Initial inequality and subsequent growth.

150

Source: Alesina and Rodrik [1994]. Note: Figures in parentheses denote t values.

Their regression results indicated a substantial negative relationship between initial inequality and subsequent growth. Particularly strong was the influence of the Gini coefficient that represents the initial inequality in land holdings. Their results suggested that an increase in the land Gini coefficient by 1 standard deviation (which is only an increase of 0.16 in this case) would decrease subsequent economic growth by as much as 0.8 percentage points per year. Table 7.3 summarizes the results of some of the regressions using Gini coefficients for initial land distributions.

The independent variables are per capita income in 1960 (GDP60), primary enrollment rates in 1960 (Prim60), the Gini coefficient on income in 1960,12 the initial Gini coefficient for land distribution (LandGini), and a dummy variable for democracy (Dem).

The first regression pools all countries for which data are available without regard to their political system. Note that the Gini coefficient for land is especially significant and negative (the Gini coefficient on initial income is less so; it is only significant at the 10% level). It is also of interest to note, in passing, that the original Barro findings continue to be upheld: initial per capita income enters negatively, whereas the human capital measure is positive.

These results are unaltered once we allow for structural differences across democratic and nondemocratic political systems. What is more, the democracy dummy is insignificant both by itself (version 3) and when interacted with the Gini coefficient on land (version 2). It does appear that political systems play little role in this relationship.

The Alesina–Rodrik findings are confirmed with the use of a more comprehensive data set in Deininger and Squire [1996b]. Initial land inequality is more significant than initial income inequality and stays that way even under several variations on the basic regression exercise (such as the use of regional dummies, which you recall, wrecked the Kuznets inverted-U hypothesis).13 The insignificance of the political system also holds up under the Deininger–Squire investigation.

Viewed in this light, it is perhaps no surprise that East Asian countries such as Korea and Taiwan have some of the highest rates of investment in the world. Early land reforms in these countries placed them among the lowest in land inequalities, and surely must have promoted economic inequality overall, given the importance of agriculture in all developing countries around 1960. The Gini coefficients for land distribution in Korea and Taiwan were 0.34 and 0.31 in 1960, and these numbers are very low even by relatively moderate Asian standards. For instance, for India and the Philippines, the corresponding numbers are well over 0.5, and for Latin America, the Gini coefficient skyrockets to well above 0.8 for countries such as Brazil and Argentina.

There seems to be little doubt, then, that there is a strong and negative relationship between initial wealth inequality (as proxied by the distribution of land, anyway) and subsequent economic growth. The question is, What drives this relationship? Might it be that lower inequality encourages savings and investment, leading to higher rates of growth along the lines discussed in Section 7.2.4, or is it that the political redistribution effect is at work?

It is hard to answer these questions at the level of existing data. For instance, we might argue, as Deininger and Squire [1996b] did, that the redistribution explanation is unsupported by the regression because the democracy dummy is insignificant. After all, the political demands for redistribution should matter more in a democracy, but it is unclear whether this is necessarily the case: dictators like to remain in power just as much as democratic governments do, and they might react to high inequality with high taxes at the margin, just as a democratic government might.

Thus we must be content (for now) with the intriguing possibility that there might be a robust and negative empirical relationship between inequality and subsequent growth.14 What drives this relationship is still very much

151

an open question, but hopefully this whets our appetite to learn some more about possible connections between inequality and development. We turn now to some other aspects of this relationship.15