Mohnish Pabrai on Avoiding Commitment Bias, Fairfax and Constructing a Portfolio

Author's Avatar
Aug 07, 2011
Article's Main Image
I was very lucky to receive a copy of notes from the 2010 Pabrai AGM. These notes provide very honest answers from a hugely successful concentrated value investor. Below, Mohnish discusses two important concepts — commitment bias and portfolio concentration:


Q: My question is regarding Fairfax. One of your major holdings is Fairfax. I just wanted to know what your thoughts were. Why you like the company? I wanted to hear why it has good future or if it doesn't have a good future?


A: We own it so we must like it. In fact, Fairfax has celebrated its 25th year in business. Just earlier this week, Prem Watsa, the CEO, asked me to come for the meeting. I have never been to their Toronto headquarters. I told him, "There's really no need for me to ever see it." I said this because they keep doing such a good job.


We don't talk much about existing holdings. But there's a guy in the room who has Fairfax running through his veins. Sanjiv, can you just stand up for a second? You could just direct all your questions to him and he can answer them for you. I just don't want to talk a lot about our existing holdings. We'll talk about Fairfax in a few years when we get to the point where we don't own it.


Q: My question is about your disclosure policies. A good segue from Fairfax. You have many friendly policies. I just want to understand better your philosophy on why it's important not to discuss an existing holding after you've completed trading. Why you feel that this benefits either you as the manager or the limited partners?


A: That's a good question. This has to do with human psychology. If you recall Buffett used to say when he had Coke (KO, Financial) that he would never sell it. In fact, even in 1999, 2000 when Coke was trading close to 35 or 40 times earnings, when he was asked about Coke he said he would never sell Coke. It was something that he would keep forever. Then a few years later, he said that it was a mistake to not to have sold Coke at that time. It was clearly overvalued — ridiculously overvalued. It was clearly the time to sell it.


In fact, I saw that in spades when we did our work on the checklist. If I talk publicly about things that we currently own, I get what Charlie Munger would call commitment bias. I could tell someone all kinds of great things about Fairfax. Then there's a change in circumstances, a change in the analysis that I've done on Fairfax. But I already said all these things about it, and it causes a distortion in the thinking. With Berkshire and with Warren, if he'd never talked about Coke through the period he held it, I don't know whether he would've sold it or not. But the odds are higher that he would've been more free to sell the business when he felt like it.


It's a huge mistake for investors to discuss their holdings while they still own them because these are living, breathing entities that are going through changes every day. Fairfax is a good example where the company seems to be doing well. It's a well-run company, but we don't know what will come at them tomorrow or day after or a month from now. Or what insurance they wrote three years back that is coming to haunt them today. There are a whole bunch of variables that could change the thesis. For me to stand here and commit myself saying it's kumbaya, best thing since slice bread, it's just digging a grave. That's where I'm coming from.


Q: I'm not concerned about the discussion of the holdings, what about disclosure of the holdings?


A: We have quite a bit of disclosure because we go through an audit by Price Waterhouse every year. Our audit reports are GAAP compliant, and they have to list holdings that are more than five percent of the fund. Since we don't own 200 things, most of the holdings get listed. We also have to file 13Fs every quarter where our U.S. holdings over a certain threshold get reported.


Q: Mohnish, could you discuss the construction of your portfolio in terms of percentages as opposed to when I first joined you when you owned 10 securities pretty much 10 percent of each.


A: That's right. That's a good question, Ted. I made a change. This change really came about towards the end of 2008 when I was looking at Teck Cominco type bets. I wasn't comfortable putting 10 percent into them but I also knew that there was a huge upside potential with slightly elevated risks. Also, there were large numbers of them coming at me. The second thing that was going through my mind was that we had made investments in the past, which had a total wipeout of capital. When you make concentrated bets at the 10 percent level and you have a couple of those go to zero, it's very hard to recover from them. That's a significant hit.


In fact, the checklist has reinforced where I have repeatedly seen investors look at companies and have a lot of confidence in the businesses. Then they've gone south after that. These are very smart people doing very thorough analysis. The combination of all that led me to not want to own 100 stocks but I wanted to increase the number. We used to have 10 stocks make up let's say 80 percent of the portfolio. We bumped that up to a typical bet of five percent which means that we would end up with somewhere between 20 and 30 stocks in a portfolio.


First change we made is that a full-size position for us went from ten percent to five percent. The second change is that we introduced these basket bets, which was some of these bets at two percent. That's been in general, good for us. It's worked all right. In fact, I brought this up at a lunch I had with Charlie Munger in 2009. I said, "You know Charlie, I made this change in my portfolio where I've gone from ten percent to five percent." He stopped me. He didn't even let me finish. He said, "You know, you're going to ask me something which goes against all my natural tendencies. I'm a guy who likes to own three stocks in a portfolio and that's it. There's no need for a fourth stock."


I said, "Yeah, but let me finish my sentence." I explained to him how we had these big blow-ups with these things going to zero, etc. I said, "I'm not going to 100 stocks but maybe pushing it to 20 seems to be more effective." He then changed his mind. He said, "Well, if you look at Berkshire Hathaway (BRK.A)(BRK.B) and if you think of every holding as a stock, … if you think of See's Candy as a stock, Geico as a stock or even Coke as a stock …." He said, "You can look at our top 20 holdings, and by the time you get to the 20th holding, you'll get into about 75, 80 percent of capital."


He felt that in effect Berkshire was following that type of a concentration model. He said, "It seems okay." In the end it was okay to get Charlie to at least say it was okay. In hindsight, it was the right decision. It seems to be working out fine.