Research

I study the effect of cigarette tax increases on cross-state border shopping in the United States. To estimate this relationship, I use high-resolution census block group-by-month cellphone tracking data from Safegraph. I estimate a Callaway and Sant’Anna (2021) difference-in-differences model that accommodates my unique setting in which the tax increases I consider become effective at different times throughout the full length of the study. I find that the median census block group (CBG) sent 0.53 more cross-border shoppers per month in response to a cigarette tax increase (19% of the pre-tax mean). I also estimate that the increase in cross-border shopping is substantially larger for those that live closer to a lower-tax border. Finally, I find that CBGs with many low educated adults and rural CBGs send substantially more cross-border shoppers than their respective counterparts. In sum, these results suggest that cross-border shopping remains an ongoing challenge for tobacco control policy efforts and for reducing tobacco-related disparities.

In this paper, we examine how cigarette taxes and indoor smoking restrictions (ISR) have impacted maternal prenatal smoking. We use the National Vital Statistics System birth certificate data, which provides the near universe of births in the USA, from 1995-2018. This data also provides intensity of maternal cigarette use for each trimester during pregnancy. We use a stacked difference-in-difference (Cengiz et. al. 2019) approach for this research topic by exploiting only large expansions in state-level ISR coverage and cigarette taxes over this period. While we find that large ISR expansions can decrease the intensity of prenatal smoking, we estimate that this decrease is small. We further find little evidence that cigarette taxes can impact maternal prenatal smoking. This paper contributes to a growing literature whose conclusion is that cigarette taxes have little impact on adult or youth consumption. 

In this paper, I examine how cigarette taxes and indoor smoking restrictions (ISR) can impact the time use of probable cigarette users. To study this relationship, I use the American Time Use Survey (ATUS) linked with the Tobacco Use Supplement of the Current Population Survey (TUS-CPS) from 2001-2019. This linked data not only allows me to see how individuals use their time on a particular day, but also whether individuals are smokers or not. By estimating a logit model on the overlap sample between the ATUS and TUS-CPS, I estimate the predicted probability that a respondent to the ATUS consumed cigarettes or not. I then used this predicted probability in a stacked (Cengiz et. al. 2019) triple difference model for only large expansions in ISR coverage and cigarette taxes over this period. Overall, I find that these policies didn’t significantly impact time use separately for those likely to be smokers versus those who were not likely. However, I did find a slight increase in time spent exercising for those more likely to be smokers in response to a large increase in ISR coverage.

In their seminal paper, Lang and Manove (2011) argue that conditional on a test measurement of ability, Black individuals attain more years of education than whites. The paper goes to show that, despite this, Black workers are not paid more than white workers when conditioning on this measurement of ability. We explore whether Black individuals with the same measurement of ability as a similar white individual have more years of education because they completed their high school and/or college degree in more time than is expected for the degree. We speculate that this may be a mechanism that is at least partially driving the non-difference in wages when comparing Black and white workers conditional on a measurement of ability.