Just the One: Time Inconsistency and the Greek Bailout(s)

As EU decision-makers grapple with their response to an imminent Greek debt default or bailout, they need to consider not only their current decisions but also their likely future decisions. It is critically important that they not deceive themselves into thinking that they (or the Greek government) can commit to making all their future decisions now. There are strong grounds for positing time-inconsistency in EU and Greek government decision-making concerning the Greek bailout. This is a simple point, but critically important to good policy planning in this situation. Acknowledging time-inconsistency does not proscribe any particular policy choice, but it encourages policy makers to act cautiously.

There are two distinct sources of time-inconsistency in this environment: one is game-theoretic and the other is decision-theoretic.  The game theoretic source arises from the precommitment problem, also called the problem of non-credible threats. Often in an intertemporal game a player would benefit if he could credibly commit to a particular future move, even though that move will not be optimal at the later date.  So for example a monopolist will benefit now if it can credibly commit to lowering prices later after a new entrant chooses to pay a set-up cost to enter the market.  If this were a credible commitment by the monopolist it would dissuade any entrants. However new entrants know that the precommitment is not credible – after the entrant has paid the cost to enter the market, the monopolist is better off keeping prices relatively high and sharing duopoly profits.  The monopolist might claim that it will lower prices against new entrants — “we will never be undersold!” — even though this is not an incentive-compatible claim. Entrants may ignore these incentive-incompatible claims depending upon the nature of the game. Game theorists call this type of publicly-communicated threat about a planned strategy “cheap talk.”   Cheap talk may affect the equilibrium of a dynamic game, or it may not.

The EU made a solemn promise never to bail out any member government, in order to prevent any member government from over-spending in anticipation of an EU bailout.  In game-theory terminology, this was “cheap talk” by the EU – the term sounds flippant but is not meant that way.  This solemn promise will soon be broken in the case of Greece.  Now, the EU would like to commit to bailing out the Greek government only once, thereby forcing Greece painfully to restructure its public finances quickly in the face of the severe consequences for Greece of an all-out default at a later date. The EU is likely to make a solemn promise along these lines after this first bailout. In the eventuality that Greece does not painfully restructure and faces all-out default at a later date, the EU will then want to bail it out a second time, but will try to commit to not bail it out a third time, etc.  I predict that there will be either two or three rounds of this game played over the next few years (counting the first bailout as round one).  Meanwhile, the Greek government will be playing a somewhat similar game strategy of cheap-talk-without-credible-threats against its own claimants (public sector unions, pensioners).   

There is a second, equally-important, source of time-inconsistency in this environment.  The Greek government, and perhaps the EU as well, cannot be modelled as decision-makers with intertemporally consistent preferences obeying the Bellman principle of optimality.[1]  The Bellman principal requires that if we ask an intertemporal decision-maker a time t about his preferred choices at time t+k, he gives us the same answer now as he would give at time t+k if he had the same information. Anyone (myself included) who has made plans to visit the pub on Friday night and have “just the one” and then wakes up Saturday morning, having had many more than one, has intertemporally-inconsistent preferences.[2] In 2010, the Greek government may sincerely want to force through painfully tough budgets for 2011 – 2012, but this does not mean it will want to do that in 2012.  At that point, it may sincerely want to force through tough budgets for 2013 – 2014 but not 2012.  This kind of intertemporal inconsistency is a feature of human nature, as well as of social and political institutions.