| Summary: | Reported bid-ask spread decompositions vary in exchange structure, for example quotedriven, order-driven or hybrid systems. The majority however, focus on quote-driven markets. This dissertation empirically studies the decomposition of spread components in a order-driven market and identifies relations among the individual components and firm-specific attributes such as trading volume, trade size, share price, volatility of returns, information arrival and market capitalisation. The data used are from a sample of FTSE 100 stocks since the 1997 movement from a quote-driven system to the order-driven system. Based on the sample observations, I implement two decomposition methodologies to estimate the order processing, asymmetric information and inventory holding costs. Each component is then regressed on the six characteristics variables to examine their relationship.
The Stoll(1989) method yields an average order processing cost of 58.17 percent, an average asymmetric information cost of 31.98 percent and an average inventory holding cost of 9.96 percent. The Huang and Stoll (1997) approach estimates mean value of 80.86, 14.84, and 4.29 percent for the order processing, asymmetric information and inventory holding components respectively. Furthermore, the regression results show that transaction size, frequency, price and market value directly relate to the component costs. On the other hand, volatility and information arrival exerts negative affects on the three components.
|
|---|