Search results

The purpose of this paper is to evaluate skilled labor demand determinants in Brazil, considering alternatives explanations: changes in relative wages, non-homothetic technology output growth and skill-biased technical change.

This study relies on a rich and unique matched employer-employee data set for manufacturing sector, from 1996 to 2003. The analysis considers a translog functional form labor demand system estimated using seemingly unrelated regression and instrumental variables to control for possible measurement errors and wages and output endogeneity.

The demand function estimates suggest that: labor demand underlying technology is non-homothetic, research and development investment is biased toward skilled workers, the non-homothetic technology is not skill biased so output changes contributed positively for skilled labor increase, relative wages played a significant role and international trade has little explanatory power explaining labor demand shifts.

This is the first paper that considers alternative explanations for the increase in the demand of skilled workers for manufacturing in Brazil simultaneously: changes in relative wages, output changes with non-homothetic technology, skill-biased technical change and, to a lesser extent, international trade. The study challenges current empirical evidence that considers trade and trade liberalization as the main factor explaining labor demand shifts.

Baumol has argued that problems experienced by cities in providing metropolitan services might be the product of an uneven growth process. Slow productivity growth sectors, he argues, could experience increased real costs and, unless demand for their service is highly inelastic, their outputs could decline severely. Looks at these issues in a framework which allows explicitly both for demand, for examination of the sources of productivity growth, and for substitution between productive factors. This extended framework allows for verification and extension of Baumol’s results. For instance, it is found that with uneven growth the relative price of the non‐progressive sector’s output will rise if that good is normal.

This paper revisits the derivation and properties of the Allen-Uzawa and Morishima elasticities. Using a Swiss dataset, this paper empirically estimates various elasticities both in a dual and primal framework using a production theory open economy model and tests for linear homogenous technology. In addition to reporting elasticity at the mean, the standard practice in the literature, this paper also calculates nonparametric distribution of various elasticities. The paper aims to discuss these issues.

To assess the effect of price change on input, the paper estimates a translog cost function and to assess the effect of quantity change on price, the paper estimates the translog distance function using the data on Swiss economy. The paper estimates Allen-Uzawa and Morishima elasticity both under homogenous and non-homogenous technology using the Swiss dataset of one aggregate gross output and four inputs (resident labor, non-resident labor, imports, and capital) over 1950-1986. Elasticities are reported and compared at the mean as well as explored by looking at the range and nonparametric distribution.

This paper shows that constant returns to scale are easily rejected in this dataset and that the elasticities, both qualitatively and quantitatively, are very different under homogenous and non-homogenous technology. These elasticities can switch from complements to substitutes or vice versa when one moves away from the mean of the sample. The equality of the nonparametric elasticity distributions under homogenous vs non-homogenous technology is rejected in all cases except one.

This paper gives a clear derivation and interpretation of different elasticities as well as demonstrates using a dataset how to systematically go about empirically estimating these elasticities in a dual and primal framework. It shows that linear homogenous technology can be easily rejected and the elasticities, both quantitatively and qualitatively, are very different under homogenous and non-homogenous technology. This paper is also very valuable because it shows that the standard practice of reporting elasticity at the mean might not be adequate and there is a possibility that these elasticities can switch from complements to substitutes or vice versa when one moves away from the mean of the sample.