View article

Using the ECB Survey of Professional Forecasters to characterize expectations at the micro-level, we emphasize two new facts: forecasters (i) fail to systematically update their forecasts and (ii) disagree even when updating. It is moreover found that forecasters have predictable forecast errors. These facts are qualitatively supportive of recent models of inattention and suggest a setup where agents imperfectly process information due to both sticky information à la Mankiw–Reis, and noisy information à la Sims. However, building and estimating such an expectation model, we find that it cannot quantitatively replicate the error and disagreement observed in the data.