We provide the first national analysis of mortality in the U.S. homeless population by linking 140,000 homeless individuals from the 2010 Census to twelve years of all-cause mortality data from the Social Security Administration (SSA). Non-elderly homeless individuals face 3.5 times the mortality risk of the housed after accounting for demographic differences and geography, with a time pattern suggesting that persistently poor health, rather than homelessness itself, primarily drives this disparity. Employment, higher income, and more extensive recent family connections are associated with lower mortality, underscoring the persistence of health disparities into the extreme lower tail of socioeconomic disadvantage.

We provide evidence that similar price effects occur from new multifamily rental housing on surrounding owner-occupied property values regardless of whether the development was subsidized. These effects were on average negative in higherincome communities, but became either non-distinguishable from zero or positive in higher-income communities with sufficient population density. These results imply that previous opposition to all new rental housing by homeowners is misguided as developments could raise property values in some higher-income neighborhoods.

We propose a new inference method in high-dimensional regression models and high-dimensional IV regression models. The method is shown to be valid without requiring the exact sparsity or Lp sparsity conditions. Simulation studies demonstrate superior performance of this proposed method over those based on LASSO or random forest, especially under less sparse models. We illustrate an application to production analysis with a panel of Chilean firms.

There is significant interest in understanding labor market consequences of the opioid epidemic, but little is known about how opioid use affects on-the-job performance. We analyze the impact of opioid initiation on job performance using linked medical and personnel data for active-duty military members. Exploiting quasi-random assignment of patients to physicians in the emergency department, we find that military members assigned to high-intensity opioid prescribing physicians have a higher likelihood of long-term opioid use, are less likely to receive promotions, and are more likely to receive disciplinary actions and leave their jobs. Our results demonstrate productivity costs of opioid use.

Stigma of welfare participation is important for policy and survey design, because it reduces program take-up and increases misreporting. Stigma is also relevant to the literature on social image concerns, yet empirical evidence is scant because stigma is difficult to empirically identify. We use a novel approach to studying stigma by examining the relationship between program participation in a recipient's local network and underreporting program participation in surveys. We find a robust negative relationship and provide evidence against explanations other than stigma. Stigma decreases when more peers engage in the stigmatized behavior and when such actions are less observable.

We document the effect of unemployment insurance generosity on divorce and fertility using an identification strategy that leverages state-level changes in maximum benefits over time and comparisons across workers who have been laid off and those that have not been laid off. The results indicate that higher maximum benefit levels mitigate the effects of layoffs. In particular, they mitigate increases in divorce associated with men’s layoffs; increases in separations associated with women’s layoffs; reductions in fertility associated with men’s layoffs; and increases in fertility associated with women’s layoffs.

We examine editors’ influence on the scientific content of academic journals by unpacking the role of three major forces: journals’ stated missions, the aggregate supply of and demand for specific topics, and scientific homophily via editorial gatekeeping. In a sample of top biomedical journals, we find the first two forces explain the vast majority of variation in published content. The upper bound of the homophily effect is statistically significant but practically much less important. Marginal changes to the composition of editorial boards do not meaningfully impact journals’ content in the short run. However, we cannot rule out persistent or pervasive frictions in the publication process.

We propose principled prediction intervals to quantify the uncertainty of a large class of synthetic control predictions (or estimators) in settings with staggered treatment adoption, offering precise non-asymptotic coverage probability guarantees. From a methodological perspective, we provide a detailed discussion of different causal quantities to be predicted, which we call causal predictands, allowing for multiple treated units with treatment adoption at possibly different points in time. We illustrate our methodology with an empirical application studying the effects of economic liberalization on real GDP per capita for Sub-Saharan African countries. Companion software packages are provided in Python, R, and Stata.