It’s lately become very fashionable to contrast the concurrent trend towards ever-larger new houses and ever-smaller new apartments. Recently, Richard Florida,* Kaid Benfield, and even the Streetsblog podcast have all mentioned this peculiar contradiction.

But scratch a little bit at the surface and you’ll find that it’s an artifact of the Great Recession, not an existential conundrum. Consider that the Census’ Survey of Construction, the source of said statistic, only measures a universe consisting of new single family (usually for-sale) houses. However, these share of total new housing construction that is in single-family houses continues to fall. Just from 2011 to 2013, multi-family houses went from 29% to 33% of all new housing completions. Meanwhile, their size hasn’t budged: it’s consistently been around 1,100 square feet (+/- 50) since 1999.

The chart above also comes from Census data, and tells a more complete story. The average size of new houses built in the USA, both single- and multi-family, declined from 2011 to 2012, then ticked up to a record in 2013. It hasn’t been on an upward march forever. Another reason why this data point isn’t quite as important as it might seem at first glance: many fewer Americans are in new houses, period.

Construction may be up slightly, but it’s still well below its peak. Completions fell 70% from the 2004-2006 rate, and now we’re back up to 60% below that (completely unsustainable) rate. From builders I’ve heard at ULI panels and the like, new move-up product is still selling in the suburbs, but first-time buyers are staying in rentals longer, largely due to tighter financing. Hence, fewer small homes are getting built, as younger households stay in existing stock for much longer, and the average home size is increasing. (Oh, and remember: we’re talking about fewer smaller single-family houses. The number of multi-family houses, small and medium-sized, is increasing.)

Thus, the real story is that the for-sale single-family starter home, as a product type, is dying off. For example, when my brother recently purchased his first new house (at 35, considerably older than the 28-year-old median buyer at, say, Park Forest), he didn’t get a starter home. He went directly from an existing multi-family unit to a move-up sized house.

* Making the additional methodological error of taking the definition of “central cities” seriously.