KnowHow® News

November 2015: In the Know

How the S&P Case/Shiller Index Should (or Should Not) be Used in Forecasting

By John Jagerson, Learning Markets Analyst1

The news is good for home sellers. According to the S&P/Case-Shiller 20-City Home Price Index data released on April 30, 2013, the price of an average home in the U.S. increased by 9.3% in the 12-month period ending February 2013 and by 0.3% in the month of January 2013 alone. This may be valuable information for potential home buyers as well. But what does it tell those of us who may only be interested in investing in housing-related securities? A closer look at the S&P/Case-Shiller Indices is needed to answer this question.

The Case-Shiller Index Family

The Case-Shiller index series was developed to capture the average market prices of residential real estate throughout the U.S. There are indices for each of 20 individual metropolitan statistical areas (MSAs) as well as 3 composite indices--the 10-City, the 20-City, and the National. The individual MSA indices and the 10-city and 20-city composite home price indices, which aggregate the data drawn from the 20 MSAs, are calculated monthly. The National Home Price Index is based on data gathered from nine U.S. Census regions and is calculated only quarterly. All three composite indices are designed to reflect the market prices of existing single-family homes within the U.S. and track each other fairly closely, as illustrated in the graph below:

View Graphic

Separate indices measure the average price of condominiums in 5 major areas—namely Boston, Chicago, New York, San Francisco, and Los Angeles.


The index calculations utilize the repeat sales method developed by Karl Case and Robert Shiller in the 1980s. Actual sales price data are collected on single-family homes (or condos) within a given region. When that property is resold--regardless of whether it is months or years later—the new price is matched with the previous price to create what is referred to as a “sale pair,” and the difference in the prices is recorded. Thus, a property must have been sold at least twice to be included in the index.

The sale pairs are screened to weed out any transactions that would not accurately reflect the true market value of the properties. For example, non-arms-length transactions, such as transfers of property between members of the same family, are excluded, as are any transactions in which the reported values are unrealistic and a data error is suspected. Bank-owned properties are considered arms-length and are, therefore, included in the calculation.

All the sale pairs that make the cut in a specific MSA are aggregated to create the home price index for that region, and data from the individual MSAs are combined to form the 10-city and 20-city composite indices. Price tier indices are also calculated for 17 of the MSAs. These further separate the sale pairs within the respective region into low, medium, and high-priced single-family homes.

Every effort is made to ensure that the indices capture the average change in home prices within each specific region, assuming constant quality. If the change in price of a sale pair seems abnormal relative to other price changes recorded in that same area, a lower weight will be given to that transaction in calculating the index since the price anomaly may be a result of significant remodeling or neglect.

For the same reason, a lower weight is also applied when the time interval of a sale pair is lengthy. The longer the time between transactions, the more likely it is that the home has undergone some substantial physical changes in the interim.

The monthly indices employ a 3-month moving average algorithm in their calculations, and there is a 2-month reporting lag. The data becomes publicly available on the last Tuesday of each month. In other words, the indices published on Tuesday, April 30, 2013 only incorporate data through February 2013, and the February 2013 index point is calculated by averaging sale pair data collected in December 2012, January 2013, and February 2013.

The Correlation With Housing-Sector Stock Returns

The correlation of the monthly returns of the Case-Shiller 20-City Home Price Index with the returns on the PHLX Housing Sector Index (HGX) is very low: 0.124.

View Graphic

This isn't surprising. Many of the companies included in the PHLX index, such as DR Horton (DHI), Owens Corning (OC), and Vulcan Materials Company (VMC), are in the homebuilding and construction supplies industries. The Case-Shiller indices focus on existing single-family homes; new construction is specifically excluded.

In fact, the correlation of the Case-Shiller 20-City Index returns with the PHLX Index returns is weaker than its correlation with the returns on the S&P 500 Index over the same time period: 0.223.

View Graphic

A slightly stronger relationship is observed between the Case-Shiller 20-City Index returns and the monthly returns on the Wilshire U.S. Real Estate Securities Index. Over the same time period, October 2002 through January 2013, the correlation coefficient of the returns is 0.304.

View Graphic

The Wilshire U.S. Real Estate Securities Index incorporates the prices of Real Estate Investment Companies (REITs) and Real Estate Operating Companies (REOPs) and specifically excludes homebuilders in its calculation. The fact that the holdings of many of REITs and REOPs include residential real estate probably explains the higher correlation of the returns. Neither, however, focuses on single-family homes, which is why the relationship is still not very strong.

Bottom Line

The relationship between the returns on the S&P/Case-Shiller 20-City Home Price Index and both the leading housing sector and real estate securities indices is very weak. Albeit positive, the correlations are low, and a correlation coefficient of zero indicates there is no relationship between the returns.

View Graphic

Given the database that the S&P/Case-Shiller 20-City Home Price Index uses—i.e., price differences in existing single-family home sale pairs—these results are not unexpected.

The results suggest that investors looking to invest in stocks related to the housing industry can derive very little value from the S&P/Case-Shiller home price indices. Recall, too, that there is a 2-month lag between the last index point calculation and the publication of the Case-Shiller index. Numerous studies have concluded that stock prices adjust rapidly to new information that enters the market; they can definitely be expected to have adjusted to information that is already two months old.

There is one message that can be gleaned from the return correlations, however. Given the weak relationship between the returns on the Case-Shiller 20-City Home Price Index, investors can diversify their investment portfolios by investing in single-family housing.

1This content was created and is being presented by an independent party not employed by or affiliated with Scottrade or its affiliates. Scottrade, Inc. and its affiliates have not created and are not responsible for such materials. The content of such materials has not been adopted, endorsed or approved by Scottrade or its affiliates and does not reflect the opinion, belief or recommendation of Scottrade.

The information and content provided in the Scottrade® Knowledge Center is for informational and/or educational purposes only. The information presented or discussed is not, and should not be considered, a recommendation or an offer of, or solicitation of an offer by, Scottrade or its affiliates to buy, sell or hold any security or other financial product or an endorsement or affirmation of any specific investment strategy. You are fully responsible for your investment decisions. Your choice to engage in a particular investment or investment strategy should be based solely on your own research and evaluation of the risks involved, your financial circumstances and your investment objectives. Scottrade, Inc. and its affiliates are not offering or providing, and will not offer or provide, any advice, opinion or recommendation of the suitability, value or profitability of any particular investment or investment strategy.