Is There Enough Housing Production? It Matters Which Indicators Are Used to Answer

Abstract

McClure and Schwartz argue that there is no US housing shortage for the 2000-2020 period because: (1) we added more dwellings than households; (2) vacancy rates are not low by historical standards; and (3) current headship rates are unrelated to the ratio of recent growth in dwellings to growth in households. I argue that these indicators offer misleading guides for addressing policy-relevant questions related to the supply of housing because household formation is endogenous to changes in the housing stock. I demonstrate the shortcomings of these indicators with a stylized graphic/numerical analysis of short-term and long-term housing market equilibrium and adjustment dynamics in two hypothetical metropolitan areas. Changes in household size (adjusted for demographic and income distributional alterations) and changes in housing rents and prices are more appropriate indicators for making housing policy choices, implying that current concerns over the responsiveness (long-run elasticity) of housing supply are indeed justified.

Keywords:

• Housing supply
• housing affordability
• housing indicators

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Notes

1 “Price” means a generic market valuation, monthly rent or sales price. Here I will keep things as simple as possible, so I will not distinguish renter- and owner-occupied markets, or distinctions among housing quality submarkets. I also assume all people and owners are identical, respectively. For more complex formulations that take such features into account, see Rothenberg et al. (1991) and Galster (2019, chs. 3-4).

2 Although it may not be as obvious as with the other indicators, vacancy rates also fall prey to this problem because they, too, are a function of the (endogenous) number of households seeking housing in a short-run equilibrium context.

3 Note that from 2000 to 2020 the mean household size across the U.S. declined slightly from 2.67 to 2.57 persons per household. This should not be interpreted as evidence of an adequate aggregate housing supply response since the 2020 figure is not adjusted for population composition changes over the period in the manner described above.

4 Such as with repeat-sales samples, as in the Case-Schiller and other indices.

5 Its growth also correlates inversely with headship rates, as McClure and Schwartz documented.

Additional information

Notes on contributors

George C. Galster

George C. Galster earned his PhD in Economics from M.I.T. and is the Clarence Hilberry Professor of Urban Affairs (emeritus) at Wayne State University. He has published 177 peer-reviewed articles, 10 books and 44 book chapters on a wide range of urban topics. The Urban Affairs Association awarded him the prestigious “Contributions to the Field of Urban Affairs” prize in 2016.