Re-examining the Feldstein-Horioka Puzzle in Few Asian Countries: A Nonlinear ARDL Approach

Abstract

This study investigates the validity of the Feldstein-Horioka (FH) hypothesis in four Asian advanced economies countries. The hypothesis claims a positive (weak) connection between domestic investment and domestic saving, implying a low (high) capital mobility. The FH hypothesis is reconfirmed and revalidated by examining the twin deficits (TD) hypothesis. The annual data are from 1982 to 2018. The nonlinear autoregressive distributed lags (NARDL) model is utilized to compare with the benchmark linear ARDL model. The empirical results reveal a cointegrating relationship between savings and investments in all countries. The NARDL model is preferred with better model fitting and its ability to capture asymmetric effects. Both FH and TD hypotheses validate the FH hypothesis (a high positive correlation in investment-saving) in China but with a puzzle (negative relationship) found in Korea based on short and long run estimates. On the other hand, a high positive correlation is also found in Japan and Singapore in the long run. The results implies a low capital mobility condition in these advanced countries. Trade openness and export are the main determinants of the investment-saving relationship for Japan, Korea, and Singapore, while economic growth is the only determinant in the relationship for China.