A tax on sugar won’t work, as the shipwreck of the Danish ‘fat tax’ shows

Bad ideas die hard in the world of ‘public health’, especially when the government can make money out of them. The shipwreck of Denmark’s notorious ‘fat tax’ should have served as a lighthouse for politicians for a generation.

The negative consequences of the Danish experiment could not have been clearer. After levying a tax on all food products which contained more than 2.3 per cent saturated fat, Danes fled over the borders to Germany and Sweden to stock up on their butter and bacon.

Those who did not shop abroad switched to cheaper supermarkets and bought budget brands of the same products. The tax required an immense bureaucracy to implement and it placed a heavy administrative burden not only on food companies in Denmark but in countries where food was exported to Denmark. Higher taxes inevitably led to inflation and raised the cost of living. Predictably enough, it was the poor who were clobbered hardest. By the time the fat tax was repealed 15 months later, it had become a byword for well-meaning government folly.

Researchers have struggled to find any positive effect on health. There were, at best, single figure percentage declines in the sale of high-fat products such as butter, but 80 per cent of Danes said the tax didn’t make them change their shopping habits at all. If it had any effect on obesity, it was far too small to measure.

The only thing the fat tax did successfully was raise money for the government. Indeed, it was so ineffective in deterring Danes from buying fatty products that it raised much more money than had been expected. Since this money came disproportionately from the pockets of those on low incomes, the fat tax managed to unite hardline socialists and classical liberals in opposition to it. It went unlamented into the dustbin of history in January 2013, along with plans for a sugar tax.

That, you might think, would be the end of it, but you would be wrong. George Freeman, the minister for life sciences (no, me neither) has told an audience at the Hay Festival that he is in favour of a sugar tax. ‘I don’t think heavy-handed legislation is the way to go,’ he said, as he prepared to announce his support for heavy-handed legislation. ‘But I think that where there is a commercial product which confers costs on all of us as a society, as in sugar, and where we can clearly show that the use of that leads to huge pressures on social costs, then we could be looking at recouping some of that through taxation. Companies should know that if you insist on selling those products, we will tax them.’

If Mr Freeman was really concerned about the burden on taxpayers, he would not be supporting higher taxes. If there is one thing guaranteed to confer ‘costs on all of us a society’, it is a tax on a basic food ingredient. Besides, there is already a de facto tax on most sugary products with fizzy drinks, chocolate, cakes and biscuits being subject to VAT while fruit and vegetables are zero-rated.

Sin taxes on food and soft drinks are extraordinarily inefficient and ineffective, which is why they always and everywhere operate as stealth taxes. Food is such a basic staple of the household budget that people either find ways to get around them – by switching from Waitrose to Aldi, for example – or they simply pay up.

Demand for tasty food is extremely inelastic and this makes it a very tempting target for taxation. The problem is that such taxes are also incredibly unpopular, as the Danish experience showed. No country has ever reduced obesity through taxation and the scale of taxation that would be required to make even a dent on obesity rates would be so vast that any government that attempted it would not be in government for very long.


  • DiarmuidOBrien

    The swingeing alcohol tax in Sweden has worked to decrease overall consumption, but over a longer period than 15 months. The Danish experiment may have been flawed but 15 months was hardly a long enough time for a proper analysis of results. I think any food tax is ultimately likely to be an inefficient instrument anywhere it’s implemented but the Danish experiment surely provokes as many questions as answers?