The post referenced here describes an approach for predicting the outcome of the Electoral College vote for president in 2020. The approach relies on two measures and the history of the last twenty-two elections. The two measures are the percentage of 1st ballot votes received by the incumbent party candidate at that party’s nominating convention and the year-over-year change in unemployment rate at the time of the election. The incumbent candidate has to meet or exceed criteria for both measures to have a substantial probability of winning. The incumbent is likely to lose if either criteria is not satisfied. An analysis of election history shows how these measures cluster the outcomes within a 95% binomial confidence band of 72.7% to 99.9%.
Figure 1 provides an assessment of these measures against the criteria as of May 2020. The horizontal axis is the percentage of 1st ballot votes received by the incumbent party candidate at that party’s nominating convention. Anything 75% and above favors the incumbent party candidate. Anyone who pays any attention at all to the news knows that Trump is in solid control of the Republican party, so I have placed the status cross well to the right of 75%. This measure clearly favors the incumbent candidate, who at this point in time is Donald Trump, Republican.
Figure 1 May 2020 Presidential Election Status
The year-over-year change in unemployment rate at the time of the election is measured from the October-to-October change. The October 2019 value was 3.6%. The Bureau of Labor Statistics report for April of this year had the unemployment rate at 14.7%, with the increase due to Covid-19 inflection fears and government forced lockdowns. The Congressional Budget Office (CBO) projects the unemployment rate will peak at 16% this year, average 11.4% for the year, and average 10.1% in 2021. Using any of these values places the year-over-year change well off the top of the chart. The -40% and +50% extremes on the vertical axis are selected to capture the extremes from the prior 22 elections. The 48% year-over-year increase in unemployment before the 1932 election between the Republican incumbent Hoover and the Democratic challenger Roosevelt set the upper extreme. Hoover received 98% of first ballot votes at his convention that year, but his re-election was doomed by the unemployment rate.
Right now, there is a high probability that Trump and the Republicans will lose this November’s presidential election.
Could Trump and the Republicans Still Win?
The short answer is YES. To get into the win box, the unemployment rate would have to come back down to 3.8% by this October. This means the employment level in October would have to be at least 159 million people, which was the level in January and February of this year. The employment level in April was 133 million people, so 26 million people would have to get back to work. It is hard to see this happening by October.
However, this analysis of the historical record does say Trump and the Republicans have a some probability of winning. The change in unemployment rate measure is indicative of the direction of the economy and people’s lives. Roosevelt and the democrats won re-election in 1936 even though the unemployment rate was 16.9%, but that was a huge improvement from the 20.1% unemployment rate in 1935. Even if the unemployment rate does peak at 16% or more in the coming months, if there is a strong negative trend by the time of the election, maybe that will be enough.
We’ll wait and see, but at this point Biden and the Democrats have an high probability of winning in November.
References:
Unemployment Report
Congressional Budget Office projection