Yale Professor Ray Fair recently updated his two-party vote share model on April 25, 2024, with the latest economic data. He now projects Biden to receive 51.72% of the two-party vote share in the 2024 General Election.
We previously looked at Professor Fair’s model from Fall 2023 and Winter 2024. This model relies on economic data (GDP growth, inflation) and fundamentals (incumbency, war) to accurately predict vote percentages. Polling, approval ratings, and fundraising data are not used at all.
“Two party vote share” means that this only counts Republican and Democratic general election candidate votes excluding all independent candidate and third-party votes:
Democratic Two-Party Vote Share = Democratic Votes / (Democratic Votes + Republican Votes)
After reaching his nadir in Fair’s model forecasts in April 2023 (49.79%), Biden’s vote share increased for the past four quarters and is now at a new high for this election. At the same time, the Iowa Electronic Markets’ (IEM) 2024 Presidential Vote Share market has Biden at the lowest point in the four forecasts we’re comparing. This was also the first time since IEM launched its presidential markets in May 2023, that Fair’s model has given Biden a higher vote share than IEM. Here are the two forecasts below:
| Date | Fair (%) | IEM ($) |
| April 25, 2024 | 51.72 | .5070 |
| January 25, 2024 | 51.46 | .5800 |
| October 25, 2023 | 51.03 | .5700 |
| July 27, 2023 | 50.06 | .5800 |
| April 30, 2023 | 49.79 | N/A |
The two articles mentioned at the beginning both look at Professor Fair’s model in more detail, but here is a quick summary of the model:
Democratic Vote Percentage ~ I + DPER + DUR + WAR + (G * I) + (P * I) + (Z * I)
The variables represent:
- I = Democratic presidential incumbent at the time of the election
- DPER = Democratic presidential incumbent is running again
- DUR = Either party has been in the White House for one term
- WAR = A World War is occurring
- G = growth rate of real per capita GDP in the first 3 quarters of 2024 (annual rate)
- P = growth rate of the GDP deflator in the first 15 quarters of the Biden administration, 2021:1-2024:3 (annual rate)
- Z = number of quarters in the first 15 quarters of the Biden administration in which the growth rate of real per capita GDP is greater than 3.2 percent at an annual rate
You can read more about IEM in this article, but here is a quick overview:
First started in 1988, the IEM is the oldest operating prediction market in the world and one of the few platforms in the United States where it is legal to bet on U.S. elections. Traders invest up to $500 in various markets, such as which party will win the 2024 Presidential election. In IEM’s Vote Share market, traders buy and sell shares based on each major party’s vote share in the General Election. Past studies have shown IEM market prices to be more accurate than polling.
Manifold also has a Democratic candidate two-party vote share market cloned from the IEM market. Those prices aren’t being used here because there have only been 3 trades from 2 total traders in the nearly 12 months of existence. (For the record, it was at 53% on April 25th).
A few things to keep in mind:
- IEM is using “Last Price” which is the last trade before midnight of that day. The trade generating this price may have occurred on a previous day if there was no trading of the contract on that specific day.
- This is somewhat of an apples-to-oranges comparison. The Fair model uses economic data released quarterly while IEM operates continuously. For simplicity, we’re taking IEM prices from the day Fair’s model is released.
- Biden’s two-party vote share in 2020 was 52.27%. The Fair model predicted 47.9% and IEM 53.4%
- A simple average of IEM prices and Fair model forecasts is a more accurate predictor of vote share than either forecast alone.
Further analysis could include:
- Rather than use a specific day’s “Last Price” for IEM, how would a moving average compare?
- We’ve looked at how Nate Silver’s FiveThirtyEight model moved prediction market prices, and it would be interesting to see if the Fair model has a similar impact on IEM vote share markets.
- Speaking of FiveThirtyEight, how much impact will the first “538” model have on prediction markets now that Silver is gone?