Wednesday 2 September 2009

Of trolley cars and health redistributions

Chris Worthington has an interesting piece on the Kiwi folic acid debate. (HT: Drinkwater) The outgoing Labour government apparently committed New Zealand to fortification of bread and flour with folic acid through a trade agreement with Australia; National, in response to public opposition to the move, fueled partially by overstated worries about links to prostate cancer, backed down.

Worthington likens the problem to the standard trolley car one. Is adding folic acid to bread more like pulling a lever to kill one where five would otherwise die, or is it more like pushing a bystander into the path of the trolley car to achieve the same end? Experiments suggest more folks are comfortable with the former than with the latter even though both cases have you being responsible for killing one innocent person to save five. I'd tend to think more the former, just because of the remoteness of consequences; nevertheless, the policy was opposed.

More interestingly, Worthington notes:
If the bread example seems too uncommon to warrant much thought, an earlier academic paper suggested a broader application. High-income groups in New Zealand have a lower mortality risk than low-income groups, although there are diminishing marginal returns to extra income. If the relationship is causal, redistributing money from rich to poor would also redistribute some of that mortality risk, producing a longer-lived society (on average).

Of course, the direct result of that policy intervention would be that some high-income people would die earlier than they would otherwise do so.

So, are such welfare interventions more or less palatable when phrased in terms of life and death as opposed to dollar signs?
Worthington here is almost surely talking about the paper by Blakely and Wilson, two of the public health folks at U Otago at Wellinton's School of Medicine. I'd previously discussed that paper during a guest stint at EconLog where I noted two big problems.

First, they ignore the deadweight costs of taxation, so it's very likely that their proposal for massive income transfers to increase average lifespan and reduce mortality inequalities would instead reduce average lifespan.

Second, they ignore that both health and income may be caused by an underlying variable, like IQ. If IQ determines both health and income, gains from their proposal are again overstated.

More fundamentally, though, I can't see any logical difference between a policy using forced income transfers to try and reduce mortality inequalities and a policy forcing healthy people to give kidneys to sick people. The risks of live kidney donation are very low and the policy would almost certainly save lives: I'd bet that it would be far more effective than Blakely and Wilson's income redistribution scheme. Would Blakely and Wilson find such a policy repugnant? If so, what usefully distinguishes stealing a man's income to aid someone else from stealing a man's kidney to do the same? Maybe the healthists have an easier time pulling the income lever than pushing somebody onto the operating table?


  1. Are you saying that policies should be based on whether they pass the repugnancy test?

  2. Certainly not. But I have a tough time seeing huge differences between the two.

  3. Well, I find the kidney idea repugnant, but that emotion isn't what I'd want to judge a policy on. On that basis, arguments can easily be made against interracial marriage, etc., etc.

  4. I'd advocate against the forced kidney transfer policy more on the basis of rights and liberty than on consequentialism. Yes, you could build a consequentialist case against it by talking about how folks will have reduced incentive to keep their kidneys in good nick, or maybe by talking about the disutility induced by the fear of a forced kidney removal, but the real issue is that my kidney is mine, not the state's.

  5. Hi Eric, thanks for the write-up. It was indeed the Blakely/Wilson paper I was referring to.

    I agree with your comments on their paper. The only thing I'd add is that the assumption of diminishing marginal health returns to income, while quite plausible, woulld be hard to estimate precisely (and their results are very sensitive to it). From memory, I doubt one could reject the possibility of constant returns in their sample.