ThreeBearsBalanced11Feb2018_UseEvenSmaller-Edited
contact
sitemap
origins
[long_equity_returns] [origins]

Towards the end of the 1970s, the “new style” of performance assessment concentrated upon short-term measurements of asset value movements and many trustees felt they had to take account of the numbers they were being sold. What struck me as bizarre was that this topic could then be immediately followed, or preceded, by a discussion of the results of the actuarial valuation. For the latter, at least back then, the actuary would have been emphasising that the assets were to be looked at over the long-term, with short-term fluctuations being virtually irrelevant.

As a simple example, suppose that two consecutive TWRs (market related time-weighted returns, the statistic normally published) are 20% pa and (10% pa), respectively. Ignoring compounding, the average return is 5% pa. The first year's statistic was a very poor indicator of what was to come.

Worse, I felt the results were positively misleading. The essential point is that trustees normally had a very much longer planning timeframe than other investors. If the trustees can't be sure that they have that long ahead of them for freedom of investment manoeuvre, then they probably shouldn't be investing in equities (or other real assets) at all (see here).

A sports analogy may be helpful. In advising trustees, I think investment managers tend to play basketball, very close to the net, which is all they can see. Trustees need advisors who play rugby, with the posts in the far distance, conversion being the eventual aim.

It is commonly stated that market values are the best estimate of “value”, with which I totally disagree for the long term. Markets are far from perfect, with many participants having far less information than others. More important, markets are about transactions carried out by traders who have little interest in, or knowledge of, the long-term. The statistics presented here, and others on “long-term inflation”, should make that clear. Indeed, if last year's (or yesterday's) was the best long-term indicator available, why is today's any better?