We measure the relative contribution of the deviation of real activity from its equilibrium (the gap), "supply shock" variables, and long-horizon inflation expectations for explaining the U.S. inflation rate in the post-war period. For alternative specifications for the inflation driving process and measures of inflation and the gap we reach a similar conclusion: the contribution of changes in long-horizon inflation expectations dominates that for the gap and supply shock variables. Put another way, variation in long-horizon inflation expectations explains the bulk of the movement in realized inflation. We also use our preferred specification for the inflation driving process to compute a history of model-based forecasts of the inflation rate. For both short and long horizons these forecasts are close to those observed from surveys.