Explaining the dynamics of the NIKKEI 225 stock and stock index futures markets by using the SETAR model

Theory predicts that, if stock and stock index futures markets operate efficiently, price movements in these markets should follow a first-order vector error correction model in which the error correction term represents the basis, and in which there are no regimes. However, following Brooks and Garrett (2002), by using the self-exciting threshold autoregressive (SETAR) model, in this article, we show that there are three regimes in the dynamics of the basis of the NIKKEI 225. In addition, in the central bound, autocorrelation exceeding the first-order variety is observed. This indicates that the basis is persistent and predictable, and triggers no arbitrage in the central bound. For Japan, the basis is successfully explained by a SETAR model with two thresholds, as suggested by Brooks and Garrett. However, the adjustment pattern of the basis outside the central bound differs from that observed for the United Kingdom.