Arnold Kling  

Impossible Mission Commission

The Worst Solution to the Fina... Taking Ezra Klein's Health Car...

Keith Hennessey writes

Yesterday Senate Minority Leader Mitch McConnell (R-KY) appointed me to be a member of a new Financial Crisis Inquiry Commission...The purpose of the Commission is "to examine the causes, domestic and global, of the current financial and economic crisis in the United States."

I can remember my father telling me that a question such as, "What caused the first World War?" cannot be answered scientifically. We cannot run a controlled experiment to verify a hypothesis The financial crisis poses a similar challenge.

I could come up with a very long list of controlled experiments that would help sort out difficult issues. Off the top of my head....

1. An experiment where we change housing policy to get rid of the housing bubble. Does a bubble simply emerge somewhere else?

2. An experiment where monetary policy follows the Taylor Rule exactly. Does it make much difference?

3. An experiment where there are no bailouts. What happens to the financial sector, and what is the impact on the nonfinancial sector?

4. An experiment where government does not promote mortgage securitization. In particular, bank capital regulations are not written to allow banks to hold less capital for mortgage credit risk when the loans are bundled into securities and still held by the bank. How much of the fancy financial engineering that took place over the past decade was driven by regulatory capital arbitrage?

5. An experiment where exchange rates adjust to reduce capital inflows to the United States. Does that prevent a bubble from emerging?

6. An experiment where European banking is less concentrated. To what extent was the crisis exacerbated--and the bailouts motivated--by the fragility of banking systems in other countries?

7. An experiment in where Glass-Steagall restrictions are maintained, rather than gradually eroded starting in the 1970's. I don't think that this would have made much difference, but some people are obsessed with the issue.

8. An experiment in where we have a "systemic risk regulator." I don't think it would have made one ounce of difference, but the conventional wisdom in the regulatory community seems to be that if we just go back to where we were five years ago and add a systemic risk regulator, we'll be fine.

Comments and Sharing

COMMENTS (6 to date)
rod writes:

All such attempts go hard aground on the Law of Large Numbers, no?

fundamentalist writes:

"I can remember my father telling me that a question such as, "What caused the first World War?" cannot be answered scientifically. We cannot run a controlled experiment to verify a hypothesis The financial crisis poses a similar challenge."

Hayek and Mises would say your definition of science is to narrow. Science is more than controlled experiments. Check out Hayek's "The Counter-Revolution in Science".

mark writes:

The rest of the commission is not very impressive. Wouldn't a better name be "A Committee Composed of a Proportionate Amount of People Liked by Democrats on Another Committee and a Proportionate Amount of People Liked by Republicans on such other Committee To Spend Taxpayer Money Synthesizing What Has Already Been Written by Somebody Else Somewhere on the Web."?

Mark Seecof writes:

I like your #3 (no bailouts). I suggest you add:

#9-- An experiment in which bond-rating agencies are not an oligopoly sponsored and guarded by government, and are paid by bond insurers and buyers instead of issuers.

#10-- An experiment in which the government does not prompt lenders, whether by stick or carrot methods, to relax underwriting standards for anyone, by race, income level, neighborhood, or otherwise.

Colin K writes:

#11: An experiment in which investment bankers' bonuses are paid entirely in equity vesting over a 10-year period.

Mike Rulle writes:

Terrific topic. This applies to all social science. What does one do? Some look to history and comparable situations. But here one can easily be accused of cherry picking the data and the historical comparability--(witness the Romer-Barro debate during Stimulus debate). Are there any Macroeconomic events that are probabilistically believed to be caused by "X" to economists of opposite ideological views?

Hayek's way of thinking has always appealed to me too. But isn't it possible that "spontaneous order" is just a statement of the obvious? I don't think so, but saying the sun rises from the east is also a prediction. But how do we know that free markets supported by consistent and enforced property laws etc., produce the best result and that under certain conditions, one should just leave it to the entrepreneurs? When do we not?

We do have some experiments on the mega scale that are very compelling. East versus West Germany. North versus South Korea. HK/Taiwan versus China (circa 1980s and before). We know extreme Socialism versus relatively free market capitalism is likely to lose. So why not on a continuum too?

I think the best macro-arguments are made on the micro plane. So, why is government intervention worse than free markets (need operational definition, but thats not that hard)? Well, we compare the "returns" of the 2 forms of investments. Simple ones are US Post Office versus UPS; Or Amtrak versus Greyhound. Less precise might be "energy investments by government versus private sector".

It is hard for me to imagine how the government ever looks good in these comparisons. Ever since the Clintons started using the word "investment" this seemed like a "Gary Hart" style open invitation to make such comparisons. But I never see these arguments in political or economic literature. Perhaps they are there, but they never make it to the public arena. Why make theoretical arguments, when one can just make actual empirical comparative ones?

The recent crisis was the sum of everything put together. That is tautology. But, if "micro" style arguments can be made against (or for, which I am obviously dubious of) govt. interventions, then perhaps logical deductions could also be made on specfic events, like the financial crisis.

I referenced Jeff Friedman's recent essay on this site before. He discusses, among other things, the enormous complexity of tracking down ultimate causes for specific events, like the recent crisis. I would rather we focus on issues that can be studied which can have an effect on at least the policy debate. My thoughts above are what strike me as plausible.

Comments for this entry have been closed
Return to top