Seemingly unrelated regression (SUR) method is applied to the instrumental variable (IV) estimation of the canonical contagion models. A finite sample Monte Carlo experiment shows that the resulting estimator, IV-SUR estimator, is substantially better than the existing IV estimator in terms of both bias and mean squares error under diverse circumstance of instrument, conditional heteroscedasticity, and cross-section correlation.
|Number of pages||10|
|Journal||Communications in Statistics: Simulation and Computation|
|State||Published - 2 Jan 2016|
Bibliographical notePublisher Copyright:
© 2016 Taylor & Francis Group, LLC.
- Instrumental variable estimator
- Seemingly unrelated regression