Betterment’s Quest for Behavior Gap Zero
A new analysis shows that Betterment is helping investors optimize their behavior and reduce the notorious behavior gap, which compromises investor returns.
Market timing and active trading, on average, cost investors up to 6.5% per year.
Betterment’s behaviorally intelligent platform helps reduce this behavior gap significantly.
Investors make countless choices designed to increase returns and minimize losses. No mystery there. Yet, if most investor behavior is oriented toward growth, why is growth typically compromised by investor behavior?
Enter the so-called behavior gap: a term coined by author and illustrator Carl Richards, and one of the most notorious phenomena in the investing world.
A number of academic studies have shown that because of bad timing decisions, investors sacrifice a significant portion of their returns. The chart below reflects a comprehensive survey of researchers’ attempts to quantify this behavior gap.
Interest in the behavior gap is no longer purely academic, yet the investment industry has been slow to address this problem and the hugely negative effect it has on investor returns. This likely comes down to a combination of factors: misaligned incentives; the lack of innovation in an archaic industry; and the inherent difficulty in addressing some of the most innate and pervasive quirks of human behavior.
So the behavior gap presents a potent challenge, particularly for an online investing advisor like Betterment: Is it possible to use technology to help improve investor behavior on a wide enough scale to make a significant difference to people’s actual returns?
At Betterment, not only are our incentives aligned with our customers’ goals, but we are the only investment advisor that we are aware of to track—and actively try to reduce—the gap in performance between the return that we provide in our model portfolios and what our clients actually achieve in their accounts.
In fact, by tracking this metric, we find that our customers are taking home on average 2.66% more per year than the investors examined by other studies. Moreover, we are continuing to use investor data to improve our product and further reduce behavioral drag on returns.
The behavior gap: a primer
Betterment customers have a behavior gap that is 1.25 percentage points fewer than that of an average investor.Our job is to provide our clients with the best take-home returns possible. We do this optimally in our portfolio by employing the so-called free lunch of diversification along with value and small-cap tilts and tax-efficient rebalancing. We also make sure our clients are adopting the appropriate risk levels in their portfolios, using our proprietary data-driven advice algorithms.
But we also want to ensure that our clients actually attain the returns that our investment product generates. Thus we use various techniques to promote good investor behavior, while discouraging the market timing and narrowly-focused decision making that leads to the behavior gap.
To understand the behavior gap, it is helpful to think about two kinds of returns—an investment return and an investor return. The Betterment portfolio seeks to maximize investment returns, while the Betterment product and the investing methodology it encourages are designed to maximize investor returns.
Imagine a young professional: Pete. Pete invests half of his $100,000 retirement fund at the start of the year in Fund X. After six months, he finds that the fund is up 10.5%, which compares favorably to the 9.0% return of the S&P 500. Pete is happy and speculates that these returns will continue, so he moves the rest of his investable assets to Fund X.
But in the second half of the year the market dips, taking Fund X with it—and it falls 5.0%. The fund can still boast a benchmark-beating return of 5.0% for the year: this is the investment return. Unfortunately Pete cannot.
Pete’s return—his investor return—for the year was 0.0%. The first $50,000 he invested in Mutual Fund X saw the 5.0% increase for the whole year, becoming $52,500. The second $50,000 however only saw the -5.0% that came in the second half of the year, becoming $47,500. As they were equal investment amounts, Pete saw a net return of 0.0%.
In other words, the investment return speaks only to the performance of the investment vehicle (i.e. the manager or the index tracked by the fund). But only a one-time, buy-and-hold pattern will actually achieve exactly that return.
In reality, we trade investments based on changes in our financial circumstances, our investment goals, and on our opinion (or emotion) about what will happen in the markets. Each of these factors contributes to a difference between the investment return and investor return—and usually not in the investor’s favor.
First, transactions generally incur costs. While there are some exceptions, mutual funds and hedge funds have their load fees, discount brokers have their trade fees, and some financial advisors will charge fees in proportion to trading activity.
At Betterment, our technology allows us to offer unlimited free trades and fast access to money at all times, and the ability to seamlessly invest as you earn, without penalty.
The second way investors underperform is through mistiming market movements. We’ve written before about the various ways in which our cognitive biases lead us to trade too often; buying high, selling low, and costing ourselves important time in the market. As a result, most investors would have been better off doing nothing.
But how significant an effect do all these have on investor returns?
As mentioned above, numerous studies have been published in peer-reviewed journals over the past 10 years attempting to quantify the behavior gap, and the results are quite staggering. Individual investors lose on average between 1.2% and 4.3% to the behavior gap. For the most active investors, the gap increases to 6.5%. This is a wide range, and there are a number of ways to extrapolate a single reasonable benchmark. We chose the findings of the Friesen and Sapp study, which observed a behavior gap of 1.56%—one of the more conservative estimates.¹
If Pete’s ill-timed investing behavior caused his $100,000 to underperform by 1.56% per year, after 20 years he’d leave about $118,000 on the table, ending up with 25% less than with a simple buy-and-hold strategy, as shown below.
The Betterment gap
What about the behavior gap at Betterment? Is it possible to determine whether the Betterment platform conveys a significant behavioral advantage that helps to close the gap and provide better returns? Can we use this data to continue to enable and encourage good investment behavior?
Our investing team applied the same rigorous analysis used in academic papers cited below to our investor results—the first case of an investment manager systematically monitoring client behavior that we know of.
Our analysis was based on tens of thousands of Betterment customers investing between 2010 and 2013, factoring in cash flows (deposits and withdrawals) as well as allocation changes over that period.
For each customer, we first calculate the Internal Rate of Return (which takes into account all investor actions and provides the investor return). We then calculate the weighted average stock allocation for each customer over the life of the account, and determine the relevant Betterment portfolio’s Time Weighted Rate of Return (the buy-and-hold, or investment return).
We then annualize these to make different durations comparable, and subtract the investor return from the investment return thus:
Betterment Behavior Gap = Investment Return – Investor Return
Combining the results gives us the distribution of the behavior gaps for each Betterment account, the average of which represents the overall effect.
Result: Betterment’s customers, on average, have a behavior gap of only 0.31% per year. As compared to our benchmark of 1.56%, Betterment customers have a behavior gap that is 1.25 percentage points fewer than that of an average investor—a gap that’s one fifth of the average.
Though we chose a conservative benchmark, we recognize that any comparison available to us would be imperfect. Our data covers a different and shorter time period and uses individual account level data for a range of portfolios, rather than aggregated fund data. Yet it’s notable that our result is substantially lower than every study we’ve seen (in some cases, by an order of magnitude).
More importantly, our analysis provides us with our own, internal benchmark, that we can now measure ourselves against. We’re able to show the average difference between our customers’ actual take-home returns and their equivalent buy-and-hold returns. This illustrates the significant potential for gains that can be made thanks to the Betterment method and portfolio, compared to the average behavior gap.
For an online investment manager striving for the best possible investor returns, this analysis is a useful gauge of the success of our methodology and product. We cannot safely assume perfect causation with respect to these findings – it’s possible that an elegant, low-cost platform designed to enable good investing behavior attracts some investors with already better than average behavior. Still, these findings help us to see that our efforts are fruitful, and our ongoing analysis will help us measure the impact of new product features going forward.
How Betterment minds the gap
We were pleased but not surprised by our findings. From the ground up, we’ve designed our product to chip away at the behavior gap at every opportunity.
1. Asset selection and portfolio display
We do not invest in single-line stocks or equities, or allow our customers to select specific ETFs and not others. Rather than a limitation, this is one of the most crucial aspects of our portfolio design.
Under-diversifying, falling prey to the disposition effect, and focusing on the trees rather than the forest (narrow framing) can trigger investors to make less-than-optimal choices. A carefully targeted portfolio, by contrast, enables customers to reap the benefits of our more impartial asset selection and built-in diversification.
Another part of our strategy is a non-traditional portfolio presentation in customer accounts that keeps people focused on the bigger picture, rather than daily market movements. For example, we’ll show you the constituents of your portfolio as the parts of a whole – but never show you the return of each individual component. To do so would be to sacrifice the behavioral advantage of diversification.
Taken together, these factors simultaneously increase expected risk-adjusted returns and reduce the customer’s own behavioral risk of overreacting to idiosyncratic asset movements.
2. No transaction fees
For behavior gap calculation purposes, the benchmark investor makes a single deposit on day one, and nothing else. In the real world, we don’t usually start out with all the principal we’re ever going to have. Making regular deposits is how many of us do (and should) invest, but that necessarily means more trades than the benchmark. This, too, is “behavior”, but we don’t think you should be penalized for it. There are no trade fees at Betterment, ever: not if you are depositing money and buying fractional shares of 12 ETFs, or if you are selling the same, or if you are changing your asset allocation as your circumstances change.
This enables investors, where appropriate, to invest monthly or biweekly, without thought for the size or volume of the transaction. You make deposits when you can, and have no need to aggregate them, which would encourage market timing.
3. Focus on the future
Imagine a weather channel that constantly reported yesterday’s weather, or the past week’s. Would that help you decide what to wear tomorrow? Sadly, this focus on the recent past is one of the most common (and useless) features of an investing interface: ”What was yesterday’s return?”
Our intelligent design helps clients focus on the decisions that matter—the ones about the future. You will never see the day’s performance of your portfolio. Our minds assign a disproportionate significance to daily volatility—it virtually never has an impact on our long-term outlook. You shouldn’t care that the Dow dropped by 1.37% if your personal goals remain on track. Thus, you’ll only be made aware when your personal goal seems to be off track, which is much less often than one might think.
4. We automate everything
Most online products want to maximize the time you spend using them. We always knew that our mission was different—to build a site you would use less. We minimize the number of decisions you have to make, which de facto reduces how much you interact with your investments, which on average hurts returns.
Based on your goal, we recommend the stock allocation and savings amount to reach your target, and we manage the portfolio for you. You don’t even have to decide when to buy—we encourage and make it easy to schedule deposits to automatically invest your money into the market, a feature unheard of for traditional investment management. That goes a long way to taking behavioral risk out of our customers’ returns.
We’ve engineered Betterment to be a more efficient way for customers to enjoy all the benefits of investing, while stripping out the behavioral variables that are proven to compromise returns. This analysis isn’t just a one-off, nor is it merely to satisfy our curiosity. It’s part of our dynamic, data-driven approach to ongoing product development in our quest for behavior gap zero—so expect to hear more.
¹After careful consideration, we elected to use the Friesen & Sapp (2007) results of 1.56% as a benchmark for our study. We made this choice after determining that overall, the data used by Friesen & Sapp most closely resembles the nature of the data we have at Betterment, making their analysis most comparable. While there are still some differences, we sought a fair marker against which to gauge our ongoing results. Friesen & Sapp was the only study that relied on the CRSP Survivorship-Bias-Free US Mutual Fund Database for its analysis. We felt that Betterment more closely resembles a U.S. equity mutual fund than a discount brokerage or hedge fund, the source of data for other studies. Also, we have previously worked with this CRSP dataset, and believe it to be the most robust available. Finally, the article and underlying methodologies appeared in the Journal of Banking & Finance, indicating that it passed the most rigorous of peer-review processes. We also took into account that Friesen & Sapp’s behavior gap finding was one of the lowest surveyed. Choosing a conservative benchmark could, if anything, underplay the beneficial impact of our platform on the behavior gap. We note that our review of existing literature on the topic may not have been exhaustive, and invite further discussion from any interested parties.
“Dalbar” – DALBAR (2012) Quantitative Analysis of Investor Behavior
“Clare” – Clare & Motson (2010) Do UK retail investors buy at the top and sell at the bottom? (working paper)
“Friesen” – Friesen & Sapp (2007) Mutual fund flows and investor returns: An empirical examination of fund investor timing ability
“Sinha” – Sinha & Jog (2005) Fund Flows and Performance
“Bullard” – Bullard, Friesen & Sapp (2009) Investor Timing and Fund Distribution Channels
“Dichev” – Dichev & Yu (2010) Higher risk, lower returns: What hedge fund investors really earn
“Barber” – Barber & Odean (2000) Trading Is Hazardous to Your Wealth
Our Evidence-Based Approach to Improving Investor Behavior
Betterment’s rigorous testing helps to ensure that we know what works (and what doesn’t) when it comes to improving design and investor outcomes.
Using Investment Goals at Betterment
Goal-based investing. The idea is prized among financial advisors—and our team at Betterment—but to the everyday investor, it’s often difficult to put into practice.
The Difference Between Vanguard and Betterment
I’ve been asked a lot in the last few months in the transition from my old role at Vanguard to my new role as Chief Growth Officer at Betterment – what’s the difference between the two companies?