alessandro nuvolari and mattia ricci
economic Growth in england, 1250-1850: 
some new estimates using a demand  
side approach
1. Introduction
Following the pioneering work of angus maddison 
(2001; 2003; 2007), economic historians have devoted 
significant research efforts to the construction of statistical 
appraisals of the performance of european economies 
since the end of the middle ages by attempting to work 
out more and more reliable estimates of GdP per capita. 
Just to name only a few of the most recent contributions: 
Broadberry et al. (2011) have produced estimates of GdP 
per capita for england and Britain over the period 1270-
1870; clark (2010) for england for the period 1200-1870; 
van Zanden and van leeuwen (2012) for Holland over 
the period 1347-1807; malanima (2011) for central and 
northern italy for the period 1300-1913; Álvarez-nogal and 
Prados de la escosura (2007) for spain and Pfister (2011) 
for Germany over the period 1500-1850. as all these 
authors would readily admit, these estimates ought still to 
be considered as highly conjectural. still one cannot avoid 
the impression, that these ongoing efforts of statistical 
reconstructions of GdP per capita have the potential to 
put our understanding of the patterns of economic growth 
during the pre-industrial revolution period on a more 
secure footing.
Broadly speaking, so far these contributions have 
produced two opposed accounts of european economic 
rivista di storia economica, a. XXiX, n. 1, aprile 2013
 We are particularly grateful to Paolo Malanima for very helpful comments and en-
couragement. We would like also to thank Greg Clark, Alexia Delfino, Bas van Leeu-
wen, Bob Allen, Francisco Beltran Tapia and Florian Ploeckl for useful discussions. 
Alessandro Nuvolari gratefully acknowledges the Jemolo Fellowship Scheme and the 
hospitality of Nuffield College, Oxford whilst writing this paper.
32 alessandro nuvolari and mattia ricci
performance in the 1200-1800 period. in a recent paper, 
clark (2011) has eloquently described these conflicting 
interpretations as a debate between «malthus» and the 
«revolt of the early modernists». the malthusian view, 
which has been fully articulated by clark (2007a), contends 
that all societies (including england) before the industrial 
revolution were characterized by a malthusian dynamics. 
the implication is that in these societies GdP per capita 
exhibited fluctuations around a subsistence income (defined 
as the level of income at which birth and death rates are 
equal), but without any systematic growth trend. the «revolt 
of the early modernists» interpretation, instead, argues that, 
since the end of the middle ages, it is possible to discern 
a small but steady acceleration in the rate of economic 
growth in at least two european countries: england and 
the netherlands. 1 the current estimates suggest that the 
«cruising speed» for these two successful early modern 
economies was represented by an average compound growth 
rate of GdP per capita of about 0.2 per cent per year. 
Broadberry et al. (2011) estimate an average growth rate of 
GdP per capita of 0.17 per cent per year for england over 
the period 1270-1690 and van Zanden and van leeuwen 
(2012, p. 123) reckon an average growth rate 0.19 per cent 
per year over the period 1347-1807 for Holland. these 
growth rates of GdP per capita may be seen as broadly 
consistent (albeit slightly lower) with those emerging 
from maddison’s comprehensive dataset on the historical 
development of the world economy (maddison 2007, p. 383: 
0.27 per cent for the united Kingdom and 0.28 per cent 
for the netherlands for the period 1500-1820). interestingly 
enough, these estimates also resonate well with the early 
speculative assessment of david landes (1969, pp. 13-14): 
Western europe was already rich before the industrial 
revolution [...] it seems clear that over the near millennium from 
the year 1000 to the eighteenth century, income per head raise 
appreciably – perhaps tripled and that this rise sharply accelerated 
in the eighteenth century even before the introduction of the new 
industrial technology. 
the present state of the debate can therefore be 
described as follows: we have two conflicting sets of 
estimates of GdP per capita for england, one consistent 
 economic Growth in england, 1250-1850 33
with the malthusian view (clark 2010) and one supporting 
«the revolt of the early modernists» (Broadberry et al. 
2011). it should be noted that these estimates of GdP 
per capita have been reconstructed using two different 
approaches and different types of data. clark’s estimates 
are based on an income approach (GdP is computed as 
the sum of all property incomes plus wages) whereas the 
estimates of Broadberry et al. (2011) have been constructed 
using an output approach (GdP is computed as the sum of 
the outputs of all economic sectors). the next step in the 
debate seems to be the analysis of the relative compatibility 
of these estimates with other pieces of empirical evidence. 
this, for example, is precisely the approach of clark 
(2011). in this paper, however, we adopt a different and 
more modest research strategy. We take as starting point 
the price and wage data that clark (2010) has used for 
his estimates and we construct a new time series of GdP 
per capita for england over the period 1250-1850 using an 
alternative (indirect) method: the demand side approach. 
this approach has been effectively used by malanima 
(2011) for constructing estimates of GdP per capita 
for central and northern italy. our aim is to check the 
consistency of clark (2010) and Broadberry et al. (2011) 
series of per capita GdP with new estimates constructed 
using a different approach with the hope to contribute to 
their further refinement. since the patterns underlying the 
GdP per capita series of clark (2010) and Broadberry et 
al. (2011) are so strikingly divergent, it seems appropriate 
to provide a preliminary assessment of these contributions 
by reconstructing a new series of GdP per capita for 
england employing an alternative method. the remaining 
of the paper is structured as follows. in the next section 
we introduce the demand side method for estimating 
GdP per capita and we discuss its main advantages and 
limitations. subsequently, we describe the data and sources 
we have used. in section 3 we set out our estimates and 
compare them with those of Broadberry et al. (2011) and 
clark (2010). section 4 concludes.
34 alessandro nuvolari and mattia ricci
2. Methods and Materials
2.1. The Demand Side Approach
our approach to the estimation of GdP per capita 
follows the one developed by malanima (2011) for central 
and northern italy. the approach is based on a two-step 
procedure. the first step consists in the estimation of the 
output of the agricultural sector using the demand-side 
method. this method has been used, among others, by allen 
(2000) and Federico and malanima (2004) for constructing 
estimates of agricultural output in the early modern period 
respectively for a number of european countries and for 
italy. the starting point is the following equation defining 
total agricultural output (YA):
(1) YA = r · c · N
In equation (1) r
C
Y
A
A=  is the ratio of domestic agricultural
production (YA) to agricultural consumption (CA) in the 
country in question (if r = 1 the trade position of the
country in agricultural goods is perfectly balanced), c
N
CA=  
is the consumption per capita of agricultural goods and N is 
the total population. dividing both sides of equation (1) by 
N we get per capita agricultural output yA: 
(2) yA = r · c
next, we assume that per capita consumption of 
agricultural goods c will depend on the level of wages and 
prices according to the following equation (allen 2000, p. 
13; Federico and malanima 2004, p. 438):
(3) c W P Pa m$ $= a b c
equation (2) is a demand function where W is the real 
wage, Pa a price index of agricultural products, Pm a price 
index of manufactures and a, b and c are the (constant) 
elasticities of demand of agricultural products with respect 
to wages and prices. standard microeconomic consumer 
theory suggests that a + b + c = 0 (malanima 2011, p. 171). 
By making sensible assumptions on the magnitudes of the 
 economic Growth in england, 1250-1850 35
elasticity coefficients, a, b and c, and using the available 
data on wages and prices it is possible to get an estimate of 
c, and from there, after having made an evaluation of r, it 
is possible to obtain the value of the per capita agricultural 
output yA.
the second step of the procedure consists in using the 
estimated values of per capita agricultural output to get to 
an estimate of real GdP per capita. in order to do this, it 
is necessary to formulate an assessment of the share of total 
agricultural output in aggregate GdP. in the case of italy, 
malanima (2011) estimates the share of agriculture output 
in total production using two different approaches. the first 
method is based on the extrapolation of the share of output 
in the secondary and tertiary sectors from urbanization levels 
on the basis of the results of a regression equation relating 
the share of output in the secondary and tertiary sectors 
to urbanization levels in post unification italy. the second 
approach consists in a backward extrapolation of the non-
agricultural output share for the period 1851-1860 using the 
trend displayed by the share of non-agricultural employment 
in the total working population. this second method relies 
on the assumption that the relative labour productivity in 
agriculture compared to that of the rest of the economy did 
not change. Here, we adopt a more simple procedure. We 
define the share of agricultural output in total output (Y) as 
(4) S
Y
Y
A
A=
equation (4) can be written as: 
(5) S
L
L
A
A A
$
$
r
r
=
where LA is the numbers of workers in agriculture, L is 
the total number of workers in the economy, 
L
Y
A
A
Ar =  and
L
Y
r =  represent respectively the labour productivity of 
agriculture and of the entire economy, measured in terms
of output per worker.
if we assume competitive labour markets, real wages in
agriculture 
p
wAc m and in the entire economy 
p
wc m
will track closely labour productivity. 2 We can then
36 alessandro nuvolari and mattia ricci
use the ratio of real wages between the two sectors as 
a proxy for the relative productivity of agriculture with
the respect to the entire economy 
p
w
p
wA A,
r
r
c m.
in other words, equation (5) is transformed into
(6) S
p
w L
p
w
L
A
A
A
$
$
=
using equation (6) we can estimate SA provided that we
have data on real wages in agriculture and in the total 
economy and the share of working population employed in the
agricultural sector 
L
LAc m. it is worth noting, that the main
difference between our approach and that of malanima 
(2011) is that rather than assuming a constant level of 
relative labour productivity between agriculture and the rest 
of the economy, we have preferred to make the assumption 
of (nearly) competitive labour markets.
We shall adopt this method for estimating the share 
of agriculture in total output only for the period before 
1690. For the more recent period, instead we shall base 
our assessment on other existing estimates. once we have 
estimated agricultural output per capita and the share of 
agriculture in aggregate output, we can calculate per capita 
GdP (y) using the formula: 
(7) y
N
Y
S
y
A
A= =
clearly, the demand side approach is an indirect method 
of estimating agricultural output and GdP. compared to 
the more conventional and «direct» output and income 
approaches, the main advantage of the demand side 
approach is represented by being less exacting in terms 
of the data necessary for constructing the estimates (as we 
have seen, the data needed for implementing the approach 
are time series of wages and prices). the main limitation 
of the approach is the need of relying on a number of 
rather stringent assumptions both for the computation of 
per capita agricultural consumption and of the share of 
agriculture in total income. it is also important to note that 
the geographical scope of the estimates constructed with this 
method will reflect the geographical scope of the original 
time series of wages and prices employed.
 economic Growth in england, 1250-1850 37
2.2. Data and Sources
the average wage and the farm wage time series we 
employ are derived from clark (2010, pp. 54-55). these 
wage time series are based, in turn, on clark (2007b) for 
farm wages and clark (2005) for non-farm wages. Farm 
and non-farm wages are then weighted by their respective 
employment share to compute an aggregate average wage 
series (clark 2010). all these series are for daily wages and 
they are constructed as 10-year averages. real wages are 
computed using the cost of living index proposed by clark 
(clark 2010, pp. 98-99).
the estimation of per capita agricultural output (equation 
3) is based on two time series of prices: agricultural goods 
and manufactures. the price index of manufactured goods is 
a geometric index constructed by clark (2010, pp. 138-139) 
using as weights the expenditure shares and it covers the 
prices of products such as: pottery, glass, woodware, pewter 
goods, brass goods, cutlery, paper, etc. the price index of 
agricultural goods is derived from clark (2004). it is also a 
geometric index employing output shares as weights and 
comprising the price series of 26 products: wheat, barley, 
oats, rye, peas, beans, potatoes, hops, straw, mustard seed, 
saffron, hay, beef, mutton, pork, bacon, tallow, eggs, milk, 
cheese, butter, wool, firewood, timber, cider and honey. the 
index has been constructed by aggregating these individual 
series in four main categories: arable products, pastoral 
products, wood products and cider/honey. these four main 
sub-components are then aggregated into a composite index 
of agricultural products. also these series represent 10-year 
averages. since we are interested in relative prices, both price 
series of agricultural products and manufactures have been 
deflated using the cost of living index constructed by clark 
(2010, pp. 98-99).
the data on the share of agricultural workers in the total
working population 
L
LAc m are taken from allen (2000, p. 8).
allen constructs this estimate of the distribution of the 
labour force following the method originally developed by 
Wrigley (1985). this approach consists in assessing the size 
of population engaged in non-agricultural occupations on 
the basis of the rates of urbanization. this estimate is then 
adjusted in order to take into account the share of rural 
population engaged in non-agricultural occupations (allen 
38 alessandro nuvolari and mattia ricci
2000, pp. 4-13). allen provides estimates for the following 
years: 1300, 1400, 1500, 1600, 1700, 1750, 1800.
clark (2010, pp. 56-57) has also recently reconstructed 
alternative estimates of the share of employment in 
agriculture. Figure 1 compares allen and clark’s estimates. 
Figure 1 shows that clark’s estimates for the period before 
1700 are much lower than those of allen’s. in this paper, 
we use allen’s estimates as they seem to be consistent with 
those elaborated independently by crafts using the social 
tables for 1688 and 1759 (crafts 1985, p. 14). Furthermore, 
in the italian case, malanima has also found allen’s estimates 
of the employment structure for the period 1300-1800 fairly 
plausible and consistent with other pieces of empirical 
evidence (malanima 2011, p. 184). 
using allen’s estimates of the employment share in 
agriculture, we can estimate the share of agriculture in total 
output using equation (6) for the benchmark years 1300, 
1400, 1500, 1600, 1700, 1750 and 1800. We calculate the 
intervening values between these benchmark estimates 
by interpolation. in this way, we are able to construct a 
complete time series of the agricultural share in total output. 
to check the reliability of our estimates, in figure 2 we 
compare them with alternative estimates of the agricultural 
share in total output constructed by deane and cole (1969, 
p. 156 for the years 1688 and 1770 and p. 166 for the period 
after 1800-1850) and crafts (1985, p. 16 for the period 
1690-1760; p. 45 for the period 1780-1801 and the period 
1801-1831 and crafts, 1983, p. 191 for the year 1780). We 
FIgure 1. employment share in agriculture, 1300-1850.
0.85
0.75
0.65
0.55
0.45
0.35
0.25A
gr
ic
ul
tu
ra
l e
m
pl
oy
m
en
t s
ha
re
1250 1350 1450 1550 1650 1750 1850
Clark Allen Crafts
 economic Growth in england, 1250-1850 39
should note that these estimates of the sectoral share of 
agriculture developed by crafts have been employed also 
for the construction of the revised estimates of real GdP 
growth proposed by crafts and Harley (1992, p. 715).
in figure 2 the series labeled «ricci-nuvolari» represents 
our estimation of the share of agriculture in total output 
computed using equation (6) whereas the series labelled 
«deane & cole» and «crafts» represent deane and cole and 
crafts estimates. Figure 2 shows that our estimates for the 
early years 1690-1700 are fully in line with those proposed 
by crafts and deane and cole (in particular our calculation 
using equation (6) yields an estimate of the agricultural share 
in total output in 1700 of 36 per cent while crafts considers 
this to be 37 per cent). For the period after 1700 instead the 
decline of the agricultural share in total output computed 
using equation (6) is more rapid than that shown in crafts 
and deane and cole estimates. allen (2009) has argued 
that in the period 1770-1840 real wages stagnated whereas 
output per worker increased: a prolonged divergence 
leading to a significant shift in income distribution. clearly, 
this means that estimating the agricultural output share 
using equation (6) is not likely to be an accurate procedure 
for this historical period. For this reason we have decided 
to revise our estimates for the interval 1700-1850 adopting 
a series closer to the estimates of crafts and deane and 
cole which are based on more direct assessments of the 
nominal value of output in different sectors in benchmark 
FIgure 2. agriculture share in total output, 1690-1850.
0.60
0.50
0.40
0.30
0.20
0.10
0A
gr
ic
ul
tu
ra
l s
ha
re
 in
 to
ta
l o
ut
pu
t
1600 1650 1700 1750 1800 1850
Crafts Deane & Cole
Ricci & Nuvolari Ricci & Nuvolari amended (1700-1850)
40 alessandro nuvolari and mattia ricci
years. in particular, we have used the following procedure. 
We adopt as a «compromise» estimate of the agricultural 
share in total output in 1800 the average of the crafts 
and deane and cole estimates. then we compute the new 
estimates for the period 1700-1800 by interpolation between 
the value of our time series in 1700 and the new value of 
1800. For the interval 1800-1830, we construct our estimates 
by interpolation between our new value of 1800 and the 
value for 1830 estimated by deane and cole. Finally the 
observations for 1840 and 1850 are derived directly from the 
deane and cole estimations. in this way we obtain a new 
series of agricultural share for the period 1700-1850 which 
in figure 2 is labeled as «ricci-nuvolari amended (1700-
1850)». to sum up, our final estimates of the agricultural 
share in total output are represented by the «ricci-nuvolari» 
series for the period before 1700 and by the «ricci-nuvolari 
amended» series for the period after 1700.
the data on the ratio between agricultural production 
and the domestic consumption of agricultural goods (r) are 
taken from crafts (1985, p. 127) for the period 1800-1850. 
intervening values were obtained by interpolation. We 
should note that alternative estimates provided by thomas 
(1985, p. 148) are also available for the period 1800-1850. 
However, they are very close to those proposed by crafts. 
For the period before 1700, there are no sources of data 
readily available. However, according to allen (2000, p. 14), 
«[t]here is no indication that r differed from one before 
the middle of the seventeenth century». Here we follow the 
same type of assumption and consider r = 1 up to the year 
1750. 3 values for the period 1750-1800 were computed by 
interpolation.
the price and income elasticities of the demand function 
(3) are taken from allen (2000, p. 14). allen, on the basis of 
studies of modern developing countries, assumes the demand 
elasticity with the respect to the price of manufactures (Y) 
to be equal to 0.1. the demand elasticity with respect to 
the price of agricultural goods (b) is taken to be –0.6. as 
a result, the condition a + b + c = 0, suggests that a must 
be equal to 0.5. experiments with alternative values of 
these elasticities have produced very similar estimates of 
agricultural output. 4
Finally the data on population are taken from Wrigley and 
schofield (1997) for the period after 1541. For the previous 
period the data are taken from clark (2010, pp. 64-65).
 economic Growth in england, 1250-1850 41
3. Estimating Agricultural Output and GDP Per Capita
the complete time series of our estimates of per capita 
agricultural output and of real per capita GdP calculated 
using the procedures explained in the previous section are 
presented in the appendix. Figure 3 shows our estimates of 
per capita agricultural output (yA) and of total agricultural 
output (YA). total agricultural output has been computed by 
multiplying per capita agricultural output by total population. 
all these estimates are reported using the index 1700 = 100. 
Figure 4 compares our estimates of total agricultural output 
with those of allen (2000) and Broadberry et al. (2011). 
the estimates of Broadberry et al. (2011) refer to england 
in the period 1270-1700 and to Great Britain for the period 
1700-1850. our estimates appear to be in broad agreement 
with those of allen (2000). For the period before 1550, the 
estimates of agricultural output of Broadberry et al. (2011) 
are instead somewhat lower. the implication is that these 
estimates will display faster growth throughout this period. 
For the period after 1550-1750, our estimates and those of 
allen (2000) and of Broadberry et al. (2011) appear to be 
broadly consistent. taking into account that our estimates 
have been constructed following the same approach used 
by allen (the only difference are the wages and prices series 
used in the computation), these findings are not completely 
surprising. the obvious implication is that a possible 
avenue for further research will be to search for the factors 
accounting for this divergence in agricultural output estimates 
between the output and the demand side approach in the 
period before 1600. 
Figure 5 contains our estimates of GdP per capita 
compared with those of clark (2010), Broadberry et al. 
(2011), maddison (2001, p. 247) and malanima (2011, p. 
189). again, all the series are reported using the index 
1700 = 100. it is important to take into account that 
maddison’s estimates refer to england, scotland and Wales, 
the estimates by Broadberry et al. (2011) refer to england 
for the period 1270-1700 and to Great Britain for the period 
1700-1850, while the estimates of all the other authors 
concern only england. the yearly series of Broadberry et al. 
(2011) has been converted into 10-year averages centered on 
each decade (so that it is directly comparable with clark’s 
and our estimates). overall, our estimates display a pattern 
that seems to be an intermediate case between the estimates 
42 alessandro nuvolari and mattia ricci
FIgure 3. agricultural output (1700 = 100).
300
250
200
150
100
50
0A
gr
ic
ul
tu
ra
l o
ut
pu
t (
17
00
 =
 1
00
)
1250 1350 1450 1550 1650 1750 18501300 1400 1500 1600 1700 1800
Per capita agricultural output Total agricultural output
FIgure 4. comparative estimates of total agricultural output (1700 = 100).
350
300
250
200
150
100
50
0A
gr
ic
ul
tu
ra
l o
ut
pu
t (
17
00
 =
 1
00
)
1250 1350 1450 1550 1650 1750 1850
Ricci & NuvolariBroadberry et al. Allen
FIgure 5. comparative estimates of per capita GdP (1700 = 100).
200
180
160
140
120
100
80
60
40
20
P
er
 c
ap
ita
 G
dp
 (1
70
0 
=
 1
00
)
1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850
Ricci & NuvolariClark
Maddison
Broadberry et al.
Malanima
 economic Growth in england, 1250-1850 43
proposed by clark (2010) and those of Broadberry et al. 
(2011).
the different patterns of economic growth implicit 
in these time series estimations become apparent when 
we consider long run growth rates. table 1 compares 
the average compound growth rates of GdP per capita 
of the different estimates using various subperiods. We 
have also included in the table the estimates of crafts 
and Harley (crafts and Harley 1992; Harley 1993, p. 178) 
covering the years 1700-1830. For the period 1250-1580, 
our estimates are consistent with those of clark showing 
no positive growth. in particular, as it can be seen from 
figure 5, both our estimates and those of clark exhibit a 
large «malthusian cycle» of growth and decline beginning 
around 1300 and ending around 1600 (although the cycle 
is clearly more nuanced in our estimates than in clark’s). 
the peak of the cycle occurs around 1450. this behavior 
of GdP per capita mirrors the fluctuations of the real wage 
series constructed by clark (2007a, p. 41). For the interval 
1580-1780, our estimates are instead consistent with those 
of Broadberry et al. (2011) indicating that the english 
economy was able to attain an annual growth rate of about 
0.2 per cent. Hence our estimates suggest that by the end 
of the sixteenth century the english economy was probably 
already beginning to break away from the malthusian 
constraints. 5 this is in contrast with the estimates of clark 
(2010) showing the english economy attaining a «sizable» 
positive rate of economic growth only from the end of the 
eighteenth century. thus, in terms of the overall pattern, 
our estimates appear consistent with those of Broadberry 
et al. (2011), suggesting the historical relevance of a 
three-stages periodization, rather than the two-stage one 
underlying clark’s estimates. in particular, our estimates 
show that the period 1250-1580 can be characterized as a 
«malthusian-phase» with a generalized stagnation in GdP 
per capita; the period 1580-1780 can be perhaps seen as a 
«smithian-phase» of positive economic growth preceding 
the industrial revolution (although we should also note 
that our estimates of the rate of economic growth during 
this period are somewhat lower than those of Broadberry 
et al. (2011)). Finally, all estimated time series are in broad 
agreement in indicating that the beginning of the process 
of «modern economic growth» with steady positive growth 
rates significantly above the threshold of 0.2 per cent per 
44 alessandro nuvolari and mattia ricci
year should be probably located at the end of the eighteenth 
century.
interestingly enough, we can note in figure 5 that both 
the clark’s series and, to a more significant extent, our series 
suggest a phase of relatively sluggish growth performance in 
the period of the revolutionary and napoleonic wars (1790-
1810). this is probably to be ascribed to the disrupting 
effects of the wars on prices. of course since the approach 
adopted here is relying heavily on the wages and prices 
series, the erratic behavior of prices in this historical phase 
may introduce some spurious effects in our estimates for this 
specific period.
Figure 6 contains our estimates of GdP per capita 
measured in 1990 «Geary-Khamis» PPP dollars, which 
is the unit of measurement adopted by maddison (2001, 
2003, 2007) and it is frequently used for international 
comparisons. this series has been computed projecting 
backwards maddison’s value of GdP per capita in england, 
Wales and scotland for the year 1850 (maddison 2001, p. 
247) using our estimated time series of per capita GdP. 6 
in order to put our estimates in a comparative context, in 
figure 6 we have also plotted the series of GdP per capita 
(also expressed in 1990 PPP $) estimated by malanima 
(2011). in order to be fully comparable with our series, the 
yearly series of malanima has been converted into 10-year 
averages centered on each decade. as we have mentioned, 
also malanima has constructed his estimates using a very 
similar approach to the one adopted here. there are two 
Table 1.  Rates of growth of GDP per capita (annual average compound growth rates 
(%)
Period Broadberry 
et al.
clark maddison malanima crafts & 
Harley
ricci & 
nuvolari
1250-1580 0.12* –0.01 –0.02
1580-1780 0.29 0.03 0.22
1780-1850 0.53 0.51 0.36
1780-1820 0.24 0.46 0.12
1500-1700 0.19 –0.11 0.28 0.14 0.05
1700-1820 0.25 0.21 0.26 0.21
1700-1800 0.06 0.10
1820-1850 0.91 0.57 1.05 0.68
1800-1870 0.79
1700-1780 0.26 0.09 0.24 0.25
1780-1830 0.32 0.46 0.45 0.33
* 1270-1580.
 economic Growth in england, 1250-1850 45
points of interest arising from figure 6. the first point is 
that our estimates show that england throughout the late 
middle ages and very early modern period was considerably 
richer than the picture emerging from maddison’s estimates. 
our estimates suggest that over the period 1250-1500 
income per capita fluctuated in an interval ranging from a 
minimum of about 1.000 $ to a maximum of about 1.700 
$, whereas maddison reckoned GdP per capita in england 
to be 400 $ in year 1000 and 762 $ in year 1500. in fact, 
several authors (Federico 2002; clark 2009) have argued 
that the basic «subsistence» income level of 400 $ per capita 
that maddison considers as characteristic of not particularly 
sophisticated societies is far too low. Following this cue, lo 
cascio and malanima (2009), on the basis of a number of 
considerations concerning the price of foodstuffs, clothing 
and fuel and the level of real wages, have proposed that 
700 $ and not 400 $ should be regarded as the minimal 
«bare bones» threshold for GdP per capita in pre-modern 
europe. in this perspective, we can see that in figure 6 in 
the period 1250-1580, english per capita GdP is fluctuating 
in a range comprised between 1.5 and 2 times this basic 
level identified by lo cascio and malanima. We should also 
note that the GdP per capita level estimated by Broadberry 
et al. (2011) for the period 1250-1350 seems to fluctuate 
around the minimum «bare bones» threshold of 700 $. 
the second point that is worth noting is that, even if our 
estimates show a relative high level of income per capita 
for england in the period 1250-1600, they still indicate the 
FIgure 6.  GdP per capita in england and in central/northern italy (1990 $), 1250-
1850.
3500
3000
2500
2000
1500
1000
500
0
G
D
P
 p
er
 c
ap
ita
 (1
99
0$
)
1250 1350 1450 1550 1650 1750 18501300 1400 1500 1600 1700 1800
England (Ricci & Nuvolari) Italy CN (Malanima)
46 alessandro nuvolari and mattia ricci
existence of a noticeable gap in GdP per capita between 
italy and england lasting at least until the second half of 
the fifteenth century. Figure 6 shows that the moment in 
which england is decidedly overtaking italy is the second 
half of the seventeenth century. this may perhaps be seen 
as consistent with interpretations that have pointed to the 
importance of international trade (in particular the successful 
challenge mounted by english traders to italian producers in 
wool textiles during the seventeenth century) in accounting 
for the patterns of economic divergence in early modern 
europe (allen 2002). 7 
Finally, it is possible to provide a rough assessment of the 
reliability of our estimates of GdP per capita, by computing 
the implied number of working days during the year that are 
«implicit» in our estimates of GdP per capita. Formally the 
number of working days during the year (d) is equal to:
(8) d
w L
GDP
$
$ v
=
where GdP is the nominal GdP, v the share of wages in 
total income, w the nominal wage and L the total number of 
workers. equation (8) tells us how many days it is necessary 
to work in order to obtain an yearly earning corresponding 
to a certain level of GdP per capita (given the prevailing 
wage and patterns of income distribution). in order to 
implement empirically the formula we proceed as follows. 
First, we construct an index of nominal GdP by multiplying 
our estimates of real GdP per capita by the cost of living 
index constructed by clark (2010, pp. 98-99) and by total 
population, again taken from clark (2010, pp. 64-65) and 
Wrigley and schofield (1997). then, following clark (2010, 
p. 59) we assume that the nominal value of the labour share 
in total income around 1860 is equal to 420 millions of 
pounds. 8 We project backwards this estimate of the nominal 
value of the labour share in total income using the index of 
nominal GdP and assuming v to be equal to 0.6. this value 
for v is consistent with the dynamics of the labour share 
in total income emerging from clark’s estimates (2010, pp. 
81-82). in this way, we are able to construct a time-series 
of the nominal value of labour income (the numerator of 
formula 8). in order to estimate L we compute the share 
of working population in total population by dividing the 
total working population given by deane and cole (1969, 
p. 143) for the period 1801-1861 by total population 
 economic Growth in england, 1250-1850 47
(clark 2010, p. 65). For the period 1801-1861 we obtain an 
average value of 0.53. We use this value for calculating the 
number of workers in each period. We can now compute 
the denominator of formula (8) by multiplying the average 
daily wages (series taken from clark (2010, pp. 54-55)) by 
the total number of workers.
Figure 7 compares the number working days during 
the year computed using equation (8) (the series is labeled 
«implied working days»), with a number of independent 
estimates of the actual working year for different periods 
assembled by allen and Weisdorf (2011). the original 
sources for these estimates are Blanchard (1978), clark 
and van der Werf (1998) and voth (2001). overall, figure 
6 suggest that our estimates of the working days during 
the year implicit in our GdP per capita series are able to 
track rather closely the available independent estimation 
of working days. Given the admittedly crude procedures 
adopted for the computation of the working days, we 
consider this result as a promising preliminary corroboration 
of our GdP per capita estimates. 
4. Conclusions
In this paper we have presented a new set of estimates of 
agricultural per capita output and real per capita GdP for 
FIgure 7.  number of working days implied by ricci-nuvolari GdP per capita es-
timates, 1250-1850.
400
350
300
250
200
150
100
N
um
be
r 
of
 w
or
ki
ng
 d
ay
s
1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850
Implied working
days
Blanchard Voth (farmers)
Voth (non farmers) Clark & van
der Werf
48 alessandro nuvolari and mattia ricci
the english economy in the period 1250-1850 constructed 
using the demand side approach. as we have seen, this 
approach to the statistical reconstruction of per capita 
GdP relies on a number of exacting assumptions and we 
believe that it should be regarded as nothing more than 
an useful framework of inquiry for formulating reasoned 
conjectural assessments of the historical performance of an 
economy. still, we think that in the case in question, despite 
its inherent limitations, the implementation of the demand 
approach has performed reasonably well, generating some 
interesting findings and producing a rather plausible picture 
of the long run evolution of the english economy. Further, 
our estimates have received some further corroboration by 
being consistent with some empirical evidence concerning the 
number of working days in the year at different time periods.
our estimates suggest that the growth experience of 
english economy over the period 1250-1850 can be suitably 
interpreted using a three-stage account. in particular, our 
estimates indicate the existence of a malthusian phase 
covering the period 1250-1580. this phase is followed 
by an «intermediate» stage preceding the industrial 
revolution covering the seventeenth and eighteenth century 
during which the economy is able to dissipate some of 
the malthusian constraints and attain a sustained positive 
growth rate. the third phase corresponds to the industrial 
revolution and by a further significant acceleration in the 
rate of economic growth. in this perspective, the pattern of 
economic growth underlying our estimates appears broadly 
consistent with the «revolt of the early modernists» view. 
However, it is important to take into account that our 
estimates also provide three important qualifications to this 
interpretation. the first is that the malthusian phase of 
generalized stagnation is protracted well after the end of 
the middle ages lasting approximately until the end of the 
sixteenth century. the second is that the rate of economic 
growth reached after the conclusion of the malthusian phase 
is somewhat lower than that emerging from the ongoing 
statistical reconstructions based on the output approach 
(Broadberry et al. 2011). the third is that the levels of 
GdP per capita during the late middle ages (1250-1400) 
estimated with the output approach by Broadberry et al. 
(2011) may be too low. 
sant’anna school of advanced studies, Pisa
 economic Growth in england, 1250-1850 49
Table a1. Per capita agricultural output and per capita GDP in England, 1250-1850
Year Per capita agricultural 
output (1700 = 100)
Per capita GdP 
(1700 = 100)
1250 111.10 83.92
1260 102.57 77.47
1270 87.47 66.07
1280 84.09 63.52
1290 89.29 67.45
1300 87.29 65.93
1310 87.56 65.15
1320 77.83 56.20
1330 89.87 63.03
1340 96.36 65.70
1350 114.06 75.65
1360 113.51 73.30
1370 113.03 71.10
1380 122.60 75.19
1390 129.12 77.25
1400 128.28 83.37
1410 131.12 85.42
1420 132.28 86.38
1430 136.18 89.14
1440 136.65 89.66
1450 146.65 96.46
1460 144.45 95.24
1470 142.03 93.87
1480 137.37 91.01
1490 142.65 94.73
1500 136.20 90.67
1510 138.01 92.52
1520 126.38 85.33
1530 115.83 78.76
1540 128.56 88.04
1550 119.54 82.44
1560 108.65 75.47
1570 118.05 82.57
1580 112.56 79.29
1590 103.37 73.33
1600 93.00 66.45
1610 92.38 68.41
1620 92.41 70.92
1630 92.04 73.21
1640 92.87 76.56
1650 93.21 79.64
1660 97.60 86.43
1670 102.95 94.48
1680 100.88 95.96
1690 105.78 104.29
1700 100.00 100
1710 106.75 109.91
1720 109.52 116.10
Appendix
50 alessandro nuvolari and mattia ricci
Table a1. (continued)
Year Per capita agricultural 
output (1700 = 100)
Per capita GdP 
(1700 = 100)
1730 111.22 121.39
1740 113.37 127.40
1750 114.57 132.56
1760 108.44 129.18
1770 97.40 119.47
1780 96.68 122.09
1790 95.43 124.09
1800 85.99 110.66
1810 83.23 114.72
1820 86.92 128.31
1830 90.86 143.66
1840 91.88 155.60
1850 86.61 157.09
1 the label «revolt of the early modernists» for this interpretation of the dyna-
mics of growth in preindustrial europe is due to J. de vries, «the industrial 
revolution and the industrious revolution», Journal of Economic History, vol. 
54, 1994, pp. 249-270. the term was meant to define a new perspective chal-
lenging the traditional belief of the european economy before the industrial re-
volution as an inherently stagnating, growthless system, see e. leroy ladurie, 
«motionless History», Social Science History, vol. 1, 1977, pp. 115-136. For an 
insightful discussion of this view in the case of the netherlands, see J.l. van 
Zanden, «the ‘revolt of the early modernists’ and the ‘First modern eco-
nomy’: an assessment», Economic History Review, vol. 55, 2002, pp. 619-641. 
2 this assumption is also made in some recent formal models of pre-industrial 
european economies, see for example P. sharp, H. strulik, J. Weisdorf, «the 
determinants of income in a malthusian equilibrium», Journal of Development 
Economics, vol. 97, 2012, pp. 112-117, and n. voigtlander, H.J. voth, «the th-
ree Horsemen of riches. Plague, War, urbanization in early modern europe», 
mimeo, 2011, available at http://papers.ssrn.com/sol3/papers.cfm?abstract_
id=1029347&http://www.econ.upf.edu/~voth/working_papers.html. 
3 For the period 1700-1750 the assumption of r = 1 is based on the consideration 
that england was a net exporter of corn and an importer of exotic foodstuffs 
such as sugar, coffee and tea. overall the balance of this segment of trade ap-
pears to have been as roughly in equilibrium, see r. davis, «english Foreign 
trade 1700-1774», Economic History Review, vol. 15, 1962, pp. 285-303.
4 Following P. malanima, «the long decline of a leading economy: GdP in 
central and northern italy, 1300-1913», European Review of Economic History, 
vol. 15, 2011, p. 179, we have experimented with values of b rang ing from –0.3 
to –0.7 and values of a ranging from 0.2 and 0.6 (with c = 0.1).
5 it is worth noting that our estimates are consistent with Wriglely’s view who 
considers the late sixteenth century as a fundamental turning point marking the 
emergence in england of an «advanced organic economy» leading to an accele-
ration in the rate of economic growth lasting throughout the seventeenth and 
eighteenth centuries: e.a. Wrigley, Poverty, Progress and Population, cambridge, 
cambridge university Press, 2004, pp. 44-67.
6 maddison does not provide an estimate for 1850, so the value for 1850 has 
been computed assuming a constant growth rate between the 1820 and 1870 
observations.
 economic Growth in england, 1250-1850 51
7 For a perceptive analysis of england’s «forging ahead» and italy’s «falling be-
hind» in this historical phase, see P. malanima, «la perdita del primato», Rivi-
sta di Storia Economica, vol. 13, 1997, pp. 131-172.
8 this estimate is actually based on l. levi, Wages and Earnings of the Working 
Classes, london, John murray, 1867.
References
r.c. allen, «economic structure and agricultural Productivity in 
europe, 1300-1800», European Review of Economic History, vol. 
3, 2000, pp. 1-25.
r.c. allen, «Poverty and Progress in early modern europe», Eco-
nomic History Review, vol. 56, 2003, pp. 403-443. 
r.c. allen, «engel’s Pause: technical change, capital accumula-
tion and inequality during the British industrial revolution», 
Explorations in Economic History, vol. 46, 2009, pp. 418-435. 
r.c. allen, J. Weisdorf, «Was there an ‘industrious revolution’ 
before the industrial revolution? an empirical exercise for 
england, 1300-1800», Economic History Review, vol. 64, 2011, 
pp. 715-729. 
c. Álvarez-nogal, l. Prados de la escosura, «the decline of 
spain (1500-1800): conjectural estimates», European Review of 
Economic History, vol. 11, 2007, pp. 319-366.
i. Blanchard, «labour Productivity and Work Psychology in the 
english mining industry, 1400-1600», Economic History Review, 
vol. 31, 1978, pp. 1-24. 
s. Broadberry, B. campbell, a. Klein, m. overton, B. van leeu-
wen, «British economic Growth, 1270-1870: an output Based 
approach», 2011, available at http://www2.lse.ac.uk/economic-
History/pdf/Broadberry/BritishGdPlongrun16a.pdf.
G. clark, «the Price History of english agriculture, 1209-1914», 
Research in Economic History, vol. 22, 2004, pp. 41-123. 
G. clark, «the condition of the Working class in england, 1209-
2004», Journal of Political Economy, vol. 113, 2005, pp. 1307-
1340.
G. clark, A Farewell to Alms, Princeton, Princeton university 
Press, 2007a. 
G. clark, «the long march of History: Farm Wages, Population 
and economic Growth in england, 1209-1869», Economic Hi-
story Review, vol. 60, 2007b, pp. 97-139. 
G. clark, «a review of angus maddison’s contours of the World 
economy, 1-2030 ad», Journal of Economic History, vol. 69, 
2009, pp. 1156-1160.
52 alessandro nuvolari and mattia ricci
G. clark, «the macroeconomic aggregates for england, 1209-
2008», Research in Economic History, vol. 27, 2010, pp. 51-140. 
G. clark, «malthus versus ‘the revolt of the early modernists’: 
european Growth 1200-1800», paper presented at the 9th eu-
ropean Historical economics society conference, dublin, 2011, 
mimeo.
G. clark, Y. van der Werf, «Work in Progress? the industrious 
revolution», Journal of Economic History, vol. 58, 1998, pp. 
830-843. 
n. crafts, «British economic Growth: a review of the evidence», 
Economic History Review, vol. 36, 1983, pp. 177-199. 
n. crafts, British Economic Growth During the Industrial Revolu-
tion, oxford, clarendon, 1985.
n. crafts, K. Harley, «output Growth and the British industrial 
revolution: a restatement of the crafts-Harley view», Econo-
mic History Review, vol. 45, 1992, pp. 703-730.
r. davis, «english Foreign trade 1700-1774», Economic History 
Review, vol. 15, 1962, pp. 285-303.
P. deane, m. cole, British Economic Growth, 1688-1959, cam-
bridge, cambridge university Press, 2nd edition, 1969. 
J. de vries, «the industrial revolution and the industrious revo-
lution», Journal of Economic History, vol. 54, 1994, pp. 249-270.
G. Federico, «the World economy 0-2000 ad: a review arti-
cle», European Review of Economic History, vol. 6, 2002, pp. 
111-120. 
G. Federico, P. malanima, «Progress, decline, Growth: Product 
and Productivity in italian agriculture, 1000-2000», Economic 
History Review, vol. 57, 2004, pp. 437-464. 
K. Harley, «reassessing the industrial revolution: a macro view» 
in J. mokyr (ed.), The British Industrial Revolution. An Econo-
mic Perspective, Boulder, Westview, 1993, pp. 160-205.
d.s. landes, The Unbound Prometheus, cambridge, cambridge 
university Press, 1969. 
e. leroy ladurie, «motionless History», Social Science History, 
vol. 1, 1977, pp. 115-136. 
l. levi, Wages and Earnings of the Working Classes, london, John 
murray, 1867.
e. lo cascio, P. malanima, «GdP in Pre-modern agrarian eco-
nomies (1-1820 ad). a revision of the estimates», Rivista di 
Storia Economica, vol. 25, 2009, pp. 391-419. 
a. maddison, The World Economy: A Millennial Perspective, Paris, 
oecd, 2001.
a. maddison, The World Economy: Historical Statistics, Paris, 
oecd, 2003.
a. maddison, Contours of the World Economy, 1-2030 AD, ox-
ford, oxford university Press, 2007. 
 economic Growth in england, 1250-1850 53
P. malanima, «la perdita del primato», Rivista di Storia Econo-
mica, vol. 13, 1997, pp. 131-172. 
P. malanima, «the long decline of a leading economy: GdP in 
central and northern italy, 1300-1913», European Review of 
Economic History, vol. 15, 2011, pp. 169-219.
u. Pfister, «economic Growth in Germany, 1500-1850», mimeo, 
2011. 
P. sharp, H. strulik, J. Weisdorf, «the determinants of income in 
a malthusian equilibrium», Journal of Development Economics, 
vol. 97, 2012, pp. 112-117. 
B. thomas, «Food supply in the united Kingdom during the in-
dustrial revolution», in J. mokyr (ed.), The Economics of the 
Industrial Revolution, new York, rowman & littlefield, 1985, 
pp. 138-150.
J.l. van Zanden, «the ‘revolt of the early modernists’ and the 
‘First modern economy’: an assessment», Economic History 
Review, vol. 55, 2002, pp. 619-641.
J.l. van Zanden, B. van leeuwen, «Persistent but not consistent: 
the Growth of national income in Holland 1347-1807», Explo-
rations in Economic History, vol. 49, 2012, pp. 119-130.
n. voigtlander, H.J. voth, «the three Horsemen of riches. Pla-
gue, War, urbanization in early modern europe», mimeo, 2011, 
available at http://papers.ssrn.com/sol3/papers.cfm?abstract_
id=1029347&http://www.econ.upf.edu/~voth/working_papers.
html.
H.J. voth, «the longest Years: new estimates of labour input 
in england, 1760-1830», Journal of Economic History, vol. 61, 
2001, pp. 1065-1082. 
e.a. Wrigley, «urban Growth and agricultural change: england 
and the continent in the early modern Period», Journal of 
Inter disciplinary History, vol. 15, 1985, pp. 638-728. 
e.a. Wrigley, Poverty, Progress and Population, cambridge, cam-
bridge university Press, 2004. 
e.a. Wrigley, r. schofield., English Population History from Family 
Reconstruction, cambridge, cambridge university Press, 1997.