WPS4649 Policy ReseaRch WoRking PaPeR 4649 Bank Involvement with SMEs: Beyond Relationship Lending Augusto de la Torre María Soledad Martínez Pería Sergio L. Schmukler The World Bank Development Economics Research Group & Latin America and the Caribbean Region June 2008 Policy ReseaRch WoRking PaPeR 4649 Abstract The "conventional wisdom" in academic and policy of bank involvement with SMEs in various emerging circles argues that, while large and foreign banks are markets documented in this paper is neither led by small generally not interested in serving SMEs, small and niche or niche banks nor highly dependent on relationship banks have an advantage in doing so because they can lending. Rather, all types of banks are catering to overcome SME opaqueness through relationship lending. SMEs and larger, multiple-service banks have in fact This paper shows that there is a gap between this view a comparative advantage in offering a wide range of and what banks actually do. Banks perceive SMEs as a products and services on a large scale, through the use of core and strategic business and seem well positioned to new technologies, business models, and risk management expand their links with SMEs. The recent intensification systems. This paper--a product of the Development Economics Research Group and the Office of the Chief Economist, Latin America and the Caribbean Region--is part of a larger effort in the departments to understand financial development. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The author may be contacted at sschmukler@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Bank Involvement with SMEs: Beyond Relationship Lending Augusto de la Torre María Soledad Martínez Pería Sergio L. Schmukler * JEL Classification Numbers: G21, G28, L25, O12, O16 Keywords: small and medium enterprises, bank finance, financial constraints, banking market structure *This paper is part of a broader project to understand bank financing to SMEs, undertaken in the World Bank Group through the Latin American and the Caribbean Region, the Finance and Private Sector Development Network, the Development Economics Research Group, and the International Finance Corporation. We thank many colleagues at the IFC and the World Bank and are particularly grateful to: Sonja Brajovic Bratanovic and Constantinos Stephanou, who gave us data on Serbia and Colombia, respectively, and to Arantxa Veraza López, Manuel López Hernández, Melina Mirlsmulstein, Ary Naim, Camila Rodríguez, who interacted with us at different phases of this project. We are also indebted to the IFC SME team and FRS (Inmark Group) for sharing with us unique data. We received very helpful comments and suggestions from Luis Ahumada, Pablo Carbajo, Stijn Claessens, Kevin Cowan, Rogelio Marchetti, Hernan Lacunza, Margaret Miller, Ary Naim, Cristina Pailhé, Gastón Repetto, José Rutman, Sophie Sirtaine, and Greg Udell. We also received useful comments at presentations held at the Asobancaria Conference on SME Lending (Bogota, Colombia), the CASIN-IFC Forum on SMEs (Geneva, Switzerland), the Central Bank of Argentina (Buenos Aires, Argentina), the Central Bank of Chile (Santiago, Chile), the Latin American Congress on Banking and SMEs (Cartagena, Colombia), the World Bank Small Business Finance Conference (Washington, DC, USA), and the World Bank (Washington, DC, USA). Maria Bernarda Dall'Aglio, Noemí Soledad López, Mira Olson, Mercedes Politi, and Victoria Vanasco provided excellent research assistance at different stages of the project. We thank all the banks that participated in the study for their generosity in sharing information and for interacting with all the members of the team. The views expressed in this paper are entirely those of the authors and do not necessarily represent the opinions of the World Bank Group. Email addresses: adelatorre@worldbank.org, mmartinezperia@worldbank.org, sschmukler@worldbank.org 1. Introduction The financing of small and medium enterprises (SMEs) has attracted much attention in recent years and has become an important topic for economists and policymakers working on financial and economic development. This interest is driven in part by the fact that SMEs account for the majority of firms in an economy and a significant share of employment (Hallberg 2001).1 Furthermore, most large companies usually start as small enterprises, so the ability of SMEs to develop and invest becomes crucial to any economy wishing to prosper.2 The recent attention on SME financing also comes from the perception among academics and policymakers that SMEs lack appropriate financing and need to receive special assistance, such as government programs that increase lending. Various studies support this perception. A number of papers find that SMEs are more financially constrained than large firms.3 For example, using data from 10,000 firms in 80 countries, Beck, Demirgüç-Kunt, Laeven, and Maksimovic (2006) show that the probability that a firm rates financing as a major obstacle is 39% for small firms, 38% for medium-size firms, and 29% for large firms. Furthermore, small firms finance, on average, 13 percentage points less of their investments with external finance when compared to large firms.4 Importantly, lack of access to external finance is a key obstacle to firm growth, especially for SMEs (Beck, Demirgüç-Kunt, and Maksimovic 2005). On the policy side, there are a large number of initiatives across countries to foster SME financing including 1 According to data collected by Ayyagari, Beck, and Demirgüç-Kunt (2007) for 76 developed and developing countries, SMEs, on average, account for over 50% of manufacturing employment. 2 See, for example, Beck and Demirgüç-Kunt (2006) and references therein. 3 See, for example, Schiffer and Weder (2001), IADB (2004), Beck, Demirgüç-Kunt, and Maksimovic (2005), and Beck, Demirgüç-Kunt, Laeven, and Maksimovic (2006). See Cressy (2002) for a review of the literature on funding gaps. 4 See Beck, Demirgüç-Kunt, and Maksimovic (2008). 1 government subsidized lines of credit and public guarantee funds. One example that has been deemed as relatively successful is Chile's Fondo de Garantía para Pequeños Empresarios (FOGAPE), a fund created to encourage bank lending to SMEs through partial credit guarantees. This fund has many features that make it attractive, including some incentives to reduce moral hazard, promote competition among banks, and encourage self sustainability.5 The "conventional wisdom" argues that the inadequate financing of SMEs is to a significant extent rooted in "supply-side" features. That is, the way in which financial institutions operate is biased against offering financing to SMEs; as a consequence, many banks and other financial institutions become uninterested in being involved with SMEs. The conventional view highlights a number of factors that might constrain financial institutions. To start, financing SMEs is difficult because they are opaque.6 Opaqueness means that it is difficult to ascertain if firms have the capacity to pay (they have viable projects) and/or the willingness to pay (due to moral hazard). This opaqueness particularly undermines lending from institutions that want to engage in the more impersonal or arms-length financing that requires hard, objective, and transparent information.7 5 Other examples abound. For instance, the Colombian Fondo Nacional de Garantías (National Guarantee Fund) provides similar partial credit guarantees. The structured finance transactions arranged by FIRA, a Mexican development financial institution focused on the agricultural sector, are another example of a government effort to provide financing to rural SMEs. Also, the Mexican development bank NAFIN has initiated a reverse factoring program to provide working capital financing to SMEs through a process of on- line sale of receivables from large buyers. See de la Torre, Gozzi, and Schmukler (2007). 6 See, for example, Berger and Udell (1998) and Cole, Goldberg, and White (2004). 7 For example, lack of audited financial statements prevents banks from engaging in what is known as financial-statement lending, by which the loan contract terms are set on the basis of the company's expected future cash flow and current financial condition as reflected in audited statements (see Berger and Udell 2006). Kano, Uchida, Udell, and Watanabe (2006) show that firms without audited financial statements benefit the most from relationship lending (bank-borrower relationships). 2 Second, relative to large firms, SMEs are more likely to be informal, particularly in developing countries. This not only makes opaqueness worse, but it also poses additional obstacles and risks to SME lending.8 For example, banks cannot lend to SMEs as much as would be warranted if firms do not report reliably their full financial activity on their financial statements.9 Furthermore, informality implies that the firm has unrecorded, contingent senior liabilities to the government and its own employees.10 Faced with the risk that tax and/or labor authorities could cause such liabilities to materialize, banks would lend less to SMEs or charge a higher risk premium. Third, capital markets do not compensate for these deficiencies in the banking sector as they do not have a comparative advantage in dealing with opaque and small firms. In effect, capital market financing rests on comparatively high accounting and disclosure requirements which, by definition, opaque SMEs lack. In addition, capital markets are typically not a source of direct funding for SMEs, given that these firms are unable to issue debt or equity in amounts sufficiently large to attract investors (who prefer liquid issues and are not willing to take too large a share of a single asset) and amortize the large issuance-related transaction costs (including compliance with complex 8See Gatti and Honorati (2007) for evidence that SMEs are more informal (measured as self-reported lack of tax compliance) and that this informality affects negatively access to credit and external finance. 9 In the case of banks in Argentina, for example, the regulation states that a bank's total lending to a borrower cannot, in general, exceed 100% of the equity of the borrower, as stated in its financial statements. This limit can be increased to 300% if the total amount of financing is less than 2.5 % of the capital of the bank and the increase is approved by the board of directors of the institution. 10 The government and employees have senior claims on a firm's assets in most legislations. The government can seek to collect unpaid taxes before the firm fulfills its obligations to repay other creditors. And employees can typically sue the firm for "unpaid" benefits, even if these benefits were agreed upon in an informal manner, for which no record exists. 3 legal, regulatory, accounting, and disclosure requirements). These factors normally render unfeasible the direct access to capital market financing for SMEs.11 In sum, there are good arguments behind the conventional view that supply-side factors--particularly opaqueness--play a key role in the problems of access to finance for SMEs. To the extent that "opaqueness" has received special attention in the literature on SME financing, so has "relationship lending," with the latter seen by the conventional view as the obvious--if not the only--way to cope with the former. The common perception is that "relationship lending" can overcome "opaqueness" because it relies primarily on "soft" information gathered by the loan officer through continuous, personalized, direct contacts with SMEs, their owners and managers, and the local community in which they operate, to mitigate opacity problems (See Berger and Udell 2006). Relationship lending has various implications on the ways that banks lend to SMEs. First, while banks can compensate for opaqueness through relationship lending, other things being equal, banks lend less to SMEs than they would if SMEs were more transparent and, hence, subject to arms-length lending technologies. Moreover, relationship lending implies a cost that makes dealing with SMEs less attractive and, in equilibrium, raises the required returns for SME loans. The likely high cost associated with relationship lending stems from the labor intensive process of collecting soft information. 11See, for example, Stoll and Whaley (1983), Stoll (1984), Berger and Udell (1998), Halling, Stomper, and Zechner (2005), and Gozzi, Levine, and Schmukler (2008). 4 Second, because of the personalized, community-based nature of the contacts, the literature has argued that it is difficult for large and foreign banks to engage in relationship lending. These banks are perceived as relatively less capable of processing "soft" information. It is, moreover, difficult for them to quantify and transmit this type of information through the formalized communication channels of large/complex organizations or financial institutions for which the headquarters are far away.12 Therefore, large and foreign banks are seen by the conventional wisdom as not well suited to lend to SMEs, leaving this segment to small or niche small banks, which are close to the relevant sector, community, or neighborhood and, therefore, are typically domestic.13 Third, in better institutional environments (typical of developed countries) relationship lending is expected to be less common, because banks would need to rely less on soft information and could count on well-functioning contract enforcement institutions, making it easier for them to screen SME debtors, originate and monitor loans, and recover loans in case of default. Although not conclusive, the academic literature has found evidence consistent with the idea that banks (mainly small and niche) engage with SMEs through relationship lending.14 A number of studies on the US find that large banks allocate a much lower proportion of their assets to small business loans than small banks.15 Other studies on the 12See Berger, Klapper, and Udell (2001), Stein (2002), and Mian (2006). 13 See DeYoung (2000), DeYoung and Hunter (2003), Carter, McNulty, and Verbrugge (2004), and DeYoung, Hunter, and Udell (2004) for a discussion of the comparative advantages that small community banks have in lending to small firms through relationship lending. 14As described in Cull, Davis, Lamoreaus, and Rosenthal (2006), access to finance to SMEs throughout the nineteenth century was primarily provided by local financial institutions that relied heavily on informal means to acquire information and allocate credit. 15 See, for example, Keeton (1995), Berger, Kayshap, and Scalise (1995), Berger and Udell (1996), and Strahan and Weston (1996). In contrast, Berger, Rosen, and Udell (2006) argue that there is not a small- 5 US find that large institutions lend to larger, older SMEs with stronger financial ratios. By contrast, small financial institutions are found to rely more on soft information and lend to SMEs with which they have stronger relations.16 Moreover, some studies show that SMEs are relatively less likely to receive loans from foreign banks, which are typically large.17 Other studies focus on how an improvement in the institutional environment enhances lending to SMEs, with the underlying belief that better institutions (associated with more transparency and a better enforcement of the rule of law, typical of developed countries) reduce the need to engage in relationship lending. Consistent with this, Beck, Demirgüç-Kunt, and Maksimovic (2005) show that SMEs tend to benefit more than proportionally as the overall institutional environment improves. Similarly, Galindo and Micco (2005) find that the difference in bank financing between SMEs and large firms is higher in countries with worse creditor protection and less efficient judicial systems. In this paper, we explore whether and to what extent the main tenets of the conventional wisdom--that banks in general are not interested in dealing with SMEs (due to their opaqueness) and that the small and niche banks that do engage with SMEs do so mainly through relationship lending--holds in reality. To do this, we use new data from bank surveys for a total of 48 banks (plus 1 leasing company) in 12 countries to characterize the degree, determinants, and types of bank involvement with SMEs. We use hard evidence collected via bank questionnaires as well as anecdotal evidence obtained bank bias in lending to SMEs, as indicated by the fact that the probability that a small firm borrows from a large bank is proportional to the market share of large banks. 16See, for example, Haynes, Ou, and Berney (1999), Cole, Goldberg, and White (2004), Scott (2004), and Berger, Miller, Petersen, Rajan, and Stein (2005). 17See Berger, Klapper, and Udell (2001), Mian (2006), and Berger, Klapper, Martinez Peria, and Zaidi (2008). 6 through interviews with bank officials to describe ongoing changes in the relation between banks and SMEs. To complement the information coming from banks, we also analyze data obtained from existing SME surveys. The data come from 3 different sources. First, we use data gathered for this project across banks for a group of countries via a specially designed questionnaire, which was completed during on-site interviews with bank business and risk managers in Argentina, Chile, Colombia, and Serbia by different teams of the World Bank during late 2006 and early 2007.18 This type of information provides a good picture of the market environment and structure in which banks serve SMEs in specific countries. In all of the countries included in the sample, surveyed and interviewed banks account for a large share of each country's banking sector assets (normally around or over 70 percent). Second, we analyze data collected during 2006 by the International Finance Corporation (IFC) via surveys conducted across 8 developed and developing countries: Australia, Brazil, India, the Netherlands, Poland, Thailand, the UK, and the US.19 This information complements well the mentioned within-country surveys because it depicts how banks that are leaders in the SME business segment of the market operate, especially in terms of business models and risk management processes. Third, we use surveys of SMEs conducted annually during 2002-2006 in 7 economies: Argentina, Chile, Colombia, Mexico, Peru, Puerto Rico, and Venezuela. These surveys, which were undertaken by FRS (Inmark Group), a consulting firm specializing in financial services research and 18Complete case studies are available for each of these countries describing the institutional and macroeconomic contexts, their banking industries and trends, and the data in detail. See World Bank (2007a and 2007b) and Stephanou and Rodriguez (2008). 19See IFC (2007) for a complete description and analysis of the IFC surveys. 7 strategy, give a different perspective by capturing the demand side of bank products and services. The analysis in this paper shows that there is a gap between the prevalent conventional view upheld in academic and policy circles and how banks are in practice interacting with SMEs. In particular, the evidence we present characterizes a different pattern of bank involvement with SMEs, which is consistent with the arguments first articulated by Berger and Udell (2006). This new pattern goes well beyond pure relationship lending and is increasingly observable even in relatively less developed countries, where relationship lending would be expected to be more prevalent. In particular, we find the following main stylized facts. First, contrary to the perception that banks in general are not interested in lending to SMEs, we find that most banks do indeed want to serve SMEs and find this segment profitable, especially as margins in other banking markets narrow due to intensified competition. In particular, as the public sector and large corporations have gained access to local and international capital markets (de la Torre and Schmukler 2006), and as competition in the retail sector (among banks and retail chains) has increased substantially, banks have greater incentives to incur the switching costs needed to pursue new business in the "middle" market of SMEs. As a result, SMEs have emerged as a strategic sector for most banks--including large and foreign banks and not just small and niche banks. As a consequence, the SME market in the sample countries has become competitive, yet it is still far from saturated. Second, part of the interest in SMEs comes from the fact that, as argued by Berger and Udell (2006), relationship lending is not the only way in which banks can extend 8 financing to these firms. Different transactional technologies that facilitate arms-length lending (such as credit scoring and significantly standardized risk-rating tools and processes, as well as special products such as asset-based lending, factoring, fixed-asset lending, and leasing) are increasingly applied to SME financing. For example, hard information on the SME or its owner obtained from credit bureaus allows banks to infer future loan performance and thus enables the use of credit scoring to process and approve small loans at a scale that makes costs sufficiently low.20 Moreover, the pledging (as collateral) of assets that do not lose much value over time and are relatively easily liquidated (e.g., equipment and real estate), provide greater assurances of repayment, even when contract enforcement processes are relatively imperfect. Also, reliable accounts receivable can underpin factoring, while the renting of tangible and marketable assets through leasing can help overcome costly contract enforcement processes including ambiguous commercial laws and inefficient bankruptcy procedures (as the creditor retains the property rights over the asset). Thus, when good financial information is not available to gauge capacity or willingness to pay, banks can use other types of hard information and incentive-compatible mechanisms to increase the likelihood of repayment. In this way, banks can compensate for weaknesses in the institutional environment. Moreover, these mechanisms free banks from having to rely on government subsidies to lend to SMEs. Third, lending is just one part of a larger overall package that banks provide to SMEs. Banks find SMEs profitable through a combination of services; and this places 20See Jappelli and Pagano (2003), Kalberg and Udell (2003), and Miller (2003) for evidence on the value of credit registries. 9 cross-selling at the heart of the banks' SME business strategy.21 In effect, banks have developed a wide range of fee-based non-lending products and financial services for SMEs. These products and services can be very attractive in terms of profitability; in fact, the evidence suggests that lending is not always the main or the first product offered to SMEs and that it is often offered as a way to eventually cross-sell other lucrative fee- based products and services, including payments, saving, and advisory services. Cross- selling is a way for banks to maximize their scarce resource (capital). Moreover, selling products and services to SMEs deepens the engagement of banks with SMEs, is part of the efforts of banks to become the principal bank the SME engages with, and may thus facilitate increasing the amount of lending to the same SME while attracting other clients (like the SME employees and the owner's family). To the extent that these products and services gain importance, the institutional environment relevant to credit contract writing and enforcing becomes less of a constraint. Fourth, under this new model of engagement with SMEs, large and international banks have several comparative advantages and, as a consequence, are leaders and relatively more aggressive in this business segment. Some of the technologies applied to lending to SMEs (other than relationship lending) benefit from the effects of economies of scale and scope. For example, credit scoring models that rely on statistical properties to assess risk need a large number of clients and loans, which tend to increase with bank size.22 Also, dealing with large corporations allows banks to reach out and offer loans to good SMEs that have long-term relations with those corporations (thereby reducing 21Prager and Wolken (2007) provide evidence that small firms in the US use a variety of financial services from banks. 22For a recent discussion on the use of credit scoring in the US, see DeYoung, Glennon, and Nigro (2008). And for a primer on credit scoring methods and their contrast (and potential linkages) with credit rating methods see Marquez (2008). 10 principal-agent problems and improving risk management). Moreover, large banks can seize the benefits from scale in supplying non-lending products and services to a large number of firms, taking advantage of their service platforms, technical expertise, and IT and back-office infrastructures. Finally, large banks are better able than small banks to use sophisticated business models (e.g., business centers, branches, SME account managers, and marketing campaigns) and risk management systems, so as to combine and integrate centralized and de-centralized processes as appropriate to realize efficiency gains in managing both costs and risks. In sum, the ability to serve many SMEs (and for international banks, the ability to serve also many countries) through large multi-service platforms and branch networks and through sophisticated business models and risk management systems gives large banks a competitive edge, enabling them to compensate more easily for the fixed costs and switching costs of developing products and services to engage SMEs while exploiting economies of scale and scope. The rest of the paper is organized as follows. Section 2 describes the data. Section 3 documents the degree to which SMEs constitute an important business for banks. Section 4 describes the business models that banks use to engage with SMEs. Section 5 presents evidence from the SME side. Section 6 concludes. 2. Data In this section we describe the data used to characterize the degree, determinants, and type of bank involvement with SMEs. In particular, we discuss the bank surveys performed in Argentina, Chile, Colombia, and Serbia by the World Bank, the cross- 11 country bank surveys conducted by the IFC, and the SME surveys undertaken by FRS (Inmark Group).23 The data are briefly summarized in Table 1. The cross-bank data for Argentina, Chile, Colombia, and Serbia was gathered via an especially designed questionnaire administered to banks during on-site, detailed interviews conducted with banks' top management in late 2006 and 2007.24 For each country study, different types of banks were interviewed by different teams. The samples include both banks that were ex-ante thought to be lending to SMEs and banks that were perceived ex-ante (by the authorities and World Bank experts) to have little or no relation with SMEs. Banks were selected to represent the domestic financial sector and cover a large fraction of it. The banks interviewed per country account for about 70% of banking sector assets at the time of the interviews. A total of 37 banks and 1 leasing company were interviewed across these 4 countries.25 In all cases, the interviews and data processing were confidential and conducted only by World Bank staff. All banks were informed that the data would be reported only in an aggregate way, without disclosing the strategy or positions of individual banks. The individual bank data are not available even to the country authorities, so banks had in principle no constraints in sharing information. The questionnaire includes a maximum of 92 questions and was designed to gather information on 3 broad areas.26 The first area addresses banks' strategic interest in SME business. The second area assesses the determinants of bank financing to SMEs, in 23The countries and banks interviewed were selected based on the interest of financial experts at the IFC and the World Bank. Within countries the sample of banks was chosen with the agreement of the respective central banks. 24The questionnaire for each bank was filled in the context of interviews that would typically take at least 5 hours. The interviews would normally involve a discussion on strategy with the top bank manager, as well as discussions with the managers of the business units and risk management units. 25Total assets stands for liquid assets, public and private securities, loans, and other bank assets. 26The questionnaire used to interview banks is available upon request to the authors. 12 terms of demand factors, competition, corporate strategy, and macroeconomic, regulatory, and institutional factors. The last area attempts to understand the business model and risk management processes used by banks when working with SMEs, specifically asking among other things how financing is conducted, what products and services are offered, how banks organize themselves to serve SMEs, to what extent banks engage in cross-selling, and how credit risk is managed. In Argentina, 14 banks were interviewed: 6 foreign, 6 domestic private, and 2 public, accounting for 75% of the banking system's total assets.27 In Chile, 8 banks were interviewed: 4 foreign, 3 private domestic, and 1 public, representing 79% of the banking system's total assets. In Colombia, the team interviewed 7 private banks and 1 leasing company, capturing roughly 66% of the system's total assets.28 In Serbia, 8 banks were interviewed: 5 subsidiaries of major European banks, 1 international bank specialized in SME lending, and 2 large, locally owned banks, accounting for approximately 70% of the total SME credit market and 60% of the banking system's total assets. Although countries' definitions of what constitutes an SME for legal or statistical purposes are typically based on the number of employees, banks generally define SMEs in terms of average annual sales--an indicator that is more easily observable, a good proxy of an SME level of business activity, and, thus, more useful to banks' business and risk management purposes. The threshold of annual sales used by banks varies by country, according to the size of the economies and structure of their corporate sector. In Argentina, a company is considered to be an SME when its average annual sales are 27The 14 banks are used to construct Figure 1. However, since 1 bank did not work with SMEs at the time of the interview, it was not used for the other figures. 28Since leasing is an important financing instrument for SMEs in Colombia but it cannot currently be provided directly by banks, the interviews included a major leasing company as well. Virtually all leasing companies belong to banks. 13 approximately between 300,000 and 30 million US dollars. In Chile, the range goes from around 90,000 to 24 million US dollars. In Colombia, banks consider SMEs those firms with annual sales between 400,000 and 13 million US dollars (although for most domestic banks the range is between 100,000 and 5 million). In Serbia, SMEs are typically defined as having annual sales between 500,000 and 10 million euros. The banks with more developed business models for SMEs typically distinguish between small enterprises (SEs) and medium-size enterprises (MEs). Companies with average annual sales below the mentioned ranges are considered to be micro enterprises and those above are defined to belong to the corporate sector. In this study, we ignore the heterogeneity observed in the definition of SMEs and use the definition used by each bank interviewed. This complicates to some degree the comparison across banks and across countries. Nonetheless, it is useful to analyze how banks conduct business with what they consider to be SMEs. It would be even more problematic to construct a unique working definition of SMEs for this study. The IFC questionnaire was conceived to identify "best practices" in bank involvement with SMEs, including key factors and links among business models, processes, tools, as well as the actual performance in SME banking. The IFC surveyed 11 banks assumed to be leaders in the SME business during 2006 in the following countries: Australia, Brazil, India, the Netherlands, Poland, Thailand, the UK, and the US. There are important differences in the purpose and scope of the IFC surveys compared to those of the World Bank. The World Bank surveys try to capture the main features of the entire banking system in each country with respect to SME banking, covering the most important banks in terms of assets and a wide range of banks in terms 14 of types (e.g., small, large, foreign, domestic, and niche banks). They also intend to identify institutional and policy constraints to SME banking. The IFC surveys, by contrast, focus on the banks (1 or 2 per country) thought to be leaders in SME banking in a diverse set of 8 countries. Whereas the results from the World Bank surveys can be reasonably taken as representative of what most banks in the surveyed countries do, the results from the IFC surveys can be thought as representative of "leading cases" of banks operating with SMEs. Both the World Bank and IFC surveys collect qualitative and quantitative data. However, IFC interviews of bank managers were more exhaustive than those conducted by the World Bank, as the IFC tried to capture in more detail the business processes and other operational details of SME banking. The IFC team spent 2 days with each bank (as opposed to the 1 day or sometimes ½ day spent by the World Bank team) trying to identify the factors that might lead to their successful involvement with SMEs. The IFC collected information both on site and through the completion of submitted data requests. The IFC questionnaire consists of 32 questions and some additional data sheets, and tries to evaluate comprehensively 5 major areas: segmentation and business models, products, sales and delivery channels, credit risk management, and information technology and management information systems. Finally, the FRS (Inmark Group) surveys collect data from SMEs, as opposed to banks. These are annual surveys conducted throughout 2002-2006 in a series of Latin American economies: Argentina, Chile, Colombia, Mexico, Peru, Puerto Rico, and Venezuela. Fortunately for our purposes, there is overlap with the countries surveyed by the World Bank, hence allowing us to analyze bank engagement with SMEs from the 15 perspective of both the demand and supply sides. Furthermore, the overlap in the country coverage also enables us to assess whether there is consistency in the responses provided separately by firms and banks. The FRS (Inmark Group) surveys cover around 900 SMEs per country per year, except in Puerto Rico, where approximately 600 firms were interviewed. In each country, the surveys cover the main cities, provinces, or states, where about 90% of the SMEs are located. The surveys contain 76 questions subdivided into 7 parts. The first part collects information about the firm: its location, the economic activity it is involved in, its ownership type, and age. The second part contains questions on the number of banking relations the firm has and whether and why the firm has intensified or reduced its operations with the banks identified in this section. The third section asks information about the type and number of banking products used by the SMEs. In particular, the survey looks into specific types of deposit and saving products, financing products, and banking services. The remaining sections ask detailed questions regarding how firms finance foreign trade, whether they use internet banking, and what forms of payment they accept. 3. SMEs as Bank Clients This section shows that the overall data from the surveys of banks within and across countries suggest a very similar and consistent pattern--that SMEs have become a strategic sector for most banks. Put differently, even if SMEs were ever a niche bank business, they no longer are. In fact, foreign banks and large banks have become very aggressive in targeting SMEs. Moreover, the market structure is competitive; that is, 16 virtually all banks are trying to attract SMEs and more competition is expected in the future as banks get more involved in this segment. But despite being competitive, banks in most surveyed countries see the sector as far from saturated and thus see profitable growth prospects. Several factors seem to be driving these patterns, suggesting that these new developments are not simply a cyclical phenomenon, associated with the recent period of credit buoyancy, but that they reflect structural changes and are, hence, part of a long-term trend. Factors that seemed important prior to the completion of the questionnaire-based surveys (like the institutional and macroeconomic policy environments) do not appear to be binding constraints for banks pursuing businesses with SMEs in the countries under study. Nonetheless, these factors seem to limit the range of products that banks offer and may become binding as banks try to deepen their relation with SMEs through more complex products, including long-term lending. In the rest of the section, we document these findings in more detail. Bank involvement with SMEs is significant in all the countries under study (Figure 1). Almost all the banks interviewed have SMEs among their active and important clients. Even the very few banks that did not work with SMEs at the time of the interview planned to enter the segment. Many of the interviewed banks--particularly in Argentina and Chile--have matched the identification of the SME sector as a strategic one with significant investments, organizational reforms, and active recruitment (mainly over the past 2 to 3 years) in order to penetrate systematically and broadly the SME market. This is a clear indication that the rising significance of SMEs to banking business is unlikely to be a simple reflection of a cyclical upturn in credit markets. 17 The quantitative data that we were able to obtain suggest that banks have a significant level of credit exposure to the SME segment. For example, the ratio of SME loans to total outstanding private sector loans (including retail) was 37% in Argentina and 14% in Chile in 2006.29 The level of credit exposure to SMEs of the most involved and medium involved banks is very high, representing on average 62% and 28% of the banks' loan portfolios in Argentina and Chile, respectively.30 In the case of Colombia, the national bankers' association (Asobancaria) estimates a rapid increase in SME lending from a low base in 2003, with the share of SME lending in the total commercial loan and lease portfolio of credit institutions almost doubling in less than 4 years to reach 25% by 2006. Estimates for Serbia indicate that bank exposure to SMEs reached 34% of the total term credit to entrepreneurs, small, medium, and large firms in 2006. For the developing countries covered by the IFC surveys, the exposure is estimated to be around 16%. Despite the increasing importance of SMEs in the lending portfolio, most banks are not yet able to measure the relative importance of the segment in terms of its contribution to income, costs, or risks. In the cases of Argentina and Chile, almost half of the banks do not keep track of the share of total net income generated by SMEs or the revenue share arising from deposits/account management, credit, and other transactions and fee-based services to SMEs. Some estimates indicate that SMEs contribute to 17% of banks' net income in Argentina and 24% in Chile (but we only received this information 29The banks that provided this information and that are considered in this average account for 64% of total private sector loans in Argentina and 80% of private sector loans in Chile. Future analysis would be useful to explain the differences across these countries. 30These are simple averages of the ratio of SME loans to private sector loans for the banks that belong to each level of involvement. The most (least) involved banks are the top (bottom) third of banks with the highest (lowest) share of SME loans as a percentage of total loans. The medium involved banks are the rest. If the sample were larger, it would be useful to compute statistics for the most involved large banks, because small banks cannot make large loans to large corporations due to diversification and legal lending- limit restrictions. Thus, because of their size small banks will have a large SME allocation. 18 from half of the banks surveyed in those countries).31 Furthermore, most banks (around 70% of the banks interviewed) do not have an assessment of how much SMEs contribute to the banks' total credit and other risks or to the banks' total costs. Many banks claim not to have adequate unit cost and risk accounting mechanisms in place to obtain this type of information, but are actively trying to develop them. This suggests that banks are on the steep part of the learning curve on how to deal with SMEs. Through this process, banks adapt their structures to increase profits, reduce the associated costs, and better manage the risks of the segment. The same applies to Colombia and Serbia. The interest and participation in the SME segment is not a small bank or niche bank phenomenon, as the relationship lending hypothesis would predict (Figure 2). Banks were asked to provide their views regarding the main players in the SME market.32 Universal large banks (including foreign ones) were the most frequently mentioned banks. Niche banks seem to play a role only in Argentina and Chile, but a less important role than public banks. Available quantitative data for Argentina and Chile show that, on average, private domestic banks are the most exposed to the segment, with a level of exposure of 56% in Argentina and 16% in Chile. In Argentina, private domestic banks are followed by public banks (31%) and foreign banks (27%), while in Chile they are followed by foreign banks (12%) and public banks (6%).33 The important participation of foreign banks is a significant finding, given the priors in some discussions that these banks are "cherry pickers," interested mainly in high net worth individuals and large 31The banks that provided this information and that are considered in this average account for 23% of total private sector loans in Argentina and 41% in Chile. 32Since each bank can mention more than 1 type of bank as being the leader in the segment, the numbers across bank types do not have to sum to 100%. 33These ratios are calculated as the sum of SME loans over the sum of private sector loans considering the banks belonging to each category of bank type (public, private domestic, and foreign). 19 firms, and the fact that foreign banks were the last ones to enter this segment. Indeed, from the interviews, foreign banks appear to have a very aggressive strategy to expand their activity in the SME sector. Moreover, some foreign banks that were interviewed aspire explicitly to be recognized as the clear leaders in the segment as part of their corporate objectives. Hence, they can be expected to become even more important over time. Contrary to our initial expectations, public banks have not been leading the market in SME involvement. The market structure for SME-related financial services is competitive (Figure 3). None of the banks interviewed stated that the market is not competitive. Except in Serbia, most banks thought that the market is also not saturated. That means that banks are expected to continue competing in this market segment in the years ahead. Almost all (interviewed) banks are planning to expand their operations in the SME market, with quite ambitious growth goals (e.g., to double the SME business in 1 to 2 years) in a significant number of banks in Argentina and Chile. Naturally, large banks with sizeable branch networks or a tradition of involvement with SMEs enjoy an initial advantage and tend to be leaders. These banks have an edge in their ability to offer services based on economies of scale and scope (like reaching out to customers, developing loan and deposit products that are specific to SMEs, and offering IT, payroll, accounting, and other services). However, laggards often enter the market and leapfrog through poaching and imitation. For example, new entrants and banks embarked on rapid expansion actively seek to hire SME account managers from other banks, with the expectation that such managers would bring along their SME clients. 20 Despite the competitive market structure, banks see profitable growth prospects (Figure 4). This points to a likely intensification of competition in the SME market in the future. Banks expand their engagement with SMEs by both deepening relations with existing clients and targeting untapped pools of new clients. Furthermore, the most proactive banks do not appear to have a sector-specific or regional focus when targeting SMEs (Figure 5). They tend to cover as broad a basis as possible. They try to gain market share in every sector and region and they have aggressive growth targets, which is useful for developing products on a large scale and better diversifying and managing risks, as described in the next section. (However, competition has been putting downward pressure on lending interest rates leading to an apparent under-pricing of risks in the hottest markets.) Again, all this is contrary to the idea of niche banks being the only ones interested in serving SMEs, with segmented markets (where each bank specializes in a set of clients) and with interactions based mainly on relationship lending. There is an array of factors driving banks' desire to become involved with SMEs (Figure 6). The most relevant aspect mentioned by banks is the perception of high risk- adjusted profitability of the sector. This means not only that profits in the SME sector are attractive, but importantly that they are attractive relative to the alternatives controlling for risk. For example, banks have experienced a thinning of margins in the corporate sector because of intensified competition from local and international capital markets, and in the consumer sectors because of strong competition from other financial and non- financial institutions (such as department stores). Similarly, with more stringent fiscal policies improving government access to capital markets, the opportunities for lending to the government at a spread over the cost of funds have shrunk significantly, particularly 21 in Argentina and Chile.34 At the same time, the cost of funds to banks has decreased substantially, making lending to SMEs at more affordable interest rates viable.35 The reduction in lending margins across segments has prompted banks to increase fee-based revenue and product cross-selling, with SMEs becoming a natural target for expansion. Moreover, large domestic and foreign banks with high fixed costs are more willing to pay the switching cost to work with SMEs, even if that was not part of their original intention.36 Other factors that banks mention as important drivers of their involvement with SMEs are their relation with large corporate clients and the fact that SMEs are a strategic sector for growth in their economies (Figure 6). These factors are partially linked. First, large banks use (to the extent possible) their relations with large firms to try to identify and sort out the SMEs that are worth approaching. Significant dealings with the corporate sector allows banks to go downstream, partly using "chain" relations, that is SMEs that act as suppliers or outsourcers to large corporations. The large corporations might gain in ensuring that the SMEs with which they work are offered appropriate financial products and services and, thus, operate more efficiently. In turn, banks benefit from the knowledge embodied in large corporations about the quality of SMEs, as this knowledge can help reduce substantially the problem of asymmetric information that banks face when approaching new SMEs.37 34 Argentina improved the fiscal stance after the 2001-2002 crisis, while Chile put in practice in 2000 a fiscal rule, which means that the structural fiscal surplus should be 1% of GDP. 35The lower domestic cost of funds reflects the combination of higher domestic savings leading to current account surpluses and substantial capital inflows. 36Anecdotal evidence suggests that many foreign banks entered new Latin American markets to work with large corporations, but decided to switch to SMEs only after the limits to the corporate business became apparent. 37An established literature exists on trade credit and the role of large corporations as relationship lenders. See, for example, McMillan and Woodruff (1999) and Boissay and Gropp (2007). 22 Second, ongoing changes in the organization of production appear to be placing SMEs in a new strategic place. It appears that the most dynamic SMEs are those connected to large firms via supply or outsourcing chains. The interviews with banks confirm the idea that large firms use SMEs increasingly to outsource certain activities. This gives them more flexibility, by allowing them to focus on their own business, limit litigation risks (like labor disputes), and reduce fixed operating costs. The evidence from the surveys--that banks are increasingly engaged in the SME sector--is consistent with the hypothesis that the mentioned changes in industrial organization, with less vertical integration and greater modularity and network economies via supply chains and outsourcing, generate new demands of financial products and services for SMEs that large banks are better able to provide.38 In effect, banks seem particularly well-placed to take advantage of the new demand for products and services from SMEs. Unlike large corporations, SMEs do not have the size to efficiently undertake certain financial and administrative activities in house (such as payroll, payments to suppliers, collection of receivables, import-export paperwork, and even some accounting and book-keeping activities).39 Therefore, as large firms outsource activities to SMEs and engage them as suppliers, causing SMEs to expand, SMEs in turn outsource some activities to banks, especially those activities that banks can supply more efficiently, given the economies of scale that banks can exploit through their large service platforms, back-office and IT infrastructures, and broad base 38See Rajan and Zingales (2000) and Rajan and Wulf (2003) for a description of some of the forces that lead to the break-up of vertically integrated firms. 39Factoring can be particularly helpful to SMEs. For example, in the US, factors often take responsibility for credit decisions (extending trade credit) and collection activities for bad debts on receivables (Udell 2004). 23 of technical expertise. This, in turn, raises the scope for banks to carry out cross-sales and offer SMEs fee-based services, tapping into non-traditional banking businesses. There are many ways in which banks can exploit scale effects, synergies, and linkages. For example, leveraging on its relation with a large corporation, a bank can incorporate the SMEs with which the large corporation works, then proceed to cross-sell an integrated package of products to the SME, including fee-based services, and then move on to attract as clients the employees of the SME and the family of the owner of the SME. Moreover, banks can offer similar products across SMEs and sectors, with a small customization to meet the firm-specific or sector-specific needs. Banks can also offer software packages to manage the SME accounts online, as well as advisory services. As a consequence, banks not only broaden their sources of income, but also diversify risk, in terms of lending to a new type of firms and deriving income from non-lending activities. This also limits the burden that lending imposes on banks' limited capital. In this context, government programs become less essential for banks to reach out to SMEs, and are usually not deemed by interviewed banks as decisive factors. Many of these changes in the relation between banks and SMEs are connected to technological advances that allow banks to offer products and services at a scale and cost that they were not able to offer before. In terms of lending, credit scoring allows banks to reduce the costs and time of making small loans, process a larger volume of such small loans more efficiently, and be better able to monitor the default risk of the portfolio. Studies such as Frame, Srinivasan, and Woosley (2001) and Berger, Frame, and Miller (2005) show that at least in the US the use of scoring models is associated with an increase in lending to SMEs. However, credit scoring works best for micro and very 24 small enterprises. As the size of loans get larger, banks find it harder to use automatic credit scoring methods, as mentioned in the interviews and as discussed in detail in Marquez (2008). However, banks that are actively engaged in the SME sector using more sophisticated business models and risk management systems (see below), have also developed standardized and streamlined ways of screening and rating the creditworthiness of the larger SMEs, for which automatic scoring methods do not apply well. Another way in which technology has helped banks sell services to SMEs is through online banking. Firms can now use banks' websites to, for example, process payments, input accounting and administrative data, monitor receivables, manage their liquidity, make investments, make international money transfers, buy insurance products, and more generally hedge risk. Aside from helping identify factors that drive bank engagement with SMEs, the surveys also allow us to identify the obstacles that banks perceive as limiting their engagement with SMEs (Figure 7). As expected, the responses in this regard are heterogeneous across countries and the patterns are less clear, but several aspects are worth highlighting. SME-specific factors seem important and, among those, banks tend to highlight the informality of SMEs. High competition is perceived by some banks as another obstacle in Argentina and Serbia. Interestingly, while banks would welcome further improvements in macroeconomic stability as well as in the quality of the informational, regulatory, legal, and contractual environments, they do not see deficiencies in these areas as major impediments or binding constraints at present in their involvement with SMEs. These responses, which are not consistent with our initial expectations, may have a number of explanations. First, these countries are middle- 25 income countries that might have already undertaken the basic macro, institutional, and contractual reforms needed for SME financing to occur. Therefore, in other less developed countries, where minimal reforms have not taken place, macroeconomic and institutional factors might be important binding obstacles and government action in these areas might be needed to enable significant lending to SMEs.40 Second, banks have learned to cope with a less than perfect macroeconomic, informational, and contractual environment by using products that help to deal with these deficiencies and thus limit risk exposure. For example, banks use short-term loans (mostly for working capital), require collateral and broad guarantees from the SME owners, and focus on products that tend to reduce the risk of SME lending like check discounting, leasing, and factoring. Third, it is also possible that macroeconomic factors are not perceived to be constraining because of the relatively strong growth and benign liquidity and interest rate conditions prevalent in the countries in the sample at the time of the surveys. If the macroeconomic situation in the surveyed countries were to deteriorate significantly, banks might change their perceptions on the prospects for the SME segment. Finally, one would expect that deficiencies in the contractual environment would become more binding as banks try to engage in long-term lending.41 To be sure, the completion of the questionnaires did help identify important areas for policy action specific to each country. These are discussed in 40Beck, Demirguc-Kunt, and Martinez Peria (2008) survey 80 banks operating in 32 developing countries and find that macroeconomic factors are rated as the top obstacle in this larger sample of countries. However, competition appears as the most important obstacle among the 11 banks from the 7 developed countries included in their study. 41Note, however, that the availability of general purpose long-term bank financing for SMEs might be a problem even in good contractual environments. For example, in the US, only mid-sized firms (among SMEs) appear to be able to obtain private long-term funds, mostly via private placements purchased by life insurance companies (Carey, Prowse, Rea, and Udell 1993 and Berger and Udell 1998). 26 detail in the papers on Argentina and Chile, Colombia, and Serbia, which are an integral part of the broader study.42 Moreover, the lending technology (with the exception of Serbia) and other bank- specific factors do not appear to be significant obstacles for banks to expand their engagement with SMEs. This suggests that banks' capacity to serve SMEs is greater than commonly believed, which runs contrary to the conventional idea that supply-side factors are the main explanation behind the frequently mentioned problems for SMEs to obtain financing at affordable terms. 4. Business Model and Risk Management Systems In this section we describe aspects of the business model and risk management systems that banks use to serve SMEs. We first analyze the organizational arrangements that underpin the emerging business models as well as the products offered to SMEs. Second, we discuss how banks conduct risk management. The analysis illustrates a point made above--namely, that banks' high level of interest in dealing with SMEs is matched, especially in the most proactive banks, with major changes in organizational structures and new, more sophisticated approaches to measure and manage the risks involved in doing business with SMEs. These emerging business models go well beyond the logic of pure relationship lending and suggest that banks plan to be involved with SMEs for the long haul. 42The interviews with the banks highlighted different areas for policy action, for example: (i) the "stamp tax" (which raises the cost of lending and hinders SME mobility between banks) and the legal difficulties for different creditors to share pro-rata a claim on the same fixed collateral as areas for possible policy action in the case of Chile; (ii) the ceiling on bank lending interest rates (which favors unregulated credit institutions) as a policy issue in Chile and Colombia; and (iii) the costly and protracted judicial process of credit contract enforcement as an area for potential policy action in Argentina and Colombia. 27 Regarding organizational arrangements, almost all of the banks interviewed in the World Bank surveys mention that they have separate, dedicated units to manage their relations with SMEs (Figure 8). That also seems to be a common feature among banks that are leaders in the SME business, since this characteristic is found in all the banks that the IFC interviewed in developed and developing countries. Although not displayed in our graphs, the World Bank questionnaires indicate that many of the large domestic and foreign banks--particularly in Chile and, to a somewhat lesser extent, in Argentina-- have separate, dedicated units for small enterprises (SEs) and medium enterprises (MEs). Most importantly, the units concerned with SMEs are different from consumer and corporate units, and in most cases also separate from the micro enterprise business, which banks tend to house either in a unit of its own or as part of the consumer lending unit, especially in Chile and Argentina. The dedicated business units approach SMEs in an integrated way, offering them a wide array of products and services. Almost all of the banks interviewed offer both deposits and loan products (Figure 9). It is simply not the case that banks offer primarily deposits and cash management products, avoiding the risk of lending to SMEs. The number of products offered is also relatively large, with the average number of different deposit products ranging between 5.3 and 10.6 for developed and developing countries, respectively (Figure 10). The number of credit products ranges between 9.3 and 18.7, while payment and other transactional products are between 7.7 and 16.9 for developed and developing countries, respectively. Each SME client utilizes, on average, around 5 products if we consider both deposit and credit products (Figure 11). 28 As another way to grasp the extent and types of products offered by banks to SMEs, we collected information on the products listed on the banks' websites (focusing only on the banks interviewed). This naturally is a lower bound of the number of listed products, as many products offered to SMEs are not explicitly listed on the banks' websites. Table 2 shows that the number of products and services is large. It can typically be divided into 14 categories: checking or savings accounts, investment products, term loans, credit cards, factoring, leasing, international trade financing, foreign exchange, international payments and collection, employee payments, supplier payments, tax payments, collection of receivables, and insurance products. The table shows that not only the number of products is large, but that the proportion of banks that offer this range of products is also large. On the financing side, there are more products than just the typical loans offered through relationship lending. Certainly, there is heterogeneity across countries. Of the countries interviewed, Serbia is the country with the least number of products offered by banks, followed by Colombia. The diversity of products offered is associated with the revenues that these products generate. Figure 12 displays the revenue breakdown by type of product collected by the IFC in developed and developing countries. The figure shows that credit generates only part of the revenue, 32% and 38% for developed and developing countries, respectively. The rest is divided between deposits and other products and services. In the case of developing countries, 29% corresponds to deposits and 32% to other products. This is consistent with separate evidence obtained from the World Bank surveys in Argentina and Chile. In those countries too, all types of products seem to have a significant importance in the total revenue generated in the banks' SME segment, with 29 a slightly higher proportion coming from credit products. In Argentina and Chile, credit represents, on average, 38% of the banks' SME revenue, while deposits and account management represent 25% of the revenues, and other transactions and fee-based services account for 29% of the revenues.43 The fact that SME-related bank revenues come from different sources might prompt banks to sometimes offer lending at a subsidized rate to attract SMEs to the bank and then profit from the revenues that other products and services generate. Banks use both branches and headquarters to reach out to SMEs. Headquarters typically design the strategy and the campaign in terms of which SMEs banks will target and what products they will offer them. As Figure 13 shows, the products tend to be standardized, or combined with some tailored products, with the importance of tailoring rising with the size of the firm. But it is not the case that banks use mostly tailored products. SMEs might perceive that they are offered tailored products; however those products do not tend to be tailored to a particular firm. Banks design products tailored to a group of SMEs with similar needs. For example, banks design special products for schools, fishing companies, and agricultural producers, taking into account their particular business needs, such as paying teachers, buying insurance, or getting credit to purchase inputs during the production cycle. Although the products differ across sectors, the individual SME perceives them as tailored to its specific needs. From the point of view of banks, they are frequently the same products with some type of customization, for example changing the features of a basic credit line to adjust it to the business cycle of, say, soy producers. 43These percentages do not sum up to 100% due to the fact that each value is the average of the percentages reported by banks for each product category, and some banks have not provided answers for all categories. 30 The types of products described above are different from what one would expect from relationship lending.44 Banks that can sell these products on a large scale tend to benefit the most, by having branches and SME account managers that act as the "personalized" point of contact for SMEs but that deliver mostly generic (yet somewhat customized by sector or group) products that have been planned and designed centrally, at headquarters. In Chile and Argentina many, especially large, banks have created business centers that service various branches within a geographic area. This helps them reduce costs by centralizing some functions that are subject to economies of scale, such as back-office functions. The account manager in a branch reaches out to new SMEs and manages the daily operations with existing SMEs, relying on the business center for specific back-office work. As a consequence, account managers are in high demand because their human capital is highly portable, generating job hopping and poaching. In Colombia, by contrast, the concept of a "business center" is embryonic and the figure of "SME account manager" is incipient. Under this model for engaging SMEs, headquarters have an advantage in designing the strategy of which group or sub-sector of SMEs to target. Headquarters are better equipped at taking advantage of synergies arising with "supply chains" and outsourcing arrangements, given the close connection with large corporations, which help them sort out which SMEs are worth approaching. Moreover, headquarters can design the array of products to be offered to SMEs to exploit cross-selling potential. These products 44We are not saying that relationship lending implies that only loans are offered to SMEs. US studies focusing on relationship lending by community banks show that banks tend to offer multiple products (Prager and Wolken 2007). Our point is that large and foreign banks, which are less likely to engage in relationship lending, have a comparative advantage in offering the breadth and types of products described here. Also, our interviews of banks reveal that cross-selling is at the heart of their strategy in catering to SMEs and not a by-product of the lending relationship. 31 could be advertised through centrally designed campaigns to reach out to a large spectrum of SMEs. Headquarters can also use existing databases like business registries to perform data mining and screen SMEs. Furthermore, in the case of international banks, national headquarter offices can obtain information and guidance from the global headquarter offices, who have acquired greater experience in dealing with SMEs worldwide. In this context, branches need to work with headquarters to generate new SME clients, as the relationship manager is not the only person central to the relation with SMEs and her effectiveness is boosted by information and support from headquarters. Moreover, branches do not operate as separate banks (or niche banks) within the bank. Again, in this context, universal banks have an advantage in exploiting economies of scale compared to small, specialized banks. As they learn to deal with SMEs, banks are reorganizing their credit risk management systems, with a greater degree of sophistication among international banks and the leading, large domestic banks. Figure 14 shows some aspects of how banks are organizing their risk management processes related to SME lending. In most large banks, and with the exception of pure credit scoring, credit risk management is not automated. In most cases, it involves a credit risk analyst. Typically, risk management is a function that is organizationally separated from sales and is done primarily at headquarters. The risk management department is given independence and strong approval and veto powers. (This type of arrangement is not typically found among small, niche` and public banks.) While maintaining independence in judgment, risk analysts and managers work cooperatively with those who sell products and originate loans (i.e., the SME account managers, in countries where business models are more advanced). In effect, risk analysts 32 endeavor to train SME account managers and raise their risk awareness, so that the credit approval process is streamlined and the loan application has a higher likelihood of not being rejected later on by risk analysts. The bank surveys suggest that large banks, particularly in Chile and Argentina and much less so in Colombia, use well-developed screening tools to sort out "good" debtors from the loan applicant pool. These screening tools are differentiated by firm or loan size. The size threshold for the applicability of a given screening tool is typically determined by the effectiveness of the tool itself, as gauged by repeated experience. Thus, automatic scoring methods are usually applied to small companies with small loans, for which the owner and SME information is combined (see also Marquez 2008). Moreover, back testing or statistical analyses of the effectiveness of automated scoring is used to determine the threshold size beyond which it is deemed to lose potency, although efforts are continuously made to improve the scoring technique to apply it to incrementally larger loans or firms. Streamlined and substantially standardized rating tools are used to screen larger SMEs applying for larger loans, for which automatic scoring is deemed to be not effective. Such tools include quantitative and qualitative information and are typically developed by adapting (simplifying, streamlining, and standardizing) to the SME business the rating methods applied to large corporations. SME ratings do not lead to the automatic approval of loans, but they rather provide the basis for the risk analyst to evaluate loans and decide on their approval. After loans are approved, banks continuously monitor the loans and the SMEs (the larger ones in particular), plus they have an early warning system, with triggers, to anticipate and detect potential problems. 33 Some of the larger and more sophisticated banks--particularly those interviewed in Chile and Argentina--are embarked in medium-term plans to link screening tools (automated scoring and ratings) to the banks' provision policies (for expected losses) and capital policies (for unexpected losses). They are also developing or perfecting systems and procedures to generate risk-adjusted pricing unit cost accounting per product or service line. Other plans include greater use of stress testing, quantitative analysis, and improved estimates of loss given default and post-default recovery costs. However, despite these plans, in the short-term, banks cope with the difficulties in lending to SMEs by hedging risk, using instruments like short-term loans, offering document discounting, and demanding collateral. The business models to serve SMEs described in this section can be better pursued by large universal banks, especially foreign ones, which can be more aggressive in reaching out to SME clients. These banks can better capture economies of scale and economies of scope (within and across countries) and move beyond reliance on relationship lending (which is better conducted by niche banks). Because of their substantial branch network, large universal banks are better positioned to develop low- cost approaches to give SMEs a closer, "personalized" service (or the appearance of it), without moving into costly, full-fledge relationship lending. Business centers capture decreasing costs in certain activities (including risk management and back-office functions), combined with SME account manager to "personalize" the service. Also, large and foreign banks have more large corporate clients and are well positioned to get information about the valuable SMEs with which the large corporations work, through 34 supply or outsourcing connections. This helps these banks overcome the asymmetric information problems that relationship lending tries to solve. In terms of risk management, large, universal banks are better suited to conduct lending based on automated scoring models for small loans (since they have the know- how and models to do so) and template-type rating systems for larger loans (based on streamlined, standardized versions of corporate rating). Also, they generally have more advanced risk management systems. Due to their size and to their presence in many different markets, large and foreign banks are better able to diversify away the idiosyncratic risks of SME lending. Also, they probably have advanced methods to assess the value of collateral, better recovery units, and more efficient ways to execute collateral. 5. SME Side To complement the information analyzed so far from the bank side, we present evidence on the demand side from surveys of SMEs. This rich dataset, collected independently by FRS (Inmark Group), gives a different but complementary perspective of the relation between banks and SMEs, corroborating many of the findings obtained from the bank surveys. In particular, the data show that SMEs interact with banks using a variety of products. Furthermore, SMEs do not exclusively obtain financing via "relationship loans" but also access financing products that do not depend on the bank processing "soft" information on the firm. Figure 15 shows data on the average number of "distinct" bank products and services used by SMEs in Latin American countries collected by FRS (Inmark Group) via 35 successive surveys conducted during 2002-2006. Here, "distinct" means that if an SME uses 2 products of the same type (e.g., 2 checking accounts) with the same bank or uses the same product with 2 different banks, the product will only be counted once. The figure shows that the number of banking products used by SMEs from all the banks they deal with ranges from 4.5 in Venezuela to 8.5 in Argentina. The average number of products and services across the 7 economies in the sample is 6.8. This evidence is consistent with the evidence collected by the IFC and shown in Figure 11, which describes separately the number of deposit and credit products used by SMEs, but not the number of other products and services used.45 Figure 16 provides information on the number of products used by SMEs disaggregated by product type for 2006. The top panel shows the average number of deposits and savings products used by SMEs. The middle panel shows the average number of financing products used by SMEs, while the bottom panel shows the average number of services and other products used by SMEs. On average, SMEs use 1.7 deposit and savings products, 1.9 financing products, and 4 services and other products. The numbers are relatively similar across countries, with Venezuela ranking lower in all cases. Interestingly, the number of services and other products is typically larger than the number of deposits, saving, and financing products combined. Tables 3, 4, and 5 provide more detailed information on the type of products and services used by SMEs in 2006. The tables report the fraction of surveyed SMEs that claim to be using a given bank product or service. Table 3 focuses on the deposit and 45The World Bank Investment Climate Assessments also show evidence consistent with this finding. According to these surveys of firms across countries, most SMEs have a bank deposit account, but a much smaller share use bank credit. The difference shrinks with firm size; the number of companies having bank credit rises as firms are larger. The combined evidence thus suggests that non-lending products are quite common in the relation of banks with SME clients. 36 savings products, Table 4 on the financing products, and Table 5 on the services and other products. Since the products offered by banks vary by country, not all rows have information for all countries. Table 3 shows that almost all SMEs have a current or checking account. The use of savings accounts is also relatively high (71.1% in Chile, 52.5% in Peru, and 62.9% in Puerto Rico), although it is merely 34.3% in Venezuela. Other deposits and savings products are relatively less used. But all SMEs report using some form of deposit or savings products. Overall, Chile is the country in which SMEs appear to use deposits or savings products to a larger extent. Table 4 shows that fewer SMEs use financing products than deposits and savings products. Around 40% of SMEs use term loans in Colombia and Puerto Rico and short- term loans in Chile. Lines of credit are also frequently used, with 75% of SMEs using them in Chile, while in the other countries the percentages are substantially lower (ranging between 18% for Peru and 43% for Puerto Rico). Consistent with the data collected from banks, check and document discounting appear to be important products in Argentina, with 35% of SMEs using them. Interestingly, the use of loans supported by public programs or guarantees is low. The highest usage of public programs is observed in Chile, where 8% of SMEs report using them and where the FOGAPE guarantee program is regarded as successful.46 In Argentina and Puerto Rico, only 3% of the sampled SMEs use public programs or guarantees. Again, Venezuela has the overall lowest fraction of SMEs using any type of financing products, while Chile is at the other 46For an analysis of the FOGAPE program of partial credit guarantees, see de la Torre, Gozzi, and Schmukler (2007). 37 extreme. The proportion of SMEs not using financing products at all is significant, however, ranging from 13% in Chile to 51% in Venezuela and 65% in Mexico. Table 5 shows that the use of other banking services and products is substantial; almost all SMEs use them across countries. SMEs seem to take advantage of several products including, in order of importance, payment of taxes, internet banking, insurance products, transfers, other payments (including payments to suppliers and employees), automatic debit payments, and debit cards. To a lower extent they use foreign exchange, credit card for executives, and collection of receivables. Venezuela again appears to be an outlier; SMEs there basically just use internet banking. Finally, the FRS (Inmark Group) surveys provide data that permit gauging the degree of involvement of SMEs with the banking system across countries, as measured by the percentage of SMEs that use a different number of banking products (Figure 17). FRS (Inmark Group) classifies as "involved" clients those that use between 7 and 8 products, as "valuable" clients those that use between 9 and 10 products, and finally as "star" clients those that use more than 10 products. The evidence suggests that Peru and Chile are ahead of the Latin pack in terms of SME involvement with banks; about 51% and 42% of surveyed SMEs, respectively, used 7 or more banking services or products (from 1 or more banks) in 2006. At the other extreme are Colombia and Mexico, where only 30% and 18% of surveyed SMEs, respectively, used 7 or more banking products and services in 2006. 38 6. Conclusions The evidence presented in this paper significantly questions the common wisdom that SMEs are underserved because their chronic opacity makes them substantially (if not entirely) dependent on relationship lending, for which niche banks have a natural comparative advantage. In fact, the new evidence in this paper from 12 developed and developing countries suggests that all kinds of private banks (large, small, domestic, and foreign) view SMEs as a strategic sector and are expanding or planning to expand their operations aggressively in this segment. As a consequence, the market for SMEs is becoming increasingly competitive, although far from saturated. Banks are developing new business models, technologies, and risk management systems to serve SMEs. Lending is only a fraction of what banks offer to SMEs, as banks try to serve SMEs in a holistic way through a wide range of products and services, with fee-based products rising in importance. Large banks and foreign universal banks are leading the process, capitalizing on their ability to exploit economies of scale and scope. They can lend on a large scale and provide a wide range of complementary products and services that are attractive to SMEs. They can sort out well-functioning and promising SMEs via their corporate clients with which SMEs maintain supply and outsourcing relations. Once they establish a client relationship with SMEs, large banks can use their well-established retail and consumer units to more easily extend services to the individuals (workers, owners, and their families) linked to those SMEs. Multi-service large banks can also manage risk better through diversification, better data, and more sophisticated risk management tools. International banks, moreover, can learn relatively fast from their successful experiences in SME banking elsewhere in the world. 39 The evidence presented in this paper is novel and unique. It comes from various sources, covers a diverse group of banks and countries, and captures both demand and supply-side dimensions. The different data sources show a consistent pattern across and within countries, confirming our conclusions. To be sure, we are not arguing that relationship lending is unimportant but that it is not the only way in which banks interact with SMEs, and that other interactions seem at least as or even more important. Moreover, while SMEs are having increasing interactions with banks and purchasing from them several products and services, they seem yet unable to obtain access to crucial products such as loans secured by certain forms of collateral (e.g., accounts receivable, inventories, equipment, cattle, and intangible assets) or long-term fixed-interest rate loans in domestic currency. However, it is still unclear how much SMEs in developing countries would be able to rely on banks to obtain those products. As the US literature has shown (Carey, Prowse, Rea, and Udell 1993 and Berger and Udell 1998), SMEs might have to rely on private placements and non-bank institutions. Bank financing for certain SMEs such as start-ups (in particular those in high-tech or research-based industries) is also likely to remain limited, as has proven to be the case in developed markets such as the US. Although we were able to advance our understanding of the relation between banks and SMEs, much work remains for future research. First, the collection of more data across more countries and over time would help better understand the nature of the changes in SME banking across countries, as well as the depth of the phenomenon characterized in this paper. The developing countries considered in this paper are middle- income and, thus, tend to have macroeconomic, institutional, and contractual 40 environments sufficiently developed for at least some forms of SME financing to take off. Also, in these countries, competition for corporate and retail customers seems to have reached a sufficient intensity to induce banks to target SMEs as a strategic sector. Conditions might be different in other less-developed countries, where further research is needed. However, preliminary evidence gathered by Beck, Demirguc-Kunt, and Martinez Peria (2008) for 91 banks operating in 45 countries is largely consistent with our findings.47 Second, considering the substantial investments and major changes in organizational arrangements being undertaken to serve SMEs, we have argued that this wave of banking SMEs is part of a structural change, and not just a reflection of a cyclical bonanza. However, more research is needed to ascertain if and when this new trend would hit binding constraints. The patterns we have observed cover a short time span, where the banks' intensified engagement with SMEs was in part boosted by the recent benign environment of high liquidity and rapid growth, and the associated low default rates. As conditions change, the new patterns of engaging with SMEs will be tested. Third, since the new model to engage SMEs emphasizes cross-selling a relatively wide range of products and services, more data (which were not yet produced by most of the banks interviewed under the World Bank surveys) would be needed to obtain a more reliable measure of the revenues, costs, and risk-adjusted profits associated with banks' SME business. For example, SMEs might generate revenues indirectly through the 47Their study collects information through a questionnaire similar to the one used in our paper, but that focuses only on the 5 largest banks within each country. Their questionnaire is completed via e-mail so there is no face-to-face interaction with banks and, hence, there is no way to capture anecdotal evidence or to clarify information. 41 accounts that their employees open. Also, the new banking structures with activities done both centrally at headquarters and at the branch level make it difficult to estimate unit cost accounting properly. Moreover, because new financing products are being offered with different types of risk-mitigating guarantees but at relatively low interest spreads to gain market share, it is difficult to assess the overall risks that this new engagement entails, especially considering that recent default rates are deceivingly low inasmuch as they reflect an unusually benign environment. Fourth, future research could fruitfully analyze the consequences on the banking sector, including on its structure, stemming from the new ways in which banks are interacting with SMEs. For example, if it is true that universal banks have an advantage in serving SMEs (including lending to them), the viability of small, niche banks specialized in SMEs would be in question. Fifth, the application of questionnaire-based surveys (such as those documented in this paper) to other countries would help identify policy priorities appropriately tailored to individual countries, on the one hand, and adequately informed by international experience, on the other. As an example, the surveys examined in this paper point to the importance of information and suggest that the governments might play a more active role in generating and maintaining more information on SMEs as a public good, including more information on the universe of SMEs and their financial condition and repayment history. This information could be used by the banking system not only to improve risk management but also to asses the size of the market and better target SMEs of particular economic sectors and geographic regions. However, the implications for the credit market of a new balance between public and private information on SMEs would 42 need to be assessed, including its distributional effects across banks--as banks with the best information would lose in relative terms. Moreover, evidence from surveys like these could help policymakers understand where their action is needed (and where it is not needed), and what types of institutions and regulations might need revision to foster SME financing. Finally, the question of whether SMEs are receiving "adequate" financing remains unanswered. In particular, we have not collected information on the costs of the services banks are offering. Futhermore, the data reviewed in this paper say nothing about the universe of SMEs. Hence, it may well be that the patterns documented in this paper reflect only an embryonic "bancarization" process and that, as a result, the fraction of viable SMEs being served by banks is still low. Alternatively, the SMEs that are not receiving adequate financing might simply be too risky or not have good projects for banks to finance. Put differently, the evidence presented in this paper does not dispel the commonly bifurcated perceptions that SMEs are unduly underserved, at the one extreme, and that SMEs are too opaque and/or lack bankable projects, at the other. It may be that these bipolar perceptions are outdated and that the developments documented in this paper will eventually lead to a perceptions revision. Or it may be that the perception that SMEs are underserved is dominated by the particular fact that they are not receiving long-term loans at affordable fixed interest rates. Future work is needed to have a more complete picture of how banks engage with SMEs and how that squares with the bifurcated perceptions. The type of data presented in this paper constitutes only a first step in this direction. 43 7. References Ayyagari, M., Beck, T., and Demirgüç-Kunt, A., 2007. Small and Medium Enterprises across the Globe. Small Business Economics 29, 415-434. Beck, T. and Demirgüç-Kunt, A., 2006. Small and Medium-Size Enterprises: Access to Finance as a Growth Constraint. Journal of Banking and Finance 30, 2931­2943. Beck, T., Demirgüç-Kunt, A., Laeven, L., and Maksimovic, V., 2006. The Determinants of Financing Obstacles. Journal of International Money and Finance 25, 932-952. Beck, T., Demirgüç-Kunt, A., and Maksimovic, V., 2005. Financial and Legal Constraints to Firm Growth: Does Firm Size Matter? Journal of Finance 60, 137- 177. Beck, T., Demirgüç-Kunt, A., and Maksimovic, V., 2008. Financing Patterns around the World: Are Small Firms Different? Journal of Financial Economics, forthcoming. Beck, T., Demirgüç-Kunt, A., and Martinez Peria, M.S., 2008. Banking SME Around the World: Lending Practices, Business Models, Drivers and Obstacles. World Bank. Mimeo. Berger, A., Frame, W., and Miller, N., 2005. Credit Scoring and the Availability, Price, and Risk of Small Business Credit. Journal of Money, Credit, and Banking 37, 191-222. Berger, A., Kayshap, A., and Scalise, J., 1995. The Transformation of the US Banking Industry: What a Long Strange Trip It's Been. Brookings Papers on Economic Activity 2, 155-219. Berger, A., Klapper, L., Martinez Peria, M.S., and Zaidi, R., 2008. Bank Ownership Types and Banking Relationships. Journal of Financial Intermediation 17, 37-62. 44 Berger, A., Klapper, L., and Udell, G., 2001. The Ability of Banks to Lend to Informationally Opaque Small Businesses. Journal of Banking and Finance 25, 2127-2167. Berger, A., Miller, N., Petersen, M., Rajan, R., and Stein, J., 2005. Does Function Follow Organizational Form? Evidence from the Lending Practices of Large and Small Banks. Journal of Financial Economics 76, 237-269. Berger, A., Rosen, R., and Udell G., 2006. Does Market Size Structure Affect Competition? The Case of Small Business Lending. Journal of Banking and Finance, forthcoming. Berger, A. and Udell, G., 1996. Universal Banking and the Future of Small Business Lending. In: Saunders, A., Walter I., (Eds.), Financial System Design: The Case for Universal Banking. Irwin (Richard D), Burr Ridge, IL, 559-627. Berger, A. and Udell, G., 1998. The Economics of Small Business Finance: The Roles of Private Equity and Debt Markets in the Financial Growth Cycle. Journal of Banking and Finance 22, 613-673. Berger, A. and Udell, G., 2006. A More Complete Conceptual Framework for SME Finance. Journal of Banking and Finance 30, 2945-2966. Boissay, F. and Gropp, R., 2007. Trade Credit Defaults and Liquidity Provision by Firms. European Central Bank Working Paper 753. Carey, M., Prowse, S., Rea, J., and Udell, G., 1993. The Economics of Private Placements: A New Look. Financial Markets, Institutions, and Instruments 2, 1- 66. 45 Carter, D., McNulty, J., and Verbrugge, J., 2004. Do Small Banks Have and Advantage in Lending? An Examination of Risk-Adjusted Yields on Business Loans at Large and Small Banks. Journal of Financial Services Research 25, 233-252. Cole, R., Goldberg, L., and White, L., 2004. Cookie-Cutter versus Character: The Micro Structure of Small Business Lending by Large and Small Banks. Journal of Financial and Quantitative Analysis 39, 227-251. Cressy, R., 2002. Funding Gaps: A Symposium. Economic Journal 112, F1-F16. Cull, R., Davis, L., Lamoreaus, N., and Rosenthal, J.L., 2006. Historical Financing of Small- and Medium-Size Enterprises. Journal of Banking & Finance 30, 3017- 3042. de la Torre, A., Gozzi, J.C., and Schmukler. S., 2007. Innovative Experiences in Access to Finance: Market Friendly Roles for the Visible Hand? World Bank Policy Research Working Paper 4326. de la Torre, A. and Schmukler, S., 2006. Emerging Capital Markets and Globalization: The Latin American Experience, Stanford University Press and the World Bank. DeYoung, R., 2000. Mergers and the Changing Landscape of Commercial Banking (Part II). Federal Reserve Bank of Chicago, Chicago Fed Letter 150. DeYoung, R., Glennon, D., and Nigro, P., 2008. Borrower-Lender Distance, Credit Scoring, and Loan Performance: Evidence from Informational-Opaque Small Business Borrowers. Journal of Financial Intermediation 17, 113-143. DeYoung, R., and Hunter, W., 2003. Deregulation, the Internet, and the Competitive Viability of Large and Community Banks. In: Gup, B. (ed.), The Future of Banking. Quorom Books, Westport, CT, 173-202. 46 DeYoung, R., Hunter, W., and Udell, G., 2004. The Past, Present, and Probable Future for Community Banks. Journal of Financial Services Research 25, 85-133. Frame, W., Srinivasan, A., and Woosley L., 2001. The Effect of Credit Scoring on Small Business Lending. Journal of Money, Credit, and Banking 33, 813-825. Galindo, A. and Micco, A., 2005. Bank Credit to Small and Medium-Sized Enterprises: The Role of Creditor Protection. Inter-American Development Bank Working Paper 527. Gatti, R. and Honorati, M., 2007. Informality Among Formal Firms: Firm-level, Cross- country Evidence on Tax Compliance and Access to Credit. CEPR Discussion Paper 6597 and World Bank Policy Research Working Paper 4476. Gozzi, J.C., Levine, R., and Schmukler, S., 2008. Patterns of International Capital Raisings. Brown University and World Bank. Mimeo. Hallberg, K., 2001. A Market-Oriented Strategy for Small and Medium-Scale Enterprises. International Finance Corporation Discussion Paper 40. Halling, M., Stomper, A., and Zechner, J., 2005. Capital Market-Oriented Financing Prospects for Austrian SMEs. Oesterreichische Nationalbank (Austrian Central Bank) Financial Stability Report 10, 83-94. Haynes, G., Ou, C. and Berney, R., 1999. Small Business Borrowing from Large and Small Banks. In: Williams, J.L., Rhine, A. S.L.W. (Eds.), Business Access to Capital and Credit, Blanton, A Federal Reserve System Research Conference, 287-327. IADB, 2004. Unlocking Credit: The Quest for Deep and Stable Lending. The Johns Hopkins University Press. 47 IFC, 2007. Benchmarking SME Banking Practices in OECD and Emerging Markets. Jappelli, T. and Pagano, M., 2003. Information Sharing, Lending and Defaults: Cross- Country Evidence. Journal of Banking and Finance 26, 2017-2045. Kallberg, J. and Udell, G., 2003. The Value of Private Sector Business Credit Information Sharing: The US Case. Journal of Banking and Finance 27, 449-469. Kano, M., Uchida, H., Udell, G., and Watanabe, W., 2006. Information Verifiability, Bank Organization, Bank Competition and Bank-Borrower Relationships. Research Institute of Economy, Trade, and Industry. Mimeo. Keeton, W., 1995. Multi-Office Bank Lending to Small Businesses: Some New Evidence. Federal Reserve Bank of Kansas City Economic Review 80, 45-57. Mian, A., 2006. Distance Constraints: The Limits of Foreign Lending in Poor Economies. Journal of Finance 61, 1465-1505. Marquez, J., 2008. Introduction to Credit Scoring for Small and Medium Enterprises. Banco de Mexico. Mimeo. McMillan, J. and Woodruff, C., 1999. Interfirm Relationships and Informal Credit in Vietnam. Quarterly Journal of Economics 114, 1285-1320. Miller, M., 2003. Credit Reporting Systems and the International Economy. MIT Press, Cambridge, MA. Prager, R. and Wolken, J., 2007. The Evolving Relationship between Community Banks and Small Businesses: Evidence from the Surveys of Small Business Finances. Federal Reserve Board. Mimeo. 48 Rajan, R. and Wulf, J., 2003, The Flattening Firm: Evidence from Panel Data on the Changing Nature of Corporate Hierarchies. Review of Economics and Statistics 82, 591-629. Rajan, R. and Zingales, L., 2000. The Governance of the New Enterprise. In: Vives, X., (Ed.), Corporate Governance, Cambridge University Press, Cambridge, 201-227. Schiffer, M. and Weder, B., 2001. Firm Size and the Business Environment: Worldwide Survey Results. International Finance Corporation Discussion Paper 43. Scott, J.A., 2004. Small Business and Value of Community Financial Institutions. Journal of Financial Services Research 25, 207-230. Stein, J., 2002. Information Production and Capital Allocation: Decentralized Versus Hierarchical Firms. Journal of Finance 57, 1891-1921. Stephanou, C. and Rodriguez, C., 2008. Bank Financing to Small- and Medium-Sized Enterprises (SMEs) in Colombia. World Bank Policy Research Working Paper 4481. Stoll, H. R., 1984 Small Firms' Access to Public Equity Financing. In Horvitz, P. and Richardson Pettit, R., (Eds.), Small Business Finance: Problems in the Financing of Small Businesses, JAI Press, 187-238. Stoll, H. R. and Whaley R. E., 1983. Transaction Costs and the Small Firm Effect. Journal of Financial Economics 12, 57-79. Strahan, P. E. and Weston, J., 1996. Small Business Lending and Bank Consolidation: Is There Cause for Concern? Current Issues in Economics and Finance, Federal Reserve Bank of New York, 2:1-6. Udell, G., 2004. Asset-Based Finance. Commercial Finance Association, New York. 49 World Bank, 2007a. Bank Financing to Small and Medium Enterprises: Survey Results from Argentina and Chile. World Bank, 2007b. Bank Lending to Small and Medium Enterprises: The Republic of Serbia. 50 Table 1 Description of Databases Used Source Year Countries Surveyed Population World Bank survey 2006-2007 Argentina 14 banks 2006-2007 Chile 8 banks 2007 Colombia 7 banks and 1 leasing company 2007 Serbia 8 banks IFC survey 2006 Brazil 2006 India 6 banks in total operating in Brazil, 2006 Poland India, Poland, and Thailand 2006 Thailand 2006 Australia 5 banks in total operating in Australia, 2006 Netherlands the Netherlands, the United Kingdom, 2006 United Kingdom and the United States 2006 United States FRS (Inmark Group) survey 2002, 2004-2006 Argentina 918-928 SMEs, depending on the year 2002-2003, 2005-2006 Chile 920-963 SMEs, depending on the year 2002-2003, 2004, 2006 Colombia 920 SMEs 2002-2006 Mexico 604-1,015 SMEs, depending on the year 2003-2006 Peru 920 SMEs 2002, 2004-2006 Puerto Rico 617-627 SMEs, depending on the year 2002, 2004-2005 Venezuela 908-923 SMEs, depending on the year Table 2 Bank Products Offered to SMEs This table shows the different types of products offered by banks as listed on their websites and the percentage of banks that explicitly mention each product. Banks included in the calculations are those covered by the World Bank survey. Banking Products Percentage of Banks Argentina Chile Colombia Serbia Deposit or savings products 1. Checking or saving accounts 85% 67% 83% 75% 2. Investments 77% 89% 83% 0% Financing products 3. Terms loans 85% 100% 100% 100% 4. Credit cards 85% 89% 83% 100% 5. Factoring 15% 56% 33% 0% 6. Leasing 77% 78% 17% 0% 7. International trade financing 77% 78% 17% 0% Services and other products 8. Foreign exchange 15% 56% 0% 25% 9. International payments and collection 69% 78% 33% 75% 10. Payment to employees 85% 78% 83% 25% 11. Payment to suppliers 100% 78% 67% 25% 12. Tax payments 23% 67% 50% 0% 13. Collection of receivables 77% 78% 83% 25% 14. Insurance products 46% 33% 0% 0% Table 3 Deposit and Savings Products Used by SMEs The table shows the percentage of SMEs that use each of the deposit and savings products listed. Empty cells mean that the data is not available. The data come fromsurveys of SMEs conducted by FRS (Inmark Group) in 2006. Percentage of SMEs Deposit/savings products Argentina Colombia Chile Mexico Peru Puerto Rico Venezuela Current account 100.0 86.9 100.0 95.8 89.7 98.0 84.6 Savings account 71.1 52.5 62.9 34.3 Term deposits 12.5 11.0 22.8 11.9 6.0 19.0 1.4 Mutual funds 2.0 4.1 27.9 6.7 0.9 2.2 0.5 Other investment products 1.4 2.9 4.4 5.4 0.4 11.0 0.5 None 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Table 4 Financing Products Used by SMEs The table shows the percentage of SMEs that use each of the financing products listed. Empty cells mean that the data is not available. The data come from surveys of SMEs conducted by FRS (Inmark Group) in 2006. Percentage of SMEs Financing products Argentina Colombia Chile Mexico Peru Puerto Rico Venezuela Term loans 40.5 39.2 Short-term loans 38.7 2.8 Working capital loans 40.0 Medium-long term loans 23.4 6.5 7.5 Investment loans Term loans with fixed asset guarantees 4.4 18.8 2.7 0.3 Loans supported by public programs or guarantees 2.7 8.1 3.1 0.3 Lines of credit 25.7 29.4 75.1 29.8 18.0 43.3 Overdrafts 28.8 4.3 20.6 40.0 0.7 Check/document discounting 35.4 2.7 5.1 1.4 10.1 19.7 3.4 Leasing 4.3 8.9 12.6 1.2 5.9 12.3 0.3 Factoring 1.9 1.8 7.5 1.1 1.7 0.3 Foreign trade financing 2.9 5.6 13.2 2.0 5.2 4.4 3.0 Credit card 13.9 11.2 Letter of credit 2.1 1.0 14.6 1.5 7.8 13.8 0.3 None 30.8 29.8 13.2 64.9 29.1 21.7 51.3 Table 5 Services and Other Products Used by SMEs The table shows the percentage of SMEs that use each of the services and other products listed. Empty cells mean that the data is not available. The data come from surveys of SMEs conducted by FRS (Inmark Group) in 2006. Percentage of SMEs Services and other products Argentina Colombia Chile Mexico Peru Puerto Rico Venezuela Insurance 63.1 48.3 45.0 23.5 62.3 64.7 0.0 Payment of taxes 57.2 59.7 60.1 48.7 90.9 0.7 Payment of wages 52.7 45.2 23.8 37.5 12.4 37.6 8.2 Payment to suppliers or third parties 22.5 36.9 23.6 49.7 56.0 38.2 0.0 Other payments done at branches 49.8 36.3 26.2 45.8 34.1 Internet banking 53.9 61.7 73.0 50.9 38.0 60.2 98.1 Transfer 49.8 53.2 35.6 36.1 92.0 36.5 0.6 Automatic debit 40.6 18.5 35.0 19.0 27.3 22.1 2.1 Debit card 28.6 20.3 29.2 32.2 22.2 1.1 Foreign exchange 16.4 10.0 17.1 12.7 22.0 2.1 1.6 Credit card for executives 14.0 13.5 14.6 9.6 26.2 0.1 Collection of receivables 13.3 4.9 5.8 4.7 None 2.0 5.5 4.4 5.9 0.1 3.0 0.0 Figure 1 Bank Involvement with SMEs This figure shows the percentage of banks that responded yes/no to the following question: "Does the bank currently have SMEs among its clients?" Information for this figure was gathered through bank interviews conducted by the World Bank. 100% 100% 93% Argentina Chile Colombia Serbia 100% 88% sknabfoegatnecreP 80% 60% 40% 20% 7% 12% 0% 0% 0% With SME clients Without SME clients Figure 2 Main Players in SME Financing This figure shows the SME lending market structure and the main players in it, as perceived by banks. Banks were asked "Who are the main players in SME financing?" Information for this figure was gathered through bank interviews conducted by the World Bank. 100% 100% Argentina Chile Colombia Serbia 100% 92% 88% 80% 75% sknabfoegtanecr 62% 60% 50% 54% 38% 38% 40% Pe 23% 20% 13% 13% 0% 0% 0% 0% 0% 0% 0% 0% Large banks Public banks Niche banks Other financial Small banks intermediaries Figure 3 Extent of Competition in SME Market This figure shows the percentage of banks that indicated each of the options given in the question: "How competitive is the market for lending SMEs?" Information for this figure was gathered through bank interviews conducted by the World Bank. Argentina Chile Colombia Serbia 100% 100% 77% 75% sknabfoegatnecreP 80% 63% 60% 38% 40% 23% 20% 13% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Not competitive / High Not competitive / Low Competitive / Saturated Competitive/ Not saturated entry costs entry costs Figure 4 Size and Prospects of the SME Lending Market This figure shows the percentage of banks that selected each of the statements proposed in answering the question: "What is your view on the size and prospects for the SME market in general?" Information for this figure was gathered through bank interviews conducted by the World Bank. 100% Argentina Chile Colombia Serbia 100% sknabfoegatnecreP 80% 69% 63% 60% 50% 38% 40% 31% 25% 20% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Small market/Bleak prospects Big market/Bleak prospects Small market/Good prospects Big market/Good prospects Figure 5 Bank Focus in Serving SMEs This figure indicates the extent to which the banks interviewed have a sector-specific or a geographic focus in their relation with SMEs. Information for this figure was gathered through bank interviews conducted by the World Bank. A. Does the bank have a sector-specific focus in dealing with SMEs? 100% 88% 88% 88% 90% 77% Argentina Chile Colombia Serbia 80% sknabfoegatnecreP 70% 60% 50% 40% 30% 23% 20% 13% 13% 10% 0% 0% Yes No B. Does the bank have a specific geographic focus in dealing with SMEs? 100% Argentina Chile Colombia Serbia 90% 80% 75% 75% 75% sknabfoegatnecreP 70% 62% 60% 50% 38% 40% 30% 25% 25% 20% 13% 10% 0% Yes No Figure 6 Drivers of Bank Involvement with SMEs This figure shows the degree to which bank involvement with SMEs is driven by different factors presented to the banks. The options available to qualify the importance of these factors vary from not significant to extremely significant/crucial. The figure shows the percentage of banks that consider these factors significant, very significant, or extremely significant/crucial. Information for this figure was gathered through bank interviews conducted by the World Bank. Responses for Colombia are not displayed because this question was not asked as part of the survey for this country. Argentina Chile Serbia 100% 92% sknabfoegatnecreP 80% 75% 75% 69% 63% 60% 54% 50% 50% 38% 38% 40% 31% 25% 25% 25% 20% 15% 8% 0% 0% 0% Perceived Relations with Strategic Competition Exposure Social objective profitability large clients sector in other in other segments segments Figure 7 Obstacles to Bank Involvement with SMEs This figure shows the degree to which different factors are important obstacles to banks in their exposure to SMEs. The figure shows the percentage of banks that consider each factor significant, very significant, or extremely significant/crucial. Information for this figure was gathered through bank interviews conducted by the World Bank. Argentina Chile Colombia Serbia 100% 88% 88% 88% sknabfoegatnecreP 80% 75% 75% 75% 69% 63% 63% 63% 63% 60% 50% 50% 46% 46% 46% 38% 38% 38% 40% 38% 31% 25% 23% 23% 23% 20% 13% 13% 13% 13% 0% 0%0% 0% SME- Competition Macro- Regulations Legal and Lending Lack of Bank- specific in SME economic contractual technology adequate specific factors segment factors environment to SMEs demand factors Figure 8 Organizational Structure to Serve SMEs This figure shows the percentage of banks that responded "yes" to the question: "Does the bank have a separate unit managing the banking relation with SMEs?" The percentages do not add to 100 because for some banks the information could not be obtained. Information for this figure was gathered through bank interviews conducted by the World Bank and the International Finance Corporation (IFC). Argentina Chile Colombia Serbia Developed countries Developing countries 100% 100% 100% 88% 88% 83% 77% sknabfoegatnecreP 80% 60% 40% 17% 20% 13% 8% 0% 0% 0% 0% With dedicated SME units Without dedicated SME units Figure 9 Type of Products Banks Offer to SMEs This figure shows the extent to which banks serve SMEs through deposit, loan, and transactional products. Banks were asked: "What type of involvement do you have with SMEs?" Information for this figure was gathered through bank interviews conducted by the World Bank. Data from Colombian banks was not collected during the interviews. 100% 88% 88% Argentina Chile Serbia 85% sknabfoegtanecreP 80% 60% 40% 13% 20% 8% 0% 0% 0% 0% 0% Both deposits Primarily loan products Primarily deposits and and loan products transactional products Figure 10 Average Number of Products Offered to SMEs This figure shows the average number of products offered to SMEs by banks. Information for this figure was gathered through bank interviews conducted by the International Finance Corporation (IFC). 18.7 20 Developed countries Developing countries 16.9 stcudorpforebmunegarevA15 10.6 9.4 10 7.7 5.3 5 0 Deposit products Credit products Transactional products Figure 11 Number of Products Used per SME Client This figure shows the number of deposit and credit products used per SME client. Information for this figure was gathered through bank interviews conducted by the International Finance Corporation (IFC). Developed countries Developing countries 5 tneilcrepd 4 3.3 2.8 seustcudorpforebmuN3 2.4 2.3 2 1 0 Deposit products Credit products Figure 12 Breakdown of Revenue from SME Segment by Product Type This figure shows the revenue breakdown for credit, deposits, and transactional products for the banks interviewed by the International Finance Corporation (IFC). The percentages across product types do not add up to 100 because we take averages across banks and within banks (across their small and medium enterprises units). Developed countries Developing countries 100% euneverlatotfoegatnecreP 80% 60% 38% 42% 32% 40% 29% 32% 24% 20% 0% Credit Deposits Other Figure 13 Standardization of SME Products This figure shows the percentage of banks that selected each of the answers proposed when they were asked to "Indicate the most relevant statement regarding the standardization of your SME products." The three options presented in this figure were given to banks. Information for this figure was gathered through bank interviews conducted by the World Bank and the International Finance Corporation (IFC). 93% 100% Argentina Chile Colombia Serbia Developed countries Developing countries 80% sknabfoegatnerceP 80% 63% 60% 50% 36% 31% 40% 33% 25% 29% 20% 23% 21% 17% 20% 8% 7% 0% 0% 0% 0% Mostly standardized Similar proportion of Mostly tailored products products standardized and tailored products Figure 14 Risk Management Practices This figure shows the percentage of banks that answered affirmatively or negatively to different options available regarding the structure of their credit risk management practices for the SME segment. Information for this figure was gathered through bank interviews conducted by the World Bank and the International Finance Corporation (IFC). A. Is it largely automated? B. Is it done by a credit risk analyst? 100% 100% 100% 100% 88% 88% 83% 75% 75% nks 77% 80% ba 80% 63% 60% of 60% 60% geantecreP 60% 46% 40% 40% 38% 40% sknabfoegatnecreP 40% 31% 25% 25% 17% 20% 8%13% 13% 20% 0% 0% 0% 0% Yes No Yes No C. Is it separate from sales? D. Is it done primarily at headquarters? 100% 88% 83% 100% 88% sknabfoegatnecreP 80% 77% 80% 75% 75% 75% 80% 75% 63% 69% 67% 60% 60% 38% 40% 40% 33% 25% sknabfoegtanecreP 25% 25%25% 20%17% 15% 20% 13% 15% 20% 13% 0% 0% Yes No Yes No Argentina Chile Colombia Serbia Developed countries Developing countries Figure 15 Average Number of Banking Products Used by SMEs This figure shows the average number of products used by SMEs. Data comes from a survey of SMEs conducted by FRS (Inmark Group). The data reported are averages of the annual surveys conducted over the period 2002-2006. 12 10 8.5 stcudorpgniknabforebmuN 8.0 7.7 8 6.9 7.0 5.3 6 4.5 4 2 0 Argentina Chile Colombia Mexico Peru Puerto Rico Venezuela Figure 16 Average Number of Banking Products Used by SMEs by Product Type This figure shows the average number of products used by SMEs by product type. The data come from surveys of SMEs conducted by FRS (Inmark Group) in 2006 A. Average number of deposit and savings products used by SMEs 2.5 2 1.5 1 0.5 0 Argentina Chile Colombia Mexico Peru Puerto Rico Venezuela B. Average number of financing products used by SMEs 3 2.5 2 1.5 1 0.5 0 Argentina Chile Colombia Mexico Peru Puerto Rico Venezuela C. Average number of services and other products used by SMEs 6 5 4 3 2 1 0 Argentina Chile Colombia Mexico Peru Puerto Rico Venezuela Figure 17 SMEs' Degree of Involvement with the Banking Sector This figure shows the percentage of SMEs that use a certain number of banking products or services considering all banks in the system. The data come from surveys of SMEs conducted by FRS (Inmark Group) in 2006. "Basic Client" refers to an SME that uses 1 to 3 banking products or services. "Ascendent Client" is an SME that uses between 4 and 6 products/services. "Involved Client" refers to an SME that uses 7 to 8 products/services. "Valuable Client" is an SME that uses 9 to 10 products/services, while "Star Client" is one that uses more than 10 products or services. Basic Client Ascendent Client Involved Client Valuable Client Star Client Peru 18% 21% 14% 12% 25% Chile 35% 23% 15% 11% 16% Venezuela 24% 35% 21% 8% 12% Argentina 40% 25% 14% 9% 12% Puerto Rico 34% 32% 14% 9% 12% Colombia 41% 29% 15% 8% 6% Mexico 45% 37% 9% 5% 4%