David Henderson has an excellent essay on “GDP Fetishism” up at EconLib. Of the problem of valuing government spending, he writes:
Take the first inaccuracy—the valuing of government-provided goods and services at cost rather than at market prices. Many government programs actually destroy value rather than create it.
This highlights the fact that knowing how much the government spends (as a percentage of GDP, say) is to know something of questionable significance. Henderson talks about the waste created by the TSA. More generally, the United States spends a spectacular amount of money on war ships, fighter planes, and bombs. Most of this is simply unnecessary for the security or defense of the United States, and so amounts to little more than a combination of upward wealth redistribution and throwing money in a hole. That’s domestically. (I’ll pass over the problem of bringing the lost lives of bombed foreigners into the equation.)
Now, we dynamist market-liberal types are right that the government does many things, such as primary education, that competitive markets could do better and more efficiently. If all government spending on education became equivalent private spending overnight, GDP would stay the same (in the short term, before gains in education quality started to pay off in terms of higher productivity), though people would be better off. However, government spending on education, as inefficient as it may be, is giving people something they want and directly benefit from. Much (most?) military and security spending is not like that. My sense is that, despite the U.S.’s historically relatively modest level of government spending, the composition of U.S. spending is such that U.S. taxpayers get less of value in return for their tax dollars than do taxpayers in many places with higher taxes and higher levels of government spending. Which is to say that when using GDP per capita as a proxy for welfare, the U.S. comes off better than it should relative to, say, Canada or Sweden.
Now, one could argue that U.S. military spending and hegemonic U.S. military power, makes it possible for the Canadian government to spend on health-care, from which the Canadian people derive some benefit (even if the service is not provided as efficiently as it could be), rather than on aircraft carriers, from which the Canadian people would derive basically no benefit (since the Americans have already taken care of it). I think this is almost certainly true, but it’s just another way of making the same point: Canadians are better off and Americans are worse off than their GDP stats suggest.
If one additionally takes into account things like the U.S.’s unusually high level of spending on keeping its citizens in prison cells (which we could do much less of without getting more crime), it becomes even clearer that U.S. GDP figures overestimates the U.S. standard of living relative to other wealthy liberal democracies. The composition of spending matters.
The U.S. is an notable anomaly in the happiness data. Average self-reported life satisfaction rose with GDP per capita over the last several decades in almost all wealthy liberal democracies, but not so much in the U.S. The idea that the unusual composition of U.S. government spending gives Americans unusually poor value for their tax dollars might help explain this.