COMMERCIAL REVOLUTION, ANONYMITY
AND TECHNOLOGICAL PROGRESS
Stephen Broadberry, Sayantan Ghosal and Eugenio Proto
July 2008 (Preliminary)
Abstract
We provide a model of the links between commercialisation and technological
progress, which is consistent with the historical evidence and places market relations
at the heart of the industrial revolution. First, commercialisation as a growing reliance
on impersonal labour market transactions in place of customary relations with a high
degree of monitoring led to the adoption of e¢ ciency wages. Second, the resulting
rise in the wage led to the adoption of a more capital-intensive technology. Third,
this led to a faster rate of technological progress through greater learning by doing on
the capital intensive production technology. An empirical analysis showing a positive
and signi…cant relationship between the lagged degree of urbanization and wages in
the period between 1300 and 1750 lends support to our theory.
JEL classi…cation: N13, O14, O43
Keywords: Commercialisation, Anonymity e¢ ciency wages, technological progress
1
Introduction
A long tradition in economic history links the transition to modern economic growth to the
widespread commercialisation of northwest Europe between the late medieval period and
Department of Economics, University of Warwick, Coventry CV4 7AL, United Kingdom
[email protected] [email protected] [email protected]
1
the Industrial Revolution (Toynbee, 1884; Polanyi, 1944; Britnell and Campbell, 1995).
However the precise nature of the links has remained obscure. We argue in this paper that
the growing commercialisation of the late medieval and early modern periods led to the
technological progress of the Industrial Revolution via its e¤ects on factor prices.
We begin, therefore, with a de…nition of commercialisation, which we take to mean
more than simply an increase in the proportion of output passing through the market
(Britnell and Campbell, 1995: 1). Commercialisation is seen here as a growing integration
of markets, where individuals exchange goods competing with a larger number of other
individuals. The commercial revolution thus generated the need for a point of geographical
aggregation of the economic agents, which resulted in the urbanization process witnessed
in Europe between 1300 and 1800.
An implication of this aggregation of traders in the same market is the growing reliance
on anonymity in all economic relations and, especially, in labor relations. In this respect,
Greif (1994) has already established a link between anonymous market trading relationships
and e¢ ciency wages, and emphasized that less reliance on personal relationship reduces
moral hazard as in the system created by the Maghribi traders, but has a cost in terms
of keeping the size of the market ine¢ ciently low. However, if the commercial revolution
and the lower reliance on personal relationships allowed the economy a better allocation
of resources, the link between commercial revolution and the most striking feature of the
industrial revolution–i.e. the extraordinary impetus of technological innovations–has never
been formally analyzed.
Our argument proceeds as follows. First, the growing reliance on impersonal labor
market transactions in place of customary relations with a high degree of monitoring led to
the adoption of e¢ ciency wages. Second, the resulting rise in the wage led to the adoption
of a more capital-intensive technology. Third, this led to a faster rate of technological
progress through greater learning by doing on the capital intensive production technology.
In particular, we analyse a model where workers can supply labour only on their orignal
island or in the islands connected with the original island. Accordingly we de…ne degree
of urbanization of an economy as the the number of connected islands in the economy.
Entrepreurs in each islands are committed not to hire a shirking worker. However, they
cannot observe shirking workers in the other islands and, therefore, a similar commitment
is not possible among entrepreneurs belonging to di¤erent islands.
2
Accordingly, in a completely isolated market a shirking worker will never be hired again.
On the contrary, in an island connected with a large number of other islands the capacity of
being rehired after shirking is almost una¤ected. Hence, in oder to full…ll the non-shirking
constraint, wages have to be higher the larger the number of connected islands, therefore
the more urbanized is the economy.
Entrepreneurs, beside setting the wages must also choose the tecnology. The capital
intensive technology is more productive in the long run only if a conspicous number of other
entrepreneurs choose it, because of an exterality generated by learning by doing. If wages
are su¢ ciently low, both labour intensive tecnology and capital intensive technology are
Nash equilibria. In this case, we assume that entrepreneurs cannot coordinate on usig the
capital intensive technology and use the labor intensive traditional technolgy by default.
For higer wages the capital intensive technology becomes a stricly dominant equilibrium;
therefore, the more connected are the islands the higher the wages the more likely is that
entrepreneurs switch to the capital intensive technology.
The model also emphasizes a relationship between economic cycle and technology adoption. In an anonymous market characterized by a labor intensive technology, a boom in
the economic cycle leads to an increase in wages –due to the fact that individuals can …nd
a new job more easily if …red–. and, consequently to a more capital intensive production
process. This is not true when the economy is completely isolated, since …red workers are
not hired again. This can explain why, in the pre-comercial revolution period, very a- uent
economies failed to generate the technological switch typical of the industrial revolution.
Our model suggests that the cause might be related to the labor market based on personal
relationships that kept wages low, sterilizing, at least in part, the e¤ects of the economic
cycle.
The last section of the paper is devoted to the empirical analysis where we use some
recent unpublished data on European urbanization collected by Paolo Malanima, and historical wage data from Robert Allen to show that urbanization is a good predictor of future
wages.
The approach taken here draws on ideas which have been used in the literature on the
importance of high wages in stimulating the innovations of the Second Industrial Revolution
in late nineteenth century America (Habakkuk, 1962, David, 1975). Until recently, there
has been a reluctance to cast Britain in the role of a high wage producer at the time of the
3
Industrial Revolution, since the vast literature on the standard of living debate emphasized
the slowness of real wages to rise. However, recent work has emphasized international
comparisons of the level of real wages and other factor prices, pointing clearly to Britain’s
unusual situation of high wages (Allen, 2001; 2006; Broadberry and Gupta, 2006; 2008).
This is important not only in explaining the adoption of modern technology, but also its
non-adoption in other countries with di¤erent factor prices, a point emphasized in the
theoretical literature by Zeira (1998) and in the historical literature by Broadberry and
Gupta (2008), Allen, (2006) and Fremdling (2000).
The idea that the structure of information in the traditional economy is very di¤erent
from that in the modern urbanized economy has been emphasized by Banerjee and Newnan
(1998), where life in a small community increases access to credit but may hamper the more
e¢ cient process of migration to the city. In Banerjee and Newnan, howevere, there is no
link between urbanization and technological progress.
It should also be noted that our approach provides a more direct link between commercial development and economic growth than that provided by Acemoglu et al. (2005), who
emphasize an interaction between the growth of Atlantic trade and institutions in the form
of constraints on the executive. Our approach also maps into variables which are more
objectively measurable than institutional quality.
In section 2 we present a benchmark model in a …rst-best environment. In section 3 we
introduce frictions in the labor market and coordination failure in the choice of technology.
In section 4 we present some empirical evidence while section 5 concludes the paper.
2
MODELLING THE TRANSITION TO MODERN
ECONOMIC GROWTH
In what follows we present a model where the choice of capital intensive technology has a
positive externality because of a learning by doing process and it is e¢ cient only if a large
number of entrepreneurs adopt it. In this section we present the model with perfect labor
market and perfect coordination among the entrepreneurs on the most e¢ cient technology.
In section 3, we introduce coordination failure among entrepreneurs and an imperfect labor
market. We will see that entrepreneurs may fail to coordinate on the e¢ cient technology
when the labor market based on personal relationships features low wages; while in a
4
less anonymous market, high wages rule out the ine¢ cient labor intensive technology and
entrepreneurs coordinate on the more e¢ cient capital intensive technology.
3
The model in a …rst best world
We consider an economy populated by individuals divided between worker types with mass
1 and entrepreneur types with mass e, who produce and consume a perishable consumption
good. Time is indexed by t = [0; 1; : : :], all individuals are risk neutral, have the same
probability d 2 [0; 1] of dying in each period and discount the future at rate
< 1 for each
period. At each period a …xed fraction of individuals (both entrepreneurs and workers) die,
this is replaced by a same fraction of new individuals with the same proportion between
entrepreneurs and workers so that the population and the proportion of types is constant.1
Individuals and entrepreneurs are equally subdivided in N di¤erent islands, so that there
are
1
N
individuals and
e
N
entrepreneurs per island. N
1 islands are disposed circularly
around the N T H that we will call the center. Islands can become connected to each other
through a connection with the center. When n
1 islands are attached to the the center,
so that n islands are connected, we say that the economy has a degree n of urbanization.
Workers can migrate from one village-island to another only if there is a connection, but
entrepreneurs share information about workers’reliability only within each island.
Entrepreneurs do not change island and are committed by a social rule not hire workers
who shirked in the past.2 The act of migration, the real age and original island of workers cannot be observed by the entrepreneurs, therefore entrepreneurs cannot discriminate
between new and migrating workers.3 For this reason shirking workers always have an
incentive to change island if this is linked with the other. If they do not do so, they will
be unemployed until they die.
1
This demographic structure is standard in the macroeconomic literature and it is called BlanchardYaari model , e.g. Blanchard 1985.
2
If all entrepreneurs do not hire a shirking worker, this action is dominant. If an entrepreneur deviates,
the workers will shirk again knowing that his reputation is already destroyed. Therefore, an entrepreneur
has no incentive to deviate. (for a formal proof we refer to Grief (2006), Proposition 3.2 pag 77)
3
This is an extreme assumption with the aim of simpli…ng the model. All results would be qualitatively
una¤ected if we more realistically assumed imperfect observability of these characteristics.
5
3.1
Production
The one unit of workers are endowed with a unit of labor that can be supplied costlessy to
the production process, the e entrepreneurs organise and monitor production using one of
the technologies and employing labor. In order to be an entrepreneur, an individuals needs
to have a business idea at the beginning of her life. We can therefore think that the mass
of entrepreneurs is determined by the long term cycle of the economy. In rich economies
there are more business ideas to exploit and therefore a larger mass of e: For simplicity we
rule out the possibility of the entrepreneur supplying labor.
There are two technologies to produce the consumption good with …xed factor proportions. Every entrepreneur choose the technology at the beginning and does not change
technology until she dies. Workers can switch from one technology to another without any
adaptation cost.
Technology is indexed by i, where the modern technology M is capital intensive, while
the traditional technology T is labor intensive; we omit the time subscript when it is not
needed. We denote capital by k i and labor by li , with lM < lT and the capital requirement
k i in the M technology is higher than in the T technology, where for simplicity we assume
0. Moreover, let elT < L, so there is always excess supply of labor.
Accordingly, entrepreneurial pro…ts are given by:
Ai
li wi
k i (1 + r)
(1)
where wi is the wage rate in sector i, and r is the rate of interest. The user cost of capital
is (1 + r) in our framework, since the price of capital is normalised to unity, there are no
capital gains and depreciation occurs completely within the period. Ai measures the level
of technological progress.
T
M
For ease of notation, we assume that AM
0 = A0 = A0 and, more simply, de…ne At = At .
Noting that mt
1
is the share of entrepreneurs simultaneously adopting the capital intensive
modern technology at time t
1, we assume that:
At = At 1 (1 + (mt 1 ))
where
( ) is a technological progress function with
0
(m)
(2)
0 for all m 2 (0; e). This
can be interpreted as a process of learning by doing, in the spirit of Arrow (1962). Each
6
entrepreneur who adopts the new technology contributes an in…nitesimal improvement to
the level of technology at time t, while at time t + 1 all these small improvements add up
linearly as in (2). We assume that the capital requirement in the M sector increases with
technological progress in the following way:
k M = At
where
< 1 is a time invariant parameter. Given that in the economy there is only one non
storable good produced at the end of each period, this can either be consumed or invested
for next period production.
Finally, we assume that entrepreneurs can borrow from a perfect international credit
market at the beginning of each periods. Individuals are risk neutral and the market
interest rate is determined by the following equilibrium condition
1+r =
Where we de…ned
=
1
(3)
d (i.e. the product between the intertemporal discount factor and
the probability of dying).
3.2
First best equilibrium
In order to provide a benchmark, we start by solving the model under the assumption that
workers’labor supply is perfectly observable and entrepreneurs can perfectly coordinate on
the Pareto e¢ cient equilibrium. The …rst assumption implies that wages w = 0, given that
there is excess supply of labor. Under these assumptions the intertemporal pro…t under
the T technology is simply:
1
X
A0
t
=
t=0
A0
1
The intertemporal pro…t under the M technology is:
1
X
A0 (1
(1 + r ))((1 + (mt 1 ) )t ;
t=0
7
(4)
where, in order to avoid trivialities, we assume that this is always less than in…nity, i.e.
(1 + (e)) < 1 or:
(5)
(e) < r
In our …rst best benchmark model, entrepreneurs are able to coordinate so that mt
1
=e
and the intertemporal pro…ts are:
A0 (1
(1 + r ))
1
(1 + (e))
(6)
We are interested in analyzing a situation where the adoption of the T technology is inef…cient in the …rst best world. Accordingly, expression (4) is always larger than (6), which
is true whenever:
(7)
r ;
>
compatible with (7) if:
> :
Note that the coordination among entrepreneurs is fundamental to achieving this result.
If the “default” technology is the traditional one and if entrepreneurs fail to coordinate,
the dominant equilibrium will be to stay in the traditional technology given that the long
run pro…ts from choosing the M technology,
A0 (1
1
)
are strictly smaller than the pro…ts
from choosing the T technology, determined by expression (4).
In what follows we assume that entrepreneurs cannot coordinate to use the M technology, and we show that an imperfect labor market may still lead the entrepreneurs to choose
the M technology. Crucially, this happens only when the imperfection in the labor market
is exacerbated by the increase in urbanization, that as we argued earlier, corresponds in
our model to an increase in the number of connections between islands, n:
4
Imperfect labor market
We now assume that the labor supply is not observable anymore. Once hired, an individual
laborer decides whether to work honestly by supplying a unit of e¤ort to the …rm or to
shirk. Following Legros, Newman and Proto (2007) we assume that a shirking worker
supplies her unit of e¤ort in a sideline activity, which is hidden in the …rm and yields an
8
amount Ai . A shirking worker is always dismissed without receiving a wage. A dismissed
worker however is never rehired by all entrepreneurs of the same island. His only option is
to change island if this is connected with the central island.4
It is reasonable to require that the sideline activity is not more e¢ cient than the entrepreneurial production by assuming:
lT
1
and
lM +
1:5
All workers are hired in a …rm using technology i with a contract specifying the wage wti .
Contracts last one period and entrepreneurs strictly prefer to hire from the pool of nonshirking workers who worked in the previous period. Since this information is available
only for workers in the same islands, they will strictly prefer to hire them (possibly hiring
the same workers hired in the past period if they survive). This implies that in each
island, surviving non-shirking workers will always be re-employed and entrepreneurs need
to replace the share d eli of workers that died in the last period by new workers, chosen
from the pool of new born, previously unemployed and workers migrating from another
island. The act of migration, the workers’real age and original island cannot be observed
by the entrepreneurs. Entrepreneurs cannot therefore discriminate between new workers,
who are selected randomly from the pool of the unemployed.6
In order to simplify the exposition, we analyze only the pure strategy equilibria in terms
of technological choice, i.e. mt 2 f0; eg. Given our technological evolution determined by
(2), mixed strategy equilibria are not sustainable in the long run and eventually converge
to the use of the M technology.
For a surviving non-shirking worker at time t, the probability of working at t + 1 is 1.
4
The strategy of not rehiring a previoulsy shirking is always weakly dominant since we assumed that
there is excess labour supply, eli < L; to the extend that shirkinng workers are not too many. As it will be
clearer below this is alaway the case in equilibrium where the dominant strategy is non shirking. Clearly
we rule out unrealistic possibilities of coordination among workers on the shirking equilibrium strategy.
5
On the contrary the entrepreneurs may reorganize the …rm in a way that each worker is employed in
the sideline activity.
6
This is an extreme assumption with the aim of simpli…ng the model. All results would be qualitatively
una¤ected if we more realistically assumed imperfect observability of these characteristics.
9
Accordingly, we have:
1
X
VNi S = w0i +
wti
t
(8)
t=1
denoting the intertemporal payo¤ for a non-shirking worker at an initial time t = 0; in a
…rm with technology i.
A shirking worker can only …nd a job in one of the other n
1 connected islands trying
to replace one of the deli dead workers. Accordingly, the probability of …nding another job
is:
i
n
=
n 1
deli :
;
i
(1 d)el n
1
(9)
which is 0 in a completely non-urbanized environment where n = 1.
The intertemporal utility for workers shirking at t = 0 is then:
VSi (n)
= A0 +
i
n
1
X
t=1
wti t
+
i
n (1
i
n)
1
X
wti
t
+ :::
t=2
+
i
(n)(1
i z
n)
1
X
wti
t
+ ::: (10)
t=z+1
from which we note that in an unconnected island with n = 1, the continuation value is 0,
the incentive to shirk is weakest. On the contrary, the bigger is n, the higher is the utility
from shirking.
Accordingly, for any level of urbanization n, entrepreneurs set the wage w0i , in such a
way that:
VNi S
VSi (n):
(11)
Since in the labor market there will always be excess labor supply; constraint (11) must
always hold strictly; from expressions (8) and (10), we note that the equilibrium wages
consistent with the inequality (11) holding strictly could be characterized by several wage
patterns. We will focus here on the most plausible one where the wage share of total
10
production is stable, wti = At ! i for all t:7 We can then rewrite (10) more compactly as:
VSi (n) = A0
+
i
n
1
X
t
wti
+
t=1
i
n
1 X
1
X
! i ((1 + (m)) )t (1
z=2 t=z
i t 1
n)
!
;
(12)
where if m = e; i = M and if m = 0; i = T . Henceforth, we will consider a generic
i
2f
4.1
M
;
T
g, where
T
= 0 and
M
= (e).
Equilibrium Analysis
Before proceeding, we de…ne a rational expectations equilibrium for any level of urbanization n and economic cycle e as:
A set of equilibrium variables mt ; wti ; r with i = fM; T g; and workers’and entrepreneurs’
decision such that for each t = [0; 1 : : :] :
1. Given mt0 , with t0 > 0, workers do not shirk, so Lt0 = 1:
2. Given mt0 with t0 > 0, entrepreneurs choose i and wti to maximize pro…t under
non-shirking constraint (11) at any time t.
3. mt is the sum of all entrepreneurs choosing i = M .
4. r is determined by expression (3).
In what follows, we determine the rational expectations equilibria of the model. Let
us start our analysis at a given time t = 0, where we assume that there are no unreliable
workers. In an economy determined by the pair (n; e), each of the
worker from the pool of laborers,
ne
N
entrepreneurs hires a
n
.
N
Recalling that in the labor market there will always be excess labor supply, constraint
(11) must always hold strictly. Hence using the expressions in (12) and (8), we can derive
a function ! i ( in ;
i
): We note from (12) that VSi (n) is monotonically increasing in
@! i ( in ; i )
> 0, and recalling that from expression
@ in
that ! i ( in ; i ) is increasing with urbanization.
therefore
0, so
7
(9)
@ in
@n
> 0, we have
@! i ( in ;
@n
i
n,
i)
>
Note that this is the only non cyclical pattern. A ! t increasing all the time is not sustainable in the
long run since pro…ts would fall to 0. Also a decreasing ! t is not sustainable given that wages will converge
to 0. Accordingly, Levin (2003) established that a incentive compatible contract in a repeated setting is
likely to be stationary.
11
We can therefore prove the following:
Proposition 1 If
(1
=
) <
lM
level Tn ;
trepreneurs e and a threshold
, there exists a level of urbanization n and en-
<
lT
such that the T technology is a pure strategy Nash
equilibrium if and only if
T
n
T
n
<
(13)
Proof. By solving (12) recursively and recalling that (1 + ) < 1; we have:
0
i
n
VSi (n) = A0 @ +
(1 +
i
i
n)
) 1
(1
(1
(1 +
i ))
(1 +
i
(1
(1
2
)
i
n)
(1
i)
n
(1 +
i
n
(1 +
i 2
)
!i
i ))2
1
A
(14)
while we can write (8) as:
VNi S = A0
!i
(1 +
1
i)
(15)
;
from which we derive our expression:
! i ( in ; ) =
(1
(1
(1
(1 + )) (1
(1 + ) 1
) (1 + ))2
(1
(1
) (1 + ))2
) (1 + )
2
(1
)
(1 + )2
(16)
The choice of T is a dominant strategy it must be true that:
1
X
A0 (1
lT ! T (
T
n;
T
)
t
>
t=0
1
X
A0 (1
lM ! M (
T
n;
T
)
= )
t
(17)
t=0
Expression (17) implies that when every e chooses T; no single entrepreneur has an incentive
to deviate. Note that ! M (
T
n;
T
) must be paid by the deviating entrepreneur and is
determined by the following equation:
A0 !
M
+ A0
1
X
! Tt
t
= VST (n);
(18)
t=1
while the wage share ! T (
T
n ; 0)
is determined by expression (16) with
T
= 0.
The LHS of equation (18) is the intertemporal utility from working at t = 0 for a
deviating entrepreneur (and working for a non- deviating …rm in the future, given that
12
the probability that the same worker …nds employment in the same …rm is 0).8 The RHS
is determined by expression (12) with i = T and
!T (
T
n ; 0)
= ! M : Accordingly, expression (17) is true if:
!T (
Since ! T (
= 0:Therefore it must be true that
T
n;
T
T
n,
) is increasing in
rium is that (19) holds for
T
n
T
n;
T
T !1
n
lT
lM
:
(19)
a necessary condition for the existence of this equilib-
= 0 or:
(1
while, noticing that lim
=
)<
!T (
T
n;
T
)<
=
lT
lM
) = ; a necessary condition for the non-existence of
a pure strategy in T is:
>
Accordingly there will be a
T
n
=
lT
lM
2 [0; 1] :
!T (
T
n;
T
)=
such that the T strategy is an equilibrium if
T
n
=
lT
<
lM
T
n
.
From (13) that, given 9, we can rewrite as:
T
n
<
1
n 1
delT :
;
(1 d)elT n
(20)
and we note that a the choice of traditional technology is more likely for low values of
urbanization n and economic cycle e: For higher levels of n and e, all entrepreneurs will
eventually adopt the M technology, as in the …rst best choice, and this switch is generated
by the high level of wages. Since the M technology is e¢ cient in a …rst best world with
no imperfections in the labor market and perfect coordination among entrepreneurs, while
the T technology is chosen when entrepreneurs fail to coordinate, we can then argue that
adding a more serious labor market imperfection can lead to the adoption of the more
productive technology by ruling out the less e¢ cient technology.
8
Even assuming that the deviating entrepreneur prefer to hire the same workers in the future, she will
not be able to do that for a smaller wege since worker will then prefer to move to the higer wage …rm.
13
Moreover, always from (7), it directly follows that:
Corollary 2 In a perfectly isolated economy, with n = 1; the choice of the T technology is
independent of the economic cycle. While for n > 1 the economic cycle and the level of n
are complementary in switching to the M technology.
This may explain why before the commercial revolution led to the process of urbanization, a capital intensive technology was never adopted even in very a- uent economies
in the past. This switch was only possible when urbanization hampered the capacity of
providing strong incentives to workers and generated an increase in the wage share.
5
Empirical Evidence
The model predicts that an increase in urbanization should be followed by an increase
in wage rates due to the e¤ects of anonymity. In this section we test empirically the
existence of this link. We use data on European countries between 1300 and 1850, with each
data point referring to a 50 year average. Our underlying hypothesis is that urbanization
led to high wages throughout Europe, but our data cannot assess the existence of the
link between high wages and technolocical progress. We refer the reader to a companion
paper (Broadberry, Ghosal, Proto 2008) for a complete historical analysis arguing how
commercialization through high wages could lead to the industrial revolution at the end of
the so called “Little Divergence”of the English and Dutch economies.
Urbanization
As already noted, one way in which the growing commercialisation of the economy
can be captured quantitatively is in the share of the population living in urban areas,
since towns were the centres of commerce. Table 1 provides data on the share of the
population living in towns of at least 10,000 inhabitants. For Europe as a whole, the trend
is unmistakably upwards from 1400. Looking at regional trends, however, urbanization
displays the Little Divergence pattern. In the late medieval period there were two main
urban centres of commerce in north Italy and in the Low Countries. While urbanization
stalled in north Italy after 1500, there was a brief surge in Portugal and to a lesser extent
Spain during the sixteenth century, following the opening up of the new trade routes to
Asia and the New World. However, the most dramatic growth of urbanization in the early
14
modern period occurred in the Netherlands in the sixteenth and seventeenth centuries and
in England during the seventeenth and eighteenth centuries as those countries displaced
the Iberian powers in long distance trade and commercialized their domestic economies to
an unprecedented extent.
INSERT TABLE 1 HERE
Wages
The Little Divergence was …rst identi…ed from the comparison of levels rather than
simply growth rates of wages in Europe. Table 2 sets out the pattern of silver wages in
Europe. The silver wage is the silver content of the money wage in the local currency,
and is useful for comparing wages across countries on a silver standard. Note …rst that
Northwestern Europe saw substantial silver wage growth in the century after the Black
Death of the mid-fourteenth century and again during the early modern period after 1500,
as well as during the Industrial Revolution period from the mid-eighteenth century, when
Britain …nally overtook the Netherlands decisively. Second, note that although southern
Europe shared in the rise in the silver wage following the Black Death, from the mid …fteenth
century the region was characterised more by ‡uctuations than by trend growth in the silver
wage. Third, central and eastern Europe were also characterised more by ‡uctuations than
by trend growth in the silver wage from the mid-fourteenth century. If commercialisation
leads to rising wages, this is the pattern that we would expect, with southern Europe
playing an important part in the commercialisation of the continent during the …rst half of
the millennium, but with northwest Europe playing the leading role after 1500.
INSERT TABLE 2 HERE
The patterns depicted in tables 1 and 2 suggest a positive link between urbanization
and wages. Given that di¤erent regions have di¤erent patterns of urbanization this link
seems to be independent of a general trend. In order to assess this link more quantitatively
and control for other factors, we estimate the following equation:
wagei;t = ci +
gdpi;t 1 : +
urbanizationi;t
where index i indicates the city, and t indexes time.
15
1
+ Y eart
1
+ "i;t 1 ;
All non binomial data are expressed in logarithmic terms. Data on wages are aggregated
at city level and are taken from table 2, while data on urbanization; aggregated at country
level, are from table 1. The variable gdp is the per capita gross domestic product, taken
largely from van Zanden ( table 3), but with additional observations for 1500-1800 for
France, Austria and Germany taken from Maddison. The van Zanden and Maddison data
are not strictly comparable with the other data. To cope with this problem– and at the
same time to provide a control for city heterogeneity–, we introduce city …xed e¤ects,
represented by the variables ci : Finally, the linear trend Y ear takes into account the e¤ect
generated by the common patterns of urbanization. We consider a lag of 50 years between
urbanization and wages (i.e. wage data of 1350-99 are related to urbanization of 1300).
The full sample includes 62 observations on wages and urbanization, but this is reduced to
54 when we include gdp.
In column (1) we present the results for the regression without gdp; so that we can base
our regression on a larger number of observation. Consistent with our model, an increase
in urbanization is followed by an increase in wages, and this coe¢ cient is signi…cant at
the 5 percent level. This result still holds when we introduce per capita GDP as a control
variable. From regression (2) we note that this coe¢ cient is still positive and signi…cant,
while its magnitude has actually increased.
INSERT TABLE 3 HERE
6
References
References
[1] Acemoglu, D., Johnson, S. and Robinson, J. (2005), “The Rise of Europe: Atlantic
Trade, Institutional Change and Economic Growth”, American Economic Review,
95, 546-579.
[2] Aghion, P. and Howitt, P. (1998), Endogenous Growth Theory, Cambridge, Mass.:
MIT Press.
[3] Allen, R.C. (2000), “Economic Structure and Agricultural Productivity in Europe,
1300-1800”, European Review of Economic History, 4, 1-26.
16
[4] Allen, R.C. (2001), “The Great Divergence in European Wages and Prices from the
Middle Ages to the First World War”, Explorations in Economic History, 38, 411447.
[5] Allen, R.C. (2006), “The British Industrial Revolution in Global Perspective: How
Commerce Created the Industrial Revolution and Modern Economic Growth”,
Nu¢ eld College, Oxford,
[6] http://www.nu¤.ox.ac.uk/General/Members/homepage.aspx?nu¢ d=ALLERC00.
[7] Arrow, K.J. (1962), “The Economic Implications of Learning by Doing”, Review of
Economic Studies, 29, 155-173.
[8] Banerjee A.V., and A.F. Newman (1998). “Information, the Dual Economy, and Development,”Review of Economic Studies 64, 631-653.
[9] Braudel, F.P. and Spooner, F. (1967), “Prices in Europe from 1450 to 1750”, in Rich,
E.E. and Wilson, C.H. (eds.), The Cambridge Economic History of Europe, Vol. IV:
The Economy of Expanding Europe in the Sixteenth and Seventeenth Centuries,
Cambridge: Cambridge University Press, 374-486.
[10] Brenner, R. (1982), “The Agrarian Roots of European Capitalism”, Past and Present,
97, 16-113.
[11] Britnell, R.H. and Campbell, B.M.S. (eds.) (1995), A Commercialising Economy: England 1086 to c.1300, Manchester: Manchester University Press.
[12] Broadberry, S.N. (1997), The Productivity Race: British Manufacturing in International Perspective, 1850-1990, Cambridge: Cambridge University Press.
[13] Broadberry, S.N. and Gupta, B. (2006), “The Early Modern Great Divergence: Wages,
Prices and Economic Development in Europe and Asia, 1500-1800”, Economic
History Review, 59, 2-31.
[14] Broadberry, S.N. and Gupta, B. (2008), “Lancashire, India and Shifting Competitive
Advantage in Cotton Textiles, 1700-1850: The Neglected Role of Factor Prices”,
Economic History Review, (forthcoming),
[15] http://www2.warwick.ac.uk/fac/soc/economics/sta¤/faculty/broadberry/wp/.
[16] Clark, G. (1988), “The Cost of Capital and medieval Agricultural technique”, Explorations in Economic History, 25, 265-294.
17
[17] Cole, A.H. and Crandall, R. (1964), “The International Scienti…c Committee on Price
History”, Journal of Economic History, 24, 381-388.
[18] Crafts, N.F.R. (1985), British Economic Growth during the Industrial Revolution,
Oxford: Clarendon.
[19] David, P.A. (1975), Technical Choice, Innovation and Economic Growth, Cambridge:
Cambridge University Press.
[20] Deane, P. and Cole, W.A. (1967), British Economic Growth, 1688-1959, (2nd edition),
Cambridge: Cambridge University Press.
[21] Dutton, H.I. (1984), The Patent System and Inventive Activity During the Industrial
Revolution, 1750-1852, Manchester: Manchester University Press.
[22] Epstein, S.R. (2000), Freedom and Growth: The Rise of States and Markets in Europe,
1300-1750, Abingdon: Routledge.
[23] Fremdling, R. (2000), “Transfer Patterns of British Technology to the Continent: The
Case of the Iron Industry”, European Review of Economic History, 4, 195-222.
[24] Greif, A. (1994), “Cultural Beliefs and the Organization of Society: A Historical and
Theoretical Re‡ection on Collectivist and Individualist Societies”, Journal of Political Economy, 102, 912-950.
[25] Habakkuk, H.J. (1962), American and British Technology in the Nineteenth Century,
Cambridge: Cambridge University Press.
[26] Harley, C.K. (1993), “Reassessing the Industrial Revolution: A Macro View”, in
Mokyr, J. (ed.), The British Industrial Revolution: An Economic Perspective,
Boulder, Colorado: Westview, 171-226.
[27] Herlihy, D. (1997), The Black Death and the Transformation of the West, Cambridge,
MA: Harvard University Press.
[28] MacLeod, C. (1988), Inventing the Industrial Revolution: The English Patent System,
1660-1800, Cambridge: Cambridge University Press.
[29] Mendels, E. (1972), “Proto-industrialisation: The First Phase of the Industrialisation
Process?”, Journal of Economic History, 32, 241-261.
[30] Neal. L. (1990), The Rise of Financial Capitalism: International Capital Markets in
the Age of Reason, Cambridge: Cambridge University Press.
18
[31] Pamuk, S. (2007), “The Black Death and the Origins of the ‘Great Divergence’across
Europe, 1300-1600”, European Review of Economic History, 11, 289-317.
[32] Phelps Brown, H. and Hopkins, S.V. (1981), A Perspective of Wages and Prices,
London: Methuen.
[33] Polanyi, K. [1944] (1957), The Great Transformation, London: Bacon Hill.
[34] Pomeranz, K. (2000), The Great Divergence: China, Europe, and the Making of the
Modern World Economy, Princeton, NJ: Princeton University Press.
[35] Rothbarth, E. (1946), “Causes of the Superior E¢ ciency of USA Industry as Compared
with British Industry”, Economic Journal, 56, 383-390.
[36] Sullivan, R.J. (1989), “England’s ‘Age of Invention’: The Acceleration of Patents and
patentable Invention during the Industrial Revolution”, Explorations in Economic
History, 26, 424-452.
[37] Toynbee, A. (1890), Lectures on the Industrial Revolution of the 18th Century in
England, (third edition), London: Longmans, Green.
[38] Voth, H.-J. (1998), “Time and Work in Eighteenth-Century London”, Journal of Economic History, 58, 29-58.
[39] de Vries, J. (1994), “The Industrial Revolution and the Industrious Revolution”, Journal of Economic History, 54, 249-270.
[40] de Vries, J. and van der Woude, A. (1997), The First Modern Economy: Success, Failure, and Perseverance of the Dutch Economy, 1500-1815, Cambridge: Cambridge
University Press.
[41] van Zanden, J.L. (1999), “Wages and the Standard of Living in Europe, 1500-1800”,
European Review of Economic History, 3, 175-97.
[42] van Zanden, J.L. (2006), “Cobb-Douglas in Early Modern Europe: Simulating Preindustrial Growth”, Paper presented at International Economic History Congress,
Helsinki, http://www.helsinki.…/iehc2006/papers3/vanZanden122.pdf.
[43] Zeira, J., ‘Workers, machines, and economic growth’, Quarterly Journal of Economics,
113 (1998), pp.1091-1117.
19
TABLE 1: European urbanisation rates (%)
1300 1400 1500 1600 1700 1750 1800 1870
Scandinavia
–
–
0.7
2.1
4.3
4.6
4.6
5.5
England (Wales)
4.0
2.5
2.3
6.0
13.2
16.4
22.1
43.0
Scotland
–
–
2.3
1.5
5.3
11.5
23.9
36.3
Ireland
0.8
2.1
–
1.0
5.1
5.1
7.3
14.2
Netherlands
–
–
17.1
29.5
32.5
29.6
28.6
29.1
Belgium
18.2
21.9
17.6
15.1
20.2
16.5
16.6
25.0
France
5.2
4.7
5.0
6.3
8.7
8.7
8.9
18.1
Italy CN
18
12.4
16.4
14.4
13
13.6
14.2
13.4
Italy SI
9.4
3.3
12.7
18.6
16.1
19.4
21
26.4
Spain
12.1
10.2
11.4
14.5
9.6
9.1
14.7
16.4
Portugal
3.6
4.1
4.8
11.4
9.5
7.5
7.8
10.9
Switzerland
3.0
2.0
2.8
2.7
3.3
4.6
3.7
8.2
Austria (Czech, Hung) 0.6
0.5
0.8
1.6
1.7
2.6
3.1
7.7
Germany
3.4
3.9
5.0
4.4
5.4
5.7
6.1
17.0
Poland
1.0
1.3
5.4
6.6
3.8
3.4
4.1
7.8
Balkans
5.2
4.6
7.7
13.3
14
12.3
9.8
10.6
Russia (European)
2.1
2.3
2.0
2.2
2.1
2.5
3.6
6.7
EUROPE
5.4
4.3
5.6
7.3
8.2
8.0
8.8
15.0
Source: Paolo Malanima (private communication).
20
TABLE 2: Daily silver wages of European unskilled building laborers
(grams of silver per day)
1300
1350
1400
1450
1500
1550
1600
1650
1700
1750
1800
3.4
4.5
3.8
3.2
4.6
7.1
9.7
10.5
11.5
17.7
3.1
4.7
7.2
8.5
8.9
9.2
9.2
3.0
5.9
7.6
7.1
6.9
6.9
7.7
2.8
5.5
6.6
6.9
5.1
5.2
9.9
4.2
6.6
8.8
6.9
5.7
5.1
–
–
6.3
8.0
–
5.1
5.3
8.0
2.9
3.8
4.7
4.1
3.2
2.9
3.1
3.3
3.5
5.3
4.8
4.8
3.8
3.8
Gdansk
2.1
2.1
3.8
4.3
3.8
3.7
4.8
Warsaw
–
2.5
3.2
2.7
1.9
3.4
4.9
Northwestern Europe
London
2.9
Amsterdam
Antwerp
3.5
3.1
Paris
Southern Europe
Valencia
5.6
5.2
Madrid
Florence/Milan
2.2
4.5
3.8
3.5
Naples
Central eastern Europe
Krakow
2.7
2.1
1.9
2.9
3.4
2.9
2.2
2.9
2.4
Vienna
4.0
3.2
2.7
2.6
4.4
3.5
3.2
3
2.1
Leipzig
–
1.9
3.5
3.9
3.7
3.1
4.4
Augsburg
2.1
3.1
4.0
4.7
4.2
4.3
–
Source: Broadberry and Gupta (2006: 7);
derived from the database underlying Allen (2001: 429).
21
Table 3: Wages and Urbanization
urbanization
Y ear
(1)
(2)
0.207*
.316*
(.105)
(.156)
11.15**
9.96**
(3.69)
(3.61)
per cap: GDP
.264
(.288)
cons:
-21.2**
-20.1**
(7.36)
( 7.11)
City …x e¤ect
Yes
Yes
R2
0.318
0.477
Observations
62
54
Cities
14
14
Std errors in parentheses;
*Signi…cant at 5% level; ** signi…cant at 1% level;
22