The Polls May Not Love Trump, But The Gambling Markets Sure Do. HUH.
Think Donald Trump is a lock in 2020? Wanna bet on it? Well you can't, because betting on politics is illegal in the United States. And we would never advise you do anything illegal. But if you were going to bet on the presidential election, you might do it through one of the many offshorewebsites that will happily take your money. Or so we hear!
An aggregate of betting sites makes Donald Trump (-134) a moderate favorite over Joe Biden (+140). For those who are not familiar with betting lines, this means that a $100 bet on Donald Trump will pay out $74.60. Conversely, a $100 bet on Joe Biden will net $140 if he takes the White House. In other words, these websites, AKA "The House," project a 57.3 percent implied probability of a Trump victory, and only 41.7 percent odds of a Biden win. Or at least, they'd like you to bet as if that were the case.
And while there are certainly plenty of Trump supporters confident he will cruise to re-election, from a polling perspective, the swing state numbers and Trump's negative 8.7 net approval rating don't suggest that level of certainty. So why are the betting sites so gung ho on Trump when the data shows a tight race with Biden holding a slight lead?
For example, the Real Clear Politics polling average shows Biden leading by 4.9 points nationally, 3.3 percent in Florida, 4 percent in Arizona, 6.5 percent in Pennsylvania, 5.5 percent in Michigan, and 2.7 percent in Wisconsin. And yet, RCP's betting odds have Trump with an 8.6 percent edge in the average implied probability.
Why is Biden up 5 points in the polls, and down 8 points in the betting odds? Sure, you need to win the electoral college, not the popular vote. But is there a reason that the bookies responsible for paying out on these bets might project an outcome untethered to the actual data? Hmmmmm, it's a mystery!
Many will point to polling inaccuracy in 2016. And while the aggregate polls did lean 3 percent more Democratic across the board that year, that doesn't necessarily mean that this error is guaranteed to repeat itself. According to FiveThirtyEight, polling bias is "small and unpredictable."
Sometimes the pollsters paint a slightly rosier picture for Democrats, as in 2016, 2014, and 2002. Sometimes they get it wrong in the other direction, as in 1998, 2000, and 2012. You can't restrict the dataset to 2016 only and make a generalization that all the polls have a leftward bias, a fact which professional oddsmakers know perfectly well.
And yet, here's betting site Oddsshark.com making exactly that argument just yesterday:
If United States voters learned one thing following the 2016 election, it's that presidential polling should be taken with a boulder-sized grain of salt.
The voluminous amount of polling data shared by the media all but guaranteed Hillary Clinton would win the presidency and she didn't. Donald Trump flipped the script with the biggest political upset in history and if you were one of the select few who saw value on oddsboards in the days leading up to the November 2016 election, a significant profit could be made. Trump's odds to win the 2016 election were near 5-to-1 the day before the election. Wagering on political outcomes seems a lot more enjoyable, and financially wise, than arguing with strangers on social media.
Ignore the media and those stupid polls, make a "financially wise" decision to gamble online instead!
The reality is that we're five months out from the election, no one knows what's going to happen with coronavirus, and the only reliable tool to forecast the outcome in November is the polling. If you simply declare that this data is all junk, then your projection isn't a calculation — it's a hunch.
So why might the betting sites benefit from catering to those who support Donald Trump? Could it be that they know what makes their customers happy and are simply giving it to them? Or could it be that The House is nudging its Trump-loving customers to bet with their hearts and not their heads? Remember, if Donald Trump loses, the bookie gets to keep all that MAGA cash.
According to Sportsbettingdime.com, almost 70 percent of betting money is on the president being reelected. Trump boasts a giant enthusiasm edge, and a majority of Americans expect him to win again, or they did before the pandemic locked us all in our houses. So the oddsmakers have very little incentive to emphasize data that their customer base has decided is fake news.
This is almost a perfect storm for betting sites: Trump is a popular brand with enthusiastic fans, and Biden's support is basically tepid. The bettors themselves like Trump, and crapping all over traditional forecasters is never going to alienate them. Meanwhile, if Trump loses, the site pockets the cash on all those losing bets from people who decided — with the site's encouragement — that polling data was bullshit. What a deal!
Which isn't to say that the oddsmakers are fudging the numbers here. But it is ... interesting.
Follow Data Goon on Twitter RIGHT HERE!
Please click here to support your Wonkette. And if you're ordering your quarantine goods on Amazon, this is the link to do it.