Voting in this Thursday's UK referendum on membership of the EU takes place between 7am and 10pm. Unlike last year's General Election there will be no official exit polls, partly because of the difficulty of extrapolating from a sample to a wider population in a referendum. Nonetheless, the Financial Times reports that some hedge funds have commissioned their own exit polls in order to trade on early indications of the result. The movement in the value of sterling during the count of votes from 10pm on Thursday evening will provide one signal of market sentiment about the outcome.
Instead of parliamentary constituencies, the geographical area for the referendum will be 382 counting areas, one for each local government area in Great Britain, and one each for Gibraltar and Northern Ireland. Each counting area will make a separate declaration which will be aggregated by eleven regional counting officers. The national result will be announced in Manchester Town Hall by the Chief Counting Officer.
Counting will start when the polls close at 10pm. The Electoral Commission estimates that the first result will be declared at 1230am and the last at 7am, with most results coming in around 4am. By 5am a number of news organisations and pollsters are likely to have produced projections for the final result. The Electoral Commission says that the official results should be in by "breakfast time" on Friday 24th June.
A year ago most opinion polls showed a significant lead for Remain, of the order of six-ten percentage points. Since then the Remain lead has narrowed. As of Sunday evening at 10pm the average of the latest six opinion polls excluding Don't Knows showed Remain and Leave neck and neck at 50% each. Remain's lead has improved since the middle of last week when it was on 47% and Leave was showing 53%.
The bookmakers' odds, based on the flow of bets, suggest a significantly lower probability of a Leave vote. As of Sunday evening the odds offered by bookmaker Paddy Power implied a 31% chance of Brexit.
Why is there such a divergence between the neck and neck picture painted by the polls and the bookies' odds of Leave of just 31%?
Betting markets may be assuming that Don't Knows, averaging around 10% in June's polls so far, will go with the status quo on the day and vote for Remain. This would be a repeat of the 2014 Scottish independence referendum where Don't Knows tended to support the Union.
The betting odds may also reflect a view that telephone polls, which tend to show a stronger position for Remain, give a more accurate picture than the on-line polls which have been rather more favourable for Leave. One theory is that by asking people to choose between Remain and Leave telephone polls confront people with the choice they will face in the polling booth, avoiding the escape clause of "Don't know" widely offered by on-line polls. Including a Don't Know option seems to siphon support from Remain and, on this theory, provides a misleadingly strong showing for Leave.
This may be the case, though in the last fortnight Leave's position in telephone polls has strengthened. Between January and May just three telephone polls, 10% of all such polls, showed Leave ahead. Since the start of June four, or 50% of the total, have done so.
It is often said that the bookies' odds give a better guide to the outcome of a vote but this is not always the case. In the run up to last year's General Election bookies consistently forecast that the Conservatives would win the largest number of seats, but they did not predict a Conservative majority. In April of 2015 it was possible to get odds on a Conservative majority of 7/1, an implied probability of 13%.
Economists have taken a close interest in the referendum debate and their views are worth noting. In its June issue of economic forecasts, Consensus Economics polled economists on the likely outcome of the EU referendum. UK economists on average assign a 37% probability to Brexit; economists covering the euro area out the probability at 42%.
The outcome will hinge critically on voter turnout. Some of the most enthusiastic Brexiteers, the over-65s, members of the Conservative Party and members of UKIP, are most likely to vote. Conversely, those aged 18-39, who tend to be in favour of Remain, have a relatively low propensity to vote. Getting the younger vote out, and ensuring a high turnout, is key to the Remain strategy.
At last year's General Election turnout was 66.2% - a better showing than the three previous General Elections, but well below the 73.3% that prevailed for General Elections between 1918 and 2010. Turnout for the Scottish independence referendum was high at 84.6%, but other referendums, local elections and mayoral elections tend to be lower. My own guess is that the length of the campaign and the prominence of the issue will generate a turnout in excess of last year's General Election.
Many will stay up into the small hours to follow the results. For me the absence of official exit polls and a trickle of results before 3am points to an early to bed early to rise strategy.