DRAFT SYSTEMATIC REVIEW OF SME BANKING AND BUSINESS REGULATION

Transcription

1 DRAFT SYSTEMATIC REVIEW OF SME BANKING AND BUSINESS REGULATION November 2013

2 CONTENTS ABBREVIATIONS... 4 EXECUTIVE SUMMARY... 5 Background... 5 SME Banking... 5 Business Regulation... 8 Business Entry... 8 Business Operations... 9 Corporate Taxation...10 Common Patterns and Gaps...10 Specific Implications for IFC...12 BACKGROUND...14 OVERVIEW OF METHODOLOGY...15 SME BANKING...17 Defining SME Banking...17 SME Banking...18 Financial Institutions...18 Effects of Accessing Finance on SMEs...23 Overall Conclusions on SME Banking...27 Overview of Gaps in Methodologies...29 BUSINESS REGULATION...30 Business Entry...30 Varying Trends in Firm Registration...30 Character of New Registrants...32 Volume and Character of Reforms...33 Effects on Firm Performance...34 Business Operations...34 Institutional Performance...34 Effects on Firm Performance...35 Corporate Taxation...37 Progress on Tax Simplification...37 Tax Registration and Informality...37 Taxation and Growth...39 Overall Conclusions on Business Regulation...41

3 Business Registration after Reforms...41 Effects on Firms Performance...43 Informality...46 Other Influences...49 Women-Owned Enterprises...51 Overview of Gaps in Methodologies...51 FILLING THE GAPS AND IMPLICATIONS FOR IFC...55 Filling the Gaps...55 Specific Implications for IFC...56 APPENDIX A Summary of Methodology...58 Defining the Areas of Focus...58 Questions and Indicators...59 Criteria for Selection of Documents...60 Approach Paper...61 Identification and Appraisal Strategy...61 Synthesis...63 APPENDIX B - REFERENCES...64

4 ABBREVIATIONS AS CCS DFIs EBRD EU FIs GDP GPFI IC ICT IDB IEG IFC IFIs IS OECD OSS RCT RIA ROA SARE SCM SIMPLES SME TFP TIN WB WBG WDR Advisory Services Compliance cost savings Development finance institutions European Bank for Reconstruction and Development European Union Financial institutions Gross domestic product Global Partnership for Financial Inclusion Investment climate Information and communication technology Inter-American Development Bank Independent Evaluation Group (of World Bank Group) International Finance Corporation International financial institutions Investment Services Organisation for Economic Co-operation and Development One stop shop Randomized control trial Regulatory Impact Analysis Return on assets Sistema de Apertura Rapida de Empresas (Mexico) Standard Cost Model Sistema Integrado de Pago de Impuestos y Contribuciones (Brazil) Small and medium enterprise Total factor productivity Tax identification number World Bank World Bank Group World Development Report

5 EXECUTIVE SUMMARY Background Small and medium enterprises (SMEs) are a key priority of the International Finance Corporation (IFC) given their importance in growing economies. To effectively support the growth of SMEs requires an understanding of which interventions can produce results and under what conditions. The Development Impact Department has been undertaking a series of meta-evaluations in strategic areas of IFC s work to address these issues. The first metaevaluation analyzed job creation effects of private sector interventions. A second metaevaluation was undertaken on private sector interventions in agribusiness and food security focusing on access to finance and farmer/business training. This systematic review adds to this knowledge base and focuses on two other key areas of IFC s work SME Banking and Business Regulation. Unlike literature reviews, systematic reviews are undertaken using strict methodological guidelines to ensure that the evidence presented is of high quality. This systematic review followed a structured process to assess the evidence of the effects of SME Banking and Business Regulation and which approaches, tools or combination of instruments produce results. The review was structured around IFC s products in order to be relevant to IFC management but it was not an evaluation of IFC s interventions or performance. The review was based on secondary sources only. A rigorous process was used to select the best research and evaluations to synthesize the experience globally in these two areas. IFC s Advisory Services (AS) and Investment Services (IS) have SME Banking as a priority for FY Working through financial institutions (FIs), IFC provides assistance to expand the FIs SME banking products and reach. Business simplification is also a priority area for AS. Working with government organizations, AS has products aimed at streamlining the burden on firms for starting and operating a business. This is done through providing assistance to governments at the national and sub-national levels in a range of areas to streamline processes. These reforms are aimed at benefiting all enterprises within the economy. Since IFC works through intermediaries for both areas, the analysis of the evidence focused on understanding the results that were emerging on two levels. One captures the changes at the FI or government level; the second focuses on the longer term changes at the enterprise level. This allows questions to be asked such as: Are FIs reaching SMEs? If so, what effect does this have on the firms performance? SME Banking Access to finance is one of the primary constraints stifling SME development. One snapshot of the extent of the constraint shows that 13.8 to 20.4 million formal SMEs 1 in developing economies may be unserved or underserved by the formal financial sector. This credit gap represents an opportunity for FIs but one that requires FIs to strengthen their current business models and expand their reach. IFC s AS and IS provide assistance to achieve this. 1 Enterprises that have 5 to 250 employees. 5

6 It was important for this review to clearly define what is meant by SME Banking in order to distinguish it from microfinance. This was done using three filters. First, the focus was on the core SME banking products namely with commercial banks. Second, care was taken to ensure that the firms covered by the review reflected the experience of SMEs not microenterprises. While definitions varied across studies, the results for small enterprises were clearly separated in the review from microenterprises. Third, where a loan size was mentioned in an evaluation or research, the review used IFC s loan proxy measure as the method to distinguish between SME Banking and microfinance. IFC has developed a proxy based on the loan size given by the FI at origination to the enterprise and covers loans above $10,000 and below $1 million or $2 million. Using these three factors, the review was able to ensure that the evidence gathered and analyzed was applicable to IFC s SME Banking AS and IS work. Two problems were encountered in undertaking the review. First, it proved difficult to find good quality evaluations and research that provided insights into the effect of lending to SMEs through financial intermediaries. Beyond stocktaking exercises and gathering of lessons, limited work has been done on systematically assessing the interventions with FIs and the conditions under which the interventions can have the largest effects. This severely limited the ability to identify the effects at the FI level and required focusing on broader research on commercial banks and SMEs. Second, by focusing on loans that fit within the SME Banking definition, most of the research that has been done to date on the effects of lending on firms was not appropriate. This was true for the growing number of randomized control trials (RCT) and all impact assessments that were covering lending to micro and small enterprises. In addition, most international research confuses finance for micro and small enterprises and combines the two under an SME Finance heading. This lack of distinction made it difficult to find clear evidence on the effects of the financial intermediation on SMEs versus microenterprises. Most fell into the latter category. As a consequence, the firm level effects from SME Banking could only be identified from research on financing constraints not from research or evaluations that tracked SMEs accessing finance through FIs. Given these constraints, a number of findings did emerge that can provide insights on SME Banking interventions. In general, the types of interventions that IFIs, including IFC, are using with FIs in terms of advisory and investment services appear to be in line with how FIs are building their SME financial portfolios. A wide range of business methods are now being used by FIs including dedicated SME units, credit scoring and innovative banking methods. This includes greater use of existing inter-firm linkages such as value chains and factoring. Many of these models were developed and tested by IFIs with financial institutions. The lack of well-done evaluations or research on the intermediation model and its effects on FIs means that it is difficult to pinpoint which interventions with FIs had an effect and under what conditions. Evaluations of intermediation would provide more information on issues such as the conditions for sustainability of SME lending and leverage that comes from IFI investment participation. This would allow a better targeting of interventions even by SME target group and methods to track performance. A wide range of banks are now interested in SME Banking based on the profitability of cross-selling products. This profitability is generated from a portfolio 6

7 of services, however, not just loans. The research reviewed showed the wide variation in the extent to which the revenues were generated from credit services ranging from a low of 22 percent to a high of 73 percent. This cross-selling of products drives the interest but also means the extent to which SMEs actually have access to credit as part of the relationship with the FIs varies widely across countries. This explains why 80 to 88 percent of the credit gap of SMEs is actually for firms with accounts with FIs. Having an existing relationship with an FI should enhance the chances of an SME accessing loans but a wide range of factors make this difficult to realize. The G-20 stocktaking report indicated that closing the credit gap for formal SMEs will be more manageable than for informal firms since a vast majority of the formal SMEs have an existing relationship with a financial institution even if they do not have credit. Leveraging this relationship may be more difficult than anticipated given the wide range of factors influencing the lending patterns. These include factors such as: the extent and character of SMEs in the economy; competition from other borrowers such as consumers, corporations and government; levels of informality within a sector; the innovation of the banks in terms of the development of methodologies such as credit scoring; presence of different types of banks within the country (e.g., foreign banks); competition in the banking sector; and the state of the financial sector infrastructure such as credit information systems and collateral registries. These conditions vary on a country by country basis and influence the extent to which an existing relationship between a FI and SME can be extended to include credit. Reforms to the financial infrastructure are becoming increasingly important since this influences the type of techniques FIs adopt for their SME portfolios. A number of recent studies have reviewed the effects of reforms in these areas. Collateral registries allow FIs to use different types of collateral such as movable assets in dealing with SMEs. Reforms have facilitated an increase in access to finance for SMEs for a range of credit products and decreased lending rates. Similar patterns are seen with the introduction of credit information systems such as credit bureaus. Here the research shows that not only do SMEs have greater access to finance but banks also increase their profitability and decrease their risk. In both cases, however, the quality of the implementation of reforms at the country level dictates whether this will take place which is an important consideration for AS. The available research focused on credit constraints of SMEs and confirmed that the lack of access to finance is a constraining obstacle to growth, investment and employment. SMEs are more likely to see finance as a key constraint to their development and more likely to be disadvantaged in economies with underdeveloped financial systems. This puts a constraint on firms growing from small to medium and affects the prospects for high growth firms. Despite this research on the credit constraints facing SMEs, little is actually known about the effects of having access to finance through FIs on SMEs. The research available did not verify the effects at the SME level in sales, productivity or investment with having access to finance nor did it explain the channels that lead to these outcomes. While the research showed a link between finance and these effects, the causality is not clear. The studies also did not adequately separate out different types of firms and differing impacts, such as on women-owned enterprises or high 7

8 growth firms. The special challenges in terms of access faced by different firms need to be investigated further. Without this work, interventions cannot be better targeted. The gaps in the evidence for SME Banking are substantial and need to begin to be addressed. There is a need to undertake more primary research and evaluations that directly look at SME banking and financial intermediation. This is particularly true around the issue of the extent that SMEs improve their performance with the availability of finance and under what circumstances. This includes separating the analysis by types of firms including women-owned and fast growing. With the growing interest in SME finance globally, it is possible that more focus can be placed on these types of issues. Business Regulation Business Entry Streamlining business registration processes is one of the most common reforms undertaken by countries to improve the investment climate. The establishment of one stop shops and revisions to regulations have resulted in substantial drops in the average time to start a business globally. One stop shops are now operational in 96 countries. This streamlining has produced some evidence on the effects of the reforms. In most cases, the number of businesses registering increases initially after the reforms are implemented. These increases have ranged from 5 percent to over 100 percent depending on the country context. However, these increased levels of registration drop off after a period of time as the backlog of firms entering dries up. The level of registrations often settles slightly above that seen before the reforms. The temporary nature of the increases reflects other factors that influence the start-up and formalization of firms beyond the specific entry reforms. Simplifying the regulations is necessary but not sufficient for new firm entry. Country specific factors come into play in the level of response by entrepreneurs. These factors include variables such as: the overall levels of entrepreneurship within the country; pent up demand in situations such as post conflict periods; and state of the economy. The approach to entry reforms has an influence on how firms will respond. Entry reforms need to have a critical mass to have an effect on registration. Small reforms defined as less than 40 percent reduction in costs, days or procedures have limited effect on registrations. Tackling one or more areas of reform has an effect. Countries with weaker environments require larger reforms. The approach taken to the reforms has an influence on the extent to which gains are seen in registrations. In some cases, there may be a temporary decline in registrations due to the way reforms are implemented. Distinctions need to be made between the types of firms registering since this reveals the prospects for growth or exit of the new firms. Difficult entry environments affect firms in sectors with low entry barriers such as retail, blocking their 8

9 entrance and sometimes forcing them to operate informally. Firms in sector that have higher entry costs such as pharmaceuticals are less affected by difficult entry environments. Easing entry requirements may draw in more marginal firms, resulting in higher exit and increased competition to the existing firms. High growth firms are encouraged with a reduction in entry restrictions but to maximize their growth potential means that the overall regulatory environment, including rule of law, needs to be conducive. Entry reforms have limited effect on bringing informal subsistence firms into formality. Firms operating informally that are more business oriented may be more draw in by the reforms. Some evidence on job creation and investments was seen from the entry of the firms. It was not clear, however, how sustainable the jobs were and the extent to which the investment levels are above the minimum capital requirements seen in many countries. Business Operations A wide range of tools has been developed for improving licenses and inspections. Countries are beginning to adopt these reforms, although they are proving complex in many cases. The evidence around licensing and inspections reform is far more limited than entry reforms, however. Few good quality evaluations or research were found in this area. The evidence that was available shows the following trends. Licensing reforms can substantially decrease the time required, costs and steps in the process. This can have a positive effect on the number of firms obtaining licenses. However, the implementation processes for reforms can be lengthy and require substantial time before registrations increase. In countries moving quickly from a highly regulated licensing environment, reforms can have more immediate and obvious effects. The classic case is India where for certain sectors setting up factories and subsequent production activities were all subject to extensive licensing requirements. The removal of the restrictions increased the number of factories and the productivity of firms. This was true for both existing and new entrants. A reduction in compliance costs has been seen with a number of licensing reforms but these are relatively small in scale. Evidence of decreases in the time required and fees paid has resulted in some cost savings for firms. However, the evidence to date shows the savings are not as dramatic as other reforms such as entry reforms and may depend on whether or not the requirements are one-off or annual. The overall effects of reforms on firms performance are not clear at this point and need further research. The Peru licensing evaluation showed no improvements in performance after obtaining a license but the study only focused on informal firms that that obtained a license not new entrants. Some of the informal firms that did enter were not interested in growing. A comparative evaluation of four African countries concluded that business licensing did not result in increases in investments. They suggest that the 9

10 outcomes may militate against placing a strong emphasis on business licensing reform in future operations. Further research is needed to better understand the spectrum of firms that respond to these reforms and the conditions under which this takes place. Corporate Taxation The review focused on a small portion of overall tax reforms, namely corporate taxes. This is an area where significant progress has been made globally in number of tax payments required, time required to comply and total tax rates. The simplification of taxes has a positive effect on firm creation and formalization but the results vary widely by type of firm. Some unofficial small firms may respond to tax reforms since it provides them an opportunity to expand their customer base by doing more advertising and issuing tax receipts. The more subsistence informal firms do not respond to incentives to register and see few benefits regardless of the extent of reforms. Some larger firms that appear to operate formally but in fact are underrepresenting sales are not enticed into full compliance even with a streamlining of the tax processes. They have found methods to circumvent the system and full tax compliance has limited influence on the ability to access credit. The results also vary by sector with sectors with low entry costs (e.g., retail) having greater compliance after reforms. Little effect was seen with high entry cost sectors. Registration for taxes can improve a firm s profitability under certain circumstances. This is particularly the case when formalization of firms allows them to expand their sales base, have more options for purchasing inputs and reorganize their workforce. A number of cross-country studies have concluded that higher tax rates result in lower investment, productivity and overall growth of firms. This research indicates that the level of tax rates has an influence on a firm s performance. Firms of all sizes are affected, although small firms have lower profitability and therefore taxes. High tax rates also stifle investments in new technology and impact manufacturing firms more than services. Recent evidence shows that tax rates may not be the biggest drag on growth but the complexity of the tax administration has a larger impact. The hardest hit region is Africa in terms of overall drag on the economies. Administrative burdens are more important for domestic firms than international investors. Tax rates are more important for the latter group. In some situations, increased tax enforcement is more effective than tax reforms in terms of increasing tax revenues. If enterprises do not think they will be found, they will avoid registration. In situations where tax inspectors have visited firms, the rates of registration have increased regardless of the reforms undertaken. Common Patterns and Gaps 10

11 A number of common patterns are seen across the three reform areas. A few are highlighted here. Different types of firms respond to reforms in differing ways. The evidence of responses to reforms varies by type of firm and sector. This is true both in terms of performance and likelihood of new firms surviving. This is an important consideration when designing reform programs with governments. While research is starting to make these differentiations, more work is needed to distinguish what reforms trigger both registrations and firm level effects, by which type of firms and under what conditions. The expectation that large numbers of informal microenterprises will formalize in response to reforms has not proven to be the case. Reforms may result in a small number formalizing that are more business oriented and see opportunities to grow. Few of the subsistence microenterprises will formalize. The benefits for this group from formalizing are few and therefore they will not be influenced by reforms. Focusing efforts at this level to bring firms into the formal sector will likely not increase registrations and bring the benefits from formalization that are anticipated by governments. Potential exists for encouraging formality with firms that are not subsistence level but more successful and larger scale. These firms are more likely to be competing with formal firms with similar characteristics and have an advantage given the lower costs. There is some evidence that reforms that lower costs and complexity of being formal will entice these firms into registering and expanding their operations. The differential effects of reforms on women-owned enterprises need to be further investigated. While project may integrate gender into their implementation, the research reviewed here provides few insights into the effects on women. Overall the quality of the research available was good but some challenges were seen that should be addressed in future research agendas. Much of the research was on the same reforms and countries, with an overrepresentation on the Americas. The varying conditions in each country and region bring into question the extent to which some of these findings can be more broadly applied. A study may have validity for one context but not be relevant for another. While a wide range of projects have been undertaken to support reforms, few evaluations have been of a quality where they can inform decision makers. IFC is at the forefront here with impact evaluations that are providing more solid evidence. Other donors should be encouraged to improve the quality of evaluations and monitoring systems. The causal links underlying much of the research need to be reassessed based on the emerging experience. Reforms will have differing effects on firms, with business entry being different from licensing or inspections. In some cases, it appears that results can be seen more quickly at a firm level, for example with business entry reforms. In other cases, the chain is more complex and the reforms may not be that important for encouraging growth and jobs. As more evidence emerges, these need to be better integrated into the approaches to interventions. 11

12 The research needs to be better disaggregate by type of firm including ownership. This includes distinguishing between new entrants, formalizing firms and firms that exit. Specific Implications for IFC This review did not evaluate interventions implemented by IFC. Instead the focus was on gathering credible evidence on the experience to date in SME Banking and Business Regulation that could inform IFC s decisions. The following are some implications of the findings for IFC s work in these two areas. SME Banking The lack of evaluations and research on the intermediation model with FIs makes it more important for IFC to further develop methods for results tracking of interventions with FIs. Currently IFC is tracking indicators such as the overall loan portfolios. Further analysis of the current indicators should begin to provide insights into trends across countries and FIs. This will allow the development of data that can be used to gain a better understanding of the effects at the financial institutions of the AS and IS interventions and conditions under which these take place. The movement by IFC to undertake impact assessments on the SME banking portfolio is important given the lack of information currently on the effects at the SME level. These assessments should look at SME banking under different conditions and by different types of firms in order to help fill in the current knowledge gaps. IFC could leverage its leadership position on the global stage in SME Finance to encourage more evaluations and research on SMEs receiving loans of a size comparable to the loan proxy. An increase in the volume of evaluation and research on different types of SMEs will provide insights into how to better target groups such as women owned enterprises and gazelles. Building on the existing relationships between FIs and SMEs can continue to be an effective technique for expanding access to credit but the conditions within a country will dictate the feasibility of the approach. Evidence is showing the complexity of factors that influence and deter an FI with an existing relationship with an SME from including credit as part of its package of services. New approaches may be needed under some circumstances by AS and IS to better address these differences. IFC can continue to play an important role in supporting the development of financial infrastructure collateral registries and credit information systems. The awareness of the importance of these reforms for allowing increased access to credit is at a high level currently. The effects from these types of interventions to date show the importance of these reforms on increasing SMEs access to finance. But the interventions need to be carefully planned since the method of implementation across countries has an effect on whether access to finance for SMEs improves. These effects are not seen across the board. 12

13 Business Regulation The business regulation reforms that are supported by IFC need to be decided based on the extent to which the outcomes of growth, jobs and investments can be achieved. As more evidence emerges, some products may be less important for achieving these results. This will require a careful tracking of what has worked, what has not and why. IFC should make clear to government partners what expectations are realistic from reforms particularly in areas such as formalization. For some groups of enterprises, other support programs may be more important for promoting formalization than business regulation reforms. Being able to reach a critical mass of reforms is important and should be integrated into the decisions on a country basis. If the potential for a critical mass of reforms is not evident, the likelihood of a response from enterprises is more limited. The causal links between certain reforms and outcomes need to be better understood. The paths from the reform areas of entry, permits and taxation to changes at the enterprise level vary. Some are more straightforward such as reforms around one stop shops. Areas such as inspection are more complex. The types of firms also influence the potential effects from a reform. Understanding the differences will allow better targeting of interventions and tracking of effects. IFC should continue to do more impact assessments that provide evidence from a broader range of reforms and geographic locations. This will begin to fill in the gaps of knowledge around the conditions under which enterprise effects may take place or not. 13

14 BACKGROUND The International Finance Corporation (IFC), along with the World Bank Group (WBG), has made small and medium enterprises (SMEs) a key priority given their importance in growing economies (Ayyagari, Demirguc-Kunt et al. 2011). Particularly important is the role played by SMEs in the creation of formal jobs. SMEs account for more than half of formal employment worldwide and in low income countries average 66 percent of permanent full time employment (IFC 2013a). Despite the potential of SMEs, they continue to face a series of constraints including access to finance, inadequate infrastructure (e.g., reliable power supply) and an investment climate that hinders both entry and expansion. IFC has a series of advisory and investment products targeting some of these constraints. To effectively support the growth of SMEs requires an understanding of which interventions can produce results and under what conditions. A growing emphasis internationally on effectiveness has generated research and evaluations that provide insights into these areas. IFC has been undertaking a range of evaluations to review these type of questions. What has not taken place to date is to have this evidence systematically analyzed. To address this missing gap, the Development Impact Department s evaluation strategy has included a series of meta-evaluations in strategic areas of IFC s work. The first meta-evaluation analyzed job creation effects of private sector interventions (IFC 2012b). A second metaevaluation was undertaken on private sector interventions in agribusiness and food security focusing on access to finance and farmer/business training (IFC 2013b). This systematic review adds to this knowledge base and focuses on two other key areas of IFC s work SME Banking and Business Regulation. 2 IFC s Advisory Services (AS) and Investment Services (IS) have SME Banking as a priority for FY Working through financial institutions (FIs), IFC provides assistance to expand the FIs SME banking products and reach. Business simplification is also a priority area for AS. Working with government organizations, AS has products aimed at streamlining the burden on firms for starting and operating a business. This is done through providing assistance to governments at the national and sub-national levels in a range of areas to streamline processes. These reforms are aimed at benefiting all enterprises within the economy, not just SMEs. This systematic review analyzes findings from research and evaluations between 2000 and 2013 that can provide evidence for SME Banking and Business Regulation on what has worked and not worked in the past and why. The intention is to provide a systematic assessment to assist IFC in making decisions, developing strategies and designing products that are based on sound evidence. 2 The original intention was to undertake a meta-evaluation but this proved difficult due to the lack of good quality evaluations for SME Banking and Business Regulation. The scope was therefore expanded to include other high quality research that could inform the review questions. 14

15 OVERVIEW OF METHODOLOGY Systematic reviews have proven to be an effective method for identifying what works and does not work internationally. Unlike literature reviews, systematic reviews are undertaken within strict methodological guidelines to ensure that the evidence presented is of high quality. They methodologically map out available evidence, critically appraise it and synthesize the results. This systematic review followed a structured process that is summarized in Appendix A. The objective of the systematic review was to assess the evidence of the effects of SME Banking and Business Regulation and which approaches, tools or combination of instruments produce results. The review was structured around IFC s products in order to be relevant to IFC management but it was not an evaluation of IFC s performance. The review was based on secondary sources only and a rigorous process was used to select the best research and evaluations to synthesize the experience globally in these two areas. The key questions for the systematic review were: 1. What types of IFC-related work with financial institutions to increase provision of financial services to SMEs (SME Banking) are most likely to have the largest effects? The effects were measured by performance indicators such as changes (i) at the FI level in number of loans disbursed to SMEs, number or value of outstanding loans, profitability of the SME portfolio at the FI and average nonperforming loan rate and (ii) growth of SMEs in terms of sales/revenue, assets, productivity and number of employees. 2. What types of IFC-related public sector interventions aimed at reforms in Business Regulation, are most likely to have the largest effects? The effects were measured by performance indicators such as changes (i) at the institutional level in areas such as number of businesses registering, number of businesses benefitting from reforms (registration, licensing, permits, taxes) and average number of days/procedures to comply with regulation and (ii) at the enterprise level in areas such as changes in costs to comply, levels of investment, revenue, productivity, employees and formalization. 3. What methodologies/approaches have been used to conduct the research and evaluations? Were they appropriately applied and relevant to the subject? 4. What lessons, conclusions and recommendations can be extracted from the research and evaluations that can inform management decision making? As noted above, for both SME Banking and Business Regulation, IFC works with intermediaries, with enterprises being indirect clients. For this reason, the analysis of the evidence focused on understanding the results that were emerging on two levels institutional in terms of the direct partners with which IFC s programs operate (e.g., FIs and governments) and indirectly with enterprises. One captures the changes at the FI or government level; the second focuses on the longer term changes at the enterprise level. This allows questions to be asked such as: Are FIs reaching SMEs? If so, what effect does this have on the firms performance? Figure 1 summarizes the indicators that were used for the systematic review. The institutional indicators are based on ones that IFC s SME Banking and Investment Climate (IC) groups are tracking as part of their results measurement frameworks. The firm level results are based on the theory of change that takes place when the FIs provide products to SMEs or the government 15

16 reforms the regulatory environment. As a consequence, they differ between the two areas under review. The review looked at the evidence at both of these levels and identified the process of moving from institutional indicators to enterprise to extract lessons on what factors produced effects at both levels. Figure 1 SME Banking FIs provide financial services to SMEs Business Regulation Governments reform business registration, regulation and taxation Indicators to be Assessed at Institutional/Government Level # Loans disbursed to SMEs # or value of outstanding loans Profitability of the SME FI Average nonperforming loan rate # of businesses registering # of businesses benefitting from reforms (registration, licensing, permits, taxes) Average # of days/procedures to comply with regulation Indicators to be Assessed at Enterprise level SMEs access financial products resulting in increases in: Sales/Revenue Assets Productivity Employees Investment Firms benefit though: Reduced costs to comply Increased investment Increased revenue Increased productivity Increased # of employees Formalization Source: Institutional/government indicators from various IFC results measurement documents Strict criteria were established in selecting the research and evaluations to ensure that only good quality work was included. 3 Many documents were eliminated since they did not provide solid evidence. The focus of the review on the indicators listed above acted as a key filter in the selection process. Therefore, many documents were excluded from the review that might be included in a broader literature review. The evidence from the resulting documents is presented in this synthesis. It should be noted that the limited number of impact evaluations or even good quality evaluations, particularly for SME Banking, meant that greater reliance had to be placed on methodologies such as crosscountry, country level and firm level studies. The methodological issues and gaps with the types of research that were available for each area is covered below. 3 Good quality means that the evaluations and research had an appropriate methodology, addressed the key questions, demonstrated a linkage between the evidence and the conclusions and clearly presented the findings, conclusions and recommendations. 16

17 SME BANKING While it is recognized that access to finance is one of the primary constraints stifling SME development, the extent of the credit gap remains unknown. Some research is beginning to provide some insights. One snapshot of the extent of the constraint shows that 13.8 to 20.4 million formal SMEs 4 in developing economies may be unserved or underserved by the formal financial sector (Stein et al. 2013). Another analysis shows the complexity of the gap distinguishing between voluntary versus involuntary exclusion from using loans (World Bank 2014). Based on Enterprise Surveys, less than a quarter of SMEs in low income countries applied for loans. Of those that did not apply, 31 percent did not need a loan and 44 percent were excluded by factors such as interest rates, complex application procedures and collateral requirements. To response to this need, SME Banking is an area where both IFC s Advisory Services and Investment Services have products that are targeted to SMEs. In both cases, all the work is done through financial institutions. IFC s Advisory Services provide support to FIs to assist them in implementing a strategy to serve the SME sector. This could include a range of services such as: assistance with standardizing products and processes; optimizing deliverycost of products and services through appropriate delivery channels; use of credit scoring and rating tools for risk appraisal, underwriting, process streamlining; risk-based pricing; and leveraging technology to reduce the cost of services. IFC s Investment Services provide financing to financial institutions in various forms depending on client needs and IFC product offerings. This can include loans, equity investments, quasi investments and risk management products. Defining SME Banking It was important for the purposes of this review that SME Banking was clearly defined in order to distinguish it from microfinance. Three factors were taken into account when reviewing evaluations and research to ensure that the review reflected IFC SME Banking products. First, the review focused on financial products through FIs, specifically commercial banks. Other areas such as equity funding, leasing and trade finance were not covered. 5 The criteria of focusing on commercial banks and their SME finance products was an important filter for determining applicability of research and evaluations. Second, IFC has a broad definition of micro and small enterprises as shown on Table 1. An enterprise qualifies as a small or medium enterprise if it meets two out of the three elements in the definition below number of employees, total assets or total annual sales. As was anticipated, most of the evaluations and research used slight variations in definitions from the IFC one. Given these variations, care was taken to ensure that the firms covered by the review reflected the experience of SMEs not microenterprises with definitions in various studies being noted as applicable. 4 Enterprises that have 5 to 250 employees. 5 See Appendix A for the rationale of what is included and why. 17

18 Table 1 IFC Definition of Micro, Small and Medium Enterprises Elements Proxy Indicator Employees Total Assets US$ Total Annual Sales US$ Loan Size at Origination Micro < 10 < $100,000 < $100,000 < $10,000 Small $100,000 - < $3 million $100,000 - <$3 million > $10,000 - $100,000 Medium $3 million $15 million $3 million $15 million >$100,000 - < $1 or 2 million 6 Source: IFC July 2013 Third, for SME Banking, IFC uses a proxy measure to define its reach. This is required since few financial institutions collect data in a consistent form that would allow an analysis by factors such as number of employees or assets across FIs. IFC developed a proxy based on the loan size given by the FI at origination to the enterprise as shown in the last column of Table 1. The veracity of the proxy measure was verified through an analysis of 1,978 enterprises that received SME loans from 34 IFC client FIs (Ernst and Young 2013). These FIs were located in 25 countries in six regions. The conclusion was that the size of loans is a reasonable proxy for the size of the recipient enterprise in terms of employees, sales, and assets. IFC is using this proxy measure in tracking its lending to SMEs through financial intermediaries. Where loan sizes were mentioned in evaluations and research, the proxies were used as a filter, eliminating evidence based on small loan sizes (e.g., $1,000). Using the proxy as a filter was not possible across the board, however, since much of the work did not include the loan sizes covered. This review of SME Banking used these three filters as a guide for separating SME Banking from Microfinance. This ensured that the evidence gathered and analyzed was applicable to IFC s SME Banking AS and IS work. By separating SME Banking from microfinance, however, this eliminated much of the research that has been done to date especially the randomized control trials (RCT) and all impact assessments that were covering lending to micro and small enterprises. Most international research confuses finance for micro and small enterprises and combines the two under an SME Finance heading. This lack of distinction between micro and SME made it difficult to find clear evidence on the effects of the financial intermediation on SMEs. In fact, the firm level effects could only be identified from research on financing constraints not from research that tracked SMEs accessing finance through FIs. SME Banking Financial Institutions International financial institutions (IFIs) and development finance institutions (DFIs) have been actively involved in supporting SME banking through financial institutions and substantially expanded the outreach to SMEs (IFC 2010b and IFC 2011b). Stocktaking exercises have identified the extent of outreach that has been achieved including the emergence of global and regional funds that allow a broader scale of coverage of SMEs. The common approach is to provide both technical assistance and investment funding (e.g., loans, 6 US$2 million for more advanced countries, including Argentina, Brazil, Chile, China, Colombia, India, Korea, Mexico, Morocco, Peru, Russia, South Africa, Thailand, Tunisia, Turkey and all EU accession countries (Poland, Hungary, Czech Republic, Baltics and Slovenia. 18

19 equity) to FIs. The combination is aimed at ensuring that the products established at the FIs are sustainable and the banks continue to offer them to SMEs. This has included the development of new delivery models for banks. Specific programs and IFIs have tracked their portfolios demonstrating some of the institutional indicators under review here such as loans disbursed to SMEs. For example, the European Union (EU) and European Bank for Reconstruction and Development (EBRD) SME Finance Facility created in 1999 provided loans to FIs and technical assistance to develop sustainable SME finance (IFC 2010b). To date, over 100,000 loans have been provided to SMEs with an average individual loan size of $30,000, and a total of $3.3 billion in SME financing facilitated. FIs are graduated from the program after five years and the majority have continued to provide SME finance. Despite this wide range of interventions by IFIs and DFIs, only one agency evaluation has been done that systematically reviews this intermediation model with financial institutions. The Independent Evaluation Group (IEG) assessed IFC s experience with financial intermediaries in frontier countries in Reviewing 36 FIs that were providing support to SMEs, the evaluation concluded that using FIs as intermediaries allowed a large number of enterprises to be reached. It resulted in strong performance by the FIs. The average return on equity at 19.4 percent for for the FIs was higher than the 15 percent considered to be satisfactory for the banking industry in developed countries. Average return on assets (ROA) was 1.8 percent. The equity funding provide by IS allowed the FIs to expand their lending and maintain prudent capital adequacy ratios. A new evaluation is currently underway in 2013 by IEG on targeted support to SMEs that includes banking. This could provide further evidence of the potential for IFIs to intermediate access to finance and lessons emerging. With the lack of high quality evaluations on the intermediation model, limited evidence is available on the effect of lending to SMEs through financial intermediaries. Beyond stocktaking exercises and gathering of lessons, IFIs have not systematically assessed their interventions and the conditions under which they can have the largest effects on SME s access to finance. The insights into the performance of FIs in areas such as profitability must then come from the research on banks and why they want to engage SMEs. From the commercial banks perceptive, interest has been increasing in SME banking as domestic and foreign banks have realized that SME banking can be profitable. One of the first studies to identify the extent to which banks viewed SMEs as profitable was Beck et al (2008). Using data from 91 banks in 45 countries, they found that banks perceived the SME segment as very profitable. In developing countries, 72 percent of the banks indicated that profitability was the most important reason for involvement with SMEs. Banks were establishing dedicated units to serve this segment and overall 32 percent of loans were going to SMEs. A 2010 survey of banks indicated that serving SMEs was more profitable than serving large companies (WEF 2012). The average pre-tax return on regulatory capital in emerging markets was 44 percent for small enterprises, 41 percent for medium and 33 percent for large enterprises. While SME banking is profitable for commercial banks, it is not necessarily through the provision of credit but from a spectrum of services. Cross-selling of products is a key part of the commercial bank s strategy for SMEs. Only a small portion of revenue may come from lending products. Other deposit and transactional products generate the rest. This explains why 80 to 88 percent of the value of the credit gap for formal SMEs in emerging markets comes from SMEs with a bank account (Stein et al. 2010). 19

20 Table 2 shows the wide range of revenue from SME banking that is derived from credit services. De la Torre et al (2010) identified the extent to which FIs saw credit services as only a part of an overall package of fee based products that the banks could provide to the SMEs. A recent study of four African countries highlights the wide variations by country to these patterns (Berg and Fuchs 2013). The contribution of SMEs to banks net income was higher than the share of SME lending in all countries, underlining the high profitability of SME portfolios overall. Table 2 Proportion of SME Banking Revenue from Credit Services Study Methodology Location % of SME Banking Revenues from Credit Services De la Torre et al. (2010) Survey of 48 banks in 12 countries Developed countries 32% Berg and Fuchs (2013) Survey of Latin American SMEs Survey of 62 commercial banks in four countries Developing 39% Countries Kenya 68% Nigeria 22% Rwanda 71% South Africa 27% Ayyagari et al (2012) used Enterprise Surveys from to further review this issue. They found that approximately 84 percent of SMEs had bank accounts but access to overdraft facilities ranged from slightly over one percent in Guinea Bissau to 99 percent in Brazil. The percentage of SMEs with access to a line of credit or loan ranged from 3 percent in Afghanistan to 70 percent in Peru. Levels of banks accounts, lines of credit and overdraft access for SMEs increased monotonically from low to high income countries. The existing relationships of SMEs with FIs, through other non-lending products, make the credit challenge of formal SMEs easier to tackle than firms without any contact with FIs (IFC 2010b). These clients could be an entry point for banks to expand their credit products lines. However, leveraging this relationship may be more difficult than anticipated. A wide range of factors influence the revenue models of FIs and their decisions about extending credit. For example, the spread in the percentage of revenue from credit seen on Table 2 among the four African countries is due to a series of factors including the extent of SMEs in the economy, extent of Government borrowing, the degree of innovation mainly as introduced by foreign entrants to financial sectors and the state of the financial sector infrastructure such as credit information and collateral registries (Berg and Fuchs 2013). Competition from other market segments also plays a role both negatively and positively in the extent that SMEs can access finance. The bank surveys used by de la Torre et al. (2010) showed that SMEs were becoming more attractive credit clients as the banks were facing increasing competition in other segments of the market such as large corporations. This was true with large and foreign banks, not just small and niche banks. Bratanovic (2012) looked at the experience of eight banks in Serbia that were foreign subsidiaries from the EU. The banks were rapidly expanding their SME lending because the profit potential for working with SMEs was high, again due to cross selling of other products. Data from the Superintendent of Banks in Peru were used by Nieder (2007) to assess SMEs access to credit through FIs. In 2006, SMEs were receiving 45 percent of business lending but their market share of loans 20

21 began to drop with the increasing profits being seen from consumer lending and a growing microenterprise business for banks. Both of these latter groups were charged higher interest rates for loans, allowing larger spreads for the FIs and decreasing the portion of lending to SMEs. The key for FIs to successfully undertaking SME financing is adapting the delivery models normally followed. Research indicates that banks are starting to increasingly adopt new approaches in areas such as screening, monitoring and risk management. Many of these have been advocated and modelled by IFIs. The banks are developing new technologies such as credit scoring and new products such as asset based lending that allow them to effectively deal with SMEs (de la Torre et al. 2010). Factoring and reverse factoring can play an important role. Factoring is a type of asset based financing where a borrower s accounts receivable is used as the underlying asset to secure a loan. The decision for financing is based on the quality of the receivable not the borrower. In reverse factoring, the bank makes payment to the supplier based on invoices that a buyer has qualified basically assuring they are guaranteed. Klapper (2006) concluded that reverse factoring was an effective solution to weaknesses in credit information systems. Inter-firm linkages and value chains are increasingly being used to facilitate access to finance by FIs. The link of SMEs to FIs can be based on the larger firm in the value chain providing references for the SME to FIs, confirming future cash flow or orders with the SME that can be used as collateral for lending or with the larger firm providing a guarantee to the FI. De la Torre et al. (2010) found that some of the larger banks in the 12 countries surveyed were using their long term relationships with corporations to reach the corporate SME suppliers or customers. The knowledge of the corporations reduced substantially the problem of asymmetric information that banks faced when approaching new SMEs. Similar arrangements were seen in Serbia where FIs were using major customers and suppliers to identify SMEs and reduce the FIs risk (Bratanovic 2012). Not all value chains are conducive to facilitating access to finance, however. Navas- Alemin et al. (2012) used enterprise level data to review the role of value chains in financing in agro-industry in Argentina, furniture in Brazil and information and communications technologies (ICT) in Costa Rica. The three case studies provided insights into the conditions under which value chains can facilitate finance for SMEs with FIs. Variations were seen depending on how the value chain operated. The nature of the ICT value chain did not make it conducive to improving the SMEs access to finance. Customers were rarely repeat and the work in progress had limited value unlike manufacturing. The agro-industry linkages, however, did facilitate financing given the long term arrangements between the smaller suppliers and the larger firm. The reputational effect of the larger firm made SMEs more attractive to FIs and in some cases the larger firms were willing to guarantee the loans. The small furniture suppliers were able to leverage their accounts receivable from a large well known firm into a loan with a bank. The relationship between the suppliers and the SMEs and the governance structure between them dictated whether and how the finance could be facilitated through value chains. The models the FIs are using fit the environment within which they work and the financial infrastructure that is available. For example, Wells Fargo in the United States relies heavily on transaction lending through various routes such as using technology and credit scoring systems and leveraging credit bureau information (IFC 2010b). ICICI Bank in India faces weak financial infrastructure and must rely more on relationship lending. 21

22 The importance of reforms in financial infrastructure is being widely recognized. Reforms have begun to be undertaken in developing countries in collateral registries. Many SMEs are constrained by the range of assets that can be pledged for a loan. Movable collateral such as inventories, machinery and accounts receivable are often not widely recognized by financial institutions. Yet 78 percent of the capital stock of business is in movable assets not in immovable assets such as land or real estate (Alvarez de la Campa 2011). Collateral registry reforms are increasing globally and are showing strong effects on SMEs access to financing. Love et al. (2013) reviewed firm level surveys in 73 countries to explore the effect of introducing collateral registries for movable assets on firms access to finance. Using a difference-in-difference approach, they reviewed access pre and post introduction of movable collateral registries in seven countries against three control groups. 7 They found that the introduction of registries increased access to finance (either loan, line of credit or overdraft) by almost 8 percentage points and access to loans by 7 percentage points. The effect is larger among small firms (5-19 employees). The introduction of a movable collateral registry increased the percentage of working capital and fixed assets financed by banks by 10 and 20 percent respectively. It had an impact on interest rates (a 3 percentage point reduction) and loan terms (6 month extension of the maturity of a loan). The small sample of reformers in the research, however, meant that further work is needed to test differences among countries. An evaluation of an IFC project supporting reforms to China s secured transactions systems showed substantial increases in access to finance for SMEs (Dalberg 2011). The reforms triggered an increase in movables lending by an average of 21 percent annual growth rate between 2008 and 2010 for SMEs. In addition, 63 percent of SMEs that obtained new loans using accounts receivable as collateral had female ownership, with 20 percent being majority owned by women. The introduction of credit bureaus and systems for information sharing is having an influence on access to finance. The opaqueness of SMEs in terms of information and transparency has often been cited as an issue in terms of FIs being able to assess creditworthiness. Women owned enterprises, in particular, are disadvantaged when there is a lack of credit information systems (IFC 2011a). Brown et al. (2009) used panel data from 24 transition countries to assess the effect that information sharing had on the availability and cost of credit for firms. They found that increased information sharing was associated with improved credit access for firms and lower cost of credit, particularly in countries with weak creditor protection. The credit access was stronger for opaque firms than transparent ones that rely on external auditors and have adopted international accounting standards. Banks appeared to benefit from access to credit information. Houston et al (2010) reviewed data on nearly 2,400 banks in 69 countries. They found that information sharing lead to higher bank profitability and lower bank risk. Better credit information also significantly mitigates the negative effects of low competition or high concentration of banks within a country, facilitating greater access to finance (Love and Martinez-Peria 2012). 7 The three control groups were: countries that had not implemented collateral reform; countries that implemented other reforms besides movable collateral registries; and countries that matched the location and income per capita of the countries introducing movable collateral registries. 22

23 The effects from implementation of collateral registries and credit information systems are not uniform across countries. One study on collateral registries indicated that substantial differences can be seen between countries in terms of implementation that can have an effect on the improvement in access to finance by SMEs (Alvarez de la Campa et al. 2012). The level of bank competition and concentration influences the extent to which voluntary credit information sharing systems are effective (Bruhn et al. 2013). Mandatory credit registries are not influenced by bank competition or concentration. Effects of Accessing Finance on SMEs Evidence of the effects of lending by FIs on SMEs in the category of loans covered in this review is very limited. Unlike microcredit that has a wide range of impact and other evaluations that review the effects of the interventions, this is basically non-existent when looking at SME Banking. No project level evaluations of work with FIs were found that were of good enough quality to provide reliable information on SMEs performance as a consequence of their access to finance. Most evidence was unsubstantiated or used methodologies that were not completely sound. The research done on the effects that access to credit has on SMEs growth, productivity and investment has been primarily cross country or country level studies focusing on credit constraints not evaluations or assessments of specific interventions by FIs. While this research is informative, it does not provide evidence of actual effects on firms. The studies are not always clear on the causal links in order to be able to determine why the differences exist. They do not deal with loan sizes, only firm sizes in most cases. These, and other, methodological issues are discussed further below. This research does, however, provide some insights into the potential effects on SMEs from accessing finance. A summary of the effects identified in the research is shown on Table 3. Table 3 Summary of Effects on Firms from Research Country Study & Methodology Database Potential Effects Cross country (98) Dinh et al World Bank Enterprise Access to loan or overdraft increased growth of employment by 3.1 percent Regression using growth diagnostics Surveys panel data Access to sales credits increased growth of employment by 2.6 percent approach of Hausmann, Access to investment funds increased Rodrik, and Velasco growth of employment by 4.2 percent Biggest gains were for small firms-loans increase employment by 9 percent Young firms (less than 5 years) increased employment by 9 percent by having loan or overdraft Mexico Love and Sánchez 2009 Surveys of 2,027 Removal of credit constraints would result in: enterprises 3-9 percent increase in the number of Regression analysis enterprises investing using survey data to measure constraints 4-19 percent higher investment to capital ratio Indonesia, Vietnam Shinozaki 2012 Regression using Ordinary Least Squares (OLS) method World Bank Enterprise Surveys 2009 Indonesia 1 percentage point credit increased resulted in sales value increased 0.76 percentage points for large firms and 0.58 for small firms. Vietnam - 1 percentage point increase in 23

24 Country Study & Methodology Database Potential Effects credit resulted in sales value increases by 0.40 percentage points in large firms and 0.68 percentage points in SMEs. Bulgaria Gatti and Love 2006 Survey of 548 enterprises Total factor productivity and credit are positively associated Pooled OLS regression No effect of credit access on sales per Cross country (90) Cross country (54) Aterido et al OLS regression Beck et al 2005 OLS regression World Bank Enterprise Surveys panel data World Bank Business Environment Survey panel data fixed capital Positive effect from access to finance on employment growth for medium and large enterprises No significant effect from access to finance on small (11-50 employees) Financing constraints reduced growth by 10 percentage points for small firms (5-50 employees) compared to 6 percentage points for large Cross country (116) Kuntchev et al Ordered logit regression Brazil D Erasmo 2013 Uruguay Firm dynamics model Gandelman and Rasteletti 2013 Econometrics model based on unbalanced panel World Bank Enterprise Survey database Firm level survey Brazil Firm panel data , Economic Activity Survey Uruguay SMEs are more credit constrained than large firms Firms from 5 to 9 years old were not more likely to be credit constrained than older firms Higher performance firms are less credit constrained across firm sizes 37 percent reduction in cost of funds for FIs and 44 percent reduction in cost issuing loans associated with a 15.4 percent increase in aggregate total factor productivity One percentage point increase in overall credit growth translated to one half percent increase in investment rate A number of these studies have concluded that smaller firms are more credit constrained and having access to finance would translate into increased performance. Dinh et al. (2010) analyzed Enterprise Surveys from covering 98 countries and looked at the variations in effects of access to credit by firm size. Overall, they found that having a loan or overdraft facility increased the growth of a firm s employees by 3.1 percent. Having sales credit increased employment by 2.6 percent and external investment funds increased growth by 4.2 percent. They found that micro (1-10 employees) and small (11-50) were the most credit constrained and therefore benefitted the most from gaining access to credit. Young firms (less than 5 years) increased their employment by 9 percent by having a loan or overdraft, while mature firms (5-15 years) only increased employment by 6 percent. They conclude that financing constraints play a role in the failure of small firms to grow into medium size firms. A study by Love and Sánchez (2009) reviewed the credit constraints on registered enterprises in rural areas in Mexico. The firms covered had more than 6 employees in agribusiness and fishing. Enterprises with more assets, longer time in operation and with exporting experience were more likely to have formal loans. They found that the incidence of investment was higher among enterprises that had access to financing from formal lenders with this group being 6.4 percentage points more likely to invest. Removing the credit constraints would result in

25 percent increase in the number of enterprises investing and in 4-19 percent higher investment to capital ratio. Shinozaki (2012) found that in Indonesia and Vietnam the funds raised by SMEs through external financing directly correlated with an increase in assets such as machinery, equipment and buildings. In Indonesia, if credit increased 1 percentage point, sales value increased 0.76 percentage points for large firms and 0.58 for small firms. In addition, in Vietnam, a 1 percentage point increase in credit resulted in sales value increases by 0.40 percentage points in large firms and 0.68 percentage points in SMEs. Gatti and Love (2006) reviewed the experience of Bulgarian firms and the impact of access to credit on their productivity. A survey in 2004 covered manufacturing firms (60 percent) and the service sector (40 percent) and firms from micro (up to 10 employees) to large (over 100 employees). They found a strong correlation between credit and total factor productivity across firms. The impact of credit on sales per fixed capital was not significant, however. The cost of credit can have an influence on firm entry and productivity. D'Erasmo (2013) reviewed the link between credit conditions, level of formalization of firms and firm size distribution in Brazil. Conditions within the country changed from 2003 to 2010 with substantial expansion of credit to Brazilian firms and a decrease in the cost. D Erasmo studied the effect of a 37 percent reduction in the cost of funds for FIs and a 44 percent reduction in the cost of issuing loans comparable to what actually took place. The reduced costs lowered the entry threshold for firms to move into the formal sector and this resulted in more entrants from the informal sector, more competition in some sectors, a reduction in the average firm size and more exits. Overall, however, the increased access to funding and lower cost increased aggregate total factor productivity (TFP) by 15.4 percent and weighted firm level productivity by 16 percent. The potential effect of increases in credit varies by the character of the firm. Aterido et al. (2011) reviewed data on enterprises in 90 countries and found a positive effect on employment growth from access to finance for medium and large enterprises but no significant effect among micro (less than 10 employees) and small enterprises (11-50). Their explanation for this result was the character of the micro and small enterprises. Those firms that were entrepreneurial and productive could grow; those with less talent did not. The additional cost of financing further hurt those firms that were less productive by raising input costs. The overall increase in access to finance also increased the competition the firms face. Reviewing data from 4,000 firms in 54 countries, Beck et al. (2005) found that financing constraints affected firms differently depending on size. Small firms (5-50 employees) and medium firms (51-500) were more affected by a series of factors related to finance including collateral requirements, bank paperwork, and high interest rates than large firms. Financing constraints reduced growth by 10 percentage points for small firms but only 6 percentage points for large. Another study concluded that higher performance firms, measured by labour productivity, were less likely to be credit constrained (Kuntchev et al. 2013). This was true across firm sizes, although it was weaker for small firms than medium and large. They also concluded that firms between 5 and 9 years old were not more likely to be credit constrained than older firms. The extent of the constraint by sector varied across regions. For example, in East Asia and the Pacific manufacturing firms were more credit constrained while in South Asia it was the retail firms. 25

26 The extent of informality in a sector can have an influence on access to finance even for formal firms in the sector. Gandelman and Rasteletti (2013) reviewed panel data on firms between 1997 and 2008 in Uruguay. They found that a one percentage point increase in overall credit growth translated into a one half percent increase in the investment rate. They then reviewed whether the extent of informality present within a sector affects firms investment decisions and their access to finance. They found that the formal firms decisions on investment within a sector were not directly affected by the extent of informality in competitors. They did find an indirect effect in terms of the perceptions of lenders towards the sector. If a sector had higher degrees of informality overall, there was less credit available. A negative perception toward a particular sector can have a spillover effect to banks when they are assessing all firms in the sector, even formal enterprises. Women-owned SMEs face additional barriers to accessing finance. A range of studies have now been done on women entrepreneurs, documenting in detail the issues they face in access financing and operating their businesses (IFC 2011a, Powers and Magnoni 2010, Klapper and Parker 2011, Hallward-Driemeier 2013, OECD 2012). The evidence clearly shows additional obstacles faced by women in starting and running their businesses including less access to credit. The characteristics of the firms, which are often very small, informal and concentrated in less productive sectors, explain some of the difference in access to finance (IFC 2011a and Hallward-Driemeier 2013). The difficulty is that the research on financial constraints facing female entrepreneurs currently does not distinguish between women-owned microenterprises and womenowned SMEs. What information is available suggests that the current research provides an incomplete picture of the issues facing women-owned SMEs. The Enterprise Surveys indicate that access for women owned SMEs is more equitable (IFC 2011a). When the environment is conducive to business, male and female owners of SMEs, particularly larger and formal firms, have similar rates of borrowing. A study of women entrepreneurs in Latin America reviewed credit practices of banks and found lending to SMEs showed no difference between women and men in terms of loan size for SMEs (Powers and Magnoni 2010). On the other hand, loans for microenterprises were sometimes 50 percent smaller for women than for men. In Africa, similar patterns are seen where financial access depended more on the size of the firm and registration status than gender of the owner (Hallward-Driemeier 2013). The difficulty is that women-owned SMEs overall represent a small share of the formal SMEs. Baseline estimates done by IFC (2012a) of the client base of its partner banks showed that overall for loans between $10,000 and $1-2 million, the representation by women-owned SMEs was 22.4 percent. 8 There were wide variations by regions with East Asia and Pacific having 26.6 percent women-owned SMEs and Middle East and North Africa 5.7 percent. Some constraints facing women may be the same across firm sizes such as legal or regulatory barriers in land or property rights. Others may be very different in terms of dealing with FIs, particularly for women targeting growth of their firms. More research is needed to clarify this and further investigate the differing challenges including the constraints to growing a business for a women. 8 The definition used in this study was women-owned SMEs were defined as enterprises with (a) > 51% owned by woman/women; or (b) > 26% owned by woman/women and have > 1 woman as CEO/COO. The Enterprise Survey uses a different definition. For this, a firm is women owned if one of the owners is female and it is women run if the top manager is female. 26

27 Overall Conclusions on SME Banking In general, the types of interventions that IFIs, including IFC, are using with FIs in terms of advisory and investment services appear to be in line with how FIs are building their SME financial portfolios. Lessons on intermediation have been gathered and a series of case studies have been completed (e.g., IFC 2010b and IFC 2011b). These provide an overview of the approaches being taken and examples of success stories from the intermediation. The coordinated efforts that are now taking place through the Global Partnership for Financial Inclusion (GPFI) should provide more lessons and models on both practices for FIs and financial infrastructure such as credit bureaus and collateral registries. What is lacking, however, are well done evaluations or research on the intermediation model and its effects on FIs. Besides one completed by IEG (2008) on IFC s work, no other high quality evaluations were found. No other work has focused on the impact of lending to SMEs through financial intermediaries. This is an important issue for IFC since both AS and IS are working through FIs to build access to finance for SMEs. This work with the FIs will establish or expand SME products, enhance the FIs profitability and eventually create competition in the market. Through this work, SMEs will gain greater access to finance. It is clear from the research that conditions within a country have an influence on the extent of access by SMEs but without more research and evaluations, it is difficult to pinpoint which interventions with FIs have an effect and under what conditions. Evaluations of intermediation could provide more information on issues such as the conditions for sustainability of SME lending and how to effectively leverage the IFI investment participation to expand the finance beyond the IFI support. Since the interventions are intended to be catalytic and promote future financing to SMEs, these issues are important. With more information, there could be better targeting of interventions by country and SME target group. This lack of evaluations meant that the performance of FIs in terms of the indicators reviewed here for SME Banking had to be drawn from broader research on banks and their SME portfolios. This evidence primarily focused on issues around profitability. A number of surveys indicated that banks were interested in SMEs due to the potential profitability of serving this sector. This profitability was generated from a portfolio of services, however, not just loans. The research reviewed showed the wide variation in the extent to which the revenues were generated from credit services ranging from a low of 22 percent to a high of 73 percent. This cross-selling of products drives the interest but also means the extent to which SMEs actually have access to credit as part of the relationship with the FIs varies widely across countries. This explains why a large portion of the credit gap of SMEs is actually for firms with accounts with FIs. The G-20 stocktaking report indicated that closing the credit gap for formal SMEs will be more manageable than for informal firms since a vast majority of the formal SMEs have an existing relationship with a financial institution even if they do not have credit (IFC 2010b). This may be more difficult than anticipated given the range of factors influencing the lending and expansion of credit services to SMEs. While SMEs are more credit constrained than large firms around the world, this is exacerbated in low income countries. What determines the proportion of credit within the overall service package of FIs depends on a complex set of variables. These include factors such as: the extent and 27

28 character of SMEs in the economy; competition from other borrowers such as consumers, corporations and government; levels of informality within a sector; the innovation of the banks in terms of the development of methodologies such as credit scoring; presence of different types of banks within the country (e.g., foreign banks); competition in the banking sector; and the state of the financial sector infrastructure such as credit information systems and collateral registries. The financial infrastructure is particularly important since this influences the type of techniques FIs adopt for their SME portfolios. A number of recent studies have reviewed the effects of reforms in these areas. Collateral registries allow FIs to use different types of collateral such as movable assets in dealing with SMEs. Reforms in this area have facilitated an increase in access to finance for SMEs for a range of credit products including lines of credit and fixed asset financing. It also allowed a decrease in lending rates. The greatest effect was for smaller firms (5-19). However, variations are seen across countries in terms of the implementation of reforms and the subsequent effects from it. Similar patterns are seen with the introduction of credit information systems such as credit bureaus. Here the research shows that not only do SMEs have greater access to finance but banks also increase their profitability and decrease their risk. Evidence on the leveraging of inter-firm relationships to access financing for SMEs is also growing. Research on value chains shows that some forms can facilitate FIs lending based on the presence of a larger firm in the chain. Even outside value chains, using long time corporate clients for references on SMEs has increased by banks. Other lending techniques are being increasingly considered such as factoring and reverse factoring. Little is actually known about the effects of having access to finance on the target SMEs reviewed here. A wide range of impact assessments and other research has been done on the effects of access to finance on microenterprises, the segment of SME finance that is covered in this review has limited coverage. No research of high quality could be found on the firm level effects that can be seen from FIs providing credit services to this specific target group. The research that is available relies primarily on assessments of credit constraints at the crosscountry and country level. This research confirms that the lack of access to finance is a constraining obstacle to growth, investment and employment, particularly for small firms. SMEs are more likely to see finance as a key constraint to their development and more likely to be disadvantaged in economies with underdeveloped financial systems. They are also more likely to benefit from changes in financial systems. The credit constraints faced by smaller firms place an impediment to their growth from small to medium enterprises and affects the prospects for high growth firms. These studies do not verify the effects at the SME level in sales, productivity or investment with having access to finance, however. Despite being cited repeatedly in various studies on SME finance, the research does not actually explain the channels that lead to these outcomes. Nor do they adequately separate out different types of firms and differential impacts, such as women-owned enterprises or high growth firms. The special challenges in terms of access faced by different firms need to be investigated along with the channels for having an effect. Without this work, interventions cannot be better targeted and tracked in terms of effects. 28

29 Overview of Gaps in Methodologies The indicators for this review covered performance at both the FIs and firm level from financial intermediation. As described above, both levels faced challenges in finding solid evidence of effects. There is little evidence besides case studies about the specific interventions by IFIs with FIs. This gap means that it is unclear why some interventions work and others do not. Understanding the effect that the increased credit and financial services has on SME s performance is solely reliant on analyses of cross-country data, panel data or country level data. This research confirms that a relationship exists between access to finance and higher growth rates for firms. The research does not provide evidence of what the effects of having credit from FIs actually are. In addition, the credit constraint research on SMEs faces some methodological issues. Causality is not always clear. Does access to credit foster productivity or does the productivity of a firm attract credit? Studies suggest both routes are possible but not the conditions under which this is the case. Cross country data also make it difficult to understand the combination of conditions within a country that can foster greater access. In many cases, the panel data are not detailed enough to begin to distinguish different size and types of firms and how these factors influence the effects at the firm level. Without this type of differentiation, the ability to understand why one intervention might have an effect and another may not will not be possible. So while this research provides valuable insights into the broad issue of access and financing constraints, it does not provide evidence of effects of FI lending to SMEs. This gap highlights the importance of starting to undertake impact evaluations on the type of SME credit interventions reviewed here. The methodologies used to date for evaluations of interventions supporting SME banking have not been rigorous, with limited evidence of actual gains for the SMEs that are documented, along with the causal links between the access and the changes at the firm level. Unlike microfinance, for SME Banking there have not been impact assessments undertaken. 29

30 BUSINESS REGULATION The World Development Report (WDR) 2005 highlighted the importance of the investment climate on growth of firms and the creation of jobs (World Bank 2004). A wide range of government policies and behaviors raise the cost and risks associated with doing business for firms and consequently depress profits and investments. The WDR indicated that small and informal firms are often hit hardest by inconsistent interpretation and implementation of regulations by various levels of government. Studies since that time have confirmed the importance of changing business regulatory regimes to support growth. A joint evaluation of the World Bank Group of IC work in 2006 indicated the complexity of assisting governments with improving their investment climate (IEG 2006). One challenge cited was the lack of knowledge within the World Bank Group about what types of institutional arrangements worked in different countries and how this impacts the process of change. Since that time, more research has been undertaken to determine the effects of various reforms and approaches under different conditions. The business regulations work of IFC currently focuses on advisory services to governments at the national and sub-national level. 9 Three areas of focus for IFC in business regulations are covered here business entry, operations and corporate taxation. The evidence for each area is presented separately to capture different effects for each area. However, it is recognized that they are closely linked and this is dealt with in the overview section. The concluding section analyzes some of the shortcomings and gaps in the current research. Business Entry Research by Djankov et al. (2002) was influential in raising awareness of the impact of entry regulations by documenting the number of procedures, time and cost required to start a business in 85 countries. The conclusion reached was that more procedures and longer delays facilitated more corruption. They found a strong association between regulations and the extent of the informal economy. Entrepreneurs wanting to register faced cumbersome, time consuming and expensive processes. More regulations were seen with less democratic governments. Since that time, streamlining business registration processes is one of the most common reforms undertaken by countries to improve the investment climate. In the past five years, Doing Business (World Bank 2013) has recorded 244 business regulation reforms in 135 countries. The average time overall to start a business has dropped since 2009 by 13 days. One stop shops (OSS) are now operational in 96 countries. Varying Trends in Firm Registration Research undertaken to date on entry reforms in various countries indicates that most see an increase in the number of firms that register immediately after the reforms. One of the most studied reform processes was in Mexico. The Sistema de Apertura Rapida de Empresas (SARE) program was started in Operating at the local level, SARE reduced the time to open a micro or small business from 30 to two days by starting one stop shops and 9 Unlike the World Bank, IFC does not provide investment services to governments for investment climate reforms. 30

31 reducing the number of procedures from 8 to two. Only certain industries were eligible, however, which meant that the retail sector and services were over-represented in the eligible category. The staged implementation of the SARE reforms allowed a number of researchers to establish control groups and determine the effect of the reforms on the number of firms that registered. Bruhn (2008 and 2011a) found that registered businesses increased by 5 percent in the industries covered by the reforms. Kaplan et al. (2011) set up a counterfactual that also indicated that new firm start-ups increased by 5 percent. Similar patterns were seen in Colombia. Cárdenas and Rozo (2009) reviewed the experience with the introduction of a OSS in Colombia located in local chambers of commerce. The implementation of that program as well was gradual allowing the authors to study the effects in the first six cities. They found that new firm registration increased by 5.2 percent in those cities. The increase in registrations, however, can be temporary depending on other factors in the enabling environment. For example, Kaplan et al. (2011) indicated that the increased flow of new registrations in Mexico was concentrated in the first 15 months of implementation. This pattern is seen elsewhere and is not surprising. The factors influencing the start-up or formalization of firms are complex and not simply related to ease of registration. While an initial surge of registrations may be seen, other factors will influence the level of on-going registrations. In other cases, limited change in registrations may be seen from the registration reforms. Bruhn and McKenzie (2013a) reviewed the experience in Brazil with rolling out a business startup simplification program to more remote municipalities. The original program reduced the steps required of firms registering as companies from 8 to 4 and the time from 28 days to 9 days. Using difference-in-difference methodology, they reviewed data from 822 municipalities to see the impact of the reforms over a 56 month period. The municipalities that opened the Minas Fácil Expresso during the period under review had a reduction of 1.3 firms registering per month, the equivalent of a 10.4 percent reduction. The largest effects were seen in the first two months after the offices opened with registration suffering less in subsequent months. One of the possible reasons given for the initial decline was that the reforms consolidated registration for the municipal, state and federal levels. Some firms may not have liked the removal of flexibility to only register with one level, and possibly avoid taxes with the others. A range of factors have an influence on the extent of registrations seen after the initial period of reforms. An evaluation of reforms supported by IFC in four African countries saw increases in all countries, but of varying magnitude (Economisti Associati 2011a). Rwanda had increased registrations over three years from 1,070 firms to 5,808. Sierra Leone had a doubling of registrations during this period. Liberia increased by over 40 percent and Burkina Faso had negligible increases. The wide variation among the countries was due to a combination of the approaches taken by IFC in each locale and the country specific conditions. The jumps in Liberia and Sierra Leone reflect the fact that they were in a post conflict period. A very large informal sector represented pent up demand for registration. Firms formalizing made up percent of the registrations. Rwanda was experiencing a buoyant economy that was much stronger than the other three countries. It was estimated that percent of the registrants were new firms. Burkina Faso had an increase in registrations before the reforms were in place, so there was less pent up demand here. It had an overall low rate of registrations compared with the other three countries and a low rate of growth. A majority of firms were new entrants. 31

32 Character of New Registrants The broad conclusions about entry of firms mask the importance of understanding which firms are benefiting from the reforms and registering. To understand the effects it is important to separate out the proportions of firms registering as start-ups, informal firms that were operating previously and decided to formalize or ones that registered but never started operation. A few studies have been able to separate this, although it is not always straight forward. In Mexico, the districts where Bruhn (2008 and 2011a) focused showed the increase in new businesses came primarily from wage earners opening businesses. There was limited indication that informal firms were registering overall. Kaplan et al. (2011) concluded that the firms entering after the reforms were primarily informal enterprises that formalized. Bruhn (2012) explored this issue of entry of informal firms further. Using discriminate analysis, the informal business owners studied were separated into two groups. The first group had characteristics similar to wage workers and the second similar to formal business owners. The informal business owners from the second group were more entrepreneurial and therefore more likely to register after the reforms in municipalities with high pre-reform constraints to formal entrepreneurship. By separating the pool of informal sector firms by characteristics, the differences among the firms were better understood. While the overall effect of the reforms in Mexico on the informal firms showed they were not more likely to register, the separation of the pool began to identify the type of owners that may respond to entry reforms. Both pools together, however, saw overall low formalization indicating other factors are important in the formalization decision. In , Bangladesh implemented reforms to registration that saw registration times going from 42 days to 1 day, only one visit required instead of 4 or 5 and a reduction in the payment process from 4 weeks to 15 minutes. A randomized control trial reviewed the issue of whether the reforms that substantially reduced the time, complexity and costs of registering a business would encourage informal firms to register (De Giorgi and Rahman 2013). An information campaign was undertaken that focused on both the procedures required to register and the benefits from registration. While the campaign increased the awareness of the new process, it did not increase actual registrations. There was no difference between the levels of registration of the informal firms in either the treated or control group. Information on the process and direct costs were not the main issue for staying informal. Understanding the types of firms that register with entry reforms provides insights into the prospects for growth or exit of those firms. Branstetter et al. (2010), using a differencein-difference methodology, reviewed the implementation of a reform program in Portugal. The Empresa na Hora program was started in 2005 and was aimed at establishing one stop shops with significantly reduced fees and simplified procedures. The time required for legal incorporation went from several months to as little as one hour and the fees from 2,000 euros to 400 euros. The program was rolled out across the country over several years and the authors were able to separate out the types of firms entering compared to the previous entrants. Registrations of start-ups in Portugal increased by 17 percent. However, the start-ups established after the reforms were smaller and 4.4 percent less likely to survive two years than start-ups prior to the reform. The more marginal firms that entered had proprietors who were older, more women and less educated. The firms were typically in sectors with low entry costs such as retail, construction and agriculture. The regulations prior to reforms were not an impediment to higher productivity firms entering. Dropping the barriers encouraged more 32

33 marginal firms to register. It was noted that the reforms took place during an economic downturn when unemployment was high and growth was weak. With better economic conditions, the results could be different. Klapper et al. (2006) used a comprehensive database of European firms to study the effects of entry regulations on the creation of limited liability firms. Using a difference-in-difference methodology, they assessed the impact that entry regulations had on two types of industries, those with naturally low barriers to entry such as retail and those with high barriers to entry such as pulp and pharmaceuticals. Their conclusion was that the entry of corporations in industries that had low entry barriers was more affected by bureaucratic requirements for incorporation and start up than those with high entry barriers. This was particularly the case in countries where regulations were more strictly enforced. They further found that the average size of the firms entering industries such as retail needed to be bigger to compensate for the higher entry costs. This meant that these firms were often forced to grow without the protection of a limited liability corporation, effectively screening out many small young firms. This had the effect of decreasing the competition of the existing firms, with the incumbent firms growing more slowly in countries with higher entry barriers. Reducing the registration costs from the 75 th to the 25 th percentile of the Doing Business ranking was associated with a 10 percent increase in new firms in industries with low barriers compared to those with high barriers. Reforms can foster high growth firms but their potential can only be reached when other elements of the rule of law are present. A study using a six year panel for 64 countries reviewed the question of high growth entrepreneurs and the impact of the regulatory environment on their formation (Levie and Autio 2011). They reviewed the data for nascent entrepreneurs (in the process of setting up an enterprise) and new entrepreneurs in business less than 42 months. They separated out those firms expecting to have 20 or more workers in 5 years (high growth ones). They found that both low growth and high growth firms entered more when the regulatory environment was less burdensome. In countries where the burden was light, a greater proportion of growth entrepreneurs were evident. High growth entrepreneurs were more sensitive to the rule of law, however. The absence of red tape did not encourage high aspiration entrepreneurs if entry regulations were not enforced. In these environments, they may still form but their expectations for growth were more conservative. Certain entry reforms can also foster entrepreneurship. The relationship of entry regulations and entrepreneurship was reviewed by Dreher and Gassebner (2013). Using panel data from 43 countries during 2003 to 2005, they found that the number of procedures and the minimum capital requirements for starting a business reduced the amount of entrepreneurship within a country. The number of days required to register and the out of pocket costs had limited effect. In countries with high levels of entry barriers, corruption is used to reduce the burden. Volume and Character of Reforms The scale and character of the reforms undertaken is an important factor in their effects on firm entry. The World Bank Group has been annually collecting data from company registrars on the number of newly registered firms. These data have been used in a number of studies to review what types of firms are being created and the relationship between entrepreneurship and the business environment. Klapper et al. (2009) reviewed the data from from 100 countries to assess the effect that modernization of business registries had on the business creation process. The conclusion was that a strong relationship existed between a cheap and fast incorporation process and the number of new businesses per capita 33

34 in countries. The data suggested a relationship between the implementation of electronic registration and new businesses registered. Some countries saw substantial increases with the electronic registries as part of the reform package. Klapper and Love (2010b) used panel data from the World Bank entrepreneurship database to look at firm registrations in 92 countries. They found that small reforms, defined as less than 40 percent reduction in costs, days or procedures required for business registration, did not have a significant effect on new firm registration. Tackling two or more of these areas time, cost or procedures allowed important synergies and a greater effect on registration. The initial conditions in countries mattered. Countries with weaker environments required larger reforms to effect a significant amount of new firm growth. For countries with high initial registration costs, long delays or a large number of procedures, the benefits of registration were outweighed by costs. Kaplan et al. (2011) also concluded that bigger programs of reform could have a greater impact. Effects on Firm Performance Some evidence indicates that the entry reforms can trigger firm level job creation and investment. Kaplan et al. (2011) estimated that the firms registering after the SARE reforms in Mexico were larger, increasing the effect on job creation by 7-8 percent. Bruhn s (2008 and 2011a) results showed an increase in job creation by 2.2 percent after the reform. These jobs were more likely to go to people who were previously unemployed or were out of the workforce. Branstetter et al. (2010) found that the reforms in Portugal resulted in an increase of 21.7 percent in jobs created by the new start-ups. However, the nature of the enterprises were in low technology sectors with a higher failure rate than start-ups before the reforms. Panel data for from Doing Business on regulations and procedures was used in another study to look at the effect of reforms on investment and overall growth rates (Eifert 2009). The study found that a decrease in 10 days to start a business had an effect of increasing the investment rate by 0.3 percentage point and the growth rate by 0.36 percent in relatively poor countries. Reforms can increase the competition among firms as new ones enter. Bruhn (2008 and 2011a) found that competition from start-ups lowered prices by 0.6 percent. This resulted in a decrease in the incomes of the existing firms by 3.2 percent. These industries were low risk and in the non-tradeable goods sector. Business Operations A wide range of tools have been developed to assist countries in improving how licenses and inspections are handled including a series of toolkits developed by the World Bank Group s Investment Climate Advisory Services (2010a and 2011). These tools for regulatory reforms have included Regulatory Impact Analysis (RIA) and Standard Cost Model (SCM). Lessons from this experience have been well documented (see for example World Bank 2010b and OECD 2010). Institutional Performance 34

35 A number of studies have verified the gains in institutional performance that can be achieved in licensing reforms but highlighted its complexity. License reforms in Peru are an example. After the reforms, which were assisted by IFC, the median number of days to obtain a license fell from 40 days to 16 days, with the number of requirements going from 8 to 4 (IEG 2013). The median cost declined from $188 to $91. The number of visits to obtain a license fell along with the number of inspections. As a consequence of the reforms, the registration went up dramatically the first year from 1,711 to 8,457. By the third year the registrations levelled off at 1,978. Achieving these types of results can be difficult. A global program evaluation of IFC s business operation product covering licenses, permits and inspections was completed in 2012 (Economisti Associati 2012). A total of 59 projects were reviewed aimed at streamlining processes and improving the overall quality of regulations. The most common reform area was general business licenses that targeted a range of licenses and permits. Overall the projects were able to achieve their targets in terms of recommendations made or laws drafted, with Europe and Central Asia performing best. Translating this into outcomes such as reduced days for processes, however, proved challenging in many circumstances. Short timeframe of many projects and difficult conditions within countries meant that the business license reforms took longer to implement. Business inspections displayed more positive results than other areas. Effects on Firm Performance In countries moving from a highly regulated environment, the reforms can have more obvious and immediate effects. Two studies reviewed the experience with dismantling a licensing regime in India that regulated both the setup of factories and their subsequent production activities. Both studies used a difference-in-difference approach. Aghion et al. (2008) estimated that the reforms increased the number of factories by 6 percent, showing that the deregulation encouraged formation. The effects on output levels were unequal across states within India, however. Output increased more in pro-employer states than pro-worker states. Labour regulations affected how firms responded to the delicensing. Chari (2011) reviewed the same reforms in India and estimated the effect to be an overall increase in total factor productivity of 5-6 percent due to easing of constraints on size of firms and a nearly percent TFP improvement due to the easing of entry restrictions. Most of the improvement was due to relaxation of entry constraints but a portion was related to relaxing the size restrictions under the previous regulations. Previously, bureaucrats would decide whether a sector needed more capacity with every project requiring a license. This process suppressed both investment and the size of firms. The initial increases after the reforms were due to expansion of existing firms, with new firms entering later. Other research suggests more limited effects on firms from licensing reforms. Much of the information available comes from case studies and project databases and reports. One study that is available looked at the effects at the firm level of formalizing by obtaining a license in Peru (Alcázar and Jaramillo 2011). Using a randomized control trial with encouragement, a series of microenterprises were followed over a period of two and a half years. The initial pool were firms that were operating without a license. The results in Peru showed that having a municipal license had little effect on the firms performance in areas such as sales, profits, number of employees, access to credit and investment. No significant difference in outcomes was seen between the firms that registered and those that did not. These findings were verified in a subsequent impact evaluation (IEG 35

36 2013). It is unclear, however, the role that the selection of firms for the study had on the findings. The sample consisted only of informal firms not new firm entrants. Some of the informal firms were subsistence enterprises that were not interested in growing or investing. Neither study separated out the types of firms to better understand whether more performance oriented informal firms that got a municipal license did have an effect in terms their sales or employment. A comparative evaluation of reforms in four African countries also identified limited investments from business licensing reforms (Economisti 2011a). Their conclusion was that these reforms were not a major obstacle to the private sector and this may militate against placing a strong emphasis on business licensing reform in the future. Some methods are now being developed that will begin to provide insights into some aspects of the effects on the ground. One is calculating compliance cost savings (CCS). This is a methodology that the Investment Climate Advisory Services of the World Bank Group is developing that measures the effect of reforms aimed at improving administrative and institutional procedures. It calculates the direct compliance costs before the reform and then after the reform to see the benefits to the private sector from the reforms. 10 The rollout of the CCS has started and could provide valuable information for future analysis. A few studies have begun to estimate compliance cost savings. Some preliminary calculations were included in the Peru tracer study which calculated that $3.02 million had been saved by firms over a three year period, 59 percent due to the reduction in fees and 41 percent due to the reduced number of visits (Alcázar and Jaramillo 2011). Another preliminary calculation was done as part of the evaluations of reforms in four African countries (Economisti Associates 2011a). That series of evaluations estimated a cost savings of $13 million across the four countries over a three year period from over 70 reforms. The reforms, however, covered a range of areas. A vast majority of these cost savings were from business registration and trade reforms. Business licensing only accounted for 1 percent, showing the difficulty in seeing results from licensing reform. The Africa study calculations had clear shortcomings, however, since it was implemented within a before-and-after framework executed after the reforms and the quality of the data, particularly for the baseline, was not always reliable. The calculation of CCS will provide interesting information for future analysis in terms of the types of reforms that generate the most savings. However, more work will be needed in terms of understanding how those savings will translate into behavior by the firms in areas such as investments, jobs or sales. Varying effects on different size firms is apparent in dealing with inspections. Using data on 56,000 enterprises in 90 countries Aterido et al. (2011) identified substantial differences across different size firms in terms of the management time required to deal with government regulations. 11 Microenterprises reported a lower share of time dealing with inspections than the other three groups. Medium enterprises spent the most time, more than even the large firms. 10 Note that the CCS methodology can be applied to a range of reforms including business entry, licenses/permits, inspections, taxation and trade. 11 The study defined microenterprise as less than 10 employees, small enterprises employees, medium and large more than 200 employees. The sample had 39 percent micro, 34 percent small, 16 percent medium and 11 percent large firms. 36

37 Large firms spent three times more management time than small. Despite this, the study found that the regulations did not appear to affect the growth of the larger firms (medium and large). Microenterprises that had been in existence for a while benefited from lower enforcement. A poor regulatory environment increased the survival of low productivity firms by penalizing higher productivity firms. Stringent regulations may lead small firms to grow less as a method to hide from authorities. Corporate Taxation Progress on Tax Simplification Business taxes represent an important source of revenue for governments around the world. A delicate balance needs to be struck, however, between the generation of revenue and the encouragement of investment by domestic and foreign firms. Firms have been concerned about both tax rates and administration and see these as an obstacle to their growth. A review of the World Bank Enterprise Surveys from indicated that tax rates and administration were constraints on firms, ranking them higher than access to finance and licensing issues (Bruhn 2011b). A Handbook for Tax Simplification has been developed that outlines the approaches for tax reform that work based on experience (World Bank 2009). Doing Business 2014 outlines the progress being made in three areas number of tax payments required, time required to comply and total tax rate (World Bank 2014). Since 2009, the time required to comply with the three major taxes (profit, labor and consumption) has dropped by an average of 20 hours due to tax simplification. The biggest gains have been in Eastern Europe and Central Asia with time reduced by 80 hours and number of payments by 20. Sub-Saharan Africa had the largest reductions in the total tax rates, at 17.5 percentage points. Tax Registration and Informality Formalization by registering for taxes can increase firm profitability and investment levels but the outcome varies by type of firm. Those that are able to expand their customer base by issuing tax receipts benefit most. In Bolivia, McKenzie and Sakho (2010) used the distance from a tax office as an instrumental proxy for informality. They found that formalisation, in the form of registration with the tax authorities, increased firm profitability but with differences by size of firm. Firms with 2-5 workers increased their profits by 40.8 percent by registering for taxes. The increased profits were through an expansion of the customer base by being able to issue tax receipts. Smaller firms (less than 2 workers) did not increase profits and in fact decreased profits by 61.6 percent after they registered. This was due to the fact that they were too small to take advantage of tax receipts and therefore did not increase their customer base. The firms with more than 6 workers also saw a decrease in profits from registering for taxes. In their case, the larger informal firms have developed methods to ways to avoid inspections and access many of the potential benefits of formality, without having to pay taxes. Medvedev and Oveido (2013) used a survey of 1,200 small manufacturing and service companies (ranging from 1-50 employees) in Ecuador to assess the relationship between informality and firm profitability. A random sample was constructed to cover eight sectors in four regions. The survey captured the extent of formality of the firms compliance with four 37

38 mandatory regulations and two requirements that were non-mandatory for most firms. More than 80 percent of the firms surveyed were operating somewhere on the informality continuum. The survey showed that higher levels of tax formality were positively and significantly associated with higher levels of profitability for micro and small firms in urban areas. It found that firms that comply with a large number of requirements had higher profits. In particular, having a tax identification number (TIN) increased monthly profits by 21.5 percent compared to unregistered firms. The TIN was considered particularly important since it allowed the issuance of receipts to customers. Using a difference-in-difference approach, Rand and Torm (2012) reviewed the outcomes for micro, small and medium manufacturing firms in Vietnam that formalized by registering for taxes. Two surveys were undertaken. One quantitative survey covered 2,500 enterprises in 2007 and The other was interviews with a random sample of ten firms that had shifted from informal to formal. Firms that became formal (by registering for taxes) between 2007 and 2009 had a 16 percent higher gross profits than comparable firms that remained informal. Formalization increased investment by between 4.2 to 5.4 percentage points. The improved performance was partially related to a decrease in the use of causal labour and a shift to more formally contracted workers. A randomized experiment in Sri Lanka tested whether incentives would be effective in encouraging informal firms to formalize (de Mel et al. 2012). A sample was constructed of firms with 1-14 paid employees that had not registered for taxes but many had registered as a business (68 percent). Offering information on tax registration and reimbursement of the cost of registration had little effect on having the firms register for taxes. Adding payments equivalent to one-half to 1 month s profits of the median firm led to registration of approximately one-fifth of the informal firms. Increasing this to two months profits triggered half to register. Those firms that decided to register, taking advantage of incentives, overall saw an increase in mean profits months later. However, these profits were generated by only a few firms. In interviews it was established that the increased profits were generated primarily by increased advertising, more options for purchasing inputs and use of a receipt book. There was no significant effect on the relationship with the financial sector, however. Female owned enterprises had no increased return to capital. Large informal firms often have limited incentive to register for taxes. This sets up competition between firms in the formal and informal economy and can be detrimental to those choosing formalization. These firms that continued to operate outside the tax system develop methods to avoid tax inspections and access the benefits from formalization without registering, thus maintaining higher profits (McKenzie and Sakho 2010). For these larger firms, limited incentive exists to formalize, particularly since 25 percent of larger informal firms had tax inspections while 77 percent of formal firms of the same size had inspections. They found that owners of large informal firms had higher entrepreneurial abilities than owners of equivalent sized formal firms. Having a tax registration had no effect on the firms ability to access credit. The incentives for formalizing vary by sector and can relate to the level of competition between formal and informal firms. Gonzalez and Lamanna (2007) analyzed Enterprise Survey data of 14 Latin American countries to review the effect that informal competition had on formal enterprises. The size of the informal sector was not a determinant of competition. Firms with high cost of entry such as chemicals and electronics had less concern regarding informal firms since few could compete or had access to the financing required. In this case, the business environment had little influence. In sectors with lower fixed costs such as retail and food products, the informal sector posed more of a concern. This threat was exacerbated in 38

39 environments with high regulatory requirements. The higher the regulatory burden of being formal, the higher the savings to informal firms. The formal firms most affected had characteristics similar to the informal firms including underutilized productive capacity, served smaller customers and were credit constrained. The authors tested the effect of tax regulations and their enforcement. In situations where the tax regulations were onerous and the enforcement poor, formal firms were more affected. An increase in enforcement, even without changes to the regulations, reduced the percentage of formal firms that are affected by informal competition. Increased enforcement can have a greater effect in some cases than tax reform. An evaluation of reforms in three African countries related to business taxation indicated that many firms did not pay taxes because they knew they would not be found out by the tax authorities (Economisti Associati 2011a). In this environment, changes to tax administration to streamline procedures and lower the administrative burden may not be as effective as increasing tax collection efforts. The effect of increasing enforcement of regulations was tested in a randomized control trial by de Andrade et al. (2013). Brazil had introduced a one stop shop in Minas Gerais that covered business registration, tax and municipal licenses. Informal firms were randomized into one control group and four treatment groups. Giving information on the tax registration process, paying the registration costs or providing free accounting services for a year had little effect on level of registrations. Having an inspector visit a neighbouring business did not trigger formalization of firms. Having an inspector visit the firm increased the likelihood of registering the business by 21 to 27 percentage points. Their conclusion was that sticks rather than carrots were a more effective tool to increase formalization. Reforms in taxation can have an effect on how firms deal with other regulatory issues. Brazil had a successful reform process for streamlining tax payments by small enterprises. The Sistema Integrado de Pago de Impuestos y Contribuciones (SIMPLES) began in 1997 and combined several federal taxes and social security contributions into a single payment for eligible small firms. Using a quasi-experimental approach, Fajnzylber et al. (2011) found that shortly after the reforms there was an increase in: licensing rates (11.6 percent); number of firms registered as legal entities (7.5 percent); microenterprise registration (6.3 percent); registration with tax authorities (7.2 percent); and increase in tax payments (3.1 percent). They found that firms that opened right after the reforms were more likely to be in a permanent location and were larger. This could be either an effect of the reforms or a reflection of the firms that registered. It is not clear whether the effect is from the reduced registration costs, reduction in number of transactions or the overall level of taxation. Reviewing the same program in Brazil, Monteiro and Assunção (2012) found that the effects varied by type of sector. A 13 percent increase was seen in formal licensing among retail firms created after the program when compared to firms in ineligible sectors. No differences were seen in construction, transportation, services and manufacturing sectors. The retailers invested more after obtaining the license. Taxation and Growth A number of cross-country studies have looked at the relationship between corporate taxes and growth and concluded that higher tax rates result in lower investment and growth of firms. This research provides some insights into the possible effect that reforms can have at the firm level, although the causal links are not always clear. 39

40 Lee and Gordon (2005) analyzed the effect of the tax structure on gross domestic product (GDP) per capita growth using an instrumental variables strategy. They found that higher corporate taxes were associated with lower GDP growth both across countries and within countries over time. A decrease in corporate tax of 10 percentage points increased GDP growth by between 1.1 and 1.8 percentage points. A series of OECD reports looked at the effect of corporate taxation in OECD economies. One focused on productivity and investment in manufacturing and service sectors using the Amadeus database of 41 European countries between (Schwellnus and Arnold 2008). Using a difference-in-difference methodology, they found that a reduction of corporate tax rate from 35 percent to 30 percent would increase the annual total factor productivity growth by 0.4 percentage points relative to the median growth. This result was consistent across firms of different sizes and age, except for small and young firms that had low levels of profitability and therefore taxes. Tax rates disproportionately affected firms that were catching up to the technological frontier through new capital investments and fast growth. Another study divided the European Union by new entrants (E12) and older members (E15) and reviewed the experience from 2002 to 2008 in terms of productivity (Dall Olio et al. 2013). A one standard deviation improvement in Doing Business tax indicators resulted in a 4.8 percentage increase in labour productivity in manufacturing in the E12 countries and 3 percentage increase in E15 countries. The same change produced a 7.1 percentage increase in productivity in services in E12 countries and 2 percentage in E15 countries. Tax rates can have different effects by type of firm. Djankov et al. (2010) reviewed the effective tax rates in 85 countries in 2004 to determine what a standardized mid-size domestic firm would be exposed to if it was formal and the impact on investment and firm formation. They found that a 10 percentage point increase in 1 st year effective tax rate reduced investment by 2.2 percentage points, business density by 1.9 firms per 100 people and average entry by 1.4 percentage points. Higher corporate tax rates were associated with lower investments in manufacturing but not for services where firms can operate more easily in the informal economy. They found that reducing the number of hours needed to comply with requirements or the number of tax payments had little effect on investment. In certain countries, tax rates may matter less than the accounting standards in the country. Da Rin et al. (2011) used a panel dataset from 17 European countries to test whether tax policy can foster the creation of new firms. They found that corporate tax rates have an effect on entry but it only happened below a given initial threshold tax level. A reduction in corporate taxes from the median of percent to percent implied a 0.88 percentage point increase in entry rates of firms. They found that countries with lower accounting standards allowed more tax elusion or evasive and therefore the level of corporate taxes mattered less since firms could shield part of their income from taxation. This latter effect may mean that lowering taxes in these circumstances will be ineffective in enticing more firms into the tax system. Simply looking at overall tax rates, however, only provides a portion of the picture. Preliminary results of research on the Doing Business 2013 data across 166 countries begins to separate out the effects of tax rates, administration and number of tax payments on growth and investment (Sentance 2013). Both tax burden and complexity have an effect through a series of routes that appear to act as a tax drag equivalent to 1.15 percent per annum on average. Tax administration burden, as reflected in the number of tax payments businesses make, placed a 40

41 drag on growth of about one percentage point per annum across the countries. The reduction in complexity through reforms over the period mitigated this effect somewhat by resulting in an average of a quarter of a percentage point increase in growth per annum. Total tax rate as a percentage of profits averaged 50 percent across countries placing a drag on economic growth of 0.4 percent per annum. The overall effect of these three factors is greatest in Africa. Administrative burdens did not have an effect on inward investment, possibly due to the fact these companies are large and have systems to deal with complex tax regimes. Total tax rates did have an effect, however, with a 10 percent cut in total tax rate associated with an increase of inward investment of 0.7 percent per year. Overall Conclusions on Business Regulation The evidence above has been presented by area of reform but certain common patterns are seen across the three areas. These are summarized here to highlight some broad topics that cut across business regulation. Business Registration after Reforms Table 4 summaries the evidence on the effects of various reforms on business registration levels as noted above. Country Mexico Table 4 Summary of Effects on Business Registration Study and Methodology Bruhn (2008 and 2011a) and Bruhn (2012) OLS with treatment/control group Mexico Kaplan et al. (2011) Colombia Fixed effects with treatment control group Cárdenas and Rozo (2009) Reforms Effects and What Types of Firms (where available) Introduction of OSS Increase in new businesses registering by 5 percent New entrants primarily from wage earners (2011a) Informal business owners with traits similar to formal business owners entered but low levels (2012) Introduction of OSS Increase in new businesses registering by 5 percent Primarily informal firms formalizing Introduction of OSS Increase in new businesses registering by 5.2 percent Brazil OLS with treatment/control group Bruhn and McKenzie (2013a) Difference-in-difference Portugal Branstetter et al. (2010) Difference-in-difference India Aghion et al. (2008) Difference-in-difference Peru IEG 2012 Experimental with treatment/control Rollout of OSS that integrated municipal, state, and federal entities involved in registering a firm Reduction in volume of firms entering by 10.4 percent for first two months and then decline tapering off for another four months Introduction of OSS Increase in new businesses registering of 17 percent Primarily more marginal firms in sector with low entry costs Elimination of restrictive 6 percent increase in number of license regime factories License reforms Increase in license registration from 1,711 to 8,457 in first year and levelled at 1,978 third year 41

42 Country Study and Methodology groups Reforms Effects and What Types of Firms (where available) Rwanda Sierra Leona Economisti (2011a) Evaluation without counterfactuals Economisti (2011a) Evaluation without counterfactuals Business registration reform Business registration reform Increase in registration level in three years from 1,070 firms to 5, percent were new firms 28 percent were firms formalizing 12 percent non-operational Increase in registration level in three years from 1,800 firms to 3,800 firms 45 percent new businesses 55 percent formalizing Burkina Faso Liberia Economisti (2011a) Evaluation without counterfactuals Economisti (2011a) Evaluation without counterfactuals Brazil Fajnzylber et al. (2011) Brazil Bangladesh Regression discontinuity Monteiro and Assunção (2012) Difference-in-difference De Giorgi and Rahman (2013) Randomized control trial Sri Lanka De Mel et al Brazil Field experiment with treatment De Andrade et al. (2013) Randomized treatment/control groups Business registration reform Business registration reform Streamlining tax payments (SIMPLES) Streamlining tax payments (SIMPLES) Business registration reform Registration for taxes by informal firms Increase in registration level in three years from 3,600 firms to 4,000 firms 58 percent were new firms 25 percent formalizing 17 percent non-operational Increase in registration level in three years from 5,200 firms to 7,400 firms 45 percent new firms 55 percent formalizing Increase in licensing rates of 11.6 percent Increase of 7.5 percent in number of firms registered as legal entities Increase in 7.2 percent of registration with tax authorities (7.2 percent) Increase of 13 percent in licenses for retail firms No increase in licenses in construction, transportation, services and manufacturing No change in registration by informal firms when given information on registration process and benefits Offering information on tax registration and reimbursement of the cost of registration had little effect on informal firms registering Adding payments equivalent to onehalf to 1 month s profits of the median firm led to registration of approximately one-fifth of the informal firms Increasing to two months profits triggered half to register RCT with inspection visits Giving information on the tax registration process, paying the registration costs or providing free accounting services for a year had little effect on level of registrations of informal firms Inspector visiting the firm increased 42

43 Country Cross-country (European) Study and Methodology Klapper et al. (2006) Regression with Tobit model Reforms Reduction of registration costs from the 75 th to the 25 th percentile in Doing Business ranking Effects and What Types of Firms (where available) the likelihood of registering the business by 21 to 27 percentage points Increase in new businesses by 10 percent A number of patterns are seen across reform areas. A range of reforms can trigger an increase in firms registering including OSS, tax registration and license reforms. However, the patterns vary by country context including different mixes of new firms and those formalizing and level of increase in firm formation. The increased levels of registration may be temporary or even negligible. A number of studies showed an immediate increase and then a decline, with levels of registrations settling slightly above those seen prior to the reforms. This does not negate the initial effect but does demonstrate that a sustained increase in the number of firms registering relies on other factors besides streamlining business regulation processes. The type and number of reforms has an influence. For example, the negative result in terms of registration in Brazil (Bruhn and McKenzie 2013a) is likely related to the integration of the municipal, state, and federal registrations and concern by firms that their exposure to taxes overall would increase. Other studies have shown that a critical mass of reforms may be required before firms enter (Klapper and Love 2010b and Kaplan et al. 2011). Different types of firms are affected differently by barriers to entry. Sectors with low entry requirements such as retail are more affected by entry barriers than firms in sector with higher requirements such as pharmaceuticals. High growth firms enter more when the regulatory environment is less burdensome and grow less in difficult regulatory environments. Marginal firms are more affected by entry barriers and enter more when reforms are made to the regulatory environment. The firms that enter immediately after the reforms may be in sectors where firms can be smaller at start-up such as retail. More productive firms in these sectors are better able to withstand the increased competition from these new entrants. This could mean lower survival rates of the less productive entrants. Reforms do not necessarily attract informal firms to formalize. Even with incentives, few informal enterprises enter. Those firms that do register tend to be more business oriented, as opposed to subsistence, and are enticed by factors such as the ability to increase their customer bases through advertising or tax receipts. Effects on Firms Performance 43

44 The evidence reviewed above indicates the potential effects of reforms on areas such as productivity, employment and compliance cost savings as shown on Table 5. Country Study and Methodology Mexico Kaplan et al (2011) Fixed effects with treatment control group Mexico Bruhn (2008 and 2011a) OLS with treatment/control group Portugal Branstetter et al. (2010) Peru Rwanda Sierra Leona Burkina Faso Liberia Table 5 Summary of Effects on Firms Performance Difference-in-difference Alcázar and Jaramillo Experimental with treatment/control groups Economisti (2011a) Evaluation without counterfactuals Economisti (2011a) Evaluation without counterfactuals Economisti (2011a) Evaluation without counterfactuals Economisti (2011a) Evaluation without counterfactuals India Chari (2011) Difference-in-difference India Aghion et al. (2008) Difference-in-difference Reform Introduction of OSS Effects Increase jobs by 7-8 percent Firms entering were larger Introduction of Increase jobs by 2.2 percent OSS Decrease of 3.2 percent revenue of existing firms due to competition Firms primarily formalizing Introduction of Increase jobs by 21.7 percent OSS Failure rate of firms higher by 4.4 percent so job growth may not be permanent License reforms Compliance cost savings of USD 3.02 million over three year period No change in sales, investment, employees for firms formalizing (confirmed by IEG 2013) Business Cost savings over three years of USD 2.39 registration million reforms Investment over three years of USD million Total jobs created 15,950 over three years Business taxation Cost savings over three years of USD 0.5 reforms million Business registration reforms Business taxation reforms Business registration reforms Business taxation reforms Business registration reforms Elimination of restrictive license regime Elimination of restrictive license regime Cost savings over three years of USD 0.8 million Investment over three years of USD million Total jobs created 15,100 over three years Cost savings over three years of USD 0.12 million Cost savings over three years of USD 0.4 million Investment over three years of USD 5-6 million Total jobs created 1,800 over three years Cost savings over three years of USD 0.29 million Cost savings over three years of USD 0.8 million Investment over three years of USD million Total jobs created 18,350 over three years Increase in total factor productivity of 5-6 percent due to easing of constraints on size of firms Increase in total factor productivity of percent due to easing of entry restrictions Increase in real output 12 of 17.8 percent in industries in pro-employer states versus pro-worker states 12 Registered manufacturing output. 44

45 Country Study and Methodology Brazil Fajnzylber et al. (2011) Regression discontinuity Sri Lanka De Mel et al. (2012) Field experiment with treatment Vietnam Rand and Tom (2012) Ecuador Matched double difference Medvedev and Oveido (2013) Reform Streamlining tax payments (SIMPLES) Registration for taxes by informal firms (field experiment) Registration for taxes by informal firms Registration for taxes Effects Increase of 3.1 percent in tax payments Profits increased but only for a few firms that formalized Returns to capital for female owned enterprises were zero Sales and employment were not higher after formalization Firms that became formal had 16 percent higher gross profits than informal and increased investment of 4.2 to 5.4 percentage points Formal firms had 21.5 percent higher monthly profits than unregistered firms Regression analysis Bolivia Cross- Country Cross-country (Europe) Cross-country (Europe) Cross-country (85) Cross-country (Europe) McKenzie and Sakho (2010) Instrumental variables Eifert (2009) Regression using panel data Klapper et al. (2006) Regression with Tobit model Oberhofer and Vincelette (2013) Regression using Gibrat model Djankov (2010) Regression analysis Da Rin et al. (2011) Regression analysis Cross-country Lee and Gordon (2005) Cross-country (Europe) Instrumental variables strategy Schwellnus and Arnold (2008) Difference-in-Difference Registration for taxes Reduction of 10 days in time to register Reduction of registration costs from the 75 th to the 25 th percentile in Doing Business ranking Unit improvement in firms perception of the business environment Effect of complex tax system Increase in 1 st year effective tax rate of 10 percentage points Reduction in corporate tax from percent to percent Decrease of corporate tax by 10 percentage points Decrease of corporate tax rate from 35 percent to 30 percent Firms from 2-5 workers increased profits by 40.8 percent Firms less than 2 workers saw drop in profits by 61.6 percent Firms with greater than 6 workers saw drop of profits by 54.6 percent Increase in investment rate of 0.3 percentage point Increase in growth by 0.36 percent Decrease of 0.7 percentage points in real growth rates of value added per worker differentials among naturally high and low entry industries (high entry industries benefit more) Increase in job creation by 1.1 percentage point annually Reduction in employment growth by 1.9 percentage point annually Reduces investment by 2.2 percentage points Business density by 1.9 firms per 100 people Average entry by 1.4 percentage points Increase of 0.88 percentage point in entry rates of firms (12.5 percent of the 7.02 percent mean) Increase GDP growth by 1.1 to 1.8 percentage points Increase in total factor productivity of 0.4 percentage points 45

46 Country Cross-country (Europe) Study and Methodology Dell Olio et al (2013) Regression analysis Reform One standard deviation improvement in the Doing Business tax indicator Effects Increase in productivity of manufacturing firms by 4.8 percentage in newer EU countries and 3 percentage in older EU members Increase in productivity of service firms by 7.1 percentage in newer members and 2 percentage in older members Overall, the evidence is positive regarding the potential effects of the reforms in areas such as job creation and productivity. However, the wide variations seen from reforms need to be carefully assessed. Unfortunately, many of the studies do not provide enough information to be able to determine why effects were seen in some countries and not others. This highlights the need for further investigation in a number of areas. The job generation from business registration reforms may reflect actual growth of jobs or simply firms moving into the formal sector. This is often not clear from the studies and could overinflate the job results. Part of the challenge here is knowing whether the new registrations are from existing firms, new start-ups or firms that register but do not operate. Few database provide this level of detail. Few of the studies have focused on licensing and inspections and those that have been done showed mixed results. Peru shows little effect from the simplification process but it is unclear whether this is related to the type of firms tracked. The changes to the licensing regime in India were so large that this likely influenced the effects there. More work needs to be done to better understand the theory of change around licensing and inspections in particular given the complexity of the reforms being implemented. What is realistic to expect under what circumstances? The types of firms that have higher performance after formalizing need to be more closely examined. Sectors, characteristics of the firms and approaches to the business seem to play a role in how the firms react to reforms. More detailed analysis is needed to better understand how this varies by factors such as womenowned enterprises or high growth firms. The cross-country studies provide insights into possible effects from registration and tax reforms but the causal links are not necessarily clear in terms of which direction they go. These need to be used carefully in making conclusions about the effects of reforms. Informality The anticipated effect of informal firms becoming formal has not materialized with the registration and other business regulation reforms. In some cases, a few informal firms moved to register. The evidence here provides some insights into why this may be the case. The debate about the role of regulations in the level of informality of firms started with De Soto (1989). He argued that high entry barriers and red tape caused higher levels of informality. If reforms were made, firms would formalize. He concluded that barriers such as information and 46

47 costs hindered informal firms from registering. More recent research shows that these are not necessarily the constraining factors (de Mel et al. 2012, De Giorgi and Rahman 2013, de Andrade et al. 2013). Other theories focus around the costs and benefits to formalization. The argument is that informal firms weigh the advantages and disadvantages of formalizing and then decide whether the reforms shift this balance. If they do not perceive benefits, they will not formalize even with incentives. The evidence presented here shows that these cost benefit calculations vary substantially based on which portion of the informality continuum a firm is on and the characteristics of the firm. There is a need to separate out the informal firms that have interest and potential for growth from other firms. A number of different approaches have been developed to better understand the spectrum of informal firms. 13 A stylized version of the informality continuum that assists in understanding business regulation reform effects is contained in Table 6. DCED (2009) provides a summary of the issues seen at each level and the likely costs and benefits. It outlined subtleties that need to be considered in this cost benefit equation. For example, tax evasion is an important consideration for the unofficial and official enterprises but not relevant for subsistence enterprises. Access to finance may only be applicable to those firms at the end of the continuum which are mostly registered since those further down may not be bankable even with formalization. Much of the research and conclusions on informality and reforms to date have focused on subsistence microenterprises. Convincing these firms to formalize is difficult as seen with some of the experiments presented here. The experience with entry reforms to date has shown that only a small number of informal firms register after the reforms. Even with enticements, few are interested and see benefits to their operations. Those that do formalize tend not to grow due to their approach to business, skill sets and other factors such as interest in advancing the firm. Focusing business regulation reform efforts at this level to bring firms into the formal sector will likely not increase registrations and bring the benefits from formalization that are anticipated by governments such as increased taxes. While a fiscal incentive may exist for other firms, with these subsistence enterprises, Bruhn and McKenzie (2013b) question whether there is a public rationale for trying to formalize. Stein et al. (2013) suggests that other incentives to formalize are needed such as access to market opportunities, access to finance and capacity development. 13 See for example, Stein et al 2013 which outlines four types of informal firms necessity firms, opportunity firms, defensive evaders and wannabe formal firms and provides an overview of the spectrum of support for each group to be enticed into formality. This support extends beyond business regulation reforms to areas such as training and credit. 47

48 Least Dynamic Table 6 - Typology of Informality Highly Dynamic Characteristic Degree of Informality Type of Activity Completely Informal Informal Economy Subsistence Enterprises Mostly Unregistered 100 percent High proportion of sales undeclared and workers not registered Single street Small traders, cottage/ manufacturers, micro enterprises, service providers, subsistence distributors, farmers contractors Technology Labor intensive Mostly labour intensive Owner Profile Poor, low Poor and nonpoor, education, low well level of skills educated, high Markets Low barriers to entry, highly competitive, high product homogeneity level of skills Low barriers to entry, highly competitive, some product differentiation Finance Needs Working capital Working capital, some investor capital, supplier credit Other Needs Personal insurance, social protection Personal and perhaps business insurance Unofficial Enterprises Source: DCED 2009 adapted from Djankov et al Mostly Registered Partially Formal Formal Economy Official Enterprises Some proportion of sales undeclared and workers unregistered. May use outside the official purview (e.g., internet to deliver software) Small and medium manufacturers, service providers, software firms Knowledge and capital intensive Non-poor, highly educated, sophisticated level of skills Significant barriers to entry, established market/product niche Investment capital and working capital, letters of credit, supplier credit Personal and business insurance, business development services At the other end are firms that are registered and appear to be operating according to all the rules but are not. These firms are likely to have developed approaches to avoiding complete formality and are unlikely to change due to reforms unless pressured. They have developed systems to cope with taxes, inspections and regulations that allow them to stay partially in the shadows. Reforms to licensing or taxation will likely have limited effect on their behavior since they have the systems to operate outside the rules and benefit from this through higher profitability. For these firms, enforcement may be the only route to have full compliance. The firms in the middle part of the continuum (highlighted in red) are the more interesting and represent a range of firms where reforms can have both a negative and positive effect. They may formalize in order to take advantage of a business opportunity such as providing tax receipts to customers but may face the most competition from informal firms. This is true where firms with similar characteristics are competing and have the same productivity 48

49 and credit constraints, but differing degrees of formality and compliance. Formal firms are at a disadvantage and may grow slower than if the competition were not informal. Receiving inspectors was raised in a number of studies as an influence on formalization in this group since formal firms may receive more inspectors and have a greater tax burden. Better understanding of how to facilitate formalization of these firms with potential for growth should be reviewed in future research. These tendencies can be heightened in more restrictive regulatory regimes. One study looked at a firm level data set from 27 countries in Eastern Europe and Central Asia (Dabla- Norris and Inchauste 2007). They defined informal firms as those that report less than 100 percent of their sales, wage bill or workforce. The informality was not limited to small firms since some large companies were evading taxes as well. The study found that growth in formal firms was negatively impacted by both tax rates and weaknesses in tax administration. It was impacted by a higher regulatory burden in terms of the cost of dealing with licenses. By acting outside tax and regulatory framework, informal firms could be less productive than their formal sector competitors but still grow faster. The move can be in either direction from informality to formality or the reverse. A study of four African countries (Kenya, Cote d Ivoire, Nigeria and Senegal) reviewed the issue of why some firms move from formality back to informality (Gajigo and Hallward-Driemeier 2012). A survey was undertaken of 1,967 firms in It covered: microenterprises (1-4 workers and 38 percent of sample); small (5-19 workers and 39 percent of sample); and medium and large (20 and above workers and 23 percent of the sample). Two-thirds of the sample were formal. Part of the sample were firms that had begun as formal enterprises and then reverted to being informal. The firms shifting back to informality were more likely to have started small with low levels of capital and workers and have lower levels of productivity than they had originally anticipated. The likelihood of remaining formal was negatively influenced by the costs of bribes since formal companies paid more in bribes. Firms going from informal to formal were more likely to have higher levels of productivity, higher education levels of the owners, greater access to finance at start-up, and have been an employee of a formal firm. Firms that made the transition to formality had 43 percent higher labour productivity than firms that remained informal, reinforcing the idea that these firms were ones with a productivity premium. Other Influences A wide range of factors influence how various reforms are implemented and their resulting effects on firms performance such as profits and productivity. A few examples are provided here. The list is not intended to be exhaustive but to highlight some of the research on influences on reforms. The overall environment within a country may override the effects of reforms for firms. One study pinpointed the broad differences in treatment across firms within some countries (Hallward-Driemeier et al. 2010). They looked at the role that making deals had in the way policies were implemented in some African countries. Firms that were connected politically might have less tax audits and receive licenses more rapidly. Looking across 24 countries, they saw wide variations within countries in actual times to receive a license despite official policies. Some firms that were obtaining licenses received them almost instantly, while other firms got them much more slowly. This deal making environment made small firms (less than 6 employees) in particular more skeptical about the predictability of rules and enforcement. It stifled employment growth, with the effect on SMEs being stronger. They raised the question 49

50 about whether changing policies and reforming procedures could have an effect in situations with weak institutions. The recent financial crisis has obviously had an influence on business entry. Reviewing data from from 95 countries Klapper and Love (2010a) found that almost all countries experienced a slowdown in business entry. The data, which covered formal firms only, indicated that developed countries slowed first with contraction taking place in In parallel to the spread of the financial crisis, developing countries were more impacted in The biggest drops were seen in the high and upper middle income countries and countries with more developed financial markets. While some countries have returned to their pre-crisis levels of firm entry, others continue to languish with more limited prospects for firm in the immediate future (Xavier et al. 2013). A report from the Organisation for Economic Co-operation and Development confirms that the crisis is continuing to have an affect (OECD 2013). Start-ups in OECD countries are still below pre-crisis levels in many countries. Some countries such as Australia and the United Kingdom are seeing increases but are facing higher failure rates. There is evidence that many of the entrants are starting businesses due to the lack of jobs not a desire for entrepreneurship. This could mean that the number declines when the economies begin to recover and jobs are available. Oberhofer and Vincelette (2013) reviewed industry and firm specific determinants of job creation in 11 European Union economies. The dataset covered , allowing an analysis of the impact of the financial crisis. They found that in industries that were heavily regulated, employment growth prior to the crisis was lower. A unit improvement in the firms perception of their business environment in that industry could have generated approximately 1.1 percentage points annually in job creation. After the crisis, unfriendly business environments were more vulnerable to job losses. Institutional barriers correlated with a lower probability of having high growth firms gazelles prior to the crisis. After the 2008 crisis, the likelihood a firm would grow fast declined further. The interrelationship between regulatory and financial constraints facing firms is important to consider. Gatti and Honorati (2008) used Investment Climate surveys in 49 countries to investigate the relationship between tax compliance and access to credit. They concluded that the lack of tax compliance limits the extent to which balance sheets of firms are informative for banks in countries with high levels of formality. The lack of tax compliance in these circumstances signals that the firm may not have reliable financial information. In countries with high degrees of informality, however, banks develop other ways of assessing creditworthiness. Hallward-Driemeier (2009) reviewed panel data from 27 Eastern European and Central Asia countries to study patterns of firm exit. The study found that higher regulatory burdens raised the probability that more productive firms would exit. However, one of the variables studied was access to finance measured by the share of firms that had a loan and the share of financing that came from formal sources. They found that in countries with stronger financial institutions more productive firms had greater probability of survival. In countries with underdeveloped financial institutions, rates of exit were lower but productive firms had a higher probability of exit while less productive firms did not exit. These latter firms relied more heavily on their own resources rather than access to formal finance. 50

51 A series of studies show that formalization does not automatically lead to greater access to finance given other factors such as the productivity of the firm, state of development of the financial system within a country, behavior of banks and ownership structure (IEG 2012, Klapper and Love 2010b, Medvedev and Oviedo 2013, McKenzie and Sakho 2010 and de Mel et al. 2012). A firm is not bankable just because it has registered as a firm, received a license or registered for taxes. Women-Owned Enterprises Gender specific challenges within investment climate work have been identified in a number of studies. A few studies have concluded that the regulations may be more or less gender neutral but other factors such as legal rights, traditions such as land inheritance and lack of transparency in areas such as tax collection make those regulations more onerous for women (see for example Hampel-Milagrosa 2011, Simavi et al. 2010). Similar issues were identified in Africa where property rights, access to finance and harassment were impediments to starting and operating businesses (Hallward-Driemeier 2013). Klapper and Parker (2011) found that the differences in business entry and performance of men and women owned enterprises were not explained by regulations. Different barriers were seen such as the concentration of women in low capital intensive industries being driven by lack of access to finance. The differential effects of reforms on women have not been fully investigated. While gender issues are being integrated more into business regulation projects, the research included in this reviews provide few insights into the issues facing women entrepreneurs. Branstetter et al. (2010) analyzed the data on the Portugal reform by gender of the owner. As noted above, women-owned enterprises that were formed after the reforms had a lower survival rate than male owned enterprises, likely because they were in more marginal sectors. Almost all of the other studies did not separate results by gender of owners. In some cases such as the Enterprise Surveys, ownership is included in the data but this information has not been used by researchers to review the effects of reforms on female entrepreneurs. In other cases, the data sets are not sex disaggregated in terms of firm ownership. This gap in the current research needs to be addressed in order to better understand how the reforms are having an effect on women and their enterprises. Overview of Gaps in Methodologies The evidence on the effect of business regulation reforms is slowly growing. Business Regulation research was more robust than that available for SME Banking. Included in the review of Business Regulation were experimental design such as randomized control trials and encouragement design. Quasi-experimental methodologies were available such as differencein-difference and instrumental variables. While the quality of research was good overall, the review faced a few challenges. The concentration on specific reforms is high. Some of the research dealt with the same reforms with different authors reviewing the reform process. In some cases, the conclusions were similar or complementary; in others they were contradictory. For example, Bruhn (2011a) said the entrants in Mexico were primarily new firms and Kaplan et al. (2011) found the same reforms increased registration from existing firms. This highlights the complexity of the reform 51

52 processes being reviewed even using experimental techniques. difficulty in separating out effects by types of firms. It further highlights the Many of the studies focused on a small group of countries with an overrepresentation of the Americas. The conditions in each country and region are different. The challenge is how the country experience can be generalized more broadly. The unique combination of factors in each country makes this difficult. Mexico and Brazil are likely not representative of countries in Africa or Asia. A study may have validity for one context but not be relevant for understanding another context. The quality of regulatory governance within countries appears to play an important role and needs to be factored into the analysis. A wide range of projects have been funded by various donors and organizations that focus on or include business regulation but few good quality evaluations have been done. With the exception of the evaluations used here (Peru and Africa) by IFC, none were able to pass the quality standards established for this review. Most were not clear about how the conclusions were reached and about the causal lines. There seemed in some cases to be an over-estimation of the results from a project a problem that has been identified previously (DCED 2009). The causal links underlying much of the literature to date need to be reassessed based on the emerging experience. This review looked at the evidence related to two levels of indicators government and enterprise. At the government level, as shown on Table 4 above, all three areas of reforms showed potential for increasing the number of businesses registering. Reforms aim at simplifying and streamlining processes in order to decrease the costs, procedures and time required to comply. These changes can trigger the entry of firms both existing and new. The conditions under which this takes place and the types of firms that appear to respond have been outlined in this review and are influenced by a range of country level and implementation factors. What is less clear, however, are the potential effects at the enterprise level from the reforms. Ideally, the reforms will trigger growth in areas such as jobs, investment and productivity. While there is evidence of some of these effects with each of the three reform areas as shown on Table 5 above, the causal links vary by reform area, location and type of firm. They are clearer with some types of reforms than others. The theory of change differs by the type of firm including the degree of formality, sector, character and other variables. Figure 2 provides some simplified examples of the causal links seen in the evidence presented here. This is done to illustrate some issues. The business entry reforms have the clearest evidence. In response to reforms such as the establishment of OSSs, firms register creating jobs and generating investments. While these effects do not always occur, they are relatively common across countries. The magnitude of the response from entrepreneurs varies widely, however, as does the eventual level of firm entry after the initial reforms are implemented. Where the data are lacking here is in the ability to consistently breakdown new firms from existing, the character of the firms (including women-owned enterprises) and the extent to which the investments generated are above the minimum capital requirements by firms within a country. 52

53 Figure 2 Possible Effects at Enterprise Level from Evidence to Date Business Entry Reforms Decreases in: Days to register Cost of registering # of procedures Increase in firms registering (new and existing) Jobs Investments Licensing Reforms - India Removal of controls on entry and production activities New Firms Enter Existing Firms Expand Increases in productivity by new and existing firms Licensing Reforms - Peru Decreases in: Day required Costs Time # of inspections Increase in existing firms registering for licenses??? Business Taxation Informal sector Response Some informal firms registering for taxes Increase customer base (by receipts, advertising, etc.) Increase: Sales Profits Business license reforms appear to be more difficult to have an immediate effect on firms. The expected causal links here are not clear and the evidence to date shows strikingly different effects. The two examples often cited are India and Peru. The dismantling of the license control regime in India was a dramatic shift in policy, moving from a controlled environment to freer entry and growth within a sector. This clearly resulted in firms responding to the changes and entering and expanding, albeit not consistently across the states of India. What is not clear is whether this example of reforms can provide insights for other countries today on the possible effects reforms can have. The licensing in India was so highly controlled and the reform so dramatic that it may not provide lessons. The Peru license simplification process resulted in decreases in compliance costs for firms which have been documented. Informal firms registered for licenses but no evidence was seen that this registration had an effect on sales, investment or other firm level indicators for firms that were formalizing. No evidence was gathered on new firms that were registering. These two examples highlight the difficulty in establishing how license reforms can influence firm performance. It is not clear from the evidence around licensing what the actual effects could be and the links to achieving those effects. How much of a reduction in compliance costs would be required to generate a response from firms in terms of investment, productivity? Do different types of firms respond differently? More research needs to be done to better establish potential causal links and expected effects from interventions and the conditions under which they might occur. For taxation, the evidence showed a portion of the picture namely relating to informal firms registering for taxes. Here, a small group of informal firms saw a benefit to the registration in terms of being able to reach new clients. Those that did take advantage of the opportunity saw increases in sales and subsequent profits. What is not clear from the evidence is the effects of tax reforms on new entrants. What reforms draw in high potential firms? What are the factors that influence this? While some of the cross-country research provides a few insights here, they 53

54 focus primarily on the fact that a causal link exists not how it actually operates. There appears to be differing responses depending on the type of firm and the reforms. Further research is needed to determine the character of this. Other assumptions being made on possible effects also need to questioned and tested. For example, it is often assumed that formal registration of informal firms will lead to better access to finance. This has not been proven. Without other changes at the firm, a firm is likely no more bankable than it was before. Informality may be a barrier to commercial lending but the removal of the barrier does not solve the bankability issue. The same is true for markets. Some of the evidence here indicates that firms have registered in order to expand their customer base through being able to give tax receipts. In other cases, new entrants were smaller and less viable after the reforms particularly in sectors with low entry barriers. This increased competition and resulted in a lowering of survival rates. The complexity of the causal links needs to continue to be explored so the expectations are realistic and based on experience not on theories. The final issue revolves around the data available. Much of the country level data in areas such as business registration is often not broad enough to allow full analysis. One of the difficulty with the data available is that few sources are able to track exits of companies. For example, this was a problem in Vietnam in terms of estimating the effects from the introduction of the Enterprise Law in The official number of firms registering increased substantially. The business registration database of the government was accurate in terms of registrations but it did not provide a real picture of the robustness of the growth (Freeman et al. 2005). Several issues were seen including a large number of registered companies that never started and others that had not succeeded and exited. Many databases do not cover exit of firms which then means this factor is not integrated into the analysis. 54

55 FILLING THE GAPS AND IMPLICATIONS FOR IFC Filling the Gaps The research and evaluations included in this systematic review were carefully selected, based on specific criteria. Focusing around access to finance for SMEs and Business Regulation reforms, evidence on indicators of effects at two levels were reviewed. For SME Banking this included indicators for financial institutions and SMEs accessing financial products. For Business Regulation it included the effects of the reforms in business entry, operations and taxation and the benefits for firms from those reforms. Different challenges were faced in each of the areas in terms of finding solid evidence. Overcoming these challenges in the future will be important to better understand what has worked and not worked and why. Currently there are gaps in knowledge that prevent important questions from being answered in terms of interventions that support either SME Banking or Business Regulation. For SME Banking, severe limitations were seen in what evidence was available on both financial intermediation with SMEs and the effects of this financial access in terms of SMEs performance. While there is extensive research and impact evaluations on microfinance, no research or impact evaluations are available on SME Banking in terms of either the intermediation model or the firm level effects. Evaluations that were available for this category of lending were not of high enough quality to be used in this review. There is a need to undertake more primary research and evaluations that directly look at SME banking and financial intermediation. Gaps are seen in knowledge around both issues. Little is known beyond general lessons of what works, what does not and why in terms of intermediation. This is also true around the issue of the extent that SMEs improve their performance with the availability of finance and under what circumstances this takes place. This includes separating the analysis by types of firms including women-owned and fast growing. With the growing interest in SME finance globally, it is possible that more focus can be placed on these types of issues. For Business Regulation the gaps are seen in terms of better understanding the differential effects for categories of reforms on different types of firms. For example, much of the research to date on informality has been focused on subsistence enterprises. The results of this work, however, is not very applicable to other categories of informal firms that have the desire and potential to grow. The research around formality should begin to shift focus from subsistence enterprises to the unofficial enterprises outlined above. This latter group contains more firms that could benefit from specific reforms. For the different areas of reform, the theories of change need to be more closely investigated to better understand what kind of reforms can trigger what effects in terms of entry and performance of firms of different types. This means that research using panel data should be disaggregated by type of firm including ownership by women. It also means investigating the differential effects seen to date with some reforms such as licensing and the magnitude of reforms required to trigger an impact. 55

56 Specific Implications for IFC This review did not evaluate interventions implemented by IFC. Instead the focus was on gathering credible evidence on the experience to date in SME Banking and Business Regulation that could inform IFC s decisions. The following are some implications of the findings for IFC s work in these two areas. SME Banking The lack of evaluations and research on the intermediation model with FIs makes it more important for IFC to further develop methods for results tracking of SME Banking interventions. Currently IFC is tracking a number of indicators on its SME Banking portfolio such as loan portfolios. Analyzing the trends in these indicators over time provides information on the overall growth of the IFC portfolio. It could also provide insights into other types of trends including which types of banks are expanding their portfolios and which are not. These data could be further analyzed to gain a better understanding of the effects at the financial institutions of the AS and IS interventions and conditions under which these take place. These conditions include country level factors and characteristics of banks. IFC should develop more extensive methods for analyzing the indicators being currently tracked to ensure that the information informs decision making around SME Banking. The movement by IFC to undertake impact assessments on the SME banking portfolio is important given the lack of information currently on the effects at the SME level. IFC has faced challenges in obtaining information on the SMEs that are being funded by the FIs. Undertaking impact assessments could begin to fill this knowledge gap and should look at SME banking under different conditions and by different types of firms. IFC could leverage its leadership position on the global stage in SME Finance to encourage more evaluations and research on SMEs receiving financing from banks. The focus to date of the international research has been on microfinance which does not inform the effects of lending for SMEs. The development of the Impact Assessment Framework: SME Finance (World Bank and GPFI 2012) is an important step forward and outlines some approaches that are suitable for SME Banking. An increase in the volume of evaluation and research of different types of SMEs will provide insights into how to better target groups such as women owned enterprises and gazelles. It will also allow the development of the possible theory or theories of change that can happen when SMEs access funding. Building on the existing relationships between FIs and SMEs can continue to be an effective technique for expanding access to credit but the conditions within a country will dictate the feasibility of the approach. Evidence is showing the complexity of factors that influence and deter an FI with an existing relationship with an SME from including credit as part of its package of services. Country context is proving particularly important. New approaches may be needed under some circumstances by AS and IS to better address these differences. IFC can continue to play an important role in supporting the development of financial infrastructure collateral registries and credit information systems. The 56

57 awareness of the importance of these reforms for allowing increased access to credit is at a high level currently. The effects from these types of interventions to date show the importance of these reforms on increasing SMEs access to finance. But the interventions need to be carefully planned since the method of implementation across countries has an effect on whether access to finance for SMEs improves. These effects are not seen across the board. Business Regulation The business regulation reforms that are supported by IFC need to be decided based on the extent to which the outcomes of growth, jobs and investments can be achieved. As more evidence emerges, some products may be less important for achieving these results. This will require a careful tracking of what has worked, what has not and why. IFC should make clear to government partners what expectations are realistic from business regulation reforms particularly in areas such as formalization. The extent of informality within many economies makes it a critical issue for governments to tackle. However, business regulations reforms have limited effects on a vast majority of the informal firms. For some groups of enterprises, other support programs may be more important for promoting formalization than business regulation reforms. This requires a more multi-faceted approach to tackling informality including coordinating with other IFC products outside of the investment climate work. Being able to reach a critical mass of reforms is important and should be integrated into the decisions on a country basis. If the potential for a critical mass of reforms is not evident, the likelihood of this response from enterprises is more limited. More work is required in understanding what is meant by critical mass in differing circumstances. To date, the focus has been more on the volume of reforms around a specific area such as business entry. How various reforms can interact and reinforce each other is not well understood to date. The causal links between certain reforms and outcomes need to be better understood. The paths from the reform areas of entry, permits and taxation to changes at the enterprise level vary. Some are more straightforward such as reforms around one stop shops. Areas such as inspection are more complex. Variations are also seen when looking at different types of formal and informal firms. Understanding the differences will allow better targeting of interventions and tracking of effects. IFC should continue to do more impact assessments that provide evidence from a broader range of reforms and geographic locations. This will begin to fill in the gaps of knowledge around the conditions under which enterprise effects may take place or not. Some of those gaps have been identified here. 57

58 APPENDIX A Summary of Methodology Systematic reviews have proven to be an effective method for identifying what works and does not work internationally. Unlike literature reviews, systematic reviews are undertaken within strict methodological guidelines to ensure that the evidence presented is of high quality. They methodologically map out available evidence, critically appraise it and synthesize the results. This systematic review has followed a highly structured process that is outlined below. Defining the Areas of Focus The first step in the process was to clearly define the parameters for the review in the two areas of focus SME Banking and Business Regulation. This was done in conjunction with stakeholders at IFC to ensure that the focus was relevant to IFC interventions. Figure 3 provides an overview of what encompasses SME Banking in general. The broad definition includes not just commercial banks but leasing companies, microfinance institutions, investment banks and equity funds. After consultations with relevant stakeholders from AS and IS, it was determined that the focus of the strategic review should be on the core SME banking products namely with commercial banks. This is where the majority of work and funding is being done by IFC with FIs and are most relevant for SMEs. Figure 3 IFC Interventions through Financial Institutions Long Term Finance Private Equity Core of systematic review Equipment Finance Leasing Working Capital Microfinance Trade Finance Only includes enterprise loans not loans to individuals Informal /Smaller Formal/ Larger Source: Graphic from IFC The SME Banking Knowledge Guide with adaptations The trade finance area was not included since IEG has just completed an evaluation of this area which covers IFC s work. Leasing is a declining priority within IFC s portfolio so was deemed to be less important for review. A portion of the private equity funds from IS goes to FIs that invest in SMEs but this portion is very small. Since SMEs are not a priority target group under the private equity funding, this was not included in the systematic review. Microfinance was not 58

59 included because its focus is on micro enterprises and not SMEs with the loan sizes being too small. To distinguish between SME Banking and microfinance three other factors were taken into account to ensure that the review reflected IFC SME Banking products. First, the review focused on financial products through FIs, specifically commercial banks. Second, the size of enterprises was reviewed for each study to ensure that the review reflected the experience of SMEs not microenterprises with definitions in various studies being noted as applicable. This process took into account the widely varying definitions for micro and small. Third, for evaluations and research where an average loan size was indicated, the IFC loan proxy of a minimum of $10,000 was used to filter out studies where loan sizes were substantially lower than this and more applicable to microenterprises. For Business Regulation, after consultations, it was agreed that the systematic review would focus on specific Business Regulation products being offered by IFC aimed at business registration, business licensing and inspections and corporate taxation. Business registration focused on initiatives such as the establishment of one stop shops. Business operations covered a range of different procedures aimed at simplifying licensing and inspections. For taxation, a narrow approach was taken given the wide range of reforms to potentially cover. The focus was on corporate taxation and the efforts to simplify tax procedures such as registration filing and payment. The review did not include areas such as value added tax, tax incentives and labor taxes. Questions and Indicators The key questions to be answered by the systematic review were developed and are: 1. What types of IFC-related work with financial institutions to increase provision of financial services to SMEs (SME Banking) are most likely to have the largest effects? The effects were measured by performance indicators such as changes (i) at the FI level in number of loans disbursed to SMEs, number or value of outstanding loans, profitability of the SME portfolio at the FI and average nonperforming loan rate and (ii) growth of SMEs in terms of sales/revenue, assets, productivity and number of employees. 2. What types of IFC-related public sector interventions aimed at reforms in Business Regulation, are most likely to have the largest effects? The effects were measured by performance indicators such as changes (i) at the institutional level in areas such as number of businesses registering, number of businesses benefitting from reforms (registration, licensing, permits, taxes) and average number of days/procedures to comply with regulation and (ii) at the enterprise level in areas such as changes in costs to comply, levels of investment, revenue, productivity, employees and formalization. 3. What methodologies/approaches have been used to conduct the research and evaluations? Were they appropriately applied and relevant to the subject? 4. What lessons, conclusions and recommendations can be extracted from the research and evaluations that can inform management decision making? As noted above, for both SME Banking and Business Regulation IFC works with intermediaries, with the SMEs being indirect clients. For this reason, the analysis of the evidence from IFC and 59

60 outside sources focused on understanding the results that are emerging on two levels institutional in terms of the direct partners with which IFC s programs operate (e.g., FIs and governments) and enterprises. One captures the changes at the partner FI or government level; the second focuses on the longer term changes at the enterprise level. This allows questions to be asked such as: Are the FIs reaching SMEs? If so, what impact does this have on the firms performance? Figure 4 summarizes the indicators that were used for the systematic review. The institutional indicators are based on the ones that IFC s SME Banking and IC are tracking as part of their results measurement frameworks. The firm level results are based on the theory of change that takes place when the FIs provide products to SMEs or the government reforms the regulatory environment. The review looked at the evidence at both of these levels and identified the process of moving from interventions to outcomes. Figure 4 SME Banking FIs provide financial services to SMEs Business Regulation Governments reform business registration, regulation and taxation Indicators to be Assessed at Institutional/Government Level # Loans disbursed to SMEs # or value of outstanding loans Profitability of the SME FI Average nonperforming loan rate # of businesses registering # of businesses benefitting from reforms (registration, licensing, permits, taxes) Average # of days/procedures to comply with regulation Indicators to be Assessed at Enterprise level SMEs access financial products resulting in increases in: Sales/Revenue Assets Productivity Employees Investment Firms benefit though: Reduced costs to comply Increased investment Increased revenue Increased productivity Increased # of employees Formalization Source: Institutional/government indicators from various IFC results measurement documents Criteria for Selection of Documents Criteria were established for assessing the research and evaluations that would be included in the review. 1. For SME banking, the focus was on private sector interventions similar to IFC s work with financial institutions through advisory or investment services. 2. For Business Regulations, the focus was on public sector interventions similar to IFC s advisory work in business regulation. 60

61 3. Evaluations and research should be of good quality and including experimental, quasi and qualitative types of evaluations (e.g. before and after), cross country studies, and research. Good quality meant that the evaluations and reports had an appropriate methodology, addressed the key evaluation questions, demonstrated a linkage between the evidence and the conclusions and clearly presented the findings, conclusions and recommendations. They had to be completed between 2000 and Interventions for SME Banking with FIs targeted SMEs (enterprises) and not individuals or microenterprises. 5. The evaluations and research needed to show evidence or support for the key indicators that are being evaluated. Approach Paper An approach paper was developed outlining the purpose, methodology, criteria and approach for the review. This was reviewed by experts and stakeholders and feedback was integrated into the work plan for the systematic review. Identification and Appraisal Strategy The systematic review gathered and analyzed a wide range of research and evaluations on SME Banking and Business Regulation. To gather the initial group of documents, the following websites and databases will be reviewed, among others: IFC evaluation database; World Bank Development Impact Evaluation Initiative (DIME); World Bank Policy Research Papers and economic and sector work; Asian Development Bank (ADB); European Bank for Reconstruction and Development (EBRD); Inter-American Development Bank (IADB) OVE; African Development Bank (AfDB); Donor Committee for Enterprise Development (DCED); Global Partnership for Financial Inclusion; European Banking Center; Organisation for Economic Co-Operation and Development (OECD) Development Assistance Committee Evaluation Resource Center (DERec); Millennium Challenge Corporation; International Labor Organization (ILO); UN Industrial Development Organization (UNIDO); 3ie; Abdul Latif Jameel Poverty Action Lab (JPAL); JSTOR; National Bureau of Economic Research (NBER); Network of Networks on Impact Evaluation (NONIE); Science Direct; Google Scholar; Social Science Research Network (SSRN); Partnership for Economic Policy; and 61

62 A range of individual academic journals. A series of key words were used to retrieve relevant work from these sources. The following Table lists those used. Table 7 Key Words Used for Review Search SME Banking Business Regulation SME banking Business registration Access to finance Business regulation SME finance Business environment SME financing Investment climate SME financial institutions Regulatory environment Commercial banks Regulatory simplification Specialized financial institutions Regulatory impact analysis Small business financing Regulatory guillotine SME loans Standard cost model SME credit Regulatory governance SME banks Business license Savings Business inspections Financial products Red tape Financial services One stop shop Financial inclusion E-registration Corporate taxes Taxes Small business tax Tax reform Tax rate Tax compliance Tax policy Tax evasion Informal sector Informal economy Formalization This review produced over 630 documents to be considered. From this, an initial selection was made that met the criteria listed above. Then a further review and narrowing of the list was done as shown on Figure 4. 62

63 Figure 5 Process of Selecting Evaluations Steps Results # of Studies Step 1: Identify research and evaluations in SME Banking and Business Regulation from wide range of databases List of potential research and evaluations to consider 264 Step 2: Exclude research and evaluations not relevant to IFC interventions or not covering indicators Pool of relevant research and evaluations 127 Step 3: Exclude research and evaluations that are not of good quality or that provide insufficient evidence for the conclusions reached Database of qualified research and evaluations 91 The first level of review (step 2) focused on whether the research was appropriate given the nature of IFC interventions. For example, for SME Banking the focus needed to be on commercial banks not microfinance institutions or public lending programs. The targets groups needed to fall into the small and medium category not microenterprise. Where loan sizes were reported, they needed to follow the IFC loan proxy definitions for SME Banking namely be between USD10,000 and USD1-2 million depending on the country. In addition, the research and evaluations needed to specifically provide evidence on indicators being reviewed at either the institutional or firm level. Many documents were eliminated that were of a general nature and did not provide information on the indicators. The remaining pool was then narrowed by reviewing in-depth the quality of the resulting documents and excluding those that are deemed to be of insufficient quality to use (Step 3). Those evaluations that pass through these series of filters were analyzed to see what conclusions can be drawn on the review questions. Synthesis Once all the material to be assessed was collected, it was reviewed and analyzed. Gaps in the evidence were assessed and attempts made to find additional evidence. All this quantitative and qualitative information was assessed and synthesized and a draft report prepared. The report was then peer reviewed and revised. 63