The current stand off in talks to renegotiate the Greek debt bailout shows the difficulty of achieving an outcome that could be good for both sides.
Why lenders sometimes rationally forgive part of a loan
There is plenty of excellent commentary on the Greek debt situation, arising from a newly elected Greek government which has a popular mandate to relieve the terms of the bailout and of the onerous budget conditions that came with it. I want to note something of wider interest and applicability in this sort of complex negotiation. It is often the case that a borrower is unable fully to repay a debt. It may be possible to repay some of the debt, in which case a voluntary renegotiation can make sense for both sides. The creditor would prefer full repayment but would prefer partial repayment over none at all, which could arise if the debtor goes bust or, as in the case of a sovereign, collapses into economic chaos.
It is quite often the case that a creditor will accept a reduction in the terms of a loan. Suppose a bank has lent €1m to a company which at the time was fully able to pay, on the basis of the shared expectations of the future. Then something happens – a recession, an unpredicted change in the industry, a natural disaster – that was beyond the control of the company and was not foreseen by it or by the bank at the time of the loan. The company cannot now repay the full loan because it doesn’t have the financial resources. We assume that there is no prospect of raising new equity capital nor of selling any assets. The firm can only pay if it stays in business but with a lower level of debt. What should the bank do?
If the bank agrees that the company’s troubles stem from outside forces beyond its control then it would be foolish to insist the company pays without any change to the loan conditions. If the company goes bust without realisable assets then the expected value of the loan repayment is zero. But there is some reduced value of the loan, after writing it down or reducing the interest rate, both of which cut the present value of the debt, which would be affordable by the company. It may be difficult to get agreement on that exact amount because the company would prefer an even lower debt. But if the bank has good access to the company’s operating information and there is common information on the industry or other external factors causing the problem, then it should be possible to reach some sort of agreement which is better for both the bank and the company. Perhaps the deal reached is €0.7m. That is 30% less than the original loan but a lot more than zero.
Note the conditions under which a mutually valuable deal can be done: i) it’s a one-off situation; ii) the cause is agreed to be external; iii) the bank has good access to information about the company; iv) there is no advantage to the bank in forcing a liquidation of the company.
Banks often do restructure loans in this way, since it is obviously in their interest to do so. It is widely regarded as less likely to happen in the world of centralised bank credit decisions than in the days when local bank managers had both a closer relationship with corporate borrowers and better information on them. I’m not sure if this is true but it’s plausible. One claim made by a former MFin speaker on Islamic finance is that Islamic or, as he prefers to call it, ethical finance means being reasonable and fair to borrowers when they are in difficulty in a way that combines justice with better overall debt repayments. Again, I don’t know if there is evidence for this but it’s an intriguing argument.
The problem of asymmetric information
Now imagine a situation where a bank has a loan to a company that has previously asked for, and received, a write down in the loan. So it’s no longer a one-off situation. The bank recalls the past restructuring and now has a different perspective. There is always incomplete information in any such situation and moreover it is asymmetric, a pervasive problem in economics. The company will always know more than the bank about what is really going on in the business. A second problem may start to look less like misfortune and more like carelessness.
On top of that the bank must consider the effect of any leniency on repayment on the incentives of this company in future and on other such companies. So long as there is incomplete and asymmetric information, companies thinking entirely rationally may judge that it is best for them to pretend that things are worse than they really are, in order to get a loan write down. The bank may in turn judge that it had better be more harsh than is strictly warranted by the individual company case to discourage other companies from playing a game of pretend-we-can’t-pay.
The more asymmetric the information, the more likely the bank departs from what would be the best actions for maximum repayment in order to protect itself against future opportunism by borrowers. It trades off short term losses (relative to what it might have been able to get back from the optimal renegotiation) against the long term benefits of being seen not to be a pushover.
More generally we are in the land of game theory – the theory of interdependent decision making. Life is relatively straightforward in a one-off game such as the famous prisoners’ dilemma (see appendix). But it gets more complicated when games are repeated. A great deal of social life is about repeated interactions with people. We each act, observe and learn. In particular we learn about people’s behaviour, which earns those people a reputation. A reputation is a signal to other people about how you are likely to act in future when there is uncertainty. A reputation for honesty, even when it would incur short term costs, is very valuable because other people will trust you. A reputation for claiming debts, or punishing failure to repay, at all costs is a valuable one for a gangster. It may be worth investing in building a reputation by taking actions that are costly in the short term. Only costly decisions will tell other people you are serious. If you do what is in your short term interest then people will assume you will always do that. But if you stick to a rule, whether for honesty or ruthlessness, even when it’s costly, then eventually you will have the advantage that in any negotiation other people will assume you’ll always stick to that rule, without you having to take costly actions to keep proving it.
Greece: a game that has already been repeated
Let’s return to the Greek debt problem. If Greece was for the first time unable to repay its debt and everyone agreed this was a result of external bad luck, then a deal would be done, because the creditors (the IMF and other EU countries) would prefer to get some fraction of their debt back than none. It is fairly widely agreed that Greece’s total sovereign debt is, at about 176% of GDP, beyond the point where a country could normally be expected to repay it. (Bear in mind that Japan’s government debt is, at around 237% of GDP, far higher but most of that debt is owed to Japanese residents and institutions, not to foreigners.) So a deal seems sensible for both sides.
But this is not, alas, a one off situation. Greece’s governments have a record of distorting their public finance figures, of failing to collect tax and of failing to implement agreements on economic restructuring. Greece has defaulted on previous foreign debts at least five times (1826, 1843, 1860, 1894 and 1932). There are other Eurozone countries that have also received bailouts with conditions attached and if Greece gets special treatment they will want the same. Even if the current Greek government is entirely sincere and the people who elected it are committed to future changes in policy (like radically improving tax collection) that would benefit both debtor and creditors, it’s very hard to reach agreement because of a lack of trust.
It is also entirely possible that some of the creditor countries (not only or particularly Germany) want to push Greece to the limit in order to reinforce the discipline on other bailout countries. I suspect there are people in the Netherlands and Finland who want to see Greece “choose” to exit the Euro so that the country is finally out of the system for good and can be left to manage itself however it can (perhaps selling itself to Russia, assuming Russia’s stricken economy can afford to help). While a Greek exit from the Eurozone would bring serious risks of contagion and might threaten the entire EU, the risk might be worth it, if it leaves a more cohesive Eurozone that is more likely to survive in the longer term.
It is also not clear that the cause of the problem is external. On the contrary, many people see the Greeks as having brought their debt misfortune on themselves. Others see the mechanism of the Eurozone as having at the very least inflamed the situation and made it difficult for the Greeks to escape their debt trap, whatever their guilt in originally causing it. But in northern Europe the story has become something of a morality tale in which Greece must face the consequences of their earlier actions, or more precisely the actions of a corrupt and self-serving upper class that dominated previous governments.
Unlike corporate bankruptcies, sovereigns to some extent choose not to repay their loans. There is no equivalent bankruptcy process for a nation, only a sketchy international legal framework and plenty of complicated and varied precedents for both creditors and debtors to choose from. Each sovereign debt negotiation is therefore unique and complex. This is a risk that needs to be taken into account when investors are considering buying government bonds.
Appendix: the prisoners’ dilemma
In the prisoners’ dilemma, a police chief holds in separate cells two prisoners who have jointly committed a crime. The chief has only enough evidence to convict them on a minor charge but wants to get one or both of them to confess to the more serious crime they actually committed. The chief offers each prisoner the same deal: if you confess you get off on a light charge; but if you don’t confess and the other prisoner does then you get a very long sentence. Given the incentives it is rational for each prisoner to confess because he (most prisoners are men) would infer that the the other prisoner would rationally confess in the same situation. So they both confess and both receive long sentences, compared with the much lighter one they would receive if they kept quiet.
In a one off game the logic is inescapable, you are rational to confess. (The game excludes the possibility of credibly promising in advance to not confess, which is of course the basis of omerta or secrecy among mafia members. But to the extent that this credibility is enforced by the threat of extreme punishment for confessing then that changes the payoffs of the game.) But in a famous piece of research done in the early 1980s by Robert Axelrod, it was shown that in a repeated game cooperation could emerge as a rational and consistent strategy. The research involved a tournament between various decision rules that were executed by a computer simulation of a repeated prisoners’ dilemma game. The winner was a simple “tit-for-tat” rule – do to the other prisoner what he did to you last time. Over repeated iterations this rule evolved into a consistent no-confess stream of outcomes.
This example is one in which repetition of the game leads to a better outcome (for the participants, not for the police chief). But in the debt forgiveness case and in many other examples, it is easier to get a mutually beneficial outcome in a single game. Life is, for most people, more like a repeated game most of the time.