I'm not convinced about the low optimum. There seems to be a clear correlation between employment and ability, but then they bend the data into a model that se

I'm not convinced about the low optimum. There seems to be a clear correlation between employment and ability, but then they bend the data into a model that seems to exaggerate the differences between >35 and <35 cohorts. I have worked to much to have energy to read the details, but if the "unemployed" group does not include retired people, then the causality might be in the other direction...