Singer on the Value of Lives

Peter Singer had a piece in The Guardian last week comparing the way that we value lives around the world.  He points out that when NATO accidentally kills Afghan civilians, it pays out compensation.  This never goes above about £5000 per death.  In comparison, NICE’s threshold for funding a treatment on the NHS is £20-30k per QALY.  But, he reminds us,

that sum is per QALY, not per life saved. So if we take the bottom end of this range, Nice recommends that the NHS pay up to four times as much to extend the life of a British citizen by just one year, as the MoD is prepared to pay in compensation for killing a child or young person. That young person could – even allowing for Afghanistan’s dismal life expectancy – expect to live another 40 reasonably good-quality years. That suggests an answer to the question with which I started: it takes about 4 x 40, or 160 Afghan lives, to be worth the same as one British life.

But that would not be the right answer, because £5,000 will buy much more in Afghanistan than it would buy in Britain – according to international price comparisons, perhaps four or five times as much. Let’s say five times. Even with that adjustment, it is going to take 32 Afghan lives to be worth the same as one British life.

At the same time, the US government’s compensation has paid on average $1.8m to the family of each victim of the destruction of the World Trade Center.

Singer’s claim is that NATO doesn’t take seriously the idea that all human lives are valuable.  But while it’s easy to get swept along by his claims (and they are powerful) I’m not sure that they stack up – especially in relation to the NHS; this is a shame, because the basic point he’s making is important.

When he compares the value of the lives lost in the US and Afghanistan, he may have some kind of point, because what we’re comparing here is like with like: straightforward compensation for the loss of a family member.  How should we put a price on that loss?  Well, I suppose that one thing we have to consider is the lost opportunity cost that a life represents.  Like it or not, the loss of a broker at the WTC may well cause, in brute cash terms, more damage than the loss of a trader in Mazar-i-Sharif.  On the other hand, of course, it’s unlikely that the family of the broker is going to end up destitute; the cash loss they suffer may be bigger, but what it means may well be a lot smaller.  I have no idea how one might actually go about making the assessment of the due compensation here – but that US compensation is higher than Afghan doesn’t seem necessarily unjust in its own terms.  Having said that, the gulf between the figures offered is somewhat astounding, and that, my gut tells me, is probably indefensible.  It seems politically daft, too: were compensation higher – and £5000 seems almost laughably low – the local Afghans may take more seriously the idea that NATO, though fallible, is on their side.

(Just for the sake of completeness, it isn’t an objection to this kind of comparison to say that we shouldn’t be trying to put a dollar value on lives: were that the case, no amount of compensation would be right, and so no compensation would not seem to be wrong.)

OK: but what about the amount of funding that NICE will approve for the NHS?  Here, I think Singer gets a bit unstuck.  The problem seems to be that he’s not comparing like with like: he’s comparing the value of a life with the value of a life year.  I’m not sure that you can do that: I have a vague suspicion that it’d be like trying to compare amplitude and frequency.  (I think that John Harris has made the same error, too, in his rejection of QALYs as a metric – or at least come close to it.)  But even if this hunch about comparing unlike things is incorrect, then there is one more important difference that Singer has missed – and I’m more sure of this.

Dead relatives stay dead, and the dead are not economically active.  This means that compensating their families is an absolute cost.  (I’m sure there’s a proper economists’ word to use here, but I don’t know what it is.)  Just as you can calculate the value of a dropped vase, so you can calculate the value of an exploded body.  But that’s not the calculation the NICE is making.  NICE’s calculation might involve a bit of thinking about the value of a life, or the value of life in specie, but there’s more to it than that.

This is because NICE is hoping that the people on whom it spends its money will continue to be economically active for a good number of years.  And because economic activity is good for the economy, then each extra QALY will go at least some way to paying for itself in a way that compensation for the dead won’t.  That is: Smith alive is much more valuable than Smith dead, and so the price of his continued life can be expected to be higher than the price of his death; and the longer he will be alive, the better.

Go back to the broken vase.  The Portland Vase, when dropped, would have had a certain economic value: let’s say, for the sake of simplicity, £10.  You drop it, you’re £10 down; if you’re the owner, you chalk it up to experience, and if you’re not, you pay the owner £10.  Everyone should be happy.  But now imagine that someone says that he can fix it, good as new, for £15.  It might actually be worth you paying that if you expect that at least 15 people would be willing to pay £1 each to see it, even though it’s inherent value is only £10.  Indeed, if you think that it has an inherent value of £10, then you might be willing to spend £15 even if only five people would pay a pound each to see it: it won’t have earned back the cost, but – well, inherent value is inherent, and doesn’t have to be earned.

(And, along these lines, compensation for deaths caused by failures of the NHS are not always all that high; this case concerns a payout not much higher than the guideline price of one extra QALY.  Compensation is also payable for non-fatal mistakes, as when, for example, people suffer brain damage during birth.  But a major factor here is the cost of future care – which the dead conspicuously don’t need.  Look at your holiday insurance cover, too: chances are, losing an eye or a limb will generate a higher payout than losing your life, because once you’ve lost your life… well, there’s not any continuing cost to that.)

That’s why a QALY gets funded more than an Afghan life: the dead will never repay any of what has been spent on them.  The living, even if all they can do is stay at home buying stuff online, will still be contributing to the economy by paying bills and VAT.  If they shuffle on for long enough, they could well make a significant dent in the cost of the operation.  And that’s why comparing the price of a dead Afghan with a QALY is not as simple as it seems.

No, I don’t much like it, either.  But if you’re going to have a sensible debate about how much compensation should be paid to someone, and how worthwhile it is to save their lives, you have to be brutally honest about the terms of the debate.  Noone ever said that NICE had to be nice.

  • Lindsay

    I think you are right Iain. The problem is that the Afghan has already been killed at the time compensation is paid, whereas in the case of NHS resource allocation we are paying to (partially? temporarily?) avoid a death. The economic term I think you are looking for is “sunk costs”. If a cost has already been irretrievably borne, then decision theory says it should pay no part in future decision-making. Of course, there are other reasons why you might want to pay compensation, and why (perhaps) existing sums might be too low.

  • Hi Iain

    I pretty much agree with you here. One comparator which would be more relevant for Singer's argument than NICE's cost-effectiveness threshold is the amounts that the Criminal Injuries Compensation Authority pay out.

    If you are murdered, CICA provide your relatives with compensation of £11,000. See

    Of course £11,000 is not absurdly larger than the £5,000 – so if we think the £5000 is obviously too low, we might wonder if our CICA payments are too low. Or alternatively we might wonder if there is a morally relevant difference between a case where the police service allows the murder of an innocent person to occur, and one in which an army actively kills an innocent person.

    (NB If you survive the attack and are left paraplegic, then you will receive £175,000 from CICA – presumably, as you argue in the case of NICE, the purpose is to pay for future care).

  • Yep: future care, and the plausible Lucretian idea that being chronically injured is (other considerations about quality of life aside) worse than being dead, so merits higher compensation.