Safe containing euro notes
VoxEU Column

A two-tier system of reserve requirements to fight inflation more fairly and effectively: Answers to our critics

The proposal to introduce a two-tier system of minimum reserve requirements in order to reduce the size of transfers to banks has received some criticism from observers. This column responds to these criticisms. It argues that unremunerated minimum reserve requirements would not be unfair to banks, are unlikely to lead to large-scale displacements of banking activities, and would not create systemic liquidity problems for banks. Furthermore, introducing these requirements would increase the effectiveness of monetary policy to fight inflation, by limiting the positive effect of higher interest rates on bank equity.

As part of its policy of fighting inflation, the ECB and the central banks of the Eurosystem transfer large amounts of money to banks. At this moment this transfer amounts to €140 billion a year. This is almost as much as total yearly spending by the EU which amounts to €168 billion. The latter is the result of an elaborate political decision process; the former has been decided ‘in smoke-filled rooms’ without any political debate. In addition, EU spending is loaded with conditions that recipients have to satisfy, whereas the transfers to banks have no strings attached. Similarly large transfers also exist in the US and the UK.

Such a large transfer of money to bankers raises issues of fairness. We have proposed to introduce a two-tier system of minimum reserve requirements (MRRs) that would allow policymakers to reduce the size of these transfers while at the same time enhancing the effectiveness of its monetary policies in reducing inflation (e.g. De Grauwe and Ji 2023, 2024).

Our proposal has been subject to criticism by several observers, which we believe reflects popular views in the financial sector and may concern policymakers. Four points of criticism have been raised: (1) imposing unremunerated minimum reserves is an unfair tax on banks; (2) this tax will lead to large displacements of bank activities; (3) due to the heterogeneity of banking sectors, our proposal will be felt very differently in different countries; and (4) minimum reserve requirements affect the transmission of monetary policies and may weaken its effectiveness in fighting inflation. In this column, we intend to answer these criticisms.

Imposing minimum reserve requirements is an unfair tax on banks

Bofinger (2023) and McCauley and Pinter (2024a) claim that imposing unremunerated minimum reserve requirements is an unfair tax on banks. The essence of the argument runs as follows. In the context of quantitative easing (QE), banks sold government bonds to the central banks because they expected future increases in the interest rate on bank reserves that they accepted to hold in exchange for the bonds. Thus, if today the central banks were to decide to stop remunerating these bank reserves, they would unfairly ‘tax’ the banks.

The fact that these transfers now take vast proportions is perfectly all right to these authors because this large increase in the interest rate since 2022 was expected by the bankers during 2015-19 when quantitative easing was at its height and conditioned their willingness to sell the bond to the central banks at that time. It would be unfair to the bankers to deprive them of this €140 billion, even if this deprivation were only partial, as we proposed in our contributions.

The problem with this argument is that there is no evidence that, during the periods of quantitative easing (2015-19) and (2020-2021), bankers, or anybody else, were expecting dramatic increases in the interest rate that we have seen since 2022. In Table 1, we notice that, during 2015-2021, yields of the long-term government bonds for most euro area countries were below 1%. For example, the average German government bond yield was 0.03%. This implies that from 2015 to 2021 the forecasts in financial markets of the short-term interest rates for the next ten years were close to zero. Bankers sold the bonds to central banks freely because the central banks offered a high price for these bonds, making these transactions profitable for the banks even under the prevailing expectations that the interest rates would remain low during the duration of these bonds.

Table 1 Ten-year government bond yields 2015-2021 (%)

Table 1 Ten-year government bond yields 2015-2021

Source: Eurostat
Note: Greece is not included as it was not qualified for the QE programme (2015-2019)

The unexpected increase in the interest rate since 2022 therefore created a large windfall profit of €140 billion for banks (on a yearly basis), at the expense of the taxpayers. It is therefore quite misplaced to suggest, as McCauley and Pinter do, that the central banks won twice (‘heads I win, tails you lose’), while the bankers lost in both cases. The bankers sold the bonds to the central banks and made a profit doing so, otherwise they would not have engaged in such a transaction. When the interest rates increased unexpectedly, bankers collected the manna that fell from heaven and won a second time. To argue the opposite without providing any empirical evidence, like the authors do, is surprising. It leads to the equally surprising conclusion that a ‘tax’ on banks would be unfair to the bankers and their shareholders.

But let us accept that our proposal of only remunerating part of the bank reserves is an unfair tax imposed on the banks. Banks routinely do not remunerate the demand deposits held by their customers (except for big holders of these demand deposits). If the non-remuneration of the deposits held by commercial banks at the central bank is an unfair tax, then the non-remuneration of demand deposits issued by banks and held by the non-banking sector is an equally unfair tax. And a larger tax because the size of these demand deposits is larger than the bank reserves.

Why is it unacceptable that central banks ‘tax’ the banks by not remunerating their deposits and is it acceptable that banks ‘tax’ their customers by not paying interest on their demand deposits? In both cases, the services provided are the same. The central banks provide a highly liquid asset to the banks, and the latter provide a highly liquid asset to the non-banking sector. There is a difference, though. The liquid asset provided by the central bank is not only the ultimate liquid asset but also the safest possible one; safer than the demand deposits provided by the banks to the non-banking sector. If anything, the demand deposits should be remunerated more than the bank reserves because they are riskier than bank reserves. Today this is not the case in the euro area.

Minimum reserve requirements and footloose banks

Bofinger (2023) and McCauley and Pinter (2024b) argue that the imposition of unremunerated MRRs would lead to large-scale displacements of banking activities. In particular, euro area banks that would face larger unremunerated MRRs would move the deposits held by their customers to countries with no, or lower, MRRs and perform their lending activities from these countries. This would have dramatic effects on the banking sectors in the euro area.

First, some empirical perspectives. There is a long tradition of the use of MRRs in Europe. Prior to the creation of the euro area, several countries like Germany, France, and Italy used MRRs, sometimes exceeding 10% of deposits. No such terrible displacements of banking activities took place. Today, Switzerland uses a 2.5% MRR (in contrast to the 1% used in the euro area) and one is still waiting for the large displacement effects.

Second, every regulation leads to attempts to evade these. Is this a reason not to impose the regulation? Take the example of minimum capital ratios. Most economists agree that minimum capital ratios are essential for maintaining a stable banking system. But, bankers dislike minimum capital ratios, and therefore also try to evade this regulation. That does not mean that we should abstain from imposing minimum capital ratios. What we should do instead is to design a regulatory system that minimises the evasion. Here is how to do this.

If these displacement effects following the imposition of a two-tier system of MRRs were to occur, the ECB could easily counter these by using an asset-based system of reserve requirements (Schobert and Yu 2014). This would consist in computing minimum reserves as a percent of total bank reserves. Thus, if bank A has total bank reserves of 100 and bank B of 200, the ECB could tell these banks that, say, 20% of these bank reserves are unremunerated MRRs. For bank A this would mean that 20 of their 100 of bank reserves would be MRR and unremunerated, and for bank B this would be 40. No amount of displacement of deposits to London, or elsewhere, would help these banks in reducing their unremunerated MRRs.

Heterogeneity of the banking sector

It has been noted by some observers (Deuber and Zobl 2023, Kwapil 2023, and Standard & Poors 2023) that the use of a two-tier system of reserve requirements in an environment of heterogeneity of the banking sector could create liquidity problems for some banks that have relatively few bank reserves. These would be forced to borrow funds in the interbank market to satisfy the minimum reserves. In this connection, these observers have pointed at Italian banks that could face liquidity difficulties.

We do not think there would be a systemic problem under reasonable MRRs. We show the evidence in Table 2. This presents the minimum required reserves (that today are 1% of outstanding deposits) as a percent of the total reserves of the euro area banks. We observe indeed heterogeneity in the distribution of bank reserves across countries in the euro area. If the MRR were to be raised from 1% to 10% (quite a large increase) all euro area countries (except Malta) should have enough reserves to satisfy the MRR while maintaining some excess reserves.

Take the case of Italy. In 2022, these minimum reserves represented 9.2% of total bank reserves of Italian banks. If the MRRs of outstanding deposits were raised to, say 5%, this would imply that these minimum reserves would represent 46% of the total reserves of Italian banks. The Italian banks would still have 54% of their bank reserves as excess reserves. Hence, we can conclude that as long as the MRRs remain below 10% of outstanding deposits Italian banks would have enough reserves to satisfy these minimum requirements. As long as there are excess reserves in the system as a whole, borrowing liquidity by a few banks to satisfy MRRs does not create a systemic issue.

Table 2 Minimum required reserves as percent of total reserves (at 2022 level)

Table 2 Minimum required reserves as percent of total reserves

Source: ECB, Disaggregated financial statement of the Eurosystem. We use the reserve level of each national banking system in 2022 as the total reserve base.
Note: MRR is defined as the percent of deposits issued by banks that have to be held as required reserves at the respective central banks.

But if it turned out that significant numbers of banks (in Italy or elsewhere) were to experience serious liquidity problems to satisfy MRRs, the ECB could define these MRRs on an asset base as defined in the previous section. In such an asset-based system, banks would be told to keep a given percent of their total bank reserves in the form of unremunerated minimum reserves. All banks would be able to satisfy such a requirement without encountering liquidity problems. An asset-based system would solve both the foot-loose and the heterogeneity problems.

Minimum reserve requirements and the transmission of monetary policies

Clearly, the use of unremunerated minimum reserve requirements will influence the transmission of monetary policies. The question is in which direction this influence goes. Prima facie, one would expect that adding an increase of unremunerated MRRs to an interest rate hike to fight inflation would strengthen the effectiveness of such a policy compared to a policy of just increasing the interest rate. But that is not how some observers see this (e.g. Kwapil 2023). These observers note that an increase in unremunerated MRRs (now 1% in the Eurosystem) could weaken the transmission of monetary policies. The reasoning is as follows. A higher unremunerated MRR raises the margin between loan and deposit rates, leading banks to raise the loan rates and to lower the deposit rates. The former strengthens the monetary transmission towards a reduction of inflation; the latter does the opposite as it leads agents to save less. If the latter effect is larger than the former, unremunerated MRRs may reduce the effectiveness of monetary policy in the fight against inflation. 

This analysis has led to a perception that the use of the unremunerated MRRs has ambiguous effects on the transmission of monetary policies, which in turn has led policymakers to be cautious about the use of unremunerated MRRs. In a letter to the European Parliament, Christine Lagarde, President of the ECB, cautioned the parliamentarians against the use of unremunerated MRRs (Lagarde 2023) and wrote: “ Limiting the remuneration on reserves held in the deposit facility could thus affect the effective transmission of the monetary policy stance,” without specifying whether the monetary policy stance would be reinforced or weakened by the use of unremunerated MRRs. The president of the ECB promised to study the issue.

In fact, there is little ambiguity about how unremunerated MMRs affects the transmission of monetary policies, for at least two reasons. First, if a raise in the unremunerated MRR leads to a decline in the deposit rate, its (positive) effect on aggregate demand is likely to be small compared to the (negative) aggregate demand effect of an increase in the loan rate. Low-liquidity households, typically borrowers, suffer from two effects following an increase in the loan rate. The first is a direct (substitution) effect: a higher interest rate leads them to borrow less. The second is a debt burden effect: an interest rate increase raises their debt burden, which also leads them to reduce their borrowing. Both effects reinforce themselves and lead to a negative effect on aggregate demand. The situation is different on the deposit side. For high-liquidity households, typically creditors, there are also two effects. The decline in the deposit rate leads to a direct (substitution) effect, inducing these creditors to reduce their deposit holdings and to spend more. But at the same time there is an income effect working in the other direction: these creditors experience a decline in disposable income leading them to consume less. Both effects work in opposite direction, weakening the potential positive aggregate demand effect of the deposit channel (Holm et al. 2021) who analyse this more formally). We conclude that the increase in the loan rate and the decline in the deposit rate induced by an increase in unremunerated MRRs is likely to reduce aggregate demand.

Second, when the central banks raise the interest rate while remunerating bank reserves, they increase the transfers to banks, thereby increasing bank profits and improving the banks' equity position. With a higher equity ratio, banks will be more willing to supply loans to households and firms. (For an analysis of this ‘equity effect’ on bank loans see Shin 2015, Gambacorta and Shin 2016, Vanden Heuvel 2002). As a result, the expected negative effect of a rate hike on loans is (partly) offset by the positive equity effect on bank loans when bank reserves are remunerated. The transmission mechanism is made less effective, i.e. increases in the policy rate have a lower effect on the loan supply and ultimately on inflation. Conversely, by not remunerating bank reserves, this ‘perverse’ equity effect is eliminated, and the monetary transmission mechanism is made more effective. This theory has been confirmed empirically by Frick et al. (2023) and De Grauwe and Ji (2024), leading to the conclusion that the use of unremunerated MRRs increases the effectiveness of monetary policies to fight inflation.


The EU struggles to find funding for Ukraine, for the energy transition, and for compensating farmers who are hit by the need to change farming to reduce global warming. The EU spends about €50 billion on Ukraine (2024-27) and €50 billion a year on the farmers. Lots of conditionality is imposed on recipients of these funds. In the meantime, the bankers in the euro area now receive €140 billion in one year, no strings attached. It is surprising that this extraordinary priority given to bankers over farmers and Ukraine remains relatively unnoticed both in political circles and in the media. It is equally surprising to find economists who defend the large transfers to bankers with the argument that the bankers are entitled to these transfers, and that taking away, even only a fraction, would be unfair.

A two-tier system of reserve requirements proposed in De Grauwe and Ji (2023, 2024) reduces these transfers and makes the fight against inflation both fairer and more effective. We have addressed the different points of criticism regarding this proposal in this column. Some of these criticisms make more sense than others, but all can be overcome relatively easily. The obstacles to implementing our proposal, or other proposals that aim at fighting inflation more fairly, are not technical. They have to do with vested interests and the political power these exert.


Bofinger, P (2023), “Banken profitieren zu Recht von den steigenden Zinsen”, Handelsblatt, 2 February.

De Grauwe, P and Y Ji (2023), “Monetary policies with fewer subsidies for banks: A two-tier system of reserve requirements”,, 13 March.

De Grauwe, P and Y Ji (2024), “How to conduct monetary policies. The ECB in the past, present and future”, CEPR Discussion Paper 18801.

Deuber, G and F Zobl (2023), “ECB Minimum Reserves – 10% or 10% less Government Bonds”, SUERF Policy Brief 741, November.

Fricke, D, S Greppmair and K Paludkiewicz (2023), “Excess Reserves and Monetary Policy Tightening”, Discussion Paper, Bundesbank, Frankfurt.

Gambacorta, L and H S Shin (2016), “Why bank capital matters for monetary policy”, BIS Working Paper 558.

Holm, M B, P Paul and A Tischbirek (2021), “The transmission of monetary policy under the microscope”, Journal of Political Economy 129(10): 2861-2904.

Kwapil, C (2023), “A two-tier system of reserve requirements by De Grauwe and Ji(2023): A closer look”, SUERF Policy Brief No 702, October.

Lagarde, C (2023), “Letter to members of the European Parliament”, ECB, 22 September.

McCauley, R and J Pinter (2024a), “Unremunerated reserves in the Eurosystem, part 1: Heads I win, tails you lose”,, 15 January.

McCauley, R and J Pinter (2024b), “Unremunerated reserves in the Eurosystem, part 2: Tax incidence and deposit relocation risks”,, 16 January.

S&P Global (2023), “Eurozone Banks: Higher Reserve Requirements Would Dent Profits And Liquidity”, 30 October.

Schobert, F and L Yu (2014), “The role of reserve requirements: the case of contemporary China and postwar Germany”, in F Rövekamp and H Günther Hilpert (eds), Currency Cooperation in East Asia, Springer International Publishing.

Shin, H S (2015), “On book equity: why it matters for monetary policy”, mimeo.

Van den Heuvel, S (2002), “Does bank capital matter for monetary transmission?”, Economic Policy Review 8(1).