Thursday, 30 May 2019

In defence of food advertising

The Department of Health's consultation on banning adverts for HFSS (high in fat, sugar and salt) food and drink before 9pm on television closes on June 10th. The government's various allies and sockpuppets have been encouraging people to make a submission so I thought I would too.

Banning advertising is a hunch-based policy, but the Department of Health's impact assessment is obliged to make it look evidence-based. It acknowledges that there will be costs from a TV ban but reckons that it will have a net benefit of £2.3 billion over 25 years, close to £100 million per annum. Most of the putative benefit (£1.4bn) comes from health benefits due to lower calorie consumption, plus £600 million from 'reinvesting cost savings back in the NHS'. Food companies are also expected to save £300 million 'from no longer advertising on TV' (how strange that food companies haven't opted to make these savings voluntarily).

The government starts from three premises, all of which are false.

Problem and justification for action

1. Childhood obesity is one of the biggest health problems this country faces. Around one fifth of children in England are obese or overweight by the time they start primary school aged five, and this rises to more than one third by the time they leave aged 11.

2. Obesity is a major determinant of ill health. This imposes a substantial burden on the NHS, with overweight and obesity estimated to have cost the health service £6.1bn in 2014/15. Obesity causes further costs to society through premature mortality, increased sickness absence and additional benefit payments.

3. It’s clear from the evidence that marketing and TV advertising can be effective at influencing children's food and drink consumption, preferences and purchases. Although food habits are not perfectly stable over life, there is potential scope for influencing lifetime habits by intervening in childhood.

1. Regular readers know that the true rate of childhood overweight/obesity among five year olds is nowhere near 20 per cent and the rate among 11 year olds is nowhere near 33 per cent. Britain uses a measure of childhood obesity that has no scientific or clinical justification. The true rate of childhood obesity is not known because no one has bothered to send a representative sample of children to a clinician to be diagnosed, but it is almost certainly below five per cent for both age groups.

2. While obesity can certainly cause ill health and has associated costs, the £6.1 billion figure is a gross cost whereas the useful measure is the net cost. When the net cost is calculated, by subtracting savings to other departmental budgets and acknowledging the costs that would be incurred if people died of something else, the cost is greatly reduced. It could, in fact, be less than zero.

3. The evidence linking advertising to children's food and drink consumption is far from 'clear'. The impact assessment cites this review by the half-mad anti-capitalist Gerard 'the corporations!' Hastings to support its claim, so forgive me if I'm sceptical. I'll come back to the evidence in a moment.

In short, the government wants to tackle something that is far less common than it thinks by restricting something that is far less important than it thinks, in order to make financial savings that probably don't exist.

There is a moment of unintentional humour on page 24 of the impact assessment when the Department of Health expresses its concerns about children watching television after 7pm:

..children’s viewing time peaks between 6-9pm, when the programmes most likely to be broadcast are not children’s programming, but instead ‘family’ or adult programmes. This means that some of the shows most watched by children, such as X Factor, Saturday Night Takeaway, or Great British Bake Off, are not captured by the current restrictions.

Heaven forbid that kids should see sugary food promoted during the Great British Bake Off. Or, indeed, see adverts for takeaways during Saturday Night Takeaway.

The impact assessment's projections of health benefits and related financial savings rely on the assumption that '4.4 minutes of food advertising results in an additional 60kcal of consumption'. The Department of Health acknowledges that the 95% confidence interval for this estimate is 3.1kcal-116.9kcal, ie. it is barely significant, but nevertheless uses it as the basis for its projections on health and spending.

The 60 calorie statistic comes from a review published last year by Russell et al. (which was co-authored by activist-academic Russell Viner). It is unreliable, to say the least.

The figure comes from a meta-analysis of eleven studies, all of which were randomised experiments in which children had their food consumption monitored during or after being shown food/non-food advertising. These experiments do not reflect children's natural living environment in two crucial ways. Firstly, the children are allowed to consume an unlimited quantity of food at no cost. Secondly, there was no parental supervision.

Furthermore, all but one of the studies involved children aged 12 or younger (the other involved children aged 9-14). In three of them, the children were no older than seven. Research into advertising has shown that children are indeed susceptible to advertising under the age of seven - in so far as they tend to want what they see advertised - but that by the age of 12 they have acquired a 'global distrust' of it. Older children are, as Michael Schudson notes in Advertising: The Uneasy Persuasion, 'very much like adults in their skepticism'.

So whilst advertising directly to very young children is morally dubious, it doesn't greatly matter because they have very little purchasing power and don't do their own grocery shopping. Older children have more purchasing power but they are as sceptical about advertising as adults, so that doesn't greatly matter either.

What the government intends to do is ban HFSS food advertising in the evening when adults and older children are watching because research shows that advertising has an effect on younger children.

What does the research actually show? As mentioned, the experiments focus on the most suggestible age group and remove the two barriers that prevent children from acting on their urges - parental restraint and a lack of purchasing power - but even if we ignore these shortcomings, the evidence is not strong enough for the 60 calorie claim to be taken seriously.

Five of the eleven studies come from the same research team in Liverpool (Halford et al. 2004, Halford et al. 2007, Halford et al. 2008, Dovey et al. 2012, Boyland et al. 2013). The first of these studies found that children under the age of 12 tended to eat more food after watching 'unhealthy food adverts'. The subsequent studies have broadly replicated that result, but it has not been strongly replicated by research groups elsewhere.

Taken together, the other six studies in the Russell analysis provide very little evidence that food advertising has any consistent effect on calorie intake, even when food is given free and parents are absent. They are...

Anderson et al. (2015). This study of 9-14 year olds found that seeing food advertising was associated with higher food intake among overweight girls but not among normal weight girls or any group of boys.

Anschutz et al. (2009). This study of 8-12 year olds found that when the children had seen food commercials, food intake was higher among boys, but lower among girls.

Anschutz et al. (2010). This study of 8-12 year olds found no association between food advertising and food consumption ('there was no main effect of exposure to the food commercials on food intake').

Norman et al. (2018). This study of 7-12 year olds in Australia found that the children 'ate comparable amounts after both food (1933 ± 619 kJ) and non-food (1929 ± 678 kJ) advertising exposures'.

Only Harris et al. (2009) and Emond et al. (2016) support the advertising hypothesis, but the children in the latter study were very young indeed (2-5 years)  and the study was not well randomised; the group that watched the food adverts contained two obese children and six overweight while the group that watched the non-food adverts contained no obese children and four overweight children.

If one looks at the studies themselves, rather than rely on Russell et al's interpretation, it becomes clear that the 60 calorie claim is based on some very mixed evidence, dominated by a series of studies from one research group whose members have strong feelings about the subject. From this dubious figure, the impact assessment extrapolates health benefits worth over £2 billion pounds, based on the belief that it will reduce the number of children who are obese - most of whom are not actually obese to begin with.

The whole thing is a fantasy. As with so much in the modern 'public health' racket, it is a model within a model. The sunshine of reality is never allowed to creep through the curtains.

There is real world evidence to consult for those who are minded to seek it out. The impact assessment notes that there were 12.1 billion 'HFSS child impacts' in 2005. In 2009, after HFSS advertising was banned on children's television, this fell to 7.7 billion. By 2017, it had fallen to 3.6 billion, of which 2.6 billion were before 9pm.

Children's 'exposure' to HFSS food advertising on television has therefore declined by more than two-thirds in the last 14 years. Given the government's assumptions about the link between HFSS food advertising and childhood obesity, this should have led to a notable decline in childhood obesity. In reality, the rate of what the government defines as 'childhood obesity' has been basically flat since 2008 when the ban came into effect.

This is the elephant in the room. The government expects there to be a measurable decline in childhood obesity from HFSS 'exposure' dropping to 8% of the 2005 rate and yet there has been no apparent impact on childhood obesity from such exposure dropping to 30% of the 2005 rate. What is it they say about people who keep doing the same thing and expecting different results?

Although the Department of Health mentions the huge decline in 'child impacts', it never asks the obvious question and therefore never draws the obvious conclusion - that food advertising has a negligible effect on the number of calories children consume.

'But if advertising doesn't work', cry the simple-minded critics of advertising, 'why do businesses spend so much money on it?!' This reflects a basic misunderstanding of what advertising is trying to achieve.

Economists have long understood that advertising does not have the mystical powers that its critics attribute to it...

Borden's (1942) massive study of the effects of advertising on demand concluded that advertising is not generally an important determinant of industry sales. Exceptions arise in new and growing sectors, where advertising can serve to accelerate growth that would occur in any case. Recent work seems generally to support these conclusions...

Overall, advertising does not emerge from the empirical literature on consumer demand as an important determinant of consumer behaviour...

The government is on the brink of passing a law preventing biscuits being advertised until after dark. This should strike any reasonable person as absurd. It is also profoundly illiberal. Fundamentally, this is an issue of free speech. Under Advertising Standards Authority rules, commercials cannot be run unless they are decent, honest and truthful. Why would we want to ban something that is decent, honest and truthful?

We have got to the point where we are banning advertisements because the government and its cheerleaders don't like the product that is being advertised. Well, tough. As Seldon and Harris said in Advertising in a Free Society sixty years ago...

Because advertisements are used to sell almost every conceivable product and service, they offer a large target for those whose real objection is to the thing advertised. People who disapprove of betting, smoking, drinking, hire purchase, self-medication, birth control, Roman Catholicism or ‘Billy Graham’ campaigns, all find advertisements to condemn; and they are joined by those who object to the intrusion of commerce into their comfortable lives. No doubt advertising mirrors the imperfections of human society, but we shall not waste much time on critics who aim at the reflected image instead of declaring openly against smoking or gambling or hire purchase or whatever it is they dislike. While the law permits such activities, their advertising must be tolerated.

There is no meaningful difference between the state banning an advert because it disapproves of the product being advertised - or the views of the person featured in it - and the state censoring an article, speech or play being it doesn't like what is being said. If a ban on cakes, ice creams and jam being advertised before the watershed becomes law, it will give licence to every obsessive, single-issue fanatic to press for further prohibitions. Once we start banning inoffensive adverts for harmless products, the censoriousness will become a runaway train.

No comments: