Borrowing Constraints, North-South Trade and Sustainable Current Accounts in Neoclassical Growth

What are the consequences for per capita incomes, convergence rates, and capital accounts under conditions of free trade and borrowing frictions between rich and poor countries? In Andic (2024), I extend the standard neoclassical growth model by assuming free trade in intermediate goods and imperfect capital flows. Comparative advantage, based on productivities, sets the pattern of trade as the South specializing in less sophisticated intermediates and the North in more sophisticated ones. Southern countries are capital poor and borrowing constrained by assumption. In this setting, I first show that the South attains a lower per capita GDP in the long-run compared to the North. Yet, its rate of convergence is relatively higher. Second, I demonstrate that the constraint on borrowing is endogenous. Third, the South runs a current account deficit in the long-run with the North. Finally, I discuss trade can lead to gains in terms of per capita capital compared to autarky and improve the world average income. Quantitatively, my analysis reveals that long-run per capita income of the South is 15 percent of the North. The conditional convergence rate is around 9 percent in the South and 3 percent in the North. The borrowing limit is 24 percent leading to a current account deficit of 2.5 percent in the South and a surplus of 0.4 percent in the North. Both blocks have static gains from trade in the constrained open economy compared to an autarky, if not dynamically better off. The relative gain of the North is found to be higher. However, the world average income more than doubles when the borrowing constraint is eliminated.
This work employs publicly available data from the IMF Balance of Payments and International Investment Position, World Bank World Development Indicators, World Bank International Debt Statistics and PWT10. It uses Stata15 and Python PyCharm to generate all the figures and tables, which can be replicated with the codes and data files provided below.
Note that in the second version, Stata .do file named "fig1_2_3_15_16_table1_2_3" is updated. Now, it produces a different Figure 2 compared to the first version.