Abstract
This article proposes a nonlinear model of bivariate Generalized AutoRegressive Conditional Heteroscedasticity with mean (GARCH-in-Mean) to construct time-varying exchange-rate uncertainty and estimate the effects of exchange-rate uncertainty on unemployment in three developing Asian countries. The effect that increasing exchange-rate uncertainty has a positive but unobvious impact on unemployment is verified for Taiwan and Singapore. The shock of exchange-rate uncertainty provides a large positive stimulus to unemployment initially, but the stimulus gradually falls in Singapore. In Taiwan, this shock on unemployment is relatively small than that in Singapore and will die out eventually.
Notes
1The significance levels of Chow test for the group of break points are below 4%. Thus, the values for the group of break points are all significant at the 5% significance level.
2Data of Singapore are not available from 1983Q1 to 1986Q4. Data of South Korea are not available in 1983Q1.
3Augmented Dickey–Fuller (ADF), Dickey–Fuller Generalized Least Square (DF-GDLS) and Zivot–Andrews tests are employed to test whether the variables are stationary. In Zivot–Andrews test, structural break points for the three countries are different.