| Abstract: | The purpose of this study is to extend the bid-ask spread decomposition literature into the order-driven environment. The use of electronic limit order books, combined with order-driven market making, has been increasing rapidly in recent years because of improvements in information technology and financial market deregulation. To date, reported bid-ask spread decompositions rely almost exclusively on quote-driven or hybrid systems. This study provides bid-ask spread component estimates from one of the world's largest order-driven markets, the Stock Exchange of Hong Kong. Based on a sample of over six million observations, we estimate a median adverse selection component of 33 percent and a median order processing component of 45 percent of the spread. Dollar-volume-based decile portfolios show significant cross-sectional variation for adverse selection costs but insignificant variation for order processing costs. Finally, order persistence is consistently positive for all deciles and displays a direct relation with the level of trading activity. |
