Mumbai upzoning

Geetika Nagpal‘s job market paper, written with Sahil Gandhi, shows that a 2017 increase in allowed floor area ratio in Mumbai had a tremendous impact on affordability by accidentally improving the economics of smaller apartments.

(Note that the authors are updating the paper, so some of the following may change).

ABSTRACT:

Does relaxing zoning regulations increase affordable housing or simply trigger the building of new luxury units? This paper exploits a rule-based relaxation of the regulatory cap on building height and floorspace, the Floor Area Ratio (FAR), to answer this question in Mumbai, India. Leveraging granular panel data and exploiting variation in time and space, we find that the reform increased housing supply in treated areas by 28%, implying an elasticity of housing supply to the FAR of 1.59. The FAR relaxation increases the scale of development, resulting in higher investment in shared amenity space within the building. This increased public good provision facilitates an 18% decline in unit sizes, leading to a 29% decrease in apartment prices that allows lower-income households to access housing. We develop a structural model of housing supply and demand that incorporates the provision of amenity floorspace and shows that after the relaxation, average home buyer incomes are 3.18% lower. We use the estimated model to show that a further 5% rule-based relaxation would amplify the scale economies and increase the affordability gains from deregulation. Taken together, our results show that concentrating FAR relaxation can improve affordability.

Some quick notes:

• The authors have access to excellent data, including detailed permit applications complete with floor plans and some mortgage applications.
• The standard errors for all estimates are very large. The point estimates are the centers of large possible ranges, so don’t take them too seriously. The imprecision also weakens the total picture presented.
• The authors’ most novel finding is that lower-income buyers prefer amenity space, such as children’s play areas & gyms. Shared amenities make more sense in bigger buildings. So bigger buildings get smaller apartments, which are cheaper.

The last point left me with two questions the authors could answer pretty easily:

• How do the “lower” and “higher” income purchasers relate to the overall income distribution?
• What’s the distribution of amenity space as a function of building size? Are there big discrete jumps in (non-hallway) public space, or is it more of a continuum?

Excellent work and I look forward to its refinement.