1. How to use GuruFocus - Tutorials
  2. What Is in the GuruFocus Premium Membership?
  3. A DIY Guide on How to Invest Using Guru Strategies
Chandan Dubey
Chandan Dubey
Articles (150) 

Investing Basics: Avoiding Confirmation Bias

April 19, 2012 | About:

I am reading the excellent book, "The Little Book of Behavioral Investing" by James Montier. In this book, he talks about pitfalls every human encounters when making decisions when there is either a) insufficient information b) the problem is complex or ill-structured c) the goal is not well defined d) high stress/high stakes e) decision relies on interaction with others.

We recognize that investing has each of these characteristics and to become a better investor we need to learn how to control our baser instincts. Sometimes, it is near impossible to do so unless you know the pitfalls.

The pitfall that caught my eye is confirmation bias. I have succumbed to this more often than I would like and the reason was that I never realized that I am doing it.

Confirmation bias is the tendency to favor information that confirms our belief or hypothesis. This means that you ignore information which does not sit well with what you think, remember information selectively or interpret it in a biased way. You fall to confirmation bias when you do these things (it is easy to come up with more elements on this list):

  • You think that a company is cheap, so you Google for articles that have a buy recommendation. Or you look for reports by analysts who think that the stock is going to go up.
  • You flock to the club your views agree with and shun those which have a different tone. For example, value investors ignoring momentum investors because they think that the momentum investing is somehow beneath them or is clearly not “investment.” Even less drastic is value investing vs dividend investing. If you look at the people who are into dividend investing, they like to read/talk to people who are into dividend investing too and vice versa.
  • Ignoring information that is counter to your thesis. You may justify the information by saying that this is an anomaly and will not happen again (say when a company reports bad earnings for a quarter). You can choose to justify a second bad earning in the next quarter as an extended bad spell and a time to buy further, ignoring the warnings of people who clearly do not fit your view of the world.
  • Democrats reading NY times and watching MSNBC and Republicans doing the same thing with Wall Street Journal and Fox News.
  • When you think a stock is cheap and it jumps 10-20% in a week or so, you like it more and maybe buy it, if it drops again.

Montier describes a simple yet ingenious test to show this bias in practice.

Imagine you are faced with the following sequence of numbers: 2-4-6. Your job is to figure out the rule I used to construct this sequence. To uncover the rule you may construct other sets of three numbers and I will give you a feedback as to whether they satisfy the rule I used. If you are sure you have the solution, you may stop testing and tell me what you think the rule is.

Given this test, most people let their emotional side of the brain take over. They ask for a feedback on 4-6-8 and say 100-102-104 and when the answer for both of them is affirmative, they stop and say “any numbers that increase in increments of two.”

This is the wrong solution as the rule was “any ascending number sequence.” Can you recognize the confirmation bias here? As soon as you see the sequence 2-4-6 you come up with the hypothesis that this is a sequence where each consecutive number is an increment of 2 from the previous one. You test your hypothesis by polling for sequences which fit your hypothesis.

This is clearly the wrong way to go about it. To test a hypothesis, it is best to look for information that disagrees with the hypothesis and not information that agrees with it. To explain it better, I will describe another experiment where you see it more clearly.

Let us suppose that after investing a significant amount of time you come up with a winning formula for investing (there is one by Joel Greenblatt, for example). The formula is quite simple and says that if some conditions are true then the stock will double in three years. Now, how do you test if your formula works? Let us consider these three options:

  1. You look at stocks that satisfied your formula and doubled within three years.
  2. You look at stocks which doubled within three years and verified that your formula holds true for them.
  3. You look at stocks that satisfy your formula and did not double within three years.

The general tendency is to check 1) and maybe 2) (which you will probably notice what most analysts and sometimes book writers do). But clearly the test for this formula is 3). If you find a company that satisfied your formula but did not double then obviously your formula does not work. It does not matter if your formula works on thousands of stocks! If it fails on one then it is not true.

Karl Popper, an English philosopher, observed that the only way to test a hypothesis is to look for all the information that disagrees with it. This process is called falsification.

The question now is: Can the confirmation bias be sidestepped? The first step towards this actually is to step back from your decision and recognize that it suffers from confirmation bias. For example, suppose Vivendi was on your watchlist for stocks to buy if it dropped below €11. A few days back you read that Seth Klarman bought shares. You have read Klarman’s "Margin of Safety" and have seen his track record. You think that if it is good enough for Klarman in terms of return, it is good enough for you. Note that you have done your due diligence before putting the company in your watchlist and you are not blindly following Klarman.

This still is confirmation bias. You must have come up with a margin of safety calculation and had felt that €11 is the cutoff price given your investment goals. Recognizing that this is the case is half the battle. Once you recognize that you are suffering from confirmation bias, you should start with a clean slate. Ignore that Klarman bought shares. Re-evaluate the company and see if it makes sense to invest at the current price.

A good example is also (as Montier points out) the checklist by Bruce Berkowitz on killing the company. Think of all the things that can go wrong with your investment. If you still can’t kill the company and there is a sufficient margin of safety, only then you should make an investment decision.

About the author:

Chandan Dubey
I invest because I want to be free by the time I reach 40 years of age i.e., 2025. My investment style is to find a small number of bets with large margins of safety. I pay a lot of attention to management and their incentive. Ideally, I like to buy owner operator businesses. I am fortunate to have a strong inclination towards studying. I aid my financial understanding by extensive reading in psychology, economic, social sciences etc.

Rating: 4.3/5 (25 votes)


Benethridge - 5 years ago    Report SPAM
Excellent article! Good advice for life in general.
Cornelius Chan
Cornelius Chan - 4 years ago    Report SPAM
Good article! Analytical tools for the investing mind.
Kfh227 - 4 years ago    Report SPAM
Apply this concept to RIMM. Ignore the negative press. Seek out the data. All of the data conflicts with over 90% of the articles written about RIMM. The irony is that I was seeking out confirming evidence on what articles tend to say about RIMM and there was none.

Please leave your comment:

More than 500,000 people have already joined GuruFocus to track the stocks they follow and exchange investment ideas.

Performances of the stocks mentioned by Chandan Dubey

User Generated Screeners

pbarker46PB & PS
jjcarlin99Utilities - non-regulated
jjcarlin99MidStream Oil & Gas
pbarker46EV/EBITDA & P/Med PS
pbarker46High short
Taekgyun Kopbr
pbarker46PB, E10 & Hist. PB
Get WordPress Plugins for easy affiliate links on Stock Tickers and Guru Names | Earn affiliate commissions by embedding GuruFocus Charts
GuruFocus Affiliate Program: Earn up to $400 per referral. ( Learn More)