Does drinking affect long-term sickness absence? A sample selection approach correcting for employment and accounting for drinking history

This article studies the effect of alcohol consumption on the probability of long-term sickness-related absenteeism for women. Using Swedish matched survey and register data, we apply sample selection models to correct for non-random sampling into paid employment. There are three main findings of the study. First, diverging from the most prevalent consumption group (long-term light drinkers) is associated with an increased probability of long-term sickness, ranging from 10% for long-term heavy drinkers to 18% for former drinkers. Second, controlling for former consumption errors (especially former drinker and former abstainer errors) and sample selection into employment are important for unbiased, consistent estimations. Third, by predicting the effect of changes in consumption on long-term sickness-related absence, we find that alcohol only explains a small part of the overall picture of long-term sickness-related absenteeism. Notwithstanding this fact, long-term sickness-related absenteeism due to alcohol adds up to substantial productivity loss for society. Our conclusion is that the commonly found U-shaped relationship between current alcohol consumption and labour market outcomes remains for women, after controlling for past consumption and selection effects. A change in consumption level increases probability of long-term sickness-related absence, compared to individuals with constant consumption levels.