Look forward, not backward, with Warren Buffett

Guess who gets the last laugh. / CBS News

(MoneyWatch) So-called hindsight bias is the inclination to see events that have already occurred as being more predictable than they were before they took place. It leads us to believe that even events that "experts" failed to foresee were obvious or even inevitable.

For example, we know today that Warren Buffett has produced great returns for investors in Berkshire Hathaway (BRK). But was it obvious back in 1965 when he took over the company? In 1989, noted economist Paul Samuelson, who called Buffett a genius, wrote that it was anything but clear that an investment in Berkshire would turn out to be so enormously profitable. He also noted that by that time, even Buffett had acknowledged that with the growing size of his portfolio he doubted he could get the same results going forward.

Most investors rely on past performance when making investment choices. Despite the SEC's standard warning, they believe that past performance is prologue. However, the academic evidence demonstrates that distinguishing skill from luck is very difficult. Consider the following.

Fidelity Magellan had beaten the S&P 500 Index in 11 of the 13 years ending in 1989. However, there was a 99.8 percent chance that a fund manager would beat the market in 11 of those 13 years. Unfortunately, for those believing Magellan's performance was prologue, the fund has gone on to produce poor returns, and investors have fled in droves.

Today, we know that Buffett is a great investor. However, there are many fund managers who invest using the same type of value-oriented strategy espoused by Buffett's mentor, Benjamin Graham. "Confirmation bias," or the tendency to favor information that confirms their beliefs, causes us to ignore information that is contrary to our views of the world. In a 2001 paper, "Buffett in foresight and hindsight," the authors discuss a few other funds that invest the "Buffet way." Among them were the Weitz Value fund (WVALX) and the Legg Mason Focus Fund (now the Legg Mason Capital Management Growth Trust, LMGTX). With the benefit of hindsight, we can now see how investors in those funds fared.

Using Morningstar's data, for the 10-year period ending March 20, 2012, WVALX had underperformed its benchmark by 0.9 percent per year and ranked in the 76th percentile of funds in its category. LMGTX fared no better. The fund underperformed its benchmark by 1 percent a year and ranked in the 71st percentile of funds in its category. And as a demonstration of another risk of investing in actively managed funds, the fund migrated (as the new name indicates) from a value to a growth strategy.

Many other examples could be provided. Of course, there will be some funds that did go on to outperform. For example, another fund that was mentioned in the "Buffett in Foresight and Hindsight" paper was the Sequoia Fund (SEQUX). For the 10 years ending December 2011, the fund earned 5.6 percent and outperformed the comparable passively managed DFA Large Cap Value Portfolio (DFLVX), which returned 4.6 percent. The problem is that without the benefit of hindsight bias we have no way to distinguish the future winners from the future losers. In other words, as economist Eugene Fama and finance professor Kenneth French demonstrated in their 2010 paper, "Luck versus skill in the cross section of mutual fund returns," it's very hard to distinguish between luck and skill.

We know with hindsight that investors in Berkshire who bought in 1965, or even in 1999, outperformed the market. But do we know that with foresight? When it comes to actively managed funds, the evidence demonstrates that the past is not prologue.

The bottom line is that hindsight bias is very dangerous. It causes investors to recall their successes, but not their failures. It also causes investors to believe that investment outcomes are far more predictable than they actually are. Santa Clara University finance professor Meir Statman puts it this way: "Hindsight bias makes it easy to believe not only that the future is preordained, but that anyone with half a brain could have seen it." Hindsight bias promotes both overconfidence and a perception that investing entails far less risk than it actually does.

[LarryswedroeCBS]

rntevans
We use historical evidence that is supported by logic. Preferably we should see a logical risk story explaining the premiums, or perhaps even a behavioral one that you believe will continue to exist (perhaps because of limits to arbitrage and fear of margin, or costs which prevent arbitrage). When there is no logical story we should be highly skeptical of data mining.

With commodities I have written that there is logical explanation for the low/no correlation with stocks and negative correlation with bonds. And one should consider adding commodities if one has high exposure to risks of unexpected inflation, or seeking to hedge some of the risks of stocks. If that is the case one should also consider adding some duration risk, earning the term premium, when adding commodities. The longer bonds and the commodities work nicely together.

As to longer government bonds, as I explain in The Only Guide to The Right Financial Plan, the right maturity varies depending on your asset allocation to equities. High equity allocations work better with longer term bonds--the portfolio's volatility will be dominated by the equity portion anyway, and you can then earn the term premium. With lower equity allocations, the reason for that is you want low volatility and adding longer term bonds would defeat that purpose. I should add that using TIPS changes the picture vs. using nominal bonds.

With energy stocks, I don't recommend them because they have higher correlation to equities than they do to commodities themselves, for good reason as they are stocks, and you likely already own them in well diversified portfolio. Health care is a sector that doesn't explain returns well, not an asset class like small or value stocks or REITs.

I hope that is helpful

Best wishes
Larry

[rntevans]

Hi Larry,

I am a big fan and owner/reader of your many books. This article touches on something I find confusing in trying to design a diversified portfolio.

Often I/we use historical returns to justify an asset to be included in a portfolio. But are we not also biased by this inclusion. For example you might tilt your portfolio towards small cap and/or value equities because historically the inclusion of these assets add something favorable to your hypothetical portfolio, namely higher returns with less volatility. In previous CBS new articles you talked about adding commodities such as (PCRX) since history shows that having a relatively small amount invested in commodities as part of a diversified portfolio is a net positive. Is this historical bias?

How about what you don't recommend. Historically having percentage of your portfolio invested in long term government bonds has had a net positive effect and yet you are not a fan of these assets. Or perhaps small sector plays such as energy (VGENX) or health care (VGHCX)these too seem to historically effect ones portfolio positively much like commodities or REITs but you don't as I recall recommend the former but you do recommend the latter. Can you see how I am confused. Are we not hopelessly biased.