Wednesday, January 12, 2011

Speculative Conjecture on Regulation and Crisis

. Wednesday, January 12, 2011

Finally someone else is writing about the politics of financial regulation, so you know I'm all over this one. Riffing off of Reinhart & Rogoff (R&R) at The Monkey Cage, David Andrew Singer asks how we could know whether this time is different:

Financial crises happen regularly throughout history; indeed, R&R's analysis makes the enduring pattern of boom and bust perfectly clear. However, the book explains very little about the patterns it presents. The authors "select on the dependent variable" by examining only cases of countries in crisis, and as a result they are unable to make causal inferences. Are large capital inflows the root cause of the current financial crisis, and did they cause earlier crises throughout history? With no variation in the dependent variable, we cannot discern whether the alleged macroeconomic triggers are the real culprits, or whether other factors are at work. Examinations of past financial crises tell us little about whether or when current account deficits lead to financial instability, or about the political and institutional factors that might militate against systemic market failures.

The point here is that because R & R only look at crisis periods, and not non-crisis periods, it is impossible to know for sure whether the macro variables are really causing the crisis. For example, even if most crises occur when countries have high current account deficits, there are many periods when countries have high current account deficits and yet do not have financial crises. Looking only at crisis periods, we might conclude that current account deficits cause crises. Looking at all periods, not just crisis periods, we (might) conclude that they do not. Or we might find a conditional effect: rapid depreciations in the current account cause crises, but more gradual depreciations do not. Or maybe levels matter more than changes, or maybe there is an interactive effect between the current account balance and levels of national debt. Etc. This is why research is hard.

But then Singer goes on a harangue:

Despite the challenges of causal inference, many social scientists seem content to attribute the financial crisis to underlying macroeconomic imbalances. In my view, these arguments provide useful cover for financial regulators who might otherwise be held accountable for their rule-making. If systemic failure is the periodic and ineluctable result of global capital cycles, then there is little reason for regulators to enact tougher regulations. Why should central banks and regulatory agencies impose more stringent capital requirements and prohibitions against risky investments? If the roots of the crisis are macro-structural, regulators feel no incentive to alter the rules. How else can we explain why U.S. regulators, when facing the public or their overseers in Congress, speak of recent bank failures as if they were exogenous acts of nature? ...

From a research design perspective, a reasonable way forward is to test hypotheses about the conditional impact of capital inflows on the probability of financial crises in the developed world. The scope and quality of regulation are likely contenders for inclusion in such a model. The cases of Australia and Spain suggest that large capital inflows might be less destabilizing if the banking system faces strict capital requirements and prohibitions against non-traditional banking activities. Other possible conditioning variables include, inter alia, resource endowments, partisanship, and corporate governance.

Singer has done a lot of good research on regulation (as I am trying to do, now), so it makes sense that he thinks it's important. I do too. But as far as I know there is no persuasive research showing that crises are the result of poor or lax regulation either. At least not of the "large-N", quantitative variety. One of the reasons for this is that cross-national time series data on financial regulations are rare and incomplete. But even a cursory look across the past few years of history suggests it's more complicated than that. India was recently regarded as a paragon of financial stability... until it wasn't. In fact, the U.S. had stricter regulations pre-crisis (at least in terms of capital adequacy and leverage) than many other major economies, including Germany and Japan, and it's been the U.S. (and U.K.) pushing for tougher standards in the Basel negotiations.

Despite some harmonization through the Basel Accords, national regulatory standards still vary quite a lot. Nevertheless, the subprime crisis spread throughout the system in ways that make national regulatory standards look largely irrelevant. At least, there's no clear pattern to me. What best predicts exposure to crisis (to me) seems to be how integrated a country is in the international financial system, especially to the U.S. and U.K., and how large the banking sector is relative to GDP. Countries that were very tightly connected and had large financial sectors (e.g. Iceland, Ireland, the U.S. and U.K.) have suffered quite a lot. Those that were less integrated into the system or had smaller financial sectors (e.g. Australia, Brazil) have done better.

Singer mentions Spain as a positive example, which is curious because Spain is in the middle of a banking crisis right now that has already required massive public sector bailouts, and that appears to be intensifying and now is threatening to turn in a sovereign debt crisis. (Which is exactly what R&R would predict, by the bye.) This despite the fact that, as Singer notes, Spain has a relatively strict regulatory structure.

Considering a time series leaves us even more unsure. In the U.S., Glass-Steagall didn't prevent the crises revolving around the Latin American debt, savings and loans, and dot-com bubbles. The Basel Accords obviously didn't do their ostensible job either, either in the 1990s or 2000s. The U.K. had more financial instability before "light touch" than after, until 2008. It's not that crises are exogenous shocks; it's just that the regulatory structure doesn't appear to tell us too much about how susceptible to crisis countries are either. Of course we don't know that for sure, because data limitations have limited our ability to rigorously tests these claims. But it isn't obviously true, even if it is intuitive and intellectually appealing.

I think Singer is correct about his main point: R&R doesn't tell us everything we need to know. There are certainly intervening variables that are important, and strength of regulation may be one of them. In fact, I would be very surprised if regulation had no effect on stability or instability. But we also need to remember that regulations are political creations; there is no ex ante reason to believe that regulatory codes are even primarily intended to promote stability. Regulations affect distribution, and that makes them inherently political rather than technocratic. Indeed, much of the Dodd-Frank reform bill was about either punishing financial institutions, protecting consumers, shuffling regulatory authority, or winding-up failed institutions. Not a whole lot of it really concerned stability, and it didn't say much of anything about capital, liquidity, or leverage.*

Jeffrey Friedman has argued that specific parts of the regulatory code -- those privileging mortgage debt, asset-backed securities, and OECD sovereign debt -- made the system less stable. These parts of the regulatory code had an explicitly political purpose: to encourage home-ownership, especially among the lower classes, and to make access to bond markets incredibly cheap for governments. This ground is ripe for political economists.

Singer concludes with this:

Until we conduct more rigorous tests, social scientists will remain in the uncomfortable position of offering only speculation and conjecture. The availability of hundreds of years of data on previous crises should not give us a false sense of confidence about our capacity to explain. Indeed, we simply do not know whether this time is different or not.

I completely agree. I would argue that this applies equally to claims about the effects of regulation as well which are, as of now, at least as speculative and conjectural.

*Title I concerns financial stability, but mostly reorganizes already-existing regulators and tasks them with monitoring and addressing systemic weaknesses. It doesn't create any substantive new statutory requirements of banks.


Emmanuel said...

Pretty interesting except for one thing:

Problems over (by definition small) microloans in one Indian province (Andhra Pradesh) do not constitute a nationwide banking crisis.

Keywords: MICROloans; provincial ≠ national regulation

Kindred Winecoff said...

Fair enough, although it hasn't been contained to just one province.

To some extent India is prevented from having a nationwide banking crisis by not having a nationwide banking sector.

Speculative Conjecture on Regulation and Crisis




Add to Technorati Favorites