Taxa de desemprego no Brasil em quatro décadas: retropolação da PNAD contínua de 1976 a 2016

Abstract

In this paper we employ a retropolation method to Brazilian Quarterly National Households Survey unemployment data to generate a time series for a longer time span covering the period from 1976 to 2016. Our retropolation method consists on multivariate state space models and allows us to estimate Brazilian unemployment rate. The retropolation is performed using the information from DIEESE’s Employment and Unemployment Survey and Annual National Households Survey National Survey as well. Our work contributes to the literature by generating a series that covers forty years of Brazilian Macroeconomic History on a quarterly basis. To the best of our knowledge an exercise like this was not performed in any other work. Our estimates suggest that peak of unemployment was reached at the last quarter of 2016 for the analyzed time span.

Keywords
Labor market; Unemployment rate; State-space models; Kalman Filter