ABSTRACT
This research addresses the omitted-uncertainty bias and functional form misspecification problem when estimating the saving-investment relationship. We apply a semiparametric varying-coefficient cointegration approach to test for the presence of an ‘uncertainty-varying’ cointegrating relationship between savings and investment. Based on quarterly time series data for the U.S. from 1947Q1 to 2014Q4, we find that the saving-retention coefficient is a U-shaped function of economic policy uncertainty. This indicates that domestic investment is more constraint by domestic savings at very high and very low levels of uncertainty. Given the continued rise in economic policy uncertainty, our estimate predicts that domestic investment will become more constrained by domestic savings. Further sub-period analysis shows that the null hypothesis of the presence of an uncertainty-varying cointegrating relationship between savings and investment cannot be rejected in the recent sub-period of 1999Q1-2014Q4, suggesting that the U.S current account is sustainable.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 Using a cross-sectional regression of average investment over the period 1960–1974 on average savings for 16 OECD countries, they found that the saving-retention coefficient is close to one.
2 Li, Lin, and Hsiao (Citation2015) used the varying-coefficient cointegration approach to test the purchasing power parity hypothesis, allowing the relationship between exchange rate and price levels to depend on yield spread differentials between the two countries.
3 The EPU is built using frequency counts of ‘uncertainty’ (and its variants) in newspaper articles.
4 For the U.S. the EPU index dates back to 1947. For other countries, the EPU index dates back to 1985.