Determinants and Impact of Hidden Liquidity : Evidence from NASDAQ ModelView Data

We take advantage of a unique data set, NASDAQ ModelView, to empirically analyze the determinants and the impact of hidden liquidity on public exchanges. Our findings are as follows. First, the cross-sectional presence of hidden liquidity is well explained by observable and readily available stock characteristics. The spread captures most of the limit order book's predictive power over hidden liquidity. Second, hidden orders arrive in large volumes, at few points in time and cluster around few price quotes. Third, these hidden order submissions have an economically significant impact on price returns. While we do not observe an impact for displayed orders, the impact of hidden orders exceeds the impact of important earnings announcements news. We find evidence that these effects do not arise from information arrival but from trading frictions. Accordingly, hidden orders impair an efficient coordination between the supply and demand side of liquidity. Ultimately, large hidden orders are forced to execute against a "thin" market leading to higher price fluctuations