[This paper outlines a procedure for estimating a general index of technical change within the context of a quite general production technology. Specifically, when panel data are available for firms in an industry, time-specific dummies can be combined in a nonlinear estimation procedure to yield a general index of technical change that may be both nonneutral and scale augmenting. This approach offers numerous advantages over the traditional time trend representation of technical change. For example, the general index can serve as the basis for analysis of the determinants of technical change. Results for a sample of 30 electric utilities over the period 1951-78 show that the productivity decline of the 1970s can be attributed primarily to sulphur oxide restrictions and secularly declining capacity utilization due to rapidly increasing peak-load demands.]

In a recent paper in this Review, Ernst Berndt and David Wood provide a useful clarification of and complementarity, pointing out that energy and capital can be substitutes in a production subfunction and yet complements in the aggregate production function. Using this theoretical proposition, they attempt to reconcile the econometric findings of energy-capital complementarity with other studies finding energy-capital substitutability. Their reconciliation rests on the finding that those studies finding energy-capital substitutability considered only capital (K), labor (L), and energy (E). By omitting materials (M) one obtains only a elasticity. The elasticity, which allows for the additional substitution between the KLE aggregate and M, can indicate energy-capital complementarity. In fact, KLEM studies generally find energy and capital complementarity. The purpose of this comment is to question whether the net and the gross elasticity distinction provides such a reconciliation among the econometric results. I agree that this explanation tends to reduce the disparity between the original Berndt-Wood elasticity estimate and Griffin-Gregory. Yet, three independent sources of evidence suggest that the omission of M is not a sufficient explanation, and probably not even a major explanation for the disparity of findings. First, the difference between the elasticity in a KLE submodel and the elasticity in a KLEM model depends critically on the elasticity of substitution between M and KLE. Let us adopt the Berndt-Wood notation and consider their mathematical example which demonstrates the possibility of capital energy substitutability as in the Griffin-Gregory study with capital-energy complementarity:

INFORMATION on patterns of structural change during modern economic growth is summarized in Kuznets (1967 and 1971). The use of secular data to identify such patterns, however, must be limited, for relatively few countries have compiled suitable long-term records. Thus, attention has focused upon an alternative data source, namely, intercountry data for evidence on past and future trends in industrial structure (Temin, 1967; Kuznets, 1967, pp. 431-436; Chenery-Taylor, 1968, p. 391; Houthakker, 1965, p. 277). Despite the frequent use of cross sections to infer intertemporal patterns, relatively little empirical and theoretical work has been devoted to the interpretation of cross-section patterns vis-a-vis intertemporal patterns. Exceptions to this rule are Kuh (1959), Houthakker (1965), Temin (1967), Maizels (1963), Kuznets (1971, chapter. 4), and Chenery-Taylor (1968). The strikingly different results obtained by Kuznets (1971) and Chenery-Taylor (1968) concerning the compatibility of intercountry versus intertemporal patterns raise the basic issue which this study addresses.'

In a 1997 paper, Borenstein, Cameron, and Gilbert (BCG) claim that gasoline prices rise quickly following an increase in the price of crude oil, but fall slowly following a decrease. This note estimates an error-correction model with daily spot gasoline and crude-oil price data over the period 1985–1998 and finds no evidence of asymmetry in wholesale gasoline prices. The sources of the difference in results are twofold. First, we use the standard Engle-Granger two-step estimation procedure, whereas BCG used a nonstandard estimation methodology. Second, even using BCG's nonstandard specification, the use of daily rather than weekly data yields little evidence of price asymmetry.

Results of earlier transcendental logarithmic (translog) production function are challenged on the basis of the limited time-series used. The same methodology is applied to a pool of international manufacturing data to test whether long-run price elasticities can be generated with intercountry samples. Both the time-series and international cross-section methods agree that there is some elasticity of substitution between energy demand and non-energy inputs and that energy forecasting should not be based on the assumption that substitutions between energy and nonenergy inputs are trivial. The international method finds, however, that capital and energy are substitutes rather than complements in the long-run and that there is a difference between short-run and long-run effects of substitution. 29 references. (DCK)

This paper estimates a dynamic demand model for liquor in the United States using panel data from 43 states. Because of taste changes over time and between states in liquor consumption, purely time series or cross sectional studies do not elicit reliable price elasticity estimates. This study makes the case for pooling and shows how one can control for individual state effects and endogeneity of the regressors using estimators suited for a dynamic demand model. Our results indicate that the long-run price elasticity is in the -0.7 range. The findings also support strong habit persistence, a small positive income elasticity, and very weak evidence of bootlegging from adjoining states. The magnitude of the long-run price effect suggests that sin taxes can serve not only as an important income source but also as a significant deterrent effect.