Limitations of the Sugar Tax: A Selected Literature Review

There’s no disputing excessive consumption of most food and drink may have long-term negative health impacts including obesity, cardiovascular disease and diabetes. Particularly, obesity is prevalent in many upper- to middle-income economies including South Africa by the consumption of nutrient-poor foods. This has created a huge public health burden around the world.

It’s in this context the tax on sugar-sweetened beverages (SSB) was proposed, which South Africa will introduce from April 1 this year.  The Rates and Monetary Amounts and Amendments of Revenue Laws Act passed approved on 5 December 2017 set the levy at 2.1 cents per gram content that exceeds four grams per 100ml. The first four grams is levy free. A can of Coca Cola will increase by about 11%.

A 2014 academic paper, which reportedly proposed the sugar tax for South Africa, said a 20% tax is “projected to reduce obesity in men by 3.8% and 2.4% in women. The number of obese adults would reduce by 220 000”.  Note its title says it’s a “mathematical model” and its findings are “predicted”. The proposed levy of 20% was double the eventual tax.

Pro-tax advocates applaud it as a significant step in the fight against non-communicable diseases and a “victory for public health” (Tracey Malawana of the Healthy Living Alliance).  However, as much as I welcome efforts to improve public health, I’m sceptical of the claimed long-term benefits that the tax will materially reduce consumption of SSBs and its related positive health impacts especially obesity.

There’s a large amount of literature about the effectiveness of using fiscal policy to change consumer behaviour in order to reduce their consumption of unhealthy goods like sugary drinks, alcohol and tobacco, and I‘ve just begun to scratch the surface.  I’ve come across many articles that while they support the motivation of these policies, caution along the lines like in this paper, Cost-effectiveness of Fiscal Policies to prevent Obesity by Marj Moodie et al (2013):

“To date, the literature on the cost-effectiveness of fiscal measures for obesity prevention is very limited, and more well-designed economic evaluation studies are required. Uncertainty in the effectiveness of fiscal policies for obesity prevention needs to be addressed first.”

But in many other articles and opinions including by the local pro-sugar tax lobby there’s nary a mention of these exceptions, provisos and uncertainties to their – for now – theoretical models and predictions. Instead, they’re adamant the tax shall work on consumption as predicted and shall have the beneficial long-term medical and health benefits despite evidence on the latter not being available, and they’re not averse to personalising the issue.  Healthy Living Alliance’s Malawana on the recent announcement: “We applaud Members of Parliament for putting the health of millions of South Africans before the narrow interests [emphasis added] of the beverage and sugar industries.”

Excise taxes on alcohol and tobacco are used as examples of the efficacy of taxing consumption. In fact, the World Health Organisation and World Bank support tobacco taxation as a “win-win policy measure to achieve public health goals by increasing prices, reducing smoking, and preventing initiation among youth”.  However, referring to the 2015 WHO Report on the Global Tobacco Epidemic, it says only 33 countries impose the recommended tax of 75% of the retail price of a pack of cigarettes that will have an impact on consumption. 

Smoking prevalence has declined in South Africa from 32% in 1994 to 16.4% in 2011 (P Reddy et al, 2015; note other sources say 20% in 2014.)  According to World Bank data, worldwide tobacco prevalence has declined from 44% in 2000 to 35% in 2015 of adult males (SA: 35% to 31%). This has been attributed to tobacco control policies, specifically, excise tax (for SA in 2017 it was 52%, or R14.30 for a pack of 20).  

Clearly, SA’s tobacco control policies have made an impact on smoking but it’s arguable the decline is entirely due to the tax.  Reddy et al observe, “Among current smokers, 29.3% had been advised to quit smoking by a healthcare provider during the preceding year, 81.4% had noticed health warnings on tobacco packages, and 49.9% reported that the warning labels had led them to consider quitting. Conclusion: a large proportion of adult South Africans continue to use tobacco”.  But “despite initial successes [of policy interventions including “hikes in excise duty”], there have been increases in tobacco use between 2008 and 2011 among SA youth, particularly girls.”

Among the key findings of the WHO Report on The Global Tobacco Epidemic is taxation accounts for only 10% of tobacco control policies for 2014 (Executive Summary, page 3).  Their graph reproduced below shows that at 55%, mass media campaigns accounts for the bulk of inventions. Mentioning “notable progress in global tobacco control”, they further note, “Only one in 10 of the world’s people live in the 33 countries that levy taxes of more than 75% of the cigarette retail price, making it the least-implemented MPOWER measure and the one with least improvement since 2007. More than 80% of countries have no tobacco taxation in place at the highest level of achievement despite clear evidence that increasing taxes to a sufficiently high level [emphasis added; 75% of the retail price, although as a pro-tax economist pointed out to me, lower rates are still effective] is an extremely effective intervention”. 

But how does the taxation lobby explain that despite more than 80% of countries having no taxation at the highest level and taxation accounting for only 10% of the policy measures, worldwide the prevalence of smoking has declined?  The graph partially explains it: taxation is not the complete and only remedy.

South Africa has long had an excise tax on alcohol.  “Besides revenue raising objectives, the rationale for excise taxes on alcoholic beverages is to reflect their harmful external costs”, i.e., “social and health concerns of alcohol abuse” (Treasury, 2014, A Review of the Taxation of Alcoholic Beverages in South Africa). For 2017 the excise tax burdens for wine, beer and spirits are 11%, 23% and 36% of the weighted average retail price respectively. 

However, World Bank data show South Africa’s per capita consumption increased from 5 litres in the mid-1960s, to 11.5 in 2015 (11.0 in 2014).  It’s almost double the Sub-Saharan African region and world average of 6.3 litres, and is placed among the top-20 biggest drinking countries, tied with Poland at 19 out of 194 countries.  Again, it’s hard to accept the argument the tax, at its present levels anyway, is helping reduce consumption. So, confirming Treasury’s stance, it exists almost entirely to raise revenue because its other rationale – to mitigate the social and health impacts of which South Africa is unfortunately a world beater – is not being met.

My purpose here is not to dispute the “consensus in literature” taxation does not change and reduce bad consumption habits, but question the sugar tax lobby’s bald assertion that on its own (and set at its proposed moderate rate) and without simultaneous, collateral interventions it will lead to “healthier lifestyles”. My reservations, especially applied to the South African situation, are essentially:

1. The tax must be significant

The tax must be significant otherwise it will have little effect on consumption. Or to put it another way, a moderate tax will have negligible to moderate effect on consumption and ultimately, obesity and health.

Kelly Bronwell et al (2009): “Currently, 33 [American] states have sales taxes on soft drinks (mean rate, 5.2%) [note not excise taxes; for SSBs, SA will have both], but the taxes are too small to affect consumption and the revenues are not earmarked for programs related to health”. While SA will have a tax of about 11%, it’s still below the 20% deemed effective.

Note Bronwell et al support taxes in terms of providing revenue for health and to address external costs related to consumption but still qualify that with “as with any public health intervention, the precise effect of a tax cannot be known until it is implemented and studied”.  In other words, the positive consumption and health outcomes predicted by the sugar tax lobby, while compelling – we want to believe, as Agent Fox Mulder says – are unclear and unproven.

2. The substitution effect is unknown

The substitution effect (substituting something else for sugary drinks) is unknown, hasn’t being considered or adequately studied, especially in the South African context. 

Nikoloas Maniadakis et al (2013) state, “The effectiveness of a taxation policy to curb obesity is doubtful and available evidence in most studies is not very straightforward due to the multiple complexities in consumer behaviour and the underlying substitution effects”.  Also, Cornelsen et al (2015), Why Fat Taxes Won’t Make Us Thin: “Although these taxes are possibly efficient in reducing by a small amount [emphases added] the consumption of targeted products if the tax is fully transmitted to the consumer, there is too little available evidence on what will be consumed instead and whether these food substitutions undermine the hoped-for health benefits of the tax.”

3. Sugar tax revenue will not go toward health

The revenue will not go toward addressing health problems related to excessive consumption of sugary drinks.  At the risk of “misrepresenting and lying” about the research and literature, advocates are more concerned about imposing the tax despite its unproven long-term effects than insisting to government that as a deal-breaker for their support the tax must be directed to health. 

As the plastic shopping bags levy prove, despite starting out with the best of intentions, but eventually faltering under official ennui, revenue of R1.1 billion went to government and not the environment, recycling and job creation as intended by the then environmental minister Valli Moosa.  Presently, only about 20% of purchased bags are recycled. It’s fair to say the sugar tax revenue will go Treasury and among other things help fill the R50 billion budget hole and R20 billion free fees fiasco.

4. Models are not proven empirical evidence

The 2014 academic paper that proposed the tax for South Africa clearly stated it’s a mathematical model rather than empirically based.  Absolute certainty is not required for the implementation of policy but there must be a reasonable assurance the outcomes will be as intended otherwise it’s just guessing. The plastic bag levy, wage subsidy and other failed policy interventions (it’s likely the minimum wage will go the way of the others) show the country has a very bad record – F for “fail” – when it comes to policy outcomes. There’s no reason to expect, and the sugar tax’s proposers have not addressed that issue, that this measure will be different.

5. Future health benefits speculative

The lobby’s insistence about future health benefits is misplaced or at worst disingenuous because conclusive evidence linking causality from reduced consumption to long-term, sustainable health and obesity benefits is not yet available or known.  Maniadakis et al (ibid), “There is need to investigate in-depth the potential underlying mechanisms and the relationship between price-increase policies, obesity, and public health outcomes.”  Cornelsen et al (ibid):  “We know very little on how the food supply side will respond and what overall impact this will have. Without a proper appreciation of the potential indirect impacts we do not know the overall impact of taxes foods on unhealthy foods and beverages and further that there is a very real possibility that they may not be beneficial for health after all.”

6. Confirmation bias

The lobby, at least the locally sourced articles I read, shows confirmation bias by typically citing papers and research that confirms their position.

Mexico, which implemented the SSB tax in 2013, is often mentioned in studies. Mexico is a useful indicator but not absolute predictor of how South Africans consumers will react because of a different, or even vastly different, social, cultural, economic and health environment. But putting that quibble aside, studies liked this one by M Arantxa Colchero et al (2016) showed that during the first year of the tax the average volume purchased was 6% lower than without the tax.  But they note:

“It will be critical to continue monitoring purchases to note whether the trend continues or stabilizes; consumers substitute cheaper brands or untaxed foods and beverages for the taxed ones, or adjustments occur in the longer term. This will allow for an understanding of the long term effects of taxes on both sugar sweetened beverages and non-essential energy-dense food on purchases, diets, and ultimately health outcomes.”

Incidentally, one study in Mexico followed 11 218 women over 2 years, and compared to those who didn’t consume SSBs, found “moderate changes of 1-serving-per-day increase in consumption of sugar-sweetened soda was associated with a 1-kilogram weight gain over a 2-year period”.  In a similar study in the US among the white population there was only a 0.4-kg increase over a 4-year period for the [one-serving-per-day] increase in SSBs”. (Also see here.)

What is clear from this and other articles cited is that despite cautious support (except Cornelsen) for the principle of the sugar tax in order to raise revenue for health and reduce SSBs’ negative health externalities, many researchers are uncertain of the long-term outcomes, say data is absent and it must still be observed and studied.  Researchers like Emily Wolff and Michael Dansinger (2008) stated higher consumption of SSBs was associated with a weight gain and an increased risk of developing type-2 diabetes among women but “clinical trials would be necessary to conclusively demonstrate a causal link between sugar-sweetened soft drink consumption and weight gain. The magnitude of the weight gain and adverse health effects caused by soft drinks are poorly understood due to a paucity of clinical trial data”.  Vasanti S Malik et al confirm the association between SSB consumption and weight gain and obesity but “further research ... is needed to provide more convergence in the data”.

Compare their reticence about the possible future, long-term effects to particularly local researchers who with unshakeable conviction claim the sugar tax “moves South Africa one step closer to a healthier lifestyle” (Karen Hofman and Aviva Tugendhaft, 2017).

So, the tax will be implemented on April 1 and we go from the modelling to observation stage. We will almost immediately see if it makes a difference to consumption, but it will take years – decades even – to fully understand the health impact, if any.  If it makes even a moderate difference, well and good, but if it doesn’t its proposers in the academic and health community, government (except for the billions in revenue) and middle-class will not be affected but, as always, the poor shall be and disproportionately so. And if it doesn’t work, government will not abolish the tax. (I wonder if the pro-tax lobby researched the economic impact on the country in that eventuality as part of their risk analysis.)

My literature review was limited and only to open-source material. I deliberately cited papers that expressed reservations, even if reasonably saying the sugar tax may work if applied in specific ways but more data is needed to fully understand its impact. I did so to prove there’s no unanimity within the academic and research community. Those who disagree are not “charlatans”, “hacks” and the “ignorant masquerading as intelligent”, and by expressing scepticism, are not kowtowing to de rigueur political correctness.

However, the lobby has not explained, and I have put the question out there, why they didn’t propose a law that set the amount of sugar in SSBs as South Africa has done for salt in processed foods.  This can easily be done at source – the manufacturer – with stiff fines for those who don’t comply.  But that’s too easy and there’s no incentive because no taxes can be collected and politically, the SSB industry is too easy a mark to forego. 

Disclosure: I have no connection to the SSB industry, private sector or research groups.  I am, though, a fully independent “armchair activist”.