Knowledge that Transforms

Christopher Green, Alfred Tella, Effect of Nonemployment Income and Wage Rates on the Work Incentives of the Poor: Reply, The Review of Economics and Statistics, Vol. 52, No. 4 (Nov., 1970), pp. 449-450

T HE theory of induced invention along Kennedy-Weizsacker lines has recently been incorporated into growth models with factor-augmenting technical progress by Samuelson [9], Drandakis and Phelps [4], Amano [1] and Fellner [5]. In these models they have shown that under the assumption of a proportional saving function, the condition that the elasticity of substitution (C-) be less than unity is sufficient for global stability of the long-run equilibrium. In this paper we attempt to generalize their results by employing a more general saving function and to examine the role of saving in a growth model of this type. We assume that the marginal propensity to save out of profits (s1) and out of wages (s2) are two different constant proportions. It turns out that, in the presence of a C-ambridge saving function, the condition C< 1 is not always sufficient to assure global stability of Kennedy's economy. The stability of the system, in general, depends upon not only the elasticity of substitution, but also the saving propensities. Furthermore, the position of the invention possibility frontier as well as its shape is an important element in stability analysis. This is due to an important feature of the generalization in which changes in income distribution between profits and wages resulting from factor substitution along the isoquants, and technological substitution along the invention possibility frontier, are taken into account in determining the changes of the rate of growth of capital. In section I we set up the model and derive the required dynamic equations. In section II we analyze the existence and global stability of the steady-state growth path. In particular, a local stability condition will be derived. Finally, in section III, after summarizing the principal findings, we interpret our stability condition and compare it with the condition which we found elsewhere in a model with exogenous Harrod-neutral technical change [2].

CEVERAL years ago I proposed a theory to explain how the supply in a developing country is determined which allows for factors other than the traditionally assumed control by a central bank [6]. Its validity was tested with an econometric quarterly model for Pakistan, 1953-1961. The passage of time makes it possible to reappraise the model which is the purpose of this paper. With the benefit of hindsight and reflections about economic model building, the reappraisal suggests that what I proposed as a general description of in a developing economy, in fact, more closely resembled a stationary one. Although some differences of opinion exist, economists agree that Pakistan's development during the 1950's was far less dynamic, albeit certainly not nonexistent, than what has since occurred.' My purpose, however, is not to propose a revised theory of how the supply is really determined in a developing economy, but rather to appraise quantitatively the extent to which the model's predictive capacity after 1961 differs from its performance during the original period. Though this is a more modest objective than attempting to reconstruct a theory which would incorporate the important developments of the 1960's, appraising the predictive capacity of econometric models during periods beyond that which was used for the original fitting, is a worthwhile endeavor which economists should more frequently practice. The postwar era has seen a tremendous growth in econometric studies. Generally, they have two features in common. First, although they really belong to the realm of economic history because the objective is to obtain the best fit to explain a set of for some past period, the hope is usually present that the estimated relationships should be useful to understand (i.e., predict) future changes, given a new set of new exogenous conditions. Second, except for narrowly conceived forecasting models which are often revised annually, we rarely have an opportunity to judge whether history repeats itself in the sense that the model's predictive capacity remains high for a new period beyond the time for which the original parameters were estimated.2 As has been suggested elsewhere, the infrequent publication of the predictive performance of econometric models is due to the fact that few investigators are willing to choose a specification on the basis of less than a complete set of the available data [3, p. 11].' Here, I hope to make amends for not having practiced in my original article what is here preached about the testing for predictive performance. My earlier article showed that the predictive performance of the model's structural equations was generally good and, furthermore, the model's capacity to explain the supply was demonstrated to be superior to a simple money multiplier model where certain controllable assets of the central bank were used to predict the supply. One simulation experiment which used the original initial conditions and the values of the exogenous variables produced a new set of predicted values which suggested that the model was quite stable, at least over the 34 quarters between July 1953 and December 1961, for which the parameters were estimated. In this paper I test, during 24 additional quarterly observations for the years 1962

T HIS paper reports results of a time series study of recent wage behavior in manufacturing. The principal conclusions reached are that labor market variables are of prime importance in determining wage movements and that the wage guideposts did retard wage advances between 1962 and 1966. A novel feature of the paper is that it utilizes wage and unemployment series not previously used in studies of wages. The paper has three main parts. Section I briefly examines recent money wage literature. Some measurement issues are discussed in section II and the main regression results appear in section III.

In the February 1967 issue of the Review we presented estimates of a relationship between changes in help-wanted advertising normalized for growth in the labor force, and changes in the unemployment rate, and new hires, and a dummy variable reflecting the phase of the business cycle [11]. Two notes by Burch and Fabricant [2] and Gujarati [3] have expanded upon our findings. Each paper presents an alternative model and finds that in its own model the relationship between the unemployment rate and the amount of help-wanted advertising is not stable over the period 1951 through 1966 or 1968. The Burch-Fabricant paper tests for the shift in the coefficient of the reciprocal of the unemployment rate before and after 1957. The Gujarati paper tests for the difference in the slope of the unemployment coefficient in ten different business cycle phases from 1951 to 1968. Both papers seem to explain a greater fraction of the variance of the dependent variable than does our model. However, our dependent variable is the change in the normalized help-wanted advertising index; in the other two papers it is the level. By the usual standards, our ]R2 of 0.82 is at least as good as the somewhat higher correlations produced in the level equations. When the two tests suggested respectively by Burch and Fabricant, and Gujarati were applied to our model, the coefficient of the unemployment rate change variable proved to be stable. We first added a variable which is zero up to 1957 and equal to the change in the unemployment rate after 1957; in effect, the regression coefficient of AU is allowed to change its value in 1957. If this variable is estimated along with the other variables in our model, a simple t test will indicate whether the coefficient did in fact change after 1957. Our original equation for the period 1951-1966, second quarter, is reproduced as (1) below. The estimated equation with the dummy variable added is shown as (2):

for using input-output tables to make intercountry comparisons, including a type of graphic presentation, and to use these procedures to compare the economic structure of the United States with that of the Soviet Union, suggesting possible causes of differences. The United States table used is the 200-sector table for 1947.' Although this is earlier than other available tables, it is larger and thus more flexible to align with another table. The U.S.S.R. table used was derived by Vladimir Treml after extensive estimation of the portions missing in the published version.2

CONOMETRIC studies of absolute prices have generally ignored the corporation income tax as a possible explanatory variable; the public finance literature on the other hand contains much speculation that pricing might be affected by the tax, though it offers little empirical evidence. This paper draws on work in both fields to develop and test tax-price hypotheses against data from the textile and rubber industries. Analysis of price change involves linking the influences of supply and demand. The standard sources of hypotheses about the links are the theories of competition and monopoly. In both cases, the parameters of cost and utility ultimately determine price and quantity, but the emphasis differs as to what are the important decisions of firms in the process. The competitor is a price-taker who decides what quantity to supply; the monopolist may be either a price-taker, or a price-maker whose decision is what price to charge. It seems a common fact of observation that most industrial firms are price-makers who supply the resulting demand, and not vice versa, so for an analysis of prices of such products, monopoly theory is likely to be more fruitful of hypotheses than competitive theory. Our procedure therefore is to regard the industry as a monopolistic firm, and to deduce what industry data would show if this firm followed alternate simple pricing rules: profit maximization, target rate of return, percentage markup over variable costs, and sales maximization subject to a minimum rate of profit. Each rule is formulated in both taxfree and tax-influenced form. Textiles and rubber were chosen both for reasons of data convenience, and because they differ considerably in structure. Yearly data for 1924 through 1962 were drawn from the Treasury Department's Statistics of Income for Corporations, supplemented by the Bureau of the Census Annual Survey of Manufactures and Census of Manufactures, and by Bureau of Labor Statistics indexes of wholesale prices, wages and employment. The time period and data sources were dictated by the need to include both financial and operating information, and periods of both high and low tax rates. Several adjustments were made to the data, the most important of which was to remove, in the interests of sample homogeneity, Miscellaneous Plastics from rubber and Apparel from textiles. Hence the industries are somewhat narrower than those of the published two-digit Standard Industrial Classification. The war years 1942 through 1946, 1951 and 1952 were omitted from all regressions. Lags, differencing, and lack of early-year information restricted the final sample to 1928 through 1962 for rubber, and 1927 through 1962 for textiles. Yearly data for industry aggregates have obvious drawbacks for investigation of this sort. The aggregates are quite broad, and if pricing patterns are discernible in them, the patterns must be generated by average behavior over heterogeneous groups of products. Yearly observations provide little basis for studying lags. But it is an unfortunate reality that most econometric work on prices has been at even higher levels of aggregation than this. One can only caution that all results, including these, must be assessed in the light of the data weaknesses.

T HIS paper presents an econometric analysis of the behavior of wages and prices in the American steel industry and the experience with steel imports during the 1950's and 1960's. The results presented are a portion of a larger, and as yet unfinished, effort to explain profit in the steel industry by estimating an equation for each economically meaningful component of the industry's income statement and then combining the equations to form a complete system. Modern empirical investigation of the determinants of wages, prices, and imports has developed in two distinct contexts. First, in response to widespread public concern over rising wages and prices during the 1950's, economists derived and tested a series of new formal models (generally embodied in a single central regression equation) to describe the processes at work. As time has passed more models have been proposed, early formulations have been elaborated and extended, and more data have become available for testing. In general, however, these models have stayed at the economywide level, and little has been done to disaggregate them by industry classification. Second, the wage-price subsections of large macro-econometric models have attempted to provide a complete explanation of the inflationary process, but as with the single equation studies, there has been little analysis of the mechanism in any individual sector. This paper draws upon the theories and models which have been developed for economy-wide studies, modifies them where necessary, and applies them to the steel industry. The next three sections present the formulation and estimation of equations for the steel wage rate, the wholesale price index for steel, and the ratio of imports to domestic shipments. Then, with the aid of the estimated relationships, some short-run projections are made; finally, the analysis is summarized.

The problem was reduced to the apparently simple one of estimating the elasticity of substitution between capital and labor in the production functions of differing industries and determining whether these elasticities varied widely or were basically the same. After a lapse of eight years, however, the existence of factor reversal is still an unresolved question. In his 1967 summary article Marc Nerlove could write: