University of Ioannina
The use of refined asymptotic methods to correct the size of t and F tests in systems of econometric equations
Abstract
dc:descriptionSeemingly Unrelated Systems of Econometric Equations (S.U.S.E.E.) are made of sets of Simultaneous Equations Systems (S.E.S.) with contemporaneously---but not serially---correlated disturbances. Finite-sample size corrections of the t and F econometric tests are developed for both S.U.S.E.E. Models and Simultaneous Equations Systems. These corrections are based on the degrees-of-freedom adjustments of the asymptotic tests, the Edgeworth corrections of the critical values and the Cornish-Fisher corrected test statistics.Since Edgeworth approximations may assign negative “probabilities” in the tails of the distributions, theory suggest the use of Cornish-Fisher corrected statistics, which are proper random variables.Monte Carlo simulations are conducted to evaluate the relative performance of the proposed correction methods, given that all these corrections have an approximation error of the same order of magnitude.
Degree
thesis:*- Grantor dc:publisher
- University of Ioannina
- Year dc:date
- 2014
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kakarantza, Eudoxia
- Κακαράντζα, Ευδοξία
Subjects
dc:subject × 24- Υποδείγματα φαινομενικά ασυσχέτιστων παλινδρομήσεων
- Συστήματα ταυτόχρονων εξισώσεων
- Επιτροπή cowles
- Αποτελεσματικότητα εκτίμησης
- Διορθώσεις του μεγέθους των ελέγχων σε μικρά δείγματα
- Edgeworth διορθωμένες κριτικές τιμές
- Cornish-Fisher διορθωμένοι έλεγχοι
- Monte Carlo προσομοιώσεις
- Βαθμός υπερταυτοποίησης
- Στοχαστικά αναπτύγματα
- S.U.R. models
- Simultaneous equations systems
- Cowles commission
- Estimation efficiency
- Small sample size corrections
- Edgeworth corrections of the critical values
- Cornish-Fisher corrected tests
- Monte Carlo simulations
- Degree of overidentification
- Sochastic expansions
- Κοινωνικές Επιστήμες
- Οικονομικά και Επιχειρήσεις
- Social Sciences
- Economics and Business
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/38894