Saturday, September 5, 2009

France Drags Its Heels

The Financial Times sees a rift opening in the G20 between the US and the UK on the one hand and France on the other. The issue is how to regulate the financial system to prevent a repeat of the crisis of 2008-9. To simplify, France favors regulating individual incentives by limiting bonuses to traders and bankers. The US and UK want to shore up bank capital positions by requiring higher reserves and lower leverage ratios.

There is a problem with the US-UK position: it tends to be procyclical. In a downturn, bank assets diminish in value, and with stringent capital regulations in place they are more likely to be forced to sell to raise capital, thus further decreasing asset prices and making everyone in the system worse office. Yet there is no doubt that before the current crisis, capital reserves were too low, so that even a modest fall in housing prices raised doubts about bank solvency. So there is ample reason for the US and UK to take the stance they do.

By contrast, France, in the person of Christine Lagarde, is arguing that the Basel II regulatory framework, agreed to by banks after 20 years of negotiation, should suffice. Basel II relies less on capital requirements and more on so-called Value at Risk (VaR) models to set limits on bank borrowing. The problem with Lagarde's position is that the VaR models in widespread use did not perform well in the crisis. These models, which rely in essential ways on the banks' own representations of the risks involved in their financial instruments, depend heavily on statistical analysis of historical performance data. The banks hit hard by the crisis were already using VaR models for their internal risk management, and these models predicted that they had nothing to worry about. In a widely quoted statement, one Goldman Sachs risk manager said that market movements at one particularly bad stage of the crisis were "5 standard deviations beyond what the model predicted" for several days running. Hence there was something wrong with the model.

Obviously improvements will be made in light of post-crisis data, but still it seems imprudent to rely, as Lagarde seems to want to do, on such modeling alone. Capital requirements are conservative, old-fashioned, potentially pro-cyclical, and increase the cost of capital, but perhaps one needs to cling to them until the VaR concepts at the heart of Basel II have been more thoroughly tested.

See also Nicolas Véron, here.

UPDATE: Looks like the US-UK version prevailed.


Leo said...


there is some truth to your contention that higher capital requirements may be pro-cyclical but the advantages far outweigh the risks. In fact, higher capital requirements especially the kind which is considered now (higher levels for higher risk assets) will have a deterrent effect and limit the development of risky, potentially toxic products. Together with a better regulation of some products, such as CDOs, they should put a severe brake on the inflation of banks balance sheets.

The salient points of the recent crisis were the mad leverage the system used to boost its profits (especially in Europe) and the moral hazard which prompted the taxpayer to shoulder the risk while banks and their executives continued to reap huge rewards.

Higher capital requirements (and other measures as so called living wills) will hopefully reduce the excess risk built in the system.

As Lord Adair Turner (the British banking regulator) recently stated, the size of the financial sector has been preying on the rest of the economy by crowding out the high skills jobs market and capturing an excessive portion of the rewards. Higher capital requirements should foster a reallocation of resources that will benefit society.

In this respect, Sarko's focus on bonuses is trying to cure the symptoms (disgusting as they sometimes may look) without addressing the root causes of the disease. As for madame Lagarde's promotion of Value at Risk models, les bras m'en tombent.

Arthur Goldhammer said...

Yes, I agree completely. Thanks for your comment. I wasn't defending the French position, just trying to say that it did have some potential rational basis.