Following the recent referendum on switching voting systems, we turn our attention this week to the Alternative Vote (AV). Having just been bombarded by arguments as to the pros and cons of voting systems, Figure It Out wonders about how people approach voting and how tactical voting can produce surprising consequences.

The media has been full of scare stories about how complicated voting could be, and how it undermines this or that fundamental principle (“one person one vote”, or “reducing unfairness in voting” to name two). There is no doubt that tactical voting takes place in the current system, especially in marginal seats, and is used as a prime reason for its deficiencies. By tactical voting we mean voting for a candidate to achieve an overall outcome, which would be different from voting for your preferred candidate. This tactic is often used to prevent an undesirable (in your eyes) candidate from being elected.

But what about tactical voting under the AV system, where second preference votes count? There is no doubt that people behave in different ways depending on the system they have, and will inevitably seek to exploit that system for their own preferred outcome.  There are clearly plenty of psychological drivers in these situations but how do you apply an analytical approach to such complex decision-making? The reason why it is complex is that the outcome depends on what many people do or intend to do when they vote, not just you.

This is a typical real-life problem:  where there are several players involved and the decisions they make depend, to some degree, on other players’ actions or decisions.   If we think about this in a business context, then the key players are often competitors in a market space, so strategies about product marketing, pricing, take-overs, and new market entry dynamics would need to adopt this more sophisticated strategic decision-making.

So how do you  tackle this type of problem? We turn to a technique called game theory to tackle this more analytically. Game theory is the formal study of decision-making where several players must make choices that potentially affect the interests of the other players.

Tactical voting is something we come across all the time, especially with the rise in popularity of TV talent shows. A classic example is the Eurovision song content (see previous Figure It Out entry).  We also know of infamous cases of the public voting for acts in talent shows, such as X-Factor or Strictly Come Dancing, that are clearly of poorer ability than others and so force better acts to be eliminated.  The rise of social networking means that there is now a way for groups of fans/voters to adopt a coordinated strategy and so sway the outcome, whether through well-organised fan groups, or a general mood to stick two fingers up to the organisers.

So what about unintended consequences in elections and in AV in particular, resulting from tactical voting?

Game theory has studied the dynamics of voting systems over many years (see summary article). One approach to tease out unexpected behaviour is to think about the different decision makers and understand what drives them and so predict their behaviour.

–          Mainstream Political parties: Likely to be the top 3 positions in voter popularity in almost every UK constituency election – the order of the 2nd and 3rd placed candidates in final run-off is crucial in deciding the outcome

–          Minority parties: Unlikely to be elected but have power to influence if they can persuade their loyal voters to adopt a strategy for their second preference, especially if they come a credible 4th

–          Voters: They are extremely non-uniform group, but influenced by many factors. Trying to predict their behaviour is where the fun starts

a)      Closely contested AV ballot can be determined by the order of 2nd and 3rd candidates

Imagine a scenario where the Conservatives are the favourite but Liberal Democrat and Labour candidates are closely matched. Because the Labour and Liberal Democrat antipathy to the Conservatives is very different, the identity of the 3rd placed party is really critical. When they are eliminated their second place votes will be very important to the outcome. Labour voters are unlikely to support a Tory candidate but some Liberal Democrats will. So what would a tactical Conservative supporter do? Stop Labour coming 3rd of course. So how do they do that? Well if they get a few loyal supporters to vote Labour then Liberal Democrats get eliminated in 3rd place, their votes are transferred and Conservatives stay on top – obvious really! (See article for a fuller analysis)

b)      A credible 4th placed party changes the result compared to just a 3 party playoff

This is a build on the previous example where the number of votes available from 4th place plays a critical role. Say the scenario includes a credible showing from the Green party. How they put their 2nd preferences affects the order of the top 3 and so affects the scenario above. So if Greens promote Labour as a 2nd preference they may inadvertently be handing the result to the Conservatives – confused yet?

c)       The popularity of the coalition at election time influences voting strategy

This analysis considers some different voting strategies depending on whether or not the Coalition is popular and so the Liberal Democrat 2nd preference vote changes from strongly pro-Conservative or pro-Labour. If the coalition is popular it might suit Conservatives to put Liberal Democrats 2nd and vice versa, then whoever gets eliminated the other stands a good chance of winning;  a spoiler tactic just as in the current system – so who says tactical voting disappears under AV!

You might be interested in a BBC blog  where the writer has created a scenario whereby depending on which of 4 voting systems is used, a different result is obtained for each one!

OR has experience of developing analytical approaches in business problems such as these such as pricing strategies for pharmaceuticals, or business tactics for new market entrants, and where competing strategies need to be taken account of. One of our successful approaches is in the area of business games, where individuals or teams work within a controlled business environment created by our OR team to explore and understand the impact of competitive decision-making. We have used this approach to deliver projects to help new entrants to a mobile telephony market, and helped utility operators understanding pricing strategies within new operating frameworks.