A third qualification is that markets are rarely perfectly competitive. Miyagiwa (1991) has examined the validity of the Baldwin-Richardson ineffectiveness proposition in imperfectly competitive situations. It emerges that if goods are perfect substitutes, then the proposition continues to hold under various types of industrial organization. However, if the government pays the domestic supplier a premium proportional to the import price, discriminatory procurement results in increased imports. The intuition is the following. The domestic supplier has a strategic incentive to raise the import price since this leads to an increase in the procurement price and therefore an increase in revenue from sales to the government. The increase in import price can be accomplished by curtailing its sales to the private sector which prompt a (less than offsetting) increase in sales by the foreign supplier. In the new equilibrium, the domestic firm's loss in the private sector is balanced at the margin by the increased profits from sales to the government. When goods are differentiated, the same counterintuitive mechanism still operates, but the results are less clear-cut.
An important assumption in the above analysis is that the government purchases at a price dependent on the market price. The results are somewhat different if government procurement is decoupled from private-sector demand, i.e. firms bid a price for government demand independent of the market price. Discriminatory procurement then effects imports both directly and indirectly, through a change in marginal costs. Three cases can be distinguished depending on whether marginal costs are increasing, constant or decreasing. When marginal costs are increasing, shifting government purchases from foreign to domestic firms weakens the latter's competitive position in meeting private demand. Thus, private sector purchases from foreign firms increase. The aggregate impact on imports is ambiguous. If marginal costs are declining, shifting government purchases towards domestic firms enhances the latter's competitive position in the private sector. Imports by both the government and the private sector are therefore likely to decline. If marginal costs are constant, shifting government purchases do not affect the private sector equilibrium, and the decline in imports is equal to the shift in government purchases.
Where does all this leave us? We can conclude that from the international trade point of view, discriminatory procurement is ineffectual in competitive markets only when government demand is small relative to domestic demand and domestic supply, and domestic and foreign goods are perfect substitutes. In imperfectly competitive markets, a further