This paper examines the price discovery role of KOSPI 200 index futures for its cash index using a five-minute for the period of 1996. 6. 14 ~ 2003. 6. 12, intraday data. Since the cash index and futures are non-stationary and cointegrated, we use the error correction model with the non-filtered data. Using the cointegration and error correction model, lead from the stock index futures to the stock index has been confirm with previous studies, such as Kim and Kim(2000). In this paper, the data was filtered to multi-scale by wavelet analysis. Since the filtered data are stationary, the VAR model was used to examine the difference of the price discovery role in the different scale.
First, the cash and futures markets were cointegrated and futures prices tended to lead the cash index by forty minutes and the lead effect of the cash index was disappeared in five minutes with the expanded period of data.
Second, sub period analysis has been performed to discover improvement of futures market efficiency as the futures market matured. Futures price led the cash index more intensively with the trading volume of future market and cash market increasing. But, the lead time become shorter in 1% significant level. We can consider this result from the decrease of infrequent trading effect.
Third, the VAR analysis with wavelet filtered data shows that the cash index and futures lead each other in scale level 1 and the futures leads the cash index in scale level 4 and 7. The result with non-filtered data shows unidirectional lead effect of futures but, there is bidirectional lead effect of cash index and futures in very short term and unidirectional lead effect of futures in longer term. This supports that investors, who have different investing time horizon, has different lead-lag relationship.
