Optimal robust and consistent active implementation of a pension fund's benchmark investment strategy

  • PDF / 203,430 Bytes
  • 12 Pages / 595 x 765 pts Page_size
  • 56 Downloads / 159 Views

DOWNLOAD

REPORT


Tim van Hest* is a part-time PhD student at Tilburg University, The Netherlands, and the Head of credit risk modelling ABB department of Rabobank Netherlands. His current research interests include robust portfolio optimisation and credit risk modelling.

Anja De Waegenaere is an associate professor of actuarial sciences and accountancy at Tilburg University. She obtained her PhD in mathematics in 1993. Her current research interests include actuarial modelling, robust portfolio optimisation, longevity risk and fair valuation. *Department of Econometrics and OR, Tilburg University, Tilburg 5000 LE, The Netherlands. Tel: þ 31 13 466 29 13; Fax: þ 31 13 466 32 80; E-mail: [email protected]

Abstract The benchmark investment strategy of a pension fund typically consists of a number of benchmark categories, each of which is assigned a weight in the overall investment budget. In this paper we assume that the benchmark strategy is given, and determine a model for its optimal active implementation. Active implementation involves a number of investment managers each of whom are assigned a specific benchmark category. We present a mean–variance approach to determine, for each investment manager, the optimal budget as well as the fraction of that budget that can be used for deviations from the benchmark. The emphasis is on robustness of the optimal allocation with respect to parameter misestimation, and on consistency in terms of risk-return preferences between active implementation and benchmark investment strategy. Journal of Asset Management (2007) 8, 176–187. doi:10.1057/palgrave.jam.2250072 Keywords: asset liability management, pension funds, active investment decisions, robust optimisation

Introduction In the context of asset liability management (ALM), two types of investment decisions can be distinguished: 1. The choice of the ALM benchmark investment strategy: passive risk management. 2. Tactical and operational investment decisions: active risk management. The ALM benchmark yields the allocation of the total investment budget over the different benchmark categories such as equity, bonds, real estate, etc. Usually, this allocation is given

176

in (monetary) weights. Since for a pension fund the match between assets and liabilities is of utmost importance, benchmark weights are typically determined by means of largescale simulation models, see for example Boender (1997). Practical implementation of the benchmark investment strategy usually involves several investment managers. Each of these investment managers is assigned a specific benchmark category, and is responsible for part of the total investment budget. In order to be able to benefit from short-term opportunities, investment managers would typically be allowed to spend

Journal of Asset Management Vol. 8, 3, 176–187 & 2007 Palgrave Macmillan Ltd, 1470-8272 $30.00 www.palgrave-journals.com/jam

Benchmark investment strategy of a pension fund

a fraction of the budget assigned to them on active management, that is deviations from the benchmark are allowed. The