A Proposal of Flexible Trend Specification in DSGE Models
|Článek ve sborníku
|Proceedings of 30th International Conference Mathematical Methods in Economics
|Fakulta / Pracoviště MU
|DSGE model; trend specification; Bayesian estimation
|In this paper I propose flexible trend specification for estimating DSGE models on the log differences. I demonstrate this flexible trend specification on a New Keynesian DSGE model of two economies, which I consequently estimate on the data of the Czech economy and Euro Area 12, using Bayesian techniques. The advantage of the proposed trend specification is that the trend component and the cyclical component are modeled jointly in one model and the model itself decides which part of the data belongs to the trend component and which part belongs to the cyclical component. The proposed trend specification is flexible in the sense that smoothness of the trend can be easily modified by different calibration of the trend parameters. Results suggest that this method is able to find very reasonable trend in the data. Moreover, according to the Bayes factor the proposed specification decisively outperforms the original specification of the model estimated on the demeaned log differences with the same number of shocks.