To begin at the beginning with actuarial science: it ain’t perfect. Most pension managers have probably already twigged this after years of data revisions by their appointed scheme actuary. In spite of any appearances to the contrary, actuaries are human, not divine. Yet they qualify and are consequently paid to divine, literally to predict, the future.
Here is an elegant alternative of the same sentiment, expressed by Wolfgang Lutz, one of the world’s leading demography experts and leader of the International Institute for Applied Systems Analysis’ Population Project: “There can never be a population projection without personal judgement. Even models largely based on past time-series are subject to a serious judgemental issue of whether to assume structural continuity or any alternative structure.” (1)
Actuarial science is never perfect because it’s a social rather than a natural science, like chemistry. Hypotheses do not need to be tested in the laboratory before they can be put into practice. Instead actuaries are perpetually working from field observations alone rather than commencing from universally agreed laws (little comfort to pension managers that investing too is a social science, in spite of all the mathematicians and former astral physicists supervising portfolios at asset management houses. The conclusion is that there is almost no certainty surrounding pension plans).
In spite of the considerable mathematical modelling to enhance the quality of forecasting and control for human error, actuarial work can never escape the social and demographic trends that deny it certainty. To be more precise, it is the unknowable duration and intensity of these trends which necessitates frequent recasting of assumptions. This is the ‘structural continuity’ of which Lutz et al write. In other words, has the world changed since the last actuarial survey?
An opening example is smoking in the Netherlands. By the 1950s, nine in 10 Dutchmen were smokers. Alders and de Beer (2) demonstrate how the expectation surprised the actuarial profession. By the time they factored in the adverse effects of smoking, the habit among men was in decline. By the late 1970s, just four in 10 Dutchmen smoked. The number of Dutch women who smoked rose in those later decades as equality manifested itself in this way. Broadly speaking, smoking was a stronger brake on the increase in male longevity than expected until the 1980s; but only acted in the same way from the 1980s for women.
Here is an international example far more dramatic in terms of unforeseen under and overestimation. By 1985, 12,000 Americans were dead or dying from AIDS-related diseases. Medical notice of the disease had only really begun four years earlier and even in the world’s most intellectually-resourced nation, understanding was slow to emerge (Apart from the ‘gay scourge’ myth, one pie-eyed suggestion was that the Egyptian pharaoh, Tutankhamun, who died in 1327BC, had released the virus in revenge for the transportation of some of his grave goods to the US for an exhibition some 3,300 years later (3) ).
By 1992, AIDS was the biggest cause of death for young US males. Between 1985 and 1996, half a million Americans died from AIDS-related diseases. During this long period, while medical knowledge was still formative and no antiretroviral drugs created, actuaries factored in the trend to their forecasts as best they possibly could.
It is no surprise to learn that they overestimated the future effect of AIDS. By the year 2000, an estimated one in hundred of the world’s sexually-active population were HIV positive.
However, scientifically-aware, wealthy places such as the US and Europe were already learning how to contain AIDS by means of more cautious sexual behaviour. If the actuaries were too conservative here, they could hardly be blamed for being over-optimistic elsewhere.
In South Africa, more than half the deaths of those aged under five are AIDS-related. It has cut life expectancy for children born in Botswana today from 70 to 30. Neither drought nor famine but AIDS is Africa’s biggest killer today.
The actuarial weighting of smoking and AIDS has not been highly accurate for understandable reasons. The dilemma for the forecasters of today will be familiar to any gambler ‘mid-session’: how to employ lessons learned and data gathered in the remainder of the session. Or as Alders and de Beer put it: “In order to be able to assess whether new forecasts are ‘really’ better than older ones, we need to know the reasons why the forecaster chose a specific method for a certain period. This information enables us to conclude whether a certain forecast was accurate, because the forecaster chose the right method for the right period, or whether the forecaster was just more lucky in one period than in another.”
This is detective work and it is unsurprisingly easier in more recent times, where greater explanation of working methods has been made, although Alders and de Beer counsel against the danger of favouring more recent surveys merely because they are better explained. This does not per se make them more accurate, merely more understandable.
Any pension fund that has employed a custodian bank or analytical firm to peruse every individual ‘buy’ and ‘sell’ decision made by their portfolio managers will also have exactly the same goal in mind outlined in the quote above.
The uncertainties that congregate around pension fund liabilities and assets begin to look more and more similar. Perhaps this explains why actuaries so commonly move into investment consulting.
1 Lutz, Goldstein and Prinz (1996), Alternative approaches to population projection
2 Alders and de Beer, Perspectives on Mortality Forecasting II
3 Steve Jones, Almost Like A Whale