While I was reading the transcripts of Berkshire’s 1994 annual meeting, I bumped into a striking discussion on risk, which should really be common sense but is often ignored by many investors. So today I will share the Q&A part of the annual meeting specifically related to the topic and my thoughts on the subject.
Question: I’d like to expound on your view of risk in the financial world, and I ask that against the background of what appear to be a number of inconsistencies between your view of risk and the conventional view of risk. I mention that in a recent article you pointed out inconsistency in the use of beta as a measure of risk, which is common standard. And I mention that derivatives are dangerous, and yet you feel comfortable playing at derivatives through Salomon Brothers. And betting on hurricanes is dangerous, and yet you feel comfortable playing with hurricanes through insurance companies. So it appears that you have some view of risk that’s inconsistent with what would appear on the face of it to be the conventional view of risk.
Warren Buffett (Trades, Portfolio): Well, we do define risk as the possibility of harm or injury. And in that respect we think it’s inextricably wound up in your time horizon for holding an asset. I mean, if your risk is that you’re going – if you intend to buy XYZ Corporation at 11:30 this morning and sell it out before the close today, that is a very risky transaction. Because we think 50% of the time you’re going to suffer some harm or injury.
If you have a time horizon on business, we think the risk of buying something like Coca-Cola (KO, Financial) at the price we bought it a few years ago is essentially, is close to nil, in terms of our perspective holding period. But if you asked me the risk of buying Coca-Cola this morning and you’re going to sell it tomorrow morning, I say that is a very risky transaction. Now, as I pointed out in the annual report, it became very fashionable in the academic world, and then that spilled over into the financial markets, to define risk in terms of volatility, of which beta became a measure. But that is no measure of risk to us.
The risk, in terms of our super-cat business, is not that we lose money in any given year. We know we’re going to lose money in some given day, that is for certain. And we’re extremely likely to lose money in a given year. Our time horizon of writing that business, you know, would be at least a decade. And we think the probability of losing money over a decade is low. So we feel that, in terms of our horizon of investment, that that is not a risky business.
And it’s a whole lot less risky than writing something that’s much more predictable. Interesting thing is that using conventional measures of risk, something whose return varies from year to year between plus-20% and plus-80% is riskier, as defined, than something whose return is 5% a year every year.
We just think the financial world has gone haywire in terms of measures of risk. We look at what we do — we are perfectly willing to lose money on a given transaction, arbitrage being an example, any given insurance policy being another example. We are perfectly willing to lose money on any given transaction. We are not willing to enter into transactions in which we think the probability of doing a number of mutually independent events, but of a similar type, has an expectancy of loss. And we hope that we are entering into our transactions where our calculations of those probabilities have validity. And to do so, we try to narrow it down.
There are a whole bunch of things we just won’t do because we don’t think we can write the equation on them. But we, basically, Charlie and I by nature are pretty risk-averse. But we are very willing to enter into transactions — we, if we knew it was an honest coin, and someone wanted to give us seven-to-five or something of the sort on one flip, how much of Berkshire’s net worth would we put on that flip? Well we would — it would sound like a big number to you. It would not be a huge percentage of the net worth, but it would be a significant number. We will do things when probabilities favor us. Charlie?
Charlie Munger: Yeah, we, I would say we try and think like Fermat and Pascal as if they’d never heard of modern finance theory. I really think that a lot of modern finance theory can only be described as disgusting.
I write this as a reminder to myself, as well as to all the readers. When dealing with risks, we are all familiar with the concept of “worst case” or permanent capital losses. But what’s often missing is the time horizon element.
As Buffett said, risk is inextricably wound up in your time horizon for holding an asset. So are worse cases and permanent capital losses. Some investors talk about how they try to avoid permanent capital losses without giving a time horizon and some realized permanent capital losses precisely because their holding period is too short.
Furthermore, some may miss the great compounders because the “worse case” in one year seems always high based on historical valuations. But as the “worse case” goes up every year as fundamentals improve, the risk of buying a great business is close to nil over a 10-year period of time, if you don’t pay a ridiculous multiple. The key question is, can you understand the business and can you hold it for 10 years?
The practical implication of Buffett’s view of risk being inextricably wound up in the time horizon for holding an asset is that having a defined time horizon is also critical in risk assessment and management, besides the business and management team elements. A mismatch between the intended time horizon for the risk involved and the actual or practical time horizon for the risk can be the source of permanent capital loss.
We can easily see how this issue is avoidable, yet it happens more often than people think in the investment world, particularly when the time horizon of the clients is different from that of the portfolio manager.
Read more here:Ă‚
John Bogle: Why Index Funds Outperform
