The Colden Age of European growìh

reconsidered
PETER TEMIN
Department cf Eccncmtcs, Massachusetts Instttute cf Technc/cgy, Cambrtdge
M¬ o:I¡:-I¡¡;, US¬
I reconsider ìhe growìh of Wesìern Europe during ìhe Colden Age of
European Economic Crowìh afìer ìhe Second World War. The preceding
ìhirìy years of conflicì and depression impeded ìhe normal paìh of
indusìrialisaìion in ìhese counìries, and ìhey had ìoo much labour in
agriculìure for ìheir level of income and sìage of developmenì aì ìhe end
of ìhe war. The disequilibrium added ìo oìher more ordinary forces ìo
produce unusually rapid economic growìh. This hypoìhesis explains ìhe
speed of economic growìh during ìhe Colden Age, differences beìween
growìh raìes in ìhese years, and ìhe end of ìhis hisìorical episode.
Iì is hardly news ìhaì ìhe years following World War II were far differenì
from ìhose following World War I. Economisìs wriìing during ìhe war
anìicipaìed repeìiìion of some of ìhe depressing forces and evenìs ìhaì fol-
lowed ìhe Creaì War (Samuelson 1ç(¸). Buì ìheir predicìions were noì
accuraìe, aì leasì parìly because of ìheir sìudies. Policymalers had ìhe
experience of ìhe inìerwar years ìo reflecì on, and iì is comforìing ìo ìhinl
ìhaì ìhey learned from experience (Ieinsìein, et a/. 1çç().
The good ìimes came ìo an end in ìheir ìurn during ìhe oil crises and
‘sìagflaìion’ of ìhe 1ç;os. We lool bacl on ìhese ìimes perhaps wiìh more
nosìalgia ìhan may be warranìed, giving ìhem names lile ìhe Colden Age of
European Crowìh and /es Trente C/crteuses. Economisìs since ìhen have
been ìrying ìo undersìand boìh ìhe sources of ìhe rapid growìh immediaìely
afìer ìhe Second World War and of ìhe slowdown in ìhe 1ç;os. This arìicle
is a conìribuìion ìo ìhaì liìeraìure.
An explanaìion of ìhe Colden Age of European Crowìh should answer
ìhree quesìions. Why was European economic growìh so rapid beìween ìhe
world war and ìhe firsì oil crisis? Why did differenì counìries grow aì differenì
raìes during ìhis ìime? And why did ìhe rapid growìh come ìo an end? I argue
ìhaì ìhese quesìions can be answered in a unified frameworl by bringing econ-
omic hisìory ìo bear on ìhis quesìion of economic growìh. The preceding
years of wars and depression impeded ìhe process of indusìrialisaìion ìhaì had
engaged ìhe economies of Wesìern Europe since aì leasì ìhe mid-nineìeenìh
cenìury. There was as a resulì a disequilibriumìhaì has noì been noìed before,
ìhaì was ìhe source of ìhe rapid and varied growìh during ìhe Colden Age.
Eurcpean Reztev cf Eccncmtc Htstcry, 6, ¸–zz. Prinìed in ìhe Lniìed Kingdom © zooz Cambridge Lniversiìy Press
I proceed by describing ìhe phenomenon ìo be explained and reviewing
earlier aììempìs ìo explain iì. I add insighìs from recenì research in econ-
omic hisìory ìo propose a new explanaìion. I ìhen formulaìe ìhis hypoìhesis
expliciìly and ìesì iì againsì ìhe daìa. Iinally, I reìurn ìo ìhe ìhree quesìions
posed above and summarise ìhe new answers.
t. The phenonenon to be expluined
The phenomenon ìo be explained is shown in Table 1. The difference
beìween ìhe growìh raìe of CDP and CDP per captta comes from ìhe grad-
ual slowdown of populaìion growìh, and ìhe growìh raìe of CDP per captta
in recenì years is very close ìo iìs raìe before ìhe Creaì War. In beìween,
Wesìern Europe had firsì slow growìh and ìhen rapid growìh. Iì is ìhe laììer
I am ìrying ìo explain.
Slow growìh from 1ç1¸ ìo 1ç¸o was ìhe resulì of ìwo world wars and ìhe
Creaì Depression. Iì is common ìo regard ìhe Creaì Depression as a failure
of aggregaìe demand. Prices fell aì ìhe same ìime as indusìrial producìion,
indicaìing a movemenì along an aggregaìe supply curve raìher ìhan a shifì
of ìhaì curve (Bernanle 1çç¸). Alìhough ìhe wars had many effecìs on ìhe
supply side, ìheir primary impacì was also on demand (Ieinsìein, et a/.
1çç;). To a firsì approximaìion, ìherefore, ìhe slow growìh was ìhe resulì of
deficienì aggregaìe demand.
Iì follows ìhaì ìhe overall paìh of CDP per captta in Table 1 can be seen
as a sìeady growìh of ‘poìenìial CDP’ wiìh a deviaìion from ìhis poìenìial
during ìhe world wars and Creaì Depression. Toìal facìor producìiviìy in
ìhis view conìinues on iìs way, independenì of all ìhe demand-side acìiviìy
in ìhe wars and inìerwar ìurbulence. This exìreme version of Solow growìh
ìheory ignores all flucìuaìions in ìhe raìe of growìh of lnowledge and of cap-
iìal, buì iì does noì seem ìo be ìoo far from ìhe experience of ìhe Lniìed
Sìaìes where we have ìhe daìa ìo lool aì ìhe early ìwenìieìh cenìury (Solow
1ç¸;). Slow growìh from 1ç1¸ ìo 1ç¸o ìhen lefì Wesìern Europe below iìs
poìenìial CDP, and rapid growìh ìhereafìer broughì iì bacl ìo iìs growìh
paìh.
( Eurcpean Reztev cf Eccncmtc Htstcry
Table 1. Eccncmtc grcvth tn 1estern Eurcpe at dtfferent ttmes ;per
cent per year).
Period CDP CDP per captta
18ço–1ç1¸ z.z 1.(
1ç1¸–1ç¸o 1.( o.ç
1ç¸o–1ç;¸ (.8 (.o
1ç;¸–1çç( z.1 1.;
Scurce: Ieinsìein, et a/. 1çç;, p. ;. Iifìeen counìries, daìa from Maddison (1çç¸).
The problem is ìhe lengìh of ìime in Table 1. The business cycle can pro-
duce a paìh lile ìhis wiìh a ìime span of a year or ìwo. New growìh ìheory
and condiìional convergence can produce a hisìory lile ìhis wiìh conver-
gence wiìh abouì ¸¸ years or so ìo half-way convergence (Manliw, et a/.
1ççz). Iì is harder ìo find a good explanaìion for ìhis inìermediaìe ìime
frame, for quiìe compleìe convergence in abouì ìhirìy years. Business cycles
generally are a demand phenomenon (Temin 1çç8), condiìional conver-
gence involves supply phenomena. Iì would noì be surprising if ìhe expla-
naìion of ìhe inìermediaìe case involved boìh demand and supply.
While ìhis simple model of deviaìion froma smooìh ìrend is appealing, I do
noì wanì ìo suggesì ìhaì ìhe disequilibriumsìudied here was ìhe only phenom-
enon ìaling place afìer ìhe war. As noìed in several sìudies, several European
counìries did noì reìurn ìo ìheir prewar ìrend paìhs of growìh afìer ìhe war or
evenafìer ìhe ColdenAge (Crafìs andMills 1çç6). Ior ìhose counìries, ìhe end
of ìhe Colden Age represenìed a reìurn ìo a more durable growìh paìh, buì noì
necessarily ìhe same one ìhey had experienced before ìhe war. I reserve for
fuìure worl ìhe inìegraìion of ìhe Colden Age and ìhe subsequenì growìh paìh.
In addiìion ìo ìhe ìime-series quesìions abouì ìhe beginning and end of ìhe
Colden Age, ìhere also is a cross-secìion quesìion: Why did some counìries
grow so much more rapidly in ìhis period ìhan oìhers? The spread of growìh
raìes among Wesìern European counìries in ìhis period is shown in Table z.
Annual raìes of growìh varied fromìwo ìo five per cenì a year. Iì is a wide range
and needs ìo be explained. Naìional hisìories always conìain developmenìs
ìhaì can be used ìo explain rapid or slow growìh, ìhe more challenging ques-
ìion is wheìher ìhere is a unified explanaìion for ìhe varieìy shown in Table z.
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered ¸
Table z. ¬nnua/ rates cf grcvth tn 1estern Eurcpe, Iµyy–;y ;per cent
per year).
Ausìria ALT (.¸
Belgium BEL ¸.¸
Swiìzerland CHE z.(
Cermany DEL ¸.(
Denmarl DNK ¸.z
Spain ESP ¸.z
Iinland IIN ¸.8
Irance IRA ¸.ç
Creaì Briìain CBR z.1
Ireland IRL ¸.z
Iìaly ITA (.¸
Neìherlands NLD ¸.(
Norway NOR ¸.¸
Porìugal PRT ¸.¸
Sweden SWE ¸.o
Scurce: Penn World Tables ¸.6.
An early conìribuìion ìo ìhe liìeraìure on posìwar growìh was provided
by Kindleberger (1ç6;) using ìhe Lewis (1ç¸() model of excess labour
supply ìo explain boìh differences in growìh raìes beìween counìries and ìhe
slowdown in growìh he could deìecì in ìhe mid 1ç6os. Kindleberger’s argu-
menì was simple: an elasìic labour supply promoìes economic growìh by
leeping wages low and preserving indusìrial peace. Iì was ìhe exhausìion of
cheap labour ìhaì caused economic growìh ìo slow. This arìicle builds upon
and exìends Kindleberger’s view of ìhirìy years ago.
The slowdown of growìh in ìhe 1ç;os, lnown aì ìhe ìime as sìagflaìion,
was ìhe subjecì of myriad papers and bools. Many people argued ìhaì
movemenìs in aggregaìe supply led ìo ìhe slowdown of growìh as well as
higher inflaìion. The ìwo shocls mosì ofìen idenìified were ìhe rise in oil
prices in 1ç;¸ and increasing rigidiìy in indusìrial labour marleìs (Bruno
and Sachs 1ç8¸). The oil shocls have faded inìo hisìory while remaining ìhe
mosì popular candidaìes for causing ìhe end of ìhe Colden Age.
Characìerisìics of ìhe labour marleì conìinue ìo be acìive ìopics in ìhe
explanaìion of European economic difficulìies.
The focus on supply condiìions led ìo new growìh ìheory, which sìressed
ìhe role of supply in ìhe long run. Solow’s frameworl had provided a way
ìo organise hisìorical daìa on economic growìh. Populaìion, invesìmenì and
TIP could be lisìed as deìerminanìs of growìh, and growìh accounìing was
born. This proved ìo be an enormously illuminaìing way ìo summarise a
vasì body of lnowledge and begin ìhe process of explaining economic
growìh (Solow and Temin 1ç;8, Criliches 1çç6). Buì Solow’s growìh
model did noì include any oìher variables, iì could noì accounì for ìhe wide
differences beìween counìries ìhaì we observe, and iì predicìed ìhaì all
counìries would converge ìo ìhe same raìe of growìh. This limiìaìion led
people ìo lump all oìher differences beìween counìries inìo TIP and pro-
vide explanaìions ouìside ìhe ìheory why ìhey differed (Denison 1ç6;).
The limiìaìions of ìhe Solow growìh model were aììacled in ìurn, giving
rise ìo new growìh ìheory. Romer (1ç86) argued ìhaì TIP growìh was
endogenous, noì exogenous. Lucas (1ç88) inìroduced human capiìal ìo ìhe
model as an addiìional deìerminanì of growìh, as had been done informally
in growìh accounìing and in economic hisìory (Denison 1ç6;, Easìerlin
1ç81). Differences in educaìion beìween counìries eliminaìed ìhe predicìion
of uncondiìional convergence (ìhaì is, convergence ìo ìhe same raìe of
growìh by all counìries), alìhough ìhey sìill lefì room for condiìional con-
vergence for groups of similar counìries, someìimes called ‘convergence
clubs’. Wide differences beìween counìries now could be explained wiìhin
ìhe model by differences in educaìional aììainmenì (Manliw, et a/. 1ççz).
New growìh ìheories provided exìensions ìo geì around ìhe limiìaìions of
old growìh ìheory aì ìhe expense of Solow’s simpliciìy and elegance, edu-
caìion is only ìhe mosì prominenì of many puìaìive inpuìs ìo growìh.
Empirical invesìigaìions flowered in ìhe form of growìh equaìions, buì few
6 Eurcpean Reztev cf Eccncmtc Htstcry
of ìhese regressions aclnowledged anyìhing special abouì ìhe Colden Age
of Economic Crowìh. The regressions focused on idenìifying ìhe equilib-
rium growìh raìe ìo which counìries were converging raìher ìhan esìimaìing
condiìional convergence iìself. The laììer by ìhe 1çços was simply assumed
as a facì of economic life.
Baumol (1ç86) provided evidence of convergence over a cenìury for a
sample of mosìly Wesìern European counìries. The claim ìhaì ìhis was a
universal paììern did noì sìand up (De Long 1ç88), and ìhe field ìurned ìo
a prolonged invesìigaìion of ìhe facìors ìhaì deìermine ìo whaì raìe of
growìh counìries will converge. Crowìh regressions ìypically are done for as
many counìries as possible, which means over a hundred in ìoday’s world.
The ìime period chosen is much shorìer ìhan Baumol’s in order ìo exploiì
ìhe plenìiful daìa afìer World War II. Two recenì surveys of ìhis liìeraìure
describe ìhe diversiìy of approaches ìalen ìo idenìify ‘convergence clubs’,
buì ìhey do noì remarl on any special ìreaìmenì of ìhe Colden Age of
Economic Crowìh (Durlauf and Quah 1çç8, Temple 1ççç).
Barro’s Robbins Lecìures, for example, were based on regressions for
periods sìreìching from 1ç6¸ ìo 1çço wiìh no aclnowledgmenì ìhaì ìhe
process of growìh mighì be differenì aì ìhe beginning and end of ìhe period.
He commenìed ìhaì ìhis was an improvemenì on his prior pracìice of using
a single cross-secìion, buì noì because ìhe daìa came from ìwo separaìe
economic periods (Barro 1çç;, pp. 1z–1¸). The common pracìice sìill is ìo
lump ìhe posìwar period inìo one cross-secìion, as done in Young’s famous
dissecìion of economic growìh among ìhe Asian NICs and ]ones’ survey of
ìhe world income disìribuìion (Young 1çç(, 1çç¸, ]ones 1çç;). The period
ìypically sìarìs in 1ç6o – Barro sìarìed laìer so he could use 1ç6o CDP as
an insìrumenì – boìh ìo exploiì easily available daìa and ìo avoid ìhe recov-
ery period jusì afìer ìhe war. Dowricl and Nguyen (1ç8ç) provide a soliìary
excepìion ìo ìhis rule. They examine wheìher ìhe convergence found by
Baumol (1ç86) conìinued afìer 1ç;¸, ìesìing earlier informal resulìs wiìh
growìh equaìions. The focus was on convergence raìher ìhan ìhe raìe of
growìh.
Economic hisìorians also have ìurned ìheir aììenìion ìo ìhe posìwar years.
Crafìs and Toniolo opened a volume of essays on ìhe period by asserìing,
‘ìhe years 1ç¸o–;¸ wiìnessed a unique episode in ìhe hisìory of European
“modern economic growìh” ’ (Crafìs and Toniolo 1çç6, p. ¸). They argued
ìhaì rapid growìh in ìhis period was parìly a consequence of slow growìh in
ìhe previous period, buì ìhey did noì dwell on ìhe mechanism of such a
reacìion. The bool as a whole is a survey of ìhe experience of abouì a dozen
Wesìern European counìries during ìhe Colden Age in a compaìible
formaì. This exercise of fiììing diverse hisìories inìo a common mould, how-
ever, was overwhelmed by ìhe sìrengìhs of parìicular issues in ìhe debaìes
abouì individual counìries, and ìhe essays are quiìe diverse (Temin 1çç;).
One counìry sìudy ìhaì anìicipaìed ìhe approach here concepìualised ìhe
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered ;
Cerman 1trtschaftsvunder as a disequilibrium phenomenon. Dumle (1çço)
argued ìhaì greaìer warìime desìrucìion generaìed fasìer posìwar growìh
and provided evidence for ìhis proposiìion in growìh regressions for OECD
counìries. His inquiry was in ìhe spiriì of Abramoviìz (1ç86), who asserìed
ìhaì ìhe desìrucìion of physical capiìal during ìhe war was less imporìanì
ìhan ìhe mainìenance of whaì he called ‘ìhe social capabiliìy’ of growìh. I
ìale my sìarì from Dumle, buì shifì his emphasis and his sample.
Eichengreen (1çç6) offered a synìheìic view in his conìribuìion ìo ìhe
Crafìs and Toniolo volume. Sìarìing from ìhe observaìion ìhaì growìh in
ìhe Colden Age was relaìed ìo caìching up and high invesìmenì, he asled
why invesìmenì was boìh high and producìive in ìhe Colden Age. He
answered ìhaì wage moderaìion and exporì growìh made invesìmenì aììrac-
ìive and profiìable. These in ìurn were due ìo governmenì insìiìuìions and
policies ìhaì were sharply differenì from ìhose pursued before ìhe war.
Eichengreen saw an impliciì bargain beìween worlers and invesìors ìhaì is
similar ìo ìhe impliciì conìracìs Aoli (1ç88) described in whaì he called ìhe
]-firm, ìypical of posìwar ]apan. The bargain was ìhaì worlers would noì
push for higher wages if invesìors would male producìive invesìmenìs ìhaì
would, over ìime, creaìe jobs and raise wages. Invesìors would agree ìo
invesì on ìhe condiìion ìhaì ìhe worlers did noì immediaìely ìry ìo ìale all
ìhe gains in higher wages.
This bargain is ìime-inconsisìenì. If worlers moderaìe wage demands,
invesìors have an incenìive ìo pay ìhemselves ìhe resulìing profiìs insìead of
reinvesìing ìhem. And if invesìors male producìive invesìmenìs ìhaì
enhance ìhe producìiviìy of labour, worlers have ìhe incenìive ìo ìale ìhe
gains home in ìhe form of higher wages. These perverse incenìives were
counìered in posìwar Europe by a complex seì of insìiìuìions ìhaì made
reneging harder and increased incenìives for honouring ìhe long-ìerm
impliciì conìracì in ìhe face of shorì-run gains from abrogaìing ìhe conìracì.
The insìiìuìions were boìh domesìic and inìernaìional, domesìic ìo enforce
ìhe bargain jusì described, inìernaìional ìo promoìe naìional specializaìion
ìhaì increased efficiency. The domesìic insìiìuìions included naìional wage
bargaining, union represenìaìion on company boards, and condiìional
access ìo governmenì programmes. The inìernaìional ones included insìi-
ìuìions lile CATT, ECSC and EPL ìhaì appeared ìo have had liììle posi-
ìive effecì. Eichengreen emphasised ìheir role in precluding negaìive effecìs,
assuring ìhaì ìrade would remain free as condiìions changed.
This is an inìriguing and plausible hypoìhesis, iì explains how demand
could grow ìo promoìe rapid economic growìh during ìhe Colden Age. Buì
iì cannoì explain how Wesìern Europe found iìself so far from equilibrium
aì ìhe sìarì of ìhe Colden Age. This organizaìional view also does noì dis-
ìinguish beìween differenì counìries in Wesìern Europe because ìhe inìer-
naìional agreemenìs ìhaì formsuch a large parì of ìhe sìory include ìhemall.
Eichengreen lisìed many causes for ìhe end of ìhe Colden Age, revealing ìhe
8 Eurcpean Reztev cf Eccncmtc Htstcry
absence of a unified explanaìion. Among ìhe reasons offered were ìhe cap-
ìure of insìiìuìions by firms and unions, ìhe oil shocls, ìhe end of ìhe
Breììon Woods Sysìem, ìhe end of general caìch-up, and reduced incenìives
ìo leep ìhe bargains ìhaì produced ìhe Colden Age of Economic Crowìh.
z. Recent theories oI econonic growth
I approach ìhe Colden Age in ìhe conìexì of economic growìh over ìhe pasì
cenìury or ìwo, which had a large componenì of economic ìransiìion.
Naìional economies around 18oo wiìh very few excepìions were almosì
compleìely agriculìural. Sìarìing in ìhe nineìeenìh cenìury and even laìer,
producìive resources were moved ouì of agriculìure inìo manufacìuring and
services. Residenìs became urban, and ìhe share of ìhe labour force in agri-
culìure fell. Since worlers were more producìive in non-agriculìural acìivi-
ìies, naìional income grew during ìhis ìransiìion.
Theorisìs of economic growìh recenìly have begun ìo aclnowledge ìhe
imporìance of ìhis ìransiìion in ìhe process of economic growìh. There are
now several models aììempìing ìo inìegraìe sìrucìural shifìs wiìh ìhe ìheory
of economic growìh (Kongsamuì, et a/. 1çç;, Temple and Voìh 1çç8, Calor
and Weil zooo). Taylor (1ççç) used a model of ìhis ìype in his exploraìion
of convergence in seven counìries before World War I. All of ìhese papers
share wiìh ìhis one ìhe aììempì ìo bring ìhe hisìorical experience of indus-
ìrialisaìion inìo ìhe mainsìream of ìhinling abouì economic growìh.
Broadberry (1çç;) evaluaìed ìhe imporìance of ìhis ìransiìion in
Cermany’s convergence ìo Briìish levels of labour producìiviìy. The firsì
column of Table ¸ shows his esìimaìe of aggregaìe labour producìiviìy in
Cermany compared ìo ìhe Lniìed Kingdom. The familiar rise over ìhe lasì
cenìury can be seen, wiìh a dip in 1ç¸o – jusì afìer ìhe Second World War.
The second column of Table ¸ reveals ìhaì ìhe rise in comparaìive labour
producìiviìy in manufacìuring did noì echo ìhe rise in ìhe aggregaìe. In facì,
ìhere is very liììle evidence of a ìrend aì all. Cerman relaìive labour pro-
ducìiviìy was as high in 1ç1¸ as iì would geì, and ìhe ìemporary decline in
1ç¸o was eliminaìed by 1ç6o when iì sìood aì 11¸ (Broadberry 1çç;, p. z¸1).
Caìch-up, Broadberry asserìs, is noì ìhe resulì of improving efficiency in
manufacìuring, buì ìhe resulì of ìransferring resources from low-pro-
ducìiviìy secìors lile agriculìure ìo high-producìiviìy ones lile manufacìur-
ing.
Iì follows ìhaì fasìer economic growìh in Cermany ìhan in Briìain was
due largely ìo ìhe more rapid secìoral shifìs in ìhe Cerman economy.
Cermany had a larger share of iìs labour force in agriculìure ìhan ìhe Lniìed
Kingdom ìhroughouì ìhe pasì cenìury. In 1ç¸o, around ìhe sìarì of ìhe
Colden Age of European Crowìh, Cermany had z( per cenì of iìs labour
force in agriculìure, compared ìo five per cenì for ìhe Lniìed Kingdom
(Broadberry 1çç;, p. z¸z). If rapid economic growìh is ìhe resulì of ìhe
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered ç
ìransiìion from an agrarian economy, ìhen Cermany was sìill engaged in ìhe
process during ìhe Colden Age while Briìain had compleìed iìs ìransiìion.
Why was Cermany lagging behind Briìain in ìhis ìransiìion? Three
reasons come ìo mind, of which ìhe ìhird has noì been appreciaìed. Iirsì,
Cermany sìarìed iìs indusìrialisaìion afìer Briìain. Second, Cermany chose
ìo proìecì iìs farmers againsì low-priced American grain in ìhe laìe nine-
ìeenìh cenìury. Third, ìhe Second Thirìy Years War – ìhe ìurbulenì period
from 1ç1¸ ìo 1ç(¸ – inìerrupìed inìernaìional ìrade and slowed ìhe ìran-
siìion. The ìhird of ìhese reasons has been neglecìed, I wanì ìo expose iìs
imporìance.
The growing liìeraìure on globalisaìion argues ìhaì iì has ebbed and
flowed in ìhe course of ìhe ìwenìieìh cenìury. Before ìhe Creaì War, inìer-
naìional commerce and ìravel were free and open, more or less as ìhey are
ìoday. Buì in beìween ìhese ìwo end poinìs, ìhe flow of goods, finance, and
people was inìerrupìed by world wars and depression. Auìhors disagree
among ìhemselves abouì wheìher ìoday’s globalisaìion acìually exisìed a
cenìury ago, buì ìhere is no disagreemenì abouì ìhe inìerrupìion during ìhe
world wars and Creaì Depression (Bordo et a/. 1ççç, Obsìfeld and Taylor
1ççç, Temin 1ççç).
Inìernaìional ìrade was inìerrupìed by ìhe Iirsì World War. The posìwar
seììlemenì creaìed many new boundaries ìhaì provided ìhe opporìuniìy ìo
impose ìariffs on ìrade. And ìhe Creaì Depression led ìo resìricìive ìrade
policies ìhaì reversed whaìever expansion had ìalen place in ìhe 1çzos. The
volume of exporìs for ìhe major Wesìern European counìries was lower in
1ç¸8 ìhan iì had been in 1ç1¸, in sharp conìrasì ìo iìs rapid growìh boìh
before and afìer ìhis period (Ieinsìein, et a/. 1çç;, p. 1o).
Sachs and Warner (1çç¸) argued ìhaì ìrade promoìed economic growìh
in ìhe posìwar world. Their regressions showed ìhaì closed economies did
noì exhibiì convergence, while open economies did. Why did closed
economies suffer? Because ìhey did noì underìale ìhe reallocaìion of
resources needed ìo increase producìiviìy. They could noì exploiì ìheir
comparaìive advanìages, and ìhey could noì end ìheir reliance on domesìic
1o Eurcpean Reztev cf Eccncmtc Htstcry
Table ¸. Ccmparattze /abcur prcducttztty tn Cermany and the Untted
Ktngdcm ;UK ϭ Ioo).
Year CDP Manufacìuring
18;o 6o ç¸
1ç1¸ ;8 11ç
1ç¸o 66 ç6
1ç;¸ 11z 11ç
1ç8ç 116 1o¸
Scurce: Broadberry 1çç;, p. z¸o.
agriculìure. Open economies decreased ìhe proporìion of food and raw
maìerials in ìheir exporìs more rapidly ìhan closed economies. Sachs and
Warner did noì dwell on ìhe connecìion beìween ìheir ìheory and ìhe his-
ìory of indusìrialisaìion in Europe during a previous period, buì ìhe paral-
lel is clear. Before World War I, parìicipaìion in inìernaìional ìrade
promoìed indusìrialisaìion. One has only ìo recall ìhe discussion of Briìain’s
‘climacìeric’ in ìhe laìe nineìeenìh cenìury ìo see ìhe imporìance of inìer-
naìional ìrade in economic growìh. Briìain was surpassed, a prominenì
sìory asserìs, because ìhe Lniìed Sìaìes and Cermany were beììer able ìo
exploiì world marleìs (Temin 1ç66).
Iì follows from ìhis view ìhaì ìhe barriers ìo inìernaìional commerce
during ìhe world wars and Creaì Depression consìiìuìed barriers ìo ìhe con-
ìinued indusìrialisaìion of European counìries. This slowdown in ìhe
process of indusìrialisaìion creaìed a disequilibrium afìer ìhe war. As
suggesìed by Table 1, ìhe supply fronìier conìinued ìo expand during ìhe
Second Thirìy Years War. The Lniìed Sìaìes, insulaìed from ìhe wars if noì
ìhe Depression, was able ìo conìinue iìs ìransformaìion from an agriculìural
ìo an indusìrial economy. Iìs exporìs were primarily food and raw maìerials
before ìhis proìracìed conflicì, ìhey were manufacìures afìerwards (Irwin
1çç6). European counìries emerged from ìhe war wiìh a developmenìal
deficiì.
This disequilibrium is separaìe from ìhe low income ìhaì generaìes con-
diìional convergence. Low income in ìhe sìandard sìory is produced by low
levels of physical and human capiìal relaìive ìo saving raìes. The develop-
menìal deficiì highlighìed here is produced by a misallocaìion of resources.
The firsì ìales place in a single-secìor economy, ìhe second, in a disaggre-
gaìed model of developmenì.
The raìe aì which worlers lefì agriculìure acceleraìed afìer ìhe war. The
decline in ìhe share of ìhe labour force in agriculìure was ìwice as rapid in
ìhe 1ç¸os and 1ç6os as before. The variance of ìhe measured change fell as
ìhe raìe increased, wheìher because of ìhe greaìer sìabiliìy of Wesìern
Europe or because of noise in ìhe imperfecì earlier daìa. The sìandard devi-
aìion of ìhe decadal raìe of change in ìhe share before World War II was
ìhree ìimes as large as ìhe sìandard deviaìion of ìhe quinquennial change
ìhereafìer. As a resulì, changes from before World War I ìo ìhe inìerwar
period are losì in ìhis volaìiliìy (Bairoch 1ç68, as quoìed in Miìchell 1çç8,
pp. 1(¸–6z).
The misallocaìion of resources can be measured by ìhe share of ìhe
labour force in agriculìure. There are many ways ìo divide up ìhe economy,
buì ìhe division beìween agriculìure and all oìher acìiviìies appears ìo be ìhe
mosì imporìanì. Broadberry (1çç;) disìinguished nine secìors of ìhe econ-
omy, buì he concluded ìhaì mosì of ìhe effecì came from ìhe changing size
of agriculìure. Denison (1ç6;), much earlier, ìalled of ìhe misallocaìion of
resources in Europe during ìhe Colden Age of European Crowìh, and he
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered 11
ìoo meanì ìhe European counìries were growing rapidly when ìhey were
geììing ouì of agriculìure.
The misallocaìion, according ìo ìhis sìory, came from a generaìion –
ìhirìy years – of economic insulariìy. Iì is reasonable ìo ìhinl ìhaì ìhe
excessive resources in agriculìure could be moved ìo oìher secìors in
anoìher ìhirìy years. This hypoìhesis ìherefore provides a way ìo raìionalise
ìhe hisìory shown in Table 1. Warìime desìrucìion was lile a business cycle
in leading ìo a shorì-ìerm disequilibrium. Condiìional convergence mighì
explain long-ìerm disequilibria. Auìarchy in ìhe Second Thirìy Years War
can explain a disequilibrium ìhaì can be eliminaìed in ìwenìy or ìhirìy years.
This phenomenon may be more general ìhan Europe afìer ìhe Second
World War. ]ones (1çç;) concepìualised growìh as a lind of Marlov
process. Counìries drew ìheir raìe of growìh from an urn once every ìhirìy
years or so, drawing fasì, slow or medium growìh raìes. ]ones characìerised
ìhe fasì growìh as growìh miracles and asserìed ìhaì ìhey were mosì preva-
lenì among poorer counìries, alìhough noì among ìhe pooresì. Young
(1çç() showed ìhaì ìhese growìh miracles were accomplished by very high
invesìmenì raìes. They also were accomplished by rapid reducìions in ìhe
size of agriculìure in ìhese counìries.
¸. Testing the hypothesis
I ìesì ìhis hypoìhesis by formalising ìhe sìory in a simple model and ìesìing
iì againsì daìa from ìhe Colden Age of European Crowìh. The model dis-
ìinguishes ìhree linds of disequilibria ìhaì can affecì growìh:
(1) Condiìional convergence, ìhaì is, sìarìing from a level of income
low relaìive ìo ìhe counìry’s equilibrium income.
(z) Warìime desìrucìion ìhaì deranges producìion in ìhe shorì run.
Dumle (1çço) measured ìhe exìenì of ìhis dislocaìion by ìhe
percenìage gap beìween per captta CDP in 1ç(8 and in 1ç¸8. I
use ìhis measure here as well, labelling iì CAP, and recalculaì-
ing iì from Maddison (1çç¸).
(¸) Arresìed developmenì, ìhaì is, excessive labour in agriculìure. In
parallel wiìh condiìional convergence, ìhis phenomenon will be
measured by ìhe difference beìween ìhe iniìial proporìion of ìhe
labour force in agriculìure, A, and ìhe equilibrium share, A*.
The model ìhen is as follows, where g is ìhe average growìh raìe of y, per
captta CDP.
g ϭ a ϩ b(y* Ϫ y) ϩ c CAP ϩ d (A Ϫ A*) ϩ e (1)
This regression, despiìe iìs convenìional appearance, differs from growìh
regressions in ìhe liìeraìure. Those growìh regressions are designed ìo eliciì
differences beìween y* in differenì counìries. Crowìh is regressed on cur-
1z Eurcpean Reztev cf Eccncmtc Htstcry
renì income and many variables, lile educaìion, ìhaì proxy for and idenìify
y*. I assume here ìhaì y* is ìhe same for all counìries in Wesìern Europe.
This emphasis is appropriaìe in a sìudy of a single region and in ìhe inves-
ìigaìion of disequilibrium growìh during ìhe Colden Age of Economic
Crowìh. I am ìrying ìo describe ìhe process of convergence, while growìh
regressions ìypically assume ìhaì counìries are near ìheir growìh paìh and
invesìigaìe ìhe naìure of ìhe equilibrium income (y*) ìo which ìhey are con-
verging.
I also assume ìhaì ìhe equilibrium share of agriculìure, A*, is ìhe same
for all Wesìern European counìries. The influences of geography, hisìory,
and ìhe Common Agriculìural Policy are ìalen ìo be second-order effecìs.
One could noì male ìhis heroic assumpìion wiìh a wider sample, buì iì is
appropriaìe when discussing economic growìh in Wesìern Europe. The
share of ìhe labour force in agriculìure is measured aì ìhe beginning of each
period, so ìhaì iì is a predeìermined variable. Differences beìween counìries
will show up in ìhe error ìerm and in ìhe goodness of fiì.
Equaìion (1) can be rewriììen, collecìing ìhe unobserved equilibrium
levels wiìh ìhe consìanì ìerm.
g ϭ (a ϩ by* Ϫ dA*) Ϫ by ϩ c CAP ϩ dA ϩ e (z)
I esìimaìe ìhis equaìion for all 1¸ Wesìern European counìries afìer ìhe
Second World War. (This is ìhe same seì of counìries whose growìh is
reporìed in Table 1, excepì ìhaì Czechoslovalia has been replaced by
Porìugal.) They all are parì of ìhe same ‘convergence club’, harling bacl ìo
ìhe origins of new growìh ìheory. They all have sìable governmenìs, secure
properìy righìs, and universal educaìion, and iì is reasonable ìo argue ìhaì
A* is vanishingly small in Wesìern Europe ìoday.
I use ìhis regression ìo ìesì ìhree hypoìheses. Iirsì, all ìhree linds of dis-
equilibrium affecìed ìhe raìe of growìh in differenì counìries in ìhe Colden
Age of European Crowìh. The presence of ìhe second ìwo disequilibria
accounì for ìhe rapidiìy of economic growìh during ìhe Colden Age.
Second, ìhese disequilibria had ìheir main impacì aì differenì periods.
Warìime desìrucìion affecìed growìh in ìhe immediaìe posìwar years, ìhen
resource misallocaìion became imporìanì, and finally condiìional conver-
gence became cenìral. In normal periods of peace, only ìhe lasì of ìhese is
relevanì. The special, hisìorical qualiìy of ìhe Colden Age comes from ìhe
presence of ìhe firsì ìwo. The susìained rapid growìh comes from ìhe pres-
ence of ìhe second disequilibrium, resource misallocaìion.
Third, warìime desìrucìion and ìhe misallocaìion of resources ceased ìo
have an effecì on growìh some ìime around 1ç;o. In oìher words, ìhe end
of ìhe Colden Age came when ìhese unusual disequilibria were removed.
We cannoì observe ìhese effecìs aì oìher ìimes because ìhey are noì ìhere,
ìhey were presenì in posìwar Europe only because of ìhe hisìorical circum-
sìances ìhaì were unique ìo ìhis ìime and place.
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered 1¸
The regressions are shown in Tables ( and ¸. The firsì of ìhese ìables
shows regressions for ìwenìy-year inìervals which correspond closely ìo ìhe
Colden Age, ìhe second, for ìen-year inìervals, following Barro (1çç;). The
daìa sìarì in 1ç¸o because daìa for 1ç(¸ are unavailable and unreliable. The
resulìs in Table ( are useful for looling aì ìhe Colden Age as a whole, ìhose
in Table ¸, for ìracling ìhe effecì of differenì disequilibria wiìhin ìhe Colden
Age.
The firsì ìhing ìo noìe is ìhaì ìhe regressions accounì for a subsìanìial
amounì of ìhe variaìion in growìh raìes among ìhe Wesìern European coun-
1( Eurcpean Reztev cf Eccncmtc Htstcry
Table (. Regresstcns exp/atntng tventy-year grcvth rates.
Variable 1ç¸o–;o 1縸–;¸ 1ç6¸–8¸
Y
o
Ϫo.oo¸61 Ϫo.oz¸8 Ϫo.ozz¸
(Ϫo.zç) (Ϫ1.ç¸) (Ϫz.oç)
CAP Ϫo.o1;z Ϫo.oozo( o.ooz¸z
(Ϫ1.8() (Ϫo.¸z) (o.¸8)
A o.o¸;¸ o.o¸8ç o.oo¸¸(
(z.;ç) (z.(() (o.z6)
Consìanì z.¸8 ¸.;6 (.o;
(z.o1) ((.o8) (¸.8()
Adjusìed R
z
o.;o o.;ç o.¸¸
N 1¸. 1¸. 1¸.
Scurces: Posìwar CDP per captta daìa from ìhe Penn World Table ¸.6, CAP ϭ per captta
CDP in 1ç(8 over per capiìa CDP in 1ç¸8, from Maddison (1çç¸), ‘A’ from Inìernaìional
Labor Office (1ç86), IAOSTAT (on ìhe web), and Social Indicaìors of Developmenì (on
ìhe web). T-sìaìisìics are below ìhe coefficienìs.
Table ¸. Regresstcns exp/atntng decada/ grcvth rates.
Variable 1ç¸o–6o 1縸–6¸ 1ç6¸–;¸ 1ç;¸–8¸ 1ç8¸–1çç¸
Y
o
o.o¸z; o.o1;1 Ϫo.o(¸ç Ϫo.oo1¸( Ϫo.o¸1¸
(1.oz) (o.8o) (Ϫ¸.61) (Ϫo.o8) (Ϫz.¸o)
CAP Ϫo.o(çç Ϫo.o16¸ o.oo6(( Ϫo.o111 Ϫo.oo(;ç
(Ϫ¸.zo) (Ϫ1.(8) (o.8¸) (Ϫo.8ç) (Ϫo.(6)
A o.o8;z o.o;1¸ o.o1¸6 Ϫo.oo66z o.o18o
(z.¸¸) (z.¸() (o.6() (Ϫo.11) (o.¸1)
Consìanì o.16z 1.¸; 6.oz z.zz ¦¸.(¸
(o.o8) (o.8() ((.ç;) (o.8ç) (z.6;)
Adjusìed R
z
o.¸6 o.(z o.8o Ϫo.1; o.¸8
N 1¸. 1¸. 1¸. 1¸. 1¸.
Scurces: See Table (. Preliminary daìa for 1çç¸ from Roberì Summers by privaìe correspon-
dence, ; March 1ççç. T-sìaìisìics are in parenìheses below ìhe coefficienìs.
ìries during ìhe Colden Age, despiìe ìhe drasìic simplificaìions of ìhe
model. The assumpìion ìhaì y* and A* are ìhe same for all counìries in
Wesìern Europe afìer ìhe Second World War does noì appear misleading.
The second ìhing ìo noìe is ìhaì inferences from ìhe sìandard errors are
slighìly problemaìical in ìhis conìexì, as for many growìh regressions. These
counìries are noì a random sample of a larger seì, ìhey are all ìhe counìries
of Wesìern Europe. The years also are noì a random sample from a larger
seì, ìhey are ìhe years of ìhe Colden Age of European Crowìh. An inference
based on a t-ìesì would be wrong in only ¸ per cenì of cases if we could find
similar seìs of counìries afìer similarly long periods of war and economic
auìarly ìo sìudy.
The firsì ìwo columns of Table ( show regressions for ìhe Colden Age.
The effecì of excessive labour in agriculìure (A) is visible clearly. This indi-
caìes ìhaì ìhe misallocaìion of resources arising from ìhe limiìaìion of inìer-
naìional ìrade during ìhe Second Thirìy Years War was cosìly ìo Wesìern
Europe. Iì also creaìed ìhe opporìuniìy for many counìries ìo grow rapidly
as ìhey allocaìed ìheir resources more efficienìly. The effecì was large. The
share of Cermany’s labour force in agriculìure aì ìhe sìarì of ìhe Colden Age
was zo percenìage poinìs higher ìhan ìhe share of Briìain’s agriculìural
labour force. Cermany’s growìh raìe received a boosì of approximaìely one
percenìage poinì from ìhis iniìial condiìion. (The effecì is 1.z per cenì if
1ç¸o–;o is used, o.8 per cenì if 1縸–;¸ is used.) The acìual difference in
growìh raìes in 1縸–;¸ was 1.¸ per cenì as shown in Table z, ìhis effecì
explains over half ìhe difference for ìhese years.
Neiìher warìime desìrucìion nor long-run condiìional convergence were
as imporìanì as ìhe misallocaìion of resources in explaining differences
beìween growìh raìes in Wesìern Europe aì ìhis ìime. Dumle (1çço) found
warìime desìrucìion ìo be more imporìanì for ìwo reasons. Iirsì, he lumped
ìhe enìire Colden Age inìo one regression, and he did noì ìracl effecìs
wiìhin ìhe period. He also included ]apan in his sample, which – lile
Cermany – had exìensive warìime damage. The presence of ìwo heavily
damaged counìries was enough ìo male warìime desìrucìion appear
imporìanì.
The firsì column of Table ¸ conìains regression for ìhe firsì posìwar
decade for which we have reliable daìa. Warìime desìrucìion (CAP) was
imporìanì in deìermining ìhe raìe of growìh of differenì counìries in ìhis
period, as was ìhe misallocaìion of resources (A). Condiìional conver-
gence, by conìrasì, was noì an imporìanì facìor. Succeeding columns show
ìhaì ìhe influence of warìime desìrucìion faded rapidly inìo ìhe bacl-
ground. Iì was only an imporìanì facìor aì ìhe sìarì of ìhe Colden Age. The
effecì of ìhe agriculìural labour force faded away as well, buì more slowly.
As ìhe share of ìhe labour force in agriculìure approached iìs equilibrium
level, ìhe esìimaìed effecì became smaller, boìh absoluìely and relaìive ìo
iìs sìandard error. The decreasing size of ìhe coefficienì over ìime suggesìs
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered 1¸
a non-lineariìy, perhaps an ‘S-shaped’ relaìion beìween ìhe excess share of
labour in agriculìure and economic growìh.
The regression for 1ç;¸–8¸ is marledly worse ìhan ìhose for oìher
decades. No coefficienìs are esìimaìed ìighìly, and ìhe share of variance of
growìh raìes explained is negligible. This decade, of course, conìained ìhe
effecìs of ìwo oil shocls which disrupìed ìhe longer-run processes of inìeresì
here. The presence of ìhese disrupìions males iì hard ìo idenìify ìiming in
longer-run phenomena. In ìhe view I am presenìing here, ìhese disrupìions
confuse ìhe sìory, buì ìhey are noì ìhe sìory iìself. The end of ìhe Colden
Age was broughì abouì by ìhe reìurn ìo equilibrium growìh paìhs on ìhe
parì of ìhese counìries, noì by adverse economic shocls in ìhe 1ç;os. The
regression for 1ç8¸–ç¸ in ìhe final column of Table ¸ shows ìhaì condiìional
convergence as indicaìed by a significanì coefficienì of y
o
is verified, while
ìhe oìher forces presenì in ìhe Colden Age do noì show up.
Alìhough iì is hard ìo ìesì for robusìness wiìh only 1¸ observaìions, vari-
ous experimenìs show ìhese equaìions ìo be robusì. The resulìs differ only
in deìails if log(y
o
) is used in place of y
o
ìo indicaìe general convergence, and
inìermediaìe ìen and ìwenìy year ìime periods yield inìermediaìe resulìs ìo
ìhose shown here. Truncaìing CAP ìo be less ìhan zero also does noì
change ìhe resulìs. Lagging A by five years ìo avoid any hinì of simulìane-
iìy preserves ìhe resulìs in Tables ( and ¸. And ìhe resulìs are unchanged if
one or anoìher small counìry is dropped from ìhe sample, or even if ìwo
ouìliers lile Spain and Porìugal are dropped ìogeìher.
Scaììer diagrams illusìraìe ìhe regression resulìs. Iigures 1 and z show
ìhe relaìion beìween ìhe raìe of growìh and ìhe iniìial share of ìhe labour
force in agriculìure for ìwo overlapping zo-year periods sìarìing in ìhe
1ç¸os. In boìh diagrams, ìhe correlaìion shown in Table ( is quiìe apparenì.
While ìhe diagrams are similar, ìhere are a few differences worìh noìing.
Iigure 1, showing 1ç¸o–;o, reveals Cermany (DEL) ìo be an ouìlier. The
1trtschaftsvunder appears clearly. Buì in Iigure z, showing 1縸–;¸,
Cermany no longer is an ouìlier. Insìead, Belgium lools ìo be an economic
miracle. We need ìo be careful in our claims of special condiìions in indi-
vidual counìries.
Briìain is aì ìhe low end of European growìh raìes, as everyone lnows.
This model provides an explanaìion for ìhis slow growìh aì variance wiìh
ìhaì in ìhe liìeraìure (Bean and Crafìs 1çç6). Ior ìhe slow Briìish growìh is
explained quiìe fully by ìhe low share of ìhe labour force in agriculìure aì
ìhe sìarì of ìhe Colden Age. Briìain had sìarìed indusìrialisaìion wiìh a low
share of labour in agriculìure, iì indusìrialised firsì, and iì lepì ìariffs low
when American grain was able ìo be ìransporìed cheaply ìo Europe in ìhe
laìe nineìeenìh cenìury. These pasì accomplishmenìs implied slow growìh
during ìhe Colden Age.
Iì follows ìhaì posìwar Briìish economic policies were noì ìhe cause of
slow economic growìh. Iì may noì be ìoo farfeìched ìo say ìhaì poor poli-
16 Eurcpean Reztev cf Eccncmtc Htstcry
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered 1;
Iigure 1. Regresstcn cf eccncmtc grcvth rate and tnttta/ share cf /abcur
fcrce tn agrtcu/ture, Iµyo–;o.
Iigure z. Regresstcn cf eccncmtc grcvth rate and tnttta/ share cf /abcur
fcrce tn agrtcu/ture, Iµyy–;y.
cies were ìhe resulì of slow growìh raìher ìhan iìs cause. Bean and Crafìs
(1çç6) offer a mulìi-layered picìure of Briìish policies, in which Briìain’s
iniìial posiìion plays only a minor role. This view suggesìs ìhaì Briìish econ-
omic policies did noì have ìhe poìenìial effecìs aììribuìed ìo ìhem in ìhis
lind of accounì. And iì suggesìs even furìher ìhaì whaì we regard as poor
policies – because Briìain grew so slowly – were eiìher reasonable adap-
ìaìions ìo Briìain’s iniìial posiìion or ìhe resulìs of Briìain’s slow growìh.
The Colden Age of Economic Crowìh ended when ìhis disequilibrium
was eliminaìed, ìhaì is, when ìhe share of ìhe labour force in agriculìure
approached iìs equilibrium level. This can be seen in ìhe regressions in
Tables ( and ¸, where ìhe coefficienì of A is noì esìimaìed clearly in ìhe laìer
regressions. Iì can be seen also in ìhe raìe of change of ìhe share of labour
in agriculìure. If one examines ìhe change in ìhe labour force in agriculìure,
as in equaìion (1), ìhen ìhe raìe of change falls over ìime, ìhaì is, becomes
less negaìive, as shown in ìhe firsì row of Table 6. This is ìhe variable ìhaì
is relevanì for growìh, since iì is ìhe movemenì of people ìhaì generaìes
growìh. If, however, one examines ìhe raìe of change of ìhe agriculìure
share, ìhaì is, ìhe change in ìhe share divided by ìhe iniìial share, ìhen ìhere
is no change over ìime. The share approaches iìs asympìoìe aì a consìanì
raìe.
The regressions in Table 6 help us ìo undersìand why ìhis is a hisìorical
explanaìion. In peaceful ìimes, ìhe share of labour in agriculìure falls as
income rises. The change in ìhe agriculìural labour force geìs smaller (more
posiìive) over ìime, buì ìhere is no evidence of a change in ìhe proporìion-
aìe raìe of decline. The evidence is consisìenì wiìh an asympìoìic approach
ìo an equilibrium share. Various facìors and policies can inhibiì ìhis change
in individual counìries, buì ìhere are no oìher hisìorical cases where large
numbers of indusìrial counìries were in ìhe same disequilibrium posiìion.
The model of equaìion (1) may be ìhe correcì model for all ìime, buì CAP
18 Eurcpean Reztev cf Eccncmtc Htstcry
Table 6. Trends cf the agrtcu/tura/ /abcur share.
Variable Change in ìhe share Crowìh in ìhe share
(1) (z) (¸)
Trend o.(z6 o.6¸¸ o.ooz(;
(1z.() ((.o8) (1.¸¸)
Trend squared Ϫo.oz1
(Ϫ1.¸8)
Consìanì Ϫ(.;¸ Ϫ¸.1¸ Ϫo.1;
(Ϫ1z.ç8) (Ϫ11.1;) (Ϫ16.oo)
Overall R
z
o.¸¸ o.¸( o.o1
N 1¸¸ 1¸¸ 1¸¸
Scurce: See Table (. Z-sìaìisìics in below ìhe coefficienìs.
and (A-A*) ìypically are aì or near zero. Iì is impossible ìo esìimaìe ìheir
coefficienìs under ìhese circumsìances. The final regressions in Tables (
and ¸ illusìraìe ìhe inabiliìy ìo esìimaìe ìhese coefficienìs ouìside ìhe
Colden Age.
q. Conclusion
I reìurn now ìo ìhe ìhree quesìions posed aì ìhe beginning of ìhis arìicle.
Iirsì, why was growìh so rapid in Wesìern Europe during ìhe Colden Age
of Crowìh? The answer is disequilibrium. The normal caìch-up ìhaì worls
in general was noì imporìanì righì afìer ìhe war, buì oìher linds of disequi-
librium were. The mosì imporìanì of ìhese was ìhe misallocaìion of
resources ìhaì came from ìhe lacl of inìernaìional ìrade during ìhe preced-
ing ìhirìy years. In ìhis sìaìe of arresìed indusìrialisaìion, ìoo many
resources sìill were employed inefficienìly in agriculìure. The insìiìuìional
facìors ciìed by Eichengreen (1çç6) helped creaìe ìhe needed demand, real-
locaìed labour rapidly enhanced ìhe supply.
Second, why did differenì counìries grow aì differenì raìes during ìhe
Colden Age? They grew aì differenì raìes during ìhe Colden Age because
of ìheir iniìial posiìion. Naìional policies had secondary effecìs relaìive ìo
counìries’ iniìial posiìion. Labour relaìions were ìumulìuous in boìh Iìaly
and Briìain, buì Iìaly grew rapidly while Briìain did noì. Aì ìhe leasì,
policies of ìhe ìime have been blamed for evenìs ouìside ìhe conìrol of
policymalers. In addiìion, some of ìhe biììerness of policy in slowly-
growing Briìain may have been ìhe resulì of slow growìh insìead of iìs
cause.
Third, why did ìhe rapid growìh come ìo an end? Crowìh slowed in ìhe
1ç;os and 1ç8os because ìhe disequilibrium ìhaì had generaìed unusually
rapid growìh no longer exisìed. The developmenìal deficiì of a generaìion
was eliminaìed in a generaìion. The Common Agriculìural Policy may have
lepì some excessive labour in agriculìure, buì ìhe misallocaìion of resources
had ceased ìo be a large macroeconomic issue by ìhe ìime of ìhe oil crises.
These crises muddied ìhe hisìorical waìers, confusing shorì-run and long-
run facìors. Iì is only wiìh ìhe hindsighì of anoìher ìhirìy years ìhaì we can
see ìhaì while ìhe oil crises were disrupìive, ìhe slowdown of growìh would
have ìalen place even if ìhey had noì occurred.
Acknowledgenents
This arìicle is a revised version of ìhe Hicls Lecìure, Oxford Lniversiìy, 1z March,
1ççç. I ìhanl ]ohnny Chen, Bural Cuner, Dirl Niepelì, and Ayalo Tanala for
research assisìance, and audiences aì ìhe lecìure and oìher seminars for helpful
commenìs. Zvi Criliches was lind enough and well enough ìo advise me in ìhe early
sìages of ìhis arìicle. All errors are mine alone.
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered 1ç
ReIerences
ABRAMOVITZ, M. (1ç86). Caìching up, forging ahead, and falling behind. )curna/
cf Eccncmtc Htstcry q6, pp. ¸8¸–(o6.
AOKI, M. (1ç88). Infcrmattcn, Incenttzes, and Eargatntng tn the )apanese Eccncmy.
Cambridge: Cambridge Lniversiìy Press.
BAIROCH, P. et a/. (1ç68). The 1cr/tng Pcpu/attcn and Its Structure. Brussels:
Insìiìuì de Sociologie, Lniversiìe Libre de Bruxelles.
BARRO, R. ]. (1çç;). Determtnants cf Eccncmtc Crcvth. Cambridge, MA: MIT
Press.
BALMOL, W. (1ç86). Producìiviìy growìh, convergence and welfare. ¬mertcan
Eccncmtc Reztev y6, pp. 1o;z–8¸.
BEAN, C. and CRAITS, N. (1çç6). Briìish economic growìh since 1ç(¸. In
Nicholas Crafìs and Cianni Toniolo (eds), Eccncmtc Crcvth tn Eurcpe Stnce
Iµ¡y. Cambridge: Cambridge Lniversiìy Press.
BERNANKE, B. (1çç¸). The macroeconomics of ìhe Creaì Depression: a
comparaìive approach. )curna/ cf Mcney, Credtt and Ean/tng zy, pp. 1–z8.
BORDO, M. D., EICHENCREEN, B. and IRWIN, D. A. (1ççç). Is globalizaìion ìoday
really differenì ìhan globalizaìion a hundred years ago? Ercc/tngs Trade Icrum.
Washingìon DC: The Broolings Insìiìuìion, pp. 1–¸o.
BROADBERRY, S. N. (1çç;). Anglo-Cerman producìiviìy differences, 18;o–1çço.
Eurcpean Reztev cf Eccncmtc Htstcry, t: pp. z(;–6;.
BRLNO, M. and SACHS, ]. (1ç8¸). Eccncmtcs cf 1cr/dvtde Stagf/attcn. Cambridge,
MA: Harvard Lniversiìy Press.
CRAITS, N. and MILLS, T. C. (1çç6). Europe’s Colden Age: an economeìric
invesìigaìion of changing ìrend raìes of growìh. In Barì van Arl and Nicholas
Crafìs (eds), Quantttattze ¬spects cf Pcst-1ar Eurcpean Eccncmtc Crcvth.
Cambridge: Cambridge Lniversiìy Press.
CRAITS, N. and TONIOLO, C. (eds), (1çç6). Eccncmtc Crcvth tn Eurcpe stnce Iµ¡y.
Cambridge: Cambridge Lniversiìy Press.
DE LONC, ]. B. (1ç88). Producìiviìy growìh, convergence, and welfare: commenì.
¬mertcan Eccncmtc Reztev y8, pp. 11¸8–¸(.
DENISON, E. (1ç6;). 1hy Crcvth Rates Dtffer. Washingìon, DC: Broolings
Insìiìuìion.
DOWRICK, S. and NCLYEN, D.-T. (1ç8ç). OECD comparaìive economic growìh
1ç¸o–8¸: caìch-up and convergence. ¬mertcan Eccncmtc Reztev yµ,
pp. 1o1o–¸o.
DLMKE, R. (1çço). Reassessing ìhe 1trtschaftsvunder: reconsìrucìion and posìwar
growìh in Wesì Cermany in an inìernaìional conìexì. Oxfcrd Eu//ettn cf
Eccncmtcs and Stattsttcs ¶z, pp. (¸1–ç1.
DLRLALI, S. N. and QLAH, D. T. (1çç8). The new empirics of economic growìh.
Worling Paper no. 6(zz, Naìional Bureau of Economic Research.
EASTERLIN, R. A. (1ç81). Why isn’ì ìhe whole world developed? )curna/ cf
Eccncmtc Htstcry qt, pp. 1–1ç.
EICHENCREEN, B. (1çç6). Insìiìuìions and Economic Crowìh in Europe Afìer
World War II. In Nicholas Crafìs and Cianni Toniolo (eds), Eccncmtc Crcvth
tn Eurcpe Stnce Iµ¡y. Cambridge: Cambridge Lniversiìy Press.
IEINSTEIN, C. H., TEMIN, P. and TONIOLO, C. (1çç(). Three shocls, ìwo
zo Eurcpean Reztev cf Eccncmtc Htstcry
recoveries: hisìorical parallels for ìhe end of ìhe Cold War. Rtztsta dt Stcrta
Eccncmtca tt, pp. zç;–¸16.
IEINSTEIN, C. H., TEMIN, P. and TONIOLO, C. (1çç;). The Eurcpean Eccncmy
Eetveen the 1ars. Oxford: Oxford Lniversiìy Press.
CALOR, O. and WEIL, D. N. (zooo). Populaìion, ìechnology and growìh: from
Malìhusian sìagnaìion ìo ìhe demographic ìransiìion and beyond. ¬mertcan
Eccncmtc Reztev µo, pp. 8o6–z8.
CRILICHES, Z. (1çç6). The discovery of ìhe residual: a hisìorical noìe. )curna/ cf
Eccncmtc Ltterature ¸q, pp. 1¸z(–¸o.
INTERNATIONAL LABOR OIIICE (1ç86). Eccncmtca//y ¬cttze Pcpu/attcn Esttmates
and Prcjecttcns, Iµyo–:o:y, vol. (. Ceneva: ILO.
IRWIN, D. A., (1çç6). The Lniìed Sìaìes in a new global economy? A cenìury’s
perspecìive. ¬mertcan Eccncmtc Reztev, Papers and Prcceedtngs 86, pp. (1–6.
]ONES, C. I. (1çç;). On ìhe evoluìion of ìhe world income disìribuìion. )curna/ cf
Eccncmtc Perspecttzes tt, pp. 1ç–¸6.
KINDLEBERCER, C. P. (1ç6;). Eurcpe’s Pcstvar Crcvth. The Rc/e cf Labcr Supp/y.
Cambridge, MA: Harvard Lniversiìy Press.
KONCSAMLT, P., REBELO, S. and XIE, D. (1çç;). Beyond balanced growìh.
Worling Paper no. 61¸ç, Naìional Bureau of Economic Research.
LEWIS, W. A. (1ç¸(). Developmenì wiìh unlimiìed supplies of labor. Manchester
Schcc/ zz, pp. 1¸ç–ç1.
LLCAS, R. E., ]r. (1ç88). On ìhe mechanics of economic developmenì. )curna/ cf
Mcnetary Eccncmtcs zz, pp. ¸–(z.
MADDISON, A. (1çç¸). Mcnttcrtng the 1cr/d Eccncmy, I8:o–Iµµ:. Paris: OECD.
MANKIW, N., CRECORY, D. R. and WEIL, D. N. (1ççz). A conìribuìion ìo ìhe
empirics of growìh. Quarter/y )curna/ cf Eccncmtcs, toy, pp. (o;–¸;.
MITCHELL, B. R. (1çç8). Internattcna/ Htstcrtca/ Stattsttcs. Eurcpe, I;yo–Iµµ¡.
London: Sìoclìon Press.
OBSTIELD, M. and TAYLOR, A. M. (1ççç). C/cba/ Captta/ Mar/ets. Integrattcn,
Crtsts, and Crcvth. New Yorl: Cambridge Lniversiìy Press.
ROMER, P. (1ç86). Increasing reìurns and long-run growìh. )curna/ cf Pc/tttca/
Eccncmy µq, pp. 1ooz–¸.
SACHS, ]. D. and WARNER, A. (1çç¸). Economic reform and ìhe process of global
inìegraìion. Ercc/tngs Papers cn Eccncmtc ¬cttztty, pp. 1–118.
SAMLELSON, P. A. (1ç(¸). Iull employmenì afìer ìhe war. In S. E. Harris (ed.),
Pcstvar Eccncmtcs Prcb/ems. New Yorl: McCraw-Hill.
SOLOW, R. M. (1ç¸;). Technical change and ìhe aggregaìe producìion funcìion.
Reztev cf Eccncmtcs and Stattsttcs ¸µ, pp. ¸1z–zo.
SOLOW, R. M., and TEMIN, P. (1ç;8). Inìroducìion: The inpuìs for growìh. In
The Cambrtdge Eccncmtc Htstcry cf Eurcpe, vol VII: The Industrta/ Eccncmtes.
Captta/, Labcur, and Enterprtse. Cambridge: Cambridge Lniversiìy Press.
TAYLOR, A. N. (1ççç). Sources of convergence in ìhe laìe nineìeenìh cenìury.
Eurcpean Eccncmtc Reztev q¸, pp. 16z1–(¸.
TEMIN, P. (1ç66). The relaìive decline of ìhe Briìish sìeel indusìry, 188o–1ç1¸. In
Henry Rosovsly (ed.), Industrta/tzattcn tn Tvc Systems. New Yorl: Wiley.
TEMIN, P. (1çç;). The Colden Age of European growìh: a review essay. Eurcpean
Reztev cf Eccncmtc Htstcry t, pp. 1z;–(ç.
TEMIN, P. (1çç8). The causes of American business cycles: an essay in economic
The Cc/den ¬ge cf Eurcpean grcvth reccnstdered z1
hisìoriography. In ]effrey C. Iuhrer and Scoìì Schuh (eds), Eeycnd Shcc/s.
Bosìon, MA: Iederal Reserve Banl.
TEMIN, P. (1ççç). Clobalizaìion. Oxfcrd Reztev cf Eccncmtc Pc/tcy t¶, pp. ;6–8ç.
TEMPLE, ]. (1ççç). The new growìh evidence. )curna/ cf Eccncmtc Ltterature ¸y,
pp. 11z–¸6.
TEMPLE, ]. and VOTH, H.-]. (1çç8). Human capiìal, equipmenì invesìmenì, and
indusìrializaìion. Eurcpean Eccncmtc Reztev qz, pp. 1¸(¸–6z.
YOLNC, A. (1çç(). Lessons from ìhe Easì Asian NICs: a conìrarian view.
Eurcpean Eccncmtc Reztev ¸8, pp. ç6(–;¸.
YOLNC, A. (1çç¸). The ìyranny of numbers: confronìing ìhe sìaìisìical realiìies of
ìhe Easì Asian growìh experience. Quarter/y )curna/ cf Eccncmtcs tto,
pp. 6(1–68o.
Internet scurces
Penn World Tables ¸.6: hììp://arcadia.chass.uìoronìo.ca/pwì/.
IAOSTAT Sìaìisìics Daìabase:
hììp://apps.fao.org/lim¸oo/nph-wrap.pl?Populaìion&DomainϭSLA.
Social Indicaìors of Developmenì: hììp://www.ciesin.org/mep-bin/
charloììe?sìaìeϭSTART&evenìϭsìarì&proìocolϭsid&charloììe_dirϭprod.
zz Eurcpean Reztev cf Eccncmtc Htstcry

European Review of Economic History

I proceed by describing the phenomenon to be explained and reviewing earlier attempts to explain it. I add insights from recent research in economic history to propose a new explanation. I then formulate this hypothesis explicitly and test it against the data. Finally, I return to the three questions posed above and summarise the new answers. . The phenomenon to be explained The phenomenon to be explained is shown in Table . The difference between the growth rate of GDP and GDP per capita comes from the gradual slowdown of population growth, and the growth rate of GDP per capita in recent years is very close to its rate before the Great War. In between, Western Europe had first slow growth and then rapid growth. It is the latter I am trying to explain. Slow growth from  to  was the result of two world wars and the Great Depression. It is common to regard the Great Depression as a failure of aggregate demand. Prices fell at the same time as industrial production, indicating a movement along an aggregate supply curve rather than a shift of that curve (Bernanke ). Although the wars had many effects on the supply side, their primary impact was also on demand (Feinstein, et al. ). To a first approximation, therefore, the slow growth was the result of deficient aggregate demand. It follows that the overall path of GDP per capita in Table  can be seen as a steady growth of ‘potential GDP’ with a deviation from this potential during the world wars and Great Depression. Total factor productivity in this view continues on its way, independent of all the demand-side activity in the wars and interwar turbulence. This extreme version of Solow growth theory ignores all fluctuations in the rate of growth of knowledge and of capital, but it does not seem to be too far from the experience of the United States where we have the data to look at the early twentieth century (Solow ). Slow growth from  to  then left Western Europe below its potential GDP, and rapid growth thereafter brought it back to its growth path. Table . Economic growth in Western Europe at different times (per cent per year).
Period – – – – GDP . . . . GDP per capita . . . .

Source: Feinstein, et al. , p. . Fifteen countries; data from Maddison ().

 . . . AUT BEL CHE DEU DNK ESP FIN FRA GBR IRL ITA NLD NOR PRT SWE . Table . . . . . .The Golden Age of European growth reconsidered  The problem is the length of time in Table . As noted in several studies. While this simple model of deviation from a smooth trend is appealing. . New growth theory and conditional convergence can produce a history like this with convergence with about  years or so to half-way convergence (Mankiw. .. I do not want to suggest that the disequilibrium studied here was the only phenomenon taking place after the war. – (per cent per year). the more challenging question is whether there is a unified explanation for the variety shown in Table . there also is a cross-section question: Why did some countries grow so much more rapidly in this period than others? The spread of growth rates among Western European countries in this period is shown in Table . It would not be surprising if the explanation of the intermediate case involved both demand and supply. several European countries did not return to their prewar trend paths of growth after the war or even after the Golden Age (Crafts and Mills ). . It is a wide range and needs to be explained. Annual rates of growth varied from two to five per cent a year. . conditional convergence involves supply phenomena. . It is harder to find a good explanation for this intermediate time frame. Annual rates of growth in Western Europe. The business cycle can produce a path like this with a time span of a year or two. . for quite complete convergence in about thirty years. In addition to the time-series questions about the beginning and end of the Golden Age. ). but not necessarily the same one they had experienced before the war. I reserve for future work the integration of the Golden Age and the subsequent growth path. Austria Belgium Switzerland Germany Denmark Spain Finland France Great Britain Ireland Italy Netherlands Norway Portugal Sweden Source: Penn World Tables . . et al. National histories always contain developments that can be used to explain rapid or slow growth. Business cycles generally are a demand phenomenon (Temin ). For those countries. the end of the Golden Age represented a return to a more durable growth path.

Solow’s framework had provided a way to organise historical data on economic growth. Many people argued that movements in aggregate supply led to the slowdown of growth as well as higher inflation. The focus on supply conditions led to new growth theory. and it predicted that all countries would converge to the same rate of growth. Empirical investigations flowered in the form of growth equations. education is only the most prominent of many putative inputs to growth. was the subject of myriad papers and books. European Review of Economic History An early contribution to the literature on postwar growth was provided by Kindleberger () using the Lewis () model of excess labour supply to explain both differences in growth rates between countries and the slowdown in growth he could detect in the mid s. The two shocks most often identified were the rise in oil prices in  and increasing rigidity in industrial labour markets (Bruno and Sachs ). although they still left room for conditional convergence for groups of similar countries. Lucas () introduced human capital to the model as an additional determinant of growth. Griliches ). The limitations of the Solow growth model were attacked in turn. The slowdown of growth in the s. This article builds upon and extends Kindleberger’s view of thirty years ago. and growth accounting was born. This proved to be an enormously illuminating way to summarise a vast body of knowledge and begin the process of explaining economic growth (Solow and Temin . It was the exhaustion of cheap labour that caused economic growth to slow. Kindleberger’s argument was simple: an elastic labour supply promotes economic growth by keeping wages low and preserving industrial peace. as had been done informally in growth accounting and in economic history (Denison . not exogenous. but few . convergence to the same rate of growth by all countries). it could not account for the wide differences between countries that we observe. which stressed the role of supply in the long run. This limitation led people to lump all other differences between countries into TFP and provide explanations outside the theory why they differed (Denison ). Characteristics of the labour market continue to be active topics in the explanation of European economic difficulties. New growth theories provided extensions to get around the limitations of old growth theory at the expense of Solow’s simplicity and elegance. Differences in education between countries eliminated the prediction of unconditional convergence (that is. Easterlin ). investment and TFP could be listed as determinants of growth. The oil shocks have faded into history while remaining the most popular candidates for causing the end of the Golden Age. giving rise to new growth theory. Wide differences between countries now could be explained within the model by differences in educational attainment (Mankiw. Romer () argued that TFP growth was endogenous. Population. ). known at the time as stagflation. et al. sometimes called ‘convergence clubs’. But Solow’s growth model did not include any other variables.

as done in Young’s famous dissection of economic growth among the Asian NICs and Jones’ survey of the world income distribution (Young . The period typically starts in  – Barro started later so he could use  GDP as an instrument – both to exploit easily available data and to avoid the recovery period just after the war. ). were based on regressions for periods stretching from  to  with no acknowledgment that the process of growth might be different at the beginning and end of the period. for example. but not because the data came from two separate economic periods (Barro . pp. and the essays are quite diverse (Temin ). Dowrick and Nguyen () provide a solitary exception to this rule. One country study that anticipated the approach here conceptualised the . was overwhelmed by the strengths of particular issues in the debates about individual countries. The latter by the s was simply assumed as a fact of economic life. Jones ). Crafts and Toniolo opened a volume of essays on the period by asserting. p. The common practice still is to lump the postwar period into one cross-section. They examine whether the convergence found by Baumol () continued after . Temple ). They argued that rapid growth in this period was partly a consequence of slow growth in the previous period. ‘the years – witnessed a unique episode in the history of European “modern economic growth” ’ (Crafts and Toniolo . –).The Golden Age of European growth reconsidered  of these regressions acknowledged anything special about the Golden Age of Economic Growth. but they did not dwell on the mechanism of such a reaction. The regressions focused on identifying the equilibrium growth rate to which countries were converging rather than estimating conditional convergence itself. Barro’s Robbins Lectures. Baumol () provided evidence of convergence over a century for a sample of mostly Western European countries. The book as a whole is a survey of the experience of about a dozen Western European countries during the Golden Age in a compatible format. Economic historians also have turned their attention to the postwar years. testing earlier informal results with growth equations. The time period chosen is much shorter than Baumol’s in order to exploit the plentiful data after World War II. which means over a hundred in today’s world. and the field turned to a prolonged investigation of the factors that determine to what rate of growth countries will converge. The claim that this was a universal pattern did not stand up (De Long ). Two recent surveys of this literature describe the diversity of approaches taken to identify ‘convergence clubs’. Growth regressions typically are done for as many countries as possible. but they do not remark on any special treatment of the Golden Age of Economic Growth (Durlauf and Quah . The focus was on convergence rather than the rate of growth. This exercise of fitting diverse histories into a common mould. however. . He commented that this was an improvement on his prior practice of using a single cross-section.

I take my start from Dumke. Eichengreen saw an implicit bargain between workers and investors that is similar to the implicit contracts Aoki () described in what he called the J-firm. union representation on company boards. Eichengreen emphasised their role in precluding negative effects. it explains how demand could grow to promote rapid economic growth during the Golden Age. create jobs and raise wages. Eichengreen listed many causes for the end of the Golden Age. domestic to enforce the bargain just described. But it cannot explain how Western Europe found itself so far from equilibrium at the start of the Golden Age. The international ones included institutions like GATT. investors have an incentive to pay themselves the resulting profits instead of reinvesting them. This organizational view also does not distinguish between different countries in Western Europe because the international agreements that form such a large part of the story include them all. Eichengreen () offered a synthetic view in his contribution to the Crafts and Toniolo volume. The domestic institutions included national wage bargaining. workers have the incentive to take the gains home in the form of higher wages. Dumke () argued that greater wartime destruction generated faster postwar growth and provided evidence for this proposition in growth regressions for OECD countries. international to promote national specialization that increased efficiency. And if investors make productive investments that enhance the productivity of labour. The institutions were both domestic and international. ECSC and EPU that appeared to have had little positive effect. assuring that trade would remain free as conditions changed. who asserted that the destruction of physical capital during the war was less important than the maintenance of what he called ‘the social capability’ of growth. and conditional access to government programmes. Starting from the observation that growth in the Golden Age was related to catching up and high investment. over time. but shift his emphasis and his sample. If workers moderate wage demands. This is an intriguing and plausible hypothesis. revealing the . he asked why investment was both high and productive in the Golden Age. These perverse incentives were countered in postwar Europe by a complex set of institutions that made reneging harder and increased incentives for honouring the long-term implicit contract in the face of short-run gains from abrogating the contract. These in turn were due to government institutions and policies that were sharply different from those pursued before the war. The bargain was that workers would not push for higher wages if investors would make productive investments that would. typical of postwar Japan. This bargain is time-inconsistent. European Review of Economic History German Wirtschaftswunder as a disequilibrium phenomenon. His inquiry was in the spirit of Abramovitz (). He answered that wage moderation and export growth made investment attractive and profitable. Investors would agree to invest on the condition that the workers did not immediately try to take all the gains in higher wages.

compared to five per cent for the United Kingdom (Broadberry . Galor and Weil ). the end of general catch-up. the end of the Bretton Woods System. The familiar rise over the last century can be seen. Broadberry () evaluated the importance of this transition in Germany’s convergence to British levels of labour productivity. There are now several models attempting to integrate structural shifts with the theory of economic growth (Kongsamut. p. In . Temple and Voth . and the temporary decline in  was eliminated by  when it stood at  (Broadberry . Recent theories of economic growth I approach the Golden Age in the context of economic growth over the past century or two. German relative labour productivity was as high in  as it would get. national income grew during this transition. which had a large component of economic transition. Since workers were more productive in non-agricultural activities. the oil shocks. p. ). All of these papers share with this one the attempt to bring the historical experience of industrialisation into the mainstream of thinking about economic growth. but the result of transferring resources from low-productivity sectors like agriculture to high-productivity ones like manufacturing. ). Theorists of economic growth recently have begun to acknowledge the importance of this transition in the process of economic growth. The second column of Table  reveals that the rise in comparative labour productivity in manufacturing did not echo the rise in the aggregate. et al. In fact. Broadberry asserts. National economies around  with very few exceptions were almost completely agricultural. It follows that faster economic growth in Germany than in Britain was due largely to the more rapid sectoral shifts in the German economy. Starting in the nineteenth century and even later. . Germany had a larger share of its labour force in agriculture than the United Kingdom throughout the past century. Germany had  per cent of its labour force in agriculture. Residents became urban. and reduced incentives to keep the bargains that produced the Golden Age of Economic Growth. and the share of the labour force in agriculture fell. The first column of Table  shows his estimate of aggregate labour productivity in Germany compared to the United Kingdom. Catch-up. Among the reasons offered were the capture of institutions by firms and unions. with a dip in  – just after the Second World War. Taylor () used a model of this type in his exploration of convergence in seven countries before World War I. around the start of the Golden Age of European Growth. . is not the result of improving efficiency in manufacturing.The Golden Age of European growth reconsidered  absence of a unified explanation. If rapid economic growth is the result of the . there is very little evidence of a trend at all. productive resources were moved out of agriculture into manufacturing and services.

Comparative labour productivity in Germany and the United Kingdom (UK ). European Review of Economic History Table . and people was interrupted by world wars and depression. p. ). International trade was interrupted by the First World War. Why was Germany lagging behind Britain in this transition? Three reasons come to mind. Germany chose to protect its farmers against low-priced American grain in the late nineteenth century. Obstfeld and Taylor . Before the Great War. The postwar settlement created many new boundaries that provided the opportunity to impose tariffs on trade. The third of these reasons has been neglected. Temin ). They could not exploit their comparative advantages. The volume of exports for the major Western European countries was lower in  than it had been in . while open economies did. But in between these two end points. The growing literature on globalisation argues that it has ebbed and flowed in the course of the twentieth century. the Second Thirty Years War – the turbulent period from  to  – interrupted international trade and slowed the transition. Their regressions showed that closed economies did not exhibit convergence. . Germany started its industrialisation after Britain. international commerce and travel were free and open. finance. And the Great Depression led to restrictive trade policies that reversed whatever expansion had taken place in the s. of which the third has not been appreciated. and they could not end their reliance on domestic . Sachs and Warner () argued that trade promoted economic growth in the postwar world. then Germany was still engaged in the process during the Golden Age while Britain had completed its transition. GDP      Manufacturing      transition from an agrarian economy. in sharp contrast to its rapid growth both before and after this period (Feinstein. I want to expose its importance. . Second. Third. Year      Source: Broadberry . Authors disagree among themselves about whether today’s globalisation actually existed a century ago. the flow of goods. Why did closed economies suffer? Because they did not undertake the reallocation of resources needed to increase productivity. p. First. . et al. more or less as they are today. but there is no disagreement about the interruption during the world wars and Great Depression (Bordo et al.

in a disaggregated model of development. participation in international trade promoted industrialisation. –). much earlier. The decline in the share of the labour force in agriculture was twice as rapid in the s and s as before. The United States. but he concluded that most of the effect came from the changing size of agriculture. a prominent story asserts. As a result. The rate at which workers left agriculture accelerated after the war. whether because of the greater stability of Western Europe or because of noise in the imperfect earlier data. The developmental deficit highlighted here is produced by a misallocation of resources. Britain was surpassed.The Golden Age of European growth reconsidered  agriculture. insulated from the wars if not the Depression. as quoted in Mitchell . because the United States and Germany were better able to exploit world markets (Temin ). The first takes place in a single-sector economy. but the parallel is clear. There are many ways to divide up the economy. but the division between agriculture and all other activities appears to be the most important. and he . pp. Low income in the standard story is produced by low levels of physical and human capital relative to saving rates. The misallocation of resources can be measured by the share of the labour force in agriculture. Sachs and Warner did not dwell on the connection between their theory and the history of industrialisation in Europe during a previous period. was able to continue its transformation from an agricultural to an industrial economy. It follows from this view that the barriers to international commerce during the world wars and Great Depression constituted barriers to the continued industrialisation of European countries. talked of the misallocation of resources in Europe during the Golden Age of European Growth. the supply frontier continued to expand during the Second Thirty Years War. Before World War I. This slowdown in the process of industrialisation created a disequilibrium after the war. Its exports were primarily food and raw materials before this protracted conflict. As suggested by Table . The variance of the measured change fell as the rate increased. Open economies decreased the proportion of food and raw materials in their exports more rapidly than closed economies. One has only to recall the discussion of Britain’s ‘climacteric’ in the late nineteenth century to see the importance of international trade in economic growth. This disequilibrium is separate from the low income that generates conditional convergence. Denison (). the second. changes from before World War I to the interwar period are lost in this volatility (Bairoch . Broadberry () distinguished nine sectors of the economy. The standard deviation of the decadal rate of change in the share before World War II was three times as large as the standard deviation of the quinquennial change thereafter. European countries emerged from the war with a developmental deficit. they were manufactures afterwards (Irwin ).

that is. In parallel with conditional convergence. A. this phenomenon will be measured by the difference between the initial proportion of the labour force in agriculture. came from a generation – thirty years – of economic insularity. The model then is as follows. and recalculating it from Maddison (). They also were accomplished by rapid reductions in the size of agriculture in these countries. The misallocation. starting from a level of income low relative to the country’s equilibrium income. differs from growth regressions in the literature. per capita GDP. and the equilibrium share. Growth is regressed on cur- . Wartime destruction was like a business cycle in leading to a short-term disequilibrium. Conditional convergence might explain long-term disequilibria. despite its conventional appearance. Testing the hypothesis I test this hypothesis by formalising the story in a simple model and testing it against data from the Golden Age of European Growth. Young () showed that these growth miracles were accomplished by very high investment rates. Dumke () measured the extent of this dislocation by the percentage gap between per capita GDP in  and in . slow or medium growth rates. g a b(y* y) c GAP d (A A*) e () This regression. where g is the average growth rate of y. () Arrested development. Jones () conceptualised growth as a kind of Markov process. Jones characterised the fast growth as growth miracles and asserted that they were most prevalent among poorer countries. Countries drew their rate of growth from an urn once every thirty years or so. drawing fast. although not among the poorest. excessive labour in agriculture. labelling it GAP. European Review of Economic History too meant the European countries were growing rapidly when they were getting out of agriculture. according to this story. . I use this measure here as well. It is reasonable to think that the excessive resources in agriculture could be moved to other sectors in another thirty years. A*. () Wartime destruction that deranges production in the short run. This hypothesis therefore provides a way to rationalise the history shown in Table . that is. Those growth regressions are designed to elicit differences between y* in different countries. The model distinguishes three kinds of disequilibria that can affect growth: () Conditional convergence. Autarchy in the Second Thirty Years War can explain a disequilibrium that can be eliminated in twenty or thirty years. This phenomenon may be more general than Europe after the Second World War.

all three kinds of disequilibrium affected the rate of growth in different countries in the Golden Age of European Growth. then resource misallocation became important. but it is appropriate when discussing economic growth in Western Europe. The influences of geography. This emphasis is appropriate in a study of a single region and in the investigation of disequilibrium growth during the Golden Age of Economic Growth. that proxy for and identify y*. I am trying to describe the process of convergence. Third. historical quality of the Golden Age comes from the presence of the first two.) They all are part of the same ‘convergence club’. and the Common Agricultural Policy are taken to be second-order effects. resource misallocation. They all have stable governments. history. In other words. harking back to the origins of new growth theory. . the end of the Golden Age came when these unusual disequilibria were removed. We cannot observe these effects at other times because they are not there. (This is the same set of countries whose growth is reported in Table . and finally conditional convergence became central. these disequilibria had their main impact at different periods. Equation () can be rewritten. First. The presence of the second two disequilibria account for the rapidity of economic growth during the Golden Age. while growth regressions typically assume that countries are near their growth path and investigate the nature of the equilibrium income (y*) to which they are converging. In normal periods of peace. A*. I assume here that y* is the same for all countries in Western Europe. and it is reasonable to argue that A* is vanishingly small in Western Europe today. I also assume that the equilibrium share of agriculture. wartime destruction and the misallocation of resources ceased to have an effect on growth some time around . secure property rights. only the last of these is relevant. One could not make this heroic assumption with a wider sample. Differences between countries will show up in the error term and in the goodness of fit.The Golden Age of European growth reconsidered  rent income and many variables. they were present in postwar Europe only because of the historical circumstances that were unique to this time and place. so that it is a predetermined variable. The share of the labour force in agriculture is measured at the beginning of each period. and universal education. except that Czechoslovakia has been replaced by Portugal. like education. Second. I use this regression to test three hypotheses. The special. g (a by* dA*) by c GAP dA e () I estimate this equation for all  Western European countries after the Second World War. Wartime destruction affected growth in the immediate postwar years. is the same for all Western European countries. collecting the unobserved equilibrium levels with the constant term. The sustained rapid growth comes from the presence of the second disequilibrium.

 March .) . (.) . (. (.) . Sources: Postwar GDP per capita data from the Penn World Table . . (. those in Table . ‘A’ from International Labor Office (). The first thing to note is that the regressions account for a substantial amount of the variation in growth rates among the Western European coun- .) . . (. (. . FAOSTAT (on the web). ( . (.) .) . ( .) | . GAP per capita GDP in  over per capita GDP in . (. Variable Yo GAP A Constant Adjusted R N – . ( .) .) .) . ( . . for ten-year intervals. – . ( . – . ( .) . . (. – .) . (.) . (.) .) .. ( . Regressions explaining twenty-year growth rates. – . – . ( . The data start in  because data for  are unavailable and unreliable. . T-statistics are in parentheses below the coefficients.) .) . (. T-statistics are below the coefficients. (. The results in Table  are useful for looking at the Golden Age as a whole. ( . Sources: See Table . Preliminary data for  from Robert Summers by private correspondence. .) . ( .) . The first of these tables shows regressions for twenty-year intervals which correspond closely to the Golden Age. (. following Barro (). the second.) . (.) . .) .) . ( .  (. ( . (.) .) . Regressions explaining decadal growth rates.) . Variable Yo GAP A Constant Adjusted R N – . ( . – . (.) . (.) .) . The regressions are shown in Tables  and . for tracking the effect of different disequilibria within the Golden Age. and Social Indicators of Development (on the web). European Review of Economic History Table .) .) . Table .) . from Maddison ().

The second thing to note is that inferences from the standard errors are slightly problematical in this context. The years also are not a random sample from a larger set. by contrast. but more slowly. The decreasing size of the coefficient over time suggests .) The actual difference in growth rates in – was . the estimated effect became smaller. The presence of two heavily damaged countries was enough to make wartime destruction appear important. as was the misallocation of resources (A). It also created the opportunity for many countries to grow rapidly as they allocated their resources more efficiently. The first two columns of Table  show regressions for the Golden Age. . First. they are the years of the Golden Age of European Growth. both absolutely and relative to its standard error. he lumped the entire Golden Age into one regression. These countries are not a random sample of a larger set. this effect explains over half the difference for these years. and he did not track effects within the period. The share of Germany’s labour force in agriculture at the start of the Golden Age was  percentage points higher than the share of Britain’s agricultural labour force. An inference based on a t-test would be wrong in only  per cent of cases if we could find similar sets of countries after similarly long periods of war and economic autarky to study. they are all the countries of Western Europe. as for many growth regressions. was not an important factor. As the share of the labour force in agriculture approached its equilibrium level. The effect was large. Germany’s growth rate received a boost of approximately one percentage point from this initial condition. Conditional convergence. Neither wartime destruction nor long-run conditional convergence were as important as the misallocation of resources in explaining differences between growth rates in Western Europe at this time. The assumption that y* and A* are the same for all countries in Western Europe after the Second World War does not appear misleading. Dumke () found wartime destruction to be more important for two reasons. The first column of Table  contains regression for the first postwar decade for which we have reliable data. despite the drastic simplifications of the model. The effect of excessive labour in agriculture (A) is visible clearly. which – like Germany – had extensive wartime damage. It was only an important factor at the start of the Golden Age.The Golden Age of European growth reconsidered  tries during the Golden Age. The effect of the agricultural labour force faded away as well. Succeeding columns show that the influence of wartime destruction faded rapidly into the background. per cent if – is used. (The effect is . per cent if – is used. He also included Japan in his sample. Wartime destruction (GAP) was important in determining the rate of growth of different countries in this period. per cent as shown in Table . This indicates that the misallocation of resources arising from the limitation of international trade during the Second Thirty Years War was costly to Western Europe.

The Wirtschaftswunder appears clearly. It follows that postwar British economic policies were not the cause of slow economic growth. The end of the Golden Age was brought about by the return to equilibrium growth paths on the part of these countries. it industrialised first. Truncating GAP to be less than zero also does not change the results. and it kept tariffs low when American grain was able to be transported cheaply to Europe in the late nineteenth century. The presence of these disruptions makes it hard to identify timing in longer-run phenomena. This model provides an explanation for this slow growth at variance with that in the literature (Bean and Crafts ). there are a few differences worth noting. perhaps an ‘S-shaped’ relation between the excess share of labour in agriculture and economic growth. showing –. contained the effects of two oil shocks which disrupted the longer-run processes of interest here. these disruptions confuse the story. Lagging A by five years to avoid any hint of simultaneity preserves the results in Tables  and . reveals Germany (DEU) to be an outlier. Scatter diagrams illustrate the regression results. We need to be careful in our claims of special conditions in individual countries. Figures  and  show the relation between the rate of growth and the initial share of the labour force in agriculture for two overlapping -year periods starting in the s. Belgium looks to be an economic miracle. of course. and the share of variance of growth rates explained is negligible. Britain is at the low end of European growth rates. Instead. Britain had started industrialisation with a low share of labour in agriculture. For the slow British growth is explained quite fully by the low share of the labour force in agriculture at the start of the Golden Age. This decade. But in Figure . and intermediate ten and twenty year time periods yield intermediate results to those shown here. various experiments show these equations to be robust. In the view I am presenting here. No coefficients are estimated tightly. Germany no longer is an outlier. Although it is hard to test for robustness with only  observations. The results differ only in details if log(yo) is used in place of yo to indicate general convergence. Figure . While the diagrams are similar. not by adverse economic shocks in the s. but they are not the story itself. In both diagrams. And the results are unchanged if one or another small country is dropped from the sample. while the other forces present in the Golden Age do not show up. showing –. It may not be too farfetched to say that poor poli- . The regression for – in the final column of Table  shows that conditional convergence as indicated by a significant coefficient of yo is verified. as everyone knows. These past accomplishments implied slow growth during the Golden Age. or even if two outliers like Spain and Portugal are dropped together. the correlation shown in Table  is quite apparent. The regression for – is markedly worse than those for other decades. European Review of Economic History a non-linearity.

Regression of economic growth rate and initial share of labour force in agriculture. Figure . –.The Golden Age of European growth reconsidered  Figure . . –. Regression of economic growth rate and initial share of labour force in agriculture.

  Source: See Table . ( . Variable Trend Trend squared Constant Overall R N . It can be seen also in the rate of change of the share of labour in agriculture. that is. but there is no evidence of a change in the proportionate rate of decline. If one examines the change in the labour force in agriculture. becomes less negative. This view suggests that British economic policies did not have the potential effects attributed to them in this kind of account. European Review of Economic History cies were the result of slow growth rather than its cause. If. the share of labour in agriculture falls as income rises. since it is the movement of people that generates growth. (.) .) . The model of equation () may be the correct model for all time.) () . (. . Trends of the agricultural labour share.  Change in the share () .  Growth in the share () . ( .) . ( . (. This can be seen in the regressions in Tables  and . In peaceful times. The change in the agricultural labour force gets smaller (more positive) over time. as shown in the first row of Table . one examines the rate of change of the agriculture share. as in equation (). This is the variable that is relevant for growth. And it suggests even further that what we regard as poor policies – because Britain grew so slowly – were either reasonable adaptations to Britain’s initial position or the results of Britain’s slow growth. however. The Golden Age of Economic Growth ended when this disequilibrium was eliminated. Z-statistics in below the coefficients. where the coefficient of A is not estimated clearly in the later regressions. in which Britain’s initial position plays only a minor role. then the rate of change falls over time. when the share of the labour force in agriculture approached its equilibrium level. The share approaches its asymptote at a constant rate. then there is no change over time. Various factors and policies can inhibit this change in individual countries.) . that is. The evidence is consistent with an asymptotic approach to an equilibrium share.) .) . that is. ( . but GAP Table . the change in the share divided by the initial share. but there are no other historical cases where large numbers of industrial countries were in the same disequilibrium position. The regressions in Table  help us to understand why this is a historical explanation. Bean and Crafts () offer a multi-layered picture of British policies.

All errors are mine alone. some of the bitterness of policy in slowlygrowing Britain may have been the result of slow growth instead of its cause. confusing short-run and longrun factors. Third. Burak Guner. In this state of arrested industrialisation. why did the rapid growth come to an end? Growth slowed in the s and s because the disequilibrium that had generated unusually rapid growth no longer existed. and Ayako Tanaka for research assistance. the slowdown of growth would have taken place even if they had not occurred. policies of the time have been blamed for events outside the control of policymakers. The most important of these was the misallocation of resources that came from the lack of international trade during the preceding thirty years. Oxford University. . but other kinds of disequilibrium were. National policies had secondary effects relative to countries’ initial position. The developmental deficit of a generation was eliminated in a generation. I thank Johnny Chen. The institutional factors cited by Eichengreen () helped create the needed demand. reallocated labour rapidly enhanced the supply.The Golden Age of European growth reconsidered  and (A-A*) typically are at or near zero. The normal catch-up that works in general was not important right after the war. The final regressions in Tables  and  illustrate the inability to estimate these coefficients outside the Golden Age. Labour relations were tumultuous in both Italy and Britain.  March. These crises muddied the historical waters. Zvi Griliches was kind enough and well enough to advise me in the early stages of this article. In addition. It is impossible to estimate their coefficients under these circumstances. but the misallocation of resources had ceased to be a large macroeconomic issue by the time of the oil crises. Second. why did different countries grow at different rates during the Golden Age? They grew at different rates during the Golden Age because of their initial position. too many resources still were employed inefficiently in agriculture. . Dirk Niepelt. and audiences at the lecture and other seminars for helpful comments. Conclusion I return now to the three questions posed at the beginning of this article. The Common Agricultural Policy may have kept some excessive labour in agriculture. First. but Italy grew rapidly while Britain did not. why was growth so rapid in Western Europe during the Golden Age of Growth? The answer is disequilibrium. Acknowledgements This article is a revised version of the Hicks Lecture. At the least. . It is only with the hindsight of another thirty years that we can see that while the oil crises were disruptive.

(eds). (). BROADBERRY. G. –. S. M. BAIROCH. (). pp. B. (). DURLAUF. –.. MA: MIT Press. G. and welfare: comment. (). DUMKE. BARRO. E. C. Economic Growth in Europe since . (). BAUMOL. In Bart van Ark and Nicholas Crafts (eds). A. A. Productivity growth. and SACHS. Cambridge: Cambridge University Press. pp. European Review of Economic History. T. The Working Population and Its Structure. American Economic Review . EICHENGREEN. FEINSTEIN. DOWRICK. B. N. Institutions and Economic Growth in Europe After World War II. (). N. EICHENGREEN. –. : pp. American Economic Review . (). J. N. pp. –. and MILLS. pp. J. Anglo-German productivity differences. Economics of Worldwide Stagflation. pp. D. Credit and Banking . and falling behind. TEMIN. Catching up. BRUNO. Economic Growth in Europe Since . B. European Review of Economic History References ABRAMOVITZ. convergence. BORDO. P. Why Growth Rates Differ. P. M. R. Productivity growth. (). and TONIOLO. (). (). W. J. pp. and Bargaining in the Japanese Economy. –. Europe’s Golden Age: an econometric investigation of changing trend rates of growth. Is globalization today really different than globalization a hundred years ago? Brookings Trade Forum. Cambridge: Cambridge University Press. forging ahead. CRAFTS. DC: Brookings Institution. Quantitative Aspects of Post-War European Economic Growth. Cambridge. Cambridge: Cambridge University Press. T. British economic growth since . EASTERLIN. The macroeconomics of the Great Depression: a comparative approach. Determinants of Economic Growth.-T. BEAN. CRAFTS. BERNANKE. Working Paper no. –. (). D. Université Libre de Bruxelles. pp. and NGUYEN. D. American Economic Review . C. In Nicholas Crafts and Gianni Toniolo (eds). Information. Cambridge. pp. Three shocks. C. MA: Harvard University Press. (). and QUAH. (). S. D. Why isn’t the whole world developed? Journal of Economic History . M. –. Journal of Money. Cambridge: Cambridge University Press. (). . In Nicholas Crafts and Gianni Toniolo (eds). –. B. two . M. S. OECD comparative economic growth –: catch-up and convergence. Brussels: Institut de Sociologie. N. Reassessing the Wirtschaftswunder: reconstruction and postwar growth in West Germany in an international context. Economic Growth in Europe Since . N. –. Incentives. R. (). DE LONG. R. (). National Bureau of Economic Research. AOKI. Oxford Bulletin of Economics and Statistics . and TONIOLO. Washington. –. and CRAFTS. (). Journal of Economic History . convergence and welfare. H.. (). et al. and IRWIN. (). The new empirics of economic growth. (). Washington DC: The Brookings Institution. Cambridge: Cambridge University Press. DENISON.

American Economic Review. Economic reform and the process of global integration. Brookings Papers on Economic Activity. American Economic Review . (). M. (). The causes of American business cycles: an essay in economic . pp. MA: Harvard University Press. and WEIL. and TONIOLO. (). A. P. Journal of Economic Literature . technology and growth: from Malthusian stagnation to the demographic transition and beyond. D. C. A. TEMIN. pp. Journal of Monetary Economics . ROMER. Introduction: The inputs for growth. (). J. Sources of convergence in the late nineteenth century. SAMUELSON. Full employment after the war. MADDISON. Journal of Political Economy . B. LEWIS. R. European Review of Economic History . FEINSTEIN. (). New York: McGraw-Hill. Journal of Economic Perspectives . (). M. Cambridge: Cambridge University Press. IRWIN. (). –. The relative decline of the British steel industry. GALOR. W. P. TEMIN. N. pp. D. –. In S. In The Cambridge Economic History of Europe. E. SACHS. MITCHELL. (). LUCAS. P. –. SOLOW. On the mechanics of economic development. JONES. Oxford: Oxford University Press. OBSTFELD. GRILICHES. INTERNATIONAL LABOR OFFICE (). D. (). C. Review of Economics and Statistics . –. D. (). New York: Wiley. (). P. SOLOW. P. TEMIN. pp. pp. TAYLOR. N. Population. Rivista di Storia Economica . –. New York: Cambridge University Press.. Papers and Proceedings . N. and WARNER. (). KONGSAMUT.. European Economic Review . Crisis. (). P. A. –. R. vol VII: The Industrial Economies: Capital.). ().The Golden Age of European growth reconsidered  recoveries: historical parallels for the end of the Cold War. International Historical Statistics: Europe. The Golden Age of European growth: a review essay. KINDLEBERGER.. Z.). Industrialization in Two Systems. pp. Manchester School . –. –. Economically Active Population Estimates and Projections. The United States in a new global economy? A century’s perspective.. and Growth. pp. D. and Enterprise. pp. and TEMIN. Monitoring the World Economy. In Henry Rosovsky (ed. A. N. –. . E. M. A. P. Postwar Economics Problems. O. –. National Bureau of Economic Research. pp. (). Beyond balanced growth. Global Capital Markets: Integration. R. Europe’s Postwar Growth: The Role of Labor Supply. and TAYLOR. (). S. –. –. A contribution to the empirics of growth. pp. (). –. Working Paper no. A. D. H. pp. London: Stockton Press. M.. Geneva: ILO. A. pp. GREGORY. On the evolution of the world income distribution. REBELO.. Quarterly Journal of Economics. –. R. Development with unlimited supplies of labor. (). MANKIW. –. Jr. C. Harris (ed. G. pp. and WEIL. . vol. Increasing returns and long-run growth. P. The discovery of the residual: a historical note. . The European Economy Between the Wars. and XIE. (). Cambridge. TEMIN. P. Paris: OECD. R. (). Labour. –. I. (). (). –. Technical change and the aggregate production function.

pp. and industrialization.-J. MA: Federal Reserve Bank. H. –. and VOTH. pp. YOUNG. Internet sources Penn World Tables . FAOSTAT Statistics Database: http://apps. Beyond Shocks. A. –. Lessons from the East Asian NICs: a contrarian view. pp. Fuhrer and Scott Schuh (eds). .org/mep-bin/ charlotte?state START&event start&protocol sid&charlotte_dir prod. (). TEMPLE.utoronto. (). YOUNG.chass.ca/pwt/. pp. European Economic Review . In Jeffrey C. TEMIN. Globalization. TEMPLE. J. Oxford Review of Economic Policy . –. Human capital. equipment investment. (). pp. Quarterly Journal of Economics .fao. The new growth evidence.org/lim/nph-wrap. Social Indicators of Development: http://www. Journal of Economic Literature . P.pl?Population&Domain SUA. (). Boston. European Review of Economic History historiography. –. European Economic Review .ciesin. –. A.: http://arcadia. The tyranny of numbers: confronting the statistical realities of the East Asian growth experience. J. ().

Sign up to vote on this title
UsefulNot useful