Our recent post on women in the workforce included a lyric from Dolly Parton and an explanation of the composition effect. Here’s the formal definition: “The part of the observed between-group difference in the distribution of some economic outcome that can be explained by differences in the distribution of covariates.”

That’s a doozy of a definition, so let’s use a picture that’s worth 1,000 words to explain it… The graph shows the year-to-year growth rate of average home prices in the United States (blue bars) and in its 20 largest metropolitan areas (red bars). The blue bars and red bars generally extend in the same direction, although by different magnitudes. Because these top 20 metropolitan areas are part of the United States, it’s not surprising both sets of average prices move in the same direction.

But look what happened in 2010: Average home prices overall decreased while average home prices in the 20 largest metropolitan areas increased. Why? Because average home prices in smaller metropolitan areas and rural areas decreased more than average home prices in large metropolitan areas increased. That’s the composition effect: Looking at the big picture sometimes masks what’s going on with the individual parts.

Want to learn more about the composition effect in housing prices? Read “A Guide to Aggregate House Price Measures” by Jordan Rappaport of the Kansas City Fed.

How this graph was created: Search for “S&P home prices” and select the “U.S. National” series (FRED series ID CSUSHPISA). From the “Edit Graph” panel, use the “Add Line” feature to search for and select the “S&P 20-City” series (FRED series ID SPCS20RSA). From the “Format” tab, select “Bar” for graph type. From the “Edit Bar 1” tab, select “Percent Change from Year Ago” for units and “Annual” under “Modify frequency.” Do the same from the “Edit Bar 2” tab. Adjust the time period shown with the slider beneath the graph or with the start and end date boxes above the graph.

Suggested by Diego Mendez-Carbajo.