Sie sind auf Seite 1von 20

The Colden Age of European growh

reconsidered
PETER TEMIN
Department cf Eccncmtcs, Massachusetts Instttute cf Technc/cgy, Cambrtdge
M o:I:-I;, US
I reconsider he growh of Wesern Europe during he Colden Age of
European Economic Crowh afer he Second World War. The preceding
hiry years of conflic and depression impeded he normal pah of
indusrialisaion in hese counries, and hey had oo much labour in
agriculure for heir level of income and sage of developmen a he end
of he war. The disequilibrium added o oher more ordinary forces o
produce unusually rapid economic growh. This hypohesis explains he
speed of economic growh during he Colden Age, differences beween
growh raes in hese years, and he end of his hisorical episode.
I is hardly news ha he years following World War II were far differen
from hose following World War I. Economiss wriing during he war
anicipaed repeiion of some of he depressing forces and evens ha fol-
lowed he Crea War (Samuelson 1(). Bu heir predicions were no
accurae, a leas parly because of heir sudies. Policymalers had he
experience of he inerwar years o reflec on, and i is comforing o hinl
ha hey learned from experience (Ieinsein, et a/. 1().
The good imes came o an end in heir urn during he oil crises and
sagflaion of he 1;os. We lool bacl on hese imes perhaps wih more
nosalgia han may be warraned, giving hem names lile he Colden Age of
European Crowh and /es Trente C/crteuses. Economiss since hen have
been rying o undersand boh he sources of he rapid growh immediaely
afer he Second World War and of he slowdown in he 1;os. This aricle
is a conribuion o ha lieraure.
An explanaion of he Colden Age of European Crowh should answer
hree quesions. Why was European economic growh so rapid beween he
world war and he firs oil crisis? Why did differen counries grow a differen
raes during his ime? And why did he rapid growh come o an end? I argue
ha hese quesions can be answered in a unified frameworl by bringing econ-
omic hisory o bear on his quesion of economic growh. The preceding
years of wars and depression impeded he process of indusrialisaion ha had
engaged he economies of Wesern Europe since a leas he mid-nineeenh
cenury. There was as a resul a disequilibriumha has no been noed before,
ha was he source of he rapid and varied growh during he Colden Age.
Eurcpean Reztev cf Eccncmtc Htstcry, 6, zz. Prined in he Lnied Kingdom zooz Cambridge Lniversiy Press
I proceed by describing he phenomenon o be explained and reviewing
earlier aemps o explain i. I add insighs from recen research in econ-
omic hisory o propose a new explanaion. I hen formulae his hypohesis
explicily and es i agains he daa. Iinally, I reurn o he hree quesions
posed above and summarise he new answers.
t. The phenonenon to be expluined
The phenomenon o be explained is shown in Table 1. The difference
beween he growh rae of CDP and CDP per captta comes from he grad-
ual slowdown of populaion growh, and he growh rae of CDP per captta
in recen years is very close o is rae before he Crea War. In beween,
Wesern Europe had firs slow growh and hen rapid growh. I is he laer
I am rying o explain.
Slow growh from 11 o 1o was he resul of wo world wars and he
Crea Depression. I is common o regard he Crea Depression as a failure
of aggregae demand. Prices fell a he same ime as indusrial producion,
indicaing a movemen along an aggregae supply curve raher han a shif
of ha curve (Bernanle 1). Alhough he wars had many effecs on he
supply side, heir primary impac was also on demand (Ieinsein, et a/.
1;). To a firs approximaion, herefore, he slow growh was he resul of
deficien aggregae demand.
I follows ha he overall pah of CDP per captta in Table 1 can be seen
as a seady growh of poenial CDP wih a deviaion from his poenial
during he world wars and Crea Depression. Toal facor produciviy in
his view coninues on is way, independen of all he demand-side aciviy
in he wars and inerwar urbulence. This exreme version of Solow growh
heory ignores all flucuaions in he rae of growh of lnowledge and of cap-
ial, bu i does no seem o be oo far from he experience of he Lnied
Saes where we have he daa o lool a he early wenieh cenury (Solow
1;). Slow growh from 11 o 1o hen lef Wesern Europe below is
poenial CDP, and rapid growh hereafer brough i bacl o is growh
pah.
( Eurcpean Reztev cf Eccncmtc Htstcry
Table 1. Eccncmtc grcvth tn 1estern Eurcpe at dtfferent ttmes ;per
cent per year).
Period CDP CDP per captta
18o11 z.z 1.(
111o 1.( o.
1o1; (.8 (.o
1;1( z.1 1.;
Scurce: Ieinsein, et a/. 1;, p. ;. Iifeen counries, daa from Maddison (1).
The problem is he lengh of ime in Table 1. The business cycle can pro-
duce a pah lile his wih a ime span of a year or wo. New growh heory
and condiional convergence can produce a hisory lile his wih conver-
gence wih abou years or so o half-way convergence (Manliw, et a/.
1z). I is harder o find a good explanaion for his inermediae ime
frame, for quie complee convergence in abou hiry years. Business cycles
generally are a demand phenomenon (Temin 18), condiional conver-
gence involves supply phenomena. I would no be surprising if he expla-
naion of he inermediae case involved boh demand and supply.
While his simple model of deviaion froma smooh rend is appealing, I do
no wan o sugges ha he disequilibriumsudied here was he only phenom-
enon aling place afer he war. As noed in several sudies, several European
counries did no reurn o heir prewar rend pahs of growh afer he war or
evenafer he ColdenAge (Crafs andMills 16). Ior hose counries, he end
of he Colden Age represened a reurn o a more durable growh pah, bu no
necessarily he same one hey had experienced before he war. I reserve for
fuure worl he inegraion of he Colden Age and he subsequen growh pah.
In addiion o he ime-series quesions abou he beginning and end of he
Colden Age, here also is a cross-secion quesion: Why did some counries
grow so much more rapidly in his period han ohers? The spread of growh
raes among Wesern European counries in his period is shown in Table z.
Annual raes of growh varied fromwo o five per cen a year. I is a wide range
and needs o be explained. Naional hisories always conain developmens
ha can be used o explain rapid or slow growh, he more challenging ques-
ion is wheher here is a unified explanaion for he variey shown in Table z.
The Cc/den ge cf Eurcpean grcvth reccnstdered
Table z. nnua/ rates cf grcvth tn 1estern Eurcpe, Iyy;y ;per cent
per year).
Ausria ALT (.
Belgium BEL .
Swizerland CHE z.(
Cermany DEL .(
Denmarl DNK .z
Spain ESP .z
Iinland IIN .8
Irance IRA .
Crea Briain CBR z.1
Ireland IRL .z
Ialy ITA (.
Neherlands NLD .(
Norway NOR .
Porugal PRT .
Sweden SWE .o
Scurce: Penn World Tables .6.
An early conribuion o he lieraure on poswar growh was provided
by Kindleberger (16;) using he Lewis (1() model of excess labour
supply o explain boh differences in growh raes beween counries and he
slowdown in growh he could deec in he mid 16os. Kindlebergers argu-
men was simple: an elasic labour supply promoes economic growh by
leeping wages low and preserving indusrial peace. I was he exhausion of
cheap labour ha caused economic growh o slow. This aricle builds upon
and exends Kindlebergers view of hiry years ago.
The slowdown of growh in he 1;os, lnown a he ime as sagflaion,
was he subjec of myriad papers and bools. Many people argued ha
movemens in aggregae supply led o he slowdown of growh as well as
higher inflaion. The wo shocls mos ofen idenified were he rise in oil
prices in 1; and increasing rigidiy in indusrial labour marles (Bruno
and Sachs 18). The oil shocls have faded ino hisory while remaining he
mos popular candidaes for causing he end of he Colden Age.
Characerisics of he labour marle coninue o be acive opics in he
explanaion of European economic difficulies.
The focus on supply condiions led o new growh heory, which sressed
he role of supply in he long run. Solows frameworl had provided a way
o organise hisorical daa on economic growh. Populaion, invesmen and
TIP could be lised as deerminans of growh, and growh accouning was
born. This proved o be an enormously illuminaing way o summarise a
vas body of lnowledge and begin he process of explaining economic
growh (Solow and Temin 1;8, Criliches 16). Bu Solows growh
model did no include any oher variables, i could no accoun for he wide
differences beween counries ha we observe, and i prediced ha all
counries would converge o he same rae of growh. This limiaion led
people o lump all oher differences beween counries ino TIP and pro-
vide explanaions ouside he heory why hey differed (Denison 16;).
The limiaions of he Solow growh model were aacled in urn, giving
rise o new growh heory. Romer (186) argued ha TIP growh was
endogenous, no exogenous. Lucas (188) inroduced human capial o he
model as an addiional deerminan of growh, as had been done informally
in growh accouning and in economic hisory (Denison 16;, Easerlin
181). Differences in educaion beween counries eliminaed he predicion
of uncondiional convergence (ha is, convergence o he same rae of
growh by all counries), alhough hey sill lef room for condiional con-
vergence for groups of similar counries, someimes called convergence
clubs. Wide differences beween counries now could be explained wihin
he model by differences in educaional aainmen (Manliw, et a/. 1z).
New growh heories provided exensions o ge around he limiaions of
old growh heory a he expense of Solows simpliciy and elegance, edu-
caion is only he mos prominen of many puaive inpus o growh.
Empirical invesigaions flowered in he form of growh equaions, bu few
6 Eurcpean Reztev cf Eccncmtc Htstcry
of hese regressions aclnowledged anyhing special abou he Colden Age
of Economic Crowh. The regressions focused on idenifying he equilib-
rium growh rae o which counries were converging raher han esimaing
condiional convergence iself. The laer by he 1os was simply assumed
as a fac of economic life.
Baumol (186) provided evidence of convergence over a cenury for a
sample of mosly Wesern European counries. The claim ha his was a
universal paern did no sand up (De Long 188), and he field urned o
a prolonged invesigaion of he facors ha deermine o wha rae of
growh counries will converge. Crowh regressions ypically are done for as
many counries as possible, which means over a hundred in odays world.
The ime period chosen is much shorer han Baumols in order o exploi
he pleniful daa afer World War II. Two recen surveys of his lieraure
describe he diversiy of approaches alen o idenify convergence clubs,
bu hey do no remarl on any special reamen of he Colden Age of
Economic Crowh (Durlauf and Quah 18, Temple 1).
Barros Robbins Lecures, for example, were based on regressions for
periods sreching from 16 o 1o wih no aclnowledgmen ha he
process of growh migh be differen a he beginning and end of he period.
He commened ha his was an improvemen on his prior pracice of using
a single cross-secion, bu no because he daa came from wo separae
economic periods (Barro 1;, pp. 1z1). The common pracice sill is o
lump he poswar period ino one cross-secion, as done in Youngs famous
dissecion of economic growh among he Asian NICs and ]ones survey of
he world income disribuion (Young 1(, 1, ]ones 1;). The period
ypically sars in 16o Barro sared laer so he could use 16o CDP as
an insrumen boh o exploi easily available daa and o avoid he recov-
ery period jus afer he war. Dowricl and Nguyen (18) provide a soliary
excepion o his rule. They examine wheher he convergence found by
Baumol (186) coninued afer 1;, esing earlier informal resuls wih
growh equaions. The focus was on convergence raher han he rae of
growh.
Economic hisorians also have urned heir aenion o he poswar years.
Crafs and Toniolo opened a volume of essays on he period by assering,
he years 1o; winessed a unique episode in he hisory of European
modern economic growh (Crafs and Toniolo 16, p. ). They argued
ha rapid growh in his period was parly a consequence of slow growh in
he previous period, bu hey did no dwell on he mechanism of such a
reacion. The bool as a whole is a survey of he experience of abou a dozen
Wesern European counries during he Colden Age in a compaible
forma. This exercise of fiing diverse hisories ino a common mould, how-
ever, was overwhelmed by he srenghs of paricular issues in he debaes
abou individual counries, and he essays are quie diverse (Temin 1;).
One counry sudy ha anicipaed he approach here concepualised he
The Cc/den ge cf Eurcpean grcvth reccnstdered ;
Cerman 1trtschaftsvunder as a disequilibrium phenomenon. Dumle (1o)
argued ha greaer warime desrucion generaed faser poswar growh
and provided evidence for his proposiion in growh regressions for OECD
counries. His inquiry was in he spiri of Abramoviz (186), who assered
ha he desrucion of physical capial during he war was less imporan
han he mainenance of wha he called he social capabiliy of growh. I
ale my sar from Dumle, bu shif his emphasis and his sample.
Eichengreen (16) offered a synheic view in his conribuion o he
Crafs and Toniolo volume. Saring from he observaion ha growh in
he Colden Age was relaed o caching up and high invesmen, he asled
why invesmen was boh high and producive in he Colden Age. He
answered ha wage moderaion and expor growh made invesmen arac-
ive and profiable. These in urn were due o governmen insiuions and
policies ha were sharply differen from hose pursued before he war.
Eichengreen saw an implici bargain beween worlers and invesors ha is
similar o he implici conracs Aoli (188) described in wha he called he
]-firm, ypical of poswar ]apan. The bargain was ha worlers would no
push for higher wages if invesors would male producive invesmens ha
would, over ime, creae jobs and raise wages. Invesors would agree o
inves on he condiion ha he worlers did no immediaely ry o ale all
he gains in higher wages.
This bargain is ime-inconsisen. If worlers moderae wage demands,
invesors have an incenive o pay hemselves he resuling profis insead of
reinvesing hem. And if invesors male producive invesmens ha
enhance he produciviy of labour, worlers have he incenive o ale he
gains home in he form of higher wages. These perverse incenives were
counered in poswar Europe by a complex se of insiuions ha made
reneging harder and increased incenives for honouring he long-erm
implici conrac in he face of shor-run gains from abrogaing he conrac.
The insiuions were boh domesic and inernaional, domesic o enforce
he bargain jus described, inernaional o promoe naional specializaion
ha increased efficiency. The domesic insiuions included naional wage
bargaining, union represenaion on company boards, and condiional
access o governmen programmes. The inernaional ones included insi-
uions lile CATT, ECSC and EPL ha appeared o have had lile posi-
ive effec. Eichengreen emphasised heir role in precluding negaive effecs,
assuring ha rade would remain free as condiions changed.
This is an inriguing and plausible hypohesis, i explains how demand
could grow o promoe rapid economic growh during he Colden Age. Bu
i canno explain how Wesern Europe found iself so far from equilibrium
a he sar of he Colden Age. This organizaional view also does no dis-
inguish beween differen counries in Wesern Europe because he iner-
naional agreemens ha formsuch a large par of he sory include hemall.
Eichengreen lised many causes for he end of he Colden Age, revealing he
8 Eurcpean Reztev cf Eccncmtc Htstcry
absence of a unified explanaion. Among he reasons offered were he cap-
ure of insiuions by firms and unions, he oil shocls, he end of he
Breon Woods Sysem, he end of general cach-up, and reduced incenives
o leep he bargains ha produced he Colden Age of Economic Crowh.
z. Recent theories oI econonic growth
I approach he Colden Age in he conex of economic growh over he pas
cenury or wo, which had a large componen of economic ransiion.
Naional economies around 18oo wih very few excepions were almos
compleely agriculural. Saring in he nineeenh cenury and even laer,
producive resources were moved ou of agriculure ino manufacuring and
services. Residens became urban, and he share of he labour force in agri-
culure fell. Since worlers were more producive in non-agriculural acivi-
ies, naional income grew during his ransiion.
Theoriss of economic growh recenly have begun o aclnowledge he
imporance of his ransiion in he process of economic growh. There are
now several models aemping o inegrae srucural shifs wih he heory
of economic growh (Kongsamu, et a/. 1;, Temple and Voh 18, Calor
and Weil zooo). Taylor (1) used a model of his ype in his exploraion
of convergence in seven counries before World War I. All of hese papers
share wih his one he aemp o bring he hisorical experience of indus-
rialisaion ino he mainsream of hinling abou economic growh.
Broadberry (1;) evaluaed he imporance of his ransiion in
Cermanys convergence o Briish levels of labour produciviy. The firs
column of Table shows his esimae of aggregae labour produciviy in
Cermany compared o he Lnied Kingdom. The familiar rise over he las
cenury can be seen, wih a dip in 1o jus afer he Second World War.
The second column of Table reveals ha he rise in comparaive labour
produciviy in manufacuring did no echo he rise in he aggregae. In fac,
here is very lile evidence of a rend a all. Cerman relaive labour pro-
duciviy was as high in 11 as i would ge, and he emporary decline in
1o was eliminaed by 16o when i sood a 11 (Broadberry 1;, p. z1).
Cach-up, Broadberry assers, is no he resul of improving efficiency in
manufacuring, bu he resul of ransferring resources from low-pro-
duciviy secors lile agriculure o high-produciviy ones lile manufacur-
ing.
I follows ha faser economic growh in Cermany han in Briain was
due largely o he more rapid secoral shifs in he Cerman economy.
Cermany had a larger share of is labour force in agriculure han he Lnied
Kingdom hroughou he pas cenury. In 1o, around he sar of he
Colden Age of European Crowh, Cermany had z( per cen of is labour
force in agriculure, compared o five per cen for he Lnied Kingdom
(Broadberry 1;, p. zz). If rapid economic growh is he resul of he
The Cc/den ge cf Eurcpean grcvth reccnstdered
ransiion from an agrarian economy, hen Cermany was sill engaged in he
process during he Colden Age while Briain had compleed is ransiion.
Why was Cermany lagging behind Briain in his ransiion? Three
reasons come o mind, of which he hird has no been appreciaed. Iirs,
Cermany sared is indusrialisaion afer Briain. Second, Cermany chose
o proec is farmers agains low-priced American grain in he lae nine-
eenh cenury. Third, he Second Thiry Years War he urbulen period
from 11 o 1( inerruped inernaional rade and slowed he ran-
siion. The hird of hese reasons has been negleced, I wan o expose is
imporance.
The growing lieraure on globalisaion argues ha i has ebbed and
flowed in he course of he wenieh cenury. Before he Crea War, iner-
naional commerce and ravel were free and open, more or less as hey are
oday. Bu in beween hese wo end poins, he flow of goods, finance, and
people was inerruped by world wars and depression. Auhors disagree
among hemselves abou wheher odays globalisaion acually exised a
cenury ago, bu here is no disagreemen abou he inerrupion during he
world wars and Crea Depression (Bordo et a/. 1, Obsfeld and Taylor
1, Temin 1).
Inernaional rade was inerruped by he Iirs World War. The poswar
selemen creaed many new boundaries ha provided he opporuniy o
impose ariffs on rade. And he Crea Depression led o resricive rade
policies ha reversed whaever expansion had alen place in he 1zos. The
volume of expors for he major Wesern European counries was lower in
18 han i had been in 11, in sharp conras o is rapid growh boh
before and afer his period (Ieinsein, et a/. 1;, p. 1o).
Sachs and Warner (1) argued ha rade promoed economic growh
in he poswar world. Their regressions showed ha closed economies did
no exhibi convergence, while open economies did. Why did closed
economies suffer? Because hey did no underale he reallocaion of
resources needed o increase produciviy. They could no exploi heir
comparaive advanages, and hey could no end heir reliance on domesic
1o Eurcpean Reztev cf Eccncmtc Htstcry
Table . Ccmparattze /abcur prcducttztty tn Cermany and the Untted
Ktngdcm ;UK Ioo).
Year CDP Manufacuring
18;o 6o
11 ;8 11
1o 66 6
1; 11z 11
18 116 1o
Scurce: Broadberry 1;, p. zo.
agriculure. Open economies decreased he proporion of food and raw
maerials in heir expors more rapidly han closed economies. Sachs and
Warner did no dwell on he connecion beween heir heory and he his-
ory of indusrialisaion in Europe during a previous period, bu he paral-
lel is clear. Before World War I, paricipaion in inernaional rade
promoed indusrialisaion. One has only o recall he discussion of Briains
climaceric in he lae nineeenh cenury o see he imporance of iner-
naional rade in economic growh. Briain was surpassed, a prominen
sory assers, because he Lnied Saes and Cermany were beer able o
exploi world marles (Temin 166).
I follows from his view ha he barriers o inernaional commerce
during he world wars and Crea Depression consiued barriers o he con-
inued indusrialisaion of European counries. This slowdown in he
process of indusrialisaion creaed a disequilibrium afer he war. As
suggesed by Table 1, he supply fronier coninued o expand during he
Second Thiry Years War. The Lnied Saes, insulaed from he wars if no
he Depression, was able o coninue is ransformaion from an agriculural
o an indusrial economy. Is expors were primarily food and raw maerials
before his proraced conflic, hey were manufacures aferwards (Irwin
16). European counries emerged from he war wih a developmenal
defici.
This disequilibrium is separae from he low income ha generaes con-
diional convergence. Low income in he sandard sory is produced by low
levels of physical and human capial relaive o saving raes. The develop-
menal defici highlighed here is produced by a misallocaion of resources.
The firs ales place in a single-secor economy, he second, in a disaggre-
gaed model of developmen.
The rae a which worlers lef agriculure acceleraed afer he war. The
decline in he share of he labour force in agriculure was wice as rapid in
he 1os and 16os as before. The variance of he measured change fell as
he rae increased, wheher because of he greaer sabiliy of Wesern
Europe or because of noise in he imperfec earlier daa. The sandard devi-
aion of he decadal rae of change in he share before World War II was
hree imes as large as he sandard deviaion of he quinquennial change
hereafer. As a resul, changes from before World War I o he inerwar
period are los in his volailiy (Bairoch 168, as quoed in Michell 18,
pp. 1(6z).
The misallocaion of resources can be measured by he share of he
labour force in agriculure. There are many ways o divide up he economy,
bu he division beween agriculure and all oher aciviies appears o be he
mos imporan. Broadberry (1;) disinguished nine secors of he econ-
omy, bu he concluded ha mos of he effec came from he changing size
of agriculure. Denison (16;), much earlier, alled of he misallocaion of
resources in Europe during he Colden Age of European Crowh, and he
The Cc/den ge cf Eurcpean grcvth reccnstdered 11
oo mean he European counries were growing rapidly when hey were
geing ou of agriculure.
The misallocaion, according o his sory, came from a generaion
hiry years of economic insulariy. I is reasonable o hinl ha he
excessive resources in agriculure could be moved o oher secors in
anoher hiry years. This hypohesis herefore provides a way o raionalise
he hisory shown in Table 1. Warime desrucion was lile a business cycle
in leading o a shor-erm disequilibrium. Condiional convergence migh
explain long-erm disequilibria. Auarchy in he Second Thiry Years War
can explain a disequilibrium ha can be eliminaed in weny or hiry years.
This phenomenon may be more general han Europe afer he Second
World War. ]ones (1;) concepualised growh as a lind of Marlov
process. Counries drew heir rae of growh from an urn once every hiry
years or so, drawing fas, slow or medium growh raes. ]ones characerised
he fas growh as growh miracles and assered ha hey were mos preva-
len among poorer counries, alhough no among he poores. Young
(1() showed ha hese growh miracles were accomplished by very high
invesmen raes. They also were accomplished by rapid reducions in he
size of agriculure in hese counries.
. Testing the hypothesis
I es his hypohesis by formalising he sory in a simple model and esing
i agains daa from he Colden Age of European Crowh. The model dis-
inguishes hree linds of disequilibria ha can affec growh:
(1) Condiional convergence, ha is, saring from a level of income
low relaive o he counrys equilibrium income.
(z) Warime desrucion ha deranges producion in he shor run.
Dumle (1o) measured he exen of his dislocaion by he
percenage gap beween per captta CDP in 1(8 and in 18. I
use his measure here as well, labelling i CAP, and recalcula-
ing i from Maddison (1).
() Arresed developmen, ha is, excessive labour in agriculure. In
parallel wih condiional convergence, his phenomenon will be
measured by he difference beween he iniial proporion of he
labour force in agriculure, A, and he equilibrium share, A*.
The model hen is as follows, where g is he average growh rae of y, per
captta CDP.
g a b(y* y) c CAP d (A A*) e (1)
This regression, despie is convenional appearance, differs from growh
regressions in he lieraure. Those growh regressions are designed o elici
differences beween y* in differen counries. Crowh is regressed on cur-
1z Eurcpean Reztev cf Eccncmtc Htstcry
ren income and many variables, lile educaion, ha proxy for and idenify
y*. I assume here ha y* is he same for all counries in Wesern Europe.
This emphasis is appropriae in a sudy of a single region and in he inves-
igaion of disequilibrium growh during he Colden Age of Economic
Crowh. I am rying o describe he process of convergence, while growh
regressions ypically assume ha counries are near heir growh pah and
invesigae he naure of he equilibrium income (y*) o which hey are con-
verging.
I also assume ha he equilibrium share of agriculure, A*, is he same
for all Wesern European counries. The influences of geography, hisory,
and he Common Agriculural Policy are alen o be second-order effecs.
One could no male his heroic assumpion wih a wider sample, bu i is
appropriae when discussing economic growh in Wesern Europe. The
share of he labour force in agriculure is measured a he beginning of each
period, so ha i is a predeermined variable. Differences beween counries
will show up in he error erm and in he goodness of fi.
Equaion (1) can be rewrien, collecing he unobserved equilibrium
levels wih he consan erm.
g (a by* dA*) by c CAP dA e (z)
I esimae his equaion for all 1 Wesern European counries afer he
Second World War. (This is he same se of counries whose growh is
repored in Table 1, excep ha Czechoslovalia has been replaced by
Porugal.) They all are par of he same convergence club, harling bacl o
he origins of new growh heory. They all have sable governmens, secure
propery righs, and universal educaion, and i is reasonable o argue ha
A* is vanishingly small in Wesern Europe oday.
I use his regression o es hree hypoheses. Iirs, all hree linds of dis-
equilibrium affeced he rae of growh in differen counries in he Colden
Age of European Crowh. The presence of he second wo disequilibria
accoun for he rapidiy of economic growh during he Colden Age.
Second, hese disequilibria had heir main impac a differen periods.
Warime desrucion affeced growh in he immediae poswar years, hen
resource misallocaion became imporan, and finally condiional conver-
gence became cenral. In normal periods of peace, only he las of hese is
relevan. The special, hisorical qualiy of he Colden Age comes from he
presence of he firs wo. The susained rapid growh comes from he pres-
ence of he second disequilibrium, resource misallocaion.
Third, warime desrucion and he misallocaion of resources ceased o
have an effec on growh some ime around 1;o. In oher words, he end
of he Colden Age came when hese unusual disequilibria were removed.
We canno observe hese effecs a oher imes because hey are no here,
hey were presen in poswar Europe only because of he hisorical circum-
sances ha were unique o his ime and place.
The Cc/den ge cf Eurcpean grcvth reccnstdered 1
The regressions are shown in Tables ( and . The firs of hese ables
shows regressions for weny-year inervals which correspond closely o he
Colden Age, he second, for en-year inervals, following Barro (1;). The
daa sar in 1o because daa for 1( are unavailable and unreliable. The
resuls in Table ( are useful for looling a he Colden Age as a whole, hose
in Table , for racling he effec of differen disequilibria wihin he Colden
Age.
The firs hing o noe is ha he regressions accoun for a subsanial
amoun of he variaion in growh raes among he Wesern European coun-
1( Eurcpean Reztev cf Eccncmtc Htstcry
Table (. Regresstcns exp/atntng tventy-year grcvth rates.
Variable 1o;o 1; 168
Y
o
o.oo61 o.oz8 o.ozz
(o.z) (1.) (z.o)
CAP o.o1;z o.oozo( o.oozz
(1.8() (o.z) (o.8)
A o.o; o.o8 o.oo(
(z.;) (z.(() (o.z6)
Consan z.8 .;6 (.o;
(z.o1) ((.o8) (.8()
Adjused R
z
o.;o o.; o.
N 1. 1. 1.
Scurces: Poswar CDP per captta daa from he Penn World Table .6, CAP per captta
CDP in 1(8 over per capia CDP in 18, from Maddison (1), A from Inernaional
Labor Office (186), IAOSTAT (on he web), and Social Indicaors of Developmen (on
he web). T-saisics are below he coefficiens.
Table . Regresstcns exp/atntng decada/ grcvth rates.
Variable 1o6o 16 16; 1;8 181
Y
o
o.oz; o.o1;1 o.o( o.oo1( o.o1
(1.oz) (o.8o) (.61) (o.o8) (z.o)
CAP o.o( o.o16 o.oo6(( o.o111 o.oo(;
(.zo) (1.(8) (o.8) (o.8) (o.(6)
A o.o8;z o.o;1 o.o16 o.oo66z o.o18o
(z.) (z.() (o.6() (o.11) (o.1)
Consan o.16z 1.; 6.oz z.zz .(
(o.o8) (o.8() ((.;) (o.8) (z.6;)
Adjused R
z
o.6 o.(z o.8o o.1; o.8
N 1. 1. 1. 1. 1.
Scurces: See Table (. Preliminary daa for 1 from Rober Summers by privae correspon-
dence, ; March 1. T-saisics are in parenheses below he coefficiens.
ries during he Colden Age, despie he drasic simplificaions of he
model. The assumpion ha y* and A* are he same for all counries in
Wesern Europe afer he Second World War does no appear misleading.
The second hing o noe is ha inferences from he sandard errors are
slighly problemaical in his conex, as for many growh regressions. These
counries are no a random sample of a larger se, hey are all he counries
of Wesern Europe. The years also are no a random sample from a larger
se, hey are he years of he Colden Age of European Crowh. An inference
based on a t-es would be wrong in only per cen of cases if we could find
similar ses of counries afer similarly long periods of war and economic
auarly o sudy.
The firs wo columns of Table ( show regressions for he Colden Age.
The effec of excessive labour in agriculure (A) is visible clearly. This indi-
caes ha he misallocaion of resources arising from he limiaion of iner-
naional rade during he Second Thiry Years War was cosly o Wesern
Europe. I also creaed he opporuniy for many counries o grow rapidly
as hey allocaed heir resources more efficienly. The effec was large. The
share of Cermanys labour force in agriculure a he sar of he Colden Age
was zo percenage poins higher han he share of Briains agriculural
labour force. Cermanys growh rae received a boos of approximaely one
percenage poin from his iniial condiion. (The effec is 1.z per cen if
1o;o is used, o.8 per cen if 1; is used.) The acual difference in
growh raes in 1; was 1. per cen as shown in Table z, his effec
explains over half he difference for hese years.
Neiher warime desrucion nor long-run condiional convergence were
as imporan as he misallocaion of resources in explaining differences
beween growh raes in Wesern Europe a his ime. Dumle (1o) found
warime desrucion o be more imporan for wo reasons. Iirs, he lumped
he enire Colden Age ino one regression, and he did no racl effecs
wihin he period. He also included ]apan in his sample, which lile
Cermany had exensive warime damage. The presence of wo heavily
damaged counries was enough o male warime desrucion appear
imporan.
The firs column of Table conains regression for he firs poswar
decade for which we have reliable daa. Warime desrucion (CAP) was
imporan in deermining he rae of growh of differen counries in his
period, as was he misallocaion of resources (A). Condiional conver-
gence, by conras, was no an imporan facor. Succeeding columns show
ha he influence of warime desrucion faded rapidly ino he bacl-
ground. I was only an imporan facor a he sar of he Colden Age. The
effec of he agriculural labour force faded away as well, bu more slowly.
As he share of he labour force in agriculure approached is equilibrium
level, he esimaed effec became smaller, boh absoluely and relaive o
is sandard error. The decreasing size of he coefficien over ime suggess
The Cc/den ge cf Eurcpean grcvth reccnstdered 1
a non-lineariy, perhaps an S-shaped relaion beween he excess share of
labour in agriculure and economic growh.
The regression for 1;8 is marledly worse han hose for oher
decades. No coefficiens are esimaed ighly, and he share of variance of
growh raes explained is negligible. This decade, of course, conained he
effecs of wo oil shocls which disruped he longer-run processes of ineres
here. The presence of hese disrupions males i hard o idenify iming in
longer-run phenomena. In he view I am presening here, hese disrupions
confuse he sory, bu hey are no he sory iself. The end of he Colden
Age was brough abou by he reurn o equilibrium growh pahs on he
par of hese counries, no by adverse economic shocls in he 1;os. The
regression for 18 in he final column of Table shows ha condiional
convergence as indicaed by a significan coefficien of y
o
is verified, while
he oher forces presen in he Colden Age do no show up.
Alhough i is hard o es for robusness wih only 1 observaions, vari-
ous experimens show hese equaions o be robus. The resuls differ only
in deails if log(y
o
) is used in place of y
o
o indicae general convergence, and
inermediae en and weny year ime periods yield inermediae resuls o
hose shown here. Truncaing CAP o be less han zero also does no
change he resuls. Lagging A by five years o avoid any hin of simulane-
iy preserves he resuls in Tables ( and . And he resuls are unchanged if
one or anoher small counry is dropped from he sample, or even if wo
ouliers lile Spain and Porugal are dropped ogeher.
Scaer diagrams illusrae he regression resuls. Iigures 1 and z show
he relaion beween he rae of growh and he iniial share of he labour
force in agriculure for wo overlapping zo-year periods saring in he
1os. In boh diagrams, he correlaion shown in Table ( is quie apparen.
While he diagrams are similar, here are a few differences worh noing.
Iigure 1, showing 1o;o, reveals Cermany (DEL) o be an oulier. The
1trtschaftsvunder appears clearly. Bu in Iigure z, showing 1;,
Cermany no longer is an oulier. Insead, Belgium lools o be an economic
miracle. We need o be careful in our claims of special condiions in indi-
vidual counries.
Briain is a he low end of European growh raes, as everyone lnows.
This model provides an explanaion for his slow growh a variance wih
ha in he lieraure (Bean and Crafs 16). Ior he slow Briish growh is
explained quie fully by he low share of he labour force in agriculure a
he sar of he Colden Age. Briain had sared indusrialisaion wih a low
share of labour in agriculure, i indusrialised firs, and i lep ariffs low
when American grain was able o be ranspored cheaply o Europe in he
lae nineeenh cenury. These pas accomplishmens implied slow growh
during he Colden Age.
I follows ha poswar Briish economic policies were no he cause of
slow economic growh. I may no be oo farfeched o say ha 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, Iyo;o.
Iigure z. Regresstcn cf eccncmtc grcvth rate and tnttta/ share cf /abcur
fcrce tn agrtcu/ture, Iyy;y.
cies were he resul of slow growh raher han is cause. Bean and Crafs
(16) offer a muli-layered picure of Briish policies, in which Briains
iniial posiion plays only a minor role. This view suggess ha Briish econ-
omic policies did no have he poenial effecs aribued o hem in his
lind of accoun. And i suggess even furher ha wha we regard as poor
policies because Briain grew so slowly were eiher reasonable adap-
aions o Briains iniial posiion or he resuls of Briains slow growh.
The Colden Age of Economic Crowh ended when his disequilibrium
was eliminaed, ha is, when he share of he labour force in agriculure
approached is equilibrium level. This can be seen in he regressions in
Tables ( and , where he coefficien of A is no esimaed clearly in he laer
regressions. I can be seen also in he rae of change of he share of labour
in agriculure. If one examines he change in he labour force in agriculure,
as in equaion (1), hen he rae of change falls over ime, ha is, becomes
less negaive, as shown in he firs row of Table 6. This is he variable ha
is relevan for growh, since i is he movemen of people ha generaes
growh. If, however, one examines he rae of change of he agriculure
share, ha is, he change in he share divided by he iniial share, hen here
is no change over ime. The share approaches is asympoe a a consan
rae.
The regressions in Table 6 help us o undersand why his is a hisorical
explanaion. In peaceful imes, he share of labour in agriculure falls as
income rises. The change in he agriculural labour force ges smaller (more
posiive) over ime, bu here is no evidence of a change in he proporion-
ae rae of decline. The evidence is consisen wih an asympoic approach
o an equilibrium share. Various facors and policies can inhibi his change
in individual counries, bu here are no oher hisorical cases where large
numbers of indusrial counries were in he same disequilibrium posiion.
The model of equaion (1) may be he correc model for all ime, bu CAP
18 Eurcpean Reztev cf Eccncmtc Htstcry
Table 6. Trends cf the agrtcu/tura/ /abcur share.
Variable Change in he share Crowh in he share
(1) (z) ()
Trend o.(z6 o.6 o.ooz(;
(1z.() ((.o8) (1.)
Trend squared o.oz1
(1.8)
Consan (.; .1 o.1;
(1z.8) (11.1;) (16.oo)
Overall R
z
o. o.( o.o1
N 1 1 1
Scurce: See Table (. Z-saisics in below he coefficiens.
and (A-A*) ypically are a or near zero. I is impossible o esimae heir
coefficiens under hese circumsances. The final regressions in Tables (
and illusrae he inabiliy o esimae hese coefficiens ouside he
Colden Age.
q. Conclusion
I reurn now o he hree quesions posed a he beginning of his aricle.
Iirs, why was growh so rapid in Wesern Europe during he Colden Age
of Crowh? The answer is disequilibrium. The normal cach-up ha worls
in general was no imporan righ afer he war, bu oher linds of disequi-
librium were. The mos imporan of hese was he misallocaion of
resources ha came from he lacl of inernaional rade during he preced-
ing hiry years. In his sae of arresed indusrialisaion, oo many
resources sill were employed inefficienly in agriculure. The insiuional
facors cied by Eichengreen (16) helped creae he needed demand, real-
locaed labour rapidly enhanced he supply.
Second, why did differen counries grow a differen raes during he
Colden Age? They grew a differen raes during he Colden Age because
of heir iniial posiion. Naional policies had secondary effecs relaive o
counries iniial posiion. Labour relaions were umuluous in boh Ialy
and Briain, bu Ialy grew rapidly while Briain did no. A he leas,
policies of he ime have been blamed for evens ouside he conrol of
policymalers. In addiion, some of he bierness of policy in slowly-
growing Briain may have been he resul of slow growh insead of is
cause.
Third, why did he rapid growh come o an end? Crowh slowed in he
1;os and 18os because he disequilibrium ha had generaed unusually
rapid growh no longer exised. The developmenal defici of a generaion
was eliminaed in a generaion. The Common Agriculural Policy may have
lep some excessive labour in agriculure, bu he misallocaion of resources
had ceased o be a large macroeconomic issue by he ime of he oil crises.
These crises muddied he hisorical waers, confusing shor-run and long-
run facors. I is only wih he hindsigh of anoher hiry years ha we can
see ha while he oil crises were disrupive, he slowdown of growh would
have alen place even if hey had no occurred.
Acknowledgenents
This aricle is a revised version of he Hicls Lecure, Oxford Lniversiy, 1z March,
1. I hanl ]ohnny Chen, Bural Cuner, Dirl Niepel, and Ayalo Tanala for
research assisance, and audiences a he lecure and oher seminars for helpful
commens. Zvi Criliches was lind enough and well enough o advise me in he early
sages of his aricle. All errors are mine alone.
The Cc/den ge cf Eurcpean grcvth reccnstdered 1
ReIerences
ABRAMOVITZ, M. (186). Caching up, forging ahead, and falling behind. )curna/
cf Eccncmtc Htstcry q6, pp. 8(o6.
AOKI, M. (188). Infcrmattcn, Incenttzes, and Eargatntng tn the )apanese Eccncmy.
Cambridge: Cambridge Lniversiy Press.
BAIROCH, P. et a/. (168). The 1cr/tng Pcpu/attcn and Its Structure. Brussels:
Insiu de Sociologie, Lniversie Libre de Bruxelles.
BARRO, R. ]. (1;). Determtnants cf Eccncmtc Crcvth. Cambridge, MA: MIT
Press.
BALMOL, W. (186). Produciviy growh, convergence and welfare. mertcan
Eccncmtc Reztev y6, pp. 1o;z8.
BEAN, C. and CRAITS, N. (16). Briish economic growh since 1(. In
Nicholas Crafs and Cianni Toniolo (eds), Eccncmtc Crcvth tn Eurcpe Stnce
Iy. Cambridge: Cambridge Lniversiy Press.
BERNANKE, B. (1). The macroeconomics of he Crea Depression: a
comparaive approach. )curna/ cf Mcney, Credtt and Ean/tng zy, pp. 1z8.
BORDO, M. D., EICHENCREEN, B. and IRWIN, D. A. (1). Is globalizaion oday
really differen han globalizaion a hundred years ago? Ercc/tngs Trade Icrum.
Washingon DC: The Broolings Insiuion, pp. 1o.
BROADBERRY, S. N. (1;). Anglo-Cerman produciviy differences, 18;o1o.
Eurcpean Reztev cf Eccncmtc Htstcry, t: pp. z(;6;.
BRLNO, M. and SACHS, ]. (18). Eccncmtcs cf 1cr/dvtde Stagf/attcn. Cambridge,
MA: Harvard Lniversiy Press.
CRAITS, N. and MILLS, T. C. (16). Europes Colden Age: an economeric
invesigaion of changing rend raes of growh. In Bar van Arl and Nicholas
Crafs (eds), Quantttattze spects cf Pcst-1ar Eurcpean Eccncmtc Crcvth.
Cambridge: Cambridge Lniversiy Press.
CRAITS, N. and TONIOLO, C. (eds), (16). Eccncmtc Crcvth tn Eurcpe stnce Iy.
Cambridge: Cambridge Lniversiy Press.
DE LONC, ]. B. (188). Produciviy growh, convergence, and welfare: commen.
mertcan Eccncmtc Reztev y8, pp. 118(.
DENISON, E. (16;). 1hy Crcvth Rates Dtffer. Washingon, DC: Broolings
Insiuion.
DOWRICK, S. and NCLYEN, D.-T. (18). OECD comparaive economic growh
1o8: cach-up and convergence. mertcan Eccncmtc Reztev y,
pp. 1o1oo.
DLMKE, R. (1o). Reassessing he 1trtschaftsvunder: reconsrucion and poswar
growh in Wes Cermany in an inernaional conex. Oxfcrd Eu//ettn cf
Eccncmtcs and Stattsttcs z, pp. (11.
DLRLALI, S. N. and QLAH, D. T. (18). The new empirics of economic growh.
Worling Paper no. 6(zz, Naional Bureau of Economic Research.
EASTERLIN, R. A. (181). Why isn he whole world developed? )curna/ cf
Eccncmtc Htstcry qt, pp. 11.
EICHENCREEN, B. (16). Insiuions and Economic Crowh in Europe Afer
World War II. In Nicholas Crafs and Cianni Toniolo (eds), Eccncmtc Crcvth
tn Eurcpe Stnce Iy. Cambridge: Cambridge Lniversiy Press.
IEINSTEIN, C. H., TEMIN, P. and TONIOLO, C. (1(). Three shocls, wo
zo Eurcpean Reztev cf Eccncmtc Htstcry
recoveries: hisorical parallels for he end of he Cold War. Rtztsta dt Stcrta
Eccncmtca tt, pp. z;16.
IEINSTEIN, C. H., TEMIN, P. and TONIOLO, C. (1;). The Eurcpean Eccncmy
Eetveen the 1ars. Oxford: Oxford Lniversiy Press.
CALOR, O. and WEIL, D. N. (zooo). Populaion, echnology and growh: from
Malhusian sagnaion o he demographic ransiion and beyond. mertcan
Eccncmtc Reztev o, pp. 8o6z8.
CRILICHES, Z. (16). The discovery of he residual: a hisorical noe. )curna/ cf
Eccncmtc Ltterature q, pp. 1z(o.
INTERNATIONAL LABOR OIIICE (186). Eccncmtca//y cttze Pcpu/attcn Esttmates
and Prcjecttcns, Iyo:o:y, vol. (. Ceneva: ILO.
IRWIN, D. A., (16). The Lnied Saes in a new global economy? A cenurys
perspecive. mertcan Eccncmtc Reztev, Papers and Prcceedtngs 86, pp. (16.
]ONES, C. I. (1;). On he evoluion of he world income disribuion. )curna/ cf
Eccncmtc Perspecttzes tt, pp. 16.
KINDLEBERCER, C. P. (16;). Eurcpes Pcstvar Crcvth. The Rc/e cf Labcr Supp/y.
Cambridge, MA: Harvard Lniversiy Press.
KONCSAMLT, P., REBELO, S. and XIE, D. (1;). Beyond balanced growh.
Worling Paper no. 61, Naional Bureau of Economic Research.
LEWIS, W. A. (1(). Developmen wih unlimied supplies of labor. Manchester
Schcc/ zz, pp. 11.
LLCAS, R. E., ]r. (188). On he mechanics of economic developmen. )curna/ cf
Mcnetary Eccncmtcs zz, pp. (z.
MADDISON, A. (1). Mcnttcrtng the 1cr/d Eccncmy, I8:oI:. Paris: OECD.
MANKIW, N., CRECORY, D. R. and WEIL, D. N. (1z). A conribuion o he
empirics of growh. Quarter/y )curna/ cf Eccncmtcs, toy, pp. (o;;.
MITCHELL, B. R. (18). Internattcna/ Htstcrtca/ Stattsttcs. Eurcpe, I;yoI.
London: Soclon Press.
OBSTIELD, M. and TAYLOR, A. M. (1). C/cba/ Captta/ Mar/ets. Integrattcn,
Crtsts, and Crcvth. New Yorl: Cambridge Lniversiy Press.
ROMER, P. (186). Increasing reurns and long-run growh. )curna/ cf Pc/tttca/
Eccncmy q, pp. 1ooz.
SACHS, ]. D. and WARNER, A. (1). Economic reform and he process of global
inegraion. Ercc/tngs Papers cn Eccncmtc cttztty, pp. 1118.
SAMLELSON, P. A. (1(). Iull employmen afer he war. In S. E. Harris (ed.),
Pcstvar Eccncmtcs Prcb/ems. New Yorl: McCraw-Hill.
SOLOW, R. M. (1;). Technical change and he aggregae producion funcion.
Reztev cf Eccncmtcs and Stattsttcs , pp. 1zzo.
SOLOW, R. M., and TEMIN, P. (1;8). Inroducion: The inpus for growh. In
The Cambrtdge Eccncmtc Htstcry cf Eurcpe, vol VII: The Industrta/ Eccncmtes.
Captta/, Labcur, and Enterprtse. Cambridge: Cambridge Lniversiy Press.
TAYLOR, A. N. (1). Sources of convergence in he lae nineeenh cenury.
Eurcpean Eccncmtc Reztev q, pp. 16z1(.
TEMIN, P. (166). The relaive decline of he Briish seel indusry, 188o11. In
Henry Rosovsly (ed.), Industrta/tzattcn tn Tvc Systems. New Yorl: Wiley.
TEMIN, P. (1;). The Colden Age of European growh: a review essay. Eurcpean
Reztev cf Eccncmtc Htstcry t, pp. 1z;(.
TEMIN, P. (18). The causes of American business cycles: an essay in economic
The Cc/den ge cf Eurcpean grcvth reccnstdered z1
hisoriography. In ]effrey C. Iuhrer and Sco Schuh (eds), Eeycnd Shcc/s.
Boson, MA: Iederal Reserve Banl.
TEMIN, P. (1). Clobalizaion. Oxfcrd Reztev cf Eccncmtc Pc/tcy t, pp. ;68.
TEMPLE, ]. (1). The new growh evidence. )curna/ cf Eccncmtc Ltterature y,
pp. 11z6.
TEMPLE, ]. and VOTH, H.-]. (18). Human capial, equipmen invesmen, and
indusrializaion. Eurcpean Eccncmtc Reztev qz, pp. 1(6z.
YOLNC, A. (1(). Lessons from he Eas Asian NICs: a conrarian view.
Eurcpean Eccncmtc Reztev 8, pp. 6(;.
YOLNC, A. (1). The yranny of numbers: confroning he saisical realiies of
he Eas Asian growh experience. Quarter/y )curna/ cf Eccncmtcs tto,
pp. 6(168o.
Internet scurces
Penn World Tables .6: hp://arcadia.chass.uorono.ca/pw/.
IAOSTAT Saisics Daabase:
hp://apps.fao.org/limoo/nph-wrap.pl?Populaion&DomainSLA.
Social Indicaors of Developmen: hp://www.ciesin.org/mep-bin/
charloe?saeSTART&evensar&proocolsid&charloe_dirprod.
zz Eurcpean Reztev cf Eccncmtc Htstcry

Das könnte Ihnen auch gefallen