1,877
Views
0
CrossRef citations to date
0
Altmetric
Discussion

Narrative Restrictions and Proxies

, ORCID Icon &

Figures & data

Fig. 1 Cumulative (null) distribution function of Wald statistic under narrative proxy.

NOTE: Dashed lines represent 68th and 95th percentiles; data-generating process assumes T = 1000, η21=0.4 and vech(Σtr)(0.7,1.1,1.4); null distribution approximated using 100,000 Monte Carlo replications.

Fig. 1 Cumulative (null) distribution function of Wald statistic under narrative proxy.NOTE: Dashed lines represent 68th and 95th percentiles; data-generating process assumes T = 1000, η21=0.4 and vech(Σtr)≈(0.7,−1.1,1.4)′; null distribution approximated using 100,000 Monte Carlo replications.

Table 1 Weak-proxy robust inference—Monte Carlo results for α=0.95.

Table 2 Weak-proxy robust inference—Monte Carlo results for α=0.68.

Table 3 Robust Bayesian inference—Monte Carlo results for α=0.95.

Table 4 Robust Bayesian inference—Monte Carlo results for α=0.68.