Regulation Is More Complicated Than You Think

Some of my research involves the regulation of bank capitalization, including the use of risk-weights. One of the things I constantly emphasize is that there are many misconceptions in the academic and policymaking communities about the relationship between banks and regulations. Here, for example, is Per Kuwoski complaining about the international Basel capital accords:

The first fact is that since banks are allowed to hold less capital, and therefore to leverage the risk-adjusted margins more on their capital, and therefore to obtain much higher expected returns on equity when lending to what is perceived as “safe” than when lending to what is perceived as “risky”, current regulations are completely distorting our financial system. 

That has caused banks to create excessive exposures to what was erroneously perceived as risky [ed.: I think he means "safe"], like in AAA rated securities, Greece, real estate, and to refrain from lending to those in the real economy perceived as “risky”, like small businesses and entrepreneurs.

The second fact is that the first fact is not even mentioned, much less discussed.

The point is that prior to the crisis banks were doing what they were supposed to do. The epicenter of the crisis was located in some of the safest categories of financial instruments: OECD sovereign debt and highly-rated securities, both of which were privileged in the risk-weighting scheme under the Basel accords. The crisis didn't occur because of junk bond trading; it occurred in part because everyone (including the banks themselves, apparently) thought banks were acting safely when they were actually concentrating evermore risk at the center of the global financial system. This means that increasing capital requirements under the current regulatory structure is likely to increase the exposure financial institutions have to these types of assets, these types of risk, and thus the sort of crisis that we've just experienced.

Was the risk-weighting system gamed? Of course it was*, particularly by the mid-2000s when the world's demand for "safe" financial assets denominated in dollars massively out-stripped supply. Safe financial assets were defined by the regulatory code, so there was quite a lot of money to be made by creating assets which would be considered safe by regulators and selling them. Was there outright fraud? Some, yes, although it's hard to find solid evidence that this was as pervasive a feature of the financial system prior to the crisis as many assume; it's even harder to demonstrate that fraud led to (or even contributed to) the crisis. It's much easier to see how gaming of the system could have.

But what banks were not doing is "racing to the bottom". That is, they were not bumping up against their minimum regulatory requirements by taking on as much risk as they were legally allowed to do*. Instead, they were piling into "less risky" assets because those were rewarded by the regulatory code. And the more that these types of assets are rewarded (or required) by the regulatory code, the more banks will game the system in increasingly opaque ways. It's what they're being asked to do, after all.

What is to be done? Some, like Kurowski and also Vice-Chairman of the FDIC Thomas Hoenig, want to abandon the risk-weighting system altogether. Their views are represented by new bipartisan legislation which was introduced into Congress by Brown and Vitter last week, which has the support of everyone from community bankers to Simon Johnson. The gist: force banks to fund some percentage of their investments with equity rather than debt, but let them (and their investors and counterparties) determine what assets are risky and what are not. Keep the system simple; the more complicated it gets, and the more beneficial it is to pursue profit through regulatory arbitrage, the more it will be gamed.

But so far the political discussion of this has often reduced to a simple lobbying story: banks don't want to be regulated, and they've got the power, so the regulation will be weak. This is both an incredibly simplistic view of regulatory politics, it is also wrong in at least some cases. Banks don't like some kinds of regulation, it is true, but bank preferences are not homogenous. All regulations have distributional consequences, so some firms will support them while others oppose them. We're seeing that with Dodd-Frank and Brown-Vitter and we've seen it in previous rounds of the Basel accords.

The point is that banks (and other financial firms) have asymmetric interests, and that these interests are conditioned by the regulatory environment. Just making the regulatory environment "tougher" won't necessarily punish large financial institutions (it'll often help them), and may concentrate risk in opaque parts of the financial system. This is arguably the worst possible result. Unfortunately, it may be the most likely under current policy.

*If by "game the system" you mean doing essentially what the regulatory code wanted them to do: invest in sovereign debt and asset-backed securities.

**If you doubt me, I can prove it. Part of the work is forthcoming in peer-reviewed form; other parts will hopefully find a home soon.