Environmental degradation, and climate change in particular, are increasingly recognized to be among humanity's largest threats. Calls to protect our ``Common Home'' are becoming more frequent and resolute. While firmly rooted in science, these declarations often build on the implicit assumption that we are all in the same boat. We are not. Environmental degradation has profoundly unequal impacts on individuals depending on where they were born, how much money they inherited, or the color of their skin. Some people will drown, and others are buying 400-feet yachts.
The dangers posed by environmental degradation and the assurance of unequal impacts present colossal challenges for public policy. Even if mitigation actions can still reduce the worst impacts, scientists warn us that massive damages are already inevitable. Yet, because these damages will not be borne by everyone, enacting effective and fair public policies needs to account for variation in these impacts as well as in people's capacities to adapt to them. How should societies best protect their people against environmental risks?
Figuring out the set of policies to best protect individuals presents a unique opportunity for research, and economics in particular. For all the powerful tools that economics offer, tackling this question requires two substantial departures from the current practice of neoclassical economics. First, recognizing that the concept of (potential) Pareto efficiency as a theoretical guide can only take us so far. Policies always create losers; given that the people who lose from a policy are seldom compensated for their troubles, it is not nearly enough to know that a policy could potentially improve the welfare of society if the right transfers were implemented. Pareto efficiency is not a morally neutral concept -- it always gives more weight to a subset of individuals, even if we often do not know who these individuals are. More complete analyses require accounting for the distribution of costs and benefits within the population in order for moral ideals to be an integral and explicit part of policy discussions.
The second departure is the empirical counterpart of the first one. Most work in applied economics and policy analysis focuses on the estimation of average effects, where the average is either taken over a large group of individuals exposed to the policy, over a small group of individuals around some cutoffs (with regression discontinuities), or over a sample of "compliers" that we cannot identify (for instance with instrumental variables). Yet, the very moment we start thinking about who gains and who loses from a policy, average effects become quite inadequate. It is not enough to know that the average effect of a policy is positive. For instance, subsidies for individual solar panels could lead to more installations, but these installations might be concentrated at the very top of the income distribution, which could leave disadvantaged households to pay higher energy prices (due to the rising share of fixed costs in the distribution system, as this is the case in California). Taking compensation and distributional concerns seriously requires the use of less common empirical tools.
I am fortunate to have started my PhD at a time when economics is undergoing a major shift in both directions. The explicit focus on distributional concerns is becoming more legitimate within the neoclassical framework, and this trend is reflected in environmental economics. While sociologists and epidemiologists have studied environmental justice for over 40 years, environmental economists are only starting to routinely take these distributional concerns into account. Similarly, an explosion of recent work in econometrics is pushing the frontiers regarding the estimation of heterogeneous treatment effects, the development of quantile estimators, and the identification of individual-treatment effects to better understand the distribution of policy impacts within a population. This dissertation builds on both trends to investigate environmental inequalities and social insurance in the United States.
The first chapter investigates the public provision of climate risk information and its distributional impacts on the demand for residential flood insurance. Flooding is among the costliest disasters in the United States, but the demand for federally-subsidized insurance is extremely low. I find that inaccurate flood maps explain a substantial share of low insurance take-up. Using a novel approach based on unit-specific synthetic controls, I estimate that official map updates over the past two decades caused substantial declines in the demand for flood insurance, primarily in neighborhoods with a higher share of African Americans. I quantify the welfare costs of these inaccurate map updates, and estimate the distribution of gains that would accrue from the provision of better flood risk information. The second chapter highlights stark mobility inequalities between income groups during the early stages of the COVID-19 pandemic, at a time when media coverage routinely used guilt to convince people to stay home. Guilt may work, but if sustained mobility reflects labor constraints rather than lifestyle preferences, money or a functioning safety net would probably have been better to keep people home. The third chapter emphasizes major shortcomings with the methodology used to estimate the impacts of public policies during the pandemic, and shows that standard robustness tests associated with fixed effects models should be viewed with considerable caution. Finally, the fourth chapter presents early findings regarding the impacts of housing choice vouchers on the residential outcomes and environmental exposure of tenants.