Frameworks for Monetary Stability

27 Government Versus Central Bank Securities in Developing Open Market Operations

Carlo Cottarelli, and Tomás Baliño
Published Date:
December 1994
  • ShareShare
Show Summary Details

Financial liberalization in general, and the adoption of indirect instruments of monetary policy in particular, has a major impact on the relationship between the government and the central bank at both the institutional and operational level.

At the operational level, the activities of the central bank and the government (or treasury in particular) become more intertwined than under a system of direct administrative monetary controls. This situation calls for close coordination at the operational level between monetary and fiscal authorities. Without such coordination, the central bank’s operational independence—i.e., the ability to control the growth of its balance sheet, enabling it to achieve its policy goals effectively and to neutralize any movements that may interfere with their pursuance—may be impaired. Operational autonomy itself is an indispensable building block on the way to the central bank’s institutional autonomy, which is now generally considered as desirable to helping achieving the central bank’s main policy goal: price stability.

One particular area where coordination is needed at the operational level between monetary and fiscal authorities is the selection and use of a financial instrument to conduct the central bank’s open market interventions. Typically, in an indirect monetary policy framework, open market operations are the central bank’s main instrument of monetary policy, supported by other instruments (such as central bank credit to commercial banks and reserve requirements). The underlying financial instrument for open market interventions can be private, government, or central bank securities. Alternatively, some countries use transfers of government deposits between the central bank’s books and the commercial banks as the main monetary policy instrument. These transfers can, in principle, yield similar results as genuine open market operations. They are therefore sometimes used in the absence of well-developed financial markets.

In developed financial systems, open market operations typically take place in secondary markets for government securities—a practice that reduces the interaction between monetary and domestic government debt management.1 Economies embarking on financial liberalization usually lack wide and deep financial markets. This situation necessitates that central banks conduct their open market operations in the primary (or issue) markets of the selected security—at least in the initial stages, until financial markets are sufficiently developed to use secondary markets. Such operations are usually called “open market-type operations.”

Irrespective of the financial instrument chosen—government or central bank securities—their use for open market operations creates an interface between monetary and debt management and, therefore, makes coordination at the operational level between monetary and fiscal authorities a necessity in order to preserve the central bank’s autonomy. Good coordinating arrangements will strengthen the central bank’s operational autonomy, its credibility in the market, and financial market development in general.

This paper reviews the relative merits and drawbacks of government and central bank securities as the underlying financial instruments for such open market-type operations, as well as the necessary features that the supporting coordinating arrangements should take on, in order to ensure the central bank’s operational independence when using these instruments.2 The paper will primarily focus on these issues from a monetary management point of view. A detailed analysis from the perspective of debt management is largely outside its scope.

Although the need for coordination between monetary and fiscal authorities is not exclusively for economies in transition, the unknown and unfamiliar nature of the issues in the initial stages of financial reform and the need to develop financial markets make it necessary to put great emphasis on proper coordinating arrangements from the onset of the reform process.

The topics under discussion have received little attention in literature as well as in practice. However, their importance should not be underestimated because, as the paper illustrates, lack of solid supporting arrangements has in some cases led to stalling (or even backtracking) in the transition to indirect instruments of monetary policy. On the other hand, good arrangements have proven beneficial for the efficiency and effectiveness of monetary management. The design of coordinating arrangements should therefore be a core element of every reform agenda.

The paper is structured as follows. It first provides the analytical background and sets forth the issues. Next it reviews selected country experiences with respect to the choice of a financial instrument for open market-type operations and the design of the supporting coordinating arrangements. The paper then assesses, on the basis of the country experiences, the relative merits of intervention instruments and reviews the supporting arrangements as the critical elements for a successful transition to indirect monetary control. Finally it sets out some general conclusions.

Background and Issues

In order to identify the issues at stake and place them in the proper context, it seems useful to recall first the salient features of an indirect monetary policy framework. Subsequently, the interactions between fiscal and monetary operations are identified and the need for coordinating arrangements is spelled out.

The Indirect Monetary Policy Framework

When analyzing the interactions between fiscal and monetary operations in a market-based financial system, it is important to bear in mind two key aspects of an indirect monetary policy framework. The first one is the nature of the monetary policy interventions; the second is the instrument mix to conduct those interventions.

Types of Interventions

From an analytical point of view, two types of monetary interventions can be distinguished (Meek, 1989). The first type relates to the central bank’s macroeconomic management: dynamic monetary policy actions that influence bank reserves to achieve broader monetary targets in line with the country’s macroeconomic economic strategy.

The second type of intervention is short-term liquidity management aimed at smoothing out sharp and undesirable liquidity fluctuations. Too large and too frequent fluctuations would endanger the achievement of the central bank’s monetary policy goals and would hamper both the smooth development and functioning of the financial markets and, in emerging markets, commercial banks’ liquidity management as well. These interventions take on a more defensive role—i.e., they neutralize autonomous changes in bank reserves to reduce movements in liquidity and hence in interest rates.

The Instrument Mix

In order to conduct both types of interventions—the dynamic policy actions and the defensive interventions, the central bank needs a mix of instruments to inject and absorb liquidity in a flexible way and to support financial market development. The importance of the latter in the financial reform process cannot be overemphasized. Well-functioning financial markets give central banks the opportunity to absorb liquidity without draining the market altogether. In turn, open market interventions themselves support financial market development as they provide liquidity to the particular market of intervention. Therefore, particularly in the early stages of financial liberalization, financial market development and the development of indirect instruments are two narrowly intertwined goals of the monetary authorities. Emerging government securities markets often provide the main impetus for financial market development, and hence, government securities are very suitable intervention instruments for the central bank’s emerging indirect monetary management. Well-functioning financial markets, in turn, facilitate domestic public debt management.3

The Fiscal-Monetary Interface

In financial systems dominated by direct controls over credit and interest rates, coordination between monetary and fiscal authorities is usually an empty notion because central banks generally function in a subordinate position. Direct controls, as well as other instruments of monetary policy in place, such as reserve and/or liquid asset requirements, are often dictated—or at least heavily influenced—by the government. In other words, monetary policy is typically a subset of fiscal policy measures.

In such settings, financial markets and instruments are also generally underdeveloped or nonexistent because, among other reasons, public debt management is guided by non-market-based principles, such as low interest rate financing and the targeting of captive markets. There is by definition, in the absence of financial markets, no need for the central bank to engage in short-term monetary management. Thus, the short-term impact of fiscal operations on domestic liquidity is not important and, as the central bank is not using any financial instrument to intervene in financial markets, there is no interference with debt management. In these circumstances, the need for coordination at the operational level between fiscal and monetary authorities is minimal.

In the new market-based environment, where direct administrative controls are replaced by an indirect monetary policy framework, two narrowly related issues in the interaction between fiscal and monetary operations come to the forefront.

First, interactions between fiscal and monetary operations become more important for the day-to-day conduct of monetary policy and market-based public debt management. With indirect monetary policies, daily government receipts and payments have a direct short-term impact on domestic liquidity and hence on interest rates and on the development and functioning of financial markets.4 These interactions require the kind of defensive monetary operations discussed earlier.

Typically, government transactions are a major source of supply of reserves. Tax collections and government payments often go through the government account at the central bank, thus directly influencing reserve money movements. Without a mechanism to neutralize these transactions in order to prevent unforeseen variations in base money and short-term interest rates, the shocks caused by these cash flows could place a heavy burden on financial markets and instruments, particularly fledgling ones. Interest rate movements, in turn, directly affect government debt servicing in an environment where public debt management is also conducted according to market-based principles.

Second, as said before, the central bank’s main instrument of monetary policy will become its open market operations. Therefore, one of the key pillars of a successful transition to indirect market-based monetary policy is the availability to the central bank of a reliable financial instrument to conduct these open market operations. Government securities are often a natural candidate, particularly if they already have an established market presence. This was usually the case in industrial countries, where financial markets were well developed prior to the introduction of indirect monetary policies.5

Well-developed government securities markets allow indirect monetary policy to be conducted from the onset in secondary markets, thereby insulating monetary policy from debt management operations, which take place for the most part in the primary market for government securities. In these circumstances, with distinct intervention markets, or even distinct instruments (e.g., different maturities of securities are used for monetary and debt management operations in the United Kingdom), the need for coordination at the operational level between fiscal and monetary authorities is minimized.

However, in other circumstances where financial markets are either nonexistent or underdeveloped, as in most developing countries undergoing financial reforms, the introduction of indirect monetary policy techniques poses more challenges. Government securities are still a natural choice for the intervention instrument, particularly if they already exist. However, given the typical underdeveloped state of markets, it is likely that fiscal and monetary authorities are initially required to operate in the same market—the primary market—with the same instrument, for both debt and monetary management.

The central bank has to rely on primary issues of government securities (open market-type operations), until such time as the markets are deep enough to switch to secondary market interventions or genuine open market operations. The use of the same instrument and market for monetary and debt management may lead to conflicts—for instance, when the central bank’s intentions with primary issues of government securities are at odds with the treasury’s goals. In these circumstances, the operational independence of the central bank is in danger. The establishment of coordinating arrangements between fiscal and monetary authorities is needed to preserve the central bank’s operational autonomy and thus the effectiveness of monetary policy.

Alternatively, in an attempt to circumvent the use of the same instrument and market and the related need for coordination, central banks may opt for a separate intervention instrument, such as central bank securities or private paper, to conduct open market-type operations, thereby largely separating monetary from debt management operations. The use of private paper, however, remains rather the exception. Private paper, while an alternative in theory, is not a real option for most countries in the early stages of financial reform. First, private paper markets are either absent or only poorly developed at the time a decision has to be made regarding the intervention instrument. In addition, the situation of excess liquidity that often occurs at the onset of financial reforms requires the central bank to have a large portfolio of private paper that it could sell to absorb liquidity, which is usually not the case. Thus, private paper would only be a valid alternative when a country’s financial situation is more or less in equilibrium.6

Central bank paper, therefore, is a more common alternative intervention instrument. However, it is a misunderstanding to think that the use of central bank paper makes coordination between fiscal and monetary authorities redundant. If not properly coordinated, the coemergence of central bank and government securities markets could hamper the development of both markets, resulting in poor transmission of monetary policy actions in general.

The choice and use of an intervention instrument for monetary policy appears to be one of the major stumbling blocks in the transition to indirect instruments of monetary policy. Loss of monetary control, owing to premature reliance on indirect instruments or lack of workable supporting arrangements, can lead to backtracking or even reverting to direct controls.

Coordinating and Supporting Arrangements

The above interactions between fiscal and monetary instruments call for coordination at the operational level between the fiscal and monetary authorities in the transition process—at a minimum to avoid conflicts and, more particularly, to achieve effective monetary and debt management. Such coordinating arrangements typically will consist of: (1) the establishment of a coordination committee at the operational level between the treasury and the central bank to forecast the effects of fiscal operations on monetary conditions and to coordinate policy actions accordingly; and (2) given the choice of a financial instrument for open market operations, the design of supporting arrangements between monetary and fiscal authorities to reduce or eliminate the potential for conflicts and ensure policy effectiveness. When government debt instruments are selected, these arrangements would stipulate how monetary and debt management are to be coordinated in practice. When the central bank prefers to use its own debt instruments for open market-type operations, arrangements need to be established to reduce or eliminate interference between central bank debt instruments and government debt instruments to preserve or enhance the effectiveness of monetary and debt management and financial market development.

Among the wide variety of issues in coordination between fiscal and monetary authorities that come to the forefront in the transition to indirect instruments of monetary policy, this paper addresses the selection of the financial instrument to conduct open market operations and, given this choice, the design of supporting arrangements that would ensure the central bank’s operational autonomy. Other papers have dealt with other aspects of coordination. For instance, Leite (1992) and Sundararajan and others (1993) discuss the need for a coordinating committee as well as its composition. Cottarelli (1993) reviews the arguments for, and the terms of, a complete halt to government borrowing from the central bank in a liberalized financial system as a condition for central bank operational independence.

Country Experiences

The following country experiences, though not exhaustive, provide a representative overview of the range of approaches that have been attempted in selecting the underlying financial instrument for open market-type operations and in designing supporting arrangements. These approaches have met with varying degrees of success and have evolved through periods of changes and adjustments along the way.

The typical approach to introducing open market interventions as part of indirect monetary management has been to use open market-type operations, as defined above, often supported by direct instruments in the early stages. While the eventual aim is to conduct open market operations in secondary government securities markets, the speed of the transition has varied. The type of supporting arrangements between fiscal and monetary authorities for conducting debt management and monetary operations in the same market also varies.

Country experiences have been divided into four main categories, mainly for presentational purposes. The first group of countries includes those where the central banks have consistently used government securities in the development of open market operations. The countries discussed are Pakistan, Kenya, The Gambia, Israel, and Mexico. A second category consists of those countries that have been experimenting with both central bank and government securities and where the latter are currently the preferred instrument. Here are reviewed the Philippines, Sri Lanka, Poland, and Nepal. The third group covers countries that have also experimented with both types, but their central banks are at present mainly relying on central bank securities—i.e., Mauritius, Ghana, Chile, and New Zealand. Finally, the fourth category includes countries that in practice have only resorted to central bank securities: Costa Rica, Indonesia, and Korea.

Operations with Government Securities

Countries Always Using Government Securities

Kenya, Pakistan, The Gambia, Israel, and Mexico are typical examples of countries that have used primary auctions of government securities as the main instrument of monetary policy in newly introduced indirect monetary policy environments. In all of these countries, some types of government securities were already in circulation prior to the introduction of indirect instruments of monetary policy. The relations between fiscal and monetary authorities, including the central bank’s role in public debt management at the time of the introduction of indirect monetary intervention, however, have varied in each country, thus influencing the initial success of the monetary operations.

The transition to indirect monetary policies in Pakistan was intended to start with central bank intervention in primary government securities markets in 1991.7 Thus far, the government has continued to maintain ultimate responsibility for determining auction volumes and cut-off rates in the primary market, often to the detriment of monetary management. No explicit coordinating arrangements have been established to ensure that monetary policy considerations are included in determining the terms of primary market issues.

In the other countries, the introduction of open market-type operations in government paper was supported explicitly by arrangements aimed at operational independence for monetary policy. In Kenya, the central bank had been actively involved in the management of government securities markets since 1985.8 The central bank conducted weekly treasury bill auctions and was fully responsible for issuing and redeeming treasury bills. At the onset of the move to indirect monetary policies in 1990, the Central Bank of Kenya was also given responsibility for determining the volume and terms of primary issues of government securities, in line with monetary policy requirements, thereby receiving a high degree of monetary policy autonomy.

In The Gambia, the central bank began to rely actively on treasury bill auctions to manage cash reserves in 1990.9 As the government budget was close to balance, the central bank was given full autonomy to adjust auction volumes in line with its monetary policy requirements. The proceeds of the sales are credited to the government’s account with the central bank and the latter is not required to pay any interest on this account.

Open market interventions in Israel also take place in primary markets for government securities.10 The Bank of Israel uses a specifically designed government security—the Government Short-Term Loan, which is issued at weekly tenders and available in 6-month and 12-month maturities. The central bank has full autonomy in deciding on the volumes to be issued within a limit set by law. The proceeds from the sales of this paper are not used to finance the government deficit. Instead, they are deposited in a government account at the central bank.11

The Central Bank of Mexico began open market-type operations through primary issues of government securities (CETES) in 1983-84, at a time when direct controls were still in place. In 1989, the central bank abandoned its direct controls and started open market operations in secondary markets for government securities in 1990.

Of the countries mentioned above, Mexico is the only one that has moved to genuine open market operations in the secondary market. The other central banks are still using primary issue of government securities as their main intervention tool, along with other instruments, such as reserve requirements and central bank credit to the banks.12

Countries Later Switching to Government Securities

The Philippines, Sri Lanka, Poland, and Nepal are examples of countries where, early in the transition to indirect instruments, the central bank expressed a preference for using government securities in developing its open market operations. The government, however, showed some reluctance in allowing the central bank the use of its debt instruments, thus forcing the central bank to issue its own securities in an attempt to keep up the transition to indirect instruments of monetary policy. Later in the process, the governments agreed to the use of government securities for monetary management purposes and the issuance of central bank securities was discontinued. Nepal is a special case among these countries because, while treasury bills are the main instrument, the central bank also intervenes with its own securities to fine-tune the interventions.

The Philippines has tried several arrangements since the early 1970s.13 At the time of the move to indirect monetary management (early 1980s), the volume of primary issues of government securities continued to be determined solely on the basis of the government’s borrowing requirements, leaving little room for the central bank to use these issues for monetary policy purposes; secondary government securities markets were not developed at the time. In light of these shortcomings, the central bank was forced to issue its own debt instrument in 1983 to improve monetary control.

Over time, however, there were complaints that competition with central bank securities pushed up interest rates in the government securities market, unnecessarily complicating domestic public debt management. At the same time, high interest costs associated with central bank securities were contributing to central bank losses, complicating domestic liquidity management.

Given the above complications, by the mid-1980s, the government and the central bank agreed to phase out central bank securities and use the issues of government securities for monetary management. To ensure that primary issues of government securities satisfied both fiscal and monetary objectives, a coordinating committee was established to determine the auction volumes of government securities on the basis of both deficit financing requirements and monetary policy considerations. Even then, the central bank was still required on occasion to issue its own securities to fine-tune its monetary management.

In 1987, an additional and important arrangement was established between fiscal and monetary authorities allowing for the possibility of overfunding the fiscal deficit on account of monetary policy considerations. Proceeds from any excess sales of treasury bills are now placed in fixed-term deposits with the central bank. Remuneration of the deposits is determined during periodic meetings between the Ministry of Finance and the central bank.14

The Central Bank of Sri Lanka had to issue its own securities in its first attempt at indirect monetary operations because the government was not prepared in 1984 to develop a treasury bill market.15 In 1987 the authorities started concentrating on the development of a treasury bill market. Weekly auctions of treasury bills have since replaced the use of central bank bills in the conduct of open market-type operations. Despite the existence of a joint forecasting committee, auction volumes were for a long time largely based on volumes of maturing bills, thereby subordinating monetary policy considerations to debt management requirements. In 1992 the authorities agreed on new operating procedures whereby the rates at auction are left to market forces and the quantities of treasury bills sold at auction are set according to a reserve money program, thereby greatly enhancing the central bank’s operational autonomy.

In Poland, at the time of the introduction of indirect instruments of monetary policy in 1990, the ministry of finance was not prepared to allow the central bank to intervene in primary government securities markets for monetary operations, prompting the central bank to issue its own debt instruments.16 The success of central bank securities and their direct competition with treasury bills, however, eventually led the government to agree in 1991 on the central bank’s using the primary treasury bill market for monetary management. Consequently, central bank bills were phased out. However, coordination between the central bank and the treasury never proved solid enough for the central bank to meet its monetary management needs. The central bank therefore started, as soon as it was feasible, using repurchase operations as its main indirect intervention tool.

In Nepal, an auction system for treasury bills was introduced in 1988.17 However, this system did not assist the central bank in its transition to indirect monetary policies because the terms of treasury bill auctions (both volume and cut-off rates) were based solely on deficit financing requirements and, initially, treasury bill auctions were held only infrequently. Therefore, the central bank started issuing its own securities in 1991 to gain control over monetary developments. Eventually, treasury bill auctions became more regular, and central bank bills are now issued, with the same terms and characteristics as treasury bills, in side-by-side auctions, with the central bank adjusting the combined volume at issue in line with monetary policy objectives. While both central bank and treasury bills carry identical terms, the central bank has shown greater tendency than has the government to allow rates on its bills to move upward, and this primary rate differential has hampered market development.

Operations with Central Bank Securities

Countries Later Switching to Central Bank Securities

Changes from government securities as the underlying intervention instrument for open market operations to central bank securities have mainly resulted from growing reluctance on the part of the government to further share its debt instruments with the central bank for monetary management, either because (1) the cost of the arrangement was deemed too high for the government (Mauritius), (2) a favorable fiscal situation reduced the need for issuing government debt instruments (Ghana, Chile), or (3) both government and central bank wanted a clear separation between monetary and debt management (New Zealand).

Starting in 1988, the Bank of Mauritius began to remove its direct monetary instruments.18 At the same time, the banking system started facing significant liquidity injections through the external accounts. The government began to issue on tap short-term, fixed-price government securities to sterilize this excess liquidity. The growing interest cost of this operation became a major concern and in 1991, the government announced its intentions to discontinue the issue of treasury bills for liquidity sterilization purposes. Thus, beginning in July 1991, the Bank of Mauritius started issuing its own bills on tap and moved to auctions of these bills in November of the same year. Since then, the central bank decides on the amounts to be auctioned in close consultation with the treasury. When the treasury needs cash, treasury bills are issued and at other times, Bank of Mauritius bills are issued. Both securities have the same characteristics—so that there is in fact only one market—but they are never issued simultaneously.

In Ghana, the introduction of indirect monetary policies during 1988-89 took place against a background of excess domestic liquidity, forcing the central bank to intervene quickly and in vast amounts.19 The concern to develop financial markets, combined with the reluctance of the government to issue treasury bills in sufficient amounts to absorb the excess liquidity, prompted the central bank to issue its own debt instruments. Over time, central bank securities became the main instrument of monetary policy, primarily because of rigidities in the weekly treasury bill auction procedures. Furthermore, when fiscal conditions improved, the government became increasingly reluctant to issue treasury bills, forcing increased reliance on central bank securities for monetary management. To stimulate market development in general, the central bank broadened both the maturity range of its own bills and the eligible holders. The concern to stimulate financial markets is further reflected in the fact that central bank bills have exactly the same features as treasury bills.

The experience of Chile bears some resemblance to that of Ghana.20 As early as 1975 the Central Bank of Chile moved to indirect monetary policy as part of broader financial reforms. The central bank started auctions of central bank promissory notes and treasury bills. Throughout the years the emphasis increasingly shifted to central bank paper. This move was accelerated by the fact that, as part of the financial sector rescue operation in the early 1980s, the Central Bank of Chile bought bad loans from the banks in exchange for central bank bills. By the end of the 1980s when the financial situation returned to normal, the government lost its interest in issuing government paper because of the favorable fiscal position, so the central bank had to rely entirely on its own securities for policy interventions and for market development. To foster markets, the central bank broadened the range of maturities of its promissory notes, which are now up to ten years. As in Ghana, these operations have had an unfavorable impact on the central bank’s profit and loss account.

Since 1988, Reserve Bank of New Zealand (RBNZ) bills have played a major part in the conduct of monetary policy in New Zealand.21 The RBNZ targets the quantity of primary liquidity (defined as the sum of settlement cash and discountable RBNZ bills) by limiting the quantity of RBNZ bills available to the banks and targeting settlement cash.22 Between 1985 (start of financial reform) and 1988, treasury bills were used instead of RBNZ bills. However, lack of central bank control over the supply of treasury bills in conjunction with the government’s desire to achieve a clear separation between monetary policy and the government’s debt management and banking operations led to the introduction of RBNZ bills in 1988.

Under the present system, a change in primary liquidity—reflecting a change in monetary policy stance—is initiated by the RBNZ’s offering of a greater or smaller quantity of its own bills. These bills are issued solely to commercial banks through twice-weekly auctions. Secondary market interventions by the RBNZ take place only when settlement cash injections are required. A variety of techniques, including securitized, short-term loans (or sellbacks), repurchases of government bonds and, occasionally, outright purchases of government bonds or commercial paper, are used to achieve this aim.

On the other hand, when the RBNZ intervenes to neutralize the impact of government cash flows on reserves (as opposed to changing the stance of monetary policy), it intervenes with a specific kind of treasury bill (“seasonal treasury bills”), solely designed for these purposes. The central bank autonomously determines the terms of these securities, taking into account forecasted liquidity flows.23 It was thought that the use of a different type of security would improve the transparency of monetary policy actions and thus enable the market to clearly distinguish between neutralization operations and changes in the stance of monetary policy.

The features, terms, and purpose of central bank securities differ from those of government securities. Furthermore, in its role as the government’s advisor in domestic debt management, the central bank makes recommendations on the structure, amount, and timing of issues and the coupon rate of government bonds, thereby ensuring consistency with monetary policy and reducing the potential for interference of debt management with monetary policy purposes to a minimum.

Countries Always Using Central Bank Securities

Central banks in Costa Rica, Indonesia, and Korea, for various reasons, have always used central bank securities as the intervention instrument for open market-type operations. Bank Indonesia, in addition, also uses private paper to enhance the flexibility of its interventions.

In Costa Rica, the central bank’s main monetary instrument is primary issues of its own securities.24 By lack of coordinating arrangements, the central bank and the treasury have been competing to place their respective securities and have experienced problems placing the desired amounts.

Government securities are not available in Indonesia, since, according to the “Balanced Budget” rule, the government is effectively prohibited from issuing any domestic debt instrument.25 Therefore, since 1983, at the initiation of the transition to indirect instruments, central bank securities or SBIs (Sertificat Bank Indonesia) have been the central bank’s main open market instrument. Because SBIs mainly served liquidity draining purposes and could not be effectively used to inject liquidity, the central bank also created a market for SBPUs (Surat Berhaga Pasar Uang—money market securities) to inject liquidity.26 These SBPU instruments can be rediscounted at the initiative of commercial banks.27 Issue procedures of SBIs and maturities of both SBIs and SBPUs have been undergoing frequent adjustments to improve their effectiveness, depending on the prevailing macroeconomic monetary conditions.

In Korea, central bank securities called Monetary Stabilization Bonds (MSBs) have been used since 1961 to deal with massive liquidity absorption.28 Treasury bills were introduced a few years later and, officially, were also available for monetary policy purposes. For a long time, however, the central bank continued to rely on direct credit controls as its main monetary policy instrument, supplemented with MSBs for absorbing excess liquidity. The dramatic expansion of the market for MSBs, needed to absorb the liquidity overhang, at times hampered expansion of other money markets by draining away funds. More recently, steps are being taken to revise the terms and operating procedures of treasury bills to allow government securities markets to become the principal vehicle for a move to full indirect monetary operations in Korea.

Evaluation and Lessons

Table 1 summarizes the survey of country experiences. The following observations pertain to these experiences.

Table 1.Selected Country Experiences with the Use of Treasury Bills and Central Bank Bills in Conducting Open Market-Type Operations
CountryStarting Date of Open Market OperationsFinancial Instrument UsedIntervention MarketSpecifics of Coordinating ArrangementsIf Central Bank Bills Are Used, Do Government Securities Markets Exist?Is Secondary Market in Intervention Instrument Developing?Date Direct instruments Were DiscontinuedRemarks
Kenya1990Treasury billsPrimaryCentral bank has large autonomy in the auction as the government’s debt manager.Slowly developingStill usedTreasury tries to influence primary market (particularly rates et issue).
Pakistan1991Treasury billsPrimaryNone. Government decides autonomously on auction.Some activityStill usedGovenment interference in auctions influences secondary market activity.
The Gambia1990Treasury billsPrimaryCentral bank has full autonomy to adjust auctions in line with monetary requirements; proceeds of auctions go to account at central bank.No1990Government budget close to balance.
IsraelEarly 1970sSpecial type of government securityPrimaryCentral bank has full autonomy over this instrument, which is not used to finance government deficit. Proceeds are deposited in special account at central bank.No1985
Mexico1984Treasury billsPrimaryYes1989
1990Treasury billsSecondary
Poland1990Central bank billsPrimaryNoneYes1993Success of central bank bills instigated government to allow use of treasury bills for monetary management.
1991Treasury billsPrimaryNoneYesCentral bank switch in 1992 to repos (based on treasury bills) as coordination for open market-type operations was never successful.
Philippines1983Central bank billsPrimaryNoneYes1981Complaints of authorities that the rate on government securities had to be too high to compete with central bank paper.
1965Treasury billsPrimaryCoordinating committee decides on volumes.Central bank had to fine-tune with central bank bills.
1987Treasury billsPrimaryThe above, plus proceeds of excess sales of treasury bills on account of monetary policy are to be placed on deposit at central bank.Secondary market activity developingCentral bank sometimes still intervenes with own securities in small amounts.
Sri Lanka1984Central bank billsPrimaryNo1987Government not interested in development of treasury bills.
1987Treasury billsPrimaryJoint forecasting committee but volumes mainly based on maturing volumes.
1992Treasury billsPrimaryVolumes based on monetary program.
Nepal1988Treasury billsPrimaryNone. Deficit financing needs prevailed.
1991Central bank billsPrimaryYes
1992Treasury bills and central bank billsPrimaryCentral bank bills and treasury bills have identical features, issued in side-by-side auctions. Central bank bills are used to adjust volumes at issue for monetary policy purposes.YesRates are not identical, tending to hamper market development.
Mauritius1988Treasury billsPrimaryMassive intervention needed to sterilize capital inflows.
1991Central bank billsPrimaryFirst on tap; later auctions; amounts to be issued are decided in consultation between treasury and central bank.YesTreasury no longer willing to issue treasury bills for monetary management; both securities have same characteristics but are not issued in same auction.
Ghana1988Treasury bills and central bank billsPrimaryCentral bank bills have same characteristics as government securities.Yes1989Central bank bills issued in several maturities to facilitate transmission of monetary policy and stimulate financial market development.
Chile1975Treasury bills and central bank billsPrimaryNoneInitially yes. In late 1980s disappearing due to favorable fiscal position.1976 (reintroduced during banking crisis 1982–87)Central bank bills also massively used to rescue financial sector in early 1980s.
New Zealand1985Treasury billsPrimarySome activity1985RBNZ was not satisfied with degree of autonomy for monetary management.
1988Central bank billsPrimary (Secondary)Reserve Bank of New Zealand (RBNZ) bills are main intervention instrument for monetary policy. RBNZ has an active role as the government’s debt manager. Special treasury bills are used to neutralize impact of government cash flows on reserve money.YesSome activityBoth RBNZ and government wanted clear separation between monetary and debt management.
Costa RicaSince late 1970sCentral bank billsPrimaryNo coordinating arrangements.YesBoth authorities are competing and face problems placing desired amounts.
Indonesia1983Sertifcat Bank Indonesia (SBIs)PrimaryNo coordination necessary.No1988SBIs only worked one way (absorption of liquidity). Bank Indonesia therefore introduced Surat Berhaga Pasar Uang (SBPUs) to inject liquidity. SBI market not really integrated in rest of money market, making transmission of monetary policy difficult.
Korea1961Monetary Stabilization Bonds (MSBs)PrimaryNoneYesLimited activityStill usedMSBs issued mainly at below market rates to captive markets.

Instruments for Policy Intervention

With respect to the type of financial instrument used for policy intervention—government or central bank securities or a combination of both—there is a greater variety of approaches than in the industrial countries. In industrial countries, there has been a certain convergence in both instruments and techniques used for open market interventions during the 1980s. Newly liberalizing countries, on the contrary, have tested a variety of options under specific local and historical conditions, and central bank securities have been tested and are used more frequently than in financially advanced countries. Reasons therefore are: (1) the lack of government securities markets at the onset of the transition, when the central bank urgently needed a financial instrument for its open market operations;29 (2) unwillingness of the government to establish such markets or to continue its support for the central bank’s use of government securities in monetary operations; (3) shortcomings in previous supporting arrangements that allowed the central bank to use government securities for its open market-type operations; and (4) a deliberate choice of the authorities to separate monetary policy from debt management as was the case in New Zealand (and Fiji).

Supporting Arrangements

There is also variety in the types of supporting arrangements between monetary and fiscal authorities, ranging from good arrangements to very weak or nonexistent arrangements. It becomes clear from the overview that, even more than the type of financial instrument used for open market operations, the quality of the supporting arrangement seems to have a bearing on the success and smoothness of the transition to indirect monetary policy. In several countries in the analyzed sample, full transition to indirect instruments has been delayed owing to the lack of good supporting arrangements. In some of them direct instruments have not even been abolished completely.

Development of Secondary Markets

Irrespective of the financial instrument used for open market operations and the supporting arrangement in place, the development of active secondary markets in the intervention instrument in general seems to be a slow process. Of the countries surveyed, only the Central Bank of Mexico has moved to genuine open market operations in the secondary market, while the Central Bank of Poland uses repos and reverse repos. In a few other countries, secondary market activities have begun to take place sporadically. A host of other factors that fall outside the scope of this paper are also responsible for this lack of development, but the quality of the initial coordinating arrangements definitely plays a part.

Government Securities Market

Finally, government securities market development is in general the least advanced in those countries where central bank securities play the leading role in the development of open market operations (with the exception of New Zealand). Quite often, the central bank is forced to introduce its own securities, because underlying the government’s unwillingness to allow the central bank the use of government securities, there is a reluctance to move to market-based debt management techniques, for fear of higher funding costs.

This section draws some general lessons from the country experiences regarding the above observations. More particularly, two questions will be addressed. First, whether the nature of the underlying financial instrument as such (government or central bank securities) matters in the adoption of indirect instruments. The second concerns the basic requirements for supporting arrangements, to achieve successful coordination of debt and monetary management.

Does the Financial Instrument Matter?

In light of the above observations, the question to be addressed is whether the nature of the instrument used for open market intervention matters. That is, for a country embarking on financial reform, does it make a difference in the medium or long term whether treasury or central bank securities are used for monetary policy interventions?

General Observations

In theory, the origin of the financial instrument used for market development and policy intervention seems irrelevant. What matter are the characteristics of the instrument. Irrespective of their origin, these instruments should be designed in such a way that they foster the development of a free, well-functioning market in this instrument; this, in turn, would facilitate the development of money markets in general and market-based monetary policy operations in particular.

Important ingredients in this respect are:30 (1) interest rates should be freely determined by the market; (2) holdership should be defined as broadly as possible to stimulate competition as more competition is likely to improve the process of interest rate setting, and broad holdership will also facilitate the transmission of monetary policy impulses; (3) the maturities of the paper should be such that they stimulate trading; (4) transfer of ownership of the instrument must be easy; (5) taxation should be simple and transparent to stimulate trading and holding;31 and (6) proper rules for rediscounting the financial instrument should be established. If access to the central bank’s rediscount window is automatic and easy, secondary market development will be hampered.

The above key features can in principle be adopted for both government and central bank paper. So, from this perspective, the choice between both instruments is indifferent. The next question is whether, based on country experience, there are other arguments in favor of or against either of these two instruments.

Central Bank Paper: Specific Arguments

The main advantage of the central bank’s use of its own debt instruments for open market operations is that monetary management will be largely separated from debt management, thereby giving the central bank operational independence from the early stages of the transition onward.

On the other hand, country practices also reveal two types of problems—or limitations—with respect to the use of central bank paper. The first is the risk of central bank losses when solely relying on the issuance of central bank paper to absorb excess liquidity. The second is the lack of spillover to foster money market development in general.

The risk of central bank losses. In economies in transition, one major danger that has been associated with the use of central bank bills is the risk of central bank losses. Given the adverse features common to many economies in transition, including situations of extreme liquidity overhang of a domestic origin, central banks often have to intervene in massive amounts to absorb the excess liquidity. If this is done by exclusively relying on central bank securities, these interventions may eventually affect the institution’s profit and loss account. Cases in point are the Philippines, Ghana, and Chile, where the issuance of central bank paper has contributed to such losses, while in Indonesia and Korea central bank bills have a significant negative impact on the profit and loss account.

Central bank losses lead to two major problems. First, they undermine the central bank’s prestige and autonomy and second, there is a danger that the huge interest payments due on the outstanding debt and, perhaps, the accompanying losses themselves, will undermine monetary policy effectiveness. In other words, they neutralize the very effects sought by the issuance of central bank securities.32

Admittedly, the issuance of central bank securities is just one of several factors that have an impact on central bank profits and may lead to losses.33 Sales of central bank paper reduce central bank profits as much as the use of any other instrument on the central bank’s balance sheet to absorb liquidity from the markets, such as government paper held in the central bank’s portfolio (the exception being required reserves when they are not remunerated at all or only in a small part). Also, if the authorities have agreed on an arrangement whereby the government budget deficit is allowed to be overfunded to meet monetary management requirements, and the central bank pays market-related interest on these government deposits, the impact on the central bank’s profit and loss account will be similar to the case wherein the central bank replaces this overfunding with its own debt instruments.

The risk of central bank losses resulting from the use of central bank securities is greatest in the initial stages of the transition to indirect monetary policies, because of the need to mop up large amounts of excess liquidity that have been building up under the direct monetary control regime. In light of the risk of central bank losses, it would be recommended that any excess liquidity of a permanent nature be absorbed not by entirely relying on short-term central bank paper, but through a combined intervention of increasing reserve requirements and issuing medium-term and long-term government bonds (which would also stimulate market development).

Contribution to market development: mixed results. A second phenomenon is that markets for central bank securities often remain very thin, which may have implications for the flexible use of central bank bills for monetary policy intervention and for financial market development in general. Often, monetary authorities seem to have a tendency to design central bank securities by focusing only on those features that would serve their immediate monetary policy purposes (usually, absorbing excess liquidity). The instrument often receives too much government support (eligibility for reserve requirements, rediscountables, etc.) for a viable market to develop and, in addition, holdings are often restricted to a small group (e.g., banks). As a result, the market never widens and secondary market activity never materializes. Such design features are not conducive to money market development in general.

However, the seriousness of the above implications depends on the particular circumstances under which central bank paper is used. The implications seem most damaging in those cases where no government securities are in circulation. As long as government securities are in circulation (and these securities have the proper market-stimulating features), there is no real problem when a central bank bill market remains small. The torch of financial market development can be carried by government securities and, in addition, the central bank can fully enjoy the benefits of a separation between debt and monetary management to gain operational independence.34

The situation becomes more difficult when no government paper exists. In that case, central bank securities also carry the weight of financial market development and should be designed accordingly. The degree of financial market development will ultimately have an influence on the efficiency and effectiveness of the central bank’s policy interventions.

A first, though perhaps minor, problem in this regard is that, in the absence of a well-developed central bank securities market, liquidity injections through this market are difficult because the market would dry up easily. However, central banks can overcome this limitation to one-way interventions (liquidity absorption) by introducing other instruments to inject liquidity, such as credit auctions (operations that most central banks already conduct). Bank Indonesia, for instance, introduced a separate instrument, the Surat Berhaga Pasar Uang (SBPU), based on private paper, for injecting liquidity.

The second implication seems more serious in the long run. If the intervention market remains underdeveloped, it will often remain a nonintegrated part of the money market and therefore fail to stimulate money market development in general. A major drawback associated with segmented money markets is that transmission of monetary policy impulses through the intervention market will remain imperfect, tending to weaken the impact of the central bank’s policy actions and possibly forcing the authorities to delay abolishing their direct instruments of monetary control.

Two countries where the central bank bill market has (yet) failed to integrate in the rest of the financial market are Indonesia and Korea. These countries have, respectively, nonexistent and underdeveloped government securities markets.35 In both countries, the intervention markets have developed significantly (although secondary market activity is still weak), but they have failed to play a leading role in money market integration because the respective central bank instruments did not bear the desirable features. In both cases, for a long time, interest rates on central bank securities were set by the central banks at lower than market clearing levels. By contrast, both countries have other money market segments that are well developed and function freely (e.g., in Indonesia, the unregulated commercial paper (CP) market has grown independently from and in contrast with the SBI and SBPU market, while in Korea, CP and certificates of deposit (CD) markets also developed successfully (Emery, 1991)).

In Korea, the administratively set interest rate on Monetary Stabilization Bonds (MSBs) has followed market trends with a lag, and as such there has been no spillover to those market segments where rates are market determined (Chart 1A). Despite the relatively low rates, the MSB market grew dramatically in the late 1980s—becoming the largest money market segment in Korea—owing to its largely captive nature (Chart 1B). Its size notwithstanding, secondary market activity is still very low. Both factors constrain the capacity of the central bank to conduct indirect monetary policy.

Chart 1A.Korea: Interest Rates at Issue on Monetary Stabilization Bonds (MSBs) and Corporate Bonds

(Quarterly data, end-1986 to end-1992)

Chart 1B.Korea: Outstanding Balances in the Money Market

(1980–92, end of period)

1Data on repos not available for end-1992.

In Indonesia, financial markets have reached a fairly developed stage but the system is still plagued by several shortcomings. From the beginning, the central bank tried to promote the SBI interest rate as a benchmark rate for the system. However, its rigidity and the lack of stimulus that radiated from the SBI and SBPU markets to the other markets resulted in the foundering of this objective. Chart 2A compares the SBI rate with the interbank and CD rates. Although the chosen periodicity (quarters) does not always result in obtaining a precise idea of the leading/lagging character of the respective interest rates, there are instances (1985-IV, 1986-IV, 1988-III, and 1988-IV) where the SBI rates lagged, rather than led, the money market or where the relationship was weak.

Chart 2A.Indonesia: SBI, CD, and Interbank Interest Rates


Market development in general, and secondary market activity in particular, has been hampered by the very short maturities of SBIs (for which Bank Indonesia was forced to increase the flexibility of its interventions in the absence of a secondary market) and the demand-determined nature of the volumes which led to large fluctuations in the outstanding SBI amounts (Chart 2B). The fact that only banks and nonbank financial institutions could buy directly from Bank Indonesia, and that certain key features such as auction frequency and maturities frequently changed, also constrained the development of the SBI market.

Chart 2B.Indonesia: SBI and SBPU Volumes Outstanding and Interbank Transactions

Some of the above measures also affected the SBPU market. In 1987 Bank Indonesia, in an effort to tighten monetary policy, moved to a system of daily auctions of seven-day SBIs. This, in conjunction with the withdrawal of central bank support, led to a drying up of the SBPU market (Chart 2B). The chart also demonstrates that the creation of the SBPU market in 1985 had a negative impact on the developing rupiah interbank market. The features of SBPUs (short maturities, rediscount facilities, the proceeds of borrowing via SBPU being exempt from reserve requirements) made these securities very attractive. As a result, transactions in the rupiah interbank market diminished significantly after 1985. In 1988, when the SBPU market dried up, the interbank market revived.

The above problems have recently been recognized by the Indonesian and Korean authorities, and in both countries plans have been devised to modify key features of the central bank bills to overcome these problems and improve the effectiveness of indirect monetary policy.

It is worth emphasizing that the above-mentioned shortcomings should not be considered as inherent flaws in the use of central bank bills. What they reveal is that, particularly in those circumstances when no government securities markets exist, the proper design of central bank paper is a crucial element for the reform process to be successful.

Government Securities: Specific Arguments

The main argument in favor of the use of government securities as the underlying instruments for monetary management is their comparative advantage to act as catalysts for financial market development. Typically, government paper—when well designed—is generally accepted and is therefore a very reliable vehicle for financial market development. Usually, short-term instruments will be issued first, in an effort to foster money market development and the maturity spectrum will be extended gradually to encourage capital market development.

The specific advantage of government paper over central bank paper in this regard is the potential of the government to issue large amounts of different maturities, given its typical debt-funding problems. These large volumes will help in deepening markets in a minimum of time. In theory, a central bank could also issue paper of varying maturities, but it is not in the nature of its business and most likely would lead to financial problems, as the examples of Ghana and Chile to some extent illustrate.

When using primary issues of treasury bills, the central bank will accelerate a cumulative development process. The central bank contributes to the deepening of the market, allowing itself to move quickly to secondary market interventions. These interventions will enhance the liquidity of the market and thus stimulate its further development. In addition, the switch to secondary market interventions considerably reduces the need for coordinating arrangements to preserve the central bank’s operational independence.

The country survey indicates that in several cases the cost to the government, resulting from massive issues of government paper to absorb excess liquidity—particularly when these issues are not needed for deficit-funding operations—reduces the government’s enthusiasm for further allowing the central bank to issue government securities for monetary management. In Mauritius, where large capital inflows forced the authorities to conduct massive absorption operations, the treasury no longer allows the central bank to use government securities in its monetary operations. Excess liquidity is also an issue in Kenya and the treasury has shown some reluctance to issue treasury bills in amounts that meet monetary management needs, although it has not changed its position.36 In contrast, in The Gambia, the treasury still allows the use of government securities for open market-type operations, despite the government’s favorable fiscal position.37

The cost-for-the-budget argument as a deterrent for the government to allow the use of government securities in support of open market-type interventions needs to be considered carefully. Absorption of excess liquidity always involves a cost that ultimately needs to be borne by the government (the cost involved being mainly the interest to be paid on it). When government securities are used for the absorption, the cost will be directly visible in the government budget. When central bank securities are used, the operation will affect the central bank’s profits, and in the worst case will lead to losses. Under the typical institutional arrangement whereby the larger part of the central bank’s profits are transferred to the government, the budget will be affected, either through forgone revenue or higher expenditures to bail out the central bank. So, in terms of cost for the government budget, both approaches are equivalent.

The true size of the cost of the absorption operation depends on the origin of the excess liquidity. If the excess liquidity had been building up under direct credit controls, there will inevitably be a cost for the authorities, whether government or central bank securities are used. If, on the other hand, the authorities have to sterilize excess liquidity resulting from capital inflows, the interest cost from issuing treasury or central bank securities will largely be offset by the interest earned on the central bank’s holdings of foreign assets. The net cost to the government will be approximate, whether government or central bank securities are used. With government securities as the intervention instrument, the interest cost to the government budget will be partially compensated by higher central bank profits. When central bank paper is used, the higher interest revenue on foreign assets will partially compensate for the interest paid on those liabilities.

Thus, the cost to the government resulting from the use of treasury bills in open market-type operations is not a valid argument against their use, because ultimately the government will always have to bear the cost of the absorption. On the contrary, reliance on government securities for absorption operations would improve the transparency of the operation. The use of government securities would make overt the cost of the exercise and put pressure on the authorities to address the underlying cost factors, which often are related to excessive fiscal deficits. In addition, the greater potential of government securities to stimulate financial market development is a strong argument in favor of using government securities in large absorption operations. Debt management considerations will dictate the proper maturity mix of these securities at each point in time. Finally, the use of government securities would take away the risk for central bank losses, as discussed earlier.


While in principle the origin of the financial instrument used for open market-type operations does not matter—provided the instrument is endowed with the proper characteristics to stimulate financial market development—it is argued that there is a preference for the use of government securities in monetary management. This preference is primarily motivated by the fact that government securities are better placed to play a catalyst role in financial market development and that their use increases the transparency of the operation.

The use of government securities in monetary management will assist in establishing a liquid and deep market in the intervention instrument. The expansion of the government securities markets (toward the longer end of the maturity spectrum) will also improve the transmission of policy impulses and thus the effectiveness of monetary policy. Also, by not using a separate instrument for monetary policy intervention, market segmentation and confusion will be avoided in the early stages of financial reform. However, to be successful, the use of government securities does require good coordination, especially in its initial stages. This requirement may make both the central bank and the government decide that it is better to introduce a separate instrument for monetary management. However, the introduction of a separate instrument does not relieve the authorities from coordinating their respective operations, as will be illustrated and argued in the next section.

On those rare occasions where a government securities market is not established and no plans exist to develop such a market, the central bank debt instruments should be carefully designed so that they can contribute to financial market development in general.

Finally, siphoning off of large amounts of excess liquidity in the initial stages of financial reform should ideally take place before indirect instruments are introduced, through the issuance of certain medium-term to long-term instruments or a combination of monetary policy instruments. Exclusive reliance on central bank securities is bound to lead to financial problems for the central bank under such circumstances.

The Search for Solid Supporting Arrangements

Without underrating other causes of delays, the transition toward an effective indirect instrument framework has often been rendered difficult or interrupted, because of the lack of proper arrangements between the monetary and fiscal authorities to support the use of the selected financial instrument for open market operations.38 The required degree of coordination basically depends upon a combination of two elements: the instrument chosen for monetary intervention and the level of sophistication of the financial markets (i.e., whether the central bank can operate in the secondary market for its open market interventions or has to plan its interventions in the primary markets—open market-type operations).

Focusing on government and central bank paper, five cases can be distinguished, each requiring a different degree of coordination, with case 1 being the arrangement needing the lowest degree of coordination:

Intervention Instruments
Intervention MarketGovernment SecuritiesCentral Bank Securities
Primary MarketHigh (Case 5)Medium (Case 4)1

or none (Case 1)2
Secondary MarketLow (Case 2)Low (Case 3)

Medium, in case government securities co-exist.

None, in the absence of government securities.

Medium, in case government securities co-exist.

None, in the absence of government securities.

Case 1

There is no need for supporting arrangements in those situations where no government securities exist. Coordination between monetary and fiscal authorities is necessary for the impact of daily government cash flows on reserves, but that is not different from any of the other cases. Here, central bank bills are also the main vehicle for financial market development. Hence, proper design of the instrument is crucial.

Case 2

Open market policy through secondary markets for government securities is the most common case in industrial countries with developed securities markets. As the danger for debt management interfering with monetary management is minimized under these circumstances—and thus the central bank’s operational independence largely guaranteed—the requirements for coordination at the operational level are minimal.39

This type of arrangement is still exceptional in countries with developing financial markets because it requires the existence of sufficiently deep and wide securities markets, which can take considerable time to develop. The Central Bank of Mexico is the only one that entirely relies on genuine open market operations. Sporadic secondary market interventions have started in a few other countries.

Case 3

The need for coordination at the operational level is equally low when the central bank conducts open market operations in the secondary market for central bank securities. Although this is a feasible arrangement in theory, to date in the countries under review, a secondary market in central bank bills has not developed sufficiently to allow the central bank to conduct the larger part of its open market operations in that market.

Case 4

A higher degree of coordination at the operational level is required when the central bank uses primary issues of its own debt instruments for monetary management, while treasury bill markets exist simultaneously. Coordinating arrangements between fiscal and monetary authorities are required to avoid mutual interference that would hamper the central bank’s operational independence and delay financial market development. Therefore, the arrangements should ensure that government securities stimulate financial market development and that a viable market develops in the intervention instrument as an integrated part of the money market in general.

The country survey indicates that only a few countries have succeeded in elaborating solid arrangements that make the achievement of the above goals possible. This does not come as a surprise because quite often, failure on the part of the authorities to share primary issues of government securities for monetary and debt management is at the origin of the resort to central bank bills. In the aftermath of these events, both parties also fail to elaborate coordinating arrangements in the new environment, or are under the false impression that they do not need such coordination.

Two problems that have emerged as a result of a lack of coordination are competition between the two authorities to sell the (separately) planned amounts of securities, and segmentation of the markets in the early stages of the reform that hampers the growth of both markets. To achieve a successful co-existence of both types of financial instruments, it is necessary to (1) coordinate the issues of both financial instruments and (2) arrive at an arrangement that avoids market segmentation in the early stages. Depending on the particular situation in each country, these conditions can be met through various arrangements.

For instance, market homogeneity can be preserved if central bank and government securities are endowed with exactly the same features (interest rate determination, maturity, eligible holders). This would be the preferred option when both markets, taken separately, are expected to remain thin for a long time, as it would add some volume to the market. The arrangements in Ghana and Mauritius are along these lines and, thus, meet conditions (1) and (2). In both countries, central bank and treasury instruments have identical characteristics, and the issues are coordinated. In Mauritius, for instance, the treasury and central bank alternate issuance dates, depending on the treasury’s needs for fresh funds.

However, success of the above type of arrangements depends to a large extent on the market’s perception of both instruments. It has turned out that, in some countries, the government has to pay a higher risk premium than the central bank on otherwise similar securities, leading to segmented markets and causing confusion in the market. Such was the case in the Philippines in the early 1980s and Poland in 1990, and it still is the case in Costa Rica. In Nepal, the central bank has a tendency to pay higher interest rates on its bills than the treasury (in side-by-side auctions). As a result of this practice, the markets remain also segmented.40

If the establishment of a homogeneous market proves to be impossible, both authorities should seek to clearly differentiate their instruments from one another.41 In such a situation, the central bank securities should have shorter maturities at issue than the treasury bills, with all the other features of the instruments being identical. Central banks always operate at the short end of the market to achieve the strongest feed-through to interest rates. Therefore, this option would increase the central bank’s intervention flexibility. In terms of remaining maturity, both types of instruments will eventually merge into one market.

The New Zealand arrangements stand out as a unique example. The Reserve Bank of New Zealand’s bill market is limited to a key market of financial institutions but this market is well integrated in the rest of the money market so that the transmission of monetary impulses is ensured. The clear separation between government and Reserve Bank securities and the clear definition of their purposes ensure that there is no confusion in the marketplace that could possibly hamper financial market development. The main difference between New Zealand and most of the other countries discussed here is that the latter often do not have established government securities markets that will ensure financial market development. In the absence of such well-established markets, market segmentation should be avoided.

Case 5

To ensure the central bank’s operational independence, the most elaborate type of coordinating arrangement is needed when the same instrument and market are used for monetary and debt management. The country survey contains a number of typical examples, demonstrating good (or potentially good) arrangements as well as the kind of problems that may arise when arrangements are missing.

To safeguard (or achieve) the central bank’s operational autonomy, the coordinating arrangement supporting the use of primary issues of government paper for both debt and monetary management should contain the following ingredients: (1) the instrument should have the proper characteristics, as discussed earlier, and more particularly, the government should allow the interest rate at auction to clear the market; (2) a mechanism should be established whereby the volume at issue incorporates the monetary management needs; and (3) an arrangement should also be established to “sterilize” the proceeds from paper sold over and above the debt management needs, and to determine the interest payments on these proceeds.

Developments in Pakistan (since 1991), Nepal (before 1992), and Sri Lanka (before 1992) point to problems in terms of monetary policy effectiveness and delays in the transition to indirect policies, when these requirements are not met. A common characteristic of these countries’ procedures was (or is) that the volumes at issue are determined by the government without due regard for monetary management objectives and the interest rates at auction were (are) not at market-clearing levels, owing to government interference.

Recent experience in Pakistan highlights the problems associated with a move to an indirect monetary control framework in the absence of clear coordinating arrangements between fiscal and monetary authorities regarding the use of government paper. From the introduction of the auction system, the volumes offered at issue did not reflect monetary management considerations but in general remained unchanged at each auction. Initially, interest rates at the auctions moved steadily higher, closer to market-clearing levels (although the government used post-auction cut-off rates). In the wake of these developments, limited secondary market activity began (Charts 3A and 3B). Later on, however, there was some hesitation on the part of the treasury to allow interest rates to move higher, which has been reflected in a slowing down of market expansion. In addition, there have been cases where all bids were rejected and the government consistently forced the State Bank of Pakistan to buy the amounts rejected at auction.

Chart 3A.Pakistan: Average Rate at Treasury Bill Auctions

Chart 3B.Pakistan: Trading Volumes in the Secondary Treasury Bill Market

(March 1991—November 1992)

The latter practices have rendered monetary management more complicated than it was before the introduction of the auction system for two reasons. First, the forced purchases by the State Bank significantly add to money creation and introduce additional variability in reserve money. Second, the central bank has to counteract the above-mentioned liquidity fluctuations but has no developed secondary market to execute such transactions.

These distortions have also affected the development of other money market segments.42 In particular, the uncertainty introduced by rejecting bids not only has affected the bidding pattern at the auction and the trading activity in the secondary market (Charts 3A and 3B), but has also introduced a high degree of volatility in the interbank market as a consequence of the large and sudden fluctuations in domestic liquidity (Chart 4).

Chart 4.Pakistan: Auction Results and Interbank Rate Developments

(September 19 to December 9, 1992)

In contrast, the arrangements currently used in the Philippines, The Gambia, and Israel seem to meet the above requirements. Israel has a unique arrangement that bears resemblance to both the use of treasury bills and central bank bills. Kenya has elaborated a potentially good arrangement, but other factors have lately been intervening.

The current arrangements in the Philippines—allowing for overfunding of the deficit—require a high degree of coordination between fiscal and monetary authorities. As part of the arrangement, the government has to commit itself to not using these funds, which in times of emergency financing needs has led to tensions or conflicts.43 Under the current arrangements in the Philippines, the cost for the central bank is slightly smaller than the cost it would incur if it had to issue its own securities to supplement the issues of treasury bills, because the central bank does not pay interest on the entire deposit. In The Gambia, unlike in the Philippines, the treasury bears the entire interest rate cost.

The arrangement in Kenya is different in the sense that the central bank was granted large autonomy as the government’s debt manager. In principle, such an arrangement leads to a large degree of central bank autonomy in monetary management and requires less day-to-day coordination than a Philippine-type arrangement. While most central banks are fiscal agents of the government, the role of debt manager is only rarely given to central banks.44 Yet, the fact that in general central banks are closer to the markets than ministries of finance is a valid argument for giving the central bank broad responsibilities for debt management.

Against the assignment of the debt management function to the central bank stands the argument that this type of function is not really at the core of central banking, and therefore may distract the latter from its main tasks.45 The validity of this argument is a practical matter because, in the circumstances of countries in transition, making the central bank the government’s debt manager may give the central bank exactly the operational autonomy it needs to pursue its main objective. Therefore, as part of reform programs, this type of arrangement deserves more attention. Such an arrangement may still require an additional mechanism to deal with overfunding of the budget.

Finally, an alternative that relies on government securities and yet has the added advantage of largely separating monetary from debt management has been adopted in Israel. The Bank of Israel uses a specific government debt instrument that is not used by the government for debt-funding purposes. This arrangement in fact bears great resemblance to the use of central bank paper, but the cost is directly borne by the government. Such an arrangement also deserves more attention for reform programs because it is fairly easy to implement, gives monetary management its operational autonomy, only marginally interferes with debt management and requires no day-to-day coordination between both authorities.

When comparing the above arrangements, distribution of the cost of the arrangement comes out as the distinguishing factor. Under the arrangements in The Gambia, Israel, and Kenya (and Mauritius before 1992), the government bears the interest cost of the operation. In the Philippines, the cost is shared, with the central bank bearing the larger part of it. However, as discussed earlier, on a consolidated basis, all solutions approximately yield the same results, provided all central bank profits are transferred to the government, and the central bank does not artificially depress profits to dampen the transfer.


The transition to indirect monetary policies, as part of broader financial reforms, causes drastic changes to take place for the monetary and fiscal authorities at the operational level. The interactions between the operations of both agents become more intertwined than under direct controls. For the central bank to achieve operational independence in the execution of its monetary policy, coordination of fiscal activities with its actions becomes highly necessary. One particular area where coordination at the operational level is needed—and which is at the heart of the transition to an indirect instrument framework—is the selection of a financial instrument for the conduct of the central bank’s emerging open market-type operations.

Based on a survey of selected countries, this paper has extracted some general lessons on the use of government securities versus central bank securities in the development of open market operations and on the design of coordinating arrangements between fiscal and monetary authorities to support the use of these instruments in the context of countries that initiate a financial liberalization process. For the central banks of those countries, it is important for the central bank to avail itself of monetary policy instruments that ensure its operational autonomy from the early stages of the reform process onward.

In countries with developed financial markets, government securities are the most common financial instrument in the conduct of the central bank’s open market interventions. These interventions usually take place in secondary markets for these instruments, which reduces the need for coordination between fiscal and monetary authorities at the operational level to a minimum. In newly liberalizing countries, the choice for government securities is not as evident as in financially developed systems. Often, central banks start issuing their own securities as the intervention instrument for open market operations. The country survey also indicates that this part of the reform agenda often appears to be a major stumbling block on the way to genuine indirect policies.

The paper has addressed two questions: first, does the origin of the financial instrument used for open market operations matter (more particularly government securities versus central bank securities)?46 Second, given the financial instrument, what are the necessary features of the coordinating arrangements between fiscal and monetary authorities to support the use of this instrument in open market operations, i.e., to preserve the central bank’s operational autonomy?

With respect to the first question, it was argued that, while in principle the origin of the intervention instrument (government securities or central bank securities) does not matter (as long as these securities are endowed with the proper characteristics that allow establishing a viable market in that particular instrument and that foster financial market development in general), two considerations tip the balance in favor of government securities. First, government securities have greater potential to foster financial market development than central bank securities. Smooth and steady financial development, in turn, will facilitate the central bank’s monetary operations and the transmission of its monetary policy activities. Second, there is the risk of central bank losses when central bank securities need to be issued in large amounts to absorb excess liquidity. Central bank losses will undermine the institution’s autonomy.

On the second issue, country experience indicates that the establishment of solid supporting arrangements is, after all, even more critical to a successful transition to indirect instruments than the choice of the appropriate financial instrument itself. The paper has identified, as the essential ingredients, that (1) when central bank securities circulate alongside government securities, the characteristics of both types of securities should be such that confusion in the marketplace is avoided and ideally, one single market emerges. The issue procedures (tuning, volume) should be coordinated to avoid unnecessary competition between both agencies; and (2) when treasury securities are used for monetary management, the arrangement should provide for coordination of the tender volume, so as to allow the central bank to issue more securities than is strictly necessary for debt management purposes and to decide on mechanisms to bear the (interest) cost of the overfunding of the government’s budget.

Taken together, the theoretical and practical arguments developed in this paper argue in favor of the early development of government securities markets during financial reforms and the establishment of coordinating arrangements between the central bank and the treasury, enabling the central bank to use government securities in developing its open market-type operations. Even though the use of the same instruments and markets in the initial period requires a higher degree of coordination than any other arrangement, the efforts will result in faster and smoother financial market development and transition to genuine open market operations.


    Bank IndonesiaReport for the Financial Year1983/84–1990/91.

    Binhadi“Monetary Policy in Indonesia”(unpublished paper presented at the 18th SEANZA Central Banking CourseBombay1990).

    Binhadi and P.Meek“Implementing Monetary Policy in Indonesia”(unpublished paper presented at the 17th SEANZA Central Banking CourseSydneyOctober/November1988).

    CottarelliCarloLimiting Central Bank Credit to the Government: Theory and PracticeOccasional PaperNo. 110 (Washington: International Monetary FundDecember 1993).

    DuesenberryJamesS. and MalcolmMcPherson“Monetary Management in Sub-Saharan Africa: Key Issues” (Cambridge, Massachusetts: Harvard Institute for International DevelopmentOctober1991).

    Economist28August1993 pp. 16 and 67–68.

    EmeryRobertF.The Money Markets of Developing East Asia (New York: Praeger1991).

    FreedmanC.“Implementation of Monetary Policy,” in Monetary Policy and Market Operations (Sydney: Reserve Bank of Australia1990) pp. 2749.

    GoldsbroughDavid and IqbalZaidi“Monetary Policy in the Philippines During Periods of Financial Crisis and Changes in Exchange Rate Regimes: Targets, Instruments, and the Stability of Money Demand,”IMF Working PaperNo. 89/98 (Washington: International Monetary FundDecember1989).

    HarperIan and JulianPearce“Implementing Monetary Policy in an Era of Budget Surpluses,”Australian Economic ReviewNo. 90 (2nd Quarter 1990) pp. 5365.

    HarrisonIan“Reserve Bank Bills,”Reserve Bank of New Zealand BulletinVol. 51 (September1988) pp. 17376.

    KangMoon-soo“Money Markets and Monetary Policy in Korea,”Working PaperNo. 9020 (Seoul: Korea Development InstituteDecember1990).

    KapurIshan and othersGhana: Adjustment and Growth 1983–91 Occasional PaperNo. 86 (Washington: International Monetary Fund1991).

    KimSe-Jin“A Study of Indirect Monetary Policy Control in Korea” (unpublishedSeoul: Korea Institute of FinanceOctober1992).

    KleinDavid“Financial Deregulation in Israel: Policy and Results,” in Frameworks for Monetary Stability (Washington: International Monetary Fund1994).

    LeiteSergioP.“Implications of Financial Reform for Public Debt Management” (unpublishedWashington: International Monetary Fund1989).

    LeiteSergioP.“Coordinating Public Debt and Monetary Management During Financial Reforms,”IMF Working PaperNo. 92/84 (Washington: International Monetary Fund1992).

    LeoneAlfredo“Institutional and Operational Aspects of Central Bank Losses,”IMF Paper on Policy Analysis and AssessmentNo. 93/14 (Washington: International Monetary FundSeptember1993).

    MacArthurA.“Monetary Operations, Financial Market Development and Central Bank Independence”unpublished paper prepared for IMF Central Banking Seminar (Washington: International Monetary FundNovember1990).

    MeekP.“Government Financing: The Role of the Central Bank,” (unpublished paper presented at the SEANZA Central Banking SeminarSydneyNovember 71988).

    MeekP.“Monetary Policy and Debt Management” (unpublishedWashington: World Bank1989).

    NascimentoJean-Claude“The Crisis in the Financial Sector and the Authorities’ Reaction: The Case of the Philippines,”IMF Working PaperNo. 90/26 (Washington: International Monetary Fund1990).

    QuintynMarc“Managing Monetary Policy Reforms: Lessons from the French Experience,”Quarterly Review Banca Nazionale del Lavoro (September1993) pp. 27598.

    Reserve Bank of Australia“Singapore: Use of Foreign Exchange Swaps as an Instrument of Monetary Policy,”in Pacific Region Central Bank Conference on Domestic Monetary Policy Country Papers1990.

    Reserve Bank of Fiji Research Department“Reserve Bank of Fiji Notes: A New Monetary Policy Instrument,”Pacific Economic BulletinVol. 5 (December1990) pp. 1520.

    Reserve Bank of New ZealandMonetary Policy and the New Zealand Financial System (Wellington: Reserve Bank of New Zealand1992).

    San JoséArmidaS.“Monetary Instruments and the Control of Liquidity in the Philippines: Focus on Open Market Operations,”Journal of Philippine DevelopmentVol. 17 (First Semester 1990) pp 133–46.

    San JoséArmidaS. and C.Polvorosa“CB Bills as Instrument of Open Market Operations,”CB Review (Central Bank of the Philippines) Vol. 36No. 6 (June1984) pp 4–5 23.

    SundararajanVasudevan and LazarosMolho“Financial Reform and Monetary Control in Indonesia,”IMF Working PaperNo. 88/4 (Washington: International Monetary Fund1988).

    SundararajanVasudevan and LazarosMolhoC.McAuliffe and M.Quintyn“Institutional Issues in Coordination of Monetary Policy—A Survey,”Appendix to the Advisory Report on China (1991).

    SundararajanVasudevan and LazarosMolho and others “Coordination of Debt and Monetary Management in Economies in Transition: Issues and Lessons of Experience,”unpublished paper prepared for OECD Workshop on Coordination of Debt and Monetary Management (Paris: Organization for Economic Cooperation and DevelopmentJuly1993).

    TsengWanda and RobertCorkerFinancial Liberalization Money Demand and Monetary Policy in Asian Countries Occasional PaperNo. 84 (Washington: International Monetary FundJuly1991).

The author extends special thanks to Catherine McAuliffe, who participated in the initial stages of the project and provided valuable suggestions and comments throughout the entire project. He is also indebted to V. Sundararajan. who stimulated this project and made helpful comments during the drafting process. In addition, D. Bigman, C. Brown, C. Cottarelli, P. Dattels, D. Dueñas, C. Enoch, D. Hardy, D. Hoelscher, G. Iden, S. P. Leite, A. Leone, H. Mehran, H. Nishikawa, A. Op de Bake and M. Saal provided useful comments. Finally, special thanks go to the authorities of several countries for their cooperation in the author’s questionnaire.

Hereinafter the term “debt management” invariably refers to government domestic debt management.

The analysis is based primarily on a survey conducted in a cross-section of approximately twenty countries at different stages of financial liberalization and development. While not exhaustive, the survey covers a broad spectrum of countries, highlighting features and circumstances that have led to both successful and disappointing experiences in the use of government and central bank securities in the conduct of monetary policy. As used in this paper, the term “country” does not in all cases refer to a territorial entity that is a state as understood by international law and practice; it also covers some territorial entities that are not states but for which statistical data are maintained on a separate and independent basis.

Typically, the goal of financial market development is also shared by the fiscal authorities, as they would want to broaden their funding resources to move to less inflationary types of funding and lower their funding costs. Whereas, for a central bank, the need for market development is immediate and of a technical nature, the fiscal authorities’ eagerness to assist the central bank may depend upon the fiscal stance. Often, governments coping with huge fiscal deficits are not willing to pay market-related interest rates on their debt, thereby slowing down the market development process and the transition to indirect monetary policy instruments.

As a matter of fact, the macroeconomic effects of government transactions occur under both a direct and indirect control regime, but their short-term impact on monetary variables tends to be blunted when direct credit controls are used.

France, for instance, in the period 1985–87, is a good example. In 1985, when the decision was taken to move toward indirect instruments, marketable treasury bills were introduced. By end-1986 the authorities thought the market had the proper characteristics to allow open market operations and all direct controls were discontinued at that time. See also Quintyn (1993).

Indonesia is probably the only developing country where private paper is used in open market operations. The country had a fairly developed private paper market and the central bank has sought a mixed approach whereby both its own debt instruments and private paper are used for open market interventions.

Based on country survey for Pakistan.

Based on country survey for Kenya.

Duesenberry and McPherson (1991). Treasury bill auctions had been held since 1986, but the volumes at issue were not really targeted for monetary management purposes.

Based on country survey for Israel and Klein (1994).

The government can borrow from this account under very specific conditions, and the credit has to be repaid before the end of the fiscal year. Under the present conditions of liquidity shortages, the dominant monetary policy instrument is monetary loans to the banking system (central bank credit) and not so much the above-described liquidity-absorbing tool.

In Kenya, some open market operations through the secondary market have lately developed, facilitated by the securitization of the government’s overdraft with the central bank, which gave the latter a stock of government securities.

Based on country survey for the Philippines, San José and Polvorosa (1984), and San José (1990).

In recent years, it has normally been agreed that some portion of these deposits would not receive interest in order to limit the losses of the central bank.

Based on country survey for Sri Lanka and Tseng and Corker (1991).

Based on country survey for Poland.

Based on country survey for Nepal.

Based on country survey for Mauritius.

Based on country survey for Ghana and Kapur and others (1991).

Based on country survey for Chile.

Based on country survey for New Zealand, Harrison (1988), and Reserve Bank Bulletin (1987 and 1992).

As the Reserve Bank of New Zealand does not allow settlement accounts to go into overdraft, the quantity of discountable bills is important for settlement banks’ cash management.

The government controls the other debt management instruments (government bonds of various maturities and “Regular Maturity Treasury Bills”). If the system is short of cash owing to government cash flows, the RBNZ would offer to inject liquidity via short-dated loans known as sellbacks.

Based on country survey for Costa Rica.

Based on country survey for Indonesia. See also Binhadi (1990), Binhadi and Meek (1988), Sundararajan and Molho (1988), and Tseng and Corker (1991).

SBPUs include promissory notes issued by customers in connection with borrowing from banks or nonbank financial institutions (NBFIs); promissory notes issued by banks and NBFIs in connection with interbank borrowing; trade bills drawn by one party and accepted by another in connection with a specific transaction, where either the drawer or the drawee is a customer of a bank or NBFI; and trade bills drawn by a bank or NBFI customer and accepted by the bank or NBFI in connection with a credit extension to finance a specific transaction.

Rediscounting usually takes place through a market maker which works under instructions of the central bank. However, under certain circumstances banks and NBFIs could also directly rediscount with the central bank.

Based on country survey for Korea, Kang (1990), Emery (1991), and Kim (1992).

Indonesia is a special case in this regard because there is no need for the government to issue its own domestic debt instruments.

Listed are only those characteristics that pertain to the financial instrument. To foster market development, other requirements, discussion of which is beyond the scope of this paper, need also to be fulfilled. These pertain, for example, to the market infrastructure, the legal and regulatory framework, and prudential supervision. For a more in-depth discussion, see for instance Emery (1991).

However, taxation should be the same across all types of financial instruments, including private paper, to avoid any discrimination.

With respect to the treatment of central bank losses, it can be argued that it suffices that the law relating to the central bank clearly stipulates that the government should carry any operating losses the central bank may incur. While this stipulation should in any case be part of the law, the weakness of this argument is that, by the time the government intervenes, part of the damage in terms of loss of central bank integrity and monetary policy effectiveness has already taken place. Thus, because the government has to back the central bank in any event, it is preferable to do this in another way—for instance by agreeing on arrangements regarding the use of government securities for massive liquidity absorption operations.

For a comprehensive overview, see Leone (1993).

The problems that can emerge in these circumstances, and that, in fact, have emerged in some countries (namely, market segmentation and competition between both markets) are discussed in the above section on supporting arrangements.

Similar problems have been experienced in Ghana and Chile, where government securities markets have virtually disappeared throughout the years. In these countries, central banks have tried to foster financial market development by issuing central bank securities in a wide range of maturities.

This reluctance was reflected in the fact that weekly auction volumes remained unchanged from the introduction of indirect monetary policy intervention in 1990 until mid-1993, when massive liquidity absorption became necessary.

In Ghana and Chile the central banks also had to scale down the use of government securities as part of their monetary management, as a result of the diminishing funding needs of the government.

Another reason for failures, which is closely related but outside the scope of this paper, is the lack of a reserve money programming framework, enabling the central bank to inject or withdraw the proper amount of liquidity. Several central banks, even with good coordinating arrangements, are handicapped on this account and therefore cannot make the final leap to indirect policies.

As in every other arrangement, there will be a need for coordination at the policy level—for instance, to avoid the situation where primary market activities (for debt management purposes) interfere with monetary policy actions in the secondary market.

The Central Bank of The Gambia has recently begun to issue central bank bills to fine-tune its monetary management Although these securities are identical to treasury bills, they are trading at a higher premium than the treasury bills, leading to a segmented market.

In response to the above problems, the authorities of the Philippines and Poland agreed to discontinue the use of central bank paper and concentrate on treasury securities. As argued before, this would be the preferred alternative.

Other reasons hampering secondary market development are related to the taxation of government securities and the lack of infrastructure for a secondary market.

Recently, the central bank has resorted to issues of its own securities to fine-tune the interventions via treasury bills, because the government had expressed reluctance to bring the volume of treasury bills at issue fully in line with monetary management requirements.

In several industrial countries, central banks act as advisors in debt management, but only in exceptional cases are central banks fully in charge of debt management. Central banks have far-reaching debt managemenl responsibilities in Australia, Canada, New Zealand, and the United Kingdom. Only the Bank of Italy has received full responsibilities in government debt management.

This argument has recently been raised in the debate on greater independence for the Bank of England (The Economist, 1993).

The case for private paper has not been considered closely because markets in these instruments are usually underdeveloped in countries engaging in financial sector reforms.

    Other Resources Citing This Publication