Up until 2008 I had a sort of "Whig view" of Fed history. They made many mistakes, but learned enough from those mistakes to gradually improve. Now I'm not so sure.
Here are 6 policy regimes, all excessively procyclical:
1. Real bills doctrine (early years): Lend money as needed for productive investments. Unfortunately, more money is "needed" during booms, and hence monetary policy becomes more expansionary during booms and more contractionary during recessions. The drop in the monetary base from October 1929 to October 1930 is a perfect example.
2. Interest rate pegging (1942-51): If you hold rates absolutely fixed, then the money supply will move in tandem with money demand. Because money demand tends to rise during inflationary booms, the money supply will also rise and policy will be procyclical.
3. Interest rate targeting (1951-79): This is better, in principle. Adjust the interest rate as needed to stabilize the economy. In practice they relied too much on nominal interest rates, and hence misjudged the stance of policy. They also misjudged the natural rate of unemployment. Thus policy became too expansionary, as inflation reduced real interest rates and the Fed overestimated the economy's potential.
4. Money supply targeting (1979-82): This did bring inflation down, but velocity tends to fall during deep recessions like 1982, which makes the policy overly contractionary.
5. New Keynesian (1982 - 2008): The "Taylor principle" fixed the inflationary bias of the 1960s and 1970s. Until 2008 this was as good as it gets--even Milton Friedman praised the policy before he died. But interest rate targeting does not work at the zero bound, so the policy was abandoned in late 2008. And there were weaknesses exposed even before it was abandoned--particularly the lack of level targeting.
6. Unconventional policies (2008-??): These included interest on reserves, QE, and forward guidance. Here the weakness was passivity. The Fed was less willing to use unconventional than conventional policy tools. Because unconventional tools are needed at the zero bound, and because the zero bound tends to occur during deep recessions, policy ends up being too contractionary during deep recessions.
7. NGDP futures targeting, level targeting (?? - infinity and beyond): This represents the state of the art in terms of monetary stability. First we need to get an NGDP futures market up and running. Investors, policymakers and pundits will then begin using NGDP futures prices as proxies for expected growth in AD. This market forecast will gradually assume a larger and larger role in policy formation. Eventually its role will be formalized.
PS. Of course I was kidding about infinity. NGDPLT will run for a few decades, and then get replaced by something even better, which is as it should be.
PPS. Policy was also procyclical under the classical gold standard (1879-1913.) During recessions interest rates tended to fall. This increased the real demand for gold (due to a lower opportunity cost of holding gold), which was deflationary.