Wednesday, November 13, 2013

Investment Survey Troubles

In Rob Carrick’s latest roundup of personal finance links on the web he asked his readers to take a short Qualtrics survey to “help build a better investor risk assessment tool.” Unfortunately, the survey has problems that will muddle its results.

The main question on the survey asks which of 4 investments you’d be most comfortable with. Here is the exact wording:
“Investments with higher potential returns typically involve greater risk. The following chart shows four hypothetical investments of $10,000, each with a different potential best and worst outcome at the end of one year. Which investment would you be most comfortable with?”
You are presented with 4 bar graphs giving ranges of possible portfolio returns going from safest to riskiest. Here were the ranges I was presented.

Portfolio A: 0 to $10,900
Portfolio B: -$9600 to $11,200
Portfolio C: -$8900 to $11,800
Portfolio D: -$8400 to $12,400

Just based on the numbers, it’s hard not to choose Portfolio A because its midpoint is more than double the midpoints of the other portfolios. So, even though I tend to be comfortable with volatility, I had to choose the least risky choice in this survey because it offered the highest expected return.

Another problem is that the bar graphs shown were hopelessly out of scale. For example, it shows the downside of Portfolio D as about 4 times bigger than the downside of Portfolio B, even though B actually has the greater potential for loss.

I suspect that the survey results will have some meaning for respondents who look only at the picture and ignore the numbers on it. However, crazy people like me who actually look at numbers will mess up the survey results.

This isn’t the end of the problems, though. When I went to look back at the survey again, I was presented with a different chart. After a few browser reloads, I saw that there appeared to be 5 different charts. Three of these charts had big mismatches between the numbers and the bar graph sizes. The other two were to scale, but still showed the safest investment as having the highest median return.

Overall, I seriously question whether this survey can produce any meaningful results about investor attitudes towards investment volatility.

15 comments:

  1. The varying graphs is probably exactly what was being surveyed. Do you check the math, or just choose based on a (possibly confusing) graph

    ReplyDelete
    Replies
    1. @Paul T: If the graphs made sense I'd say it was possible they are seeing how people react to different graphs. However, it's hard to see what they could be researching when using graphs that are just plain wrong.

      I'm not sure I understand your question. Respondents aren't asked to check the math if that's what you mean.

      Delete
    2. What I mean is:
      You're asked which investment return would you prefer. The graphs presented don't represent the numbers give (gain / loss possibilities). Maybe it's gauging whether the reader uses the graph or underlying data to make their investment decisions.

      So if a MF company wants to "sell" it's product, the type of data (mis)represented in graphs will influence the investment decision, ignoring the underlying hard numbers.

      Delete
    3. Paul and Michael, I think you are right. I'm convinced that the point of the survey is to see if people are mislead by the misleading graphs.

      Delete
    4. @Greg: You may be right, but I'm not sure how they can distinguish between people who are risk-averse and chose Portfolio A based on the bar graph and people who are not risk-averse but crunched the numbers and determined that Portfolio A was best. I think it's more likely that the survey is messed up, but who can tell?

      Delete
  2. I am not sure how to interpret the investment results:"Portfolio A: 0 to $10,900", does this mean that I will either earn $900 or lose the whole $10,000?

    ReplyDelete
    Replies
    1. @AnatoliN: I think it means that you invest an unspecified amount and some time later your investment results are somewhere between break-even and gaining $10,900.

      Delete
    2. I don't know if they have changed the survey since you looked at at, but right now it is pretty clear that you invest $10,000 and the various graphs show absolute $ gains/losses or $ gains/losses.

      Delete
    3. @Greg: You're right that the investment is $10,000 as noted in the text. I mistakenly said in a comment above that the investment amount was unspecified. Four of the graphs show dollar gains and losses, one shows percentages.

      Delete
    4. and how exactly can one lose $9,600 on a $10,000 investment except by going leveraged, which is, imho, boneheaded? This survey makes no sense to me

      Delete
    5. @AnatoliN: I suppose the survey's purpose may be to see if people notice such crazy things, but I'm not sure how the results will say anything useful.

      Delete
  3. Personally, I question the logic of the statement "higher risk = higher returns." Maybe that's true over a long period of time. During the short term, though, I've seen some amazing capital gains for very low risk stocks because so many income-investors have been buying blue chips trying to have a decent return (3-6% in dividends) with less perceived risk (e.g. banks, telcomms, utilities.)

    It sounds like whether he meant to or not, Rob Carrick has shown how confusing it can be to look at financial literature and make a choice that actually reflects what you meant to choose!

    ReplyDelete
  4. Like Michael, I couldn't help but load it up a few times. There are a few scenarios it presents, in a few different ways (percentages, a range of gains and losses, a range of absolute amounts). It also asks which you'd be "most comfortable" with which would likely also bias to option A in most cases.

    I can't figure out what Rob is trying to glean from this experiment. It's interesting to present basically the same choices to people and see how they respond differently, but with only a single question posed at a time (no repeated measures) he's going to be posting that survey link for a long time to get the kind of power he might need...

    ReplyDelete
  5. There's a bigger problem. By putting in dollar amounts, the survey is mixing up two aspects of risk assessment - risk tolerance, which is a personality trait aka a person's attitude towards risk and risk capacity the objective financial capacity of the person to withstand a negative outcome. i.e. if you are rich a $10k loss has no discernible effect on your life but if your net worth is only $10k you are laying every penny on the line. Properly built risk attitude questionnaires, such as one developed by Oxford University and made available on the Standard Life website here - http://www.standardlife.co.uk/1/site/uk/financial-education/financial-calculators-and-tools/risk-questionnaire - don't mix the two.

    ReplyDelete
    Replies
    1. @Canadian Investor: I agree that there are two separate things to measure. I always seem to get into trouble when I talk about this. Attitude toward risk is one thing and rational capacity for risk is another. The degree to which they differ is one measure of a person's irrationality. That's not a message many people want to hear.

      Delete