تامین تجهیزات استراتژیک, آزادسازی و ساختار بازار

We examine strategic procurement behaviour by governments and its effect on market structure in sectors, such as defence and pharmaceuticals, where the government is the dominant consumer. In a world economy with trade between producer countries, and between producers and non-producers, we use a modified Dixit–Stiglitz utility function with an independent taste for variety. There is free entry and exit by firms, but by anticipating their participation constraint governments can indirectly choose the number of domestic firms and their size through its choice of procurement price. Unlike the standard model with no independent taste for variety and no external sector of non-producer countries, there are incentives for subsidies, openness impacts on industrial structure and procurement coordination between producer countries affects firm numbers.

Government procurement constitutes an important share of a typical country's GDP (up to 20% in some cases). In some industries, domestic government procurement is also the most important source of sales and this is clearly the case in the defence and pharmaceutical industries (see e.g., Achilladelis and Antonakis, 2001 and Kyle, 2007). As the World Trade Organization expands the restrictions over traditional protectionist trade policies, procurement practices could be used as a less obvious trade policy tool to promote strategic domestic industries. We refer to this as strategic procurement. The government's preference for maintaining a domestic provider base within ‘sensitive industries’ can provide a justification for strategic procurement. 1 The defence industry provides a clear example of domestic firms survival directly depending on government purchasing commitments and regulatory environment (see Dunne et al., 2003). An interesting illustration of this fact is the 1993 merger wave of US military firms.2 In the pharmaceutical industry, according to Kyle (2007), in many producer countries the price for prescribed drugs to be paid by domestic health authorities is set high enough to support the local pharmaceutical industry, which is a big employer and important export earner.3 Interestingly, in both industries, there has been a recent tendency towards an increase in concentration. In the defence industry, for the top 100 firms, Dunne et al. (2003) report falls in the inverse Herfindahl index from 49 to 22, between 1990 and 1998. For the pharmaceutical industry, Matraves (1999) reports an increase in global market shares of the top 10 pharmaceutical companies from 25% to 31% between 1988 and 1995, also firms in ranked places from 11th to 20th saw increases in their market shares. Changes to procurement policies may be behind these trends.4 The above mentioned industries share a number of additional characteristics which may influence procurement decisions. First, procurement authorities usually have a preference bias for the consumption of domestic goods (‘home bias’), which could arise from concerns about security of supply in conflict or a desire to maintain a domestic industrial base in these sectors. Such concerns are subject to change across industry and time (see Achilladelis and Antonakis, 2001, Pharmaceutical Industry Task Force, 2005 and NHS procurement review, 1998). Second, procurement authorities are interested in purchasing a variety of products that gives an aggregate of either military capability or medicines provision. Variety is important to the procurement authority so as to cover a spectrum of health and security risks. Third, there is a relatively well-established set of producer countries. Most countries cannot afford the massive R&D required to set up a major weapon systems or innovative drugs industry and, therefore, there is a small number of producer countries serving both themselves and the non-producer countries. This means that there is scope for producer countries to interact strategically. Our main objective is to examine the impact of strategic procurement behaviour on the market structure of producer countries. More specifically, we analyze the impact that changes in home bias, taste for variety or the relative size of the non-producer market may have on procurement prices and market structure. We also study whether governments get a ‘better deal’ from their domestic producers, that is, whether the price paid for domestic procurement is lower or higher than the price at which domestic producers sell internationally. In addition, we investigate the potential impact of international coordination of procurement decisions on concentration. Such analysis is relevant to the industries we have in mind in this paper. As discussed by Hartley (2006), there is an ongoing debate within EU countries about the possible gains from coordinating defence procurement decisions. Different possible levels of coordination are being suggested; the lowest level would just imply a coordinated decision on domestic procurement. It is the consequence of this type of cooperation that we aim to model in our paper. There has also been an increase in the coordination in the regulation of the pharmaceutical industry in the EU which is relevant to the present paper (see Vogel, 1998). We construct a model of strategic public procurement and international trade. There are both producer and non-producer countries. Governments in producer countries buy products from the domestic firms and also import from the rest of the world, governments in non-producer countries cover their public procurement needs through imports. Governments endogenously determine the number of domestic firms by committing to a domestic procurement price that ensures their existence. Our focus in this paper is the decision of the military or public health authority on how best to utilize the budgets they have, which are assumed to be exogenous. Endogeneizing the expenditure choice would be interesting, but far from straightforward. Military capability is just a component in a measure of national security. The choice of how much to spend on military capability is affected by a number of factors such as the impact that this expenditure may have on the behavior of potential adversaries (see García-Alonso and Levine, 2007 for a discussion). Even for the health sector, we are really only considering medicines provision; a complete measure of health provision would include many other factors such as numbers of GPs, preventive health, etc. which we do not analyze in this paper. An important feature of our model is the existence of producer and non-producer countries. A big domestic market and strong home bias (see Levine et al., 2000) may affect governments decision to initially support a domestic industry. Other factors, such as export controls, limit the access that importer countries have to sensitive military technologies. Also, regulations on prices and patents are important determinants of why some countries may become drug producers (see Achilladelis and Antonakis, 2001). However, our model does not aim to explain the reasons for countries to become producers; that is, we take the producer or non-producer status as given. The interaction between the procurement authorities across countries and firms is modelled as a four-stage game with a subgame perfect equilibrium (henceforth, SPE). First, given the procurement budget, producer governments choose the volume and price of domestically procured goods. Second, given procurement decisions firms enter or leave the sector. Third, firms in producer countries simultaneously compete in export prices and finally, governments in both producer and non-producer countries procure imports. We then use backwards induction to find the SPE. In a standard Dixit–Stiglitz monopolistic competition model of trade only involving producers (see Dixit and Stiglitz, 1977), the procurement price turns out to be the world market price, the bias for domestic rather than imported procured goods, the inverse of ‘openness’ in our terminology, has no effect on market structure and the non-cooperative procurement equilibrium is efficient (from the viewpoint of producers). As a result of two features of our model, these results no longer hold: first, the existence of an external market of non-producers importing goods from producers and second, we use a modified Dixit–Stiglitz utility function as in Benassy (1996) to incorporate a taste for variety effect that is independent of the elasticity of substitution. The latter allows the choice as to whether to procure an additional variety to be different from the choice of quantity procured. In this paper, we find that the price that procurement authorities pay to their domestic firms may be higher or lower than the imports price. Procurement authorities are more likely to ‘overpay’ their domestic firms if they have a high taste for variety and the external market is relatively small. In addition, we show that an increase in openness, a reduction in taste for variety and an increase in the relative size of the external market reduce the number of firms in equilibrium. This result provides a theoretical explanation for the recent increases in concentration in both the defence and the pharmaceutical industry. Increased development costs and a reduction in health and military budgets may be among the other factors determining such trends (see Achilladelis and Antonakis, 2001). Our paper also shows that cooperation between the governments of producer countries in setting procurement policies affects firm numbers. Such impact will depend on the size of the external market and the additional taste for variety effect. For instance, with a non-producer fringe but, no additional taste for variety factor, cooperation would lead to an increase in concentration. For the case of no external market but, additional taste for variety factor, the opposite holds. Our research is linked to a branch of the procurement and trade literature starting with Baldwin (1970).5 This literature studies the impact of unilaterally home-biased procurement on the patterns of international specialization. Baldwin, 1970 and Baldwin, 1984 shows that a unilateral home bias in favour of domestic producers is inconsequential to the patterns of specialization under the assumption of perfect competition. Later papers prove that this neutrality result does not necessarily hold with imperfect competition. Brulhart and Trionfetti (2004) prove that if a country has a unilateral home bias towards a domestic monopolistic sector, it will also have more firms in that sector relative to the other country (see e.g., Miyagiwa, 1991 for impact on trade volumes). Our framework differs from this literature in a number of aspects. Brulhart and Trionfetti (2004) consider a general equilibrium model in which there is both private consumption and public procurement. Government's bias in favor of domestically produced goods means that a given proportion of government purchases are reserved to domestic producers (a higher proportion means a higher home bias). In our paper, there is no private demand. Also, we have a multiple stage game where governments foresee the impact that their domestic procurement decisions have on firm numbers. In order to make this model more tractable, we use a partial equilibrium model which focuses on the simultaneous decisions of the procurement authorities across countries, each with an exogenous procurement budget. However, unlike Brulhart and Trionfetti (2004), we allow for the existence of a group of non-producer countries and we generalize the standard Dixit–Stiglitz preferences to include an additional like-for-variety element as in Benassy (1996). Given the differences between the two papers, comparing them is not straightforward. Using the symmetric general equilibrium as the baseline, they analyze the impact of a unilateral change in home bias on specialization. In that case, monopolistic competition with Dixit–Stiglitz preferences are enough to make changes in home bias affect the number of domestic firms relative to total firm numbers across the world. Whilst their focus is to analyze the impact of asymmetric changes in home bias on international specialization (domestic number of firms relative to the rest of the world), our comparable objective is to analyze the impact of a symmetric multilateral change in home bias across producers on the absolute numbers of firms. For the standard Dixit–Stiglitz preferences, we show that this would not have an impact on total firm numbers. When the additional like-for-variety element or the external producer market are introduced, more home bias will lead to an increase in the number of firms. Interestingly, in the cooperative case, the optimal number of firms is independent of home bias. The present paper is specifically designed to capture government home bias. In a separate paper (Coto-Martínez, García-Alonso and Levine, 2005), we develop a model in which the bias lies within the private consumers. In that paper, there is no government procurement, only private consumption. Although a parallel could be established between the two papers, the modeling requirements are quite different. Even if there is a home bias, with private consumers one must ensure that the international price arbitrage conditions are met. Also, since there is only private consumption, governments have different policy instruments such as domestic consumption subsidy, imports tariff and a fixed cost subsidy to the domestic firms. In the present paper, we focus on the impact that government procurement decisions alone have on industry structure. The rest of the paper is organized as follows. Section 2 provides the basic set-up and the sequence of moves in the procurement game with governments and firms as players. Section 3 solves the SPE with non-cooperative at the procurement stage one of the game. Section 4 studies the cooperative procurement equilibrium at stage one and compares it with the non-cooperative equilibrium. Section 5 provides some concluding remarks. Detailed proofs of Propositions can be found in García-Alonso and Levine (2005).6

This paper has explored the strategic procurement behaviour by governments who can, in effect, choose the number of firms and their size by adjusting the procurement price. In a standard Dixit–Stiglitz monopolistic competition model where all countries are producers, the procurement price coincides with the world market price, openness has no effect on market structure and cooperation between producer governments does not affect firm numbers. With an external market and a modified Dixit–Stiglitz utility function to incorporate a taste for variety effect, this is no longer the case. In a symmetric, non-cooperative equilibrium, the number of firms is influenced by a number of factors. We show that the number of firms increases as the taste for variety by producer countries increases. In addition, an increase in openness, in the form of a reduction in preferences of producer countries for domestic supply, and an increase in the relative size of the external market, results in a decrease in the number of firms and therefore an increase in concentration. We also show that producer government cooperation in procurement decisions may now affect firm numbers. For example, for a large external market, coordination would lead to more concentration. The main implication of our results is that the marked increase in concentration in the military sectors of the US and the EU can be explained by a increase in openness and the increased importance of the external sector of arms importers. To some extent, this also helps to explain concentration trends in the pharmaceutical industry. In addition, our result on the impact of cooperation in procurement decisions on firm numbers is specially relevant to the military industry in the EU context where different levels of cooperation in military procurement are being discussed (Hartley, 2006 discusses current proposals for cooperation in the military industry). Our paper makes a number of simplifying assumptions that allow us to use a model specification that can be applied to a number of industries: home bias is exogenous, we only analyze the symmetric equilibrium, we do not endogeneize the decision of countries to become producers or non-producers and procurement budgets are exogenous. As already discussed, the motivation of the home bias may change across industries and also across time and therefore, it is not easy to endogeneize whilst keeping the generality of the model. The focus on the symmetric equilibrium is made to help us obtain explicit results (Brulhart and Trionfetti, (2004) analyze the asymmetric case for the standard Dixit–Stiglitz framework). Also, endogeneizing the government's decisions on the size of the procurement budget would again require us to make assumptions on the government's objective function. These would include the more complex decision on how much national security or health to provide; we just focus our analysis on military capability and pharmaceuticals provision which are only elements of national security and health provision. Similar issues arise when endogeneizing the decisions of countries to be producers or non-producers. The above issues have been partially addressed in the defence and health literatures. Our paper presents a simple analysis of the impact of government procurement policies on market structure and prices. Finally, the regulation of the industries that motivate our paper tends to be quite elaborate and contains other elements that may affect industry structure. For instance, in the case of the pharmaceutical industry, price regulation of medicines is complex and varies across countries. For the military industry other policies such as export controls influence the market structure. Our paper aims to present a simple model that may encompasses the idea that government's monopsony power can have an impact on market structure that is applicable to at least these two industries.