ension fund investment is extraordinarily difficult. Trustees are faced with two variables. How much will be needed to provide an adequate pension for all the claims of the beneficiaries? The second variable is consequent on the first: can the investments in the pension fund provide this adequate return, going forward?

The first question can be answered through objective assessment. The plethora of different entitlements and life expectations of the beneficiaries can be expressed as a single discounted present liability.

The difficulty arises in the second variable: how to invest the assets to provide the necessary returns? Some propose a theoretical answer to this question and create a benchmark to be followed. Therein lies the difficulty for the pension fund, since this is, in fact, a practical question, not a theoretical one. Whatever the probabilities, once a single collection of assets is gathered together as an investment portfolio, the actual performance will bear little correlation to the assumptions which moulded it.

Instead, it should naturally fall to the investment management industry to determine the correct, practical spread of investments, since their area of expertise is exactly centred on the assessment of future specific risk. Alas, repeated failure to deliver the goods has sapped the confidence of the investment houses. It has become horribly obvious that the role of investment manager requires the ability to foretell the future: a manifest impossibility. However, if someone can be found to provide a benchmark, it makes available the defence of all middle managers who serve a tyrant: "I'm only obeying orders". The tragedy of the investment industry is that they have long abdicated taking responsibility for their actions and have shown over many years and in many circumstances that, collectively, they represent the cork bobbing about on the maelstrom of average investment returns.

At Ruffer, we believe that if there is no hiding place for the pensioner in stormy markets, why should there be for the fund manager? This is why we target absolute returns, not comparative performance based on indices and benchmarks. We aim to return twice the Bank of England base rate (after fees) and never to lose money in any rolling 12 month period.

Many of our pension fund clients have eschewed their benchmarks completely for our approach. Others, usually larger funds, find Ruffer's returns conveniently uncorrelated to their traditional benchmark returns and employ us in a 'satellite' capacity to diversify the market risk implicit in their benchmarks.