Comments

Nancy Birdsall: Analyzing a society's social mobility can answer a fundamental distributional question typically posed by economists: are economic opportunities broadly shared in that society, independent of an individual's luck in the parents he or she has? The authors' analysis of intergenerational schooling mobility in Latin America adds to a small literature assessing equality of opportunity in Latin America. Their paper contributes to our understanding of the link between income inequality and social mobility, and it begins what should be a deeper assessment of the effects of market reforms and other policy changes on equality of opportunity in the region.

Inequality, Mobility, and Growth

Income inequality has traditionally been viewed as a regrettable byproduct of a process of growth and structural change, which might better be accepted than addressed by distributional efforts that would introduce distortions and undermine the dynamic of growth itself. An alternative view is that income inequality signals deeply entrenched inequality of opportunities, particularly when it occurs at the high levels found in Latin America and when it is accompanied by widespread absolute poverty. Given the weak capital and other markets and the inadequate educational and regulatory institutions typical of many developing countries, income inequality reinforces the likelihood that the poor will be crowded out of jobs, credit, and other productive opportunities, which ultimately undermines their productive potential and the economy's overall efficiency and growth.

In principle, social mobility could be high even in countries with high income inequality, as long as the high inequality of income in one generation is not replicated in the next generation. Evidence on social mobility across generations thus provides more convincing evidence about equality of opportunity than a typical cross-section-based measure of inequality itself.¹ Furthermore, information on social mobility provides evidence on whether inequality is good or bad for growth. Inequality may be detrimental to growth, for example, by combining with credit market failures to reduce investment in schooling. On the other hand, as some might argue about inequality in the United States, it may enhance growth by providing incentives for hard work, for innovation, or indeed for parents to invest in their children's schooling.² Where social mobility is high, high inequality is more likely to be growth enhancing.

For Latin America, the authors' analysis indicates that social mobility is not offsetting high measured inequality. Schooling mobility is much lower across Latin America than in the United States, and the differences among countries in Latin America are sufficient to suggest that country characteristics, history, and social and economic policies matter. The authors also refer to limited evidence from Asia showing that by comparison Latin America is less mobile (in terms of schooling).

Their analysis is only a beginning, however. Several puzzles and problems remain. First, the ranking of countries within Latin America shifts depending on the measure of schooling mobility. The authors provide a healthy reminder that the profession is still short of a definitive, trustworthy measure that is clearly independent of correlated unobserved factors over time and across countries or other units.

Second, the authors' measures of immobility may simply reflect inequality of schooling in each country. Across countries, schooling immobility is closely related to a measure of schooling inequality (see figure 7 herein and the authors' table 4). Also, schooling inequality, like measures of immobility, is closely associated with mean levels of schooling (see figure 8). High inequality of schooling in Latin America, which is even higher in some countries than others, seems to be all too good a predictor of immobility—and a good proxy for unequal opportunity. That result is underlined by the authors' troubling finding that within countries over time, the increasing mean level of education has not been associated with increasing mobility (see table 6).³ In other words, in Latin America, at least compared with other regions, current inequality across adults of schooling is replicating itself, even as overall means rise. But is the close association of schooling inequality and immobility an empirical finding or an outcome of the relative imprecision of the available measures of one or both?