Drivers of structural convergence: Accounting for model uncertainty and reverse causality

Objective: The objective of the article is the examination of factors that affect structural convergence and assessing their robustness. Research Design Methods: Determinants of structural similarity are examined using the Bayesian model averaging with dilution prior to establishing robust drivers in the long run. The short-run analysis is conducted using Bayesian model averaging within a dynamic panel framework with weakly exogenous regressors. Findings: The application of Bayesian model averaging allowed for the identification of 12 variables associated with more similar production structures, among others, the bilateral total and intra-industry trade, the level of development, geographical distance, foreign direct investment flows, technology, corruption, and membership in the EU. Accounting for reverse causality showed that trade induces divergence in the short run – in line with predictions of neoclassical theories – but is associated with more similar production structures in the long run. Interestingly, even though old EU countries are characterised by more homogenous production structures, EU membership is associated with structural divergence once differences in income are included in the model. Even more unexpectedly, countries with more similar production structures are characterised by more similar and generally lower levels of corruption. Implications Recommendations: The analysis shows that policies aiming at the promotion of FDI and technological transfers can speed up the process of structural convergence. Contribution Value Added: The paper presents the first systematic analysis into the sources of structural similarity.