Arnold Kling  

My Paper on Housing Finance Reform

Systematically Biased Beliefs ... Ken Rogoff and Others...

I write,

Any attempt to reengineer a housing-finance system with a new set of government-guaranteed entities would entail all of the risks of restoring the existing GSEs, plus more. The taxpayers would be exposed to similar potential hazards, but with new and inexperienced organizations engaged at the level of enterprise management and regulatory oversight.

A sentence that may surprise:

Before we bury Freddie Mac and Fannie Mae, we should praise them.

Overall, I suggest either going back to some version of the S&L model or some version of the GSE model. In either case, the focus should be on minimizing taxpayer risk, not on trying to tilt the housing market in some particular direction. If you are going to comment, first read the paper.

The paper is part of a compendium called Five Proposals for New Housing Finance System in the United States. Perhaps it should say six, given that my paper includes two.

Of course, what a lot of the papers are doing is saying, "Let the market work," which means no real concrete proposal. We can make guesses as to how the market will adjust, but we cannot play the game that a more statist reformer could play, of saying exactly what our plan is supposed to do.

For example, Dwight Jaffee thinks that private securitization will come back stronger than ever. Peter Wallison is equally confident that the mortgage market will do fine without a government guarantee.

However, the way I see it, too many investors, particularly foreign investors, have been spooked by mortgage securitization. Yes, in theory, all that the securitizers need to do is tighten underwriting standards, but will the investors trust them to do that? My guess is that if we do not have Freddie, Fannie, or some other government guarantor, the advantage in mortgage lending will shift back to banks and away from Wall Street. Nothing wrong that. In my paper, it's the preferred outcome.

Unlike me, Lawrence J. White would be willing (although this is not his first choice) to try a new form of government guarantee. As I understand it, the government would basically match private mortgage insurance, with the government's portion priced in line with the private portion. I can think of plenty of ways that the private mortgage insurers could game this, and what worries me is that there are ways that I have not thought of that also would work.

Last but not least, Michael Lea and Anthony B. Sanders offer an eclectic vision for a post-GSE future. That is, they see a role for banks, for private securitization, and for changes in mortgage products.

We wrote our papers independently. As a result, there is considerable overlap and redundancy. All of us see the arguments for a government role in the mortgage market as very weak. All of us see the private sector as capable of adapting to fill the gap left as Fannie Mae and Freddie Mac are phased out. All of us see a return of higher down payments and stricter underwriting standards as the key to creating a mortgage market that investors can trust.

There is even a fair amount of overlap between our papers and The Treasury Report provided by the Obama Administration. All of us can envision a world without Freddie and Fannie. However, I worry that some new guarantee mechanism will be created that is ultimately even more dangerous.

All of these proposals run counter to the narrative that the private sector cannot be trusted to deliver a safe mortgage system. Our proposals assume that the private sector has an incentive to avoid taking unreasonable risks. One could argue that this is simply a reversion to the view that Alan Greenspan said that he held mistakenly prior to the crisis.

I would put it slightly differently. I have very little faith that the private sector will always avoid taking unreasonable risks. However, I have even less faith that regulators will have the wisdom to see the risks that private agents are missing.

Remember that up until very late in the game, the criticism that Congress and regulators were directing at lenders was for turning down too many borrowers. What the government was pushing for was more lending to the "under-served" segments of the market.

Yes, there can be market misbehavior. But regulators may well, as in this case, amplify the problems rather than dampen them. This illustrates the aphorism that "Markets fail. That's why we need markets."

Comments and Sharing

COMMENTS (4 to date)
fundamentalist writes:
Of course, what a lot of the papers are doing is saying, "Let the market work," which means no real concrete proposal.

Good point! The market can "work" properly only if it is free. The less free it is, the less it works the way economics says it should and the way people want it to work. Today no free market in finance or housing exists and we will never have a free market in them. The regulations are simply too numerous. Getting rid of a few guarantees does not create a free market in housing or finance. And it's impossible to predict what the next unintended consequence of regulation will be or when it will hit. But when it does, I can guarantee that everyone will blame the market (as Greenspan did) and not government regulation.

I have very little faith that the private sector will always avoid taking unreasonable risks.

I have full confidence that the private sector avoids unreasonable risk. People are naturally risk averse. There are too many studies out there showing that people fear loss ten times more than they enjoy gain. What I don't have confidence in is the ability of people to know when they are taking unreasonable risks.

In a free market, it's difficult enough to when one is taking unreasonable risk. But in a market in which the government has distorted prices beyond recognition through its intervention, it's nearly impossible to calculate risk at all.

"Slapped by the Invisible Hand" by Gary B. Gorton is an important book, and he has many papers on the net summarizing his work. If Gorton knows anything at all, then it's clear the investment banks did not know they were taking unreasonable risks. They had done their best to reduce risk and thought they had done so. Their problem was schooling in bad monetary theory.

Chris Lemens writes:


Could you explain why you believe that the market response to phasing out the GSE's as suggested in your second proposal would not be something like the following?

A small set of large national banks hire in the people who ran the GSE's underwriting functions. They create strict underwriting guidelines like those that you propose for the GSE's. They make loans using them, then securitize most of the loans, retaining a tranche for themselves. Over time, they build brand value with investors in the securitized instruments.

One objection would be that the tranche to be retained would presumably be have the most credit risk. (That is, tranche 1 would soak up with first 10% of losses, tranche 2 gets the next 10%, etc.) But, since these are deposit-taking institutions, they can rely on the government deposit guarantee to attract funding for that riskiest of tranches. Thus, the objection would seem not to have much force. The investors in the securitized assets won't care if the government has to pay off the bank's depositors.

(If you wanted to stave off that problem regulatorily, one could also require that the issuer notionally securitze the entire set of loans in a single tranche. Then, either the government or the market could require that the issuer buy some percent of it. But that's not really my question.)

Clearly, you have thought about this stuff more then me. So I'm curious why you believe that something like the above would not evolve.


Various writes:


I agree with everything you say. You do, however, go on at some length to proactively defend your recommendations, both in this post and others. I interpret this as a signal that you are somewhat unsure about their efficacy. As someone with a slightly different point of view, I can say that I think this is a case where you can err on the side of being bold, and state a bit more forcefully what you believe the solutions are. I think it is somewhat obvious that many of the folks opining on this subject (other than you and other bloggers of integrity) are stakeholders in this fight seeking to protect their turf. I believe the remedies are relatively obvious, and are some derivative of the following:

1. Reduce leverage for gov't guaranteed loans (i.e. larger down payments)
2. More stringent capital requirements for lenders (i.e., "dumb regulation")
3. Break up some of the larger private institutions.
4. Keep Fannie/Freddie, but subject them to greater oversight ("smart regulation")

In summary, we have a political problem here, not a theoretical one.

Arnold Kling writes:

I believe that the market will always seek to concentrate the risk with the taxpayers. The scenario you describe is one way to do that. However, if the regulators were to employ a stress test methodology, rather than mechanical "risk buckets," the stress tests would expose the level of risk that banks are taking. This could then be used to assign appropriate deposit-insurance premiums and capital requirements. But if regulators set themselves up to be gamed, as they did under the Basel accords, then you are correct.

Another key, as I point out in the paper, is separating government guarantees from high-risk mortgages. If the underlying mortgages require 20 percent down payments and do not include cash-out refis, the securities are very unlikely to become laden with risk.

Comments for this entry have been closed
Return to top