Sharp instrument: A stab at identifying the causes of economic growth

We shed new light on the determinants of growth by tackling the blunt and weak instrument problems in the empirical growth literature. As an instrument for each endogenous variable, we propose average values of the same variable in neighboring countries. This method has the advantage of producing variable-specific and time-varying—namely, “sharp”—and strong instruments. We also introduce “bias norms” to test the sensitivity of the estimates to the potential invalidity of our instruments. The estimations show that export sophistication is a relatively robust determinant of growth compared to other standard growth determinants such as years of schooling, trade openness, private credit to the economy, and institutions as measured by law and order. Other growth determinants such as human capital quality and technological level of production may be important to the extent they help improve export sophistication.