Arguably the most important issue facing humanity as a whole is when and how to reopen countries which have largely been in lockdown for the past 6-8 weeks, and what will happen when they do. While governments will influence these decisions, ultimately each business will have to decide for itself when and how to reopen.
I argue that a key element to this decision is the ability to identify regions which are on the same pandemic trajectory as the United States, but further along. If such a region can be identified, we can analyze the results of this region as it opens up, to assess our chances of success.
Our data source is Our World in Data (The link is provided at the end).
A key insight into the trajectory of the United States outbreak is the importance of finding other countries which have had similar trajectories and are “ahead”. By finding such a country, we can look at the path the other country’s outbreak took and glean insight into the likely trajectory of the USA outbreak.
To assess the progression of the disease we use a metric which we call mortality rate. It is calculated as the total death count divided by the total case count, calculated daily. Admittedly, this is a very crude measure of the mortality rate of the disease, but we are using it only to follow the trajectory of the disease, not as a true guage of fatality.
We can see clearly that, while the magnitude of the mortality rate of South Korea is significantly lower than that of the United States, the trajectories are very similar. They both start low, then jump alarming, then fall abruptly, and then grow gradually before eventually levelling out.
We next take the first differences of the mortality rate and compare the outbreaks.
South Korea appears to be experiencing a similar outbreak trajectory to the USA (admittedly with a lower mortality rate, as mentioned), but are “ahead” of the USA.
Can we quantify how similar the outbreak trajectories are and how far apart they are in time?
We do this using the Cross-Correlation Function, which is plotted below. In the following plot, A lag of k indicates that the USA outbreak has been lagged behind by k steps.
At a lag of 10, the correlation coefficient is 0.581. This seems to be significant evidence that South Korea is leading the USA by 8-10 days.
If this is the case, we need to assess carefully what happens as South Korea reopens its economy. Businesses should look at South Korea for insight into what may happen when the USA reopens. The resurgence in cases now being seen in South Korea is a worrying sign for plans to reopen in the USA.
The indications from South Korea are not positive, but what is the situation in the USA currently?
We plot the current death count in the USA, as well as forecasts from two well-known time series methods: ARIMA and Holt.
We see from the above plot that these time series methods are predicting that the United States will cross 100 000 deaths on approximately 25-26 May.
Some experts have suggested that there need to be 20 million tests administered daily in this country in order to safely reopen fully. The minimum threshold that has been suggested is 500 000 covid-19 tests per day.
The following graph shows the current progress and ARIMA and Holt forecasts for when we may expect to reach the goal of 500 000 tests per day.
These forecasts indicate that the United States is not projected to reach 500 000 tests until June 17th.