Who Should Decide?

In Chapter 7, Section I of “The Open Society and Its Enemies” Popper critically discusses Plato’s theory that the wise should rule, partly by attacking Plato’s implicit presumption that “Who should rule?” is a good question. Popper points out that it is difficult to find a government on whose goodness and wisdom we can rely and so that we should set up our institutions to allow us to prevent bad government from doing too much damage. Many people seem unaware of this argument and its implications for political and moral philosophy.

For example, in debates over voting systems many people hanker after proportionality or ‘fairness’. That is, political parties should be ranked in order of how many people prefer them. This seems like a “Who should rule?” idea and on those grounds we should already be a bit suspicious of it. But this demand should be completely and decisively rejected when we realise that it’s incompatible with getting rid of bad governments. Why? Let’s suppose that you cast a vote in which you rank candidate A above candidate B. You would hope that your vote will increase A’s representation rather than decrease it – let’s call this monotonicity. If this doesn’t happen then your vote literally counts for nothing. It’s not just that you’re only one person voting out of millions. Rather, if you vote then you might actually hurt your favoured candidate. Conversely if you want to keep B out of power, you might as well not vote at all since ranking B lower might give him more representation. It turns out that you can’t have both monotonicity and proportionality. The economist Kenneth Arrow showed that a voting system in which there isn’t a single person who decides the outcome of elections can’t have all of the following properties:

• Pareto – if all voters prefer A to B then A should be ranked above B.
• Binary independence – if two profiles agree on ranking of A and B then voting procedure should give same ranking of A and B for both profiles. (A profile is the set of preferences of all of the voters.)
• Mononicity – if you rank A above B this increases A’s representation or leaves it the same.

Proportional representation satisfies binary independence, as does alternative voting (you rank candidates in order of your preference for them) and so it doesn’t satisfy monotonicity. The first past the post (FPP) voting system doesn’t respect binary independence but it does respect monotonicity and so it allows people to vote bad governments out of power.

I also often see libertarians say that democracy is the rule of the people and so that it is no better than tyranny. This is badly misleading. The people don’t rule in Western liberal democracies. Politicians and government employees rule. Voters get a chance every few years to try to get rid of government employees who suck worse than their rivals by voting them out of office. This system is flawed in a variety of ways. When you select a politician you have to “buy” all of his policies: you can’t say “I’ll have A’s domestic policy and B’s foreign policy”. In addition, you only get to have the policies you want if lots of people agree with you. Government is funded by taxation – that is, threatening to use force against people who refuse to pay for policies with which they disagree. However, it’s better than having no ability to get rid of incompetent or horrible politicians. It also provides a relatively peaceful way for one government to succeed another: you don’t need a civil war to change governments.

I should also say that libertarians who decide not to vote because they interpret it as endorsing the current system despite its flaws are wrong, and their error is related to the “Who should rule?” mistake. Voting for a politician doesn’t imply consent. If there are two men A and B threatening to rob you and use your money to do things you dislike and they ask you to vote for one of them, if you vote for robber A that doesn’t imply you want to be robbed. Nor does it imply that you think robbery is good. Taxation and many other horrible policies will continue whether you vote or not. There’s no point in worrying about endorsing it because it’s going to happen anyway, and you can always say you’re voting in self defense against a bigger thug than the person for whom you are voting. Getting hung upon the fact that you are voting is just concentrating on the identity of the person making the decision. You should concentrate on whether voting in a particular way will be better or worse for what you value. (By the way, I’m in favour of anarchocapitalism – I think all of the activities the state currently undertakes should be stopped or should be funded by voluntary subscription. I’m not doing this to have a go at libertarianism.)

I also hear left wing people make bad political arguments. For example, about five to ten years ago I heard somebody say that Margaret Thatcher was a bad Prime Minister because she broke the power of trade unions in Britain and the workers should be in charge.

Furthermore, the “Who should rule?” fallacy shouldn’t be interpreted as being limited to politics. Everybody makes mistakes, so you can’t rely on anybody to make good decisions all the time. So when people make decisions in their personal lives they ought not to make decisions that hang on the identity of the person making the decision, but people do this a lot. “We did something you wanted to do last time, this time we should do something I want to do.” What’s the alternative? When two people have preferences that clash they should try to find a common preference: that is, a set of options they both prefer to their original position. The common preference need not involve people doing something together, all that has to happen is that they both agree that the common preference is better than what they wanted to do originally.

To summarise, concentrating on who is making a decision at the expense of the substance of the options available is bad whether you’re doing it in politics or your personal life.