"I was never under any illusion about the theoretical possibility that the Phillips curve is vertical. In fact, I could cite passages that precede Friedman and Phelps, which say that it could well be. But all the empirical findings before 1966 suggested that the feedback term from price inflation to wages was considerably less than one so I accepted a nonvertical Phillips curve. I was probably overoptimistic about the degree to which you could expand employment in the '60s. In fact, I have acknowledged that in print. We at the council had stated 4 percent as what would now be called the natural rate of unemployment, I suppose, with a nonaccelerating rate of inflation. We did get the 4 percent before the Vietnam War, and we didn't have any accelerating rate of inflation. Maybe that was because we had an incomes policy. It was a rather weak one; it was sometimes called "open-mouth" policy. It was an informal guide-post policy, but in retrospect it looks like it made more difference than people credited it for. At any rate, the acceleration of inflations came when Johnson disobeyed his economic advisers. We went down to 3 percent unemployment which was probably well below the natural rate. So, I could admit to having been too optimistic."