The Economics of Life and Death

This dissertation is a collection of empirical essays encompassing labor, development, and population economics. Together, they seek to understand how economic forces and incentives shape individual choices around family formation, health, and the quality of life, using rigorous identification strategies. My first two chapters study the gender-bias of technological change, and how these changes influence choices regarding fertility and mortality- some of the most consequential choices individuals make in their lives. These chapters cover settings in the developing and developed worlds, emphasizing how the prevalence and impacts of gender-specific technological change have universal relevance. My third chapter studies how individual voluntary behavior can influence health and the quality of life in local communities by focusing attention on environmental problems and changing the way people approach climate-relevant decisions over their lifetimes.

In my first chapter, Agricultural Technological Change, Female Earnings, and Fertility: Evidence from Brazil, I study the bias of technological change in developing countries. In these settings, the technological frontier is in agriculture, where tasks and occupations are starkly divided along gendered lines. I turn to the context of Brazil, which became an early adopter of new agricultural technologies when it legalized Genetically Engineered (GE) soy in 2003. I first establish new evidence of the gender-bias of technological change in agriculture, finding that the adoption of GE soy constituted a large negative demand shock to women’s work in agriculture, without any reallocation of their labor into other sectors of the economy. GE soy adoption had opposite effects on male employment in agriculture, increasing men’s and overall family earnings.

Then I turn to understanding how these differential effects on men’s and women’s earnings affect family decision-making. I write down a simple Beckerian model of fertility to show how these labor market changes can incentivize higher fertility. Assuming that children are normal goods and that women bear most of the time cost of childcare, GE soy adoption generates positive income effects for children by increasing male and overall family earnings. By reducing female earnings, it lowers the opportunity cost of time, generating positive substitution effects that further reinforce the income effects. Using administrative records from birth certificates in Brazil, I confirm the predictions of the model, and show that this female-labor saving technological change led to economically meaningful and persistent increases in fertility. I interpret these results as showing that economic development and technological change may not improve economic opportunity for women or promote fertility decline, contrary to historical experience.

In my second chapter, Blue Collar Booms and American Mortality: Evidence from US Counties, which is joint work with Paul Shaloka, we study whether a long-run technology-induced increase in economic opportunity for prime aged men can reverse rising mortality trends in the U.S. Around the turn of the 21st century, the U.S. has experienced a precipitous rise in mortality driven by suicides, drug overdoses, and alcoholic liver disease for lower-skilled men, coined ‘deaths of despair.’ While research is untangling the role of economic decline in causing this rise in mortality, there is little evidence of whether expansions in economic opportunity can reverse these trends. We investigate this by exploiting the large increase in hydraulic fracturing (fracking) across the country. The fracking boom represents significant, positive economic changes to local economies, chiefly by providing well-paying jobs for lower skilled male workers. It therefore represents another example of a gender-biased technological shock in a different setting that can have consequences on important life decisions.

We combine two sources of proprietary data in order to carry out our strategy. We first use restricted all-cause mortality data from the Centers for Disease Control and Prevention (CDC) to identify the race, sex, and primary cause of death from all registered death certificates in the U.S. from 1990-2018. To implement our empirical strategy, we purchase shapefiles on how amenable the underling geological features of different regions of the U.S. are to fracking technologies. Using this data, we compare mortality outcomes in counties best suited to take advantage of the fracking boom based on its pre-existing geological characteristics with those less able to. We find that fracking led to large increases in earnings and employment for prime-aged men, and that fracking led to reductions in mortality, driven by suicides, for young men. This suggests that some components of ‘deaths of despair’ are economically driven and speaks to the efficacy of using policies that promote local employment growth as a tool to combat rising mortality.

My final chapter, Every Day is Earth Day: Evidence on the Long-term Impact of Environmental Activism, is joint work with Dr. Daniel Hungerman. In this paper, we are interested in whether ordinary people’s voluntary actions matter in improving societal well-being. We estimate the benefits of activism, and particularly environmental activism, by considering the original Earth Day, April 22, 1970. We study how participation on Earth Day affected a community’s environmental attitudes, air quality, and children’s health. We circumvent concerns over reverse causality by exploiting the weather on the original Earth Day celebration. Unexpectedly good or bad weather are essentially random events, unrelated to other factors that affect these outcomes. Combining weather data from 1970 with survey data and newspaper archives, we provide both quantitative and qualitative evidence that worse weather on Earth Day is associated with lower participation. We then provide evidence that communities with more participation on the original Earth Day had more favorable attitudes towards the environment, particularly for those who were school aged at the time of Earth Day. Then, we show these communities experienced better air quality, driven by lower levels of Carbon Monoxide, decades after the event took place. We finally show that they had improved infant health, measured by lower incidences of congenital abnormalities. These results show that voluntary social behavior can shape the trajectories and improve the quality of life within communities in consequential and long-lasting ways.