Science for profit:

Two prominent professors at the service of the tobacco industry

The UZH study was led by two researchers, Michael Wolf, professor of econometrics and applied statistics at the Department of Economics at UZH and Ashok Kaul, professor of economics at Saarland University (Germany). PMI proclaimed that prominent professors at prestigious universities had demonstrated the ineffectiveness of plain packaging.¹³ The results of the study were widely disseminated by PMI and other tobacco companies.

These two professors, as well as the two other members of the research team, worked for the Institute for Policy Evaluation (IPE), a consultancy based in Germany, of which Kaul was the founder and research director. According to a document which remained confidential until May 2023,¹⁴ this institute was the main contributor to the research project: it received two thirds of the budget.

Why did the role of this institute remain undisclosed? IPE played a considerable part in the legal proceedings against Australia, during which various actors claimed that this country’s introduction of plain packaging was in breach of World Trade Organization (WTO) agreements. IPE was, in particular, engaged by one of the complainantsd to produce reports supporting its complaint, also written by Kaul and Wolf. It turned out that, in this action, the complainant had received financial support from PMI.^15-17 It may thus be assumed that IPE had been engaged, directly or indirectly, by PMI to provide the necessary expertise.

A conflict of interests covered up by the Rectorate

In 2015, Pascal Diethelm called on Professor Michael Hengartner, Rector of UZH, to retract the two articles published on the institution’s website, on the grounds that they were problematic in several respects.^18-20 However, UZH failed to recognise its responsibilities. It merely commissioned an external expert to evaluate the articles by the two professors, without providing access to the underlying documents and excluding questions of (research) ethics from the expert review.

Following the publication of the external expert report,²¹ UZH stated: “For the University of Zurich, the accusations that the working papers^e are flawed have thus been refuted. Based on the expert’s recommendation, the university management has decided to leave the working papers on the website of the Department of Economics[...]”^{f, 22}

Suspecting interference by PMI in the research process, Pascal Diethelm then requested access to the contract signed between UZH and PMI. The contract itself was supplied,²³ but an Annex mentioned therein was missing. Despite a further request, access to the latter was not granted:

“The Annex is simply the research proposal. We do not hand out research proposals from our professors, as they contain trade secrets, including for example their research plans for the coming years. Please note that this is a standard policy that is shared by essentially all (perhaps even all – I do not know of any exception) institutions of higher learning.” (Professor Michael Hengartner, Rector of the University of Zurich).

In March 2023, commissioned by the Tobacco Control Fund to document the affair, Pascal Diethelm wrote to UZH, seeking to obtain all the documents available. In May 2023 – almost ten years after the start of the affair – he was surprised to note that, among the documents sent to him by UZH was the Annex to which he had previously been denied access.^24,g This document¹⁴ turned out to be explosive: it showed that PMI had complete control over the research project, thus blatantlly violating the principle of academic freedom.

Adverse impact on public health policies

Carried out within an international context of tensions for the tobacco industry, the study needed to “demonstrate” that plain packaging was ineffective in Australia, in order to win the various legal actions undertaken and to avoid other countries following Australia’s example.

• In 2011, Philip Morris Asia took legal action against the Australian government, arguing that the introduction of plain packaging violated a bilateral investment treaty between Hong Kong and Australia, damaging the investments of Philip Morris Asia and undermining intellectual property rights.
• In 2012, the UK launched a consultation to assess whether plain packaging could be an effective way of combating tobacco use. However, no decision was made following the consultation, and on 12 July 2013 – four days before UZH signed its contract with PMI – the British government announced that it would await the outcomes of plain packaging in Australia before proceeding with legislation.^h
• In the same year, three countries supported by the tobacco industry (Ukraine, Honduras and the Dominican Republic, later followed by Cuba and Indonesia) initiated a WTO dispute resolution process against Australia, arguing that plain packaging violated WTO treaties and claiming that this measure undermined intellectual property and trademark rights and constituted an illegal barrier to trade.

By signing the research contract with PMI and thus affording it the “scientific credibility” which it required, UZH allied itself with the tobacco industry in these procedures. It supported the industry’s tactic of denying the effectiveness of plain packaging with the aim of sabotaging the implementation of this measure in countries eager to follow the recommendations of international health authorities.ⁱ

PMI had complete control over the research

The Annex to the contract between UZH and PMI clearly shows that the tobacco multinational had established mechanisms to ensure that it had a tight control over the research project (excerpts from the contract and from the Annex are shown between quotation marks):

• Regular meetings are to be held between the researchers and PMI. The research work includes “regular meetings of PMI team members and our team members, regular conference calls, and frequent email communication [which] is inevitable for reaching our project goals.”
• Of greater concern, the researchers indicate that the long-term strategic goals of the research will be defined according to the following criteria:
  • “Which directions for future research are worth pursuing from a PMI point of view?”
  • “Which developments in the tobacco control research on plain packaging (in Australia) pose a threat to the interest of PMI?”
• In addition, PMI will decide whether or not the results are to be published and in what format: “Depending on the internal assessment of PMI, it will have to be determined whether and in which format the results of the analysis should be presented to a broader audience”; moreover, “whether the document will eventually be published and in which form shall be decided by PMI.” In addition, it is specified that, at each phase of the study, a report “will be compiled for PMI internal use only”.
• In the contract, it is also specified that UZH is bound to secrecy. The university is not authorised to disclose “the existence or terms of this Agreement or any other aspect of the relationship between the Parties” without the advance written agreement of PMI, nor to respond to media enquiries without its agreement: If contacted by any news organization, UZH is not allowed to answer questions related to the project and must notify PMI of the contact.^j

Researchers’ conflicting interests were not declared

The researchers were subject to a conflict of interests, but this was not clearly indicated in any of the documents which were made public. It is known, as mentioned above, that Kaul was the founder and Research Director of the IPE, the consulting firm which partnered with PMI, and that Wolf was a member of its research network.²⁶ The classified Annex indicates that the IPE was the main contributor to this project, with a third of the budget going to UZH and two thirds to the IPE. It may thus be assumed that there was a separate contract between PMI and the IPE for this second (majority) engagement.

The second conflict of interests is more fundamental: it is impossible to accept an assignment from a tobacco company concerning a question of public health without – inevitably – being subject to a conflict of interests. There is an essential incompatibility between the mission of the two professors, which is to defend scientific knowledge and to assume responsibilities towards society, and defending the interests of a tobacco multinational.

On this question, the position of the United Nations and the World Health Organization (WHO) is clear: “There is a fundamental and irreconcilable conflict between the tobacco industry’s interests and public health policy interests.”^27-29

UZH sacrificed its scientific independence by lending its name to the project

The fact that the entire research team was actually working for the IPE and that two thirds of the budget was allocated to this institute indicates that UZH was simply lending its name so as to endow the research project with a degree of credibility. The Annex, moreover, concludes with the following statement: “In the unlikely event that UZH is not willing to contract with PMI, IPE will take over all services described in this proposal. The staffing and project work as well as the total fee budget would be unaffected in this case.” The IPE was thus prepared to take on the entire project, with the same budget and staffing.

If PMI had simply commissioned the IPE to conduct this research, the results would have been regarded as just one more study produced by an organisation paid for by the industry, with virtually no scientific credibility.

PMI thus had a pressing need to create the impression that the study had been conducted by serious and independent researchers. The fact that UZH, a major international academic institution,^m agreed to collaborate on this project was thus a godsend to PMI, which did not fail to exploit the prestige of UZH in its communications. This is illustrated by an excerpt from a letter PMI sent to Michael Moore, CEO of the Public Health Association of Australia, in 2014: “Perhaps you missed that paper when it was published in the prestigious University of Zurich Department of Economics Working Paper Series […]. Had you seen this publicly available report, you would know that […] the experts could find no evidence of a plain packaging effect.”³⁰

The two articles have not been published in a peer-reviewed journal

The articles were published on the UZH website without having been examined by independent experts. Independent peer review is, however, standard academic practice, attesting to the serious nature of an article.

When Kaul and Wolf were interviewed in March 2014 by representatives of the UK Chantler Review, the aim of which was to assess the state of knowledge on the effectiveness of plain packaging, they stated that their results would be published in a peer-reviewed journal: “We have published all our papers in peer-reviewed journals so there’s no reason to stop here.”³² So why have these articles not been published in such a journal? The authors have never communicated the precise reasons for their failure to do so.

The study includes (numerous) methodological flaws

Omission of explanatory variables

In the Annex to the contract, the researchers state: “It is crucial to thoroughly describe the regulation under consideration and other relevant policy measures in the past. If policy interventions are combined, it is desirable to disentangle the effects of each measure. As part of phase 2 we would therefore - based on PMI material and information - document regulatory interventions in order to identify significant changes to the regulatory environment in prior years, including tax and price increases as well as changes in tobacco control policies. These regulatory changes would then be coded in a way that the relevant information can be included in a statistical analysis.”

However, these variables were not taken into account in the final statistical model, with no explanation being given for this omission, even though the authors subsequently acknowledged the weaknesses of such an approach.

Farley and Diethelm, for their part, reanalysed the same data, taking into account the three other main regulatory measures adopted in Australia – increasing the price of cigarettes, banning smoking in public places, and using graphic images on packets. Their model clearly refutes the findings of Wolf and Kaul, showing a statistically significant plain packaging effect. Their results were published in the peer-reviewed journal Tobacco Prevention & Cessation in 2015 and 2017.^33,34 They were used as the main reference in a 2017 Cochrane Systematic Review on “Tobacco packaging design for reducing tobacco use”, which excluded the Wolf and Kaul studies for methodological reasons.³⁵

Wolf and Kaul did not follow their research protocol, even though it was perfectly reasonable. Had they done so, they would have obtained results contrary to the interests of their funder: this provides a plausible explanation for their change of strategy.

Lack of self-criticism

The authors of the two articles never question the methodology employed in their study, even though it includes numerous flaws, but on the contrary represent it as the best possible approach. The deficiencies should at least have been discussed in the two articles, as is standard practice in any high-quality scientific publication.

Unsuitable statistical model

The researchers base their analysis on the assumption that the reduction in smoking prevalence follows a linear time trend. This assumption is, however, invalid, as shown by other studies, such as those of Wakefield et al. published in 2008 and 2014.^36,37

Low statistical power

The model used by the two researchers is severely lacking in what statisticians call “power”: their analyses were not capable of detecting a significant plain packaging effect,^38-40 making their claim that they found no evidence of a plain packaging effect meaningless.