VoxEU Column Financial Markets Financial Regulation and Banking

Regulatory treatment of sovereign debt: Risk weights vs large exposure limits

Sovereign risk and its treatment by European banks is a frequently debated topic. In particular, regulators are focusing on zero risk weighting and large exposure limits. This column argues that redesigning the macroprudential framework for sovereign risk management will be a key theme in the years to come. Depending on the exact outcome, the structure of the EZ bond market might look very different from its current shape. This could have far-reaching consequences for both the ECB’s monetary policy strategy and investors alike.

Doubts about the appropriateness of a 0% risk weighting of European government bonds (EGBs) are not new. Several officials across Europe have repeatedly addressed this issue in the past. For example, Danielle Nouy, now Chair of the Supervisory Board of the ECB, stressed back in April 2012 that “more capital charge against sovereign risk and less incentives for the purchase of sovereign debt should especially be considered in a context where this asset class can no longer be considered as a low-risk or risk-free asset class” (Nouy 2012). However, her plea went unnoticed and European government bond holdings of the Eurozone’s monetary financial institutions (MFIs) increased by another €375 billion in market value during the ensuing three years.

Sabine Lautenschläger, a member of the ECB’s executive board and vice-chair of the Supervisory Board, has also expressed her preoccupation about 0% risk weights: “For me it is quite clear that our system of risk weightings has some shortcomings; just think of the treatment of government bonds, which allegedly carry zero risk” (Lautenschläger 2013).

However, we should not forget that sovereign issuers were facing a completely different market environment during the Crisis, when the need to provide governments with a stable source of funding was a top priority. In this context, we appreciate the reasons for the EU’s Capital Requirements Directive (CRD) to deviate from the core philosophy of the Basel Accord and treat EGBs as 0% risk-weighted assets. Times have changed, though.

More recently, the ECB’s Vice President, Vitor Constâncio, has expressed his disagreement with the idea of introducing a so-called ‘large exposure limit’ to banks’ domestic eligible government bonds holdings, while sounding more confident about a review of the current regulatory treatment of sovereign exposure for Eurozone’s banks (Constancio 2015).

This view has been challenged by a recent European Commission report reminding us about a key lesson from the Crisis, i.e. that government bonds are not riskless (Figure 1). The report has concluded that “zero-risk weighting of sovereign debt in the EU, as well as the exemption from existing large exposure requirements, are a source of vulnerability” (EPSC 2015). Also, top executives at the European Commission have voiced their concerns about the relationship between banks and government bonds: “In the medium term, it may make sense to review the treatment of bank exposures to sovereign debt, for example by setting large exposure limits.”1

Figure 1. Bond and equity market volatility

Source: Citi Research, Bloomberg

Eventually, the ESRB argues in favour of a reform to the current 0% risk weighing practice that should “design adjustments to regulatory treatment, with the aim of enhancing banks’ incentives to take account of sovereign risk and increasing their resilience to such risk” (ESRB 2015).

From a systemic risk perspective, 0% risk weights should be questioned for various reasons:

  • Large exposure risk - increased solvency risk at the individual bank level due to over-investment and risk concentration;
  • Systemic risk for the financial system in the low-probability case of a realisation of sovereign risks (i.e. the ‘doom loop’); and
  • Crowding out of the private sector (Figure 2) - regulation allows a bank to hold quasi-default free, 0% risk-weighted sovereign paper, thus reducing the appetite for private-sector loans that carry not only a high risk weight, but also a significant risk of not performing.2

Figure 2. Loan-to-deposit ratios and EGB holdings

Source: ECB, Citi Research

The demand for 0% risk weight assets

We think that banks and insurance companies are very likely to continue playing a significant role in the demand for European government bonds, even in a scenario of non-0% risk weights. Furthermore, we also think that European officials should establish a level regulatory playing field between sovereign, quasi-sovereign and non-sovereign risk under the assumption that government bonds are not risk-free any longer.

Timing is of key importance. We argue that introducing non-0% risk weight in the current market environment would be more disruptive to bond valuations than during the Global Crisis. Not only are European banks now better capitalised and their balance sheet risks better understood, but European government bonds markets are also benefiting from the ECB’s support.

A risk-based, country-by-country approach and a non-country-specific large exposure limit are the most discussed schemes. There are various permutations of a risk-based weighting scheme, but in general, any weighting scheme will imply a trade-off between the precision of risk mapping and the implied pro-cyclicality of applying a risk weight.

Figure 3. The effect of regulatory scenarios on banks’ holdings of EGBs

Note: Euros in billions.
Source: ECB; Citi Research.

To analyse the effect of a regulatory change on banks’ holdings of European government bonds, we glance at a snapshot of the data together with alternative distributions of portfolio weights (Citi Euro Weights 2015). In Figure 3, we show different schemes and define the ‘home bias’ as the ratio of domestic to total European government bonds held by banks. The underlying thought is that positive risk weights would force a reallocation of sovereign risk across banks. Of interest are the two limit cases of a ‘change in benchmark’ and the more controversial case of a strict 25% large exposure threshold on domestic holdings.

The extreme examples below are illustrative of the reallocation of European government bonds:

  • Proportionality to total MFI assets: Italian and Spanish banks would have to reduce their total government bonds holdings by €221 billion and €124 billion, respectively. By contrast, French banks would have to increase holdings by €220 billion. Total government bond holdings across the banking sector would remain at €1.76 trillion, in other words, this policy option is a pure redistribution of risk.
  • Proportionality to EGBI: Same redistribution of government bonds risk as above, albeit less aggressive.
  • Large exposure cap for domestic holdings: Total domestic bond risk across Eurozone’s banks would decline sharply by €870 billion (i.e. 66% of domestic EGB holdings). This policy measure afflicts all countries, in particular Italy (-€305 billion) and Spain (-€178 billion) due to their significant home bias.

Again, note how the ECB’s public sector purchase programme (PSPP) acts as a smoothing factor. In that sense, one has to wonder about the negative effect on European bonds valuations of a transition from 0% risk weights to non-0% risk weights in the absence of the monetary policy support.

Conclusions and policy recommendations

  • Non-0% risk weights for sovereigns are likely to be introduced in European financial regulation. We have a hard time finding arguments in favour of treating government bonds as risk-free assets.
  • The discussion about the modality is well alive and there does not seem to be a consensus between the ECB/SSM (risk weighted approach) and the European Commission (large exposure approach).
  • If managed carefully – especially when it comes to timing and modality – we don’t think that the regulatory innovation will be a negative shock to bond valuation.
  • The potential European government bonds rebalancing flow, both in terms of regional allocation and total investment, could end up being sizeable. Regulators are strongly urged to perform impact studies taking into account the regional bias of bond holdings.
  • The process of regulatory tightening continues and adds layers of complexity to an already complex environment. The interplay between the leverage ratio and risk-based measures is just one example.3
  • The new regulation should take the form of a new Basel Accord, otherwise risking a cross-border regulatory arbitrage that would ultimately harm the profitability and international competitiveness of European banks.
  • Changes to the risk treatment of sovereign bonds would imply changes for quasi-sovereign debt as well.
  • The process of reassessing public sector risk should not stop with government bond markets. Public sector loans might also come under regulatory pressure. In this case, German banks might end up being affected more than peripheral credit institutions.

Authors' note: Please refer to important analyst disclosures and disclaimers at the end of the published research here and here.


Citi Euro Rates (2012), “Eurozone’s Lost Decade”.

Citi Euro Rates (2014), “Low-for-Longer: Not Even at Half Time”.

Citi Euro Rates (2015), “On the Likelihood of Non-0% RW for European Government Bonds”.

Constâncio, V (2015), “The Role of Stress Testing in Supervision and Macroprudential Policy”, speech, 29 October.

EPSC (2015), “Further Risk Reduction in the Banking Union”, Five Presidents’ Report Series.

ESRB (2015), ESRB Report on the Regulatory Treatment of Sovereign Exposures.

Lautenschläger, S (2013), “The Leverage Ratio: A Simple and Comparable Measure?”, speech, 21 October.

Nouy, D (2012), “Is Sovereign Risk Properly Addressed by Financial Regulation?”, Financial Stability Review No. 16, Banque de France.


[1] “EU Plans to Review Rules on Banks’ Exposure to Sovereign Bonds”, Reuters, 24 November 2015.

[2] Crowding out of the private sector, the evolution of bank holdings of sovereign risk and the effect on EGBs’ pricing has been a key theme for us in the past, see Citi Euro Rates (2012, 2014).

[3] Fender / Lewrick (2015), “Calibrating the Leverage Ratio”, BIS Quarterly Review, December 2015.

4,829 Reads