Pre-industrial Inequalities

1
Pre-industrial Inequalities
Branko Milanovic World Bank Training Poverty and
Inequality Analysis Course March 3, 2011
2
Questions

• ? Is inequality caused by the Industrial
  Revolution?
• Or, has inequality been pretty much the same
  before and after?
• ? Is inequality in poor pre-industrial economies
  today pretty much the same as in ancient
  pre-industrial economies?
• ? Was inequality augmented by colonization?
• ? Have some parts of the world always had
  different levels of inequality than others?

3
Constraints on the Elite in Ancient
Pre-Industrial Societies

• ? Fact Ancient pre-industrial societies had
  average income levels usually twice, but
  sometimes 4-5 times, the subsistence level.
• ? Fact Low average income, combined with the
  requirement that few fall below subsistence,
  meant that the elites surplus (and thus
  inequality) could not be very large.
• ? Query What happened when average income and
  the potential surplus rose? Did the poor
  subsistence workers get any the added surplus or
  did the elite grab it all?

4
A New Measure the Inequality Possibility Frontier

• Divide society into 2 groups people with
  subsistence income and elite (fraction e of total
  population) that shares the surplus equally among
  themselves.
• There is no overlap between the two classes, and
  no inequality within each.
• Then, the Gini simplifies to

5

• Per capita income of the elite is

where Ntotal population, µoverall mean income,
ssubsistence. Per capita income of people is
s and respective population shares are e and
(1-e). Substituting all of this into Gini
gives
6
If, for simplicity, we express µ as so many (a)
subsistence minimums, the Gini becomes
IMPORTANT The expression gives the maximum Gini
compatible with mean income of as e fraction of
the elite, and no inequality among either elite
or people. When e tends to 0 (one Mobutu), G
(a-1)/a. With a1, G0 a2, G0.5 if a100
(like in the US today), G0.99. Introduction of
inequality among the elite does not affect the
maximum Gini.
7
Other interpretations

• This is the maximum inequality which may exist
  at a given income level when the entire surplus
  income is appropriated by (at the extreme) one
  individual.
• The size of the overall income (the pie) limits
  the level of measured inequality (measured by the
  synthetic measures like the Gini where all
  incomes matter).
• It is a new and realistic generalization of the
  Gini index since it requires that the society be
  sustainable.

8
New Measurement of Inequality

• The ratio between the actual Gini and the maximum
  Gini (a point on the IPF) is the inequality
  extraction ratio.
• The inequality extraction ratio shows what
  percentage of maximum feasible inequality an
  elite is able (or wishes) to extract ratio A/B
  (next slide).

9
The locus of maximum inequalities is inequality
possibility frontier
B
A
Note Vertical axis shows maximum possible Gini
attainable with a given a.
10
How are we going to study ancient inequalities

• There are no household survey data, but..
• There are social tables akin to Kings 1688
  table.
• We shall use mostly the social tables that have
  already been produced or the data that can allow
  us to produce such tables (in some cases from
  professional censuses). Plus Ottoman censuses of
  settlements (2 cases)
• Inequality (Gini) calculated from such tables
  assumes that (i) all members of a group have the
  same income, and (ii) groups are non-overlapping
  (i.e, all members of an upper group have higher
  incomes than all members of a lower group). This
  is our lower-bound Gini1.

11

• We relax assumptions (i) by calculating maximum
  feasible inequality within the income ranges of
  the groups (thus allowing for an estimate of
  within-group inequality). But we have to keep
  (ii) although we know that there are members of
  (say) nobility who may have lower income than
  some merchants. This is our upper-bound Gini2.
• The ratio between Gini2/G estimates inequality
  extraction ratio for a given country.

12
What countries do we include?

• Wherever we could find a social (class) table
  with estimated mean class income and population
  shares.
• We set time limits for the developed world,
  1810 for the rest, 1929 (with India 1947 as an
  exception).
• Difficult decision to decide what is a country
  an officially distinct territory with autonomous
  or foreign government (the latter is a colony).
• We do not include cities (Jerez, Paris, Amsterdam
  for which data exist).

13

• This leaves us with 30 data points, ranging from
  Rome 14 to India 1947.
• Four data points from England (1230, 1688, 1759,
  1801) and three from Holland though (1561, 1732,
  1808)
• Number of social classes mostly in double digits
  except in Nueva España and China (3 classes
  only), Moghul India (4) and England 1290 (7).
  Median number of classes 20, but Tuscany (1427)
  almost 10,000 households, Levant (1596) 1415
  settlements.
• Does number of classes matter? Sensitivity
  analysis suggests Not (see below).
• Estimated per capita incomes in 1990 PPP almost
  all from Maddison if not, use the ratio between
  the estimated mean LC income and estimated
  subsistence (a) and price the latter at PPP 300
  (Byzantium paper)
• In the sample, a ranges from 1.6 to 6.7 (based on
  a subsistence minimum of PPP 300).

14
An example of a social table France 1788
Social Group Population (in 000) Per capita income (livres per annum) Population
Nobles and Clergy 540 724.1 1.9
Bourgeoisie 2160 724.1 7.7
Shopkeepers and artisans 3240 150.0 11.6
Workers (non agricultural) 1500 66.7 5.4
Servants (non agricultural) 1080 92.6 3.9
Small scale farmers 5250 64.6 18.8
Large scale farmers 2250 219.6 8.0
Agricultural day laborers and servants 10150 39.4 36.3
Mixed workers 1800 75.0 6.4
Total 27970 143.3 100
Source Morrisson and Snyder (2000)
15
Data Sources, Estimated Demographic Indicators
and GDI Per Capita(Contd.)
Country/territory Source of data Year Number of social classes Population (in 000) Estimated GDI per capita
Roman Empire Social tables 14 11 55000 633
Byzantium Social tables 1000 8 15000 533
England Social tables 1290 7 4300 639
Tuscany Household survey 1427 9,780 38 978
South Serbia (w/o foreign) Census of settlements 1455 615 80 443
Holland Tax census dwelling rents 1561 10 983 1129
Levant Census of settlements 1596 1,415 237 974
England and Wales Social tables 1688 31 5700 1418
Holland Tax census dwelling rents 1732 10 2023 2035
Moghul India Social tables 1750 4 182000 530
Old Castille Income census 1752 33 1980 745
England and Wales Social tables 1759 56 6463 1759
16
Data Sources, Estimated Demographic Indicators
and GDI Per Capita
Country/territory Source of data Year Number of social classes Population (in 000) Estimated GDI per capita
France Social tables 1788 8 27970 1135
Nueva España Social tables 1790 3 4500 755
England and Wales Social tables 1801-3 44 9277 2006
Bihar (India) Monthly census of expenditures 1807 10 3362 533
Netherlands Dwelling rents 1808 20 2100 1800
Kingdom of Naples Tax census dwelling rents 1811 12 5000 637
Chile Professional census 1861 32 1702 1295
Brazil Professional census 1872 813 10167 721
Peru Social tables 1876 9 2469 653
China Social tables 1880 3 377500 540
Java Social tables 1880 32 20300 661
Japan Tax records 1886 21 38622 916
Java (w/o foreign) Social tables 1924 12 34984 909
Siam Social tables 1929 21 11607 793
British India Social tables 1947 8 346000 617
Notes GDI per capita is expressed in 1990
Geary-Khamis PPP dollars (equivalent to those
used by Maddison 2003 and 2004).
17
18th century included countries
18
19th century included countries
12 countries before the French revolution, 18
countries after No social tables for the United
States (!), Russia, Africa (except Kenya and
Maghreb)
19
but more may be coming

• American colonies 1776/1800 (Lindert and
  Williamson working on it)
• Czarist Russia (Mironov)
• Poland
• Mehmet Alis Egypt
• More Ottoman defters
• Madagascar
• Audiencia de Quito

20
Kingdom of Naples around 1810
21
Map of Levant 1596-97 (yellow areas included)
22
Inequality Measures
Country/territory/ year Gini1 Gini2 Maximum feasible Gini with s300 Actual Gini as of the maximum
Roman Empire 14 36.4 39.4 52.6 75
Byzantium 1000 41.0 41.1 43.7 94
England and Wales 1290 35.3 36.7 53.0 69
Tuscany 1427 46.1 69.3 67
South Serbia (w/o foreign) 1455 19.1 20.9 32.2 65
Holland 1561 56.0 73.4 76
Levant (w/o foreign) 1596 39.8 69.1 67
England and Wales 1688 44.9 45.0 78.8 57
Holland 1732 61.0 61.1 85.2 72
Moghul India 1750 38.5 48.9 43.4 113
Old Castille 1752 52.3 52.5 59.7 88
England and Wales 1759 45.9 45.9 82.9 55
France 1788 54.6 55.9 73.5 76
23
Inequality Measures
Country/territory/ year Gini1 Gini2 Maximum feasible Gini with s300 Extraction ratio Actual Gini as of the maximum
Nueva España 1790 63.5 62.0 105
England/Wales 1801-3 51.2 51.5 85.0 61
Bihar (India) 1807 32.8 33.5 43.7 77
Netherlands 1808 56.3 57.0 83.3 68
Naples 1811 28.1 28.4 52.9 54
Chile 1861 63.6 63.7 76.8 83
Brazil 1872 38.7 43.3 58.3 74
Peru 1876 41.3 42.2 54.0 78
China 1880 23.9 24.5 44.4 55
Java 1880 38.9 39.7 54.6 78
Japan 1886 39.5 67.2 59
Java 1924 31.8 32.1 66.9 48
Siam 1927 48.4 48.5 62.1 78
British India 1947 48.0 49.7 51.3 97
24
Estimated Gini Coefficients and the Inequality
Possibility Frontier
Note The IPF is constructed on the assumption
that sPPP300. Estimated Ginis are Ginis2 unless
only Gini1 is available
25

• At alt3, Ginis range from 25 to low 60s and are
  clustered around the IPF. These countries
  extract quite a large share (on average 80
  of maximum inequality).
• With higher mean income, as the IPF becomes
  higher, Gini does not rise to the same extent,
  and the extraction ratio goes down.
• This is true when we compare ancient and modern
  societies, but true within ancient as well as
  within modern (application of IPF methodology to
  the contemporary societies see below)
• All countries with the extraction ratio around
  100 were colonies Moghul India 1750 (112),
  Nueva España 1790 (105), Maghreb 1880 and Kenya
  1927 (100), Kenya 1914 (96). 4 different
  colonizers.

26

• For the ancient, if alt3, the median Gini is 42
  and median extraction 78 (n18). If agt3, the
  median Gini is 49 and median extraction 64
  (n12). Ho of ? extraction accepted (p0.999), Ho
  of ?Gini accepted (p0.972 Kuznets).
• Thus, Gini alone is not a sufficient measure of
  inequality.
• A Gini of (say) 40 in Rome and in the US does not
  mean the same thing. In Rome, that Gini extracts
  75 percent of maximum inequality, in the US less
  than 40 percent.

27
Ginis and the Inequality Possibility Frontier for
the Ancient Society Sample and Selected Modern
Societies
Note Modern societies are drawn with hollow
circles. IPF drawn on the assumption of sPPP
300 per capita per year. Horizontal axis in logs.
28
Inequality extraction ratio for the ancient and
the same modern societies
All but one, colonies!
Based on the subsistence minimum PPP300.
29
Highlight colonies extraction ratio
30
Distribution of the extraction ratio across three
types of society
Use Figure25.do file (bottom graph)
31
Relationship between GDI per capita and
extraction ratio for ancient societies only
Note 95 percent confidence interval
32
Can we try to explain determinants of ancient
inequalities and extraction ratio?

• Paucity of data points (30 in total) and possible
  explanatory variables
• However, we have some income per capita
  (Kuznetsian relationship), urbanization rate,
  population density, dummy for being a colony

33
Gini determinants
First cut Is Asia different? Drop 2 Javas
Ln GDI pc 360.5 366.7 360.2
(Ln GDI pc)2 -25.0 -25.5 -25.0
Urbanization 0.349 0.354 0.353
Pop. density -0.105 -0.100 -0.107
Colony 12.63 12.93 12,41
Asia -1.28
No foreign -9.59 -9.97 -9.26
No. of groups -0.009 -0.01 -0.01
Tax survey -4.86 -4.85 -4.85
Adjusted R2 (N) 0.75 (28) 0.73 (28) 0.73 (26)
34
And the extraction ratio
Parsimonious Add pop density Drop 2 Javas
Ln GDI pc -20.92 -6.48 -6.45
Urbanization 0.677 0.229 0.236
Pop. density -0.188 -0.200
Colony 16.12 25.52 25.35
No foreign -25.28 -39.20 -39.23
Adjusted R2 (N) 0.34 (28) 0.65 (28) 0.60 (26)
35
Drawing together Gini and the extraction ratio

• Kuznets quadratic relationship relatively strong
  for Gini, but income negatively associated with
  the extraction ratio (as we saw before)
• Asynchronism in the behavior of the Gini and
  extraction ratio as societies get richer Gini at
  first ?, but the extraction ratio ? throughout
• Population density puts downward pressure on both
  Gini and the extraction ratio. The effect on the
  latter particularly strongso much that both
  urbanization and income lose significance
• Colony very significant adds 12-13 Gini points,
  and twice as many extraction points throughout
• Controls for different types of surveys and
  number of social groups not significant

36
Other implications

• Asia (absence of economies of scale in the
  cultivation of rice) does not appear to have been
  more equal in Gini terms population density more
  important (although high population might have
  been made possible by the absence of extreme
  inequality)
• No causality can be proven.
• 2 possibilities (i) less extractive regimes
  however they might have arisen-- allow
  population to increase (ii) greater population
  density ---however it happened-- threatens the
  rulers more so the extraction ratio goes down
  (Campante and Do). Think why Louis XIV moved from
  Louvre to Versailles.
• Most likely both effects operate and impossible
  to disentangle them
• IMP Why and how population density limits
  elites predatory power?

37
Other implications (cont.)

• Re. Engelman-Sokoloff Ho If Western Europe was
  as unequal as Latin America, why were the
  trajectories of the two so very different in
  19th-20th century?
• W. European mean Gini (1500ltyearlt1810 8 obs)
  LA mean Gini (4 obs) in 19th century 53. But
  Europes extraction ratio 70 vs. LA 85.
• Their Ho should be recouched in extraction, not
  Gini terms

38
Two propositions

• Proposition 1. While the estimated Ginis for
  pre-industrial societies fall in the same range
  as inequality levels observed today, ancient
  inequality was much greater when expressed in
  terms of the maximum feasible inequality.
• Proposition 2. Under conditions of economic
  growth, particularly in poor or middle-income
  societies, constant inequality reflects great
  restraints on exploitation because the inequality
  extraction ratio is falling. The reverse is true
  during periods of economic decline (e.g., Russia
  under Yeltsin).

39
Global inequality and poverty

• If we take all 12 countries within years 1750 and
  1880, we have 583 income groups representing
  incomes of almost 650 million people.
• Over that period, average world population was
  around 900 billion.
• These LC incomes are converted into PPP
  (Geary-Khamis, 1990)
• What is inequality among world citizens, and
  poverty rate?

40

• Gini for these individuals is 38.2. This is only
  about a half of global Gini today (70 with the
  new PPP data 65 with the old PPP data).
• The poverty headcount (with the PLPPP410) is 85
  percent. Crucially depends on China.

41
Global inequality then and now
MLW data
1820, 1870 from Bourguignon and Morrisson, 2005
from Milanovic
42
Global poverty then and now(much more dependent
on the assumption re. income of the poor in China
than inequality calculations)
MLW data
1820, 1870 from Bourguignon and Morrisson, 2005
from Chen and Ravallion
43
Who were the people with the highest incomes then?

• European colonizers in Java about 2,500 people
  had per capita incomes in excess of PPP90
  100,000.
• Also a few hundred people in England 1759 and the
  Netherlands 1808. (English top income group in
  1801-3 is broader.)
• Incidentally, the rich British in 1947 India had
  an average per capita income in excess of PPP90
  50,000 which would place them in the 2nd richest
  percentile in the US today.
• Little wonder colonies were good for colonizers!

44
An added dimension the share of top 1

• Recent work (Piketty etc) implies that there is a
  strong correlation between the top 1 (and fewer)
  income share and inequality.
• Is it true in ancient societies?
• Caveat these are not true distributions of
  people or families but of social classes.
• Estimate the top share using Pareto interpolation
  (assumes Pareto distribution at the top).

45
Estimated top of income distribution ancient and
modern counterparts
Top 1 share in total income (in ) The cut-off point (in terms of mean income) Gini coefficient
Byzantium 1000 30.6 3.7 41.0
Chile 1861 25.7 11.8 63.7
China 1880 21.3 5.6 24.5
Nueva España 1790 21.1 9.8 63.5
Japan 1886 19.1 39.5
Netherlands 1808 18.1 9.8 57.0
France 1788 16.8 6.9 55.9
Rome 14 16.1 12.4 39.4
England 1801 8.9 6.2 51.5
England 1688 8.7 6.1 45.0
Old Castille 1752 7.0 6.2 52.5
Siam 1929 6.7 5.1 48.5
Average ancient 14.6 7.4 45.4

Average modern counterparts 8.6 5.4 42.1
Chile 2000 14.6 7.9 54.6
UK 1999 7.0 4.3 37.4
India 2004 5.2 4.2 32.6
46
Weak correlation (?0.45) between Gini and top 1
income share
twoway scatter top_percent gini if sample1,
msize(vlarge) mlabel( country)
47
Top five percentiles of income distribution in
Rome 14, Byzantium 1000, and England 1688
Note All data points except for the top 1
percent are empirical. The top 1 percent share is
derived using Pareto interpolation.
48
Embourgeoisement of England increasing share of
top 5 and declining share of top 1
Based on per capita transformation of King,
Massie and Colquhoun social tables
49
Third proposition (re the top shares)

• ? Fact The share of the top percentile in
  ancient societies is not tightly connected with
  overall inequality in contrast with modern
  societies.
• ? Proposition 3. What drove ancient inequality
  was not the top share, but rather the size of the
  income gap between average income (y) and the
  average income of poor (w) y/w.

50
(No Transcript)
51
Five take-away observations

• ? Measured annual inequality is not very
  different in pre-industrial societies today than
  it was in ancient societies.
• ? New measure of inequality maximum inequality
  compatible with preservation of a society the
  inequality possibility frontier.
• ? The extraction ratio how much of potential
  inequality was converted into actual was much
  bigger in ancient societies.
• ? In contrast with modern societies, the top 1
  share was not correlated with overall inequality
  in ancient societies. But the gap between elite
  or average income and poor peoples incomes was
  correlated with overall inequality.
• ? Can we contrast equal societies with a very
  small and very rich elite (Oriental despotism)
  vs. those with a more graduated (diversified)
  income structure?

52
Moving to the present the use of the IPF and
extraction ratio

• Maximum Gini a new upper bound on the Gini such
  a society is sustainable in the long-run.
• More realistic Gini.
• Extraction ratio reflection both of the level of
  development and rapacity of the elites (or their
  ability to appropriate the surplus).

53
The extraction ratio and GDI per capita (year
2002)
54
Gini and GDI per capita (year 2002)
Using ineq_frontier.do file
55
Probability of civil war (1990-97) as function of
inequality or extraction ratio in the period
1970-1990
Mean HBS income (ln) -0.319 (0.002) -0.238 (0.000) -0.410 (0.000) -0.321 (0.000)
Gini (in ) 0.0004 (0.82) 0.0015 (0.49)
Extraction ratio (in ) 0.0075 (0.00) 0.012 (0.000)
Democracy(Polity2) 0.056 (0.000) 0.058 (0.000)
Ethnolinguistic fract. 0.765 (0.000) 0.675 (0.000)
Pseudo R2 0.042 0.046 0.097 0.104
No. of obs 427 427 381 381
Civil war within-war variable from CoW
project my gdppppreg.dta file weighted probit
probit civil_warCoW Giniall lngdpppp if
yeargt1970 yearlt1990 whhh