He differentiates two types bubbles: bank-centered bubbles and bubbles not based on bank lending. Central banks have more and better information on the bank-centered bubbles and have a lot of tools to deal with it rather than raising interest rate. For the non-bank-based bubble, central banks have no information advantage and monetary policy tools rarely "fit the crime". (source: NYT)
I would argue that the central bank's proper role is fundamentally different in the two types of bubbles. Here's why:
When bubbles are not based on bank lending, the Fed has no comparative advantage over other observers in distinguishing between rising fundamentals and bubbly valuations. It may see bubbles where there are none, or fail to recognize them until it's too late — or probably both.
Indeed, at the Fed, I recall Mr. Greenspan thinking that he saw a stock market bubble as early as 1995, when Internet stocks barely existed and the Dow was under 5,000. Fortunately, he did not make the mistake of trying to burst it. Conversely, the tech bubble became obvious only in 1999 — by which time it was already enormous.
That's the first problem, and it's a huge one. Here's the second:
Once a central bank correctly recognizes a bubble's existence, what is it supposed to do? The Fed has no instruments aimed directly at, say, tech stocks, and practically no instruments aimed at stock prices more broadly. (Those who argued that higher margin requirements would have worked were engaged in deeply wishful thinking.)
Of course, the Fed could have raised interest rates. But why would raising the federal funds rate by, say, two to three percentage points have ended the stock market mania when investors were expecting 19 percent annual returns in the stock market? That much monetary tightening, however, might well have stopped the economy in its tracks. If that strikes you as a good bargain, you might enjoy reading Cotton Mather's autobiography.
But a bank-centered bubble is starkly different in both respects.
As long as the central bank is also a bank supervisor and a regulator, it is extraordinarily well placed to observe and understand bank lending practices — much better positioned than almost anyone else. Beyond merely knowing more, part of a bank supervisor's job is to make sure that banks don't engage in unsafe and unsound lending, and to scowl at or discipline them if they do. We know that America's bank regulators fell down on the job as the housing-mortgage bubble inflated. But that was a failure of bank supervision, not of monetary policy.
AND what about instruments specifically aimed at the bubble? Whereas the Fed's kit bag is pretty much empty when it comes to stock-market prices, it is stuffed full when it comes to taking aim at bank lending practices. Escalating upward from a sternly arched eyebrow to an outright prohibition of certain types of lending — for example, subprime loans with no documentation for 100 percent of a home's appraised value — bank supervisors have a broad range of finely calibrated weapons at their disposal. Like the Mikado, they can "let the punishment fit the crime."
