Investigation of portfolio analysis under the stable distribution and an application at ISE

in ABSTRACT In this study distributions of stock returns, fitting of non-normal stable distributions to stock returns and stable portfolio analysis are examined. Because of departures of stock returns from normal distribution, it is focused on modelling of return series with alternative distributions and searceh for appropriateness of non-normal stable distributions to stock returns. Since the distribution of stock returns is not appropriate to normal distribution, Markowitz's mean-variance model is not valid. Hence stable portfolio analysis that has been formed under stable distribution assumption gives more more accurate results about expected return and risk of investor's portfolios. In study primarily Portfolio Theory, Markowitz's mean-variance portfolio analysis, statistical and distributional properties of stock returns, alternative distributions of stock returns, stable (non-normal) distributions and stable portfolio analysis are examined. Appropriateness of daily returns of 15 stocks that are found in ISE 100 index to stable distributions is examined. Under assumptions that stock returns have distributed stable and normal, structure of portfolios that have been formed according to mean-variance analysis and stable portfolio analysis are different from each other. Under the both assumption, daily looses of these portfolios are calculated by value-at-risk method. It is found that risk of stable portfolio is much less than the risk of portfolio which is formed under normal assumption.