Page 42 - IDEA Study 3 2015 Duchodova studie

.75 .75 .5 .5 .25 .25 GRAPH 5.4 Retire ent p obabil ties in t 00 m r i c lic and Austri a in 20 3 5 5 0 Proportion WORKING by age: AT - WOMEN 55 Proportion WORKING by age: AT - MEN 70 5 6 0 6 5 7 he Cze h Repu b 1 0 5 Real CZ Simulated CZ Real AT 11 .75 .75 .5 .5 .25 .25 Real CZ Real AT Simulated CZ 00 50 55 60 65 70 50 55 60 65 70 Proportion RETIRED by age: AT - WOMEN Proportion RETIRED by age: AT - MEN The intermediate case simulating an institutional shift from the Czech Republic to the Netherlands' policy context is shown in the Appendix. The Impact of Unemployment For the Czech Republic and Sweden, we were able to match district specific unemployment indicators. In particular, we constructed district-level unemployment rates, gender specific unemployment rates, and the share of 50+ year-olds in the district population. Since all these variables turned out to be highly correlated, we estimated model (1) augmented only by total unemployment rate and unemployment rate interacted with age. In the Czech Republic we identified that unemployment has a significant effect on employment and retirement among women but no effect among men. As expected, the impact of unemployment on the probability of female retirement is positive and its effect on the probability of employment is negative. In particular, for 50 year-old women, one percentage point difference in the unemployment rate implies an employment rate difference of around 3.3%. Each additional 10 years of age lower the marginal impact of an additional year of age by about half (by 1.7 pp. to about 1.7%). The corresponding impact on retirement probability is smaller (1% difference in probability at the age of 50). It should be noted that the unemployment rate across Czech districts ranges from 5 to 18%. This implies that the maximum difference in the impact on employment rate is about 20 pp. (13*1.5) and on retirement probability 10 pp. (13*1.0). In Sweden, we did not find any statistically significant impact. These estimates are summarized in table 5.1 below. 40 0 60 6 5 Share of people RETIRED Share of people WORKING Share of people RETIRED Share of people WORKING