Top 5 Ranking and heterogeneity in monthly inflation expectations in the Focus Survey: an agent-based computational analysis

We propose an agent-based model (ABM) to represent the strategic environment of reputation gain induced by the short-term Top 5 Ranking (RT-5) regarding inflation measured by the IPCA of the Focus Survey, which is carried out by the Central Bank of Brazil. In this computational model, we treat the formation of inflation expectations as a discrete choice process of forecasting strategies, in which agents have the incentive to choose over time the best forecasting strategy due to the reputation gain arising from the RT-5. Based on the proposed ABM, we found that the market expectations system becomes more effective with respect to the forecast of inflation in the presence of the short-term RT-5 of the Focus Survey, when compared to the case in which the incentive of reputation gain generated by the RT-5 is not considered. The results of the computer simulations also showed that, in general, the proposed ABM can generate aggregate properties of the respondents’ inflation expectations that have been detected by the empirical literature, among them the persistent heterogeneity of these expectations over time.

Keywords:
Agent-based computational model; Heterogeneous expectations; Focus Research; Inflation