A threshold model of the US current account

What drives the US current account imbalances? Is there solid evidence that the behavior of the current account is different during deficits and surpluses or that the size of the imbalance matters? Is there a threshold relationship between the US current account and its main drivers? We estimate a threshold model to answer these questions using the instrumental variable estimation proposed by Caner and Hansen (2004). Rather than concluding that the size or the sign of (previous) external imbalances matters, we find that time is the most important threshold variable. One regime exists before and another one exists after the third quarter of 1997, a period that coincides with the onset of the Asian financial crisis and the Taxpayer Relief Act of 1997. Statistically significant determinants in the second regime are the fiscal surplus, productivity, productivity volatility, oil prices, the real exchange rate, and the real interest rate. Productivity has become a more important driver since 1997.