As outlined in Appendix B, the calculation of unemployment rate less educated resulted in a remarkable finding: the lowest score of that variable was close to zero (0.1%). In reality, such a low unemployment rate among the less educated will never occur. Moreover, the highest score on unemployment rate less educated is lower than the highest score on unemployment rate (12.30 < 16.20, see Table B1). This is also very unlikely, as the unemployment rate among the least skilled is always higher than among the most skilled. Both of the above-mentioned peculiarities of unemployment rate less educated indicate that it underestimates the very employment level it aims to measure. This is why all of the analyses on that variable in Chapters 3 and 5 are replicated in this appendix with analyses on the most conceptually and empirically closely-related indicator: the general unemployment rate (from now on referred to as unemployment rate).

Unemployment rate and unemployment rate less educated are very highly correlated (see Table D1), which suggests that the latter is particularly reliable as an indicator of the differences in unemployment rates of the less educated across years and cities. As previously noted, it merely underestimates that level; however, as the findings in Table D1 indicate, it does so consistently across all years and cities.

Robustness checks on the findings in Table 3.1

Table D2 replicates the analysis in Table 3.1 on the impact of the producer services on the unemployment rate of less-educated urbanites, and yields similar results. Just like Table 3.1, it indicates that in cities with the highest employment shares in the advanced producer services, the unemployment rate is the lowest, as predicted on the basis of the polarization thesis. Consequently, Figure D1 is very similar to Figure 3.1, although the predicted unemployment rates are higher in the former than in the latter. This is in line with the previously mentioned underestimation of the unemployment rate among the less educated by unemployment rate less educated.