Hysteresis vs. natural rate of unemployment in Brazil and Chile

Abstract

This article examines the hysteresis hypothesis in the unemployment rates of Brazil and Chile using an LM unit root test with two endogenous breaks. The phenomenon is confirmed for both countries. However, the hysteresis hypothesis is able to account for only a small part of the unemployment evolution.

Notes

¹ A third theory of unemployment is described by Phelps (1994). It suggests that most shocks to unemployment are temporary with occasional (but permanent) changes in the natural rate. As a result, the unemployment rate can be defined as a stationary process around a small number of (permanent) structural breaks.

² Another argument for the presence of hysteresis in unemployment has to do with human capital depreciation when an individual is unemployed for a long period of time.

³ Recently, Mikhail et al. (2005) found evidence that both the aggregate and sectoral Canadian unemployment exhibit persistence.

⁴ As usual, we define a k-max to choose k and use the (approximate) 10% value of the asymptotic normal distribution, 1.645, to assess the significance of the last lag.

⁵ We decided not to extend the period due to changes in IBGE's methodology.

⁶ For Chile, we decided to use the series for the metropolitan area because it is longer than the national rate and the figures are quite close.

⁷ An ADF test was performed previously, as a benchmark. Hysteresis was found in both series at a 10% level.

⁸ We also use k-max = 8, but the results don’t change.

⁹ These results don’t change if we use a test with just one break in level and trend.