Chapter

III. An Ideal Model

Author(s):
Carlo Cottarelli
Published Date:
September 1993
Share
  • ShareShare
Show Summary Details

A good starting point for discussing the ideal constraints on central bank credit to the government is the “Maastricht model”—which the governments of the EC have proposed to regulate the relations between the European central bank and the national central banks, on the one side, and the (community, national, and local) EC governments, on the other. Such a model presents several of the features of an ideal model of the financial relations between the central bank and the government.

The Maastricht Model: General Features

Table 1 summarizes the constraints on central bank credit to the government included in the Maastricht Treaty. The following features stand out.9

  • Central bank credit to the government is entirely discretionary, that is, the central bank cannot be forced to provide such credit;
  • Only indirect credit (defined as the acquisition of government securities by the central bank from a third party) is permitted; direct credit (through overdraft facilities, advances or purchases of government paper on the primary market) is prohibited;
  • Indirect credit is unconstrained;
  • National parliaments (or any equivalent political entity at the EC level) have no direct involvement in monetary policy;
  • Government deposits at the central bank are unconstrained;
  • The central bank may act as government fiscal agent.

The rationale for the above provisions is discussed below.

Table 1.Constraints on Central Bank Credit to the Government as Embodied in Maastricht Treaty
InstrumentDegree of

Control
Allowed

Maturity
Quantitative

Ceiling
Interest

Rate
Overdraft on current accountProhibitedNot applicable
Fixed-term loans and advancesProhibited
Purchase of securities (primary market)Prohibited
Purchase of securitiesDiscretionaryUnconstrainedNoneUnconstrained
(secondary market)
(excluding repos)
Repurchase agreements (repos)DiscretionaryUnconstrainedNoneUnconstrained
Government deposits at central bankDiscretionary1Not applicableNoneUnconstrained
Source: Amendments to the European Economic Community treaty as agreed in the European Council at Maastricht, December 10, 1991.

According to article 21.2 of the protocol on the statute of the European Central Bank, the latter (as well as national central banks) may act as government fiscal agent, which usually is intended to include the holding of government deposits.

Source: Amendments to the European Economic Community treaty as agreed in the European Council at Maastricht, December 10, 1991.

According to article 21.2 of the protocol on the statute of the European Central Bank, the latter (as well as national central banks) may act as government fiscal agent, which usually is intended to include the holding of government deposits.

All Credit Is Discretionary

The attribution of full discretion in the control of central bank credit is essentially a matter of principle that cannot be justified purely as a measure to avoid excessive monetary creation. Indeed, although it is true that unlimited automatic access to central bank credit would put monetary policy in the government’s hands (Leone, 1991), relatively little and limited automatic access does not seem to be particularly harmful. The central bank could offset the undesired creation of base money by using other available instruments (for example, by selling government or central bank paper in the open market). Moreover, if the prohibition had been concerned purely with the effect on base money creation of such automatic access, changes in government deposits held at the central bank should also have been restricted as they have the same effect on liquidity as do changes in central bank credit (see Section III).

Two factors can explain the absolute prohibition of automatic access to central bank credit, whose inclusion in the Maastricht Treaty was considered important enough to require amendments to the central banking legislation of most EC member countries. First, endowing the central bank with what can be called an unlimited monopoly power (that is, full discretion) in the control of base money enhances its formal independence vis-à-vis the government. It is equivalent to a formal statement that the government should have no direct control of central bank money.10 Seen in this perspective, the prohibition, although of limited relevance when taken in isolation, can be important as part of a model of a central bank with a large amount of independence, a model guaranteed by a number of other provisions, including those on the central bank statutory objectives (article 2 of the statute of the European Central Bank), on its formal independence (article 7), and on the appointment of its management (articles 10-11).

Second, in countries with developed financial markets, the traditional rationale for automatic access to central bank credit no longer holds. Such access has traditionally been defended on the grounds that, without it, the elasticity of government debt and expenditure management would be seriously impaired (Ulrich, 1931). In principle, this argument is valid. The attempt to cover any liquidity shortfall purely by issuing treasury bills may be extremely costly for the government and may create turbulence in monetary and financial markets (Spaventa Commission, 1989). However, in practice, this argument must be judged against the degree of development of monetary and financial markets. The more developed these markets, the less the recourse to central bank credit is justified. First, large treasury bill issues can be absorbed by the market more easily. Second, short-term liquidity services can also be provided by commercial bank deposits and loans. The more organized and efficient bank and interbank markets are, the more easily commercial banks can accommodate the needs of a borrower as large as the government even in the short run.

Finally, there seems to be no valid reason why the principle of discretion should be sacrificed to increase the flexibility of government debt management in countries in which that elasticity could be provided by alternative means.

Only Indirect Credit Is Permitted

The Maastricht Treaty not only bans any form of automatic access to central bank credit, it also prohibits any form of direct credit even if it is voluntary. The central bank can finance the government only indirectly, that is, by acquiring government securities from the market.11

Indirect credit has the advantage of being securitized. This is an advantage (with respect to overdraft facilities or fixed-term loans) because marketable securities can be more easily sold at a later stage, thus making central bank operations more flexible. Indeed, in many countries (such as Italy, Portugal, and Hungary), the practice of extending credit through overdraft facilities has severely reduced the liquidity of central bank portfolios.

Indirect credit, when it takes the form of open market operations, that is, of regular transactions, on a voluntary basis, and possibly through auctions (rather than through bilateral transactions), typically occurs at market clearing rates. The market rate condition, although not sufficient to guarantee that the government is indifferent to borrowing either from the market or from the central bank (see Appendix I), is important. First, it reduces the advantage for the government of borrowing from the central bank. Second, it represents an additional formal statement of the separation between the central bank and the government and puts private and government sectors on an equal footing in their access to financial markets. Third, and most important, if a large share of central bank assets bear lower than market rates, central bank revenues are squeezed possibly to the point of generating losses. Even if in principle all central bank profits had to be transferred to the government, in practice the collection by the central bank of sufficiently high revenues enhances its budgetary autonomy (for example, in determining its operating costs, such as central bank wages), and therefore its actual independence from government.12 A central bank that has to rely on government transfers to balance its budget (even if these transfers are legally required) is certainly in a weaker negotiating position vis-à-vis the government.13

The above arguments do not, however, explain why intervention in the primary market for government paper is prohibited. As long as this intervention occurs through organized procedures (such as regular auctions), purchases on the primary and the secondary markets are equivalent. Therefore, this prohibition must, again, be justified as a matter of principle, that is, as an institutional signal of the separation between the central bank and the government. Whether this principle should be sacrificed on more pragmatic grounds depends on a host of factors. Allowing the central bank to intervene in the primary market for government paper, particularly in the treasury bill market, may have advantages. The treasury bill tender rate in many countries is a key market rate with a direct influence on bank deposit and loan rates, and the central bank may be interested in influencing it directly, particularly in countries in which secondary markets are not well developed. But there are also some drawbacks. For example, the treasury bill tender rate at specific maturities may be used as the reference rate for government floating rate bonds. Allowing the central bank to intervene in the tender may subject it to undue pressure from the government aimed at reducing the cost of servicing the debt by altering the term structure of interest rates (Italy is a case in point). In conclusion, whether the central bank should be allowed to intervene in the primary market should be determined by the above advantages and drawbacks.

Indirect Credit to the Government Is Unconstrained

The choice of excluding indirect credit from the constraints can be justified because, in the countries admitted to stage III of economic and monetary union (EMU), not only will indirect credit occur at market rates, but government and private assets are likely to be close substitutes, and the interest rate paid on government and private securities will therefore be close.14

As shown in Appendix I, under the above conditions, whereas it remains true that the government benefits from a monetary expansion, the source of that expansion becomes irrelevant.15 Constraining credit to the government sector alone would therefore be irrelevant. The only meaningful constraint would be on total domestic credit of the central bank. Whether that credit should be constrained is of course an aspect of the never-ending “rules versus discretion” debate, discussion of which goes beyond the scope of this paper. Following their traditional approach to monetary policy, the EC members opted to avoid rules and to rely on discretion, although in the context of an independent central bank, which, of course, leaves the door open to some government pressure.16 However, in the conditions prevailing in countries similar to those admitted to stage III of EMU, the real issue is not whether credit to the government should be restrained, but whether total domestic assets of the central bank should be restrained.17

In addition, to bolster central bank independence, the central banking law could set a formal constraint on the purpose of all open market operations-stating that they can be performed only “for reasons of monetary policy” (as in Austria), “to regulate the money market” (Germany), or “in execution of monetary and exchange rate policy” (Portugal). A stronger solution would be to prohibit the transfer of seigniorage to the government. Indeed, as shown in Appendix I, as long as credit occurs at true market rates (and there is no interest premium on government paper), such a prohibition would eliminate the above-mentioned revenue motive for putting pressure on the central bank. Even in this case, however, the central bank would not be completely isolated from government pressure aimed at exploiting the inflation-income trade-off (the unemployment motive), nor would it eliminate the revenue motive entirely in the presence of nonindexed government debt that could be eroded by inflation.

Parliament Is Not Directly Involved

The prohibition of direct credit to the government included in the Maastricht Treaty is absolute in the sense that it cannot be overruled by the EC or by national parliaments. Although the need for the central bank to be independent of the government is widely accepted, there is less consensus on whether it should also be independent of parliament, which is the ultimate holder of power (Grove, 1952, and Franzen, 1991).

The ultimate responsibility for policy decisions, including economic policy, certainly belongs to parliament. But the issue is whether parliamentary control of the central bank should be direct or indirect (that is, via the definition of its mandate and the accountability of its managers). Indirect control is more appropriate because the arguments in Section II in favor of central bank independence hold not only with respect to the government but also with respect to parliament, particularly in present-day democracies.18

Assigning to parliament the decision to determine the extent of central bank credit to the government may severely undermine monetary control, particularly when this assignment is routine (for example, when the amount of credit to be extended each year is determined in the budget law, as in Indonesia). It is less worrisome when the parliament is only requested to intervene, in exceptional circumstances, to remove a constraint imposed by the law (as in Italy). However, the best possible option is for parliament to have no direct involvement.19

Government Deposits at the Central Bank Are Unconstrained

With respect to government deposits at the central bank, five issues are relevant. First, should the central bank be forbidden from receiving government deposits at all? Second, if government deposits at the central bank are allowed, should they be constrained? Third, should the central bank be the sole holder of government deposits? Fourth, should the constraint on central bank credit be defined in terms of net outstanding credit (credit minus deposits)? Fifth, should government deposits be remunerated?

On the first question, although prohibiting the central bank from receiving government deposits would enhance the separation between it and the government, such a prohibition would be inconsistent with the role of fiscal agent that the central bank is often asked to play and, for good reason, should continue to play (see below).

On the second question, the Maastricht Treaty has not introduced any constraint on government deposits at the central bank, which reflects the tradition prevailing in almost all EC countries. In principle, this may weaken monetary control, as running down government deposits at the central bank has the same liquidity effect of central bank borrowing. The government could, for example, accumulate liquidity through overfunding when market conditions are favorable and use this liquidity later to oppose a liquidity tightening by the central bank. Of course, the central bank could offset the liquidity created by running down government deposits, but there may be shortrun problems if the initial stock of government deposits is large.

Some central banks have been given some control of government deposits. The Bank of Canada, for example, can transfer government deposits from commercial banks to itself, and vice versa. In Germany government deposits can be held outside the Bundesbank only with its authorization. However, control of the allocation of government deposits (as in Canada and Germany) does not prevent the government from increasing liquidity by raising expenditure. To contain this possibility, the central bank and the Government in Belgium have recently agreed to introduce a ceiling (equal to BF 15 billion) on the amount of government deposits held at the central bank. Such a ceiling, expressed, for example, as a percentage of government revenue in the preceding year, would contain the maximum amount of liquidity that the government could create to finance its deficit.20 As an alternative to a ceiling, the amount of deposits that the government is allowed to withdraw in each period (say, a month) could be limited to a level that the central bank can easily offset with market instruments.21

On the third question, the case for concentrating all government deposits at the central bank is argued, for example, by Kisch and Elkin (1932) in their classic book on central banks (pages 37-39), in which they refer to the danger that large movements in government deposits may destabilize local banks. However, this danger does not seem to be serious in industrial countries, where the interbank market is well organized. Therefore, there seems to be no overriding reason to prevent the government from keeping deposits with commercial banks (on the contrary, it may be necessary if a relatively low ceiling is set on government deposits held at the central bank). Indeed, in about 80 percent of the sample countries considered in Appendix III, the government is allowed to keep deposits outside the central bank (notable exceptions are France, Greece, Japan, and Switzerland).22

On the fourth question, it is sometimes argued (for example, Martínez Méndez, 1993) that the constraints on central bank credit to the government should be imposed on a net basis, that is, be related to the net outstanding amount of central bank credit to the government minus government deposits at the central bank. This does not seem appropriate. Allowing the government to borrow from the central bank up to the outstanding amount of deposits would weaken the separation between government and central bank. Moreover, if government deposits are liquid, such a provision would not really benefit the government and would simply add an unnecessary complication to bookkeeping. Finally, it would allow the government to borrow from the central bank any arbitrarily large amount of money as long as a balancing deposit was created. It is not clear why the government should be given this possibility (which may tempt it to use the deposits created by borrowing).23

On the fifth question, government deposits at the central bank are often unremunerated (including in Germany, Japan, and the United States), usually as a counterpart to the fact that the services of the central bank as fiscal agent, together with the credit granted, are also free of charge. However, as market conditions should apply to all financial relations between the government and the central bank, government deposits at the central bank should also bear market rates.

Central Bank May Act as Government Fiscal Agent

The Maastricht Treaty stipulates that the central bank may act as government fiscal agent but does not specify what is involved in this activity. It is usually intended to cover two tasks: first, receiving and making payments on behalf of the government as government cashier;24 and, second, organizing and promoting the placement of government paper, as well as managing arrangements to handle government debt in book-entry form. Most central banks are involved in one or both activities.25

In principle, acting as fiscal agent may be in conflict with the need to preserve central bank independence. For example, promoting the placement of government paper, interpreted broadly, may involve purchases of government paper only to stabilize its price. Consequently, it would be appropriate to define specifically the tasks that the central bank may perform as fiscal agent and to specify that these are secondary to the primary mandate of the central bank. More specifically, they must be consistent with the constraints set on central bank credit to the government. As far as prohibiting direct credit to the government is concerned, this prohibition implies that the central bank’s obligation as government cashier to honor government payment orders holds only as long as the government account at the central bank is not overdrawn (see Section IV). If direct purchases on the primary market are prohibited, the central bank could acquire government paper directly from the government only on behalf of the government itself.26 As to open market operations on secondary markets, it would be better to avoid a specific commitment for the central bank to stabilize government paper prices, as this commitment may conflict with monetary control.27

Within the above boundaries, the role of the central bank as government fiscal agent should not interfere with its monetary policy tasks and could be permitted. The central bank is a public organization and, within the limits of its statutory mandate, it should cooperate with the government in all its activities. In many countries it has more expertise than government officials in dealing with monetary and financial markets, which justifies its role as manager of government debt issues. Moreover, the information it acquires as government fiscal agent (particularly concerning the timing of government inflows and outflows) may facilitate liquidity management.

According to many central banking statutes (including those of some independent central banks such as the Bundesbank and the Netherlands Bank), the services provided by the central bank as fiscal agent are not remunerated. This is a legacy of the past, and cost transparency requires some form of remuneration to be given for the services provided. The remuneration of these services, although not quantitatively crucial, would be immaterial only if central bank profits were transferred entirely to the government (which is unusual) and the government was committed to cover all central bank losses (which is rare). The remuneration should cover the costs met by the central bank as fiscal agent, and, again, market prices (prices on similar services provided by the banking system) should in principle apply. If the law gives the central bank a monopoly in the provision of these services, the price cannot be decided by the central bank alone, as this would eliminate any incentive to provide the services efficiently, and must be agreed between the central bank and the government.

Summary of Recommendations

Based on the preceding discussion, Table 2 summarizes the main features of the ideal model, which centers on two guidelines. First, all allowed credit is discretionary. Second, only indirect credit is allowed. Both features are to be considered components of a central bank model characterized by a high degree of independence within a mandate to pursue price stability as the primary objective of monetary policy. Consideration could, however, be given to allowing the central bank to purchase government paper on the primary market, as long as the purchases are made at public auctions, with the central bank participating on an equal footing with other market agents.

Table 2.Constraints on Central Bank Credit to the Government: Summary Recommendations for an Ideal Model
InstrumentDegree of

Control
Allowed

Maturity
Quantitative

Ceiling
Interest

Rate
Overdraft on current accountProhibitedNot applicable
Fixed-term loans and advancesProhibited
Purchase of securitiesDiscretionary or Prohibited
(primary market)
Purchase of securitiesDiscretionaryUnconstrainedNoneUnconstrained
(secondary market)
(excluding repos)
Repurchase agreements (repos)DiscretionaryUnconstrainedNoneUnconstrained
Government deposits at central bankDiscretionaryUnconstrainedExpressed as

percentage of

government revenue

in preceding year
Unconstrained

Linked to rate

paid on

government

short-term

paper

A third feature, implicit in the previous two, is that all credit is securitized and at market rates, which guarantees the flexibility of central bank asset management and its profitability. An additional feature that is implicit in the first two is that the central bank should not provide credit guarantees on any component of public debt.

Fourth, and with reference to indirect credit, it has been argued that, for countries similar to those admitted to stage III of EMU, the real choice is not whether indirect credit should be constrained, but whether all domestic sources of base money creation should be constrained. In countries in which the overall guarantees of central bank independence appear sufficient, such a constraint is unnecessary. A formal statement that open market operations should be used only to achieve the targets of monetary policy may, however, be useful.

Fifth, parliamentary control over central bank credit to the government is discouraged (and highly discouraged if it is on a routine basis).

Sixth, although there are no overwhelming reasons to prohibit the central bank from receiving government deposits, the introduction of some constraint on government deposits at the central bank is recommended to enhance the shortrun control of base money.

Finally, the involvement of the central bank as fiscal agent of the government may be appropriate, although it is not considered strictly necessary in countries in which alternative financial intermediaries could perform the same role. The services provided by the central bank, as well as the deposits kept by the government at the central bank, should be remunerated at market rates.

    Other Resources Citing This Publication