ERB building
VoxEU Column Financial Regulation and Banking Inflation Monetary Policy

Monetary policies with fewer subsidies for banks: A two-tier system of minimum reserve requirements

The ECB subsidises commercial banks massively against inflation, which constitutes a more than €1 trillion transfer of money from taxpayers to private banks over the next ten years. This column argues the subsidies are exorbitant and proposes a two-tier system of non-interest-bearing minimum reserve requirements on part of bank reserves. This will reduce the excessive subsidies to banks while maintaining the current operating procedure used by the central bank. It will also alleviate the burden on taxpayers and avoid making the ECB's operating procedure unsustainable.

In a previous contribution (De Grauwe and Ji 2023), we argued that the massive subsidies given to commercial banks by the ECB (and other central banks) in its fight against inflation are exorbitant and unsustainable. These subsidies constitute a large transfer of money from taxpayers to private banks orchestrated by the central banks. From March 2023, this transfer will amount to €129 billion on a yearly basis in the euro area. In addition, we argued that the ECB could raise the interest rate without having to make these large transfers. It can achieve this by raising minimum reserve requirements. This makes it possible to go back to a reserve scarcity regime that existed prior to the financial crisis, which allows the ECB to raise the interest rate without subsidising banks.

We are aware that this proposal is quite intrusive and is resisted by the banks, which will see an easy source of profit disappear at once. It is likely to be resisted by central banks also because it implies a return to operating procedures that existed in a reserve scarcity regime prior to the financial crisis. The ECB, together with other central banks, now embrace their new operating procedure (arising from the abundant reserve regime), which consists in raising the rate of remuneration on bank reserves as an instrument to increase the market interest rate in their fight against inflation. This has also led to a widespread conviction, one could even use the term ‘dogma’, among central bankers and economists, that this is the only reasonable operating procedure. It is hard to fight dogmas.

It could also be, of course, that central banks make these large transfers to banks out of a concern that the recent interest rate increases make banks more vulnerable. It is indeed the case that when interest rates increase banks’ profits tend to decline, because their liabilities have a shorter maturity than their assets. An interest rate increase is transmitted faster to the liabilities than to the assets, leading to a decline in banks’ profits. By remunerating bank reserves, central banks solve this problem of commercial banks at the expense of their own profits.

If this is the motivation of central banks it is certainly ill-advised. The way to deal with the problem of declining profits when interest rates increase is the regulators and supervisors have to make sure banks have sufficient buffers, like capital and liquidity ratios. Subsidising banks in advance is not the way to deal with this. The large transfers of money from taxpayers to banks are mainly protecting the shareholders of the banks rather than the banks themselves. This should not be part of the central banks’ objectives.

Can one design a system that will avoid having to make massive transfers to banks while maintaining the current operating procedure used by the central banks, and in doing so (hopefully) gaining their backing? We believe it is possible to design such a system. It is a two-tier system.

The two-tier system consists in imposing non-interest-bearing minimum reserve requirements on part of the bank reserves. The bank reserves exceeding the minimum requirement (excess reserves) would then be remunerated as they are today (for similar proposals for a two-tier system, see Whelan  2021 and Buetzer 2022).

In order to show the implications of our proposal, we first replicate Figure 1 from our previous contribution. This figure 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 an 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. The only way to raise the market interest rate is to increase the remuneration on bank reserves, rD,  (shown by the green line). This remuneration rate then acts as a floor for the market interest rate.

Figure 1 Demand and supply of reserves in reserve abundance regime 

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 which acts as a ceiling and is raised together with the deposit rate.

We can now use this figure to analyse the implications of our proposal for a two-tier system. The imposition of minimum reserve requirements leads to a horizontal displacement of the demand curve to the right. The minimum reserve requirement would apply only to part of the total bank reserves (in contrast with our previous contribution, where we assumed that the whole of the existing bank reserves would be transformed into required reserves). We show this in Figure 2. As a result of this partial displacement of the demand curve, we remain in the abundant reserve regime. The central bank then remunerates the excess reserves with the rate rD (the horizontal green line). As before, this rate of remuneration acts as a floor for the market rate, and the central bank can raise the market rate by increasing the interest rate on (excess) bank reserves.

Figure 2 Demand and supply of reserves: two-tier system

Figure 2 Demand and supply of reserves: two-tier system

The advantage of this two-tier system is that the operating procedure so cherished by central bankers can be kept unchanged. The central bank continues to use the interest rate on bank reserves as its monetary policy instrument. However, the transfer of central banks’ profits to commercial banks can be reduced significantly. To give an example: in February 2023, bank reserves in the Eurosystem amounted to €4.3 trillion. The ECB could block 75% of this (€3.2 trillion) in the form of non-remunerated minimum reserves. There would then be €1.1 trillion of excess reserves that are remunerated. The implications for the interest transfers are shown in Table 1 for the year starting in March 2023. We show two scenarios for the interest rate on bank reserves: 3% (the rate that will apply from March 2023 on) and 3.5% the rate that is likely to be applied later in 2023 when the ECB raises the interest rate again. The left-hand part of the table shows what the interest transfers to banks are in these two interest rate scenarios in the present system, where all bank reserves are remunerated. The right-hand part shows the implications of our proposed two-tier system. We observe that in our proposed system, there would be a significant reduction in interest transfers to banks. With the interest rate of 3% that will apply from March on, the Eurosystem would have to transfer only €32 billion instead of €129 billion in the present system. In our two-tier system, the banks would continue to profit: they would receive €32 billion on what is essentially a risk-free asset. This would be less than today, but surely less “exorbitant”. It appears to us that if the central bank cares about the general interest, in particular the interest of the taxpayer, a two-tier system that allows the central bank to maintain its operating procedure intact but that reduces the massive subsidies to banks should be agreeable.

Table 1 Interest transfers in 2023 (bn €)

Table 1 Interest transfers in 2023 (bn €)

Source: own calculations based on ECB Statistical Data Warehouse

This is all the more true since the ECB used a two-tier system in the recent past when it charged a negative interest rate of -0.5% on bank reserves. When the ECB introduced this negative rate, the bankers did not like it. This led the ECB to introduce a two-tier system in 2019 in which the banks were exempted from paying this interest rate on part of their reserve holdings (six times the minimum reserves). As a result, banks paid interest on only a fraction of their bank reserves (Boucinha et al. 2022). The total cost for the banks from the negative interest rate amounted to approximately €17 billion in 2021. This compares to the subsidy of €129 billion that the banks will receive in 2023 in the present system. It would be incomprehensible if the ECB today, when the taxpayers are hurt, is not willing to introduce a similar two-tier system that would alleviate the burden on taxpayers, in the same way as it was willing to alleviate the burden on banks when they were hit by a negative interest rate.

Failure to move to a two-tier system risks making the operating procedure of the ECB unsustainable in the medium run. From March 2023, the ECB will be transferring €129 billion to banks on a yearly basis (see Table 1). The expectation is that to fight inflation, the ECB will further raise the interest rate on deposits, probably to 3.5%. If it does so, the yearly transfer would jump to €150 billion, which is almost equal to the total yearly expenditures of the EU (€168 billion). This would put immense pressure on the ECB to justify why it deems it necessary to transfer so much free money to banks and not to taxpayers in its fight against inflation.

The ECB has announced a gradual decline of its government bond holdings at a pace of €15 billion a month. At this pace, it will take more than 20 years for the Eurosystem to eliminate its government bond holdings from its balance sheet and, in so doing to reduce the supply of bank reserves so as to return to a scarce reserve regime. During this period, it is likely that interest rates will settle into a new normal. What this is, is difficult to predict, but it is not unreasonable to expect that this may be around 3% if the ECB manages to stabilise the rate of inflation at 2%. 

We would like to know how much the ECB will pay out to banks under these assumptions in the future. In order to find out, we assumed, first, that the interest rate settles at 3%, and second, that the Eurosystem continues to shrink its stock of government bonds at the rate of €15 billion a month. Under these assumptions, the Eurosystem will pay out cumulatively more than €1 trillion to banks during the next ten years (see Table 2). Note that our assumption that the interest rate stabilises at 3% implies that the ECB is successful in bringing the rate of inflation back to its target level of 2% rather quickly. If the ECB is not successful, the interest rate is likely to be higher, thereby leading to even higher future transfers to banks.

It is illuminating to compare the size of these future transfers to banks to the size of the grants to be distributed to governments in the context of the multi-annual NextGenerationEU programme. These NextGenerationEU grants amount to €340 billion. It will be remembered that this programme was decided after long and sometimes acrimonious political discussions and that numerous conditions were imposed on governments receiving these grants. In contrast, no political discussion has preceded the decision made by the ECB to hand out grants to banks that are three times the size of the NextGenerationEU grants. In addition, the ECB did not impose any conditions when granting this money to banks.

These massive transfers to banks are substantially reduced in our proposed two-tier system, as can be seen in Table 2. Our proposal would lead to cumulative transfers to banks of ‘only’ €262 billion compared to €1,047 billion if the Eurosystem perseveres in handing over large grants to banks for the next ten years. Thus, while our proposal still leads to substantial transfers and a source of easy profits for banks in the next ten years, it is a dramatic reduction compared to the transfers that are now in the pipeline. In addition, it also implies an even larger gain for taxpayers who will receive a cumulative amount of €785 billion (€1,047 billion – €262 billion) in our proposal. 

The decision of the Eurosystem to hand over to banks a large part of its profits which over the next ten years is likely to exceed one trillion euros, is quite extraordinary. It is also an unreasonable decision considering that these exorbitant transfers can largely be avoided without changing the operating procedure of central banks, as we have argued in this paper.

Table 2 Cumulative interest transfers, 2023-33

Table 2 Cumulative interest transfers, 2023-33

Source: own calculations based on ECB Statistical Data Warehouse.
Notes: * We start with the initial amount of bank reserves observed in February 2023, which amounted to €4.3 trillion, and reduce this amount yearly by €180 billion (15 billion a month). We assume the same 3% on this declining stock up to 10 years into the future.  ** In the 2-tier system, we assume that the gradual decline in government bond holdings by the Eurosystem leads to a gradual relaxation of minimum reserve requirements.


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

Boucinha, M, L Burlon, M Corsi, G Della Valle, J Eisenschmidt, I  Marmara, S Pool, J Schumacher, O Vergote (2022), “Two-tier system for remunerating excess reserve holdings”, ECB Occasional Paper No. 302.

Buetzer, S (2022), “Advancing the Monetary Policy Toolkit through Outright Transfers”, IMF Working Paper No. 22/87.

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

De Grauwe, P and Y Ji (2023), “Monetary policies that do not subsidise banks”,, 9 January.

Whelan, K (2021), “Are Central Banks Storing Up a Future Fiscal Problem?”,  blog.