Standard texts of price discrimination see it as beneficial in a social welfare sense: consumer surplus is transferred to producer surplus, but the fact that quantity increases is good for social welfare (as it reduces the deadweight loss triangle). Interestingly, this isn’t the whole story. Tullock pointed out, a number of years ago, that social welfare loss can originate from the rent-seeking activities of those that are trying to capture these rents.
For instance, in a monopoly setting, standard economic theory would say that the cost of monopolies comes from the reduction in quantity sold (as a result of higher prices), leading to a loss of social welfare. This is the (Harberger) deadweight loss triangle. But applying Tullock’s point, we see that social welfare costs are actually higher, as the monopoly firm will likely expend resources fighting to preserve this monopoly. They might, for example, lobby regulators and politicians to prevent their monopoly position from being removed. This is obviously not efficient, and can be thought of as a further social welfare loss (greater than the deadweight triangle normally considered).
In the case of perfect price discrimination (first degree), remember that we have different consumers, each with different price elasticities. In a non-perfectly competitive market, a price above marginal cost prevents a socially optimal amount of good being consumed. One way that this can be resolved, is by the firm charging different consumer segments different prices, depending on their elasticities. By charging a lower price to people with higher elasticities, and higher prices to those with lower elasticities, the firm can increase output (which is welfare enhancing) whilst also increasing profits as it better captures the consumer surplus of the lower elasticity consumers.
Leeson and Sobel (2007) apply the Tullock principle to the case of price discrimination, pointing out that price discrimination is costly to firms as they have to segment consumers and prevent resale of goods between consumer segments. The firm is actually willing to expend a large degree of effort in price discriminating, equal to the value of the deadweight loss, plus the consumer surplus that is transferred. However, the social welfare improvement only comes from the reduction in deadweight loss, so it is easy to see that such a move would be welfare reducing.
In other words, when price discrimination is costly, society may be worse off from its implementation. Specifically, if the costs of price discrimination are greater than the size of the deadweight loss, then society would be better off if price discrimination was prevented.
In fact, the authors find that for pure monopolists perfect price discrimination is sometimes socially inefficient, whilst for monopolistic competitors is always socially inefficient (under general demand and marginal cost functions).
In conclusion, we have seen that perfect price discrimination may not actually be welfare-enhancing, when there are high costs to enacting price discrimination. Essentially, this stems from the fact that the monopolist firm absorbs consumer surplus (as well as deadweight loss), and is thus willing to expend resources to capture the consumer surplus, which do not benefit society as a whole.