Dynamics of liquidity in order book markets

Following the LTCM collapse and the Asian crisis, liquidity has become a key issue for practitioners, regulators, and the academic community. Meanwhile, the limit order book trading platform has spread around the world, either to replace market marker systems, or to compete against them and prevent market failures (such as the collusion between dealers at the Nasdaq). The growing importance of order books, also called automated auction markets, has spurred a growing academic literature. The present thesis belongs to this new stream of literature and gives insights into the provision and consumption of liquidity by traders in automated auction markets. All chapters are empirical. Chapters 3 and 4 use order book data from the Frankfurt Stock Exchange, operating with the Xetra platform. In the last chapter the methodology is applied to data from Euronext. The rules implemented in both trading platforms, Xetra and Euronext, are the classical rules applying to limit order book systems. After introducing concepts and measures in Chapter 2, our research agenda has pursued three different paths, that we very briefly summarize hereafter. Starting from a descriptive study of liquidity dynamics (Chapter 3), we continued with the analysis of the interaction between, on one hand, liquidity and trade informativeness (Chapter 4), and, on the other hand, between liquidity and volatility (Chapter 5). Our main conclusions read as follows. First, liquidity supply at the best quotes (the spread) does not comove with the liquidity supplied by the book (the so-called price impacts). However, the two sides of the book, buy and sell, partially comove. Second, when facing private information, traders behave more cautiously and the book becomes more shallow. Finally, contrary to Cassandra's expectations, the limit order book is resilient to large changes in prices, despite the absence of a committed provider of liquidity.Doctorat en sciences économiques (ECON 3)--UCL, 200