Free votes and referendums

This is inspired by Gowers’ blog entry ‘If politicians were mathematicians’, although it gets away from mathematics almost immediately!… Gowers raises the following problem in democratic politics. Say a party wins a general election, with 55% of the popular vote, and proposes a controversial policy, which only 70% of its MPs support. If party discipline is sufficiently strong, all the MP’s from that party may end up voting for the measure, and it may be passed, when in fact under 40% of MPs support it. Gowers puts forward the following idea, which might help overcome this problem; I take the liberty of quoting verbatim…

`Votes are made electronically and then counted. After they are counted, the way people voted is made public. However, before that happens, each vote is changed, independently, with a certain probability such as 10% (but the precise value could be argued about, and might even vary from vote to vote, being lower for especially important votes). If you feel strongly that your party is wrong on a certain issue, then you can vote against it, and if that annoys the party whips, you can tell them that you voted for it but your vote was flipped. However, you cannot play this game too much, or the number of times your vote appears to be against the party line will be so far above 10% that it will be clear that you are not a loyal party member.’

What I have been thinking is, maybe there are other solutions to this problem as well. First of all, a government will not want to propose too many policies which are unpopular with a majority of voters, if only because this would endanger its chances of re-election. (Naturally, one would hope that they also wish to represent the best interests of the voters!) Neither will it want to alienate a majority of MPs, as if it does this repeatedly it may even risk been brought down by a no-confidence vote. Governments will often allow MPs a free vote on contentious issues: in the UK, this tends to be popular with constituents, as it (hopefully) gives MPs more freedom to vote as a majority of their constituents would wish. But the problem is: what happens when an MP finds themselves at variance with their party in a whipped vote? Ideally they should be able to vote against the party line without severe consequences. One would hope that a party that issues a two or three line whip on every vote would alienate its MP’s and also become gradually less popular with the electorate due to perceived authoritarianism, but in reality this may not happen. Gowers’ solution does indeed provide a way of occasionally escaping the penalties for breaking party lines. But if an MP chooses this way of escape, they are unable to participate openly in the debate on this issue, which is a sad loss.

An alternative solution would be to lobby governments to relax party discipline in more instances. Before a general election, parties could be asked to publicize their policy on whipped votes. (Certain rules could even be enforced by an independent parliamentary commission.) For example, parties could promise (or be forced!) to allow a free vote if enough MPs signed a petition to allow one. This blog advocates a free vote on every single issue. This would not necessarily make government impossible – the government would just have to spend more time building consensus within its own party (and maybe even across parties).

All this begs a much larger question: what about direct democracy? Why should we delegate so much decision-making power to MPs? Wouldn’t it be possible to have a nationwide referendum instead of a parliamentary vote, on many issues? A friend of mine has proposed a system where everyone has the option to vote on every new measure, but can choose to delegate their vote to an elected representative if they wish (for example if they were apathetic, or felt they were uninformed about the issue).

I’m not sure that the majority of British voters would support a direct voting system. I’ll try to outline some (reasonably well-known) potential problems with it below.

For one thing, there are theoretical problems with allowing a referendum on everything separately. Suppose there are 3 people, A, B and C, voting on 3 economic decisions, decisions 1, 2 and 3. A ‘yes’ to decision 1 will cost A £1000 pounds, but it will earn B and C £100 each; a ‘no’ means that nobody loses or gains. Similarly, a ‘yes’ to decision 2 will cost B £1000 and earn the other two £100, and a ‘yes’ to decision 3 will cost C £1000 and earn the other two £100; a `no’ means that nobody loses or gains. If each person votes on each decision separately to maximize their earnings, two out of three of them will vote ‘yes’ on every decision, so each will end up making a net loss of £800. Presumably, if none of them communicate with the others before or after their vote, each will vote ‘yes’ to every decision, if they are guided only by self-interest. What happens if beforehand, there is a general election? There are two parties, one which promises to decide `no’ every time, and one which promises to hold a referendum on each decision separately. All three of A, B and C should be in favour of the first party, as it would reduce their losses to zero. So it looks like it could be in the best interests of everyone to appoint a ‘government’ which will carry through a unified policy.

The same kind of thing happens in the well-known ‘prisoner’s dilemma’. Here, the situation is even simpler. There are two prisoners locked in different cells, accused of being partners in crime. The police know that they can convict both on a minor crime, for which the maximum prison sentence is 2 years, but to get one of them on the big crime, they need the co-operation of the other. A policeman visits each of them separately and makes them an offer. ‘If you refuse to co-operate, we’ll give you 2 years for the minor crime. But if you implicate the other guy, we’ll just give you 1 year, and he’ll get 10 years.’ The prisoner, who isn’t a fool, asks the policeman what happens if they both implicate each other. ‘In that case’, the policeman replies, ‘you both get 5 years.’ Whatever his partner does, it’s better for each prisoner to rat: if his partner does too, his sentence has gone down from 10 years to 5; if his partner doesn’t, his sentence has gone down from 2 years to 1. Without the possibility of influencing the other prisoner’s decision, in the absence of the proverbial ‘honour among thieves’, both prisoners are likely to implicate each other and get 5 years, whereas if they both refused to co-operate, they would both get only 2 years. But what happens if the two prisoners had another option, to appoint a lawyer who would prevent the police from making either prisoner any kind of offer? They’d be better off going for the lawyer.

Even allowing each person a certain amount of ‘credit’ which they can use at each vote, as discussed in Gowers’ blog entry above, won’t prevent the above situation from happening, where a pre-agreed, fixed strategy is better for every single person involved than when everyone votes independently for selfish reasons. It is true that many people will not vote for a policy which they know will cause overall harm, even if it benefits them slightly, but there are probably enough people motivated mainly by self-interest to make it dangerous to rely on altruism.

What happens, though, if the ‘prisoner’s dilemma’ above is repeated again and again, this time with financial penalties rather than jail-time? This would be more similar to what happens when different measures are voted on. One would expect two prisoners to learn rather quickly that the best thing for them to do is to consistently refuse to co-operate, establishing trust and a stable pattern, and indeed, this tends to happen in sociological experiments.

One could argue that over time, people would learn that voting for the bests interests of the country as a whole will build trust and lead to reciprocal benefits, but again, there is a certain amount of risk in relying on this. Delegating some decision-making power to MPs, provided it is accompanied by genuine debate involving the whole cross-section of society, seems to be a way of getting around this risk.

But what if a majority of the electorate becomes convinced that a particular policy proposed by the government is detrimental to the country as a whole? A compromise solution might be to oblige a government to hold a referendum on an issue provided enough people signed a petition for it. The government would then have the opportunity to put its case to the electorate before the referendum. It’s interesting to speculate on what would have happened if there had been referendums on capital punishment in the UK. (Only in 2006 did popular support for capital punishment as a penalty for adult murder drop below 50%, for the first time.) Would the majority of those who backed capital punishment in opinion polls feel strongly enough about it, and be sufficiently convinced of their position, to press for a referendum on it? Maybe if there had been several referendums on this issue, there would have been a real chance for an informed debate, and opinion would have shifted much more quickly?…

Do we have the right balance between delegating authority (and consequently losing some influence over the decision-making process) and being able to intervene in the process? Maybe we’ve delegated too much… I’ll return to this at some point!


One Response to “Free votes and referendums”

  1. firefighter wedding Says:

    Hi there, just became alert to your blog through Google, and found that it is truly informative.

    I’m gonna watch out for brussels. I will appreciate if you continue this in future. Lots of people will be benefited from your writing. Cheers!

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: