Risk management used to be an exclusive province of hedge fund quants. Due to tremendous increases in the computing power, the risk management toolkit is now utilized not only by advisers, but by many retail investors through “robo-advisers” like Wealthfront. Wealthfront brought the Markowitz optimization algorithm to the masses and in many ways the company is a symbol of the democratization of risk management that’s now going on.
However, in order to make risk tools useful for wealth management, they have to be adapted for the purpose. Most such tools that are used by financial advisers or by retail investors through companies like Wealthfront focus on subjective risk tolerance as the metric that sets the appropriate risk level for the portfolio. That makes sense for hedge funds, because they are targeting a risk level and an absolute return, but don’t have any specific investment goals. However, clients of robo and human financial advisers have crucial investment goals and risk tolerance is simply not enough to arrive at a suitable portfolio for them.
Let’s use Wealthfront’s questionnaire as an example. I absolutely do not mean to pick on Wealthfront here; rather I am using them as an example as one of the leaders in the industry. Here are the questions that Wealthfront asks:
- What is your current age?
- What is your annual after-tax income?
- What are your savings?
- When deciding to invest, do you care more about maximizing gains or minimizing losses?
- If your portfolio lost 10 percent of your market value, what would you do?
• The options are Sell All/Sell Some/Keep Some/Buy More
And after this, Wealthfront will recommend a portfolio based on your answers. My risk tolerance according to all tests I have ever had is around the median (it is 6.5 out of 10 in Wealthfront). My portfolio looked like this:
- 20% – Vanguard VTI (US Stocks)
- 17% – Vanguard VEA (Foreign Stocks)
- 13% – Vanguard VWO (Emerging Markets)
- 15% – Vanguard VIG (Dividend Stocks)
- 12% – Vanguard VNQ (Real Estate)
- 15% – iShares LQO (Corporate Bonds)
- 8%- iShares EMB (Emerging Market Bonds)
Now this portfolio contains some very risky assets like 38 percent of combination of foreign and emerging stocks and bonds. Real estate stocks at 12 percent are almost as volatile as the emerging markets stocks. Let’s run this portfolio through series of stress tests and compare it to S&P 500.
Below are the results in a bar chart. My Wealthfront recommended portfolio is on the left and the S&P 500 SPDR is on the right. Note, that despite my having roughly median (or slightly above median) risk tolerance, the portfolio that I end up with is virtually identical to S&P 500 in its risk profile. The biggest loss for both portfolios is in the Federal Reserve Dodd Frank 2015 Severe Scenario (marked as “A” on the chart), with Wealthfront-recommended portfolio losing 46.4 percent and SPY losing 48.8 percent (virtually identical estimates of an uncertain future event). A 2008-like stock and credit collapse (marked as “B” on the chart) is also similar with Wealthfront suggested portfolio actually losing more at 38.3 percent vs. 36.7 percent for the SPY. In a Euro meltdown (marked as “C” on the chart) again the portfolios are virtually identical.
This portfolio is clearly too risky for the risk tolerance of 6.5 out of 10, which Wealthfront assigned to me. Because if 6.5 is the level of risk for S&P 500, then what is 10? I am perplexed by complete absence of U.S. Treasuries which provide such an effective risk hedge due to negative correlation with stocks in the extreme periods.
But there are more issues beyond riskiness and here is where we are getting into an area where I believe current robo-advisers are really lacking. Let’s suppose for a second that this portfolio did suit my risk tolerance.
The key point is that none of this addresses my investment goals, namely why I am investing my money. Let’s do a simple illustration of the performance of the portfolio over the investing horizon. Let’s assume that I have $200,000 to invest, a horizon of 30 years, and I would like to withdraw $40,000 per year in nominal terms starting in year 15. I can also save $15,000 and add it to the portfolio every year between now and year 14. As you can see from the chart above the long-term return of the Wealthfront portfolio which we are going to use for this illustration is 8.5 percent annually (marked as “D” on the chart).
We will see how this portfolio will grow if it grows at the long-term return rate. We add $15,000 per year in years one through 14 and we withdraw $40,000 per year in years 15 through 30. Because risk happens, we will also assume two crashes for this portfolio, one in year 1 and one in year 15. The crash will be equal to the average of the worst three stress test losses (“A” through “C” on the chart), so for the Wealthfront suggested portfolio on the left that number is equal to: (46.4+38.3+22.3)/3=35.7%.
The outcome of this scenario is illustrated below. Under those assumptions the Wealthfront suggested portfolio ends up with more than $366,000 at the end of the 30-year horizon (marked as “E” on the chart below). Note that it is not a prediction, but an illustration using some conservative (two major crashes) assumptions.
However, do I really need to take all that risk to achieve my goals? We don’t know because Wealthfront does not take into account my goals, only subjective risk tolerances. What if I sold all of the risky REIT and Emerging Market holdings and bought a Lehman Aggregate ETF (AGG). Clearly, my portfolio would become a lot safer. Let’s look at the crash test to confirm this. The new safer portfolio is shown in blue bars below. It loses only $60,000 in the Fed Dodd-Frank Severe Scenario (a 1/3 loss reduction, marked as “F” on the chart below) and ~$47.5,000 vs. $76,000-plus in the stock and credit collapse (marked as “G” on the chart below).
OK, we’ve reduced the risks to something that I think I can tolerate better, but the key question is: how does it help with my investment goals?
Looking at our projection illustration we can see that the safer portfolio with a blue line ends up with approximately $270,000 (marked as “H” on the chart below). But I am still left with a surplus and I am taking much less risk to get there. So, here is the peril of risk tolerance approaches. Even if they are accurate and academically verifiable, as is the case with the Finametrica questionnaires, there is still something missing. Just because I can tolerate the risk it doesn’t mean that I should take it given my objectives.
And therein lies the opening for financial advisers to out compete robos and many of their peers. Create a framework that addresses both the subjective risk tolerance and the objective capacity to take risks and you will be giving people something unique. As long as we are all in a multi-year bull market, methodologies that assign very risky portfolios based on a questionnaire may do very well. However, when inevitable volatility hits, investors will need to know why it is that they are tolerating risks. And if they don’t see the connection between risk and retirement goals, they might give up on a portfolio at exactly the wrong time—like at the bottom of the market—and that will ruin their chances of reaching the goals that they do have. Advisers who make this connection for clients and explain to them why exactly they are or are not taking certain risks will do very well in the long run.
We have used this approach to show a major weakness in the current robo-advisory offerings. We are certain that in due time robo-advisers will turn into robo-planners and will correct this deficiency. But in the years that it will take them to do that, this approach will help advisers connect risk to clients goals, dreams, and fears and out compete robo and human competition.
Daniel Satchkov, CFA
New York, NY
Editor’s note: To read more about robo-advisers, their shortcomings and their good aspects, check out the Journal of Financial Planning’s archives. Some we recommend are:
- Darren Tedesco’s “I, Robo-Adviser? Creating the Blended Adviser Experience“
- A roundtable moderated by Michael Kitces titled “Advisers + Technology: Better Than Either Alone?”
- “Now Anyone Can Be a Robo-Adviser” by Rick Ferri
- Dan Moisand’s “Why Robo-Advisers Are a Problem for the Profession”