Public goods are undersupplied in markets (private provision is not optimal)
Lighthouses in the UK were publicly allocated
No.3 is admittedly wrong, but No.s.1&2 are debatable.
No.1: Lighthouses are not pure public goods. Of course the services they provide are for all purposes nonrivalrous and nonexcludable; however not everyone really cares to use them, or benefits directly from them. Those with direct benefits--the shipping and whaling interests-- place a much higher value on lighthouse services than, say, farmers. But the latter also benefit through increased (and lower-cost) trade that lighthouses make possible. So lighthouses are first and foremost club goods to those with direct interests, and to a lesser extent public goods.
No.2: Market undersupply of pure public goods is not entirely relevant to the discussion of lighthouses, as lighthouses are demanded by admixture of club and public interests. How does the club character of club-public goods affect their "optimal" provision? We don't know--there is no theory to guide us here. The empirical evidence in the case of lighthouses seems to demonstrate that club interests provided substantial pressure (effective demand) for private participants and government to create the necessary institutions to secure private allocation of lighthouses with government enforcement. This does not mean markets--i.e. Trinity House--were, are, or will be perfect.
Last week, the statistical agencies of the federal government were given the OK to share amongst themselves data collected from businesses. The National Association for Business Economists has been supporting such a move for a long time:
The bill (H.R. 2458), which authorizes the sharing of business data among the Bureau of Economic Analysis, the Bureau of Labor Statistics, and the Census Bureau, has been NABE's chief legislative priority for the past seven years. Data collected from households are not covered by this legislation.
The new law will protect the confidentiality of propriety information companies supply to the federal government on surveys, will reduce companies' reporting burden by eliminating duplicate surveys, and will improve national economic information. Major statistical releases, such as the gross domestic product and productivity, should be more accurate and less prone to dramatic revisions because of more complete and consistent source data.
Over the next few months, I will be reviewing the theoretical and empirical evidence that supported lobbyist claims of increased efficiency and statistical accuracy; a more long-term project will identify (and quantify) the resulting quality changes in U.S. government statistics.
But many doubts remain: Is this a feasible research program? Do we have the capability of assessing the impact of policy changes to statistical agencies? If we do, are we successful? If we do not, what exactly is the basis for our policy recommendations?