Picture: ISTOCK
Picture: ISTOCK

THE policy paper issued by the Treasury proposing an excise tax on sugar-sweetened beverages, equivalent to 20%, on Coca-Cola, which contains 35g of sugar per 330ml serving, claims that such a tax will have a significant effect on obesity rates. However, a detailed look at the science shows this is not true.

This emerges from a submission by the South African Institute for Race Relations on the Treasury policy paper (to which I contributed substantial research).

In support of its contention, the Treasury relies on a paper that attempts to quantify the effect of a 20% tax on such drinks. It was written by researchers including Mercy Manyema and Karen Hofman of Wits University with the explicit aim to "enable the [Department of Health] to consider [a sugar-sweetened beverage tax] as a lever to prevent and reduce the burden of disease resulting from obesity-related [noncommunicable diseases]".

The Manyema paper claims to "provide evidence on the potential effect of fiscal policy on sugar-sweetened beverage consumption and obesity in SA", but it provides no such thing. It consists of a mathematical model that relies on a number of assumptions linking an excise tax to an increase in price, a corresponding decrease in sugar-sweetened beverage consumption, a consequent reduction in body mass index and a lower national obesity rate. Therefore, it provides not evidence but theoretical speculation about a causality chain that is only as strong as its weakest link.

A review of 880 studies by the Institute for Economic Affairs points out the problem with such models: "There is a striking contrast between theoretical studies, which generally predict that such taxes ‘work’, and studies of hard data in places that have actually implemented them, which generally show the opposite. Lacking real world evidence that sugar taxes are effective as health measures, campaigners continue to cite findings from crude economic models that do not adequately account for the ability of consumers to choose cheaper or discounted brands, to shop at cheaper shops, or to switch to alternative high-calorie food and drink products."

International experience shows a sugar tax has less of an effect on sugar-sweetened beverage consumption than expected, and no effect on obesity rates at all. In a study by the McKinsey Global Institute, which notes that there are more than 100 variables that affect obesity outcomes, a tax on sugar proved to be the least effective of 44 obesity-related interventions. The Treasury policy paper does point out that it is the most "cost-effective" intervention but that is only true in a trivial sense, because the cost of levying an excise tax is negligible.

The Manyema paper does not adequately consider the potential for substitution, such as switching to coffee and tea, which can be sweetened to taste at hardly any cost, or fruit juice. The latter is an interesting case. The Treasury has excluded 100% pure fruit juice from the sugar-sweetened beverage tax, reinforcing the perception that it is healthier than sweetened soft drinks. The surprising reality is that even unsweetened fruit juice contains as much or more sugar than typical soft drinks.

Assuming, for the sake of argument, that all the Manyema paper’s assumptions hold true, it still fails to justify a tax on sugar-sweetened beverages. It says an increase in daily energy intake of 94kJ accounts for a 1kg gain in weight, and assumes that the opposite is also true.

However, it concludes that the proposed tax will only reduce daily energy intake by 36kJ on average, which represents a tiny 0.34% of the recommended intake of an adult male, and only 0.55% of that of a child. If 94kJ accounts for 1kg of body weight, then a cut of 36kJ in daily energy intake would account for 383g of weight loss — too small to detect on an average bathroom scale.

The only statistic that the Treasury policy paper lifts from the Manyema study is that the obesity rate is predicted to decline 3.8% in males and 2.4% in females, but that deceptively overstates the real effect.

Using obesity rates from the South African National Health and Nutrition Examination Survey study (SANHANES-1), this would only marginally reduce obesity rates, from 10.6% in males to 10.1%, and from 39.2% in females to 38.2%.

The number of people who would no longer be classified as obese, according to the Manyema study, would decline by between 32,610 and 412,323 people. This extraordinarily wide range indicates very high statistical uncertainty. Even so, the average of 222,669 fewer obese people accounts for only 0.4% of the population, and those would be the people for whom a reduction of 36kJ per day, or 383g in body weight, would be enough to take them from just above the "obese" line to just below it. In context, the predicted decline in absolute obesity numbers conceals how limited the effect is really predicted to be.

The SANHANES-1 study also finds that only 19.7% of South Africans have a high sugar intake to begin with. On average, South Africans do not consume too much sugar, whether in the form of sugar-sweetened beverages or otherwise. Moreover, it finds that although 25% of South Africans are classified as obese, only 15.7% of the population reports being unhappy with their weight.

In the real world, the effect of a tax on sugar-sweetened beverages would be negligible, even if all the assumptions and predictions of the Manyema paper’s mathematical model hold true, which is far from a given. In public health terms, there is no justification whatsoever to impose an onerous tax on an arbitrarily selected product category that will affect not only major producers, but also everyone in the value chain from sugar farmers to small-scale retailers, who derive as much as 20% of their income from sugar-sweetened beverage sales.

One is left with only one possible conclusion: despite the Treasury’s protest that it’s not about a potential R10.5bn in annual revenue, that can be the only possible motive. The proposed tax on sugar-sweetened beverages is an unjustified tax grab that will have a negligible effect on public health, if any at all.

Cronje is CEO of the IRR.