The Effect of Policy and Social Interaction Variables on Youth Smoking in Low and Middle Income Countries
thesisposted on 2012-12-13, 00:00 authored by Silda Nikaj
This dissertation investigates the effect of policy and social interaction variables on youth smoking among countries which conduct the Global Youth Tobacco Survey (GYTS). This is the first study to use GYTS data to investigate the effect of peer smoking on individual smoking. In the second part of the dissertation, I estimate the effect of cigarette prices on youth cigarette smoking among a group of less developed countries from the GYTS. This first chapter investigates the effect of peer smoking on individual smoking among middle-school students in 109 low and middle income countries. I address endogeneity in school choice and unobserved school-level characteristics through the use of school fixed-effects. Identification arises by comparing students in different grades within the same school. I instrument in order to address the simultaneity in individual and peer behavior. I correct for measurement error in the peer variable through the use of instruments and address the upward bias in IV estimates by modeling the measurement error through the use of GMM. The evidence suggests that the measurement error is in fact mean-reverting. On average, increasing the share of one’s peers who smoke by 10 percent increases the probability that an individual will smoke by 2-4.5% percentage points. Peer effects vary when the analysis is conducted among five geographic sub-regions and at the country level. This second chapter investigates the effect of cigarette prices on smoking prevalence and consumption among youth in 38 Global Youth Tobacco Survey (GYTS) countries. I employ two-way fixed effects to address country level unobservables which may be correlated with smoking use among youth over time. I control for local level variables that affect smoking among youth. The estimated total price elasticity is -1.5, -0.6 through participation and -0.9 through conditional demand. This estimate suggests that a 10 percent increase in price reduces cigarette use among youth by 15 percent among high income and less developed countries. Restricting the analysis to only less developed countries produces a total price-elasticity of cigarette use of -2.2, -0.6 through participation and -1.6 through consumption.