An emerging and influential literature finds a large and significant decline in macroeconomic volatility since the middle of the 1980's. In this paper, I examine the extent to which the decline in annual and quarterly real output volatility since the onset of this period of Great Moderation can be attributed to changes in macroeconomic uncertainty and macroeconomic predictability. I use point forecasts of future real output growth from the Survey of Professional Forecasters (SPF) between 1969 and 2003 as a proxy for the predictable component of real output growth. The results indicate that declining predictability has played a significant role in the Great Moderation. Prior to the Great Moderation, professional forecasts explained roughly 30 percent of the variance in output growth. Post-moderation, the predictive ability of professional forecasts quickly vanished. This decline in predictability implies that interpreting the decline in the volatility of output shocks identified from a fixed parameter autoregressive model overstates the decline in macroeconomic uncertainty by between 20-40 percent. I also examine forecasts of the probability of a decline in real output from the SPF. Consistent with the findings from the point forecast data, these probability forecasts indicate that the decline in macroeconomic uncertainty as measured from an autoregressive model is overstated. While both the average probability of a decline in output and the uncertainty surrounding future declines in output computed from an autoregressive model decrease sharply after the mid-1980's, the SPF probability forecasts exhibit no such decrease. I assess the economic significance of the overstatement in the decline of macroeconomic uncertainty in terms of its effects on forecasts of the future equity premium. These results indicate that using the decline in the total volatility of real output growth along with the standard CCAPM model overstates the decline in the future equity premium by roughly 20 percent.