Ahead of tonight's curtain raiser to the Democrat primaries, Paul Krishnamurty explains why we should treat early polls with caution...
"In-play betting in US primaries is often a remarkable spectacle. Results come in district by district...the market invariably over-reacts to the latest result."
The 2020 US Election starts tonight, with the Iowa Caucuses. As always, Betfair markets will stay open throughout the counting process until a winner (most delegates) is declared. It could well be a very long night!
Polls contain huge margin of error
Polling signals are far from clear. The psephological and political betting communities were on tenterhooks over the weekend, awaiting the sainted Des Moines Register poll - only for it to be cancelled due to errors. Naturally, being the 21st century, conspiracy theorists had a field day.
The final seven surveys all showed Bernie Sanders ahead, with his lead ranging from a top end of 7% to tied with Joe Biden. That is reflected by the betting, with the Vermont Senator rated odds-on at [1.6]. However those with experience of caucuses, particularly this one, urge caution.
The margin of error for polls of the Iowa caucuses is about 10 percentage points on each candidate's vote share, **ON ELECTION DAY**. Don't be surprised by an upset on Monday.? G. Elliott Morris (@gelliottmorris) January 31, 2020
Sanders outperformed 2016 polls
That message is reinforced by events in 2016. Hillary Clinton defeated Sanders for the Democrat caucus by a miniscule 0.3%, having led in five of the last six polls, three of which were by eight points or more. Ted Cruz won the Republican race, despite trailing Donald Trump in the polls.
The previous presidential cycle - when there was only a Republican primary - produced a political betting market for the ages. Rick Santorum - rated a no-hoper just a fortnight from polling day - and Mitt Romney swapped favouritism all night, swinging back and forth literally every few minutes. If memory serves, the result came in around 7.30am UK time and then took months to resolve.
Big rise in turnout would add to unpredictability
There are numerous logical explanations for the uncertainty. First, as ever, differential turnout is liable to completely blindside the polls. Based on the evidence of numerous special elections during the Trump era plus the mid-terms, I expect turnout among women and young people to soar.
With no Republican race this year, expect plenty of them to switch over, whether in protest against Trump or to back the Democrat they'd like him to face (that person is Bernie Sanders).
Second, these early primary voters are notoriously late deciders. People start to truly engage in the final fortnight, as each candidate holds multiple rallies across the state. Something like half of all voters have suggested they are either undecided or open to change.
Caucus system is all about transfers
Third, the voting system. Caucuses are different from regular primaries or conventional elections with one non-transferable vote. Here, voters attend a local meeting and try to win votes from rival campaigns. If a candidate fails to reach a threshold - minimum 15% - in the first round, they can switch to another candidate.
That dynamic is highly pertinent in this Democrat race, because candidates to some extent fall into ideological 'lanes'. Few expect Andrew Yang or Tulsi Gabbard to hit 15% anywhere. Amy Klobuchar may also struggle to do so and possibly even, according to some estimates, Elizabeth Warren. In a particular precinct, one or more of Bernie Sanders, Joe Biden and Pete Buttigieg could fall short as well.
Therefore, successfully predicting this race not only involves predicting share of vote and delegates for each candidate but whom these voters would award a second preference to. It is generally believed, for example, that Biden, Buttigieg and Klobuchar supporters will transfer to one another.
On the other side, I'm confident that Yang and Gabbard supporters will transfer to Sanders. Their support is driven by a very similar, Kremlin-amplified social media phenomenon.
The interesting one is Warren. Some polls show her very competitive but others being eclipsed by Sanders among her progressive base. There is some suggestion that she'll thrive among late deciders. I still regard her as the potential unity candidate but she'll need to finish top-three at worst tonight.
In-play betting on US politics can produce carnage
In-play betting in US primaries is often a remarkable spectacle. Results come in district by district, with a live leaderboard (CNN will have live coverage on Sky Channel 506) updating by the minute. The market invariably over-reacts to the latest result.
A fundamental rule to keep in mind throughout the entire 2020 process is that the cities and areas with the largest populations declare last. Rural counties go first but they won't be reflective of the overall result.
So far as tonight is concerned, that could offer an angle on Bernie Sanders. He will not thrive in those rural counties but will sweep younger voters in the cities and the university areas. I expect him to win - on the basis of his superior ground game and motivated supporters - but hope to get a lot bigger than [1.6] at some stage in-running tonight. Place an order to back him at [2.2].
Place order to back Bernie Sanders @ [2.2]