[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Understand the short-run inflation dynamics is essential for conducting the fiscal or monetary policies, and the New Keynesian Phillips Curve (hereafter, NKPC) has been widely used for charactering it in the past two decades. The mixed results arise from estimating NKPC with constant parameters. I argue that the timevarying features of parameters in NKPC help reconcile the conflicting conclusions in the empirical NKPC studies. Moreover, it is useful for policymakers to estimate the effect of government spending on private spending and aggregate output. However, the precise estimate of such effects is hard to pin down since the researchers use very different theoretical models, ranging from a frictionless Real Business Cycle model to a medium-scale New Keynesian model with many nominal and real frictions. I take a top-down approach by generalizing the Dynamic Stochastic General Equilibrium (DSGE) model of Smets and Wouters (2007), in which many DSGE models can be viewed as simpler versions of it after removing certain nominal or real frictions. I take a Bayesian approach to estimate the fiscal stimulus in different models, which are obtained by imposing a tight prior on a single parameter or a combination of tight priors on multiple parameters. I pick up an appropriate model via Bayes factor and then use it to forecast the effect of government spending. I find a positive short-run effect but a negative long-run consequence of fiscal stimulus.
Fiscal Stimulus in a Monetary Union: Evidence from US Regions
