Tyler Cowen is getting rightfully bashed for presenting both non-scaled and nominal federal spending projections (apparently his critics aren't satisfied with "but the numbers are from a well known Washington think tank" - go figure).
But this paragraph in his response provides a good example of why I hate references to "public goods" when talking about the appropriate function of government:
"Some MR commentators raise the issue of per capita measures of discretionary spending and whether they will decline. It might be nice to have growing public sector per capita quality with growing population and growing wealth. But if the good in question is a public good (and is it not supposed to be?), adding extra people to the mix, ceteris paribus with no spending boost, is compatible with those additional people getting more or less the same services as the previous consumers. Falling per capita expenditures on public goods, if it is not too big a fall, still means a greater real quantity of public goods enjoyed, given non-rivalry of consumption."
Are there public goods? Sure there are. But there are also lots of normal goods with externalities, and those are the really important ones. Those expenditures also grow with population and the size of the economy.
For the record I also don't like "market failure" although I probably use it somewhat more than the phrase "public good". "Market failure" implies there's something inadequate or bad about what the market is doing under certain circumstances. That seems like a strange way to think about it. It's like saying that the fact that hammers don't bang in screws well implies "hammer failure". I don't think most economists think the market is "bad" when they think about "market failures" but it's still sloppy language and it opens you up to sloppy critics who would rather presume you're insulting the dignity of the market than formulate a real argument.