The Demand for Alcohol, Tobacco and Marijuana: International Evidence ‐ by Saroja Selvanathan and Eliyathamby A. Selvanathan
Xueyan Zhao
DOI: https://doi.org/10.1111/j.1475-4932.2008.00518.x
2008-01-01
Economic Record
Abstract:The Demand for Alcohol, Tobacco and Marijuana: International Evidence , by Saroja Selvanathan and Eliyathamby A. Selvanathan ( Ashgate Publishing , Burlington, VT , 2005 ), pp. 384 . * The authors do not suggest a suitable audience for the book from the outset. After reading the book, I conclude that two groups of readers would benefit the most from the book. The book could be a handy reference for policy-makers and researchers needing quick access to cross-country comparisons of key statistics and demand elasticities on food and popular drug consumption. It assembles for an impressive list of countries such figures as budget shares of food (51 percent for the Philippines and 8 per cent for the USA), per capita consumption of alcohol and tobacco, and consumption and price trends for food and drugs over time, presented with numerous tables and graphs. It has a broad brush list of alcohol and tobacco policies across countries. It also provides estimates for income and price elasticities for food and drugs for many countries. Another group of readers who would find the book useful are those who want to estimate a demand system model following Theil (1965), Barnett (1979), and Theil and Clements (1987) differential approach. The book offers technical details of the model and illustrates estimation and tests with applications. However, this is not a book that covers thoroughly the current international debate on drug policies, the consumer theory of addictive goods, or the large empirical literature on drug consumption for various countries. The book consists of three parts. Chapters 2–5 present detailed statistics on the consumption of food, soft drinks, tobacco, alcohol and marijuana for over 40 countries, compiled from various government and international organisational sources, such as the Australian Institute of Health and Welfare (AIHW), the World Health Organisation (WHO), and the European School Survey Project on Alcohol and Drugs (ESPAD). Countries are ranked according to their budget shares for food and drugs, per capita consumption of beer, wine and spirits, proportions of daily smokers for high school students, and so on, although only limited information is presented on marijuana. More detailed information is also provided for 4 countries: Australia, New Zealand, the UK and the USA. I expect that politicians and researchers would find these cross-country comparison statistics handy for motivating their speeches or research. However, the book does not go very far beyond the presented statistics in discussing cross-country differences. The second part of the book, Chapters 6–10, presents the theory and applications of the Theil–Clements differential approach to demand system analysis. Various versions of the Rotterdam model are presented and testing procedures are detailed for economic constraints. Rotterdam models in absolute price format are estimated for each country using aggregated time-series data, and homogeneity and symmetry assumptions are tested. A very restrictive type of additive utility for ‘preference independence’ (PI) is also tested. These hypotheses are accepted for most countries, so final estimates for price and income elasticities are presented for each country under PI assumption. This is not the first time I have tried to digest the theoretical part of the Theil–Clements demand system approach, and I am finding this material challenging. Unlike the other commonly used demand systems, such as the translog or the almost ideal demand system, where a particular functional form is specified for the indirect utility or cost function, the differential demand system approach starts with a utility function and the derived demand system equations in generic functional form. The demand equations are then totally differentiated, and finite discrete time approximation with time-series data is used in applications to approximate the infinitesimal changes. Unlike working with a particular functional form for the objective function where it is clear which are coefficients (thus are constant) and which are variables from the outset, in the differential approach, new coefficients, such as marginal shares and income flexibility, are defined along the way. I cannot help but think that there needs to be a proof for the recoverability of a utility function with all these measures defined as constants. The estimation and tests of such models are relatively easy to implement with time-series data, and I expect one can do so with the standard econometric software for system equations as well as the specialised packages mentioned in the book. For the empirical application on the authors’ part, I feel somewhat uneasy when they state on page p. 152: ‘In contrast to the other approaches to generating demand equations, the differential approach requires no algebraic specification of the utility function. Consequently, the coefficients of the demand equations (5.10) need not be constant; they may be functions of income and prices’ (there is a similar statement on p. 135). I wonder if they really meant that the equation is a nonlinear function of the exogenous variables pjs, as the coefficients νij and θi are constants. In addition, while in one version of the conditional demand equation in (2.16) (p. 166) νij is constant, in another version in equation (3.1) (p. 167), ν′ij = νij/Wgt is defined as constant (and estimated as constant) when Wgts are endogenous variables or data varying by t. It is interesting to note that the book never actually estimates a demand system with all five goods (food, alcohol, tobacco, soft drinks and marijuana) included. In fact, the title of the book is misleading, as the book does not estimate a demand system model for alcohol, tobacco and marijuana either (and the book covers food and soft drinks as well). The empirical part of the book estimates three demand system models: one 5-good unconditional demand system (for food, alcohol, tobacco, soft drinks, and ‘other’), one 3-good conditional demand model (for beer, wine and spirits), and one 4-good conditional model (for beer, wine, spirits and marijuana, with part of the coefficients, marginal shares, specified a priori). Although the theory for two-step budget allocation, block independence and conditional demand are discussed in Chapters 6 and 7, it is a pity that the book never estimates a two-stage budgeting model estimating the two demand systems simultaneously and endogenising group allocation, DQgt. For example, in the first step, total budget is allocated to total alcohol and other groups, and then in the second step total alcohol budget is allocated to beer, wine and spirits (see Edgerton (1991) for applications of multi-stage budgeting demand system). It is certainly not ideal to estimate an unconditional demand system with budget shares of 0.01 (soft drinks) and 0.80 (‘other’) at the same time, or a conditional demand system alone for beer, wine and spirits assuming total alcohol budget is exogenous. I wonder how the estimated elasticities would be different if a different allocation structure were assumed for the national budget. I am also interested to see that the four rather closely related goods (alcohol, soft drinks and tobacco especially) are given an additive utility PI structure with no specific cross-price effects. The final part of the book, Chapters 11 and 12, is dedicated to a discussion on alcohol and tobacco control policies for selective countries. As the authors point out, the countries are selected according to availability of information on such policies. In Chapter 11, for alcohol, almost 30 countries are discussed, one by one. The discussion includes details regarding liver cirrhosis deaths and alcohol-related road accidents over time, alcohol sale and advertising restrictions, drink driving regulations, taxation, and government preventative programs. For tobacco in Chapter 12, smoking prevalence is compared for selective countries. Control measures on both demand and supply are then discussed, including tobacco advertising, health warning labels, restrictions on smoking in public places, taxation and educational programs. Again, this part of the book would be a useful reference and entry point to other sources for policy-makers and researchers who are interested in different policy measures across the world for the two most popular recreational drugs. The book does not provide much discussion or analysis beyond brief summary of policy measures (which is covered in just two chapters). With a quick Google search, one can find other books and survey papers that provide more details on international comparisons of particular aspects of tobacco and alcohol policies. Alcohol, tobacco and marijuana are addictive goods, and there is economic literature on the rational addiction theory of consumers of addictive goods. Much has also been discussed in terms of the relationship of different drugs and the so called ‘gateway’ effect postulating that consumption of softer drugs leads to the use of harder drugs. While the book uses macro time-series data and a traditional demand system approach, there is also much empirical literature using micro survey data that considers socio-demographic as well as economic factors relating to the consumption of these popular drugs. The book discusses little of any of these. I also find myself becoming frustrated while reading because I cannot find certain things. The book relies on data from many sources, but it is often difficult to find references to data sources for a particular piece of analysis. Sources are not presented in the footnotes of tables and graphs. Also, in at least five cases, I am unable to find references in the bibliography that have been referred to in the text. There are also obvious typos in the text and one table does not even have a title. One wonders whether the publisher has proofread the manuscript. The authors are experts in using the Theil–Clements differential demand systems for the analysis of international consumption. They have an impressive list of recent books and papers on related topics, including between them at least four books on international consumption and five published papers on alcohol consumption (including one on beer, wine and spirits published in this journal in 2005). The main contribution of the book is the estimated price and income elasticities for the demand of food and drugs for many countries. These economic coefficients are often needed by researchers for policy analysis and there is in general a scarcity of such econometric work. Two pieces of work in the guest chapter by Clements and Daryal on marijuana are also significant. They provide an estimate of the size of the marijuana market in Australia, and they provide some new evidence on the potential effect of marijuana legalisation on consumption for the young population using a student survey at their own university. Readers might not find all chapters in this the book useful, but many will find particular chapters of the book good reference to have on the shelves. I wish to thank Keith McLaren for helpful discussions regarding demand systems.