Friday, April 30, 2010

Russ Roberts and Joe Stiglitz on the Crisis

Russ Roberts, an economist at George Mason University, recently finished an essay on the incentive structures that caused the financial crisis, which he's been working on for several months now. While I don't agree with Roberts's policy recommendations, I think he presents a very good and clear way of thinking about financial regulation. It's a typology that I've also heard Joe Stiglitz use (he comes to different conclusions), so I thought it was worth sharing with you. This is my spin on what these two men have said, of course - it's extremely general. Please take it as a policy typology and not a policy position.

We can simplify and think of two possible types of policy for finance: "regulating the upside" and "insuring the downside". Of course many policy options fall into these categories, and it's a lot more complicated - but this is their basic framework. "Regulating the upside" is far dicier an endeavor and a lot more falls under that umbrella. Restrictions on interstate banking, capital requirements, disclosure requirements, distinctions between commercial and investment banking, and regulation of specific products like derivatives all fall under this category. So I want to make clear that just because Stiglitz (and I) advocate "regulating the upside", that doesn't mean that everyone in this camp supports all the same policies. It's just a good way of thinking about our options. "Insuring the downside" is somewhat easier to think about. This includes things like deposit insurance (FDIC), arguably the Fed discount window's "lender of last resort" function, post hoc bailouts (and, as Roberts emphasizes, the presumption of post hoc bailouts), and even the "Greenspan Put". So with these two policy types, we can think of four basic options: (1.) don't regulate or insure, (2.) regulate, but don't insure, (3.) insure but don't regulate, and (4.) regulate and insure.


Both Roberts and Stiglitz have argued that one of the major causes of this crisis was that we insured against losses for financial institutions in a variety of ways, leading them to take dangerous risks. Roberts would have us eliminate these insurance measures so to reduce excess risk taking. Stiglitz would keep the insurance measures but regulate the activities of financial institutions more stringently. They each made their case to the Joint Economic Committee in December, which you can watch here.

What's interesting and nice is that there is such broad agreement that a disproportionate emphasis on insuring losses is dangerous. But there are reasonable critiques of the other options as well. Roberts's solution of removing is rejected by almost all economists, and for good reason. Policies like deposit insurance were put in place in the first place because excessive losses tend to exacerbate crises. Roberts is right that when someone covers your losses you will take more risks than you otherwise would. But he cheerfully ignores or glosses over the fact that when you risk suffering large losses from the mistakes of others or systemic downturns, that dampens investment activity and encourages people to take too few risks.

Stiglitz argues, contra Roberts, that the solution isn't to unlearn all of the lessons about the importance of insuring certain downside risks. His point is that you insure, but then you regulate the excessive behavior that a few will take, knowing that their losses are insured. This avoids the biggest dangers of an erratic credit cycle that Stiglitz suggests (and I agree) would happen with Roberts's plan. The danger, of course, is that there are a lot of regulations of the upside that we could conceive of, and not all of them are good. Government employees and private employees are made of the same stuff, and both make mistakes. Excessive regulation can dampen economic growth.

And of course there's also a lot of common ground. Debt is privileged over equity in the tax code, and I think fixing that would have broad appeal for both camps. "Too big to fail" distorts the idea of loss insurance, substituting ad hoc corporate welfare. Insurance is about covering expected losses - "too big to fail" is about "it's easier to ask forgiveness than permission". Both sides want to take measures to guarantee that the implicit bailout is no longer implicit.

Ultimately, though, Stiglitz and Roberts come to wholly opposite conclusions from roughly the same initial critique of our current regulatory apparatus. The common ground only goes so far, but I think it's interesting that they're working from the same basic typology. The choice between Roberts and Stiglitz ultimately boils down to a question of the volatility of the market vs. the heavy handedness of government. Both are quite real concerns, and we do ourselves a disservice by ignoring either one of them. While I personally fall in Joe Stiglitz's lower-right hand quadrant of the typology I've presented, I definitely have my differences with him, as I certainly do with Roberts as well. Generally speaking, I worry that Roberts doesn't pay adequate attention to the market failures that I described, and Stiglitz doesn't pay adequate attention to the potential government failures.

Before closing, I'd like to provide a link and a disclaimer. First, I want to refer people back to my post on Hyman Minsky (which now has some very insightful comments from F&OST guest, Sebastian). My basic critique of Roberts is that he misses the Minsky insight of financial fragility. Indeed, to the extent he recognizes this sort of fragility, he lauds it as a virtue. In many cases it is a virtue. There is absolutely no doubt that failure is functional - that destruction is creative. The market is successful precisely because firms and individuals fail. That's not where Roberts disagrees with economists like Stiglitz (although Roberts would probably like you to think that's where the disagreement lies!). The point of disagreement isn't the necessity of letting people fail. The point of the disagreement is that Stiglitz sees both functional and unfunctional kinds of failure, whereas Roberts rarely mentions the type of destruction and failure that isn't functional or creative.

My disclaimer is that I haven't read Roberts's new essay yet. However, he's been working on it for a while, and I've read and extensively commented on numerous blog posts from Roberts on the issue and even on this essay while it was in its formative stages. I also did listen to his entire December testimony. So take this more as general thoughts on Russ Roberts's position, rather than on this essay in particular.

No comments:

Post a Comment

All anonymous comments will be deleted. Consistent pseudonyms are fine.