What is Your Confirmation Bias Quotient?

Most thoughtful investors know and understand the concept of confirmation bias. Very briefly put, we selectively perceive and choose evidence that supports our existing beliefs. It is a powerful natural process. Everyone is susceptible.

 

Morgan Housel has a good challenge: “What’s something you strongly believe that’s likely wrong?” He has a wonderful description of the key problem:

 

And while most of us are OK being told we don’t know everything, being told we have a lazy thought process is hard to interpret as anything but an insult.

So we have the ultimate cognitive dissonance: Fully aware that we’re wrong about something but unable to admit being wrong about anything.

 

Unfortunately for the decision maker, fixing the process is the key to better results. There are various discussions about how to avoid confirmation bias, but they are pretty general and not well-linked to investment decisions. Even worse, many investment discussions descend into an argument about who is biased, instead of an intelligent discussion of the facts.

 

Since it is not easy to detect your own biases, I have devised a Confirmation Bias Quotient to help. I have scaled the test so that high is good.

 

  1. Anecdotes. If you pay a lot of attention to specific stories and examples, give yourself -3. Illustrations can add color to conclusions, but when used as the basic level of analysis if is too easy to find supporting narratives.
  2. Specific examples. Similar to #1 but probably even more common. How do you interpret information during earnings season? If you pay a lot of attention to news reports on specific companies, give yourself -3. (It does not matter whether the stories are positive or negative; -3 either way).
  3. Symbols. If you find yourself drawn to colorful or graphic symbols of events – new paradigm, stall speed, stagnation, or anything similar pointing in any political direction – give yourself -2. If you completely reject analysis of data, take an additional -2.
  4. Demonstrably biased data. Examples are things like ShadowStats, where there has been compelling and responsible refutation, without response, on several occasions. Or like the idea that over 90 million people in the U.S. are without work. There is a legitimate debate about some data, but a general rejection of this type indicates a preference for conclusions before evidence. Take -2 if you find these arguments credible.
  5. Emphasizing unimportant data. Choosing to use data rather than stories is a good step. The problem is that there are so many indicators, and most of them have little significance. If you are looking at the Markit PMI (for Europe, China, or the U.S.), or regional diffusion indexes like Empire or Dallas, give yourself -1. There are so many of these that you can find anything you want, and none of them are established as really important.
  6. Embracing biased interpretations. This happens so frequently that I can only give examples. Suppose that a source complains about seasonal adjustments one month, but not another. Or emphasizes sentiment measures only when pointing in the preferred direction. Or emphasizes some specific factor (birth/death adjustment, core measure versus headline) only when it fits their message. It is pretty easy to spot such sources if you look for them. If you find yourself in this camp, take another -1.
  7. Relying upon biased or weak sources. Mr. Buffett said that you should not ask your barber if you need a haircut. Why ask a bond guy about stocks? Or an emerging market manager about bonds? Or a hedge fund manager, who is not really there to help you, about anything? If you do not have a high level of skepticism about sources, take another -1.

 

If you are really mired in bias, you could have a score of negative 15 at this point. Let us turn to the positive factors. Each is worth a possible +5 points, for a total of +20.

 

  1. A willingness to separate your evaluation of the economy and investments from your personal political beliefs.
  2. Finding the most important economic indicators and sticking with them, even when they convey a message that feels wrong to you.
  3. Discovering sources that have demonstrated expertise and track records in the relevant subject.
  4. Being willing to read carefully the analysis of experts with differing viewpoints.

 

The Test is one of Process, not Conclusions

 

A crucial point: You may well reach a consistent bearish or bullish conclusion without significant confirmation bias. The test is about your information, method, and process — not about conclusions. Different experts can look at the same data and reach different conclusions. In my weekly WTWA column I carefully follow all four of the positive factors listed, and strive to maintain a high CB quotient. It happens that my conclusions have been correctly bullish. Some erroneously believe that this reflects bias. Not so. If the evidence changes, so will my conclusions. Why shift from a winning method for “cosmetic” reasons?

 

Scoring the Test

 

If your score is negative, your biases are costing you money. My estimate is that 70% of investors would have a negative score on this test.

If you have even a small positive score you are actively seeking objectivity – probably in the top 20% of all investors.

If your score is above +10, you are doing a very good job of seeking evidence. Your investment results probably reflect this!