PORTFOLIO PROBLEM

• How do we construct, plan and manage the relationship between assets and liabilities over time?

• How do we optimise the allocation of resources between investment research and strategy and client portfolio management?

A wealth manager's research and strategy should be focussed on the management of excess risk and return at a point in time and on applying efficient selection of securities and markets (allocation) to portfolios in general.

It is not efficient to have to focus on risk/return allocation, strategy and security selection at the same time as multiple portfolio objectives.  

THE ONE PERIOD PROBLEM

If we were only looking at simple objectives over short periods of time, for example a year, the portfolio problem would be much simpler.

In this context, we are only looking at the relationship between asset allocation, risk/return and simple liability objectives.

Within modern portfolio theory, the risk/return equation is one which looks at the most efficient combination of assets to meet the client’s return objectives, return being synonymous with both liabilities and performance.

Because the point in time liability is more often than not small relative to the overall portfolio, the rationale of the risk/return relationship in liability space is not stretched. Within a one period problem we do not need to look at longer term issues within portfolio construction.

Looking at the one period problem within the traditional portfolio context (where the portfolio is structured so that income is met by interest and dividend yield), this is also not a problem. We are only looking at one period and any capital expenditure is easily accommodated. Even here, there is a risk/return rationale and objectives are met.

The one period problem is not a problem and it is also not the reality most investors face. 

THE MULTIPLE PERIOD PROBLEM

It gets more complicated once we start to increase the time horizon since the size and timing of income and capital inflows to and from the portfolio become more complex.

In the context of modern portfolio theory, liabilities are not an input into portfolio structure. At this point the limitations of the MVO (mean variance optimisation) structure are well apparent. All that can be done to reflect liabilities is to determine the average return needed by the client’s financial needs and use this to structure the portfolio. However, the return profile is only an average and provides no information regarding the size and timing of income and capital needs. The limitations of MVO in an asset liability modelling and management context are discussed in the Technical documents.

In the traditional portfolio construction approach, it is impossible to incorporate all this change into portfolio structure. The portfolio will be moving in all directions if you attempt to do so. There are frankly too many variables.

What the traditional approach will do is to have a limited time horizon, say a three year horizon where the portfolio is structured to meet the yield/growth objective and known liabilities are planned for. As needs change and liabilities enter the equation the portfolio manager has to react and make changes within this 3 year window.

The multi period problem is closer to reality and may be manageable.  But is it the optimal situation and does it meet the individual investor's objectives?

THE LONG TERM PROBLEM

Not only does the complexity of the relationship between asset allocation and liability profiles increase over the very long term, but we are faced with a number of other decisions that we were not faced with in the single and multi period problems.

In this context we also find that we may need to understand the disposition of all current and future assets and, all current and future financial needs.

We then start to have to assess the effect of current demands on the ability of assets to meet needs over time.