Will the U.S. Trade Deficit End Badly?

In today's NY Times, Paul Krugman renews his prediction that we are heading for a crisis:

the answer to the question, "Why haven't we paid a price for our trade deficit?" is, just you wait.

On the other hand, in a new NBER working paper, Sebastian Edwards says that this outcome, while possible, is unlikely:

In this paper I use a large multi-country data set to analyze the determinants of abrupt and large “current account reversals.” The results from a variance-component probit model indicate that the probability of experiencing a major current account reversal is positively affected by larger current account deficits, lower prices of exports relative to imports, and expansive monetary policies. On the other hand, this probability is lower for more advanced countries, and for countries with flexible exchange rates. An analysis of the marginal effects of current account deficits and of the predicted probability of reversal indicates that both have increased significantly for the U.S. since 1999. However, the level of this probability is still on the low side. I estimate that the predicted probability of a current account reversal in the U.S. has increased from 1.7% in 1999, to 14.9% in 2006.

Update and clarification: The "current account" is a broad measure of the trade balance. Click here for details on the definition.