Endogenous Demography and Welfare Gains from Trade: Evidence from British India [Job Market Paper]

The trade literature emphasizes many channels through which market integration enhances aggregate welfare. However, for societies characterized by a Malthusian equilibrium, in which fertility and mortality respond to living standards, the long-run impact of falling trade costs on real wages is muted by population growth. I study the relationship between railroads and demography in colonial India, a context in which recent scholarship has identified large gains in real income associated with falling trade cost reductions over the late 19th and early 20th centuries. I augment a benchmark neo-Ricardian trade model with Malthusian features and show that local population levels are functions of market access as well as local agricultural productivity. Exploiting variation in trade costs associated with the gradual expansion of the railroad network between 1881 and 1931 and a new dataset of district-level vital statistics, I examine local population responses to the arrival of the railroad in both the short- and long-run. I find that mortality falls steeply, by 8-10%, and fertility rises moderately, by 6%, in the wake of railroads’ arrival. Most of this mortality-reducing effect of railroads is driven by the diminished responsiveness of mortality to local rainfall, consistent with prior findings in the literature on colonial India. My estimate of the long-run elasticity of population with respect to this trade cost reduction is around 6%. I use my model to conduct a simple back-of-the-envelope counterfactual exercise, which suggests that wages would have been 3% higher in the absence of these responses, approximately as much as measured total agricultural real wage growth over 1873-1941. My work suggests that population is an important margin for conceptualizing and measuring the impact of colonial policies on colonized countries, and has important implications for understanding the long-run implications of market integration in all pre-industrial economies.

Midnight’s Elders: Reconstructing District-Level Population Statistics in Colonial India, 1881-1931
We know very little about demographic change at high spatial and temporal resolutions for the developing world before the mid-20th century. Colonial India represents a unique setting in which colonial authorities began to collect detailed data on vital events in the late 19th century. Focusing on 150 districts of British India over five major provinces (accounting for over 200 million people), I construct annual measures of fertility, age-specific mortality, and population counts. I employ newly-collected data on district-level births and deaths by age and Census measures of district-level age-distributions. The two datasets are inconsistent with each other due to under-registration. I show that they can be reconciled, using tools from empirical demography to impose structure on the age-mortality relationship and assumptions about the temporal stability of these systematic errors within district-age-gender-decade cells. I recover underlying district-specific life-table, fertility, and under-registration parameters to match the data, and reconstruct district-level population and vital series that are demographically coherent (i.e., are stock-flow consistent). These data shed new light on a number of features of India’s demographic history that will constitute the subject of future work, including the timing and causes of its mortality transition and the long-run evolution of regional heterogeneity in fertility regimes.

Markets and Famines: Agricultural Market Efficiency in Colonial India, 1873-1883
The role of markets in exacerbating or relieving local scarcities remains contested in the literature on famine history. Prior work on famine-prone colonial India relies on annual price datasets that are too highly aggregated to discern the high-frequency spatial arbitrage that we expect in response to disequilibrium price gaps. I document a new dataset of fortnightly commodity prices during the 1870s. First, I employ an error-correction model in order to provide a new description of the properties of spatial price transmission for 19th-century India. My estimates suggest more efficient markets than has previously been understood, comparable to those for early modern Europe. Second, I use this model to study changes in these parameters during the Great Famine of 1876-1878. I find evidence that the responsiveness of price improves for most district-pairs during this time, suggesting that markets generally operated to reduce spatial price dispersion and therefore to alleviate suffering in a time of extreme scarcity. This result stands in contrast to existing findings in the literature on Europe, which finds little evidence of market efficiency changing during famines.

Pressed Hard Against the Limits: Weather and Vital Rates in the Madras Presidency, 1871-1931
The Malthusian model has strong predictions about the signs of short-run responses of fertility and mortality to economic shocks. I validate the usefulness of the Malthusian framework for understanding colonial Indian history by by analyzing the short-run dynamics of vital rates to economic shocks, employing a novel district-level panel dataset of births, deaths, and local monsoon rainfall for the Madras Presidency over 1871-1931. I find that mortality responded to shocks in a manner essentially consistent with the basic Malthusian model, although the results suggest that the rainfall-mortality relationship is complex and that more factors are needed to explain spatiotemporal variation in mortality shocks. Despite the stress contemporary observers and historians place on the lack of a preventive check in South Asia, I find that fertility is also responsive to winter monsoon shocks. I compare the magnitudes of these responses to those of early modern France and England to grain price shocks, and show that Madras is not a dramatic outlier in either fertility or mortality responses, conditional on the underlying volatility of rainfall shocks. I conclude that a fuller investigation of the causes of fertility and mortality behavior is warranted by these findings, and outline some next steps.