Monetary policies that do not subsidise banks

Paul De Grauwe is the John Paulson Chair in European Political Economy at the London School of Economics and Political Science, and Yuemei Ji is Associate Professor at University College London

Central banks pay interest on commercial banks’ holdings of cash reserves at the central bank. Thus, recent rate increases imply larger interest payments to commercial banks and loss of revenue for national governments. This column argues that a better policy would be to combine sustained sales of government bonds with higher minimum reserve requirements. This would avoid transferring central bank profits to commercial banks, which essentially amounts to a subsidy paid by the central bank.

The recent increases in the interest rates have important implications for the profits and losses of central banks. Since major central banks pay interest on commercial banks’ holdings of bank reserves (held at the central bank), interest rate increases also lead to larger interest payments by the central banks to these commercial banks1.

Taking the example of the Eurosystem: bank reserves held by credit institutions at the national central banks and the ECB amounted to €4.6 trillion at the end of 2022 (ECB, Statistical Data Warehouse). In December 2022 the remuneration rate on these bank reserves held by commercial banks was raised to 2%.

This means that the Eurosystem will be paying out €92 billion interest to credit institutions during 2023. These interest payments are probably going to be even larger as the ECB has announced further interest rate increases.

One way to give an indication of the size of these interest payments is the following. The extra €92 billion interest payments to the banks means that the Eurosystem will have to reduce its profit transfers to the national governments by €92 billion. This loss of revenue of national governments amounts to 0.75% of euro area GDP and will lead to an increase in the budget deficit of 0.75% of euro area GDP, requiring additional fiscal austerity in the future.

Given that the short-term interest rate is likely to be raised further, the fiscal squeeze will likely reach 1% of euro area GDP in 2023. And this is likely to happen when the euro area enters a recession.

Several issues arise here. First, why should commercial banks be remunerated for holding liquid reserves at the central bank? Second, is this remuneration necessary to conduct monetary policy? Third, do there exist alternative policy procedures that avoid making large interest payments to banks?

Why should commercial banks be remunerated for holding liquid reserves?

Many economists today take it for granted that bank reserves are remunerated. Yet this remuneration is a recent phenomenon. Prior to the start of the euro area in 1999 most European central banks, with the exception of the Bundesbank, did not remunerate banks’ reserve balances.

Under pressure from the Bundesbank, the ECB started this practice in 1999. The Federal Reserve introduced the remuneration of banks’ reserve balances only in 2008. Thus prior to 2000 the general practice was not to remunerate banks’ reserve balances.

This made good sense: commercial banks themselves do not remunerate demand deposits held by their customers. These demand deposits have the same function as bank reserves at the central bank: they provide liquidity for the non-bank sector. These are not remunerated.

It is difficult to find an economic justification why bankers should be paid when they hold liquidity while everybody else should accept not to be remunerated.

The lack of economic foundation for paying interest on banks’ liquid reserves becomes even more striking when considering the following. When the central bank makes interest payments to commercial banks it transfers part of its profits to the banking sector. Central banks make profit (seignorage) because they have obtained a monopoly from the state to create money. The practice of paying interest to commercial banks thus amounts to transferring this monopoly profit to private institutions.

This monopoly profit should in fact be returned to the government that has granted the monopoly rights. It should not be appropriated by the private sector, which has done nothing to earn this profit. The present situation of paying out interest on banks’ reserve balances amounts to a subsidy to banks paid out by the central banks.

The paying of interest on banks’ reserve accounts has another unfortunate consequence. It transforms long-term government debt into a short-term debt. Most of the government bonds held by the Eurosystem (and other major central banks) have been issued at very low interest rates, often even zero or negative. This implies that governments are immune for some time from the interest rate rises.

It is difficult to find an economic justification why bankers should be paid when they hold liquidity while everybody else should accept not to be remunerated

By paying an interest rate of 2% on bank reserves and thus reducing government revenues in the same amount, the central bank now transforms this long-term debt into highly liquid debt forcing an immediate increase in interest payments on the consolidated debt of the government and the central bank. There is no good economic reason why a central bank should do this.

We show the cost for the national treasuries of this operation in Table 1. With the exception of Italy and Portugal, euro area countries had government debts with average interest payments of less that 2%. By the payment of interest on bank reserves is now transformed into a debt with an interest burden of 2%.

When the ECB raises interest rates further in 2023, these governments will be facing even higher interest burdens while their long-term debt on the balance sheet of the Eurosystem is kept unchanged at the low levels shown in Table 1.

Table 1. Average interest rate on government debt, 2021

Note: Data for Greece are not available.

Source: Eurostat

Is the remuneration of bank reserves necessary to conduct monetary policy?

The standard answer of many economists is positive. Here is the conventional argument. Today, there is an oversupply of bank reserves thanks to the large-scale quantitative easing (QE) operations of the past.

There is, in other words, no scarcity of liquidity; on the contrary, there is an abundance. This creates a problem for the central banks when they want to raise the interest rate.

We show this in Figure 1. This represents the demand for reserves (by banks) and the supply (by the central bank). The demand is negatively related to the money market interest rate (interbank rate). The supply is determined by the central bank. The latter increases (reduces) the supply by buying (selling) government bonds.

Figure 1 presents the regime of reserve abundance: the central bank has bought large amounts of government bonds in the past and thereby created excess supply of reserves. As a result, without remuneration of bank reserves the interest rate is stuck at 0% and the central bank cannot raise the interest rate.

In order to raise the interest rate in this reserve abundance regime the central bank can remunerate bank reserves. In the context of the Eurosystem this amounts to raising the interest rate on the deposit accounts held by credit institutions (banks).

In doing so, the demand curve becomes horizontal at the level of the deposit rate, ie. the deposit rate, rD, acts as a floor for the interbank interest rate. The reason is that banks will not be lending in the interbank market at an interest rate below the (risk-free) deposit rate. Given the abundance of bank reserves this is the only way to raise the money market interest rate.

An increase in the interest rate on bank reserves (deposit rate) is then transmitted into an increase of the money market interest rate and to the whole structure of interest rates (Ihrig and Wolla 2020, Baker and Rafter 2022). Today such an increase in the interest rate is necessary to fight inflation.

Therefore, in the present regime of reserve abundance, the only way to raise the interest rate is to remunerate banks’ reserves and to increase this rate of remuneration.

There are other possibilities for the central bank, however, to raise the interest rate without having to transfer its profits to the commercial banks.

Figure 1. Demand and supply of reserves in reserve abundance regime

Note: This is a stylised representation of the market for bank reserves. It does not show the marginal lending rate (in the case of the ECB) which acts as a ceiling and is raised together with the deposit rate.

Alternative policies that avoid making large interest payments to banks

The first method is to sell government bonds (in today’s parlour, quantitative tightening, or QT). This has two effects. First, the sales of government bonds reduce the amount of bank reserves, and therefore the amount of liquidity in the system. We show this in Figure 2.

By selling a sufficient amount of government bonds the supply of reserves shifts to the left until it intersects the demand curve in the downward sloping part. The interbank interest rate is then determined by the intersection point of demand and supply of reserves.

This recreates the situation that existed prior to quantitative easing. This was a regime of reserve scarcity. The central bank would set a target interbank interest rate and would guide the market rate towards this target by manipulating the supply of reserves.

This operating procedure would then determine the interbank rate without the need for the central bank to remunerate bank reserves (see Ihrig and Wolla 2020 for more detail).

Figure 2. Demand and supply of reserves in reserve scarcity regime: No remuneration

The problem with this approach today is that the central banks would have to sell large amounts of government bonds. For example, the ECB today holds €4.9 trillion of bonds (mostly government bonds). This has led to reserve balances of the banking system of €4.4 trillion, 99% of which are reserves in excess of minimum reserve requirements (of 1%).

In order to bring back the supply curve in the range given by the downward sloping part of the demand curve, the ECB would have to sell almost all the government bonds it holds. An operation that would create havoc in government bond markets.

The ECB has announced that it will gradually reduce its holdings of government bonds by not reinvesting in new bonds when old bonds come to maturity. This will lead to a gradual decline of the amount of government bonds on its balance sheet.

It will take many years, however, to reach the point where we are back in the reserve scarcity regime as illustrated in Figure 2. Thus, we will remain in a reserve abundance regime for many years to come. What happens to the interest rates during this period of transition?

In order to answer this question, we analyse the second effect of the sales of government bonds by the central bank. These sales lead to a drop of bond prices and a surge in the yields. Since the central bank holds bonds with different maturities, the whole spectrum of interest rates is pushed up. Precisely what the central bank aims at would depend on its anti-inflationary stance.

The question that arises, however, is whether this can be achieved without remunerating bank reserves. Let us assume that during this process of bond sales, the central bank stops remunerating bank reserves. The problem that will then arise is that banks will have a strong incentive to buy all the bonds sold by the central bank so as to avoid having non-interest-bearing assets on their balance sheet. As a result, the supply of bonds by the central bank will be met by eager buyers.

The upshot is that there will be little downward pressure on bond prices so that bond yields will not, or only barely, increase. The central bank will fail to raise the interest rates and is forced to remunerate bank reserves during the whole process of return to a scarce reserve regime. Is there a way out of this conundrum? The answer is positive. The use of minimum reserve requirements will do the trick.

Central banks could decide today to raise minimum reserve requirements while paying no interest on bank reserves. The ECB has minimum reserve requirements in its toolkit, ie. the statutes of the ECB make it possible to use it as an instrument of monetary policy.

The ECB, however, has chosen not to use this instrument and has kept it constant most of the time. Today it stands at 1%. (The Federal Reserve has abolished minimum reserve requirements). Thus, the ECB could decide to raise minimum reserve requirements so that the excess reserves banks hold today become required reserves on which no interest is paid. What would be the effect on the interest rates?

We show the effect on the interest rate in Figure 3. As a result of the increase in minimum reserve requirements, the demand for reserves shifts horizontally to the right. We are back in the reserve scarcity regime: the interest rate is determined by the intersection of the new demand curve with the unchanged supply curve.

Banks are not remunerated on their bank reserves and the central bank can manipulate the supply of reserves to guide the money market rate by relatively small open market operations. For example, if it wishes to raise the money market rate it can reduce the supply of reserves by relatively small sales of government bonds thereby shifting the supply of reserves to the left.

Note also that banks would now have a larger proportion of their balance sheet in the form of assets that have no return. In order to restore their overall interest spread (the difference between the interest earned on their assets and the interest paid on their liabilities) they would have to increase the interest rate they apply on their loan portfolio.

This would lead to a generalised increase in interest rates. This is exactly what central banks today pursue in their strategy to fight inflation.

At the end of 2022 total assets reported by euro area credit institutions stood at €39.2 trillion. Their reserves held in the form of deposits at the central banks of the Eurosystem amounted to €4.4 trillion. In the limit minimum reserve requirements could be raised to encompass the whole of these bank reserves.

This would imply that 11% of the balance sheet of these credit institutions would be tied up in non-interest-bearing assets. This is a percentage that is not unusual in countries that apply minimum reserve requirements as a policy tool (see the IMF Integrated Macroprudential Policy (iMaPP) Database).

An often-formulated objection to the use of minimum reserve requirements is that these amount to an implicit tax on the banking sector. Thus, minimum reserve requirements introduce a distortion which should be avoided.

The answer is that all taxes introduce distortions. We have to evaluate whether the cost of these distortions is offset by gains. The gains here are double. First, the authorities can eliminate another distortion which is the subsidy that is granted to the banks today. Second, the use of reserve requirements is an additional policy tool of the central bank that can be used to stabilise the economy when reserves are abundant.

Figure 3. Demand and supply of reserves with reserve requirement: No remuneration

A combination of sustained sales of government bonds and minimum reserve requirements would probably be the best policy option. Thus, the central bank would raise minimum reserve requirements to move into the scarce reserve regime as in Figure 3.

It would then start gradually reducing its bond holdings allowing the supply curve to shift to the left. This then also would make it possible for the minimum reserve requirements to be relaxed gradually.

In Figure 3 both the supply and the demand curves would then shift to the left, maintaining a regime of reserve scarcity and allowing the central bank to use its monetary policy tools without subsidising banks.

We conclude that it is perfectly possible for central banks today to raise the interest rates to reduce inflation without having to transfer large parts of their monopoly profits to commercial banks. These profits belong to society as a whole and should be transferred to governments.


1. They also lead to valuation losses of the central banks. To the extent that these losses are realised on government bonds they do not matter as they are compensated by equal gains of the national treasuries that have issued these bonds (see Gali 2020, Muellbauer 2016).


Baker, N and S Rafter (2022), “An international perspective on monetary policy implementation systems”, Bulletin, Reserve Bank of Australia.

Gali, J (2020), “Helicopter money: the time is now”,, 17 March.

Ihrig, J and S Wolla (2020), “Closing the monetary policy curriculum gap”, Federal Reserve, FEDS Notes.

Muellbauer, J (2016), “Helicopter money and fiscal rules”,, 10 June.

This article was originally published on

Translate »