business cycle insurance

I am trying to write a paper on the economic impact of an EU-wide unemployment insurance mechanism (joint with Stephane Moyen and Nikolai Staehler from the Bundesbank). Such a mechanism has been suggested, among others, by EC President van Rompuy and IMF staff. The basic idea is that unemployment insurance might be a simple and implementable way of sharing cross-country risk. Country-specific booms and recessions could be mitigated through transfers among EU member states which take place through differences in contributions paid to and benefits drawn from a common unemployment insurance scheme.

But after a few discussions, I have had to learn that the appeal of cross-country insurance is difficult to sell. This could be a result of my meager sales skills, but I suspect there is something deeper in the resistance to this type of business cycle insurance.

That insurance in general is desirable is the same as saying that people are risk averse, which is not all too disputable. If you know that your house can be destroyed in a fire, then you probably like taking out a fire insurance contract. Most of the time, you pay an insurance company money without getting anything in return, but in the event of a fire you get a lot of money to compensate for the damage to your house. This prevents you from having to make big cuts in your consumption and to run down your lifetime savings when you need to rebuild your house. On the other side, the insurance company is able to provide you such a contract because it pools risk: It receives small payments from a large number of people whose house is not burning and makes large payments to a few people whose house is burning.

The last paragraph is so commonplace it's almost a waste of space. But now replace "house" with "country" and "fire" with "recession", and it becomes a highly controversial issue. Why? What's wrong with building a cross-country insurance mechanism in the EU, for example through a common unemployment insurance system, which every year pays out transfers to countries in a recession and collects premia from countries with healthy economies?

There are numerous attacks on the proposal: the difficulty of measuring recessions as opposed to structural changes, the political viability, the size of welfare gains, and correlation of business cycles across countries, to name just some of the most frequent ones. But the one argument that dominates any debate about a European "transfer union" is moral hazard: Germans don't want to pay for Spaniards and Greeks because they suspect that doing so prevents reforms in those countries and perpetuates the alleged laziness of their workers and politicans.

Moral hazard in general is a powerful argument against insurance. It can also be applied to fire insurance, where it amounts to saying that taking out insurance will make homeowners set fires on purpose or make them less willing to take precautionary measures to prevent fires . The first problem is not very large: few people want to commit insurance fraud if it means setting their own house on fire. The second problem is more relevant, and insurers write clauses that preclude payment when obvious fire prevention measures haven't been taken. But nobody questions the usefulness of fire insurance in general. This is because fire is regarded as an exogenous event: we believe it generally happens unexpectedly to people and without them being able to affect its occurence.

Modern macroeconomic modeling practice sees a recession much in the same way as a fire: it is caused by an exogenous shock, outside of the control of agents in the model. Unexpectedly, productivity drops, credit tightens, people become thrifty or lazy. Yet many people, including economists I have spoken to about our paper, are worried that introducing insurance against business cycle risk would somehow lead to more, longer or deeper recessions, as if this was something in people's control.

So do economists really believe that business cycles are caused by exogenous shocks? I don't think so. It's simply too much to stand up and say that a recession is just "bad luck". It's also not what economic advisors are paid for. Instead, economists spend most of our times explaining the latest business cycle episode as an endogenous build-up: now, the financial crisis was caused by reckless overborrowing and taking on systemic financial risk, whereas before the crisis, the preceding boom was said to be the endogenous product of successful financial deregulation, "anchored expectations" and lower trade costs, among other things. To get more in the European context, the fact that Germany is not in a recession but Spain is is most often attributed to differences in labour market, industrial and general fiscal policies. It is not attributed to Spain getting some random bad shock and Germany getting a random good shock. And if business cycles are the products of the vice and virtue of governments and citizens, then obviously you shouldn't insure them against bad outcomes!

The problem is that any narrative about endogenous causes of business cycles runs into problems with modern macroeconomic and macroeconometric methods. There, we model business cycles not as "cycles", but as outcomes of a small set of random shocks. Econometrically, it is hard to predict business cycles, hence the estimations attribute them to random shocks. Theoretically, endogenous business cycles are really hard to produce as well: them being endogenous implies that their existence is due to agents' choices. But standard theories abound with rational expectations, optimising agents, perfect information and efficient markets, which means that agents disliking large fluctuations will not make such choices.

In principle, the same urge for endogenisation applies to houses on fire. If your house burns down because of a leaking gas pipe, even if this gas pipe was properly and regularly checked, you will probably blame it on the gas engineer, or on your decision to have a house with gas heating. But at some point, at least from a social perspective, we accept that fires just sometimes happen, that it is bad luck, and that people should be able to insure against it. The only place where we think otherwise is insurance fraud. But as mentioned, this is a rare thing. Likewise, it would probably be strange to argue that the main moral hazard issue in cross-country insurance were that it would induce the governments of Southern Europe to trigger recessions on purpose in order to claim transfers from the North. After all, how many people would want to set their own house on fire even if they get some insurance money for it?

But of course, it's not only about what caused the fire, but what you did to prevent it, or mitigate its spreading through your house. In the same spirit, one can argue that recessions are indeed caused by exogenous shocks, but that the reaction of some country's economy to those shocks is endogenous. For example, Spain could make itself more resistant to business cycles by reforming its labour market towards more flexibility; a European insurance mechanism might destroy the incentives to do so. That is a valid point.

There is a great Econometrica paper by Persson and Tabellini that makes this point theoretically, using a stylised and static political economy model. Unfortunately though, it is hard to make it within the context of business cycle macroeconomics, which hardly ever considers endogenous government policies. Bridging this gap between political economy and business cycle theory seems a daunting task. And how would one go about calibrating moral hazard of national governments to data? The deeper problem here, I think, is that economics offers no good framework how governments actually make choices; and the only framework of how they should make choices, namely as benevolent, rational, optimising social planners, is inadequate to the problem.