RESEARCH BULLETIN NO. 42

Sub-sovereign bonds in banks’ portfolios: A role for political connections?

11 January 2018

By Alexander Popov[1]

This article shows that German savings banks appear to increase their holdings of bonds issued by their respective Bundesland (federal state) government if as a result of an election, the local governments at Bundesland and at Kreis (county) level are no longer dominated by the same party. This behaviour is not consistent with other known reasons why banks hold government debt, such as compliance with regulation, the tendency to accumulate risky assets when close to bankruptcy, or political pressure. Instead, we argue that in the wake of a post-election loss of political connections along party lines, local government-owned banks use purchases of sub-sovereign bonds to keep communication channels with state politicians open, a mechanism akin to lobbying.

Why political connections matter to banks

Political connections are valuable to firms. Research shows such connections can increase firms’ stock returns, improve their access to credit, strengthen their ability to influence legislation, and lead to more favourable regulatory treatment and to more generous government support when in distress.[2] At the same time, while beneficial to the individual firms, close relationships with politicians can generate substantial economic distortions, such as a gain in market power and a reduction in the provision of public goods.[3] This applies particularly to banks, because the government both regulates them and provides them with a financial backstop. Therefore it is important to document how firms in general – and banks in particular – build political connections, so we can understand both the private benefits and the social costs of such connections.

There is already plenty of evidence that banks invest extensively in building political connections, but such evidence is restricted to campaign contributions, lobbying, and direct lending.[4] This article[5] presents new evidence that banks also can – and do – adjust their securities portfolios in an attempt to bridge political distance. We demonstrate a robust new fact: after a loss of political connections along party lines due to regional elections, local government-owned banks in Germany increase substantially their holdings of bonds issued by their home state government.

Setting of the empirical test

To identify the political incentives to buy regional government debt, we exploited the unique dual institutional setting of the German electoral system and of the German banking system.

First, it is important to understand how regional governments are elected. Germany is a federation that consists of 16 federal states and 438 counties. Federal states and counties hold regular elections for state parliaments and county councils, respectively. These elections take place every four to six years, but typically elections at the different levels are held at different points in time. This leads to staggered changes in the composition of ruling coalitions in the federal states and in the counties within these states. There are two major parties in Germany, the Christian Democrats (CDU) and the Social Democrats (SPD). Both are active at all levels of regional politics and elections at state or county level nearly always result in one of these two parties leading a ruling coalition.

Second, it is necessary to understand the German banking system. It comprises private banks, cooperatives, and government-owned banks. There are two types of government-owned banks: Sparkassen (local savings banks) – typically one per county – and Landesbanken or regional umbrella institutions operating at state level, but in some cases covering several states. Taken together, government-owned banks account for around a third of the German banking market, in terms of assets, so they are a key part of Germany’s banking system. What is unique about local savings banks is that they are the only segment of the German banking system that is under the direct influence of political parties. In practice, the city mayor or county administrator typically sits on the bank's supervisory board. Such a position involves considerable control over the bank. Importantly, the local politician acquires this position as a direct result of winning the local election, and loses it once his or her party is no longer in power locally. As the elections at federal state and county level are always held at different times, there is frequently potential for "political misalignment", i.e. different parties dominating at different levels of local and regional government.

Finally, States in Germany have independent fiscal authority, and all 16 German States issue their own government bonds. Local savings banks collectively hold around 25 bln euro worth of German State bonds, which is ten times more than their holdings of German federal bonds. Because sub-sovereign debt is typically placed privately, contractual prices are often not publicly observed. It is plausible that in such private placements, and as a political favour, banks offer to purchase bonds at better-than-market prices. And even in the absence of that, bond-purchasing operations can simply be an initiation of a repeated interaction between local politicians and state politicians, which makes it akin to the "face time" or "front-row access" that lobbying buys.

The impact of elections on banks’ sub-sovereign holdings

Using this set-up, we studied the home-state bond-buying behaviour of 455 government-owned banks, taking advantage of a total of 32 state parliament elections and 600 county council elections between 2005 and 2013. Our main finding is that political misalignment is positively and significantly associated with the propensity of local government-owned banks to hold sovereign bonds issued by their state. Numerically, an election that breaks a local government-owned bank connection to the state government along political party lines results in an increase in that bank's holdings of debt issued by the respective state government of on average 42% of the sample mean exposure, relative to a similar savings bank in a politically aligned county. Figure 1 describes this development graphically. "Event time" on the horizontal axis refers to the timing in quarters before and after an election that brings about political misalignment (for "Misaligned" banks) or not (for "Aligned" banks), with "0" indicating the election quarter.[6] Furthermore, local government-owned banks in politically misaligned counties are not more likely to increase their holdings of alternative but similar assets, such as German federal bonds (Bunds) and bonds issued by other German States.

Figure 1: Holdings of home-state bonds by local savings banks: politically misaligned versus aligned banks

Notes: Average holdings of home-state sub-sovereign bonds as a percentage of total assets, for politically aligned versus misaligned banks, during quarters around election events. The sample period is 2005 – 2013.

Our empirical strategy allows us to rule out other explanations for banks increasing their sub-sovereign bond holdings, either because they are unlikely or because they would have the opposite effect. In particular, neither regulation nor the need for highly liquid assets explain our finding. This is because these would lead banks to stockpile Bunds which are virtually risk-free, perfectly liquid, and widely available. They cannot be driven by asset substitution either (i.e. acquiring a riskier portfolio), as bonds issued by German federal states are typically highly rated. Finally, our results show it is unlikely that politicians from the home-state government are pressuring local savings banks to buy their bonds ("moral suasion"). If that was the case, we would see politically aligned banks increasing their holdings of home-state bonds. Finally, we find that the main effect is mostly a result of political misalignment following elections at state level. This strongly suggests that incumbent politicians on the supervisory boards of local government-owned banks prompt the banks in question to buy more bonds issued by their home state in reaction to losing political proximity to the state government. Overall our results suggest that we have discovered a genuine link between election outcomes and local government-owned banks' propensity to purchase bonds issued by their home state.

Conclusion

There are at least three macroeconomic implications of our finding that should worry policy-makers. First, if banks hold more government debt this increases the vicious circle of interdependence between banks and government in times of sovereign stress[7], and this argument is just as valid for sub-sovereign debt as it is for sovereign debt. Second, research has shown that increased investment in public debt can crowd out investment in the private sector.[8] Third, close political relationships that benefit private parties are often associated with substantial costs to society.[9] For all these reasons, it may be advisable for supervisors to pay attention to politically-motivated actions by banks, even when these are not in breach of regulatory rules.

References

Becker, B. and Ivashina, V. (2018), "Financial repression in the European sovereign debt crisis", Review of Finance, forthcoming.

Braun, M. and Raddatz, C. (2010), "Banking on politics: When former high-ranking politicians become bank directors", World Bank Economic Review, Vol. 24, pp. 234-279.

Cingano, F. and Pinotti, P. (2013), "Politicians at work: The private returns and social costs of political connections", Journal of the European Economic Association, Vol. 11, pp. 433-465.

Claessens, S., Feijen, E. and Laeven, L. (2008), "Political connections and preferential access to finance: The role of campaign contributions", Journal of Financial Economics, Vol. 88, pp. 554-580.

Cooper, R. and Nikolov. K. (2018), "Sovereign debt and banking fragility: The spreading of strategic uncertainty", International Economic Review, forthcoming.

Duchin, R. and Sosyura, D. (2012), "The politics of government investment", Journal of Financial Economics, Vol. 106, pp. 24-48.

Faccio, M. (2006), "Politically connected firms", American Economic Review, Vol. 96, pp. 369-386.

Gao, H., Ru, H. and Tang, D. (2016), "Subnational debt of China: The politics-finance nexus", NYU mimeo.

Goldman, E., Rocholl, J. and So, J. (2009), "Do politically connected boards affect firm value?", Review of Financial Studies, Vol. 22, pp. 2331-2360.

Khwaja, A. and Mian, A. (2005), "Do lenders favour politically connected firms? Rent provision in an emerging financial market", Quarterly Journal of Economics, Vol. 120, pp. 1371-1411.

Koetter, M., and Popov, A. (2018), "Politics, banks, and sub-sovereign debt: Unholy trinity or divine coincidence?", ECB Working Paper, forthcoming.

Kroszner, R. and Strahan, P. (1999), "What drives deregulation? Economics and politics of the relaxation of bank branching restrictions", Quarterly Journal of Economics, Vol. 114, pp. 1437-1467.

Lambert, T. (2018), "Lobbying on regulatory enforcement actions: Evidence from banking", Management Science, forthcoming.

Mian, A., Sufi, A. and Trebbi, F. (2010), "The political economy of the US mortgage default crisis", American Economic Review, Vol. 100, pp. 1967-1998.

Popov, A. and van Horen, N. (2015), "Exporting sovereign stress: Evidence from syndicated bank lending during the euro area sovereign debt crisis", Review of Finance, Vol. 19, pp. 1825-1866.

Shleifer, A. and Vishny, R. (1994), "Politicians and firms", Quarterly Journal of Economics, Vol. 109, pp. 995-1025.



[1] Disclaimer: This article was written by Alexander Popov (Principal Economist, DG-R/FIR). The author thanks Paul Dudenhefer, Michael Ehrmann, Geoff Kenny, Manfred Kremer, Luc Laeven, Silvia Margiocco, Zoe Sprokel and Michael Steen for useful comments. The views expressed here are those of the author and do not necessarily represent the views of the European Central Bank and the Eurosystem.

[2] See, for example, Khwaja and Mian (2005), Claessens, Feijen, and Laeven (2008), Faccio (2006), Kroszner and Strahan (1999), Braun and Raddatz (2010), and Duchin and Sosyura (2012).

[3] See Cingano and Pinotti (2013).

[4] See, for example, Mian, Sufi, and Trebbi (2010), Lambert (2018), and Gao, Ru, and Tang (2016).

[5] Based on Koetter and Popov, 2018 (forthcoming).

[6] Importantly, the statistical association between political misalignment and bond buying disappears in placebo tests where we move the timing of State elections by 1, 2, or 3 years around the true election date.

[7] See Cooper and Nikolov (2018).

[8] See, for example, Becker and Ivashina (2018) and Popov and Van Horen (2015).

[9] See Shleifer and Vishny (1994).

Media contacts