[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Institute Sessional Meeting

Practical risk management for equity portfolio managers

The above-titled paper from Giles Heywood, John Marsland, and Gordon Morrison generated a good debate at the Institute sessional meeting on 28 April 2003.

It represented a breakthrough in technology: not necessarily for the content of the paper, but for the fact that it was the first Institute paper to be produced in colour. On the face of it this might seem a small step forward, but it was important because a central theme of the paper, fully supported by the contributions from the floor, was the need to be able to communicate risk in clearly understood ways. Most of the profession may be unfamiliar with dendrograms and heat maps, but they presented clear illustrations of risk characteristics.

The debate highlighted a difficult balance. These techniques mean that the risk characteristics can be much more easily communicated to the fund managers and to external parties: they are a huge step forward from single-number tracking errors or factor contributions that lack any obvious economic meaning. However, the asset management risk specialists observed that they still conceal a degree of over-simplification. The balance of the debate suggested that the authors have managed to find a sensible path that is detailed enough to be useful yet simple enough to communicate.

The paper contained a taxonomy of modelling methods. A few contributors would have preferred the terminology to be tightened up in places. (A distinction could be drawn between models, methods, and techniques and where some methods are sub-sets of broader approaches that could have been more explicitly spelt out.) That said, it was generally agreed that this would provide a handy reference point where all the main approaches are explored, particularly for actuaries not involved in this area of work.

Greater clarity of description of risk should enable much more effective applications in risk-budgeting and risk-management rather than merely risk-monitoring. This will be a real test in determining whether risk departments in fund managers remain a backroom function or integrated into the investment process.

Several of the contributors suggested that further work was needed to link up this type of analysis of the assets to a closer understanding of the liabilities. There was also a general concern expressed that models had not necessarily worked as well as they should in the last few years. At one extreme, it was noted that just because two dice are thrown and the result is a double one does not mean that the model failed to work; however, experience has recently thrown up a few too many two-standard-deviation events. This led to a familiar discussion of the trade-offs between mathematically manageable models and more accurate fatter-tail models.

Overall the debate was particularly good, not least because there were fewer prepared speeches and entrenched positions and a greater willingness of practitioners to discuss real practical problems. The authors should be particularly thanked for this measure of the success of the paper.