This thesis investigates the distribution of the daily return in Korean stock market. The hypotheses adapted are the stable Paretian distribution and the mixture of normal distributions and the daily return data 30 stocks during 1983. 1. 4.-1988. 6. 30. period are used.
To test the hypothesis of the stable Paretian distribution, the parameters are estimated by Koutrouvelis' method. and the stability test and tail analysis are taken. The result rejects the stability and the probabilities near the tail are thinner than the stable Paretian.
The mixture of 3-7 normal distributions are best fitted and the one component distribution of all samples is the degenerate distribution.
Two kinds of the goodness-of-fit test indicate that the hypothesis of the mixture of normal distributions is better descriptive and so is the case of indices.
