Econometric testing of unemployment hysteresis in selected CEE countries: Lessons for the Serbian economy

Abstract

The paper investigates the empirical relevance of the unemployment hysteresis hypothesis in the following CEE countries: Bulgaria, Croatia, Romania, Hungary and Slovenia. Monthly time series are considered over the 2004-2015 period. The econometric analysis is based on several techniques. Apart from employing conventional unit root tests, the Fourier ADF test is computed to capture non-linear pattern of the deterministic component, and the Lee-Strazicich test to control for up to two endogenously determined structural breaks. These two tests are also used in the two-step approach to increase testing efficiency. Finally, to allow for more flexible treatment of the unemployment rate dynamics, the ARFIMA models are estimated. The unit root presence was clearly detected in Hungary and Slovenia, supporting the empirical validity of the unemployment hysteresis hypothesis and suggesting that random shocks have permanent impact on the evolution of unemployment rates. The unit root is rejected in the unemployment rates of Bulgaria, Croatia and Romania, but their trend components were under the long-lasting influence of great shocks. Impulse responses of these variables are more persistent than in purely stationary cases. Results for the analyzed CEE countries which have been part of the EU for a longer time indicate the necessity of implementing demand-side policies in order to reduce unemployment. New EU members in the same region may achieve satisfactory results in lowering their unemployment rates by combining supply-side and demand-side measures. Given the similar institutional background of the Serbian economy, we would expect that our results might be useful in considering the issue of reducing the unemployment rate in Serbia.