Counting once, counting twice: the precarious state of subsidy regulation.

• Author: Zheng, Wentong
• Position: Continuation of IV. The Subsidy Pass-Through Presumption: Economic Fallacies through VII. Conclusion, with footnotes,, p. 451-476

1. The Behavioral Impact of Subsidies

   The first step in understanding the falsity of the subsidy pass-through presumption is to note that not all subsidies cause subsidy recipients to alter their output level or price. Below, the term "behavioral impact" will be used to refer to the impact of a subsidy on the output level or price of a subsidy recipient. For a subsidy to have a behavioral impact, it must cause a subsidy recipient to change either its output level or price, or both.

   As discussed earlier, the scope of subsidies that are subject to regulation under the SCM Agreement is very broad. Under the SCM Agreement, a subsidy is defined as a financial contribution that confers a benefit on a specific enterprise or industry or a group of enterprises or industries. (172) This definition makes no distinction between subsidies that cause recipients to change productive behavior and subsidies that do not cause recipients to change productive behavior. An example of subsidies in the former category is a fixed payment to the recipient on each unit of its products. With a lower cost of production, the subsidy recipient may increase production or lower price. (173) An example of subsidies in the latter category will be a lump-sum payment that is not contingent on the recipient meeting any performance requirements. Such a subsidy will not alter the recipient's productive behavior and will only result in a transfer of benefits to the recipient. (174)

   That not all subsidies have a behavioral impact is supported by economic theory on firm behavior. In an important contribution to the economic literature on subsidies, Professors Goetz, Granet, and Schwartz laid out the general framework for analyzing the behavioral impact of subsidies. (175) They argued that for a foreign subsidy to adversely affect the interests of a U.S. producer, the subsidy must either directly or indirectly lower the foreign producer's marginal cost of production. (176) One example of such cost-reducing subsidies given by Goetz, Granet, and Schwartz is a reduction in the price of a variable input. (177)

   The economic reasoning behind Goetz, Granet, and Schwartz's proposition is straightforward. In economics, the marginal cost of production is the cost of producing one additional unit of products, whereas the marginal revenue is the increased revenue from selling one additional unit of products. (178) When the marginal cost of production is lower than the marginal revenue, it is more profitable for a firm to increase production and vice versa. A profit-maximizing firm, therefore, will choose a level of production where the marginal cost of production equals the marginal revenue, whether the firm is perfectly competitive or a monopoly. (179) Figures 2 and 3 below illustrate the productive behavior of a perfectly competitive firm and a monopoly firm respectively.

   In Figure 2, MC represents the marginal cost of production of a perfectly competitive firm and MR represents the marginal revenue of such firm. Because a perfectly competitive firm cannot influence the market price, it faces a horizontal demand curve, DD, which is also the firm's marginal-revenue curve, MR. (180) The marginal-cost curve intersects with the marginal-revenue curve at (p, [q.sup.0]), meaning that the firm produces [q.sup.0] units at price p. When a subsidy lowers the marginal cost of production, the marginal-cost curve shifts downward from MC to MC' and intersects with the marginal-revenue curve at (p, [q.sup.1]), meaning that the firm produces [q.sup.1] units at price p. The subsidy has an impact on the firm's productive behavior because it induces the firm to increase its volume of production. However, for this to happen, the subsidy must cause the marginal-cost curve to shift by lowering the firm's marginal cost of production.

   [FIGURE 2 OMITTED]

   The same also holds true for a firm with market power. Figure 3 below illustrates the productive behavior of a monopoly firm in response to a subsidy. In Figure 3, MC represents the marginal cost of production of a monopoly firm and MR represents the marginal revenue of such firm. Since the quantity of the output of a monopoly firm could influence the market price, the marginal-revenue curve facing a monopoly firm, MR, is downward sloping (181) and is always lower than the market demand curve, DD. (182) Like a perfectly competitive firm, however, a monopoly firm will choose a level of production where the marginal cost equals the marginal revenue to maximize profits. (183) This means that the monopoly firm will produce [q.sup.0] units at price [p.sup.0]. When a subsidy lowers the monopoly firm's marginal cost of production, the marginal-cost curve shifts downward from MC to MC'. As a result of the subsidy, the monopoly firm produces [q.sup.1] units at price [p.sup.1]. The subsidy has an impact on the monopoly firm's productive behavior, as it induces the monopoly to increase production and lower the price. For this to happen, the subsidy must cause the marginal-cost curve to shift by lowering the marginal cost of production.

   [FIGURE 3 OMITTED]

   Richard Diamond refined Goetz, Granet, and Schwartz's analysis and applied it to specific types of subsidies. (184) Diamond made it clear that the receipt of a subsidy by a firm does not per se change any of its actions or the action of any other participants in the relevant market. (185) Consistent with Goetz, Granet, and Schwartz's theory, Diamond argued that a subsidy will cause the recipient firm to change its productive behavior in a manner that adversely affects the "entitlement" of U.S. producers only if it lowers the recipient firm's marginal cost of production. (186) Diamond categorized various types of subsidies into three groups based on their impact on the recipient firm's marginal cost. First, some subsidies do not affect the marginal cost of the recipient firm. These subsidies include, among others, subsidies intended as a transfer of benefits to owners of the recipient firm and subsidies provided to cover operating losses, to decommission unused facilities, to help in the clean-up of existing company wastes, or to pay vested retirement allowances if the payments were unanticipated. (187) Second, some subsidies vary with production and therefore affect the recipient firm's marginal cost. Examples of these subsidies include per-unit subsidies (whereby the amount of subsidies received by the recipient firm varies with the quantity produced) and input subsidies (whereby the subsidy is based on the consumption of an input whose use varies with production). (188) Third, some subsidies affect firm decisions regarding capital assets, i.e., factors of production that remain fixed during the period for which production is set. Examples of these subsidies include grants (189) and low cost loans (190) that are used to purchase capital assets. (191) As Diamond explained, the impact of these subsidies on the recipient firm's marginal cost is uncertain: It may result in lower, equal, or greater marginal cost depending on the characteristics of the plant or machinery purchased. (192)

   The marginal-cost-of-production theory propounded by Goetz, Granet, and Schwartz and by Diamond is not without skeptics. Alan Holmer, Sosau Haggerty and William Hunter acknowledged that subsidies may not always alter the productive behavior of the subsidy recipients, but they argued that a subsidy that functions only as an income transfer to the subsidy recipient still affects the entitlements of foreign producers because the subsidy will increase the revenues of the recipient firm and will, over the long run, encourage other firms to enter the subsidized sector. (193) Alan Sykes holds a similar view that "virtually all subsidies" have the potential to cause recipients to increase output. (194) According to Sykes, a subsidy that increases the profitability of the recipient will attract more firms to enter the market, leading to increased output over the long run. (195) However, even if one were to accept the proposition that all subsidies increase output over the long run, the current subsidy regulation regime does not require a subsidy to have increased output before it could be regulated. (196) It is entirely possible, therefore, for a subsidy that does not alter the recipient firm's short-run cost structures and has yet to cause long-run effects--therefore a subsidy that has no behavioral impact yet--to be subject to regulation under the current subsidy regulation regime. (197)

2. The Price Impact of Subsidies

   Even assuming that all subsidies lower the marginal cost of production of the recipient firm, that condition alone is not sufficient for a subsidy to have an impact on the price of the subsidized product. As discussed below, for a subsidy to pass through to the price of the subsidized product, the structure of the market for the subsidized product must be such that the recipient firm has the capacity to influence the world price of the product.

   That a subsidy lowers the recipient firm's marginal cost of production is not a sufficient condition for the subsidy to have a price impact can be seen from Figures 2 and 3 above. In Figure 2, where the subsidy recipient is a perfectly competitive firm in the world market for the subsidized product, the lowering of the firm's marginal cost of production causes the marginal-cost curve of the firm to shift downward from MP to MP', but the new marginal-cost curve still intersects with the demand curve at price p. The only change in the firm's productive behavior caused by the subsidy is increased production volume from [q.sup.0] to [q.sup.1]. Suppose that [DD.sup.d] represents the domestic demand for the subsidized product. Prior to the subsidy, the domestic demand for the subsidized product at price p is [q.sup.d], resulting in a net export of [q.sup.d][q.sup.0] by the firm's. After the subsidy is conferred, the domestic demand for the product remains at...