Election 2020: quant analysis of new party registrations vs actual votes

I think we should ascribe very high uncertainty to polling results in this election, for a number of reasons including the shy Trump voter effect as well as the sampling corrections applied which depend heavily on assumptions about likely turnout. 

Graphs below are from a JP Morgan quant analysis of changes in number of registered voters by party and state, and the correlation with actual votes in subsequent election. Of course it is possible that negative covid impact has largely counteracted the effect discussed below (which is an integrated effect over the last 4 years) -- i.e., Trump was in a strong position at the beginning of 2020 but has declined since then. 

This is an unusual election for a number of reasons so it's quite hard to call the outcome. There's also a good chance the results on election night will be heavily contested.

The author of this analysis is Marko Kolanovic, Global Head of Macro Quantitative and Derivatives Strategy at J.P. Morgan. He graduated from New York University with a PhD in theoretical high-energy physics.

Anyone with high conviction about the election is welcome to post their analysis in the comments.