All this week we are running a Guide to Active and Passive Investing to help you, the investor, make smart choices for your portfolio.
Veteran investor and founder of the Vanguard Group Jack Bogle's least-appreciated great insight is the quiet virtue of relative predictability.
Conventional risk statistics measure absolute predictability. Cash is regarded as the riskless investment because its price is completely predictable from one day to the next. One pound today sells for one pound tomorrow.
In contrast, the price of a bond that fluctuates with changes in interest rates cannot be known in advance, and the price of a stock that fluctuates widely is more uncertain yet. Stocks are absolutely less predictable than bonds, which are absolutely less predictable than cash. That is how risk is described.
Relative predictability – a phrase that Bogle mentioned in passing to me, in a conversation last week – is something altogether different. The relatively predictable asset is the asset that behaves as expected, given the performance of the financial markets. Relative predictability has nothing to do with the usual statistics that measure absolute levels of risk. It cannot be measured by standard deviation. Relative predictability is unrelated to volatility. In its realm, cash can be highly risky and stocks can be fully safe.
Consider the Reserve Primary Fund. On September 16, 2008, that money market fund cut its net asset value from $1 to $0.97, as it got caught holding Lehman paper when that bank was seized by the government. By absolute measures, that 3% daily decline was nothing much. By relative measures, of course, the loss was enormous. Nobody expected a cash fund to break the buck; the fund immediately was swamped with redemption requests, such that the fund and the firm was eventually liquidated.
Meanwhile, a Vanguard US Equity tracker dropped a cool 37% in 2008, but steadily gained new assets. Each and every month of 2008, including October when it lost 17.6%, this Vanguard index tracker sold more new shares than it paid out in redemptions.
The reason was relative predictability. The fund's absolute returns were anything but predictable or desired but its relative performance was as reliable as clockwork. Whatever the index did, the fund did. It gave shareholders exactly what they had purchased.
If those examples haven't convinced you that a fund investor's risk meter ticks to something other than absolute results, then consider PIMCO Total Return. Once the world's largest fund, PIMCO Total Return has shed $180 billion in assets during the past 24 months, even as it has posted steadily positive returns.
Press reports about a corporate power struggle and performance that slightly lagged the category average were all it took to topple the king. The coup came from the loss of investors’ trust rather than of principal.
By now, you might be muttering, "Has this guy never heard of tracking error?" Well yes, I have. Tracking error, which measures how far a fund's performance deviates from that of its benchmark, assesses one aspect of relative predictability. But there's much more to the concept than that. Critically, realised tracking error does not incorporate expectations – the beliefs and hopes that investors place in their funds.
Tracking error doesn't explain why PIMCO Total Return became shunned.
Relative predictability, not tracking error, fully explains how Vanguard was constructed. Offering index funds obviously improves relative predictability.
But so do many other Vanguard policies. Using multiple managers rather than a single manager improves relative predictability. Running plain-vanilla bond funds that eschew exotic bonds and complex strategies improves relative predictability. Creating funds with tightly defined mandates improves relative predictability. Temporarily closing popular funds that are attracting hot money makes for a savvier investor base, which improves relative predictability.
As I've written elsewhere, textbook definitions of investment risk are woefully inadequate. No greater example exists than the lack of discussion about relative predictability.
