Is raising required bank capital the answer?
Basel I and II were attempts to connect the riskiness of bank assets on one side of the balance sheet, with the available capital cushion on the other side. Banks fail and become wards of the state, so the reasoning went, because they make bad loans. By making private equity holders take the first loss, regulatory capital requirements were supposed to limit the incentive to take risk ex ante, and so also the cost of government bailout ex post.
From this point of view, the recent financial crisis looks like a problem of insufficient capital buffer, and the solution looks like a larger buffer (Anat Admati et al 2010). One reason the Basel buffers were inadequate, from this point of view, is that there were a lot of off-balance sheet exposures which incurred no capital charges at all. The true risk exposure of banks was not, in the end, connected to the available capital cushion. The solution is to bring off-balance sheet exposures back on to the balance sheet, and to raise the capital cushion required to hold those exposures.
But this is not the only way to think about the problem; there are other analytical entry points that lead to different conclusions.
Basel I and II were supposed to eliminate regulatory arbitrage between more and less highly regulated banking systems; in fact they created a new incentive for regulatory arbitrage between the more regulated bank-based credit system and the less regulated capital market-based credit system. By 2007, the latter had become the source of the majority of credit in the United States (Adrian and Shin 2009).
Instead of starting analytically with the bank-based credit system, it may make more sense to start with the market-based credit system. It is true, ex post, that the collapse of the unregulated “shadow banking system” wound up absorbing the capital cushion of the regulated traditional banking system. But that was by no means clear ex ante. Indeed the whole idea was that the shadow banking system was supposed to be separately capitalized.
Just so, consider the equity tranche of a collateralized debt obligation. It was supposed to absorb the first loss, and so protect the higher tranches; that’s capital that is not measured by the bank-centered Basel approach. Indeed, when doubts began to arise about the quality of the underlying assets, it seemed likely that holders of those equity tranches, mostly credit-focused hedge funds, would be the hardest hit (Morris 2007). But those fears proved largely misplaced.
A more important source of capital turned out to be the capital of the insurance companies that sold credit insurance of various kinds on the best assets in the shadow banking system; that is also capital that is not measured by the bank-centered Basel approach. AIG is only the most famous example, because its enormous capital cushion got wiped out by the crisis. It was not the failure of the equity tranches at the bottom of the capital structure that brought the system down but rather the failure of the AAA-rated tranches at the top of the capital structure.
From this point of view, the regulatory response to the crisis should focus not on the insufficient regulatory capital in the traditional banking system, but rather on the failings of the credit insurance contracts in the shadow banking system. That is what broke. Probably AIG was underpricing its insurance; definitely it was under-reserving for it.
But the mechanics of AIG’s failure reveal also the key importance of liquidity factors, which are overlooked by Basel I and II. It was the mark-to-market collateralization feature of AIG’s CDS that brought it down, and triggered the Fed’s involvement. Had AIG been selling CDS on an exchange, rather than bilaterally, the exchange would have stood in between AIG and its counterparties (Goldman Sachs, Societe Generale, and others). As we move CDS onto an exchange, we need to be thinking about how to keep that exchange from suffering the fate of AIG.
The capitalization of the regulated banking system is not the central problem. The problem is the capitalization, and also the liquidity, of the shadow banking system.