Efficacy of Fat Taxes and Thin Subsidies

Science News reported the following results from a recent study:

Taxes on soft drinks and foods high in saturated fats and subsidies for fruit and vegetables could lead to beneficial dietary changes and potentially improve health, according to a study by experts from New Zealand published in this week's PLOS Medicine.


The authors say: "Based on modelling studies, taxes on carbonated drinks and saturated fat and subsidies on fruits and vegetables are associated with beneficial dietary change, with the potential for improved health. "

My first reaction was "duh."  Clearly if you raise prices of (say sugar or fat) high enough, you will get people to eat less.  In fact, one way researchers often model a ban on a substance is by simulating what happens when the price gets high enough that no one buys the product.  

Thus, the key question isn't whether one can change consumption and nutrient intake with sufficiently high taxes or subsidies.  The better questions are by how much? and at what cost?  An even better question still: where is the market failure that would justify the tax or subsidy?  The answer to that last question is actually much less obvious than most public health professionals presume (see here or here).

On the former question of how much?, let's turn to the original study mentioned at the first of this post.  The study is actually a literature review, pulling together the findings of previously published papers (including one that I co-authored).  Below is a graph showing some of the key results from different studies simulating how much change in consumption (or energy intake) would occur from a change in the price of a good. Pay attention to the scale of the vertical axis.  My take (see the middle chart) is that it would take very large price changes to get energy consumption to change by much (a 20-40% increase in price results in a 0.2-0.4% reduction in calories consumed).  

Stated differently, these sorts of policies are likely very costly in achieving the desired health outcomes.  Moreover, we must ask why - if these health changes are really so inexpensive and beneficial - people are not already voluntarily making them?