In this study we analyze the Standard and Poor's 500 index data of the New
York Stock Exchange far more than 32 years. Using a simple random walk mode
l we demonstrate that the proper variable to look at is the logarithmic ret
urn. In the statistical analysis we have done fittings to the Livy distribu
tion using either the index data as such or pre-processing it with ARCH, GA
RCH or IGARCH methods, which tend to remove the time-dependent variance. Fo
r short times the truncated Levy distribution is found to fit the data quit
e well. Since this is not a stable distribution, the sealing behavior obser
ved for short times should brake down for longer times. We demonstrate that
the characteristic time where this cross-over starts is of the order of on
e day. (C) 1999 Elsevier Science B.V. All rights reserved.