Lockdown Stringency Analysis

Neal Page: He says we're going the wrong way.
Del Griffith: Oh, he's drunk. How would he know where we're going?
--Planes, Trains & Automobiles

Although Twitter can devolve into a cesspool of negativity and ad hominem back biting, the CV19 situation has revealed that Twitter provide a platform for an army of analytical talent to share exceptional work. Here's another example, this one studying the relationship between lockdown stringency among EU nations and CV19 morbidity. I've pulled some of the graphics from the thread for viewing here.

First, the cumulative stringency index estimated for various EU countries. There are several stringency indexes floating around. All of them involve scales aim at rating the intensity of authoritarian measures implemented. This particular scale sums daily stringency ratings to get a cumulative total that can be applied against measures of dependent variables summed over the same time period.

Next, CV19 deaths/million population for each country. If lockdowns were 'working,' then one would expect a negative relationship between lockdowns and virus deaths. Here, not only is the relationship not negative, it is weakly positive, suggesting that more intense lockdowns result in more deaths. R-squared clocks in at 0.138, which not all that bad for a 'macro' study. This result echoes findings from other work discussed on these pages.

Finally, stringency vs change in GDP. Here we find a negative relationship as expected (i.e., the more stringent the lockdown, the larger the decline in GDP). Solid R2 of about 0.487.

Would be interesting to see this analysis applied to US states.