Presidential Elections in the United States of America in 2004CE and 2020CE

Note; the file names comprise MD5SUM's of the documents posted on github. Loading the .xlsx into excel and then saving it will change its MD5 hash.

Data

An excel spreadsheet Presidential_Election_Data_2000_to_2024.xlsx contains national and state election data for election years 2000CE to 2024CE from reputable sources. Each election year has its own tab. Then, there is an Analysis tab which contains the Voter-vs-Census curve, and calculation of novelty and their distributions for two years of interest, 2004CE and 2020CE. The spreadsheet allows user to vary outlier thresholds to test various election outcomes. Notice that all the methods rely on interpolation, so no predictions are attempted.

Executive Summary

We examine national and state-specific voter count for excess, focusing on mid-mandate years, 2004CE and 2020CE, of the turbulent mandates (2000CE-2008CE, or Bush-I and Bush-II; and 2016CE-2024CE, or Trump-I and Biden). We define turbulent mandates using national voter count ($V_t$ ) vs. census data ($C_t$ ), during which $\Delta V / \Delta C \approx 1$. We define excess operator Ê over a time-series of total national and state voters ${V_i }{i=0,1\cdots}$ as negative of a symmetric second difference operator $E_i = \hat E V_i = V_i - \frac12 (V{i+1} +V_{i-1})$, where $i$ is the index of an election year. Its positive values mean excessive votes, while negative values mean disinterested voters. For 2004CE and 2020CE we find national execess as 4M(illions of voters) and 13M. We find their census-corrected values as 5M and 11M, respectively.

The voter excess analysis of the 2004CE election suggests two criteria for identifying outliers: 1., a threshold $\delta V = 0.3$M; and 2., the excess has to flip the state election vote, $|R_i - D_i | < \delta V$ , where $V_i = D_i + R_i$, are that state totals for the number of votes in that year for (R)epublicans and (D)emocrats. In 2004CE we identify two states as outliers: Ohio and Florida, which together accounted for 0.9M excessive votes of 4M total. We refer to this behavior as anti-novelty: If excess were subtracted from the republican state winner vote count, the democrat candidate would have won the state.

The excess analysis of the 2020CE election using the same outlier threshold, identifies 13 outlier states. We show we can categorize them into three groups:

1. the battleground states (Arizona, Georgia, Michigan, and Pennsylvania), which have a total excess of voters of 1.4M;

2. the apparently republican states (Florida, North Carolina, and Texas), which have a total excess of voters of 2.4M;

3. the single-majority states (R: Ohio; D: California, Illinois, New Jersey, New York, Washington), which have a total excess of voters of 5.6M.

To the first group we refer to as novelty (if excess were subtracted from the democrat state winner vote count, the republican candidate would have won the state). The second group demonstrates anti-novelty. The states in the third group exhibit diffuse excess, meaning that the excess is insufficient to change the outcome of the state election rather, this excess only affects the popular vote count.

Interpretation

We interpret our findings using correlation between the outlier-states and the politics of the day. We remark, were these excesses random fluctuations in voter numbers and preferences, such a correlation would have been unremarkable in magnitude or in political orientation.

In re 2004CE

We surmise that if there had been actual voter injection by republicans, in its absence the Democrat challenger, Kerry, would have won the presidential election against the incumbent Republican, Bush. We recall, at that time Bush needed election victory to continue and expand the war effort against Iraq.

In re 2020CE

We surmise that in the 2020CE elections, the republicans injected votes in at least, apparently republican and battleground group of states. In the former, it was done to affirm increasingly right-wing leaders, DeSantis in Florida and Abbott in Texas. In the latter, the goal was to secure those states for Trump, but they ultimately failed. Without republican interference in the 2020CE elections Trump would have been defeated in a landslide, with a vote tally of 150 to 388 in favor of Biden. This outcome would have effectively removed Trump from politics. Instead, the actions of republicans played out as a damage control in the apparently republican states, while in the battleground states they evolved into legal and illegal attacks on victorious democrat challenger, Biden, accusing him and the democrats of election fraud. All these actions paved the way for Trump’s reprisal in 2024CE.