Journal Article Statistical cost Analysis Re-Revisited: Reply Get access James M. Griffin James M. Griffin University of Houston Search for other works by this author on: Oxford Academic Google Scholar The Quarterly Journal of Economics, Volume 96, Issue 1, February 1981, Pages 183–187, https://doi.org/10.2307/2936150 Published: 01 February 1981

I. Introduction, 107.—II. Pseudo-data generation from a process model, 109.—III. A comparison of time series and pseudo data, 112.—IV. The translog cost function, 114.—V. Statistical cost function estimation, 116.—VI. Substitution and price elasticities: Tests for an energy aggregate, 119.—VII. Applications for long-range analysis, 122.—VIII. Summary, 126.

[This paper explores a new approach to the estimation of a joint production technology. Pseudo data, which are obtained by solving a petrochemical process model for alternative relative prices, are used to estimate a price possibility frontier with 3 inputs and 6 outputs. Unlike traditional data sources, pseudo data are not constrained by historical price variation, technologies, and environmental controls. As an econometric exercise, the approximation of the process model's detailed piecewise linear production surface by a single equation "generalized" functional form, the translog, raises a host of interesting empirical and methodological questions.]

Over the period 1979-85, the petroleum industry offers a unique test of the agency theory of corporat e restructuring. A panel data set for twenty-five firms indicates con siderable support for a hybrid free cash flow model in which explorat ion expenditures appear guided by neoclassical profitability measures , but at the margin, cash flow also exerts an independent influence. The magnitude of the parameter estimates suggests that the free cash flow hypothesis does not fully explain the gains from restructuring.

THE efficiency implications of telecommunications service pricing are analogous to a puzzle whose parts are beginning to be assembled. Recently, Mitchell (1978) has examined the efficiency implications of alternative local service pricing approaches. He contrasts the welfare effects of a flat monthly rate with measured service pricing under an optimal two part tariff with an access line charge and a per call charge. Daly and Mayor (1980) have examined the efficiency implications of free directory assistance, contrasting it to marginal cost pricing and found large welfare losses. The efficiency implications of long-distance telecommunications service, which in 1975 accounted for almost one-half of the Bell System revenues,' remains to be placed in the puzzle. The primary purpose of this paper is to examine the efficiency implications of the pricing of local service vis-h-vis conventional long-distance service referred to as Message Telecommunications Service (MTS), which constitutes the bulk of long-distance service. Excluded from this analysis are WATS and private line service. As demonstrated by Rohlfs (1979), substantial cross subsidization occurs between local service, which is priced approximately 50% below marginal cost, and long-distance, which is priced two to three times above marginal cost. Nevertheless, cross subsidization need not imply inefficiency in a second best pricing framework for several reasons (Baumol and Bradford, 1970). First, subsidization of local service can be justified due to access externalities, arising because the value to potential callers is not internalized in the subscriber's price. Second, even in the absence of access externalities, if the price elasticities of both services tend to be very inelastic, cross subsidization may have negligible efficiency effects. The optimal second best pricing approach depends critically on the deviation of price from marginal costs, the extent of the access externality, and the relative price elasticities for local and long-distance service. In this exercise, the welfare effects are particularly sensitive to the price elasticity of MTS service, thereby justifying the empirical focus on the price elasticity of MTS. The MTS demand relationship estimated in this paper contains advances in several respects. First, unlike previous pure time series or cross sectional studies,2 the data set consists of pooled quarterly data (1966 to 1978) for five southwestern states. The analysis of intrastate long distance demand3 offers a much more robust data source than national interstate demand owing to the limited price variation in the latter. The pooled model features polynomial distributed lags and an error structure which corrects for both autocorrelation and heteroskedasticity. Also, the model includes a unique and superior measure of television advertising effects, an index of gross rating points, reflecting the actual frequency with which television advertising is viewed by the public. The use of gross rating points avoids the distortion implicit in the substantial volume discounts reflected in expenditure data.4 The subsequent section outlines the simple theoretical model, which is the basis for econometric estimation and the pooling techniques. Section III reports the empirical results. In section IV, we examine the price elasticity implications for optimal pricing of MTS service. Section V recapitulates the major conclusions and suggests directions for future research. Received for publication September 29, 1980. Revision accepted for publication April 20, 1981. * University of Houston. The author wishes to acknowledge the collaboration of Bruce Egan in the econometric modelling section of this paper. In addition, numerous helpful comments were provided by George Daly, Thomas Mayor, Jeffrey Rohlfs, William Taylor, and an anonymous referee. I See Rohlfs (1978), table III-1. 2 For a review, see Taylor (1980) and Lowry (1976). 3 Taylor notes that from the view of demand, the distinction between interstate and intrastate MTS is purely artificial. See Taylor (1980), p. 97, and table 5.1 4 Comanor and Wilson (1967) note that volume discounts may give rise to increasing marginal returns to advertising expenditures.

straints that are imposed, the price vectors chosen to generate the data, and the functional form used for the cost function approximation. It is also fair to say that if input-output coefficients are the ones of interest, since these are obtained for each of the points used in the generation, it may be better to seek some approximating functions for these directly rather than seek an approximating function for the cost function and then derive the input-output coefficients. All these comments do not imply that the pseudo data approach should be given up. It is, however, important to sort out the aims of the analysis. There are some problems (like studying the effects of changes in environmental regulations) where one cannot get any answers from time series and one has to use the process analysis models that have detailed specification of technology and the constraints. But for this the appropriate thing is to do a simulation analysis of the process model itself and not seek a single equation approximation of the complex technology. What distin,guishes Griffin's approach from the garden variety simulation analyses of process models is this distillation in a single equation.

The pseudo data approach to the joint production of petroleum refining and chemicals is described as an alternative that avoids the multicollinearity of time series data and allows a complex technology to be characterized in a statistical price possibility frontier. Intended primarily for long-range analysis, the pseudo data method can be used as a source of elasticity estimate for policy analysis. 19 references.