Who to push in front of the trolley?
Fri 28 Aug 2009 by Infometrics Ltd.

I admit to only vaguely following the recent furore over the decision to supplement bread with folic acid.   The underlying principle behind such decisions seems straight-forward.   But economics can only take us so far when we start to examine the moral dimension.       

The benefits to folic acid supplementation arise from fewer cases of spina bifida.   The costs are the costs of changing bread-making procedures, and a possible enhanced risk of prostate cancer.   If the benefits exceed the costs, there is a strong argument for proceeding with supplementation.

Not being a medical specialist myself, I am (like most people) in a poor position to judge whether the estimates of costs and benefits are robust.   But let’s assume that it is true that adding folic acid to bread will lead to more prostate cancers.       

The heated debate over this issue implies that there is more at stake than just the precision of the estimated values.   Clearly a cursory cost-benefit analysis is ignoring some important dimension of this question.

The most obvious complaint is that the "winners" from this decision are not the same group as the "losers".   Now this is true of almost all real-world policy decisions.   And compensation of the losers is not always possible or feasible – redistribution via the welfare system being the obvious example.   In these situations, we must base our decisions on some sort of utilitarian calculus, whereby we try and maximise total welfare over the whole population.

Since we can’t directly quantify human well-being, we are forced to make a number of value judgements.   We normally assume that an extra dollar is worth more in terms of quality of life for a poor person than a rich person; we may also value improvements in health for children higher as opposed to adults.         

The distribution of gains and losses brings a complicated philosophical dimension to our cost-benefit analysis.   Further, I’d argue that when the distribution relates to physical health, as opposed to simply income, these questions become even thornier.   Is it right to force the exchange in health status between the potential victims of spina bifida case and prostate cancer, even when the cost-benefit analysis supports the transfer?

There is no right answer to this question.  In fact, laboratory experiments tend to suggest that these decisions are highly context-critical.

One famous example is the trolley problem: an out-of-control trolley car is running down a track that has five people tied to it; but there is a switch you may pull that will divert it on to a second track which only has one person tied to it.

Faced with this situation, most people can rationalise pulling the switch: killing one to save five.

But what about this alternative scenario: the train car is heading towards five people, but this time there is no switch.   Instead, there is a fat man nearby; throwing him in front of the trolley would derail it, again saving the original five through the sacrifice of one.

This version of the problem produces far fewer people willing to kill one person to save five.   Although the calculus remains the same, it is much harder to condone the deliberate homicide of an innocent bystander.

Analysis of these problems raises a host of questions about how we make moral judgements; the relevant one for our topic is what precisely drives the differing responses?   Common answers are the intent of the decision-maker, or their control (or lack thereof) over the situation.

Returning to bread supplementation, we are still none the wiser as to whether we can justify swapping a certain number of spina bifida cases for prostate cancer cases.   Should we view this as a neither/or situation (throwing the folic-acid switch in the bread-making machine), or is this an active intervention (throwing some men, prostates and all, on to the track)?

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?

The trolley car results offer a prediction for how one would answer that question.   If one views the current income distribution as arising from the "natural" state of society, one might regard active intervention into health outcomes as immoral.   If instead, one regards the income distribution as simply being a consequence of a broader nexus of policy, the trade-off in health outcomes may be perfectly acceptable.

Economists have created a number of clever ways for assigning values to costs and benefits, even on tricky questions relating to the value of life.   But unravelling the moral dimensions of such decisions remains elusive.

Related Articles