As we explained earlier, where goods are public rather than private, resources will tend to be inefficiently allocated. The inefficiencies we have in mind manifest themselves in two principal ways:
Under provision of public good
Under protection of public goods
Under provision follows from the non-excludability property and the attendant free-rider problem. It is difficult for a producer to extract a financial return for the provision of a public good, and so many public goods will either not be provided at all by free markets, or will be under-provided. Even if a firm could extract enough revenue to make market provision worthwhile, many people will tend to understate the extent of their willingness-to-pay and so market demand will be a systematically downwardly biased representation of the true social value of the good. Under protection of public goods tends to follow for similar reasons. Why would a profit-oriented owner of a public good (such as a tropical forest) bother to protect that asset for others to enjoy if no revenues can be extracted for that action?
A final problem concerns the efficient price of a public good. If the use of the good does not impose costs on others (which it will not do if it is a public good, because of the non-rivalry property) the socially efficient price should be zero! Recall that the efficient price is equal to marginal cost, which in this case is zero (once the good exists). So even if, somehow or other, a price could be extracted, any price other than zero would be socially inefficient!
The implication of these arguments seems to be that only the public sector is in a position to install or maintain pure public goods in socially-efficient ways. We have already noted applications of this idea to environmental problems. Another important application is to knowledge, and the processes which generate it: research, development and invention. Knowledge is close to being a pure public good. Think, for example, of the case of knowledge about a new drug formula. Once available, consumption of it is non-rivalrous. Also, as a matter of practice, it is nearly impossible to exclude others from using it. Will knowledge be generated in socially-efficient quantities in a pure market economy? Our analysis of public goods suggests the answer is probably no, and explains why the state plays such a large role, in most countries, in the generation of knowledge.