City Research Online

Abstract

The problem of determining the optimal asset allocation strategies for a non-profit life company is approached from a rational decision- making framework. Initially, a number of methods for analysing investment risk are discussed, from which utility theory is felt to be the most appropriate. Stochastic simulation and numerical optimization methods are employed in order to allow more realistic assumptions to be used in these decision models.

The multiperiod consumption of dividends is dealt with by considering the expected utility of accumulated dividends, or pay-outs. At first, the case of an open fund is investigated in a static asset allocation framework. In general, the results produced are quite intuitive. At low levels of risk tolerance, the optimal portfolios seem reasonably matched in relation to the liabilities. As the risk tolerance level increases, the preference for matching is seen to reduce. If pay-outs are measured in real terms, greater proportions tend to be invested in the real asset classes. From a mean-variance perspective, the utility maximizing portfolios generally appear to be efficient. However, imposing insolvency constraints on the objective function can have the effect of shifting some of these portfolios away from the efficient frontier.

In the case of a closed fund, dynamic asset allocation strategies are investigated. Due to the restrictive assumptions it requires, the possibility of applying dynamic programming in this situation is rejected. Instead, it is proposed that the asset proportions be made functions of the duration of the liabilities, so that the expected utility may be maximized in respect of these function parameters. Overall, this appears to produce reasonable results, although the occasional emergence of less intuitive strategies leaves further scope for refining the treatment of multiperiod consumption.

Publication Type:	Thesis (Doctoral)
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HG Finance
Departments:	Bayes Business School > Actuarial Science & Insurance Bayes Business School > Bayes Business School Doctoral Theses Doctoral Theses

Export

Downloads