In this paper we apply a dynamic factor model to generate out of sample forecasts for the inflation rate in Mexico. We evaluate the role of using a wide range of macroeconomic variables with particular interest on the importance of using CPI disaggregated data to forecast inflation. Our data set contains 54 macroeconomic series and 243 CPI subcomponents from 1988 to 2008. Our results indicate that: i) Factor models outperform the benchmark autoregressive model at horizons of one, two, four and six quarters, ii) Using disaggregated price data improves forecasting performance, and iii) The factors are related to key variables in the economy such as output growth and inflation.
Factor models Inflation forecasting Disaggregate information Principal components Forecast evaluation