Two distinct questions concern the desirability of constraints on the monetary authorities. (1) To what extent should the central bank be constrained, versus being allowed full discretion? (2) To whatever extent it is to be constrained, should it be by a rule and if so what rule? With respect to the second question, a good argument for Nominal GDP targeting is that it is robust with respect to supply shocks, whereas CPI targets, for example, are vulnerable to them. But with respect to the first question, I am increasingly convinced that the constraint – whether a NGDP target or something else – must be very loose. Even the most sincere of central bankers will often fail to hit their targets, due to unforeseen shocks. I therefore propose only a mild innovation: the FOMC could include nominal GDP in its Summary of Economic Projections. A final thought concerns a different kind of constraint: if Fed independence from political influence is compromised, monetary policy will likely become more pro-cyclical.