Response Bias in Survey Measures of Expectations: Evidence from the Survey of Consumer Expectations' Inflation Module

The Survey of Consumer Expectations (SCE) infers respondents' inflation expectations from density forecasts. Using numeracy data and tests for coherence among 117,000 respondents to the SCE, I find that density forecasts suffer non-negligible reporting bias and selective nonresponse. A simple verbal question collected by the SCE suffers neither of these deficiencies and so has better properties to deliver an accurate snapshot of the population's inflation expectations than the headline measures of inflation expectations published by the SCE. I demonstrate how the verbal measure can be harnessed to improve the signal-to-noise ratio in density forecasts.