Following Europe’s plan of the week to deal with its sovereign debt crisis (and directly associated banking crisis) can make one’s head spin. It might therefore help to try to identify a few central concepts that can help put the details of this plan or that into perspective.
A key distinction is between measures intended to prevent things from getting worse, and those intended to allocate losses already incurred; the interaction between those issues is an important factor in the dynamics, and will play a huge role in determining the ultimate outcome.
In terms of preventing things from getting worse, the near-term danger Europe faces is runs on sovereigns, especially Italy. The mechanism for a run on sovereign debt is that lenders will refuse to rollover their holdings as they mature, making it impossible for a government to finance its operation; a similar mechanism is at play for bank wholesale lending, where banks with exposure to dodgy sovereign debt are facing problems rolling over their maturing funding.
A run is one of several equilibria in a coordination game: coordination games typically have multiple equilibria. I have no incentive to run if no one else runs, but if everybody else runs I have an incentive to run. In this particular instance, I will rollover my maturing Italian bonds if everybody else but I won’t if nobody else does.
Expectations play a crucial role in determining which equilibrium prevails: whether I run or not depends on what I expect others to do. Thus, crisis management plans often focus on influencing expectations. That is why the term “confidence” is bruited about so freely: you want everybody to be confident that nothing bad happens so they won’t run.
It is also why backstop funding mechanisms are typically a part of managing these types of crises. The more public money available to fund the renewal of maturing sovereign debts, the bigger the flight of private investors required to spark a full-blown run. These private investors calculate that the likelihood of a big flight is smaller than the likelihood of a small flight, so they are more confident that a run will not occur when there is a big public backstop, so they will not run themselves. This means that the backstop ideally will not have to be drawn on, because private investors continue to roll their lending to troubled countries. This is basically the Paulson “bazooka” idea.
Or another analogy. Someone with a weapon is confronted by a crowd of 100 people. If all 100 rush him, he’s had it. But someone has to start the rush, and the first mover is at greatest risk of getting shot, so each person may hold back wanting someone else to take the lead, allowing the outnumbered gunman to escape unharmed. But some in the crowd may be hotheaded and charge. Here’s where the nature of the weapon matters. If it’s a single shot pistol, once the hothead is shot the rest of the crowd knows its in no danger and will charge. If it’s a revolver, it takes six hotheads to start a run. If it’s an automatic with 10 rounds in the magazine, it takes even more hotheads. (To make the analogy fit, there has to be some advantage of being first to get to the guy with the gun.)
The European plans are all based on the idea of convincing investors that there are a lot of bullets in the gun in the hope that this will deter a self-feeding run. This is why Europe is so focused on expanding the size of backstop funds, including the use of ECB leverage. The bigger the country that is at risk of a run, the bigger the fund must be to reduce the likelihood of a run. This is why Italy’s problems are a real challenge to Europe. Italy’s debt is so much larger than Greece’s or Portugal’s that commensurately bigger sums are needed to reassure investors that a run cannot get started.
The expectations idea also helps explain why policy mistakes can have devastating consequences. Lehman is a good example. Based on the Bear Stearns experience, investor expectations were that a big financial institution would be bailed out. This made them less likely to run. But the decision not to bail out Lehman changed those expectations, and runs began on every major bank. A policy misstep can cause things to go non-linear in a hurry.
This expectations management strategy is most likely to work, the more credible the promise of the provider of the backstop. Here is where politics comes in, and where the design of the backstop matters. If funding is not locked in, but it is based on promises by governments to pay in the breach, the backstop is less credible: when called on, governments may find it in their self-interest to renege on their commitments, or voters may force them to do so. This is a major problem in Europe given its current structure. Stability is a public good (in the economic sense of the term) and individual nations have incentives to free ride. But if the backstop is not credible–if investors believe the gun is really empty–it is of limited value in stopping runs.
So the creation of a backstop creates a meta-”global game”* in which there is uncertainty about the credibility of the backstop, and the backstop mechanism may be subject to a run.
Moreover, the public good nature of stability makes it difficult to negotiate the loss sharing agreements in the first place. There are gains from trade to crafting an agreement that reduces the likelihood of runs, but any such agreement exposes the participants to credit risk. Yes, an agreement may forestall a run on Italy, but the reason for a run in the first place is the recognition that Italy’s risk for insolvency in the future is greater. Which means that any backstop fund that acquires Italian debt is at risk of loss as the price of the debt changes to reflect changes in the likelihood of ultimate repayment. Each Eurozone county wants to reduce the likelihood of a run, but would like all the other countries to bear the cost of doing that. Similarly, in the event a run on the backstop does occur, each country would like other countries to pay a disproportionate share of the associated costs. Negotiating such a loss sharing arrangement, particularly in a polity like Europe, is extremely difficult.
This is complicated even more by the fact that some losses have already been incurred. This is why Greece matters. Greece is already a lost cause, and everybody knows that holders of Greek debt have suffered mark-to-market losses. (The same is true for Portuguese debt too.) These losses threaten the financial condition of banks and insurers throughout the Eurozone. The failure and imminent nationalization of Dexia (the Belgian-French bank) and the tenuous condition of major French banks are symptoms of the problem. The run dynamic threatens the banks, which means that they need to be recapitalized in order to mitigate that potential: this raises the question of where the capital will come from. Public support may be required here too.
But the outstanding losses are not distributed equally across countries. This leads to disputes among EU members over how to provide the public support. Those facing bigger losses want to use EU-wide mechanisms that spread the losses to other countries facing smaller losses. The latter want those facing bigger losses to contribute the lion’s share of any public support.
This is quite clear in the case of France (whose banks are acutely exposed) and Germany (whose banks are less vulnerable). Sarkozy wants to use European Financial Stability Fund monies to recapitalize: that would spread the costs around to all the countries that contribute to the EFSF. Merkel says “nein”: private capital first, French government money next, and EFSF money only as a last resort. In other words, she’s telling Sarko if you want to recapitalize your banks, use your own effing money. Sarko, in essence, is trying to buy insurance from Merkel after receiving a cancer diagnosis: Merkel is not interested.
These distributive conflicts between countries, and the nature of EU governance, are combining to make any resolution an agonizingly long process, and one that may well fail. It makes it devilish hard to negotiate an agreement. It makes the credibility of any agreement doubtful.
Moreover, there is a strong incentive of those governments negotiating an agreement to obfuscate and conceal. They don’t want their voters to know how big the cost is and who is bearing it. This helps explain the convoluted financial structures being mooted to provide the backstop support. These structures would make Enron’s Andy Fastow quite proud. It is virtually impossible to figure out where the losses would actually ultimately land. Which is the point. But convoluted structures raise questions about their credibility as a backstop. Political self-interest and collective interest are again at odds.
Add all of this up, and you have a situation that is virtually impossible to predict. Coordination situations are often unstable and result in jumps from good to bad equilibria in the blink of an eye, in response to seemingly trivial information. The ability to prevent runs on individual countries through the creation of collective backstop mechanisms just transfers the run risk to the backstop mechanism. The backstop’s run risk depends on the credibility of the commitments made, which depends on political dynamics in numerous countries. Distributive/rent seeking/free riding problems make it very difficult to negotiate a sharing and commitment arrangement in the first place, and raise doubts about the credibility of commitments. All of these things are fiendishly difficult to predict, and the resulting uncertainty itself contributes to the likelihood of a run.
Coordination problems. Problems making credible commitments. Collective action problems. Agency problems. Individually, the knottiest problems in economics and institution design, all in one ugly package.
My read is that although one can construct scenarios in which things play out well, these scenarios are analogous to drawing several inside straights in succession. Everything has to go right. Given the natures of the linked problems–coordination problems, the difficulty governments face in making credible commitments, the difficulty of negotiating agreements with distributive as well as efficiency effects–there are many things that can go wrong. The negotiation and credibility problems are acute given the European political structure, and the uninspiring leadership in place there. Moreover, feedback loops are positive here, and that’s always very dangerous because small events can have huge consequences that spin out of control. Put it all together, and there is reason to sweat.
Bottom line. High risk of things ending badly. Virtually impossible to predict the timing of the bad end, or the exact way the bad end will play out.
* The literature on global games formalizes the logic of equilibria in coordination games.
Leave a Reply