# Secondary market.

SECONDARY MARKET

Every November, mortgage lenders all around the country anxiously await the result of the Federal Housing Finance Board's (FHFB) survey indicating the October-to-October increase (or decrease) in the national average one-family house price. The result of the October survey by the FHFB (formerly done by the Federal Home Loan Bank Board) provides the basis for calculating the statutory conventional mortgage limits for Freddie Mac and Fannie Mae mortgage purchases of one- to four-family residential loans (i.e., "conforming loan limits.") The FHFB surveys the purchase price reported on conventional loans closed in the first five days of the month by commercial banks, savings and loan associations, savings banks and mortgage companies responding to the FHFB survey.

In the 10-year period since the first conforming loan limit adjustment became effective January 1, 1981, there have been eleven adjustments. The average increase over the period has been 6.8 percent with the highest increase being 15.6 percent in 1985 and the lowest being a .01 percent decrease in 1990.

The FHFB survey includes both fixed- and adjustable-rate loans, loans on newly built and previously occupied houses and loans with various loan-to-value ratios. The "average purchase price" used for conforming loan limit adjustments is not an average for the year. Nor is it the median (i.e., mid-point between the high and low) purchase price often cited in the newspapers. It is simply the average purchase price reported by those surveyed of all conventional loans closed during the first five days of the month in October.

When the FHFB issues its press release on contract interest rates in mid-November, the following method can be used for calculating the new mortgage limits: * Determine the October- to-October change in the FHFB's average purchase price in dollar terms. * Divide that dollar change by last year's base and add one to the resulting figure. * To get the new mortgage limit, take the result obtained in step two and multiply by the current limit and round down to the nearest \$50. Apply the result from step two to the current limits for two to four properties as well. Example
```Step 1: FHFB's average purchase price, October 1990, \$139,600
FHLBB's average purchase price, October 1989, -\$136,800
\$2,800
Step 2: \$2,800/\$136,800 (base year 1989) = .02046 + 1 = 1.02046
Step 3: 1.02046 X 187,450 = \$191,286.7
```

Rounded down to the nearest \$50 = \$191,250 (Limit for 1991).

Typically, one can expect an increase in the average purchase price for October if the year- to-year changes for the other months (e.g., January to January, March to March, etc.) point in the same direction. For example, increases of 10.2 percent (1988) and 14.9 percent (1989) came in years where the annual change for the other months compared favorably to the October figures (based on 1986-87 and 1987-88 comparisons, respective). Such was not the case, however, for the 1988-89 figures when the October 1989 results showed a decline in the average purchase price. In every month of calendar year 1989 (except October) the annual change was positive, even for the following November and December. One FHFB staff person familiar with the survey believes that changing demographics underlying the survey resulted in the unexpected decrease.

So far this year the 1990-91 year-to-year dollar percentage changes for January through September are as follows: January -\$3,400 (-2.3 percent); February \$4,300 (+3 percent); March \$1,800 (+1.3 percent); April \$3,700 (+2.5 percent); May \$3,300 (+2.3 percent); June \$8,500 (+6 percent); July \$4,000 (+2.7 percent); August \$2,600 (+1.8 percent); and September \$7,800 (+5.7 percent). The average increase through September is \$3,622 or 2.56 percent. Every month since last November of 1990 has revealed a higher average purchase price than the one reported in October of 1990. Based on the trend over the past 12 months one could reasonably expect an average purchase price for October to be in the range of a 2.6 percent to 3.6 percent increase. This translates into higher 1992 conforming limits in the range of \$196,200 to \$198,000.

Traditionally, mortgage bankers have been under-represented in the FHFB sample. This has probably led to lower loan limit increases over the years because mortgage companies typically finance higher priced homes compared to banks and S&Ls. The demise of a large number of S&Ls and the greater market share of mortgage companies will make it even more important that mortgage companies respond to the FHFB survey if asked to participate. (For example, the average purchase price reported for September 1991 reveals \$166,500 for mortgage companies; \$131,000 for S&Ls; \$104,500 for commercial banks and \$180,100 for mutual savings banks.)

The FHFB has taken steps to ensure that the sample is representative. Beginning in November 1991, the FHFB will survey loans closed during the last five days of the month because the higher volume of loan closings during that time should be more representative of the universe compared to loans closed in the first five days of the month. The FHFB has hired a private contractor to assist it in manipulating the sample and appropriately weighting the data. Hopefully, efforts like these by the FHFB will ensure that mortgage bankers' growing lending activities are reflected in the survey results.
COPYRIGHT 1991 Mortgage Bankers Association of America
No portion of this article can be reproduced without the express written permission from the copyright holder.