Tactical Voting: Some Problems

There was a lot of anger yesterday about an online tactical voting tool launched by the pro-remain group Best for Britain. The tool appeared, judging by the results of the 2017 general election, to be suggesting Labour voters tactically vote Lib Dem in what were Labour/Conservative marginals. If the objective is to maximise seats held by parties in favour of a second referendum, this would obviously be an unfortunate recommendation.

The problem is, deciding how to vote tactically is a trickier business than it may seem at first sight. As Tony Yates shows here, tactical voting requires some sense of fixed, or only marginally changing background of other voters who are not voting tactically. The easiest way of attempting to do this is by looking at previous election general election results. But party allegiance at the moment is unusually fluid, and this is not the only data which could plausibly used to get a sense of how things look in particular constituencies. Best for Britain appeared to have based their calculations not just on the 2017 general election result, but more recent opinion polls and the European Election results.

How much weight to give these different types of data is a difficult question, but the answer may well not be zero. It is also not entirely clear who the tactical voters in these elections are. If, for the sake of illustration, it turned out that Lib Dem voters were much more willing to tactically vote Labour than vice versa, it is possible to imagine a situation in which the correct strategy for maximising remain seats were to recommend a Labour vote in a constituency, even if current predictions showed a greater Lib Dem likely vote share in that constituency than Labour without tactical voting. 

As Phil Syrpis points out, these problems may only really apply to a small number of marginal seats. But in the event of a close election, this could cause trouble. Moreover, disputes of this nature breed distrust between supporters of the various remain parties. This is a problem, as tactical voting is unlikely to work if there is too much hostility. People need to trust their second choice party at least somewhat, and are most likely to think of tactical voting as a useful strategy is they think of themselves as part of something bigger, where others might be returning the favour. 

Unfortunately, disputes over how to model the current state of the electorate are to be expected. This is partly, as may have happened with Best for Britain, due to differing opinions about methodology. But in other instances, these will also be mediated by partisan interests. Party campaigners are very likely to select prediction methodologies which show a more favourable picture for their party. This may be down to psychological biases, or just the role they fulfil as party campaigners. But either way, it is likely to cause problems down the line if and when candidates put out contradictory campaign material about the state of the race in their constituency. 

In a better world, perhaps the various remain parties might have circumvented these difficulties via an electoral pact. But not only there have always been near insurmountable political barriers to such an endeavour, that could also have caused a counter measure between the leave parties (though it may well still be the case that the latter happens). None of this is to suggest that voters should not attempt to vote tactically. We are where we are, and we have to make the best of this situation. But the difficulties may turn out greater than anticipated. 

No comments:

Post a Comment