A Stochastically Optimised Senate

Why do we end up with so many politicians that we dislike implementing policies that we disagree with?

I would argue that this is a natural outcome of the fact that we vote for politicians. Voting implies having election campaigns and these require money. An aspiring politician has to align themselves with a party to gain access to this funding. So, in order to get elected, you need to be preselected. To be preselected requires years of ingratiating oneself within a party and following the party line as an undying loyal servant. Thus our parliaments are filled with people who have been carefully preselected for their complete willingness to let the party think for them. No wonder the media is full of opinion pieces bemoaning the lack of political leadership. Indeed, no wonder parties often have so much trouble finding someone from within their ranks to be their leader.

You think this is harsh? Just watch Mr Shorten agreeing, absolutely, with his (then) party leader even though he did not know what she had said. Another, more recent, demonstration of our politicians' inability to think for themselves was illustrated when Liberal senators voted for the motion "It's OK to be white" by 'mistake'.

If our parliaments are filled with people who let their parties do the thinking for them, what then do the parties 'think'? Well, parties have to 'think' in a way that will not hinder their party's funding base otherwise they cannot fund their election campaigns. Corporate influence on government policies is now accepted as inevitable. Property developers, the mining, financial, gambling, and food industries wield extraordinary power over us. We now know that it only takes $22 million to get rid of an Australian prime minister.

This state of affairs is a natural and inevitable outcome of the fact that we vote for our politicians. The behaviour of our politicians and the parties is not the problem, this is just the symptom. Voting is the problem, this is what has to change.

I believe the best alternative to voting is sortition, that is, our representatives should be selected by lot. Since its inception by the ancient Greeks sortition has been used by many societies with success (http://en.wikipedia.org/wiki/Sortition). Today we use sortition to select our juries. More recently John Burnheim champions these ideas in his book "Is Democracy Possible? The alternative to electoral politics" (Sydney University Press)

Applying sortition in the lower house would be difficult, and is perhaps a social experiment we would rather not try (how would you obtain a treasurer and construct the budget?). However, sortition would be ideal for the senate. It is the house of review, it has to act as a jury to pass judgment on the legislation produced by the lower house. A group of people selected from Australian society by sortition so that vested interests are excluded from the review process is just what we want.

My vision is as follows: In each state people could nominate to sit in a pool from which they may be randomly selected to serve in the senate. Unlike jury duty nomination to sit in the pool would need to be optional due to the length of service that is required. Thus, the pool may not be as representative as we might wish, however as long as we use random selection to remove the influence of vested interests much of what we desire can be achieved.

At this point I want to add an enhancement to the concept of sortition, a vote of retention.

When the senate term ends we will hold a vote. The ballot paper will just have the names of the existing state senators choosing to stand for retention. Against each name voters can write "Yes" or "No". If the yes vote dominates the member is retained to serve another term. Otherwise the standing member is rejected and a new representative is selected at random from the pool.

Thus, this is not a vote to elect people, rather it is a vote whether to retain people who have previously been chosen by lot. With time the senate will progressively accumulate competent people that we have chosen to retain.

Hopefully the atmosphere in the senate will be more collegial than adversarial and, with an election campaign no longer needed, a greater variety of people will be attracted to serve. We will have a senate with a better gender balance, indigenous representation, and a wider variety of community skills, experiences and ages.

Such a senate will moderate and improve the behaviour of the lower house. Leglislation will have to be framed for the acceptance by a senate that now represents a much broader, more represenatiive, section of society. The government may find this frustrating at times but an advantage for the lower house is that any legislation that is approved by the senate will now have much greater authority. For example, imagine that the senate approves legislation for a mining tax. Rather than campaigning against the governing political party the companies that might want to oppose this now have to campaign against a decision approved by a group of people representing a cross section of society. They will, in effect, be campaigning against the whole of Australian society. This would put things in rather a different light.

The title of this essay? Those with a training in computer science will see this method of obtaining our representatives in government as being a stochastic optimization algorithm. Solutions to the problem (of finding representatives in government) are generated at random. Better performing solutions are retained and poor solutions rejected to be replaced by new candidate solutions generated at random.

How can we make this happen? I have no idea. The main point of the proposed system is to reduce the influence of vested interests and to change the behaviour of parties. Thus, every vested interest and every party will work to ensure that this can never happen.

Peter Kovesi
April 2014