As the Greek crisis approaches some sort of resolution, it's worth asking whether the euro ever had a chance of working. While we may never know the answer to this question, we do know that a well-designed euro would have worked far better than the actual euro. And we also know that it actually would have been pretty easy to design a much better system, even assuming that fiscal union was politically infeasible.
Let's start with what we knew even before the euro was created:
1. The one-size-fits-all problem.
2. The need for strict monetary policy rules, to prevent the monetary authority from monetizing public debt.
Normally there are trade-offs in economic policymaking. But not in this case. The optimal policy rule to overcome the one-size-fits-all problem is also best for preventing monetization of the debt.
To minimize the one-size-fits-all problem, you need to minimize eurozone-wide business cycles. That's because there will always be country to country variation in recession intensity. But if you start from a position where the entire eurozone is deeply depressed, then you'll end up with the worst performers being catastrophically depressed. In contrast, if the entire eurozone has stable NGDP growth, even the worst performers will probably be only modestly below normal. So the worst monetary regime (of the plausible alternatives) is actually pure inflation targeting, which puts zero weight on stabilizing the business cycle.
But wasn't inflation targeting supposed to prevent monetizing the debt? Yes it was, but NGDP targeting does just as well. Even better, NGDP level targeting prevents debt monetization even more effectively than NGDP growth rate targeting. And even, even better, NGDPLT avoids severe recessions more effectively than NGDP growth rate targeting.
And note that this is not Monday morning quarterbacking; all of this was easy to anticipate, even back in 1999. That's not to say the actual crisis was anticipated---it wasn't. Rather it was easy to anticipate which type of policy regime would make this sort of crisis least likely to occur.
NGDPLT would not paper over problems like Greece's reckless fiscal splurging; rather it would prevent the contagion effect from occurring. Greece would still struggle, but speculators would not anticipate the crisis spreading to the other eurozone countries. And Greece would only be punished for its own fiscal errors, not both its fiscal irresponsibility and the ECB's overly tight monetary policy.
Just to be clear, I am not endorsing the euro. The system is what it is. It's a bad system precisely because we cannot trust central banks to do the right thing. And in a world where central banks don't do the right thing, individual countries need the flexibility to devalue or revalue.