Article ID: | iaor19932122 |
Country: | United States |
Volume: | 38 |
Issue: | 11 |
Start Page Number: | 1586 |
End Page Number: | 1609 |
Publication Date: | Nov 1992 |
Journal: | Management Science |
Authors: | Turner Andrew L., Weigel Eric J. |
Keywords: | financial, statistics: empirical, time series & forecasting methods |
This paper examines the daily return variability of the S&P 500 and the Dow Jones indices over the 1928-1989 period. The authors use the traditional close-to-close standard deviation of returns, two alternative estimators incorporating the daily high and low of the index, and a robust estimator to measure the volatility of stock index returns. The 1980s were the third most volatile decade behind the 1920s and 30s. To a large extent, this was caused by the anomalous behavior of the fourth quarter of 1987. Returns in the 1980s had far more skewness and kurtosis than in any other decade studied; these results were not entirely due to 1987, as returns in 1988 and 1989 had large measures of both skewness and kurtosis. The frequency of extreme-return events increased in the 1980s, but was still dramatically less than the 1920, and 30s. When extreme negative days occurred in the 1980s the losses tended to be more severe than in the previous four decades. Extreme-return days are preceded by significant losses and are intertemporally clustered. There is no evidence of short-term market reversals after either positive or negative jumps in stock index returns.