Multimodality Test and Mixture Distributions: An Application to the Central Bank Expectation Survey

This paper presents the Silverman multimodality test and mixture distributions methodology, applying both approaches to the Economic Expectations Survey of the Central Bank of Chile. The main results reflect the importance of permanently monitoring the implicit kernel distributions in each survey. Private analysts’ forecasts are found to have systematically been in line with the inflation targeting range, although during episodes where effective inflation fell outside the target range, the analysts encounter forecasting difficulties. Meanwhile, we observe the relevance of the Bank’s inflationary report for anchoring expectations, as the distributions become systematically more homogenous after publication.