To understand what the latest US house prices mean, it helps to know something about the data.

It is very difficult, actually, to measure house price changes. Houses differ - in age, location, number of bedrooms, building style, house size, lot size, etc. A 300 sq. m. penthouse in New York City isn't a 300 sq. m. Florida beach front villa.

Technical issues abound. An increase in the quality of new houses may increase average reported prices, even if existing home values remain unchanged. In addition, property sales in a given year can be skewed say, toward high-quality houses. That messes the results.

Another problem is infrequency of sales. Houses are usually sold once in several years. Less expensive houses, as it happens, tend to be sold more frequently, again skewing results.

Because of these complexities, different house price measures show different results. The three commonly used approaches are:

  • Averaging technique Takes the average (mean, median or mode) price of a large number of transactions.

The averaging technique is the simplest way to construct a house price index. Data are updated on a timely basis. However, its simplicity is also its main weakness. The method ignores the issues of heterogeneity and infrequency of sales.

  • Repeat-sales methodology Focuses on houses that have been sold more than once, excluding newly constructed homes. The repeat sales approach assumes that the quality of these houses remains approximately the same over time.

One of the criticisms of this method is that it is wasteful of data. Since only some percentages of the total number of houses sold in a given period are repeat sales, it uses a relatively small sample size, as compared to the averaging method.

  • Hedonic methodology Controls for differences in the quality of houses by establishing a constant-quality new house. The idea behind the model is that the price of a given house can be precisely estimated from its attributes. It treats a house as a bundle of attributes, each with its corresponding price that changes over time.

Though the hedonic method solves the issue of heterogeneity, it is relatively complex to implement because it requires high-degree of statistical expertise. In addition, it requires a large amount of data.

Leading US house price indicators

The two most widely-used housing indicators are the house price index of the US Federal Housing Finance Agency (FHFA) and the Standard &Poors (S&P)/ Case-Shiller index.
The FHFA house price index (previously produced by the Office of Federal Housing Enterprise Oversight) uses repeat mortgage transactions of single-family properties purchased or securitized by the Federal Home Loan Mortgage Corporation (Freddie Mac) and the Federal National Mortgage Association (Fannie Mae).

The FHFA publishes:

  • All-transactions HPI data obtained from mortgages used for both home purchases and refinancing, based on more than 36 million transactions.
  • Purchase-only HPI derived only from home purchase mortgages, excluding refinancings.
  • Individual indices for the nine US Census Divisions
  • Individual house price indices for the 50 US states and Washington DC
  • Purchase-only indexes for the 25 most-populated metropolitan areas in the US (added recently).

The S&P/Case-Shiller is another repeat sales index, and has:

  • Indices for selected metropolitan areas, calculated monthly. Published with a two month lag, the indices accurately track the price changes of typical single-family homes in each of the included US metropolitan areas.
  • Two composite indices, value-weighted aggregates of their component metropolitan areas, published monthly.

the S&P/Case-Shiller Composite of 20 Home Price Index (SPCS-20)
the S&P/Case-Shiller Composite of 10 Home Price Index 10 (SPCS-10).
National home price index a broader composite of single-family home price indices for the nine census divisions in the US, published quarterly.

Although both the FHFA/OFHEO and S&P/Case-Shiller indices employ the same fundamental repeat-valuations approach, they differ in their data sources.

  • The FHFA HPI uses mortgage originations data from Freddie Mac and Fannie Mae, so it excludes cash house purchases, jumbo loans, and Federal Housing Administration (FHA) and Veterans Administration (VA) loans. In addition, the FHFA HPI does not adequately represent the subprime and nontraditional loan markets.
  • The S&P/Case-Shiller national index includes only 20 metropolitan areas and these are among the highest priced areas in the US. Since these areas are also the hardest hit by the crisis, the index results are skewed toward dramatic price swings. Because of this, some economists feel the S&P/Case-Shiller is biased toward bad news. The S&P/Case-Shiller also has no price cutoff, does not include refinancings, and excludes houses that undergone improvements.

The US Census Bureau has a Constant-Quality House Price Index (CQHPI), updated quarterly, which employs the (more complex) hedonic methodology. Attributes included are geographical location, lot size, number of bedrooms, number of bathrooms, type of parking facility, construction method, types of heating and air-conditioning systems.

  • National CQHPI an aggregate of the US regions
  • Regional CQHPI includes home price indices for Northeast, Midwest, South, and West regions.

The Census Bureau's CQHPI is however unpopular, because it uses only a very small sample of transactions, and the method is relatively complex to interpret.

The US Census Bureau also publishes a monthly mean and median sales price of new single-family dwellings, using the averaging technique, based on a monthly survey of residential construction activity in the US.

Other house price series

The National Association of Realtors (NAR) reports the sale prices of existing homes sold through the Multiple Listing Services (MLS).

The NAR's house price series strongly reflects the mix of houses sold in a given period. Therefore the index is criticized for its high volatility. The mean and median house prices rise when a lot of high-end houses are sold, and fall when a lot of low-end houses are sold. Also, NAR sales prices are not seasonally adjusted.

Trulia, a popular real estate search engine, provides information on the average listing price, median sales price, and average price per sq. ft in each of the major cities in the US. It also reports property sales figures, foreclosures and price reductions. The median sales price per city is computed based on licensed county assessor records. On the other hand, the average listing price is computed based on the prices of all homes listed in Trulia for a particular period.

Zillow publishes monthly home value index which tracks changes in the value of different property types in various cities across the US. The index is the mid-point (or the median) of the Zestimate (Zillow's estimated market value, computed using a proprietary formula; not an appraisal) home valuations.

All these house price indicators employ the averaging technique. They are timely, but over-volatile. Repeat transactions of the same house yield better estimates.

Each method has its own strengths and weaknesses. In order to have a realistic view of the state of the US housing market, one should look at several housing indicators and know the basic premises underlying them.


HOUSE PRICES Standard and Poor's S&P/Case-Shiller Index
  • National home price index published quarterly
  • Indices for selected metropolitan areas updated monthly
  • S&P/Case-Shiller Composite of 20 Home Price Index (SPCS-20)
  • S&P/Case-Shiller Composite of 10 Home Price Index 10 (SPCS-10)
Federal Housing Finance Agency (FHFA), previously OFHEO FHFA house price index
  • Purchase-only HPI: national, per census divisions, states, and 25 largest MSAs
  • All transactions HPI: national, per census divisions, states, and 25 largest MSAs
FHFA, previously FHFB House price
  • Purchase price, national average, all homes
National Association of Realtors (NAR) Existing home sales price
  • Median sales price of existing homes: national and regional
  • Mean sales price of existing homes: national and regional
US Census Bureau Sales price for new homes sold
  • Median and average sales price of new one-family houses sold
AFFORDABILITY INDEX NAR Housing affordability index
  • Monthly and quarterly housing affordability indexes
  • Quarterly affordability index for first time buyers
HOMEOWNERSHIP US Census Bureau Homeownership rates
  • Homeownership rates: by states, MSAs, and national
HOME SALES NAR Existing homes sales
  • Sales of existing single-family homes for the nation and the four regions
CONSTRUCTION & HOUSING SUPPLY US Census Bureau New and existing home sales
  • Quarterly sales by price and financing
US Census Bureau Housing Permits
  • Monthly and annual permits for new-privately owned housing units
US Census Bureau Housing starts
  • New privately owned housing units started purpose and design: national and regional data
US Census Bureau Housing Completions
  • New privately owned housing units completed by purpose and design: national and regional data
US Census Bureau Total housing stock (Table 11)
  • Estimates of the total housing inventory: owner-occupied and renter-occupied
NAR Inventory
  • Monthly inventory of existing homes for sale
NAR Months' supply
  • Months' supply of all resale sales
INTEREST RATES US Federal Reserve Fed funds rate
  • Historical data for the fed funds rate
Freddie Mac Mortgage rates
  • 30-year fixed mortgage rate
  • 15-year fixed mortgage rate
  • 5-year adjustable mortgage rate
  • 1-year adjustable mortgage rate
MORTGAGE LOANS US Federal Reserve Total mortgages
RENTAL MARKET US Census Bureau Median asking rents (Table 11)
  • Median asking rent for the US and Regions: Annual and quarterly historical data
US Census Bureau Vacancy rates
  • Rental vacancy rates: by states, by MSAs, and national
FORECLOSURES Realty Trac Foreclosures
  • Total foreclosure filings - default notices, auction sale notices, and bank repossessions - from nearly 2,500 counties
Source: Global Property Guide