CFE-CMStatistics 2024: Start Registration
View Submission - CFECMStatistics2024
A1636
Title: A test of exogeneity in structural vector autoregressions with external instruments Authors:  Giovanni Angelini - University of Bologna (Italy)
Giuseppe Cavaliere - University of Bologna (Italy)
Luca Fanelli - University of Bologna (Italy) [presenting]
Abstract: A novel test is introduced for instrument exogeneity that is robust to proxy strength and does not require auxiliary information beyond SVAR restrictions and the instruments themselves. The test uses a Waldt-type logic to assess whether specific moment conditions implied by the proxy-SVAR are consistent with the hypothesis of no contamination of the instruments by the non-target shocks. To construct the test statistic, a consistent estimator of key elasticity parameters is required, independent of the instruments. This estimator is derived from the dynamic structure of the SVAR, under a set of overidentifying restrictions similar to `Byron-type' constraints used in simultaneous equation systems. The test is reliable only if the Byron-type restrictions hold within the SVAR. Conditional to the non rejection of the Bayron's restrictions, the asymptotic inference on instrument exogeneity based on this estimator is valid. Extensive Monte Carlo simulations are conducted to evaluate the finite-sample performance of the test. Its practical utility is demonstrated by revisiting some empirically prominent proxy-SVARs from the existing literature.