Chaotic structure in intraday data of JGB futures price

The purpose of this study is to examine the deterministic and chaotic structure in intraday data of Japanese government bond (JGB) futures price. In the first place, the characteristics of the price scales and of the time scales of the data are considered. In JGB futures market, the contracted prices are discrete with the tick size, and there is no transaction while the market is closed (e.g. in the early morning, at lunch time, and at night). Given these factors, two price scales, the price and the deviation from the moving average, and two time scales, the number of transactions and the trading time, are adopted in this paper. And then, the chaos characteristics such as the correlation dimension and the maximal Lyapnov exponent are estimated in the various cases. Finally, the significance of the estimated chaos characteristics are tested statistically with the method of surrogate data. The result implies the possible existence of chaos when the number of transactions is used as the time scale. This possibility is attributed to the characteristic features of the JGB futures market.