Page 43 - IDEA Study 3 2015 Duchodova studie

TABLE 5.1 Employment probability model: estimated coefficients on unemployment p-values in parenthesis denote statistical significance of estimated parameters Overall, unemployment's explanatory power and its interaction with age on the two models we estimated for the Czech Republic is relatively low (the adjusted R2 statistics are almost unaffected). Given these results and the fact that the average unemployment rate in the Czech Republic (5.9%) is the 4th lowest within the EU countries (with the exception of Austria (4.8%) and Germany (4.7), all other countries in our study have a higher unemployment rate, for example Denmark (6.2%), Sweden (7.8%), and the Netherlands (7.2%)), it does not seem that the incidence of unemployment in the Czech Republic represents a significant omitted variable in our simulations. Going beyond the evidence provided by our analysis, empirical literature does suggest, however, that early retirement has negative long-term effects on health, cognitive ability and socio-economic conditions (see Mazzonna and Peracchi (2012)). Comparing the Czech Republic in 2007 and 2013 Finally, we explore recent developments in the Czech Republic. In particular, we estimated model (1) for the Czech Republic on the 2013 sample. We used the estimated coefficients and personal characteristics in 2007 to predict hypothetical employment and retirement probability in year 2007. Graph 5.5 shows the real profile in the Czech Republic in 2007 (denoted as Real CZ2007 and marked by “+”), the real profile in the Czech Republic in 2013 (denoted as Real CZ2013 and marked by “o”), and the simulated profile in the Czech Republic in 2007 using estimated parameters from 2013 (denoted as Simulated CZ2007 and not marked). There is very visibly no effective difference between the Simulated CZ2007 and Real CZ2013 profiles, which implies that the changes observed in employment and retirement between 2007 and 2013 in the Czech Republic were not driven by changes in personal characteristics, but by the institutional setup. It should also be noted that in the case of men, the section of employment profile bounded by ages 58-62 moved from right to left between 2007 – 2013, probably due to a one-off influx, or a larger proportion of men retiring early, due to the incentivising effect of the “small pension reform” in 2011, as discussed at the end of section 1. Czech Republic 2013 Women Men Sweden 2013 Women Men Unemployment rate (in %) -.0337961 -.0050227 .0157916 -.0082 (0.000) (0.484) (0.364) (0.634) Unemployment rate * (age – 50) .0017539 .0002334 -.0010848 -.0011167 (0.000) (0.656) (0.366) (0.344) adjR2 .43 .41 .44 .44 41