Latest news



The Money Says Prepare for a Long Night, and Week

Points of Return

Coming Soon: 2020 Hindsight

This is a blissful moment; I am about to write my last ever piece about the U.S. elections of 2020 without the benefit of hindsight. More or less everything seems inevitable once it's happened. From here, nothing looks inevitable. Any number of strange phenomena may turn out to be important, or may soon be forgotten. So, while the U.S. elections remain like a Whodunnit novel that is missing its last page, let me try to review what is going to happen Tuesday and the rest of the week.

There are a number of ways to try to predict this. Data-based models, using aggregations of polls, make victory for Joe Biden in the presidential election close to a certainty. Market-based models, in which people risk real money buying futures or placing spread bets tied to the results, agree that he is the favorite but suggest that the election is far tighter, and that Donald Trump would win at least one time in three. This chart from Macro Hive summarizes the latest predictions:

Spread betting markets, as aggregated by RealClearPolitics, see the odds almost identically to the Predictit market:

When we look at individual states we see further differentiation. Again, the prediction market bets at state level are consistent with a Biden victory, but a much narrower one:

Election Night:

The pollsters think Biden wins Florida and (much more narrowly) North Carolina and Georgia. The bettors think Trump wins them (very narrowly in the case of North Carolina). Results for all three states should be clear by midnight on the eastern seaboard (he wrote, without the aid of hindsight). Should Biden win them all, his victory would be a virtual certainty and markets would respond accordingly. If Trump wins them all, everything gets much tighter. All of this should be in the bag by the time European markets start to trade.

But be careful of the results as they come out. Because both Florida and North Carolina announce all their postal votes (which tend to skew Democratic) at the outset, these states will likely start with a big advantage for Biden to be followed by a "red shift" toward Trump as on-the-day ballots are counted. That could lead to some market head fakes. There are no obvious red or blue shifts in the way Georgia will report. FiveThirtyEight has an invaluable guide to these technicalities.  

Election Week:

If he fails to wrap up the three Southern states on election night, Biden could still win if he holds on to the states won by Hillary Clinton, and adds the three rust-belt states that she lost by a whisker: Michigan, Pennsylvania and Wisconsin. Of these, none will have results by midnight. Wisconsin hopes to have a result not too long after daybreak Wednesday; the guidance from the other two is to wait until Friday. So if Biden has failed to tie up the states in the south, there is every chance that the saga of counting the votes drags on all week, while the world reacts.

If Biden wins those three states, he can lose all the other "battleground" states, and still win the presidency. That is indeed exactly what Predictit is currently projecting — the prediction market's electoral map at this point shows Biden losing all the southern battleground states, and also Arizona, but still winning thanks to the rust belt. The snag is that we would have to wait at least until the end of the week to be sure that this had happened. 

So if Biden wins some of the southern states on election night, it's a virtual certainty that he wins; if he doesn't win any of them, then nothing is certain, and a disputed election becomes far more likely. 

For those in Europe planning to stay up for the results, it will probably be safe to go to sleep by the time the three key southern states have announced. 

The Senate:

That covers the presidency. When it comes to the Senate, there is again a difference between the modelers and the markets, although it is tighter. FiveThirtyEight puts the Democrats' chance of winning the Senate at 76%, while Predictit has odds of 59%.

Again, there is unlikely to be clarity on election night. The Democrats need a net pickup of four seats for an overall majority, which in effect means flipping five seats as the Republicans are likely to reclaim Alabama. Two "must-wins" for the Democrats are Colorado and Arizona, in the Mountain time zone. Both should be more or less clear on election night, but possibly not until after Europe has started trading. The other must-win for Democrats is Maine, which should be available by midnight; if the moderate Republican Susan Collins holds on to her seat, that would presage disaster for Democrats, and the pollsters. 

The other key states that the Democrats want to pick up are Iowa (where the betting is now that Republican Joni Ernst will prevail); North Carolina (where results should be available on election night, and the Democrats are now thought likely to prevail); and Georgia, where two seats are up for election thanks to the retirement of a Republican incumbent. If all states go as currently expected, that will leave the Democrats with 50 seats before counting Georgia.

Now, another catch. Both Georgian races are very close. Under Georgian election law, if no candidate wins 50%, there needs to be a run-off election, in January. It is likely that at least one of the Georgian races will have to wait until next year. Amazingly, the Predictit odds at this point suggest that this is exactly what will happen, with one Georgia seat held by the Republicans, and the other going to a runoff. That would mean control of the Senate would have to wait until next year:

Now, take on board that Senate control requires 50 seats for the party that controls the presidency (because the vice president breaks ties), and 51 for the party that isn't in the White House. Also, the chances of the party that loses the presidency in any Georgia run-off will be much improved; they can act as a check on the presidency. Also, the Georgia races promise to be polarized. One of them features a Black pastor who preaches from the pulpit in Atlanta once used by Martin Luther King, against a Republican businesswoman who describes herself as more conservative than Attila the Hun. 


Nightmare Scenarios

The nightmare scenarios are, indeed, nightmarish. This is more because of flaws in the U.S. constitution than the behavior of the current president — although his willingness to cast doubt on an election result before it has happened is alarming.

The constitution is ambiguous on several possible points of conflict over the next few months. For example, it is not clear exactly who gets to appoint electors from each state (the legislature or the governor?), and who gets to count the electoral votes once they have been submitted. One reading is that this would be the job of the sitting vice president, who is running for re-election. For a brilliant recap of the potential problems, I recommend this podcast, which goes through a series of role-playing games to assess the risk of a contested election. One hinges on a National Guard member in Michigan taking it upon himself to destroy a sack of ballots. Without laboring the point, some of the scenarios, and the risks of violence and disorder should they play out, are truly alarming. It should be great listening to while away the time before the results start to roll in. 


Prediction Markets vs Pollsters

Who should we trust more? Pollsters have actually asked people how they're going to vote, and believe they have dealt with the methodological problems from 2016; prediction markets are expressions of opinion, but they are opinions with the significant backing of money at stake to make them disciplined. Over history, they tend to be surprisingly similar. From 1988 to 2004, the National Bureau of Economic Research found that the Iowa Electronic Markets tended to predict U.S. election outcomes very slightly better than the Gallup poll — over that time, markets had an average error on presidential elections is 1.4%, while Gallup's was more than 2%.

What's driving the difference between the two two pollsters and the markets? Mainly, the previous presidential election. Rightly or wrongly, the perception is that there was a major polling error in 2016. The prediction markets capture this, accurately showing that very few Americans regard the election as a foregone conclusion regardless of what the polls show. A key weakness of prediction markets is that they incorporate the biases of their bettors. For example, in sports prediction markets, the chances of African-American competitors tend to be slightly underestimated. In the same way, liberal angst and conservative confidence might be giving Trump a chance that isn't justified by the empirical evidence. Either that, or the betting markets have spotted a systematic flaw in opinion polls.



Either way, prediction markets are more likely to capture the implicit assumptions of investors in financial markets and suggest that they are straddling two options — either a disastrously inconclusive result with conflict and crisis in the aftermath, or a "Blue Wave" that should add a lot of oomph to reflation trades.

One obvious manifestation of this is in implied volatility, with both equity and bond volatility indexes unduly elevated for a foregone conclusion Biden victory, but still looking subdued if the nation is on the verge of civil unrest and constitutional crisis:

As for individual markets, the theme is consistent — with positioning also affected by the recent worrying news about the virus. They aren't fully positioned either for a Blue Wave or for a constitutional earthquake.

Citigroup Inc.'s summary is that parts of the credit world remain vulnerable to a spike in risk aversion after the election. The same is true of the bid to invest in value over growth, which is contingent on reflation and also on an early vaccine.  It may be credit in Europe and emerging markets that bears the brunt of an all-out post-election fight. As for equities, even though indexes are still not that far away from records, exposure is relatively limited:

If Biden sweeps through the South on Tuesday night, expect people to start buying equities, and betting on reflation. If he doesn't, the VIX is ready to go much higher. As for investors, the secret will be to keep calm and carry on.


Survival Tips

It's unfortunate that this section is called "survival tips" at a time like this. After watching the local news, which features businesses across New York boarding up to guard against election night disorder, along with similar scenes across the country, surviving begins to seem like a challenge. The apprehension is palpable. And it's hard to come up with anything to meet the challenge.

Here is my best effort. Looking back to the early days of the lockdown, we lost Bill Withers, and a lot of us discovered that he'd been one of our favorite singers and songwriters without our ever knowing it. His music lit up a particularly dark period in the spring. Now maybe he can lighten up a dark autumnal season. Try listening to Lean On Me, either live or on record; or Just the Two of Us; or perhaps best of all in current circumstances, Lovely Day. There is a warmth and honesty to his music that is very rare and touching. Now to brace for a long night…

Like Bloomberg's Points of Return? Subscribe for unlimited access to trusted, data-based journalism in 120 countries around the world and gain expert analysis from exclusive daily newsletters, The Bloomberg Open and The Bloomberg Close.


Before it's here, it's on the Bloomberg Terminal. Find out more about how the Terminal delivers information and analysis that financial professionals can't find anywhere else. Learn more.


Post a comment