[Skip to content]

Sign up for our daily newsletter
The Actuary The magazine of the Institute & Faculty of Actuaries

Spurious accuracy and the curious actuary

THIS MONTH we haveincluded an articleby Robert Shapiroon the dangers ofplacing too much relianceon sophisticated financialmodelling tools. It mightstrike you as vaguely ironicthat there appears alongsideit an article by Shyam Mehtaadvocating the use of modernasset-pricing theory,which by all accounts fallsinto the category ‘sophisticatedfinancial modellingtools’.Publishing these two articlesside by side brings intofocus one of the dilemmasfacing many of us in our daily working lives. With abit of time and thought, most of us could come upwith a complex, theoretically sound way of estimatinga price, a future claim cost, or an investment allocation,but time constraints often mean that we have tocome up with ‘something sensible’ instead.I imagine that it is very satisfying to solve a problemusing a complex yet beautiful approach that nobodyelse has thought of before. (I say ‘I imagine’, because Ican’t say I’ve done so myself.) And so it should be.Without the ingenuity of our predecessors, therewould be no actuarial profession, and I’d probably bethe editor of Practical Cheesemongery magazine.However, as Robert Shapiro points out, there is littlevalue in finding solutions that rely on too manyassumptions holding true. We must keep in mind thefact that our assumptions are almost certainly notgoing to be borne out – in other words, the only certaintyis that we’ll be wrong, no matter how complexa model we use.So what is the right balance? Robert has gone someway towards answering that question in his article,where he points out the human risks involved in theuse of models. As long as we can understand the traps,we have no excuse for falling into them.We could also learn a thing or two from the work ofbehavioural psychologists. Well-documented flaws ofhuman decision-making include:°ª Anchoring, ie latching onto an idea or fact andusing it as a reference point for future decisions.°ª Confirmation bias, ie searching for, treating kindly,and being overly impressed by information that confirms your initial impressions or preferences.°ª Choosing the middle option from a range of alternatives.°ª Choosing the course of least resistance.But hang on, we’re members of the actuarial profession.Our reputation for being logical is the basis formost actuarial ‘jokes’. Surely we can’t be susceptible tothe same types of irrational decision-making error aseveryone else?Oh, yes we can. And that may be a good reason forusing sophisticated financial modelling tools such asmodern asset-pricing theory. If we restrict ourselves tothe use of ‘common sense’ methods, we risk fallingfoul of human irrationality.As with most things, there is no universal solution.My own preference is to use a variety of methods totackle a problem, and use the differences between theresults to help challenge the assumptions of eachapproach, and investigate which is more suitableunder the particular circumstances. In the end, actuaries,not formulae, make the judgements, and that’swhy we’re here.