9

A recent analysis of the health-related harms associated with red and processed meat has hit the news with headlines like this (from the Guardian):

Taxing red meat would save many lives, research shows

The cost of bacon and sausages would double if the harm they cause to people’s health was taken into account

One of the authors of the research has explained it thusly in The Conversation:

The cancer agency of the World Health Organisation (WHO) classifies the consumption of red meat, which includes beef, lamb, and pork, as carcinogenic – or having the potential to cause cancer if eaten in processed form. This includes hot dogs (frankfurters), ham, sausages, corned beef, and beef jerky – as well as canned meat and meat-based preparations and sauces...

...Like taxes on other products that can harm health, a health tax on red and processed meat could encourage consumers to make healthier choices. And our new research, which looks at the benefits of a health tax on red and processed meat has found that such as tax could prevent more than 220,000 deaths and save over US$40 billion globally in healthcare costs every year.

But there are two sources of uncertainly in the underlying data that deserve some skeptical analysis. How big is the true effect of meat eating on cancer and other diseases? And how certain are we that increased tax has the effect climbed on actual consumption?

For example, David Spiegelhalter, an expert on risk analysis and risk communication was not going to give up bacon sandwiches on the basis of the original harm estimates as he explains in this video.

So would higher taxes on meat save large numbers of lives?

(note this breaks down into two questions: can we trust the risk estimates? and can we believe the impacts of tax on consumption levels?)

Community
  • 1
matt_black
  • 56,186
  • 16
  • 175
  • 373
  • Relevant question: https://skeptics.stackexchange.com/questions/38533/is-eating-processed-meat-as-bad-as-smoking-for-causing-cancer?rq=1 – Jon.G Nov 08 '18 at 09:15
  • Jokes, political opinions, and half-answers deleted from comments. – Oddthinking Nov 08 '18 at 23:53
  • 1
    As you say, there are two questions here. I would be tempted to break it into two posts. Does eating red & processed meat cause increase mortality to the degree this study claims? (The related question partially addresses that.) Would [Pigovian taxes](https://en.wikipedia.org/wiki/Pigovian_tax) on red & processed meat reduce consumption as much as claimed? – Oddthinking Nov 09 '18 at 00:24
  • This second subquestion is likely to suffer the same challenges as other economic model questions here. It is likely the study specifies the simplifying assumptions it makes, making the study "correct" (sans any spreadsheet errors!) whether the assumptions are actually true or not. We often seem to get into a knot trying to answer these questions. – Oddthinking Nov 09 '18 at 00:25
  • @oddthinking Fair point but there are still useful ways to approach them. First are the assumptions in their model as solid as they claim and have they quoted the error bars? Second is there any empirical test of anything similar (eg how did models of sugar taxes do in predicting their effect on consumption). I previous price elasticity models didn't work well, why trust this one (or what trust it without any error bars)? – matt_black Nov 09 '18 at 09:52
  • The study seems pretty thorough, although confidence intervals are pretty large, due to high uncertainty. Because it looks at the impact globally, numbers tend to be very large. – redleo85 Nov 09 '18 at 16:09
  • In my country they started [taxing lemonades a few years ago](http://www.brusselstimes.com/belgium/politics/4954/tax-on-sweetened-drinks-belgian-government-under-fire-from-coca-cola). The result: people now go and buy their sodas in bulk just over the border once a month, and the income for the state has significantly lessened in that market, so much that they abandoned extending it to other sugar-holding products as they originally intended. – Codosaur Nov 10 '18 at 11:29

0 Answers0