First off, this post is not pro either candidate. If you read it as such, it’s your bias, not mine. Truth is, I am anti both parties.
Today’s topic is one that has received a lot of attention of late – political betting markets. I thought I would use some of the information in them (and the polls) to help explain what the odds imply about the election outcome.
I know there are a lot of people who want to believe these markets are manipulated. While that is possible, it is also harder than people realize.
This is for the very simple reason that someone who disagrees can bet the other side and profit!
I am not saying that makes these markets gospel. Markets are wrong all the time and they can be manipulated.
However, it is harder to manipulate a market with a near term end point, like an election or a football game.
Gamestop still has a $10B market value. That is absurd, but unless it files for bankruptcy, the price can remain artificially high for years.
Election markets, unlike stocks, resolve to a known outcome in a short window. Thus, while it can be dangerous to short Gamestop, there is no risk of a short squeeze in political markets.
If you think the price is wrong, you can bet the other side and get paid your winnings in just days.
Popular vs. Electoral
With that out of the way, let’s move on to what you need to know.
The biggest mistake most people make is looking at the national polls and think they mean anything. As most of you know, all that matters is the seven swing states (PA MI WI NC NV AZ GA).
In 2016, Hillary was up by 4% in the days before the election, won the popular vote by 2% and still lost the election.
In 2020, Biden was up 7% a week before the election, won the popular vote by nearly 5% and yet if Trump “found a few more votes” in a few states he could have won.
So, realistically Harris needs to win the popular vote by at least 3% to feel confident in the electoral college. Poll aggregator sites like RealClearPolitics and Nate Silver have her up <1%.
Of course, it’s possible this time is different. There are two reasons to believe the electoral college may be less of an advantage for Trump this time.
The big one would be if Harris does a better job at turnout than Biden or Hillary. The other would be if the pollsters have finally corrected for the persistent Trump undercount in previous campaigns.
But, it seems more likely than not, that a 1% popular vote margin will not win Harris the electoral college. This is why the betting markets are favoring Trump.
What The Markets Are Saying
Most markets favor Trump to win the election with his odds being +/-55% (with Kamala being ~70% likely to win the popular vote). This is down from as high as 67% Trump.
The correction means markets are working as they should. When price momentum goes too far in one direction, people will start taking the other side. That is what makes market information valuable.
What’s more interesting is the betting sites have the most likely electoral outcome as Trump winning by 65-104 votes. This is given a ~20% chance at both PredictIt and Polymarket.
Note, Trump won by 74 in 16 where he took 6 of the 7 swing states. Similarly, Biden won by 68 last time, also taking 6 of 7 swing states (PA WI MI GA AZ flipped while NV was blue twice and NC red twice).
So markets are suggesting Trump will take 6 of the 7 swings again which is pretty consistent with the betting on individual state outcomes.
Markets are pretty confident in Trump winning NC, AZ, and GA, as he is over 65% likely in each state. If he takes those three, he’s at 262 electoral votes and only needs to win one of the three northern states to win the election.
Interestingly, WI & MI have, in recent days, flipped from 60% Trump to 60% Kamala. They surely could flip right back, but there may be some important signal there that momentum has shifted.
That would mean it would all come down to PA. Trump is at 55-60% there. Nate Silver’s aggregate polling has Trump up 0.4% in PA and RealClear is Trump +0.3%, so well within the margin of error.
Are State Outcomes Independent or Correlated?
Stepping back for a moment, recall Trump took all three northern states in ’16 while Biden took all 3 in ’20. Does that mean we should expect a winner take all this time?
Perhaps. If state poll outcomes were independent, then, if each state was 50/50, 1/8 of the time Harris would win all three, 1/8 Trump would win all three, 3/8 Trump would win two, and 3/8 Harris would win two.
In other words, 7/8 of the time Trump would win the election (assuming he also takes NC AZ & GA). However, nobody is saying Trump is 87% likely to win the election.
Why not? First, because it’s not guaranteed Trump takes all of NC, AZ, & GA (see below).
But also because each state’s outcome is not an independent event. If there is an undercount of either candidate’s support in the polls, it likely applies more broadly. That is why the two prior elections saw the winner take 6 of 7.
A more accurate prediction might be 1/4 Kamala sweeps, 1/4 Trump sweeps, 1/4 Trump takes two, and 1/4 Kamala take two.
This would still make Trump 75% likely to win the election…if he wins NC AZ & GA.
The Leading Indicators
Of course, those are not guarantees either. If each of the southern three are 67% likely to be Trump, then there is just a 30% chance he wins all three (if results are independent, so it’s probably more like 40-50% he takes them all assuming they’re not).
If Kamala were to win one of the three southern states, then whoever wins two of three among MI PA & WI wins the election.
So, the way I will watch results is to first focus on NC & GA. If Trump takes both, he likely takes AZ and then it’s at least 75% he wins one of the northern states to clinch the election.
On the other hand, if Harris takes one of NC or GA (if she takes both she is almost certain to win the election), then it likely comes down to whoever gets two of three up north (though there are scenarios where AZ or NV could be the deciding state) with a lean towards Harris given she would be more likely to outperform the polling.
If I were going to be “live betting” the election, I would wait for the NC & GA results before making a decisive call. If Trump takes both, he is a strong favorite. If Harris steals one, she is the strong favorite.
Margin of Error
It’s worth remembering that just because the markets are probably more accurate than the polls doesn’t mean they are highly accurate! Nobody really knows what will happen.
The day before the election Hillary was up 6.5% in WI polls, 3.6% in MI, and 2.1% in PA and lost all three (in each case, the polls were accurate on her support but undercounted Trump).
Similarly, Biden was +6.7% in WI (won by 0.7%), 4.2% in MI (won 2.8%), and 1.2% in PA (which was accurate!).
Markets aren’t suggesting the polls are that wrong again – otherwise, Trump would be well ahead in the WI & MI odds.
All the markets are saying is that, if the polls are reasonably accurate, Trump is a little more likely to win than Harris.
Tail Events
Finally, don’t assume the outcome of this race will be predictable. It’s unlikely we’ll know who won Tuesday. Probably won’t know Wednesday either. In fact, it may not be this month!
If Kamala wins by taking PA MI & WI and losing the rest, she will take the electoral college by one vote (270-268). That means Trump only needs to find one faithless elector to throw a tie to the House.
I’ve written before about other unexpected things that could happen this cycle. If you’re not familiar with the 1876 election, you may hear a lot about it in the coming days.
That was a contested election where nobody could really be sure who won several states. Between the potential for fraud, malfeasance, and legal challenges, the potential for another 1876 is high.
I don’t expect either side to accept the result of a close election without challenges. And even if the losing candidate concedes, that doesn’t mean the voters will.
If you write US political violence or domestic terrorism, I hope your underwriting modeled a scenario where 1876 repeats in the current heated political environment and managed your exposure accordingly.
The one prediction I will make is that people are far too complacent about the risks of the outcome not being accepted. My hope is that one side wins decisively to remove that risk.