The debate between Jason Sorens and Greg Howard & Jack Liebersohn regarding housing supply deregulation forms a key segment of the scholarly commentary published in Volume 22, Issue 2, September 2025 of Econ Journal Watch. This issue also includes discussions on asset bubbles (Miao & Wang vs. Hirano & Toda), factory productivity in Imperial Russia (Lychakov vs. Gregg), and online grocery shopping in Russia.
The debate specifically centers on the policy implications of Greg Howard and Jack Liebersohn's 2021 article, "Why Is the Rent So Darn High? The Role of Growing Demand to Live in Housing-Supply-Inelastic Cities," with Sorens arguing for deregulation and offering critiques of the findings in the H&L paper and two others.
Sorens's Argument: For Deregulation Against Recent Papers
Jason Sorens's piece, "For Housing Supply Deregulation Over Against Three Recent Papers," advocates for deregulation, such as eliminating local zoning restrictions that increase housing development costs. He states that advocates believe liberalization will allow developers to meet housing needs, thereby reducing rents and house prices.
Sorens frames his critique against a "new wave of sophisticated economics papers" that challenge this pro-deregulation conventional wisdom by making two claims: (1) Increasing housing supply doesn't reduce prices because local housing demand is perfectly elastic, and (2) Zoning regulation increases demand by creating amenities.
Sorens focuses heavily on the first claim as posited by Howard and Liebersohn (H&L).
Critique of Howard and Liebersohn's Findings:
- Perfectly Elastic Demand and Policy Implication: Sorens notes that H&L (2021) estimate local housing demand is perfectly elastic over a 17-year horizon (2000 to 2017). This logic implies that an infinitesimal drop in rents (caused by deregulation) results in a surge of migrants, which cancels out the initial rent drop. The policy implication drawn from this finding, Sorens states, is that local efforts to liberalize zoning would not affect national or local rents much after a couple of years, except perhaps in the very largest cities.
- Endogeneity Bias: Sorens suggests that H&L's use of nominal wages in their estimation may lead to endogeneity, as higher rents might force employers to offer higher nominal wages to retain workers in highly regulated, low-elasticity cities. He argues that the critical estimate ($\beta_1$, the coefficient on the interaction term between wage growth and housing supply elasticity) is likely biased downward, leading to potentially infinite estimates of $\mu$ (the elasticity of population to rents).
- Inclusion of Shrinking Cities: Sorens criticizes the inclusion of shrinking cities (like Detroit or Youngstown) where location demand is falling and housing prices are below replacement cost. Since housing supply is always inelastic downward (demolition is slow), these cities would exhibit a correlation between wages and rents similar to growing, inelastic places, potentially skewing the findings on whether the wage-rent correlation varies by upward housing supply elasticity.
- Counter-Evidence: Sorens points to evidence contradicting H&L's findings, suggesting that pro-supply regulatory reforms do reduce local housing costs over a period of years. Examples include a 15% drop in Austin rents after a construction surge and a 21-24% reduction in Auckland rents six years after upzoning. He concludes that city-level upzoning should benefit local affordability over a timeframe of two to about 20 years, eventually fading as spatial equilibrium is reached nationally.
Howard and Liebersohn's Reply: Not Against More Housing
Howard and Liebersohn (H&L) respond by asserting that their paper should not be taken as an argument against housing supply deregulation.
Scope of Affordability: Local vs. National
H&L agree that, under their model's parameters, a local increase in housing supply in one city will have negligible effects on local rents because demand is estimated to be very elastic. However, they clarify that this is because the rent decrease is spread across many places nationally.
- If the goal is to make housing affordable in general (nationally), building more housing is effective regardless of the population elasticity.
- They argue that the welfare benefits of increased housing supply are present—and of similar magnitudes—regardless of whether populations are highly elastic or inelastic; the main difference is whether the rent declines are concentrated locally or dispersed nationally.
Addressing Criticisms:
- Endogeneity/Identification: H&L state they are sympathetic to Sorens's criticism regarding identification, but they note that their empirical evidence using Bartik shocks actually provided stronger evidence of a high population elasticity (suggesting wages have larger effects on rents in more elastic cities), contrary to Sorens's expectation that the bias would suggest a smaller elasticity.
- Shrinking Cities: While sympathetic to the concern, H&L doubt that dropping shrinking cities (which contained only 7% of the 2000 U.S. population in their sample) would significantly impact the main quantitative results.
- Time Horizon: H&L observe that a substantial difference between Sorens's case studies and their analysis is the time frame; Sorens's examples are shorter, while their analysis spans 18 years. They share the prior belief that a longer time frame would lead to a higher population elasticity, meaning the analyses might not be contradictory.
H&L conclude by emphasizing that understanding the population elasticity to local rents ($\mu$) is a critical parameter for understanding rent increases and the scope (local or national) of the decreases resulting from construction. They welcome continued research to develop better estimation strategies for this crucial parameter.
