Building the Ultimate Risk Tolerance Assessment
This article was originally published by Advisor Perspectives.
Investor risk tolerance drives portfolio decisions, yet many financial advisors are rightly concerned about the accuracy of risk tolerance assessments. Why is it so hard? How can we get it right?
These questions came up in a recent podcast discussion I had with Doug Heikkinen. We talked about the events that led me to start Andes Wealth Technologies, how the risk tolerance tools were disconnected from the investor’s preferences and the portfolio decision, and how my training at MIT formed the foundation for a new approach to investor risk assessment.
Also, many people believe that risk tolerance is not stable, because an investor’s attitude towards equity changes when market conditions change. If risk tolerance is indeed unstable, why do we bother measuring it, not to mention using it to drive investment decisions?
In my article, Next Generation of Risk Measurement Methodologies, I dispelled this myth and made the case that the investor’s inherent risk preference is stable, and the changing attitude towards certain asset classes is caused by changes in risk perception, which can be realigned using Andes’ newly patented deep analytics (example).
But how can we build a tool that measures investor risk tolerance accurately, while directly mapping the investor’s choice to a portfolio decision?
A direct approach to investor risk assessment
Traditional risk questionnaires are easy to understand but lack precision. For example, how do you know that it should be a 60/40 portfolio, not 50/50 or 70/30? More importantly, many questionnaires don’t tell investors that they could lose money, causing problems down the road.
Some tools ask investors to choose an upside/downside tradeoff, which is good, but the tradeoffs presented to investors are artificial, not real, in that they don’t reflect the models used by financial advisors. If the end goal is to assign a model portfolio to the client, why don’t we show the gains and losses of the advisor’s models and ask the client to pick the tradeoff that they feel most comfortable with? This way, the client’s choice maps directly the advisor’s models. This is the approach I take.
In figure 1, each bar represents one of the advisor’s model portfolios, showing the range of upside (green) and downside (red). Together, the model set covers the entire risk range from conservative (low risk, low return) to aggressive (high risk, high return). When the investor picks a tradeoff that they feel most comfortable with, it maps directly to one of the advisor’s models.
It is direct, transparent, and defensible. The investor explicitly says that they feel comfortable with potentially losing 7.4% in six months in a relatively normal market condition, as indicated by the 80% confidence interval.
This risk tolerance test has been awarded a patent by the U.S. Patent and Trademark Office (patent id: 11403708). It was also a winner of the WealthManagement.com Industry Award in the Risk Tolerance & Client Profiling category.
It doesn’t give a risk score of 1-100. This is intentional because volatility is directly measurable, and it is the same scale as VIX that many people are familiar with. If you want a risk score of 1-100, you can simply multiply volatility by five (see article).
Understanding the upside and downside
In figure, the range of gain and loss for the chosen 60/40 model, [-7.4%, 14.1%], were calculated based on a target return of 6.80% and target volatility of 11.90. It assumed a normal distribution and an 80% confidence interval, i.e., in a relatively normal market condition.
People who are not trained in statistics may not fully understand that it is a statistical range that the return will fall into 80% of the times, and the chance of hitting either -7.4% or 14.1% is small. Figure 2 shows the normal distribution, telling us that the chance is highest around the mean (i.e., average), which is the target return (6.8% in this case), or equivalently, 3.4% for six months.
Other tools may show the upside and downside, and they typically use two standard deviations, which is 95% confidence interval. It covers more scenarios, but the range is wider than the 80% confidence interval, which is 1.28 standard deviations. The wider range gives an exaggerated sense of upside and downside. The 80% confidence interval strikes the right balance between covering enough scenarios while not exaggerating the upside and downside, eliciting a more accurate response.
There are still three unanswered questions:
The set of three buttons in figure 1, “10-Year Range,” “Growth of $1,” and “Drawdown,” are designed to answer these questions, so investors are fully aware of what they are getting into.
Telling the long-term story
Figure 3 shows how the return of the 60/40 portfolio can vary greatly from year to year, but if you look at longer time periods, the outcomes become more and more certain. If the investor has a 10-year or longer time horizon, they don’t have to worry about the fluctuations from year to year.
Both sides of the coin
The drawdown chart, showing the peak-to-trough decline, and the “growth of $1” chart, showing the growth over a period, help clients see both sides of the coin, so they are fully comfortable with the investment decision.
Recommended by LinkedIn
Figure 4, the drawdown chart, shows how the 60/40 portfolio has a much smaller drawdown compared to S&P 500, offering protection in down markets.
If a client chooses an aggressive portfolio, this chart will encourage them to think twice. If they confirm that they are indeed fine with, say, a 50% drawdown, this information will be captured in the investment policy statement, which they will sign.
Figure 5 tells the story that the 60/40 portfolio, while offering protection in a down market, doesn’t provide the same growth opportunity.
If someone in their 30s and 40s is choosing a 40/60 or even 20/80 portfolio, their growth will be lower than the orange line above, and the risk of losing out on growth is real. Showing this chart helps them see the opportunity cost, so they won’t lament on the missed opportunity five years later.
Ensuring accuracy
Now we have looked at the risk and return tradeoff from all angles and believe that we have got it right. But how can we know for sure?
In science and engineering, a common way to verify a result is to use a different method to cross validate. In this case, we can use a traditional risk questionnaire to cross validate the result from the risk tolerance test. Dr. Ruth Lytton and Dr. John Grable developed a popular risk questionnaire, which is perfect for this purpose.
If the result of these methods match, you know that you get it right. If they don’t, the investor might not have fully understood the risk tolerance test, or they focused too much on the upside but not enough on the downside. You will explain all aspects of the decision and invite them to take the test again.
Making It personalized
If an advisor has more than one model set, say an ETF model set for small clients and a more sophisticated model set for higher net-worth clients, they can indicate the right model set for each household, which will power the risk tolerance test. When a client takes it, their choice will be mapped directly to one of the models in the intended model set. Now it is personalized.
Infusing with behavioral finance
We can also infuse this analysis with behavioral finance. For example, the investor type helps advisors better understand how clients behave when the market goes up and down. Adapted from the research by Dr. Andrew Lo from MIT, I use a mini questionnaire to determine if an investor is passive, a trend follower, or a contrarian. Advisors can discover if a prospect is a good fit and customize their conversations.
From risk tolerance assessment to asset allocation
Now that we have a risk tolerance assessment to our satisfaction, what else shall we consider in making portfolio decisions?
While risk tolerance – how much risk that the investor is inclined to take – is the main driver for asset allocation, risk capacity – how much risk they can afford to take – is also an important factor that needs to be incorporated.
How can we measure risk capacity? I believe that risk capacity is primarily determined by the investment horizon; clients with long time horizons can afford to ride out the ups and downs. Some people say that how much money the investor has relative to their goals determines the risk capacity, and it is wrong. A well-funded client can afford to lose some, but they can also afford to be conservative because they don’t need the extra return. An underfunded client can’t afford to lose any, but they also can’t afford to be conservative because they need the return.
Using the chosen model portfolio from the risk tolerance test as the anchor, I propose a simple algorithm to incorporate risk capacity:
Summary
Risk tolerance assessment is a cornerstone in wealth management. Getting it right ensures appropriate asset allocation that matches the client’s risk appetite and financial circumstances.
More importantly, by having robust conversations about risk and reward, and documenting it in the investment policy statement, advisors can set expectations and tell a compelling long-term story, which will prevent problems during market turmoil.
Dan Song, CFA, partner and portfolio manager at Westfield Financial Planning, a NJ-based RIA, said,
When the markets are down, you come to us and perhaps look to change your investment portfolio at the worst possible time. We will pull up your IPS, show you the questionnaires we went through, and remind you that you were comfortable with, say, the 60/40 portfolio.
To date, we haven't had to do it. But we let clients know ahead of time that if the market drops and they want to get out at a bad time, we will remind them we did this IPS when cooler heads prevailed.
Helen Yang, CFA, is the founder and CEO of Andes Wealth Technologies, a Lexington, MA-based provider of technology solutions for financial advisors.