83150 Microfinance Consensus Guidelines GUIDING PRINCIPLES ON REGULATION AND SUPERVISION OF MICROFINANCE July 2003 Building financial systems that work for the poor African Development Bank Germany: Federal Ministry for Economic Cooperation and Development Asian Development Bank Kreditanstalt für Wiederaufbau European Bank for Reconstruction Die Deutsche Gesellschaft für and Development Technische Zusammenarbeit European Commission Italy: Ministry of Foreign Affairs, Directorate General for Development Inter-American Development Bank Japan: Ministry of Foreign Affairs/ International Bank for Reconstruction and Japan Bank for International Development (The World Bank) Cooperation/ Ministry of Finance, Development Institution Division International Fund for Agricultural Development (IFAD) Luxembourg: Ministry of Foreign Affairs/Ministry of Finance International Labour Organization United Nations Development The Netherlands: Ministry of Foreign Affairs Programme/United Nations Capital Development Fund Norway: Ministry of Foreign Affairs/ Norwegian Agency for Development United Nations Conference on Cooperation Trade and Development Sweden: Swedish International Australia: Australian International Development Cooperation Agency Development Agency Switzerland: Swiss Agency for Belgium: Directorate General for Development and Cooperation Development Cooperation, Belgian Development Cooperation United Kingdom: Department for International Development Canada: Canadian International Development Agency United States: U.S. Agency for International Development Denmark: Royal Danish Ministry of Foreign Affairs Argidius Foundation Finland: Ministry of Foreign Affairs Ford Foundation of Finland France: Ministère des Affaires Etrangères France: Agence Française de Développement Microfinance Consensus Guidelines GUIDING PRINCIPLES ON REGULATION AND SUPERVISION OF MICROFINANCE by Robert Peck Christen Timothy R. Lyman Richard Rosenberg July 2003 © 2003 by CGAP/The World Bank Group 1818 H Street, N.W., Washington, D.C. 20433 USA All rights reserved Manufactured in the United States of America First printing June 2003 Photograph front cover: Women carrying buckets on top of their heads, Sénégal. (© 1993 The World Bank Photo Library/ Curt Carnemark) TABLE OF CONTENTS ACKNOWLEDGMENTS 3 INTRODUCTION 5 I TERMINOLOGY AND PRELIMINARY ISSUES 5 What is “Microfinance?” 5 Vocabulary of Microfinance Regulation 6 Prudential vs. Non-Prudential Regulation, and Enabling Regulation 7 Regulation as Promotion 8 “Special Windows” and Existing Financial Regulation 9 Regulatory Arbitrage 9 II NON-PRUDENTIAL REGULATORY ISSUES 10 Permission to Lend 10 Consumer Protection 10 Protection against Abusive Lending and Collection Practices 11 Truth in Lending 11 Fraud and Financial Crime Prevention 12 Credit Reference Services 12 Secured Transactions 13 Interest Rate Limits 13 Limitations on Ownership, Management, and Capital Structure 14 Tax and Accounting Treatment of Microfinance 14 Taxation of Financial Transactions and Activities 14 Taxation of Profits 14 Feasible Mechanisms of Legal Transformation 15 Regulation and Supervision 1 III PRUDENTIAL REGULATION OF MICROFINANCE 15 Objectives of Prudential Regulation 15 Drawing the Line: When to Apply Prudential Regulation in Microfinance? 16 Timing and the State of the Industry 16 Sources of Funding 16 Rationing Prudential Regulation, and Minimum Capital 18 Drawing Lines Based on Cost-Benefit Analysis 19 Regulate Institutions or Activities? 20 Special Prudential Standards for Microfinance 20 Minimum Capital 21 Capital Adequacy 21 Unsecured Lending Limits, and Loan-Loss Provisions 22 Loan Documentation 23 Restrictions on Co-Signers as Borrowers 23 Physical Security and Branching Requirements 23 Frequency and Content of Reporting 24 Reserves against Deposits 24 Ownership Suitability and Diversification Requirements 24 Who Should These Special Standards Apply To? 25 Deposit Insurance 26 IV FACING THE SUPERVISORY CHALLENGE 26 Supervisory Tools and Their Limitations 27 Costs of Supervision 28 Where to Locate Microfinance Supervision? 28 Within the Existing Supervisory Authority? 28 “Self-Regulation” and Supervision 29 Delegated Supervision 30 V KEY POLICY RECOMMENDATIONS 30 NOTES 33 2 Microfinance Consensus Guidelines ACKNOWLEDGMENTS These Guiding Principles were formally adopted by CGAP’s 29 member donor agencies in September 2002. The document was written by Robert Peck Christen, Timothy R. Lyman, and Richard Rosenberg, with input from more than 25 commentators who have worked on regulation and supervi- sion of microfinance in every region of the world. Mr. Christen and Mr. Rosenberg are Senior Advisors to CGAP. Mr. Lyman is President and Executive Director of the Day, Berry & Howard Foundation and chairs its Microfinance Law Collaborative. Regulation and Supervision 3 4 Microfinance Consensus Guidelines Microfinance Consensus Guidelines GUIDING PRINCIPLES ON REGULATION AND SUPERVISION OF MICROFINANCE INTRODUCTION Many developing countries and countries with transitional economies are con- sidering whether and how to regulate microfinance. Experts working on this topic do not agree on all points, but there is a surprisingly wide area of consen- sus. CGAP1 believes that the main themes of this paper would command gener- al agreement among most of the specialists with wide knowledge of past experi- ence and current developments in microfinance regulation. We hope this paper will provide useful guidance not only to the staff of the international donors who encourage, advise, and support developing- and tran- sitional-country governments, but also to the national authorities who must make the decisions, and the practitioners and other local stakeholders who par- ticipate in the decision-making process and live with the results. On some ques- tions, experience justifies clear conclusions that will be valid everywhere with few exceptions. On other points, the experience is not clear, or the answer depends on local factors, so that no straightforward prescription is possible. On these lat- ter points, the best this paper can do for the time being is to suggest frameworks for thinking about the issue and identify some factors that need special consid- eration before reaching a conclusion. Part I of the paper discusses terminology and preliminary issues. Part II out- lines areas of regulatory concern that do not call for “prudential” regulation (see the definition and discussion below). Part III discusses prudential treatment of microfinance and MFIs. Part IV briefly looks at the challenges surrounding supervision, and Part V summarizes some key policy recommendations. I TERMINOLOGY AND PRELIMINARY ISSUES “What is “Microfinance?” As used in this paper, “microfinance” means the provision of banking services to lower-income people, especially the poor and the very poor. Definitions of these groups vary from country to country. Regulation and Supervision 5 The term “microfinance” is often used in a much narrower sense, referring principally to microcredit2 for tiny informal businesses of microentrepreneurs, delivered using methods developed since 1980 mainly by socially-oriented non- governmental organizations (NGOs). This paper will use “microfinance” more broadly. The clients are not just microentrepreneurs seeking to finance their business- es, but the whole range of poor clients who also use financial services to manage emergencies, acquire household assets, improve their homes, smooth consump- tion, and fund social obligations. The services go beyond microcredit. Also included are savings and transfer services.3 The range of institutions goes beyond NGOs and includes commercial banks, state-owned development banks, financial cooperatives, and a variety of other licensed and unlicensed non-bank institutions. Vocabulary of Microfinance Regulation and Supervision Varying terminology used in the discussion of microfinance regulation some- times leads to confusion. This paper uses the following general definitions: Microfinance institution (MFI)—A formal organization whose primary activity is microfinance. Regulation—Binding rules governing the conduct of legal entities and individu- als, whether they are adopted by a legislative body (laws) or an executive body (regulations). Regulations—The subset of regulation adopted by an executive body, such as a ministry or a central bank. “Banking” law or regulations—For the sake of simplicity, the paper uses “bank- ing” in this context to embrace existing laws or regulations for non-bank finan- cial institutions as well. Prudential (regulation or supervision)—Regulation or supervision is prudential when it governs the financial soundness of licensed intermediaries’ businesses, in order to prevent financial-system instability and losses to small, unsophisticated depositors. Supervision—External oversight aimed at determining and enforcing compliance with regulation. For the sake of simplicity, “supervision” in this paper refers only to prudential supervision. Financial intermediation—The process of accepting repayable funds (such as funds from deposits or other borrowing) and using these to make loans. License—Formal governmental permission to engage in financial-service delivery that will subject the license-holding institution to prudential regulation and supervision. 6 Microfinance Consensus Guidelines Permit—Formal governmental permission to engage in non-depository microlending activity that will not subject the permit-holding institution to pru- dential regulation and supervision. Self-regulation/supervision—Regulation or supervision by a body that is effec- tively controlled by the entities being regulated or supervised. Prudential vs. Non-Prudential Regulation, and Enabling Regulation Regulation is “prudential” when it is aimed specifically at protecting the financial system as a whole as well as protecting the safety of small deposits in individual institutions. When a deposit-taking institution becomes insol- vent, it cannot repay its depositors, and—if it is a large institution—its failure could undermine public confidence enough so that the banking system suffers a run on deposits. Therefore, prudential regulation involves the government in overseeing the financial soundness of the regulated institutions: such reg- ulation aims at ensuring that licensed institutions remain solvent or stop collect- ing deposits if they become insolvent. This concept is emphasized because great confusion results when regulation is discussed without distinguishing between prudential and non-prudential issues.4 Prudential regulation is generally much more complex, difficult, and expen- sive than most types of non-prudential regulation. Prudential regulations (for instance, capital adequacy norms or reserve and liquidity requirements) almost always require a specialized financial authority for their implementation, where- as non-prudential regulation (for instance, disclosure of effective interest rates or of the individuals controlling a company) may often be largely self-executed and can often be dealt with by other than the financial authorities. Thus, an important general principle is to avoid using burdensome pru- dential regulation for non-prudential purposes—that is, purposes other than protecting depositors’ safety and the soundness of the financial sector as a whole. For instance, if the concern is only to keep persons with bad records from owning or controlling MFIs, the central bank does not have to take on the task of monitoring and protecting the financial soundness of MFIs. It would be sufficient to require registration and disclosure of the individuals owning or con- trolling them, and to submit proposed individuals to a “fit and proper” screen- ing. Some non-prudential regulation can be accomplished under general com- mercial laws, and administered by whatever organs of government implement those laws, depending on the relative capacity of those agencies. Even where it has hundreds of thousands of customers, microfinance today seldom accounts for a large enough part of a country’s financial assets to pose serious risk to the overall banking and payments system. Thus, the rest of this discussion assumes that at present the main justification of prudential regulation of depository microfinance is protection of those who make deposits in MFIs. (On the other hand, the development of the microfinance is not static. Wherever depository microfinance reaches significant scale in a particular region or coun- Regulation and Supervision 7 try, systemic risk issues must be taken into consideration, in addition to deposi- tor protection issues. The failure of a licensed MFI with relatively small assets but huge numbers of customers could be contagious for other MFIs.) Certain regulation is aimed at correcting perceived abuses in an existing industry. Other regulation is “enabling”: its purpose is a positive one—to allow the entry of new institutions or new activities. Most of the microfinance regula- tion being proposed today is enabling. But what is the activity being enabled? Where the purpose is to enable MFIs to take deposits from the public, then pru- dential regulation is generally called for, because the return of depositors’ money cannot be guaranteed unless the MFI as a whole is financially solvent. If, on the other hand, the regulation’s purpose is to enable certain institutions to conduct a lending business legally, then there is usually no reason to assume the burden of prudential regulation, because there are no depositors to protect.5 The general discussion of microfinance regulation worldwide tends to emphasize prudential issues—how to enable MFIs to take deposits. However, in some countries, especially formerly-socialist transitional economies, the most pressing issues are non-prudential—how to enable MFIs to lend legally. Regulation as Promotion For some, the main motivation for regulatory change is to encourage formation of new MFIs and/or improve performance of existing institutions. In the case of both prudential and non-prudential regulation, providing an explicit regulatory space for microfinance may very well have the effect of increasing the volume of financial services delivered and the number of clients served. The right type of non-prudential regulation can frequently have the desired promotional effect with relatively low associated costs (see, for example, the discussion of permis- sion to lend on page 10). In the case of prudential regulation, however, experi- ence to date suggests that opening up a new, less burdensome regulatory option—particularly if existing MFIs are not yet strong candidates for transfor- mation—can sometimes result in a proliferation of under-qualified depository institutions, and create a supervisory responsibility that cannot be fulfilled. In several countries, a new prudential licensing window for small rural banks result- ed in many new institutions providing service to areas previously without access, but supervision proved much more difficult than anticipated. As many as half of the new banks turned out to be unsound, and the central bank had to devote excessive resources to cleaning up the situation. Nevertheless, many of the new banks remained to provide rural services. Whether the final outcome was worth the supervisory crisis is a balancing judgment that would depend on local factors and priorities. Any discussion of providing an explicit new regulatory space in order to devel- op the microfinance sector and improve the performance of existing MFIs should weigh carefully the potential unintended consequences. For instance, the politi- cal process of regulatory change can lead to reintroduction or renewed enforce- ment of interest rate caps (see the discussion of interest rate limitations on page 13). In addition, over-specific regulation can limit innovation and competition. 8 Microfinance Consensus Guidelines “Special Windows” and Existing Financial Regulation Discussion and advocacy regarding microfinance regulation often focuses on whether or how to establish a “special window”—that is, a distinct form of license and/or permit—for microfinance. The range of regulatory approaches possible, whether or not they are understood as special windows for microfinance, is limit- ed. It is important to be clear about which of these is being pursued: • Enabling non-bank microlending institutions, which should not require pru- dential regulation and supervision • Enabling non-bank financial intermediaries taking retail deposits, which gen- erally does require prudential treatment • Enabling a combination of these two If a new special window is to be established, should it be done by amendment of the existing financial sector laws and regulations, or should separate legislation or regulation be proposed? As a general proposition, incorporation within the existing framework will better promote integration of the new license and/or permit into the overall financial system. This approach may increase the likeli- hood that the regulatory changes are properly harmonized with the existing reg- ulatory landscape. Inadequate attention to harmonization has often led to ambi- guities about how the various pieces of regulation fit together. Moreover, adjust- ing the existing framework may be technically easier, and may be more likely to facilitate the entry of existing financial institutions into microfinance. However, local factors will determine the feasibility of this approach. In some countries, for example, policymakers may be reluctant to open up the banking law for amend- ment because it would invite reconsideration of a whole range of banking issues that have nothing to do with microfinance. Regulatory Arbitrage In any event, the content of the regulation involved is likely to be more impor- tant than whether it is implemented within existing laws and regulations, or whether it is specifically designated as new “microfinance regulation.” In either case—but particularly if new categories of institution are added to the regulato- ry landscape—critical attention must be paid to the interplay between the new rules and the ones already in place. If the new rules appear to establish a more lightly or favorably regulated environment, many existing institutions and new market entrants may contort to qualify as MFIs. Such regulatory arbitrage can leave some institutions under-regulated. Several countries have carefully crafted a special regulatory window for social- ly-oriented microfinance, only to find that the window is later used by types of businesses that are very different from what the framers of the window had in mind. This is particularly the case with consumer lending, which generally goes to salaried workers rather than self-employed microentrepreneurs. In some cases, these lenders could easily have gotten a banking license, but they opted to use the microfinance window instead because minimum capital and other require- ments were less stringent. Regulation and Supervision 9 II NON-PRUDENTIAL REGULATORY ISSUES Most of the current discussion of microfinance regulation focuses on prudential regulation. Nevertheless, this paper will treat non-prudential issues first, to underscore the point that there are many regulatory objectives that do not require prudential treatment. Non-prudential (“conduct of business”) regulatory issues, relevant to micro- finance, span a wide spectrum. These issues include enabling the formation and operation of microlending institutions; protecting consumers; preventing fraud and financial crimes; setting up credit information services; supporting secured transactions; developing policies with respect to interest rates; setting limitations on foreign ownership, management, and sources of capital; identifying tax and accounting issues; plus a variety of cross-cutting issues surrounding transforma- tions from one institutional type to another. Permission to Lend In some legal systems, any activity that is not prohibited is implicitly permissible. In these countries, an NGO or other unlicensed entity has an implicit authori- zation to lend as long as there is no specific legal prohibition to the contrary. In other legal systems, especially in formerly-socialist transitional countries, an institution’s power to lend—at least as a primary business—is ambiguous unless there is an explicit legal authorization for it to conduct such a business. This ambiguity is particularly common in the case of NGO legal forms. In still other legal systems, only prudentially licensed and regulated institutions are permitted to lend, even if no deposit taking is involved. Where the legal power to lend is either ambiguous or is prohibited to institutions that are not prudentially licensed, a strong justification exists for introducing non-prudential regulation that explicitly authorizes non-depository MFIs to lend. Where the objective is to enable lending by NGOs, modification of the general legislation governing them may be needed. Regulation of permission to lend should be relatively simple. Sometimes not much more is needed than a public registry and permit-issuing process. The scope of documents and information required for registration and the issuance of a permit should be linked to specific regulatory objectives, such as providing a basis for governmental action in case of abuse (see the discussion of fraud and financial crime prevention on page 12) and enabling industry performance benchmarking. Consumer Protection Two non-prudential consumer-protection issues are particularly relevant to microfinance and are likely to warrant attention in most, if not all, countries: pro- tecting borrowers against abusive lending and collection practices, and provid- ing borrowers with truth in lending—accurate, comparable, and transparent information about the cost of loans. 10 Microfinance Consensus Guidelines Protection against Abusive Lending and Collection Practices There is often a concern about protecting microcredit clients against lenders who make loans without enough examination of the borrower’s repayment capacity. This can easily lead to borrowers becoming over-indebted, resulting in higher defaults for other lenders. In a number of countries, consumer lenders have proved particularly susceptible to this problem, and governments have found it necessary regulate against such behavior. In addition, there is often con- cern about unacceptable loan-collection techniques. Regulation in these areas does not necessarily have to be administered by the prudential supervisory authority. Truth in Lending As discussed in the section on interest rate limits (page 13), the administrative cost of disbursing and collecting a given amount of portfolio is much higher if there are many tiny loans than if there are a few large loans. For this reason, microlending usually cannot be done sustainably unless the borrowers pay inter- est rates that are substantially higher than the rates banks charge to their tradi- tional borrowers. Moreover, different combinations of transaction fees and interest-calculation methods can make it difficult for a borrower to compare interest rates of lenders. In many countries, lenders are required to disclose their effective interest rates to loan applicants, using a uniform formula mandated by the government. Should such truth-in-lending rules be applied to microcredit? Microlenders usually argue strongly against such a requirement. It is easy to be cynical about their motives for doing so, and certainly the burden of proof should lie with anyone who argues against giving poor borrowers an additional tool to help them evaluate a loan’s cost—especially when this tool will promote price competition. Moreover, the mandated discipline of disclosing effective interest rates may help to focus microlenders on steps they can take to increase their efficiency and thus lower their rates. So there ought to be a presumption in favor of giving borrowers full and usable information about interest rates. But the issue is not always simple. In many coun- tries, the public prejudice against seemingly exploitative interest rates is very strong. Even where high interest rates on tiny loans make moral and financial sense, it may still prove difficult to defend them when they are subjected to broad (and uninformed) public discussion, or when politicians exploit the issue for polit- ical advantage. Micro-borrowers show again and again that they are happy to have access to loans even at high rates. But if MFIs are required to express their pric- ing as effective interest rates, then the risk of a public and political backlash becomes greater, and can threaten the ability of microlenders to operate. Obviously, the seriousness of this risk will vary from one country to another. In some places, this risk can be dealt with through concerted efforts to educate the public and policymakers about why loan charges in microfinance are high, and why access is more important than price for most poor borrowers. But pub- lic education of this sort takes significant time and resources and will not always be successful. Regulation and Supervision 11 Fraud and Financial Crime Prevention Two types of concern related to fraud and financial crimes predominate in con- nection with microfinance regulation: (1) concerns about securities fraud and abusive investment arrangements such as pyramid schemes, and (2) money-laun- dering concerns. In addressing these, the same rules should apply to MFIs as to other economic actors. It should not be assumed automatically that the best body to deal with these concerns is the one responsible for prudential regulation. In many countries, the existing anti-fraud and financial crime regulation will be adequate to address abuse in the case of MFIs, or will need amendment only to add any new categories of institution to the regulatory landscape. Often the most pressing need is to improve enforcement of existing laws. Credit Reference Services Credit reference services—called by a variety of names including credit bureaus—offer important benefits both to financial institutions and to their cus- tomers. By collecting information on clients’ status and history with a range of credit sources, these databases allow lenders to lower their risks, and allow bor- rowers to use their good repayment record with one institution to get access to new credit from other institutions. Such credit bureaus allow lenders to be much more aggressive in lending without physical collateral, and strengthen borrow- ers’ incentive to repay. Depending on the nature of the database and the condi- tions of access to it, credit information can also have a beneficial effect on com- petition among financial service providers. In developed countries, the combination of credit bureaus and statistical risk- scoring techniques has massively expanded the availability of credit to lower- income groups. In developing countries, especially those without a national iden- tity-card system, practical and technical challenges abound, but new technologies (such as thumbprint readers and retinal scanners) may offer solutions. Experience suggests that when MFIs begin to compete with each other for customers, over- indebtedness and default will rise sharply unless the MFIs have access to a com- mon database that captures relevant aspects of their clients’ borrowing behavior. Does the government need to create a credit bureau or require participation in it? The answer will vary from country to country. A common pattern in devel- oping countries is that merchants participate voluntarily in private credit bureaus, but bankers are more reluctant to share customer information unless the law requires them to do so. Especially when banks participate in them, credit information services raise privacy issues. Sometimes these issues can be handled simply by including in loan contracts the borrowers’ authorization for the lender to share information on their credit performance with other lenders. In other circumstances, laws will need to be amended. Credit information services can provide clear benefits, but such data collec- tion can entail risks. Corrupt database managers may sell information to unau- thorized parties. Tax authorities may want to use the database to pursue unreg- 12 Microfinance Consensus Guidelines istered microenterprises. Borrowers can be hurt by inaccurate information in the database, although guaranteeing them access to their own credit histories can lower this risk. For donors wanting to help expand access to financial services for both poor and middle-class people, development of private or public credit information systems that include micro-borrowers could be an attractive target of support in countries where the conditions are right. Among these conditions are a national identity system or some other technically feasible means of identifying clients, a fairly mature market of MFIs or other firms that lend to low-income borrowers, and a legal framework that creates the right incentives for participation as well as protecting fairness and privacy. Secured Transactions Borrowers, lenders, and the national economy all benefit when not only real estate but also moveable assets can be pledged as collateral for loans. But in many developing and transitional economies, it is expensive or impossible to create and enforce a security interest in moveable collateral. Sometimes there are also con- straints that make it hard for lower-income people to use their homes and land as collateral. Legal and judicial reform to support secured transactions can be very worthwhile, although these matters tend to affect the middle class more than they do the poor. Such reform typically centers on the commercial and judi- cial laws, not the banking law. Interest Rate Limits To break even, lenders need to set loan charges that will cover their cost of funds, their loan losses, and their administrative costs. The cost of funds and of loan loss varies proportionally to the amount lent. But administrative costs do not vary in proportion to the amount lent. One may be able to make a $20,000 loan while spending only $600 (3 percent) in administrative costs; but this does not mean that administrative costs for a $200 loan will be only $6. In comparison with the amount lent, administrative costs are inevitably much higher for microcredit than for conventional bank loans.6 Thus, MFIs cannot continue to provide tiny loans unless their loan charges are considerably higher in percentage terms than normal bank rates. Legislatures and the general public seldom understand this dynamic, so they tend to be outraged at microcredit interest rates even in cases where those rates reflect neither inefficiency nor excessive profits.7 Therefore, if the government takes on control of microcredit interest rates, practical politics will usually make it difficult to set an interest rate cap high enough to permit the development of sus- tainable microcredit. Interest rate caps, where they are enforced, almost always hurt the poor—by limiting services—far more than they help the poor by lower- ing rates. Some international donors assume too easily that the argument over high interest rates for microcredit has been won. But recently there have been back- Regulation and Supervision 13 lashes in many countries. Before donors and governments commit to build- ing an enabling regulatory framework for microfinance, they need to con- sider the possibility that the process may unavoidably entail political dis- cussion of interest rates, with results that could damage responsible micro- credit. Experience shows that this risk is real, although it is certainly not relevant in all countries. Limitations on Ownership, Management, and Capital Structure In many legal systems, citizenship, currency, and foreign-investment regulations create hurdles for some forms of MFI. Common problems include prohibitions or severe limitations on the participation of foreign-equity holders (or founders or members in the case of NGOs), borrowing from foreign sources, and employment of non-citizens in management or technical positions. In many countries, the microfinance business will not attract conventional commercial investors for some years yet. Since alternative sources of investment—particular- ly equity investment—tend to be international, limitations on foreign investment can be especially problematic.8 Tax and Accounting Treatment of Microfinance Taxation of MFIs is becoming a controversial topic in many countries. Local fac- tors may call for differing results, but the following approach is suggested as a starting point for the analysis. It is based on a distinction between taxes on finan- cial transactions and taxes on net profits arising from such transactions. Taxation of Financial Transactions and Activities With respect to taxes on financial transactions, such as a value-added tax on lend- ing or a tax on interest revenue, the critical issue is a level playing field among institutional types. In some countries, favorable tax treatment on transactions is available only to prudentially licensed institutions, even though the favorable tax treatment bears no substantive relationship to the objectives of prudential regu- lation. In other countries, financial-transaction taxes affect financial cooperatives differently from banks. Absent other considerations, favorable transaction tax treatment should be based on the type of activity or transaction, regardless of the nature of the institution and whether it is prudentially licensed. To do otherwise gives one form of institution an arbitrary advantage over another in carrying out the activity. Taxation of Profits It can reasonably be argued that not-for-profit NGO MFIs ought to be treated the same as all other public-benefit NGOs when the tax in question is a tax on net profits. The reason for exemption from profits tax is the principle that the NGO is rendering a recognized public benefit and does not distribute its net sur- pluses into the pockets of private shareholders or other insiders. Rather, it rein- vests any surplus to finance more socially-beneficial work. To be sure, there are always ways to evade the spirit of this non-distribution principle, such as exces- 14 Microfinance Consensus Guidelines sive compensation and below-market loans to insiders. However, these potential abuses probably occur no more commonly in NGOs engaged in microlending than in other types of NGOs. For any institution subject to a net income or profits tax, rules for tax deductibility of expenses (such as reasonable provisioning for bad loans) should apply consistently to all types of institutions, regardless of whether they are pru- dentially licensed. Moreover, if it is appropriate to provision a microloan portfo- lio more aggressively than a conventional loan portfolio, then the microlender’s profits tax deduction should also vary accordingly. For licensed institutions, pru- dential regulation will normally dictate the amount of loan-loss provisioning. In the case of unlicensed lending-only institutions, the tax authorities may need to regulate allowable amounts of provisioning in order to prevent abuse. Feasible Mechanisms of Legal Transformation Legal transformations in microfinance—from one institutional type to another— raise a variety of crosscutting non-prudential regulatory issues. The simplest and most common type of transformation occurs when an existing MFI operation is transferred to the local office of an international NGO as a new, locally-formed NGO. Such a transfer can face serious regulatory obstacles, including limits on foreign participation, ambiguous or prohibitive taxation of the portfolio transfer, and labor law issues created by the transfer of staff. A second, increasingly com- mon type of legal transformation involves the creation of a commercial compa- ny by an NGO (sometimes together with other investors), to which the NGO contributes its existing portfolio (or cash from the repayment of its portfolio) in exchange for shares in the new company. Such transformations often raise addi- tional issues, including how to recapture or otherwise make allowance for tax benefits that the transforming NGOs have received; restrictions on the NGO’s power to transfer what are deemed “charitable assets” (its loans) to a privately- owned company; and restrictions on the NGO’s power to hold equity in a com- mercial company, particularly if this will become its principal activity as a result of the transformation. Ordinarily, these disparate bodies of regulation do not contemplate, and have never been applied to, microfinance transformations. Harmonizing their provi- sions and creating a clear path for microfinance transformations can be an impor- tant enabling reform. On the other hand, such reform may be a lower priority if there are only one or two microfinance NGOs who are likely candidates for transformation.9 III PRUDENTIAL REGULATION OF MICROFINANCE Objectives of Prudential Regulation The generally agreed objectives of prudential regulation include (1) protecting the country’s financial system by preventing the failure of one institution from leading to the failure of others, and (2) protecting small depositors who are not Regulation and Supervision 15 well positioned to monitor the institution’s financial soundness themselves. If prudential regulation does not focus closely enough on these objectives, scarce supervisory resources can be wasted, institutions can be saddled with unneces- sary compliance burdens, and development of the financial sector can be con- strained. Drawing the Line: When to Apply Prudential Regulation in Microfinance? Timing and the State of the Industry New regulatory windows for microfinance are being considered in many coun- tries today. In a few of these countries, a somewhat paradoxical situation exists. The expectation is that, over the medium term, the new window will be used mainly by existing NGO MFIs that want to change to deposit-taking status. But at the same time, none or almost none of the existing MFIs have yet demon- strated that they can manage their lending profitably enough to pay for and pro- tect the deposits they want to mobilize. In such a setting, the government should consider the option of waiting and monitoring microlenders’ performance, and open the window only after there is more and better experience with the finan- cial performance of the MFIs. Developing a new regulatory regime for microfi- nance takes a great deal of analysis, consultation, and negotiation; the costs of the process can exceed the benefits unless a critical mass of qualifying institutions can be expected. In this context, the actual financial performance of existing MFIs is a crucial element that often gets too little attention in discussions of regulatory reform. Whenever there is an expectation that existing MFIs will take advantage of a new regulatory window, there should be a competent financial analysis of at least the leading MFIs before decisions are made with respect to that window. This analysis should focus on whether each MFI’s existing operations are profitable enough so that it can pay the financial and administrative costs of deposit-taking without decapitalizing itself. Naturally, this analysis will have to include a determination of whether the MFI’s accounting and loan-tracking sys- tems are sound enough to produce reliable information. Sources of Funding Both the objective to prevent risk to the financial system and the objective to protect depositors, of prudential regulation, are served when retail deposits of the general public are protected. Thus, raising funds from this source will usual- ly call for prudential regulation. Are MFIs that fund their lending from other sources of capital also engaged in financial intermediation that needs to be pru- dentially regulated? This question needs close analysis, and its answer will often depend on local factors.10 Donor grants. Historically, donors of one type or another, including bilateral and multilateral development agencies, have supported MFIs with grants. The justi- fications for prudential supervision do not apply in the case of MFIs funded only by donor grants. The government may have an interest in seeing that donor 16 Microfinance Consensus Guidelines funds are well spent, but microfinance is no different in this respect from any other donor-supported activity. Cash collateral and similar obligatory deposits. Many MFIs require cash deposits from borrowers before and/or during a loan, in order to demonstrate the bor- rower’s ability to make payments, and to serve as security for the repayment of the loan. Even though these deposits are often called “compulsory savings,” it is more useful to think of them as cash collateral required by the loan contract, rather than as a true savings service. This cash collateral is some- times held by a third party (such as commercial bank), and thus is not inter- mediated by the MFI. Even where the MFI holds its clients’ obligatory deposits, and even if it intermediates them by lending them out, the ques- tion of whether to apply prudential regulation should be approached from the standpoint of practically weighing the costs and benefits. If cash collat- eral is the only form of deposit taken by the MFI, then most of its customers owe more to the MFI than the MFI owes to them, most of the time. If the MFI fails, these customers can protect themselves by simply ceasing repayment of their loan. It is true that some of the MFI’s customers will be in a net at-risk position some of the time, so that the MFI’s failure would imperil their deposits, but this relatively lesser risk needs to be weighed against the various costs of prudential supervision—costs to the supervisor, to the MFI, and to the customer. Several countries have taken a middle path on this issue, requiring prudential licensing only for MFIs that hold and intermediate their clients’ cash collateral, but not for MFIs that keep such collateral in low-risk securities or in an account with a licensed bank. Borrowing from non-commercial sources, including donors or sponsors. Increasingly, donors are using loans rather than grants to support MFIs. Although the loan proceeds are intermediated by the MFI, their loss would pose no substantial sys- temic risk in the host country, and the lenders are well-positioned to protect their own interests if they care to. The definition of deposit-taking that triggers pru- dential regulation should therefore exclude this type of borrowing. Commercial borrowing. Some MFIs get commercial loans from international investment funds that target social-purpose investments, and from locally licensed commercial banks. Here, too, the fact that commercial loan proceeds are intermediated by the MFI should not lead to prudential regulation of the borrowing MFI. Where the lender is an international investment fund, the loss of its funds will not pose systemic risk, and the lender should be able to look out for its own interests. Where the lender is a locally-licensed commercial bank, it should itself already be subject to appropriate prudential regulation, and the fact that an MFI borrows from the bank does not justify prudential regulation of the MFI any more than would be the case for any other borrower from the bank.11 Wholesale deposits and deposit substitutes. In some countries, MFIs can finance themselves by issuing commercial paper, bonds, or similar instruments in the Regulation and Supervision 17 local securities markets. Similar issues are presented by the direct issuance of large certificates of deposit. Unlike deposits from the general public, all these instruments tend to be bought by large, sophisticated investors. There is not a consensus on how to regulate such instruments. Some argue that the buyers of these instruments ought to be able to make their own analysis of the financial soundness of the issuing business. Therefore, they would subject the issuer only to normal securities regulation, which generally focuses on insuring complete disclosure of relevant information, rather than giving any assurance as to the financial strength of the issuer. Others, less impressed by the distinction between wholesale and retail deposits or skeptical about the local securities law and enforcement, insist that any institution issuing such instruments and intermedi- ating the funds be prudentially regulated. Members’ savings. Much of the current discussion of microfinance regulation focuses, implicitly or explicitly, on NGO MFIs that have begun with a credit- based model and now want to move to capturing deposits. But in large parts of the world, most microfinance is provided by financial cooperatives that typically fund their lending from members’ share deposits and savings. It is sometimes argued that, because these institutions take deposits only from members and not from “the public,” they need not be prudentially supervised. This argument is problematic. In the first place, when a financial cooperative becomes large, its members as a practical matter may be in no better a position to supervise man- agement than are the depositors in a commercial bank. Secondly, the boundaries of membership can be porous. For instance, financial cooperatives whose com- mon bond is geographical can capture deposits as extensively as they want by the simple expedient of automatically giving a membership to anyone in their area of operations who wants to make a deposit. Often such financial cooperatives are licensed under a special law, and their supervision may be lodged in the government agency that supervises all cooper- atives, including cooperatives focused on production, marketing, and other non- financial activities. While these agencies may be legally responsible for prudential supervision of the safety of depositors, they almost never have the resources, expertise, and independence to do that job effectively. Absent strong local rea- sons to the contrary, financial cooperatives—at least large ones—should be prudentially supervised by a specialized financial authority. In countries with a large existing base of financial cooperatives, securing effective regulation and supervision of these cooperatives may be a more immediate priority than devel- oping new windows for NGO microfinance. Rationing Prudential Regulation, and Minimum Capital As discussed below, prudential supervision is expensive. When measured as a per- centage of assets supervised, these expenses are higher for small institutions than for large ones. Furthermore, supervisory authorities have limited resources. As a practical matter, there is a need to ration the number of financial licenses that will require supervision. The most common tool for this rationing is a minimum cap- 18 Microfinance Consensus Guidelines ital requirement—the lowest amount of currency that owners can bring to the equity account of an institution seeking a license. In theory, setting of minimum capital could be based on economies of scale in financial intermediation: in other words, below a certain size, an intermediary can- not support the minimum necessary infrastructure and still operate profitably. However, there is an increasing tendency to downplay the utility of minimum cap- ital as a safety measure and instead to treat it more straightforwardly as a rationing tool. The lower the minimum capital, the more entities will have to be supervised. Those who see regulation of microfinance primarily as promotion will want low minimum-capital requirements, making it easier to obtain new licenses. On the other hand, supervisors who will have to oversee the financial soundness of new deposit-taking institutions tend to favor higher capital requirements, because they know there are limits on the number of institutions they can supervise effec- tively. To put the point simply, there is a trade-off between the number of new institutions licensed and the likely effectiveness of the supervision they will receive. The most common tool for drawing the balance is minimum capital. However, minimum capital is not necessarily the only tool available to limit new market entrants. For example, licensing decisions can be based in part on qualitative institutional assessments—though qualitative standards leave more room for abuse of official discretion.12 Whatever rationing tools are used, it would seem reasonable to err on the side of conservatism at first, as long as the requirements can be adjusted later, when the authorities have more experience with the demand for licenses and the prac- ticalities of microfinance supervision. Obviously, such flexibility is easier if the requirements are placed in regulations rather than in the law. Drawing Lines Based on Cost-Benefit Analysis— The case of small community-based intermediaries Some member-owned intermediaries take deposits but are so small, and some- times so geographically remote, that they cannot be supervised on any cost- effective basis. This poses a practical problem for the regulator. Should these institutions be allowed to operate without prudential supervision, or should minimum-capital or other requirements be enforced against them so that they have to cease taking deposits? Sometimes regulators are inclined to the latter course. They argue that insti- tutions that cannot be supervised are not safe, and therefore should not be allowed to take small depositors’ savings.13 After all, are not small and poor cus- tomers just as entitled to safety as large and better-off customers? But this analysis is too simple if it does not consider the actual alternatives available to the depositor. Abundant studies show that poor people can and do save. Especially where formal deposit accounts are not available, they use savings tools, such as currency under the mattress, livestock, building materials, or infor- mal arrangements like rotating savings and credit clubs. All of these vehicles are risky, and in many if not most cases, they are more risky than a formal account in a small unsupervised intermediary. Closing down the local savings and loan Regulation and Supervision 19 cooperative may in fact raise, not lower, the risk faced by local savers by forcing them back to less satisfactory forms of savings. Because of these considerations, most regulators facing the issue have chosen to exempt community-based intermediaries below a certain size from require- ments for prudential regulation and supervision. The size limits are determined by number of members, amount of assets, or both. (Sometimes the exemption is available only to “closed bond” institutions whose services are available only to members of a pre-existing group.) Once the limits are exceeded, the institu- tion must comply with prudential regulation and be supervised. If small intermediaries are allowed to take deposits without prudential super- vision, a good argument can be made that their customers should be clearly advised that no government agency is monitoring the health of the institution, and thus that they need to form their own conclusions based on their knowledge of the individuals running the institution. These issues presented by very small intermediaries illustrate a more general principle that applies to many of the topics discussed in this paper. Depositor protection is not an absolute value that overrules all other considerations. Some rules that lower risk can also lower poor people’s access to financial services, an equally important value. In such cases, the regulator’s objective should be, not the elimination of risk, but rather a prudent balancing of safety and access. Regulate Institutions or Activities? When trying to open up regulatory space for microfinance, there is a natural ten- dency to think in terms of creating a new, specialized institutional type. In some settings this is the best option. But alternatives should be considered, including the possibility of fine-tuning an existing form of financial license. There is some danger that too exclusive a focus on a particular institutional form may cramp innovation and competition, encourage regulatory arbitrage, or impede the inte- gration of microfinance into the broader financial sector. These considerations lead some in the field to argue that policy makers should focus more on regulating microfinance as a set of activities, regardless of the type of financial institution carrying them out, and less on particular institutional forms. This is a healthy emphasis. The following section discusses special regula- tory adjustments needed for microfinance; almost all of these adjustments would be applicable no matter what type of institution is doing microfinance. At the same time, a few of the necessary regulatory adjustments will have to do with the type of institution rather than the activity itself. For instance, microlending arguably presents a lower risk profile when it is a small part of the portfolio of a diversified full-service bank, compared to microlending that constitutes the majority of a specialized MFI’s assets; thus, it can reasonably be argued that these two institutional types ought to be subject to different capital-adequacy rules. Special Prudential Standards for Microfinance Some regulations common in traditional banking need to be adjusted to accom- modate microfinance. Whether microfinance is being developed through spe- 20 Microfinance Consensus Guidelines cialized stand-alone depository MFIs, or as product lines within retail banks or finance companies, the following sets of regulations will commonly need reex- amination, at least for those products that can fairly be categorized as “micro.” Other rules may require adjustment in some countries, but the list below includes the most common issues. Minimum Capital The kind of investors who are willing and able to finance MFIs may not be able to come up with the amount of capital required for a normal bank license. Furthermore, it might take a specialized MFI a long time to build a portfolio large enough to leverage adequately the amount of equity required for a bank. The trade-offs involved in setting minimum capital requirements for microfi- nance were discussed on page 18. Capital Adequacy14 There is controversy as to whether the capital adequacy requirements for special- ized MFIs should be tighter than the requirements applied to diversified com- mercial banks. A number of factors argue in the direction of such conservatism. Well-managed MFIs maintain excellent repayment performance, with delin- quency typically lower than in commercial banks. However, MFI portfolio tends to be more volatile than commercial bank portfolio, and can deteriorate with surprising speed. The main reason for this is that microfinance portfolio is usu- ally unsecured, or secured by assets that are insufficient to cover the loan, once collection costs are included. The borrower’s main incentive to repay a microloan is the expectation of access to future loans. Thus, outbreaks of delin- quency in an MFI can be contagious. When a borrower sees that others are not paying back their loans, that borrower’s own incentive to continue paying declines, because the outbreak of delinquency makes it less likely that the MFI will be able to reward the borrower’s faithfulness with future loans. Peer dynam- ics play a role as well: when borrowers have no collateral at risk, they may feel foolish paying their loans when others are not. In addition, because their costs are high, MFIs need to charge high interest rates to stay afloat. When loans are not being paid, the MFI is like any bank in that it is not receiving the cash it needs to cover the costs associated with those loans. However, the MFI’s costs are usually much higher than a commercial bank’s costs per unit lent, so that a given level of delinquency will decapitalize an MFI much more quickly than it would decapitalize a typical bank. Another relevant factor is that in most countries, neither microfinance as a business nor individual MFIs as institutions have a very long track record. Management and staff of the MFIs tend to be relatively inexperienced, and the supervisory agency has little experience with judging and controlling microfi- nance risk. Furthermore, many new MFIs are growing very fast, which puts heavy strain on management and systems. Finally, as will be discussed below, some important supervisory tools do not work very well for specialized MFIs. Regulation and Supervision 21 For all these reasons, a prudent conservatism would seem to suggest that spe- cialized MFIs be subject to a higher capital-adequacy percentage than is applied to normal banks, at least until some years of historical performance have demon- strated that risks can be managed well enough, and that the supervisor can respond to problems quickly enough, so that MFIs can then be allowed to lever- age as aggressively as commercial banks. Others argue that applying a higher capital-adequacy requirement to MFIs, or an equivalent risk-weighting requirement to microloan portfolios in diversified institutions, will tend to lower the return on equity in microlending, thus reduc- ing its attractiveness as a business and creating an uneven playing field. On the other hand, the demand for microfinance is less sensitive to interest rates than is the demand for normal bank loans, so that microlenders have more room to adjust their interest-rate spread to produce the return they need, as long as all microlenders are subject to the same rules and the government does not impose interest rate caps. Applying capital-adequacy norms to financial cooperatives presents a specific issue with respect to the definition of capital. All members of such cooperatives are required to invest a minimum amount of “share capital” in the institution. But unlike an equity investment in a bank, a member’s share capital can usually be withdrawn whenever the member decides to leave the cooperative. From the vantage of institutional safety, such capital is not very satisfactory: it is imperma- nent, and is most likely to be withdrawn at precisely the point where it would be most needed—when the cooperative gets into trouble. Capital built up from retained earnings, sometimes called “institutional capital,” is not subject to this problem. One approach to this issue is to limit members’ rights to withdraw share capital if the cooperative’s capital adequacy falls to a dangerous level. Another approach is to require cooperatives to build up a certain level of insti- tutional capital over a period of years, after which time capital adequacy is based solely on these retained earnings. Unsecured Lending Limits, and Loan-Loss Provisions In order to minimize risk, regulations often limit unsecured lending to some percentage—often 100 percent—of a bank’s equity base. Such a rule should not be applied to microcredit because it would make it impossible for an MFI to leverage its equity with deposits or borrowed money. Bank regulations sometimes require 100-percent loan-loss provisions for all unsecured loans at the time they are made, even before they become delinquent. However, this is unworkable when applied to microcredit portfolios. Even if the provision expense is later recovered when a loan is collected, the accumulated charge for current loans would produce a massive under-representation of the MFI’s real net worth. To meet these two problems, a common regulatory adjustment is to treat group guarantees as “collateral” for purposes of applying such regulations to microcredit. This can be a convenient solution to the problem if all microlenders use these guarantees. However, group guarantees are less effective than is often 22 Microfinance Consensus Guidelines supposed. Many MFIs do not enforce these guarantees, and there is no evidence that group-guaranteed microloans have higher repayment rates than nonguaran- teed individual microloans. The most powerful source of security in microcred- it tends to be the strength of an institution’s lending, tracking, and collection procedures, rather than the use of group guarantees. Whatever the rationale used to justify the adjustment, competent lenders should not be required to automatically provision large percentages of micro- credit loans as soon as they are made. But once such loans have fallen delinquent, the fact that they are unsecured justifies requiring them to be provisioned more aggressively than conventionally collateralized portfolio. This is especially true in countries where microlending tends to be short-term. After sixty days of delin- quency, a three-month unsecured microloan with weekly scheduled payments presents a higher likelihood of loss than does a two-year loan secured by real estate and payable monthly. In some countries, MFIs are effectively prevented from borrowing from banks because the MFIs cannot offer qualifying collateral, and without such col- lateral the bank would have to provision 100 percent of the loan. In such coun- tries, consideration should be given to adjusting the banking rules so that the loan portfolio of an MFI with a strong track record of collection can qualify as collateral for a bank loan. Loan Documentation Given the nature of microfinance loan sizes and customers, it would be excessive or impossible to require them to generate the same loan docu- mentation as commercial banks. This is particularly true, for instance, with col- lateral registration, financial statements of borrowers’ businesses, or evidence that those businesses are formally registered. These requirements must be waived for micro-sized loans. On the other hand, some microlending methodologies depend on the MFI’s assessment of each borrower’s repayment ability. In such cases, it is reasonable to require that the loan file contain simple documentation of that assessment of the client’s cash flow. However, when it makes repeated short-term (for instance, three-month) loans to the same customer, the MFI should not be required to repeat the cash-flow analysis for every single loan. Restrictions on Co-Signers as Borrowers Regulations sometimes prohibit a bank from lending to someone who has co- signed or otherwise guaranteed a loan from that same bank. This creates prob- lems for institutions using group-lending mechanisms that depend on all group members co-signing each others’ loans. Physical Security and Branching Requirements Banks’ hours of business, location of branches, and security requirements are often strictly regulated in ways that could impede service to a microfinance clien- tele. For instance, client convenience might require operations outside normal business hours, or cost considerations might require that staff rotate among Regulation and Supervision 23 branches that are open only one or two days a week. Security requirements such as guards or vaults, or other normal infrastructure rules, could make it too cost- ly to open branches in poor areas. Branching and physical security requirements merit reexamination—but not necessarily elimination—in the microfinance con- text. Clients’ need for access to financial services has to be balanced against the security risks inherent in holding cash. Frequency and Content of Reporting Banks may be required to report their financial position frequently—even daily. In many countries, the condition of transportation and communication can make this virtually impossible for rural banks or branches. More generally, reporting to a supervisor (or a credit information service) can add substantially to the admin- istrative costs of an intermediary, especially one that specializes in very small trans- actions. Reporting requirements should usually be simpler for microfinance institutions or programs than for normal commercial bank operations. Reserves against Deposits Many countries require banks to maintain reserves equal to a percentage of cer- tain types of deposits. These reserves may be a useful tool of monetary policy, but they amount to a tax on savings, and can squeeze out small depositors by raising the minimum deposit size that banks or MFIs can handle profitably. This latter drawback should be factored into decisions about reserve requirements. Ownership Suitability and Diversification Requirements15 The typical ownership and governance structure of MFIs tends to reflect their origins and initial sources of capital. NGOs, governmental aid agencies, multi- lateral donors, and other development-oriented investors predominate over those who have the profit-maximizing objectives of typical bank shareholders. The individuals responsible for these development-oriented investments are usu- ally not putting at risk money from their own private pockets. Investors of this kind, and their elected directors, may have weaker personal incentives to moni- tor the risk-taking behavior of MFI management closely. This does not imply that private, profit-maximizing owners of commercial banks always do a good job of supervising commercial bank management. But experience does indicate that such owners tend on the average to watch the management of their invest- ments more carefully than do the representatives of donors and social investors. Typical banking regulation mandates the nature of permissible share- holders, as well as the minimum number of founding shareholders and a maximum percentage of ownership for any shareholder. Both types of rules can pose obstacles for depository MFIs, given their ownership and gover- nance attributes. These rules serve legitimate prudential objectives. Mandates as to the nature of permissible shareholders aim to assure that the owners of a depository finan- cial institution will have both the financial capacity and the direct interest to put 24 Microfinance Consensus Guidelines in additional funds if there is a capital call. Ownership diversification require- ments aim at preventing “capture” of bank licenses by single owners or groups, and building checks and balances into governance. But together these require- ments can cause serious problems in the common case, where the assets of the new licensed MFI come almost entirely from the NGO that has been conduct- ing the microfinance business until the creation of the new institution. First, laws or regulations sometimes prohibit an NGO from owning shares in the licensed institution. While such a prohibition may serve a legitimate purpose, it generally poses too heavy an obstacle to the eventual licensing of microfinance that originated in an NGO: consideration should be given to amending it. Even if the NGO is permitted to own shares of the new institution, diversification requirements may pose an additional challenge. For instance, a five-owner min- imum and a 20-percent maximum per shareholder would force the transforming NGO to seek out four other owners whose combined capital contribution would be four times as much as the NGO is contributing. This can be an impractical burden for a socially-oriented business whose profitability is not yet strong enough to attract purely commercial equity. The only alternative has sometimes been to distribute shares to other owners who have not paid in an equivalent amount of equity capital. This arrangement does not tend to produce good over- sight by the other owners. Given the legitimate objectives of shareholder suitability and ownership diver- sification requirements, there is no easy or universal prescription for how to modify these types of rules to accommodate MFIs. However, the solution may in some instances be as simple as permitting the licensing agency the discretion to consider the particular situation of microfinance applicants and their proposed backers, and waive shareholder suitability and diversification requirements on a case-by-case basis. Who Should These Special Standards Apply To? It is worth reiterating that most of the adjustments mentioned in this section should ideally apply, not only to specialized MFIs, but also to microfinance operations in commercial banks or finance companies. Some of them are also relevant to unsecured lending by financial cooperatives. Even if a country’s commercial banks have no interest in microfinance at pres- ent, those attitudes can change once specialized MFIs credibly demonstrate the profit potential of their business. If a full-service bank decides to offer microfi- nance products, or to partner with an MFI to offer those products, it should have a clear regulatory path to do so; otherwise, continued fragmentation of the financial sector is guaranteed. Regulators and supervisors should have a special incentive to encourage such developments: when microcredit is a small part of a diversified commercial-bank portfolio, the risk and cost of supervising the micro- finance activity become much lower. Moreover, a level playing field in terms of the prudential standards applied to an activity helps to stimulate competition. Regulation and Supervision 25 Deposit Insurance In order to protect smaller depositors and reduce the likelihood of runs on banks, many countries provide explicit insurance of bank deposits up to some size limit. Some other countries provide de facto reimbursement of bank depos- itors’ losses even in the absence of an explicit legal commitment to do so. There is considerable debate about whether public deposit insurance is effective in improving bank stability, whether it encourages inappropriate risk-taking on the part of bank managers, and whether such insurance would be better provided through private markets. In any event, if deposits in commercial banks are insured, the presumption ought to be that deposits in other institutions pru- dentially licensed by the financial authorities should also be insured, absent com- pelling reasons to the contrary. IV FACING THE SUPERVISORY CHALLENGE Decades of experience around the world with many forms of “alternative” finan- cial institutions—including various forms of financial cooperatives, mutual soci- eties, rural banks, village banks, and now MFIs—demonstrate that there is a strong and nearly universal temptation to underestimate the challenge of supervising such institutions in a way that will keep them reasonably safe and stable. When the various stakeholders start discussing legal frameworks for microfinance in a country, it is relatively easy and interesting to craft regulations, but harder and less attractive to do concrete practical planning for effective supervision. The result is that supervision sometimes gets little attention in the process of regulatory reform, often on the assumption that whatever superviso- ry challenges are created by the new regulation can be addressed later, by pump- ing extra money and technical assistance into the supervisory agency for a while. This assumption can be wrong in many cases. The result may be regulation that is not enforced, which can be worse than no regulation at all. Microfinance as an industry can never reach its full potential until it is able to move into the sphere of prudentially regulated institutions, where it will have to be prudentially supervised.16 While prudential regulation and supervision is inevitable for microfinance, there are choices to be made and balances to be drawn in deciding when, and how, this development takes place. Those balances are likely to be drawn in the right place only if supervisory capability, costs, and consequences are examined earlier and more carefully than is some- times the case in present regulatory discussions. The crucial importance of early and realistic attention to supervision issues stems from the fiduciary responsibility the government assumes when it grants financial licenses. Citizens should be able to assume, and usually do assume, that the issuance of a prudential license to a financial intermediary means that the government will effectively supervise the intermediary to protect their deposits. Thus, licenses are promises. Before deciding to issue them, a government 26 Microfinance Consensus Guidelines needs to be clear about the nature of the promises and about its ability to fulfill them. Supervisory Tools and Their Limitations Portfolio supervision tools. Some standard tools for examining banks’ portfolios are ineffective for microcredit. As noted earlier, loan-file documentation is a weak indicator of microcredit risk. Likewise, sending out confirmation letters to verify account balances is usually impractical, especially where client literacy is low. Instead, the examiner must rely more on an analysis of the institution’s lend- ing systems and their historical performance. Analysis of these systems requires knowledge of microfinance methods and operations, and drawing practical con- clusions from such analysis calls for experienced interpretation and judgment. Supervisory staff are unlikely to monitor MFIs effectively unless they are trained and to some extent specialized. Capital calls. When an MFI gets in trouble and the supervisor issues a capital call, many MFI owners are not well-positioned to respond to it. NGO owners may not have enough liquid capital available. Donors and development-orient- ed investors usually have plenty of money, but their internal procedures for dis- bursing it often take so long that a timely response to a capital call is impractical. Thus, when a problem surfaces in a supervised MFI, the supervisor may not be able to get it solved by the injection of new capital. Stop-lending orders. Another common tool that supervisors use to deal with a bank in trouble is the stop-lending order, which prevents the bank from taking on further credit risk until its problems have been sorted out. A commercial bank’s loans are usually collateralized, and most of the bank’s customers do not necessarily expect an automatic follow-on loan when they pay off their existing loan. Therefore, a commercial bank may be able to stop new lending for a peri- od without destroying its ability to collect its existing loans. The same is not true of most MFIs. Immediate follow-on loans are the norm for most microcredit. If an MFI stops issuing repeat loans for very long, customers lose their primary incentive to repay, which is their confidence that they will have timely access to future loans when they need them. When an MFI stops new lending, many of its existing borrowers will usually stop repaying. This makes the stop-lending order a weapon too powerful to use, at least if there is any hope of salvaging the MFI’s portfolio. Asset sales or mergers. A typical MFI’s close relationship with its clients may mean that loan assets have little value in the hands of a different management team. Therefore, a supervisor’s option of encouraging the transfer of loan assets to a stronger institution may not be as effective as in the case of collateralized com- mercial bank loans. The fact that some key supervisory tools do not work very well for microfi- nance certainly does not mean that MFIs cannot be supervised. However, regu- Regulation and Supervision 27 lators should weigh this fact carefully when they decide how many new licenses to issue, and how conservative to be in setting capital-adequacy standards or required levels of past performance for transforming MFIs. Costs of Supervision Promoters of new regulatory windows for MFIs are rightly enthusiastic about the possibility of bringing financial services to people who have never had access to them before. Supervisors, on the other hand, tend to concentrate more on the costs of supervising new, small entities. Good microfinance regulation needs to balance both factors. In relation to the assets being supervised, specialized MFIs are much more expensive to supervise than full-service banks. One supervisory agency with sev- eral years of experience found that supervising MFIs cost it 2 percent per year of the assets of those institutions—about 30 times as expensive as its supervision of commercial bank assets. Donors who promote the development of deposito- ry microfinance should also consider providing transitional subsidy for supervising the resulting institutions—particularly in the early stages when the supervisory staff is learning about microfinance and there are a small num- ber of institutions to share the costs of supervision. However, in the long term, the government must decide whether it will subsidize these costs or make MFIs pass them on to their customers. Even if a donor pays the additional cash costs of MFI supervision, there is a further cost in terms of the time and attention of the managers of the supervi- sory agency. In some developing and transitional economies, the national econ- omy is at serious risk because of systemic problems with the country’s commer- cial banks. In such settings, serious consideration should be given to the cost of diverting too much of agency management’s attention away from their primary task, by requiring them to spend time on small MFIs that pose no threat to the country’s financial systems. The administrative costs within the supervised MFI are also substantial. It would not be unusual for compliance to cost an MFI 5 percent of assets in the first year or two and 1 percent or more thereafter. Where to Locate Microfinance Supervision? Given the problem of budgeting scarce supervisory resources, alternatives to the conventional supervisory mechanisms used for commercial banks are frequently proposed for depository MFIs. Within the Existing Supervisory Authority? The most appropriate supervisory body for depository microfinance is usually (though not always) the supervisory authority responsible for commercial banks. Using this agency to supervise microfinance takes advantage of existing skills and lowers the incentive for regulatory arbitrage. The next question is whether to create a separate department of that agency. The answer will vary from country 28 Microfinance Consensus Guidelines to country, but at a minimum, specially trained supervisory staff is needed, given the differing risk characteristics and supervisory techniques in the case of MFIs and microfinance portfolios. The question whether to house microfinance regulation within the existing supervisory authority becomes more complicated when both non-depository microlending institutions and depository MFIs are to be addressed within a sin- gle, comprehensive regulatory scheme. The tasks involved in issuing permits to non-depository microlending institutions have relatively little to do with the prudential regulation and supervision of depository institutions. In some con- texts, lodging both of these disparate functions within the same regulatory body might be justified on pragmatic grounds—such as the absence of any other appropriate body, or the likelihood that the permit-issuing function would be more susceptible to political manipulation and abuse if carried out by another body. In other cases, non-depository MFIs are required to report to the bank- ing supervisor in order to make it easier for them to move eventually into more services and more demanding prudential regulation. Often, however, the risks of consolidating prudential and non-prudential regulation of microfinance within the supervisory body responsible for banks will outweigh the benefits. These risks include the possibility of confusion on the part of supervisors as to the appropriate treatment of non-depository institutions, and the possibility that the public will see the supervisory authority as vouching for the financial health of the non-depository institutions, even though it is not (and should not be) mon- itoring the health of these institutions closely. “Self-Regulation” and Supervision Sometimes regulators decide that it is not cost-effective for the government financial supervisor to provide direct oversight of large numbers of MFIs. Self- regulation is sometimes suggested as an alternative. Discussion of self-regulation tends to be confused because people use the term to mean different things. In this paper, “self-regulation” means regulation (and/or supervision) by any body that is effectively controlled by the regulated entities, and thus not effectively controlled by the government supervisor. This is one point on which historical evidence seems clear. Self-regulation of financial intermediaries in developing countries has been tried many times, and has virtually never been effective in protecting the soundness of the regulated organizations. One cannot assert that effective self-regulation in these settings is impossible in principle, but it can be asserted that such self-reg- ulation is almost always an unwise gamble, against very long odds, at least if it is important that the regulation and supervision actually enforce financial discipline and conservative risk management. Sometimes regulators have required certain small intermediaries to be self- regulated, not because they expect the regulation and supervision to be effective, but because this is politically more palatable than saying that these depositor-tak- ers will be unsupervised. This can be a sensible accommodation in some settings. While self-regulation probably will not keep financial intermediaries healthy, it Regulation and Supervision 29 may have some benefits in getting institutions to begin a reporting process or in articulating basic standards of good practice. Delegated Supervision “Delegated supervision” refers to an arrangement where the government finan- cial supervisor delegates direct supervision to an outside body, while monitoring and controlling that body’s work. This seems to have worked, for a time at least, in some cases where the government financial supervisor closely monitored the quality of the delegated supervisor’s work, although it is not clear that this model reduces total supervision costs. Where this model is being considered, it is important to have clear answers to three questions. (1) Who will pay the sub- stantial costs of the delegated supervision and the government supervisor’s over- sight of it? (2) If the delegated supervisor proves unreliable and its delegated authority must be withdrawn, is there a realistic fallback option available to the government supervisor? (3) When a supervised institution fails, which body will have the authority and ability to clean up the situation by intervention, liquida- tion, or merger? Because many MFIs are relatively small, there is a temptation to think that their supervision can be safely delegated to external audit firms. Unfortunately, experience has been that external audits of MFIs, even by internationally-affiliat- ed audit firms, very seldom include testing that is adequate to provide a reason- able assurance as to the soundness of the MFI’s loan assets, which is by far the largest risk area for microlenders. If reliance is to be placed on auditors, the supervisor must require microfinance-specific audit protocols that are more effective, and more expensive, than the ones now in general use, and must reg- ularly test the auditors’ work. V KEY POLICY RECOMMENDATIONS Discussion of microfinance regulation and supervision is necessarily complex and filled with qualifications and caveats. For the sake of clarity and emphasis, this paper concludes with a brief reiteration of some of its more important recom- mendations. ➨ Powerful new “microfinance” techniques are being developed that allow for- mal financial services to be delivered to low-income clients who have previ- ously not had access to such services. In order to reach its full potential, the microfinance industry must eventually be able to enter the arena of licensed, prudentially supervised financial intermediation, and regulations must even- tually be crafted that allow this development. ➨ Problems that do not require the government to oversee and attest to the financial soundness of regulated institutions should not be dealt with through prudential regulation. Relevant forms of non-prudential regulation, including 30 Microfinance Consensus Guidelines regulation under the commercial or criminal codes, tend to be easier to enforce and less costly than prudential regulation. ➨ Proponents of microfinance regulation need to careful about steps that might bring the topic of microcredit interest rates into public and political discus- sion. Microcredit needs high interest rates. In many countries, it may be impossible to get explicit political acceptance of a rate that is high enough to allow viable microfinance. In other contexts, concerted education of relevant policymakers may succeed in establishing the necessary political acceptance. ➨ Credit-reference services can lower lenders’ costs and expand the supply of credit for lower-income borrowers. However, they are not technically feasible in all countries. ➨ A microlending institution should not receive a license to take deposits until it has demonstrated that it can manage its lending profitably enough so that it can cover all its costs, including the additional financial and administrative costs of mobilizing the deposits it proposes to capture. ➨ Before regulators decide on the timing and design of prudential regulation, they should obtain a competent financial and institutional analysis of the lead- ing MFIs, at least if existing MFIs are the main candidates for a new licensing window being considered. ➨ Prudential regulation should not be imposed on “credit-only” MFIs that merely lend out their own capital, or whose only borrowing is from foreign commercial or non-commercial sources or from prudentially regulated local commercial banks. ➨ Depending on practical costs and benefits, prudential regulation may not be necessary for MFIs taking cash collateral (compulsory savings) only, especial- ly if the MFI is not lending out these funds. ➨ As much as possible, prudential regulation should be focused on the type of transaction being conducted rather than the type of institution conducting it. ➨ Where possible, regulatory reform should include adjusting any regulations that would preclude existing financial institutions (banks, finance companies, etc.) from offering microfinance services, or that would make it unreasonably difficult for such institutions to lend to MFIs. ➨ Where cost-effective prudential supervision is impractical, consideration should be given to allowing very small community-based intermediaries to continue taking deposits from members without being prudentially super- vised, especially in cases where most members do not have access to safer deposit vehicles. ➨ Minimum capital needs to be set high enough so that the supervisory author- ity is not overwhelmed by more new institutions than it can supervise effec- tively. Regulation and Supervision 31 ➨ Most microlending is for all practical purposes unsecured. Limits on unse- cured lending, or high provisioning of unsecured portfolio that has not fallen delinquent, are not practical for MFIs. Instead, risk control needs to be based the MFI’s historical collection performance, and analysis of its lending sys- tems and practices. ➨ Loan documentation and reporting requirements need to be simpler for microfinance institutions and operations than for normal commercial bank operations. ➨ Limitations on foreign ownership or maximum shareholder percentages may be inappropriate, or need flexible application, if local microfinance is at a stage where much of the investment will have to come from transforming NGOs and other socially-motivated investors. ➨ Designers of new regulation for microfinance need to pay much more atten- tion to issues of likely effectiveness and cost of supervision than is usually done. Financial intermediation licenses are promises. Before issuing them, a government needs to be clear about the nature of the promises and its prac- tical ability to honor them. ➨ Design of microfinance regulation should not proceed very far without esti- mating supervision costs realistically and identifying a sustainable mechanism to pay for them. Donors who encourage governments to take on supervision of new types of institution should be willing to help finance the start-up costs of such supervision. ➨ Supervision of microfinance—especially portfolio testing—requires some techniques and skills that are different from those used to supervise commer- cial banks. Supervisory staff will need to be trained and to some extent spe- cialized in order to deal effectively with MFIs. ➨ Financial cooperatives—at least large ones—should be prudentially supervised by a specialized financial authority, rather than by an agency that is responsi- ble for all cooperatives. ➨ In developing countries, “self-supervision” by an entity under the control of those supervised is extremely unlikely to be effective in protecting the sound- ness of the supervised financial institutions. ➨ External auditors cannot reliably appraise the financial condition of MFIs unless they test portfolio with microfinance-specific procedures that go well beyond normal present practice. 32 Microfinance Consensus Guidelines NOTES 1. CGAP, the Consultative Group to Assist the Poor, is made up of 29 international donor agencies that support microfinance. A version of this document was approved and adopted by the Consultative Group members in September 2002. 2. Average microcredit loan balances tend be below per-capita national income. 3. This paper does not embrace insurance and leasing, even though these financial serv- ices have important potential for lower-income people. For these services, there is almost no experience yet with specialized regulation focused on the needs of poorer clients. 4. The term “non-prudential regulation” poses some problems. The distinction between prudential and non-prudential regulation is not always crisp—sometimes a rule serves both prudential and non-prudential objectives. For example, regulation aimed at prevention of financial crimes (see the discussion on fraud and financial crimes, on page 12), also contributes to prudential objectives. Moreover, defining non-prudential regulation simply by reference to what it is not leaves open the ques- tion of the scope of the concept. The term “conduct of business” regulation is some- times used to denote the non-prudential rules applicable to financial institutions. However, this term is also problematic since prudential regulation also affects the conduct of a financial institution’s business. 5. This is not to say, of course, that the failure of a lending-only MFI has no adverse consequences. If customers lose access to loans from an MFI, it may become a severe problem particularly if the failing microlender is the only available source of much- needed capital. However, the same is true of any other important supplier. The fact that a good or service is important to customers has not been held to justify pru- dential regulation of the supplier’s business. 6. The point here is not that cost of funds and loan losses is always the same for MFIs and commercial banks, but rather that both types of lenders face higher administra- tive costs per dollar lent when they engage in microlending. 7. Sometimes, of course, high interest rates do reflect inefficiency (excessive adminis- trative costs) on the part of MFIs. However, competition has proved to solve this problem better than interest-rate caps. 8. In the case of prudentially regulated financial institutions, these types of restrictions are sometimes exacerbated by other, prudentially motivated, limitations on owner- ship. (See the discussion of ownership suitability and diversification requirements on page 24.) 9. The discussion of MFI transformation in this section has dealt with non-prudential issues only. Such transfers are also affected by prudential rules, especially the rules about suitability and diversification of owners, discussed on page 24. 10. A useful analysis of microfinance regulation along these lines can be found in Hennie van Greuning, Joselito Gallardo, and Bikki Randhawa, “A Framework for Regulating Microfinance Institutions,” Policy Research Working Paper 206 (Washington, D.C.: World Bank, 1999). 11. In some countries, banks have to provision 100% of unsecured loans, except for loans to other licensed intermediaries. In such a context, banks may be more willing to lend to MFIs who are themselves licensed, and thus licensed MFIs might be able to borrow at a low interbank rate. These are reasons why an MFI borrowing from com- mercial banks might want to be prudentially licensed, but they do not justify impo- sition of a licensing requirement. Regulation and Supervision 33 12. Rationing licenses is not cost-free, because barriers to entry hinder competition. 13. Regulators sometimes hope that, instead of shutting down, these small intermedi- aries will merge to form larger ones that can be supervised more easily. But the prac- tical economics of branch operation can often make this impossible. 14. Capital adequacy has to do with maintaining a prudent relationship between an insti- tution’s risk assets and its “cushion” of owner’s funds. This relationship is affected, not only by the equity/assets ratio, but also by the rules for risk-weighting and pro- visioning. Capital adequacy in the broad sense is managed using a combination of these measures. 15. This section discusses prudentially-driven ownership requirements. These tend to overlap with non-prudential ownership requirements discussed on page 14. 16. This statement does not imply that prudential regulation will eventually embrace all institutions providing microfinance services. 34 Microfinance Consensus Guidelines