Perhaps my favorite paper delivered at the American Economics Association meeting—in part because it had no math—was a historical essay from Christina Romer and David Romer on "The Most Dangerous Idea in Federal Reserve History."
The most dangerous idea, they say, is excessive pessimism about monetary policy. If you look back at the two key eras where we say monetary policy went awry—during the deflation of the 1930s and the inflation of the 1970s—the interesting thing that Romer and Romer find is that if you dig into the archives of the Federal Reserve minutes there weren't really "mistakes" as you might think of it. Policymakers in the '30s knew there was a deflationary slump, and they knew it was bad, just as policymakers in the '70s knew there was an inflationary spiral, and they knew it was bad. But in the '30s, policymakers persuaded themselves that with interest rates already low there was nothing more they could do, while policymakers in the '70s persuaded themselves that inflation represented a purely structural phenomenon that they couldn't cure. So you got a lot of talk about how other people need to step up.
The funny thing, they say, is that in both cases it turned out the problems could actually be solved quite quickly once you put someone in office who thought it was possible to solve them. The Volcker recession, in particular, though painful, was actually remarkably short and achieved its objective decisively. There's every reason to think you could have done the same thing in 1971 or 1978, and it probably would have been even quicker and less painful.
I would add to the Romer and Romer analysis a little international perspective from my reading of Hjalmar Schacht's memoir. Schacht played a key role in halting the Weimar inflation, and then after a period in the political wilderness, he was the mastermind of Adolf Hitler's early economic recovery strategy. Schacht's key retroactive insight in both cases is that he didn't really need to do very much at all. What he had to do was accept responsibility and step up. In neither case was resolve a panacea for economic problems, but it did essentially solve the monetary aspects of the problem with rapid speed and near-zero costs.
If I have any disagreement with the Romers, it's that they emphasize a disagreement about efficacy (as David Romer put it, "fear of impotence is bad for performance"), but I would emphasize an issue of accountability. Economists tend to worry a lot about incentives except when it comes to the behavior of economists. But the operational independence of the Federal Reserve means its personnel have a strong incentive in any troubled time to engage in a lot of blame-shifting and ducking of responsibility. But in the '30s and '70s and today, you're essentially facing problems of expectations management, and a central banker can't steer expectations effectively unless he's willing to put his reputation on the line.