GPA actually prevents countries from maximising their national welfare. It is the purpose of this Section to examine whether the non-discriminatory provisions of the GPA survive these theoretical assaults.
Implications of preferences for trade and output
The general presumption that discriminatory procurement reduces imports and increases domestic output in comparison to free trade is the primary motivation for the GPA. However, Baldwin (1970, 1984) and Baldwin and Richardson (1972) are among those who have argued against this presumption in a perfectly competitive context. They have shown that if private demand is sufficiently large, and domestic and foreign goods are perfect substitutes, then extending preferential treatment to domestic industry neither reduces imports nor increases domestic price, output and employment. Discriminatory procurement is ineffectual because shifting government demand towards domestic products tends to increase their prices and, therefore, generates an equal and opposite shift in consumer demand toward imports. Any price preference is simply a transfer from the government to domestic producers.
This result can be qualified in three important respects. First, in certain sectors, government demand may be a large part of total domestic demand. More precisely, if government demand is larger than the quantity supplied domestically in the non-discriminatory equilibrium, then shifting domestic demand towards domestic producers can clearly have real effects - i.e. increased domestic output and reduced imports. Thus, a revealing empirical test would involve comparing the magnitude of procurement with the quantity supplied domestically at the notional free trade price to determine when discriminatory procurement is likely to affect trade and output.8
Data on procurement, which is sufficiently disaggregated to make the homogeneity assumption plausible, is difficult to find. At a somewhat aggregated level, Francois et al. (1995) find that even though public purchases in the United States are large ($440 million in 1993), they tend to be concentrated in a few sectors. Government accounts for 80 per cent
(In principle, it may be possible to directly estimate the impact, if any, of the GPA on trade flows. Such an exercise has not yet been attempted, and is beyond the scope of this paper. There is some evidence to suggest that the share of domestic sources in aggregate procurement has declined since the inception of the GPA (Hoekman, 1995). But the extent to which even these changes can be attributed to the GPA is not clear.