which factoring is not available to SME due to factoring companies' refusal to take on SME risk. Currently, there is a single reverse factoring fund experience inPeru. This is managed by COMPASS and has had a slow start, but, as a number of issues have been resolved and awareness has increased, i s likely to continue growing. This Fondo de Inversihpara Pymes i s a close-end investment fund that was created in April 2004 with a duration of two years (untilApril 2006) and which has been extended for four more (until April 2010). Its objective i s to provide working capital financing to SME with no particular sector preference at competitive rates, through the purchase of their corporate receivables. At the time of creation, the COMPASS fund received subscription commitments of up to U S D 63 million. It was finally opened on April, 2004 with a capital of USD55 million, the maximum amount. However, finding corporate customers took longer than expected, therefore the General Assembly decided on October, 2004 to reduce the fund's paid up capital to U S D 10 million and allow capital calls to its maximum U S D 55 million. Currently there are 5 corporations and 1,600 SME participating in this scheme." *Notably, on July 2006, Compass launched a new fund called Fondo de Inversionpara PYMES I1with the same objective of the first one and with subscription commitments of up to Soles. 40 million. This fund may be absorbed by Fondo de Inversion para PYMES Iby July 2008. 4.27 Table 1and 2 inthe Appendix, summarize the legal barriers/obstacles encountered in this specific transaction. In great detail the nature of the obstacle i s described together with the authority responsible for the issue. The obstacle i s classified as a: (i) relevant restriction, (ii) restriction or (iii) suggested improvement. Then the type o f legal regulation required addressing the issue and the suggested approach to overcome the problem. 4.28 Below we look at. three key issues: (i) the minimum size required to cover costs; (ii) the length of time necessary to structure the instruments, and; (iii) the cost associated with taxation. 4.29 Minimum size. A recent quantification of the costs of these activities was provided by BCR, in its 2007 Financial Stability Report, (see Table 7). Market practitioners conclude that an issuance amount of $10 million can provide significant coverage of fixed and transaction- specific costs. This i s primarily the case for companies that offer a conservative risk. Comparatively, Zervos (2004)38 concludes that the minimum amounts of issuance to cover transaction costs for corporate enterprises of local investment grade rating (BBB minimum rating) are $18 million for Mexico, $40 million for Chile and $50million for Brazil. 38 Zervos, Sara ,2004, "The Transactions Costs of Primary Market Issuance: The Case of Brazil, Chile and Mexico", World Bank Policy Research Working Paper No. 3424. 48 Preparation cost 77,463 77,463 Costs for primarypublic offering 35,707 217,582 I Total Cost 113,170 220,045 Annual maintenance cost 40,250 88,100 4.31 Clarity in the tax treatment of instruments, markets and investors. The lack of clarity as to how the law applicable to capital gains taxation will be interpretedi s one o f the main determinants of the costs o f issuance. The relevant Income Tax legislation establishes a temporary exemption for capital gains that thus far has been renewed annually however; this approach detracts from the predictability needed to facilitate the development o f long-term instruments. 39 4.32 Policy implications: efforts are needed towards the progressive reduction of the transaction costs, legal barriers and duplicate requirements to structure the instruments. Another improvement could be CONASEV realizing a central role in ensuring greater transparency and legal clarity, with the market assigning prices corresponding to the risk - thus allowing the market to play a greater role in risk evaluation. Finally, greater clarity in the tax treatment between instruments, markets and investors, with a better interpretation o f the law on taxation 39 In early 2007, a tax reform was launched that amended the tax burden on financial instruments through differential rates of income tax. The DL 972 tax on capital gains states that different rates will apply depending on the nature of the gain, the type of investor, the nature of issuance - creating incentives for the transactions to be made by an offshore company due to the rate differential. In the case of bonds, an exemption i s granted until December 2008, while government bonds are not affected by this tax. Additionally, the pension fund industry has an explicit exemption, based on the rationale that the income of affiliates has already been taxed and should not be subject to double taxation. Finally, a remaining aspect is the regulation of structured instruments, which, depending on the underlying asset involved, require some clarification for its implementation. In the case of securitized instruments there i s a need to clarify details related to taxation in order to ensure the success o f this new type of instrument. In the treatment of securitizations, the LMV presents little clarity in establishing mandatory enrollment in the public register of "guarantees and other covenants relating to the rights granted by the trust" and several interpretations have been made regarding this issue. Notably, the requirement to register the assignment of mortgage lending in the land register would make securitization of such claims unviable. 49 with regard to capital gains, would remove market asymmetries between investors and instruments. Box 3 Benefits of the CMAC SME Loansecuritization - For the seller (CMACs) is the enhancedfinancing opportunity in obtaining funding, reducing financing costs, and diversifying funding sources. Furthermore, the transaction can also provide better asset-liability management through matching maturities, diversified sources of income, capital relief. This is so only when the originator does not hold a subordinated participation or when its subordinated participation i s smaller as a proportion of the securitized pool than the regulatory capital requirements for the whole securitized pool. In addition an initial contact with the market, this could lead to easier market access and less credit enhancement in further issuances. For the CMACs the option of securitization was attractive as they required funding at better prices and longer terms than the prevailing ones. The alternatives were: (i)attracting deposits from the public and others, traditionally short-term, partially in dollars and with interest rates greater than the average rates of banks; (ii) taking up traditional debt in the form of shordmedium term loans in market conditions; (iii) capital private inflows to shareholders, permitted by law since 1996 but not practiced; (iv) issuing securities like corporate bonds, not yet explored. Box 4 Benefits of the TREFImodel - For SME: (i) Increased amount and maturity to obtain credit with purchases, as it no longer has a bad impact on the working capital liquidity for corporations, but a positive impact on sales; (ii)o additional administration for N the SME. For Corporations, in addition to advantages offered by capital markets factoring: (i) Ability to increase amount of finance to SME customers, without compromising their working capital. This allows them to increase sales to SME customers with adequate credit; (ii) Increased certainty to obtain finance of SME receivables, if compliant with structure requirements as compared to issuing corporate bonds, CP or loans. This results from the fact that TREFI finance is not correlated with the rating of the corporation. If the corporation manages its credit well adequate credit quality portfolios can be sourced from the corporation; (iii) Low barriers to entry when compared to issuing bonds, as no public ratings are required for the corporation and lower legal expense; (iv) Lower set-up and operational expenses. Minimal administrative impact. For private sector investors: (i) Increase of the amount of high quality investment product supplied. Non-rated and smaller corporations by themselves have highly diversified portfolios of SME receivables, out of which large amounts of high quality rated investment product can be obtained; (ii) Due to low correlation with corporate debt, ability to diversify portfolio. For government compared to alternative SME finance stimulation: (i) the model generates true economic As value by improving SME risk assessment, improving risk management and reducing transaction costs, SME finance can be stimulated in a budget neutral way; (ii) Price and selection mechanism can be used to direct any government support to specific groups of SME (e.g. regional, industry type, SME size); (iii) The method i s easy to administer for government as compared to SME guarantee programs; (iv) As the credit capabilities of suppliers are used inthe process, a vast new resource i s applied to stimulate SME finance. 50 Recommendations regarding institutional, legal and regulatory constraints 4.33 In this section we summarize general constrains and outline suggestions to tackle them supplementing some of the considerations referred to in previous chapters. The team responsible for the study uncovered these issues, while reviewing documentation, procedures and market practices as well as in conversation with Government Officials and private sector representatives. 4.34 There is a perception among market participants that regulators, especially CONASEV, have not kept pace with the rapid development of the capital markets. Although CONASEV has recently streamlined several processes, the institution i s perceived to be bureaucratic, its tariff structure as too burdensome, and slow in approving new issuances. For example, SAFIs (Sociedades Administradoras de Fondos de Inversion, Mutual Funds Companies) that recently entered the market took on average 200 days to receive the authorization to operate. In addition, coordination with SBS in key issues such as unifying criteria for risk ratings and methods for assessing the value o f instruments remains inadequate. Although reforms are being undertaken, the rapid expansion of the stock market in upcoming years will require accelerating the reform o f the institutional framework. 4.35 In order to further the development of new instruments there needs to be a concomitant development of complementary instruments. The market for instruments in New Soles i s rapidly growing and will need new instruments to hedge against inflation and exchange rate risk. Most companies that issue debt in the stock market remain heavily exposed to the exchange rate risk. Despite generating most o f their revenue in New Soles, most companies continue issuing relatively short tenor and instruments denominated in US Dollars. An indirect way to address the above is to base the concessions of new infrastructure projects on local inflation. This way, price adjustments would be indexed to soles and dealers would be more likely to seek financing in Soles. 4.36 The size and the efficiency of the capital market can be further enhanced by developing the market for securities lending. This market mechanism would extend the capital market and allocate more efficiently the risks o f the market and o f the underlying assets. Other measures for improving the infrastructure of the market include the introduction o f an internet-based system of electronic trading and mechanisms that ensure good practices intrading and inthe use of privileged access to information. 4.37 The costs for origination and transaction are high for most issuers and investors, virtually excluding small participants from the stock market. Given the relatively small size o f Peru's capital market, fixed costs are shared by few customers and operations. In order to reduce costs, regulators need to simplify procedures, eliminate unnecessary requirements and disseminate information regarding the services and commissions charged by different administrators. 4.38 The development of the capitalmarket faces several challenges on the demand side. Key challenges include the high concentration of securities in few investors, a restrictive regulatory framework for investors, the strong discretionary power o f the regulators and the lack o f a centralizedmechanismfor pricing securities. 51 4.39 The demand for securities is highly concentrated in very few investors, affecting competition and stability of the market. By the end of 2007, the four AFPs represented 59 percent of the demand for securities, whereas the rest of the demand was composed by relatively small investors. As a result, AFPs could determine the performance of a stock in the capital market. The power of the AFPs goes beyond the capital market. By September 2007, the funds managed by these 4 AFPs represented 117% o f private investment and 20% of GDP. Other countries have dealt with the concentration issue by allowing other financial institutions, such as banks or insurance companies, to receive voluntary savings funds and by requiring AFPs to subcontract part of the funds under their management. 4.40 Although the regulatory framework has become more flexible in recent years, and despite being one of the least restrictive in the region, it possesses significant weaknesses. Currently there i s a lack of regulation allowing investments in derivatives or hedging techniques such as swaps and forwards or other instruments such as private equity. Limits on the shares that AFPs can holdby instrument type or for investing abroad remainlow. 4.41 The selection of the specific instruments in which AFPs can invest in, is decided by SBS. Some currently approved instruments do not meet the criteria (liquidity) set by the SBS, raising questions on the quality and the impartiality of the selection process. Market participants agree that the decision process would benefit from greater transparency and a more automatic process in selecting investment instruments. 4.42 The pricing of securities and of illiquid financial instruments needs to be further improved. The development o f the array o f prices in 2005, the recent publication of the vector o f prices of AFPs and the amendment to the Securities Market Law in June 2008 are three important improvements. However, the mechanism for setting prices o f securities remains weak. The SBS and CONASEV have yet to establish clear and standardized criteria for the valuation of assets o f institutional investors. Moreover, the prices o f financial instruments are not set in one consolidated and liquid market and as a result it i s difficult to obtain a single price for each financial asset. 4.43 On the supply side, the offerings of securities in the market remains fairly limited. Inaddition to the barriers described above, many businesses remain reluctant to obtain financing through the capital market because o f the fear o f revealing competitively sensitive business information to third parties or the fear o f being treated differently by tax authorities. Inaddition, the high liquidity in the traditional financial sector makes the capital markets less attractive. In order to increase the number of issuances, the regulator needs to reduce the regulator-related costs for entering and remaining in the market, simplify procedures and make the regulatory framework more flexible. Some governments have taken a more active role in promoting new securities by providing tax incentives or by promoting sectors such as venture capital. 52 Appendix to Chapter 3 An exampleof Multi-originator,securitizationof Trade ReceivablesinBrazil-FIDC Zoomp A3.1. Zoomp S.A., a Brazilian clothing label, entered the local securities markets for the first time in 2005 through a R$48 million (roughly US$23 million) three-year securitization program o f trade receivables, using the popular FIDC structure. The transaction achieved a rating o f brAAAf, the highest in the Brazilian local market, which translated into a yield only nine basis points above the interbank reference rate. a. The StructureandParticipants A3.2. The assets backing the transaction are trade credit receivables generated through the sale o f the originators' products to multibrand retailers and franchised shops. The mechanism o f the transaction (see figure A.1): The transaction starts with the discounted sale of existing receivables from Zoomp to an FIDC, which operates as a bankruptcy-remote, special-purpose vehicle, managed by a professional independent asset manager. The discount on the face value of the receivables i s intended to be the mainsourceof interest payments for investors. The FIDC issues senior shares sold to institutional investors and subordinated shares kept by the originator. A large part of the proceeds from senior shares i s used to pay to the originator (Zoomp) for the receivables sold. Part i s maintained by the FIDC to create a liquidity reserve to repay investors. At maturity, trade receivables are paid to a custodian-in this case, a highly rated commercial bank.The custodianisincharge oftransferring the funds to the FIDC.This eliminates the riskthat payments of securitized receivables made directly to Zoomp could be diverted to other uses. One part of the funds collected through the custodian i s allocated to the reserve account, the source of payment of interest andprincipalto investors. Another part i s used to acquire additional receivables from Zoomp as a substitute for those matured. At the end of the transaction, the reserve account is usedto pay to investors the final installment and the remainder, after payment of operating expenses for participants, is paidto Zoomp as the holder of the subordinated notes. 53 A. 1. The Structure of the FIDC Zoomp 7 1 Discounted I F'DCZoomp (Issuer) I Pavmentfor I Institutional Investors (Fund Manager) (Senior-note receivables holders) T Payhatof subordinated Fundsto acquire bonds (R$? .. more receivables(R$) Paymentfor principal on bonds (R$) the products (R$) Client D Client N b. Mitigating the Risks A3.3. The creditworthiness of the fund shares sold to institutional investors i s achieved through subordination. In this transaction, the main risk i s default by Zoomp's clients. The assessment o f this risk is determined through analyzing the historical performance of Zoomp's trade credit portfolio. Once assessed, this risk i s covered by issuingtwo types of notes: senior notes offered to institutional investors and subordinated notes kept by the originator (20 percent o f the total issuance). Because senior-note holders have the first claim on the fund's assets, the securitized portfolio has to suffer losses beyond 20 percent of its value before senior-note holders are affected. Holding subordinated notes, Zoomp receives payments only after full repayment of principal and interest to senior-note holders. Zoomp assumes any losses below 20 percent, creating incentives to generate creditworthy trade receivables. A3.4. The possible deterioration o f portfolio credit quality also poses a risk to investors. Given its revolving nature and the short term of the portfolio securitized, the transaction relies heavily on Zoomp's ability to originate receivables o f similar credit quality and characteristics to those originally securitized. It i s thus essential that the originator can demonstrate a good track record inoriginating trade receivables. Strict eligibility criteria ensure that matured assets in the transaction portfolio are substituted by other assets o f similar characteristics and quality. Inthis case, the custodian i s incharge of constantly monitoring the quality of the portfolio. 54 Government Support through DevelopingS M E SecuritizationInfrastructure A3.5. In Germany, a government program reduces transaction costs for securitizations. The Programme for Mittelstand-loan Securitization (PROMISE), managed by Kreditantstalt fiir Wiederaufbau (KfW) in Germany, aims at enhancing SME financing by providing the means for lenders to transfer the credit risk of existing SME loans to the market. A3.6. The capital relief through these transactions is expected to allow Germanbanks to expand their loan origination without additional capital. The underlying benefit o f the program i s that it drastically reduces the transaction costs in securitizations through the use of an existing platform and a tested legal structure that investors, intermediaries, and regulators are familiar with. A3.7. This program i s based on synthetic transactions4' because KfW was seeking explicitly a mechanism to help lenders only with capital relief. A similar program for true sale transactions would provide the same main benefits inreducing transaction costs to originators. A3.8. InPROMISE transactions, KfW acts only as an intermediary, to help originating banks transfer to the market the unexpected losses o f the reference portfolio. Banks pay a premium and retain the expected losses (the first loss position, typically about 3 percent). Since originators maintain the propefty o f the underlying loans, the size o f the pool i s well above that o f the actual bonds issued. KfW takes the credit risk of the reference pool from the originating bank (after first loss) and transfers it to financial institutions (through credit default swaps) and institutional investors (through credit-linked notes). Note that KfW does not retain any credit risk or provide any guarantees on the transactions. A3.9. The results of the program have been positive so far, with close to 15 billion issued. From 2000 to 2004, 11 transactions were issued under the PROMISE program, ranging from 200 million to 3.7 billion. Pension and investment funds have demonstrated considerable interest inthese securities. A3.10. The benefits of the program are manifold for market participants. For lenders, besides the benefits of capital relief, through this mechanism, KfW provides equal access to capital markets to banks of different sizes. The common infrastructure o f the program reduces significantly the transaction costs and facilitates the execution o f transactions. In terms o f deepening financial markets, the PROMISE program has helped to develop a secondary market for SME credit risk, with the associated price discovery benefit. For investors, PROMISE notes were a new asset class o f high credit quality. For SMEs, the program i s expected to translate into enhanced access to bank credit, though this impact has been difficult to quantify 40 Securitization of SME-related assets can be done through true sale or synthetic transactions, with the main difference being whether the ownership of the underlying assets i s transferred to the market or remains within the originator. Intrue sale transactions, the SME loans are segregated from the originator's portfolio and their ownership and credit risk i s transferred to a special-purpose vehicle. In synthetic structures, the ownership of the loans remains within the originator's portfolio. Accordingly, the originator does not obtain liquidity, but uses these transactions only to transfer portfolio credit risk to the market. 55 Appendixto Chapter4 The TREFI model A4.1. The modelexaminedremovesconcerns arising by corporatesuppliers. The model is based on TREF141, a method for medium and large corpopate suppliers that (i) removes the negative working capital effects o f providing credit to SME by enabling effective refinance of the SME credits through the capital markets, and; (ii) removes SME risk concerns o f providing credit to SME by generating effective credit indicators for SME using SME payment behavior. Use of this mechanism allows corporate suppliers to provide additional finance to stimulate the sale of their products and services to selected SMEs and therewith allowing the SMEs to increase their sales and thus increase the growth o f Peruvian economy. The model may be introduced in the Peruvian market by 2008. Easing the re-finance of credits provided by SME to corporations was also examined. A. 2 The functioning of TREFI Financingof I Account SPV Receivables Funds ~~ -II MarketsCapital cn TREFI n (Back Office) Bonds Banks v) a '0 s0 CORPORATION Receivables n A4.2. The scheme operates through a SPV acting like a clearinghouse, purchasing receivables from Corporations on a revolving basis and financing these by obtaining loans or notes financing from financiers in the capital markets. TREFI can operate in Peru as a bankruptcy remote special purpose entity or another similar vehicle authorized by the Peruvian legislation. The SPV purchases receivables from corporations on a revolving basis. The corporations continue to service the receivables and provide a sub-participation through which the first loss on the portfolio i s transferred to the corporations, incentivising them to administer and risk manage the receivables well. The SPV finances the purchases by obtaining loans or notes financing from financiers in the capital markets in Peru. Government entities can participate in the financiers to stimulate financing o f SME. The SPV may also obtain funds directly from banks to purchase receivables. The SPV acts like a clearinghouse. The financiers determine the pricing, revolving 41 TREFI (www.trefi.com) is held by the not for profit foundation The ReceivableStandards Board (TRSB). It i s run by a set of replaceable service partners that must comply with minimum quality standards. TRSB has an independent board that controls that the model operates in compliance with the TRSB standards and ensures that upgrades do not negatively impact existing finance. In Europe TREFI i s currently run by Getronics (www.eetronics.com), Meespierson Intertrust a Fortis Subsidiary (www.meesDiersonintertrust.com), KPMG (www.kDmE.com), Pricewaterhouse Coopers (www.DricewaterhousecooDers.com), Capital Tool Company Agency Services Limitedand others. 56 period and risk level at which the receivables are transferred to SPV and therefore the pricing to the corporations. A4.3. To performits duties, the special vehicle needs a PayingAgent, an Administration Agent and a Calculation Agent. The paying agent (PA) executes the payments calculated by the calculation agent; the administration agent (AA) performs the accounting of TREFI, and; the Calculation Agent (CA) plays an important role in the TREFI model, as: (i)acts as an independent rating agency to the corporation calculating the risk on each debtor and each receivable; (ii) provides the financiers necessary tools to adequately price, manage and finance the portfolios the corporation selects to finance; (iii) selects SME receivables from the corporate, flags them as financed and calculates on a net basis the payments to be made by corporations, the SPV and the financiers (funds or banks) on a daily basis. A4.4. The TREFI model needs to include a portfolio management function.42The model operates as a highly efficient mid and back office enabling active portfolio management, therefore it helps to segregate an important pool o f clients by their payment behavior, economical sector or specific characteristics. Splitting the portfolio management function from the risk assessment and payment operations improves segregation of duties and therefore integrity of the investments to the end investors. LegalObstacle Authority Classify Norm Recommendation General If there is an economic link MEF Include cost of Technical between the SPV and the Study in transaction costs or entities transferring credit structure the transaction in a rights (represented or not by way as to avoid any economic invoices), rules regarding linkbetween the transferor and Transfer Prices apply. If this is transferee of credit rights. the case, it would be necessary to prepare a Technical Study supporting value set for transferred assets. Exemptions to income and Include this premise in the value added taxes which projection of future flows used currently apply to securities as basis for structure. issued in public offerings or in organized market mechanisms are temporary. It is not possible to predict if these exemptions will be maintained incoming years. The tax exemption applied to Include this premise in Pension Funds-and noti o other estimating demand for issued 42 Note that banks would generally invest directly into TREFIfinances. Banks in that way use the TREFI model as a highly efficient mid and back office enabling active portfolio management. It also provides Basel I1compliant risk management and reporting to the financier (and the corporations), therefore financing provided in this way i s compliant with Basel 11. 57 institutional investors places I Isecurities. these at a disadvantage in private placements. Applicableto CMACs CMAC require specific SBS Resol SBS The SBS might be able to authorization to act as establish criteria CMAC must originators or managers of meet to act as originator or securitization funds manager of securitization fund, instead of requiring authorization for every operation. Entities not considered part of BCR Circular The BCR should specify that financial system (such as BCR the maximum limits- to interest securitization funds) are rates do not apply in subject to limits established for securitization of assets from interest rates companies that are part of the financial svstem. CMAC are subject to fiscal 2 None The operation structure should control from Contraloria not include CMAC obligations General de la Repdblica. If the which would require the securitization structure opinion of the Contraloria. includes CMAC guarantees for a term longer than a year, prior opinion from Contraloria i s required. Capital gains derived from MEF 2 Law Include this cost in the transfer of credit portfolio transaction costs for CMAC. "without recourse" to securitization fund are subject to income tax Impossible to restrict SME 3 Interp Include risk in transaction debtor's right of prepayment INDECOPI structure. CMAC' s principal Municipal 2 Contract Local governments must shareholders continue to be Gvts. demonstrate willingness to local governments, thus adopt corporate governance complicating implementation standards. Ensure compliance of good corporate governance by introducing covenants in securitization contract. SBS recommends CMAC to SBS 3 None Ask for SBS to make maintain leverage ratio under information relative to CMAC the legal limit of 11 times leverage requirements public. equity. However, there i s no single limit that applies to all CMAC and each one individually i s unaware of the limit the SBS will apply to them. Yassification: 1= RelevantRestric m; 2 = Restric an; 3 = SuggestedImprovement 58 Legal Obstacle Authority Classify Norm Recommendation General If there is an economic link MEF 3 None Include cost of Technical Study between the SPV and the in transaction costs or structure entities transferring credit the transaction in a way as to rights (represented or not by avoid any economic link invoices), rules regarding between the transferer and Transfer Prices apply. If this i s transferee of credit rights. the case, it would be necessary to prepare a Technical Study supporting value set for transferred assets. Exemptions to income and CONASEV 3 Law Include this premise in the value added taxes which projection of future flows used currently apply to securities as basis for structure. issued in public offerings or in organized market mechanisms are temporary. It i s not possible to predict if these exemptions will be maintained in coming years. The tax exemption applied to MEF 3 Law Include this premise in Pension Funds and not to other estimating demand for issued institutional investors places securities. these at a disadvantage in orivate olacements. Investment funds are currently CONASEV Resol. A norma de cardcter general not allowed to make CONASEV that authorizes Investment investments involving: (i) Funds to invest in this type of purchase of SME receivables; operations i s required. y, (ii)grant loans to third parties, except in the case of debt instruments issued by companies that comply with requirements set by Investment FundRemlation. v Applicable to Receivables Capital gains derived from MEF 2 Include this cost in structure as transfer i f receivables to an transaction costs for company SPV are subiect to income tax Capital and guarantee CONASEV 2 requirements as established by CONASEV terms of current capital and Ley de Fondos de Inversidn guaranteerequirements could result in considerable financial burden. I The transfer of rights CONASEV 2 Law or It i s necessaryto establish norms (including credit rights Contract similar to those actually in place represented by SME invoices) regulating the communication of 59 must be communicated to the transfer of rights in the debtor in order to be valid. The framework of a securitization costs of communicating the fund. Nevertheless, it is possible transfer might result in to overcome this obstacle considerable financial burden through the structuring of the given the large number of transaction, by obtaining debtors. previous accepted from a transferred debtor. Capital gain resulting from sale MEF Law Take this issue into account of participation certificates when estimating demand from registered in the RPMV or investors. from return of public placed participation shares i s exempted from income tax. However, investment fund profits or losses are allocated to fund participants and are thus taxed as income. The Reglamento de MEF Reglamento Include this issue in Comprobantes de Pago trainingpresentations that will requires corporates that transfer take place for implementation of invoices to an investment fund Product 11. or SPV through a factoring contract to issue a document specifying the total credit amount transferred on the date of the transfer. This generates confusion among corporates that transfer invoices (since effectively invoicing sale of invoices). Pension Funds are authorized SBS Resolution Greater clarity i s needed to invest in Investment Funds SBS regarding whether or not A F P s that purchase "titulos de are allowed to purchase deuda" (debt instruments), participation shares in whether publicly or privately investment funds that purchase placed. Regulation for Pension invoices with SMErisk. Funds, however, does not define these "titulos". It would be convenient to clarify whether or not invoices are considered debt instruments, in order to determine if A F P s can buy participation shares in an investment fund that purchases SMEinvoices. Classification: 1= Relevant Restric on; 2 = Restri ion; 3 = SuggestedImprovement 60 BIBLIOGRAPHY Alles, L.2001.Asset Securitization and Structured Financing: Future Prospects and Challenges for Emerging Market Countries. IMFWorking Paper. Bank for International Settlements. 2005. The role of ratings in structured finance: issues and implication. BIS Report from the Committee on the Global Financial System. Beck, Thorsten and Ash Demirgiig-Kunt. 2006. Small and Medium-Size Enterprises. Access to Finance as Growth Constraint. Journal of Banking and Finance 30, 2931-43. Beck, Thorsten, Ash Demirgiig-Kunt and Sole Martinez Peria. 2007. Reaching Out: Access to and Use of Banking Services across Countries. Journal of Financial Economics 85,234- 66. Beck, Thorsten, Ash Demirgiig-Kunt and Sole Martinez Peria. 2008. Banking Servicesfor Everyone? Barriers to Bank Access and Use around the World. World Bank Economic Review, forthcoming. Beck, Thorsten, Erik Feijen, Alain Ize, and Florencia Moizeszowicz. (2008). Core Financial Development Indicators, World Bank mimeo. Bigsten, Arne, Paul Collier, Stefan Dercon, Marcel Fafchamps, Bernard Gauthier, Jan Willem Gunning,MAns Soderbom, Abena Oduro, Remco Oostendorp, Cathy Pattillo, Francis Teal, and Albert Zeufack. (2000). Credit constraints in manufacturing enterprises in Africa. Centre for the Study of African Economies Working Paper WPS/2000.24. Borensztein, E., Eichengreen, B. and U.Panizza. 2006. Debt Instruments and Policies in the New Millennium: New Markets and New Opportunities. Inter-American Development Bank Working Paper. Cheikhrouhou, H and Mendoza, JC. 2006. SME Financing: The missing piece of the puzzle. Presentation Banco de la Republica de Colombia Conference on Access to Finance, 9-10 November. Cheikhrouhou, H., Gwinner, W.B., Pollner, J., Salinas, E., Sirtaine, S., and D. Vittas. 2007. Structured Financial in Latin America. Channeling Pension Funds to Housing, Infrastructure and Small Businesses.Washington, DC: The World Bank. De la Torre, A., Gozzi, J.C. and S.L. Schmukler. 2006. Financial Development in Latin America :Big Emerging Issues, Limited Policy Answers. World Bank Policy Research Working Paper 3963. De la Torre & Schmukler (Eds.). 2004. Whither Latin American Capital Markets? Chapter 2. Developments inCapital Markets". Washington, DC: The World Bank. Iacobucci, E.M. and R.A. Winter. 2005. Asset Securitization and Asymmetric Information. The Journal of Legal Studies,vol. 34. Ketkar, S. and Ratha, D. 2004. "Recent Advances inFuture-Flow Securitization". TheFinancier, vol. 11, 12. Levenson, A.R. and W.F. Maloney. 1998. TheInformal Sector, Firm Dynamics and Institutional Participation. World Bank Policy Research Working Paper 1988. 61 LoveInessa andNataliyaMylenko. 2003. Credit Reporting and Financing Constraints. World Bank PolicyResearchWorking Paper 3142. Srinivas, P.S. and J. Yermo. 1999. Do Investment Regulations Compromise Pension Fund Performance? Evidencefrom Latin America. Washington,DC: TheWorld Bank. Superintendenciade Banca, Seguros y AFP (Eds.). 2006. TheNew Strengths of Peru. Financial, Insurance and Private Pension Systems. The World Bank. 2002. Financing development throughfuture-flow securitization. PREM notes 69. The World Bank. 2007. Developing the Domestic Government Debt Market: From Diagnostics to Reform Implementation. Washington,DC: The World Bank. The World Bank.2007. Polastri, R., Saavedra, J., Loayza, N.,Humphrey,C., Shankar, R., Vakis, R. andLecussan. Peru. Casting Light in the Shadow Economy. Draft Document. Wright, DavidL.and DewanA. H.Alamgir. 2004. MicrocreditInterest Rates in Bangladesh: Capping Versus Competition. CGAPPaper producedfor the Donors' Local Consultative Group onFinance. Zervos, S. 2004. The Transactions Costs of Primary Market Issuance: The Case of Brazil, Chile, and Mexico. World Bank PolicyResearchWorking Paper 3424. 62