Fixed link
0
vote

Persistent Intraday Correlations Create Skews in Daily-Scale Distributions

T. J. Mazurek

posted on 19 September 2017

download (113 views, 57 download, 0 comments)

Abstract.  This study shows that the known extreme persistence of sign-correlations in intraday order-flow creates a positively skewed daily-scale distribution of bull-bear sentiment. The latter is defined by the difference of daily buyer minus seller trade totals.  This study then introduces the intraday price-mobility and shows that it also exhibits extreme persistence in sign-correlations of even greater strength than those of order-flow.  It then shows that the correlations of price-mobility also create a positively skewed daily-scale price-change distribution.  This work also introduces the normal inverse Gaussian representation (NIG) of empirical price-change and bull-bear sentiment distributions to examine their behaviors with increasing scales.  It shows that the shape and skew parameters of both distributions initially have parallel behaviors with increasing scales that sharply contrast with the behaviors of NIG independent identical distributions (NIG IIDs): the skews grow instead of decaying while the shape parameters decay much more slowly than those of NIG IIDs.  The extremely persistent intraday correlations create these non-IID scaling behaviors.

This revised paper corrects errors present in its prior version that was posted in July 2017.