If we get fairer votes, we need some better stats too

If AV is passed by referendum in a month’s time – which I hope it is – voters will be able to give much more information about their real preferences than they currently can.  That is, of course, the point: by giving second (and third, and so on) preferences, voters will no longer have to choose between heart and head.  They can choose both, all on one ballot paper.

Against that, ‘No’ campaigners pit the fear of more coalition governments.  I don’t accept the premise that these are a bad thing; better a compromise government, reflecting a plural electorate, than winner takes all on 35% of the vote.  But that’s an argument for another post; AV is perfectly capable of delivering majority governments, and in Australia it generally does.  In a British context, majority governments would be more legitimate, too: by allowing the expression of subsequent preferences, most governments would be shown to have a base of (partial) support extending well beyond their total share of first preferences.  The fact that most Liberal Democrat (and SNP and Plaid Cymru) voters preferred a Labour government to a Conservative one in 2005 would have been made explicit and measurable in second and third preferences: and Labour’s mandate would have looked much stronger as a result.

But even where AV produces hung parliaments, it will produce hung parliaments where the will of the electorate is made a little clearer. Looking at the spread of their voters’ preferences, the party or parties which held the balance would have a much clearer idea of which partner their voters would prefer; and if the second-placed party had a much wider support base among the electorate as a whole than the first-, we’d be much more likely to know about it. So an AV hung parliament would be more democratic than the FPTP version: it would be that bit clearer if a party defied the popular will and chose the ‘wrong’ coalition partner, and that bit easier to punish them later.

The Australian election of 2010 provided a good example of how AV can help clarify an unclear result.  A strong Green vote cut the Australian Labor Party’s voteshare down to 37%, well behind the Liberal/National Coalition’s 44%. But when the full preferences were tallied and allocated to one of the two major groupings, the ALP had 50.12% of the vote: a very small lead, but clear. When given a choice between the two options, the Australians’ voting system allowed the politicians to see which option they preferred. FPTP would have given no such clarity.

The slight snag is that, under British counting procedures, we might well not be given the information which AV creates and makes available. The Australians record the full preference data for their elections, allowing much more information to be gleaned from them – even if you vote for a leading party in your constituency and your other preferences never need to be used, they’re still kept on record. In Britain, by contrast, we tend to perform the minimum checks on our ballot papers compatible with making sure we get the right winner – which means that a lot of that preference data would probably be lost in our version of an AV count.  We wouldn’t necessarily get an equivalent of two-party-preferred voting data (in 2010 this would presumably involve a Labour-Conservative, Labour-Liberal Democrat and Liberal Democrat-Conservative analysis) here.

Technocratic though it sounds, the time could come when that information matters.  It matters because, when no one party has won power and politicians therefore have to strike bargains, this information leaves them with fewer places to hide.  A party whose voters split 70% in one direction but then went the other way in its choice of coalition partner would live to regret it: and they’d have been warned in advance, too.  But equally, if the people expressed a clear preference for one leading party or another overall, they’d still stand warned: and if they ignored it, again, they would be likely to pay for it.  The democratic basis of a coalition would be much better spelled out for all parties: and that would help to make sure that coalition did reflect the broad wishes of the electorate.

So hung parliaments and coalitions shouldn’t necessarily be feared; AV doesn’t necessarily create more hung parliaments; and when it does produce hung parliaments, it provides more information than FPTP about the outcome voters would prefer.  But if we want to have that information to hand, we need to make sure it’s recorded.  So if there’s a ‘Yes’ vote, a polite word in the ear of the Electoral Commission might be no bad thing …