A Mediation Analysis of the Relationship Between Land Use Regulation Stringency and Employment Dynamics

Shaoming Cheng; Uche Oluku

arxiv: 2305.02159 · v1 · pith:LL4N2QEEnew · submitted 2023-05-03 · 💰 econ.GN · q-fin.EC

A Mediation Analysis of the Relationship Between Land Use Regulation Stringency and Employment Dynamics

Uche Oluku , Shaoming Cheng This is my paper

Pith reviewed 2026-05-24 08:56 UTC · model grok-4.3

classification 💰 econ.GN q-fin.EC

keywords land use regulationhousing cost burdenemployment growthmediation analysisWRLURIsectoral employmentUS local jurisdictions

0 comments

The pith

Stricter land use rules raise renter housing costs, which then reduce employment growth in professional and information sectors.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether land use regulations affect job growth mainly by pushing up housing costs for renters. Using data from 878 U.S. jurisdictions, it applies a mediation model that separates direct effects of regulation from effects that run through the share of cost-burdened renters. The results show the entire link between higher regulation and slower employment growth operates through the housing-cost channel. A one-standard-deviation rise in the regulation index lifts the share of cost-burdened renters by roughly 0.8 percentage points, and each extra percentage point of such renters trims employment growth by 0.04 points in the professional sector and 0.017 points in the information sector.

Core claim

The relationship between land use regulation stringency, measured by the WRLURI index, and employment growth in the Retail, Professional, and Information sectors is fully mediated by the proportion of cost-burdened renters. A one standard deviation increase in the WRLURI index raises the share of cost-burdened renters by about 0.8 percentage points, while a one percentage point rise in that share lowers employment growth by 0.04 percentage points in the Professional sector and 0.017 percentage points in the Information sector.

What carries the argument

Mediation analysis framework that decomposes the total effect of the WRLURI regulation index on sectoral employment growth into the portion that travels through the housing cost burden mediator.

If this is right

Higher regulation stringency produces a measurable rise in the share of cost-burdened renters.
The increase in cost-burdened renters accounts for the observed slowdown in Professional sector employment growth.
The same housing-cost channel accounts for the slowdown in Information sector employment growth.
No statistically detectable direct effect of regulation on employment growth remains once the mediator is accounted for.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

If the mediation pattern holds in other sectors or time periods, easing land use rules could raise employment growth by lowering renter housing costs.
The size of the employment effect may differ across metro areas that vary in the share of workers in professional or information occupations.
Policy changes that target housing affordability without altering regulation stringency could produce similar employment gains.

Load-bearing premise

No unmeasured factors simultaneously influence both the share of cost-burdened renters and employment growth, and the WRLURI index is independent of local economic conditions.

What would settle it

A dataset or natural experiment in which jurisdictions change land use rules but the share of cost-burdened renters stays constant, yet employment growth in the Professional or Information sectors still declines.

Figures

Figures reproduced from arXiv: 2305.02159 by Shaoming Cheng, Uche Oluku.

read the original abstract

The paper examines the effects of stringent land use regulations, measured using the Wharton Residential Land Use Regulatory Index (WRLURI), on employment growth during the period 2010-2020 in the Retail, Professional, and Information sectors across 878 local jurisdictions in the United States. All the local jurisdictions exist in both (2006 and 2018) waves of the WRLURI surveys and hence constitute a unique panel data. We apply a mediation analytical framework to decompose the direct and indirect effects of land use regulation stringency on sectoral employment growth and specialization. Our analysis suggests a fully mediated pattern in the relationship between excessive land use regulations and employment growth, with housing cost burden as the mediator. Specifically, a one standard deviation increase in the WRLURI index is associated with an approximate increase of 0.8 percentage point in the proportion of cost burdened renters. Relatedly, higher prevalence of cost-burdened renters has moderate adverse effects on employment growth in two sectors. A one percentage point increase in the proportion of cost burdened renters is associated with 0.04 and 0.017 percentage point decreases in the Professional and Information sectors, respectively.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper builds a usable WRLURI panel and reports small mediation coefficients through renter cost burden, but the full-mediation claim rests on untested no-confounder assumptions with no identification strategy shown.

read the letter

The two things to know are that the authors created a balanced panel of 878 jurisdictions from the 2006 and 2018 WRLURI waves and matched it to 2010-2020 sectoral employment data, and that they apply a standard mediation decomposition claiming the regulation-employment link runs entirely through the share of cost-burdened renters. The reported numbers are a 0.8 percentage point rise in cost burden per standard deviation of the index, followed by 0.04 and 0.017 percentage point drops in professional and information employment growth per percentage point rise in cost burden. Retail shows no effect. That panel construction and the sector-specific estimates are the concrete addition here. The method itself is off-the-shelf mediation applied to this data source, which gives a reader a descriptive channel worth noting if they already work on housing-labor links. The soft spot is exactly what the stress-test note flags. Mediation requires that nothing else jointly moves the mediator and the outcome once regulation is controlled, and that WRLURI itself is not driven by local employment or housing conditions. The abstract supplies no instruments, no regression-discontinuity design, no discussion of fixed effects that would absorb those confounders, and no robustness tables. Without those pieces the “fully mediated” language stays associational. The paper is aimed at urban and regional economists who track land-use effects on local labor markets. A reader already interested in the WRLURI data or in housing-cost channels could pull the panel construction and the coefficient magnitudes as starting points. It is not going to reset the broader literature. The work deserves a serious referee because the data effort is real and the policy question is live; the identification gaps are fixable in revision rather than fatal. Send it out and ask for the full specification, standard errors, controls, and sensitivity checks on the mediation assumptions.

Referee Report

2 major / 1 minor

Summary. The paper examines the effects of land use regulation stringency (WRLURI) on employment growth in Retail, Professional, and Information sectors (2010-2020) across 878 US jurisdictions using a mediation framework with housing cost burden as mediator. It reports a fully mediated relationship, with a one SD increase in WRLURI associated with a 0.8 pp rise in cost-burdened renters and subsequent negative effects on employment growth in two sectors (0.04 pp and 0.017 pp per pp increase in cost burden).

Significance. If supported by a credible identification strategy, the mediation decomposition would add to the literature on regulatory impacts on local labor markets by quantifying the housing-cost channel. The construction of a balanced panel from the 2006 and 2018 WRLURI waves is a constructive data contribution.

major comments (2)

[Abstract and §3] Abstract and §3 (Empirical Strategy): The reported point estimates and 'fully mediated' conclusion are presented without any model equations, list of control variables, standard errors, fixed effects, or robustness checks, so the central claim cannot be evaluated from the supplied information.
[§3] §3 (Mediation Framework): The decomposition into direct and indirect effects rests on the untestable assumptions of no unmeasured confounders jointly affecting the cost-burden mediator and sectoral employment and of WRLURI exogeneity; the manuscript supplies no instruments, regression-discontinuity design, or sensitivity analysis to support these assumptions, which are load-bearing for any causal interpretation of full mediation.

minor comments (1)

[Abstract] Abstract: The phrase 'approximate increase of 0.8 percentage point' should be replaced by the exact coefficient and its standard error once the model is specified.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments. We address each major comment below and note planned revisions.

read point-by-point responses

Referee: [Abstract and §3] Abstract and §3 (Empirical Strategy): The reported point estimates and 'fully mediated' conclusion are presented without any model equations, list of control variables, standard errors, fixed effects, or robustness checks, so the central claim cannot be evaluated from the supplied information.

Authors: We agree that the abstract and §3 omit the mediation model equations, the full list of control variables, standard errors, fixed effects, and robustness checks. In the revised manuscript we will add the complete set of mediation equations, enumerate all controls, report standard errors and fixed effects, and include robustness checks so that the central claims can be evaluated. revision: yes
Referee: [§3] §3 (Mediation Framework): The decomposition into direct and indirect effects rests on the untestable assumptions of no unmeasured confounders jointly affecting the cost-burden mediator and sectoral employment and of WRLURI exogeneity; the manuscript supplies no instruments, regression-discontinuity design, or sensitivity analysis to support these assumptions, which are load-bearing for any causal interpretation of full mediation.

Authors: The referee correctly identifies that the mediation decomposition relies on the standard sequential ignorability and exogeneity assumptions, which are untestable, and that the manuscript contains no IV, RD design, or sensitivity analysis. The paper presents the results as a mediation decomposition rather than a strong causal claim. In revision we will explicitly state these assumptions in §3, discuss their plausibility given the panel structure and controls, and add sensitivity analyses where feasible with the data. revision: partial

Circularity Check

0 steps flagged

No circularity: standard empirical mediation on observational panel data

full rationale

The paper applies a mediation framework to decompose associations between WRLURI, housing cost burden, and sectoral employment growth using jurisdiction-level panel data. All reported quantities are fitted regression coefficients from the data; no algebraic identity, self-referential definition, or self-citation chain reduces any coefficient to an input by construction. The central claims rest on the usual mediation assumptions (no unmeasured confounding) rather than on any tautological reduction of the reported effects.

Axiom & Free-Parameter Ledger

2 free parameters · 2 axioms · 0 invented entities

Empirical mediation study; central claim rests on standard econometric assumptions for identifying indirect effects and on the validity of the WRLURI index as a regulation measure. No new entities postulated.

free parameters (2)

WRLURI-to-mediator coefficient
The reported 0.8 pp association per SD is an estimated parameter from the mediation model.
mediator-to-employment coefficients
The reported -0.04 and -0.017 pp associations are fitted from the data.

axioms (2)

domain assumption No unmeasured confounding between mediator and outcome
Required for causal interpretation of the indirect effect in mediation analysis.
domain assumption WRLURI index validly and exogenously measures regulation stringency
Central to interpreting the first-stage association as a policy effect.

pith-pipeline@v0.9.0 · 5737 in / 1483 out tokens · 42430 ms · 2026-05-24T08:56:54.227376+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

3 extracted references · 3 canonical work pages

[1]

WRLURI2006

We merged both the 2006 (“WRLURI2006”) and 2018 (“WRLURI2018”) indexes in the analysis and identified 878 localities that exist in both surveys. The extent of the jurisdictions that appear in both rounds of surveys are comparable to the estimates in Gyourko, Hartley, and Krimmel (2021) and Mleczko and Desmond (2023). All 878 jurisdictions are matched with...

work page 2006
[2]

strong selection effects in who answered the underlying questions in both surveys that could bias our conclusions

They concluded that productivity shocks due to misallocation of labor in New York, San Francisco and San Jose affected wages and labor productivity not only in those cities, but elsewhere. Relationship between Land Use Regulation and Housing Unaffordability: Glaeser & Gyourko, (2003), Lin & Wachter (2019), and Landis & Reina (2021) examined the relationsh...

work page doi:10.1080/10511482.2022.2065328 2003
[3]

Retrieved on May 21, 2022, from https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/margin.html Evenson, B., Wheaton, W

Stern School of Business at New York University. Retrieved on May 21, 2022, from https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/margin.html Evenson, B., Wheaton, W. C., Gyourko, J., & Quigley, J. M. (2003). Local Variation in Land Use Regulations [with Comments]. Brookings-Wharton Papers on Urban Affairs, 221–260. http://www.jstor.org/stable...

work page doi:10.1016/b978-0-444-59531-7.00019-3 2022

[1] [1]

WRLURI2006

We merged both the 2006 (“WRLURI2006”) and 2018 (“WRLURI2018”) indexes in the analysis and identified 878 localities that exist in both surveys. The extent of the jurisdictions that appear in both rounds of surveys are comparable to the estimates in Gyourko, Hartley, and Krimmel (2021) and Mleczko and Desmond (2023). All 878 jurisdictions are matched with...

work page 2006

[2] [2]

strong selection effects in who answered the underlying questions in both surveys that could bias our conclusions

They concluded that productivity shocks due to misallocation of labor in New York, San Francisco and San Jose affected wages and labor productivity not only in those cities, but elsewhere. Relationship between Land Use Regulation and Housing Unaffordability: Glaeser & Gyourko, (2003), Lin & Wachter (2019), and Landis & Reina (2021) examined the relationsh...

work page doi:10.1080/10511482.2022.2065328 2003

[3] [3]

Retrieved on May 21, 2022, from https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/margin.html Evenson, B., Wheaton, W

Stern School of Business at New York University. Retrieved on May 21, 2022, from https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/margin.html Evenson, B., Wheaton, W. C., Gyourko, J., & Quigley, J. M. (2003). Local Variation in Land Use Regulations [with Comments]. Brookings-Wharton Papers on Urban Affairs, 221–260. http://www.jstor.org/stable...

work page doi:10.1016/b978-0-444-59531-7.00019-3 2022