Trade, education, and income inequality

ABSTRACT

This paper examines the relationship between countries’ bilateral trade with the United States that is not due to gravity (non-gravity trade) and the distribution of income within countries. In countries where only a small share of the population are educated, an increase in non-gravity trade is associated with a significant increase in income inequality. As education of the population increases the correlation between non-gravity trade and income inequality becomes smaller. Non-gravity trade has no significant effect on income inequality in countries that are world leaders in education.

KEYWORDS:

• Non-gravity trade
• inequality
• education

JEL CLASSIFICATION:

• F1
• E2

Acknowledgements

We are grateful for constructive comments by Simon Evenett, Harris Dellas, Bob Gregory, Warwick McKibbin, Jorge Miranda Pinto, Rabee Tourky, Josef Zweimueller as well as seminar participants at the Research School of Economics Macro-Series (ANU), and Crawford School of Public Policy/Arndt-Corden/ACDE Trade and Development joint seminar (ANU), the Silvaplana Workshop in Political Economy, and Modelling Macroeconomic Shocks Workshop.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 In the online appendix https://www.joaquinvespignani.com/online-appen, we make the argument formally in an open economy model. In this footnote, we sketch the main idea for an economy with a tradable good, YT, and a non-tradable good, YNT. In this economy perfect competition equates the wage, W, to the value marginal product of labor, d(PY)/dL. Consider now two extremes: (i) labor is sector-specific, i.e. labor is immobile between the tradable and non-tradable sector; (ii) labor is perfectly mobile across sectors. In the first case, it holds that.

(i) $W^T/W^{NT}=\left[d\left(P^T{Y}^T\right)/d{L}^T\right]/\left[d\left(P^{NT}{Y}^{NT}\right)/d{L}^{NT}\right]$

An increase in trade openness – P^TY^T/(P^NTY^NT+ P^TY^T) – increases the right-hand side of the above equation. (For the Cobb Douglas production function, the right-hand side is [P^TY^T/P^NTY^NT]*[L^NT/L^T]).Hence, an increase in trade openness increases inequality (under the standard assumption that productivity in the tradable sector is higher than in the non-tradable sector). In the other extreme case of perfect labor mobility across sectors, perfect competition forces the wages to equalize:

(ii) $W^T=W^{NT}$

From equation (ii) it follows that trade openness has no effect on income inequality.

2 For the multilateral trade data, we cannot use tariffs as an instrument for non-gravity trade. This is because the exogeneity assumption is not plausible for all countries; only for a large economy like the US is the assumption that a tariff imposed by the US on manufacturing products of country i is exogenous to inequality of country i.

3 We also estimated the model with the share of income held by the poorest 10% as dependent variable. Results are similar in terms of sign, magnitude, and statistical significance to those reported in Table 5. Results are not reported and are available upon request from the authors.