Paper for Combined APSAD and Cutting Edge Addiction Conference 2007 (Aotea Centre, Auckland, New Zealand, 4-7 November 2007). This reports on a contract of SHORE (Centre for Social and Health Outcomes Research and Evaluation) at Massey University.
One area of addiction is gambling. Over the last year SHORE, the Centre for Social and Health Outcomes Research and Evaluation, with its considerable experience in alcohol studies, has been working on a Ministry of Health funded project to measure the socioeconomic impacts of gambling. The work is in progress, so today I am going to report on some of the conceptual problems we have faced and Kay Hammond, also of SHORE, will talk about the survey.
I am assuming that the WHO-published International Guidelines for Estimating the Costs of Substance Abuse is reasonably well known. I was one of the members of the panel which prepared it. Two of my Australian panel colleagues, David Collins and Helen Lapsley, extended the analysis to the economic issues of the social costs and benefits of gambling, although there are some issues to be resolved and – of course – still data to be collected.
At the heart of a social cost study is a comparison between the actual situation that exists, and the counterfactual scenario which is typically some ideal situation. This arises because economists always define a cost (and benefit) as an opportunity cost (or benefit), that is, the difference between two scenarios.
But what is the right counterfactual scenario: one in which there is no gambling whatsoever, or one in which there is no addictive gambling (and how would we define that)?
Having got the two scenarios – the ‘actual’ and the ‘counterfactual’ – we compare them and then value the differences. The general rules for valuation, derived from economic theory, have been known for almost fifty years, although sometimes they are applied to a new situation. This is the case whether we are investigating substance abuse or gambling. However there are some issues particular to gambling which also throw light upon substance abuse.
First, mortality is not important in gambling. Its human costs are those of the misery – the poorer quality of life – that it causes to the living. One of the purposes of the SHORE survey is to collect data to enable us to assess these morbidity costs suffered by the addicted gambler and her or his friends and families. This is going to be pioneering – highly innovative – research and if we can get it right we will be able to apply the approach to the living who suffer from substance abuse.
A second problem, which is especially important in gambling, is that of the income redistribution it causes. The standard economic framework treats such transfers as unimportant. There is a good reason for doing this, but it is not compelling for a gambling study where transfers are so important. There are some sophisticated ways of taking income redistribution into account – my favourite is the Atkinson-Stiglitz index – but we dont have the data. Instead we will follow the International Guidelines recommendation and do a sectoral breakdown. It wont capture all the redistribution but it will provide a framework to discuss the issue. Both Collins-Lapsley and I do sectoral breakdowns in our substance abuse studies, but this study may use more sectors and do so with greater precision.
A third issue, not covered by the second edition of the International Guidelines, is how to value the outlays of the addicted gambler (or substance abuser). Since the guidelines were promulgated, economists’ understanding of the issue has progressed.
Economists know how to treat the consumption of rational consumers, but have been unsure how deal with addicts, who may not be rational. Recently, drawing on psychological theory, economists have found a rigorous formulation for addiction. In economists’ jargon addicts show ‘time inconsistent behaviour’. Such behaviour occurs when a person plans to restrict their expenditure on something, changes their mind when the opportunity to spend occurs even though there is no fundamental change in the situation, and regrets the spending afterwards, again despite no fundamental change in the situation.
This may seem obvious, but when the notion is mathematised, it proves to be a powerful way to adapt the standard model of rationality. It suggests that consumption for which there are subsequent regrets should be valued at less than rational consumption. The liquor which causes the hangover you do not want is not as valuable as a more restrained indulgence.
Collins and Lapsley have used an ad hoc adjustment in the past for this ‘irrationality’, deducting some of the consumption from consumer value. It turns out that their adjustment is close to the theoretical ideal. I followed them. We will apply their approach with more confidence in the future.
All I have been able to do in my session is to alert you to some of the theoretical issues we are challenged with. In order to apply the theory we need some data. Over to Kay.