Fifth Annual Grant Winners 2004-2005

Charles Collver, Ph.D., H. Wayne Huizenga School of Business and Entrepreneurship

Dean Randolph Pohlman, H. Wayne Huizenga School of Business and Entrepreneurship

Title: Does Liquidity Improve When Traders Need It Most? An Examination of the Impact of Decimalization on NYSE Specialist Behavior During Periods of Large Price Changes

Abstract:

Both the New York Stock Exchange and the Nasdaq Stock Exchange have recently switched the way that stock prices are quoted by their respective market makers. Prior to 2001, stocks were quoted in “teenies” or sixteenths of a dollar. Now, stock prices are quoted in decimal form; that is, in pennies. Both the Securities and Exchange Commission and the Securities Industry Association rationalized the switch as a necessary step to make US exchanges more competitive with international stock exchanges, arguing that decimalization would improve the market quality of US exchanges.

Recent research suggests that decimalization has resulted in lower transaction costs, measured as the average bid-ask spread quoted by the market maker. However, this improvement in one measure of market quality is accompanied by decreased depth; also known as a reduction in liquidity, another measure of market quality. To date, researchers have focused on overall average market quality without addressing the issue most traders worry about: buying and selling large quantities of shares when prices are changing rapidly. Additionally, current research fails to recognize that market makers strategically select both the bid-ask spread and the quoted depth simultaneously. Failure to properly model these two endogenous variables properly could result in spurious conclusions regarding the impact of these variables on each other and the impact of other trade-related variables on the market maker’s choice set.

This study seeks to address both concerns. This study presents and defends a bivariate simultaneous equations model that treats both spread and depth as endogenous variables. The trade and quote data will be two years of TAQ data – one year before decimalization and one year after. The model allows for testing of specific hypotheses regarding the impact of decimalization on market quality both before and after decimalization. Additionally, the model allows for assessment of market quality during periods of high information flow and during periods of more normal market behavior.