Dissertations, Theses, and Capstone Projects

Date of Degree


Document Type


Degree Name





David Jaeger

Committee Members

Michael Grossman

Partha Deb

Subject Categories

Econometrics | Health Economics


alcoholism, health behavior, substance abuse, treatment effect models, Scotland


This dissertation consists of two chapters, and it evaluates the short-term impact of Scotland’s Minimum Unit Pricing (MUP) policy on a variety of outcome measures. Excessive alcohol consumption has long been considered one of Scotland’s greatest public health challenges, and in an attempt to change this harmful relationship with alcohol, the Scottish Government proposed an alcohol strategy comprising of over 40 measures, a central part of which was the introduction of a statutory minimum price of 50 pence per unit (10ml/8gr) of alcohol known as Minimum Unit Pricing (MUP). This policy aims to reduce alcohol consumption, and in turn, alcohol associated health and social harms, by targeting alcohol products that are cheap relative to their alcohol content. Though variations of minimum pricing had been successfully implemented in other regions before, Scotland is the first country to introduce a minimum price per unit of alcohol – a measure focusing solely on alcohol content, and not the type of product. Chapter 1: In May 2018, Scotland became the first country in the world to enforce a Minimum Unit Pricing (MUP) of 50 pence ($0.63) on alcoholic beverages as a measure to control alcohol consumption and alcohol-related harms. This paper presents early estimates of the impact of MUP on three outcomes: average price per unit (8 gr/10 ml) of alcohol, liters of alcohol sold per adult, and liters of alcohol sold per adult drinker, in three different settings: off-premise, on-premise and both combined. Using yearly alcohol price and sales data from 2011 to 2019, I find that the average price per unit of alcohol has significantly risen post-MUP across all beverage categories (beer, spirits, wine, cider and alcohol-all types), whereas liters of alcohol sold per adult and per adult drinker have significantly reduced for beer, spirits, cider and alcohol-all types. For all outcomes, the impact of MUP on off-premise alcohol prices and sales is more pronounced than under the combined case, while no significant effects are found on-premise. These results are robust to model specification, and to the addition of various control variables. Falsification tests changing the timing of the policy were performed and no significant results were found. Chapter 2: In this paper, I explore individual level data from the Scottish Health Survey (SHeS) and the National Survey for Wales (NSW) for the period 2016-2018, and use a Difference-in-differences approach with country and year fixed effects, to estimate the immediate impact of MUP on four self-reported measures of alcohol consumption separately for moderate, hazardous and harmful drinkers across different population subgroups. Findings suggest that MUP has been effective in reducing the prevalence of drinking in the last 7 days for both moderate and harmful drinkers, by 10 and 8.4 percent respectively – an indicator of policy success at the extensive margin; and in reducing the number of drinking days in the last 7 days, and the number of alcohol units, as well as the prevalence of exceeding the daily limit on the heaviest drinking day (HDD) in the last 7 days for harmful drinkers by 13, 26 and 12 percent respectively – all three indicators of policy effect at the intensive margin. Given that MUP is a targeted intervention aimed at reducing alcohol consumption and related harms of those drinking in excessive levels particularly, its early significant impact on harmful drinkers is promising for the future of MUP. Estimates from subgroup analysis generally indicate anticipated differential MUP effects with respondents who are males and aged 16-44 and 65 plus being affected the most. Analysis by income quintile refutes the assertion that MUP is a regressive policy affecting moderate drinkers with lower incomes more, whereas analysis by the Index of Multiple Deprivation quintile shows that male moderate drinkers whose income falls in the first quintile will be affected the most. Overall, no significant policy effect is observed among hazardous drinkers and this could be somewhat concerning. The main results are robust to the exclusion of sampling weights, explanatory variables and fixed effects, and estimation under different models such as logit and Probit.