The effectiveness of the order-splitting strategy: an analysis of unique data

Abstract

This letter first examines the effectiveness of the order-splitting strategy by analysing the unique intraday dataset of the KOSPI200 futures market, which contains high-quality information on the classes and identification of investors. The empirical finding indicates that a significant number of traders fragment their trades, and these split trades are generally more informative than nonsplit trades. When trades are fragmented, the price impact of buy trades is significantly larger than the price impact of sell trades. By conducting an analysis of investor type, we also find that the order-splitting strategy of domestic individuals, banks and foreign institutions is quite successful in the futures market; on the other hand, the split orders submitted by domestic money managers typically incur a substantial loss.

Keywords:

• order splitting
• stealth trading
• KOSPI200 futures market
• investor type

JEL Classification:

• G15
• G19

Notes

¹ After the pioneering work by Barclay and Warner, most of the following studies interpret the order-splitting behaviour of informed investors under the framework of the stealth-trading hypothesis and link the higher information content or price impact of ‘split’ trades (thus, their size is relatively small) to this hypothesis.

² Lee (2008) also examines the data containing the identification of investors. However, she covers the part of equity transactions for only 4 months in the Korean stock market.

³ The dataset yields the account number of each investor, which is encrypted as nine special characters. All encrypted values have a one-to-one correspondence with their original account numbers.

⁴ In the KOSPI200 futures market, the trading volume of nearby futures contracts always dominates that of other maturity contracts. During the sample period, 5 trading days are excluded due to serious defects in the raw data.

⁵ The daily average number of domestic individuals who do not split their trades is 2777.6; about 355.7 individual investors fragment their trades per day. Thus, we can say that about 11.4% (=355.7/(355.7 + 2777.6)) of the domestic individual investors fragment their trades. For foreign institutional investors, this figure is calculated as 20.4% (=15.0/(15.0 + 58.6)).

⁶ More specifically, in the case of split trades initiated by domestic individuals, the raw Weighted Price Contribution (WPC) value of buy trades is about 3.2 times greater than that of sell trades. If we gauge the price contribution of trades using the relative WPC measures, this asymmetric effect between buy and sell trades becomes more prominent.

⁷ The continuous trading session of the KOSPI200 futures market starts at 9:00 am and ends at 3:05 pm each trading day. There are 1-hour long pre-opening sessions (from 8:00 to 9:00 am) and 10-minute long closing sessions (from 3:05 to 3:15 pm) where the accumulated orders are transacted under the uniform pricing rule.