An empirical study on the lead-lag relationship between individual share futures and spot markets: focused on NHN and GS Construction futures

This study tests the lead-lag relationship between spot and futures markets of NHN and GS construction company. We introduced the daily near by futures price and spot price of the two individual share futures data covering from December, 14, 2009 to January, 4, 2013. For this purpose we employ both Granger causality test and impulse response analysis based on the vector autoregressive model(VECM). Considering the co-integration relationship between spot and futures markets’ level data of NHN and GS construction company. The main empirical results are as follows. First, in terms of basic statistics NHN spot and futures markets are up-ward trend but GS spot and futures markets are down-ward trend during the whole sample period. The standard deviation of spot markets of both NHN and GS construction are relatively greater than those of each futures markets. Second, we estimated whether the level variables and returns of cash and futures price of NHN and GS construction company are stationary or not. For this we estimated the ADF(Augmented Dickey-Fuller) and PP(Phillips Perron) which is used frequently in financial econometrics areas. According to the unit root test results, we find that there is a unit root in the level variables but not in the returns of cash and futures price of NHN and GS construction company. Therefore, we use the returns data to estimate the lead-lag relationship between cash and futures price of NHN and GS construction company. Third, we also investigated whether there is a long-run relationship among the level variables in cash and futures price of NHN and GS construction company. For this, we employ Johansen co-integration test. According to the empirical results, we find that there is a co-integration relationship between cash and futures price of NHN and GS construction company with a statistically significant level. Considering these empirical results, we added the error correction term in the vector autoregressive model. Fourth, in order to the short-run information transmission between cash and futures price of NHN and GS construction company , we estimated Granger causality test based on VECM(4) and VECM(3) model. We selected lag four(4) and three(3) based on estimates of bayesian information criteria for NHN and GS construction futures, respectively . According to the Granger causality test based on VECM(4), we find that there is a feed back lead-lag relationship between cash and futures price of NHN company with a statistically significant lever, but NHN futures’ price discovery effect is dominant. In case of the cash and futures markets of GS construction company, there is a bilateral influence between the spot and futures markets of GS construction with the statistically significant levels. Finallly, we estimated the impulse response analysis based on VECM(4) and VECM(3) to study the persistence of impact between the spot and futures markets of NHN and GS construction company. The empirical results show that the futures markets of NHN and GS construction company reply to the information shock arising from spot markets of the two companies after one day and these impact persistent more than 10 days. but the spot markets of NHN and GS construction company reply instantly the new impact took placed in the spot markets of NHN and GS construction company. From these empirical results we infer that the change in the futures markets of NHN and GS construction company have much more information than those of the spot markets of NHN and GS construction company. We also infer that the futures markets of NHN and GS construction company are more efficient than those of spot markets.