Abstract:
<p>This dissertation asks whether frequency misspecification of a New Keynesian model</p><p>results in temporal aggregation bias of the Calvo parameter. First, when a</p><p>New Keynesian model is estimated at a quarterly frequency while the true</p><p>data generating process is the same but at a monthly frequency, the Calvo</p><p>parameter is upward biased and hence implies longer average price duration.</p><p>This suggests estimating a New Keynesian model at a monthly frequency may</p><p>yield different results. However, due to mixed frequency datasets in macro</p><p>time series recorded at quarterly and monthly intervals, an estimation</p><p>methodology is not straightforward. To accommodate mixed frequency datasets,</p><p>this paper proposes a data augmentation method borrowed from Bayesian</p><p>estimation literature by extending MCMC algorithm with</p><p>"Rao-Blackwellization" of the posterior density. Compared to two alternative</p><p>estimation methods in context of Bayesian estimation of DSGE models, this</p><p>augmentation method delivers lower root mean squared errors for parameters</p><p>of interest in New Keynesian model. Lastly, a medium scale New Keynesian</p><p>model is brought to the actual data, and the benchmark estimation, i.e. the</p><p>data augmentation method, finds that the average price duration implied by</p><p>the monthly model is 5 months while that by the quarterly model is 20.7</p><p>months.</p>
