Flicking through Financial Adviser (FT Business) after it dropped through the letterbox this morning, I came across this article from Distribution Technology’s Actuarial Director, Danny Quant.
Having initially marvelled at his parents foresight, my eye was caught by the bold proclamation that “It will be reassuring for investors to know that the model used by their adviser has held up in the toughest markets for a generation”. Really?
Danny’s argument is that, because 9 out of 10 model portfolios in his three and a half year example performed better than the 1st percentile prediction, the “outcomes have been in line with expectations”.
I have two problems with this. First, I recall another eminent actuarial fellow telling me that stochastically modelled outcomes below the 5th percentile tended to be so erratic as to be untrustable, and therefore not to be given much legitimacy. Maybe that’s the reason general guidance on presentation of stochastically modelled outcomes suggests that output should show the extremes at the 5th & 95th percentile.
Second, Distribution Technology’s own Financial Plan output from a report I ran in September 2006 is in line with the guidance mentioned above, i.e. showing the 5th, 95th and 50th percentile only. As a result, any investor in possession of such a report probably won’t be able to share Danny’s confidence as they won’t have seen the 1st percentile predictions!
Instead, they’ll be comparing actual performance versus Danny’s 5th percentile, and half of the portfolios have performed below that mark. Indeed, portfolio 4 (which unfortunately is probably one of the more widely used – sitting in that “Balanced” range between portfolios 3 and 7) is just below the 1st percentile!
So, no, I don’t agree with you Danny, that “actual performance has been show to have occurred within the range of expected outcomes”, and I doubt many investors or indeed advisers will take that much comfort from this piece of nonsense!
No related posts.
Mike, I am no actuary, but my basic statistics suggests that the expectation would be that 9 out of 10 portfolios would do at least as well as the 10th percentile, not the 1st percentile. I’d expect 99 out of 100 to perform better than the 1st percentile. So I’d describe the performance of these funds as being below expectations (although I’m not sure that the original authors provided enough data in the article you reference to reach a definitive conclusions.)
Mike, you need to remember that Distribution Technology are outfit who created the FSA and the ABI’s pension calculators, which were then taken down from the websites when I (and others) pointed out to the FSA that the calculations appeared to be wrong and they were therefore in danger of being breach of at least one of the statutory objectives. At the time they said the calculations would be corrected and the calculator would then be put up on the sites again. If memory serves me correctly, this was over 2 years ago and nothing on the sites now.
DT were also the people who created the investment portfolio planner for Sinfonia, which drew an efficient frontier and then plotted the clients existing portfolio on the graph. Good idea, the only problem is that their system plotted all my model protfolios high left of their efficient frontier. More efficient than efficient? Go figure.
Makes you wonder if they can use a calculator to do basic mathematics. Perhaps this explains why they are laying off so many staff?