By Electoral Reform Society:Turnout is a vital measure for understanding our relationship with politics, whether it be low or high. If  people do not vote, it tells us they are being failed by politics.  If people do vote in large numbers, then it can be argued the Government which emerges after an election has a much stronger mandate to run the country as it sees fit.

Given its importance, a new paper from a team of academics at the Universities of Oxford and Manchester which challenge many of our assumptions on turnout, is deserving of significant attention.

The issue lies in that turnout – the number of people that vote in any given election – is not actually that easy to measure. In the UK it is given as a percentage of the electoral register, so if 44m people are registered and 22m ballots are cast, the turnout figure will be given as 50%.

Counting the ballots on election day is relatively easy, but working out how many people are actually on the electoral register is trickier than it sounds.

This is because there are perfectly legal duplicates on the register. For instance, second-home owners and university students may register at multiple addresses though they may only vote in one place per election. If you split your time between two homes you could vote in two sets of local elections, but only once in a general election.

A further reason is redundancies and inaccuracies – voters who move house and register at their new address but aren’t removed from their old one.

These can build up as the UK does not a have a centralised electoral roll. Each local authority holds their local electoral roll and manages their own elections. This makes it very hard to centrally rig an election in the UK, but also leads to people being on multiple electoral rolls.

To work out the size of the roll you have to apply some complicated maths to estimate how many errors there are and remove the duplicates. You then use this number as the official size of the electoral roll and then work out the turnout.

This is where the problems started. 

The paper authors estimate that, on average, UK turnout is underestimated by around 9.4%, on average, which would boost UK turnout north of 70%.

This is both positive and dramatic. It is often said that people are not interested in politics, and this argument is grounded in turnout figures which may have been hugely underestimated.

Such inaccuracy also throws into question the way the register works and voters access to it in general. Whilst you can register online there is no way to check if you are registered already. Voters end up inadvertently registering multiple times and can’t easily correct this themselves. Some of the inaccuracies in the register will also work against voters being able to vote at all – if, for instance, they are registered in the wrong place.

Furthermore, the electoral register is used for the drawing of electoral boundaries. As duplicates are often produced by people regularly moving house, and this regular movement is usually concentrated in urban areas. Constituencies where duplicates and errors are concentrated end up with fewer voters. It would be far simpler to use population and nationality data from the census to design constituencies with equal amounts of eligible voters. To ensure citizens are fairly represented in Parliament, ensuring boundaries are based on sound data is vital.

As has been identified, where there are mistakes in electoral data, it can have very real consequences for democracy. Findings such as these should cause us to pause for thought and wonder how we can have the fullest possible register.