How should governments regulate
microfinance?
Richard Rosenberg1
INTRODUCTION
Powerful new microfinance techniques are being developed that allow formal financial services to be delivered to low-income clients who have long had no access to such services. But the microfinance industry will not reach its full potential unless many of its service providers can eventually enter the arena of licensed, prudentially supervised financial intermediaries. Regulations must eventually be crafted that allow this to happen.
Dozens of developing and transition country governments are now at earlier or later stages of addressing this challenge.
Many different actors are pushing for regulatory adjustments, from microfinance institutions themselves (MFIs), to international development agencies, to government officials who want to democratize finance or protect against perceived risks for the financial system (or perhaps clamp down on annoying non-governmental organizations—NGOs). The interests and objectives of these actors diverge considerably. Thorny technical and practical issues are involved. We do not have decades of experience with regulated microfinance to guide us—most of the countries with microfinance regulations have only a few years of experience with implementing them. And country-specific circumstances loom large, so there can be no standard model for microfinance regulation.
Nevertheless, among people working on these topics there are surprisingly wide areas of consensus on some general principles that should bear on regulatory design for microfinance. The author is confident that most of the material in this chapter is consistent with the views of most of the technical advisors who have multi-country experience and who do not represent the interests of any particular combatant in the
fray.
The discussion will begin with an important definitional distinction between “prudential” and “non-prudential” regulation. Non-prudential regulation will be treated in the third section, prudential regulation in the fourth, and the challenges of prudential supervision in the fifth. The sixth section concludes.
PREAMBLE: PRUDENTIAL AND NON-PRUDENTIAL REGULATION
Governments regulate the behavior of all businesses. Such regulation may be aimed at protecting consumers, or employee safety, or the environment, but it usually does not try to protect the financial health of the business that concern is generally left to the owners, at least where the owners are private. But in almost every country in the world, banks are treated differently. Governments impose an elaborate regime of “prudential” regu - lation whose aim is to protect the solvency of banks. Various reasons are advanced for this. The main one is that financial systems depend critically on confidence, so that the failure of one bank can hurt many other banks and provoke a systemic crisis, with dire effects for the economy at large. Another reason is that banks are financed predominantly by money of people other than the shareholders, which creates incentives for bank managers to take imprudent risks: when the gamble succeeds the bank and its shareholders capture the gain, but when the gamble fails, others especially depositors—may bear a large part of the loss. Finally, governments believe that small, unsophisticated depositors need protection because they are in no position to appraise the riskiness of a bank on their own. Examples of prudential requirements include capital adequacy rules (how much of other people’s money a bank can use), restrictions on risky uncollateralized lending, limits on insider lending, or requiring main tenance of reserves for loans that are likely to go bad.
“Non-prudential” regulation is an inelegant name for all the other banking rules the ones that don’t involve the government in assessing and protecting the financial health of banks. Such rules are sometimes referred to as “conduct of business” regulation. Examples include limits on interest rates charged to borrowers, other consumer protection like truth-in-lending laws, or anti-money-laundering rules that require screening and reporting of customers. Securities regulations are another non-prudential example: these rules usually require banks, just like other firms, to disclose all material information about their business to potential investors. The rules have been complied with when all the information about the bank’s business is disclosed. Investors are left to fend for themselves when it comes to weighing their risks.
Implementing prudential regulation, where the government in some sense is vouching for the financial soundness of each licensed bank taking deposits, tends to be much more complex, difficult, expensive, and intrusive than implementing non-prudential regulation. Enforcing prudential regulation always requires a specialized banking authority, whereas many nonprudential regulations apply to non-deposit-taking firms as well, and might not necessarily require a specialized banking supervisor to enforce them.
The reason for emphasizing this distinction is that in a number of countries, governments are applying burdensome prudential rules to nondeposit taking MFIs whose failure would cause neither loss of depositors’ funds nor material disruption of the national financial system. Prudential regulation has high costs not only for the supervisory authority but also for the supervised institution, which will eventually pass these costs along to its customers. Prudential regulation of credit-only MFIs uses a cannon a very expensive cannon—where a rifle would be more than adequate in view
of the risks involved.
It is especially important to focus on the implications of regulation for the administrative costs of MFIs. In the centuries-old effort to improve financial access for poor and low-income people, the critical factor is cost, more than the motivations of financial service providers. Major, long-lasting improvements in access are usually associated with new ways to lower costs. Concern for the poor has played an important part in the microfinance revolution of the last three decades. But concern for the poor has been around for a long time. The revolution became possible when Grameen Bank and other pion -eers in Indonesia and Latin America discovered less costly ways to deliver and collect tiny uncollateralized loans, and mobilize and manage small savings. Some of the regulatory requirements discussed in this chapter have significant cost implications for microfinance providers. Decisions about such practices need to be handled carefully.
NON-PRUDENTIAL ISSUES
Usury Limits
Lending a million (US) dollars in 10 000 loans of $100 each entails administrative costs that are hugely greater than the cost of lending out the same amount in one or two big loans. As a result, it is usually impossible to do micro-lending on a financially sustainable basis without charging interest rates that are very substantially higher than what banks charge to larger borrowers. In 2005, the median annual interest rate collected by the hundreds of MFIs reporting to the MIX Market (www.themix.org) was about 31 percent. Rates above 50 percent are not uncommon, and a few MFIs charge more than 80 or even 100 percent. In most cases, these rates reflect not high profits but high costs of micro-lending: the smaller the loan size, the higher the administrative costs are for lending a given amount. But this analysis of lending costs is fine print that is usually too small to show up on the screens of politicians or the general public. Charging poor borrowers 30 percent when fat cats pay only 10 or 15 percent shocks most consciences.
Not all microfinance interest rates can be explained by the costs of lending. In a recent well-publicized instance, the Mexican MFI Compartamos was charging interest of about 100 percent a year and producing annual profits that gave its shareholdersmore than a 50 percent return on equity. This cause célèbre, despite being a highly exceptional case, has fanned the flames of a growing backlash against high micro-credit interest rates, a backlash that has already been underway in Latin America and the rest of the world for several years now.
Limits on interest rates can hurt rather than help low-income borrowers if the interest cap is set too low for certain types of lending to be profitable: providers will withdraw from the business and potential borrowers will lose access to services. In theory, an interest rate cap would avoid this result if it were set at just the right level. As a practical matter, however, finding the right level is hard, not least of all because loan products, clienteles, and costs vary. Moreover, it is politically difficult for governments to set interest caps at appropriate levels: a reasonable interest rate for tiny, high-cost micro-loans will inevitably seem exploitative to most people, because they do not understand the reasons for the high rates.
Consumer Protection and Borrowers’ Rights
When governments are concerned about microfinance interest rates that sound abusive, they are sometimes advised to avoid interest rate caps and focus instead on other borrower protection issues such as truth-in-lending (disclosure of the full cost of loans in a format that makes it easy to compare rates offered by various lenders) or prohibition of certain unacceptable lending and collection practices.
Most MFIs today do not quote their loan charges in the form of an annualized effective interest rate that includes all costs. In some cases there may be a legitimate concern that explicit quotation of rates this way would lead to a political backlash, resulting in interest rate caps that would make it impossible to continue serving their clients. In most cases, though, microlenders’ opposition to truth-in-lending policies and requirements probably stems mainly from normal, less noble motives. Loan-cost disclosure may not be a panacea, however: there are some indications that low-income
clients have trouble understanding and using the information. This concern has led to scattered efforts to provide financial education for consumers, but most of these programs are still too young to be assessed, at least in developing and transition countries.
Consumer protection regimes may also include restrictions on the way loans can be made and collected. Obvious examples would include prohibiting the use of violence or other heavy intimidation to collect loans, but other less dramatic practices may also be deemed abusive. In South Africa, for instance, so-called “micro-credit” consists mainly of firms making high-interest consumer loans to a clientele consisting mostly of salaried formal-sector employees.3 Taking possession of a borrower’s ATM card or requiring delivery of a post-dated check for the loan amount were common practices, which were prohibited under a non-prudential regulatory regime created by the Micro Finance Regulatory Council (MFRC), a government agency lodged outside of the banking authority. MFRC rules were given teeth by stipulating that any loans issued in violation of those rules would be legally unenforceable.
The Bolivian microfinance sector suffered huge losses when profligate Chilean-backed consumer lenders started marketing to unsalaried microentrepreneurs, passing out loans that bore no relation to the borrowers’ repayment capacity. Many borrowers got in over their heads, and since a large percentage of these were also borrowing from more responsible MFIs, those sound MFIs were badly hurt by the ensuing wave of defaults, not to mention borrowers who lost their credit rating. The Bolivian Superintendency of Banks responded by requiring all uncollaterized lending to include an assessment of repayment capacity.
A the time of this Bolivian crisis, the government’s credit reference bureau was not working well, so it was hard for lenders to know whether a potential borrower had loans outstanding from another source, or had a history of repayment problems. After the crisis, all of the actors found themselves considerably more enthusiastic about the credit bureau, and unlicensed lending-only MFIs were allowed to participate for the first time. As a general matter, credit reference bureaus are a powerful tool for extending credit access to previously excluded groups, because the bureaus significantly lower the costs of appraising borrower creditworthiness, and strengthen borrowers’ motivation to repay. Credit bureaus make it possible to lend to customers who would have been unprofitable otherwise.
Other consumer protection measures include privacy protection and accessible dispute-resolution systems.
AML/CFT Regulation4
The Financial Action Task Force (FATF) is an international body that recommends standards for national legislation on anti-money-laundering and countering the financing of terrorism (AML/CFT). The FATF standards do allow room for adjustment to fit individual country circumstances, but developing and transition countries are often nervous of straying far from the standards, because winding up on the list of non-complying nations can have severe consequences. Among the FATF standards are know-yourclient rules (ascertaining and documenting the customers’ true identities and addresses), heightened surveillance of transactions, preserving transaction records, and reporting suspicious transactions to national authorities. Many banks complain loudly about the additional costs generated by these requirements when dealing with their normal customers. In the world of microfinance, where transactions and balances are much tinier, full enforcement of regular AML/CFT standards would make it uneconomic to serve large groups of customers. The additional administrative costs are particularly problematic when transactions are so small, and compliance can be impractical for some kinds of clients. For instance, it is challenging to document identity and address for people who have no national identity card, are illiterate, and have never seen any document that specifies where they live.
In its early years FATF was dominated by people coming from a lawenforcement perspective, who were not always instinctively sympathetic with concerns about how FATF rules might exclude low-income clients from services. More recently, this problem is getting more attention at international and local levels. Taking a risk-based approach to AML/CFT, it would not seem that transactions of, say, $30 or loan or savings accounts of $300 create substantial security risks. Governments should consider softer requirements, or waiving them altogether, for accounts and balances below defined limits. After this was done in South Africa, for instance, banks were able to offer basic, no-frills transaction accounts that in a few short years reached 1.6 million customers, most of whom would have remained unbanked if the AML/CFT rules had not been relaxed.
PRUDENTIAL ISSUES
Whether, When, and How to Open Prudential Licensing Regimes for
Microfinance
In more than a few countries, the microfinance sector consists mainly of weak non-governmental organizations that provide lending only, as well as credit unions or similar savings and loan cooperatives, few of which are large and stable. When a government is confronted with this situation, and wants to catalyze a large expansion of quality financial services for its lower-income population, one plausible-sounding response is to develop a
new licensing window that allows institutions to become prudentially regu - lated, and take deposits, without facing minimum capital requirements as high as those required for a full banking license. (Deposit-taking is doubly important: not only does it provide a large funding source for expansion of lending, but it also gives the institution’s low-income clients access to a savings service that is often even more valuable to them than credit.)
In such countries, this “build it and they will come” approach is based on the hope that the special licensing window will attract new private sector entrants to the business, or encourage weak existing MFIs to tighten up their operations so as to meet the prudential standards for licensing. There is considerable controversy over this approach. It is premised on the belief that the binding constraint is absence of appropriate regulation, rather than scarcity of competent retail operators. International experience to date has been too limited to produce a general answer to the question; if anything, it suggests that the answer will vary from country to country.
For instance, Tanzania spent a great deal of time, effort, and money on the development of a well-conceived licensing regime, but years afterward the results were disappointing. South Africa does not offer a micro-banking license, but the government took steps a decade ago to make it possible for microfinance institutions to offer small loans at relatively high interest rates. Despite this change, South Africa still does not have many solid institutions offering uncollateralized loans to unsalaried micro-entrepreneurs.
Pakistan has a huge unserved microfinance market, and it’s hard to find many countries with as good a licensing regime for microfinance as the one Pakistan enacted in 2001. But until very recently, institutions licensed under this law contributed hardly at all to the growth of microfinance in the country, and the overall financial condition of licensed and unlicensed providers was worse than it had been when the law was passed. Within the last year, however, things are looking brighter: some of the newly licensed, privately owned microfinance banks are expanding aggressively, and the overall financial performance of the sector is improving.
Regulation that follows, rather than leads, the market
In most of the countries with effective prudential regimes in place today for microfinance, the regulation came after, not before, the development of a critical mass of strong MFIs that were delivering loans on a sustainable basis. Bolivia has the longest and most solid record of successful microfinance regulation, and this experience is often held up as an example where a new licensing window for microfinance was a powerful contributor to the success of the industry. But the Bolivian licensing regime was put in place only after the country already had a number of strong NGO MFIs who had shown they could manage their lending business stably and profitably. BancoSol, the leading MFI, did not use the microfinance licensing window it got a regular commercial banking license well before the specialized microfinance law was passed. Most of the other MFIs who got licenses under the new law could probably have raised the money for a banking license if the easier and cheaper MFI license had not been available. In BancoSol’s first year or two, there was a certain amount of “supervision by winking,” as the Superintendent waived application of some prudential rules that didn’t fit microfinance very well. But the time the microfinance law was passed and new prudential norms formalized, the Superintendency already had experience from supervising BancoSol. It’s possible to argue that Bolivian microfinance did not need a new specialized microfinance window to reach its present vital state, and that a few adjustments to the country’s banking law and regulations would have created the necessary regulatory space.
When countries design a new licensing window for microfinance on the expectation that licenses will go mainly to existing NGO MFIs during the early years, the regulators sometimes don’t pay enough attention to the condition of those MFIs and their loan assets. In Zambia, for instance, the foreign aid agencies of the United States and Sweden financed development of a prudential regime in 1999 that would allow MFIs to take deposits. But at that time, sources say, the country had few if any MFIs whose cost recovery and loan collection would make them safe custodians of customers’ deposits. There may have been some expectation that donor-funded technical assistance would turn the MFIs into strong institutions, but it is hard to find many examples of weakly managed MFIs that have been turned into vibrant, stable MFIs by outside technical assistance. This is not to suggest that such assistance is useless: MFIs that already have strong managers make good use of such support, but technical assistance can seldom turn a
weak manager into a strong one. Political considerations prevented enactment of the Zambian law at the time. A set of microfinance regulations was finally put into force in 2006, but a review of the MIX Market database as of the beginning of 2007 shows only a single sustainable Zambian MFI, and that one had only 12 500 clients.
Adjusting Prudential Norms to Fit Microfinance Products and Institutions
Some regulations common in traditional banking need to be altered toaccommodate microfinance. Whether microfinance is being developed through specialized stand-alone deposit-taking MFIs, or as a product line within retail banks or finance companies, the following norms will usually need re-examination:
Minimum capital
The kind of investors who are willing and able to finance MFIs may not be able to come up with the minimum capital required for a full banking license, especially as minimum capital requirements trend upward around the world. Setting a low minimum capital bar is often the central objective of those pushing for new licensing regimes for microfinance.
When banking authorities set minimum capital, bank safety and soundness may not be their primary concern. Rather, minimum capital is often used as a rationing device to manage the number of separate institutions that have to be supervised. The arguments for and against low minimum capital for MFIs will be treated in the next section, which deals with supervisory challenges.
Capital adequacy
Under the Basel Accords, the relationship between shareholders’ equity and bank risk assets is the foundation of prudential regulation. Equity is treated as a cushion that protects depositors and other creditors of the bank: the more of its assets are funded by shareholders’ money, the higher the losses the bank can sustain and still be able to repay its depositors.
There has been controversy about whether solvency (capital adequacy) requirements should be tighter for specialized MFIs than for banks. If we want a level playing field in the financial sector, should microfinance be penalized with tougher solvency requirements that lower shareholder profitability?
Several theoretical arguments point in the direction of higher equity-torisk- assets ratios for MFIs. In the first place, deposit-taking microfinance is a new business in most countries, which supervisors—and some MFI managers as well—do not have decades of experience with. Second, most MFI
loan assets are not collateralized. Normally, MFI portfolio quality is very good, but if an MFI starts to have problems with loan delinquency, they can balloon out of control much faster than would be normal with collateralized loans. Third, administrative costs for MFIs are much higher than for commercial banks. When a significant part of the MFI’s loans are not being paid, the uncompensated administrative cost on those loans decapitalizes the MFI faster than would be the case with a normal bank. All of
these considerations suggest tighter solvency requirements, at least in the
early years.
Balanced against those theoretical arguments is the clear empirical fact that licensed MFIs suffer fewer loan losses than commercial banks do. There is also emerging evidence that licensed MFIs are more resilient than commercial banks in times of financial or economic emergencies. In a recent banking crisis in Bolivia, all the commercial banks went insolvent and MFIs came through in good shape. During the financial meltdown in Indonesia, repayment plummeted on the commercial loans of Bank Rakyat Indonesia (BRI), while there was hardly a blip in the repayment of its micro-loans. When times are uncertain, low-income people are especially anxious to maintain their access to a credit facility, which can be a life-saver if an unexpected shock hits. BRI’s micro-borrowers understand that the only way to keep access to a future loan if and when they need it is to faithfully repay today’s loan.
Some microfinance is delivered through credit unions and other financial cooperatives. Application of capital adequacy norms to these institutions presents a particular issue with respect to the definition of capital. All credit union members have to invest a minimum amount of “share capital” into the institution. But unlike an equity investment in a bank, share capital can usually be withdrawn whenever a member decides to leave the credit union. From the vantage of institutional safety, such capital is not very satisfactory: it is impermanent, and is most likely to be withdrawn at precisely the point where it would be most needed—when the credit union gets in 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 credit union’s capital adequacy falls to a dangerous level. A more conservative approach, now recommended by the World Council of Credit Unions, is to require credit unions to build up a certain level of institutional capital over a few years, after which time capital adequacy is based solely on those retained earnings.
Unsecured lending limits and loan-loss provisions
The experience in normal banking is that loans are more likely to default when they are not backed by collateral or guarantees. Thus, banking regulations often put tight limits on unsecured lending—for instance, capping it at no more than 100 percent of the bank’s equity base. Such a limitation would make a portfolio of uncollateralized micro-credit impossible, at least in a specialized MFI. Some regulators have avoided the problem by treating group guarantees as collateral. But not all micro-lenders use a group methodology, and group guarantees are less effective than is commonly supposed. Many MFIs do not really enforce such guarantees, and loan losses in MFIs that use such guarantees are not markedly lower than loan losses in MFIs that do not.
A more straightforward approach is to recognize the empirical evidence. Worldwide, with relatively few exceptions, uncollateralized loans in a country’s licensed MFIs suffer less delinquency and default than collateralized loans in normal bank portfolios. The reasonable response to this evidence is to put no collateral requirements on micro-loans, but instead to concentrate on close supervision of the MFI’s lending systems and repayment history.
Banks in some places have been required to book a loan-loss provision expense to cover the full value of uncollateralized loans they make, even before they become delinquent. Again, this is unworkable for micro-credit. Even if the provision expense is later reversed when the loan is collected, the accumulated charges for loans that are showing no problems would produce a massive under-representation of the MFI’s real net worth. And such a requirement has no empirical justification in the case of microcredit. Thus, general provisions (provisions booked when the loan is made, and so not based on the presence of any repayment delays) should be no more stringent for micro-credit than for normal portfolios.
The picture changes, however, once a micro-loan has actually fallen late. When one narrows the focus down to the micro-borrowers who do run into repayment problems, experience shows that ultimate collection of their loans is less likely than collection of collateralized loans that fall late by the same amount of time. As a result, provisioning of already-delinquent loans needs to be more aggressive for micro-credit than for normal collateralized loans.
Loan file requirements
Given the nature of microfinance borrowers, their informal businesses, and their loan sizes, it would be unreasonable to make micro-lenders generate the same loan documentation that is required for normal bank loans. This is particularly true in the case of financial statements for the borrower’s business, evidence that the business is formally registered, or collateral documentation. On the other hand, micro-loan files should always
contain at least the loan contract, a record of the borrower’s repayment history on prior or concurrent obligations, and a simple estimate of the borrower’s income, expenses, and repayment capacity, at least where the MFI’s methodology relies on loan officers rather than fellow group members to determine repayment capacity. However, MFIs that make repeated short-term loans, for instance every three months, should not be required to do a fresh analysis of borrower cash flow before 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. Such rules would need to be waived for MFIs that do group lending with crossguarantees among the group.
Insider lending
Loans made to owners, directors, or managers of a bank are not likely to receive the same objective scrutiny as loans to unrelated parties. In recognition of this fact, most banking authorities now restrict insider lending to some limited percentage of the bank’s assets or equity. This author’s view is that insider lending should be completely prohibited in licensed MFIs, with the exception of small welfare loan programs for employees. When specialized MFIs are receiving favorable regulatory treatment for the sole reason that they are extending financial access to low-income customers, it is hard to see any reason or need for insider lending.
Frequency and content of reporting
Banks may be required to report their financial position frequently, even daily. In many settings, the undeveloped state of transportation and communication infrastructure may make this difficult or impossible for rural banks or branches. In addition, frequent or voluminous reporting to the banking supervisor can add substantially to the administrative costs of an intermediary, especially one that specializes in very small transactions. The chief financial officer of BancoSol once estimated that compliance with the banking supervisor’s reporting requirements cost the bank 5 percent of its
assets the first year, and 1 percent or more a year thereafter. On the other, effective supervision is impossible without adequate reporting. Specialized microfinance banks or branches usually present a less complex set of risks than normal banking, so it should be possible to supervise them well based on reporting that is somewhat less burdensome and expensive than what conventional banks have to provide.
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 clientele. For instance, the convenience of clients who are running their micro-businesses all day might require operations outside normal business hours, or cost considerations might require that staff rotate among branches that are open only one or two days a week. Security requirements such as guards or expensive vaults, or other normal infrastructure rules, could make it too costly to open small-volume branches in poor areas. Branching and physical security requirements merit reexamination but not necessarily elimination in the microfinance context. Clients’ needs for financial services have to be balanced against the security risks inherent in managing cash.
Ownership requirements
Some countries have ownership-diversification rules that prohibit any single party from controlling more than 20 percent (for instance) of a bank’s shares. Also, NGOs may not be eligible to own bank shares. Both of these rules serve legitimate prudential objectives, but they can cause serious difficulty in the common case where the assets of a newly licensed MFI come almost exclusively from an NGO that has built up the business over a number of years. In recent years, commercial and quasi-commercial investors are showing greater interest in buying shares of newly licensed MFIs, but there are many transforming MFIs for whom attracting such investment is not a practical option, at least not at the time that they convert to licensed form. If the original NGO has to find new owners who will purchase 80 or 100 percent of the business, transformation into licensed form could be delayed a long time. Some banking supervisors are allowing the original NGO to own most or all of the shares of a newly licensed MFI, with a requirement that the ownership structure has to be brought into linewith normal banking rules over a reasonable period of years.
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 depositors’ losses even in the absence of an explicit legal commitment to do so. There is considerable debate about whether public deposit insu -rance 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 probably ought to be that deposits in other institutions prudentially licensed by the financial authorities should also be insured, absent strong reasons to the contrary.
Branchless banking
In a growing number of developing and transition countries, financial services are being provided outside of conventional bank branches. The use of automated teller machines (ATMs) has been spreading for years. More recently, payment, transfer, and savings services are being offered through post offices or retail outlets like groceries, pharmacies, or gas stations. Such services may be used mainly by the middle class, but they hold promise for poor people as well, especially the rural poor. Using such “retail agents,” banks can reach places where building and staffing a branch would be unprofitable because of remoteness, low client density, or low client transaction sizes. In addition, mobile phone operators in countries like the Philippines, South Africa, and Kenya are exploiting their networks to provide fast and convenient payment and transfer services to their subscribers, who include increasing numbers of low-income people.
Some branchless banking is bank-led: all of the clients’ transactions are with a licensed commercial bank, and the retail agents or mobile phone operators are acting as third-party agents to handle cash on the bank’s behalf. The bank remains responsible for any cash received. These arrangements raise some regulatory risk, including security of cash-handling and proper training of agents, but in general they do not add materially to the risks that are present in normal branch-based bankin.
More of a regulatory challenge is presented by non-bank-led models, where the client’s cash is taken and held by a company like a mobile phone operator that is not licensed and prudentially supervised by the banking authorities. When such companies are holding significant amounts of customers’ cash, should they be required to meet the same prudential standards as banks? South Africa’s answer to that question is a conservative one: any mobile phone operator that wants to provide “e-money” services is required to partner with (that is, operate under the license of) a commercial bank. This raises the mobile operator’s costs considerably, and these costs must eventually be passed along to the client.
The Philippines is starting with a more liberal approach, allowing mobile operators to operate independent of a banking license. However, the amount of such transactions and the size of outstanding balances owed an individual customer are capped at low levels. So far this arrangement has not created significant problems, but the central bank is moving to tighten restrictions further. It is not yet possible to identify best practices in dealing with this issue: so far, the European Community has not been able to agree on a common approach.
SUPERVISORY ISSUES
The Burden of Supervising Small Intermediaries
For bank supervisors in many developing countries (though certainly not all), the central fact of life is responsibility for supervising a commercial banking system with severe structural problems, often including some sizable banks teetering dangerously close to the edge of safety. The collapse of one—or a half-dozen—of these banks could threaten the country’s financial system with implosion. In trying to manage bank risk, supervisors may have to work in a political minefield, because the owners of banks are seldom under-represented in the political process. The supervisors’ legal authority to enforce compliance or manage orderly clean-ups is often inadequate. They may not have enough control over the tenure, qualification, and pay of their staff. Monitoring healthy banks is challenging enough, but the real problems come when it is time to deal with institutions in trouble. When a sick bank finally crumbles, its president can start sleeping again(though perhaps in a different country), while it is the supervisor who has to stay awake at night worrying.
If bank supervisors sometimes display resistance to adding MFIs mostly small, mostly new, mostly weak on profitability to their basket of responsibilities, we should recognize that their reasons may be nobler than narrow-mindedness or lack of concern for the poor.
The Philippines licenses hundreds of small intermediaries as “rural banks.” Originally, the minimum capital requirement for a rural bank was very low. These banks are not microfinance institutions, but their operations do include credit and deposit services for lower-income clients. They have access to the national payments system and are supervised by the central bank. As of September 1997, 824 rural banks were serving half a million clients. These banks had only about 2 percent of the banking system’s assets and deposits, but they made up 83 percent of the institutions the central bank had to supervise.
Supervising the rural banks severely stretched the resources of the central bank’s supervision department, tying up as much as one-half of its total staff and budgetary resources at times. In the early 1990s one in every five rural banks had to be shut down, and many others had to be merged or otherwise restructured. An unpublished 1996 analysis reported that about 200 inspectors were assigned to the rural banks, but even this level of resources was viewed as inadequate. Each on-site examination consumed up to three person-weeks or more. At one point the supervisory department found that this burden, combined with its budget limitations, was severely endangering its ability to function.
One of the responses to the crisis was to raise the minimum capital substantially. But a knowledgeable observer has guesstimated that as of late 2007, perhaps half of the rural banks are still technically insolvent, though these tend to be the smaller ones. The larger rural banks have most of the assets and customers are said to be strong and expanding aggressively.
The occurrence of a supervisory meltdown doesn’t necessarily mean that licensing rural banks has been a failure in the Philippines. Hundreds of thousands of people are still getting services that would otherwise have been unavailable to them. But the experience, and similar ones in Indonesia and Ghana, teaches us to be realistic about the difficulty of supervisin large numbers of small new financial institutions. Some would argue that ineffective supervision is worse than no supervision at all, because it misleads depositors and tarnishes the credibility of the banking authorities.
Minimum Capital as a Rationing Device
When regulators set minimum capital requirements for licensing MFIs, a major consideration should be limiting new licenses to a number that is consistent with available supervisory resources. Obviously, this has to be balanced against the objective of opening access to financial markets, an objective that tends in the direction of keeping minimum capital as low as possible. There is a strong argument to be made that regulators should start with relatively high minimum capital for a new licensing window, and gradually relax the requirement after there has been more supervisory experience, and supervisors are better able to judge what they can take on.
A country does not necessarily need large numbers of licensed MFIs in order to serve its market well. In most countries, a few large MFIs account for the vast majority of the outreach. In 2000, Bangladesh probably had over 1000 MFIs, but the largest ten served all but about 15 percent of the clients.
Small Community-based Intermediaries
Smaller institutions may not require as much supervision as big ones, but there are lower limits to how far supervision can be watered down. At some point, “supervision lite” is no longer effective, if effectiveness means that the supervisor can expect to flag most problems before they have gotten too serious to fix.
Some member-owned intermediaries take deposits but are so small, and sometimes 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 institutions that cannot be supervised are not safe, and therefore should not be allowed to take small depositors’ savings. After all, are not small and poor customers 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. Poor people can and do save. If formal deposit accounts are not available, they have to fall back on savings tools like currency under the mattress, livestock, building materials, or informal 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 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 requirements 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 such as employees of a company.) Once the size limits are exceeded, the institution must comply with prudential regulation and be supervised. If small intermediaries are allowed to take deposits without prudential supervision, 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 mainly on their knowledge of the individuals running the institution.
Supervisory Tools and their Limitations
Some standard tools for examining banks’ loan portfolios are ineffective for micro-credit. As noted earlier, loan-file documentation is a weak indicator of micro-credit risk. In a commercial bank, one can often capture most of the portfolio risk by examining a small number of large loans, but this is not true in a micro-credit portfolio consisting of thousands of tiny loans.Sending out confirmation letters to verify account balances is usually impractical for micro-credit, especially where client literacy is low. Instead, the examiner must rely more on analysis of the institution’s lending systems and their historical performance. Analysis of these systems requires knowledge of microfinance methods and operations, and drawing practical conclusions 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.
When an MFI gets in trouble and the supervisor issues a capital call, many MFI owners are not well-positioned to respond to it. NGOs who own shares may not have enough liquid capital available. Development agencies and development-oriented investors usually have plenty of money, but their internal procedures for disbursing it sometimes 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 a timely injection of new capital, as the Colombian banking supervisor found out when the MFI FinanSol ran into trouble.
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 period 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 micro-credit. 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 want 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.
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 commercial bank loans.
The fact that some key supervisory tools do not work very well for microfinance certainly does not mean that MFIs cannot be supervised. However, regulators should weigh this fact when they decide how many new licenses to issue, and how conservative to be in setting capital standards or required levels of past performance for transforming MFIs.
