Pro-Poor Land Reforms, Education and Fertility (Madsen, Tong & Ulubasoglu, 2026)
A global panel of 388 major land reforms in 165 countries, 1820-2010, finds that reforms aimed at the poor raised school enrollment and lowered fertility, while reforms with other motives did not; it concerns redistribution and tenure, not land taxation.
Summary
"Pro-poor land reforms, education and fertility: international evidence, 1820-2010," by Jakob B. Madsen, Lan A. Tong and Mehmet A. Ulubasoglu, appeared in the Journal of Economic Growth on 16 November 2026 (open access). Using a global panel, the authors report that land reforms whose stated beneficiaries are the poor are followed by higher school enrollment and lower fertility, whereas reforms with other motives show no such pattern.
This is not evidence about land value tax. The reforms studied redistribute land titles or secure tenure; none is a tax on land rent. Its relevance is to the Georgist diagnosis, the land-monopoly argument that concentrated land ownership and the power of landed elites hold back ordinary people's investment in themselves. It bears on that shared diagnosis, not on Henry George's remedy.
The Data and Design
- Reform data. The authors extend the land-reform dataset of Bhattacharya et al. (2019) back to 1820. It contains 388 major land reforms across 165 countries over 1820-2010 (16 in 1820-1899 and 372 in 1900-2010). Of these, 141 are classed "pro-poor" because the laws name the poor, the landless, peasants or subsistence farmers as beneficiaries. In 123 of the 141, land title is transferred from wealthy landowners or public holdings to the poor; the remainder improve tenure security (Sect. 2).
- Outcomes. Gross enrollment rates (primary, secondary, tertiary) come from Lee and Lee (2016) and overlap with the reform data for 104 countries and 4,056 observations. Crude fertility overlaps for 155 countries and 13,471 observations (Sect. 2).
- Identification. The authors use event studies, semi-parametric difference-in-differences, two-way fixed effects, an instrumental-variable check and placebo tests, with controls including income, democracy, urbanization and geography (Sect. 3).
Findings
- Enrollment. Pre-reform coefficients are insignificant. After a country's first pro-poor reform, enrollment increases are in the ranges of 26-34% (primary), 14-27% (secondary) and 16-51% (tertiary). Primary gains reach their long-run level after about 15 years, secondary and tertiary after about 25 (Sect. 3.1.1). The semi-parametric estimates are smaller for primary and broadly similar for secondary and tertiary.
- Fertility. Reforms lower the crude fertility rate by 4.6 per 1,000 population, an 11.8% decline from the pre-reform mean of 38.9, reaching a long-run level after about 35 years (Sect. 3.1.2). Delayed effects grow: 16.7% (6.5 per 1,000) after 30-35 years (Sect. 3.4, citing Table 3).
- Placebo. Expropriation and redistribution motives look like the pro-poor results. Tenure-security-only, administrative, collectivization and nationalization reforms do not; pooled non-pro-poor reforms are, if anything, associated with lower enrollment and higher fertility. Nationalization lowers enrollment at every level, which the authors read as showing that "the distribution of property rights, rather than state ownership per se, is a key mechanism" (Sect. 3.5).
- Instrument. Using CIA-supported US interventions in 1947-1989 as an instrument gives estimates the authors call "broadly consistent" with the main results (Sect. 3.3).
- Channels. The results are consistent with a credit channel (private credit), a family-farm channel and, more weakly, a compulsory-schooling channel. The credit channel is the strongest for enrollment; the compulsory-schooling coefficients are not statistically significant for secondary and tertiary (Sect. 4).
Relation to the Georgist Case
The authors' political-economy story is close to Georgist concerns: landed elites have reason to oppose public schooling because schooling raises rural wages and reduces the returns to land. They write that reform "can weaken this political resistance by redistributing economic and political power away from landed elites" (Sect. 1). That is an empirical case that land concentration has costs beyond rent extraction, namely lower human-capital investment and slower demographic transition.
The result should not be over-read. George rejected dividing up land titles and proposed taxing economic rent instead, so a result about redistribution tests a different remedy for the same problem. The paper says nothing about whether collecting economic rent through a tax would widen schooling or reduce elite power, and it has no land-tax reform in its sample. Taiwan's land-to-the-tiller programme of 1949-53 is the kind of episode such a dataset would code; the paper's reported results are pooled, not country by country.
Nuances and Limits
- Causality is argued, not randomized. Reforms are not randomly assigned, and reforming and non-reforming countries differ systematically. The identification rests on controls, pre-trend checks and an instrument whose validity is an assumption (Sects. 3.1, 3.3).
- Treatment is defined by stated beneficiaries. "Pro-poor" depends on how laws describe their aims, and 65% of enacted reforms in the sample were implemented (Sect. 3.2.1).
- Channels are suggestive. The mediation analysis finds the mechanisms consistent with the data, not proven; the supply-side (elite power) channel is inferred rather than directly measured.
- Income effects are inferred. The authors extrapolate to per capita income using elasticities from other studies; this is a back-of-the-envelope calculation from other studies' elasticities, not an estimate from their own data.
- Historical scope. The authors state that their results do not establish that today's low-income agrarian economies would see the same effects (Sect. 5).
See Also
- Land Monopoly
- Economic Rent
- Taiwan Land Reform (1950s)
- Kerspien, Madsen & Strulik: Buildings and the Labor Share (same first author, different paper)
Sources
- Jakob B. Madsen, Lan A. Tong & Mehmet A. Ulubasoglu (2026), "Pro-poor land reforms, education and fertility: international evidence, 1820-2010," Journal of Economic Growth, published 16 November 2026, open access. doi.org/10.1007/s10887-026-09274-w — open-access full text read; regression tables and the online appendix were not consulted, so only figures stated in the running text are used. Peer-reviewed journal article (Tier 1) — used for the reform dataset and sample sizes (Sect. 2), event-study enrollment and fertility estimates (Sects. 3.1.1-3.1.2, 3.4), placebo and instrument checks (Sects. 3.3, 3.5), mediating channels (Sect. 4) and the authors' stated limits (Sect. 5); B-claim.