Working Paper: CEPR ID: DP4456
Authors: Bart Frijns; Peter Schotman
Abstract: In this Paper we propose a tick time model for dealer quote interactions using ultra-high-frequency data. This model includes duration functions to measure the time dependence of volatility as well as information asymmetry. In order to assess price discovery we define several measures in tick time. These measures can be aggregated to calendar time and we define a comparable measure to Hasbrouck (1995) information shares. In our empirical part we examine the Island and Instinet Electronic Communication Networks, and three wholesale market makers for 20 actively traded stocks with varying liquidity at Nasdaq. Our results include that volatility does not increase with the duration between quote updates, and that longer quote durations lead to lower price discovery. In terms of price discovery we find that ECNs tend to dominate the liquid stocks, whereas market makers dominate the less liquid stocks.
Keywords: microstructure; NASDAQ; price discovery; tick time models; ultrahigh frequency data
JEL Codes: C32; G15
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| duration (C41) | price discovery (D47) |
| duration (C41) | volatility (E32) |
| electronic communication networks (ECNs) (D85) | price discovery (D47) |
| market makers (D40) | price discovery (D47) |
| duration (C41) | information flow (D83) |