Paul Krishnamurty, aka The Political Gambler, has been betting on elections for over a decade. A politics graduate from the University of Hull, he also works as the chief political analyst for Betfair UK.
In his September update, he looks over the factors contributing to Donald Trump’s short time as favourite, as well as how increased voter numbers will likely damage the incumbent’s chances.
The six weeks since my last update have been, if not stressful, perplexing for we Biden backers. Whilst $1.72 odds to win still represented a significant under-estimate, at least the market had moved in our direction. Then came the RNC.
The post-convention gamble on Trump was extraordinary. Biden drifted to $2.00. This without any meaningful evidence of change. The political betting community is still debating what lay behind it.
One legitimate explanation is speculation that the violence in Kenosha and Portland would work to Trump’s advantage. My initial instinct was the same but polls never supported it. In Wisconsin itself, Biden moved further ahead.
I think a more fundamental reason is simply Trump’s fame. This race is all about him and everyone has an opinion. By comparison, non-Americans know little about Biden. Moreover, gamblers remember him upsetting the odds and defying the pundits last time. Trump backers amount to around 80% of all bets, across the industry, as far as I’m aware.
That suggests this market is unbalanced and therefore potentially irrational. It won’t surprise me if further avalanches of money arrive to back him as interest grows closer to polling day. That means laying back my position has become unlikely. I won’t take what I calculate to be the wrong odds.
Let me justify in numerical terms. First, see the relatively low estimates from Fivethirtyeight’s polling model compared to implied Betfair odds. They have Biden at 78% for the presidency, compared to 56%.
8⃣ Latest odds on those four pivotal states, compared to @Fivethirtyeight forecast.
Arizona: 58% v 64%
Michigan: 72% v 85%
Wisconsin: 71% v 82%
Pennsylvania: 66% v 77%— Betfair Exchange (@BetfairExchange) September 28, 2020
Don’t believe the polls? Then see the betting. I don’t know a single forecaster who believes Trump can win without either Arizona or Pennsylvania. Yet he can lose whilst still winning one of them. In a less likely scenario, he could win both and still lose.
If you want to back Trump, do so in these two states or another permutation involving Wisconsin or other Biden targets. The outright odds simply don’t correlate. Trump is either over-rated for the presidency or under-rated for these states.
Obviously, I believe the former. Consider the bare numbers. Trump won with 63M votes, with 46.1% of the vote. He will need a higher share because, unlike 2016, there is no strong third party. There is no perfect measure for popular vote, but I estimate he needs at least 47.5%.
Turnout was 55.7% and is widely expected to rise, given the mid-term surge and record engagement levels. I think it could rise well above 60% but even 59% has him needing 5M plus more voters.
We know from all the mid-term trends that this extra turnout skews towards Democrat-inclined segments. Young voters, women, minorities. Particularly African Americans, among whom turnout fell in 2016, to ruinous effect in the key states. The Democrat advantage in new registrations in Pennsylvania is three times Trump’s 2016 margin, and that’s before counting many multiples higher of Independents, who lean Biden.
Where is the logic behind Trump gaining 5M plus extra voters? His approval rating fell almost immediately upon taking office and remains low, set against record strong disapprovals. Trump makes little effort to reach out to opponents, preferring to double down and excite his base.
It is the perfect recipe for creating the current state of play. An entrenched electorate which is roughly 55-45 against Trump, that rarely shifts. Two tribes locked into different news cycles, consuming different facts. This is why scandals that would have ruined former presidents never move the needle. There are record low numbers of persuadables.
I see no reason to distrust long-term polling trends when they merely reflect these wider societal trends. If there are 5M plus extra Trump voters, why do they *never* turn up in high-grade polls? Biden has been polled against him since 2015 and led in almost every one, often by double-digit margins. An entrenched 7-8% lead looks completely believable.
To predict 2020, we need to understand the precise context of 2016. Hillary Clinton was a damaged candidate and it is no surprise to see Biden outpolling her. He isn’t under FBI investigation and hasn’t been the subject of quarter-century-old conspiracy theories. This Pennsylvania boy has greater appeal to blue collar voters in the Rust Belt.
Prior to that incredible night, Michigan, Wisconsin and Pennsylvania were considered Democratic strongholds. Trump didn’t massively expand the GOP voter base there. He won with sub-50% totals because third parties split the vote. All the evidence, whether real elections or polls, strongly suggests all three states have since reverted to type.
Besides the main market, my bet is that Biden Electoral College Votes lands in the 330-359 band. Currently a $6.60 chance.
The best value trade, however, is to build some kind of hedging plan to exploit the mismatch between outright and state odds. For example:
BACK (WIN) — Joe Biden for 100 units at $1.79
BACK (WIN) — Trump for Pennsylvania for 22 units at $2.95
BACK (WIN) — Trump for Wisconsin for 18 units at $3.40
BACK (WIN) — Trump for Arizona for 25 units at $2.40
Here’s a challenge. If Trump’s odds for the presidency are good or even merely fair value, explain the flaw in that hedge.
Can anything change in the final month? I seriously doubt it but there are two potential worries. First, Trump has taken more hits than anyone and is still standing. It is possible, if not likely, that Biden crumbles under the spotlight of a campaign.
Second, Kremlin-led conspiracy theories such as QAnon are all over social media again – with the purpose of helping Trump, according to his own hand-picked FBI chief. I reckon US voters will be wiser to it this time around, but their effect remains a known unknown.
Such outcomes have a probability which is factored in. I wouldn’t lay Biden at anything higher than 1.35 and even then, I wouldn’t because there are better ways to have any desired saver on Trump.
Check out our interview with political betting expert Paul Krishnamurty on the Betfair Hub.
Check out our latest Pro Punter interview with Tavis Rendell as he discusses the 2020 US Election and where ...
With the QLD Election fast approaching, set to be held on Saturday 31 October, the Exchange market is starting ...
