Abstract—Real gross domestic product (GDP) and unemployment rate (UR) are basic indicators of business conditions, which are correlated with each other. It is important to understand the dynamic relationship between quarterly GDP and monthly UR. To analyze this dynamic relationship, we propose a Bayesian regression method to estimate the dynamic dependence of a stationary component of GDP on a stationary component of UR. First, we extract the stationary components from the original time series for GDP and UR using a set of state space models. Then, we construct a set of Bayesian regression models, with each model having a time-varying coefficient. As an application of our proposed approach, we analyze the dynamic relationship between the stationary components of GDP and UR in Japan, using the separate URs for men and women. Overall, we find that the movements of UR lead those of GDP by a few months during expansion phases. In contrast, movements of the UR lag those of GDP by a few months during recession phases. Moreover, there is negative correlation between the stationary components of GDP and UR. Such negative correlation is stronger in the recession phase than in the expansion phase. We conclude that the UR can.
Index Terms—Bayesian modeling, gross domestic product, unemployment rate, Japanese business cycle.
Koki Kyo is with Obihiro University of Agriculture and Veterinary Medicine, Inada-cho, Obihiro, Hokkaido 080-8555, Japan (e-mail: kyo@obihiro.ac.jp).
Hideo Noda is with Tokyo University of Science, 1-11-2 Fujimi, Chiyoda-ku, Tokyo 102-0071, Japan (e-mail: noda@rs.tus.ac.jp).
[PDF]
Cite: Koki Kyo and Hideo Noda, "A Bayesian Approach for Analyzing the Dynamic Dependence of GDP on the Unemployment Rate in Japan," International Journal of Modeling and Optimization vol. 8, no. 1, pp. 55-61, 2018.
