What is the Case-Shiller index and how is it calculated?

The big idea: same house, two sales

Most price headlines compare this month's sales to last year's sales. The problem is that this month might have sold more mansions and fewer condos, and suddenly "prices" jumped even though no single house got more valuable.

Case-Shiller avoids that by tracking the same house over time. As Inman's explainer puts it, the index follows the sale of the same house over time, comparing a home's price across two sales instead of averaging all transactions. S&P describes it as using matched price pairs from arm's-length sales of single-family homes.

Say a hypothetical house sold for $500,000 in 2019 and $650,000 in 2026. That pair says one real house rose 30%. Collect enough pairs across a metro and you get a solid read on price movement that is not fooled by what happened to sell that month.

It is the economist's version of comparing your height on the same doorframe every year.

What gets left out, and why

The index is picky about which pairs it trusts. Per Inman's breakdown:

• Non-arm's-length sales are dropped. A sale between relatives at a friendly price does not reflect the market.
• Quick flips are ignored. Homes that trade more than once in a six-month period are left out.
• Suspect pairs are removed, such as price jumps that suggest a major renovation happened between sales.
• New construction is excluded. A brand-new home has no earlier sale to pair with.
• Condos are limited. The index has condo data in only a few markets, not nationally. S&P's release mentions five condo markets with seasonally adjusted data.

All of this makes the index cleaner. It also means Case-Shiller describes the existing single-family market, not the new-home subdivision going up off the freeway with the inflatable tube man.

Timing: three-month averages and a two-month lag

Two details trip people up, and they are the most important ones for anyone pricing a house.

The average. Inman notes the index number is calculated monthly using a three-month moving average. So a "July" reading blends sales recorded in May, June and July.

The lag. The index is reported with about a two-month delay, and S&P publishes it on the last Tuesday of each month at 9:00 a.m. Eastern. The July 2026 numbers came out on September 29, 2026.

Now stack those on top of escrow. Inman points out that contracts usually close one to three months after they are signed. A closed sale in July often reflects a deal negotiated in spring. By the time you read a headline, you are looking at buyer decisions that may be five or six months old.

S&P also notes that the prior 24 months of data can be revised as more records come in. Yesterday's number is allowed to change its mind.

The 20 metros, the composites and the weighting

Case-Shiller publishes a set of indexes, now under the S&P Cotality Case-Shiller name (it was S&P CoreLogic Case-Shiller before). The main ones:

• 20-City Composite, covering 20 metro areas
• 10-City Composite, covering the 10 original metros
• U.S. National Index, built from the nine U.S. census divisions

The index base is set at 100 in January 2000. Inman also notes that results are split into low, middle and high price tiers based on each home's prior sale price.

The composites are value weighted. Inman explains that this gives higher-priced markets, like coastal California and New York, outsized influence. One expensive metro moving can tug the composite more than a cheaper one.

For example, the release for July 2026 reported annual gains of 1.9% for the National Index, 2.5% for the 20-City Composite and 3.4% for the 10-City. Chicago led at 6.9%. Seattle was lowest at minus 1.6%. San Francisco was up 3.48%, San Diego 1.63% and Los Angeles 1.24%.

What I see in the field

I like Case-Shiller as a weather report. It tells you the season. It does not tell you whether it is raining on your street right now.

An appraisal has to be as of a specific date and a specific house. A metro index that is smoothed over three months, lagged by two and value weighted across a whole region cannot tell me what buyers paid last month for a three-bedroom on your block. Neighborhoods inside the same metro can move in opposite directions.

That is why I read the market in real time. I test the shortest recent window the local sales can support, and I measure price change from the sales themselves, including repeat sales of the same homes when there are enough. The index is useful context. The local evidence sets the adjustment.

How to use Case-Shiller without misreading it

1. Check the date. A release is reporting sales from about two months earlier, averaged over three months.
2. Use your metro, not the national number, when one is available. National gains can hide a local drop.
3. Remember what is excluded. No new construction, few condos, no quick flips.
4. Compare it to FHFA's House Price Index, another repeat-sales index built on loans backed by Fannie Mae and Freddie Mac. If both point the same way, the trend is more believable.
5. Go local for real decisions. For a list price, a refinance or an estate, look at recent comparable sales near the house or get an appraisal.