Explaining the saving-investment relationship with threshold effects

There have been many attempts to explain the unreasonably high correlation between domestic saving and investment rates. The threshold testing procedure developed by Hansen (1999) provides a framework for testing the effects of key variables relating to capital mobility in conjunction with the saving-investment relationship. Ho (2003) first applied this method to the saving-investment puzzle controlling for thresholds in country size. Extending this model, this paper reports a number of significant thresholds effects for country- size, trade and financial openness measures, age dependency ratios and trade balances. After controlling for threshold effects the relationship between savings and investment is found to be statistically insignificant. Additionally, controlling for the thresholds effects in a dynamic model of the current account allows for direct comparison between the savings-investment coefficient and adjustments to a country’s external balance.