Pubali Chakraborty

This paper examines the effects of an alcohol prohibition law in Bihar, India, on intoxicant consumption. We implement a dynamic difference-in-difference estimation strategy using longitudinal data on monthly household expenses, exploiting state-level variation in policy exposure and household-level variation in alcohol use. We document that alcohol- consuming households in Bihar reduced their spending on tobacco products following the ban announcement, indicating complementarity between alcohol and other intoxicants; however, after its strict enforcement, when alcohol was unavailable, these households gradually increased their tobacco consumption. We find reallocation in healthcare spending: urgent medical expenses decrease with increased spending towards positive lifestyle changes.

Description of Project 

The consumption of alcohol and intoxicants, such as cigarettes, bidis, marijuana, and other forms of tobacco products, which are addictive or habit-forming goods, have been documented to be closely related. However, there is considerable debate regarding the substitutability or complementarity be tween them. Decker & Schwartz (2000) find that while the cross-price elasticity of alcohol consumption with respect to cigarette price changes is positive, implying substitutability, the relationship is reversed on investigating changes in cigarette consumption as a result of alcohol price changes. Shrestha (2018) finds that higher cigarette prices result in fewer young adults participating in drinking; however, condi tional on participation, individuals increase their alcohol consumption. Tauchmann et al. (2007) finds complementarity between liquor and tobacco products, whereas there is evidence of substitutability between the consumption of alcohol and marijuana or drug use (DiNardo & Lemieux, 2001; Alley et al., 2020; Reiman, 2009; Chaloupka & Laixuthai, 1997; Miller & Seo, 2018). 

Policies aimed at increasing overall market consumption of households or reducing the consumption of one of these addictive goods may have externality effects on the other goods. Therefore, it is imper ative to understand the household preferences over these goods to correctly quantify the effectiveness of a policy and to avoid unintended consequences, both in the short run as well as over a longer period of time. In April 2016, the State of Bihar implemented the Bihar Excise (Amendment) Act, which prohibited the manufacture, sale, or consumption of alcohol. In this project, we aim to quantify the dynamic effects of this ban on the consumption of intoxicants and other goods in households. 

Data Sources 

For our project, we plan to use data on the monthly income and expenditure of households in India from the Consumer Pyramids Households Survey database developed by the Centre for Monitoring the Indian Economy (CMIE). They report detailed expenditure on different categories of food, clothing, durables, alcohol, intoxicants, health, and education. CMIE conducts nationwide surveys throughout India to collect this data. This survey covers over 160,000 households in all the 640 districts in the 2011 Census. Our study, however, would be restricted to Bihar and the adjoining districts of Uttar Pradesh and Jharkhand, which share borders with Bihar. 

By design, CPHS is a continuous panel, and this feature would make it crucial for our study. The CMIE started reporting their data since 2014, and thus, we can observe the consumption behavior of households both before and after the ban. Surveyed households are selected through a multi-stage stratification process that ensures geographical diversity. Households who are in the survey are paid a visit every four months, and their expenditures and income are recorded via a detailed list for each item. This enables us to study changes in the expenditure of households at different points in time: in response to the announcement of an upcoming ban, immediately following its implementation, and a few years after the ban has been in place. 

Methodology 

We first use a difference-in-difference framework to assess the impact of the ban on households that consumed alcohol and thus were affected by the changing law relative to households that never con sumed liquor prior to the announcement of the ban. We study the resulting changes in the consumption of these goods, when the ban is announced, when the ban is implemented, and the persistence of these changes over the next couple of years. To the best of our knowledge, this dynamic investigation is novel in this literature and aims to provide insights into the effectiveness of a policy both in the short run as well as over a longer period of time. Next, we use variation across states to investigate the effect of the ban in Bihar relative to the neighboring states of Uttar Pradesh and Jharkhand (we restrict our analysis only to the districts that share a border with Bihar), where no such change in law took place. We also use a triple difference approach to compare between Bihar and its neighboring states, the expenditure of households that consumed liquor relative to those that did not, as a result of the prohibition of alcohol.