European Banks’ Sovereign Debt Home Bias: Voluntary or Involuntary?

Bálint Horváth, Harry Huizinga, Vasso Ioannidou

31 July 2015

a

A

Banks’ exposure to sovereign debt potentially reinforces the negative feedback loop between weak public finances and financial instability in a country (Acharya et al. 2014). Moreover, bank holdings of sovereign debt may transmit public default risk in one country to other countries (Korte and Steffen 2014). These perils inherent in banks’ sovereign debt portfolios motivate a closer look at why banks invest heavily in sovereign debt and, particularly, in domestic sovereign debt, leading to a sovereign debt home bias. 

Broadly speaking, the public debt home bias can be either voluntary or involuntary. Banks may voluntarily load up on domestic debt, as additional domestic sovereign exposure cannot hurt them much, because they are likely to fail anyway if their sovereign defaults.1 Alternatively, the home bias is involuntary, resulting from some form of regulatory or supervisory pressure on banks to purchase domestic sovereign debt. As an example of this, the Hungarian central bank announced a new interest swap facility in April 2014 that was designed to provide incentives to Hungarian banks to hold additional Hungarian public debt.

In a recent paper (Horváth et al. 2015), we examine the merits of these explanations of the home bias, using data on sovereign debt exposures from EU-wide stress tests conducted by the European Banking Authority and its predecessor over the 2010-2013 period. The evidence supports a partially voluntary and a partially involuntary sovereign debt home bias. Furthermore, we find that the ECB’s three-year Long Term Refinancing Operations (LTRO) around the end of 2011 enabled EU banks to significantly increase their domestic debt exposures towards levels that maximise bank valuation.

The sovereign debt home bias  

On average, banks examined in the EU stress tests had a ratio of domestic sovereign debt to total assets of 6%, compared to an average ratio of sovereign debt of any foreign EU country to total assets of 2.2%, as evidence of a material home bias. Figure 1 pictures the development of the average ratio of domestic debt to total assets for banks in three GIIPS countries (Italy, Portugal, and Spain) and in 14 non-GIIPS countries during 2010-2013.[2] Banks in both groups saw their domestic debt exposure rise during this period. Banks in the three GIIPS countries were relatively more heavily exposed to domestic debt throughout the period, and these banks increased their domestic debt exposure significantly after the ECB provided extensive liquidity to the Eurozone banking system through its three-year Long Term Refinancing Operations around the end of 2011.

Figure 1. Banks’ exposure to domestic government debt in GIIPS and non-GIIPS countries

Notes: The graph displays the average domestic government debt to total assets ratios for banks in three GIIPS countries (Italy, Portugal and Spain) and non-GIIPS countries.

Following Bracke and Schmitz (2011), we define a bank’s sovereign debt home bias as 1 minus the ratio of a bank’s sovereign debt portfolio share allocated to foreign EU countries to the share of foreign EU country sovereign debt in all EU sovereign debt held by banks in the sample. This home bias measure is positive if a bank holds its domestic sovereign debt in a higher proportion than the average bank in the EU stress tests, and vice versa. Figure 2 shows that the average home bias rose over the period, especially for non-GIIPS banks. Banks in the three GIIPS countries display a higher domestic debt home bias throughout. The ECB’s Refinancing Operations of 2011 and 2012 appear not to have affected the average home bias much, reflecting that banks increased their domestic as well as their foreign public exposures.

Figure 2. Average home bias for banks in GIIPS and non-GIIPS countries

Notes: The home bias is calculated as 1 minus the ratio of a bank’s sovereign debt portfolio share allocated to foreign EU countries to the share of foreign EU country sovereign debt in all EU sovereign debt held by banks in the sample. The included GIIPS countries are Italy, Portugal, and Spain.

Evidence on whether the home bias is voluntary or involuntary

The empirical investigation in our paper sheds light on which banks maintain a high ratio of domestic sovereign debt to total assets. Weak banks (as indicated by a high leverage) and banks located in weak countries (as indicated by a high sovereign Credit Default Swap (CDS) rate) tend to show high exposures to domestic public debt. These results are consistent with weak banks voluntarily holding risky sovereign debt in an effort to shift risk to the bank’s creditors, but also with the governments of weak countries jawboning weak banks into holding more domestic sovereign debt.

Further evidence suggests that high domestic debt exposures have both a voluntary and an involuntary basis:

  • Weak banks with shareholder-friendly corporate governance schemes in the area of, for instance, board composition invest relatively heavily in domestic sovereign debt.

These banks are likely to voluntarily concentrate on holding domestic sovereign debt, as a bank’s shareholders are the main beneficiaries of any risk shifting achieved by a high domestic debt exposure.

  • Banks located in weak countries that are government owned also have relatively high domestic public debt exposures.

Publicly owned banks can more easily be forced into holding domestic debt by weak governments, and hence may end up holding such debt involuntarily.

  • Banks with a sovereign debt portfolio biased towards domestic debt are valued more highly in the stock market, if both the bank and the sovereign are weak.

This is consistent with weak banks located in weak countries voluntarily increasing their sovereign debt home bias in an effort to shift risk with a view to increasing bank valuation. Interestingly, the valuation of the home bias turns negative if the bank is strong and the country is weak, suggesting that banks in this scenario maintain more domestic sovereign debt than is consistent with maximum bank valuation.

Overall, banks’ holdings of domestic public debt appear to be the outcome of both private and public influences.

The ECB’s long-term refinancing operations and the valuation of the home bias        

As suggested by Figure 1, additional liquidity in the form of refinancing operations made available by the ECB around the end of 2011 enabled Eurozone banks to increase their exposures to domestic sovereign debt. To see how bank valuation was affected, we first estimate the relationship between bank valuation and the portfolio home bias for each of 6 stress test dates during 2010-2013.

The resulting estimated relationships can be used to derive the marginal valuation of the portfolio home bias for each of the 6 dates as displayed in Figure 3. The marginal value of the home bias is seen to be positive in March 2010 and December 2010 before the long-term operations announcement, while it is about zero in June 2012 and December 2012 after their implementation. This pattern suggests that before the refinancing operations injections, banks could have augmented valuations by increasing their sovereign debt home bias. The absence of sufficient liquidity may have prevented them from doing this. The LTRO injections, however, expanded banks’ access to low-priced, risk-insensitive liquidity, and allowed previously constrained banks to enlarge their holdings of domestic government debt. This brought down the marginal valuation of the home bias, while the spreads of riskier government bonds began to drop.

Figure 3. Marginal value of home bias during 2010-2013

Note: The graph displays the average marginal value of the home bias during the sample period and the 95% confidence intervals around the estimated mean values.

Policies to reduce the home bias

Current risk-shifting incentives underlying high concentrations of domestic public debt should be eliminated, and any remaining pressuring of banks by governments to invest in domestic public needs to stop.

Unlike current practice, banks should be required to apply positive and risk-sensitive risk weights to their sovereign exposures in calculating their total risk-weighted assets for the purpose of determining capital adequacy. Discriminatorily high risk weights for domestic government debt would even be appropriate to address banks’ specific incentives to accumulate domestic rather than foreign government debt.

In a European context, banks and countries jointly face incentives to let domestic banks build up high concentrations of domestic sovereign debt, because part of the cost of joint failure can be exported to the ECB, the European Stability Mechanism, and individual Eurozone countries, as shown by the recent experience of Greece. Individual Eurozone countries, therefore, may not be interested in instituting a comparatively high risk weight for domestic sovereign debt, and thus an initiative to implement a relatively high risk weight for domestic debt can only be successful at the European level.

Taking the existing framework of optional zero risk weights for government debt for granted, Corsetti et al. (2015) recently proposed that banks should only be allowed to apply the zero risk weight if they hold the sovereign debts of varying countries in proportion to, say, national GDPs. This would appropriately countervail a bank’s incentive to bring about a sovereign debt home bias, even if the general incentive to overinvest in sovereign debt would remain intact.

The scope for governments to use bank regulation and supervision to pressure their banks into holding more domestic sovereign debt can be reduced by transferring bank regulatory and supervisory powers to a supranational level. As an important step in this direction, the European Central Bank has become the single bank supervisor in the Eurozone as part of the Single Supervisory Mechanism that took effect in November 2014. It remains to be seen whether this institutional change will contribute towards a reduction in the sovereign debt home bias of Eurozone banks.

References

Acharya, V, I Drechsler, and P Schnabl (2014), “A Pyrrhic victory? Bank bailouts and sovereign credit risk”, Journal of Finance 69, 2689-2739.

Battistini, N, M Pagano and S Simonelli (2014), “Systemic risk, sovereign yields and bank exposures in the euro crisis”, Economic Policy 78, 205-241.

Bracke, T and M Schmitz (2011), “Channels of international risk-sharing: capital gains versus income flows”, International Economics and Economic Policy 8, 45-78.

Corsetti, G, L P Feld, P R Lane, L Reichlin, H Rey, D Vayanos, and B Weder di Mauro, (2015), “A new start for the Eurozone: dealing with debt”, CEPR Monitoring the Eurozone Report.

Horváth, B L, H Huizinga, and V Ioannidou (2015), “Determinants and valuation effects of the home bias in European banks' sovereign debt portfolios”, CEPR Discussion Paper 10661.

Korte, J, and S Steffen (2014), “A ‘sovereign subsidy’ – zero risk weights and sovereign risk spillovers”, VoxEU.org, 7 September.

Endnotes

[1] Domestic government debt can in addition be a natural hedge against the risk of currency redenomination which is a possibility in the event of a Eurozone break-up (Battistini et al. 2014). Some evidence supporting a risk-shifting motive for the home bias is also consistent with a natural hedge explanation.

[2] The non-GIIPS countries are Austria, Belgium, Cyprus, Denmark, Finland, France, Germany, Hungary, Netherlands, Norway, Poland, Slovenia, Sweden, and the United Kingdom.

a

A

Topics:  Financial regulation and banking

Tags:  sovereign debt, financial instability, home bias, stress tests, bank valuation

PhD candidate, Tilburg University; and Junior Fellow, European Banking Center

Professor of Economics and Chairman of the European Banking Centre, Tilburg University

Professor of Finance, Lancaster University; Research Fellow and board member, European Banking Center; and Research Fellow, CEPR

Events