Estimating the effect of spillovers on exports: a meta-analysis

This study uses meta-analysis to study the effect of spillovers on exports. It collects 3291 estimated spillover effects from 99 studies. The estimated spillover effects in the literature span a large number of types and measures of both exports and spillovers. As a result, we transform estimates to...

Full description

Saved in:
Bibliographic Details
Published in:Review of world economics 2020-05, Vol.156 (2), p.219-249
Main Authors: Duan, Jianhua, Das, Kuntal K., Meriluoto, Laura, Reed, W. Robert
Format: Article
Language:English
Subjects:
Bayesian analysis
Bias
Economic Policy
Economic theory
Economics
Economics and Finance
Estimation
European Integration
Exports
Global economy
International Economics
Macroeconomics/Monetary Economics//Financial Economics
Meta-analysis
ORIGINAL PAPER
Random effects
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study uses meta-analysis to study the effect of spillovers on exports. It collects 3291 estimated spillover effects from 99 studies. The estimated spillover effects in the literature span a large number of types and measures of both exports and spillovers. As a result, we transform estimates to partial correlation coefficients. We analyze these transformed effects using four different versions of Weighted Least Squares (WLS) estimators, incorporating both meta-analytic “Fixed Effects” and “Random Effects”. Our analysis produces three main findings. First, while we estimate an overall mean effect of spillovers on exports that is statistically significant, the size of the effect is economically negligible. Second, we find modest evidence for the existence of publication bias in the empirical literature. Publication bias can arise when researchers and journals have a preference to publish articles that find positive and significant results. While some of our tests indicate the presence of publication bias, in every case the size of the effect is small. Third, using both Bayesian Model Averaging and frequentist meta-regression analysis, we find that some data, estimation, and study characteristics are significant in some regressions. However, only a few of the characteristics are robust, and none are large in size.
ISSN:1610-2878
1610-2886
DOI:10.1007/s10290-020-00377-z