In Search of the Missing Middle: Historical Trends in and Contemporary Correlates of Permitting of 2–4 Unit Structures

Abstract

Missing middle housing is an important although often overlooked housing form in America’s built environment. Although still a large component of the US housing stock, production of new small missing middle (SMM) housing—which we define as multifamily structures with two to four units—has steadily declined over the last several decades. In the early 1980s, units in SMM structures comprised ∼9% of residential building permits nationally. Today, less than 3% of new permits are for two- to four-unit structures. We document these trends and explore reasons for the current variation in new SMM production. We build a national, place-level data set combining data on building permits, current and historical census data, land-use regulations, and housing supply characteristics. We then examine the association between SMM production and three sets of community characteristics: supply, regulatory restrictiveness, and demand. Our analysis suggests that correlates of SMM permitting are similar to those for larger multifamily structures. This, we argue, helps explain the decline in SMM, because these developments are competing with and losing to larger multifamily development. We end by considering how these findings can help cities that have recently passed or are considering zoning reforms that broadly legalize missing middle housing forms.

Keywords:

• Housing
• historic
• multifamily
• location
• land use/zoning

Disclosure Statement

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

Notes

1 Note that the Northeastern US has a historically large proportion of SMM stock (An et al., 2021).

2 According to their definition, a structure qualifies as an attached single-family if it satisfies all the following criteria: a ground-to-roof wall separates each unit; each unit has a separate heating system; each unit has individual meters for public utilities; units are not stacked on top of each other. Although we limit our sample in this way to be consistent across our data sources, this context is important for interpreting our results. From both a regulatory and a practical perspective, there may be little difference between attached single-family structures and the small multifamily we study here. Both tend to be denser development forms and are likely regulated similarly.

3 In supplemental Appendix Tables 1 and 2, we present matricies showing the correlations between the independent and dependent variables in our models. There is a weak correlation between some of the demographic measures, but not at a level where we are overly concerned about multicollinearity in our regression analysis.

4 We also modeled the binary and count models in a single, zero-inflated Poisson model. This method uses a Poisson regression for a count outcome and a binary logit model to estimate the excess zeros. The interpretation of the zero-inflated Poisson model is large consistent with the results we present here. We elected to report the logit and Poisson models separately, because we found it easier to interpret these models this way in our descriptive analysis.

5 We also modeled commute distance as a cubic term. With this more complicated modeling technique, the form was more or less the same as with the simpler quadric term. We ultimately decided to report the quadratic model, as it is easier to interpret its coefficients.

6 In this calculation, 9.7 million SMM units in 1980 out of 86.8 million total housing stock reflects an SMM share of 11.2%. The 2014–2018 5-year ACS estimates show 10.9 million SMM units and 136.4 million total housing stock, reflecting an SMM share of 8.0%. Were SMM to have maintained the 11.2% SMM share, this would necessitate a net gain of 4.9 million units to the SMM stock (also increasing the total stock by 4.9 million).

Additional information

Notes on contributors

Daniel Kuhlmann

Daniel Kuhlmann, PhD, is an Assistant Professor of Real Estate and Urban Planning at the University of Arizona. He researches various housing topics, including the impact of land use reforms, housing in declining markets, rental investors, and the businesses of small rental landlords.

Seva Rodnyansky

Seva Rodnyansky is an Assistant Professor of Urban and Environmental Policy at Occidental College. He holds a PhD in Urban Planning and Development from the University of Southern California. His work focuses on housing markets and housing supply, transportation, migration, municipal finance, and affordable housing.