Ability to take risk or willing to take risk, which one will dominate?
If one investor has average level of ability to take risk and below average level of willingness to take risk, what should be his overall risk tolerance?
Take the more conservative one i.e in your example RT will be below avg.
Recently, the exception to this is that when one is above average and the other is below average, CFA Institute has been concluding an average risk tolerance, not below average.
Thanks for the update!!!
My pleasure.
I found that they like to use a lot of examples of individuals who have a high willingness to take risk and lower ability and that the ability will dictate what the portfolio should look like. E.g. Joe loves his friend Mark who invests in PE and hedge funds. Joe sees his friend Mark as rich and successful so he wants to have a similar portfolio. Oh, Joe also has liquidity needs every year of $100,000 and income of $75,000 on a portfolio worth $X (just say its not significant). Basically here it doesn’t matter how rich Joe sees Mark, what matters is that he can’t handle higher risk assets that could lead to greater fluctuations in returns and not meeting his annual liquidity requirement, so his risk tolerance is going to be lower despite having a high willingness to take risk.
In essence, in many (not all) cases the ability trumps the willingness, especially when the ability is lower than the willingness. There are cases with someone has high ability and low willingness where you meet in the middle and S2K mentioned, but the opposite is much less likely.
These problems are always a little tricky but they always give you clues as to where they should be, Time horizon, passive vs active wealth accumulation, attitude towards previous investments, etc. Once you practice them a bunch they start making more sense and you can see what the institute is laying out more clearly.
I feel ability to take risk should dominate and the investor should be educated on why their willingness should match their ability
This works to an extent, but if the investor is afflicted by emotional biases, this isn’t a game you’ll win easily with “reason”.
If you adopt this approach on the exam, you run the risk of losing several points on a simple question.
I agree with this and in an example where one is above and the other is average then go with the more conservative choice which is average
Even if the ability is below average (and willingness above avg)?!
When one (either one, it doesn’t matter which) is below average and one (either one, it doesn’t matter which) is above average, CFA Institute’s current position is that you should conclude that the overall risk tolerance is average.
That could change tomorrow, of course.
@S2000
Do you feel this is a logical position on the part of the CFA Institute?
The variables “ability to take risk” and “willingness to take risk” are certainly ordinal variables, let’s say with 3 levels, to keep it simple [below average, average, and above average]. To argue these are interval scale measurements (movement between categories is naturally ordered and “equal” differences in categories are meaningful and equal distance) is really untenable. Looking only at one variable, say ability to take risk , it’s totally unreasonable to assume that moving from below average to average is “equivalent” to moving from average to above average. So within the variable of ability to take risk , “averaging” below and above average risk tolerance to “average” is really foolish. This is often the same trap most social science people fall into with “strongly disagree” and “strongly agree” type shenanigans on Likert-type items and scales. Now setting that up, it’s even more foolish to assume that movements between levels are equal across the different variables of ability to take risk and willingness to take risk – averaging those is even more disastrous!
I think for test purposes it might be okay to say “assume [insert bad assumptions]” so people can answer test questions, but I’m highly skeptical that the curriculum actually explains more judgment should be used in real life since these assumptions are highly untenable.
Sure I do.
In the same way I think that if Bob is above average in weight and below average in height, then he’s average.
Tickersu, those statements about below average, average, and above average are just the preliminary analysis of wealth management. Just the first step. Following to this, we crunch the numbers: portfolio required return, related risk, and constraints that affect risk, return and asset allocation.
Suppose you have a US$ 41 million portfolio (high ability to take risk) and very low willingness to take risk. You are a conservative rich. As an investment advisor would you advice this client to invest 90% in AAA corporate bonds and 10% cash? I think not. This guy should have an average tolerance to take risk because 1) he has ample asset base and 2) the probability that this investor stick to a risky portfolio in the long term is low, therefore we consider he has an “average” level or risk tolerance. If the client has no emotional biases but cognitive instead, then education is the cure. Possibly, the client could realize hi has great ability to take risk and invest accordingly. On the other hand, if emotional biases are detected, an average-risk portfolio sounds really reasonable.
About models (probit or logit) fed by qualitative surveys. I think they are good enough to show us the paths common people take in decision making or opinion belief. Of course, alphas are relaxed in coefficient significance tests.
For example, assume we make a survey to 1,000 investors of all types and ages and ask them how they consider themselves about their willingness to take risk and ability to take risk (below average, average, above average). In parallel, we, as researchers, make special tests on them about their true ability to take risk and which investor types are them. A logit or probit model that contrast both data sets could tell us how mistaken are investors in self-assessment about their ability and willingness to take risk (not an exact outcome but still useful for wealth management).
Seriously!!!
Believe me for the exam you dont need to go that far. I dont think CFAI will ever give you such tricky situation. When you start solving recent past exams you will notice that Risk Tolerance is no more a very dear exam topic to test makers may be because of the diverse thought process of the candidates. Rather they have started asking whether a person has above average or below average Ability to take risk and/or Below or above average willingness to take risk i.e they ask these two elements separately for which you need to memorize a little list to be able to recognize the correct answer. So my advice is not to overcomplicate things.
May be a poor example, but willingness and ability to take risk are more arbitrary and abstract constructs that those categories are trying to measure, where as measuring someone’s height and someone’s weight are underlying continuous, ratio-scaled measurements that are reduced into an ordinal measure (not even necessarily interval scaled, depending how it’s done).
However, I would point out that in the univariate distributions he might be above “average” in one and below “average” in the other, but is that necessarily the “average” of the joint, bivariate distribution (i.e. is it “average” to have that particular “AND” condition).
Taking another perspective, too, leads to an erroneous idea that you can “average” these arbitrarily created categories-- implicitly you’re assuming they’re equally important and that’s even if we can get past all the issues of not being able to “average” categories like this; after all, what’s the arithmetic average (what was done here) of “extremely unlikely” (1) and “extremely likely” (5) on a survey from “1” to “5” where “3” is “neutral”? The knee jerk response that is incorrect is “3” “neutral”; this is a big issue in social science research. The problem in “averaging” these categories is that the distances between them aren’t logically argued to be equal despite this arbitrary numbering scale (can make the argument without the numbering scale, as well). I think to suggest that the difference between below average and average willingness to take risk is the same as average willingness and above average willingness is silly. Further, suggesting that one of those differences is equivalent to the corresponding “difference” in the ability to take risk variable is more ludicrous; a measure of psychological preparedness is suddenly equivalent to financial well being in the context of the grander life picture?
I don’t see a great argument that doesn’t involve a lot of untenable assumptions that make this a sound, real life approach; for the test they can tell you to make assumptions to answer questions, and this will test your reasoning within a framework, but in the real world, stuff like that doesn’t work (there’s research that shows this and, even from a theoretical, baseline perspective of measurement levels of data…NOIR…this doesn’t work).
Sure, I don’t disagree with any of that. I’m saying it’s not a logically sound perspective to literally take the arithmetic mean of variables that are not at least interval scaled measurements; this problem has plagued social science research for a long time.
I’m not sure an example was needed. Personally, I wouldn’t force the client into investment positions that weren’t agreed on or weren’t specified by the IPS and the desired level they wanted to participate. You said yourself that your client has low willingness to take risk despite the fortune; ability is high and willingness is low. I would talk with them and talk about ways to make sure we can meet their goals, but in the end, I’m not cranking up the riskiness of their portfolio because I think they should take on more risk-- unless they want to or it’s provided for in the IPS. Getting off track though. There is a clear distinction in willingness and ability to take risk. One is more or less monetary and goals in the context of life position and the other is largely a psychological feature that may have various causes (religion, education, etc.) You can’t take the arithmetic average of those two things. What’s the average of happy and money in the bank if I create levels of low, medium, and high for each variable? It’s absolutely and incredibly wrong to say someone who is high on money in the bank and low on happiness is average.
I’m not really sure which reports you’re referring to, but nonetheless, I don’t disagree that surveys might give a decent look at “typical” decisions.
Sure, but there needs to be a true assessment of both of those things and an ordered logit would fit under the assumption of proportional odds of moving to the next category. This still misses the key point I’m making; you’re not able to correctly “average” variables with two different meanings. It doesn’t make any sense to average shoe size in cm with hair length in cm. You can do it, but it doesn’t mean anything just because the units of measure are the same.
Right, no one is overcomplicating this for the test. If I study the CFAI curriculum it’s not hard to use that knowledge only to answer their questions. I’m taking issue with the inappropriate use of “averaging” different things (which I think is okay for them to state some assumptions for test taking purposes) and my suspicion that the curriculum doesn’t say that real life needs careful thought about these variables in the context of each client and the specific IPS. The curriculum likely treats it as a golden rule that above average on one and below average on another makes the person “average” which is only true with very specific assumptions that don’t necessarily fit with a reasonable approach.
I argue the real answer to the OP is that the overall risk tolerance depends on unique weights that are tied to the utility function of each investor to give a composite risk score that is, in general, not equal to the “middle” of the two categories.