Sunday, December 02, 2007

Mr. Buffett's Model of Circle of Competence

From: Sanjay Bakshi [
Sent: Sunday, December 02, 2007 8:39 AM
To: Ashutosh Datar
Subject: Re: Circle of competence - what is it?
Hi Ashutosh,
Hope all is well.
Here are my comments on your mail:
  1. Mr. Buffett's "circle of competence" has several meanings. One is simply the need to not stray outside an area of one's own skills. The probability of success is much greater if one's does not stray outside one's circle. Mr. Buffett likes to say he has a very large "too tough" basket where most of the things go. But occasionally, he comes across something compelling which he really understands. 
  2. Risk: Mr. Buffett's framework of risk has the circle of competence model at its heart. In his 1993 letter, in which he wrote about his thoughts on risk (defined as probability of permanent capital loss, as opposed to mere quotational losses), he mentioned that risk depends to a large degree on one's ability to accurately predict the future about the situation being examined. He connects the circle of competence model to his idea of risk by relating it to the certainty with which the long-term prospects of the business, the operating skills, the capital-allocation skills, and the integrity of the management can be ascertained by an investor. So, the same investment, may carry different levels of risk for different investors. This, as you know, is contrary to what academic finance teaches. Mr. Buffett, however, believes, that risk comes from not knowing what one is doing. 
  3. I don't think the circle of competence model is industry specific. For example, I consider my own circle of competence in the deep value space and within that space,  I have sort of found my calling in the field of special situations. Over the years, I have accumulated a substantial experience in those fields. I think its possible to find deep value in a sector which you do not have to understand to the degree required of a sell-side research analyst.
  4. My own view is that sell-side analysts often go wrong because they know too much - they have too much information to work with and as a result they often miss the obvious, which can easily, perhaps, be observed by a generalist, who maintains a distance from the noise that surrounds sell-side research. There is a very famous experiment in psychology called the "fire hydrant" experiment. A class of students is divided into two equal groups and each group is separately shown a very blurry image of a fire hydrant which is unrecognizable. Then, the image is brought into focus and the process is stopped at a point well before the image is totally clear to anyone. However, for one group, the image is brought into focus in ten increments, while in the other group its brought into focus in five increments. Note that both groups start with the same image and end with the same image. Surprisingly to many, the group which had more information (ten increments) take much longer to recognize the fire hydrant than the group which had less information. The chief reason is that the members of group which was exposed to more increments come to initial judgments (first conclusions) derived from the pattern-seeking human brain, and then in subsequent information it tend to under-weigh disconfirming evidence and over-weigh confirming evidence of their initial notions (confirmation bias). The second group, on the other hand, does not suffer from these biases. They see no relation between the first image and the next and the next and suddenly they see something that resembles a fire hydrant. The second group - the one with less information - identifies the fire hydrant- much more quickly than the first group which had too much information. The analogy with sell-side research is obvious, in my view. 
  5. Cost of capital - Circle of competence has much to do with cost of capital. The opportunity of cost an investment is the expected return of the next best investment which is available to me. The key phrase is "available to me". Now, most of the things that happen in the stock market and fall outside my circle of competence, are not "available to me" so they must not influence my thinking about opportunity costs. This is fundamentally logical, in my view, and is totally at odds with Markowitz's model of portfolio optimization. 
  6. This means that there will always be some strategy which will be doing far better that what I know and understand. Does that mean that I must run after that strategy? No! So long as my own strategy, derived out of my own circle of competence, has delivered me satisfactory results and is expected to do so in the future as well, I don't have to be envious (envy being the deadliest of the seven deadly sins), I should be contended. This is consistent with the Herb Simon's idea of satisficing as an application of his theory of bounded rationality. 
  7. One's own circle of competence expands over time through personal and vicarious experience - the latter one obtained primarily through careful reading of extreme successes and failures - the lollapalooza outcomes - and relating them to the causal factors (mental models) which combined to produce those outcomes. In this respect, vicarious experience is terribly important, according to Mr. Munger, for "you do not have to to do it to learn not to pee on an electrified." :-)
  8. You really have to choose a field that you feel passionate about - it could be a sector, it could be a strategy (deep value, special situations, growth, momentum - whatever), it could be top-down or it could be bottoms up - It does not matter. What does matter is that you have real passion for it, and you have determination to build a circle of competence in that space and to expand it over time through the accumulation of experience - both direct and vicarious.
I'd like to blog this thread because there may be other people who may benefit from this exchange. If you prefer, I can protect your identity from being revealed when I post it. Do let me know.
Thanks and best wishes
Sanjay Bakshi
On 25-Nov-07, at 11:27 AM, Ashutosh Datar wrote
Hello Sir
I am struggling with this concept for the past few weeks and so writing to you to get your perspective.
Circle of competence is a good and a fairly intuitive concept but I am not clear what it means in actual sense…

If by it we mean knowing one particular sector inside out, then the typical sell side analyst should be the best guy. He covers a particular sector, does not go beyond it and pretty much knows every small thing that is happening in it. But then he doesn’t seem to produce good results – what’s the problem there? Does he know too much? Or does he not know the right stuff
  • WB talks about buying businesses that he understands, buying businesses that are simple, what does it mean? I mean Coca-cola is a simple business but Petrochina I guess is a whole lot complicated than that… add the political risk of dealing with ppl as opaque as Chinese and it seems kinda odd to me.
  • Is circle of competence necessarily a sector thing? I mean someone could say I can call the interest rate cycle or politics for that matter, so he could use it for a wide range of sectors which are affected by that…
  • Lastly, how does one go about developing his/her own circle/niche? Is it just a natural process driven by time?
Could you help?
Best Regards
Ashutosh Datar


Rohit said...

Ashutosh, a simple reason that analysts do not produce good results is that they are short term focussed -- their quarterly/annual bonuses depend on it! (incentive caused bias ). Good results typically require having a multi-year view of a company.

If you understand an area (or company) well enough to see the long term direction and are not bogged down in predicting hundreds of variables on a quarterly basis, and are confident enough around your analysis to invest and patiently wait out your predictions you should do well..

Rana Pratap said...


The other point of view is that if I know a company really well then I can REALLY take advantage of slightly shorter duration events and happenings !!!

To put it down simply what I mean is -
If you understand an area (or company) well enough to see the 6 month direction and are not bogged down in predicting hundreds of variables on a monthly basis, and are confident enough around your analysis to invest and somewhat patiently wait out your predictions you should do well..

Now over a day, there is a lot of noise, a month, slightly lesser, and over 6 months reasonably less

I am not talking from a sell side guy's point of view but am talking about a chap looking at tripped over companies, complex corporate event happening etc etc situations

I admire all the Long Term value investors but in a world where competitive advantages are shrinking by the day and moats disappearing I wonder for how long the long term will remain long term.

There are many ways to skin a cat, and having a multi year outlook just happens to be one of them

Somebody just forgot to tell the special situation funds or Mr Buffett who dabbles in the things that I described in his personal portfolio. And he beats the pants off Berkshire's returns

"If you understand an area (or company) well enough to see the long term direction" - I can see the long term direction of MTNL everyday but that has not stopped me from generating returns from it

Good results typically do require multi year periods but nobody specified that it means your investment horizon, It can be fine tuning of the process also

Luv and Kisses
PS - Absolutely love your blog

VISHNU said...

Best example is pharmaceutical

It is very difficult to think about the sales of Lupin , Insulin (not sure about the names :) ) and their prices in 2010 ..

But definetley a doctor or pharmacist might be having a good idea about these products and companies, Hence can pick a good company or Stock

Exception to this rule is a Cash Bargain..(As Prof Bakshi says its very hard to argue with CASH)