This short paper seeks first to revisit the reasons for questioning today’s best practice, and then moves on to propose a practical alternative. A feature of the alternative proposed is that it naturally takes account of a fund’s individual characteristics, its regulatory environment and its risk preferences. The main difference between the proposed model and today’s is a more dynamic approach to asset allocation where asset allocation is driven by valuation (price of risk assets, risk premia) and wealth. While I think we can all agree (provided we can overcome our behavioural biases!) as to how we should respond to changing risk premia, there is no single answer as to how we should respond to shifting wealth. However, we can demonstrate that there are some unavoidable consequences to different wealthdriven utility functions.
When I started out in this business thirty eight years ago, investment management agreements could be encapsulated in a one and a half page letter and almost every mandate was a ‘balanced’ mandate investing across multiple asset classes. Mind you, the world was a lot simpler then and we limited our asset class definitions to gilts, UK and overseas equities, and property. The first and most important decision we made was how much to invest in our four asset classes and there was a real willingness to change the asset mix in a meaningful way.
All this began to change with the passing of ERISA (The Employee Retirement Income Security Act) in the United States in 1974. This turned out to be a powerful catalyst for the growth of the investment consulting industry which in turn set about redefining the best practice model to the standard we are familiar with today. The following five stage process is a short-hand description of that model:
• Conduct an asset/liability study to determine a strategic benchmark
• Construct an implementation plan around that benchmark – typically combining a mix of specialist managers in both active and passive strategies
• Conduct a manager search to fulfil the implementation plan
• Fund and monitor managers
• Repeat every three to five years
On the face of it, this is an appealing model built around some apparently common sense principles:
• No one can time markets, so a relatively static asset mix based around some long-run equilibrium return assumptions made good sense. Funds would rebalance periodically to the benchmark which, if you believed in a mean reversionary world, might actually add a little value as you would be selling the asset class that had performed well and buying the one that had lagged.
• No single manager could be the best at everything so instead managers would be chosen for their best of breed capabilities in specialist areas. The investment manager became a component supplier.
• While markets were not perfectly efficient, some areas were clearly more efficient than others – large cap US equities for example, compared to emerging market equities. It made sense therefore to use one’s alpha risk budget (active management risk) in areas where rewards were likely to be greatest and to buy cheap, passive beta market exposures elsewhere. Concepts of market efficiency and alpha and beta, which were previously confined to the ivory towers of academia, now migrated into the jargon of the industry.
It all seems very sensible. How could anyone find fault with this as a model? Unfortunately, it is riddled with problems.
1.Extract from 2010 Investment Perspectives