Commentary

Presidential Election Predictions Data: Where We Stand

Tomorrow is Election Day and if the predictions are accurate, we can expect an extremely close outcome.

In fact, The Economist’s Intelligence Unit (EIU) predicts a 70% probability there will be a contested, result.

As someone with a passionate interest in the science, I like to look at predictions before and after elections to see which were right.

For tomorrow's, I’m doing it publicly and sharing with MediaPost’s readers so we can look at it together.

Some of what I've included may not be a "prediction," per se, but it may be a valuable signal. It’s interesting to see which of those provide foresight, as well.

In terms of market predictions, I’m applying Media Predict’s rules: if the prediction is between 45% and 55% it’s a “no call."

advertisement

advertisement

Here we go.

The Polls

FiveThirtyEight’s polling average has Harris up by 0.9 (Harris 47.9 vs. Trump 47.0). These numbers look to be within FiveThirtyEight's error band, which equates to a tie, or too close to call, ruling.

The Markets

  • Iowa Electronic Markets (IEM)

    • 2024 U.S. presidential vote share market – Dem 54.7% vs. Rep 49.2%*

    • 2024 winner-takes-all market – Dem 74.0% vs. Rep 27.0%

  • PredictIt

    • 2024 presidential election winner?  – 54% Dem vs. 51% Rep*

    • Which party wins the presidency in 2024? –  55% Dem vs. 50% Rep*

  • Polymarket

    • 2024 election forecast – 57.7% Trump vs. 42.5% Harris

    • Popular vote winner – 26% Trump vs. 74% Harris

  • Kalshi

    • Who will win the presidential election? – 54% Trump vs. 46% Harris*

    • Who will win the popular vote? – 78% Harris vs. 22% Trump

  • Media Predict

    • Will Harris be elected as U.S. President in 2024 – 44.8% odds*

The Keys

The 13 Keys prediction system has been predicting a Democratic party win going back to my first prediction column in May.

The Democratic nominee might have changed, but Allan Lichtman’s prediction didn’t. His 13 Keys still predict a Democratic (Harris) victory.


The Betting

BetOnline:

  • 2024 presidential election – odds to win: Trump -147, Harris +127

  • Election winner to lose popular Vote: No -170, Yes +130

  • President – popular vote winner: Harris -425, Trump +315

For anyone unfamiliar with betting, odds with a negative symbol represent the favorite. The number after the negative symbol is the amount of money you need to bet to win $100.

The Stock

The NASDAQ share price for Trump Media Technology Group Corp. (DJT ticket symbol)

  • $30.56 -- up 75.13% year-to-date

Dow Jones Industrial Average (DJIA) Model as of Oct. 31 (Source: MarketWatch) based on the stock market’s year-to-date performance.

  • Predicts 70% probability that the incumbent White House party (Harris) will win.


The Misery

The Misery Index has been ticking down and was at 6.5 as of Oct. 28 (Source: Investor’s Chronicle). It was 7.02 when I last wrote about it in September. Indices below a 7.353 predict that the incumbent White House Party will retain power.

  • The Misery Index predicts Harris and the Democrats will win.


The Models

The Political Economy Model, Charles Tien, Hunter College & The Graduate Center, CUNY; Michael S. Lewis-Beck, University of Iowa

  • As of late August, either party is almost equally likely to lose.

Prospective Financial Future & Length of Time in the White House Model, Brad Lockerbie, East Carolina University

  • Forecast the incumbent party receiving 49.09% of two-party vote (based on July 5th data). Too close to call.


The Voting

Early voter registration is tracked. They know who voted and the party for which people are registered.

They don’t know actual voting behavior and each party’s distribution of early voters can skew the predicted outcome. For some states where party registration isn’t available, party affiliation is modeled.

  • 76,438,831 early mail-in and in-person votes cast nationally

  • National - 41% Democrat vs. 39% Republican

  • Blue-Belt Battleground States

    • MI - 46% Democrat vs. 43% Republican

    • PA - 57% Democrat vs. 33% Republican

    • WI - 34% Democrat vs. 26% Republican

  • Sunbelt Battleground States

    • AZ - 33% Democrat vs. 42% Republican

    • GA - 45% Democrat vs. 48% Republican

    • NC - 32% Democrat vs. 33% Republican

    • NV - 34% Democrat vs. 38% Republican

Source: NBC News analysis of TargetSmart data.

What do I think? I’m going with Allan Lichtman’s research indicating that the defining factor will be how well the incumbent party performed during the last term. The 13 Keys, The Misery Index and the DJIA Model all support the view that the Democrats did well enough to retain power. Moreover, IEM has a great track record and its Winner-Takes-All Market also supports a Democratic victory.

When the election is over, I’ll post again with the pre-post comparison.

Regardless of everyone’s political leanings, happy voting and good luck with the election.

*Too close to call.

2 comments about "Presidential Election Predictions Data: Where We Stand".
Check to receive email when comments are posted.
  1. Tony Jarvis from Olympic Media Consultancy, November 4, 2024 at 9:40 p.m.

    Ed: A terrific analysis.  Remember Tony's polling Guidelines:  Only use high quality polls, e.g., J Ann Selzer Co. in Iowa.  Many are poor and therefore misleading. Even for high quailty polls double the estimated range of error to "truly"  understand the potential spreads.

  2. Ed DeNicola from MediaLytics, November 4, 2024 at 10:26 p.m.

    Hi Tony -- Thank you for the feedback and for the polling guidelines. They make a lot of sense. FiveThirtyEight's polling average is more complicated than it looks (see https://fivethirtyeight.com/methodology/how-our-polling-averages-work/).
    They're weighting the polls in the average based on sample size and for multiple polls from the same pollster in a short window. The error band is supposed to be for the 95% confidence level (two standard error). Have a great Election Day tomorrow!

Next story loading loading..