Premium
Comparing Apples to Apples: Estimating Consistent Partial Effects of Preferential Economic Integration Agreements
Author(s) -
Egger Peter H.,
Tarlea Filip
Publication year - 2021
Publication title -
economica
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.532
H-Index - 65
eISSN - 1468-0335
pISSN - 0013-0427
DOI - 10.1111/ecca.12361
Subject(s) - propensity score matching , covariate , econometrics , economics , observable , matching (statistics) , bilateral trade , average treatment effect , similarity (geometry) , statistics , mathematics , computer science , physics , quantum mechanics , artificial intelligence , political science , law , china , image (mathematics)
The dominant paradigm of the estimation of causal partial effects of preferential economic integration agreements (PEIAs) on trade costs and trade flows is to rely on a selection on observables, with propensity‐score matching being the leading example. Conditional on some score of joint determinants of PEIAs and trade flows, causal partial effects of PEIAs on trade are obtained from a simple conditional mean comparison of trade between members and non‐members. Key for this approach to obtain consistent estimates is that the score used to form comparison groups is balanced: similarity of country pairs in the PEIA propensity core means similarity in all the observables. Otherwise, the effects may be biased, misascribing effects of differences in individual observables to PEIA membership. We employ a remedy for this bias through entropy balancing, demonstrate that there is an upward bias of PEIA effects on trade flows from lack of covariate balancing, and quantify the bias of the effects.