The pollsters and bookies actually called the referendum right

The pollsters and bookies actually called the referendum right

Pollsters have come in for a lot of criticism in recent years: first it was the 2015 General Election and most recently over their work relating to June’s referendum. However, I believe they got it right on Brexit, but their results were simply misinterpreted.

Indeed, on my own analysis of the polls the night before the referendum I was totally convinced that the Brexiteers would win. My view is backed up my more detailed analysis of the bookies’ odds, which again the pundits misinterpreted.

As far as the polls were concerned, everyone was focusing on the timeline, but this only revealed only part of picture. For each poll taken there were four key pieces of information revealed: the date of poll, the result, the polling organisation and the sample size of the poll.

Take a look at this table of results.  If you examine the last fifty polls in the two months leading up to the referendum, from 26th April, what you see is that about 50% of the polls showed Leave winning and the other 50% going to Remain. The timeline shows that the Remainers were leading the early polls but as time moved on, Leave was gaining ground. In the last few days, and particularly after the tragic death of Jo Cox, the results swung back to Remain and then reverberated.

Now look at the results tabulated according to poll sample size and it is a revelation: 76% of all polls with a sample size of 1,100 or more show Leave winning. 81% of all polls with a sample size of 1,100 or less show Remain winning. When the sample size was larger, it was quite clear, to me at least, that betting on Leave was a good bet. Coupled with my analysis of the bookies’ position, I was certain which way it was heading.

During the campaign, one often heard people say “the bookies never get it wrong” and of course the odds were in favour of Remain. However, this is because people misinterpreted how odds are calculated: they are worked out on the basis of the total revenues placed on either side of the bet.   If one person had bet £1million on Remain and one million people had bet £1 on Leave, the odds would have been even. But we are not dealing with a horse race or football match.

In this situation, the people betting are the ones voting and therefore they have the opportunity to influence the vote. As our democracy works on the basis of one person one vote, what is important here is the number of bets placed, not the size of the bet. In the event, the average bet to Remain was £475 and the average bet to Leave was £75; but there were three times as many people betting to Leave than Remain.

Whilst this may have shown that Leavers like to gamble more than Remainers, it also demonstrated that there was a greater chance that Leave would win. The pollsters and the bookies were spot on.