John Curtice: Scotland's experience of AV makes awkward reading for both camps

In discussing the merits and demerits of the Alternative Vote (AV), much has been made by both sides of the experience of the system in Australia, Fiji and Papua New Guinea. Yet it is always debatable how far the lessons in one country would apply elsewhere.

But there is no need to go halfway round the world to find out what would happen under AV in Britain, because the system is already in use here. Following the introduction in 2007 of the Single Transferable Vote (STV) – AV's close, though more proportional, cousin – in Scottish local elections, 32 local by-elections have been held using AV.

What happened in those by-elections raises questions about claims made by both sides in the referendum debate. One claim made by the No campaign is that the introduction of AV would require the use of electronic counting machines. Some have even suggested that, as a provider of electronic balloting services, the key backer of the Yes campaign, the Electoral Reform Society (ERS), might profit financially from its introduction – although the ERS says this is untrue.

Some of the 32 local by-elections in Scotland have been counted by machines already in use, but many have not. Faced with AV counts, many returning officers in Scotland have determined the use of counting machines is an unnecessary expense.

A second charge commonly made by the No camp is that, to quote Winston Churchill, AV gives undue weight to "the most worthless votes cast for the most worthless candidates". By this they mean that the winner is decided by the subsequent preferences of those whose first preference is a fringe or extremist candidate.

Yet Scotland's experience demonstrates that in practice, the candidate who secures most first-preferences usually ends up the winner. In only four out of 32 cases has that not been the case. Twice a Liberal Democrat has managed to leapfrog past a Conservative, while twice also, an Independent candidate has been enabled to overtake the SNP.

One reason why AV often fails to make a difference is that many voters eschew the opportunity to cast subsequent preferences. When counting machines have been used in Scotland, they have revealed that between only a half and two-thirds of voters cast a second preference, and only one in three casts a third.

That, however, raises a question for the Yes campaign which is keen to argue that AV ensures the winner secures 50 per cent of the vote. Not necessarily so. On 12 out of 32 occasions in the past four years, too few voters cast sufficient preferences to take anyone past the 50 per cent mark. Perhaps it is not so strange that there has been little discussion of Scotland's experience after all. For it makes awkward reading for both sides.

John Curtice is Professor of Politics at Strathclyde University