recent developments in high frequency financial ... - Index of
recent developments in high frequency financial ... - Index of
recent developments in high frequency financial ... - Index of
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
164<br />
contrast to those <strong>of</strong> Pascual and Veredas (2004). A possible explanation for these<br />
conflict<strong>in</strong>g results is that Pascual and Veredas (2004) apply a discrete-time duration<br />
model which does not allow for time-vary<strong>in</strong>g covariates. However, particularly for<br />
the processes <strong>of</strong> the <strong>in</strong>frequent <strong>high</strong>ly aggressive orders, it seems to be essential to<br />
account for changes <strong>of</strong> the order book dur<strong>in</strong>g a spell. 16<br />
Clear evidence for the existence <strong>of</strong> multivariate dynamic structures <strong>in</strong> the order<br />
arrival processes is found. We observe significant spill-over effects between the<br />
both sides <strong>of</strong> the market and—<strong>in</strong> a weaker form—between market trad<strong>in</strong>g and limit<br />
order trad<strong>in</strong>g. The fact that these <strong>in</strong>terdependencies are primarily (significantly)<br />
positive suggests that order book dynamics are driven by general market activity<br />
which simultaneously <strong>in</strong>fluences all <strong>in</strong>dividual processes rather than by economic<br />
“crowd<strong>in</strong>g out” arguments which would imply negative spill-over effects. These<br />
f<strong>in</strong>d<strong>in</strong>gs support the notion that the arrival rates <strong>of</strong> aggressive orders are basically<br />
driven by two pieces <strong>of</strong> <strong>in</strong>formation: (1) the state <strong>of</strong> the market as revealed by the<br />
open limit order book and which directs traders’ order submission strategy, and (2)<br />
general market activity which simultaneously <strong>in</strong>fluences the <strong>in</strong>dividual arrival<br />
rates. 17 Our f<strong>in</strong>d<strong>in</strong>gs show that order book <strong>in</strong>formation plays the dom<strong>in</strong>ant role <strong>in</strong><br />
expla<strong>in</strong><strong>in</strong>g order aggressiveness. In particular, we observe that <strong>in</strong> terms <strong>of</strong> its<br />
explanatory power, a model which excludes all dynamics but <strong>in</strong>cludes order book<br />
covariates significantly outperforms a completely dynamic model that does not<br />
account for the state <strong>of</strong> the market. Nevertheless, the dynamic variables are<br />
absolutely necessary <strong>in</strong> order to obta<strong>in</strong> a well-specified model. These f<strong>in</strong>d<strong>in</strong>gs<br />
provide support for advocates <strong>of</strong> greater transparency <strong>in</strong> electronic trad<strong>in</strong>g and<br />
<strong>in</strong>dicate that real benefits to traders may result from complete disclosure <strong>of</strong> the<br />
order book.<br />
Acknowledgement Special thanks are due to James McCulloch whose assistance <strong>in</strong> prepar<strong>in</strong>g<br />
the data has made this research project feasible.<br />
References<br />
A. D. Hall, N. Hautsch<br />
Al-Suhaibani M, Kryzanowski L (2000) An exploratory analysis <strong>of</strong> the order book, and order<br />
flow and execution on the Saudi stock market. J Bank F<strong>in</strong>anc 24:1323–1357<br />
Bauwens L, Giot P (2000) The logarithmic ACD model: an application to the Bid/Ask Quote<br />
Process <strong>of</strong> two NYSE stocks. Ann Econ Stat 60:117–149<br />
Bauwens L, Hautsch N (2003) Dynamic Latent Factor Models for Intensity Processes, Discussion<br />
Paper 2003/103, CORE, Université Catholique de Louva<strong>in</strong><br />
Bauwens L, Veredas D (2004) The stochastic conditional duration model: a latent factor model<br />
for the analysis <strong>of</strong> f<strong>in</strong>ancial durations. J Econom 119:381–412<br />
Biais B, Hillion P, Spatt C (1995) An empirical analysis <strong>of</strong> the limit order book and the order flow<br />
<strong>in</strong> the Paris bourse. J F<strong>in</strong>anc 50:1655–1689<br />
Bisière C, Kamionka T (2000) Tim<strong>in</strong>g <strong>of</strong> orders, orders aggressiveness and the order book at the<br />
Paris bourse. Ann Econ Stat 60:43–72<br />
Bowsher CG (2002) Modell<strong>in</strong>g Security Markets <strong>in</strong> Cont<strong>in</strong>uous Time: Intensity based,<br />
Multivariate Po<strong>in</strong>t Process Models, Discussion Paper 2002-W22, Nuffield College, Oxford<br />
16 In our sett<strong>in</strong>g, an updat<strong>in</strong>g <strong>of</strong> the <strong>in</strong>formation set occurs whenever a new po<strong>in</strong>t <strong>of</strong> the pooled<br />
process arrives. A further extension would be to account for any changes <strong>of</strong> the order book.<br />
However this would considerably <strong>in</strong>crease the computational burden <strong>in</strong> our multivariate sett<strong>in</strong>g.<br />
17 This result supports the idea <strong>of</strong> Bauwens and Hautsch (2003) to model the underly<strong>in</strong>g market<br />
activity <strong>in</strong> terms <strong>of</strong> a latent autoregressive component which simultaneously affects all <strong>in</strong>dividual<br />
<strong>in</strong>tensity processes.