On Monday, the New York Times ran an exposé claiming that “five decades of research into the role of nutrition and heart disease, including many of today’s dietary recommendations, may have been largely shaped by the sugar industry”.
As evidence for this allegation it cites a paper that claims a lobby group made a $6,500 grant to three Harvard scientists in 1967, to exonerate sugar and implicate saturated fat in obesity-related diseases.
It is true that official dietary guidelines ended up advocating low-fat diets, as I related in a previous column. It is also true that this advice was flawed, and based on what now seems to be methodologically weak research. (For a detailed account of how the original low-fat dietary guidelines of 1977 got it wrong, read The sugar conspiracy, by Ian Leslie in the Guardian. For an excellent critique of that piece, read this response. It takes no sides and, as I did, cautions against concluding the opposite of the 1977 guidelines – that fat is healthy and sugar is not.
However, academic research grants by private industry were commonplace at the time, and conventions about disclosing them were decades in the future. The food industry was not then, and is not today, homogeneous. Producers of all kinds of foods would have funded research that might strengthen their lobbying position with the government.
It is preposterous to suppose that a mere $6,500, which funded only a two-part review of the existing literature and no original research, “largely shaped” half a century worth of dietary science. Moreover, the three scientists implicated did not include Ancel Keys. He was the giant in the field of nutrition who challenged the findings of John Yudkin, the leading proponent of the hypothesis that sugar was the main culprit in heart disease and other obesity-related diseases – apparently with good cause. He was the author of the groundbreaking Seven Countries Study, which led to the low-fat dietary guidelines promoted by George McGovern, then chairman of the United States Senate Select Committee on Nutrition and Human Needs. And the sponsors of the Seven Countries Study, which unlike the small pro-sugar paper must have cost millions of dollars, were government agencies and public organisations across the world.
So, while it is clear today that sugar is implicated in obesity and related diseases, the belief that saturated fat was a bigger culprit was not the result of a secret plot by the sugar industry to keep everyone in the dark. There was broad scientific consensus based on primary research funded largely by public money.
Ironically, the low-fat diet controversy reveals that no tax is required to change people’s dietary habits. Thanks to the authority of scientific findings, government guidelines and coverage in respected media, the supermarket shelves began to fill with low-fat and fat-free products. Margarine replaced butter. Vegetable oil replaced lard. Lean cuts of meat became all the rage. Removing fat from meat, dairy and other foodstuffs cost money, as did replacing tasty fat with other taste-enhancing additives such as MSG and sugar, but the food industry gladly did so. It marketed products as healthier for it.
In last week’s column, we saw that the mathematical model on which the National Treasury based its proposal to tax sugary drinks does not withstand scrutiny. Even if its optimistic assumptions hold, which is a big if, it predicts only a very small decline of 0.34% in daily energy intake in adult males, which might lead to weight loss of 0.383kg. At best this will have a negligible impact on public health and obesity-related costs to the state.
This conclusion alone should be enough to sink a sugar tax proposal. But since it probably won’t, let’s tackle another major part of the government’s policy paper: its appeal to international precedent for a tax on sugar-sweetened beverages. That argument turns out to be just as weak as the fancy model with all the happy assumptions.
In the case of Mauritius and the UK, the Treasury policy paper claims only that a sugar tax was instituted, without making any claims that it had a measurable effect, so let’s focus on the remaining countries the paper mentions.
A recent paper in the British Medical Journal compared actual SSB consumption in Mexico at the end of 2014, a year after the implementation of a 10% tax on SSBs, with modelled consumption had the tax not been implemented based on trends established in 2012 and 2013. It found an average decrease in consumption of 6%. This number is cited prominently in the Treasury policy paper.
However, sales data do not support this claim. On the contrary: sales rose significantly, by 6.4% in 2014 and by 7.0% in 2015. Mexico’s National Institute of Public Health (NIPH), which strongly supports a sugary drink tax itself, argues that sales data cannot be used to assess the effectiveness of the tax, because of confounding factors such as population changes, changes in economic activity, seasonal changes, previously existing trends, and new marketing and promotion strategies by the soft drink industry.
Its explanation of these factors does not add up, however. Using per-capita figures to account for population change, which seems entirely reasonable, shows that consumption of taxed drinks increased from 160 litres per capita before the tax was implemented, to 162 litres per capita in 2014 and 161 litres per capita in 2015. That adds up to an insignificant consumption growth of 0.3% per year.
Mexico’s economic growth rate averaged 0.6% during 2014 and 2015. It is far less clear whether one should factor this in: after all, if people drink more sugary drinks because they get richer, they’re not going to get any thinner. But let’s humour the tax collectors, and adjust for economic growth: this results in a 159 litre per capita consumption by 2015, which represents an equally insignificant decline of 0.3% per year.
By seasonal changes, the NIPH presumably refers to the fact that 2015 was warmer than 2014. If so, it couldn’t have been by much more than 0.2°C, according to climate records. That this would have negated the claimed 6% decline in SSB consumption seems highly implausible.
It points out that prior to the introduction of the tax, sales of SSBs were on a decreasing trend, and that by not taking this into account, one risks attributing changes to this decreasing trend rather than the tax. This is simply absurd. The opposite is true: ignoring a pre-existing trend risks attributing a subsequent change to a policy intervention, rather than to the trend, which might have continued with or without that intervention.
Despite the statistical models on which supporters of the tax rely, and the claims of the Mexican NIPH, the empirical evidence shows that a tax on SSBs at best has a trivial effect on consumption. That the models show a decline while the sales data do not also suggests models do not even do a good job of approximating the past, let alone predicting the future.
Mexico’s NIPH says that marketing, promotions, new product offerings and other industry marketing strategies can lead to an increase in consumption of up to 20%. It did not specify which new marketing efforts, if any, were undertaken, and whether or not they did affect consumption. However, if industry marketing can overwhelm the claimed effect of a tax on sugary beverages by a factor of more than three, then the case for such a tax is fatally undermined.
According to Mexico’s National Beverage and Carbonated Drink Producers’ Association, the tax on SSBs has cost the industry 1,700 jobs. Mexico’s parliament voted overwhelmingly to slash the tax last year. According to a 2016 KPMG report on the prospect of a South African sugar tax, the tax had a disproportionate effect on lower-income households, and a smaller effect on obese individuals, thus missing the target market.
The Treasury policy paper mentions Ireland, which had a tax on soft drinks from 1916 to 1992. However, since the goal of this tax was to raise revenue, and it was replaced by ordinary value-added tax, this hardly makes the case that a sugar tax is anything more than a revenue grab.
Treasury notes that France introduced a levy on sugar-sweetened beverages in 2012, which led to a rise in price of 5% in 2012 and 3.1% in 2013, while demand for the products predictably declined by 3.3% and 3.4% in the respective years. How reliable these numbers are is anyone’s guess, since they are not drawn from peer-reviewed science conducted in France, but from a policy brief published by a British activist organisation that seeks “progressive change” on food policy.
What neither the lobby group or the Treasury paper reports, however, is whether the claimed decline in sugary drink consumption had any effect on individual body mass index, national obesity rates, incidence of obesity-related diseases, or public health expenditures. All of these would have to show improvement to support the idea that a sugar tax is justified. It would be astonishing if such a small decline in consumption of a product that accounts for only a small fraction of total energy intake had any significant impact on these numbers at all.
The Treasury paper cites Norway as another example. It introduced an excise duty on sugar-sweetened beverages in 1981, but notes that by the early 2000s, this had not had the desired effect. Only when taxes were increased and reinforced by complementary measures did research show a decline in sweetened drink consumption among the youth. It may very well be that the complementary measures, rather than the tax, were the catalysts driving the change. Again, no mention is made of actual health impacts resulting from lower consumption of taxed products, which leads one to suspect that there weren’t any.
For Hungary, the Treasury’s own data is contradictory. In the main body of its paper, it suggests tax reduced the consumption of affected foods by 25% to 30%. In the appendix with details, however, it turns out consumption of sugar-sweetened beverages declined by only 16.2% over three years, amounting to 6% per annum. Compounding the problem, there was evidence of a rise in demand for substitute products with similar nutritional characteristics as taxed products.
Even if none of the reported decline is offset by alternatives, a 16.2% reduction in the consumption of sugary drinks, which in South Africa account for only 3% or 4% of daily energy intake, means a daily calorie consumption reduction of about 0.2% per year, which is hardly a stunning success. Moreover, according to KPMG, the unintended consequences in Hungary were that government revenues were less than expected, jobs were lost and low-income earners were affected negatively.
Finally, Treasury notes the case of Denmark, which instituted a tax on sugary drinks in the 1930s, but recently abolished it, along with a tax on fatty foods that was barely a year old. According to KPMG, the tax resulted in hoarding, cross-border smuggling, higher calorie intake, an unacceptably high administrative burden and reduced country competitiveness.
Alain Beaumont, secretary-general of the European soft drinks association Unesda, told the media: “Soft drinks taxes are on the wane and being voted down by governments and parliaments across Europe. They have not proven to achieve any public health objectives and they destroy jobs and economic value.”
The KPMG report concludes: “While various studies provide evidence of a reduction in obesity levels following a tax on SSBs, the reduction in obesity is often not large, with the limitations of such studies often calling the results into question.”
The author of that report, KPMG’s chief economist Lullu Krugel, adds: “What is ‘good’ regulation? That is not an easy question to answer, but regulation that has a clear focus of what it needs to achieve, where the benefits outweigh the costs and where potential unintended consequences are anticipated and addressed as soon as possible, certainly goes a long way towards being considered ‘good’.”
Since a tax on sugary drinks will not have the intended effect of significantly reducing obesity or obesity-related diseases, there are no benefits to outweigh costs. There is absolutely a public health justification for encouraging overweight and obese people to limit their consumption of sugar, but there is no such justification for a tax on sugar-sweetened beverages. It simply will not work, and there is little evidence to suggest otherwise.
Treasury would be more honest about its intentions if it raised VAT by a point or two, but that wouldn’t give it the false excuse that it cares about the poor. DM
Note: This column, like last week’s piece, relies on research commissioned by the South African Institute for Race Relations (IRR) after my first column on this subject was published. This research was used to inform the IRR’s submission to Treasury on its policy paper on an SSB tax. The research and opinions expressed in this column are entirely my own.