Semivarians, ett användbart riskmått eller endast ett överflöd?

Sammanfattning: Since Markowitz presented the mean-variance model as a way of putting together a financial portfolio, variance has been the established measure of risk. Even though the measures of downside risk hasn’t gained the same attention as variance, it has been used in research for some time. Semivariance, best described as the risk of a portfolio’s return falling below a set target, was introduced as a risk measure by Markowitz during the same decennium as variance, but did not receive the same popularity, partly because of the problems of effective calculations of it back at that time. Later LPM (Lower Partial Movement) was introduced as a more general measure of downside risk, as it can be adjusted to different risk preferences, one of them turning the measure into semivariance. Earlier research has focused on comparing mean-variance to mean-semivariance and mean-LPM in holding periods no shorter than 24 months (and often much longer) and only one estimation period per holding period. Furthermore, chosen assets have always had enough data for estimation, and therefore seem to have been picked accordingly. This essay does not only compare portfolios based on different risk-measures, but also set available assets as those from the index OMXS30. The author uses the risk measures in a mean-risk model to i) compare the portfolios of downside risk measure to those of variance and ii) to see if any of these method gives better results than when investing in the index. Unlike previous research on the subject, estimation period is varied between 12, 24, 36, and 48 months and the holding period is set to 6 months (20 periods), giving the essay a more realistic approach to financial investment. The results show that even if there are noticeable differences in return and risk adjusted return between portfolios and between portfolios and index, none of them can be described as statistical significant enough to reject null hypothesizes that there are not any differences. Therefore the conclusion of this essay is that neither of the used downside risk-portfolios can be proved to perform better than the corresponding variance portfolios and no portfolio performs than return of investment in index.