We seek to improve the measurement of the dynamic causal effects of expectation shocks by addressing issues related to data uncertainty. The expectations shocks are estimated in a mixed-frequency VAR model which incorporates monthly and quarterly economic and financial indicators. The VAR is estimated on real-time data to prevent the shocks being confounded with the effects of data uncertainty. But dynamic responses are calculated using a quarterly VAR for revised data, estimated using older vintages as instruments to account for the fact that ‘true values’ of key macroeconomic variables may never be observed. We show that expectations shocks – revisions in GDP expectations unrelated to changes in current economic fundamentals and orthogonalized to other, potentially related shocks – explain 7–8% of the two-year variation of output, investment, consumption and hours. This is similar to the proportion of business-cycle variation explained by monetary shocks, for example.

我们试图通过解决与数据不确定性相关的问题来改进对预期冲击的动态因果效应的度量。在混合频率VAR模型中估计了预期冲击，该模型结合了每月和每季度的经济和财务指标。VAR是根据实时数据估算的，以防止震荡与数据不确定性的影响相混淆。但是动态响应是使用修正后数据的季度VAR计算的，使用较旧年份作为工具进行估算，以说明可能从未观察到关键宏观经济变量的“真实值”这一事实。我们表明，预期冲击-GDP预期的修正与当前经济基本面的变化无关，并且与其他可能相关的冲击正交-解释了产出，投资，消耗和时间。例如，这类似于由货币冲击解释的商业周期变化的比例。