# 14-027/III (2014-02-27)

Author(s)
Sait Ozturk, Econometric Institute, Erasmus University Rotterdam; Michel van der Wel, Econometric Institute, Erasmus University Rotterdam
Keywords:
High-frequency data, Market microstructure, Price Discovery, Kalman filter
JEL codes:
C32, G14

For many assets, trading is fragmented across multiple exchanges. Price discovery measures summarize the informativeness of trading on each venue for discovering the asset’s true underlying value. We explore intraday variation in price discovery using a structural model with time-varying parameters that can be estimated with state space techniques. An application to the Expedia stock demonstrates intraday variation, to the extent that the overall dominant trading venue (NASDAQ) does not lead the entire day. Spreads, the number of trades and volatility can explain almost half of the intraday variation in information shares.