Recently I have been getting a lot of mail from mortgage brokers and small lenders complaining that they are no longer able to get funding for loans that previously were never questioned. One of them described it as a market meltdown. I promised to take a hard look.
Fortunately, I have been developing a database on wholesale mortgage prices that lets me peer into the heart of the problem, if there is one.
Wholesale prices are those that lenders quote to mortgage brokers and small lenders called “correspondents.”
My database covers 11 of the largest wholesale lenders, and because they account for a large share of what is an extremely competitive market, it is safe to assume that their prices are “the” wholesale market prices. While most price differences are small, in every case I take the lowest of those shown.
Wholesale prices have much less statistical “noise” than retail prices because they do not include markups, which can vary widely from one transaction to another. Further, the transaction characteristics underlying the wholesale data are well-defined. I know the loan amount, property value, type of property, state location, purpose of loan, use of property, borrower’s FICO score, type of documentation, whether borrower escrows taxes and insurance, and lock period. None of the available series on retail prices contain this amount of detail on factors that affect mortgage prices.
The data are available for each of 14 separate loan programs, of which six are fixed-rate and eight are adjustable-rate. Data will be available for individual states, and for a U.S. average. Subprime mortgages are not covered, though Alt-A and other intermediate credit categories are included.
To simplify price comparisons, I adjust all rates to zero points and fees. The interest rate is the only price used.
Because this is a work in progress, I don’t yet have the daily series that will permit day-to-day monitoring of the market. However, I did do a test run on May 4 covering California, and will use that as a benchmark for assessing the state of the market three months later, on Friday, Aug. 3.
On cream-puff loans, interest rates rose by about .4 percent between May 4 and Aug. 3. On some programs it was a little more, on others a little less, but the dispersion was small.
A cream-puff loan is one with a 20 percent down payment on a $500,000 single-family home purchased as a permanent residence, by a borrower with a credit score of 720 or more, who fully documents income and assets, and escrows taxes and insurance. My respondents, however, are not complaining about cream-puff loans; their concern is the riskier niches, and the evidence provides support for their claim. The price of risk has gone up.
One of the tables I ran on May 4 showed the rate at different FICO scores. On scores ranging down to 680, rates on Aug. 3 were about .4 percent higher, but at 660 the increase was .56 percent and at 620 it was 1.4 percent. Below 620, there were no quotes on either date, that’s subprime territory.
I also looked at rates on loans of different sizes on May 4: the sizes were $75,000, $417,000, $418,000 and $2 million. The two middle sizes distinguish loans that can and loans that cannot be purchased by the two federal agencies, Fannie Mae and Freddie Mac. The rate increases, starting with the $75,000 loans, were .41 percent, .45 percent, .74 percent and .8 percent.
The only other risky niche I priced on May 3 was a cash-out refinance for investment on a 4-family property, though it had a 20 percent down payment, 720 FICO and full documentation. On Aug. 3, the rate on this loan was 1.24 percent higher.
A risky niche I decided to check out, which I had not priced in May, was stated-income loans with zero down. I found it priced close to 8 percent for a borrower with a 700 FICO, above 10 percent for a borrower with a FICO of 660, with no quotes below 660.
It is clear that the price of risk has risen substantially, the higher the risk category, the larger the increase in price. Some at least of the very highest-risk niches are no longer being offered. The lowest risk niches, what I called cream-puff loans, have not been affected at all and may even have benefited. My data don’t cover subprime loans, but the plausible surmise is that these loans have been affected most of all because they are the riskiest of all.
This is not a market meltdown, far from it. In a meltdown, lenders jettison their capacity to assess risk and flee into assets insured by the government. We haven’t seen that since the 1930s, and we won’t be seeing it now.
This market is correcting a previous tendency to underprice risk, a tendency arising from a prolonged period of house-price appreciation. Steady price appreciation virtually eliminates the difference in performance between the least risky and the most risky loans. Lenders with short time horizons and/or poor memories, who were willing to price on the assumption that price appreciation would continue forever, forced other lenders to do the same to remain competitive.
The fantasy that home prices only rise has been punctured. It is no longer rational to price loans on the assumption that rising prices will convert most bad loans into good loans. The market is now reacting to that realization.
The writer is professor of finance emeritus at the Wharton School of the University of Pennsylvania. Comments and questions can be left at www.mtgprofessor.com.