The simple version

A house price index reports that prices rose or fell by some percentage. That claim is harder to make than it sounds, because the houses sold this quarter are not the houses sold last quarter, and no two of them are alike.

The main indexes get around it by narrowing the question. Instead of comparing all houses to all other houses, they look only at properties that have sold more than once and measure how much each one changed between its own sales. A house is compared to itself, which is the only comparison that holds everything else constant.

The numbers

• The Federal Housing Finance Agency (FHFA) releases its House Price Index on Tuesday, August 25. That release is the quarterly index with monthly tables, and the latest data it includes is June 2026 and the second quarter of 2026 (Federal Housing Finance Agency)
• The index is a weighted, repeat-sales index, meaning it measures average price changes in repeat sales or refinancings on the same properties (Federal Housing Finance Agency)
• It is built by reviewing repeat mortgage transactions on single-family properties whose mortgages have been purchased or securitized by Fannie Mae or Freddie Mac since January 1975 (Federal Housing Finance Agency)
• Conforming means a mortgage that meets Fannie Mae or Freddie Mac underwriting guidelines and does not exceed the conforming loan limit, so a loan above that limit sits outside the index (Federal Housing Finance Agency)
• A purchase made entirely in cash creates no mortgage for either company to acquire, so it cannot enter the index either (definition)
• A newly built house is outside a repeat-sales index by definition, because a house selling for the first time has no prior sale to compare against (definition)
• FHFA calls the result a constant quality index, because the repeat-sales technique helps control for differences in the quality of the houses in the sample (Federal Housing Finance Agency)
• By contrast, mean and median values inherently reflect changes to both the price and the quantity of housing, which is FHFA's own explanation of why the distinction matters (Federal Housing Finance Agency)
• The Census Bureau publishes a monthly median sale price for new houses sold. It was $485,000 in June 2026, and July figures release on Tuesday, August 25, at 10:00 a.m. Eastern (U.S. Census Bureau)
• FHFA describes its calculation as a modified version of the Case-Shiller index using a geometric weighted repeat-sales procedure. Different indexes cover different populations and each publishes its own methodology (Federal Housing Finance Agency)

Why the median can move when prices do not

The simplest way to summarize house prices is the median sale price: line up everything that sold and take the middle one. It is easy to compute and easy to misread.

The problem is that it describes the houses that sold rather than the value of houses. If expensive homes make up a larger share of sales one month, the median rises even if every individual house in the country is worth exactly what it was worth before.

Nothing appreciated. The mix changed.

That happens routinely, for reasons that have nothing to do with value. When mortgage rates rise, buyers at the lower end drop out first, which shifts the mix upward and can push the median higher during a stretch when prices are flat or falling.

A repeat-sales index does not have that problem, because it never compares one house to a different house. FHFA makes the same point in its own documentation, noting that mean and median values reflect changes in both the price and the quantity of housing while its index isolates the price part.

What the method leaves out

The elegance comes at a cost, and the exclusions are specific enough to matter when reading the number.

New construction cannot appear, because a new house has no earlier sale. An index built this way is therefore blind to exactly the segment that housing supply debates are about, which is part of why the new home sales report is published separately and read alongside it.

The FHFA index adds a further filter. It is built from mortgages that Fannie Mae or Freddie Mac purchased or securitized, and conforming means a loan that both meets their guidelines and stays under the conforming loan limit. The expensive end of the market sits outside that boundary, and so does every all-cash purchase, because a cash sale produces no mortgage to acquire.

There is also a subtler gap. A house that sells twice may have been renovated in between, in which case some of the price change is improvement rather than appreciation. Index builders handle this in various ways, and it remains a live methodological question rather than a solved one.

None of that makes the index wrong. It makes the index a measurement of a defined population, which is what every index is.

The Real Cost lens on a number about your biggest asset

For most households that own a home, it is the largest asset they hold, and these indexes are the closest thing to a public estimate of what happened to it.

• A national index describes an average across a country where local markets diverge sharply, so it says very little about any particular house
• If a home was bought with cash, or with a loan above the conforming limit, it sits outside the FHFA index entirely
• If it was newly built, it cannot be in a repeat-sales index at all until it sells a second time
• A median headline and an index headline can point in different directions in the same period, and both can be accurate, because they answer different questions

None of that is a reason to distrust the numbers. It is a reason to read the label. Knowing which population a number covers is most of knowing what the number means.

What this means

When a house price headline lands on Tuesday, the first question is which kind of number it is. If it is a median, the mix of what sold is doing part of the work. If it is a repeat-sales index, the mix is removed but new construction and part of the market are missing.

The broader idea generalizes well past housing. Any index measuring things that are not identical has to solve the comparison problem somehow, and every solution excludes something. Finding out what was excluded is usually the fastest route to understanding what the number can and cannot tell you.

What this is NOT

This is not a prediction of house prices, mortgage rates, or housing activity, and it makes no claim about what Tuesday's releases will show. This is not advice about buying, selling, refinancing, or timing any housing or financial decision. This is not advice about any security, fund, or asset, including housing-related ones. This is not a claim that any index is better or worse than another; different indexes measure different populations with different methods, and each publishes its own methodology for readers to check. National figures describe averages and can differ substantially from any local market or any individual property. This is not a valuation of any home. This is not investment or financial advice of any kind.

Sources

• Federal Housing Finance Agency, House Price Index: https://www.fhfa.gov/data/hpi
• Federal Housing Finance Agency, House Price Index frequently asked questions: https://www.fhfa.gov/faqs/hpi
• Federal Housing Finance Agency, House Price Index technical description: https://www.fhfa.gov/research/papers/house-price-indexes-hpi-technical-description
• U.S. Census Bureau, New Residential Sales: https://www.census.gov/construction/nrs/index.html
• U.S. Census Bureau, Economic Indicator Release Schedule for construction surveys: https://www.census.gov/construction/soc/schedule.html