Title: The fractionally cointegrated VAR model with threshold adjustment
Authors: Chi Wan Cheang - University of Southampton (United Kingdom) [presenting]
Abstract: The fractional cointegrated vector autoregressive (FCVAR) model is extended by allowing two regimes in the speed of adjustment parameter in the error-correction term, treating the long-run cointegrating vector, the fractional and cofractional orders invariant across regimes. Since the threshold parameter is not identified under the null hypothesis of no threshold, a SupLM test for the presence of a threshold is proposed for the FCVAR model. Hence the bootstrap null distribution and p-value are derived. The size and power of the test are investigated through a Monte Carlo simulation. The asymptotic of the cointegrating vector and the threshold parameter is discussed through the smooth transition case. The proposed nonlinear FCVAR model is applied on the volatility index (VIX) and its related futures. By the fact that investors treat the VIX-related products as a security on tailed risk, the VIX products are more attractive in uncertain time that in quiet time. It is argued that the adjustment between volatility index and its futures towards equilibrium could be regime-switching. Empirical result provides a theoretical ground to explain the momentum strategy used by volatility traders.