Betting Odds and Election Outcomes

My colleague David Bell has a short paper on how opinion polls and gambling odds are predicting the outcome of the Scottish Referendum. He notes there are a number of potential limitations in using odds as unbiased predictors of outcomes particularly if markets are very thin. There is also obviously a good literature on prediction markets more generally and their relation to the Efficient Markets Hypothesis (see, for example, Robin Hanson’s excellent blogposts on this). With all that in mind, it was tempting to see what odds are available for the next general election in Ireland. One prominent alliteratively named firm has odds for the next general election as being (as of 5pm 29th May): FF/FG coalition 5/6; FG/Lab coalition 7/2; FF/SF coalition 6/1 and so on. This seems pretty consistent with other firms. Would be interested in hearing what people make of this.

15 replies on “Betting Odds and Election Outcomes”

Not a bad start from the bookies. They have the favourite right, even if the odds are very short at 5/6. FF/SF at 6/1 is a reasonable bet, but SF are unlikely to enter any govt as an appendage and would only enter at 50/50; If FG/Lab continue as before, I would not rule it out.

The FF/Lab is totally wrong.
Labour has lost the left vote. Labour has in fact left the Left.
Whether Labour elect a young turk or an old turkey, is irrelevant. Barring a volte face they are finished and a volte face means the government is also finished.
The only hope for the Labour Party to recapture the left vote, would be to insist on no more cuts and that the deficit gap be made up of tax rises (or preferably tax break removals), That and a UKIP like campaign against the euro, might reposition them. Might.

(from another thread, re changed backdrop for State funding)

The deficit gap will regrettably have to be closed, regardless of which administration is in place. However the idea of reducing State cash balances should be immediately shelved; they will be needed.
The deficit gap will have to be reduced now by tax measures only. Even the FG purists will know that this is the case.
[The EZ morons should be told to take a hike with the SGP target of 60%, a nonsensical piece of rubbish, that most of the countries have no notion of adhering to.]

There is some uncertainty as to whether the ECB will perform the minor miracle of pulling the EZ back from the deflation precipice. The odds on favourite in come circles is it will be another case of too little, too late.

Expectation is everything in betting, if the doom and gloom specter looms large, anything can happen. If the usual behaviour continues, namely “The rabbit in the headlights. Too terrified to move.” Then FF alone or FF/FG or FG/FF. May God help us in our helplessness as Johnny Nolan used to say in stressful times.
The whole organised betting scene is best avoided as the surest bet is you will lose 5 to 10 per cent of your outlay.

FF/SF at 6/1 is a great bet. I would look for odds on SF/FF though. Gerry has a decent shot of being An T in 2016 and he will probably be T in 2021 for the centenery of the Treaty.


In 2021 Gerry Adams will be 73 years – no chance as a T. In fact I’d be amazed if Gerry Adams stood again in two years time.

There is very little chance of him standing down given he has a chance of the top job in 2015. Without him the Shinners are bereft of a strategic brain.

A delightful little paper close to my heart. I always use the betting markets to inform me of the likelihood of all sorts of events and yet I rarely bet. Best of all is B*tfair. This is an exchange between punters – BF get their profit from a commission.

Let’s look at an example – this morning’s BF market on the number of UKIP seats in the next GE. There is less than 1% bid ask spread so although probably not very active this is highly liquid. They split the market between 3 categories – No seats is favourite at 5/4; One to Five is next at 7/4 and it is 3/1 more than 5. This for a party with 27% of the Euro vote and 24 seats – if you asked anyone how many seats in the 600 seat parliament UKIP would get, surely very few would say none and yet that is the favourite. So when I read headlines about earthquakes in UK politics I consult BF and conclude not even a minor tremor.

Another example. Barak Obama started to slip in the polls up to the last presidential election to the point where all the professional pundits were shouting too close to call. I consulted BF. Odds for Obama win were still 1/6.

Another example. UK election night. By the time the first few votes are in the pundits are still waffling away with their swingometers but as far as BF (betting in play) is concerned the result is in bar the shouting. I can safely go to bed.

As I write, BF are quoting 75% a Scottish No and 25% Yes (0.8% bid ask spread). It’s been like that for a while. It is difficult to interpret as a predictor. Normally if there was a referendum with opinion polls running at 60% there would scarcely be a market on the event as it would be regarded as such a slam dunk. The market is saying that the result is much closer to call than would be suggested by the polls. So what about EMH? Not so sure. The polls suggest, say, 95% certainty so it seems a lot of “smart money” is forcing that down.

One thing for sure punters rarely bet with their heart. BF go 28/1 England to win the WC even though the great majority of their patrons want that result. So EMH is not distorted by the outcomes that people want. All the same situations can arise where the main body of evidence points one way but canny observers see things the overall market is missing. The Scottish Referendum looks like one of those situations.

Another interesting BF market. Next US president – bid/ask 0.5%. Hillary Clinton 11/8 faourite, Jeb Bush 11/1, 12/1 bar

@ Tull

FF/SF have 34 seats between them at present. It would take a remarkable surge to reach 83 from there and considering any comeback by FF will likely hinder the SF train, I don’t see that happening.

Consider this: FF won 20 seats in 2011 with 17.5% of the vote. SF only got 15.2% of the Local Election first prefs. On that showing it is hard to see them getting more than 20 seats at the next GE. People are totally overreacting to the apparent success of SF in the Euros but this is about as irrelevant as UKIP’s success in the UK Euros.

My bet would be that Labour will make a partial recovery with a new leader ; the economic situation will be looking brighter and so FG/Lab will be re-elected. Fill your boots at 7/2.


If the government lasts full term, it will be a majority of 158 seats which is required, not 166 as at present. But if, as predicted , it breaks up before the five years are up then the election will be based on the previous Dail boundaries.

The new boundaries will favour FF and operate to Labour’s detriment, according to NUIM analyst Adrian Kavanagh. So FF will capitalise, particualrly in rural areas on teh strong vote it secured as well as its 25% share of teh vote nationally; though ti is now clearly a rural party rather than having any great urban strenght.

Overall, it is premature to suggest any seismic shift in national politics – or more widely in European politics – from the EP election results. The most commonly quoted argument is that since most voters don’t know how the EP works, and don’t particularly care either since it is too far remote from their own daily life concerns, votes in EP elections are about national issues where it is ‘safe’ to vote for fringe parties/candidates as it makes no difference to the national political context beyond delivering a message to mainstream national parties, especially government parties. Perhaps even more important, nobody should underestimate the resilience of established mainstream parties to see off interlopers, whether from left or right, between the EP vote and the next series of general elections. The mainstream have capacity, mature organisation and resources on their side, as well as already being securely embedded within the state system.

So the best advice might be to hold off on the betting on the likely composition of the next government folks, and wait for the dust to settle. Meanwhile, our own government will be in a ‘state of chassis’ until the outcome of the Labour leadership is decided.

In judging whether EMH applies to betting on events, we should distinguish the extent to which an event is truly random and the extent to which it seems random because of our uncertain knowledge. Consider a premiership football match. The outcome of the result would scarcely affect the odds were the two teams to meet again in similar circumstances a week later. To that extent the outcome was truly random. But consider next week’s Epsom Derby. If Australia (the horse) wins that at even money he will be well odds on for the subsequent re-run 4 weeks later at the Curragh whilst if he loses he will go out to long odds. So at this point in time the market judges on balance that it is 50/50 that Australia is a goodun’, but he either is or he isn’t.

The Scottish Referendum is more in the second category. When the result is in, let’s say they had to re-run it on a technicality, the odds second time round will be greatly influenced by the first outcome, thus it is more like the Epsom Derby than a Premiership football match.

So where does EMH come in? The result is there to be known if we could read the indicators properly – it is not like guessing the throw of a die.

A bit like the great bank share collapse – there we had bank shares falling like a stone but still many learned commentators were talking the prices up, in good faith. But the shrewdies knew. Was AIB at €5, having fallen from €20 priced consistently with EMH? The market was probably quite segmented between those who “knew” the game was up and those who were misinterpreting the available knowledge.

Financial & Corporate System:

Betting suspended: rigged to within an Angstrom of a certainty.

That 1/17 was simply fabulous value ….

…………………. a rare ‘error of judgment’ by the bookie!

Comments are closed.