Abstract
This paper investigates the assumption of normality in the distribution of cash flow ratios of listed South African industrial companies over the period 1974 to 1993. Twenty eight cash flow ratios are defined and the distributions are investigated for the assumption of normality, usually assumed in parametric statistical procedures, both before and after outlier removal. The chi-square goodness-of-fit and the Kolmogorov-Smimov tests are used for this purpose. The tests indicate that the assumption of normality is consistently rejected on the industry level, both before and after outlier removal, and frequently rejected on the sectoral and yearly levels.