When are liability rules inferior to property rights?
The preceding analysis showed that from an efficiency point of view, the liability rule is superior. But does the analysis necessarily imply that a liability rule is always superior to a property right?: after all, even in the case of zero transaction costs, both give the same outcome in the above situation. So why would a property right ever be preferable? One possible reason for preferring a property right when transaction costs are small is that courts may underestimate the entitlement holder's true valuation of the activity, instead preferring some measure of market value rather than personal subjective value.
Consider, for example, Figure 7.2.2. Suppose that a court estimates that the market marginal value of the residents' right to clean air is v, and under a liability rule awards them compensation on this basis. Then the factory will produce excessively, at Q'LR > QLR. The factory will gain A + B + C + D + E + F + G + H + I and pay compensation (D + E + F + H), for a net gain of (A + B + C + G + I). The residents will be compensated (D + E + F + H), but will lose E + F + G + H + I + J, for a net gain of D - (G + I + J).
The overall welfare gain (relative to the initial situation of no trade) is A + B + C + D - J. The welfare loss from excessive production of the factory is J. Thus, for sufficiently small transaction costs and a sufficiently severe underestimation of the subjective value that the residents place on clean air, a property right may be preferable to a liability rule.
Figure 7.2.2 Liability rules and transaction costs