Economics
416
Fall,
2014
Advanced Macroeconomics: Estimation and
Analysis of Dynamic Macroeconomic Models
The course is the first in the three-part 416 series. The
course focuses on a mixture of methodological tools and economic substance relevant
to empirical macroeconomics. The course evaluation is based on a midterm, a
final and weekly homeworks. The final may be replaced by a term paper. The
recommended computer software is MATLAB and Dynare.
1. Solution
and stochastic simulation of dynamic models (software
used to generate the graphs in the handout, a zip
file that uses Dynare to do some of the computations).
a. Perturbation methods and pruning (detailed handout
on the use of symbolic algebra in MATLAB to do second order perturbation).
b. Projection methods and dynamic programming.
c. Applications: real business cycle models, later:
models with sticky prices.
d. Extended discussion
of first order perturbation: Blanchard-Kahn conditions for determinacy.
e. References:
i.
Christiano-Fisher
(JEDC,
2000)
ii.
Ken Judd (Numerical Methods in Economics, MIT
Press, 1998).
iii.
Kim-Kim-Schaumburg-Sims
(JEDC, 2008) .
iv.
den
Haan-de Wind (2009).
v.
Lombardo
(2011)
vi.
Andreasen,
Fernandez-Villaverde and Rubio-Ramirez (2013).
vii.
Mario Miranda and
Paul Fackler, Applied
Computational Economics and Finance, MIT Press, 2002 (codes).
2. Methods
for Bayesian inference.
a. Brief overview of state space/observer representations
(see Hamilton, Time Series Analysis and Prof. Primiceri’s
416 course).
b. Bayes’ rule.
c. Integration: Monte Carlo and Quadrature.
d. The Metropolis-Hastings algorithm for computing the
posterior distributions of parameters.
e. Laplace approximation to the posterior distribution
and Geweke’s modified harmonic mean estimator of marginal likelihood.
f. Illustration of Bayesian estimation methods using
artificial data generated from simple NK model.
g. References: Smets and Wouters (AER, 2007); An and
Schorfheide (Econometric
Reviews, 2007); Zellner, Introduction
to Bayesian Inference in Econometrics (1971). For a discussion of a
Bayesian version of GMM, section 3.3.3 here.
To see how model properties such as variances and impulse responses can be
incorporated into priors, see.
For a rigorous discussion of the parameter in the jump distribution, see.
3. Simple
New Keynesian model.
a. Economic foundations and properties of the model.
b. Solution and analysis using perturbation methods (Rotemberg
and Calvo sticky price models.)
c. Extended path solution method
(manuscript,
the MATLAB code for the examples in the manuscript can be found here).
d. Implications
of the model for the zero lower bound on nominal interest rates.
i.
The vulnerability
to deep depression, the impact on the government spending multiplier.
ii.
Multiplicity
of equilibria and equilibrium selection (related
material, including exercise).
iii.
References: Christiano,
Eichenbaum and Rebelo
(2011), Eggertsson and
Woodford (2003).
e. References for the model:
i.
Gali, Unemployment
Fluctuations and Stabilization Policies: A New Keynesian Perspective, MIT
Press; Monetary Policy, Inflation and the
Business Cycle: An Introduction to the New Keynesian Framework, Princeton
University;
ii.
Woodford, Interest and Prices: Foundations of a Theory
of Monetary Policy, Princeton University Press.
i.
My handbook of
monetary economics chapter.
4. The labor market (background).
a. Motivation
for ‘sticky wages’, and a critique. (Macro
Annual discussion of Gali-Smets-Wouters, slides.)
b. Extensions
of the Diamond-Mortensen-Pissarides approach. (Manuscript.)
5. Extensions of dynamic models
i.
The ‘timeless
perspective’.
ii.
Time
inconsistency.
b. Financial frictions
i.
Hidden effort
models in banking.
ii.
Dynamic
contracts in
the Absence of Commitment (related work: Albuquerque-Hopenhayn, 2004,
RESTUD, vol. 71, No. 2; and Jonathan
Thomas and Tim Worrall, 1994, ‘Foreign Direct Investment and the Risk of
Expropriation,’ RESTUD, vol. 61, pp. 81-108).
Homework
#3, code
(qzswitch.m)