Portfolio Optimization: Our Secret to Driving Better Performance
We optimally blend funds to pursue higher expected investor returns for each asset class.
Many investors use a combination of tactics to try to get the best performance they can from their portfolios, including asset allocation, diversification and other risk management techniques. But the difference between Betterment and individuals who are trying to navigate this alone, is the complexity and the scale on which we can do this for you.
When you invest with Betterment, you’re getting a professional portfolio that has fully integrated these tactics, delivering you an investment vehicle that's already been optimized. We integrate a number of sophisticated strategies that few people can implement on their own as part of our portfolio optimization, including maximizing upside potential and minimizing the downside risk for each of your investment goals.
Building the portfolio
We know any DIY investor can choose a bunch of funds with enough personal time spent on research, whether it's through Fidelity or Vanguard or some other platform. In fact, DIY investors can and do apply the lessons of many years of research with respect to picking funds, like only sticking with index funds, or favoring a value tilt.
But for many people, spending a couple of days a month on investment research and management is either not of interest or a practical use of their time. The alternatives are paying for an advisor, or using a basic target date fund. The former is expensive, while the latter is inflexible to your needs, and can also be unnecessarily pricey.
That's where our portfolio and service are ideal. At Betterment, we offer 101 customizable allocations to customers, ranging from 100% stocks to 100% short-term Treasuries.
First, it's helpful to understand how we built our overall portfolio and what's inside. We started with a practical foundation based on Harry Markowitz's Nobel prize-winning research1 from the past seven decades.
We began with the concept of diversifying as much as possible (Markowitz, Modern Portfolio Theory, 1950s), and then tilted toward value and small-cap stocks (Fama & French, 1970s). Since we know that most active mutual-fund managers tend to underperform, we then picked low cost, index-tracking, high liquidity ETFs for our portfolio. And because people often worry about potential losses about twice as much as potential gains (Kahneman, Prospect Theory, 2002), we worked on minimizing downside risk.
Lastly, we assembled those funds in a way that gives you better performance by adding another level of analysis, or portfolio optimization. To do that, we used some of the most recent quantitative models for asset allocation and downside risk to squeeze even more performance—or diversification alpha—out of these assets.
The two modern techniques we used are the Black-Litterman model and a downside-risk optimization model. These two models complement each other like yin and yang—one model helps us optimize for the upside, while the other helps us see what the downside might look like.
The Black-Litterman Model: This model allows us to generate forward-looking returns estimates —the upside—based on actual data that includes the collective intelligence of all investors around globe. To be sure, this is a general description of this model; there is also an academic view as well.
This complex formula has a very basic insight at its core: it looks at how all investors around the world behave, and based on that information, creates a kind of global asset allocation model. This makes it a very good anchor of where all the world’s money is invested in the aggregate at any given time. The model was introduced in 1990 and refined over the next decade, and also helps make up for some of the shortcomings in the classic Modern Portfolio Theory, which can underestimate the diversification benefits of some asset classes. Read more about our diversification strategy.
In addition, Black-Litterman is the way to avoid a so-called home bias in investing. This refers to the preference investors have for favoring assets that are “close to home”, contra evidence that would suggest a more global allocation. In other words, it's a tool for using empirical evidence to make investing decisions, with no reference to regional likes or dislikes. U.S. stock markets are only about 48% of the world stock market—the remainder is international developed (43%) and emerging markets (9%). You can see this breakdown in the MSCI All-Country World Index.
Minimizing potential for loss
Downside Risk and Uncertainty Optimization: Modeling for worst-case scenarios allows us to generate forward-looking views of potential downside risk and uncertainty based on the combination of the historical returns of our chosen assets.
When we model our future expected returns we want to know two things — what is the frontier for expected outcomes and what is the frontier for worse than expected scenarios (e.g. everything from a mild downturn to a massive drawdown). With this model, we can evaluate a full range of future outcomes. We can also stress-test our allocations through negative market scenarios to get an idea of how severe a drawdown could be, and what duration. We can also factor in the role our continuous, algorithm-based investment management will play, primarily via automatic rebalancing.
An easy way to see the value-add of our strategies is to look at the difference between our efficient frontier and that of a so-called "naive" portfolio, one that is made up of only the S&P 500 and an index tracking all U.S. bonds (AGG). The expected returns of Betterment's portfolio significantly outperform a basic two-fund portfolio for every level of risk. This result is a function of portfolio optimization, along with our well-crafted selection of assets and funds.
Even if it's clear that these strategies are out of reach for virtually all DIY investors, you might ask: why doesn't every advisor do portfolio optimization? There are several reasons. One is the issue of quantitative capacity—these methods are mathematically complex with multiple moving parts (that's why our investing team includes experts in mathematics, statistics and economics.)
Second, portfolio optimization is time consuming—whenever a new asset class become available (FX-hedged international bonds, for example), or funds change their expense ratios, an advisor needs to rerun the optimization.
Lastly, there's the cost of updating portfolios—we have built a sophisticated proprietary trading platform that automates these calculations on an ongoing basis, meaning that if we update our optimization, all our customers can instantly benefit. A traditional advisor would have to process many of these changes by hand.
As you can see, investing well is not just a matter of picking the right funds—it's also a matter of applying some serious computing power to squeeze out optimal performance. For you, the result of this portfolio optimization is the security of knowing that for any level of risk you select, we've done a careful evaluation to provide the optimal risk-adjusted performance, and your portfolio is re-optimized on an ongoing basis.
1Markowitz, H., "Portfolio Selection".The Journal of Finance, Vol. 7, No. 1. (Mar., 1952), pp. 77-91.