The secret of Warren Buffett's success at Berkshire Hathaway
It's not all just about being a megabrain stock-picker
Worstall @ the Weekend We've just had the 50th anniversary of Warren Buffett and Charlie Munger taking over Berkshire Hathaway and that means we've also just had their 50th letter to shareholders. And this is the point where we try to puzzle over just how two guys managed to make so many hundreds of billions (if we include their own holdings and those gains that their shareholders made).
The truth is, there's less to the standard explanations than we might think and more to two little details from the more recondite corners of economics.
The first and most obvious possible explanation is that, well, there is variance in investing performance. The efficient markets hypothesis, that no one can beat the market over long periods of time, might be exactly right. But, however, if you roll the dice enough times, you'll get a series of 12s... roll them enough times and you'll get whatever series of 12s desired. So, it's within the realm of possibility (even probability) that someone or other will manage that feat, of beating the market over 50 years. Given enough random monkeys throwing darts at the finance pages, why not?
That is possible but not exactly a pleasing explanation: especially if you're trying to work out how one might replicate that result. The second could be that what Buffett has been saying he has been doing all along really is what should be done. That is: look for value stocks, ones with strong cashflow that appear to be undervalued, buy them and sit on them. That is, to an extent, what he has done, so why not take him at his word?
But that doesn't quite fit. Partly because it's simply a repetition of that first argument, someone has found a way to beat the market. Secondly, it doesn't explain the scale of his earnings over time. Buffett, after earlier adventures, had about $1m to play with 50 years ago. The returns since then have been good (19.4 per cent compounded in book value and I'll let you check to see whether this next statement is true). That's very good, but that shouldn't lead to $60bn for Buffett alone. Obviously there's something else going on here.
At this point we should look at that 50th letter to shareholders: the two secrets are in there. The first is this (PDF):
Berkshire’s huge and growing insurance operation again operated at an underwriting profit in 2014 – that makes 12 years in a row – and increased its float. During that 12-year stretch, our float – money that doesn't belong to us but that we can invest for Berkshire’s benefit – has grown from $41bn to $84bn. Though neither that gain nor the size of our float is reflected in Berkshire’s earnings, float generates significant investment income because of the assets it allows us to hold.
Actually, both of the reasons are in that small snippet. The first is leverage and the second is that the insurance market in the US is not competitive.
Competition - who needs it?
To take the second first. As Buffett points out, if you run an insurance company you get a float. These are the premiums that you receive and then sit on for some time before you have to pay out on the claims. How long you get to sit on them for depends upon what sort of insurance you're doing.
Car insurance claims, for example, come pretty thick and fast. Most contracts are annual and most of these premiums are going to be paid back out to the risk pool in that year. But, say, doing reinsurance on earthquake insurance (and reinsurance is a big Berkshire Hathaway line) might mean that you hang onto the premiums for 20, 30 years before having to pay out. The length of time that you've got that float for thus depends upon the time expected between premium collection and claim payout.
But, as Buffett notes, while that money doesn't “belong” to the insurance company, the profits to be made from investing it do. So, in a competitive market for insurance, we would expect underwriting losses.
Think of it this way: we have lots of firms competing to get insurance business, to underwrite policies. Everyone in the industry knows that there are those two potential profit streams. There are potential profits on the amount of premiums collected and the investment profits from carrying the float. We would expect the lure of the second to compete down the profits to be made from the first. And that's generally what we do find in such businesses. The actual insurance business itself, the underwriting, makes a small loss. This is offset by those investment profits so that the company as a whole is making around the normal rate of return on capital.
Being a futures broker, like the late lamented derivatives trader MF Global, works in very much the same manner. Speculators have to put up a cash margin against their trades; the broker is really just there to do the paperwork on the trades that the speculators decide they wish to execute. But to the broker, the paperwork side is (usually) money-losing because there are so many people competing to be able to make the investment gains from those cash margins lying around. And this is what killed MF Global: in a low-interest-rate environment, it was not making a big enough profit on that cash float. So off it went looking for higher returns, something that inevitably means taking higher risks. It got into Greek bonds....and yes, lost that float and thus went bust.
We can also run this the other way around. If we see underwriting profits in an insurance business, then we can conclude that that insurance business is not truly competitive. So it is with the US insurance market. It's not actually a national market, it's 51 different ones (50 states plus DC) and each of them has their own insurance regulator. Who determines what should be in a policy often determines rates and so on. And, of course, like all regulators usually end up, there's a certain amount of industry capture there. Very few US insurance markets are what we would call truly competitive at the detailed level.
Yes, we do think that partaking in something more like this sort of setup than a true free market is likely to be profitable.
Sponsored: Beyond the Data Frontier