On the accuracy of low-order projection methods

We use low-order projection methods to compute numerical solutions of the basic neoclassical stochastic growth model. We assess the quality of the obtained solutions, and compare them to numerical approximations derived with first and second-order perturbation techniques. We show that projection methods perform surprisingly poor when the degree of approximation is very low, and we provide some intuition behind this finding.

MoreLess

Year of publication:	2007
Authors:	Pichler, Paul
Published in:	Economics Bulletin. - AccessEcon, ISSN 1545-2921. - Vol. 3.2007, 50, p. 1-8
Publisher:	AccessEcon
Subject:	numerical accuracy

More details

Extent:	application/pdf
Type of publication:	Article
Language:	English
Classification:	C6 - Mathematical Methods and Programming ; E0 - Macroeconomics and Monetary Economics. General
Source:	RePEc - Research Papers in Economics

Persistent link: https://www.econbiz.de/10005196512