Knowledge that Transforms

As has long been recognized in the literature, questions about the effects of unions on wages may be conveniently grouped under two headings, which, following Lewis (1983), I shall refer to as wage gaps and wage gains. The wage gap refers to the difference between the pay a worker would receive were he represented by a union in collective bargaining and the pay he would receive were he not so represented. The wage refers to the difference between the pay a worker actually receives and the pay he would receive in the absence of any unionization. (When negative, as might be the case for a worker who is not represented by a union, the wage gain may actually be a wage loss.) It has generally been agreed that it is not possible to measure union wage gains, and a host of empirical studies attest to a consensus that measurement of the wage gap, while not impossible, is certainly difficult. Estimating the wage gap is particularly difficult in analyses that use aggregate (e.g., industry or occupation) cross-section data: such data typically do not contain measures of the average wage paid to unionized or nonunionized workers (Wu and W, respectively) considered separately.' The most obvious solution to this problem is to obtain aggregate data in which measures of both Wft and W, are available. In the present paper, I analyze union wage effects using just such a data set. The main point of interest that emerges from the analysis is that, contrary to what seems to be a general consensus, it is indeed possible albeit with some stringent identifying assumptions to estimate union wage gains as well as union wage gaps.

Singh, Balvir, Helmar Drost and Ramesh C. Kumar, An Empirical Evaluation of the Relative, the Permanent Income and the Life-Cycle Hypotheses, Economic Development and Cultural Change 26 (Jan. 1978), 281-305. Singh, Balvir A., and Aman Ullah, Consumption Function: The Permanent Income versus the Habit Persistence Hypothesis, this REVIEW 58 (Feb. 1976), 96103. U.S. Department of Commerce, Expenditures, and Wealth, The Statistical A bstrac t of the U.S., 1942. ______, National Product and Income, Historical Statistics of the United States, Colonial Times to 1970, Bicentennial ed., Part I (Washington, D.C., 1975), 225.

ONE of the most puzzling dichotomies in economics is that between the microeconomic theory of consumer demand and the consumption function of macroeconomics. Although there is now a highly developed literature on the specification and estimation of systems of consumer demand equations, this is basically a theory of allocation, i.e., given the level of consumption, how is it to be allocated among the various categories of consumer goods? This literature takes as given the allocation of current income between consumption and savings, and hence contributes very little to the specification of the consumption function. A notable exception to this dichotomy is the Extended Linear Expenditure System (ELES) proposed by Lluch (1973), an intertemporal model of consumer decision making which endogenizes not only the allocation of consumption across commodities, but also the consumption-saving decision, and hence generates a consumption function which is firmly based in microeconomic theory. However, ELES is based on the rather special Klein-Rubin instantaneous utility function, which will be shown below to have quite restrictive implications. In fact, when the problem is set up in terms of the indirect utility function, generalizations of the Klein-Rubin form are easily incorporated into the intertemporal optimization problem. A second generalization of ELES relates to price expectations. Most empirical applications of ELES, for example, Lluch, Powell and Williams (1977), Lluch and Williams (1975) and Powell (1973) have generally assumed static price expectations and have depended upon the continual replanning assumption to justify observed price movement through time. Where price expectations have been considered, it has usually been in an ad hoc manner. Additionally, these models have generally assumed constant (current) nominal interest rate expectations,'which would appear to be inconsistent with an assumption of stationary price expectations, since the current nominal rate of interest will include a premium for inflation. In a previous paper (Cooper and McLaren (1980a)), an initial attempt was made to integrate a model of the formation of price expectations into an intertemporal consumer decision making model at the optimization stage. Two weaknesses of this model were the restriction to the case of a Klein-Rubin utility function, and the inconsistency of the assumptions made about interest rate expectations with those on price expectations. In the following two sections we present a reasonably general model of intertemporal consumer decision making incorporating price and interest rate expectations which overcomes these two weaknesses. Section IV presents the estimating forms based on a particular choice of functional forms, and estimates using Australian data are presented in section V.

T HERE is no professional consensus about the process of aggregate wage inflation. A spectrum of opinion exists between the following extreme poles: (1) That wages are determined instantaneously in markets where participants have rational expectations and can anticipate in their behavior the long-run consequences of any consistent pattern of macro policy making;' and (2) that wages are determined largely by institutional forces including considerations of equity, normal historical wage patterns, union strength, and current bargaining conditions.2 Adherents of these polar positions have little respect for wage adjustment equations of the Phillips (1958) or Phelps (1967)-Friedman (1968) variety in which wages are related structurally to aggregate unemployment rates. These adjustment equations occupy a middle ground in the spectrum and are currently used in large scale econometric models to explain how the effects of a change in demand are distributed into real and price components. Between the poles are several alternative justifications of wage-unemployment relationships. Among these are the following views: That wages are determined as in (1) above except for the existence of long-run contracts;3 that the determination of expectations about future prices can be fairly approximated by a distributed lag on past prices;4 that it is past prices rather than expectations of future prices that are important;5 and that economic conditions are but one of a set of factors to be included in the current bargaining conditions that determine wages.6 Which of these views best describes the process of wage determination is an empirical question, but the question is quite complicated and does not appear to be capable of resolution through a single, conclusive test. Many issues are involved simultaneously and no one has been able to find a set of workable assumptions that can be agreed upon by all as being a sensible way to proceed. The question of wage determination continues to divide macroeconomic opinion more than any other single issue. This paper reports results of empirical work on wage equations in which a strategy of disaggregation has been followed. It is part of a larger project to estimate the dependence of the natural rate of unemployment on the distributions of the supply and demand for labor according to location and occupation. As an estimation strategy, disaggregation can avoid problems of identification and could, in principle, provide a way to distinguish among the many competing hypotheses in this area. In practice, the disagreements are so fundamental and the possible tests so limited that the results can be offered as no more than an extension of the wage equation literature rather than as a resolution of the issue of whether wage equations should be treated as structural relations or, of even greater ambition, what those relations might be. The extension provided follows the direction of Baily and Tobin (1977, 1978) who hypothesized that the rate of wage change in a single labor force group should depend on the unemployment rate of that group and its wage relative to the wages of other groups. The results are interesting for several reasons. First, the wage data that are used in these tests come from a survey not previously used in the estimation of aggregate wage equations. The data are from the National Survey of Professional Administrative, Technical and Clerical Pay (PATC).7 This survey is conducted annually by the Bureau Received for publication September 18. 1981. Revision accepted for publication May 4, 1982. * The University of Wisconsin. The author wishes to thank Gregory Krohn and Bruce Chapman for their excellent research assistance. Helpful comments were received from Paul Gertler and the participants at a seminar at the National Commission for Employment Policy. This research was supported by Grant Number 99-0-2289-50-11 from the National Commission for Employment Policy, and Contract Number 20-06-08-11 from the Employment and Training Administration, U.S. Department of Labor. ' See, for example, Lucas (1973) and Sargent and Wallace (1975). 2See Dunlop (1977). 3See Phelps and Taylor (1977) and Fischer (1977). 4McNees (1979) provides a way to distinguish this position from the one in the subsequent phrase. 5 See Okun (1978) for a summary of studies of this kind. 6See Hicks (1955) and (1974, pp. 59-85). Bureau of Labor Statistics (1980).

Because it concentrates on the co-movements of jointly determined endogenous variables, the traditional analysts of labor productivity does not directly address the question of the causes of productivity change.This problem is solved by a modelling approach in which productivity and other choice variables are assumed to respond optimally to five broad classes of exogenous (causal) shocks.Although these shocks are unobservable to the econometrician, maximum likelihood estimates of their relative importance in the determination of productivity change are obtained.

States. Rather, observed interregional differences in average real wages probably arise from different relative endowments of various heterogeneous labor types. Because these results conflict with findings of most previous studies, comparisons are made with the approaches taken by other investigators. Those comparisons indicate that empirical estimates of interregional differences in the structure of wage and earnings equations are sensitive to (1) the treatment of geographic cost of living differences, (2) the completeness of the specification of the regressors, particularly the human capital measures, and (3) whether part-time workers are included in the sample.

HE purpose of this paper is to analyze the behavior of the Japanese trade balance for the period of 1966-80 using a conventional elasticities approach. However, unlike previous works in this area, our approach emphasizes the following key features of Japan's foreign trade: (a) almost all exports (more than 95% in 1975) are manufactured goods: (b) except for foods (15.2% of imports) imports are either used as inputs in the manufacturing sector or manufactured goods which compete with home products. Consequently, by concentrating on the behavior of the manufacturing sector, we can explain most of the components of exports and imports. In view of this we estimate a model of the manufacturing sector in which firms in the sector use imported fuels and raw materials as inputs and produce outputs sold in both foreign and domestic markets. The demand for fuels and raw materials and the supplies of output (or output prices) in the domestic and foreign markets are all derived from a profit maximizing behavior of the manufacturing sector. Thus, imports of fuels and raw materials are treated as derived demands. The most important advantage of such an approach is that we shed light on interactions between domestic and foreign markets on the one hand, and between exports and imports, on the other. This is in sharp contrast to the traditional demand-for-imports-supply-of-exports approach which estimates each component of the trade balance separately. Another advantage of our approach is that, on the exports side, a simultaneous equation estimation method is adopted. That is, we estimate an export supply (or export price) equation as well as an export demand equation. The importance of this method has been recognized by, for example, Goldstein and Khan (1978). These advantages of the approach enable us to capture a number of important features of Japan's foreign trade. In the analysis of exchange rate effects on the trade balance, the following aspects are emphasized. First, a depreciation of the exchange rate may be partially offset by a change in export price in the opposite direction. The magnitude of this offsetting effect is determined by the impacts on export price of competitors' price and prices of imported intermediate goods. Second, the impact of depreciation on the trade balance may also be weakened because depreciation raises the price of domestic manufactured goods and decreases people's incentive to switch from foreign to domestic manufactured goods. Third, as stated above, imports of fuels and raw materials are derived demands; therefore, price elasticities of these may not be very large in the short run. Among other variables affecting the trade balance, we focus on cyclical variables such as domestic and foreign incomes. There are direct effects of domestic income on the imports of intermediate goods and of foreign income on exports. In addition, we are especially interested in testing whether or not a domestic recession creates a pressure to increase exports. Such a possibility has been studied in the literature on the capacity pressure hypothesis. (See, for example, Dunlevy (1980).) In the following, we first explain the framework of the analysis and estimate equations for the components of the trade balance. The estimates are then used to assess quantitatively the importance of each of the mechanisms discussed above. It is shown that all these interrelationships between various components of the trade balance are important. We then proceed to present some simulation results that show the effects of exchange rate, cyclical variables, and input prices on the trade balance. They indicate that exchange rate changes exert a quick but not very large impact on the trade balance. An increase in the price of fuels leads to an immediate and significant worsening of the trade balance. Most important, cyclical variables have crucial effects on the trade balance. It is Received for publication April 5, 1982. Revision accepted for publication December 10, 1982. * Osaka Universitv. This paper is a substantially revised version of the first half of chapter 4 of the author's Ph.D. dissertation at M.I.T. The author would like to thank Stanley Fischer, anonymous referees of this Review, and the participants in seminars at the University of Tokyo, the Bank of Japan, Economic Planning Agency, and the Ministry of International Trade and Industry (Japan) for useful comments on earlier versions of the paper.

The response of nominal interest rates to price increases has been intensively studied since the onset of substantial inflation and historically high interest rates in the late 1960s. Early empirical investigations (Yohe and Karnosky (1969), Gibson (1972), and Pyle (1972)) tested the Fisher hypothesis that nominal rates react one-for-one to changes in the expected inflation rate. Paralleling results in the wage-price literature, these studies often found that nominal rates appeared to adjust too little to imply neutrality with respect to expected inflation. Theoretical work by Mundell (1963) and Tobin (1965) that incorporated wealth effects, by Sargent (1972) that considered an extended macromodel, and by Darby (1975) and Feldstein (1976) that allowed for income tax effects, however, suggested that additional variables, like measures of fiscal and monetary policy, may be relevant and that nominal rates may change by either more or less than unity in response to a one unit change in expected inflation. Two puzzling features of the response of nominal rates to expected inflation remain. First, even when other factors are included, estimates of that response are unstable for the postwar period (see Cargill and Meyer (1977)). Second, empirical studies to date have failed to produce any statistically discernible impact of expected inflation on interest rates during the 1950s (see Cargill and Meyer (1974) and Cargill (1976)). Similarly striking is the sizeable residual autocorrelation exhibited by nearly all estimates that include the 1950s. These symptoms of model misspecification imply that an important factor may have been omitted from previous models. A reduced form model of nominal interest rates that allows for an additional impact, that of factor supply shocks, is advanced in section II. The empirical tests in section III show that, once this variable is included, a significant relation between interest rates and expected inflation emerges for the 1950s. We also show that our specification is stable over the entire post-Accord period. Section IV concludes.

Economies, unpublished mimeograph, Jerusalem, The Falk Institute for Research in Israel, 1962. Lindley, D. V., and A. F. M. Smith, 'Bayes Estimates for the Linear Model, Journal of the Rox'al Statistical Soc ietv 34, Series B, no. 1 (1972), 1-18. Mundlak, Yair, On the Pooling of Time Series and CrossSection Data, Econometrica 46 (Jan. 1978), 69-86. Pakes, Ariel, Economic Incentives in the Production and Transmission of Knowledge: An Empirical Analysis, Ph.D. Thesis, Harvard University, 1978. , 'On the Asymptotic Bias of Wald-Type Estimators of a Straight Line when Both Variables Are Subject to Error International Review 23 (June 1982), 491-497.