Natural Justice

Fri, 23 Jun 2017

I finished a book a couple of weeks ago that I suspect is going to be very significant in my thinking over the coming years: Binmore’s Natural Justice.  Kenneth Binmore is my kind of thinker (Keynes is another) - a mathematician turned economist, equally at home with philosophy and the real world (Binmore’s team designed the UK’s 3G spectrum auctions which raised £22bn).  

Natural Justice uses technical game theory to explore the way in which our need to coordinate within groups leads (through biological and cultural evolution) to our tendency towards justice and morality.  (Please don’t interpret this as suggesting that morality isn’t real or important in its own right.)  

At the heart of any coordination problem is the Nash equilibrium.  John Nash (portrayed by Russell Crowe in A Beautiful Mind) proposed this idea of any situation where no agent had any incentive to unilaterally change their behaviour.

The most famous example of a Nash equilibrium is the prisoner’s dilemma, but a more practical example is driving: if all drivers drive on the left of the road, it generally makes sense to drive on the left.

This example points to the insight that there may be multiple Nash equilibria, and not all equally good or stable.  Everyone driving on the right is also a stable equilibrium.  Everyone choosing randomly is also an equilibrium, although a clearly inferior one and unstable (any bias will tend to reinforce itself, leading to a new equilibrium).  Another insight from is that societies can be guided towards equilibria, for example, with traffic laws.

Binmore will appeal to many conservatives/economists when he claims that societies should not aim for coordination utopias that are not Nash equilibriums - these are unstable and unlikely to persist.  However, in a challenge to many conservatives/economists, he reminds readers that there is no good reason to assume that we are in nash equilibrium at any time, or that the equilibrium we are closest to is the only or best one.  Far from suggesting we should be laissez-faire, he endorses a big role for thinkers to help us discern the best Nash equilibria, and ways to achieve them.

This compromise appealed resonated with how I think.  While I recognise that a huge amount of good in the world is done out of altruism, I never want my society to be reliant on it, or to be able to be gamed.  And I’m far more interested in helping people see what is in their interest, than by trying to convince them to act against their interest.  

Many treatments of game theory ignore questions of enforcement; Natural Justice argues availability of enforcement is pivotal to which situations are equilibria.  For example, me working for payment on completion only makes sense if I believe the outcome can be enforced.  Trust or social norms may be a replacement for enforcement in the short term, but they are likely to be insufficient in the long run.  

The book ties this distinction into the argument between utilitarianism (that things should arranged to maximise overall utility) vs egalitarianism (that things should be arranged to improve the position of the person with least utility).  Binmore suggests that with the ability to enforce contracts, we will end up with utilitarianism, but without it we will end up with egalitarianism.

Another significant contribution of the book is its consideration of what economists call Interpersonal Comparison of Utility.  I clearly value outcomes for myself, though I don’t ignore the value of outcomes to others.  Some economists argue that we cannot objectively compare other people’s utility to our own.  The book argues, more pragmatically, not only that we do take others’ utility into account, but that this is a useful evolved mechanism.  I’m sure my understanding of this idea would benefit from a second reading of the book.

I’ve got a couple of other books on social contract and game theory to get through, so I am certain that my thinking will evolve, but I felt this introduction was a good start.