Public Availability of Freddie, Fannie Loan Performance Data

Todd W. Schneider has a write-up and some analysis.

I decided to dig in with some geographic analysis, an attempt to identify the loan-level characteristics most predictive of default rates, and more. As part of my efforts, I wrote code to transform the raw data into a more useful PostgreSQL database format, and some R scripts for analysis. The code for processing and analyzing the data is all available on GitHub.

I recommend reading the entire post.

House Prices and Incomes

The data in this story suggest that they are most out of line in San Francisco and Los Angeles. That is, the ratio of the median home price to the median income is highest in those two cities, and it is dramatically higher than in the other cities listed (where it also is high).

The price to income ratio exceed 8 in both cities. A normal ratio of price to rent is about 12. A normal ratio of rent to income is about 25 percent. So a normal ratio of price to income would be 3. To get to 8, you need, say, a ratio of rent to income of 40 percent and a ratio of price to rent of 20.

Some possibilities:

1. This is a sign that some time in the next few years home prices will fall in those two cities.

2. It is the rest of the country that is under-priced. People in other cities should be bidding up house prices where they live.

3. Perhaps there is some statistical fluke. Perhaps the discrepancies across cities go away if you break things down to individual neighborhoods. Note however, that the ratio of two medians is a harder statistic to skew than, say, the mean of income. So I doubt that this is the explanation.

4. Foreign ownership plays a larger role in those two cities than elsewhere, so that the inadequacy of median income is not a factor.

5. Supply is more constrained in California than elsewhere. Note, however, that I have a hard time specifying the demand curve that creates such a dramatic ratio of house prices to incomes.

I think that supply constraints are more likely to lead to volatility in house prices than to house prices that are permanently super-high relative to incomes. And in fact LA and SF have a history of high volatility.

Housing and the Punch Bowl

On Wednesday, I appeared on a panel discussing the state of credit underwriting in the housing market. I raised two questions:

1. Are national credit standards, set by Freddie, Fannie, and FHA, appropriate, or do they throw out too much local information?

I made a Four Forces argument that there are too many divergences in economic performance that make local information valuable. On the other hand, you could argue that simply by tracking search data, Google and Zillow have better information on local trends than would an on-site mortgage underwriter. Interestingly, the session chairman, Bob Van Order, presented information showing that after the crash loans under-performed relative to their known characteristics (including ex post home price performance) and over-performed more recently. This suggests that it is possible for underwriting to be looser or tighter than it appears based on observable characteristics, which in a way suggests that there is local information that is important.

2. Are we in 2004? That is, is the stage set for another housing bubble, and all that is needed is a loosening of credit standards?

One of the speakers, Sam Khater of CoreLogic, re-iterated what he wrote here, that “price-to-income and price-to-rent ratios are high.”

Very few mortgages originated since 2009 have defaulted. There are two reasons for this. One is that credit standards were tightened. The other is that the trend of house prices has been up. Now, there is all sorts of talk about the need to loosen standards. I pointed out that both the private sector and public officials tend to be very procyclical when it comes to mortgage credit–when the market is going up, they want to loosen standards, and after it crashes they want to tighten standards.

I would be ok with loosening standards on credit scores now, provided that the industry holds the line on down payments, meaning that we do not see an increase in the the proportion of loans with down payments below 10 percent. This is not the time for the FHA to make a big expansion in its high-LTV lending (Ed Golding, can you hear me?)

To encourage high-LTV lending now would be adding alcohol to the punch bowl just as the party is getting good.

Housing Finance and Recessions

Oscar Jorda and others write,

The rapid increase in credit-to-GDP ratios since the mid-1980s was just the final phase of a long historical process. The run-up started at the end of World War II and was shaped by a long boom in mortgage lending. One of the startling revelations has been the outsize role that mortgage lending has played in shaping the pace of recoveries, whether in financial crises or not, a factor that has been underappreciated until now.

Pointer from Mark Thoma.

When I read this, I wanted to shout “Underappreciated by who?” Maybe by the macroeconomists who were trained by Stan Fischer, Thomas Sargent, and their progeny. But until Genghis Khan pillaged macro, every macroeconomist knew that housing and mortgage credit rationing were major economic forces in the United States. Until the late 1980s, the process generating recessions consisted of interest rates rising, mortgage lenders losing deposits (because of interest rate ceilings), home buyers losing access to credit, and housing collapsing. And every macro economist knew this.

And even if you are too young to know any old-fashioned macro, you could read Ed Leamer. I would suggest that the authors of this essay try searching for Leamer Housing is the business cycle.

What this essay teaches shows to be underappreciated is Google.

Note that there is more to the essay, which Timothy Taylor found worthwhile.

Housing Policy: A Civilized Approach

Joe Gyourko writes,

They should begin by completely phasing out the FHA over some clearly defined period (for example, three to five years) and replacing it with a new subsidy program that would help the two types of households the agency is meant to serve. The new program would help prospective homebuyers amass a 10% down payment that would then allow them to obtain financing at market rates from private lenders. This could be done through a simple system in which qualified households pay into a special savings vehicle and receive some type of match from the government. These funds would accumulate on a tax-free basis until they were large enough to provide a 10% down payment on a home.

On the civilization vs. barbarism axis, this would encourage saving and responsible behavior. If you read the whole article, you will see reinforcement for this. Of course, along the oppressor-oppressed axis, the only thing holding back people from owning homes is oppression, so there is no need for requiring a down payment.

Of course, what matters here is a different axis entirely. The real estate lobby vs. everybody else. And from that perspective, Gyourko’s sensible proposal does not stand a chance. Another indication of the political impossibility of this approach is that it is one that I have long advocated.

Marginal vs. Average Debt to Equity in Housing

Alejandro Justiniano and others write,

if the relaxation of collateral constraints had been widespread, it should have resulted in a surge of mortgage debt relative to the value of real estate.

In the data, however, household debt and real estate values rose in tandem, leaving their ratio roughly unchanged over the first half of the 2000s, as shown in Figure 3. In fact, this ratio only spiked when home prices tumbled starting in 2006.

Pointer from Mark Thoma.

Suppose that back when lenders asked for 20 percent down, three families bought houses for $100,000 each and put $20,000 down each. Total mortgage debt is $240,000 and total home values are $300,000. The ratio of household to real estate debt is 80 percent.

Next, lenders allow someone to buy a house with no money down. As a result, home prices rise to $130,000. Adding $130,000 debt to the debt of the other three households (ignoring any equity they may have built up through paying down mortgage principal), we have total mortgage debt of $370,000. But total home values are $520,000, so that the average ratio of debt to equity has actually fallen, to just over 70 percent.

As long as home prices are rising, the last thing you should expect is for the average debt to equity ratio to rise. The fact that it did not fall is an indication of how powerful the boom in credit was. Only if you use a silly representative-agent model, in which there is no difference between average and marginal borrowers, would you predict something different. I have not read the paper, but I suspect that is what the authors did.

Kevin Erdmann Revisits the Housing Boom

He writes,

the commonly repeated anecdotes of janitors and checkout clerks being
handed $300,000 mortgages on a hope and a prayer do not appear to be representative. On net, all the new mortgages went to families with incomes around $45,000 and higher.

And elsewhere he writes,

growth in homeownership came from high income households and that households didn’t increase their debt payment/income ratios or their relative consumption of housing during the boom. The evidence against the standard narrative is even more stark when we look at dollar levels, because, despite frequent implications to the contrary, low income households don’t tend to take on nearly as much debt as high income households.

Consider two ratios:

1. Debt-to-income.

2. Debt-to-equity.

Many narratives of the financial crisis focus on debt-to-income ratios. The oppressor-oppressed narrative is that greedy lenders imposed inappropriately high debt-to-income ratios on innocent borrowers, who then could not meet their mortgage payments. The civilization-barbarism narratives stress the use of houses as ATMs and the forced expansion of lending toward irresponsible borrowers.

It seems to me that Erdmann is suggesting that debt-to-income ratios did not got out of line, or perhaps they only got out of line for high-income borrowers. He may be right, although I would suggest looking at the Home Mortgage Disclosure Act (HMDA) data and not just the survey of consumer finances.

Ever since Bob Van Order explained the mortgage default option to me almost 30 years ago, I have viewed debt-to-equity as more important than debt-to-income. If we define sup-prime lending in terms of debt-to-income, then I am inclined not to attach much significance to the proportion of sub-prime loans. To me, the dangerous loans are the ones with low down payments. There is some overlap between those and loans with high debt-to-income ratios, but not enough overlap to equate the two.

Let’s take Erdmann’s analysis of debt and income as accurate. I see no reason to change my preferred narrative of mortgage lending and the housing boom. That is, there was a surge in lending with low down payments. I am pretty confident that the HMDA data support this. In addition, there was a surge of lending for non-owner-occupied homes (speculators).

Think of owner-occupants with low down payments and non-owner-occupants as the speculative component in the housing market. My narrative is that the speculative component soared during the housing boom. These speculators did very well until house prices started to level off late in 2006. Then what had been a virtuous cycle on the way up turned into a vicious cycle on the way down, and the speculative buyers got hammered.

Getting from that to a recession is the more difficult part for me, because I do not allow myself to use the words “aggregate demand.” Instead, to explain the recession I have to make a case that many patterns of specialization and trade became unsustainable, or were finally perceived as unsustainable, while new sustainable patterns were difficult to discover. I might argue that the surge in government economic intervention exacerbated the difficulty of discovering new patterns of sustainable specialization and trade. TARP and the stimulus were largely efforts at redistribution, and that gave people a bigger incentive to focus on grabbing some of the loot than on developing a sustainable new business. Of course, Keynesians will tell you that the problem is that the surge in government intervention should have been bigger and lasted longer.

PG County Foreclosure Story Focuses on the N-word

This WaPo story is long, but nonetheless incomplete.

Using court and land records, The Post analyzed 173 home purchases in Fairwood that wound up in foreclosure between 2006 and 2008.

In 43 of those home purchases, borrowers financed 100 percent of the cost of the home with loans that had high interest rates and reset periods within three years. The loans were of the type that Angelo Mozilo, the CEO of defunct subprime lending powerhouse Countrywide Financial, had called “toxic” because they offered such onerous terms. He warned his own company in internal e-mails that the loans were “the most dangerous product in existence.”

Nearly all the remaining loans The Post examined contained features associated with high default rates, such as low or no down payments, interest-only payment periods and higher rates than prime loans.

Only seven out of the 173 defaulters received the most favorable lending terms, known as conventional 30-year fixed interest rate loans. These “prime” loans are the least likely to fail, experts agree.

The neighborhood is described as primarily African-American, with a median income over $170,000.

Some questions that I have:

1. Why were so many loans made with zero down payment? If the median family income is that high, should there not have been higher down payments?

2. If the borrowers put nothing down to begin with, then foreclosure cost them nothing in terms of lost equity. Presumably, if they were affluent before, they are still affluent now. If not, why not?

3. Did anyone benefit from making these loans? The companies that ended up owning the mortgages took huge losses (taxpayers also may have been involved in some way, through bailouts). Companies that originated subprime loans but did not hold them (the “originate to distribute model”) picked up some small fees, but my guess is that they competed away a lot of profits by incurring marketing costs, and in any case enough of them went out of business that you can hardly envy their franchises.

4. When did borrowers start to fall behind on their payments? If it was within a year of buying the home, then you can be sure that even if the borrowers had gotten prime, thirty-year fixed rate loans they still would have defaulted.

Remember, it was the WaPo that said on January 1st that “narrative” is “out” and “facts” are “in.” Their story is instead all about narrative (the N-word, as I call it), and I think it could use more facts.

Peter Wallison and the N-word

He says,

By 2008, before the financial crisis, there were 55 million mortgages in the US. Of these, 31 million were subprime or otherwise risky. And of this 31 million, 76 % were on the books of government agencies, primarily Fannie and Freddie. This shows where the demand for these mortgages actually came from, and it wasn’t the private sector. When the great housing bubble (also created by the government policies) began to deflate in 2007 and 2008, these weak mortgages defaulted in unprecedented numbers, causing the insolvency of Fannie and Freddie, the weakening of banks and other financial institutions, and ultimately the financial crisis.

Remember what the Washington Post Style section proclaimed on January 1st. Narrative is out. Facts are in.

Of course, in addition to the Freddie and Fannie securities, there were lots of private-sector securities backed by risky mortgages. My contention is that this boom was fueled by risk-based capital rules, which stated that once these loans were packaged into securities, divided into tranches, and blessed by rating agencies as AAA, banks could earn three times the return on such mortgages as could be earned by originating and holding an old-fashioned, low-risk mortgage.

How to Make Manhattan More Dense

Shlomo Angel and Patrick Lamson-Hall write,

densities in today’s Manhattan can increase again if we allowed its lower income residents—and lower income, given today’s housing prices, includes its middle income residents as well—to live in more cramped quarters and to consume less floor space per person. As long as public authorities can maintain acceptable elementary standards of health and safety—from access to water and sanitation, to proper ventilation and fire protection—there is no reason to restrict the housing options of lower income residents by mandating a minimum consumption of floor space. A contemporary densification policy may thus entail the removal of zoning and building standards that require minimum apartment sizes, allowing for the construction of micro apartments as well as single rooms sharing common facilities (formerly known as SROs, Single Room Occupancies). It may entail extending legal permission to subdivide larger apartments into smaller ones by furnishing them with additional kitchens and bathrooms. And it may also entail the passage of new regulations that eliminate the exclusionary restrictions now imposed by the boards of cooperatives and condominium associations on the leasing of apartments that are left empty to non-owners, as well as the prohibitions on the rental of rooms on a short or longer term basis.

Most interesting was their demonstration that Manhattan density peaked in 1910, then fell through 1980. Think of the elevator as increasing effective floor space and the subway as reducing the demand for housing right near factories.