Abstract
This study presents the empirical evidence of Kyle (Citation1985) that informed traders spread their orders over time to camouflage their information. Our proprietary data, which contain account numbers in the Korean options market, identify whether traders break up his or her order (split orders) or not. We show that split orders are associated with a larger proportion of cumulative price change than nonsplit orders are. Furthermore, nonsplit orders, even small- or medium-size ones, cause much smaller cumulative price changes. These results improve upon the tests of Barclay and Warner (Citation1993).
Notes
1The data in this study are provided by the Korea Exchange. To secure the identification of investors, account numbers are garbled.
2Hansch and Choe (Citation2007) compare the results from different methods of Barclay and Warner (Citation1993), Hasbrouck (Citation1995) and Gonzalo and Granger (Citation1995). They confirm that all results are qualitatively the same. Thus, we follow the method of Barclay and Warner (Citation1993).
3When we use weighted averages as in Chakravarty (Citation2001), we find that the results are qualitatively the same.