Set Up for Failure

W. Scott Frame, Kristoper Gerardi, and Joseph Tracy write,

The extinction of the private subprime market and the rapid rise of the government insurance programs may strike many as a largely positive development. After all, it was the subprime segment of the mortgage market that triggered the global financial crisis and subsequent Great Recession as subprime loans defaulted at an astronomical rate during the housing bust. However, while government-insured mortgages are typically underwritten with more rigor and discipline than private subprime loans, they are not low-risk loans. The combination of high leverage and low credit scores documented above translates into extremely high default rates. The table above shows that five-year cumulative default rates (CDRs) by year of origination varied between 5 and 25 percent over our sample period. To put these numbers into perspective, the five-year CDRs associated with loans insured by Fannie Mae and Freddie Mac (the housing government-sponsored enterprises, or GSEs) are typically an order of magnitude lower. According to our calculations, the 2002 and 2009 vintages of GSE loans had five-year CDRs of approximately 2 percent, while Ginnie Mae’s same vintages had five-year CDRs of almost 10 percent and 13 percent, respectively.

Pointer from Mark Thoma.

The Federal Housing Administration, FHA, sets up many borrowers to fail. One could argue that these borrowers put up so little of their own money that this is a worthwhile risk from their point of view. It is the taxpayers that are being set up to fail.

Reverse Mortgages

An NYT story says,

it may surprise you to learn that some community bankers are quietly offering the loans, too, bringing a kind of Main Street respectability to a product that has long lacked it.

Pointer from Tyler Cowen.

Here is how I think of a reverse mortgage.

Step 1. Sell your house to the bank.

Step 2. The bank rents your house back to you.

Step 3. The bank forgives the rent, but instead charges you interest that accumulates until you die or move out.

Step 4. When you die or move out, the bank adds up the accumulated unpaid rent and interest. If you want to pay it up, you can get your house back. Otherwise, if the debt is higher than the value of the house, then it makes more sense to let the bank keep the house.

In principle, whether this works out financially depends on how long you live in the house relative to expectations at the time you take out the reverse mortgage. You want to live so long that the value of living rent-free in step 3 is so high that at step 4 you or your heirs gladly hand the bank the house rather than pay all that rent and interest. But if you move out or die relatively soon, the bank will have priced the mortgage in such a way that if you or your heirs pay off the debt the bank will come out ahead–and it will come out ahead even more if you give up the house.

Thus, as in any sort of life insurance or annuity situation, you are making a bet against the financial institution. My guess is that this is unwise.

1. In general, the individual loses bets against financial institutions. I tell my daughters, “remember that insurance companies always price to make a profit.” My point is not that you should never buy insurance of any kind. But you should always at least consider self-insuring (rather than paying for extended warranties, for example) or alternative ways of insuring.

2. I think that old people are inclined to over-estimate how long they will live in their houses. They do not like to think about how they might lose the ability to climb steps, to tolerate bad weather, or to live independently.

3. I do not think that many old people need to live rent-free in the short run. In the short run, you can just take out a regular mortgage and use some of the proceeds from the mortgage to meet the payments. Ten years from now, after you have used up most of the proceeds on making the payments and financing consumption, you can think in terms of selling the house. By that time you will probably want to. See (2).

Upward-Sloping Demand Curves

Why are big cities becoming expensive places to live? One answer is that they have good jobs and restrictions on housing construction. That may be right.

But one possibility I want to throw out there is that people want affluent neighbors. If I want an affluent neighbor, and an affluent neighbor is going to live in a neighborhood with high prices, then in some sense I want to live in a neighborhood with high prices. In the extreme, this makes my demand for neighborhoods upward-sloping. Higher prices make me want to live there.

I first considered this possibility many years ago when thinking about school vouchers. I thought that if what people really want for their children is to have them go to school with affluent children, then vouchers would not work as well. Instead of allowing non-affluent parents to send their children to good private schools, the result would just be that good private schools would raise prices so that only affluent children can attend.

I also think that some colleges that are not in the top tier may face upward-sloping demand. George Washington University, which is hardly an academic icon, may benefit from charging very high tuition. Affluent parents come and see a student population that is predominantly affluent, and this gives them comfort that sending their children to GW is a high-status thing to do.

Back to cities. Suppose that an important “urban amenity” is having a lot of affluent people around. Young singles may wish to meet potential marriage partners who are affluent. People who have acquired affluent tastes (sushi, yoga, wine) may want to be around people with similar tastes.

If that is the case, then there is not much that a mid-sized midwestern city can do to lure affluent people. The cost of living there is not high enough to create a barrier to non-affluent people living there. And that means that affluent people will not want to live there.

The Construction-worker Bottleneck

Conor Sen writes,

From a labor slack standpoint, the pool of potential construction workers is probably well-represented by unemployed men under the age of 55. To get back to late ‘90s levels of male unemployment (from a level standpoint, not an unemployment % standpoint), we would need essentially every single male unemployed worker who finds a job in the coming years to go into construction.

Generic pointer from Tyler Cowen. Like Kevin Erdmann, Sen sees us having to build a lot of housing to make up for under-production from 2008-present.

I am not sure there is a housing shortage. A lot of people my age have too much house. That is why they do not mind having their kids move back in with them. It could be that what we need is not so much a surge in construction as a redistribution of housing.

Indeed, Mark J. Perry writes,

In 2015, the average size of new houses built in the US increased to an all-time high of 2,687 square feet (see dark blue line in top chart above), and the median size new house set a new record of 2,467 square feet (see light blue line in top chart). Over the last 42 years, the average new US house has increased in size by more than 1,000 square feet, from an average size of 1,660 square feet in 1973 (earliest year available from the Census Bureau) to 2,687 square feet last year. Likewise, the median-size house has increased in size by almost 1,000 square feet, from 1,525 square feet in 1973 to 2,467 last year. In percentage terms, both the average and median size of new US houses have increased by 62% since 1973.

Combined with a decline in household size, this means according to Perry that living space per person has nearly doubled.

Because we need more space to store our cookbooks, vinyl records, maps, encylopedias, radios, and photo albums.

Indulging in Confirmation Bias

For my view of the housing bubble. John Geanokoplos and others wrote,

Notice that if we freeze leverage (LTV) at constant levels, the boom gets dramatically attenuated, and the bust disappears.

This statement is based on a simulation of an “agent-based” model for house prices. Pointer from Eric Beinhocker from Mark Thoma.

Beinhocker writes,

rather than predict we should experiment. Policymaking often starts with an engineering perspective – there is a problem and government should fix it. For example, we need to get student mathematics test scores up, we need to reduce traffic congestion, or we need to prevent financial fraud. Policy wonks design some rational solution, it goes through the political meat grinder, whatever emerges is implemented (often poorly), unintended consequences occur, and then – whether it works or not – it gets locked in for a long time. An alternative approach is to create a portfolio of small-scale experiments trying a variety of solutions, see which ones work, scale-up the ones that are working, and eliminate the ones that are not.

American pragmatist John Dewey also thought that technocrats should take an experimental approach. That is not a new idea. (I learned this from Jeffrey Friedman, who sent me a draft from his forthcoming book.) Of course, my view is that I would rather see experiments come from the market than from technocrats.

Later, Beinhocker writes,

A major challenge for these more adaptive approaches to policy is the political difficulty of failure. Learning from a portfolio of experiments necessitates that some experiments will fail. Evolution is a highly innovative, but inherently wasteful process – many options are often tried before the right one is discovered. Yet politicians are held to an impossibly high standard, where any failure, large or small, can be used to call into question their entire record.

I would argue that avoidance of failure is natural in any large organization, not just government. That is why I think that markets are better able to conduct experiments to solve problems.

I found Beinhocker’s essay interesting. However, if we are going to try to improve economics, it is important to include behavioral policy-making and politics into the analysis. Do not simply assume a benevolent, rational technocrat as decision-maker.

Speaking of confirmation bias, a recent Instapundit post linked to an old essay of mine, one which speaks to this comparison between expertise mediated by markets and expertise mediated by government.

Housing Finance Policy to Maximize Harm

A reader pointed me to this Urban Institute forum on housing finance policy. It is a topic that leads me to be bitter and uncharitable. I take the view that America’s housing policy in practice is to subsidize demand and restrict supply. There is no economic model that would argue for this, but it makes perfect sense from the standpoint of public choice theory (i.e., policy driven by interest groups)

Since, as usual, nobody asked me for my opinion, I will give it on this blog. My contribution will be to list what I think are the most harmful ideas and hope that they are not implemented. These are:

1. Restrict supply as much as possible by encouraging rent controls, the strongest tenant rights possible, Nimbyism, making environmental regulations difficult to comply with, encouraging additional environmental challenges by private groups, loading builders with expensive requirements (such as setting aside a percentage of their developments for “affordable” housing), etc.

2. Encourage as much housing speculation as possible by subsidizing mortgages with low down payments and by giving the same subsidies to investors as to owner-occupants.

3. Instead of going back to the “originate to hold” model or the Freddie-Fannie model, invent a whole new complicated approach to mortgage finance that mixes government support and private sector activity. This will ensure privatized profits and socialized risks. Above all, creating a new system will ensure that no one has enough experience to be able to manage risk or even to observe where it is being concentrated.

Idiosyncratic Housing Market Perspective

Kevin Erdmann, whose comments on this blog are much appreciated, wrote

There wasn’t even a housing boom. We all just decided to freak out about the one type of homebuilding that was growing – single family units for sale – and ignore every single other category of housing supply, which included homes built by owner, multi-unit homes, and manufactured homes. All of those categories had been in decline. Of course, it was the decline that created the illusion of a boom, because it was precisely those cities where we can’t build, yet where income opportunities are available, where home prices were skyrocketing, because households were bidding up the stagnant pool of homes in those cities in an attempt at economic opportunity in a country that has become inflexible.

What I think he is saying is this (and I could be wrong in my characterization):

1. Because of natural and artificial constraints on supply in cities like SF, the housing stock stays just about fixed, so any increase in demand shows up in price.

2. People have to live somewhere. When supply is fixed in some places, some households get pushed to other places. However, the rise in supply in those other places was never much ahead of demand.

3. If there were excess supply, we would expect rents to fall, and they have not.

4. The sharp fall in house prices came from tightening mortgage credit by much more than was necessary.

My own thoughts:

1. A fact that is salient to me is that the share of mortgages for non-owner-occupied homes went from about 5 percent before 2004 to at least 15 percent in 2006. To me, this says that at the margin there was some demand that was not driven by housing needs. Also, I believe that in housing the marginal supplies and demands exert big effects on prices, even though those marginal Q’s are small relative to the stock of housing and the total number of household.

2. Another salient fact is that the average price-to-rent ratio also shot up over this period.

3. This suggests to me that something other than “pure” supply and demand was at work in driving up house prices. I am inclined to see some combination of looser credit and (unrealistic) expectations for house price increases.

4. I think that Kevin is right to stress that the characteristics of housing markets differ in different locations. In SF or DC, rapid gentrification combined with restricted supply gives you one dynamic. (I don’t think I would pin it all on people bidding for “an attempt at economic opportunity,” as if these cities offer better jobs to people of every skill, which is what Enrico Moretti has claimed. Instead, I see a shift in economic opportunity inside cities away from low-skilled workers and toward professionals in the New Commanding Heights sectors.) In rural Ohio, a long decline in economic opportunity gives you another dynamic. In Texas, a big population inflow with more elastic supply gives you yet another dynamic. I could imagine that national averages, including the national averages I tout as “salient facts,” could be quite deceiving. Perhaps to understand the whole you need to study the parts.

My Song and Dance

I am back from a trip to Norway and Denmark. The impetus for the trip was an invitation to speak to a group of alumni of the Business Institute in Oslo. My topic was property bubbles.

The morning of my talk, I took a walk along the beautiful path next to the Akers River. My mind wandered, and at one point I tried to recall the steps to this dance, which I had learned recently and only done a few times. My mind also wandered to my talk, and so when I gave it I used the opening 25 seconds of the dance while singing (details below the fold).

Anyway, a lot of the questions were good ones. My favorite was when a guy asked if Norway would be less subject to a bubble because its mortgage loans come with recourse. In most states in the U.S., you can default on your mortgage and owe the bank nothing other than the keys to the house. With recourse loans, after the foreclosure sale, you still owe the bank any deficiency between the sale price of the home and the outstanding balance on the loan. I agreed with the thrust of the question, because recourse loans make people think twice about making speculative home purchases. I certainly think that recourse loans change the dynamics of the housing market. Canada has recourse loans, and that may be one reason that they did not suffer the bubble-and-crash that we did.

In my answer, I also pointed out that the underwriting criteria shift when you have recourse loans. You can be a bit less concerned about the property itself as security for the loan and a bit more concerned with the borrower’s capacity to repay the loan (income and assets) and with the borrower’s conscientiousness (as typically measured by a credit score).

For all the Dodd-Frank legislation and all the mumbo-jumbo about shadow banking, it really would have been simple to prevent the financial crisis of 2008. Recourse on mortgage loans would have done it. In effect, recourse loans give mortgage borrowers more “skin in the game.” Instead, Dodd-Frank tried to envision giving mortgage brokers more skin in the game.

And, of course, removing government support for investor loans probably would have done it as well.
Continue reading

Ed Pinto Takes on Mom and Apple Pie

He writes,

The 30-year fixed rate mortgage should be retired — for good. Despite continued proof that it fails to build up wealth for the most disadvantaged Americans, and that mortgage debt should not be a burden as homeowners approach their 50s and 60s, misguided advocates maintain that the 30-year fixed rate mortgage should be at the core of the U.S. housing finance system.

I have been telling friends that the tragedy of this year’s election is that in my opinion this is a time when we need to change course in many ways, and the Trump phenomenon reduces the probability of that happening. One example is housing finance policy, where Democrats are always looking for an excuse for government to intervene, and I think any reasonable evaluation of history would say that it would be better to go in the opposite direction.

I am willing to claim that the whole financial crisis of 2008 could have been prevented with just one innocuous reform: take away any government subsidies for investor loans.

Subsequent research has confirmed that in the regions with the most pronounced price cycles, significant shares of home purchases were by non-owner-occupants. Their mortgages were eligible for purchase by Freddie and Fannie. They were eligible to receive favorable capital treatment when packaged into securities that were rated as safe by rating agencies.

Investor loans do nothing to promote the goal of home ownership. They are riskier than loans to owner-occupants. Why did regulators allow Freddie and Fannie to buy them, and why did they allow them to be eligible for favorable capital treatment? Probably because the powerful mortgage lobby was at work behind the scenes.

Speaking of the mortgage lobby, Pinto is probably correct that we do not need the 30-year fixed-rate mortgage. Canada does fine without it. But don’t hold your breath waiting for any reform that might be guided by his point of view. That is one example of what the Trumplosion of the Republican Party has made impossible.

What did Dodd-Frank Reform?

In this book compiled by the Heritage Foundation (long PDF), Ed Pinto writes (p. 33),

a home-purchase loan that qualifies under QM could have a 580 FICO credit score, no down payment, and a 43 percent DTI. A loan with these characteristics acquired by Freddie in 2007 had a 42 percent failure rate under the adverse conditions that prevailed between 2007 and 2012.

In case anyone thought that having the government set credit standards for mortgages would allow one to sleep at night.

At this point, I am completely pessimistic about the future of housing finance in America. There is a powerful lobby at work to maintain policies that subsidize demand and mortgage indebtedness. And the prospects for electing someone with ideological opposition to government involvement in housing finance are quite dim.

Overall, I find the book painful to read. It shows the extent to which Dodd-Frank produced costly, counterproductive policies, based on misguided diagnoses.