In part the financial sector does the equivalent of writing "naked puts," namely taking risks which usually yield extra income but occasionally blow up and bring large losses, part of which are socialized.
He is trying to respond to Kevin Drum, but I think he overlooks one of Drum's issues.
But this is mysterious. After all, not everyone is going short on volatility. In fact, by definition, only half of the punters on Wall Street are doing it. The other half are taking the other side of the bet.
He raises a question that it is instructive to answer. I think that one answer is that those who write uncovered options make there money on volume, so that when they lose the winners are very dispersed, and when they win the losers each lose very little. I would argue that with the term premium and the risk premium, we are talking about one-sided bets.
First, note the following. Suppose you and I make a bet against each other that is so large that neither one of us can pay if we lose, but if we lose the government will make good our bet. Then, I assume it is clear that making this bet is a good idea for both of us. Now, of course, the government will not do that for you and me, but if we were "too big to fail," ....
Further thoughts are below.
Concentration on one Side, Dispersion on the Other
If I offer flood insurance in New Orleans on behalf of my company, my bet might be "There won't be another Katrina in 2011." Let's say that we lose $1 billion if I am wrong, and we win $1 million (in insurance premiums) if I am right. If the chance of another Katrina is one out of 1000, that is a fair bet. But I can choose to make that bet even if the chance is 1 out of 50. The chances are 49 out of 50 that this deal will show a nice profit and I can get a fat bonus, and 1 out of 50 that I lose my job and the shareholders take big losses. A reasonable deal--at least for me.
When I do that, who is on the other side of the bet? In some sense, it is the people of New Orleans, although they are neither enriched by winning the bet nor impoverished by losing it. The people with the big stakes in the bet are me (because of my bonus) and the shareholders (because of the huge loss they may take). The people of New Orleans and the shareholders are dispersed. The effect on my income is highly concentrated.
Imagine I offer a lottery where if one ticket wins, they all win. We will pick two random numbers between one and one hundred, and if the numbers happen to match, all the tickets win. So the chance of winning is 1 in 100. Each ticket pays $1000 if it wins, and each ticket costs $1. People who buy the tickets will, on average, do well. Suppose that 10 million people buy tickets. The chances are 99 out of 100 that I will walk away with $10 million. The chances are 1 out of 100 that I will lose $10 billion, which will be dispersed among 10 million people.
Assuming I pay up if I lose, the distribution of income is still pretty unequal. There is a 99 out of 100 chance that the lottery will make me a millionaire 10 times over.
It gets even worse if you assume that I won't pay if I lose. In that case, if I lose then either the ticket-buyers get shafted or else the taxpayers get shafted as part of a bailout.
Next, consider "riding the yield curve." In the good old days, managing a savings and loan was a "three-six-three" operation. Pay 3 percent on deposits, earn 6 percent on mortgages, and hit the golf course at 3 PM. Let's say that your assets are $1 billion, and your costs are $10 million, or 1 percent. So you make 6-3-1 = 2 percent on $1 billion, or $20 million. Nice.
This works fine as long as long-term interest rates remain above short-term rates, as they typically do. But suppose that the rate on deposits goes up to 6 percent, and the rate on your mortgages stays at 6 percent. Now, you earn nothing on the spread, and you lose $10 million in expenses.
Your bet that long-term rates would stay above short-term rates lost. Who won? The mortgage borrower? You could say he won if mortgage rates went up to 8 percent and he enjoyed his 6 percent loan. But that is not what happened.
Who is on the other side of the bet when you ride the yield curve? The mortgage borrowers probably win, but not necessarily. If the mortgage rate had gone up to 8 percent, then you can say that the household with the 6 percent loan is better off. But in the example I gave, where mortgage rates stay at 6 percent, the borrower has not gained anything.
What happened was that the "term premium" (the difference between long-term rates and short-term rates) disappeared. The S&L was long the term premium, but nobody else was short the term premium. In some very deep sense, the nonfinancial sector was short the term premium, but not in a sense that shows up in cash or bonus potential.
The Risk Premium
The other big source of one-sided bets is the risk premium. Corporate bonds in the Baa range might typically trade at a spread of 200 basis points over Treasury. Some of this is "expected default cost" and some of it is a premium over and above the expected default cost. The latter is the risk premium. If you are a AAA-rated firm, you can borrow at AAA rates to buy Baa paper and earn an expected profit equal to the risk premium. You are making a bet that the risk premium will not go up. Again, this is a one-sided bet. If it turns out badly for the AAA-rated firm, there is no "other side" that makes a profit.