Investigating the information content of non-cash-trading index futures using neural networks
Introduction
It is quite common that the daily trading time of an index futures contract begins earlier and ends later than which of its underlying spot market. In this study, the time period between daily close and the subsequent day's opening of the cash index trading is defined as the non-cash-trading (NCT) period. Valuable information should be obtained through the analysis of the NCT futures and contribute to the success of exercising proper investment decisions for the underlying spot market. This research investigates the information content of SGX-DT (Singapore Exchange-Derivatives Trading Limited) Nikkei 225 and MSCI Taiwan index futures prices during the NCT period. Unlike past studies (Chan, 1992, Hiraki et al., 1995, Iihara et al., 1996, Martikainen and Puttonen, 1994, Min and Najand, 1999, Pizzi et al., 1998, Stoll and Whaley, 1990) only reported that index futures price changes lead price changes of the underlying spot market (the so-called lead–lag relationship between futures and cash market), this study tries to analyze this phenomena using a two-stage approach. It first tests if there is a lead–lag relationship between the index futures during the NCT period and the cash market index during its opening period. The obtained leading futures and previous day's cash market closing index are then used to predict the opening cash market price by the artificial neural networks (ANNs) model. The rationale underlying the analyses is to learn the information content of the NCT SGX-DT Nikkei 225 and MSCI Taiwan futures prices by comparing the opening cash price index forecast including NCT futures and the random walk1 (with previous day's cash closing index as the forecasted opening cash price index) assumption. If the former forecasts outperform the latter then the information of the NCT index futures is considered valuable. Note that it is valuable to use the lead–lag relationship analysis as a supporting tool for neural networks as we can learn more about the inner workings. Besides, as there is no theoretical method in determining the best input variables, this procedure can be implemented as a generally accepted method for determining appropriate leading futures and thus giving statistical support in deciding the input vector of the designed neural network model.
Please note that we cannot use the popular ARIMA forecasting technique here since ARIMA is an univariate model which only uses the historical data to make inferences about the variable we are interested. And in this study we will be using both the cash index and futures index in predicting the opening cash price index and hence make the ARIMA approach inapplicable here. The ANNs are adopted in building the forecasting model with its ability to capture subtle functional relationships among the empirical data even though the underlying relationships are unknown or hard to describe. Besides, no strong model assumptions (variation homogeneity and system stationarity) are required and the literature on applying ANNs in the finance area is vast and fruitful (Vellido et al., 1999, Zhang et al., 1998). In this study, the backpropagation neural networks (BPN) with various numbers of nodes in the hidden layer and different learning rates is extensively studied to address and solve the issue of finding the appropriate setup of the topology of the networks. Further studies are performed on the robustness of the constructed networks using different training and testing sample sizes. To evaluate the effectiveness of the proposed neural network model, the daily 5-min transaction data of index futures and cash prices from October 1998 to March 1999 of SGX-DT Nikkei 225 and MSCI Taiwan index futures are used as illustrative examples. Finally, analytic results compared with BPN with previous day's closing index as the input, the random walk (with previous day's cash closing index as the forecasted opening cash price index), and GARCH model forecasts are also discussed.
The rest of the paper is organized as follows. We will give a brief overview of lead–lag relationship and neural networks in Section 2. The hypotheses and assumptions for the proposed study are presented in Section 3. Section 4 describes the lead–lag relationship analysis and the development of neural networks forecasting model. To verify the robustness of the designed neural network model, the prediction efficiency is summarized using different training and testing sample sizes in Section 5. Section 6 addresses the conclusion and possible future research areas.
Section snippets
Lead–lag relationship
Based on Chan, 1992, Iihara et al., 1996, an autoregressive (AR) model is established for cash returns to account for the non-synchronization in the cash trading, one of the reasons that causes a result of futures leading cash.2 The residual of the model is considered as the proxy of the real return. Following the same process, this study fits an AR model for the futures return to remove the autocorrelation component and then takes the residual as the proxy of
Hypotheses and assumptions
For SGX-DT Nikkei 225 futures contracts and its underlying cash market, there are two trading sessions in each trading day. For futures (cash) trading, the morning session is from 07:55 to 10:15 (from 08:00 to 10:00) and the afternoon session is from 11:15 to 14:25 (from 11:30 to 14:00), Singapore time. Therefore, the information contents of NCT (from 14:00 to 14:25 in each trading day and from 07:55 to 08:00 in the following trading day) futures prices are analyzed in this study. As to the
Empirical results and discussion
The daily 5-min transaction data of futures and cash prices from October 1, 1998 to March 31, 1999, provided by the Reuters database, is used in this study. The Nikkei 225 futures dataset used in the lead–lag relationship analysis includes 12 5-min NCT futures prices from 14:05 to 14:25 in each trading day and from 07:55 to 08:25 in the following trading day. For cash prices to match in time with the futures prices, there will be six unavailable cash prices (which will be treated as missing
Robustness evaluation of the neural network model
To evaluate the robustness of the neural network model, the performance of the designed neural network was tested using different ratios of training and testing sample sizes. The testing plan is based on the relative ratio of the training data set size to the complete data set size. In this section, four relative ratios, 50, 60, 70, and 80% are considered. The prediction results for the opening cash price index by the designed BPN model are summarized in Table 8 in terms of two criteria, the
Conclusions and areas of future research
This study uses SGX-DT Nikkei 225 and MSCI Taiwan intraday 5-min data from October 1, 1998 to March 31, 1999 to analyze the information content of futures trading in the NCT period. Lead–lag relationship analysis is first implemented in obtaining the futures leading the opening cash price index. The obtained leading futures and previous day's cash closing index are then served as the input nodes of the BPN models in forecasting the opening cash price index. It is concluded that the BPN model
Acknowledgements
The authors would like to thank Prof. Chih-Chou Chiu at National Taipei University of Technology for his valuable comments that greatly improve the quality and presentation of the paper. The authors would also like to thank Capital Futures Corporation, Taipei for providing the data that makes the research possible.
References (27)
Generalized autoregressive conditional heteroscedasticity
Journal of Econometrics
(1986)- et al.
The information content of end-of-the-day index futures returns: International evidence from the Osaka Nikkei 225 futures contract
Journal of Banking and Finance
(1995) - et al.
A note on the predictability of Finnish stock market returns: Evidence from stock index futures markets
European Journal of Operational Research
(1994) - et al.
Neural networks in business: A survey of applications (1992–1998)
Expert Systems With Applications
(1999) Time series forecasting using backpropagation neural networks
Neurocomputing
(1991)- et al.
Forecasting with artificial neural networks: The state of the art
International Journal of Forecasting
(1998) - et al.
Neurocomputing: Foundations of research
(1988) A further analysis of the lead–lag relationship between the cash market and stock index futures market
Review of Financial Studies
(1992)- et al.
Neural network: A review from a statistical perspective (with discussion)
Statistical Science
(1994) Approximation by superpositions of a sigmoidal function
Mathematical Control Signal Systems
(1989)
Large sample properties of generalized methods of moments estimators
Econometrica
Neural networks: A comprehensive foundation
Neurocomputing
Cited by (23)
A novel separation-ensemble analyzing and forecasting method for the gold price forecasting based on RLS-type independent component analysis
2023, Expert Systems with ApplicationsIntegration of nonlinear independent component analysis and support vector regression for stock price forecasting
2013, NeurocomputingCitation Excerpt :To construct the forecasting model for the daily Nikkei 225 closing cash index, we used the Nikkei 225 index futures prices as forecasting variables. This is because the futures price changes usually lead to the price changes of the cash market [41,42] Nikkei 255 index futures contracts are traded on Singapore Exchange Derivative Trading Limited (SGX-DT), Osaka Securities Exchange (OSE), and Chicago Mercantile Exchange (CME) markets. Thus, three previous day's futures closing prices of Nikkei 255 traded on SGX-DT, OSE and CME, respectively, are used as forecasting variables.
Combining nonlinear independent component analysis and neural network for the prediction of Asian stock market indexes
2012, Expert Systems with ApplicationsCitation Excerpt :For evaluating the performance of the proposed NLIC-BPN forecasting model, the daily Nikkei 225 closing cash index and Shanghai B-Share stock index closing price are used in this study. In forecasting Nikkei 225 closing cash index, the Nikkei 225 index futures prices are used as forecasting variables since the futures price changes lead price changes of the cash market (Lee & Chen, 2002; Lee & Chiu, 2002). Using the leading futures as forecasting variables should contribute to the success in increasing the forecasting accuracy.
An efficient CMAC neural network for stock index forecasting
2011, Expert Systems with ApplicationsCitation Excerpt :In forecasting Nikkei 225 closing cash index, four forecasting variables are used. One is the previous day’s cash market closing index and the others are three Nikkei 225 index futures prices since the futures price changes lead price changes of the cash market (Lee & Chen, 2002; Lee & Chiu, 2002). The three Nikkei 255 index futures contracts are traded on Singapore Exchange-Derivative Trading Limited (SGX-DT), Osaka Securities Exchange (OSE) and Chicago Mercantile Exchange (CME) markets.
CAST: Using neural networks to improve trading systems based on technical analysis by means of the RSI financial indicator
2011, Expert Systems with ApplicationsCitation Excerpt :According to Chang and Liu (2008), however, these models have their limitations owing to the tremendous noise and complex dimensionality of stock price data, and besides, the quantity of data itself and the input variables may also interfere with each other. Recently, in the first decade of the 21st century, various studies using ANN have been developed in the fields of forecasting stock indexes (Chang, Liu, Lin, Fan, & Ng, 2009; Chavarnakul & Enke, 2008; Chen & Leung, 2004; Chen, Leung, & Daouk, 2003; Enke & Thawornwong, 2005; Lam, 2004; Lee & Chen, 2002; Lee & Chiu, 2002; Leigh, Hightower, & Modani, 2005; Thawornwong & Enke, 2004; Yao, Li, & Tan, 2000). The importance of further developments in soft computing led to several papers devoted to forecasting stock indexes using techniques such as support vector machines (e.g., Chiu & Chen, 2009; Huang, Nakamori, & Wang, 2005; Kim, 2003; Pai & Lin, 2005; Wen et al., 2010), fuzzy systems (e.g., Chang & Liu, 2008; Chang, Wang, & Liu, 2007; Huang & Yu, 2005; Wang, 2003), genetic algorithms (e.g., Chen et al., 2009; Oh, Kim, & Min, 2005; Oh, Kim, Min, & Lee, 2006; Potvin, Soriano, & Vallee, 2004) and mixed methods (e.g., Armano, Marchesi, & Murru, 2005; Armano, Murru, & Roli, 2002; Hassan, Nath, & Kirley, 2007; Kwon & Moon, 2007; Leigh, Purvis, & Ragusa, 2002).
Integrating independent component analysis-based denoising scheme with neural network for stock price prediction
2010, Expert Systems with Applications