The effect of tobacco- and alcohol-control policies on household spending patterns in Kenya: An approach using matched difference in differences.

Abstract This paper examines the effect of tobacco and alcohol control policies on tobacco and alcohol consumption patterns and the evolution of crowding-out effects on other household expenditure in Kenya. The current literature on crowding out does not provide a defensible instrumental variable for a system of demand equations. This paper uses Matched Difference in Differences (MDID) as an alternative strategy and data from two nationally representative surveys in Kenya conducted ten years apart (2005/6 and 2015/16). We find that tobacco control policies contributed to a decrease in the proportion of tobacco consuming households between 2005 and 2015. Alcohol-control policies were only effective in reducing the proportion of alcohol consuming households in the bottom quartile of the expenditure distribution. Overall, tobacco consuming households spent less on education, communication, and some food items. Alcohol consuming households also spent less on some food items, but expenditure on transportation was the only non-food item crowded out. Tobacco and alcohol control policies, when they result in reduced consumption of these products, can increase household expenditure on human capital development in the long run.