Vertical economies: the case of U.S. electric utility industry, 1983-87.

• Author: Hayashi, Paul M.

1. Introduction

   The major argument supporting public utilities to operate as monopolies is based on the concept of natural monopoly. Kahn states:

   . . ., the public utility industries are preeminently characterized in important respects by decreasing unit costs - or increasing returns - with increasing levels of output. That is indeed one important reason why they are organized as regulated monopolies: a "natural monopoly" is an industry in which the economies of scale - that is, the tendency for average costs to decrease the larger the producing firm - are continuous up to the point that one company supplies the entire demand. It is a reason, also, why competition is not supposed to work well in these industries [15, 123-24].

   When economies of scale exist, the average cost of producing a unit of a output is lower if the entire demand is supplied by a single producer rather than supplied by many competitive producers. Since the average cost of the product is at its lowest when the entire demand is supplied by monopoly, it is in the public's interest to allow one public utility to operate in a particular market. The public utility is regulated to pass the benefit of low cost to consumers. Consumers benefit from regulated monopoly rather than competition since regulation is substituted for competition in public utilities. The concept of natural monopoly allows utility companies to operate as regulated monopolies.

   Following the growing trend in the U.S. to substitute competition for regulation, it has been suggested that if economies of scale have been exhausted in the generation stage of the electric industry, this stage could be deregulated by removing it from the existing companies and placing it in a competitive market environment. There have been some attempts to move towards deregulation in the electric utility industry. One such attempt is the Public Utility Regulatory Policies Act of 1978 (known as PURPA). PURPA requires electric companies to purchase power from both small producers and industrial cogenerators at avoidable cost, which is equal to the utility's cost of internally generating that unit. Another attempt is made by the Federal Energy Regulatory Commission (FERC). The FERC issued a notice proposed rulemaking entitled "Regulation Governing Independent Power Producers" which questioned natural monopoly in the electric generation phase.

   Although transmission and distribution continue to be considered natural monopolies (at the plant level) today there is no such consensus for generation. In fact, empirical studies question the proposition that larger unit size generators have lower production costs.

   Furthermore, even if generation technology were to favor larger units, and therefore fewer plants, this fact would not be sufficient to conclude that a single plant can supply the market at least cost, i.e., that generation technology is a natural monopoly. Advances in transmission technology and the expansion of the transmission network have increased the size of electric power markets. As a result, the capacity needed to serve any particular market may be much greater than the capacity of a single generation plant [8, 23].

   The FERC proposed rules are based on the belief that, ". . ., while the generation sector of the electric industry was assumed to be a natural monopoly when the FPA was enacted, that assumption now appears questionable for many markets [8, 25]."

   PURPA and the Regulation Governing Independent Power Producers, however, ignores the economies due to the vertically-integrated structure of this industry. The merit of these attempts should be examined on the basis of not only the economies of scale in the generation stage but also the economies due to vertical integration. In other words, cost reduction by creating competition at the generation stage can offset cost reduction due to vertical integration. This paper examines the effect of vertical integration on the cost of electric supply.

   We begin with a review of literature on economies of scale and vertical integration in the electric utility industry. The most well-known study on economies of scale is done by Christensen and Green [5](1) in 1976. They investigated economies of scale in the generation stage of electric supply and concluded that economies of scale existed in 1955 but were exhausted in 1970 and that most firms operated in the horizontal part of their average cost curve in 1970, implying constant returns to scale.

   Heuttner and Landon [13](2) investigated scale economies in different cost categories and found that scale economies existed only in sales expense. Henderson [12](3) investigated scale economies for the generation stage, the transmission and distribution stage of electric supply, and the separability of the transmission and distribution stage of electricity from the generation stage. In order to estimate the cost function of the transmission and distribution stage, the overall marginal cost of electric generation was estimated and used as the transfer price of generated electricity. He found that no significant scale economies existed in the generation stage while substantial scale economies existed in the transmission and distribution stage of electric supply. In addition, he rejected functional separability of the transmission and distribution stage from the generation stage and concluded that the price of generated electricity affects the input ratios in the transmission and distribution stage.

   Kaserman and Mayo [16](4) investigated the existence of vertical economies by applying the concept of multiproduct cost economies. According to their study, vertical economies exist if the cost of electric supply by a vertically-integrated firm is lower than that of electric supply by separate firms specializing in each function of electric supply. Their study found the existence of vertical economies in the U.S. electric industry. Kaserman and Mayo also concluded that cost complementarities existed between the different stages and that economies of scale were exhausted in each stage of electric supply. It is interesting to note that the reason they found this result is that due to cost complementarities, multistage economies are extended beyond the exhaustion of stage-specific economies.

   Gilsdorf [9](5) employed a multiproduct translog function. The major results of his empirical work are inconsistent with the cost complementarity hypothesis, although they showed increasing returns to scale in each stage.

   These studies appear to be inconclusive concerning the issues of economies of scale and cost complementarity between different stages of electric supply. The issue of cost complementarity is crucial to vertical integration because if cost complementarity does not exist, vertical integration will not reduce the cost of electric supply. These results necessitate further investigation of these issues.

   The objective of this paper is to take a different approach to investigating the returns to scale and vertical economies by specifying the cost function of each stage and their relationships in the context of vertical integration. The first section develops a cost function of a vertically-integrated firm. This will be followed by the empirical estimation of the cost functions developed in the first section using the data from the U.S. electric utility industry 'over the 1983 to 1987 period. Finally, in the third section, a discussion of the implications of these results will be presented.

2. The Model

   The supply of electricity involves two activities; generating electricity (the upstream activity) and transmitting and distributing electricity (the downstream activity).(6) There are two possible ways to perform these activities. One possibility is to have each activity of electric supply performed by a separate firm. One firm (the upstream firm) can generate electricity and sell it as an intermediate good to another firm (the downstream firm) which transmits and distributes it to the final customers. Another possibility is to have one firm integrating both activities vertically and performing both. If a firm chooses the vertically-integrated form of production over the non-integrated form, there must be some incentive to do so. The key to answering this question is the separability of the cost function of the two activities. This implies decentralized decision making [3], which will be explained in the following section.

   To investigate separability between the cost functions, the cost function of electric supply is specified as:

   C = C(Q, [w.sub.k], [w.sub.l], [w.sub.p]) (1)

   where

   C = total cost of electric supply,

   [w.sub.i] = the price of the ith input,

   k = capital,

   l = labor,

   p = generated electricity,

   Q = delivered electricity.

   For the cost function of transmission and distribution to be separable from generating electricity, the capital-labor ratio of the transmission and distribution activity is independent of the price of generated electricity. In other words, "these planners would design the labor and capital requirements of the distribution network in the same way, irrespective of the marginal cost of generation. If it is false, planners design the network to provide better control of line losses, for example, if the price of electricity is too high [12, 81]." For the separability between the cost functions of the different stages, the following condition must be satisfied:

   ([Delta]/[Delta][w.sub.p]) [([Delta]C/[Delta][w.sub.k])/([Delta]C/[Delta][w.sub.t])] = ([Delta]/[Delta][w.sub.p])(K/L) = 0 (2)

   where

   K = capital input

   L = labor input.

   If the above condition is satisfied, the cost function of electric supply can be specified as:

   C = C[[C.sup.D]([w.sub.k], [w.sub.t], Q), [C.sup.U]([w.sub.k], [w.sub.t], [w.sub.f], K)] (3)

   where

   [C.sup.D] = cost of transmission and distribution,

   [C.sup.U] = cost of generation,

   X = generated electricity,

   [w.sub.f] = the price of fuel.

   An important implication of...