Estimation of level-I hidden liquidity using the dynamics of limit order-book

Abstract

Many invisible forms of market liquidity exist since many market participants often prefer to hide their trade intentions. Among others, hidden limit order placements are allowed in many public exchanges and generate hidden liquidity. As a result, only part of market liquidity is visible, leading to markets with incomplete information. This study investigates how hidden liquidity alters the econophysical dynamics of limit order books and price impact functions. Accordingly, this study proposes an estimation method of level-I hidden liquidity (hidden waiting orders at best prices) using only publicly available data. Though direct validation with actual hidden liquidity was not possible, this study demonstrates that estimated hidden liquidity provides two empirical benefits. First, estimated hidden liquidity enhances an existing price-impact function and achieves an R-squared value of 70.8% on average. Second, estimated hidden liquidity improves order-book pressure models that forecast the future direction of price change. Using the central notion of market liquidity, this study investigates the different subjects of high-frequency data in an integrated manner, such as the dynamics of execution, price-impact function, and order-book pressure.