The 'Smee Report' as a contribution to the tobacco advertising debate

Author: John Luik
Article Published: 1996

The UK Department of Health's consultative document on tobacco advertising (informally known as the 'Smee Report') occupies a curious position with respect to debates on public policy about advertising. While its conclusions have not been accepted by the UK government, they continue to be used, uncritically, elsewhere in the world as part of the 'justification' for bans on tobacco advertising. This article examines the basis of the Smee Report's key conclusions concerning the UK, Canada, Norway, and Finland, and argues that its claims with respect to the effectiveness of advertising bans are untrue.

Almost two-and-a-half years after the release of the Smee Report and despite the fact that the UK government has declined to use the Report's conclusions as the basis of public policy with respect to tobacco advertising, the Smee Report continues to be cited around the world both as a comprehensive analysis of the evidence about tobacco advertising and as a definitive judgement about the necessity for bans of such advertising.

Despite the anti-tobacco lobby's attempts to sell Smee to public policy-makers, for those careful crafters of public policy who care more for basing their policy recommendations on cogent analysis than on doctrinaire rhetoric, the crucial scholarly questions about Smee have remained unanswered: namely, what sort of assessment should we make about the quality of the Report's research and analysis and are the conclusions of the Report a valuable contribution to the public policy discussion about tobacco advertising? We wish to argue that while containing some important insights, the Smee Report, on the whole, is so deeply flawed that its usefulness is as yet another demonstration of how not to deal with the issue of tobacco advertising. In effect, it should not provide the basis for any public policy on tobacco control.

Let us begin with the Report's strengths.

First and most significantly, the Report's own analysis of a possible effect of advertising on tobacco consumption concludes that `advertising does not have a statistically significant effect in any form'. It is, however, regrettable that the author found it necessary not merely to bury this `politically incorrect finding in an appendix, but to explain it away in the Report's conclusions.

Second, the Report correctly notes that the direction of causal relationship between advertising bans and declining tobacco consumption is most likely the opposite to that often assumed. Rather than advertising bans lowering consumption, such bans are introduced once consumption has already begun to fall and social attitudes have changed. This leads, as the Report rightly observes, to 'an association between the two without the controls causing the lower tobacco consumption'. [2]

Third, the Report rejects the simplistic and illogical claim that recognition of advertisements leads to increased consumption by making an important distinction between the necessary and sufficient conditions for advertising to affect consumption. While young people might well recognize tobacco advertisements it does not follow that such recognition leads them to begin smoking: '[A]wareness of advertising is at most a necessary condition for coming under its influence. [3] It is not reliable evidence that advertising increases consumption.' Moreover, the Report again notes that the direction of causality is just as plausibly interpreted as running in the opposite direction: namely, young people 'disposed to smoke are more likely to react positively to tobacco advertising and show greater awareness of it'. [4]

Fourth, the Report concedes that the most important factor in the smoking behaviour of young people is the smoking behaviour of parents and siblings. More crucially, but again unfortunately buried in a technical appendix, the author carried out a detailed statistical analysis of whether tobacco advertising is likely to influence the smoking prevalence among teenagers'. [5] A separate model was developed for UK men and women (with the recognition that the data was limited to age 16 and above), and 'in neither model was advertising a significant factor'. [6]

However, despite these commendable aspects of the Report, there are three weaknesses which are decisive in judging its overall value: weaknesses with respect to the issues of evidence, causality and context.

First, the Report is plagued by problems about the nature, scope and quality of its evidence. The literature on tobacco advertising, tobacco consumption and tobacco advertising bans is enormous. While it is not necessary for a review of this sort to examine all of the evidence, it is important for it to have examined a very broad range of evidence. The Report, however, looked at only a fraction of what is available. Indeed, it is this failure to master the relevant literature that accounts for some of the Report's most significant weaknesses. For example, the International Journal of Advertising devoted an entire issuer to a detailed, scholarly examination of various aspects of the tobacco advertising controversy that are particularly relevant to the Report's inquiry, particularly as they contradict some of the Report's conclusions, yet none of these articles is cited by the Report, nor listed in its bibliography Again, the Report's section about the possible effect of advertising on the initiation of juvenile smoking contains but two references, despite the fact that the worldwide literature of articles and books numbers in the hundreds, if not thousands. Or again, though referring to the Canadian court case dealing with the Canadian ban on tobacco advertising, the Report's authors appear not to have read either the judgement or the court record. This failure is especially significant in that the trial, which lasted over a year and heard 'twenty-eight witnesses . . . the great majority of whom were experts in fields as wide-ranging as advertising and marketing, epidemiology, public health, oncology, cardio-respiratory diseases, perinatal care, psychology, neurology, statistics and econometrics' s is perhaps one of the most comprehensive 'examinations' of all aspects of the advertising debate in the world. Indeed, the trial provided the venue in which virtually all the claims made about advertising and tobacco consumption were subjected to rigorous and objective scrutiny. To take but three examples, had the Report's author looked at the trial evidence he would have found that his conclusions about:

  • (1) the reliability of much of the international data on consumption and advertising restrictions;
  • (2) the credibility of several experts on advertising and tobacco consumption, and
  • (3) the impact of advertising bans in certain key countries, might well have been different.

It is difficult to overestimate the effect which this omission had on the reliability of the Report's findings. One is left with the uncomfortable conclusion that the Report's failure to examine a comprehensive sample of the available evidence is attributable to either ignorance of the range of such evidence or a bias towards certain sorts of evidence.

However, it is not simply what evidence the Report failed to consider; it is also the Report's inability to judge the quality of the evidence which it did consider that counts against its usefulness. Although it notes that the `forms of evidence differ', the Report none the less fails to come to terms with a more fundamental distinction: a distinction not between forms of evidence but between assertions and claims and the evidence that is adduced to support such claims. It is almost as if the Report's author, at times, accepts any published claim as evidence for some assertion. Quite obviously, the relevant task is not to repeat a claim but to determine (1) whether the claim has any evidentiary backing, and (2) what the quality of that backing is.

This failure is particularly distressing as it is manifest in the Report's opening sections on the ways in which advertising works and the impact of advertising on markets in general and on specific markets. In its analysis of the Incentive Structure in the Tobacco Market the Report claims that advertising (1) creates a monopoly power for a brand, and (2) encourages smoking initiation. While these may be interesting claims or assertions, they are regrettably neither indirect nor circumstantial evidence; they are unsupported assertions. Thus, at the very outset of the Report there is not merely a failure to provide compelling evidence but something more fundamental: a failure to appreciate what precisely distinguishes assertion from evidence. Moreover, this is not an isolated instance, for the same problem occurs in the following section on the mechanisms by which advertising might increase consumption.

Where this failing to adequately assess the quality of evidence is most manifest is with respect to the Report's reliance on the cross-national study data of Laugessen in 1991 and the New Zealand Department of Health (1989). Both Laugessen's consumption figures (Laugessen was the scientific editor of Health or Tobacco) and New Zealand's Department of Health Report - Health or Tobacco - were subject to extensive scrutiny during the Canadian tobacco advertising ban trial. The judge's conclusion about the value of such evidence was decisive in that he found that 'it contains serious methodological errors and a lack of scientific rigour which renders it for all intents and purposes devoid of any probative value. It is a report with an obvious point of view and its conclusions reflect that point of view'. [9] Moreover, it is instructive to recall what the Canadian court said of one of the Crown's chief witnesses, Dr Jeffrey Harris, an econometrician and medical doctor, who used Dr Laugessen's consumption figures as the basis of his testimony about the efficacy of tobacco advertising bans.

The witness struck the court as extremely intelligent and very skilled in the manipulation of ideas and statistics. Unfortunately, he did not demonstrate the scientific objectivity that the Court is entitled to expect from an expert witness of his stature . . . . In this instance, the Court is of the opinion that the input data used by Dr. Harris were unreliable and that his methodology led necessarily to the desired result. Here again, the Court entirely agrees with the analysis of the reports and testimony of Dr. Harris made by RJR's counsel in his argument and accords no probative value to those reports or that testimony. [10]

At the very least, the author of the Report should have informed his readers that an important piece of evidence which he cited had been rejected on the only occasion that it had been subjected to independent review. An author more attentive to the quality of his evidence might, given the damning judgement above, have avoided using such evidence.

But without unduly championing the merits of Canadian jurisprudence, it is worth noting how a careful study of the Chabot decision (after Justice Jean-Jude Chabot, presiding judge in the advertising ban case) might have again saved the author from one of his most significant evidentiary errors, namely his conclusion that the `current evidence available on these four countries [Canada, New Zealand, Norway and Finland] indicates a significant effect' [11] that is, a reduction in smoking brought about by an advertising ban. After examining the statistics from Norway and Finland, not only Justice Chabot but also the three Court of Appeal judges conclude that the evidence does not support the conclusion that there is any probability that a ban on advertising will bring about a reduction in tobacco consumption. As for New Zealand, it is worth noting that the consumption figures used by the author to determine the `effect' of the advertising ban were supplied by the New Zealand Department of Health, the same department whose consumption data and conclusions were judged to be of no 'probative value'.

What then, finally, of Canada itself? has its advertising ban produced `a fall in smoking on a scale which cannot reasonably be attributed to other factors'? 12 The answer to this question is no, an answer that the Report's author might well have come to had he been more attentive to the quality of his evidence, in addition to the variables used in his model. The source of the Report's Canadian data is never revealed, though we are assured that they are official statistics plus estimate to account for cigarettes brought into the country either legally or ' y. What is quite likely, however, is that these official statistics and the auf tor's estimate significantly underestimate tobacco consumption in Canada.

For example, it is now conceded that prior to the recent reduction of tobacco taxes, at least 25 per cent of the cigarettes consumed in Ontario and 66 per cent of the cigarettes consumed in Quebec were illegally purchased and thus not counted in 'official statistics'. This unaccounted for consumption in addition to the increases in taxes since 1980 - 500 per cent provincially, 550 per cent federally - are more than enough to discredit the Report's claim of a 1.7 per cent 1989 and 1.1 per cent 1990 decline in consumption attributable to the advertising ban. Indeed, if we examine the prevalence of smoking in Canada, that is, the percentage of adults who smoke, we find that it has declined at substantially the same rate from 1965 until 1994.

More crucially, if we look at Figure 1 which reports the Government of Canada's first cycle smoking prevalence survey, it will be noted that the prevalence of smoking in Canada is 31 per cent of the adult population aged over 15, unchanged since 1991. And as the authors of the survey note, 'there has been no real change in prevalence since 1988', a year before the tobacco advertising ban took effect.13 Finally, and perhaps most significantly, prevalence for those aged 15-19 actually increased by 'four percentage points since 1991', 14 despite the absence of any tobacco advertising.

In essence, the Report is dramatically wrong about the effect of the Canadian advertising ban. Not only has the ban not reduced overall prevalence, but it has failed to reduce teenage prevalence. Indeed, if one were to postulate any relationships based on the Canadian experience it would be that there is an inverse relationship between the advertising ban and juvenile prevalence. These results are more than simply inconvenient for those who advocate advertising bans, as they suggest that advertising bans not only do not reduce overall prevalence, but they also fail to reduce prevalence among the young. And without these two arguments it is difficult to discover what policy justifications could be brought forward for such bans.

This study is unique in that it provided an opportunity to measure the effect of alcohol advertising on a population that had not previously been exposed to such advertising. On the basis of the single distribution theory the authors hypothesized that allowing alcohol advertising would lead to an increase in sales of alcoholic beverages. The results of the study did not, however, support the author's hypothesis.

The change in legislation regarding alcohol advertising produced neither an abrupt permanent nor a gradual permanent effect on the pattern of the total volume of sales in Saskatchewan . . . . In spite of what appears to be a substitution effect between beer and spirits, overall sales of alcohol did not increase between 1981 and 1987 in Saskatchewan . . . . That is, there was no increase in the overall level of sales in Saskatchewan following a change in advertising policy 21 Indeed, this is precisely the 'causal' effect of advertising on consumption that experts on advertising predict.

Sixth, the Report's conclusion about advertising bans reducing consumption in Canada, Norway, Finland and New Zealand is simply untrue, as the findings of the Canadian advertising ban case discussed earlier clearly demonstrate. In Norway, for instance, the consumption of tobacco today is virtually unchanged from its 1975 level (the year the advertising ban was introduced), while in Finland the incidence of smoking among young people has actually increased since the ban. All this is in sharp contrast to countries like the United Kingdom and the Netherlands which, even while permitting advertising, have experienced the greatest declines in consumption.

The third decisive weakness of the Report is its contextual weakness, its failure to place the entire question of controlling or banning tobacco advertising in its larger context within democratic societies. As we have argued elsewhere in this journal there is a significant dimension of this controversy that goes beyond mere econometric studies of the alleged connections between advertising and consumption. The proponents of advertising bans must invariably recommend restrictions on values like freedom of expression that are central to democratic societies, and these restrictions must in turn be carefully examined for their effects on the character of a free society Thus whatever the empirical evidence about advertising, consumption and advertising bans, this evidence cannot constitute a decisive answer, at least within a democratic society, as to the legitimacy of banning tobacco advertisements. Such evidence will at best be only one consideration in the formulation of public policy on tobacco advertising. For instance, given the enormous powers that such bans convey to government, and given the serious setbacks which such bans bring to the democratic interests of all citizens and the in significant results such bans bring, it might be suggested that the costs of a ban far outweigh the benefits to democratic society.

All this suggests that in examining the question of advertising bans and failing to place the question within its larger democratic value context, the Report appears to claim that the only evidentiary considerations worthy of discussion are econometric, indirect and circumstantial. This is conceptually on a par with a study charged with comparing the criminal conviction rates of totalitarian regimes with police forces that routinely employed torture with those of free societies with police operating under the rule of law, that failed to consider the question of human rights as a legitimate part of its remit.

What then, it might be asked, is the value of the Report for formulating sounddemocratic public policy on tobacco advertising? Democratic public policy should be founded on a combination of comprehensive, reliable and cogent evidence and compelling argument that is sensitive to the core values of a democratic society To assist in the formulation of such a policy the Report should have provided such evidence and argument on the four central questions that shape the tobacco advertising debate:

  • (1) Does advertising encourage young people to start smoking?
  • (2) Does advertising have an effect on tobacco consumption?
  • (3) Do advertising bans reduce consumption? And
  • (4) Do such bans pose significant threats to a democratic society?

Instead, we are presented with a claim that advertising does have an effect on tobacco consumption which is then contradicted by the Report's own econometric work; we have a series of suggestions based on indirect and circumstantial evidence about the possible role of advertising on initiating smoking among the young, suggestions that are directly contradicted by the extensive literature in this area that the Report did not examine and also directly contradicted by its own statistical work; we have a claim that advertising bans in four countries have reduced consumption based on data and evidence that were either rejected in court as having no probative value or that are so flawed as to be unbelievable; and we have not a single mention of the threats such advertising bans pose to free societies. Viewed from this perspective, the Smee Report is far more of a liability than an asset in an informed public policy discussion about tobacco advertising.



1 Effect Of Tobacco Advertising On Tobacco Consumption: A Discussion Document Reviewing The Evidence, Clive Smee, Chief Economic Adviser; Economics and Operational Research Division, UK Department of Health, London,1992; and p.140 in Do Tobacco Advertising Bans Really Work? A Review of the Evidence, J.C. Luik (ed), Ontario: The Niagara Institute,1994.

2 Ibid., p.136.

3 Ibid.

4 Ibid., p.128.

5 Ibid., p.151.

6 Ibid., p.152.

7 International Journal of Advertising, Special Report (1990).

8 Justice J.J. Chabot, RJR Macdonald Inc. v. The Attorney General of Canada and imperial Tobacco Ltd v. The Attorney General of Canada (1991), p. 75 in Freedom of Expression: The Case Against Tobacco Advertising. Bans, J.C. Luik (ed.), Ontario: Graymatters,1991.

9 Ibid., p.150.

10 Ibid.

11 Effect of Tobacco Advertising on Tobacco Consumption: A Discussion Paper Reviewing the Evidence, p. 137.

12 Ibid.

13 Survey on Smoking in Canada, Ottawa: Health Canada,1994, p. 1.

14 Ibid.

15 Effect of Tobacco Advertising on Tobacco Consumption: A Discussion Paper' Reviewing the Evidence, p. 137.

16 Ibid.

17 Ibid., p.144.

18 Ibid., p.141.

19 Smart, R.G. (1988) Does alcohol advertising affect overall consumption? A review of empirical studies. Journal of Studies on Alcohol, 49, 314-323.

20 Makowsky C.R. & Whitehead, PC. (1991) Advertising and alcohol sales: a legal impact study Journal of Studies on Alcohol, 52, 555-567.

21 Ibid., pp. 563-564.

22 See J.C. Luik, Tobacco advertising bans and the dark face of government paternalism. International Journal of Advertising,12, 30 324.

© Advertising Association 1996. Published by Blackwell Publishers,108 Cowley Road, Oxford OX4 1JF UK and 238 Main Street, Cambridge, MA 02142, USA.

FORCES is supported solely by the efforts of the readers. Please become a member or donate what you can.

Contact Info
Forces Contacts
Media Contacts
Ian DunbarIan Dunbar, United Kingdom

Latest Article »»  

Bill Brown, USA

Latest Article »»  

Michael J. McFadden, USA

Latest Article »»  

Joe Jackson, United Kingdom

Latest Article »»  

Virginia Day, USA

Latest Article »»  

Robert Prasker, USA

Latest Article »»
Contact Robert Prasker »»

John Dunn, MD, United States

Latest Article »»
Contact John Dunn, MD »»

Andrew Phillips, Canada

Latest Article »»
Contact Andrew Phillips »»

Pat Nurse, United Kingdom

Latest Article »»
Contact Pat Nurse »»

Elio F. Gagliano, MD, Italy

Latest Article »»  

Edmund Contoski, USA

Latest Article »»
Contact Edmund Contoski »»

John Luik, Canada

Latest Article »»  

Norman Kjono, USA

Latest Article »»
Contact Norman Kjono »»

Gian Turci, Italy

Latest Article »»  

Søren Højbjerg, Denmark

Latest Article »»
Contact Søren Højbjerg »»