Analysis of job changes in Poland using the Bayesian method
Wioletta Grzenda, Warsaw School of Economics
The primary objective of the work is to identify and estimate the impact of key factors influencing career mobility in Poland in the years 1950-1988 by using Bayesian methods. The work starts with an overview of numerous studies concerning employee mobility both in Poland and other countries. Next, the characteristics of the data set used in this work, research hypothesis and potential determinants have been presented. This provides basis for the discussion of methods used to estimate a Bayesian Poisson regression model applied in this work. The model has been estimated using Markov chain Monte Carlo method with Gibbs sampling. It allows determining factors that have a significant impact on the number of employment periods, also in the case of the small size of the third age group. Higher career mobility has been observed in big cities than in rural areas, moreover this difference is even larger for young persons. The education level has turned out to be insignificant in the first age group, while in case of respondents aged over 30 years, career mobility of respondents who had higher education was larger. The sex of the respondent plays an important role in career mobility, but the scale of this impact is also dependant on the age. Moreover, lower career mobility has been observed for young persons who did not have any children or who had a single child. Finally, in every age group, the respondent’s age when entering the workforce had significant impact on career mobility. The results of employment mobility studies for the state-socialism period of the Polish economy are frequently divergent. This clearly shows that social, demographic and economic processes of the analysed period have not been fully investigated yet. Bayesian models, used in this work, have enabled more in-depth analysis of these processes.
Presented in Poster Session 1