Valuation effects and long-run real exchange rate dynamics

• I test for saddle-path dynamics between net foreign assets and real exchange rates.
• I use the Johansen method to test for saddle-path dynamics.
• I studied five developed countries who can borrow in their own currency.
• Saddle-path dynamics is rejected because predictable valuation effects are small.
• The net foreign asset position does not determine long-run real exchange rates.

This paper uses the Johansen test for cointegration to check the prediction of a portfolio balance model that predictable valuation effects are associated with a saddle-path dynamic relationship between the net foreign asset position and the real exchange rate. The analysis uses newly constructed quarterly series on the net foreign position as a percentage of the nominal gross domestic product, together with data on real effective exchange rate indices for a sample of developed countries which borrow in their own currency. The results indicate that the net foreign asset position and the real exchange rate are not cointegrated for all the countries in the sample. The rejection of saddle-path dynamics suggests that predictable valuation effects are quantitatively small in developed countries. The rejection of cointegration suggests that the net foreign asset position is not a determinant for long-run real exchange rates in developed countries.