1 Overview

The ability to trace long-run causal effects of historical events has expanded dramatically with the availability of linked administrative data and innovative quasi-experimental designs. This issue summarises four influential papers that use rigorous identification to estimate long-run effects of infrastructure investment, educational opportunity, and health interventions. Each paper exemplifies a different identification strategy and a different sector.

2 Paper 1: Donaldson and Hornbeck (2016) Railroads and US Economic Growth

Citation

Donaldson and Hornbeck [2016]: "Railroads and American Economic Growth: A 'Market Access' Approach." American Economic Review, 106(10):2584-2625.

Research Question

What was the aggregate economic impact of the nineteenth-century US railroad network expansion? And what would have happened to the US economy if railroads had never been built—if only waterways and roads existed?

Identification Strategy

Standard regressions of county-level income on railroad access suffer from obvious endogeneity: railroads were built where economic activity was expected to grow. Donaldson and Hornbeck address this by using a market access approach. They compute for each county a measure of trade access—a CES aggregate of trade opportunities with all other counties, weighted by bilateral trade costs based on the transport network. As new railroad lines open, market access changes for connected counties.

The key identification is cross-county variation in market access improvements driven by the route of the railroad network, which was determined partly by engineering constraints and political economy rather than solely by pre-existing economic activity. They use changes in the railroad network over time as instruments for changes in market access.

Key Results

A one standard deviation increase in market access raises agricultural land values by approximately 45%. In a counterfactual where railroads are replaced by roads (no railroad construction), total US agricultural land value in 1890 would have been roughly 60% lower.

The market access framework is flexible: it allows the authors to compute general equilibrium welfare effects (the Harberger triangle plus the trade-cost savings), yielding an aggregate welfare gain from railroads of roughly 3.2% of GDP by 1890. This is substantially larger than the direct consumer surplus estimate from Fogel [1964].

Takeaway

Railroads had large general equilibrium effects on US economic geography that are not captured by partial equilibrium cost-benefit analysis. The market access approach offers a tractable way to aggregate effects across heterogeneous locations with different network connections.

3 Paper 2: Aaronson and Mazumder (2011) The Rosenwald Schools

Citation

Aaronson and Mazumder [2011]: "The Impact of Rosenwald Schools on Black Achievement." Journal of Political Economy, 119(5):821-888.

Research Question

Did the Rosenwald Fund's construction of over 5,000 schools for Black children in the U.S. South between 1913 and 1931 improve educational attainment and earnings for Black Americans?

Identification Strategy

The Rosenwald programme provided matching grants to Southern counties to construct schools for Black children, conditional on local matching funds. The timing and location of school construction varied across counties and cohorts in ways related to local fundraising capacity and state education authorities—providing quasi-experimental variation in access to the Rosenwald schools.

Aaronson and Mazumder [2011] use a difference-in-differences design: they compare Black children of school age during the Rosenwald construction period to those who were too young or too old to have attended, and compare counties that received Rosenwald schools to those that did not (or received them later), using Census microdata linked across decades.

Key Results

Counties that received Rosenwald schools saw increases in Black educational attainment of 0.4 years of schooling and a 5 percentage point increase in high school completion rates. Returns on educational attainment translated into labour market gains: earnings of Black men born in counties with Rosenwald schools were 7% higher than comparable men in counties without schools.

Effects were concentrated in counties with initially lower levels of existing school quality, consistent with an interpretation that the programme expanded the frontier of educational access rather than improving already-adequate schools.

Takeaway

Educational infrastructure investment in an underserved population with minimal access to schooling has large and lasting effects on attainment and earnings. The Rosenwald case also illustrates how philanthropic interventions with geographic variation can be studied rigorously using DiD-in-exposure designs.

4 Paper 3: Bleakley (2010) Malaria Eradication in the Americas

Citation

Bleakley [2010]: "Malaria Eradication in the Americas: A Retrospective Analysis of Childhood Exposure." American Economic Journal: Applied Economics, 2(2):1-45.

Research Question

Did the eradication of malaria in the early-to-mid twentieth century raise human capital and labour market outcomes for children who grew up in high-malaria areas?

Identification Strategy

Bleakley [2010] exploits differential exposure to malaria eradication campaigns across countries (Brazil, Colombia, Mexico, and the United States) and across cohorts within each country. The identifying variation is the interaction of pre-eradication malaria prevalence (which region had how much malaria) and birth cohort relative to the eradication campaign (who was of childhood age during the eradication).

This is a DiD-in-exposure design: the treatment is the eradication campaign, the "dose" is pre-existing malaria burden, and the comparison is between cohorts who received the eradication during childhood versus cohorts who were either born after malaria was already gone or were too old for childhood exposure to matter.

Key Results

Regions with higher pre-eradication malaria showed larger improvements in literacy, school enrolment, and income for cohorts exposed to eradication during childhood, compared to cohorts in the same regions born before or after the eradication window.

In the United States, the eradication of malaria in the South in the 1920s-1930s raised literacy by approximately 6 percentage points and income by 14-18% for cohorts exposed during childhood relative to unexposed cohorts in the same high-malaria regions.

Takeaway

Early-life disease burden has large and persistent effects on human capital formation. The DiD-in-exposure design—exploiting variation in both pre-existing burden and cohort timing—is a powerful template for evaluating disease eradication and prevention programmes when there is no randomisation.

5 Paper 4: Acemoglu, Johnson, and Robinson (2001) Settler Mortality and Institutions

Citation

Acemoglu et al. [2001]: "The Colonial Origins of Comparative Development: An Empirical Investigation." American Economic Review, 91(5):1369-1401.

Research Question

Do institutions—specifically, the security of property rights and the constraints on executive power—have a causal effect on long-run economic development across countries?

Identification Strategy

Cross-country regressions of income on institutional quality are plagued by endogeneity: rich countries can afford better institutions, and common historical factors affect both. Acemoglu et al. [2001] use settler mortality during the colonial period as an instrument for current institutions.

The logic: in colonies where European settlers faced high mortality from disease (malaria, yellow fever), colonial powers established extractive institutions rather than investing in property rights and democratic governance. These extractive institutions persisted after independence and remain associated with weaker institutions today. Settler mortality in the colonial period is plausibly exogenous to current economic performance—it reflects the disease environment of the 1700s and 1800s, not anything about current production possibilities.

Key Results

The IV estimate finds a one standard deviation improvement in institutional quality raises per capita income by approximately 1.0 log points. The OLS estimate is smaller (roughly 0.5), consistent with reverse causality (rich countries affording better institutions) attenuating the true causal effect downward.

The instrument is strong: first-stage F-statistics exceed 30 in most specifications. Placebo tests using settler mortality in neighbouring colonies (which should not affect local institutions) fail to predict institutional quality.

Takeaway

Institutions have a large causal effect on long-run income. The settler mortality instrument has become a benchmark in the comparative development literature, though subsequent work has scrutinised its validity—in particular whether the disease environment affects current income through channels other than institutions (e.g. direct effects on current labour productivity through health).

References

1. Aaronson, D. and Mazumder, B. (2011). The impact of Rosenwald schools on Black achievement. Journal of Political Economy, 119(5):821-888.
2. Acemoglu, D., Johnson, S., and Robinson, J. A. (2001). The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5):1369-1401.
3. Bleakley, H. (2007). Disease and development: Evidence from hookworm eradication in the American South. Quarterly Journal of Economics, 122(1):73-117.
4. Bleakley, H. (2010). Malaria eradication in the Americas: A retrospective analysis of childhood exposure. American Economic Journal: Applied Economics, 2(2):1-45.
5. Donaldson, D. and Hornbeck, R. (2016). Railroads and American economic growth: A "market access" approach. American Economic Review, 106(10):2584-2625.
6. Fogel, R. W. (1964). Railroads and American Economic Growth: Essays in Econometric History. Johns Hopkins Press, Baltimore.
7. Dell, M. (2010). The persistent effects of Peru's mining mita. Econometrica, 78(6):1863-1903.

‍