LISREL growth model on direct and indirect effects using cross-country data

Many papers have analyzed the factors affecting economic growth. However, these have concentrated on direct effects and ignored indirect effects through other variables in the model. This study investigates direct and indirect effects of various factors on growth with a causal growth model using LISREL (LInear Structural RELations) for a sample of 105 countries over the period of 1975-2002. Results suggest that ignoring the indirect effects of productivity growth, geography and economic development on economic growth may lead to a considerable underestimation of their 'true' total effects on growth. While the importance of the indirect effects of productivity growth and geography relative to their direct effects changes with the estimated model, the relative importance of the indirect effect of economic development on growth is found to be robust to different model specifications.

► First comprehensive LISREL model investigating indirect effects on economic growth. ► Based on theoretical models and discussions by Barro and Sala-i-Martin, and Weil. ► Indirect effects operate through other variables, some important for true total effects. ► Find large indirect effects of productivity growth, geography, development on growth. ► Find that indirect effect of development on growth is robust to model specifications.