Inequality and growth

Many development economists used to believe it likely that inequality would inevitably rise as developing countries grew, though they also thought it likely that inequality would eventually start to fall. This gave the famous “Kuznets curve” giving an inverted U relationship between inequality (on the vertical axis) and mean income (on the horizontal axis).

By and large the subsequent evidence has not born out this expectation, as discussed in EOP Chapter 8. Rather few countries have followed to course over time predicted by the Kuznets curve. (This was shown in, Equity and Growth in Developing Countries.) Roughly half the time, growing developing countries have seen rising inequality, but the other half saw inequality falling. Thus economic growth tends to be inequality-neutral on average.

But note that this refers to relative inequality. If one does not accept the “scale-independence axiom” of standard inequality measurement but instead one prefers “translation-independence” then one will view inequality in absolute terms–the absolute gap between “rich” and “poor” rather than the proportionate gap. And it seems that many people think about inequality in absolute terms. (See the page on the views of ECON 156 students about inequality means, under “TEACHING EOP.”) As EOP also shows, absolute inequality has tended to rise in growing developing economies.

For further discussion see “Inequality in the developing world.”

Poverty and growth

The impacts of growth on poverty have long been debated. For absolute poverty measures (meaning that the poverty line is fixed in real terms) higher rates of economic growth tend to be associated with higher rates of poverty reduction. And the average change in the poverty measure at zero growth is zero. This is another way of saying that one average, economic growth tends to be inequality neutral (as we heard about above). These points are among what Ferreira and Ravallion dub the “stylized facts” about distribution and growth: Poverty and inequality: The global context.

EOP Chapter 8 provides an overview of the arguments and evidence.  The chapter also points out that relative poverty measures tend to be less responsive to growth, since the poverty line rises with growth in the mean.

There also appears to be a strong reverse causation, whereby high initial poverty incidence makes it harder to grow the overall economy. The literature (reviewed in Chapter 8 of EOP) has pointed to a number of ways in which inequality is bad for growth. However, many of these arguments suggest that it is not inequality that really matters, but rather poverty. It is the lack of access to credit for poor people–to finance their education and other investments–that stalls growth.

There is a severe identification problem here, given that when one holds the initial mean income constant in a regression for the rate of growth (as is standard in growth empirics), the explanatory variable for “inequality” could simply be picking up poverty. (The two will naturally be very highly correlated when controlling for the initial mean.) In my paper “Why Don’t we see Poverty Convergence” I show that this is indeed the case: that the stronger distributional predictor of growth rates is the initial poverty rate, with only a rather weak effect of (relative) inequality when one controls for both the initial mean income and the poverty rate.

This an example of what is called “omitted variable bias” in econometrics; in this case, the poverty rate was the omitted variable in past growth regressions claiming to show that inequality is bad for growth.

New evidence for India

Cross-country comparisons needed to be augmented by repeated observations over time in specific developing countries. Since writing EOP, new evidence has appeared on the role played by economic growth in India’s progress against poverty. For their paper, “Growth, urbanization and poverty reduction in India,” Gaurav Datt, Rinku Murgai and Martin Ravallion constructed a new dataset of poverty measures for India spanning 60 years, including 20 years since reforms began in earnest in 1991. They find a downward trend in poverty measures since 1970, with an acceleration post-1991, despite rising inequality. Faster poverty decline came with both higher growth and a more pro-poor pattern of growth. Post-1991 data suggest stronger inter-sectoral linkages: urban consumption growth brought gains to the rural as well as the urban poor and the primary-secondary-tertiary composition of growth has ceased to matter as all three sectors contributed to poverty reduction.

For further details see the page “Poverty Reduction in India.”

The distribution-corrected growth rate

It is not the ordinary rate of growth but the distribution-corrected rate of growth that best predicts progress against poverty. EOP Chapter 8 summarizes the evidence on growth and distribution. Intuitively, if the poor have a low initial share of total income, and inequality does not fall with growth, then the poor will have a lower share of the benefits of growth: growth will have less impact on poverty. This is what we find empirically; the relevant rate of growth is the ordinary rate of growth times 1-inequality:

Expected rate of poverty reduction = 

constant × (1 – inequality) × growth rate

Here the constant term is negative and “inequality” is measured by an index that takes the value 0 when there is no inequality (everyone has the mean income) and 1 when the richest person has all of the income. The term “(1 – inequality) × growth rate” is called the “distribution-corrected growth rate.” (The term “constant × (1 – inequality)” is the growth elasticity of poverty reduction.)  The relationship in the above equation was found in Can high inequality countries escape poverty? and verified in Inequality is Bad for the Poor.

Poverty-inequality trade-off?

The idea that developing countries face a trade off between poverty and inequality has had considerable influence on thinking about development policy.  The experience of developing countries in the 1990s does not, however, reveal any sign of a systematic trade off between measures of absolute poverty and relative inequality. Indeed, falling inequality tends to come with falling poverty incidence.  And rising inequality appears more likely to be putting a brake on poverty reduction than to be facilitating it.

However, there is evidence of a trade off for absolute inequality, suggesting that those who want a lower absolute gap between the rich and the poor must in general be willing to see lower absolute levels of living for poor people.

Read more about the evidence in this paper, “Poverty-inequality tradeoff?”.