If effectiveness lags are long and variable, should policymakers use the current values or their best forecasts of target variables to determine policy? In evaluating the behavior of the Fed between 1994 and 2001, does it seem more likely that current or forecasted values of inflation and output determined their policies?