The recent crisis has led to a considerable number of interventions in, and resolutions of, failing banks, and we can expect many more in the coming months as the Eurozone crisis in particular continues to undermine bank fragility. The discrepancy between the geographic perimeter of banks – especially large banks’ activities – and regulators’ perimeters has added a further layer of complication, and some have called for a Eurozone-level banking regulation and resolution regime (Saurez 2011).
For example, the resolution of Fortis at the national level - undertaken separately by Dutch, Belgian, and Luxembourg authorities - has confirmed Charles Goodhart’s and Mervyn King’s point that "banks are international in life and national in death". The failure of the Icelandic banks, with wide-ranging economic and political repercussions, has shed doubts on the viability of large multinational banks in small countries.
Ideally, authorities should only intervene in a bank when the expected costs of doing so are less than the expected benefits. The expected survival likelihood – which largely determines the benefits – should be reflected in the CDS spread, which is the expected loss on the underlying asset. Yet, when we look at a sample of 54 banks that were the subject of government interventions in 2008 or 2009, we observe a significant dispersion across banks and countries (see Figure 1 from our recent research, Beck et al. 2011). This casts doubt on the above statement.
Some banks, such as the German Commerzbank, were the subject of interventions early on (when the risk of failure – and hence their CDS spread – was low). Others, such as Wachovia, rather late. Intervention into large cross-border banks came often at a late stage, and often with conflicts between home and host country supervisors. Notably, interventions into Fortis and the Icelandic banks were very late.
Figure 1. Bank CDS spreads three days before intervention
In our recent research, we explore both theoretically and empirically whether this variation across different banks can be explained by banks’ cross-border activities, which may distort supervisory incentives to intervene. This research is complementary to other research on cross-border banking, such as Loranth and Morrison (2007) who discuss the implications of capital requirements and deposit insurance for cross-border banks, and Dell’Arricia and Marquez (2006) who show that competition between national regulators can lead to lower capital adequacy standards, as well as Calzolari and Loranth (2010) who show how the organisational structure of multinational banks can influence regulatory behaviour.
A simple theoretical model
Our model shows that cross-border activities indeed distort the supervisory decision to intervene in a bank. Assume a supervisor with information about a bank’s success probability for the next period has to decide on whether to allow the bank to continue or to resolve it. In the case of purely national banks, the supervisor will compare the costs of intervening now with the possible costs and benefits of allowing the bank to continue further. In the case of cross-border banks, this becomes more complicated. Since the home-country supervisor will care primarily about the costs of resolution or failure to domestic depositors and equity holders, and to the domestic economy, she will not take into account the costs and benefits of allowing continuation accruing to foreign depositors and equity holders and to foreign economies. In particular, with a higher share of foreign deposits, the home-country supervisor becomes less likely to intervene early, as some of the costs are being borne by agents outside the supervisor’s geographic perimeter. Similarly, for the case of foreign assets, the home-country supervisor will be less likely to intervene early, as some of the costs of bank failure will be borne by foreign borrowers and economies. A higher share of foreign equity, on the other hand, will bias the home-country supervisor towards an early intervention, as only a some of the benefits of allowing the bank continue will accrue domestically.
If cross-border banking takes place through more than one channel, the net distortion depends on the strength of each channel. For example, if there are both cross-border ownership of deposits and equity, the biases created by each channel go in opposite directions and tend to offset each other. If there is mainly foreign deposit-taking but little foreign ownership, we are then likely to end up with a domestic regulator who is too lenient, and vice versa. This implies that in order to evaluate the efficiency properties of cross-border banking, we have to look at all aspects of cross-border banking jointly, and not only at one channel in isolation.
Going back to the variation of CDS spreads of banks subject to interventions discussed above, we show that the cross-border activities of these banks can indeed explain why interventions took place at different levels of fragility, that is, different CDS spreads. Specifically, we find that:
- Banks with a higher share of foreign assets or deposits are subject to interventions at a higher level of fragility.
- Banks with a higher share of foreign equity are subject to interventions at a lower level of fragility.
- Banks with an overall cross-border balance biased towards foreign ownership rather than foreign assets and deposits are subject to interventions at a lower level of fragility.
These results are robust to controlling for other bank characteristics, as well as indicators of regulatory structure.
What about supranational supervision?
Can a supranational supervisor improve on the current network of national supervisors in Europe? When domestic and efficient intervention thresholds differ, a supranational supervisor could, in principle, always increase welfare because this supervisor would also take into account the effects that materialise outside of the country in question. However, supranational supervision might itself also be subject to imperfections.
- First, a supranational supervisor, even having the correct incentives, can make wrong decisions if she has less perfect knowledge of the success probability than national supervisors.
- Second, the global supervisor may be less efficient in intervening, resulting in higher resolution costs. Such higher costs could arise from being farther away from the relevant market and thus being disadvantaged – relative to a national supervisor - in terms of arranging for a merger and acquisition or purchase and assumption operation.
- Third, intervening in and resolving a bank that is present in markets with different legal frameworks can result in extended and costly resolution, raising the external costs for the economies in question. There is, however, also a countervailing effect - a supranational supervisor might be in a better position to resolve a financial institution that dominates its home country when operating in a supranational banking market. This is because a national supervisor may not have the resources to resolve a bank that is large relative to the domestic economy.
Could European-level supervision work?
The comparison between national and supranational supervision has clear implications for the current debate on establishing a European-level failure resolution framework. To which we would add the following points:
- First, a supranational regime can only improve the failure resolution for banks whose mix of foreign deposits, assets, and equity is imbalanced and, hence, distorts the intervention decision by a national supervisor.
- Second, such a supranational framework also has to relate to the appropriate geographic area.
As shown by Osterloo and Schoenmaker (2007) and Schoenmaker (2010), the largest 25 European banks have, on average, 25% of their assets in other European countries outside of their own. This share ranges, however, from 2% in the case of BBVA (which has 31% of assets outside Europe) to the Nordea Group, with 74% of assets outside its home countries in other European countries (and no assets outside Europe). A European-level supervisor would not help to reduce distortions in the case of banks with a significant share of assets outside Europe.
- Third, such a regime can only improve on a purely national resolution framework if equipped with the necessary means and resources to resolve a bank efficiently.
- Fourth, resolution powers have to come with the necessary supervision and monitoring tools; a close relationship with national supervisors is therefore critical.
Beck, Thorsten, Radomir Todorov, and Wolf Wagner (2011), “Bank Supervision Going Global? A Cost-Benefit Analysis”, European Banking Center Discussion Paper, 2011-33.
Calzolari, Giacomo and Gyongyi Loranth (2011), “Regulation of multinational banks: A theoretical inquiry”, Journal of Financial Intermediation, 20:178-198.
Dell’Ariccia, Giovanni and Robert Marquez (2006), “Competition Among Regulators and Credit Market Integration”, Journal of Financial Economics, 79:401-430.
Loranth, Gyonguyi and Alan Morrison (2007), “Deposit Insurance, Capital Regulations and Financial Contagion in Multinational Banks”, Journal of Business Finance & Accounting, 34:917-49.
Osterloo, Sander and Dirk Schoenmaker (2007), “Financial Supervision in an Integrating Europe: Measuring Cross-Border Externalities”, International Finance, 8:1-27.
Schoenmaker, Dirk (2010), “Burden Sharing for Cross-Border Banks", Revista de Estabilidad Financiera 18, Bank of Spain, 31-47.
Suarez, Javier (2011), “A three-pillar solution to the Eurozone crisis”, VoxEU.org, 15 August.