Foresthill
EOG Addicted
Note: This analysis was posted May 13, 2020, on the site electoral-vote.com. Thus, the data on the presidential election is old, but the points are still valid.
He makes an excellent point that they attract "sucker" bets from the uninformed and shows why they sometimes aren't as accurate a predictor as some think. His analysis follows below:
Betting Markets Like Trump (as of 5/13/20)
Speaking of sports, you may have heard that some people like to bet on them. Some people also like to bet on politics. And, as Axios' Dion Rabouin points out, the betting markets are currently predicting that the next president of the United States will be...Donald John Trump. It depends on which book you deal with, but the incumbent is currently getting 10/11 in most places, which implies a 52% chance of winning. Joe Biden is getting 6/5, which implies a 45% chance of winning (they don't add up to 100% because of the books' cut, also known as the vigorish).
Is it possible that Trump is actually the favorite, and that the polls (and thus our map, above) have it all wrong? After all, political betting has been around a long time and has, historically, been a pretty accurate predictor of actual results. This is because there are a lot of data coming in, and—inasmuch as people are putting real money on the line—their "opinion" is likely to be unvarnished truth.
So, it could be that we're learning something important from the bettors. However, we doubt it. This cycle has already given us a bunch of wonky betting odds, in particular overrating Pete Buttigieg and Michael Bloomberg. You could also argue that, at 100-to-1 (or anything other than ∞-to-1), Alexandria Ocasio-Cortez was overrated too, inasmuch as she is not eligible to be elected. And there are certainly some things that are skewing the current odds. First, it's not like things got "reset" when Joe Biden became the presumptive nominee, which means that Donald Trump's current odds are at least partly an echo of the campaign phase when he was the only Republican, while Democratic money was being split across multiple candidates. Further, and in a related point, the Democratic money is still being split, as Hillary Clinton is still getting 25-to-1 (4%) at most places, while Gov. Andrew Cuomo (D-NY) is getting 35-to-1 (3%).
In addition, there are likely other dynamics at play here. Among the possibilities:
It's still fun to check in on the odds, and we will continue to do so from time to time. However, to the extent that they can be taken at all seriously as a predictive tool, that time will not arrive for several months, at the earliest. (Z) (text bolding by me)
He makes an excellent point that they attract "sucker" bets from the uninformed and shows why they sometimes aren't as accurate a predictor as some think. His analysis follows below:
Betting Markets Like Trump (as of 5/13/20)
Speaking of sports, you may have heard that some people like to bet on them. Some people also like to bet on politics. And, as Axios' Dion Rabouin points out, the betting markets are currently predicting that the next president of the United States will be...Donald John Trump. It depends on which book you deal with, but the incumbent is currently getting 10/11 in most places, which implies a 52% chance of winning. Joe Biden is getting 6/5, which implies a 45% chance of winning (they don't add up to 100% because of the books' cut, also known as the vigorish).
Is it possible that Trump is actually the favorite, and that the polls (and thus our map, above) have it all wrong? After all, political betting has been around a long time and has, historically, been a pretty accurate predictor of actual results. This is because there are a lot of data coming in, and—inasmuch as people are putting real money on the line—their "opinion" is likely to be unvarnished truth.
So, it could be that we're learning something important from the bettors. However, we doubt it. This cycle has already given us a bunch of wonky betting odds, in particular overrating Pete Buttigieg and Michael Bloomberg. You could also argue that, at 100-to-1 (or anything other than ∞-to-1), Alexandria Ocasio-Cortez was overrated too, inasmuch as she is not eligible to be elected. And there are certainly some things that are skewing the current odds. First, it's not like things got "reset" when Joe Biden became the presumptive nominee, which means that Donald Trump's current odds are at least partly an echo of the campaign phase when he was the only Republican, while Democratic money was being split across multiple candidates. Further, and in a related point, the Democratic money is still being split, as Hillary Clinton is still getting 25-to-1 (4%) at most places, while Gov. Andrew Cuomo (D-NY) is getting 35-to-1 (3%).
In addition, there are likely other dynamics at play here. Among the possibilities:
- Manipulation: Because political betting markets involve relatively small amounts of money (relative to, say, football betting markets), it's possible for one or two bettors to distort the odds. In 2012, for example, the betting markets said Mitt Romney was a shoo-in to win the White House. It turns out that one extremely pro-Romney voter dropped between $4 million and $7 million on Romney bets, as a combination gamble and PR investment. We doubt that is what is happening here, but we would be remiss if we did not mention the possibility.
- Selection Bias: Call this the "1936 Literary Digest" thesis. In that notorious example, the Digest used telephones to conduct its annual straw poll, working from lists of the magazine's subscribers, of telephone owners, and of automobile owners. Overlooking that all three categories skewed heavily Republican in 1936, they predicted a big win for Alf Landon (R), who was then swamped by Franklin D. Roosevelt in the actual election. In 2020, it's likely that the folks who have money to burn on political bets are doing ok, and have not lost their jobs or their rainy day funds to COVID-19, and so likely skew Republican. Or, at very least, they skew not-furiously-angry-with-Trump. If every person in America was given a dollar and allowed to bet it on the race, we doubt the President's odds would remain where they are.
- Emotion: As any serious bettor knows, the money is to be made when folks are betting with their hearts rather than their heads. So, bet against the Mavericks on the day of Dirk Nowitzki's final game. Bet against the Tampa Bay Buccaneers on the day that Tom Brady debuts as their quarterback. Bet on the Astros in every game, since everyone outside of Houston hates them now. Our guess is that Trump is more likely to attract "fan" bets than Joe Biden is.
- Betting as Entertainment: Speaking of serious bettors, they are not in the habit of laying bets that don't pay off for 6-12 months, and letting the books have a nice, interest-free loan. Consequently, most of the bettors right now are surely dilettante bettors, who are just interested in having something to watch and root for. This doesn't necessarily argue for a pro-Trump bias, but it does argue that the current numbers are not necessarily reflective of cold, hard, rational decision-making.
It's still fun to check in on the odds, and we will continue to do so from time to time. However, to the extent that they can be taken at all seriously as a predictive tool, that time will not arrive for several months, at the earliest. (Z) (text bolding by me)
