Earned income tax credit policy and unhealthy alcohol use: A difference-in-differences analysis of U.S. states, 2001-2019

William Tsang | 2022

Advisor: Ali Rowhani-Rahbar

Research Area(s): Social Determinants of Health

Full Text

Background: Unhealthy alcohol use is a major contributor to chronic diseases and mortality in the United States. There is variation in generosity of state Earned Income Tax Credit (EITC) which supplements the federal EITC, the largest poverty-alleviation program for low-earning working adults and their families. Prior studies have found income and employment to be associated with risk factors that lead to unhealthy alcohol use. However, none have found significant associations between EITC and unhealthy alcohol use. Our study builds on this body of evidence by controlling for state alcohol policy environment and estimating heavy episodic drinking outcomes.

Methods: We used the Behavioral Risk Factors Surveillance System survey to examine associations between presence and generosity of state refundable EITC and three measures of unhealthy alcohol use: heavy episodic (≥4 and ≥5 drinks per occasion for women and men, respectively), risky (≥7 and ≥14 drinks per week for women and men), and any unhealthy alcohol use (met criteria for either pattern in the past month). We used a generalized difference in-differences approach to compare exposed states to control states from 2001 to 2019. We used quasi-Poisson regression models with state and year-fixed effects controlling for state level covariates to obtain survey-weighted prevalence ratios for each alcohol outcome.

Results: From 2001 to 2019, 11 states enacted refundable EITCs. Median state credit as a percentage of federal credit was 10% and as high as 85%. A 10-percentage point increase in state EITC generosity was significantly associated with heavy episodic drinking (PR = 1.01, 95%CI: 1.00 – 1.01) but not for risky (PR = 1.01, 95%CI: 1.00 – 1.01) or any unhealthy alcohol use (PR = 1.01, 95%CI: 1.00 – 1.01). EITC and patterns of unhealthy alcohol use remained weakly associated after restricting analysis to the population with high school (or lower) educational attainment (since this group is more likely to receive EITC), months when credits are typically refunded, and excluding California for its high EITC generosity.

Conclusion: We found weak associations between EITC generosity and three measures of unhealthy alcohol use that were attenuated by adjustment with time-varying confounders. Consistent with previous findings, these results suggest minimal association between EITC and unhealthy alcohol use. Our study is the first to measure the association of EITC and heavy episodic drinking and adjust for state alcohol policy environment. These findings inform understanding of the impacts of EITC and other poverty-alleviation policies on alcohol-related harm and provide impetus for policymakers to examine whether generosity is adequate in current and future policies.