Liquid Asset Ratios and Financial Sector Reform

Contributor Notes

Authors’ E-Mail Addresses: AGULDEWOLF@IMF.ORG and LZAMALLOA@IMF.ORG

As a monetary, selective credit, and government debt-management instrument, a liquid asset ratio is generally inefficient and may introduce serious distortions. However, it may play a limited role as a prudential instrument, particularly in less sophisticated banking systems or in the context of currency board arrangements. Recent trends in the use of this instrument have been to either abolish it altogether or to design it so as to minimize distortions. When necessary, these changes have been part of a broader effort to make financial intermediation more efficient by relying more on markets and less on regulations.

Abstract

As a monetary, selective credit, and government debt-management instrument, a liquid asset ratio is generally inefficient and may introduce serious distortions. However, it may play a limited role as a prudential instrument, particularly in less sophisticated banking systems or in the context of currency board arrangements. Recent trends in the use of this instrument have been to either abolish it altogether or to design it so as to minimize distortions. When necessary, these changes have been part of a broader effort to make financial intermediation more efficient by relying more on markets and less on regulations.

I. Introduction

A liquid asset requirement, or ratio, is defined as the obligation of commercial banks to maintain a predetermined percentage of total deposits and certain other liabilities in the form of liquid assets. In a number of countries this requirement is calculated as a percentage of short-term liabilities. The eligible range of assets varies, but usually includes cash, deposits with the central bank, correspondent accounts, and government securities. This requirement may be maintained on a day-to-day or on an average basis.

Industrial countries have for the most part eliminated the use of a binding liquid asset requirement for monetary and prudential purposes.2 In developing countries their use mainly reflects a mix of monetary and prudential purposes. Recently, this requirement has been used in the context of currency board arrangements as a prudential instrument to help banks meet their systemic liquidity needs, given the limitations such arrangements set on the central bank’s ability to act as a lender of last resort. However, the general trend has been to reform this instrument with a view to improving banks’ liquidity management. Reform has included lowering liquid asset ratios to the minimum level necessary to manage cash flows and facilitate interbank settlements, allowing for averaging of liquid asset balances and including among the list of eligible assets those that can be realized in a relatively short time without significant loss of principal.

Several problems have typically led countries to undertake such a reform. When eligible securities carry below-market returns, financial flows tend to be diverted to markets or institutions that are exempted from the ratio. Under such circumstances, liquid asset requirements lead to inefficiencies and disintermediation and become increasingly ineffective for monetary control. In any case, liquid asset requirements have a significant monetary effect only to the extent that the assets used to satisfy them are a central bank liability or are issued and negotiated abroad. But, from a monetary standpoint, a liquid asset requirement is an inefficient instrument that may introduce serious distortions.

Typically, reform of this requirement has been part of a broader effort by countries to make financial intermediation more efficient by relying more on markets and less on regulations. In the area of monetary management, this effort involved using indirect instruments of monetary policy such as open market operations instead of direct instruments such as interest rate controls or credit ceilings. In the area of public debt management, it involved greater emphasis on selling government debt at market interest rates. In the area of banking, this effort involved implementing structural reforms in banking, including banking supervision and bank restructuring, to enhance banks’ efficiency in liquidity management and steps to enhance the effective implementation and transmission of monetary policy, including through streamlining the payment and settlement system.

This paper reviews key analytical issues and central bank practices pertaining liquid asset requirements, focusing on their effectiveness as a monetary or prudential tool. As a number of countries have reformed or removed this instrument, the paper also examines transition issues in such changes. Based on general principles developed in the first part of the paper, case studies for five countries are used to extract lessons on how to best include the abolition of a liquid asset requirement into a wider financial restructuring program. The experience of countries under review suggest that reform is most effective and most smoothly accomplished under conditions of a stable macroeconomic environment and sound fiscal policies. In addition, a market-based Government debt strategy, a sound financial system and an adequate supervisory framework also facilitate a successful reform.

Section II examines the role and effectiveness of the liquid asset requirement for a variety of purposes such as monetary control, selective credit objectives, Government debt management and prudential control. Section III addresses transition issues in reform including the conditions which facilitate the reform, possible strategic and technical problems that could arise during the reform, and the approaches to removing or lowering the liquid asset requirement.3 Section IV examines the experience of selected countries with the use and design of these requirements; and, using the analysis presented in section III, reviews the experience of the five countries (Jamaica, Malaysia, Mexico, New Zealand, and Turkey) in reform. Finally, Section V summarizes the principal conclusions.

Detailed empirical material is presented in the appendices. Appendix I tabulates liquid asset requirements in selected industrial, developing and transition economies in relation to purpose, current status and special features. Appendix II contains case studies for the five countries listed above; these highlight the conditions which facilitate the process of reform in countries with differing financial systems and traditions.

II. Role and effectiveness of liquid asset requirements

Liquid asset requirements have been used for both monetary and prudential purposes. In addition, the ability to ensure demand for certain financial instruments also makes such a requirement a selective credit instrument or even a debt-management instrument.

A. As a Monetary Instrument

Currently, the view that a liquid asset requirement is an inefficient monetary instrument is widespread. In the past, however, advocates of the credit availability theory claimed that the liquidity ratio could be used as a monetary instrument to control credit growth.4 Given that, according to this view, private investment was responsive to changes in the availability of credit, an increase in the liquidity ratio would decrease credit and therefore dampen aggregate demand. Although this viewpoint assumes that there are no offsetting increases in government expenditure, it ignores interest rate effects. An increase in the liquidity ratio would cause an increase in private sector credit interest rates, but it would also cause a decline in government securities rates, thus providing incentives for additional government expenditure. As a result, the overall effect on aggregate demand is ambiguous.

Another strand of the credit availability literature postulated that a liquid asset requirement should be used to supplement a contractionary monetary policy (such as an increase in the cash reserve requirement) during an economic expansion.5 For instance,Dean (1975) has claimed that taxing the banks by forcing them to hold larger cash reserves may actually increase their desire to hold high-income loans rather than low-interest securities. To the extent that the deposits of borrowers had a greater income velocity than the deposits given up by the purchasers of bonds, the contractionary effect was frustrated. Thus, in his view, a liquid asset requirement was necessary to avoid this effect.

In contrast to the credit availability view, monetarists postulated that liquid asset requirements were ineffective for monetary control because the money supply, instead of bank credit, was the best indicator of the thrust of monetary policy. Consistent with this view, the supply of and demand for money determined the short-term money market interest rate. This rate, in turn, affected other market interest rates, which then affected spending and output. In this way, monetary policy worked by affecting bank deposits, not bank loans. Furthermore, monetarists asserted that it was unnecessary to impose a liquid asset requirement to supplement a monetary contraction during an economic expansion. Unlike Dean, they believed that depositors and bondholders both had the same income velocity. Therefore, to ensure a contractionary monetary policy, it was sufficient to increase the cash reserve requirement.

Given the importance of the money supply in the transmission of monetary policy, numerous writers have analyzed the impact of a change in the liquidity ratio on the money supply. For instance, Myhrman (1973) claimed that, in Sweden, the liquidity ratio did not have any significant effect on the money supply and was therefore ineffective. He argued that the direct effect of a higher ratio was a shift from loans to bonds in bank portfolios, leaving the money stock unchanged. In his analysis, private sector lending was reduced, but the reduction was matched exactly by the sale to banks of bonds held by the private sector.

He also postulated that the indirect effects of an increase in the liquidity ratio on the money supply, such as the impact of changes in interest rates on the demand for currency and excess reserves, were small in absolute value.

Clark (1985) also analyzed the impact of a change in the liquidity ratio on the money supply. Although he argued that an increase in the liquidity ratio might actually lower the money supply, he was not able to corroborate his findings at the empirical level using data for Canada. In contrast to Myhrman, he gave much weight to the indirect effects on the money supply of raising the liquid asset requirement. The thrust of his argument was that by reducing the effective interest rate on the banks’ portfolio, increasing the ratio would reduce the opportunity cost of holding excess reserves. Consequently, the demand for excess reserves would increase and this effect, working through the multiplier, would help reduce the money supply.

Still another author (Hörngren, 1985) postulated that the impact of an increase in the liquidity ratio on the money supply was at best uncertain and possibly perverse in that it might lead to a monetary expansion. He analyzed the effects of a change in the liquidity ratio in a financial environment with unregulated intermediaries and where banks were free to adjust interest rates. To maintain lending to the private sector constant, an increase in the liquidity ratio might actually induce banks to raise additional funds by issuing certificates of deposits (CDs).6 In Hörngren’s analysis, induced by an increase in the CD interest rate, the public would obtain certificates of deposit from banks in exchange for bonds, which banks needed to fulfill the higher ratio.7

Notwithstanding the above views, the impact of an increase in the liquid asset requirement on the money supply is unpredictable. In contrast to a cash reserve requirement, which must be fulfilled typically through holdings of cash and deposits with the central bank, a liquid asset requirement can usually be met with a range of assets.8 This range, however, complicates the ex ante calculation of the effects on narrow and broad monetary aggregates of a given change in the ratio which would differ depending upon the type and composition of eligible liquid assets and the interest sensitivity of banks’ asset portfolio.

The overall effect is likely to fall between two extremes: the minimum effect on narrow money would be a change in a liquid asset requirement that primarily induces investment in government or private sector securities. In that case, there would not be any first-round effect on narrow or broad money aggregates. Nevertheless, a second-round effect arising from interest rate effects on currency and bank reserves could have an impact on broad money. However, as discussed above, Myhrman and Clark both claimed that these effects were small in absolute value. Conversely, the maximum effect on narrow money would be achieved when the change in a liquid asset requirement was to be exclusively satisfied through adjustments in cash holdings or deposits with the central bank—amounting, in fact, to a change in the cash reserve requirement. A similar effect on narrow money would be achieved when the requirement was satisfied with central bank liabilities, such as central bank bills, or with securities issued and negotiated abroad.9 In these cases, the change in the liquid asset requirement primarily influences the demand for base money and, in turn, the overall liquidity conditions in the economy. Nevertheless, from a monetary standpoint, a liquid asset requirement is generally an inefficient and redundant instrument that may introduce serious distortions.

B. As a Selective Credit Instrument

Although liquid asset requirements’ effectiveness for monetary policy is inadequate, they have important allocative effects on certain sectors of the economy and have thus have been used for selective credit control purposes.10 In particular, they have been used to allocate credit to the government by requiring commercial banks to hold government debt such as treasury bills, often at below-market rates.11 A binding liquid asset requirement forces an increase in the demand for these securities and limits their trading.

However, in all but the most liquid markets, such forced demand for government securities will distort the structure of interest rates. Yields on government securities will be lower than they would be if banks could allocate all assets freely. Below market returns, necessarily, imply lower returns on banks’ overall asset portfolio—amounting to an implicit tax on the institutions subject to the liquid asset requirement. Borrowing costs shift from the government to the financial sector and, indirectly, to the economy at large. Owing to the implicit tax imposed by the forced investment, banks will want to increase their spreads, which will encourage disintermediation.

The exact size of these interest effects depends on the degree of distortions created by the liquid asset requirement. The degree of distortion itself is a function of the liquid asset requirement design, the size of the spread between market interest rates and the interest rate (or yield) on eligible securities, the overall market volume and menu of securities, and the degree of competition in financial markets. Clearly, a liquid asset ratio maintained on average would create fewer distortions than one maintained on a day-to-day (continuous) basis because the former allows banks to make better use of their liquid reserves. Also, the larger the spread between market interest rates and interest rates on eligible securities, the larger the implicit tax imposed on banks, and, thus, the larger the distortion. Moreover, in sophisticated and deep financial systems with a wide range of liquid assets, such as treasury bills, bank CDs, banker’s acceptances, prime bills, and repurchase agreements (repos), forced investments cause fewer distortions than in developing markets with their limited range of liquid assets. This is because the demand for securities tends to be more elastic in the former than in the latter. Furthermore, a competitive financial market would be less likely than a noncompetitive one to transfer the implicit tax imposed by the forced investment, and the tax would therefore be less distortive.

Interest rate distortions like these just described are likely to lead to disintermediation from the regulated financial system, and, thus, the effectiveness of liquid asset requirements in inducing a higher demand for government securities remains dubious, at least in the long run (Hodgman, 1972). Applied to banks only, such a requirement would increase banks’ holdings of government securities, but it would not necessarily increase the aggregate demand for these assets. Borrowers whose loans are ineligible to meet the liquid asset requirement might shift to other financial institutions that are exempt from the requirement. As a result, banks’ relative importance in financial intermediation would diminish. Also, the other financial institutions’ demand for government securities would decrease, offsetting the banks increased, albeit forced, demand for these securities. In addition to the disintermediation effect, the interest rate distortions could hinder the emergence of an efficient and responsive financial sector. The liquid asset requirement reduces investment choices for banks, which thus have a narrower scope to assess market risks and returns.

C. As a Government Debt Management Instrument

While liquid asset requirements traditionally have been used to channel credit to the government, they have also been used to develop a market for government securities, particularly in Asian countries such as India. Even where the government is willing to pay market interest rates a liquid asset requirement is attractive to the debt manager. It creates a captive market for government securities, which facilitates the placement of treasury securities. Also, at least in the short term, it is likely to lower the budgetary costs of a given deficit by lowering interest rates on government debt.

Easier debt management and lower interest outlays, however, come at a cost. In addition to the disintermediation effect and the inefficiencies that emerge with the use of liquid asset requirements, they contribute to the thinness of the secondary bond market by forcing banks to buy and hold a substantial fraction of the outstanding stock. Moreover, they divert attention from actions necessary to improve the creditworthiness and market design of the new government debt instruments—in particular, ensuring an appropriate interest rate level as well as the market’s desired maturity structure.12 Furthermore, given the monetary effects of a liquid asset requirement, maintaining it for debt-management reasons forces the central bank to subordinate monetary and interest rate management to the budgetary cycle.

D. As a Prudential Instrument

As a prudential instrument, a liquid asset requirement has limited usefulness. Traditionally, it has been used presumably to ensure that individual banks can meet their obligations as they fall due without incurring the heavy cost of trying to sell illiquid assets. However, notwithstanding its apparent simplicity and easy applicability, a liquid asset requirement may be a poor indicator of banks’ liquidity and may be misleading.

Its use implies that the degree of liquidity of the various items included on the asset side of the balance sheet can be determined with certainty. However, even if a secondary market exists for government and private securities, these securities are not capital certain, i.e., the market value may vary over time. Swift changes in market confidence and conditions may reduce the value or marketability of certain assets, such as medium- and long-term securities, that would normally be regarded as highly liquid.

Even if the assets included are truly liquid, the rest of the bank’s assets or liabilities are not necessarily liquid. For instance, the balance sheet of a bank in compliance with a liquid asset requirement could also entail a significant maturity mismatch between assets and liabilities such that, while the bank holds a large portion of its liabilities in short-term deposits, it holds most of its assets in long-term mortgage loans. Hence, a bank that is in full compliance with the liquid asset requirement could be in a much worse liquidity situation than another that does not meet the requirement.

In addition to the above shortcoming, a liquid asset requirement applied on a day-to-day (continuous) basis does not enable individual banks to meet their obligations as they fall due, much less to absorb unforseen withdrawals of deposits. If the funds a bank holds to meet a liquid asset requirement are blocked, they constitute safe assets that make it easier for the bank to acquire assets from other sources, for example, through interbank borrowing.

However, they would not constitute “primary” liquid assets in the sense of immediate availability. In this regard, stating the liquid asset requirement as an average over some period provides individual banks with more flexibility to meet daily liquidity needs.

Liquid asset requirements can have some advantages as a prudential tool when commercial banks lack sophisticated tools for managing liquidity. The main additional benefit derives from their being easily monitored, especially if banking supervision capabilities are being established or are relatively weak. Likewise, a statutory liquid asset requirement can facilitate prudent behavior within the banking sector by ensuring a high degree of system liquidity.

However, to serve its prudential purpose, a liquid asset requirement needs to be carefully designed to mitigate some of its shortcomings. First, the level must be nonbinding; that is, it must be kept at the minimum level needed to manage cash flows and facilitate interbank settlements. Also, for reasons explained above, it is also preferable to apply the liquid asset requirement as an average requirement. In addition, the definition of “liquid asset” needs to be narrow. It should include assets that, in fact, can be realized in a relatively short time without significant loss of principal, such as cash, deposits with the central bank, correspondent and collection accounts, accounts receivable in good standing, and short-term government securities. Ideally, treasury bills and treasury bonds should be highly liquid assets. In this connection, a functioning secondary market or a central bank rediscount window would ensure that such assets can be mobilized quickly. Foreign assets issued and negotiated abroad may also be included, provided that the above requirements and the limit on the net foreign asset position are satisfied.

III. Transitional issues in reforming liquid asset requirements

Given the problems and shortcomings outlined above, a country’s first-best solution will often call for significantly lowering or abolishing an existing liquid asset requirement. In countries where the ratio is lower than banks’ voluntary holdings of eligible securities, its abolition will not lead to any changes in portfolio behavior. However, in countries with a binding and high liquid asset requirement, a number of potential complications may arise if the ratio is abolished without implementing appropriate reforms. Furthermore, the transition itself needs to be carefully designed to ensure a smooth reform process.

A. Analytical Considerations

Transitional problems may arise if initial conditions in the monetary and fiscal environments as well as in the banking sector are unfavorable—namely, excess liquidity in the banking sector, high fiscal deficits and financially troubled banks, respectively. It may thus be necessary toto abolish the requirement gradually so that the proper adjustments can be made in each of the areas highlighted above.

This section summarizes general guidelines for each of the areas noted, with more detailed discussion of possible cases given in Table 1. In the transition, adaptations to a liquid asset requirement might be possible, which can reduce the implied level of distortions (Table 1).13

Table 1.

How to Lower or Abolish a Liquid Asset Requirement

article image
Source: Gulde (1995)

For an assessment of the monetary environment, the relevant criteria will include the initial level of liquidity and the availability of alternative monetary instruments. Abolishing a liquid asset requirement is easiest when there is no excess liquidity and alternative monetary instruments are available. Even in the presence of excess liquidity, a reduction in the liquid asset requirement is possible when the demand for government securities is high. In all other cases, parallel supporting measures, such as sterilization operations, will be required. Implementing sterilization operations, in turn, might require developing alternative monetary instruments.

The overall fiscal environment—including the budget deficit and available financing instruments—is the second area of concern. If the starting deficit is low and alternative financing instruments are available, there are no fiscal impediments to a speedy abolition of the liquid asset requirement. However, reducing the fiscal deficit while removing or reforming the liquid asset requirement offsets the resulting interest rate pressures. If, in contrast, the deficit is sizable and alternative financing instruments do not exist, the case for a quick abolition of the liquid asset requirement is weakened. In this case, phased-in abolition should be coordinated with both fiscal action to lower the overall financing needs and, as required, with the development of market-based financing instruments.

Finally, before a country’s monetary authorities can determine how quickly to phase out a liquid asset requirement, the state of the banking system, including the health of the banking sector and progress in modernizing banking supervision should be examined. Straightforward advice for immediate abolition can be supported only when policy makers have other prudential instruments at hand to monitor the solvency and liquidity of commercial banks. In all other cases (discussed below), the liquid asset requirement might have to stay in place temporarily while supervision based on other instruments is being strengthened. Meanwhile, however, the liquid asset requirement might be designed more efficiently, for example, by allowing for period averaging, reducing excessively high levels, and broadening the range of eligible assets to include those assets that in fact can be realized in a relatively short time without significant loss of principal.

Sometimes, the rise in interest rates that may accompany the abolition of the liquid asset requirement reduces the value of the often large government securities portfolio of banks, and depending on the accounting rules for the valuation of securities, adversely affects banks’ balance sheets and income statements. In addition, the realized value of a bank’s government securities portfolio could be affected differently if it regularly trades the securities as opposed to maintaining them until maturity. It may be desirable to abolish the liquid asset requirement gradually while banks build up enough reserves to absorb valuation or actual losses.

B. Technical and Strategic Problems

Liberalizing interest rates while maintaining, in parallel, the liquid asset requirement poses some technical and strategic problems for monetary management. During periods of cash reserve shortages, when commercial banks hold limited liquid assets in excess of the requirement, an outright purchase of treasury bills by the central bank to relieve the cash reserve shortage could create a shortage of liquid assets (Johnston and per Brekk, 1989). Early in the reform process, the central bank could reduce the liquid asset requirement to avoid this potential complication. If this problem arises at a later stage of the reform process, lowering or phasing out the requirement could be accelerated. Alternatively, to relieve cash reserve shortages, the central bank could use repos rather than outright purchases of treasury bills while allowing those bills to continue to be counted toward meeting the liquid asset requirement. Under these circumstances, repos would not create a shortage of liquid assets. However, this may entail double-counting the cash paid by the central bank and the treasury bills purchased in a repo. In any case, repo transactions would not solve the problem of a structural shortage of discountable paper.

Introducing market-related selling techniques for government securities while maintaining the liquid asset requirement also poses some technical and strategic problems for fiscal and debt management. In particular, issuing additional securities at auction while maintaining a captive demand for securities may distort interest rate signals. Maintaining a captive demand for the eligible securities implies lower budgetary costs than those that would be incurred at truly market rates. This, in turn, may induce further government spending at a time when fiscal tightening is necessary. Another problem is that, as discussed previously, forcing banks to hold government securities may limit the growth of outright transactions and—depending on banking regulations—repos in that security. Thus, to get reliable interest rate signals and for market-development purposes, it might be preferable to make the new or additional government securities ineligible to meet the liquid asset requirement. Alternatively, lowering or phasing out the requirement could be accelerated.

C. Methods

Once a decision is made to gradually lower the liquid asset requirement, several methods to attain this outcome may be used. To reform the liquid asset requirement, the new requirement may be applied on the stock or flow (or both) of deposits as of a benchmark time, and ordinary or specially issued government securities may be used as eligible assets.

Regarding the stock or flow dilemma, several options are possible. First, the liquid asset requirement could be maintained on both the stock and the flow of deposits, but the ratio could be reduced (through applying the same ratio on both the stock and the flow of deposits).14 Second, the liquid asset requirement could be applied only on the stock of deposits as of a benchmark time. Any increase in deposits would not require additional purchases of government securities. After a reasonable adjustment period, the liquidity requirement on the stock of deposits could also be lowered. Third, the liquid asset requirement could be applied only on the flow of deposits as of a benchmark time.

These options each have a number of advantages and disadvantages. In general, maintaining the requirement on the stock of deposits, but not on the flow, would favor new banks because they would not be subject to any constraint. In contrast, maintaining the requirement on the flow of deposits, but not on the stock, would penalize banks that are actively mobilizing additional resources, thus discouraging competition among banks. Therefore, applying the liquid asset requirement across the board on both the stock and the flow of deposits gives an even treatment to new and old banks. The latter option would be particularly attractive in growing financial systems with new banks, and it has the advantage of simplicity. Although it does represent a less marked departure from the past, and thus offers the temptation of raising the coefficient whenever needed, the same effect could occur under other options.

Regarding the eligible assets, two main options are possible. If the liquid asset requirement is maintained (both on the stock and flow), banks would continue satisfying the requirement with market-based treasury bills. Alternatively, if the requirement is only maintained on the stock of deposits, the government could convert banks’ statutory holdings of treasury bills into long-term securities.15 While maintaining a captive demand for securities may distort interest rate signals, issuing long-term securities may pose some operational problems. First, there is the problem of appropriately “pricing” long-term securities. Second, conversion into long-term securities may create liquidity problems for banks that suffer a deposit loss, particularly if the secondary market for government securities is not well developed. During the transition, unless banks hold government securities at below market rates, it is thus preferable to maintain the requirement and to continue satisfying the requirement with market-based treasury bills.

IV. Country experiences

Country experiences with designing and using liquid asset requirements are analyzed below. While industrial countries have for the most part eliminated them for monetary and prudential purposes, developing countries and economies in transition are still using them. Recently, liquid asset requirements have been used in the context of currency board arrangements as a prudential instrument to help banks meet their systemic liquidity needs without resorting to borrowing from the central bank. However, the general trend has been to reform this instrument with a view to improving banks’ liquidity management. This reform has included lowering liquid asset ratios to the minimum level needed to manage cash flows and facilitate interbank settlements, allowing for averaging of liquid asset balances and including among the list of eligible assets those that can be realized in a relatively short time without significant loss of principal. Thus, in addition to analyzing the use and design of liquid asset requirements in selected countries, this section discusses the conditions that facilitated or hindered the process of reforming or removing the liquid asset requirement in Jamaica, Malaysia, Mexico, New Zealand, and Turkey.

A. The Use of Liquid Asset Requirements16

Supervisors in industrial countries mainly use an approach to measure liquidity that is based on a global analysis of a bank’s balance sheet and relevant off-balance-sheet items that can give a much more accurate picture of a bank’s liquidity.17 This approach places greater emphasis on the maturity structure of a bank’s assets and liabilities and is based on a cash-flow concept.18 Currently, industrial countries do not use a minimum liquid asset requirement for monetary or debt-management purposes, and only Austria and Iceland use it mainly for prudential purposes.19 In some countries, such as Ireland and the Netherlands, supervisors enjoy a certain discretion to modulate liquidity requirements according to circumstances and the risks experienced by individual banks. However, other industrial countries (Canada, Italy, New Zealand, Spain, and United States) do not use liquid asset requirements at all for prudential purposes.20

The use of liquid asset requirements in developing countries mainly reflects a mix of monetary and prudential purposes. In the transition from direct to indirect instruments of monetary control, developing countries have taken different measures in connection with liquid asset requirements. Some countries continue using them as a monetary instrument by including central bank liabilities among the eligible assets. For instance, Bangladesh, Botswana, Ghana, Jamaica, Malaysia, and Mauritius allow banks to satisfy the liquid asset requirement with central bank bills. Still other countries use this requirement mostly for prudential purposes (Kenya and Taiwan Province of China). Other countries have gradually abolished or are abolishing liquid asset requirements as a selective credit or debt-management instrument (India, Mexico, Tunisia). Few sample countries (Burundi, Ghana) use it for selective credit purposes.

More recently, liquid asset requirements have been used for prudential purposes in the context of currency boards (Baliño and Enoch, 1997). In 1995, the Central Bank of Argentina, which has a currency board arrangement, effectively replaced an unremunerated reserve requirement system with a liquid asset requirement, both maintained on average. This new requirement reduces the implicit tax burden imposed on commercial banks and safeguards bank liquidity against capital flight. It applies to all non-interbank liabilities and can be satisfied by holdings of interest-earning Bank Liquidity Certificates issued by the treasury, a special account at an international bank abroad, the central bank’s reverse repurchase agreements, government bonds of OECD countries with ratings of a single A or better, and Argentine securities.21 If banks are allowed to fulfill the liquidity requirements with foreign assets, the need for backing banks’ reserves at the monetary authority declines. Furthermore, a reduction in the liquid asset requirement would allow the banking system to obtain foreign assets to meet deposit withdrawals, thus serving as a substitute, albeit an imperfect one, for the lender-of-last-resort function that the Central Bank of Argentina cannot perform owing to the currency board arrangement.22

Some countries with significant foreign currency deposits such as Croatia and Honduras have used liquidity requirements on foreign currency deposits to limit foreign currency credit growth, while imposing unremunerated reserve requirements on domestic currency deposits. Given that in these countries liquidity requirements on foreign currency deposits are mainly satisfied with foreign currency liquid assets remunerated at market rates such as deposits abroad, that policy encourages foreign currency deposit growth by unduly taxing domestic currency deposits.23

Many economies in transition, in particular the Baltic countries, Russia, and the other countries of the former Soviet Union, have a number of liquidity ratios in place for prudential purposes. However, liquidity cannot be easily managed owing to the limited availability of liquid assets and the embryonic nature of secondary markets for securities. Even in countries where there is a growing secondary market for government securities, such as Russia, liquid assets cannot be mobilized as rapidly as in other countries because of the inadequacies of the payments system. Nevertheless, these ratios may be needed for some time, until the approach to liquidity supervision becomes more sophisticated. In this regard, supervisors need to develop an understanding of the issues in asset and liability management, and commercial banks need to implement liquidity indicators and cash-flow analysis systems and report on them within a specific time frame.

B. The Design of Liquid Asset Requirements

In general, countries have defined the liquid asset requirement in various forms depending on the eligible securities banks are required to hold, the base to which the liquid asset requirement applies, and the methodology of calculation.

The eligible range of assets varies, but usually includes cash, deposits with the central bank, government securities, and net short-term positions in the money market. In a few countries, such as Botswana, private sector securities are also included, but the credit risk of holding these securities is typically higher than that of holding government securities. Only in Argentina were reverse repos included as an eligible asset; this allows the Central Bank of Argentina to manage intramonth fluctuations in liquidity.24 In a few countries, such as Jamaica and Malawi, eligible assets include only government and central bank securities; however, in addition, a cash reserve requirement is imposed on commercial banks.

The base usually consists of deposit liabilities, but in a few countries it includes nondeposit liabilities, such as asset-backed securities and repos (Turkey), and thus the latter bank liabilities are viewed as close deposit substitutes. Regarding interbank borrowing, in Malaysia, liquid asset requirements apply to net interbank borrowing of individual banks, thus eliminating the effect of double taxation present when interbank borrowing is included.

Countries that have made some progress in financial sector reform, and that maintain a liquid asset requirement, usually calculate it on average to improve banks’ liquidity management (Argentina, China, Malaysia, and Mauritius). This measure lowers the operating costs of financial institutions, and dampens the volatility in interbank market interest rates caused by banks’ seeking to meet their day-to-day requirements.

Several countries have a liquid asset requirement for foreign currency deposits (Argentina, Croatia, Honduras, Kenya, Malaysia, Mexico, and Turkey). Whereas in Argentina, Croatia, Honduras and Mexico banks are allowed to fulfill these requirements with foreign assets,25 in Kenya, Malaysia and Turkey, banks must fulfill the requirement with domestic currency assets, mostly with government and central bank securities. Holding liquidity requirements on foreign currency deposits in foreign currency may limit foreign currency liquidity risk, particularly in countries with a significant amount of foreign currency deposits such as Argentina.

C. Lessons From Experience in Reform

This section analyzes the actual experience of selected countries in removing or reforming liquid asset requirements in light of the discussion in the previous section. The sample contains both industrial and nonindustrial countries, selected to include both successful and unsuccessful experiences. It comprises Jamaica, Malaysia, Mexico, New Zealand, and Turkey.26 To better illustrate the complexities of this reform, the subsequent analysis focuses on (1) institutional and macroeconomic conditions, (2) the pace of reform, (3) supporting reforms, and (4) implementation experience.

Institutional and macroeconomic conditions

Although the reform experience of the sample countries has been diverse, certain initial conditions were common to most of them. In all countries, central banks relied on a monetary framework with direct controls, including interest rate restrictions, domestic credit controls, and high cash and liquidity requirements. The use of these controls diverted financial flows to unregulated markets or institutions (see Appendix II). Moreover, the government securities market was largely captive in that the government relied on statutory investment requirements to fund part of its budgetary operations. Outright transactions in government securities were virtually nonexistent because these securities were generally held until maturity to meet the day-to-day continuous requirements. Similarly, repurchase agreements in government securities were also discouraged because banking regulations usually required eligible assets to be unencumbered. Further common features were a low level of central bank autonomy, weak and segmented money and interbank markets, and ineffective banking supervision.

All sample countries had large macroeconomic imbalances at the start of the reform, as measured by high fiscal deficits in relation to GDP (Table 2). In addition, Jamaica, Mexico, and Turkey had inflation rates of more than 20 percent; but only Jamaica had negative real interest rates. Jamaica, New Zealand, and Turkey were coping with excess liquidity in the financial system at the time of the reform. As a consequence, for the sample countries, reforming the liquid asset requirement was part of a more comprehensive reform and stabilization program. Undoubtedly, the success of the reforms and that of the stabilization program were closely linked.

Table 2.

Macroeconomic Conditions During the Reform of Liquidity Requirements

article image
Source: International Monetary Fund, International Financial Statistics and various Recent Economic Developments.

Measured using consumer price index; end of period for Jamaica, Mexico and Turkey; average, otherwise.

Measured ex-post using deposit rates for Jamaica, three-month treasury bill rate for Malaysia and Turkey, 28-day treasury certificate rate for Mexico, and five-year government bond yield for New Zealand. Choice of interest rate based on data availability. Consumer price index used as deflator.

As a percentage of GDP. For Jamaica, includes central bank losses.

Measured using excess reserves of commercial banks, that is, those reserves above the banks’ demand for statutory and precautionary reasons.

Table 3.

Country Experiences During the Reform of Liquidity Requirements

article image
Table 4.

Approaches to Reform

article image

In most cases, macroeconomic conditions improved after the reforms. All of the sample countries adopted a macroeconomic stabilization program, with Jamaica, Mexico, and Turkey having an IMF-supported program at the outset of the reform. In all sample countries except Jamaica, the fiscal deficit was consistently reduced; in Jamaica, the consolidated public sector fiscal deficit returned to prereform levels in less than three years owing to the impact of a severe hurricane on public sector expenditures and on the operating surplus of enterprises involved in rehabilitation programs after the hurricane (Table 2). The Bank of Jamaica issued its own certificates to mop up the excess liquidity caused by the fiscal imbalance, but this tactic was insufficient to control credit growth because implicit interest rate caps were placed on auctioned securities. In Turkey, the fiscal deficit was reduced, but excessively high domestic borrowing requirements remained, which the authorities were at times reluctant to finance at market rates of interest.

After macroeconomic stabilization was implemented, in all sample countries except Turkey, inflation was lowered and remained below 20 percent after the reforms. Real interest rates were negative in Mexico and New Zealand for a short time, possibly owing to capital inflows. In contrast, Turkey had very high real interest rates perhaps owing to banking distress in commercial banks (see below) and political instability. In Malaysia, during 1985-86, tight liquidity conditions, caused by a decline in the terms of trade, did not translate into high real interest rates owing to interest rate controls.27

After the stabilization program and reforms were introduced, in most cases, domestic liquidity was significantly affected by capital flows. In New Zealand, the decision to float the currency in March 1985 facilitated monetary control. During 1991-94, Malaysia experienced a surge in capital inflows that—coupled with the decision to maintain export competitiveness and threatened with a loss of monetary control—prompted the central bank to impose a number of administrative and monetary measures to manage these flows.28 During 1994-95, Turkey also experienced short-term volatile capital flows in the context of a high domestic public sector, borrowing requirement (see Appendix II).

Pace and approaches to reform

The pace of reform has varied across the sample countries. While New Zealand eliminated its liquid asset requirement overnight, the rest either took a more gradual approach, or the process was not smooth.29 Jamaica took three years to phase out the liquidity ratio, but reintroduced it in less than one year owing to unstable macroeconomic conditions and inconsistencies in implementing indirect monetary control.30 Turkey maintained a high requirement on the stock of deposits as of April 1994, but, in less than one year, reintroduced it on the flow of deposit and nondeposit liabilities—albeit at a low level—owing to unstable macroeconomic and financial conditions. Malaysia took five years to reform its liquid asset requirement in a piecemeal fashion, with a view to using it for prudential purposes; however, in recent years, there was a renewed interest in the using the liquid asset requirement for monetary purposes owing to massive capital inflows. In contrast, Mexico, which initially maintained a much reduced requirement on the stock and flow of deposits in relation to prereform levels, accelerated the reform owing to a structural shortage of government securities. In 1991, the Bank of Mexico reduced the requirement, maintaining it only on the stock of deposits as of August 1991, and in 1992 eliminated it altogether to increase the supply of loanable funds and foster a reduction in intermediation margins. Today, Jamaica, Malaysia and Turkey continue to employ liquid asset requirements.

Regarding the eligible securities, only Mexico (in 1991) and Turkey chose to issue a special long-term government security to satisfy the liquid asset requirement. Because, in both cases, the requirement was calculated on the stock of deposits (as of a specified date), domestic government debt financed by banks could be maintained at a certain level, while the necessary adjustments were made in government expenditures and revenues.

Supporting measures and reforms

In all sample countries, the reform of the liquid asset requirement was part of a broader financial reform package that included adapting monetary operations to foster both market development and monetary control, implementing or further developing a market-based government debt strategy, and improving banking supervision and prudential regulation. In addition, in Malaysia and Turkey, some banks were closed or restructured owing to their significant financial problems (see below).

Central banks in the sample countries introduced indirect methods of monetary control, each at a different pace and intensity. All countries that planned to remove the liquid asset requirement liberalized interest rates early in the reform to provide adequate scope for market-based instruments. Indeed, Jamaica, Mexico, New Zealand, and Turkey had already liberalized interest rates by the time they removed the liquid asset requirement.

In some countries, such as New Zealand, success in implementing an indirect framework of monetary control, including indirect instruments, encouraged the permanent removal of the liquid asset requirement. In Mexico, outright transactions in the secondary market for treasury certificates (CETES) and repurchase and reverse repurchase operations in the same security became the main instrument in early 1990, shortly after the beginning of financial reforms and even before the liquid asset requirement was eliminated. However, in Malaysia, a thin government securities market limited the scope for open market operations in this market and Bank Negara Malaysia chose to maintain a complementary mix of direct and indirect monetary instruments. In the early 1990s, it was using indirect instruments, such as borrowing in the interbank market, but maintained liquidity requirements. In Jamaica, in the context of a fixed exchange rate and high fiscal imbalances, inconsistencies in implementing an indirect framework of monetary control led to the reintroduction of the liquid asset requirement.

Countries that successfully removed the liquid asset requirement, such as Mexico and New Zealand, also succeeded in implementing a market-based government debt strategy. To facilitate secondary market trading, the government in these countries competitively priced its own securities through auctions.

In some cases, banking reforms also had to be undertaken. Malaysia and Turkey restructured or closed a number of commercial banks that had accumulated large nonperforming loan portfolios. However, in these countries, banking problems had different effects on the pace of reform. While in Malaysia banking problems together with macroeconomic imbalances slowed down the pace, in Turkey they actually encouraged the reform. In the latter country, some of the banks’ financial difficulties arose because they were required to hold government securities with a longer average maturity than that of their deposits. As a result of this forced interest rate risk exposure, banks suffered large losses when interest rates surged in early 1994. Because this was contributing to the financial difficulties of commercial banks, the central bank eliminated the liquid asset requirement on the flow of deposits as of April 1994.

In general, in most countries, banking supervision and regulation were strengthened to contain excessive risk taking on the part of banks and limit systemic problems. In addition, other means of supervising banks’ liquidity beyond static ratios were developed, particularly in New Zealand. Although reforms in banking supervision were implemented with some delay, reversals in reforms cannot be attributed to banking supervision shortcomings. As discussed below, other factors, such as macroeconomic instability, may have been more important in inducing reversals. Nevertheless, in some cases, rapid credit growth coupled with limited banking supervision increased commercial banks’ nonperforming loans. For instance, in Mexico, the percentage of nonperforming loans net of provisions in relation to equity increased after the implementation of financial reforms including bank privatization and the abolition of the liquid asset requirement.31

Implementation experience

While New Zealand had a fairly speedy and smooth reform, other countries have faced a wide variety of problems during the process of reforming the liquid asset requirement, mainly because certain institutional and economic conditions were often lacking. Several factors contributed to a smooth implementation of reforms in New Zealand: the curtailment of the government’s direct access to central bank financing; a monetary policy firmly oriented toward price stability; the use of indirect instruments of monetary control; the development of a market-based government debt strategy; and new prudential measures and a strengthening of banking supervision. Undoubtedly, sound commercial banking was also an important factor, because banks were able to absorb some government securities valuation losses induced in part by the abolition of the liquid asset requirement.

Jamaica, Malaysia, and Turkey experienced reversals of reforms. In Jamaica and Turkey, the reversal was a direct response to excessively high fiscal and quasi-fiscal imbalances, which at times the authorities were reluctant to finance at market rates of interest. While Jamaica reintroduced the liquid asset requirement at prereform levels, Turkey reintroduced it at a low level. In Malaysia, the renewed interest in the liquid asset requirement for monetary purposes was prompted by monetary instability caused by capital inflows. Thus, unstable macroeconomic and financial conditions and monetary instability were key factors in policy reversals. In contrast, Mexico accelerated the reform after experiencing difficulties, under tight money market conditions, in implementing monetary operations through outright purchases of treasury bills while maintaining the liquid asset requirement—a potential complication discussed earlier. To lower short-term interest rates, the Bank of Mexico injected bank reserves by purchasing large amounts of CETES at the primary auction, an operation that aggravated the shortage of assets eligible to satisfy the liquid asset requirement. As a result, commercial banks had difficulty in complying with the requirement except by holding cash.

A positive externality of the financial sector reform, including the reform in liquid asset requirements, was the growth of the interbank market and of the government securities secondary markets. Once interest rates were liberalized and the government followed a market-based government debt strategy, these markets developed quickly. This was particularly so in Mexico and New Zealand, but less so in Jamaica, Malaysia, and Turkey. In Jamaica, the development of the securities market over the period of reform was constrained by the active participation of the central bank. In Malaysia, although interbank market transactions grew in volume, the growth of the government securities market remained limited because the supply of government securities was not able to match the growing demand. In Turkey, the interbank market and the secondary market for government securities have been growing, albeit erratically.

Another positive effect of the financial sector reform, including the reform of the liquid asset requirement, was that resources shifted from nonbanks back to commercial banks and from deposit-like liabilities to deposit liabilities. In addition, the reform of the liquid asset requirement increased the efficiency of financial intermediaries by reducing distortions in the structure of interest rates. Although these developments cannot be exclusively attributed to the reform of the liquid asset requirement, developments in this area certainly played a role.

In sum, experience in the sample countries suggests that the reform of the liquid asset requirement, particularly in countries where it is an important instrument, is most effective and most smoothly accomplished in conditions of a stable macroeconomic environment with sound fiscal policies. In addition, implementing a market-based government debt strategy, encouraging the development of a sound financial system, and strengthening the supervisory framework helps to sustain the reform. More generally, the experiences of those countries also suggest that the removal of the liquid asset requirement is usually part of a process of financial sector reform and cannot be segregated from a successful transition from direct to indirect monetary control.

V. Conclusions

Liquid asset requirements have shortcomings as either a monetary, a debt management or a prudential instrument. As a monetary instrument, the impact of a change in the liquid asset requirement on the money supply is unpredictable because it depends on which assets banks choose to satisfy the requirement. If banks satisfy an increase in the liquid asset requirement by holding government or private sector securities, then there will be no first-round effect on base money and broad monetary aggregates. A second-round effect arising from interest rate effects on the demand for currency and bank reserves could have an impact on broad money, but it is to be small in absolute value. If banks satisfy the liquid asset requirement by holding central bank securities or securities issued and negotiated abroad, then the maximum effect on narrow money and broad monetary aggregates would be achieved. Nevertheless, from a monetary standpoint, the liquid asset requirement is generally an inefficient and redundant monetary instrument that may seriously impede the efficiency of the financial sector. It may introduce serious interest rate distortions that are likely to induce disintermediation from the regulated financial system. Therefore, its effectiveness in inducing a higher demand for central bank securities remains dubious, at least in the long run.

As a selective credit instrument, liquid asset requirements have been used to allocate credit to the government often at below market rates. This introduces serious interest rate distortions that are likely to induce disintermediation from the regulated financial system. As a debt management instrument, a liquid asset requirement will give a distorted view on the real borrowing cost of the government and may actually impede outright transactions of Government securities. Therefore, from a monetary, selective credit and debt management standpoint, liquid asset requirements should be replaced by more market-based instruments.

As a prudential instrument, sophisticated banking systems—characterized by, among other features, greater reliance on foreign and domestic interbank markets—require more elaborate liquidity standards than static ratios. However, liquid asset requirements may play a limited role as a prudential instrument in less developed banking systems or as indicators to be flexibly used in conjunction with other measures of liquidity. Moreover, liquid asset requirements can be useful when the monetary authority has limited lender-of-last-resort capabilities—such as in a currency board arrangement—to make the banking system more resilient as well as to reduce the burden imposed on commercial banks by unremunerated reserve requirements. In all these cases, the liquid asset requirement needs to be adequately designed to minimize distortions on the financial system. In this regard, it is preferable to set it at the minimum needed to manage cash flows and facilitate interbank settlements, to maintain the required liquid asset balances on an average basis, and to include among the eligible liquid assets a variety of assets that, in fact, can be realized in a relatively short time without significant loss of principal.

Abolishing or reforming the liquid asset requirement, particularly when they are an important instrument, is most smoothly accomplished under a stable macroeconomic environment, sound fiscal policies, and, if necessary, a broad financial sector reform package. Reducing the fiscal deficit while removing or reforming the liquid asset requirement helps offset the resulting interest rate pressures. Key supporting reforms are needed to minimize a wide variety of problems and ensure as smooth a reform as possible. These reforms include the following elements:

• Introducing or further developing indirect methods of monetary control in order to have more efficient means to implement monetary policy. Also, liberalizing interest rates early in the reform process provides adequate scope for market-based instruments.

• The removal of the liquid asset requirement must usually be supported by a comprehensive program to develop a market-based government debt strategy. The government must accept market rates of interest on its debt and refrain from pressuring the central bank to keep those rates low.

• A sound banking system facilitates the reform because banks are in a position to absorb the possible valuation or actual losses of their government securities portfolio that arise from the increase in interest rates that may accompany the abolition of the liquid asset requirement. However, for a variety of reasons, banks may be in a weak position to absorb the losses and need to be restructured to deal with nonperforming assets and other capital losses. More generally, given the new market-based environment, banks need to strengthen their managerial capacity because they may have little experience in adequately assessing credit and other market risks.

• The supervisory and regulatory framework needs to be reinforced—by introducing standards for minimum capital and for provisioning for doubtful loans—to encourage prudent behavior. Adequate prudential accounting standards are also needed to have a true and fair valuation of assets and liabilities; this, in turn, facilitates the market in playing a role in ensuring financial discipline. The institution involved in supervising banks needs to build the capacity to monitor banks’ liquidity beyond the use of static liquidity ratios.

In addition, abolishing or reforming the liquid asset requirement is easier and less likely to suffer reversals if the authorities can do it gradually, in line with the speed with which they can introduce supporting measures. Sometimes, though, a rapid removal or reform is necessary, for instance, when the liquid asset requirement creates significant distortions in the banking system and is contributing to banks’ financial difficulties. At other times, the reform may be accelerated if there are structural shortages of liquid assets or if there are conflicts with newly introduced monetary instruments.

Moreover, abolishing or reforming the liquid asset requirement within a broader financial sector reform package had significant effects on the financial system, at least judging by the experience of the sample countries. In particular, the shift of resources to unregulated financial intermediaries and markets, typically observed before the financial sector reform, reversed itself. Moreover, the reform of the liquid asset requirement reduced distortions in the structure of interest rates and, in some cases, lowered the losses of commercial banks. More generally, the reforms supported the growth of the interbank market and of the government securities market. Although these developments cannot be attributed exclusively to the reform of the liquid asset requirement, the latter certainly contributed.