Deconstructing Risk and the Magic 8-ball

Until the broad proliferation of exchange-traded funds (ETFs) it was difficult to find a low-cost option for investing in international markets. Now, not only do we have numerous options in this space and others, but there are reasonable choices for investing in segments of many markets that may provide more efficiency than a position in the broad market itself.

Disclaimer (of course)
I am not a registered investment adviser, nor do I play one on TV. No information contained in The Blog of the Flying Taco on this site is intended to be a recommendation to buy or sell securities of any kind.

Developed International Markets
Whenever I hear someone suggest an allocation to the EAFE (the Morgan Stanley Capital Indices Europe, Australasia and Far East Index; all information © 2009 MSCI Barra. All Rights Reserved), I cringe. Why? Well, let’s look at the last 10 years ending June 30, 2009. On a US dollar basis, with dividends reinvested, the EAFE has clocked in an annual return of a whopping 1.59 percent. Not exactly blowing your doors off, eh? Hell, that’s not even keeping up with inflation! Looking at the variability of monthly returns, you’ll find that the annualized standard deviation of the EAFE during that span is 17.73 percent. Worse yet, if we estimate the 10-year annualized risk-free return to be 3.26 percent, the EAFE hasn’t even beat the risk-free return and tallies up a less-than-impressive Sharpe ratio [(asset return – risk-free return)/asset standard deviation] of MINUS 0.094. The Sharpe ratio is a “bang for the buck” measure of how well you’re compensated for a given level of risk. Numbers south of zilch in this department aren’t exactly what we’re looking for. At face value, if you bought the EAFE at the end of June 1999 you’d have been better served by rolling T-bills. You’d be only a little better off than if you’d stuck your loot in a mattress.

Widely-followed domestic stock indices
The EAFE isn’t alone. The venerable S&P 500 (Standard & Poor’s 500 Index; all information Copyright © 2009 by Standard & Poor's Financial Services LLC, a subsidiary of The McGraw-Hill Companies, Inc. All rights reserved.) LOST an annualized 2.22 percent with an annualized standard deviation of 16.06 percent, for a less-than-stellar-to-say-the-least Sharpe ratio of -0.204. Widely regarded as THE benchmark for American equity investors, the S&P 500 has been worse-than-dead money for the last 10 years…ouch! Its middle brother, the S&P 400 index of mid-cap stocks, has done better over this time period, generating an annual return of 4.61 percent with an annualized standard deviation of 18.32 percent, leading to a positive Sharpe ratio of 0.074. The baby of the family, the S&P 600 index of small-cap stocks, had a slightly higher return of 4.74 percent, but with higher risk - annualized standard deviation of 20.00 percent on the nose – its Sharpe is just about identical to the 400’s at 0.074.

What HAS worked?
I’m not trying to pile on by any means; the last 10 years have been challenging for the equity markets since they included not one, but two difficult periods: the 2001-2003 recession and the global deleveraging that started in 2007 and is still in progress. Still, there have been mildly successful investment strategies in hindsight. Let’s start with the easy one first…a domestic balanced portfolio. Allocate 65 percent to an S&P 400 index fund and the remaining 35 percent to an intermediate-duration US Treasuries fund. If you rebalanced back to the original 65-35 split each quarter (ignoring transaction costs and fund expenses), where would that leave us? We end up with an annualized return of 6.13 percent with a much lower standard deviation of 11.52 percent and a Sharpe of 0.249. Not setting the world on fire by any means, but at least there’s moderate compensation for risk.

The trusty Magic 8-ball
I ran my asset allocation model to determine, in hindsight, what would have been the ideal portfolio to generate real risk-adjusted returns for the past 10 years if we had the investment choices we have today. The model includes 86 different investment options. For this exercise I’ve measured risk tolerance simply based on my age - just over 40 at June 30, 2009. To keep it relatively simple I won’t include local currency investments, because for the individual investor that would entail buying a single-country or region ETF and shorting a currency ETF against it – this isn’t always achievable or even allowable. If you started with the original allocation and let it run for 10 years, you’d end up with a return of 8.86 percent net of fund expenses (at current levels), a standard deviation of 7.67 percent and a Sharpe ratio of 0.731. If we rebalance this allocation quarterly (again ignoring transaction costs but including the current level of fund expenses for each option), we’d generate an annualized return of 9.11 percent with an annualized standard deviation of only 5.30 percent. Our Sharpe ratio on this magic 8-ball portfolio is a tidy 1.165! So what kind of stuff, you may ask, would generate such results? Here goes…

Investment (Allocation percentage)
Intermediate US Treasuries (78.93%)
Chile (9.27)
Latin America (2.95)
Nickel (2.48)
Copper (1.84)
Emerging Europe (1.67)
China (0.78)
Brazil (0.73)
Precious Metals (0.70)
Sugar (0.65)

So there you have it…the bulk of the portfolio is allocated to an intermediate duration US Treasuries fund…a whopping 79 percent! Of the rest, 15 percent is in emerging markets equities (including 13 percent in Latin America), and 6 percent goes to commodities. Now look at what ISN’T there…you guessed it…absolutely ZERO domestic equities. What else is missing? You don’t see any broad market indices here…no S&P 500, Barclays Aggregate Bond Index or MSCI EAFE or MSCI Emerging Markets indices. We’ve deconstructed the risk in these and other larger indices to select smaller segments that fit well together to provide decent, risk-adjusted real returns. Some of these investments on a stand-alone basis exhibit extreme risk characteristics, but when assembled as part of a portfolio of complementary assets we can diversify a significant portion of that risk away.

WARNING! Ummm…we don’t have a trusty Magic 8-ball
That’s right, you don’t. Y’know what? I don’t either. It may be easy to say “well, that weird portfolio of Treasuries, Latin America and commodities worked in the past, so I’ll just do that in the future.” What words do you see in literature and advertisements for just about every mutual fund and investment product?

Past performance is not a guarantee of future performance.

This is so true it’s not funny. The model portfolio above tells us what we should have done ten years ago, not what will work for the next ten years. Besides, I’ve backtested my optimization model to see if what worked in the past several years will work in the future. Guess what…it doesn’t! So what’s an innovative individual investor to do?

Is there an easy way out? Not really…
Say you do a version of the plain-vanilla portfolio I mentioned above – split between domestic equities and domestic fixed income. It’s easy, it’s regarded by a lot of people as “safe” and won’t get you dirty looks at any cocktail parties. Will it work? That’s debatable. Looking at ten years of information, the results over that time aren’t horrible, but consider the strategy’s largest drop from a previous high – the “max drawdown.” This point would have you in the hole to the tune of 31.80 percent after February 2009 from a previous high set in May 2008 – only 9 months earlier. Almost a third of your portfolio…that’ll leave a mark! The “magic 8-ball” portfolio’s hole was much shallower with a max drawdown of 12.42 percent in October 2008 from a March 2008 high – but that’s with 20-20 vision in your rear-view mirror. Easier to stomach, but since it’s in hindsight, it’s not an investable portfolio. Like Marc McGwire, “I’m not here to talk about the past; I’m here to talk about the future.”

...but there's a glimmer of hope
I mentioned that I backtested the “Magic 8-ball” portfolio to see it would work in future periods, and it didn’t work out too terribly well. So it was back to the drawing board. All that historical information has to be good for SOMETHING, doesn’t it? Well, it turns out…maybe. I’ve never considered myself a momentum investor by any means, but I looked at correlations between performance of shorter historical time periods and the following three months of performance. None of them are huge by any means, but I’ve seen a few things that may be interesting enough to discuss once I determine how best to temper the risk profile of the model portfolio.

Leave no stone (or its pebbles) unturned...the sum of the parts is greater than the whole
I’ll be honest; I’m a weird investor. If I can figure out a mathematically sound model for investing, I will be less likely to care about what the portfolio will look like to other people because I can justify it with the numbers. I can sleep at night knowing I've done my all to generate the best risk-adjusted real returns I can. That’s not how everyone works. There are political, currency and other risks associated with many segments of the investable markets. China GDP growth slows and the copper and nickel markets end up tanking, for example. Additionally, some of the investment options trade less frequently than others, so the bid-ask spreads may be wider (adding to de facto transaction costs). A lot of these risks are too much to handle for many investors. The key is to keep your mind open and be honest with yourself about the type and level of risks you’re willing to take. Then be sure to explore ALL available options within your comfort zone and investable segments of larger options. For me, it’s easy to tailor a model to include as many or as few choices as possible from the investment options that I track. By deconstructing the risk of larger indices into their smaller investable components, you’ll often find that the sum of these smaller parts can indeed be greater than the whole.