Price clustering: Evidence using comprehensive limit-order data

Abstract

Employing comprehensive limit-order data which identify investor types, this paper examines the clustering pattern of limit-order prices. First, limit orders, particularly those submitted by individual investors (IIs), tend to cluster at integer and even prices. Second, nonmarketable limit-order prices cluster more than marketable limit-order prices, indicating that aggressive limit orders generally embed more information. Third, investors choosing even-priced limit orders are not penalized by lower execution ratios. Fourth, investors (particularly IIs) strategically exhibit front-running behavior. Fifth, price clustering indeed creates price barriers. Finally, the degree of price clustering using trade data is significantly underestimated, compared to that using limit-order data. © 2009, The Eastern Finance Association.