Crowd Out Research Provides No Evidence of Subsidized Affordable Housing Development Displacing Unsubsidized Affordable Housing

An oft repeated (Eriksen & Rosenthal 2010; Burge 2011) critique of supply-side affordable rental housing programs is that they do not increase the total supply of available rental housing. The research in this area does not suggest that supply side affordable housing displaces other affordable housing development, or challenge the efficacy of these programs in adding to the overall supply of affordable housing. Rather, it is argued that additional subsidized units built with the help of supply-side housing subsidies “crowd out” market rate units. But researchers have arrived at varying quantitative and qualitative conclusions regarding this crowd out effect.

Quantitatively, the studies conducted to date suggest a level of crowd out as low as 0.2 market rate units crowded out for each affordable unit added (0.67 in areas in the top third of housing value appreciation), to as high as one market rate unit crowded out for each affordable unit added (Baum-Snow & Marion 2009; Eriksen & Rosenthal 2010). In the middle, estimates have ranged from 0.3 to 0.7 market rate units crowded out for each affordable unit added (Eriksen & Rosenthal 2007; Sinai and Walfrogel 2004; Baum-Snow and Marion 2008; Baum-Snow and Marion 2009).

The variability of these estimates is in part explained by a “radius” effect, where estimates are sensitive to the level of geography studied. For example, Eriksen and Rosenthal (2010) found that moving from 10-mile radius circles around low-income housing tax credit (LIHTC) developments to the metropolitan statistical area (MSA) level resulted in the crowd out estimates going from one, full crowd out, to indistinguishable from zero, no effect. Similarly, Sinai and Walfrogel’s (2004) estimates of crowd out decrease from approximately two thirds to one-half when moving from the census designated place level to the MSA level.

Qualitatively, some researchers have also noted positive effects on total housing supply. Chen and Ge (2013) found a “significant positive correlation” between LIHTC development and overall housing supply. Malpezzi and Vandell (2002) note that even if crowd out exists, “this would not necessarily imply inefficiency of these programs [since] there are other performance measures besides the “tightness” or “looseness” of the market that must also be considered, including rent/ price levels, housing quality levels, and the quantity of housing consumed by tenants in subsidized housing”(pg. 21). In addition, Eriksen and Rosenthal (2010) note that the fact that LIHTC housing is often placed in upper income areas may, “help low and moderate income families gain access to higher quality local schools and other local public services “(pg. 30).

Overall, the research on crowd out tends to vary according to the level of geography studied, with no robust estimates of crowd out of market rate units by affordable housing development at levels equal to or higher than the MSA level. More importantly, the research does not challenge the success of programs like the LIHTC in adding to the overall stock of quality rent-restricted affordable housing for low-income households. This is important because unlike unsubsidized but affordable housing, LIHTC housing generally must comply with affordability restrictions for a minimum of 30 years.