2 Comments

Probably not what you're looking for, but some comments on the stats:

France has pretty high productivity in terms of GDP per hour - higher than Germany depending on how you adjust it. Of course with a lower GDP per capita than the UK, this means fewer hours, hard to know exactly how to interpret this - are there any country-wide biases in hour measurements? Probably some. But if it's truly the case that France is becoming more per-hour productive and using that to just work less... Good trade. Some evidence is out showing sharp diminishing returns to hours worked, e.g. in here https://hbr.org/2015/08/the-research-is-clear-long-hours-backfire-for-people-and-for-companies If this is true, productivity growth may not really be that much slower in the UK, they're just working far too much (doesn't explain the US that well though)

For the education differences, at least in the US:

- The voting patterns of highschool non-college do change a bit over time: from just over 50% dem on average prior to 2016 to 45% dem in 2016. This is now a very big group!

- However, looking at the graph over time, it changes by far more than 5%. This is primarily driven by change at the top 10% and not as much the bottom 90%. The top 10% go from like under 40% dem in 1960 to over 70% dem in 2016

- a small part of what's happening is that the top 10% is a very different group over time. In 1976 for example most of this top 10% was people with BAs (negligible % with PhDs, 4% had masters, 11% had BAs), but in 2016 we had 2% with PhDs, 11% with MAs.

- The voting patterns of those without high school is sort of consistent over time, this group just goes from huge to small

I imagine something similar is happening in Britain - I guess these "10% of people vs 90%" in terms of voting behaviour within those groups is liable to produce weirdness. E.g. if it was balanced 50% dem in both bottom 90 and top, and then went to 48% dem in bottom 90%, the balancing % voting dem in top 10% to stay at 50% overall would be 68%, thus the gap between the groups would be 20%, even though the bottom 90% are mostly the exact same way.

I suppose my thought here is: if you care to evaluate the differences in voting behaviour of the working class, just measure that directly instead of subtracting it from a tiny weird subgroup.

Expand full comment

You managed to articulate a logic to what seemed to me mostly madness. Who knows maybe they'll pull it off.

Expand full comment