Bayesian Estimation of DSGE Models: Lessons from Second-Order Approximations
(coming soon)

This paper investigates a general procedure to estimate second-order approximations to a DSGE model and compares the performance with the estimation technique widely used for a log-linearized economy on a version of new Keynesian monetary model. It is done in the context of posterior distributions, welfare cost, and impulse response analysis. Our findings include the followings. First, we find that all the results of An and Schorfheide (2007) are confirmed with U.S. data. With the nonlinear estimation we can identify parameters that are neglected in linear estimation; the marginal data density evaluation shows that data support the nonlinear estimation procedure; and parameter estimates that are related to nondeterministic steady states are quite different from the linear estimates. Second, the estimated welfare differentials are more aggressive for the second-order approximations, that is, the posterior welfare differentials from the linear estimation may underestimate the welfare cost resulted from changes in the monetary policy. Third, the second-order approximation unveils quite different dynamics which are neglected in a log-linearized economy.

Can a Representative Agent Model Represent a Heterogeneous Agent Economy?
Joint with Yongsung Chang and Sun-Bin Kim

This Version: July 2007

Accounting for observed fluctuations in aggregate employment, consumption, and real wage using optimality conditions of a representative household often requires preferences that are incompatible with economic priors (e.g., Mankiw, Rotemberg, and Summers 1985). This discrepancy between the equilibrium model and the aggregate data is often viewed as evidence of the failure of labor-market clearing. We argue that such a conclusion is premature. We construct a model economy where all prices are flexible and all markets clear at all times but household decisions are not readily aggregated because of incomplete capital markets and the indivisible nature of labor supply. We demonstrate that if we were to explain the model-generated aggregate time series using decisions of a ˇ°fictitiousˇ± stand-in household, such a household is likely to have a non-concave or unstable utility. Our analysis suggests that the representative agent model often fails to represent an equilibrium outcome of a heterogeneous agent economy.

On Accommodation of Capital Stock Series in the Estimation of Dynamic Economy
(coming soon)

We have presented a statistical method to generate the capital stock series based on a model economy. While the implied capital stock from the baseline economy with capital utilization and investment shock does not keep track of the observed data, the augmentation of low frequency data enables us to have an improved estimates on model parameters and the implied capital series.

Refereed Publications

Bayesian Analysis of DSGE Models
Joint with Frank Schorfheide
Econometric Reviews, 26(2-4), 2007, 113-172, with discussion and rejoinder:211-219
doi:10.1080/07474930701220071
doi:10.1080/07474930701220246

This Version: July 2006
Other Versions: CEPR Discussion Paper DP 5207; Federal Reserve Bank of Philadelphia Working Paper No.06-5

GAUSS programs to analyze DSGE model
GAUSS programs to analyze DSGE-VARs
MATLAB program for nonlinear analysis

This paper reviews Bayesian methods that have been developed in recent years to estimate and evaluate dynamic stochastic general equilibrium (DSGE) models. We consider the estimation of linearized DSGE models, the evaluation of models based on Bayesian model checking, posterior odds comparisons, and comparisons to a reference model, as well as the nonlinear estimation based on a second-order accurate model solution. These methods are applied to data generated from a linearized DSGE model, a vector autoregression that violates the cross-coefficient restrictions implied by the linearized DSGE model, and a DSGE model that was solved with a second-order perturbation method.

home | research | teaching| photos | links