Teaching
Solution, Identification, and Estimation of DSGE Models
Ph.D.-Level, Prof. Michael Binder, Ph.D.
Teaching assistant in the following semesters: winter 23/24, winter 22/23, winter 21/22, winter 19/20
In this second year Ph.D. course we start by introducing and deriving a baseline DSGE model following for example Fernández-Villaverde & Rubio-RamÃrez (2006). In the first part of the course, we solve this model using various well known solution methods for rational expectations models such as Binder & Pesaran (1997), Blanchard & Kahn (1980), Sims (2002), as well as Dynare, and compare them to each other. We also investigate the differences of unique stable solutions to multiple stable solutions. The students are then introduced to the novel structural break approach introduced by Kulish & Pagan (2017) and have to analyze model behavior when deep structural parameters change in an unexpected, expected, or announced scenario.
The second part focuses on the crucial aspect of identification issues commonly seen in modern DSGE models. After defining the concept of global and local identification we start with a single equation rational expectation model set-up. Naturally, we advance back to the multi equation model established in the first part. Since identification properties are not as easily derived as before we delve into frequency domain time-series econometrics to apply the novel approach of Qu & Tkachenko (2017) to numerically check for identification using spectral densities by calculating Kullback-Leibler distance measures.
The last part of the course focuses on the estimation of structural parameters in linear rational expectation models. Since there is a separate course on Bayesian techniques, we focus on the generalized method of moments (GMM) and maximum likelihood (ML) approaches. One benefit of GMM is that the model solution is not necessary and thus no stance needs to be take on whether the model solution is unique stable or not. In order to estimate the model using ML, we set up and explain the state space representation and derive the Kalman filter.
Throughout the course, we use and program most of the code directly in Matlab. This includes the various solution and identification methods. Dynare is used to compare results in the solution part and shown for ease of use. In the last part, Stata is used for simple GMM and ML estimation.
Macroeconomics 1 (BMAK)
B.Sc.-Level, Prof. Michael Binder, Ph.D.
Teaching assistant in the following semesters: winter 23/24, winter 22/23, winter 21/22, winter 20/21, winter 18/19
This introductory macroeconomics course, we cover a wide range of topics beginning with what defines the field and methods of macroeconomics. The first part covers the measurement of macroeconomic activity in various forms base on the national income and product accounts. We start by deriving GDP with the product, income, and expenditure approach and describe their equality. This is followed by the introduction of gross national product and gross national disposable product and the current account.
The main part of the course can be divided into two parts, the macroeconomy in the short run and the macroeconomy in the long run. But first, we start with the definition of business cycles and the co-movement of important macroeconomic variables. For the real sector, aggregate demand and the IS-curve are derived from microeconomic first principles such as household and firm optimization. Furthermore, we model the government and the trade balance to derive at the IS-curve. The financial sector is part of the next section in which the TR-curve is derived from a combination of monetary policy and the bank lending. These two combined yield the short run macroeconomic outcome determined by decision making in the real financial sectors. This model set-up is then used and extended to analyze the financial crises, including the zero lower bound, the pandemic period, and the following energy crisis.
The second part builds on the first but looks at the macroeconomy in the medium run. From this point onward, price and wage changes play a crucial role. Now, firms face a trade-off between pricing their goods with a mark-up over production costs and employees bargaining for higher wages. This model also features a real monetary policy rate and effects of inflation expectations on macroeconomic outcomes and is used to analyze the same periods as in the short run before.