Who Benefits From Productivity Growth? Local TFP Growth on Wages, Rents, and Inequality

Summary

"Who Benefits From Productivity Growth?" is a paper by Richard Hornbeck (University of Chicago Booth and NBER) and Enrico Moretti (UC Berkeley, CEPR, NBER), circulated as NBER Working Paper 24661 (May 2018, revised January 2019) and published — with a broadened title and scope, from "manufacturing productivity" to "city productivity" — as "Estimating Who Benefits from Productivity Growth: Local and Distant Effects of City Productivity Growth on Wages, Rents, and Inequality," Review of Economics and Statistics 106(3), 2024 (see Sources for the full citation). The paper asks a question with a directly Georgist edge: when productivity rises in a place, who actually captures the gain — the workers who produce it, or the owners of the land they must live on?

Its method is unusually careful about causal identification. The authors use four alternative instrumental variables to isolate exogenous local manufacturing TFP shocks across US cities (a Bartik/shift-share instrument plus three others), and trace each shock through wages, housing costs, and purchasing power for renters and homeowners. Crucially, they do not stop at the shocked city: they quantify indirect effects that propagate to other cities through worker migration — a step that, they show, reverses several conclusions a direct-effects-only analysis would reach.

What It Establishes

1. Locally, productivity gains for renters are eaten by the cost of location. In a city hit by a positive TFP shock, employment and average earnings rise substantially — but for renters those earnings are "largely offset by increased cost of living," while "for homeowners, the benefits are substantial" (abstract).[1] Much of the local benefit of productivity growth thus arrives through the housing market rather than the labor market — that is, as a claim on land, which homeowners hold and renters do not.

2. The incidence split follows the classical elasticity logic. The paper's Rosen–Roback spatial-equilibrium model makes incidence "depend on relative elasticities, as is standard in economics, and which of the two factors (labor or housing) is supplied more elastically."[1] The two extreme cases are exactly the Georgist and anti-Georgist poles: if labor is completely immobile, workers capture the full gain; if housing supply is fixed, "the entire productivity increase is capitalized into land values… and worker purchasing power is unchanged."[1] Where a city actually sits between these poles is an empirical question about housing-supply elasticity — connecting this work to the housing-supply literature.

3. Indirect (mobility) effects are large — 38% of the total. Because workers move toward productive cities, a local shock raises housing costs and wages elsewhere too. The authors "estimate that 38% of the overall increase in workers' purchasing power occurs outside cities directly affected by local TFP growth."[1] These indirect effects are larger for more-skilled (more mobile) workers, which raises inequality in receiving cities even as local productivity growth reduces inequality in the shocked city.

4. The aggregate gain is real and education-neutral. Summing direct and indirect effects, manufacturing TFP growth from 1980–1990 raised the average US worker's purchasing power by roughly 0.5–0.6% per year over 1980–2000; the gains "do not depend on a worker's education" but "mainly depend on where workers live."[1]

Relation to the Georgist Case

This paper is a mainstream, quasi-experimental revival of the Ricardian rent thesis — stated in those words — which makes it valuable Georgist evidence, but its national bottom line partly cuts the other way, and the page must carry both.

On the Georgist side, the framing is explicit. Hornbeck and Moretti write that rising living costs in high-productivity cities "benefited incumbent landowners while dampening increases in purchasing power for workers," and that this "revives the classical concern of Ricardo, in which land is in fixed supply and so landowners capture all gains from productivity growth."[1] They further argue that "the split between labor and land is potentially more consequential for understanding the effects on standard of living than the split between labor and capital that has been the main focus of recent research"[1] — a direct challenge to the Piketty/capital-share frame, redirecting attention from capital to land (compare Rognlie, who reaches the land conclusion from the capital-share side, and the capital-share-is-land claim). The mechanism it documents — productivity growth capitalizing into housing/land costs and squeezing renters — is precisely the modern mechanism behind rising land costs driving poverty.

But the honest counterweight is the paper's own headline result. Once worker mobility is accounted for, "the impacts on landowners are largely a transfer from one location to another. Thus, the overall incidence of TFP growth falls mainly on workers."[1] A positive shock in one city raises rents there but lowers landowner gains in cities that lose migrants — so at the national level, landowners do not capture all the gains the way the pure fixed-land Ricardo case predicts. The strong Ricardian outcome is the local, inelastic-housing case, not the national one. For the Georgist argument this is a genuine limit: the fixed-supply logic bites hardest exactly where housing supply is most constrained (the coastal metros), and mobility disperses landowner incidence across space rather than eliminating it — but the paper does not conclude that landowners as a class capture productivity growth economy-wide. Cited fairly, it supports the mechanism (growth capitalizes into land; renters are squeezed locally) while cautioning against the strong universal form of the claim.

Limits

• Manufacturing TFP, 1980–2000. The working paper's shocks are manufacturing-sector productivity over a specific window; the published version generalizes the framing to city productivity, but the empirics remain rooted in this setting. Extrapolation to today's tech/finance-driven agglomerations (the paper's own San Francisco/New York motivating examples) is by analogy, not direct estimate.
• The strong land-capture result is the extreme, not the estimate. Full capitalization into land holds only in the fixed-housing-supply limit; the paper's actual estimates put real cities between the poles, and its national conclusion is that incidence falls mainly on workers. Do not cite the extreme case as the finding.
• Homeowner/renter, not landlord/tenant, and no asset-market complications. The model treats workers as renters who own no other assets; homeownership is layered on in the empirics. It sets aside that some workers own real estate or firm shares, and does not model absentee or corporate landownership concentration.
• Spatial-equilibrium assumptions. Results inherit Rosen–Roback assumptions (mobility, amenities held constant in the relevant sense, functional forms from Moretti 2011); the indirect-effects estimates combine reduced-form estimates with structural assumptions about migration linkages between cities.
• Not a study of land taxation. Like Albouy, Ehrlich & Shin, the paper never discusses LVT or Henry George by name; its relevance is evidentiary (the Ricardo framing is the authors' own, but as economic analysis, not policy advocacy).

Bears On

• Problem (supports the mechanism): Rising land values and housing costs drive poverty — supplies four-IV causal evidence that local productivity growth raises housing costs enough to "largely offset" renters' earnings gains, the modern engine of George's progress-and-poverty mechanism (with the caveat that net national incidence, after mobility, falls mainly on workers).
• Problem (adjacent / See-Also, not support): Public investment capitalizes into nearby land values — that claim is specifically about public investment; Hornbeck & Moretti study private TFP growth. The capitalization mechanism is the same (any inelastically-housed local demand shock prices into land), and the authors note infrastructure investment as an analogous setting, but the paper is not evidence about public investment. Adjacent, not supporting.
• Problem: Most of the modern rise in the capital share is land, not capital — the paper's labor-vs-land framing reinforces, from the incidence side, the case for looking at land rather than capital.
• Concept: Law of Rent / Ricardian rent — a modern, instrumented restatement of Ricardo's fixed-supply logic in an urban setting.

See Also

• Rising land values and housing costs drive poverty
• Public investment capitalizes into nearby land values
• What Are Cities Worth? · Metropolitan Land Values
• Housing Constraints and Spatial Misallocation (Hsieh & Moretti — companion Moretti work on the cost of inelastic urban land)
• Rognlie: Deciphering the Fall and Rise in the Net Capital Share
• Law of Rent · Economic Rent

Sources

1. Richard Hornbeck & Enrico Moretti, "Who Benefits From Productivity Growth? Direct and Indirect Effects of Local TFP Growth on Wages, Rents, and Inequality," NBER Working Paper No. 24661, May 2018 (revised January 2019). PDF — used throughout for the abstract incidence quotes (renters' gains "largely offset," homeowner benefits "substantial"), the 38% indirect-effects figure, the explicit Ricardo-revival and labor-vs-land framing, the spatial-equilibrium incidence rule and its fixed-housing-supply capitalization limit, the 0.5–0.6%/year aggregate gain, and the conclusion that net incidence falls mainly on workers with landowner impacts "largely a transfer from one location to another." Fetched and read 2026-07-10.
2. Richard Hornbeck & Enrico Moretti, "Estimating Who Benefits from Productivity Growth: Local and Distant Effects of City Productivity Growth on Wages, Rents, and Inequality," The Review of Economics and Statistics 106(3), May 2024, pp. 587–607. DOI: 10.1162/rest_a_01208 — the peer-reviewed published version (title and scope broadened from the working paper); cited for provenance and venue. A freely available author draft is posted at eml.berkeley.edu/~moretti/tfp.pdf.