Introducing the 2024 Voter Registration Index Forecast Model

Topic: 2024 Election Forecasts

I have recently finished previewing all seven decisive states plus North Carolina by reviewing the county trends through a pessimistic lens, classifying them all by their observed leans (Crimson, Republican, Competitive, and Democrat) and extrapolating their vote totals based on similar 2020 turnout and vote growth where appropriate and precedented for both candidates (particularly “fortified” voting) – even if my gut feeling is that growth won’t be there for one of them organically.  Collectively, this helps me visualize which counties, and in turn, states, pose the biggest challenges for Donald Trump in surging toward a 2024 victory.

Those produced tentative results that looked like this:

·      Pennsylvania – Trump +1.0%

·      Georgia – Harris +0.4%

·      North Carolina – Trump +1.4%

·      Michigan – Harris +0.3%

·      Arizona – Trump +1.4%

·      Wisconsin – Trump +1.3%

·      Nevada – Harris +1.1%

·      New Hampshire – Harris +6.9%

In modern history, the most predictive indicator is voter registration by party.  Here is a brief overview that features it I did on Westmoreland County, Pennsylvania, to highlight what I believe is highly suggestive statistical evidence of ballot stuffing there in the 2020 quasi-election.  Recently, while quality checking a poll from Iowa, I realized this model was highly replicable and an excellent means of conveying real information based on factual data and outcomes, which of course run counter to mainstream media narratives that certain states are not fertile for Trump.  The New York Times will be featuring some of this analysis in the near future in what I believe is an honest attempt to capture alternative means of election forecasting in lieu of widely credible polling.

The voter registration index model produces two figures – a coefficient for up to six races (from 2000 through 2020) if registration data is available going that far back, and a coefficient to measure the two races Trump has run in a given state or county.  I think this is a useful tool for observing the potential magnitude of trend progression based on the severity of registration changes, not simply a suggestion of if a county should shift right or left as the pessimistic model considers.

I intend to incorporate those findings into a final assessment of each state or county that includes:

·      The pessimistic forecast

·      The average Republican coefficient of the voter registration index

·      The average Trump coefficient of the voter registration index

If I have Harris +2 in a pessimistic forecast, an even race in the average Republican voter registration index (+0), and Trump +3 in the Trump-specific voter registration index, then an average would suggest Trump with a slight advantage of 0.3%.

---

Here is an example of the model, using Pinal County, Arizona, as my example (note: this model “bakes in” the capacity for ballot stuffing, which is useful to see if it can be overridden in 2024 if it arrives in the same magnitude as before):

2000

Pinal County, which is wedged between Maricopa and Pima Counties and is currently Arizona’s third largest county, voted for Bill Clinton twice and had an old-school working-class Democrat registration advantage at the time of the Bush v. Gore race.  Walking you through the table:

·      The Democrat lead was 16,014 registrations over the GOP membership…

·      Which equates to a 21.7% lead, expressed as D+21.7%.

·      Since I’m not including the data from 1996, I indicate N/A in the registration change from previous campaign.

·      In the results, Bush defeated Gore by 472 votes, flipping the county.

·      Subtracting the Democrat registration lead from the results, we find that Bush ran 16,486 to the right of the Democrat registration advantage.

---

2004

·      The county begins a serious Republican identity shift, and the GOP makes up 9,148 registrations and brings the county to a D+7.5% advantage.

·      This shift is indicative of a coming Republican surge in the 2004 election.

·      Bush defeats John Kerry by 15.1% in the election, a margin of 9,754 votes.

·      The Bush margin is 16,620 to the right of the Democrat registration advantage, a coefficient nearly identical to the 2000 number.

---

2008

·      Despite Arizona moving left in registrations due to the Obama enthusiasm surge and Republican fatigue, Pinal County’s coalition shift continues to the right.

·      The GOP eats up another 5,021 net registrations, bringing the Democrat advantage to a miniscule D+1.3% and signaling another move to the right.

·      McCain wins the county by 0.7% less than Bush did in 2004, indicating a slight bucking of the registration index, but his margin of victory in raw votes is greatly enhanced due to population gain and comes out to 15,167 votes.

·      With incredible consistency, McCain’s winning margin is 17,012 to the right of the Democrat voter registration advantage.

---

2012

·      Republicans flip the county registrations index with a net shift of 12,440 in their favor, putting the index at R+4.9%, which is indicative of further gains in margin.

·      Romney underperforms, adding just 2,658 votes to McCain’s 2008 total, and only slightly advancing the margin of victory against Obama.

·      Romney comes in 7,178 to the right of the Republican voter registration advantage, a clear sign that non-partisans in the county lean heavily Republican.

---

2016

·      With Trump looming as the GOP standard bearer, registrations move along to the right, but by reduced net numbers, putting the county at R+7.6% but still signaling greater Republican margins against Hillary Clinton.

·      Trump delivers a strong gain of 10,740 net new votes, and Clinton gains just 3,586, giving Trump a margin of victory of 19.2%, and an increase in raw vote margin of 7,154, in keeping with the registration indicator and 10,523 to the right of the Republican registration index.  A large third-party vote share of 6.8% is present.

---

2020

·      Republicans move out to an enormous registration advantage of R+11.3%, netting a gain of 13,738 registrations over 2016 and forecasting a bigger win in 2020.

·      Trump’s vote gain is a Republican record of 34,258 net new votes, but it is countered by a Biden gain of 27,214 net new votes, which is larger than any previous Republican single election gain.

·      Biden’s gain of real and fraudulent ballots results in the Trump margin growing by 7,044 from 2016 but sitting just 3,829 to the right of the Republican registration index, the lowest coefficient in this sample and indicative of widespread election “fortification” for Biden given the magnitude of Trump’s gains.

---

2024 – So What?

Where do these six elections point us for 2024?

·      In the modern era, Republicans always run to the right of the voter registration index.

·      The average coefficient over six cycles is 11,491 to the right.

·      The Trump coefficient, in two races, is 7,176 to the right of the registration advantage.

·      Common sense suggests the 2024 result will again run to the right of the Republican voter registration advantage. The question is, “how much?”

·      The pessimistic model for Pinal County (referenced at the front of this article) gives Trump a victory of 40,000 votes in margin, a strong improvement in its own right.

·      The July 2024 Secretary of State update shows Republicans out to a lead of 42,127 registrations, bringing the county to R+15.7% and forecasting another increase in Republican margin, at least in terms of raw votes and likely in percentage margin of victory.  Please note that the Republican advantage will almost certainly be much higher, perhaps approaching 50,000, in the final update before the election, which will require the following estimates to be revised upward.

·      Considering only the current lead of 42,127, if Trump is limited to his 2020 coefficient (+3,829 to the right of the index), his margin will be +45,957.

·      If Trump sticks with his average of 7,176 to the right of the index, his margin will be +49,303.

·      If Trump grabs the standard Republican average of six cycles, which is 11,941 to the right of the index, his margin will be +54,069.

·      If Trump hits the high point from 2008, 17,012 to the right of the index, his margin will be +59,139.

Considering an average of the following three points:

·      Pessimistic forecast: Trump +40,000 (+21.6%)

·      Trump average coefficient: Trump +49,303 (+26.7%)

·      Republican average coefficient: Trump +54,069 (+29.2%)

Trump appears likely to win Pinal County by roughly +47,791 votes (+25.8%) in margin.  That increase in margin would wipe out the 10,457 gap statewide from 2020, and when paired with other massively Republican counties showing the same trend, suggests a strong likelihood of Trump victory in Arizona.

I can replicate this data in any state or county with voter registration by party figures, and in some cases, can run comparative studies as seen in the correlation between Iowa and Wisconsin, or Pennsylvania and Michigan.  These figures will be factored into my forecasts as we enter the home stretch.

---

Seth Keshel, MBA, is a former Army Captain of Military Intelligence and Afghanistan veteran. His analytical method of election forecasting and analytics is known worldwide, and he has been commended by President Donald J. Trump for his work in the field.