A longtime expert on asset allocation strategy tells us why it didn't protect investors in 2008.
Roger Gibson is someone I've followed since entering the industry in the early 80s. He has long had a reputation as a deep thinker on the subject of asset allocation and portfolio design, particularly after publishing Asset Allocation: Balancing Financial Risk in 1989, which remains a best-selling book to this day.
I spoke to Gibson, the chief investment officer of Gibson Capital in Wexford, Pa., about the 2008-2009 market and economy.
Drucker: Do we need to somehow recognize and incorporate a new level of volatility in the asset classes we use?
Gibson: It depends on whether we're looking at short- or long-term volatility. Certainly 2008 was without much precedent in terms of how big the losses were and how far they stretched across all asset classes. For our year-end client communication, we studied Morningstar's entire database of 3,734 funds invested primarily in U.S. stocks. All but one lost money. Of the 978 funds invested in non-U.S. stocks, all lost money. Of the 144 U.S. and non-U.S. real estate funds, all had negative returns. And of the 128 natural resource funds, all lost money. So, if you add those numbers up, there were over 4,000 equity funds and all but one lost money.
Drucker: Since most portfolios are blended, one would ordinarily look to fixed income investments to help out. Did you study those, as well?
Gibson: Yes. Of 1,730 bond funds--both taxable and municipal--68% lost money, which surprised us. Those that didn't were government or short term bond funds. If you think about your distribution of possible returns from a diversified portfolio as being a bell-shaped curve, you've got this far left-hand tail in your distribution and we picked that "marble" out of the urn in 2008. I've never seen losses like 2008 but it wasn't something completely unthinkable. And when you have absolutely horrible, panic-driven significant losses, they're usually not just confined to a particular asset class. In 2008, panic fed on itself. I had more than a couple of sleepless nights in the last quarter as foreign markets were in a freefall because the U.S. markets were in a freefall, and the next morning U.S. markets were again in a freefall because foreign markets were in a freefall, and so on.
Drucker: Did you do any other studies of volatility our readers might find interesting?
Gibson: Well, I've never seen markets move as frequently and sharply as in last few months of 2008, so I had a staffer look at index movements for the S&P 500, the Morgan Stanley Far East Index, a REIT index, and the Goldman Sachs Commodity Index vis-a-vis the question: over how many trading days did these indices gain or lose 5% or more? The S&P 500, from January through August of last year, had no 5% days. From September through December, there were 18 days when the market moved 5% or more, and there were only 85 trading days, so that averages out to one out of five days, which is a lot! Similarly, the international market had no 5% days between January and August but, between September and December, had 28 days of 5% or greater movements. That's 33% or one out of three days. And the REIT market was even more volatile. During the January to August period, it moved more than 5% on seven days and, between September and December, there were 46 days, or over 50% of the index's trading days. Commodities were also more volatile with five days of 5%-plus movement from January to August and 18 days in the September-to-December period.
Drucker: How can we use this information to make educated trading decisions in our client portfolios?
Gibson: If you're building a strategy around these kinds of data, it's probably better to take a sampling of trading days. Five-percent shifts can make you a hero or a bum. The kinds of volatility we found in the REIT and commodities indexes seems extreme but I'm not sure it will make a significant difference when measuring standard deviations on an annual return basis. From a portfolio construction point of view, when looking at how much of this or that we should put in a portfolio, we base this decision on annual return data. I'm amazed at how quickly investors have accommodated themselves to 5% moves being no big deal. Over the last 50 years, these kinds of moves probably only happened once a year and, already, investors have almost gotten numb to this volatility. So I think from a volatility tolerance perspective, investors can get used to heightened short term noise. The larger concern is ... are the annual returns OK ... am I progressing?
Drucker: So it sounds like you're saying the rules going forward aren't really any different from the asset allocation assumptions we've held the past?
Gibson: In terms of volatility, the jury is out on the next 10 to 20 years. It's not a given that annual return volatility for major asset classes will be substantially different in the next 20 years as in past 20 years. Also, we need to remember it's possible to have tremendous volatility on an intraday or intraweek basis that doesn't necessarily translate into increased annual volatility. So, from an asset allocation decision-making point of view, last year was a volatility test. In our firm, we spend a lot of time up front when designing the investment strategy for a client and discussing with the client how asset class returns change over time so the client can understand the likely range of possible outcomes with any portfolio strategy we might recommend.
Drucker: Was the 2008 market decline within the realm of possibility given the historical volatility of returns?
Gibson: Using Ibbotson data for the U.S. market from 1926 forward, the simple average return of S&P has been roughly 12%. If the S&P returns are normally distributed--which they're not perfectly--then you'd expect 95% of annual returns to fall within two standard deviations. One standard deviation is 20%, so two standard deviations give us a return range of negative 28% to plus 52%. Five percent of all observations would fall outside that range, so we know 2.5% of all observations will fall outside the range on the low side, and also on the high side. So looking back to 1926, we find the market's return outside two standard deviations three times: 1931, 1937 and 2008. Of course, you would also expect to see several years when returns fell in the 2.5% portion of the distribution on the high side, and we did see this in 1933 and 1954. So my point is that if you look at what happened in 2008 in the context of the good data we have on the U.S. market, and if you assume distributions are relatively normal, we actually have had experiences in the tail ends of the distribution that coincide with what we'd expect.
Drucker: So 2008 wasn't as improbable as most of us think?
Gibson: Not really. It was a crazy year, but we know those marbles are somewhere in the urn waiting to be picked. But I'd put all of this into an even longer historical context.
Drucker: How do we do that?
Gibson: We don't need to necessarily do anything differently. What will change [going forward] isn't volatility, but capital market assumptions. In other words, relative to whatever capital market assumptions advisors have been working with, 2008 will raise expected returns--in absolute terms--going forward, the spread between equities relative to fixed income will widen, and all kinds of risk premiums will widen too as a result of last year's experience. So if we look back through time, we were in a lower absolute return and lower risk premium world and, going forward, we'll be in a slightly higher absolute return world and higher relative risk premium world. During times of excessive optimism, people overshoot markets on the high side, and during times of extreme fear and panic, markets overshoot on the downside. In 2008, people panicked and dumped securities, which sets the stage for higher-than-normal rewards for people holding on.
Drucker: How can we better assure that portfolios created to help clients reach retirement by a given date will actually accomplish that objective?
Gibson: This gets back to the question ... are you using reasonable capital-market assumptions? Toward the end of the 90s, we used a forward-looking single-digit assumption for U.S. equities. Others thought we were low-balling our assumptions since other advisors were routinely modeling double-digit return expectations. But if advisors used these double-digit assumptions, then went through the 2000-2002 markets and then had the 2008 experience, there was no relationship between their assumptions and what actually happened. [Using excessive return assumptions] destroyed retirement planning for lots of clients. If those advisors had used reasonable assumptions, then they would have been acknowledging the fact that these bad times could happen, and their Monte Carlo simulations would have been incorporating these more reasonable return assumptions--to their clients' benefits.
Drucker: So how has 2008 affected the annual return assumptions your firm will use going forward?
Gibson: We just compiled and updated our numbers, pulling in 2008. The S&P 500 for the 10 years ending February 2008 lost 1.4% annually. The EAFE achieved a positive 1.2%, and the NAREIT Equity Index was up 7.4% as was the commodity index. I would say this is a great success story for our strategy of multiple asset class investing even though U.S. and non-U.S. stocks were flat for a decade; both real estate and commodities doubled over same ten-year period. So, if an advisor had a meaningful allocation to both of those asset classes, you've got two classes that doubled in value and that's a good story to be able to tell. If you look at ten-year rolling averages between 1972 and 2008, the S&P 500 has compound return of 9.5%, EAFE 9.7%, NAREIT 11.2% and commodities 9.5% so, in the long run, equity asset classes have equity-like performance. Over these ten-year periods, the U.S. market is sometimes in the lead, but that leadership changes over time, so you just keep diversified over those asset classes and you rebalance back to target. Asset allocation doesn't eliminate all risk on the downside, but it never promised to do so. Strategic asset allocation isn't broken and never promised to sidestep a year like 2008, but what it will do is make the portfolio as a whole have less average risk. That's mathematically driven. And it will cause a portfolio to have a higher compound return than the average return of its asset classes. That said, it doesn't mean you can't get into a scary environment.
Drucker: So there will just be some years--like 2008--where asset allocation, with all its benefits, won't keep us safe.
Gibson: Let me answer that with my favorite quote, by Paul Volcker, who, 15-20 years ago, was at CFA conference and said, "You cannot hedge the world." I love that quote, if you're living on this planet, you can manage risk, but you can't eliminate it. We got that lesson in 2008.