What do high infant mortality rates really tell you?

The US and the UK have high infant mortality rates. As the Economist.com points out:

THE new UNICEF report on children in industrial countries is out. Readers will be shocked, shocked! to find out that the United States and the UK are indisputably the worst places to have been a child.

The problem with all of these reports, of course, is what computer programmers call GIGO—Garbage In, Garbage Out. They are extraordinarily sensitive to the chosen metrics. So if you think that the most important thing is for children to be as close to each other as possible in income distribution, you will decide that Danish children are living in paradise. On the other hand, if you peg material items like dishwashers, computers, and so forth as major contributors to child welfare, then the United States might be more to your taste; as the Heritage Foundation points out, the Census Bureau finds that . . .

Yet neither income inequality, nor the abundance of colour televisions, tells me what I want to know which is how happy and healthy are children in these various countries?

Even things like health statistics are fraught. African-Americans have, for reasons no one quite understands, higher levels of premature birth, infant mortality, and low-birth-weight babies, and birth complications. This is true even when obvious factors like income, prenatal care, and maternal health and age are controlled for, and substantially lowers America's performance in the statistics.

Similarly, the UN has somewhat inexplicably decided to use "deaths from accidents and injuries, 0-19", as a proxy for health among that age group, rather than the more obvious "deaths, 0-19". This statistic makes America look awful, almost entirely due to the fact that American children spend a lot of time in cars. Yet the differences run from 10 per 100,000 to 20 per 100,000, meaning that 99.98% of American children lead lives blissfully untouched by accidental death. . . . .

There is also the possibility of what is known as the Peltzman effect. Making something safer can actually encourage more dangerous behavior. Suppose that you make riding a motorcylce completely safe. What would happen to how fast that you drive? I bet that people would drive a lot faster. In general, safety features may increase or decrease the number of deaths. Airbags might reduce the number of deaths per accident, but they might also increase the number of accidents because people feel that they can drive more recklessly. You might also be more likely to accidentally kill a few pedestrians. There is always the risk that improved health care might generate a similar response. Some groups of people might respond to these changing costs more than others.

Of course, some people might engage in risky behavior generally because of the social safety net. People might be more willing to risk getting hooked on drugs because there are so many potential protections for them. But getting hooked on drugs might also mean more premature births and thus greater child mortality.