The Effect of
Does the fact that contemporaneous changes in mean
incomes and inequality do not seem to be systematically related
necessarily imply that there is no link between the two variables at
all? Could it be, for instance, that countries with higher initial
(i.e. ex ante) inequality grow more slowly (or faster) than
others? And if any such pattern emerged, what would explain it? Are
there causal links between the dispersion of an income (or
expenditure, or wealth) distribution and economic growth, or indeed
other aspects of economic performance?
In the 1990s, the classical view that distribution
(one aspect of which is measured by inequality indices) is not only
a final outcome, but in fact plays a central role in determining
other aspects of economic performance, has come back into fashion.
While many economists often start working on a topic at the same
time, much of the credit for pioneering this line of enquiry must go
Galor and Joseph Zeira, whose 1993 paper on "Income Distribution
and Macroeconomics" concluded thus: "In general, this study shows
that the distributions of wealth and income are very important from
a macroeconomic point of view. They affect output and investment in
the short and in the long run and the pattern of adjustment to
exogenous shocks. It is, therefore, our belief that this
relationship between income distribution and macroeconomics will
attract more studies in the future." (1993, p.51).
They were certainly right in their last prediction.
And most of the studies that followed concurred that they were right
on the first point as well. Empirically, the proposition that
initial inequality seemed to be associated with lower growth rates
was put forward by Persson
and Tabellini (1994) and Alesina
and Rodrik (1994). Using the data sets available to them, both
studies found that inequality variables had significantly negative
coefficients in growth regressions, when controlling for a number of
the usual right-hand side variables, such as initial income,
schooling and physical capital investment. A survey by Benabou
(1996a) listed a number of other cross-country empirical
investigations of this relationship, and reported that the vast
majority of them reached the same conclusion.
But the debate is not as settled as that survey
implied. Since then, a number of studies drawing on the
Deininger-Squire database - which is superior to those available to
Persson and Tabellini or to Alesina and Rodrik, both in number of
countries and in time-spans covered – have questioned the budding
consensus. Adding a number of econometric methodological
improvements to this better data set, Forbes (1997) actually finds a
positive and significant relationship between inequality and growth.
Others have echoed her concerns and cautioned against a premature
acceptance of the inverse relationship between initial inequality
and growth as a new stylized fact of development economics.
However, the econometric problems that seem to
beset the negative relationship in the newer data sets appear to be
specific to inequality variables defined in the income (or
expenditure) space. Deininger and Squire themselves, for instance,
find that the negative coefficient on initial income inequality in
their growth regressions becomes insignificant only when a variable
for asset inequality (the Gini coefficient for land
ownership) is introduced. Birdsall
and Londono (1997) investigate a similar relationship for the
other asset crucially important for the poor, namely human capital.
Also using a subset of the Deininger-Squire database, they conclude
that: "…initial inequalities in the distribution of land and of
human capital have a clear negative effect on economic growth, and
the effects are almost twice as great for the poor as for the
population as a whole." (p.35) Once again, once these asset
distribution variables are included, the significance of income
The empirical issue is clearly not settled.
Nevertheless, it would seem fair to report the current state of the
debate as follows: while initial income inequality may not
directly affect an economy’s aggregate growth potential, others
thing being equal, it does proxy for more fundamental inequalities
of wealth. Once measures for those are included, there seems to be a
significant negative relationship between asset inequality
Now, why would that be the