In her foreword to the first Annual Report after taking over responsibility at the ECB for supervising European banks, Danielle Nouy states: “Over the course of 2015, we made good progress in promoting the objectives of European banking supervision. We contributed to the safety and soundness of credit institutions and to the stability of the financial system. We also promoted the unity and integrity of the internal market based on equal treatment of credit institutions” (ECB 2016).
After the twin crises of the new millennium, this sounds like wonderful news, especially for European banks, except that the goal of contributing to the safety and soundness of the financial system had already been the stated goal of the Basel Accord in its original version about 30 years earlier. The Accord, now commonly referred to as Basel I, states:
“Two fundamental objectives lie at the heart of the Committee's work on regulatory convergence. These are, firstly, that the new framework should serve to strengthen the soundness and stability of the international banking system; and secondly that the framework should be in fair and have a high degree of consistency in its application to banks in different countries with a view to diminishing an existing source of competitive inequality among international banks” (Basle Committee on Banking Supervision 1988).
This raises the question of whether the Basel process of capital regulation might not have achieved its intended goals, and, hence, whether it has triggered the need to develop a new supervisory infrastructure such as the Banking Union in Europe. In a recent paper, we provide an evidence-based policy evaluation of the Basel process of capital regulation (Gehrig and Iannino 2017). In a first step, we trace the trajectories of various commonly used systemic risk measures and document important non-linearities and heterogeneity across size groups. We find that the major build up in exposure to systemic risk remains concentrated in the upper quintile of banks, while for the majority of banks no such secular increase in exposure to systemic risk can be documented. In a second step, we identify the drivers of these unequal developments, which essentially are the options to self-regulate by the use of internal models for market risk and credit risk subject to supervisory approval.
Evolution of systemic risk
On a large sample of listed European financial institutions, we analyse the evolution of commonly used systemic risk measures since the 1980s, which include exposure measures, such as SRISK (Brownlees and Engle 2017), and contribution measures such as Delta CoVAR (Adrian and Brunnermeier 2017). Our most disconcerting finding is the secular upward trend in the aggregate exposure of European Banks with respect to systemic risk (Figure 1). Since the introduction of Basel II in July 2006, the capital shortfall has almost doubled. It remains worrisome that this high level of risk exposure has not yet been substantially reduced. Even the Banking Union appears to have only had a short-lived effect on aggregate capital shortfall according to this measure.
Figure 1 Evolution of capital shortfall
Notes: the evolution of capital shortfall, as the daily average SRISK (Brownlees and Engle 2017), from 1988 to 2016. We average positive values of SRISK over a sample of 320 European listed financial institutions from 15 European countries. It represents the capital shortfall, as the total amount of recapitalisation needed in the system in case of major distress in the market, and if the transfer of capital from institutions in surplus to institutions in need is not possible. We use a capital ratio of 8%.
Applying a distributional perspective and classifying banks into different risk groups, it appears that the dominant driver of risk exposure is found in the upper quintile of institutions (Figure 2). For the majority of institutions – that is, the lower three quintiles – Basel II did not seem to exert a similarly detrimental effect on their systemic risk exposures. Moreover, for these banks, it appears that the market risk amendment in January 1996 had a small risk-reducing effect quite in line with the initial intentions of the Basel Committee.
Figure 2 Evolution of the daily average estimated SRISK for banks by degree of capital shortfall
Notes: the evolution of the daily average estimated SRISK, distinguishing five equal-size groups of financial institutions. Each institution is attributed to a group according to its percentage contribution of capital shortfall to the overall system. The top quintile corresponds to the group with the highest level of percentage SRISK, while the bottom quintile corresponds to the group with the lowest level of percentage capital shortfall. For each group, we average both positive and negative values of SRISK and trace it from 1988 to 2016.
Finally, one might ask whether the supervisory reform under the Banking Union was successful in identifying the main drivers of systemic risk exposure. Figure 3 suggests that the answer should be yes. Indeed the most systemic banks are now under the direct supervision of the ECB. Nevertheless, after a short phase of recapitalisation in the run up to November 2014, the drive towards recapitalisation seems to have lost steam; aggregate capital shortfall remains at imprudently and disconcertingly high levels of the great financial crisis.
Figure 3 Evolution of the daily average SRISK, banks under ECB supervision
Notes: the evolution of the daily average SRISK, distinguishing banks under ECB supervision from the rest of institutions. The group of interest (red line) corresponds to the group of banks that were part of the Stress Test performed by ECB in 2014, and their supervision has been taken over by ECB.
Drivers of systemic risk
While the analysis of the trajectories of SRISK is quite suggestive, a multivariate statistical analysis is required to establish causal links from instruments to effects. Applying quantile regressions and controlling for the implementation dates of internal models for credit risk on the level of banks, the type of internal models employed, the usual drivers of bank stock prices, macroeconomic indicators, as well as country and time fixed effects, we find strong evidence that the use of internal models is highly concentrated among larger, and, hence, riskier banks (Gehrig and Iannino 2017).
Based on unconditional quantile regressions, illustrative counterfactual experiments can be run as presented in Figures 4 and 5. Fitting a statistical model on the period prior to the introduction of the market risk amendment one can work out how the realisation of subsequent risk factors would counterfactually have affected the SRISK trajectory if there had not been a change in business models. Comparing this hypothetical trajectory (blue) with the realised trajectory of SRISK (green) one can assess the impact of a change in business models induced by the introduction of internal models of market risk. Likewise, fitting a model prior to the implementation of Basel II, one compares the pre-Basel II trajectory (red) with the realised trajectory that incorporates banks adjustments to the new regulatory framework under Basel II. This analysis reveals that the adoption of internal models for credit risk (Basel II) was an especially major driver of the build-up of systemic risk exposure in Europe.
Ironically, Basel II was introduced for precisely the opposite reasons, as stated publicly in the Federal Reserve Bulletin in 2003 (about five years prior to the Global Crisis):
“The existing capital regime needs to be replaced for the large, internationally active banks whose operations have outgrown the simple paradigm of Basel I and whose scale requires improved risk-management and supervisory techniques to minimise the risk of disruptions to world financial markets. Fortunately, the art of risk measurement and management has improved dramatically since the first capital accord was adopted. The new techniques are the basis for the proposed new accord. The Basel II framework is the product of extensive multiyear dialogues with the banking industry regarding evolving best-practice risk-management techniques in every significant area of banking activity. By aligning supervision and regulation with these techniques, the proposed new framework represents a major step forward in protecting the US financial system and those of other nations. Basel II will also provide strong incentives for banks to continue improving their internal risk-management capabilities and will give supervisors the tools to focus on emerging problems and issues more rapidly than is now possible” (Board of Governors of the Federal Reserve System 2003: 405).
Based on the SRISK exposure measure, the empirical basis for such a positive appraisal of Basel I remains uncertain, since significant capital shortfall was already building up after 2000 in Europe. In any case, the US has never adopted the Basel II framework formally despite its strong lobbying efforts in its favour.
Figure 4 Evolution of the historical total SRISK compared with counterfactuals
Notes: the evolution of the historical total SRISK compared with its counterfactuals. We estimate a statistical model of SRISK in two sub-periods: i) before internal models were available (prior to 1996); and ii) prior to the implementation of internal credit risk models with Basel II (1996-2006). We then predict what SRISK would have potentially been, based on this model without the Market Risk Amendment in 1996 (blue line) and without the Basel II accord in 2006 (red line).
As a control, the counterfactual analysis can be expanded to include other financial intermediaries such as insurance companies or real estate financiers (Figure 5). Not surprisingly this exercise demonstrates that the Basel process largely affected the banking sector. Surprisingly, however, there are also significant spillovers in the insurance sector suggesting a build-up in systemic risk there as well.
Figure 5. Comparison between the observed total SRISK and the predicted counterfactual SRISK in three financial subsectors
Notes: Figure 5 reports the comparison between the observed total SRISK and the predicted counterfactual SRISK in three subsectors of financial markets: Sector 1 being banks, Sector 2 insurance companies, and Sector 3 real estate companies.
We document a secular build-up of capital shortfall of the European banking industry. Ironically, and counter-intuitively, by trying to regulate minimal capital standards, the Basel process itself contributed to an ever-increasing shortfall in aggregate bank capital. Consequently, European banks have become increasingly exposed to systemic risk over the last three decades.
The main drivers of this process appear to be the self-regulatory options introduced into the Basel process with the amendment for market risk (1996) and culminating in the introduction of internal models for credit risk in Basel II (2006). Rather than providing incentives for better risk management for the larger and internationally active banks, precisely those sophisticated banks used internal models to carve out even more equity in order to increase return on equity and at the same time reduce resiliency. In light of these results, placing caps on the use of internal models seem reasonable policy options.
However, reverting the Basel process and replacing complex rules – such as Basel II, III, IV and so on – with simple rules such as the original Basel Accord (Basel I) may be an even more effective way of reducing lobbying powers and system manipulation. In this light, the suggestion of Admati and Hellwig (2013) to impose a simple floor on capitalisation are worthy of serious consideration. Moreover, the recent history of re-capitalisation of UBS demonstrates that systemic risk exposure can indeed be reduced to pre-crisis levels when it is done seriously enough.
Implicitly, our study unveils another unintended consequence of policy, namely that expansive monetary policy adversely affects the resiliency of banks. The negative effect of the policy rate on systemic risk exposure is very robust and holds across the whole spectrum of banks. This finding suggests that expansionary monetary policy is not a good instrument to aid recapitalising ailing banks.
Admati, A and M Hellwig (2013) The Bankers' New Clothes, Princeton University Press, Princeton.
Adrian, T and M K Brunnermeier (2016) “CoVaR”, American Economic Review, 106(7): 1705-1741.
Basel Committee on Banking Supervision (1988) "International convergence of capital measurement and capital standards", Basel, July.
Board of Governors of the Federal Reserve System (2003) “Capital standards for banks: The evolving Basel accord”, Federal Reserve Bulletin, September.
Brownlees, D and R F Engle (2017) “SRISK: A conditional capital shortfall measure of systemic risk”, Review of Financial Studies, 30(1): 48-79.
Danielson, J, P Embrechts, C Goodhart, C Keating, F Münnich, O Renault, and H S Shin (2001) “An academic response to Basel II”, FMG, Special paper 130, London.
ECB (2016) Annual report on supervisory activities 2015, Frankfurt, March.
Gehrig, T and M Iannino (2017) “Did the Basel Process of capital regulation enhance the resiliency of European Banks?”, CEPR Discussion Paper 11920.