 |
 |
 |
CMHPI Frequently Asked QuestionsGeneral frequently asked questions regarding the CMHPI
Frequently asked questions regarding the series extension:
Frequently asked questions regarding OMB Metro Area definitions:
Frequently asked questions regarding the purchase-transaction only CMHPI series:
The acronym "CMHPI" stands for Conventional Mortgage Home Price Index. The computation of the index is based on mortgages that were purchased or securitized by Freddie Mac or Fannie Mae since January 1975. These mortgages are "conventional" in their financing: they are not insured or guaranteed by any federal government agency such as the Federal Housing Administration or Veterans Administration. Although not specified in the name, the index is based on mortgages for single unit residential houses only; it does not reflect condominiums, multi-family or commercial properties. Finally, the mortgages are "conforming": at the time of purchase they met Freddie Mac or Fannie Mae underwriting standards, and they did not exceed the allowable loan limit set for the two companies. The conforming loan limit is revised each year based on a Federal Housing Finance Board survey. The 2007 loan limit is $417,000 except for mortgages originated in Alaska, Hawaii, Guam and the US Virgin Islands, where it is $625,500. Home loans above this dollar limit are called "jumbos"; the CMHPI does not reflect such homes. The CMHPI uses a statistical method based entirely on "repeat transactions". Any time a house's value is observed twice over time (via either a sale or an appraisal), the change in the price contributes one observation of house price growth over that time period. The index is defined to be the statistically determined set of values that most closely fits many such repeated observations. Mathematically, this is equivalent to taking complicated weighted averages of all the observations of house price growth. For a more technical description of the method, please see the
article from the Journal of Housing Research 6(3):389-418 (1995)
titled "Conventional Mortgage Home Price Index1 How large is the repeat transactions database that the CMHPI is based on? As of the current release, the combined Freddie Mac-Fannie Mae database contains more than 33 million of these "matched pairs". When or how often is the CMHPI released? New index figures are released about 2 months after the end of every quarter. Are the values in the CMHPI ever revised? Yes. Due to the nature of the statistical method used, revisions occur every time the CMHPI is released. The index values are essentially huge weighted averages. One property that averages have is that any time new observations are added, the average changes. As a result, each quarter when new index figures are computed, the values for all previous quarters may change. These revisions are not "corrections" in the usual sense (say, for seasonal effects), but rather reflect the variable nature of averages. How long has the CMHPI been in production? Freddie Mac and Fannie Mae jointly developed the CMHPI and began releasing it in the first quarter of 1994. How different are the CMHPI and OFHEO's HPI? For most purposes, the differences are immaterial. OFHEO is the Office of Federal Housing Enterprise Oversight, Freddie Mac and Fannie Mae's safety and soundness regulator. In 1996 OFHEO began publishing a data series referred to as the House Price Index (HPI). OFHEO uses the same joint database of Freddie Mac and Fannie Mae repeat transactions as a starting point and employs essentially the same statistical method used to calculate the CMHPI. As a result, differences between the CMHPI and OFHEO's HPI are extremely small relative to the differences between either index and the various other house price indices that are publically available.
Why did Freddie Mac extend the national index back to 1970? As technology has developed, our ability to match loan transactions has improved and the quality of the data has improved with better systems. Because of this, Freddie Mac now believes that we can provide a national index that extends back an additional five years, which enables researchers and others interested in housing market information to assess the impacts of an additional business cycle on home values. This information is important because between 1970 and 2007 there have only been five recessions and housing is so important to the economy.
Why aren't the regional indices extended back to 1970? While the sample size used to calculate the national series is sufficient for reasonable statistical robustness, the sample sizes get too small when the data are divided by Census division. For this reason, Freddie Mac has decided to extend only the national sample.
Why did Freddie Mac change its definition of metro areas for the MSA series? The Office of Management and Budget (OMB) defines metropolitan and micropolitan statistical areas for purposes of collecting, tabulating, and publishing Federal data. Metropolitan and micropolitan statistical area definitions result from applying published standards to Census Bureau data. In June 2003 the U.S. Census Bureau released new metro area definitions in accordance with the 2000 Decennial Census data to replace the standard metropolitan area definitions that had been based on the 1990 Decennial Census. Software vendors have recently made geocoding software available with the new definitions. More details of the OMB bulletin can be found at the following website: http://www.whitehouse.gov/omb/bulletins/fy2007/b07-01.pdf
Starting with the 2Q 2005 release of the CMHPI Freddie Mac has decided to use the new metropolitan area definitions for its MSA series – however, a micropolitan series will not be made available due to data samples that are too small for statistical robustness in those areas. Publishing the MSA series using the new metropolitan area definitions will allow researchers to match these data to other data series now being published under the new OMB definitions. The new MSA index series file has 352 metropolitan areas and 29 metropolitan divisions (smaller areas within larger metro areas). Freddie Mac is also making available a separate file with indices for the 11 large metropolitan areas from which the 29 metropolitan divisions are obtained (e.g., the metro area of Washington DC-MD-VA-WV was subdivided into two metro area divisions of Bethesda-Frederick-Gaithersburg MD and Washington-Arlington-Alexandria DC-MD-VA-WV) with a mapping between the divisions and their respective metro area.
Why do some MSA series values show “na” instead of a number value? As part of Freddie Mac’s review of the MSA series of the CMHPI, we put in place two sample size requirements to assure a minimum level of statistical robustness to the estimated values. These requirements are that there be at least 1000 matched transaction pairs for an MSA cumulatively and that there be at least 10 matched transaction pairs in each quarter. When an MSA fails either test a value of “na” is used for that quarter’s value.
Will Freddie Mac continue to release the MSA series with the old metro area definitions? No. In all releases after the 2Q 2005 release of the CMHPI, Freddie Mac will only be publishing the MSA data using the new metro area definitions. In the 2Q 2005 release only, both the new and the old metro area definition series will be released.
Why did Freddie Mac decide to publish a purchase-transaction only CMHPI series? Market transactions are the best way to measure changes in values, and Freddie Mac believes that researchers and other people with an interest in home price trends will find value in having a series based solely on home-purchase transactions. When the CMHPI was originally developed in 1994, the decision was made to include data from all Freddie Mac and Fannie Mae loan purchases, whether the loan was for a home purchase or for refinance so that there would be enough observations for the CMHPI to be statistically robust. The repeat-transactions index regression methodology that the CMHPI uses requires many more observations than most regression models for the values to be reliable. Although there are several known potential problems with appraisal information, good appraisal information is a valuable source of market prices and improves the index when samples are small. There are three reasons why samples may not be sufficient for estimation of the CMHPI in the absence of appraisal data from refinance transactions: small population in an area, a depressed market with few homes sales occurring, or a market where prices are too high, like in San Francisco, to have many home purchases backed by mortgages below the conforming loan limit, as required for loans purchased by Freddie Mac and Fannie Mae.
Why does Freddie Mac also have the CMHPI classic series? For two important reasons – First, for continuity since so many researchers and businesses use the classic CMHPI in their work and understand its properties. Second, because appraisal data is also a valid observation of market conditions. Researchers can use the two series to try to predict market turning points, or may find that one series works better for their needs. For these reasons Freddie Mac will continue to publish both series.
Why didn't Freddie Mac also create a purchase-transaction only series for states and metro areas too? At this time the sample sizes for these smaller geographical areas are not viewed as sufficient to support a purchase-only series.
|
|
|
