Belligerati

The hospital eventually turned off Habtegiris's life support because she wasn't going to get better and couldn't pay for it to continue. This was their right under Texas law. The medical care to extend her life cost at least $2,255 per day.

Dr. Steven Landsburg, an economist that writes for Slate, took a beating in the press for suggesting that this was the right thing to do. Essentially, his argument was that because she was young, before she got sick she could inexpensively have purchased insurance. The fact that she didn't is a reflection of the value she placed on her own life. Since she didn't value her life as much as consumption today, this should relieve society of the cost of caring for her. She didn't care, so why should we.

This was an interesting point, and it met with the following basic criticism (which I read in an article by Rob Frank in the NY Times, now in the archives). Moral emotions like sympathy and empathy have utilitarian value. The purpose of the economy is utility, not economic efficiency. Providing free health care to those who could afford insurance, but didn't, and now can't afford treatment seems to have more utilitarian benefits than the loss from the moral hazard of fewer people buying that insurance in the first place.

This moral valuation allows inconstant valuation of life to creep into the benefit-cost analysis. The fact that someone will die with certainty in a probabilistic sense can be valued less than someone dieing with certainty from deliberate action.

For example, let's say that NYC has 25 children who die a year in bike accidents that could be prevented by wearing helmets. NYC could give every child in NYC a free helmet for $25 million. Let's say they would wear them if they got them for free. That's $1MM per life saved with probabilistic certainty. If all lives are morally equivalent, then this should be preferred over saving $25 million to save 5 backpackers lost in the desert. If we value the probabilistically ccertain lives saved as worth less the actual lives saved, we can prefer to spend the money on the second choice.

But I'm not convinced that we should calmly accept that our morals are what they are and blindly follow them in determining how to allocate social resources. For example, Habtegiris was from sub-Saharan Africa. The cost of a single day of her treatment was equal to the annual per-capita income of four residents of that region.
A recent study of childhood disease intervention in Africa found that lives could be saved for $779. 29 children have died in Africa because money was allocated to keep Ms. Habtegiris alive instead for their vaccinations. The UN estimates that 5 - 6 million children died last year from starvation. They can be easily fed on $.17 a day or $62.05 a year. 362 could be fed for a year for the price of Ms. Habtegiris's treatment. Those aren't probabilistic deaths. We might not know their names, but there are real children out their dying for certain without very cheap food.

Why should we value their lives as worth a .25% as much as hers? If we really think that strangers have a moral claim on our income to save their lives, we should decide how big that claim is and allocate it in a fair way. That means that if two people are equally innocent of their distress, we save the cheaper one first.