The CME Group real estate marketplace covers both U.S. residential and commercial properties. These products can be used by a broad range of market participants, from property owners and manufacturers to hedge funds and investors.
The S&P/Case-Shiller Home Price Index measures the residential housing market tracking changes in the value of the residential real estate market in 20 metropolitan regions across the United States.
The S&P/Case-Schiller Metro Area Home Price Indices originated in the 1980s by Case Shiller Weiss’s research principals, Karl E. Case and Robert J. Shiller. At the time, Case and Shiller developed the repeat sales pricing technique. This methodology is recognized as the most reliable means to measure housing price movements and is used by other home price index publishers, including the Office of Federal Housing Enterprise Oversight (OFHEO).
The S&P/Case-Schiller Metro Area Home Price Indices use the “repeat sales method” of index calculation – an approach that is widely recognized as the premier methodology for indexing housing prices – which uses data on properties that have sold at least twice, in order to capture the true appreciated value of each specific sales unit.
Eligibility Criteria:
To be eligible for inclusion in the indices, a house must be a single-family dwelling. Condominiums and co-ops are specifically excluded. Houses included in the indices must also have two or more recorded arms-length sale transactions. As a result, new construction is excluded.
Index Construction
Approaches:
The S&P/Case-Shiller Metro Area Home Price Indices are based on observed changes in home prices. They are designed to measure increases or decreases in the market value of residential real estate in 10 defined MSAs. In contrast, the indices are specifically, not intended to measure recovery costs after disasters, construction or repair costs, or other such related items.
The indices are calculated monthly, using a three-month moving average algorithm. Home sales pairs are acculmulated in rolling three-month periods, on which ther repeat sales methodogy is applied. The index point for each reporting month is based on sales pairs found for that month and the preceding two months. For example, the December 2008 index point is based on repeat sales data for October, November and December of 2008. This averaging methodology is used to offset delays that can occur in the flow of sales price data from county deed recorders and to keep sample sizes large enough to create meaningful price change averages.
Index Calculations
To calculate the indices, data are collected on transactions of all residential properties during the months in question. The main variable used for index calculatiuon is the price change between two arms-length sales of the same single-family home. Home price data are gathered after that inforamtion becomes publicly available at local recording offices across the country. Available data usually consist of the address for a particular property, the sale date, the sale price, the type of property, and in some cases, the name of the seller, the name of the purchaser, and the mortgage amount.
For each home sale transaction, a search is conducted to find information regarding any previous sale for the same home. If an earlier transaction is found, the two transactions are paired and are considered a "repeat sale." Sales pairs are designed to yield the price change for the same house, while holding the quality and size of each house constant.
These sales pairs are further examined to eliminate outliers that migth distort the calculations. Outliers include non-arms-length transactions (ex. protperty transfers between family members), transactions immediatley preceding or subsequent to substantial physical changes to a property; transactions where the property type designations is changed (ex. properties originally recorded as single-family homes are subsequentially recorded as condominiums), and suspected data errors where the order of magnitude in values appears unrealistic.
Each sales pair is aggregated with all other sales pairs found in a particular MSA to create the MSA-level index. The 10 Metro Area Indices are then combined, using a market-weighted average, to create the national composite.
The Weighting of Sales Pairs
The Indices are desighned to reflect the average change in all home prices in a particular geographic market. However, individual home prices are used in these calculations and can fluctuate for a number of reasons. In many of these cases, the change in value of the individual home does not reflect a change in the housing market of that area: it only reflects a change in that individual home. The index methodology addresses these concerns by weighting sales pairs.
Different weights are assigned to different changes in home prices based on their statistical distribution in that geographic region. The goal of this weighting process is to measure changes in the value of the residential real estate market, as opposed to typical changes in the value of individual homes.
Price anomalies. If there is a large change in the prices of a sales pair relative to the statistical distribution of all price changes in the area, then it is possible that the home was remodeled, rebuilt or neglected in some manner during the period from the first sale to the second sale. Or, if there were no physical changes to the property, there may have been a recording error in one of the sale prices, or an excessive price change caused by idiosyncratic, non-market factors. Since the indices seek to measure homes of constant quality, the methodology will apply smaller weights to homes that appear to have changed in quality or sales that are otherwise not representative of market price trends.
High turnover frequency. Data related to homes that sell more than once within six months are excluded from the calculation of any indices. Historical and statistical data indicate that sales made with a short interval often indicate that one of the transactions (1) is not arms-length, (2) precedes or follows the redevelopment of a property, or (3) is a fraudulent transaction.
Time interval adjustments. Sales pairs are also weighted based on the time interval between the first and second sales. If a sales pair interval is longer, then it is more likely that a house may have experienced physical changes. Sales pairs with longer intervals are, therefore, given less weight than sales pairs with shorter intervals.
Intitial home value. Each sales pair is assigned a weight equal to the first sale price to ensure that the indices track the aggregate/average of all homes in a market.
While the indices are intended to represent all single-family residential homes within a given MSA, data for particular properties or component areas may not be available. Performance of individual properties or counties is not necessarily consistent with the MSA as a whole. County components are subject to change as a result or revisions by the U.S. Census Bureau or data insufficiencies.
National Composite
The composite home price index is constructed to track the total value of single-family housing within its constituent metro areas:
The composite home price index is analogous to a cap-weighted equity index, where the aggregate value of housing stock represents the total capitalization of all of the metro areas included in the composite.
Calculating Composite Index History
Calculating history for the composite index requires setting the base periods for weights and the aggregate values of single-family housing stock for those periods. Since the decennial U.S. Census currently provides the only reliable counts of single-family housing units for metro areas, the years 1990 and 2000 were chosen as the base periods. The housing stock measures used to calculate the aggregate value of single-family housing (for both 1990 and 2000) are the U.S. Census counts for the metro areas.
Index Construction Process
The S&P/Case Schiller Metro Area Home Price Indices are based on observed changes in individual home prices. The main variable used for index calculations is the price change between two arms-length sales of the same single-family home. Home price data is gathered after that information becomes publicly available at local deed recording offices across the country. For each home sale transaction, a search is conducted to find information regarding any previous sale for the same house. If an earlier transaction is found, the two transactions are paired and are considered a "sale pair". Sale pairs are designed to yield the price change for the same house, while holding the quality and size of each house constant.
The S&P/Case -Shiller Metro Area Home Price Indices are designed to reflect the average change in market prices for constant-quality homes in a geographic market. The sale pairing process and the weighting used within S&P/Case-Shiller repeat sales index model ensure that the indices track market trends in home prices by ignoring or down weighting observed price changes for individual homes that are not market driven and/or occur because of idiosyncratic physical changes to a property or a neighborhood. Sale prices from non-arms-length transactions, where the recorded price is usally below market value, are excluded in the pairing process or are down-weighted in the repeat sales model. Pairs of sales with very short time intervals between transactions are eliminated because observed price changes for these pairs are much less likely to be representative of market trends. Idiosyncratic changes to properties and/or neighborhoods are more likely to have occured between sales with longer transactions intervals, so these pairs are down-weighted in the repeat sales index model if they are not eliminated during the sale pairing process.
Pairing Sales and Controlling Data Quality
The automated sale pairing process is designed to collect arms-length, repeat sales transactions for existing, single-family homes. This process collects as many qualifying sale prices as possible, ensuring that large, statistically representative samples of observed price changes are used in the S&P/Case-Shiller repeat sales model. Ina an arms-length transaction, both the buyer and seller act in their best economic interest when agreeing upon a price. When they can be identified from a deed record, non-arms-length transactions are property transfers between family members and transfers of properties from mortgage borrowers to lenders during foreclosure proceedings. Although identified foreclosure transfers are excluded during the pairing process, subsequent sales by mortgage lenders of foreclosed properties are candidates to be included in repeat sale pairs.
The pairing process is also designed to exclude sales of properties that may have been subject to substantial physical changes immediately preceding or following the transaction. Furthermore, since a property must have two recorded transactions before it can be included as a repeat sale pair, newly constructed homes are excluded from the index calculation process until they have been sold at least twice. Deed records do not usually describe the physical characteristics of properties (other than the size and alignment of land parcels). However, other items listed on the deed record can be used to identify properties that may have been subject to substantial physical changes. Deeds that have been marked as transfers of land with no improvements (ex. no structures) are excluded. Transactions where the seller may be a real estate developer (based on the seller's name) are also excluded, since it is likely that this is the sale of a newly constructed home built on a previously vacant or occupied lot or a rebuilt existing home.
Finally, sales that occur less than six months after a previous sale are excluded, primarily because single real estate transactions often have duplicate or multiple deed records due to the procedures used by local deed recorders and property data vendors. It is also more likely that in cases with a very short interval between sales that: (1) one of the transactions is non-arms length (ex. A transfer between family members before selling a property), (2) the property has undergone substantial physical changes (ex. A developer has purchased and quickly sold a rebuilt property), or (3) one of the transactions is a fraudulent transaction (a “property flip”).
Although the number of excluded transactions will vary from market to market, depending on how much detailed information is available in recorded deeds, usually less than 5% of non-duplicate transaction records are identified as non-arms-length and are removed as possible pairing candidates. Similarly, typically less than 5% of non-duplicate transaction records are preceded by another transaction within the last six months. The percentage of properties identified as either new construction or rebuilt existing homes depends on local market conditions, since construction activity is cyclical and related to the strength of the market’s economy, the overall age and condition of the existing housing stock, and the balance between housing supply and demand. Depending on these factors and the completeness of deed information, the percentage of sales identified and eliminated from the pairing process because there may have been substantial physical changes to the property usually ranges from 0% to 15%.
The Weighting of Sales Pairs
Although non-arms-length transactions and sales of physically altered properties are discarded during the pairing process, it is not possible to identify all of these sales based on the information available from deed records. Furthermore, the price changes observed for individual homes may be the result of non-market, idiosyncratic factors specific to a property (which cannot be identified form the deed information) or a property’s neighborhood. For example, a buyer was in a special hurry to buy and paid too much, boosting the value of nearby properties relative to the market, or an individual property may have been well maintained, reducing its value relative to the Market. Finally, errors in recorded sale prices may cause a particular sale pair to mismeasure the actual price change of an individual property.
To account for sale pairs that include anomalous prices or that measure idiosyncratic price changes, the repeat sales index model employs a robust weighting procedure. This automated, statistical procedure mitigates the influence of sale pairs with extreme price changes. Each sale pair is assigned a weight of one (no down-weighting) or a weight less than one but greater than zero, based on a comparison between the price change for that pair and the average price change for the entire market. The degree to which sale pairs with extreme price changes are down-weighted depends on the magnitude of the absolute difference between the sale pair price change and the market price change. No sale pair is eliminated by the robust weighting procedure (ex., no pair is assigned a zero weight), and only sale pairs with extreme price changes are down-weighted. Although the number of sale pairs that are down-weighted depends the statistical distribution of price changes across all of the sale pairs, in large metro area markets, typically 85% to 90% of pairs are assigned a weight of one (no down-weighting), 5% to 8% are assigned a weight between one and one-half, and 5% to 8% are assigned a weight between one-half and zero.
The S&P/Case-Shiller repeat sales model also includes an interval weighting procedure that accounts for the increased variation in the price changes measured by sale pairs with longer time intervals between transactions. Over longer time intervals, the price changes for individual homes are more likely to be caused by non-market factors (ex., physical changes, idiosyncratic neighborhood effects). Consequently, sale pairs with longer intervals between transactions are less likely to accurately represent average price changes for the entire market.
The interval weights are determined by a statistical model within the repeat sales index model that measures the rate at which the variance between index changes and observed sale pair price changes increases as the time interval between transactions increases (time-between-sales variance). It is also assumed that the two sale prices that make up a sale pair are imprecise, because of mispricing decisions made by homebuyers and sellers at the time of a transaction. Mispricing variance occurs because buyers and sellers have imperfect information about the value of a property. Housing is a completely heterogeneous product whose value is determined byhundreds of factors specific to individual homes (ex., unique physical attributes; location relative to jobs, schools, shopping; neighborhood amenities). The difficulty in assigning value to each of these attributes, especially when buyers and sellers may not have complete information about each factor, means that there is significant variation in sale prices, even for homes that appear to be very similar.
The interval weights in the repeat sales model are inversely proportional to total
interval variance, which is the sum of the time-between-sale variance and the
mispricing variance. A statistical model within the repeat sales model is used to
estimate the magnitudes of the two components of total interval variance. The interval
weights introduce no bias into the index estimates, but increase the accuracy of the
estimated index points.
More technically, the interval weights correct for heteroskedastic (non-uniform) error variancein the sale pair data. These corrections for heteroskedasticity reduce the error of the estimatedindex points, but do not bias the index upwards or downwards.
Information obtained from http://www.cmegroup.com/trading/real-estate/index.html
|
