• by Ed DeNicola , Op-Ed Contributor, October 7, 2024

Who won the Vice President candidates' debate depends on who you ask. If you ask a Democrat, the answer is likely Walz. If you ask a Republican, it’s Vance.

But neither Vance’s or Walz’s debate performances did much to change 2024 presidential election predictions. 

If you look at prediction markets in particular, the results are based on a wisdom-of-crowds methodology and the crowd’s post-debate numbers hardly budged, except for Polymarket, which went from 50% Harris and 48% Trump before the debate to 50% Trump and 49% Harris after the debate.

While prediction market data is not normally reported within a range of error, I can tell you that based on the results we had at the Media Predict prediction market, the founder came up with the rule that any predictions within the 45% to 55%t range were considered a “no call.”

Of the three prediction markets we’re tracking -- IEM (Iowa Electronic Markets), PredictIt and Polymarket -- IEM is the only one showing significant prediction results for one party over the other. IEM’s trading prices were DEM 0.895 and REP 0.144 before the debate and DEM 0.890 and REP 0.145 after.

A couple of weeks ago, a University of Iowa professor by the name of Thomas Gruca reached out to me via LinkedIn.

As it turns out, Gruca is a George Daly Professor of Marketing and the present Director of the Iowa Electronic Markets (IEM). He’s a great connection.

To my surprise, Gruca informed me that prior to leaving Media Predict, the founder had shared a large database with the University of Iowa.

One of the unique aspects of Media Predict is that it forces people to justify why they traded the way they did before they are able to make a trade.

These justifications enable researchers to know why participants traded one way or the other and more broadly why the market moved the way it did.

Using the Media Predict data, Vahid Karimi Motahhar, Thomas Gruca and Mohammad Hosein Tavakoli published a research paper in the International Journal of Forecasting on the topic of Emotions and the Status Quo: The Anti-Incumbency Bias in Political Prediction Markets.

Motahhar is with Sabanci University in Istanbul, Turkey and Tavakoli the University of Warwick in the UK.

The study is very sophisticated. They used a VADER sentiment score to measure the existence of anti-incumbency bias (negative emotional sentiment).

VADER stands for Valence Awareness Dictionary and sEntiment Reasoner. It uses grammatical rules to produce a polarity score.

Armed with VADER (scores, they looked to see if the incumbency bias was more negative for the respondents that traded in support of challenger (non-incumbent) candidates.

Based on the data they had for the 2012 U.S. presidential election and the 2015 U.K. general election, it was the case that sentiment was more negative for those who placed trades for challengers.

The results suggest that the same linkage between anti-incumbency bias and actions against the status quo seen in voting behavior also applies to prediction markets.

It also supports the basis for the 13 Keys prediction system that presidential elections are won or lost based on the performance of the incumbent party. If the party doesn’t perform up to expectations, anti-incumbency bias will be higher, and voters will act against the status quo (party in power).

We can go on and on about debates and who won them, but at the end of the day it’s just chatter. Elections are mostly won based on the public’s perception of the incumbent party’s performance.