In a 2007 study, William Easterly used cross-country
data to examine the Engerman and Sokoloff hypothesis.
His research confirmed that “agricultural endowments
predict inequality and inequality predicts
development.” Specifically, Easterly finds that inequality
negatively affects per capita income; it also
negatively affects institutional quality and schooling,
which are “mechanisms by which higher inequality
lowers per capita income.” That the negative relationship
between income and inequality is present in the
data is clear—but how do development economists
take the step to prediction and assignment of causality
when measurement error and many confounding factors
are present, such as the possible link that underdevelopment
itself is a cause of inequality?
Sometimes development economists run field
experiments—but obviously, we cannot randomly
assign countries various levels of inequality to see
what happens! In the many cases when field experiments
are impossible, development economists frequently
try to understand causality by searching for
an instrumental variable (or “instrument”); in fact,
many researchers in development economics invest a
lot of their time in this search. This is a topic covered
in classes in econometrics. But the basic idea is that to
identify the effect of a potential causal variable c
(such as inequality) on a development outcome variable
d (such as income or educational attainment),
the hunt is on for an elusive instrumental variable e
that affects d only through e’s effect on c. So an instrument
has no independent effect on the outcome
variable of interest. You saw earlier that Acemoglu,
Johnson, and Robinson used settler mortality as an instrument
for early institutions. Easterly uses “the
abundance of land suitable for growing wheat relative
to that suitable for growing sugarcane” as an instrument
for inequality. Using this strategy, Easterly concludes
that high inequality of the Engerman and
Sokoloff variety is independently “a large and statistically
significant barrier to prosperity, good quality institutions,
and high schooling.” Schooling and institutional
quality are precisely the mechanisms proposed
by Engerman and Sokoloff by which higher inequality
leads to lower incomes. Like a leprechaun, a good instrumental
variable is hard to get hold of but when
caught can give the researcher’s equivalent of a pot of
gold. Though active debate on inequality and development
continues, the interplay between the careful
institutional analysis and economic history scholarship
of Engerman and Sokoloff and the study of
causality with larger data sets as used by Easterly gives
a window into how the field of development economics
continues to make progress.