INDUSTRIAL SOLOW RESIDUALS: AN EMPIRICAL STUDY FOR BRAZIL

Abstract

In this paper, we study the cyclical properties of total-factor-productivity (TFP) rates (“Solow residuals”) for the Brazilian industry. In doing so, we check if methodological differences related to the construction of TFP measures can affect their statistical properties. The results obtained show that such properties are sensitive to the methodology employed. In particular, in the case of our sample, we obtain three main results: (i) corrected productivity measures are less correlated with industrial production measures when compared to non-corrected measures; (ii) corrected measures, where correction is done through the use of industrial energy consumption, are less volatile than industrial production measures, while the opposite is true for corrected measures where correction is done through capacity utilization rates; (iii) corrected measures where correction is made through the use of capacity utilization rates are the only measures which obey the Hall-Evans Invariance Property and they can be characterized as exogenous. These results show the importance of methodological issues related to the construction of productivity measures as well as the importance of capacity utilization rates in the case of business-cycle horizons.