a faulty and inflammatory statistic on housing

In November, a provocative statistic got a ton of attention. The National Association of Realtors (NAR) reported that the median age of a new homeowner increased from 33 to 40 years old, from 2020 to 2025. (Before then, it had been roughly unchanged, between 28 and 33 years old, back to 1981.) The first thing to notice is that it’s odd for a median age to move by more than the passage of time. For example, if every non-homeowner simply waited five more years before they bought their first house, the median age would only rise by five years. 

To increase so substantially, it’s likely that there was a strange change in the sample. For example, imagine if all young people bought homes in 2021. New homeowners in the next few years would be much older, significantly inflating the statistic in subsequent years. Instead, the median age of a category rarely changes much. When it does, it could be influenced by odd factors (e.g., Covid)—and thus, likely to revert after the ripple effects end. 

A statistical aberration could also be caused by poor surveys and faulty statistical work. And it turns out that the NAR methodology was problematic. According to Edward Pinto and Joseph Tracy at the American Enterprise Institute, the biggest problem is that the NAR result came from a small survey of 6,000 responses—a response rate of only 3.5%. Of those, only 20% were first-time homebuyers. So, the valid response rate was less than 1%. On top of that, the sample under-selected younger people and over-selected older people.

In contrast, data from the Mortgage Bankers Association, Cotality, the NY Federal Reserve Bank, and the Census Bureau’s “American Housing Survey”—all based on data from millions of mortgage records—indicates virtually no increase in median age over the same time period. In a word, the NAR result seems to be an artifact of sample selection biases. From there, media bias toward sensationalism and innumeracy among journalists (and the general public) allowed the flawed statistic to flower in the public eye.