WHEN pension funds are performing well, few clients question how this performance has been generated. Understanding how fund managers generate performance can be complicated and the calculation methods can be obscure. Most clients only want to know whether the fund has performed well, and whether the fund manager have vindicated the decision to appoint him. As a way to monitor a fund manager, his performance is often ranked alongside competitors. Good relative performance implies a correct decision, poor relative performance implies a poor decision, according to market conventions. Last week, I noted that different funds had different objectives and were therefore likely to have different asset weightings. Therefore direct performance comparisons were often difficult. But even in the case where different funds have similar objectives and similar asset weightings, the comparison is not necessarily straightforward. Performance in a single year must surely be compared - but so should the method by which this was achieved. Different managers will have different styles which may imply greater or lesser risk exposures. Good performance may be the objective, but this needs to be delivered consistently. Volatile performance may not meet the clients' objectives. For example, two fund managers may deliver the same performance, but their decisions during the year could have been quite different. Manager A could have achieved his performance by making one big decision during the year, while Manager B may have alternatively built his performance through a succession of small incremental decisions. In terms of risk, Manager A has a higher profile. If the one decision had not paid off, the fund performance would not have been so good. Whereas Manager B did not rely on a single decision and diversified the performance over a range of decisions. In this example, Manager A could argue that his approach is no more risky as he was confident of his decision, but this overlooks the risk of his decision being either mis-timed or inappropriate. When analysing the performance statistics in greater detail one is often presented with terms such as ''asset allocation performance'' and ''stock selection performance''. In broad terms, asset allocation relates to the manager's ability to judge when to move investments from one market to another and thereby add value; while stock selection refers to the fund manager's ability to choose stocks that out-perform a market index. I have often heard in Hong Kong that one of these aspects is more important. This is fundamentally untrue. Different managers will have different skills and philosophies. The decisions a fund manager takes should exploit his strengths. Some are good stock pickers, some are good country analysts, and some are a blend of both. If all the major investment markets deliver similar performance over a given period, then the asset allocation decision will seem unimportant. Here the ability to choose high-performing stocks will be more influential. Alternatively, when different assets provide a range of different performances, the asset allocation decision can be more influential. For example, if bonds outperform equities in a given year it will appear that the decision to move from equities to bonds added the greatest performance. Much depends on the manager's philosophy, style, how often the manager reviews asset weightings, and how many stocks he holds in each country. Does the manager have a view based upon the consensus of the company's analysts and fund managers or does each fund manager follow their own views? How does the fund manager manage the liquidity requirements? Last year, a manager who did not re-balance his portfolio regularly would have increased his exposure to the markets performing well and reduced his weightings in the markets that were under-performing relatively. In 1993, markets tended to move upwards, and therefore a manager who did not re-balance would have outperformed a manager who did re-balance. This year the situation is likely to be quite different. As the stock markets become more volatile and the trend is less obvious, re-balancing is likely to be more productive. When a fund is re-balanced the manager tends to sell those assets that have increased in value and buy those that have decreased in value. In volatile conditions, this rotational effect can generate excess returns. These are just examples of how performance analysis requires more detailed consideration. The key is to communicate with the fund manager to gain a better understanding of how the performance was generated. What were the major decisions and how did they affect the total performance? What are the manager's strengths and how does he expect to add value in the future? What internal methods does he use to monitor consistency? Do not be put off by the technical aspects of performance measurement and try to relate the fund manager's decisions to his performance. Mark Konyn is associate director at Indosuez Asset Management Asia.