Constrained k-class Estimators in the Presence of Weak Instruments
-
Emma M. Iglesias
In this paper, we show how we can apply the existing theory when a parameter is on a boundary (see Andrews (1999, 2000) and Iglesias and Linton (2007)) in the context of weak instruments. When we have a priori knowledge of the signs of the relationships between the instruments and the endogenous variables, we can construct constrained k-class estimators (a special case is a constrained 2SLS estimator), where the asymptotic theory in the presence of weak instruments changes in relation to the traditional weak-instrument asymptotics of the unconstrained k-class estimators of Staiger and Stock (1997). We show theoretically and in a simulation study the advantages that we have by constraining the estimator.
©2012 Walter de Gruyter GmbH & Co. KG, Berlin/Boston
Articles in the same Issue
- Article
- Early Detection Techniques for Market Risk Failure
- Beta Autoregressive Transition Markov-Switching Models for Business Cycle Analysis
- A Computationally Practical Robust Simulation Estimator for Dynamic Panel Tobit Models
- Panel Cointegration Rank Testing with Cross-Section Dependence
- Constrained k-class Estimators in the Presence of Weak Instruments
- Stages of Economic Development in an Innovation-Education Growth Model
Articles in the same Issue
- Article
- Early Detection Techniques for Market Risk Failure
- Beta Autoregressive Transition Markov-Switching Models for Business Cycle Analysis
- A Computationally Practical Robust Simulation Estimator for Dynamic Panel Tobit Models
- Panel Cointegration Rank Testing with Cross-Section Dependence
- Constrained k-class Estimators in the Presence of Weak Instruments
- Stages of Economic Development in an Innovation-Education Growth Model
