Bookies and Pollsters

Strange Bedfellows

There are studies and surveys and polls for all kinds of stuff out there in the political realm, and some of the more wonkish ones end up turning into the most fascinating, from what I'm starting to see.

The Brookings Institute is a familiar name to lots of folks who touch into the arena of professional political sciences. They conduct some of the most reliable and re-producable polling data available, and have a track record of being quite thorough.

Back in 2016, however, they found themselves curiously on the losing side of a fight they weren't even aware that they'd been pulled into. You see, while the Institute was pulling data from their traditional sources regarding what they expected to see out of Senate, House and the Presidential races, there was a small cadre of professionals of another sort who were weighing in against them, and calling picks of their own based on metrics not normally looked to:


According to Hart Research Associates (whose works I've been diving into the last few days), an informal association of above-board and under-the-table sports bookies and money men in gambling circuits around the country had started seriously getting into taking big money bets on political races as far back as 2008. Most of their odds-calculating mechanics early on were based purely on data sets made publicly available from pollsters, prior election results, and suchlike. However, in 2010, in the run-up to the mid-terms of President Barack Obama's first term, several Atlantic City bet takers began to notice a trend among their lower-sum betters; these largely blue-collar workers were grumbling about local matters, in Jersey and New York, instead of talking national level issues.

Over the course of four years, a small group of these bookies started adjusting the odds they would offer out based on several such observations, most of which have since been tested by and concurred with by both Hart and Brookings. The basic set of principles they arrived at are as follow:

1) The lower the level of the office or initiative on the ballot, the more accurate the early polling of that voter base. In short, the seats on school boards, for town council, for city council, for local ordinances, and for mayors of small to mid-sized towns are far more predictable in outcome if you ask around ahead of time among likely voters. On the local level, folks seem a lot more inclined to arrive at a decision and stick with it all the way through the vote, and they are far less likely to dissemble in local matters.

2) Lower and Lower-Middle Class Voters don't turn out to vote on matters that don't immediately effect them. Schools are funded via local property taxes, by and large. If someone doesn't have the good fortune to own their own home, they are far less inclined to vote on school funding issues, because no matter the result, nothing changes for them on a personal financial level.

3) Weirdly, if you are FROM a small municipality, you have LESS of a chance of winning any public office race in that small area than you do a seat in the nearest major metro. By and large, this seems to stem from locals having a priori knowledge of you, and not necessarily wanting to be led by you, even if they intimitely KNOW you. This seems to lend some credence to that old bit of doggerel that 'Familiarity breeds contempt'.

Yet some of that 'small town charm' seems to really go gangbusters with the big city crowd early in one's political career.

4) There is a bookie in Vegas, colloquially known only as 'Professor Right', who has accurately predicted every Senate, House, and Presidential race since 1996. Just a neat little aside.

5) "The 15 Point Problem"- A phenomenon that this cadre of bookies started relaying to Hart and Brookings in 2016, there appears to be a high likelihood that any ballot measure or bill in Congress will invariably get shot down if the gap between early whip counts by percentage are polling with the public by a percentage gap of 15 or less. In other words, if a proposed bill has 40% of the legislature on board, it needs at LEAST 56% positive response in public polling. Otherwise, the bill dies. Every. Single. Time.

Oddly, this seems to cease working if the bill has less than 20% support in Congress; nothing gets off the launch pad then.

6) Voter Blocs treated as monoliths rebel every third election cycle. This from Brookings Institute alone, they have found that registered voters who are frequently identified as belonging to smaller blocs tend to become resentful toward their affiliatef Party every 3 cycles, feeling they are less valued because their support is assumed to automatically be there. Union groups are the least likely to succumb to this pattern, but it hits a lot of them, too.

7) Registered Republicans will vote for a conservative or 'Blue Dog' Democrat if the Republican candidate fails to be at least equally conservative to that Blue Dog along fiscal lines. Registered Democrats will vote for a liberal Republican only when or if they demonstrate a track record of supporting education reforms and funding. It's an odd dichotomy, but there it is.

8) When Atlantic City and Vegas are in agreement on odds, the favored candidate in every race wins 82% of the time. When there's disagreement greater than +/- 6 points between the two, the underdog wins 77% of the time. Again, odd, but it seems to work.

9) More gambling addicts stay afloat on political betting than any other type of betting per anonymized self-reporting from Gamblers' Anonymous, counts submitted to Brookings.

10) The last and perhaps most intriguing part of all of this for me, so far, is the unpredictability of 'Early Independents'. Bookies love when a district is reliably and staunchly either D or R, and they despise Independent and Libertarian voters. These two groups of voters throw a major wrench into odds whenever they show up in numbers, because regardless of their genuine support of a third candidate, most betting odds lines only offer binary choices in public office bets. In the rare event that a third candidate has won an upset, no formal arrangement exists to pay out bet takers.

Side-note 1: No bookie in Vegas will accept bets on Nevada-based races. No bookie in Atlantic City will accept bets for New Jersey-based races. The rationale seems obvious.

Side-note 2: Bookies in both locations since 2016 have introduced a new wager metric focused entirely on one side of the political spectrum: gamblers can try to bet on how many times a Republican candidate will be accused directly or indirectly of being a racist on mainstream media outlets (NBC, CNN and MSNBC were used that first year, along with The New York Times). In 2016, the closest wager was placed by one Derrick Eddings of North Branch, Minnesota. He bet the total would be just under 100; the total was 837 times in the six months leading to the election between just the previously mentioned 4 sources.

Side-note 3: Ever the opportunists, book makers added a new wager metric in 2018: how many times the following 6 sources labeled Democrats as socialists or Marxists- The Daily Wire, Fox News, Newsmaxx, OANN, The Blaze and The New York Post. The nearest wager was 500; the actual count for those midterms was 721.