ACCOUNTING FOR THE GREAT
DIVERGENCE
Stephen Broadberry
London School of Economics
October 2012
1. INTRODUCTION
• Great Divergence debate has had big impact
on economic history
• Traditionally, economic historians saw
Industrial Revolution as culmination of
process of gradual improvement in Europe,
with Asia stagnating
• Pomeranz questioned eurocentric bias of this
view
2
Pomeranz
• As late as 1800, richest parts of Asia (Yangzi
Delta, South India, Japan) seen as on same
development level as richest parts of Europe
(GB and Holland)
• Great Divergence seen as C19th phenomenon,
attributable to coal and colonialism
• But revisionists did not present systematic
quantitative data
3
Accounting
• The title of the lecture is “Accounting for the
Great Divergence” and I want to use the word
“accounting” in 2 ways: measurement and
explanation
• Measurement: revisionists have exaggerated
development level of most advanced Asian
economies in 1800
• Traditional view of early modern origins of
Great Divergence is hence confirmed
4
Measurement
• Nevertheless, revisionists correct to point to
regional variation within both continents
• Within Europe, Little Divergence between
North Sea Area and Mediterranean Europe, as
Britain and Holland overtook Italy and Spain
• Within Asia, another Little Divergence
occurred, with Japan overtaking China and
India
5
Explanation
• Black Death led to permanent upward shift of
p.c. incomes in NSA, which did not occur in
rest of Europe or Asia
• Long distance trade accelerated divergence
through effects on position of merchant class
• Mixed agriculture with large pastoral
component put NSA on path to high-valueadded, capital-intensive, non-human-energyintensive production
6
Explanation
• High female age of first marriage in NSA led to
lower fertility and more human capital
• Industrious revolution more important in
explaining Little Divergences within Europe
and Asia than Great Divergence: similarities
between NSA and Japan
7
2. MEASURING ECONOMIC
GROWTH BEFORE 1870
• Until recently, most accounts of economic growth
before 1870 largely qualitative
• Maddison (2001) provided first systematic
dataset, but large amount of “guesstimation”,
with many observations set at $400 in 1990
international prices
• This is equivalent to most people living at “bare
bones subsistence”, or $1 per day, with a small
rich elite
• Recent work has established estimates based on
hard data
8
2.1 Europe
• For some European countries (e.g. GB, NL), abundant
quantitative information has survived, so sectoral
HNA can be provided in great detail
• For other countries (e.g. Italy, Spain), information
more limited, short-cut method for reconstructing
GDP developed
– Agriculture: demand depends on population, real
wage and relative price of food
– Non-agriculture: urbanisation
9
Revisions for Europe
• Data sources:
– England/GB: Broadberry, Campbell, Klein, Overton
and van Leeuwen (2011)
– Holland/NL: van Zanden and van Leeuwen (2012)
– Italy: Malanima (2011)
– Spain: Álvarez-Nogal and Prados de la Escosura
(2012)
• New estimates revise upwards level of p.c. GDP in
medieval W. Europe, which was richer than
Maddison thought
10
TABLE 1: GDP per capita levels in Europe (1990 international
dollars)
1750
1800
England/
GB
759
755
777
1,090
1,055
1,114
1,143
1,123
1,100
1,630
1,563
1,710
2,080
1820
1850
2,133
2,997
1270
1300
1348
1400
1450
1500
1570
1600
1650
1700
Holland/
NL
876
1,245
1,432
1,483
1,783
2,372
2,171
2,403
2,440
2,617
1,752
1,953
2,397
Italy
Spain
1,482
1,376
1,601
1,668
1,403
1,337
1,244
1,271
1,350
957
957
1,030
885
889
889
990
944
820
880
1,403
1,244
910
962
1,376
1,350
1,087
1,144
11
Europe’s Little Divergence
• Before Black Death in 1348, p.c. incomes
substantially higher in Italy and Spain than in
England and Holland
• Reversal of fortunes between North Sea Area and
Mediterranean Europe: by 1800 p.c. incomes
substantially higher in GB and NL than in Italy and
Spain
• Italy, England and Holland all experienced
substantial increase in p.c. incomes across Black
Death, as population fell sharply
12
Europe’s Little Divergence
• Spain did not share in this Malthusian
response to Black Death
• Italian incomes increased but fell back to preBlack Death level as population growth
returned after 1450
• Little Divergence occurred with surge in p.c.
incomes in NSA, led initially by Holland with
Golden Age 1500-1650, then by GB after 1650
13
2.2 Asia
• Raw data available in abundance for some
Asian economies (e.g. China), but still needs
much processing
• For China, can produce annual estimates of
GDP from output side, with gaps during
dynastic changes
• Data for Japan and India less abundant, but
sufficient to use short-cut method developed
for Europe
14
Revisions for Asia
• Data sources:
– Japan: Bassino, Broadberry, Fukao, Gupta and
Takashima (2012)
– China: Broadberry, Guan and Li (2012)
– India: Broadberry and Gupta (2012)
• New estimates suggest upward revision of
early GDP p.c. compared with Maddison’s
estimates, but not generally on same scale as
in Europe
15
TABLE 2: GDP per capita levels in Asia (1990 international
dollars)
730
900
980
1086
1120
1150
1280
1300
1400
1450
1500
1570
1600
1650
1700
1750
1800
1850
Japan
480
520
China
India
1,328
1,244
962
600
646
688
787
834
897
814
874
933
948
946
909
898
852
843
737
639
600
682
638
622
573
569
556
16
Asian Little Divergence
• Japan had very low levels of p.c. GDP at start of
2nd millennium, then experienced modest but
steady growth at 0.06% p.a. through to mid-C19th
• Japan’s more dynamic growth after Meiji
Restoration in 1868 thus built on this earlier
progress
• Chinese p.c. GDP, by contrast, was on a
downward trajectory from its high-point during
Song Dynasty
17
Asia’s Little Divergence
• On these estimates, Japan overtook China
during the C17th
• India shared in Chinese pattern of declining
p.c. GDP from 1600, at height of Mughal
Empire under Akbar
• However, Japan already slightly ahead of India
by 1600, when our Indian series starts
18
Asia’s Little Divergence
• Asian Little Divergence thus parallels
European Little Divergence quite closely
• If North Sea Area economies of GB, NL and
Belgium are added together, they show
continuous upward trajectory from mid-C14th
to mid-C19th, much as in Japan
• Just as stagnation and decline characterises
Europe outside NSA, so too we find stagnation
and decline in Asia outside Japan
19
Regional variation
• Of course China is a large economy, and it would
be desirable to disaggregate further
• Perhaps the Yangzi Delta was on a par with Japan
until the C19th?
• Li and van Zanden have produced a comparison
of GDP per capita in Yangzi Delta and NL in 1820s
• Li and van Zanden find per capita incomes in the
Yangzi Delta 53.8% of the level in the Netherlands
during the 1820s
20
Regional variation
• This suggests a p.c. GDP figure of around
$1,050 for Yangzi Delta (in 1990 international
dollars), slightly above Japanese level
• Recent paper by Tirthankar Roy finds p.c.
incomes in Bengal were around 20% of British
level in 1760s
• This is a bit lower than average for India (34%
in 1750, falling to 27% by 1801), as would be
expected for a relatively poor region
21
2.3 Great Divergence
• Let’s now put together Europe and Asia in Table 3
and focus on the Great Divergence
• China was richer than England in 1086. However,
we have to remember that England was a
relatively poor part of Europe in the C11th
• Comparing China with the richest part of
medieval Europe, it seems likely that Italy was
already ahead by 1300
22
TABLE 3: GDP per capita levels in Europe and Asia (1990
international dollars)
England/
GB
730
900
980
1086
1120
1150
1280
1300
1400
1450
1500
1570
1600
1650
1700
1750
1800
1850
Holland/
NL
Italy
Japan
China
India
401
428
1,328
1,244
962
754
679
755
1,090
1,055
1,114
1,143
1,123
1,110
1,563
1,710
2,080
2,997
518
525
1,245
1,432
1,483
1,783
2,372
2,171
1,403
2,440
1,752
2,397
1,376
1,601
1,668
1,403
1,337
1,244
1,271
1,350
1,403
1,244
1,350
525
572
557
627
596
639
679
948
946
909
898
852
843
737
639
600
682
638
622
573
569
556
23
Great Divergence
• But need to be careful here, since a smaller
region of China such as the Yangzi Delta may
still have been on a par with Italy in 1500
• This would be consistent with accounts in
earlier, qualitative literature
• However, with rise of Holland during its
Golden Age, there can be little doubt that
Great Divergence already well underway
during C16th and C17th
24
Great Divergence
• By this stage, discrepancy between aggregates for
China and Holland too large to be bridged by regional
variation
• Pomeranz (2011) now accepts that his earlier claim
of China on a par with Europe as late as 1800 was
exaggerated, and settles for earlier date of 1700
• Japan followed similar trajectory to North Sea Area,
but at much lower level, and continued to fall behind
West until after Meiji Restoration in 1868
25
3. EXPLAINING ECONOMIC
GROWTH
• Armed with estimates of economic growth before
1870, now turn attention to accounting for Great
Divergence in sense of explanation
• Begin by explaining Europe’s Little Divergence
before turning to Asia’s Little Divergence and
Great Divergence between Asia and Europe
26
3.1 Europe’s Little Divergence
A. Black Death
• Catching-up process in NSA starts with Black Death
of mid-C14th
• GB and Holland experienced an increase in per capita
incomes
• Spain did not share in this Malthusian response to
Black Death
• Álvarez-Nogal and Prados de la Escosura explain this
by high land-to-labour ratio in frontier economy
during Reconquest
27
Black Death
• Instead of reducing pressure on scarce land
resources, Spanish population decline
destroyed commercial networks and isolated
scarce population, reducing specialisation &
division of labour
• Although Italy shared in p.c. income gains
after 1350, they disappeared after return to
population growth from 1450
28
B. Long distance trade
• Reversal of Fortunes pivots around 1500,
when p.c. incomes approximately $1,500 in
both Italy and Holland
• NSA forged ahead after 1500, led initially by
Holland during its Golden Age (1500-1650),
later by GB
• GD often linked to developments in trade:
– New routes to Asia around south of Africa
– Europe’s encounter with Americas
29
Trade and institutions
• Might have expected Spain and Portugal to
have been gainers from these changes, since
both were pioneers and had Atlantic as well as
Mediterranean coasts
• Acemoglu, Johnson & Robinson explain
success of GB & Holland (and failure of Spain
& Portugal) through an interaction between
Atlantic access and institutional constraints on
executive power
30
AJR
• In GB & Holland, constraints on rulers
sufficient to ensure rulers unable to
appropriate bulk of gains from trade
• In Spain & Portugal, rulers sufficiently strong
to exploit opportunities themselves and
prevent strong merchant class from
constraining their powers to appropriate
31
C. Agriculture
• Success of NSA may also be linked to
agriculture and structural change.
• Agriculture in NSA more animal oriented. This
did not create more kilocalories per person,
but food was more processed
• Characteristics important for future growth:
– High value added
– Capital intensive
– Non-human energy intensive
32
TABLE 4: Share of pastoral sector in English agricultural value
added, 10-year averages (%)
1270s
1300s
1350s
1400s
1450s
At current
prices
39.9
48.8
51.2
53.7
61.6
At constant
1700 prices
30.8
33.6
46.7
42.5
46.9
1550s
1600s
1650s
1700s
1750s
1800s
1850s
1860s
41.9
41.9
35.5
40.3
42.2
51.5
55.2
60.0
39.5
41.2
36.0
38.5
45.4
54.7
55.8
55.7
33
Agriculture and structural change
• North Sea Area pulled ahead of
Mediterranean Europe as high-value-added,
capital-intensive, non-human-energy-intensive
techniques spread from agriculture to industry
and services
• And as industry and services became more
important with structural change (partly as
result of trade)
34
D. Marriage patterns
• Hajnal argued that northwest Europe had
different demographic regime from rest of
continent, characterised by later marriage and
hence limited fertility
• Although he originally labelled this European
Marriage Pattern, he later realised it applied only
to NW Europe
• This linked to labour market opportunities for
females, which de Moor and van Zanden link in
turn to pastoral agriculture
35
Marriage patterns
• Fewer children associated with more
investment in human capital
• Development in North Sea Area characterised
by human as well as physical capital intensity
• These developments helpful in breaking out of
Malthusian Trap
36
E. Industrious revolution
• One other difference between NW Europe and
southern Europe which has received attention is
attitudes to work
• Idea can be traced back originally to Max Weber
and protestant ethic, but most recent version is
“Industrious Revolution”
• This term widely associated with de Vries’s work
on Europe, but actually coined by Hayami
working on Japan
37
Industrious Revolution
• Basic idea is that people worked harder to obtain
new goods made available by long distance trade and
industrial innovation
• Following Reformation, number of holidays in Europe
reduced by around 50, and during industrial
revolution St Monday disappeared, removing
another 50 holidays
• Can see this as increasing labour intensity in SR, but
as incomes increased, savings also increased,
providing funds for investment and increasing capital
intensity in LR
38
TABLE 5: Annual days worked per person in England
Period
1433
1536
1560-1599
1578
1584
1598
1600-1649
1650-1699
1685
1700-1732
1733-1736
1760
1771
1800
1830
1867-1869
1870
Blanchard/Allen
and Weisdorf
165
180
Clark and
van der Werf
Voth
257
260
210
259
266
276
312
286
295
258
280
333
336
293-311
318
39
3.2 Asia’s Little Divergence
• Although idea of Chinese decline since Song
Dynasty not new, and Japanese post-Meiji
growth widely seen as building on foundations
laid in Tokugawa Shogunate, there is no prior
literature on an Asian Little Divergence
• Here draw out parallels with European Little
Divergence
40
A. Black Death
• Black Death did not play an important role in
Asian Little Divergence, in contrast to Europe
• No signs of positive effect from Black Death on
p.c. incomes in Asia
• Not surprising in Japan, which remained
isolated from Black Death
• However, large decline in China’s population
during C14th
41
Black Death
• This did not have positive effect on p.c.
incomes since coincided with Mongol
interlude
• Mongol interlude destroyed institutional
framework that had underpinned high p.c.
incomes of Song Dynasty
• This reduced specialisation and division of
labour, closer to experience of Spain than to
that of England or Italy
42
B. Long distance trade
• While European states engaging in voyages of
discovery during C15th, Asian states turned
inwards
• Both Japan and China adopted restrictive closed
door policy to long distance trade, which suggests
won’t help in explaining Little Divergence within
Asia
• Indeed, within Asia, India was most open to trade
43
Trade and institutions
• Returning to AJR, however, India’s participation in
long distance trade did not lead to emergence of
merchant class strong enough to impose
constraints on executive
• Indian merchants thus shared similar position to
Spanish or Portuguese merchants rather than
British or Dutch merchants
• Note that it was European rather than Asian
companies that brought Asian goods to Europe
44
C. Agriculture
• Agriculture much less animal oriented in both
China and Japan than in Europe, so this factor
again sheds little light on Little Divergence
within Asia
• But this is important in explaining Great
Divergence
45
D. Marriage patterns
• Hajnal: different marriage pattern in NW
Europe compared with rest of Europe
• Although female age of marriage in China and
India much lower, Japan was an intermediate
case, closer to experience of NW Europe
• Average age 22.1 in Japan, compared with
25.4 in England, but 18.6 in China and 13.0 in
India
46
TABLE 6: Female age of first marriage
Period
England
Japan
China
India
1600-1849
1680-1860
1550-1931
1911-1931
Range
23.4 to 26.5
18.8 to 24.6
17.2 to 20.7
12.9 to 13.3
Unweighted
average
25.4
22.1
18.6
13.0
47
E. Industrious revolution
• This later marriage in Japan also linked to
labour force participation of women, which
underpinned industrious revolution
• de Vries influenced by Hayami’s work on
Tokugawa Japan when arguing for an
industrious revolution in Europe
• On closer inspection, Hayami’s interpretation
is a bit different for Japan
48
Industrious revolution
• Indeed, Hayami generalised his idea to an East
Asian industrious revolution, based on rice
cultivation, and seen as basis of an alternative
to western capital-intensive industrialisation
• This idea picked up by Pomeranz, who argues
for a Chinese industrious revolution
• However, Huang argues that this is a
misinterpretation of what he calls “involution”
49
Industrious revolution
• For Huang, Chinese over-population led to
smaller landholdings, driving women to work
in proto-industry just to remain at subsistence
• This leaves out crucial demand side of de
Vries’s notion of an industrious revolution:
working harder to consume luxury goods
• In Western Europe and Japan, harder work
brought rising income & consumption per
head
50
3.3 Accounting for the Great Divergence
• A. Black Death: led to permanent upward shift of
GDP p.c. in NSA, which did not occur in rest of
Europe or Asia
• B. Long distance trade: accelerated divergence:
– In Europe, openness strengthened position of
merchants in GB & Holland, but not in Spain or
Portugal
– In Asia, Japan and China turned inwards and although
India remained open, merchants unable to impose
constraints on executive
51
Accounting for Great Divergence
• C. Agriculture: different agricultural systems
mattered:
– Large pastoral share put NW Europe on a path to
high-value-added, capital-intensive, non-humanenergy-intensive production
– These techniques spread from agriculture to
industry and services, which accounted for a
growing share of output
52
Accounting for Great Divergence
• D. Marriage patterns also mattered:
– High female age of first marriage in NW Europe,
led to lower fertility and more human capital
– Early female marriage in India & China, Japan an
intermediate case
• E. Industrious revolution more important in
explaining Little Divergences within Europe
and Asia than Great Divergence: similarities
between NSA and Japan
53
4. CONCLUSIONS
• A. Measurement
• Traditional view of early modern origins of Great
Divergence is confirmed
• But revisionists correct to point to regional variation
within both continents
• Little Divergence within Europe: reversal of fortunes
between North Sea Area and Mediterranean Europe
• Little Divergence within Asia: Japan overtook China
and India
54
Conclusions
• B. Explaining the Great Divergence
• Black Death led to permanent p.c. income gain in
NSA, but not in rest of Eurasia
• Long distance trade accelerated divergence through
effects on merchant class
• Pastoral agriculture put NSA on capital-intensive path
• Late marriage in NSA lowered fertility
• Industrious revolution more important in explaining
Little Divergences within Europe and Asia
55
More research needed!
• I would claim that historical national accounting has
now made a substantial contribution to
understanding the Great Divergence
• But there is more to be done
• Historical national accounts needed for more
countries, reaching further back in time
• More regional disaggregation needed within large
countries
• Much more comparative data needed on the
explanatory variables
56