EMPIRICAL COMPARISONS OF DISTRIBUTIONAL MODELS FOR STOCK INDEX RETURNS

Several market studies have shown that the assumption of normally distributed log stock returns is not valid for individual stocks. A number of studies have suggested alternative finite-variance distributions and have shown that they provide better descriptions of observed stock returns than does th...

Full description

Saved in:
Bibliographic Details
Published in:Journal of business finance & accounting 1990-07, Vol.17 (3), p.451-459
Main Authors: Gray, J. Brian, French, Dan W.
Format: Article
Language:English
Subjects:
Mathematical models
Normal distribution
Securities markets
Security prices
Statistical analysis
Stock prices
Stock returns
Studies
Theory
Volatility
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Several market studies have shown that the assumption of normally distributed log stock returns is not valid for individual stocks. A number of studies have suggested alternative finite-variance distributions and have shown that they provide better descriptions of observed stock returns than does the normal distribution. The ability of the normal distribution to model the log price of returns from the Standard & Poor's 500 Composite Index is examined, and its performance is compared to the performance of 3 alternative finite-variance distributions previously proposed for individual stocks. The 3 alternative distributions - the scaled-t, the logistic, and the exponential power distributions (EPD) - demonstrate a greater ability to model log stock index returns. Of the 3 models considered, the EPD appears to offer the superior fit.
ISSN:0306-686X
1468-5957
DOI:10.1111/j.1468-5957.1990.tb01197.x