Measuring Stock and Market Sentiment from Full-displayed Limit Order Book Data

Emanuel Albin Kopp, Michael Hütl, Otto Loistl, Johannes Prix

Publication: Working/Discussion PaperWorking Paper/Preprint


One of the most important recent developments in stock exchanges’ information management
systems is the real-time provision of full-displayed order book data to market participants.
For instance, the London Stock Exchange has already introduced such services to
interested clients. This evolution, which also represents an improvement in market transparency,
offers considerably more detailed order book information to stock market traders
than ever before. This additional trading information may also work as input data into
trading algorithms and event stream processing systems. To address this current issue, we
present measurement techniques for real-time order book data, and introduce some partly
new sentiment measures that allow capturing information from the entire limit order book.
Moreover, we study interrelations between the applied variables, and their connections to
price behavior and liquidity supply. To examine to which extent idiosyncratic sentiment
is influenced by market sentiment, we extract the systematic components of each measure
using projection-pursuit principal components analysis. We furthermore consider common
factors in liquidity and discuss the relation between market sentiment and systematic liquidity
in more detail. Empirical results are presented using high-frequency order book data
from the Xetra trading system of the Frankfurt Stock Exchange.
Original languageEnglish
Publication statusPublished - 1 Jun 2008

Cite this