ANOTHER ARGUMENT AGAINST AV

Bill Myers of Leicester has this in the current London Review of Books (scroll down from here):

Whatever Ross McKibbin may say, opponents of AV are not ‘cave dwellers’ (LRB, 18 November). AV maximises the votes of extremist candidates, since anyone voting for them knows their second preference votes will still count, while the second preference votes of the last candidate to be eliminated have no impact on the result, though as many as 40 per cent of the votes may be affected. In constituencies where the Labour and Lib Dem candidates are the leading contenders, for example, only the second preferences of Conservative, UKIP and BNP supporters will matter. It is possible, however, that if their own candidate is defeated, Labour voters would prefer to be represented by an ‘honest-to-God’ Tory than a ‘pragmatic’ Lib Dem. The second preference votes of the last candidate to be eliminated should take precedence over those of the least successful candidates. Under the standard counting procedure, AV is demonstrably less democratic than first past the post.

I'm not sure that redistributed second preferences of voters whose first preferences are for extremist candidates are any more problematic than the second preferences of any other voters – why should anyone's second preference be worth than anyone else's or indeed the same as anyone else's first preference? – but there is a valid point here. Andrew Rawnsley picked up on it here in his otherwise entertainingly knockabout but vacuous piece in the Observer on Sunday, in which he fails to recognise that quite a few of us reject AV not because we love the present electoral system but because it is not a proportional system or even a step on the way to one.