This study aims to test the effect of Seattle’s discontinuous sidewalk requirement, on the number of housing units per construction permit.
This study uses discontinuity linear regression (DLR) on a database of Seattle’s housing construction permits during January-2015 to January-2018, controlled by 51 socioeconomic, planning and geographic variables. The sidewalk requirement is continuous inside the designated urban villages; however, it is spatially and quantitatively discontinuous in the rest of the city: certain blocks at certain locations require sidewalks’ design and construction in permits with six or more housing units. DLR detects the effect of the discontinuity while controlling for a vast array of confounding variables.
The primary finding is that the discontinuous requirement reduces the number of housing units in about 75% of a housing unit per permit, which at the aggregate level amounts to around 335 fewer housing units during the period of analysis.
The database is relatively small, which has limited a more thorough specification process and robustness tests.
Besides directly testing the effect of a discontinuous in-kind development contribution, the research setup allows to discuss a wider, more structural problem: the possibility of contributions avoidance due to spatial substitution. In contrast, spatially continuous (i.e. city-level) contributions cannot be avoided by performing spatial substitution, and they are internalized by the housing supply side (market-neutral).
