Conditional asset allocation using prediction intervals to produce allocation decisions
- PDF / 238,641 Bytes
- 11 Pages / 599.005 x 767.008 pts Page_size
- 26 Downloads / 226 Views
Bevan Blair* is a Quantitative Analyst at Westdeutsche Landesbank Asset Management (UK) Ltd, where he specialises in asset allocation. Dr Blair obtained his PhD in finance from Lancaster University in the UK, where he studied the forecastability of equity market volatility. Prior to joining WestAM, he was a lecturer in finance at Lancaster University. *WestLB Asset Management, 25 Basinghall Street, London EC2V 5HA, UK. Tel: ⫹44 (0)20 7020 7412; Fax: ⫹44 (0)20 7020 7350; e-mail: [email protected]
Abstract Traditional conditional asset allocation involves using key past economic and financial data to produce forecasts of expected returns for the various asset classes involved in the asset allocation decision. Ordinarily these are point forecasts and little or no use is made of the prediction interval in the decision-making process. The main reason is that forecast models are misspecified, hence error distributions are unknown and so model-based prediction intervals provide spurious measures of forecast uncertainty. Alternative approaches, such as Bayesian densities and bootstrapping, provide non-parametric forecast distributions, which are not model based and so reduce the uncertainty about the nature of the prediction interval. Here we investigate various different prediction intervals and attempt to quantify biases between model-based and non-model-based solutions. Asset allocation decisions based on prediction densities are investigated, and it is found that knowledge of a more accurate prediction interval is economically meaningful in an asset allocation context. Keywords: asset allocation; bootstrapping; prediction densities; kernel regression
Introduction Asset allocation can be loosely defined as the problem of determining a set of decision rules that enables the ‘asset manager’ to determine what proportion of their portfolio wealth should be placed in the relevant asset classes. In general, an active ‘asset manager’ will have a pre-determined benchmark with which decisions are compared. Restrictions are normally placed on how much the manager can deviate from this benchmark. For instance, a simple benchmark may be one which is
䉷 Henry Stewart Publications 1470-8272 (2002)
composed of 50 per cent equities and 50 per cent bonds. The manager may be allowed to deviate from this mixture by up to 10 per cent, implying that the mixture can be varied from 40 per cent equities to 60 per cent equities. The philosophy adopted in this paper is that the most fundamental part of determining what this mixture should be lies in being able to forecast asset returns accurately. For instance, if equities are forecast to have a higher risk adjusted return than bonds, then the asset manager should place more of the wealth
Vol. 2, 4, 325-335
Journal of Asset Management
325
Blair
of the portfolio in equities. Hence great emphasis is placed on methodologies devoted to providing forecasts of asset returns. Most forecasting models use key historic economic and financial data to forecast asset returns. Asset allocation u
Data Loading...
