Title: Estimating the Euler equation from the basic C-CAPM using a large set of possible instruments
Language: English
Authors: Goncharenko, Roman 
Qualification level: Diploma
Advisor: Pelenis, Justinas
Issue Date: 2013
Number of Pages: 49
Qualification level: Diploma
We perform the instrumental variable estimation of the Euler equation and the system of Euler equations from the basic Consumption - based Capital Asset PricingModel (C-CAPM) using a large set of possible instruments. This large set of possible instruments is due to the Rational ExpectationHypothesis. The optimal GMMestimator, which is used in the estimation, has a finite sample bias proportional to the number of instruments. This means that there is a need of the efficient instrument dimension reduction method. The two different methods of such a reduction are compared: the FIV estimator and the optimal GMM estimator that uses preselected principal components (constructed from the large set of possible instruments) as instruments. Originally, the two methods were developed for linear models. We modify the latter method to extend it to non-linear models. The Euler equation is estimated in both nonlinear and linearized formswith different utility function specifications. Estimation results do not contradict earlier research in terms of the estimated magnitude of parameters of interest but unambiguously point on the dominance of the substitution effect in the consumers' saving decision. Moreover, we find that the optimal GMM that uses preselected principal components as instruments (the second method) performs better than the FIV estimator (the first method).
URI: https://resolver.obvsg.at/urn:nbn:at:at-ubtuw:1-63972
Library ID: AC10895876
Organisation: E017 - Weiterbildungszentrum der TU Wien 
Publication Type: Thesis
Appears in Collections:Thesis

Files in this item:

Show full item record

Page view(s)

checked on Apr 24, 2021


checked on Apr 24, 2021

Google ScholarTM


Items in reposiTUm are protected by copyright, with all rights reserved, unless otherwise indicated.