From Happy Hour to Rush Hour: Effects of At-Risk Drinking on Labor Market Outcomes Among Mid-Career Men and Women.
详细信息
- 作者：Barnes ; Andrew James.
- 学历：Doctor
- 年：2011
- 导师：Ettner, Susan L.,eadvisor
- 毕业院校：University of California
- ISBN：9781267144386
- CBH：3493496
- Country：USA
- 语种：English
- FileSize：5739994
- Pages：223
文摘
This dissertation adds to the alcohol and labor literature by investigating both the potential mechanisms outlined above along with their resulting policy implications. This work also attempts to address gaps in our knowledge of the relationship between at-risk and labor market outcomes, by focusing on a U.S. representative sample of mid-career men and women, adding controls for potentially important confounders e.g. time preference and risk aversion) not addressed in past work, and testing the sensitivity of the association of at-risk drinking with labor market outcomes to endogeneity. In addition, this dissertation also defines at-risk drinking according the NIAAAs Clinicians Guide to improve translation of study findings for policymakers and clinicians. The labor market outcomes examined include wage rates, work hours, total earnings, occupational attributes i.e., physical exposure, job autonomy, and social engagement), receipt of a variety of fringe benefits, and total hourly compensation i.e. wage rate plus the hourly value of the benefits received). For the sample, data from 5,123 traditionally employed men and women ages 41--49 from the 2006 NLSY cohort were used. Missing data were multiply imputed. With the exception of fringe benefit outcomes where probit models were used, ordinary least squares were used to test the association of at-risk drinking with labor market outcomes while holding constant proxy measures of human capital, perceived opportunity cost of time, other financial resources, and local labor market demand. All regression models were stratified by gender to acknowledge the differential effects of the determinants of labor market outcomes between the sexes. Wage rates, work hours, total earnings, and total hourly compensation outcomes were log transformed to account for their right-skewed distribution and then retransformed using Duans nonparametric heteroscedastic smear factor. To test the sensitivity of the single equation estimates to endogeneity bias i.e. reverse causality and omitted variables bias), instrumental variables probit and two-stage least squares models were specified using religion growing up and having an alcoholic or problem drinking relative as instruments. Sensitivity of the estimates to selection into employment and to the use of sample weights was also conducted. This dissertation found no significant adjusted differences between at-risk and not at-risk drinkers in the wage rate or total earnings for either men or women. For women only, there was weak evidence that at-risk drinkers worked about one hour per week less p＝0.053). After adjusting for confounding factors, single equation models revealed that at-risk drinking was positively associated with employment in occupations with exposure to physical risks 0.016, p<；0.001) among men and negatively associated with employment in socially engaging occupations among both men -0.05, p<；0.001) and women -0.08, p<；0.001)； all scores were on a 5 point scale. For women, at-risk drinking, was associated with a 5-10 percentage point reduction depending on the outcome) in receiving medical, dental, and life insurance, retirement, maternity/paternity, training, child care, and paid time off benefits after adjustment p<；0.05 in all cases). At-risk drinking men were 5 percentage points less likely p<；0.01) to receive maternity/paternity benefits than not at-risk drinkers. No significant adjusted differences were found for total hourly compensation between at-risk and not at-risk drinkers among men or women. However, due to data limitations, the total hourly compensation measure captured valuations for only four medical and life insurance, retirement and paid time off benefits) of the ten possible fringe benefits employees could receive. Evidence from this dissertation suggests the at-risk drinkers do not receive earnings premia nor suffer earnings penalties. However, for mid-career women, at-risk drinking appears to be associated with non-earnings compensation penalties. The non-earnings compensation penalties found in mid-career, if they persist, suggest at-risk drinking women approaching retirement may be less likely to have had health insurance for a number of years, and have less retirement savings. In this regard, at-risk drinkers may, all other things being equal, place more strain on government health care programs for older adults, as a result of potentially decades of poor access to health care, and rely more heavily on social security as their main source of income. Abstract shortened by UMI.)