Technology is applied knowledge. Knowledge is a pure public good because it is completely non-rivalrous (my consumption does not reduce your consumption).
Pure public goods suffer from the free-rider problem. The good should be provided to the level where the marginal cost equals the marginal social benefit. But the marginal private
benefit is less that the marginal social benefit and so private provision tends to under-supply.
Patents solve this by allowing the owner to appropriate the social benefit. The licence fees or monopoly market position enable the owner to internalise the social benefit by converting it to private benefit. Therefore the marginal private benefit can be equated to the marginal social benefit, meaning that an efficient level of knowledge production can be achieved.
The efficient level defined here is a Pareto-efficient outcome. At this level of output, it is not possible to increase the welfare of one member of society without decreasing the welfare of at least one other.
But Pareto-efficient outcomes are seldom socially optimal. Market conditions do not necessarily lead to the socially preferred outcome – poverty, inequality, etc can be pareto-efficient outcomes of a market allocation of resources.
In the case of patents, the Pareto-efficient outcome is achieved by allocating the entire market surplus to the patent owner. Therefore, although we can buy high quality DVD players for less than $100, almost the entire economic profit is being appropriated by a few companies that make hundreds of millions of dollars in revenue from their licences, and some of which do not manufacture anything.
A solution to the inequality and social non-optimality may be redistribution via taxation. The patent system attempts to provide this by applying patent maintenance fees, but they are set at derisory levels that are only sufficient to cover the costs of running the patent office. How much redistribtion can take place when maintenance fees are set at several thousand dollars for a patent with a market value of hundreds of millions?
So how about a licence revenue tax policy? Coupled with a monopoly profit tax. The funds could be used to reallocate surplus to areas of technology that are under-developed (for good market reasons) such as research into malaria or other diseases affecting poor countries; or low-carbon energy technologies, etc. A tax policy such as this can achieve better social optimality without producing an inefficient outcome, as is shown by the second theorum of welfare economics.
There are, as usual, practical difficulties. What company will correctly disclose the value of its patent portfolio if it is going to be taxed on that value? And what are the transaction costs of administrating such a valuation and collection policy?
