Fermer ce champ de recherche.

The regulation of bank capital, conference from Professor Geoffrey Miller, New York University


Professor Geoffrey Miller from New York University was invited on December 13th at Escp-Europe business school to talk about a particularly burning topic, namely « the regulation of bank capital ». Some of us were here to discuss this issue and make you benefit from this conference.

  • 1. Why do we need regulation? Some reasons

Professor Miller began this conference by recalling the problems associated with debt as a contract. The bankruptcy costs are of course the first one. What he calls the ‘risk taking problem’ is another one. While managing a company, managers take risks. If the risks are to pay out, then shareholders get some money and debt holders get their money back. If the risks don’t pay out, then shareholders don’t bear all the costs: they share them with debt holders. Thus debt can be seen as an insurance: if losses arise then shareholders pay for a part of them (as much as the level of capital of the firm) and debt holders have to make up for the rest. If the market has enough power (that is, enough information on the one hand, and enough means of pressure on the other hand) then it can easily deter the company from taking too much risks by setting higher interest rates. This is perfectly the case for the general companies, Professor Miller stated: when a firm don’t have enough capital compared to its debt, it tends to pay more interests on its debt. But this is a too simple reasoning for banks, Professor Miller says.


Banks have three special features when comparing them to usual companies. First, they are the only companies having an explicit insurance on part of their liabilities, namely deposit insurance. Deposit insurance provides indeed a guarantee for the depositors to get back their funds in case of bankruptcy[1]. This scheme, largely discussed in the literature, has the shortcoming to encourage depositors not to monitor the banks (the famous problem of moral hazard deriving from insurance). A second special feature that applies for the larger banks is the implicit insurance resulting from their size: the fact that they are considered “too big to fail” makes them less sensitive to the risk of insolvency. This will result in a low elasticity of the interest rates paid on the debt to the leverage of the bank (assets / capital). A third special feature is that the leverage of banks can generally, according to Professor Miller, be higher than the one of usual companies, in so far as their expected return is, in general, easier to predict than the expected return of any other company. However, Professor Miller pointed out that this analysis doesn’t hold when banks have risky and non-transparent assets, as it is still the case currently.


These three specific features help understand why we need regulation. We could summarize as follows the main reasons for having a strong government regulation for banks:

– weak private regulation (mainly because of the insurance described above)

– high leverage

– important external costs of failure (the financial crisis demonstrated this argument…)


  • 2. From Basel 1 to Basel 3: a short review of the improvement of the regulation

Professor Miller then drew a review of the evolution of the Basel regulation.

The first leverage ratio that came up with Basel 1 in 1988 was in the following form:

Capital / (Credit) risk-weighted assets

or more formally;

capital / sum (alpha * asset i)

where just four levels of risks (α) were distinguished, depending on the nature of the asset. This oversimplification regarding the risk classification was the main deficiency of Basel 1 as a capital regulatory tool: the risk weight for every OECD country bond was the same, that is, Turkish bonds were considered to bear the same risk as German bonds. This system was to be improved and new capital regulatory agreements arose with Basel 2.


Basel 2 had many objectives, putting aside the regulation of bank capital (enhancing disclosure requirements to strengthen market discipline and improving the bank supervision framework were for example some of these other objectives). It enhanced the Basel 1 capital regulation in several ways:

– it took into account the operational and market risks in the above ratio

– it allowed for more precise weights regarding the risks of banks’ assets (more possibilities for α in the above ratio)

– it took into consideration a ‘wonderful tool’ -as considers Professor Miller- namely the analysis of banks regarding their assets’ risks. This ‘wonderful tool’ can indeed be used by the bank to compute the risk-weight α of any asset in the regulatory leverage ratio (besides, this is known as the FIRB and AIRB approach[2]).

– it added a distinction between core capital and capital by setting different ratios depending on the type of capital (details on the relevance of this distinction can be found in Acharya et al (2010))


Nevertheless, the crisis brought to light the remaining shortcomings of Basel 2. Professor Miller summarized these deficiencies as follows:

– excessive reliance on credit ratings

– inaccurate risk weighting (the crisis indeed showed that tail events[3] were not as scarce as any risk modeler could think)

– the required capital is too low: many banks encountered several difficulties during the crisis, which means that the required capital was at least ‘too low to go through this crisis’

– the ratios are pro-cyclical: it tends to be higher when the economic situation is thriving and lower when it worsens (when risks get higher, α gets larger)

– it lacks of a leverage ratio (see the exhaustive article of Blum (2008) on this issue[4])

– too much complacency associated with it, according to Pr. Miller


The Basel 3 capital regulation framework deals with most of these issues: it tries to take into account tail events in the capital requirements (with measures such as stressed VaRs[5] for example), to address pro-cyclicity (with capital cushions requirements), to deal with systemic risk (additional capital requirement for the SIFIs[6] for example), it includes a leverage ratio (3% of Tiers 1 capital) and it increases the overall capital requirements.


  • 3. What about the future?

Basel 3 seems therefore to have a lot of advantages for financial stability when looking at the past. From this view, we could be tempted to think that Basel 3 would save us once and from all from future financial stability. But this would be overestimating our knowledge about the factors affecting capital needs: we do not know what an optimal level of bank capital is for the present situation, and even if we knew it, there is no reason to believe that this optimal level of capital will depend on the same factors in the future. The conclusion is well summarized by Pr. Miller “Basel 3 fights the last war, but what about next war?”.


A short session of questions concluded the conference. We can regret that neither the debate regarding the economic costs of Basel 3 nor the political international issues regarding the adoption of Basel 3 have been addressed.


NB: we should keep in mind that capital regulation is just a single part of the regulatory framework: it must be associated with other reforms to have a deep overview of the potential efficiency of the regulation. The recent measures tackling with the size of the banks and the links between their different activities (known as Liikanen report, Vickers report, Dodd Frank Act and Volcker rule…) will have impacts that can’t be ignored while thinking about capital regulation.



To go further on this issue, here are some interesting articles:

Acharya, Gujral, Kulkarni and Shin (2011) ‘Dividends and bank capital during the financial crisis of 2007-2009’ NBER Working Paper No. 16896

Blum (2008) “Why Bassel 2 may need a leverage ratio restriction” Journal of Banking and Finance

Anat R. Admati, Peter M. DeMarzo, Martin F. Hellwig, Paul Pfleiderer Fallacies (2011) “Irrelevant Facts, and Myths in the Discussion of Capital Regulation: Why Bank Equity is Not Expensive” Stanford Graduate School of Business Research Paper No. 2065


[1]Although with a certain limit, which is set around 100 000€ per account in France for example

[2] To go more into detail, a bank using the AIRB approach calculates more factors on its own than a bank using the FIRB approach does (as the loss given default and the exposure at default)

[3]Tail events are events considered as ‘very scarce’ when modeling risks

[4]A leverage ratio can be considered as a second ‘safety net’ (if the risks happen to be wrongly modeled) and put a floor to the leverage of banks

[5]The Value at Risk (VaR) at a 5% threshold for example can be described as the value of an asset under which the real value of this asset has 5% of chance to fall beyond. There are many ways of computing VaR, the simpliest being the historical method where we use the past values of an asset to compute the VaR. A stressed VaR relies on simulations about the possible values of an asset under a ‘stressed scenario’, where the financial system as a whole is stressed.

[6] Systemically Important Financial Institutions




Plus d’analyses