Update: Is your state a hot spot?

The uncertainty surrounding coronavirus is frustrating. Is it widely prevalent and predominantly benign? Or is it rare and lethal? The answers to these questions affect the optimal public health response and how quickly we can get the economy back.

If it is rare and lethal, then testing and tracking would likely be sufficient. If it is widely prevalent, then the current lockdowns is silly (something I doubt). And if it is largely benign, then maybe we can let certain people out of their homes and monitor their health closely with reasonable public health strategies (a strategy I will have more to say about later).

I have also been asked whether there should be a national response. That seems silly to me, since there is so much heterogeneity by state. To see this point, I graph below two measures of risk. The first is (log) per capita risk of death from Covid19 on the y-axis. The second is per capita risk of Covid19 infection (y-axis). The former is a lagging indicator, the latter a leading indicator. Both are worrisome, although the infection is confounded with testing patterns in each state. However, as you can see, there is a strong correlation between the two.

Whichever of the two measures you choose, it is New York, New Jersey, Louisiana, and Connecticut that are outliers. Washington DC and Massachusetts are new hot spots. There are are some off-diagonals that might offer further clues. Michigan is more of an outlier on deaths than infections; and Rhode Island is more of an outlier on infections than deaths.

The focus of states like Oregon, Washington, and California on re-opening their economies makes sense given the relatively low risk.