Media Briefings

Inequality In Pre-Industrial Societies – And Lessons For Today

  • Published Date: March 2011

Pre-industrial societies had about the same level of measured inequality as rich, modern
societies, but their elites extracted a much greater share of the surplus than happens today.
That is the central finding of research by Branko Milanovic, Peter Lindert and Jeffrey
, which estimates inequality for 28 pre-industrial societies, ranging from first
century Rome to pre-independence India.
For example, measured inequality in the Roman Empire was about the same as in the
United States and China today. But Roman inequality was four-fifths of the maximum
possible inequality – if the elite had pushed its surplus extraction to the limit, leaving the
rest of the population at subsistence level – while today’s US and Chinese levels of equality
are less than half of maximum inequality.
The study, which is published in the March 2011 issue of the Economic Journal, concludes
that high measured inequality has to be interpreted within its social context: the same level
of inequality is much less socially destabilising in rich and/or growing economies than in
poor or declining ones. According to Branko Milanovic and his colleagues:
‘High inequality in poorer societies means that the elite has appropriated the lion’s
share of the surplus for itself. Inequality in societies experiencing declining incomes
– perhaps today’s crisis-stuck advanced economies – signals potential social trouble.
‘But societies in growing economies can accommodate higher inequality. So, for
example, growth may soften the perceived injustice of inequality in today’s China
and India.’
The researchers calculate inequality in pre-industrial societies using social tables, which
provide estimated average incomes for each salient social class – for example, nobles,
merchants, workers, peasants – and their population shares. The most famous social table
was constructed by English economist and statistician Gregory King for England and Wales
in 1688: it contains 33 classes, ranging from temporal lords to beggars and vagrants.
The new study finds that estimated inequality in pre-industrial societies is not much different
from inequality today. The most common measure of inequality is the Gini coefficient, which
ranges from 0 when all individuals have the same income to 100 when all income is
appropriated by one individual. The Roman Empire at the death of Octavian Augustus had
a Gini of 40, about the same as the United States and China today.
But, the researchers argue, while conventionally measured inequality may be similar, its
social meaning is different. In pre-industrial societies, whose average level of income was
only double or triple subsistence, most of the population lived on the same subsistencelevel
income and only a small elite made much more.
It is thus more appropriate to contrast measured inequality with the maximum possible
inequality that could exist if the elite pushed its surplus extraction to the maximum – while
allowing the populace to earn just enough for subsistence.
The extraction ratio was usually far higher in the past than it is today. Roman inequality of
40 amounted to four-fifths of the maximum inequality. In today’s United States and China,
about the same Gini is less than half of the maximum possible inequality while the
extraction ratio in egalitarian and rich Sweden is only 28%. It is only in today’s poor and
unequal societies like Congo, Nigeria and El Salvador that the inequality extraction ratio
compares with that of ancient societies.
The extraction ratio was particularly high in colonies: close to 100%, implying that the elite
there pushed inequality to its maximum. All colonisers were seemingly equally rapacious,
whether we look at inequality in French-ruled Maghreb around 1880, English-ruled Kenya
and India between the two World Wars, Moghul-ruled India around 1750 or Spanish-ruled
New Spain in 1790. Foreign elites were less interested in the welfare of local inhabitants –
and perhaps less fearful of rebellion.
The threat of rebellion may help to explain another finding of the study: that more densely
populated pre-industrial societies had both lower inequality and lower inequality extraction
ratio. Perhaps the sheer physical proximity of the masses held back the ‘grabbing hand’ of
the elite. Consider that Louis XIV thought it prudent to move from Louvre, situated in the
middle of Paris, to far-away and inaccessible Versailles.
This study of pre-industrial societies offers clues about contemporary societies. For
example, impoverished Tanzania with a relatively low Gini of 35 may be less egalitarian
than it appears since it has pushed inequality almost as far as it can go. But while much
richer Malaysia has a higher conventionally measured inequality (a Gini of almost 48), the
country’s elite has extracted only about one-half of maximum feasible inequality.
This approach suggests that the social meaning of inequality depends on economic growth.
As a country becomes richer, its feasible inequality expands. Even if recorded inequality
increases, the inequality extraction ratio need not. Growth may soften the perceived
injustice of inequality in today’s China and India.
But the opposite happens when the average income declines and inequality rises, as
occurred in Russia in the early transition. The same negative dynamics might operate in
crisis-stuck advanced economies today if inequality increases and average incomes
Notes for editors: ‘Pre-industrial Inequality’ by Branko Milanovic, Peter Lindert and Jeffrey
Williamson is published in the March 2011 issue of the Economic Journal.
Branko Milanovic is at the World Bank. Peter Lindert is at the University of California at
Davis. Jeffrey Williamson is at Harvard University.
For further information: contact Branko Milanovic on +44-(0)7580-260435 (email: or; Peter Lindert via
email:; Jeffrey Williamson via email:; or
Romesh Vaitilingam on +44-7768-661095 (email: