I've written some stuff on why value investing isn't maybe the best thing in the world, and how Berkshire's S&P 500 equity puts are threatening the company. Felix Salmon makes a really obvious point I wish I thought up about these contracts. The idea is that money is worth different amounts in different periods of time, so by taking the tail risk that things go really bad, Berkshire faces an extraordinary loss: Things go really bad, and Berkshire has to pay out on the contract. And the probability of things going really bad is tough to tell from past data.
This is indicative of a broader issue with Berkshire and value investing. Berkshire the business specializes in taking exactly these sorts of extreme catastrophic risks. Value investing in general specializes in exactly those companies which go under in times of extreme stress. Value stocks do bad in a crisis, and this fact pushes up the premium for these stocks in good times (who wants to own something that is worth so much less exactly when you need money?). Only when there's a crisis do you see--as you do now with United Health, the rail companies, GE, American Express--how bad things can get.
This can still be a smart bet to make. Thomas Sargent has some interesting Bayesian calculations in which different actors share risk. Normal people trade away risk, while other agents take these risks and make out like bandits in good states. But it's a bet that makes use of a broader story: Stock markets are as close to information-efficient as you like, and the whole story of day-to-day fluctuations is driven by differing risk premia. It's very very difficult to make money picking and choosing stocks (More on this later...), but there is a role for smart managers to pick up different sets of risk exposure. In some extreme sense, there is no other way to make sense of investing.