Colorado State University’s Everitt Real Estate Center stands by its earlier report that homes in the Denver-metro region gained only 7.1 percent from 1999 to 2009. A number of top real estate officials have been skeptical of the findings, as reported on Monday by InsideRealEstateNews.
Steve Laposa, director of the Everitt Real Estate Center, defends the study completed on behalf of the Colorado Association of Real Estate, noting that the center analyzed the data on more than a million homes sales from 1999 to 2009 and that its rigorous analysis avoids some of the “biases” by groups such as Case-Shiller, which can make the market look better than it is because it excludes certain type of home sales, such as condos. Also, a government report does not include homes with subprime loans, Laposa notes. The CSU methodology, by contrast, provides an “unbiased view of the market,” Laposa says. He added that it may shake up the “status quo” from time time.
Following is Laposa’s response to the critcisms leveled against the report, in an e-mail to InsideRealEstateNews sent Wednesday afternoon.
“This response is in regard to your recent post “CSU-CAR study draws skepticism, disdain,” dated August 9, 2010. Based on the content of the article, comments from selected interviewees, and the interpretation of the EREC-CAR HPI study results, we believe there are clear misunderstandings and incorrect interpretations of the residential home price modeling methods discussed in your posting. Specifically, we want to address the following:
1. Why it’s important to understand methodology differences between various home price indices, and
2. Why it’s important to compare ‘apples to apples’ home price trends with empirical market statistics and not necessarily anecdotal and geographic-limited data
Research methodologies differ
It is important that your readers understand the basic differences between the various indices such as the S&P/Case-Shiller (Case-Shiller), Federal Housing Finance (FHFA), CoreLogic, and the EREC-CAR HPI, as well as the differences between house price indices based on repeat-sales or paired-sales methodology versus average price and median price calculations. For example, the EREC-CAR HPI for the Denver Metropolitan Region does not include Elbert, Park, Gilpin, and Clear Creek counties, which are included in the Case-Shiller and FHFA indices for the Denver metropolitan statistical area (MSA). It is also important to note which indices: (1) include only conforming loans, (2) include or exclude foreclosures, or (3) create separate indices for attached or detached housing. Factors and methodologies, such as an index with a high composition of urban housing versus one that includes rural and mountain homes, will result in significant differences in the calculation of home price changes.
Million home sale examined
We stand by our research, which is based on a large body of empirical data, and we follow standard industry and academic modeling in developing our analysis and reports. The basis of the EREC-CAR HPI is over 1 million Colorado home sales from the past decade. The primary author of the EREC-CAR HPI, Dr. Sriram Villupuram, is a PhD in Finance from Arizona State University and an expert in housing economics and real estate research who has developed housing models based on over 4 million records from the Phoenix market. Dr. Villupuram is currently working on housing research with the Federal Reserve Branch in Chicago. Our Residential Director of research has over 15 years of residential consulting and home building experience. We are not, as one of your respondents posted, “…these students dealing with number stats they probably do not even understand.” Ironically, your subsequent posting, “Home sales fall 28%,” lends credibility to the findings in the EREC-CAR housing study. What happens to pricing if supply is basically constant and demand falls 28%? Perhaps, as your subheading states, the “…market falls off the cliff.”
An important difference between the EREC-CAR HPI and Case-Shiller and FHFA is the inclusion of certain property and transaction types. The EREC-CAR HPI incorporates condominium sales that are not included in the calculations of either the S&P/Case Shiller Index or FHFA. According to Standard & Poor’s Case-Shiller Website, “The factors that determine the demand, supply, and value of housing are not the same across different property types. Consequently, the price dynamics of different property types within the same market often vary, especially during periods of increased market volatility.”
Market rose 35% from 1999 to 2005
What the EREC-CAR HPI does illustrate is a 35 percent increase in home prices in the Denver Metropolitan Region from 1999 to 2005, which incidentally corresponds with the 38 percent increase in home prices reported by Case-Shiller and FHFA during this same time period. The variation between the EREC-CAR HPI and the other two indices is the extent of the decline in home values from 2006 to 2009.
Other studies have flaws
Why the differences over the last few years? FHFA indices are positively biased due to the fact that they create an index of home prices simply using confirming loans, which has a price limit and conservative FICO scores. As a result, these homes tend to perform well as an asset group. They do not include homes financed using subprime loans, homes that were purchased with cash, attached properties and high-end homes. The reason behind this is that FHFA has access to Fannie Mae and Freddie Mac’s data as their regulator. Case-Shiller index only uses detached single family residences and they have relatively done well compared to the condos during the last 3 years. They also filter out the sale of certain large and small homes from their sample. Even though it is not clear from their documentation of their methodology, studies have found that they filter out top and bottom 5% of the transactions for the Phoenix MSA. In the last three years, due to the restrictive nature of credit for higher-end homes, their sales prices have been significantly lower than that of their asking or previously sold price. When an index excludes those adversely affected homes, a positive bias is introduced.
We are extremely grateful for the opportunity to work closely with the Colorado Association of REALTORS® (CAR) and we respect the REALTORS® who are in the trenches every day, many with decades of experience. If home prices really appreciated 25% to 75% from 1999 through 2009 as stated in your article, we challenge your interviewees to produce empirical data to support their alleged housing appreciation rates. In his critique of the research, Mr. Lane Hornung, an owner of Cohomefinder.com and president of 8Z Real Estate, incorrectly referenced gains in home prices for the Denver area as reported by Case-Shiller and FHFA. Mr. Hornung noted, “That the widely followed S&P/Case-Shiller index puts the overall gain in Denver-area prices from 1999 to 2009 at 45 percent, while the Federal Housing Finance Agency puts it at 52 percent.” In reality, the Case-Shiller Index reports the change in home prices for the Denver-Aurora MSA was 27.94 percent from December 1999 through December 2009, far less than 45 percent. In addition, the FHFA reports a 37.97 percent change in home prices in the Denver-Aurora-Broomfield MSA from 1999 to 2009, not the 52 percent as erroneously quoted by Mr. Hornung. The differences in the numbers are partially explained by differences in geographic coverage, selection bias of sales transactions, and composition of residential property types. The impact of foreclosures and real-estate-owned (REO) property transactions is significant. Our HPI models include REO sale transactions to more accurately reflect pricing trends in the marketplace.
We acknowledge an honest mistake in our initial study that transposed two region names in the narrative portion of the report and we appreciate you bringing it to our attention. The mistake was promptly corrected and you were sent a revised report immediately with an invitation to contact us if ERECyou had any further questions. Nonetheless, a large portion of the article was focused on the minor transposition. (Editor’s note: CSU did not respond to two follow-up e-mails from InsideRealEstateNews.com detailing concerning concerns raised by area Realtors on Monday, or another request to respond on Tuesday morning.)
Research helps Realtors, consumers
The Colorado State University Everitt Real Estate Center is committed to providing unbiased real estate research for the State of Colorado; we are a land-grant university that takes our mission seriously. We recognize that at times our research will challenge the status quo, but we need to maintain our independence and high commitment to quality applied research. There will be debates about home price appreciation rates in the future; we expect to participate in that open dialogue. There are lessons to learn from the recent national, regional, and local housing crises that started in early 2007. It is our goal, with the support of CAR, to develop sound research that helps REALTORS® and consumers alike, to better understand both short and long-term trends in housing values.”
Steve Laposa, Director of Everitt Real Estate Center< class="related_post_title">Related Posts:>