Central banks are increasingly communicating their economic outlook in an effort to manage the public and financial market participants’ expectations. We provide original causal evidence that the information communicated and the assumptions underlying a central bank’s projection can matter for expectation formation and aggregate stability. Using a between-subject design, we systematically vary the central bank’s projected forecasts in an experimental macroeconomy where subjects are incentivized to forecast the output gap and inflation. Without projections, subjects exhibit a wide range of heuristics, with the modal heuristic involving a significant backward-looking component. Ex-Ante Rational dual projections of the output gap and inflation significantly reduce the number of subjects’ using backward-looking heuristics and nudge expectations in the direction of the rational expectations equilibrium. Ex-Ante Rational interest rate projections are cognitively challenging to employ and have limited effects on the distribution of heuristics. Adaptive dual projections generate unintended inflation volatility by inducing boundedly-rational forecasters to employ the projection and model-consistent forecasters to utilize the projection as a proxy for aggregate expectations. All projections reduce output gap disagreement but increase inflation disagreement. Central bank credibility is significantly diminished when the central bank makes larger forecast errors when communicating a relatively more complex projection. Our findings suggest that inflation-targeting central banks should strategically ignore agents’ irrationalities when constructing their projections and communicate easy-to-process information.

中央银行越来越多地传达其经济前景，以努力管理公众和金融市场参与者的预期。我们提供了原始的因果证据，表明所传达的信息和中央银行预测的假设可能对预期形成和总体稳定性很重要。使用主体间设计，我们系统地改变中央银行在实验性宏观经济中的预测预测，其中激励主体预测产出缺口和通货膨胀。如果没有投影，受试者会表现出广泛的启发式方法，其中模态启发式方法涉及一个重要的后视组件。产出缺口和通货膨胀的事前理性双重预测显着减少了受试者使用后向启发式方法的数量，并将预期推向理性预期均衡的方向。事前理性利率预测在认知上具有挑战性，并且对启发式分布的影响有限。自适应双重预测通过诱导有限理性预测者使用预测和模型一致预测者将预测用作总体预期的代理，从而产生意外的通胀波动。所有预测都减少了产出缺口分歧，但增加了通胀分歧。当中央银行在传达相对更复杂的预测时出现较大的预测错误时，中央银行的可信度会显着降低。