Scaling behavior in stock return distributions
<-- Return to the list
Date: 04-12-2004
Start Time:
5:30pm
End Time: 7:00pm
Speaker: Parameswaran Gopikrishnan, Goldman Sachs & Co.
Location: 412 Schapiro CEPSR, Davis Auditorium
ABSTRACT
We present results of some extensive empirical analysis on the nature of the
distribution of stock returns. By analyzing data on both intraday and longer
time scales, we find that the distribution function for stock returns is
consistent with a power-law functional form for a wide range of stocks and
indices, and also across different markets. For intra-day returns, our analysis
of 1000 largest US stocks shows that the exponent alpha characterizing this
power-law seems to have a rather "universal" value of alpha ~ 3; this value of
alpha allows for a finite variance, but higher moments seem divergent. Our
analysis of returns on larger time scales shows an interesting tiime-scaling
behavior, i.e., the distribution preserves its functional form for much larger
time horizons. We then relate the scaling behavior of this distribution to the
long-range correlations in volatility. If time permits, we will also discuss
some empirical results on the relationship between price fluctuations, market
activity and volume.