Katherine Rich: You can't tax people slim

1 April 2016

For more than a year now, public health activists around the world have been pointing to the Mexican soda tax as proof that taxes on sugar in food and drink work.

As recently as February, the British Medical Journal was quoting US and Mexican researchers as saying the 10 per cent tax had cut sales by 12 per cent in the first year, with the average person purchasing 4.2 fewer litres of sugar-sweetened drinks in that time. The BMJ article was based on data from Nielsen home panels in 2014.

But the most up-to-date Nielsen sales data, to December 2015 (home panels plus national scan data), compiled for the NZ Food and Grocery Council shows the opposite is true – that the Mexican tax has been a failure. It shows the tax has made almost no change to sales volumes in the two years since it was implemented in January 2014. It did drop early in 2014 but has since returned almost to pre-tax levels. New Zealand has seen a significantly higher decline in the sales of carbonated drinks during the same period without any tax at all.

As I mention, the data was compiled by respected research company Nielsen, and as such it’s the most accurate proxy for volumes of soda consumed in Mexico, covering 65% of sales. It’s the next best thing to getting manufacturing data from all the beverage manufacturers.

FGC asked New Zealand economist Dr Brent Wheeler to analyse the data, and he confirms my reading of the Nielsen numbers: “The clustering of sales even measured on a monthly basis is clear. There is, for example, no consistent or even noticeable change following January 2014 upon the introduction of the tax.”

The data shows that unit sales of low sugar grew 5% in 2014 then dropped 2% in 2015, while regular sugar sales dropped 3% in unit sales then gained 3%. In other words, back to pre-tax levels. Prices are now higher since the tax was imposed, while sales volumes are back to where they were. In 2013, there were 10.75 billion litres of sugar-sweetened sodas sold across Mexico. The year after the tax was introduced, that dipped to 10.39 billion litres, but a year later they had bounced back to 10.69 billion litres. That’s an infinitesimal 0.5 per cent drop on pre-tax sales volumes – certainly not sufficient to claim the tax as a success.

To put that into perspective, it’s a fraction of the more than 5 per cent decline in consumption of carbonated drinks seen in the New Zealand market over the past 12 months – which has occurred without any tax at all!

Aside from the irrefutable evidence put forward by this hard sales data, there are three other points to consider around the claims put forward by the pro-tax brigade: the economic conditions in Mexico, the lack of a price signal, and the reliability of their research.

Economic conditions

Dr Wheeler says the net change in sugar-sweetened sodas “was not material”, particularly when Mexico’s economic performance over the period, and consumer spending generally, are factored in. From consumer spending and retail sales data, it’s clear that consumers stopped spending on a lot of things and retailers stopped making sales of all goods, not just sodas as a single item. “Such changes [to sales] … are so well aligned with economic performance over the period in question as to make isolation of any change (had there been one) attributable to tax close to impossible.” There were a lot of challenging things happening at the time in the Mexican economy, but claiming that sales dipped just because of a tax on one item overlooks that.

Price signal

One aspect that surprised me when I reviewed the average cost of sugar-sweetened sodas versus low-sugar ones was that the tax did not introduce a price signal. Remember that the theory behind the tax was to send a signal to consumers, by way of higher prices of sugar-sweetened drinks, to choose low-sugar options. But this did not occur. The average price of a sugar-sweetened soda in December 2015 was 11.4 pesos and the average price of a low-sugar option was 15.05 pesos, so how has the tax sent any signal to consumers to act differently? The tax is now not noticeable to consumers, which is why sales have returned to pre-tax volumes.

Research reliability

The US and Mexico researchers used data collected from more than 6,200 households across 53 large cities. FGC’s was actual sales data. Dr Wheeler says: “The chief benefit of the [sales] data is that it reflects behaviour – not the poorer quality data inevitably resulting from all surveys in which people’s behaviour typically departs from what they report. Genuine economic behaviour can only be derived from observations of what people do rather than what they say.”

From these figures, it’s clear to me that the claim that the Mexican tax experience proves the tax works is more wishful thinking than anything. Like a lot of food issue debates, it’s important always to return to what the real sales data says.

To reduce obesity, governments, industry, communities, families and individuals need to work together to encourage people to make healthier choices. But one thing is for sure – we can’t tax people into healthy eating without causing economic hardship to those who can least afford it.

It’s time for those championing soda taxes as a magical solution to reducing obesity to accept that governments cannot tax people slim.

(as published in FoodNZ)