VoxEU Column EU institutions EU policies Financial Regulation and Banking

Assessing the effects of regulatory bank levies

In response to the Global Crisis, governments have implemented restructuring and resolution regimes backed by funds financed by bank levies. Bank levies aim to internalise system risk externalities and to provide funding for bank recovery and resolution. This column explores bank levy design by considering the German and European cases. The discussion points to the importance of structured policy evaluations to determine the effects of levies.

Banking crises cause large costs to the economy (Reinhart and Rogoff 2011, 2013). Advanced economies are not immune to these costs. Here, the output losses due to financial crises have, on average, been one third of GDP (Laeven and Valencia 2013).1 Across emerging and advanced economies, unemployment increased on average by seven percentage points in the aftermath of financial crises (Reinhart and Rogoff 2009). Compared to a ‘typical’ cyclical downturn, unemployment tends to be higher during downturns following financial crises, and shocks to employment tend to be more persistent (GCEE 2013).

Often, the resilience of the private sector in dealing with the costs of financial crises has been limited. These costs have frequently been shifted to the public sector through bailouts of bank creditors. In the G20 countries, public debt-to-GDP ratios have increased from 64% before the recent crisis to almost 90% in the year 2012,[2] with public support of distressed banks being one key driver behind this increase. In the EU, recapitalisation and asset relief measures used during the period 2008-2013 amounted to 5% of GDP in 2013 (European Commission 2016). One result has been a detrimental feedback loop between banks and sovereigns (Acharya et al. 2015, Farhi and Tirole 2016). 

In response to the crisis, governments have implemented restructuring and resolution regimes backed by funds financed by bank levies (Gottlieb et al 2012). Bank levies serve two interrelated purposes.

First, like a Pigouvian tax, bank levies attach a price to a bank’s contribution to systemic risk and thus aim at internalising negative externalities. A bank’s contribution to systemic risk depends, most importantly, on its leverage, its size, and its degree of interconnectedness with the financial system (Admati and Hellwig 2013, Laeven et al 2015). Bank levies aim at changing the incentives of a banks’ management and owners such that they take into account their bank’s contribution to systemic risk.

Second, systemically important banks enjoy an implicit guarantee by the government. The government might be inclined to bail out bank creditors because winding up a systemically important bank entails the risk of destabilising the entire financial system. This implicit guarantee creates incentives to become a systemically important bank. If the aim of the regulation is to withdraw implicit guarantees, winding down even systemically important banks must become feasible. The proceeds that are collected through bank levies thus contribute to a fund that provides enough liquidity to initiate the restructuring and resolution of ailing banks and the bail-in of creditors. In this way, resolution funds aim at increasing the credibility of the restructuring and resolution regime (Buch et al. 2015, GCEE 2014). 

The bank levies used in practice are designed as follows. The assessment base (the ‘tax base’) is usually derived from the total liabilities of a bank. Insured deposits and equity are typically excluded because the implicit guarantee is correlated with the importance of uninsured creditors of banks. The remaining liabilities are subject to a tax schedule. In Germany, for example, a progressive tax rate has been applied which varies with variables correlated with systemic risk such as size; in France, taxes have been based on minimum capital requirements; in Hungary, the tax base is derived from the size of banks’ assets (Capelle-Blancard and Havrylchyk 2013).

What have been the effects of bank levies, and what lessons can be learned from previous experience for the recently introduced bank levy at the European level? The remainder of this column discusses the case of the German bank levy and provides some lessons to be learned for the European case.

Bank levies at the national level: The German case

Germany introduced a progressive bank levy in 2011, with the purpose of financing a restructuring fund with a targeted size of €70 billion, which is roughly equal to the public capital support to banks in the period 2008-2013 (€64 billion). Larger banks, banks with a market-based funding strategy, and banks involved in derivatives trading faced a higher marginal levy. ‘Contribution-relevant liabilities’ have been total liabilities minus equity, customer deposits, profit participation rights, and reserve funds for general banking risk. Banks are exempted from the levy if their contribution-relevant liabilities were smaller than or equal to €300 million .

The levy has been capped at 20% of the annual earnings of the accounting year preceding the contribution year, a floor having been set at 5% of the tax payment. The levy has been designed such that a large fraction of smaller banks were exempt from payments. Consequently, the large commercial banks and head institutions of savings banks and credit unions contributed most (Figure 1).

Figure 1. Bank levies paid by German banks

Note: This figure shows the payments of the bank levy (in million euros) by German banks for the years 2011-2015, broken down by banking group. During the years 2011-2014, banks had to contribute to the national restructuring fund as specified in the Restructuring Fund Act. In 2015, banksnd Acyeibutions were for the first time calculated as specified in the Banking Resolution and Restructuring Directive.
Source: Deutscher Bundestag Drucksache 18/5993, BB8/5993he 18 5 December 2015.

Assessing the impact of bank levies on systemic risk is a rather ambitious empirical task. A first step is to assess the short-term impact of levies on banks. Potentially, banks can adjust to bank levies in various ways, depending on the degree of competition on product and factor markets. Banks can try to shift the burden of the levy:

  • to customers by increasing fees or interest rates on loans;
  • to employees by lowering wages and employment; and
  • to creditors and equity holders by changing the composition and the remuneration of their funding structure.

Evidence from Germany shows that the design of bank levies determines both which banks contribute and the amount that can be collected (Buch et al. 2016). Exempting smaller banks and setting an upper limit for contributions reduced the amount collected to around €600 million annually or less than 5% of banks’ profits. This number is small compared to the targeted size of the fund (€70 billion) or compared to bank capital support by the German government.

Moreover, the impact of the levy on bank behaviour has been limited. Empirical research can exploit the fact that banks could not fully anticipate the impact of the levy because it was introduced in 2011, while payments were calculated based on balance sheet components of the preceding year. This helps identifying banks’ short-run responses to the levy (Buch et al. 2016). Banks affected by the levy reduced loan supply in response to the introduction of the levy and compared to unaffected banks. Also, banks tended to increase deposit rates, probably to attract customer deposits that have been exempted from the tax base. Yet, banks did not change their balance sheet composition away from liabilities, having been taxed in order to reduce the tax burden. Overall, the German bank levy had no strong macroeconomic impact on the behaviour of German banks. It has not been set at a high level compared to banks’ profits, and many banks that are relevant for the German credit market have been exempt from the tax due to their small individual size.

Bank levies as part of the Single Resolution Mechanism

What lessons does the German experience have for the impact assessment of the bank levy established in the Eurozone as part of the Single Resolution Fund (SRF)? The Single Resolution Mechanism (SRM), as one pillar of the Banking Union, foresees the build-up of a SRF financed by bank levies. Each country has to contribute 1% of deposits covered, and this amount shall be collected until 2023. The fund has a targeted size of around €55 billion . It is comparatively smaller than the German fund because part of the losses have to be covered by a bail-in of 8% of liabilities, because the contribution of the SRF to an institution is limited to 5%, and because ex post contributions can be charged, increasing the size of the fund to around €75 billion  (Gros and De Groen 2015). The SRF aims to facilitate the restructuring and resolution of banks in distress. In 2015, for the first time, banks were required to pay according to the new design. However, the levy will only be fully applied on a supranational level from the year 2023 onward.3

Similarities and differences in design

The design of the European levy differs along some important dimensions from the German one, which will matter for its effects – both, intended and unintended.

Similar to the German levy, the European levy is composed of a quantity component which comprises banks’ total liabilities, excluding positions like equity and customer deposits. Smaller banks are not excluded but have to make lump sum payments. Even if these banks are unlikely to receive funds from the SRF, they nonetheless benefit if the funds are used to stabilise systemically important banks in distress. For example, a bank with contribution-relevant liabilities close to €200 million and with a balance sheet not exceeding €1 billion has to pay €15,000 annually for the European levy. That same bank would be exempted from the German levy.

Given that the European levy has no upper bounds, larger banks will have to contribute more in absolute terms. Total payments of German banks under the European levy were €1.58 billion in 2015, which is equivalent to 0.5% of banks’ stock of 2013 equity as obtained from the ECB Consolidated Banking Statistics, and thus higher than comparable numbers for the German bank levy (see Figure 1). Contributions across all countries amounted to €4.3 billion in 2015. In 2016, total contributions of €6.4 billion have been transferred to the SRF by the national resolution authorities.4

In contrast to the German levy, the European levy is calculated based on a quality component capturing various risk indicators. According to the Bank Recovery and Resolution Directive, the base tax rate is adjusted by a factor that captures four different ‘risk fields’, including risk exposure, stability and diversification of funding structure, systemic relevance, and other risk factors to be determined by the resolution agency. These four risks are divided into various ‘risk factors’ (European Commission 2014). The risk-adjustment of the European levy is a relative procedure. It compares different risks per bank and assigns a risk factor depending on the place in the risk distribution of a bank relative to all others.

Generally, the European levy is thus more complex than the German one. This complexity complicates an analysis of the effects of the levy, provides greater scope for (tax) evasion, and may have unintended consequences in case the adjustment factor misrepresents relevant risks. In addition, the complexity of the European levy may make it more difficult for banks to anticipate their annual payments. Also, there is scope for national resolution authorities to have some flexibility in the execution because the levy is not tax deductible in all countries. This could cause differential treatments across Eurozone countries.

Implications for policy evaluation

Whether the design features of the European levy help to mitigate systemic risk and internalise externalities to a sufficient degree is, ultimately, an empirical question. Evidence from Germany shows that careful research design can contribute to an improved understanding of how bank behaviour changes in response to regulatory changes.

Data availability and accessibility are crucial in this regard – data need to be sufficiently granular in order to model the effects of the levy and to compare different types of banks according to their ‘treatment’. Also, impact assessments should be replicable in order to check their robustness and sensitivity to specific underlying assumptions. Cross-country data on levy payments, which can be used to evaluate the levy, are available to the SRB;5 the European Commission has been mandated to review the risk factors.

In a broader sense, causal impact assessments are a key element of the accountability of regulators vis-à-vis the general public. Bank levies impose taxes on banks and intend to affect their behaviour. Regulators and the wider public need to know whether or not the intended effects of the bank levy are achieved – that is, whether systemic risks are reduced by limiting externalities arising from the combination of a bank’s ‘riskiness’ and its choice of financing. Bank levies need to set sufficiently strong incentives to internalise systemic risks and to prevent potential future losses to be borne by tax payers as has often been the case in past crises. Without providing evidence of its effectiveness and efficiency, the bank levy – like other regulatory initiatives – runs the risk of being questioned by both the financial industry and the wider public. While industry might stress the costs, the public may question effectiveness in terms of reducing costs to taxpayers.


Acharya, V, I Drechsler and P Schnabl (2015) “A Pyrrhic victory? Bank bailouts and sovereign credit risk”, Journal of Finance, 69(6): 2689-2739.

Admati, A R and M Hellwig (2013) The bankers' new clothes: Whats' wrong with banking and what to do about it, Princeton: Princeton University Press.

Buch, C M, B Hilberg and L Tonzer (2016) “Taxing banks: An evaluation of the German bank levy”, Journal of Banking & Finance, 72: 52 of

Capelle-Blancard, G and O Havrylchyk (2013) “Incidence of bank levy and bank market power”, CEPII Working paper 2013-21.

European Commission (2016) State Aid Scoreboard 2014 > Aid in the context of the financial and economic crisis.

European Commission (2014) Estimates of the application of the proposed methodology for the calculation of contributions to resolution financing arrangements, Accompanying the document Commission Delegated Regulation supplementing Directive 2014/59/EU of the European Parliament and the Council of 15 May 2014 with regard to ex ante contributions to resolution financing arrangements {C(2014) 7674 }; Part 3/3.

Farhi, E and J Tirole (2016) “Deadly embrace: Sovereign and financial balance sheets doom loops”, NBER Working paper no 21843.

Gottlieb, G, A Avanova and G Impavido (2012) “Taxing finance”, Finance and Development, 49(3): 44-47.

Gros, D and W P De Groen (2015) “The Single Resolution Fund: How much is needed”, VoxEU.org, 15 December.

Laeven, L, L Ratnovski and H Tong (2015) “Bank size, capital, and systemic risk: Some international evidence”, Journal of Banking & Finance, forthcoming.

Laeven, L and F Valencia (2013) “Systemic banking crises database”, IMF Economic Review, 61(2): 225-270.

Perotti, E and J Suarez (2009) “Liquidity insurance for systemic crises”, CEPR Policy Insight No 31, February.

Perotti, E and J Suarez (2011) “A Pigovian approach to liquidity regulation”, International Journal of Central Banking, 7(4): 3-41.

Reinhart, C M and K S Rogoff (2009) “The aftermath of financial crises”, American Economic Review, 99(2): 466cono.

Reinhart, C M and K S Rogoff (2011) “From financial crash to debt crisis”, American Economic Review, 101: 1676-1706.

Reinhart, C M and K S Rogoff (2013) “Banking crises: An equal opportunity menace”, Journal of Banking & Finance, 37(11): 4557-4573.


[1] This number is based on the median deviations of GDP from its trend for all crises periods in advanced countries during the years 1970-2011.

[2] See IMF Data Mapper, Historical Public Debt Database.

[3] An overview of legal acts is available at https://srb.europa.eu/en/content/legal-acts

[4] For details, see https://srb.europa.eu/en/content/2016-ex-ante-contributions.

[5] Contributions to the SRF are calculated by the Single Resolution Board and national resolution authorities are in charge of collecting the contributions and transferring them to the SRF.

2,205 Reads