Warren Buffett: What Most Investors Don't Understand About Risk

Can you please elaborate your views on risk? You clearly aren't a fan of relying on statistical probabilities, and you highlight the need for 20 billion dollars in cash to feel comfortable. Why is that the magic number, and has it changed over time?

Yeah, well, it isn't the magic number and there is no magic number. I would get very worried about somebody that walked in every morning and told us precisely how many dollars of cash we needed to be, you know, secured to three sigma or something like that. But Charlie and I have had a lot of discussions about this. We saw a lot of problems developing in organizations that expressed their risks in sigma, and we even argued sometimes about the appropriateness of how they calculated their risk. It was truly horrible, and they were a lot smarter than we were. That's what it is—depressing.

But we both have the same band of mind whereby we think about worst cases all the time, and then we add on a big margin of safety. We don't want to go back to go. I mean, I enjoy tossing those papers in the other room, but I don't want to do it for a living again. So we undoubtedly build in layers of safety that others might regard as foolish. But we've got six hundred thousand shareholders, and we've got members of my family that have 80 or 90 percent of their net worth in the company. I'm just not interested in explaining to them that we went broke because there was a 100th one percent chance that we would go broke, and the remaining probability was filled by the chance of doubling our money, and I decided that that was just a good gamble to take. We're not going to do that.

It doesn't mean that much. We are never going to risk what we have and need for what we don't have and don't need. We'll still find things to do where we can make money, but we don't have to stretch to do it. It is my job, and you know, Charlie thinks the same way. I mean, we don’t have to talk about it much, but it's our job to figure out what can really go wrong with this place. We've seen September 11th, and we've seen September of 2008. We'll see other things of a different nature but similar impact in the future. We not only want to sleep all of those nights; we want to be thinking about things to do with some excess money we might have around.

If you're calibrating it in some mathematical way, I would say it's really dangerous. I could give you a couple of examples on that, but unfortunately, I've learned about them on a confidential basis. Some really great organizations have had dozens of people with advanced mathematical training thinking about it daily, making computations, and they don’t really get the problem. So it's at the top of the mind always around Berkshire, and your returns in 99 years out of 100 will probably be penalized for being excessively conservative. One year out of 100, we'll survive when some other people won't.

Charlie: Yeah, but how do these super smart people with all these degrees and higher mathematics end up doing these dumb things?

I think it's explainable by the old proverb that to a man with a hammer, every problem looks pretty much like a nail. They've learned these techniques, and they just twist the problem to absolutely fit the solution, which is not the way to do it. They have a lack of understanding of history. I would say that one of the things, in 1962 when I set up our office in Keywood Plaza, where we still are, it's a different floor. I put seven items on the wall. Our art budget was seven dollars, and I went down to the library, and for a dollar each, I made photocopies of the pages from financial history.

One of those cases, for example, was in May of 1901 when the Northern Pacific Corner occurred, and it's kind of interesting in terms of being in Omaha because Harriman was trying to get control of the Northern Pacific, and James J. Hill was trying to control the Northern Pacific. Unbeknownst to each other, they both bought more than 50 percent of the stock. Now, when two people buy more than 50 percent of the stock each and they both really want it, they're not just going to resell it. Interesting things happen to the shorts. In that paper of May 1901, the whole rest of the market was totally collapsing because Northern Pacific went from 170 a share to one thousand dollars a share in one day, trading for cash because the shorts needed it.

There was a little item at the top of that paper, which we still have at the office, where a brewer in Troy, New York, committed suicide by diving into a vat of hot beer because he'd gotten a margin call. To me, the lesson that that fellow probably understood sigmas and everything and knew how impossible it was that in one day a stock could go from 170 to a thousand to cause margin calls on everything else, and he ended up in a vat of hot beer. I've never wanted to end up in a vat of hot beer.

Those seven days that I put up on the wall, life and financial markets has got no relation to sigmas. If everybody that operated in financial markets had never had any concept of standard errors and so on, they would be a lot better off. Don't you think so, Charlie?

Charlie: Well sure.

Here, have some questions. It's created a lot of false confidence, and now it has gone away. Again, as I said earlier, the business schools have improved; so has risk control on Wall Street. They now have taken the Gaussian curve and they just changed it away. They threw it away. Well, they just made a different shape than Gauss did. It's a better curve now, even though it's less precise. They talk about fat tails, but they still don’t know how fat they make them. They have no idea.

Well, but they knew that they learned through painfully. Yeah, they weren't fat enough. Yeah, they learned the other was wrong. Yeah, but they don't know what's right. But we always knew that there were fat tails.

Warren: At the Solomon meetings, we would look over at one another and roll our eyes when the risk control people were talking.

Okay, Jay, this question is on Swiss Re. Berkshire's quota share treaty with Swiss Re covering 20% of Swiss Re's property casualty risk ends in 2012. Is Berkshire planning to replace that premium volume through another transaction?

Well, we would hope to. We always hope to get more good volume, but what we do has no relationship to the expiration of that contract. I mean, that contract was a five-year contract. It's a big contract, billions of dollars a year. But the fact that that expires in our premium bomb, we've gone on by multiple billions, does not cause us to do one thing differently than we would do otherwise. We've got the capacity to write billions and billions of business, and we would love to do it if we were expanding the Swiss Re contract.

We don't want to write any dumb business when we lose that contract. It's just a non-event in terms of future strategy. It's not a non-event in terms of losing some business that we like, but it's not an event in terms of any future strategy. We regard every decision, you know, as independent. We don't do—if money comes in, that doesn't cause us to want to think about doing something today that we weren't thinking about doing the day before. We just don't operate that way.

We'll have things come along that are terrific, and that doesn't mean us that the next day we don’t want to look for something additionally that's terrific. Every decision is sort of independent. I don't think there's another large insurance operation in the world that is more cheerful about losing volume than we are. It doesn't make sense. The business has to shrink; we let it shrink. Yeah, we don’t measure ourselves in any way by size, by size except by the growth in value over time.