AEA RCT Registry

Fields Changed

Registration

Field	Before	After
Abstract	The purpose is to understand how smokers respond to cigarette taxes. Specifically, we study how smokers reallocate their budgets when cigarettes become more expensive, a subject about which little is known. This is especially surprising because the literature has found little effect of cigarette taxes on smoking. That is, given a household budget constraint, when cigarette taxes increase, something must change if smokers are to continue smoking at the same intensity. In addition to an analysis of consumer expenditure survey (CEX) data, we plan to conduct a randomized survey experiment via the survey research platform Prolific. The survey will be of 1,200 current smokers. We plan to ask baseline questions on smoking intensity and shopping behavior, as well as consumption of non-combustible tobacco products. We also ask about baseline weekly expenditures on cigarettes and other discretionary items. We then ask respondents to forecast their smoking and discretionary spending over the next 12 months. Some respondents are randomly assigned a hypothetical cigarette tax increase that randomly varies between 10% and 100%. The questions are framed relative to each respondents revealed weekly cigarette expenditures. We have two goals. First, we seek evidence that smokers do adjust on some dimension when taxes increase. Our survey captures potential changes in expenditures in a wide variety of discretionary categories, as well as labor supply. Second, we use the randomized survey experiment to better understand the dimensions on which smokers adjust consumption when taxes increase. This informs our analysis of CEX data. In addition to closed-ended questions on potential responses to cigarette taxes, we also ask an open-ended question on potential responses. We will use AI to categorized these responses and evaluate how they differ by treatment arm.	The purpose is to understand how smokers respond to cigarette taxes. Specifically, we study how smokers reallocate their budgets when cigarettes become more expensive, a subject about which little is known. This is especially surprising because the literature has found little effect of cigarette taxes on smoking. That is, given a household budget constraint, when cigarette taxes increase, something must change if smokers are to continue smoking at the same intensity. In addition to an analysis of consumer expenditure survey (CEX) data, we plan to conduct a randomized survey experiment via the survey research platform Prolific. The survey will be of 2,200 current smokers. We plan to ask baseline questions on smoking intensity and shopping behavior, as well as consumption of non-combustible tobacco products. We also ask about baseline weekly expenditures on cigarettes and other discretionary items. We then ask respondents to forecast their smoking and discretionary spending over the next 12 months. Some respondents are randomly assigned a hypothetical cigarette tax increase of 100%. The questions are framed relative to each respondents revealed weekly cigarette expenditures. We have two goals. First, we seek evidence that smokers do adjust on some dimension when taxes increase. Our survey captures potential changes in expenditures in a wide variety of discretionary categories, as well as labor supply. Second, we use the randomized survey experiment to better understand the dimensions on which smokers adjust consumption when taxes increase. This informs our analysis of CEX data. In addition to closed-ended questions on potential responses to cigarette taxes, we also ask an open-ended question on potential responses. We will use AI to categorized these responses and evaluate how they differ by treatment arm.
Trial Start Date	January 07, 2025	January 22, 2025
Trial End Date	January 09, 2025	January 23, 2025
Last Published	January 02, 2025 09:48 AM	January 17, 2025 03:31 PM
Intervention (Public)	All respondents will be asked to forecast their smoking and other discretionary expenditures over the next 12 months. Half the respondents will be asked these questions in the context of a hypothetical cigarette tax increase. Within those facing a cigarette tax increase, there will be three randomly sized increases of 10%, 50%, and 100%.	All respondents will be asked to forecast their smoking and other discretionary expenditures over the next 12 months. Half the respondents will be asked these questions in the context of a hypothetical cigarette tax increase of 100%.
Intervention Start Date	January 08, 2025	January 22, 2025
Intervention End Date	January 09, 2025	January 23, 2025
Experimental Design (Public)	We will conduct a simple experiment in which half of our sample will be asked to reveal their smoking and expenditure goals over the following 12 months. These control respondents will be asked about their hopes and expectations regarding their smoking behavior and other spending outcomes. The treatment individuals will be presented with a hypothetical cigarette tax increase that varies (randomly) between 10%, 50%, and 100%, and smoking and expenditures forecasts will be asked referencing the hypothetical tax increase.	We will conduct a simple experiment in which half of our sample will be asked to reveal their smoking and expenditure goals over the following 12 months. These control respondents will be asked about their hopes and expectations regarding their smoking behavior and other spending outcomes. The treatment individuals will be presented with a hypothetical cigarette tax increase of 100%, and smoking and expenditures forecasts will be asked referencing the hypothetical tax increase.
Planned Number of Clusters	1,200 individuals	2,200 individuals
Planned Number of Observations	1,200 individuals	2,200 individuals
Sample size (or number of clusters) by treatment arms	600 control, 200 treatment 1 (10% increase), 200 treatment 2 (50%) increase, 200 treatment 3 (100%) increase.	1,200 control, 600 treatment (100% increase).