Effects of milk marketing order regulation on the share of fluid-grade milk in the United States.(Author abstract)(Report)

Milk marketing orders are central to dairy market regulation in the United States. Since the Agricultural Marketing Agreement Act (AMAA) of 1937 and similar state legislation of the same period, federal and state governments have administered marketing orders that determine minimum prices that processors must pay farmers for fluid-grade milk. These marketing orders have significantly influenced the economic performance of milk markets. In this paper, we show empirically that marketing orders have helped increase the production of fluid-grade milk beyond the demand for fluid milk.

Because the industry is large and the policies have important economic consequences, milk marketing orders are among the most studied of all agricultural policies (see, among others, Cox and Chavas 2001; Dobson and Salathe 1979; Ippolito and Masson 1978; Kessell 1967; Sumner and Wolf 1996). This literature has established what have come to be stylized facts of milk marketing orders. Price discrimination by marketing orders raises the relative price of milk used in beverages (fluid milk), reducing consumption of fluid milk and decreasing economic surplus of consumers of fluid milk. Revenue pooling by marketing orders, together with discriminatory pricing, raises the average producer price of milk, inducing increased milk production and increasing producer surplus. By reducing fluid milk consumption and increasing milk production, revenue pooling also effectively subsidizes production of milk for manufacturing uses, resulting in a lower price for such milk and for consumers of products like cheese, butter, and milk powder. Moreover, the degree of price discrimination varies by region, as do supply and demand parameters, resulting in welfare effects that also vary by region (Cox and Chavas 2001).

However, this important literature generally abstracts from a crucial feature in the implementation of milk marketing orders. Milk sanitation standards in the United States distinguish between milk eligible for use in fluid products, known as Grade A milk, and milk eligible only for manufactured dairy products, known as Grade B milk. The milk grade designation is made at the farm level, such that all milk from a given farm is either fluid or manufacturing grade. The highest standards are set for Grade A milk because of the food safety risks associated with fluid milk products. Marketing orders regulate only Grade A milk, and do not regulate Grade B milk. By raising the price of Grade A milk relative to Grade B milk, marketing orders raise the incentive to produce Grade A milk relative to Grade B milk.

[FIGURE 1 OMITTED]

The Grade A share of total milk production in the United States rose steadily through the latter half of the twentieth century. The Grade A share of milk grew from 59% in 1948 to 98% in 2000 (figure 1). The dramatic shift in the milk grade mix toward Grade A is remarkable in the context of the uses to which Grade A milk has been put. The quantity of milk used in fluid products has changed little since World War II. However, total milk production has grown dramatically, as has Grade A milk production, so that the share of Grade A milk actually used to make fluid products has fallen from 78% in 1948 to 34% in 2000 (figure 1). Thus, the share of milk satisfying the relatively high fluid standards grew, while a shrinking share of this milk was used in the fluid products.

Grade A milk not used in fluid products is used in manufactured dairy products where it is a substitute for Grade B milk. Given the difference in sanitation standards, Grade A milk is cleaner than necessary for use in manufactured dairy products; Grade A milk exceeds manufacturing grade standards. In this sense, Grade A milk used to make manufactured dairy products is of excess quality, which comes at the cost, borne by producers, of meeting the stricter standards for Grade A milk.

Economic intuition suggests a link between milk marketing order policy, which raises the relative price of Grade A milk, and the observed shift toward Grade A milk production in the United States. Indeed, a small number of extension reports by agricultural economists suggest that these observed patterns were driven at least in part by a conversion of farms from Grade B to Grade A in response to marketing order regulation (Bartlett 1964; Cummins 1978; Frank, Peterson, and Hughes 1977; Graf and Jacobson 1973). Yet the literature dedicated to the subject of modeling and measuring the effects of milk marketing orders on prices, quantities, and economic welfare has largely ignored the interaction between marketing order regulation and milk sanitation regulation. A recent exception is Balagtas and Sumner (2005), who show that the previous literature overstates producer benefits and understates the social costs of milk marketing orders.

[FIGURE 2 OMITTED]

This article develops an econometric model to test the link between milk marketing order policy and milk grade. We model the Grade A share of milk as a dynamic stochastic process influenced by the costs and benefits of Grade A conversion. To measure the benefits of conversion, we use the premium paid for Grade A milk relative to Grade B milk. In the next section, we show how this Grade A premium captures price discrimination and revenue pooling by marketing orders. To identify the effect of marketing orders on the Grade A premium and, in turn, on the Grade A share, we exploit exogenous variation in marketing order policy over time and across states. We find strong econometric support for the hypothesis that marketing orders raise the premium paid for Grade A milk, which in turn encourages a shift toward the production of Grade A milk for manufactured dairy products. We conclude that marketing orders have contributed to excess milk quality, and that the standard welfare analysis of milk marketing orders has omitted the cost of excess quality.

Conceptual Framework

Milk marketing orders raise the price of Grade A milk through price discrimination and revenue pooling (Kessell 1967; Ippolito and Masson 1978; Cox and Chavas 2001; Sumner and Wolf 1996; among others). Figure 2 illustrates the effects of a marketing order on a milk market. (See Balagtas and Sumner 2005 for the more detailed analytics of milk marketing orders.) In the figure, fluid milk demand is [Q.sub.F]([P.sub.F]), manufacturing milk demand is perfectly elastic at price [P.sub.M], and the marginal cost of producing Grade B milk is [MC.sub.B]. Given that producers could sell milk on the manufacturing market at price [P.sub.M], the marginal cost of producing Grade A milk is [MC.sub.A] = [P.sub.M] + k([Q.sub.A]), where the additional cost of meeting Grade A standards, k([Q.sub.A]), is increasing in [Q.sub.A]. In the absence of milk marketing orders, milk sold to the fluid market gets a premium that just compensates producers for the additional costs of meeting Grade A standards. The price of milk sold on the fluid market is [P.sub.F0] = [P.sub.M] + k([Q.sub.A0]), and just enough Grade A milk is produced to meet fluid milk demand at that price ([Q.sub.A0] = [Q.sub.F0]). Total milk production, [Q.sub.T], is found where [P.sub.M] intersects [MC.sub.B], and all milk sold on the manufacturing market is Grade B milk, [Q.sub.B0] = [Q.sub.T] - [Q.sub.F0].

The marketing order sets the price of fluid milk as a fixed differential over [P.sub.M], [P.sub.F1] = [P.sub.M] + D(D > k), thus reducing fluid milk consumption to [Q.sub.F1]. The price discrimination scheme is implemented through revenue pooling, which pays a uniform "blend" price for all Grade A milk sold in fluid and manufacturing markets, where

(1) [P.sub.blend] = [P.sub.M] + ([Q.sub.F]/[Q.sub.A])D.

Grade A milk production, [Q.sub.A1], is found where [P.sub.blend] exceeds [P.sub.M] by the additional cost of meeting Grade A standards: [P.sub.A1] = [P.sub.M] + k([Q.sub.A1]). Total milk production is still [Q.sub.T], and Grade B milk production is [Q.sub.B1] = [Q.sub.T] - [Q.sub.A1]. Thus, the marketing order raises the price of Grade A milk, inducing some producers to shift from Grade B to Grade A. As a result, the quantity and share of milk meeting Grade A standards rise, although total production is unchanged and the quantity sold to the fluid market decreases. This stylized model of milk marketing orders is consistent with the data showing dramatic growth in the Grade A share of milk coupled with a reduction in the share of all milk used in fluid products (figure 1).

Previous authors have noted the welfare effects of milk marketing orders. Price discrimination and revenue pooling transfers wealth from fluid milk consumers to Grade A milk producers and manufacturing milk consumers (Ippolito and Masson 1978; Cox and Chavas 2001; Sumner and Wolf 1996). Balagtas and Sumner (2005) showed that rents created by milk marketing orders were dissipated by the costs of compliance with Grade A standards. In figure 2, wealth transferred to Grade A milk producers, ([P.sub.A1] - [P.sub.M])[Q.sub.A1] - ([P.sub.A0] - [P.sub.M])[Q.sub.F0] is partially offset by the additional cost of meeting Grade A standards for milk sold to the manufacturing market, [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]. (1) Balagtas and Sumner show that the cost of excess milk quality comprises a substantial portion of the total social cost of the milk marketing orders.

This paper develops econometric evidence of the impact of milk marketing orders in determining two variables related to milk grade. First, we consider the impact of milk marketing orders on the premium paid for Grade A milk. Then, we discuss how the Grade A premium influences the choice of milk grade by dairy farms and, therefore, the Grade A share of milk.

Marketing Orders and the Premium for Grade A Milk

A dairy farmer will switch from Grade B to Grade A if the discounted expected benefits from switching exceed the discounted expected costs. (2) Thus a key determinant of milk grade is the difference between the Grade A and Grade B milk prices, which we term the Grade A premium. Subtracting PM from both sides of equation (1) yields the Grade A premium