Because economic outcomes depend on private-sector expectations, central banks might be tempted to guide these by publishing projections of key macroeconomic variables. We find that optimal projections require misleading the public. Optimal non-misleading projections are time-inconsistent. Non-misleading time-consistent projections can only improve policy outcomes if the public’s forecasts are noisier, or inconsistent with implemented policy. Since the public only has incentives to be guided by policymakers’ projections when most vulnerable to being mislead, these cannot be trusted blindly. Consistent with this, we find statistically significant systematic deviations between FOMC projections and professional forecasts for US inflation and GDP growth.
