Asset Allocation under Threat of a Crash推荐.pdfVIP

CHAPTER 67 asset allocation under threat of a crash In this Chapter. . . • how to allocate your money to an investment when you are worried about a market crash 67.1 INTRODUCTION The following by Ralf Korn1 and myself was originally published in the International Journal of Theoretical and Applied Finance 5 171–188 (2002). We consider the determination of optimal portfolios under the threat of a crash. Our main assumption is that upper bounds for both the crash size and the number of crashes occurring before the time horizon are given. We make no probabilistic assumption on the crash size or the crash time distribution. The optimal strategies in the presence of a crash possibility are characterized by a balance problem between insurance against the crash and good performance in the crash-free situation. Explicit solutions for the log-utility case are given. Our main finding is that constant portfolios are no longer optimal ones. It is well known that the classical lognormal stock/Black–Scholes model is not able to explain large jumps in stock prices appearing in real-world security markets. In particular, it does not contain the possibility of a crash of the stock prices. Although these crashes are rare events, they do occur in real life; the October 1987 crash being the largest in recent memory. Therefore, there is a long tradition of modeling jumps in stock prices. A seemingly obvious candidate of a class of suitable stochastic processes is that of jump-diffusion processes (see Merton (1976) for an early reference). However, jump-diffusion approaches only lead to strategies that hedge a crash situation in the mean which is no real protection against the consequences of a jump at all. In particular, an investor following such a strategy will suffer large losses during a crash. As a contrast, by implementing the strategy that we will propose an investor need not be in f