Whatever the actuaries’ unique selling points (hopefully plural) have been in the past, I don’t think they're that well recognised outside the profession any longer. The concept that we predict the future is just wrong. The Carlsberg actuary would probably provide the best advice in the world.

However much one might guess, no-one on Earth actually “knows” the future, the key questions being “when”, “in which direction”, “how far” and “for how long”. As a long-term indicator, “today's” conditions are useless. While I have heard some actuaries confidently assert that “today” is the only evidence we have, that is simply not true because there is far more. Looking back at almost any interesting financial time series, we have seen huge variability. Ignoring that variability excludes virtually all of the data, leading me to conclude that “spot” can't be superior to “whole population”. If actuaries think it is their duty to assume that future experience must be worse, I find that deeply depressing.

For final salary pension schemes, actuaries should be emphasising risk PLUS reward, not just the former. Indeed, that is exactly what UK actuaries used to do for pension fund assets (the DCF approach). An essential point to bear in mind is that one can't invest assets “short” and “long” at the same time. So long as they're aware of a realistic timeframe, to be discussed with trustees and sponsors, actuaries can help to choose appropriate assets.

Short-termism is commonly seen as a systemic problem in areas other than pensions and then ignored. Further, short-term optima won't deliver the long-term optimum because path dependence is crucial (a topic upon which I intend to expand). In the past, the actuary’s unique selling point was specialising in long-term projections, with few others recognised as doing that. We must stop being fixated by “now” and focus upon strategy rather than tactics. For truly long-term entities, I believe that smoothing is an actuary's tool which can be used far more than has been common or is currently permitted in the UK. Long-term entities can use the time to soak up volatility, which is then partially removed as a problem.

If we really had a crystal ball, we would not need to make assumptions about the future. In fact, we don’t have even good foresight of the future so that we must make assumptions. The very first question must be the extent, if any, to which we think it wise to err on the side of caution. While I have seen assumptions adopted which were clearly beyond optimistic, such an approach is not advocated.

However a “best estimate” is defined, being cautious beyond that level clearly introduces some bias. Its impact can only be assessed with hindsight. For long-term funding purposes, I believe an element of bias must just be accepted but let's at least try to retain some control.