Macroeconomic Theory
Jeffrey Parker, Reed College

Paper of the Week

Assigned paper

Pritchett, Lant. 1997. Divergence, Big Time. Journal of Economic Perspectives 11 (3):3-17.

Reading suggestions

• The "convergence hypothesis" has most often been tested by comparing the growth rates of rich and poor countries to determine whether the poor countries grow faster, and therefore are catching up. This paper takes a different approach. It is not technically difficult.
• Recall that "conditional convergence" means that the catching-up of poorer countries may depend on their characteristics: whether they have the same parameters and thus the same steady-state growth path as the rich countries. "Absolute convergence" means that all poorer countries will catch up, either because they do (or will as they grow) have the same parameters or because the model that best describes the world does not require the parameters to be the same.
• The regression approach that is most common estimates a regression equation with growth on the left-hand side and initial per-capita income on the right. Tests of conditional convergence add more explanatory variables on the right to correct for potential differences in countries' characteristics; tests of absolute convergence do not. The typical find in these studies is that (with the "right" set of additional variables) conditional convergence occurs at about 2% per year. In other words, poor countries grow faster than rich countries with the same characteristics by enough to eliminate about 2% of the per-capital income gap per year.

Questions for analysis

1. Consider the convergence hypothesis using Pritchett's Table 1. Does it appear that, within this sample, countries with lower per-capita income in 1870 generally grew faster in the following century-plus than those with higher incomes? Within this sample of countries, is this evidence for or against absolute (or conditional) convergence. Is this a valid test of convergence? Why or why not?
2. Pritchett calculates an annual PPP-adjusted income of $250 to be a minimum subsistence level. Discuss and assess his method of establishing this. Does it seem reasonable to you?
3. How does Pritchett establish the existence of "divergence, big time"? Is he rejecting absolute convergence or conditional convergence? Can you reconcile his results with the conditional convergence result from the regression literature (discussed in the third bullet above)?
4. What do you conclude from Pritchett's study (and the general convergence literature) about the prospects for global convergence in per-capita income in the next 100 years? Why?