Two general approaches are frequently used to analyze the impact of infrastructure, of which PPPs are a subset. The first is the production function, with infrastructure as a key production input; the second is cross-country growth regressions that relate economic outcomes to infrastructure indicators, controlling for other critical growth determinants (Servén 2010)
This chapter uses the second approach by estimating the following growth regression:
| git = αyi,t-1 + Χ′itβ + εit'′ | (1) |
| εit = αt + μi +V it′ |
where git is the real per capita GDP growth of country i at year t , yi,t-1 captures the conditional convergence using the logarithm of real per capita GDP of country i at year t-1, and xit is a column vector of growth regressors, with PPP investment as percentage of GDP as the variable of interest. The disturbance term, εit, has orthogonal components: the country and year fixed effects, μi and αt and the idiosyncratic shocks, Vit.
Depending on the availability of data, we arrive at an unbalanced panel involving 19 developing Asian economies over the period 1985-2015. Appendix A2.1 presents the variables used in the analysis, with their definitions and sources.