Date of Degree
tobacco, tobacco cessation, alcohol, food, household economics
This dissertation was designed to: (1) estimate the relationships between recent, long-term, and relapsed tobacco cessation and dollars spent on non-tobacco goods among households in the U.S., (2) estimate the relationships between recent, long-term, and relapsed tobacco cessation and the budget shares allocated to non-tobacco goods among households in the U.S., and (3) to identify and characterize unobserved heterogeneity in the relationship between tobacco cessation and dollars spent on alcohol and food.
Using 2006-2012 Consumer Expenditure Survey (CES) data, a cohort of 6,739 tobacco-consuming households was created and followed for four quarters. Households were categorized during the fourth quarter as having: (1) recent tobacco cessation, (2) long-term cessation, (3) relapsed cessation or (4) no cessation. Generalized linear models and fractional logistic regression were used to compare the fourth quarter expenditures and budget shares of eight categories (alcohol, food at home, food away from home, housing, health care, transportation, entertainment and other) between the no cessation households and those with recent, long-term or relapsed cessation. All models controlled for potential confounding variables, including household sociodemographics and economic changes during the study’s four quarters (e.g., income change, state-level change in unemployment, household job loss). Finite mixture modeling (FMM) was used to further examine unobserved heterogeneity in the relationships between cessation and household spending on food and alcohol.
Compared to households that did not quit during the study, households with long-term and recent cessation had significantly lower fourth quarter spending on alcohol, entertainment, and transportation. The reduced spending on alcohol among households with recent or long-term cessation was driven by a significantly higher proportion of households with no alcohol spending in the fourth quarter. Recent cessation was further associated with reduced spending on food at home, while relapsed cessation was associated with higher spending on food away from home. There were no significant differences in dollars spent on health care, housing, or other goods between households that did not quit and those with any type of cessation.
The analysis of fourth quarter budget share data found that long-term cessation was associated with a significantly lower share of the budget allocated to alcohol and entertainment, yet a higher share to housing, health care and food. Recent cessation was associated with a significantly lower budget share for alcohol, yet a higher share on housing and ‘other’ goods or services. There were no significant differences between households with recent cessation and those that did not quit in the share of the budget allocated to health care, food, transportation, or entertainment. Households with relapsed cessation had a significantly higher food away from home budget during the fourth quarter than households that did not quit, despite resuming their spending on tobacco. Further, the proportion of the total food budget in particular among relapsed households shifted toward food away from home purchases.
Lastly, the FMM’s results revealed significant heterogeneity in household spending on alcohol following cessation. While most households reduced alcohol spending during or after tobacco cessation, a subgroup of households (determined by lower pre-quit alcohol spending and possibly younger age, white race, and multi-person membership) were also able to reduce or maintain a reduction in alcohol spending following relapse back to tobacco. The food at home FMM analysis suggested that the effect of tobacco cessation on food at home spending was largely homogeneous, with 96% of the sample captured by a single subpopulation, in which both long-term and recent cessation were associated with a small but significant reduction in food at home spending. Households were more likely to be part of the subpopulation that did not lower their food at home spending after quitting if they had lower pre-quit spending on food at home, were not married, and had experienced a decrease in earners during the study. The food away from home FMM analysis found that long-term and recent cessation were unrelated to fourth quarter spending on food away from home in the two subpopulations identified, but relapsed cessation was associated with higher spending on food away from home in only the smaller of the two subpopulations identified. Households were more likely to be part of the subpopulation that increased food away spending with tobacco relapse if they had higher baseline food away spending, were not married, and had not experienced job loss during the study.
Long-term and recent tobacco cessation were not associated with higher spending in any non-tobacco category. Households with long-term and recent cessation appeared to execute expenditure reductions that enabled or complemented their tobacco cessation and decreased their total expenditures – leaving a prioritization within the budget of food, housing, and health care. Overall, the relapsed household findings suggest one of a struggle between a new health priority (e.g., a quit attempt and reduced relative spending on alcohol) and an increase in eating away from home, with ultimate relapse back to tobacco. Financial strain and/or income constraints, as well as social motivations to eat at home, were the most consistent determinants of the change in a household’s spending on food away from home following cessation. Additional research is needed to understand the mechanisms of these relationships.
Rogers, Erin R., "Tobacco Cessation and the Household Budget: A Longitudinal Analysis of Consumption and Heterogeneity in the United States" (2016). CUNY Academic Works.