requirement for discrimination not to affect imports adversely is that the government buys at the private market price. While more empirical research is needed on all these questions, the limited evidence available suggests that in a wide variety of situations discriminatory procurement can adversely affect trade and the non-discriminatory disciplines of the GPA are necessary to prevent distortions.
It could, of course, be argued that there is a case for selective application of the GPA rules depending on characteristics of specific sectors, such as the relative importance of government demand, substitutability of domestic and foreign goods, market structure and the manner in which procurement price is determined. This would eliminate the administrative costs of compliance with the GPA where it is unnecessary. However, there are three reasons why there may be little reason to modify the GPA's disciplines: first, there will be considerable conceptual and empirical difficulty in establishing clear rules for inclusion of sectors; secondly, the costs of compliance with the GPA where it is not necessary are likely to be small compared to the benefits of compliance where it is; and finally, coverage of sectors and entities is in any case negotiated between countries, so there is already an element of selectivity, with countries presumably unwilling to waste negotiating currency to secure the inclusion of sectors and entities where rules are redundant.
The cost of procurement and welfare implications of price preferences