How does the alcohol industry attempt to influence marketing regulations? A systematic review

1Understanding how large corporations seek to shape health policy has been considerably advanced by the release of millions of internal tobacco industry (TI) documents following litigation in the United States (US) 1, 2, 3. More than 850 peer‐reviewed papers 4 have now been published examining TI political activity, including a growing number of systematic reviews 5, 6, 7, which provide detailed overviews of how the TI seeks to influence policy. Research on alcohol industry (AI) political activity is more case‐study based and, without access to previously confidential industry documents, awareness of the AI's political activity is less developed 8, 9. This is reflected in differences in how public institutions, such as the World Health Organization (WHO), engage with the TI and AI. With regard to the TI, Article 5.3 of the WHO's Framework Convention on Tobacco Control (FCTC, the WHO's first global public health treaty) requires all Parties to protect health policies ‘from commercial and other vested interests of the tobacco industry’ 10 and guidelines for this Article identify the need to actively monitor and expose TI conduct 11. By contrast, the WHO's approach to the AI is more ambiguous; highlighting the importance of protecting the development of health policies from ‘commercial or vested interests’ 12 on one hand, while allowing AI participation 13 on the other. This is despite research illustrating parallels between the political activities of ‘Big Tobacco’ and ‘Big Booze’ 14.

2This paper aims to systematically review the tactics (political techniques) and arguments the AI uses to influence public policy relating to alcohol marketing. This focus is chosen for two reasons. First, AI marketing is known to significantly influence drinking initiation and prevalence 15, 16, 17, 18, and restrictions on alcohol marketing are a key element of alcohol control 19, 20, 21. Second, it provides a basis for making comparisons with TI efforts to influence marketing policies which were systematically reviewed in a paper published in 2014 5. The TI review built on existing methods for categorizing corporate political activity in the management literature 22 by classifying both the strategies/tactics and frames/arguments the TI uses in efforts to oppose marketing policies. The present paper develops and applies the TI framework to the AI. It aims to provide a tool for public health advocates and policymakers to understand, predict and potentially counter the AI's efforts to influence alcohol policy. Our approach responds to a recent recommendation that ‘investigations of the strategies of alcohol industry actors may benefit from comparisons with other industries, and particularly with the tobacco industry’ 23.

3This review aimed to identify all papers (based on either primary or secondary data) that examined AI attempts to influence marketing regulation from 1990 to 2013. The AI comprises large multi‐national companies and tiny specialist brewers, and both on‐ and off‐trade businesses (sales for consumption ‘on’ the premises and ‘off’ the premises, respectively). In this review we include tactics and arguments used by alcohol producers or groups representing producers. Marketing encompasses five key variables: product, promotion, price, place and person 24. Political activity in respect of tax (which affects price) was excluded from the TI systematic review 5 as a systematic review of TI influence on tobacco tax had already been completed 6, so for comparative purposes it was also excluded from this review. However, efforts to influence minimum unit pricing of alcohol were included under price.

4The databases Web of Knowledge (which includes Web of Science, BIOSIS Previews and MEDLINE), Business Source Premier and Embase were searched using the string: (corporat* OR industr* OR compan* OR busines* OR firm*) AND (alcohol OR drink) AND (marketing OR advertis* OR sponsor*) AND (regulat* OR policy OR legislat*). The search engine Google was used to identify grey literature and experts were contacted to identify any additional literature (more information is available in the supporting information, Appendix S1). All searches were conducted between April and July 2011, and were updated in March 2013. Searches were limited to papers from 1990 to 2013 and those written in English. The search protocol was developed in conjunction with a qualified librarian.

5Initial study inclusion/exclusion criteria were piloted and were discussed extensively between all three authors. The final inclusion/exclusion criteria used in this review can be seen in Box 1. In total, 917 papers were identified, 670 of which were excluded based on their title and abstract alone. 239 papers were downloaded for full analysis (eight papers could not be located despite efforts to contact the authors). 222 papers were excluded for not meeting the inclusion criteria. The remaining 17 papers met all the criteria and were included within the review. To be included in this review, studies and individual arguments/tactics had to fulfil the following criteria: Studies must be written in English.Studies must cover the period from 1990 to 2013. In papers that cover both before and after 1990, only those tactics/arguments relating to post‐1990 will be recorded and included within this review. Political activity prior to 1990 is excluded to enable valid comparisons with the findings of our earlier review on the TI 5.Studies must look at AI efforts to influence (a) policy debates concerning marketing regulations generally, (b) new specific marketing policies or (c) broader alcohol policy within which marketing is included (information regarding how the industry attempts to circumvent existing regulation will not be included within the review).The tactics/arguments covered must be related to one or more of the following: product (for example, packaging, new products/flavours, branding), price* (for example, price promotions, minimum pricing), promotion (advertising including billboards, point‐of‐sale, sponsorship), place (for example, restrictions on advertising near schools) or person (for example, restrictions on advertising or selling to youth).Each individual claim made regarding AI tactics/arguments used to influence marketing regulation must be supported directly by verifiable evidence (either a clear citation that could be verified by the authors or a direct quote from an AI official or industry affiliated body).Tactics/arguments identified must be implemented directly by the AI or by a group where substantiated evidence suggests that they act on the AI's behalf.Tactics/arguments which are noted within the included papers are assumed to have been carried through, in the absence of evidence to the contrary. Tactics/arguments which are shown to only have been planned, and not used, will not be recorded.Only tactics/arguments directly related to marketing regulation will be recorded. For example, health warning labels are included as they influence the means of packaging as a marketing tool, but they are excluded if the study only looks at, for example, the wording of the warning, as this does not affect marketing.Only tactics/arguments that are clearly detailed in the paper(s) are coded. *Price in the form of tax has been excluded because tax‐related lobbying was excluded from the systematic review of TI political activity 5, and we aimed to make the TI and AI reviews comparable. Price in terms of price‐based promotions have been included.

6To be included in this review, studies and individual arguments/tactics had to fulfil the following criteria: Studies must be written in English.Studies must cover the period from 1990 to 2013. In papers that cover both before and after 1990, only those tactics/arguments relating to post‐1990 will be recorded and included within this review. Political activity prior to 1990 is excluded to enable valid comparisons with the findings of our earlier review on the TI 5.Studies must look at AI efforts to influence (a) policy debates concerning marketing regulations generally, (b) new specific marketing policies or (c) broader alcohol policy within which marketing is included (information regarding how the industry attempts to circumvent existing regulation will not be included within the review).The tactics/arguments covered must be related to one or more of the following: product (for example, packaging, new products/flavours, branding), price* (for example, price promotions, minimum pricing), promotion (advertising including billboards, point‐of‐sale, sponsorship), place (for example, restrictions on advertising near schools) or person (for example, restrictions on advertising or selling to youth).Each individual claim made regarding AI tactics/arguments used to influence marketing regulation must be supported directly by verifiable evidence (either a clear citation that could be verified by the authors or a direct quote from an AI official or industry affiliated body).Tactics/arguments identified must be implemented directly by the AI or by a group where substantiated evidence suggests that they act on the AI's behalf.Tactics/arguments which are noted within the included papers are assumed to have been carried through, in the absence of evidence to the contrary. Tactics/arguments which are shown to only have been planned, and not used, will not be recorded.Only tactics/arguments directly related to marketing regulation will be recorded. For example, health warning labels are included as they influence the means of packaging as a marketing tool, but they are excluded if the study only looks at, for example, the wording of the warning, as this does not affect marketing.Only tactics/arguments that are clearly detailed in the paper(s) are coded.

7Studies must be written in English.

8Studies must cover the period from 1990 to 2013. In papers that cover both before and after 1990, only those tactics/arguments relating to post‐1990 will be recorded and included within this review. Political activity prior to 1990 is excluded to enable valid comparisons with the findings of our earlier review on the TI 5.

9Studies must look at AI efforts to influence (a) policy debates concerning marketing regulations generally, (b) new specific marketing policies or (c) broader alcohol policy within which marketing is included (information regarding how the industry attempts to circumvent existing regulation will not be included within the review).

10The tactics/arguments covered must be related to one or more of the following: product (for example, packaging, new products/flavours, branding), price* (for example, price promotions, minimum pricing), promotion (advertising including billboards, point‐of‐sale, sponsorship), place (for example, restrictions on advertising near schools) or person (for example, restrictions on advertising or selling to youth).

11Each individual claim made regarding AI tactics/arguments used to influence marketing regulation must be supported directly by verifiable evidence (either a clear citation that could be verified by the authors or a direct quote from an AI official or industry affiliated body).

12Tactics/arguments identified must be implemented directly by the AI or by a group where substantiated evidence suggests that they act on the AI's behalf.

13Tactics/arguments which are noted within the included papers are assumed to have been carried through, in the absence of evidence to the contrary. Tactics/arguments which are shown to only have been planned, and not used, will not be recorded.

14Only tactics/arguments directly related to marketing regulation will be recorded. For example, health warning labels are included as they influence the means of packaging as a marketing tool, but they are excluded if the study only looks at, for example, the wording of the warning, as this does not affect marketing.

15Only tactics/arguments that are clearly detailed in the paper(s) are coded.

16*Price in the form of tax has been excluded because tax‐related lobbying was excluded from the systematic review of TI political activity 5, and we aimed to make the TI and AI reviews comparable. Price in terms of price‐based promotions have been included.

17Data extraction (supporting information, Appendix S1) was undertaken by the lead author, and 100% of the included papers were second‐reviewed by either the second or third author to check that all the inclusion criteria were met and to agree tactic and argument categorization. Any differences were discussed between all three authors. Disagreements related only to categorization, more often in relation to the categorization of arguments than tactics. Where disagreement occurred, all evidence falling under that particular category was re‐reviewed by all three authors until agreement had been reached. Narrative synthesis was undertaken to combine the evidence from the papers.

18Unlike the TI review 5 which was based solely on secondary data, this review is based on both primary and secondary data. Primary data came predominantly from a United Kingdom (UK) parliamentary inquiry into alcohol where four producers and their communications agencies were asked to provide documents relating to five brands and were questioned by Members of Parliament from the Health Select Committee; many additional companies, trade groups and social aspect organizations (SAOs) also provided written evidence 25, 26, 27. Additionally, due to a lack of evidence focusing specifically on AI efforts to influence marketing regulations, the review was expanded to include AI influence on marketing policy debates alongside their influence on specific marketing regulations, as per the TI review 5.

19AI political activity was divided into ‘strategies’ containing individual ‘tactics’ (the methods by which a corporation attempts to exert influence) and ‘frames’ containing individual ‘arguments’ (the reasons given by a corporation as to why they oppose one idea or support another). The system of classification developed in an earlier systematic review of TI political activity 5 (which, in turn, had been based partly on Hillman & Hitt's paper 22) was used as an initial framework to code AI political activity. Coding categories (strategies/tactics and frames/arguments) were amended and developed via ‘emergent coding’ 28. This was an iterative process, and the frameworks were only finalized after all of the papers had been reviewed as described above. Once the framework of political activity was finalized, the strategies/tactics and frames/arguments used by the AI were compared to those identified in the systematic review of TI political activity 5.

20The geographical distribution of where tactics and arguments were used was also recorded. If the paper included was transnational, the geography of where the individual tactics and arguments were used was listed. For example, the paper by Casswell & Thamarangsi 29 is a transnational study, but the ‘free market economy’ argument was used in France.

21In total, 17 papers met our inclusion criteria. A quarter (24%) of the papers focused on Europe, and a further quarter (24%) were transnational (Table 1). No papers focused on AI conduct in South America.

22This review identified 20 separate tactics falling under five main strategies (Table 2), which we have termed as follows: ‘information’ (providing or misrepresenting evidence), ‘constituency building’ (forming alliances with other sectors, organizations, or the public to give the impression of larger support for the industry's position), ‘policy substitution, development and implementation’ (proposing, supporting or helping to implement alternative policies), ‘legal’ (using the legal system) and ‘financial incentive or disincentive’ (offering direct or indirect monetary incentives or threatening financial withdrawal) (see further details included in the supporting information, Appendix S1).

23A variety of information strategies were used across multiple jurisdictions. These include direct 30, 32, 38 and indirect 30 lobbying of policymakers and establishing collaborative working arrangements with policymakers 30, and a variety of efforts aimed at shaping and manipulating the evidence base. The latter included commissioning, writing or disseminating research/publications 27, 32, 42 or more technical reports 32, 38, 42, the selective citation 27, 42 and omission of evidence 30, contesting the evidence used to support policy 27, 42 and the efforts to remove ‘troubling’ phrases such as ‘alcohol and other drugs’ from the official lexicon 42. The AI‐funded International Center for Alcohol Policies (ICAP) has played a key part in such efforts: commissioning and publishing a large number of books, monographs, briefing papers, in‐depth reviews of alcohol policy issues, journal papers and policy guides on all manner of alcohol‐related issues 42. These activities have populated the evidence base with non‐peer‐reviewed research which, among other things, tends to highlight the health benefits of alcohol 27, 42 and omit evidence of its negative health and social effects 30.

24Constituency Building was often linked to indirect lobbying. The AI creates front groups, astroturf organizations Astroturf organizations can be defined as ‘fake grassroots organizations usually created and/or sponsored by large corporations to support any arguments or claims in their favor, or to challenge and deny those against them’ 44. or SAOs (such as ICAP 42, the Portman Group 36, The DrinkAware Trust 27 and the Federation on Alcohol Concern of Thailand (FACT) established during the formation of an advertising ban in 200632) to lobby on its behalf 28, 32. It also forms alliances with other industry sectors or trade organizations 28, 32, 41, and civil society organizations, consumers or employees 28, 32 in order to oppose public health measures 41. In Thailand the AI worked with groups such as the Thai Retail Association, the Hotel Association, the Restaurant Association, the Tourism Association and the Marketing Association of Thailand 32, and in the US it reached out to the Federal Trade Commission 28 and built partnerships with government departments, non‐governmental organizations (NGOs), universities, researchers and physicians 28. The AI also uses media advocacy, such as press launches 38 and seminars 32, to shape the news and public agenda.

25Astroturf organizations can be defined as ‘fake grassroots organizations usually created and/or sponsored by large corporations to support any arguments or claims in their favor, or to challenge and deny those against them’ 44.

26Policy substitution, used to prevent the implementation of formal marketing regulations, appears to be a key strategy and has been documented globally 27, 28, 29, 30, 32, 35, 36, 37, 38, 40, 42. For example, in Lesotho, Malawi, Uganda and Botswana SAB‐Miller Africa was given de facto responsibility for drafting national alcohol policy documents 30. These policy documents focused on self‐regulatory measures, education campaigns, and gave responsibility for the policy's implementation to a National Alcohol Council on which AI representatives served 30. The promotion of self‐regulatory measures is designed to reduce political pressure for and pre‐empt formal regulation and was identified in numerous jurisdictions. For example, we found evidence of voluntary codes being developed and promoted by individual companies 27, 28 and by industry groups in the UK 27, 36, Ireland 38, the Netherlands 37 and transnationally 29. Another technique involves the promotion of non‐regulatory initiatives such as education programmes 27, 28, 30, 40, 42 delivered through stand‐alone websites (for example, SABMiller's www.TalkingAlcohol.com 27) or more (CSR) initiatives. For example, Diageo's Responsible Drinking Fund, which in 2009 claimed to have led or supported more than 130 prevention programmes in more than 40 countries, covering ‘education, public awareness, and responsible retail practices’ 28.

27Using or raising the prospect of legal action against a proposed regulation was documented only in Thailand 32 and France 29, but there is also evidence of the AI attempting to shape international trade and investment agreements (specifically the General Agreement on Trade in Services (GATS) with a view to reducing restrictions on alcohol distribution and advertising 29). We also documented one example of the AI using its marketing budget as a lever of policy influence (financial disincentive); in Thailand the AI ‘threatened to withdraw sports sponsorship in retaliation for [an] advertising ban’ 32.

28This review identified 20 separate arguments grouped into five main frames (Table 3): ‘regulatory redundancy’ (asserting that proposed policies are unnecessary), ‘legal’ (questioning the legality of policies (the implicit cost for government)), ‘negative unintended consequences’ (direct and indirect compliance costs associated with proposed policies), ‘complex policy area’ (policies, and the issues surrounding them, are presented as highly complicated) and ‘insufficient evidence’ (questioning the strength of evidence supporting policies) (see the supporting information, Appendix S1 for further details).

29The argument that population‐level health measures are unnecessary (regulatory redundancy frame) is developed through a wide range of mutually reinforcing arguments which rest on industry claims of its own responsibility, its ability to market alcohol responsibly, and its distinction between responsible and irresponsible consumption. This frame included arguments that the AI is responsible 27, 33, 34, 41 (for example that industry always encourages ‘responsible consumption’ 33 and recognizes ‘that responsible drinking is important both to [its] business interests and to society's interests’ 27), that self‐regulatory codes are ‘sufficient’ 31, ‘robust’ 33, ‘effective’ 32, 33, ‘extraordinarily successful’ 35, ‘faster’ 27 and ‘better’ 32, allowing the AI to deal quickly with, and rectify, any complaints or regulatory breeches 27, 34 and close regulatory ‘gaps’ 27, and that the industry markets only to those of legal drinking age 27, 28, 34, 41. Further, the AI appears to overstate the parallels between voluntary and statutory regulation for example by emphasizing the independence of their (industry‐funded) monitoring and adjudication groups 27.

30This set of arguments overlaps with claims around personal responsibility and responsible drinking. The AI frequently attempts to shift the blame for alcohol misuse to the consumer and away from their products and marketing 42, arguing that there should be an individual‐level focus on education and the promotion of responsible consumption 27, 30, 42 (and even that AI marketing itself has this aim 41) and that their SAOs such as the DrinkAware Trust and the Portman Group provide information and education so that consumers can make ‘informed judgements’ about how they use alcohol 27; ‘misuse is caused by certain drinkers who clearly misuse alcohol and by some under 18s who are clearly breaking the law. This therefore is not a problem about problem drinks but about problem drinkers’ 27. The focus on a small number of alcohol misusers provides the AI with a frame that has the potential to invalidate the current focus of health policy; the AI argues that population‐level approaches, such as taxation or restrictions on advertising, penalizes moderate drinkers because of a ‘few people’ who consume alcohol in an irresponsible way and that these approaches do not tackle alcohol misuse effectively 27, 34, 42. This supports AI claims that ‘existing regulation is satisfactory’ 27, 33 or that it simply ‘requires better enforcement’ 27, 32, 38; ‘the panoply of powers available to the police and local authorities should be used much more effectively both against individuals who misuse alcohol and those who wilfully seek to break the law in obtaining alcohol underage, as well as against those retailers who sell alcohol irresponsibly’ 27.

31The argument that regulations are disproportionate and more extensive than necessary also rests on claims of personal responsibility and the inherent health risks of alcohol. In relation to health warnings in the UK it was stated that it is not ‘necessarily appropriate to have a health warning on a drink of alcohol. Alcohol is not like cigarettes; it is capable of being misused but when drunk in moderation it is perfectly compatible with a healthy lifestyle’ 27. This argument has also been used regarding minimum pricing 27, dealing with alcohol misuse 27 and raising the legal drinking age 42.

32Additionally, the AI argues that the industry, its marketing, and alcohol itself, has a positive impact and should therefore not be regulated further: in Lesotho the AI argued that when used in moderation alcohol has a ‘positive role to play in socialisation’ and that the industry is a ‘major contributor to the economy’ 30. In Milwaukee, USA, Miller Brewing argued that a product (beer) which forms a significant part of the local history should not be ‘demonised’ 39. In Australia the AI argued that ‘responsible advertising can have a positive cultural impact’ 33 and highlighted the importance of the AI to communities as ‘funders of events’ 34. And in the UK the wine company Constellation argued that marketing could be used ‘to promote a more responsible approach to alcohol consumption’ 27 (a very similar statement was also made in the US by Philip Morris (tobacco) when they owned Miller Brewing 41).

33While actual legal action is rarely reported in the literature (see above), arguments questioning the legality of policies to curb AI marketing are more common. These arguments aim to shift the focus of the debate away from public health and consumer protection, with the AI contesting advertising regulations and minimum pricing proposals under international trade and investment agreements (UK 27 and Thailand 32), national constitutions (USA 41) and international law (UK 27).

34The AI often argues that regulation would have negative unintended consequences for manufacturers 27, 32, 34, 40, associated industries 32, 34, the public revenue 34 and public health 27, 33, 41. For example, in both Australia 33 and the UK 27 it was argued that advertising restrictions would make it impossible to introduce new, lower‐strength products to the marketplace, thereby stopping producers from developing and selling healthier products, and in the US the Beer Institute argued that mandated health warnings on alcohol products could ‘undermine the credibility of other government campaigns to provide information about serious risks which are not commonly known’ 41. The AI commonly avoids citing evidence to support such claims, suggesting that the aim may simply be to exaggerate the broader political and economic risks associated with public health measures.

35The AI also emphasizes the complications involved in addressing problems associated with alcohol consumption with a view to highlighting the value of industry–government cooperation 27, 30, 33, 38; it claims that it has ‘a unique capacity to access those responsible for promoting and selling alcohol as well as to those who consume their products’ 30 and that ‘partnership working can deliver a more responsible drinking culture’ 27. Although the AI advocate being involved closely in policymaking and alcohol harm reduction initiatives, they also stress that some issues are beyond the AI's control; for example, preventing individual retailers from offering certain price promotions 27 that may be deemed irresponsible. Similarly, the AI regularly characterizes policymakers and public health actors as authoritarian (‘the health lobby's approach is to ban everything, and if it cannot be banned, regulate it severely’ 33) with the Thai government being labelled a ‘dictatorship’ because of an advertising ban 32, and the Australian government being described as a ‘“nanny state” needlessly interfering with people's choices’ 34.

36Questioning the strength of evidence favourable to public health policies is another common technique that has been used to oppose advertising bans transnationally 41, in Thailand 32, Australia 34, Ireland 38 and in the UK 27, minimum pricing in the UK 27 and health warning labels in the US 40. This argument is used to reinforce the other arguments made by the AI.

37We identified 13 common tactics used by both the AI and TI 5 when attempting to influence marketing regulation, in addition to five tactics used only by the TI and seven unique to the AI. Similarly we also identified 13 common arguments used by both industries, along with four arguments unique to the TI, and seven (three of which formed the new frame ‘complex policy area’) which had been used only by the AI (see further details included in the supporting information, Appendix S1).

38This systematic review illustrates the varied nature of AI political activity used in attempts to influence marketing regulation or marketing‐related policy debates, and highlights similarities with TI political activity.

39The results support the findings of the TI review 5 in highlighting the varied nature of industry political activity and provides further confirmation that Hillman & Hitt's 22 model of corporate political activity, which is the most widely cited attempt to analytically categorize the tactics used by corporations, considerably under‐represents the range of tactics corporations use to shape policy outcomes and debates. By identifying tactics/strategies, such as the promotion of self‐regulatory codes and raising the prospect of litigation, the results also challenge Hillman & Hitt's 22 assumption that corporate political activity represents one side of a mutually beneficial exchange relationship in which corporations offer policymakers support and information in return for influencing policy.

40The existing literature challenges the validity of many of the arguments identified in this review (see examples in Table 4). For example, despite the AI's assertion that self‐regulatory codes negate the need for formal regulation and that industry collaboration would be beneficial for policymakers, there is no evidence that self‐regulation and industry–government partnerships lead to reductions in alcohol‐related harm 19, 20, 25, 60. Arguing that there is insufficient evidence supporting the need to curb AI marketing and that marketing does not change behaviour is also false as much research has found a significant link between AI marketing and drinking initiation and drinking prevalence 15, 16, 17, 18. Similarly, arguing that the AI does not market to children is misleading, as research shows that AI marketing often targets and appeals to youth and those below legal drinking age 19, 36, 67, 68.

41This review identifies marked similarities between TI and AI political activity 5. Differences in the observed political activities of each industry may be due to a number of factors. First, raising concerns about compensation or debating which body has power to regulate (as the TI has done), for example, are arguments likely to be made in the face of impending regulation by companies which have lost the ability to exercise insider influence over policy discussions, reflecting greater TI regulation and differences in alcohol and tobacco policy contexts. Second, differences in framing may reflect variations in how different industries make similar points. For example, while the AI may not directly contest the health impacts of alcohol consumption, questions about the degree of harm caused by alcohol consumption are implicit in claims concerning individual responsibility and the health benefits of alcohol consumption. Third, differences may reflect differences in access to data; because of the availability of TI documents, information on lower visibility political activity, such as raising the prospect of legal action, is more available on the TI. Finally, differences may reflect the broader inclusion criteria used for the AI review (i.e. covering policy debates concerning marketing regulations generally rather than just new specific marketing policies), the inclusion of both primary and secondary evidence in the AI review, and differences in alcohol and tobacco policy contexts. For example, despite not being identified in the TI review there is evidence of the TI attempting to shape the evidence base 69, 70, 71, 72, influencing international regulations 73, 74, and focusing on individual responsibility 75.

42Consistent with the TI review findings 5, many of the individual arguments fall within a larger ‘cost–benefit’ meta‐frame which promotes the economic and social costs of proposed public health policies and underplays their benefits. Arguments claiming that regulation is more extensive than necessary and likely to produce negative unintended consequences are used to increase uncertainty about the likely benefits of regulation, and highlight the potential future costs for the industry, retailers, and the public revenue. This is also observed through the omission of a ‘health’ frame 76; this review found little evidence of the AI making reference to the dangers of drinking alcohol (only in terms of references to ‘problem drinkers’), although multiple examples of the AI highlighting the potential health benefits of alcohol consumption were identified 27, 42. The review also found that many arguments were supported by CSR activities. CSR tends to be used strategically by an industry to prevent the introduction of legislation 77. By acting as vehicles for the promotion of arguments, CSR activities such as self‐regulatory codes work politically as agenda‐setting devices which frame issues and shape policy debates 77. The AI's emphasis on CSR highlights its value in maintaining industry credibility and forming relationships (CSR partnerships are likely to create further opportunities for cooperation 78) ahead of regulation.

43This review has a number of limitations. First, although a broad search strategy and search string was used when initially identifying papers it is still possible that some relevant papers may have been missed. To minimize this, we worked with a librarian, searched online research repositories, and contacted experts in the field to identify additional papers. Second, interpretive coding of arguments and tactics is ultimately subjective. To mitigate this, all three authors reviewed and re‐reviewed the coding at various points during the systematic review process and second‐reviewed all the included papers to ensure consistency. Third, the identification of tactics and arguments, and the jurisdictions in which they are used, is dependent upon the available literature, its quality, and any publication bias. As such, it is possible that some tactics and arguments are not identified in or used more frequently than the literature would suggest. Closely related to this is the fact that the review focuses only on marketing policy, and the AI may use a more diverse set of tactics and arguments in other policy areas. For these reasons the number of papers listed next to each tactic and argument (the ‘count’) should be used only as an indication of the reliance the AI places on particular tactics and arguments. Finally, due to limited information in the papers identified, it was not possible to reliably determine which tactics or arguments were most persuasive or successful in defeating marketing‐related regulations.

44The main strength of this review is its systematic approach, which provides a comprehensive and geographically diverse overview of AI tactics and arguments. Its attempt to rigorously categorize industry strategies/tactics and frames/arguments is, to our knowledge, along with our first paper 5, the first attempt to do so. While care needs to be taken in assuming that tactics and arguments used in one jurisdiction will be used elsewhere, this and our previous review 5 suggest that the findings will be broadly applicable across different jurisdictions.

45This systematic review has identified strategies/tactics and frames/arguments used by the AI between 1990 and 2013 to shape policy debates and prevent the implementation of restrictions on alcohol marketing. Policymakers need to be aware of these in order to understand how the AI may try to influence the policymaking process, and public health actors can use this information to prepare effective counter strategies and arguments. This review has also confirmed substantial commonalities between AI and TI political activity: particularly the use of obfuscating tactics such as misrepresenting the evidence base and using third parties and front groups to lobby. The similarities suggest that alcohol policy may benefit from reproducing efforts in tobacco control aimed at excluding corporate actors from the policy process and enhancing transparency. Additionally, as differences between the two industries are likely to be due, at least in part, to differences in alcohol and tobacco policy contexts, the findings from the TI review 5 may provide an indication of how AI political activity is likely to develop under conditions of increased regulatory risks.

46The current review has further developed the frameworks for classifying corporate political activity outlined in the earlier TI review 5, and shown the policy and scholarly value of applying them to other industries. Future work could apply these frameworks to other industries or policy areas. Based on limitations in the studies reviewed, we again recommend that future research on corporate policy influence should, where possible, include contextual information, ensure all claims made within the paper are supported by empirical evidence, and that the receptivity of stakeholders to and the success or failure of individual tactics and arguments are recorded.

47None.