in the US, as elsewhere, most public expenditure goes towards redistributive transfers, health and education. Table 2 shows
total government expenditures for 18 OECD countries. Most government expenditures go towards health (7 to 19 percent of government spending), education (7 to 15 percent), and ‘social protection’ programs that more directly redistribute income (19 to 45 percent).
The goods most often cited as public goods are unimportant in terms of overall government expenditure for OECD countries: defense accounts for 1 to 6 percent of spending; public order between 2 and 5 percent of spending.
Pointer from Bryan Caplan.
It is an interesting and wide-ranging essay, difficult to excerpt. Here is another:
substantial insights into the economics of non-rival goods can be gained from the analysis of clubs (for providing local goods that are non-rival but excludable), the theory of natural monopoly (for goods such as Microsoft office where there are substantial initial development costs but the additional cost of an extra Word user is (close to) zero), or the economics of information (for the development of new technologies or drugs, where the manufacture of the new drug may cost a dollar or two per user, making it close to non-rival but the drug development may cost millions). While it is interesting and useful to have a theory for the special sub-set of non-rival goods that happen to be non-excludable also, there too few such goods to justify the place such goods hold in the public economics curriculum
One point she is making is that we should not encourage students to think that the theory of public goods explains or describes the actual role of government.