CMStatistics 2017: Start Registration
View Submission - CFE
A1803
Title: Performance of volatility maximization strategies Authors:  Jan Vecer - Charles University, MFF (Czech Republic) [presenting]
Abstract: Volatility in finance is traditionally regarded negatively. The classical risk measures are increasing functions of volatility. Thus a traditional portfolio management tries to completely eliminate or minimize the risk associated with volatility. However, volatility is just a measure of dispersion, so a high volatility can result in both substantially negative or substantially positive outcomes. We show that the volatility maximization portfolio is maximizing the costs associated with insuring an actively traded portfolio. These contracts are called options on a traded account and their existence in practice is very limited precisely for the high price. Option pricing theory uses risk neutral measures (also called martingale measures) for pricing in contrast to using the real measure. The martingale measures determine the replication costs of the scenarios insured by the option. These replication costs can differ substantially from the real expectation and thus the risk neutral measures and the real measures may exhibit some discrepancies. In fact, they differ most on the scenarios associated with high volatility. In particular, the substantially negative outcomes for high volatility strategies are much less likely to happen in reality in comparison to their replicating costs. We illustrate the performance of the high volatility strategies on the portfolio of major world currencies and a portfolio of stocks from NASDAQ 100.