| Summary: | Using US quarterly data (i.e., real-time data and survey data respectively) from 1969 to 2015 through two different estimation approaches (i.e., Bayesian estimation approach and indirect inference estimation approach) to investigate the empirical performance of the standard reduced-form New-Keynesian Dynamic Stochastic General Equilibrium (DSGE) model under the condition without (i.e., full-information rationality) and with inattentive features (i.e., sticky information and imperfect information data revision), we find some consistent results. Firstly, the model of sticky Information is detected to be the preferred model to fit the real-time data behavior. Secondly, the model with sticky information is the only one can generate delay response, which is matching the evidence observed in actual data and in line with most consequences from the previous studies. Thirdly, the imperfect information data revision model performs better when we substitute the real-time data with the survey data, through which we can deduce that the survey data contains extra information to help improve imperfect information data revision model’s performance. Three main contributions are made in this thesis. The first contribution is the estimation and comparison of different types of inattentive DSGE model (sticky information versus imperfect information data revision) for US small-closed economy through Bayesian approach using the US quarterly data (i.e., real-time data and survey data) representing the main macroeconomic time series from 1969 to 2015. What the second contribution is that through comparing different inattentive New-Keynesian DSGE models basing on the full structure (relative to the single equations competition), we inspect which way of inattentive expectation is closer to the way that people form their expectation in real economy. Besides, the thesis adopts Indirect Inference approach as the robust check methodology, which delivers a new way to assess inattentive macroeconomic models, which is the third contribution.
|
|---|
