An Event-Based Approach for Dynamic Volume Return Relationships of DAX Companies

Download or read book An Event-Based Approach for Dynamic Volume Return Relationships of DAX Companies written by Roland Mestel and published by . This book was released on 2008 with total page 22 pages. Available in PDF, EPUB and Kindle. Book excerpt: Several recent papers report market returns and returns of individual securities to carry informational content about future trading volume of individual stocks. In addition some authors identify significant abnormal returns of stocks that currently exhibit high volume.This paper conducts a comprehensive empirical examination of the implications of the above outlined findings for the German stock market, concretely for the most liquid stocks.We do this by applying event based methodology, which roughly means, that the stock market as a whole and individual securities themselves are clustered into states of volume and returns. For each date we identify the prevailing level of returns and volume, which allows us to categorize days into different events. Strictly peaking events in our sense are not rarely distributed over the data sample, however each day marks an event in terms of signalling a certain state of the stock market to market participants.Dependencies between market-wide/security-specific returns and volume are separately analyzed for each cluster. We examine the performance of individual stocks in each cluster and take the whole market as a benchmark, which allows to statistically check for abnormal volume and returns.Furthermore we apply vector-autoregressive models, that do not only capture dynamic structures within market data, but also allow to check for temporal causalities between volume and return. Again for each cluster, we analyze Granger-causalities between volume and market/security returns.Our preliminary results indicate only weak relations between volume and returns, however, with our methodology and possibly due to the specific data set of the most liquid German stocks, we find little statistical significance.