Template-Type: ReDIF-Paper 1.0
Author-Name: Alberto Vindas-Quesada
Author-Name-First: Alberto
Author-Name-Last: Vindas-Quesada
Author-Email: vindasqa@bccr.fi.cr
Author-Workplace-Name: Department of Economic Research, Central Bank of Costa Rica
Author-Name: Carlos Brenes-Soto
Author-Name-First: Carlos
Author-Name-Last: Brenes-Soto
Author-Email: brenessc@bccr.fi.cr
Author-Workplace-Name: Department of Economic Research, Central Bank of Costa Rica
Author-Name: Adriana Sandí-Esquivel
Author-Name-First: Adriana
Author-Name-Last: Sandí-Esquivel
Author-Email: sandiea@bccr.fi.cr
Author-Workplace-Name: Economic Division, Central Bank of Costa Rica
Author-Name: Susan Jiménez-Montero
Author-Name-First: Susan
Author-Name-Last: Jiménez-Montero
Author-Email: jimenezms@bccr.fi.cr
Author-Workplace-Name: Department of Economic Research, Central Bank of Costa Rica
Title: Univariate inflation forecasts in Costa Rica: model evaluation and selection
Abstract: This document presents the methodology that the Central Bank of Costa Rica uses to evaluate and select the univariate models for short-horizon forecasting purposes. This methodology consists on cuantifying several properties that are deemed desirable for forecasting models, assigning scores and combining them to obtain a final score. The robustness of the model selection to the evaluation period is analyzed, given the recent inflation dynamics. The selection is sensitive to this period, leading to the recommendation of regular selection processes. ***Resumen: Esta nota presenta la metodología que usa el Banco Central de Costa Rica para la evaluación y selección de los modelos univariados de proyección de inflación en el corto plazo. La metodología consiste en la cuantificación de varias propiedades deseables en modelos de pronóstico, la asignación de puntajes y su combinación para obtener un puntaje final. Se evalúa la robustez de la selección de los modelos a cambios en el periodo de evaluación, dados los cambios recientes en la dinámica inflacionaria. La selección del modelo es sensible a este periodo, por lo que se recomienda actualizar la selección con regularidad.
Length: 41 pages
Creation-Date: 2024-10
Publication-Status: Published
File-URL: https://repositorioinvestigaciones.bccr.fi.cr/handle/20.500.12506/399
File-Format: Application/pdf
Number: 2405
Classification-JEL: E31, E37, C52, C53
Keywords: Inflation, Forecasting, Stochastic Volatility, GARCH, Evaluation, Pronóstico
Handle: RePEc:apk:nottec:2405