Content uploaded by Arnaud Deseau
Author content
All content in this area was uploaded by Arnaud Deseau on Apr 18, 2023
Content may be subject to copyright.
LAND REFORM AND
PRODUCTIVITY: EVIDENCE
FROM THE DISSOLUTION OF
THE FRENCH MONASTERIES
Arnaud Deseau
LIDAM Discussion Paper IRES
2023 / 09
Land Reform and Productivity: Evidence from the
Dissolution of the French Monasteries∗
Arnaud Deseau†
March 16, 2023
Abstract
This article uses the confiscation and auction of monastic properties during the French
Revolution to assess the effects of land reallocation on agricultural productivity. To proxy
for monastic landholdings, I construct a novel dataset using the annual income and location
of more than 1,500 French monasteries in 1768. I perform several cross-checking analyses
and demonstrate the validity of the data as a proxy for monastic landholdings both at the
monastery and arrondissement levels. I show that arrondissements with greater land real-
location experienced higher levels of agricultural productivity in the mid-19th century. I
trace these increases in productivity to the creation of larger and less fragmented farms,
leading to an increase in mechanization and the substitution of family labor with a hired
specialized labor force.
Keywords: Land Reform, Productivity, French Revolution, Monasteries, Farm Size
JEL Codes: O13, O40, Q15, N53
∗I am grateful to Hugues Annoye, Simone Bertoli, Guillaume Blanc, Pierre de Callataÿ, Eric Chaney, Francesco
Cinnirella, David de la Croix, Guillaume Daudin, Yannick Dupraz, Alice Fabre, Raphaël Franck, Oded Galor,
Cecilia García Peñalosa, Paula Gobbi, Joseph Gomes, Anne Jollet, Hélène Latzer, Preston Martin Perluss, Deirdre
McCloskey, Stelios Michalopoulos, Nuno Palma, Luca Pensieroso, Èric Roca Fernández, Marc Sangnier, Mara
Squicciarini, Peter Solar, Uwe Sunde, David Weil, Jacob Weisdorf and Alexis Wilkin for their valuable comments
and suggestions. I also thank participants at the 2022 Catholic University of Louvain conference on Religion,
Culture and Economic Growth in Historical Perspective; the 2022 Lisbon University conference on the Economic
Consequences of the Age of Liberal Revolutions, 1810-1848; the 2022 World Economic History Congress in
Paris; the 2021 University of Poitiers conference on « Les Biens Nationaux » : Une Révolution ? XVIIIe – XXIe
siècle; and seminars at CERDI (University of Clermont-Ferrand), Aix-Marseille School of Economics, IRES
(UCLouvain), Brown University and University of Namur. I also thank Sébastien Angonnet, Bernard Bodinier
and Cédric Chambru for sharing their data with me and for their valuable comments.
†CEREC, Université Saint-Louis – Bruxelles, 38 Boulevard du Jardin Botanique, 1000 Bruxelles, Belgium and
IRES/LIDAM, UCLouvain, College L. H. Dupriez, 3 Place Montesquieu, B-1348 Louvain-la-Neuve, Belgium
(e-mail: arnaud.deseau@uclouvain.be).
[...] the Benedictine abbey of Saint-Germain [...] is the richest abbey in France; the abbot has
300,000 livres a year (£13,125). I lose my patience at such revenues being thus bestowed; consistent
with the spirit of the tenth century, but not with that of the eighteenth. What a noble farm would
the fourth of this income establish! What turnips, what cabbages, what potatoes, what clover, what
sheep, what wool! Are not these things better than a fat ecclesiastic? If an active English farmer was
mounted behind this abbot, I think he would do more good to France with half the income than half
the abbots of the kingdom with the whole of theirs.
Arthur Young (1792), Travels in France During the Years 1787, 1788 & 1789.
1 Introduction
Growth in agricultural productivity has long been viewed as a necessary step for economic
development and structural change (Lewis,1955;Rostow,1990;Gollin et al.,2002). Yet,
despite the availability of modern and mechanized inputs, agricultural productivity remains
remarkably low in most developing countries. A growing body of literature identifies the
misallocation of productive resources as one of the key elements explaining the agricultural
productivity gap.1In particular, the preponderance of (very) small family-operated farms is
recognized as a critical symptom of land and labor misallocation in developing countries.
The importance of the misallocation problem for agricultural productivity is well estab-
lished in today’s developing economies, but there is scant empirical evidence of its historical
importance. In particular, there is still little evidence of the policies and reforms that have en-
abled today’s developed countries to mitigate the effects of misallocation. For instance, Polanyi
(2001, p. 325) identifies the “commercialization of the soil” as a crucial step to achieving effi-
cient allocation of the land. In particular, he highlights that the “secularization of church lands
was [...] one of the chief means of the ordered transference of land into the hands of private
1See Restuccia and Rogerson (2013,2017) for a review of the literature on production factors misallocation.
1
individuals”.
In this article, I investigate a historical case of market-based land reallocation and its effect
on agricultural productivity and farm size. I study the confiscation and auction of Church land,
known as the Vente des Biens Nationaux (the Sale of National Properties), ordered by the French
Constituent Assembly in the 1790s. As a result of the Vente des Biens Nationaux, 6% of French
land and more than 170,000 buildings were reallocated by auction from the Church to secular
owners in the course of five years. According to Lecarpentier (1908, author’s translation p.
4), this was “the most important event of the Revolution”.
I focus, in particular, on the reallocation of monastic land, which represented a substan-
tial part of Church land (Bodinier and Teyssier,2000). Extensive historical evidence suggests
that the Vente des Biens Nationaux allowed rich farmers and bourgeois to create larger farms
by merging their landholdings with land confiscated from the Church and monasteries. This
typically favored the emergence of capitalist farmers with large and mechanized production
techniques. In contrast, places with less monastic land were trapped in pre-revolutionary land-
holding patterns, as they were either not affected, or only marginally affected, by the Vente des
Biens Nationaux.
To conduct the empirical analysis, I assembled a rich dataset from historical archives and
secondary sources at the arrondissement level.2My main measure of the extent of monastic
land reallocation in a given arrondissement is its initial exposure to monastic income in 1768.
This measure captures the importance of monastic lands before the Revolution at the local
level and, consequently, the extent of land reallocation through the Vente des Biens Nationaux.
To calculate monastic income exposure, I collected data on the annual income and location of
more than 1,500 French monasteries before the Revolution. To further validate this measure, I
perform two cross-checking exercises. First, using additional monastery-level historical data, I
show that the best predictor of monastic income at the monastery level is hectares of agricultural
2An arrondissement is the first-level subdivision of French départements (NUTS3 units).
2
land. Other types of properties, such as mills, houses or barns have little or no predictive power
in relation to monastic income. Second, using additional historical data at the arrondissement
level, I show that most of the variation in the percentage of Church land confiscated and sold
at auction in 1789 is explained by monastic income exposure, and this is not confounded by
other factors.
I find that those areas with higher levels of monastic land reallocation had higher levels
of agricultural productivity in the first-half of the 19th century. Regarding magnitudes, my
preferred specification indicates that a doubling in monastic income exposure in 1768 led to an
11% increase in wheat yields in 1852.
One potential source of concern in interpreting the above findings is that the initial distribu-
tion of monastic income exposure might be correlated with other factors affecting agricultural
productivity in the 19th century. I employ several strategies to alleviate this concern. First, my
analysis accounts for a large set of confounding characteristics, such as agricultural suitability,
topography and pre-Revolution development levels. I also show the robustness of my main re-
sults to other potential confounders including market potential, the confiscation of land owned
by émigrés, upper-tail human capital, literacy and religiosity. Further, my preferred speci-
fications include region fixed effects, thus exploiting the within-region variation in monastic
income exposure.3Finally, I provide pre-trends evidence, showing that arrondissements with
higher monastic income exposure did not grow faster than their counterparts in the years before
the French Revolution.
What explains the positive relationship between monastic land reallocation and agricul-
tural productivity in the mid-19th century? Following the recent literature on farm size and
misallocation in developing countries and historical evidence of land fragmentation in pre-
Revolutionary France, I first investigate the effect of the Vente des Biens Nationaux on farm
size and land fragmentation. Using département and arrondissement-level information from the
3The regions are the French NUTS1 entities.
3
Enquête Agricole of 1852, 1862 and Legoyt (1843), I show that areas with higher levels of land
reallocation had larger and less fragmented farms in the mid-19th century.
Next, I investigate two potential mechanisms linking farm size and agricultural productivity:
(i) mechanization and (ii) labor organization. First, mechanization diffused slowly through
France in the first-half of the 19th century. A key reason for this slow pace was the substantial
cost of modern physical capital that only large landowners were able to finance. By favoring
land concentration at the right tail of the farm size distribution, the Vente des Biens Nationaux
triggered an increase in land inequality. According to Galor and Zeira (1993) and Galor and
Moav (2004), inequality supports economic development when the prime engine of growth
is physical capital accumulation, as it was presumably the case in mid-19th century France.
Using data from the Enquête Agricole of 1852, I show that land reallocation at the time of the
Revolution was positively associated with investment in agricultural machines, as measured by
the number of scarifiers and extirpators in 1852.
Second, I investigate the gradual change in the composition of the agricultural labor force
along with farm size. Large farms typically substitute family labor force by hiring a specialized
male labor force. Using data from the Enquête Agricole of 1852, I find that land reallocation
was negatively associated with the share of labor by women and children required to farm one
hectare of wheat, suggesting a greater use of a male (specialized) labor force.
This article contributes to the literature on the productivity effects of land reforms. Existing
studies show mixed results, depending on the type of land reform considered. On the one hand,
land-ceiling reforms are found to have a negative effect on agricultural productivity in India
(Ghatak and Roy,2007) and the Philippines (Adamopoulos and Restuccia,2020), for example.
On the other hand, market-assisted land reforms are found to increase agricultural productivity
in Malawi (Mendola and Simtowe,2015), for example. I contribute to this literature in two
main respects. First, I contribute directly to the scant literature studying the productivity effects
of land reform in a historical context. Second, I offer an additional case study of successful
4
market-assisted land reform.
The most closely related study is that by Finley et al. (2021), which uses Church land
confiscations of the Vente des Biens Nationaux to assess the role played by transaction costs in
delaying the reallocation of property rights. The authors find a positive effect on agricultural
productivity that dissipated over the course of the 19th century. While I share the focus on the
same historical episode and some underlying mechanisms, my research stands out with its sev-
eral original contributions. First, I focus on a specific Church related entity, monasteries, that
account for most Church-held land before the Revolution (Bodinier and Teyssier,2000). Sec-
ond, I build on the data ground, providing a new dataset enabling to proxy the land reallocation
triggered by the Vente des Biens Nationaux for the entire French territory at the arrondissement
level.4Finley et al. (2021) use data from Bodinier and Teyssier (2000) covering only about
40% of the French arrondissements. Finally, I explore additional complementary mechanisms
such as land consolidation and labor organization.
This article also contributes to the literature on the relationship between inequality, invest-
ments and economic development. Galor and Zeira (1993) and Galor and Moav (2004) argue
for the non-monotonic role of equality in the process of development. When growth is driven
by physical capital accumulation, equality is detrimental to economic development, diverting
resources from individuals with a high propensity to save. On the contrary, when growth is
driven by human capital accumulation, equality promotes economic development. Most of the
literature has focused on the detrimental effect of land inequality on human capital provision
and its consequences in the context of the Second Industrial Revolution (Galor et al.,2009;
Cinnirella and Hornung,2016;Goñi,2022). By contrast, I provide evidence of the positive
effect of land inequality for economic development in the context of the First Industrial Revo-
lution, namely when basic education of the labor force was not yet a condition for economic
4This is also notable because the majority of studies looking at the determinants of French comparative develop-
ment in the 19th century are conducted at the département level (one NUTS level above the arrondissement). For
instance, Diebolt et al. (2017), de la Croix and Perrin (2018) and Franck and Galor (2021) study the interactions
between education, fertility and long-run development at the département level.
5
growth (Galor and Moav,2006).5Consistent with this view, I provide evidence that, up to
the first half of the 19th century, the reallocation of monastic lands triggered both an increase
in land ownership inequality and an increase in physical capital and agricultural productivity.
Finally, this article also relates to a broader literature analyzing the economic consequences
of the secularization of society through the dissolution of Church related entities such as monas-
teries. In particular, my study is closely related to Heldring et al. (2021) and Cantoni et al.
(2018), who analyze the economic consequences of the 16th-century dissolution of English
and German monasteries, respectively. In both cases, the authors argue that the dissolution of
monasteries triggered an efficient reallocation of resources from religious to secular purposes,
thereby promoting economic development. I contribute to this literature by exploring the eco-
nomic consequences of the dissolution of the monasteries in the French case. Although in a
different context and epoch, I reach similar conclusions to those in Heldring et al. (2021) and
Cantoni et al. (2018), showing that the dissolution of French monasteries and the privatization
of their lands promoted economic development.
The remainder of the article is organized as follows. In Section 2, I provide the necessary
historical background, including an overview of the changes observed in French agricultural
productivity and landholding patterns before and after the French Revolution. In Section 3, I
introduce my measure of monastic landholdings, explain my empirical strate, and discuss the
main threats regarding identification. I present and discuss my main results on the effect of land
reallocation on agricultural productivity in Section 4. In Section 5, I explore the mechanisms
driving my results, and in Section 6, I conclude.
5Related to that literature and my focus on French agricultural productivity, Bignon and García-Peñalosa
(2021) show that a tariff on cereals (the Méline tariff of 1892) reduced primary school enrollment and increased
fertility, thus slowing French economic development and industrialization in the second-half of the 19th century.
6
2 Historical Background
In this section, I provide some historical background on French agriculture, monastic land-
holdings and the Vente des Biens Nationaux. I begin by discussing the evolution of French
agricultural productivity and landholding patterns before and after the Revolution. Then, I
discuss the importance of monastic land before the Revolution and the changes prompted by
the Vente des Biens nationaux in terms of farm size.
2.1 Agricultural Productivity and Landholding Patterns Before and After
the French Revolution
Evidence indicates that French agricultural productivity began to rise consistently during the
first half of the 19th century. Newell (1973), analyzing the historical series compiled by
Toutain (1961), shows that the French agricultural output per worker started to rise in the
1820s. By contrast, the pre-revolutionary and the Napoleonic periods were characterized by
stagnation in agricultural productivity (Newell,1973;Allen,2000;Hoffman,2000).6The
overall rise in agricultural productivity was rapid. Bairoch (1988) estimates its average annual
growth of 1.1% between 1830 and 1880. This is higher than the rate of growth in the United
Kingdom (0.7%) and the European average (0.6%) over the same period. The strong rise
in agricultural productivity was seen across all French départements and all major crop types
however, there were remarkable differences across regions (Newell,1973). For instance, be-
tween 1800 and 1862, labor productivity in wheat production, as measured by man-days per
hectolitre, increased 10pp faster in the north of France than in Brittany (Grantham,1993).
Historians have put forward two main hypotheses to explain the rise of French agricultural
productivity during the first half of the 19th century: (i) technical innovations and (ii) organi-
zational changes. In respect of the first, this period is marked by several important agricultural
6Hoffman (2000) finds very low growth in total factor productivity in agriculture before the Revolution, of
the order of 0.1% per year at most.
7
innovations that diffused gradually within France. More efficient crop rotation systems replaced
the three-field or two-field rotation systems established in the Middle-Ages.7More powerful
fertilizers were also available, such as Peruvian guano and, from the 1840s on, artificial fer-
tilizers. Finally, this period was also marked by the gradual adoption of the first agricultural
machines, for example, threshers and harvesters.
Despite the importance of these innovations for agricultural work, it should be noted that
their slow diffusion meant that their actual impact likely remained limited for a long time in
some parts of the French territory. Their cost, as well as reluctance to change (because of
“traditional mentalities”), seem to have presented significant obstacles to the adoption of such
innovations for a large share of agricultural exploitations. Sée (1927) estimates that modern
agricultural production techniques had only achieved total dissemination over the whole French
territory by the second half of the 19th century.
The low investment in available new technologies and productivity differences within France
after the Revolution can also be explained by landholding patterns. Pre-Revolution France was
characterized by the dominance of small landowners. Peasants, while representing 90% of the
landowners, owned only about 40% of the French land before the Revolution (Sée,1925).
Hoffman (2000) notes that in the village of Goincourt, north of Paris, only 3% of the farmers
owned more than 10 hectares in 1717; 96% owned less than 2 hectares. Similar patterns are
evident for 18th century Normandy and in the South of France. A pattern confirmed also
in the North of France, a heavily agricultural region, where 60-70% of peasants possessed less
than one hectare (Lefebvre,1972, p.37).8
It was not simply that landowners had few hectares available to farm. What also kept
agricultural productivity low before the Revolution was the fragmentation of landholdings. For
7In particular, these new systems replace the unproductive land in fallow by artificial prairies (prairies artifi-
cielles) of forage crops such as clover, alfalfa or sainfoin. This has the double advantage of fixing nitrogen in the
soil while providing forage for farm animals, allowing for better fertilization of the soil with manure.
8Vigneron (2008) gives similar figures for the Cambrésis and Lille province with properties of less than one
hectares representing more than 50% of the landholdings in 1751.
8
a farmer, owning 10 hectares of land did not mean that those 10 hectares were concentrated
in a continuous stretch of land. Rather, the average farmer was likely to own several parcels
of few hectares each, physically separated and distant from each another. Figure 1illustrates
the pre-Revolution land fragmentation by showing the agricultural plots in the village of Athis-
Mons in the Essonne département in 1750 (Moriceau,2002). Each color indicates a different
owner, revealing sizeable land fragmentation.
The influence of small landowners on French agriculture remained strong after the French
Revolution. The first comprehensive data on landholdings after the Revolution shows that
in 1862 half of all agricultural exploitations were under 5 hectares. However, it is also worth
noting that there was substantial variations in landholding patterns across départements, notably
at the top of the distribution; the same source indicates that, in the mid-19th century, 25% of the
farms were above 10 hectares, which is usually considered the threshold to be a large agricultural
exploitation.
Large and consolidated farms were key to raising agricultural productivity for several rea-
sons. There are several reasons for this. First, an increase in farm size led to a gradual change
in the composition of the agricultural labor force, from family labor toward the hiring of spe-
cialized laborers. Indeed, small agricultural exploitations traditionally relied on labor supply
from the family (the head of the family, as well as wife and children).9Task specialization was
limited, with all members of the family performing the various farm tasks required (plowing,
sowing and harvesting).
Large farms were able to employ specialists and day laborers (journaliers), each dedicated
to specific tasks: this specialization enabled larger farms to employ fewer workers per hectare,
thus increasing labor productivity. Allen (1988) shows that the change in labor-force com-
position and the size of farms explains the rise in labor productivity for 18th-century English
agriculture. In a modern context, Adamopoulos and Restuccia (2014) also find substantial labor
9Nuclear family members were the most common source of labor on small farms. Labor from extended family
was also present, but could rarely provide sufficient labor for large farms (Hoffman,2000, p.48).
9
productivity differences by farm size using the 2007 US Census of Agriculture. In particular,
they find that value added per worker is more than doubled when moving from the smallest
farms (0.5-5 hectares) to what, in the present case, would be considered a large farm (30-40
hectares).10
Large farms are also more productive because of increasing returns to mechanization with
increased farm size, allowing large agricultural exploitations further savings on labor costs and
increased labor productivity. As pointed by Foster and Rosenzweig (2011,2022), in the context
of India in the 2010s, large machines cannot be used at their full capacity on small farms or
plots. In the present case, the relationship between farm size and mechanization is supported
by the historical study of Hoffman (2000, p.36), who finds that before the Revolution, farms
under 5 hectares did not invest in basic capital, such as plows and horses.11
2.2 Monastic Land, the Vente des Biens Nationaux and Farm Size
Before the Revolution, monasteries (and the Church more generally) were among the largest
landowners in France. Historians estimate that they possessed as much as 5-6% of the French
land while representing only 1.8% of the adult male population in 1789 (Lecarpentier,1908;
Sée,1925). Church and monastic lands were unevenly distributed over the French territory.
The detailed analysis by Bodinier and Teyssier (2000) of over 40% of French districts shows
that in 1789, 4.4% of the territory of the median district was owned by monasteries and the
Church.12 The top (bottom) quartile was composed of districts with more (less) than 8%
(1.9%) of their land held by monasteries and the Church, with the maximum being reached in
the district of Cambrai (40.1%) and the minimum in the district of Tartas (0.3%).
10This pattern holds also in various developing countries (Cornia,1985).
11Hoffman (2000, p.286) cites numerous studies showing that the median farm size to own a plow was 10
hectares.
12Districts were the initial first-level subdivision of the French departments created after the Revolution. They
were replaced by the arrondissements in 1800. Districts were more numerous than arrondissements (534 districts
in 1790 for 364 arrondissements circa 1850). They were therefore smaller on average.
11
Monasteries received large parcels of agricultural land from patrons during the Middle Ages
and were a key component of the Church’s landholdings. Bodinier and Teyssier (2000, p.339)
show that, at the time of the Revolution, monasteries held around 60% of Church land. In
fact, the presence of a single powerful monastery in one district could account for as much
as 20 to 30 % of Church land in that district, with some notable exceptions reaching even
higher figures. This was the case for the famous abbey of Cluny (46%), Saint-Sever (32.2%),
Jumièges (23.7%) or Fontevraud (21.2%) (Bodinier and Teyssier,2000, p. 341). Despite the
significance of these figures, Bodinier and Teyssier (2000) recognize that they are undoubtedly
underestimated as powerful monasteries typically held additional land outside of their district
of origin. For example, the Parisian abbey of Saint-Germain-des-Prés owned land across the
Ile-de-France region and even in Normandy, that is to say, far beyond its original constituency
(Bodinier and Teyssier,2000, p. 343).
The French Revolution brought the Church’s dominance in landholding to an abrupt end.
On November 2, 1789, a law was passed to confiscate and auction all Church properties,
including monastic properties. This decision, largely unexpected by the public, came as a means
to pay off the debts accumulated by the monarchy.13 This historical event, known as the Vente
des Biens Nationaux (the Sale of National Properties), saw 6% of French land and more than
170,000 buildings reallocated from the Church to secular owners auction; more than 700,000
Church properties were sold. According to Lecarpentier (1908, author’s translation p. 4), this
was “the most important event of the Revolution”.
The Vente des Biens Nationaux triggered a vast reallocation of land, enabling rich farmers and
bourgeois to create large agricultural exploitations by merging their lands with those confiscated
from the Church and monasteries. As underlined by Tocqueville (1967, author’s translation
p. 89), most of the lands “were purchased by people who already owned other lands; so that,
13The Vente des Biens Nationaux is closely linked to the creation of bonds backed on the confiscated property,
called assignats. In 1791, these bonds became a fiat currency, before collapsing due to hyperinflation. The assignat
was finally abolished in 1797.
12
if the property changed hands, the number of owners increased much less than one might
imagine.”14 This view has been confirmed by the detailed historical analysis of Bodinier and
Teyssier (2000). In most of the districts, a small number of rich farmers and bourgeois acquired
most of the land. For example, in the district of Bernay, 27 members of the grand bourgeoisie
succeeded in buying 39% of the Church land while representing only 4% of the buyers. Small-
scale peasants farmers, on the other hand, were unable to acquire a significant amount of land
through the auction process because of their limited capacity to bid against wealthier bourgeois
and large farmers.15
The Vente des Biens Nationaux thus also represented an increase in land inequality. This ef-
fect is clearly seen in the evolution of farm-size distribution in the Artois region before and after
the Revolution (Jessenne,1987); the case study shows that the Vente des Biens Nationaux corre-
sponded to an increase in the right-tail of the farm size distribution (Figure A-1 in Appendix).
Notably, small agricultural exploitations (between 5 and 9 hectares) completely disappear.16
The implications of monastic land reallocation through the Vente des Biens Nationaux for
farm size and the concentration of landholdings are illustrated in Figure 2, the parcels held
by Alexandre Le Bourlier d’Orgeval, the largest landowner in Athis-Mons, before and after
the French Revolution. His pre-Revolution landholdings are represented in blue and green,
and the land acquired through the Vente des Biens Nationaux is represented in red. As the
map reveals, the Vente des Biens Nationaux enabled this pre-Revolution landowner to increase
his already considerable estate by acquiring almost 50 hectares from the Cistercian abbey of
Vaux-de-Cernay.17 This represented a of 30% increase and consolidation of his pre-Revolution
14This view is also defended notably by Lecarpentier (1908), Marion (1908) and Jaurès (1924).
15Another reason was the fact that buyers had to travel to the district or département administrative capital to bid.
This further limited the capacity of small landowners to acquire land as they faced a relatively high transportation
cost Bodinier and Teyssier (2000, p.228).
16Unfortunately, this study does not provide the evolution of the smallest agricultural exploitations – i.e. below
5 hectares. However, as explained in this section, there is a good chance that very small farms remained unchanged
after the Vente des Biens Nationaux because: (i) they were not confiscated and (ii) poor landowners were not able
to acquire land through the auctions.
17The abbey of Vaux-de-Cernay is located 34 kilometeres away from Athis-Mons.
13
Figure 2: Land Consolidation and the Vente des Biens Nationaux in the village of Athis-Mons
Notes: This figure shows the changes in Alexandre Le Bourlier d’Orgeval’s parcels before and after the French
Revolution (Moriceau,2002). In particular, his pre-Revolution landholdings are represented in blue and green,
while his land acquisition through the Vente des Biens Nationaux is represented in red.
agricultural domain.
14
3 Data and Empirical Framework
In my empirical analysis, I combine various datasets at the arrondissement level. An arrondisse-
ment is the first-level subdivision of French départements.18 The arrondissements were created in
1800 and replaced the districts initially created after the Revolution. Importantly, the number
and boundaries of the arrondissements were stable during the 19th century. There were 364
arrondissements at the time of our analysis (circa 1850), of which 354 are in our main sample.19
The average size of an arrondissement in our study was 1,435 square kilometers, with a standard
deviation of 573.
I begin by presenting my main explanatory variable – i.e. monastic income exposure – as
a proxy for monastic land reallocation at the arrondissement level. Then, I present my main
estimating equation and discuss the potential threats to my identification strate.
3.1 Monastic Income Exposure and Monastic Lands
My main explanatory variable is monastic income exposure in 1768 at the arrondissement level.
I use it as a proxy for the importance of monastic landholdings in a given arrondissement at
the time of the Revolution. Ideally, I would like to have information on the size and location
of each parcel belonging to a monastery in France at the time of the Revolution to study the
effect of land reallocation. Unfortunately, such data is scarce and only available for certain
monasteries. Monastic income exposure combines data on the annual income and location of
French monasteries in 1768 from three sources: the France Ecclésiastique, the Almanach Royal
and Lecestre (1902); see Appendix Cfor a complete discussion and details about monastic
income data. Monastic income exposure is defined as follows:
18The départements correspond to the NUTS3 units.
19The 10 missing arrondissements are due either to missing data from the monastic income side (the 6 Corsican
arrondissements) or from the Enquête Agricole side (Bourganeuf, Béziers, Grasse and Paris).
15
Monastic Income Exposurea=
m
1/da,m
a1/da,m
·Im,(1)
with da,m the kilometric distance between the centroid of arrondissement aand the location
of monastery m, and Imthe annual income of monastery m. In this form, the monastic in-
come exposure of arrondissement ais a weighted average of all monastic incomes, with weights
corresponding to relative inverse distance.20
My measure of monastic income exposure exploits two facts about monastic landholding
patterns to proxy the extent of land owned by monasteries in each arrondissement. First, the
probability of an arrondissement hosting the land of a given monastery decreases with distance.
Historians have pointed to the decreasing concentration of monastic properties as one moves
away from the cloister (Bodinier and Teyssier,2000;Goudot,2006;Wilkin,2011).21 This is
captured in (1) as monastic income exposure of a given arrondissement decreases with an in-
crease in distance to a monastery (∂M I Ea/∂da,m <0). Most monasteries have been founded
and endowed by the local nobility, meaning that most of the monastery’s land was located
in neighbouring arrondissements.22 Second, the amount of land owned by a monastery in an
arrondissement increases with monastic income. This is because monasteries were powerful
landowners and consequently derived a substantial part of their income from agriculture (see
Section 2.2). This is also captured in (1) as the monastic income exposure of a given arrondisse-
ment increases with respect to monastic income (∂M I Ea/∂Im>0).
To validate my proxy of monastic landholdings, I perform two empirical exercises using
additional historical data, one at the monastery level and one at the arrondissement level. In the
20This ensures that each monastery’s income is distributed at 100% across arrondissements. In spirit, this is close
to computing a spatially lagged variable with inverse distance and row-standardized weights, an approach widely
used in the spatial econometrics literature (Anselin,2001).
21The cloister is the main monastic building where the monks live.
22Nevertheless, famous monasteries were receiving donations coming from hundreds of kilometres away. For
example, the abbey of Marmoutier, one of the oldest and most famous Benedictine abbeys, had about 200 priories
in the 17th century that were located in 29 different départements (Carré de Busserolle,1882, p. 181-191). Five
priories of the abbey of Marmoutier were even located in England.
16
first cross-checking exercise, I explicitly test the link between monastic income and the amount
of land owned by monasteries in 1789 at the monastery level. To do so, I use data compiled by
Bodinier (1988) on the number, size and type of properties owned by French monasteries in the
Eure or Seine-Maritime département (NUTS3 level) on the eve of the Revolution.23 From this
dataset, I can, therefore, compute the number of hectares of agricultural land, woods, vineyards
or wasteland owned by 45 monasteries located in the two départements along with information
on other economic assets such as mills, houses, markets, justice courts, and chapels.
Table D-1 in the Appendix examines the relationship between monastic income in 1768 and
hectares of agricultural land owned by monasteries in 1789. Across all specifications, hectares of
agricultural land appears to be a strong and robust predictor of monastic income. Specifically,
hectares of agricultural land held explains half of the variations in monastic income in the
bivariate regression. On the contrary, other types of properties, such as mills, houses or barns,
have little or no predictive power in respect of monastic income. This relationship holds also
when considering the number of farms rather than hectares of agricultural land owned by
monasteries in 1789 (Table D-2).
The second cross-checking exercise is a direct test of the ability of monastic income expo-
sure to capture monastic landholdings at the arrondissement level. For that exercise, I use data
collected by Bodinier and Teyssier (2000) on the percentage of Church land redistributed in
French districts through the Vente des Biens Nationaux. These data are available for only 40%
of the French arrondissements referenced in my main analysis. As established in Section 2.2,
monasteries were large landowners, accounting, on average, for 60% of Church land in the
various districts (Bodinier and Teyssier,2000). Consequently, I expect a large and positive
relationship between the two variables.
I proceed in two steps. First, I use bivariate regressions of the percentage of Church land
redistributed in 1789 on monastic income exposure with different distance cutoffs to calibrate
23I warmly thank Bernard Bodinier for having shared his data with me. The data come from his doctoral thesis
and his personal notes for the Seine-Maritime département.
17
my inverse distance weights in (1). As shown in Appendix Table D-3, the fit between the
percentage of Church land redistributed in 1789 and monastic income exposure is maximized
for a distance cutoff of 100km.24 Specifically, monastic income exposure explains half of the
variation in the percentage of Church land redistributed in 1789 (column 5). Figure 3de-
picts that strong bivariate relationship. By contrast, the worst fit is found in column 1 where
monastic income exposure is defined as the sum of monastic income at the arrondissement level
(MIEa=m∈aIm) – that is, ignoring spatial spillovers. In that case, monastic income ex-
posure explains only 16% of the variation in the percentage of Church land redistributed in
1789.
Figure 4shows monastic income exposure for each arrondissements by decile. My proxy of
monastic landholdings is consistent with Bodinier and Teyssier (2000, p.335), who note that
Church properties were concentrated in the North-Eastern France (Brittany excluded) and, in
particular, above a line from Nantes to Belfort. Significant variations within each region are
accurately depicted.
Second, I verify the ability of monastic income exposure to explain the percentage of Church
land redistributed in 1789 in the presence of confounding variables. For instance, arrondisse-
ments with fertile land for agriculture could have given more land to the Church and hosted
richer monasteries. This is an endogenity issue that I address in more details in the next section.
As set out in Table D-4 in the Appendix, monastic income exposure remains the best pre-
dictor of the percentage of Church land redistributed in 1789 across all the specifications. In
particular, the correlation between monastic income exposure and the percentage of Church
land redistributed in 1789 stays positive and highly significant, controlling for the agricultural
potential of the land, ruggedness, urbanization, distance to bishoprics, and region fixed effects.
Figure A-2 shows that the relationship identified in column 6 is not influenced by outliers.
24In the rest of the paper, I will use the term monastic income exposure to designate monastic income exposure
computed with (1) and a 100km cutoff.
18
Figure 3: Correlation between the percentage of Church Land Redistributed during the Vente
des Biens Nationaux and Monastic Income Exposure
-10 0 10 20 30
Percentage of Church Land Redistributed in 1789
-2 -1 0 1 2
Log(Monastic Income Exposure)
coef = 6.174636, (robust) se = .77223888, t = 8
Notes: This figure plots the relationship between the percentage of Church land reallocated through the Vente des
Biens Nationaux and log monastic income exposure in 1768. Residuals and coefficient estimates from Table D-3,
column 5.
3.2 Estimation Framework
My main specification estimates the cross-sectional relationship between monastic land realloca-
tion triggered by the French Revolution and agricultural productivity in the mid-19th century.
I estimate the following ordinary least squares (OLS) specification:
Ya,1850 =β·Monastic Income Exposurea,1768 +γ′Xa+αr+εa,(2)
19
Figure 4: Spatial Distribution of Monastic Income Exposure
Regions
Arrondissements
MonasticIncomeExposuredecile
inlivrestournois(1768)
3140-16863
16863-20403
20403-24188
24188-27649
27649-34163
34163-40691
40691-49726
49726-62015
62015-98281
98281-497230
N
Notes: This figure plots monastic income exposure in 1768 by decile. See text for more informations on the
construction of monastic income exposure.
where a= 1, ..., N represents an arrondissement, and the dependent variable, Y, represents
agricultural productivity, typically measured circa 1850. The right-hand side is composed of our
variable of interest – monastic income exposure in year 1768 – proxying Revolution-era land
reallocation, a vector of control variables Xa,t, region fixed effects αr, and an idiosyncratic error
term εa. Throughout my analysis, I report robust standard errors for regression coefficients,
clustered at the département level. I address potential spatial correlation issues by systematically
reporting Conley (1999) standard errors. In particular, I use a bandwidth of 100km together
with a Bartlett kernel. Typically, I find that Conley standard errors do not differ significantly
20
from the standard errors clustered at the département level.
The main concern is the potential endogenity of monastic income exposure; that is, rich
monasteries may be located in places that were inherently more favorable to economic devel-
opment. For instance, arrondissements with land that is suited for agriculture could be more
productive and host richer monasteries. My main strate to deal with this issue is to use several
relevant control variables to capture development differences prior to 1789. First, I explicitly
take account of differences between arrondissement in the initial suitability of their land for
agriculture using the Caloric Suitability Index of Galor and Özak (2016).
I then control for the ruggedness of the terrain using elevation data from CGIAR-CSI
SRTM (Jarvis et al.,2008). Terrain ruggedness captures a broad range of factors affecting eco-
nomic development, for example, transportation and trade. Importantly for my analysis, irreg-
ular terrain is difficult to farm, making land fragmentation more likely and directly impacting
agricultural productivity. Third, I control for pre-revolutionary differences in economic devel-
opment using urban population levels in 1750 from Buringh (2021). Fourth, in my preferred
specification I include 21 region fixed effects, identifying the effect of land reallocation using
the within-region variation. The 21 regions correspond to administrative boundaries prevailing
before the 2015 reform; they are closer to Pre-Revolutionary French provincial boundaries.25
Finally, I test the robustness of my main results to several other potential confounders, includ-
ing market potential, confiscation of land owned by émigrés, upper-tail human capital, literacy
and religiosity.
4 Main Results
In presneting my main results, I first focus on the effect of land reallocation on agricultural
productivity. I then test the robustness of my main results to other potential confounders and
25For example, Lorraine, Alsace and Champagne are now part of the same region, whereas before 2015 they
were three separate regions that corresponded more closely to the provinces before the Revolution.
21
to outliers.
4.1 The Effect of Land Reallocation on Agricultural Productivity
In Table 1, I report the estimations for specification (2) using OLS. The three dependent vari-
ables used to measure agricultural productivity are wheat yields (columns 1-3), the average days
required to farm one hectare of wheat (columns 4-6), and the daily agricultural wage (columns
7-9). I first present the bivariate relationship for each dependent variable (columns 1,4 and 7)
and then include my main control variables (columns 2, 5 and 8). Finally, I add region fixed
effects to isolate the effect of land reallocation using the within-region variation (columns 3, 6
and 9).
The results show that arrondissements in which there was more extensive land reallocation,
as proxied by monastic income exposure in 1768, experienced higher levels of agricultural
productivity in 1852. The relationship remains valid across all specifications, regardless of the
measure of agricultural productivity, and is economically significant. For instance, column 3
suggests that a doubling of the monastic income exposure is associated with a 11% increase in
wheat yields, conditional on my main controls and region fixed effects.26 I find similar effects
using alternative measures of agricultural productivity. Columns 6 and 9 suggest that a doubling
in monastic income exposure leads to a 10% decrease in the average number of days required
to farm one hectare of wheat and a 9% increase in the daily agricultural wage, conditional on
my main controls and region fixed effects. The magnitude of the effect is in line with Finley
et al. (2021).
These results, taken together, are consistent with a better land allocation among farmers
following the Vente des Biens Nationaux and thus with an increase in agricultural productivity.
26A doubling in monastic income exposure corresponds to a one-standard-deviation (5.76) increase in the per-
centage of Church land redistributed in 1789. See column 5 in Table D-4.
22
Table 1: The Effect of Monastic Land Reallocation on Agricultural Productivity
Dependent variable: log(Wheat yields) log(Days per hectare of wheat) log(Agricultural wage)
(1) (2) (3) (4) (5) (6) (7) (8) (9)
log(Monastic Income Exposure) 0.26 0.24 0.11 -0.14 -0.15 -0.10 0.15 0.18 0.09
(0.029)*** (0.025)*** (0.044)** (0.039)*** (0.037)*** (0.038)*** (0.055)*** (0.057)*** (0.040)**
[0.030]*** [0.027]*** [0.040]** [0.038]*** [0.037]*** [0.036]*** [0.058]*** [0.059]*** [0.036]**
Caloric suitability No Yes Yes No Yes Yes No Yes Yes
Ruggedness No Yes Yes No Yes Yes No Yes Yes
Urban population in 1750 No Yes Yes No Yes Yes No Yes Yes
Region fixed effects No No Yes No No Yes No No Yes
Observations 354 354 354 354 354 354 354 354 354
Adjusted R20.41 0.49 0.58 0.17 0.20 0.47 0.12 0.16 0.69
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, on agricultural productivity in 1852 at the arrondissement
level. I use three different measures of agricultural productivity as dependent variable: wheat yields (columns 1-3), the average number of days required to farm one hectare of wheat (columns
4-6) and daily agricultural wage (columns 7-9). See Section Bof the Appendix for more details on the variables used. For each dependent variable, I first display the bivariate relationship in
the first column, then I include my main set of controls (caloric suitability of the land, ruggedness and urban population levels in 1750) and finally I add region fixed effects. Standard errors
clustered at the département level are in parentheses and Conley (1999) standard errors, with a Bartlett kernel and a cut-off distance of 100km, in brackets. * p<0.1, ** p<0.05, *** p<0.01.
23
4.2 Robustness
In this section, I present the results of various robustness checks. Each column of Table 2
introduces the additional control variable specified at the top of that column to the main control
variables and region fixed effects. The results for the dependent variables referenced in Table
1, namely wheat yields, labor days to farm one hectare of wheat, and daily agricultural wage,
are set out in rows 1–3, respectively.
My main concern in this analysis is the endogeneity of monastic income and, in particular,
the possibility that an omitted factor determines agricultural productivity and monastic income
simultaneously before and after the Revolution. In columns 1 and 2, this issue is addressed using
two different measures of economic development at the time of the French Revolution. First,
market potential, by representing the potential demand for agricultural products, can influ-
ence both monastic income and farmers’ incentives to supply agricultural products efficiently.
I tackle this issue by constructing a measure of the market potential of each arrondissement
at the time of the Revolution using the first comprehensive census of the French population,
conducted in 1794.27 This is a powerful measure of economic development since it is a compre-
hensive assessment of the size of French municipalities at the time of the French Revolution;
it is included as a control in column 1. In column 2, I introduce an alternative measure of the
level of economic development in 1789. Using data from Daudin (2010) and following Franck
and Galor (2022), I proxy early market integration by computing the number of firms that
sold their products outside of their home arrondissements in the 1790s. As reflected in columns
1 and 2, my main results are robust to the inclusion of these two potential confounders.
Another related concern is the endogeneity of monastic locations. It is possible that early
monasteries targeted locations with higher levels of development to attract more donations com-
27I define market potential for arrondissement aas the distance-weighted sum of the population of all French
cities: MPa= [j1/dac ·P opc], where P opcis the population of city cin 1794 and dac is the kilometric distance
between the centroid of arrondissement aand city c. I consider as a city all municipalities with 1,000 or more
inhabitants in 1794. When population is not available for year 1794, I use information of the next census of 1800.
This is the case for only 2.3% of French municipalities.
24
Table 2: Robustness Checks
Added control: Market
Potential
in 1794
Market
Integration
in 1790s
Urban
Population
in 700
%
Emigré
Literacy
in
1786
Subs.
Density
%
Refractory
Priests
Banks
in 1850
Distance
to
Paris
Distance
to
Bishoprics
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Wheat Yields 0.15 0.11 0.10 0.12 0.13 0.11 0.13 0.11 0.11 0.11
(0.041)*** (0.044)** (0.045)** (0.039)*** (0.045)*** (0.044)** (0.047)*** (0.044)** (0.044)** (0.044)**
[0.037]*** [0.040]** [0.041]** [0.035]*** [0.040]*** [0.041]** [0.042]*** [0.040]** [0.040]** [0.040]**
Days per hect. -0.11 -0.10 -0.11 -0.10 -0.08 -0.10 -0.11 -0.10 -0.08 -0.10
(0.039)*** (0.038)** (0.039)*** (0.038)*** (0.033)** (0.038)** (0.040)** (0.038)*** (0.038)** (0.039)**
[0.038]*** [0.036]** [0.036]*** [0.036]*** [0.032]** [0.035]** [0.038]** [0.036]*** [0.035]** 0.036]**
Agricultural Wage 0.05 0.08 0.08 0.09 0.10 0.08 0.08 0.08 0.05 0.09
(0.046) (0.039)** (0.041)** (0.040)** (0.039)** (0.039)** (0.041)** (0.040)* (0.038) (0.040)**
[0.042] [0.035]** [0.037]** [0.036]** [0.035]** [0.036]** [0.036]** [0.037]* [0.035] [0.036]**
Controls Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Region fixed effects Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Observations 354 354 354 354 333 354 345 354 354 354
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, agricultural productivity in 1852 at the arrondissement level.
In each column, I test the sensitivity of my results to an additional control variable that is specified in the column header. See Section Bof the Appendix for more details on the variables used.
I use the same three dependent variables as in Table 1to measure agricultural productivity. Each column controls for the caloric suitability of the land, ruggedness, urban population levels in
1750 and region fixed effects. Standard errors clustered at the département level are in parentheses and Conley (1999) standard errors, with a Bartlett kernel and a cut-off distance of 100km, in
brackets. * p<0.1, ** p<0.05, *** p<0.01.
25
ing from rich patrons. In column 3, I control for urban population levels in year 700; this is the
year of the earliest population figure available for France using the data from Buringh (2021).28
My main results remain unchanged, alleviating the selection concern. In Appendix Table E-1
and E-2, I verify further the robustness of my results to that potential issue, controlling for
urban population in years 800, 900, 1000 and 1100. I obtain similar results.
I have also to consider that the French Revolution not only confiscated and auctioned
Church properties, but also émigrés properties. Emigrés were supporters of the former regime,
mostly aristocrats and churchmen, who fled France at the time of the Revolution (Greer,1951;
Franck and Michalopoulos,2017). Consequently, there is a possibility that my results are influ-
enced by differences in the reallocation of émigrés land. I mitigate that concern by controlling
for the share of émigrés in the population at the département level, using data from Greer (1951),
as set out in column 4. My main results are unaffected.
In columns 5 and 6, I account for the possibility that my results are driven by differences
in levels of human capital before the Revolution. First, upper-tail human capital, as captured
by the density of Encyclopédie subscribers, might have affected the adoption of agricultural
innovations and, as a result, agricultural productivity (Squicciarini and Voigtländer,2015).
Another possibility is that initial literacy levels were an determining factor in the adoption of
agricultural innovations after the Revolution; in both cases, my main results are unaffected.
Another possibility is that the concentration of monasteries in certain places was linked to
religiosity. This might have pushed arrondissements to specialize in the agricultural sector by
hindering the diffusion of knowledge and innovation in other sectors. To assess that possibility,
I follow Squicciarini (2020) and include as a control variable the share of refractory priests in
1791. Refractory priests were priests who refused to swear the oath of allegiance to the Civil
Constitution of the newly formed French Republic. This expression of loyalty to the Catholic
28The year 700 also corresponds to a period when relatively few monasteries were present in France. It is
before the appearance of crucial monastic reforms which will lead to the foundation of the majority of French
monasteries, such as the order of Cluny (910), the Cistercians (1098) or the Premonstratensians (1120).
26
Church is a proxy for religiosity at the local level (Tackett,1986). As column 7 reveals, my
main results are unaffected by religiosity.
As established in Section 2.1, the diffusion of agricultural machines during the first-half of
the 19th was slow due to their cost. Consequently, access to financial services and especially
credit is another factor potentially driving my results. To test the potential influence of financial
development, in column 8, I control for the number of banks operating in each arrondissement
between 1800 and 1851; my results are relatively unaffected.
Next, I investigate whether my main results are affected by the distance of each arrondisse-
ment to Paris and to bishoprics in 1789. First, as noted by Tocqueville (1967), the admin-
istrative and economic dominance of Paris in relation to the rest of the country was evident
as early as the 17th century. This opens the possibility that proximity to the French capital
simultaneously affected economic development and monastic income exposure. Second, in the
course of the Vente des Biens Nationaux, all types of Church property were confiscated and
auctioned. Even though monastic land represented the majority of confiscated properties (see
Section 2.2), the possibility remains that my results are influenced by the confiscation of the
property of other religious institutions such as bishoprics and archbishoprics. I find that my
main results are robust to the distance to Paris (column 9) and to bishoprics and archbishoprics
(column 10).
Appendix Esets out my additional robustness checks. First, I test the robustness of my
results to the inclusion of other meaningful geographical distances capturing potential diffusion
of technologies or trade opprotunities. In Table E-3, I establish that the distance to London,
Fresnes-sur-Escaut and major French harbors (Rouen, Nantes, Bordeaux and Marseilles) are
not confounding my main results. I then check whether my results are driven by extreme ob-
servations by rerunning my analysis and trimming the top and bottom 5% of monastic income
exposure. Table E-4 shows my main results are stable.
As an additional endogeneity test, I verify whether arrondissements with higher levels of
27
monastic income exposure were on specific trends before the Revolution. My concern is that,
despite the rich set of control variables employed in the main analysis, arrondissements with
higher monastic income exposure systematically differed in key characteristics affecting eco-
nomic development and were already growing faster before the Revolution – i.e. where on a
different trend than their counterparts. The only data available to assess this possibility is urban
population data from Buringh (2021).
Table 3presents regressions of rates of urban population growth in different periods on
monastic income exposure in 1768. I find no consistent pattern systematically relating monas-
tic income exposure and urban population growth prior to the French Revolution. In all
specifications, the effect of monastic income exposure on urban population growth is small and
statistically insignificant. This provides clear evidence that arrondissements with higher monas-
tic income exposure were on the same general economic development path up to two centuries
before the French Revolution. In particular, column 1 shows that there is no statistically sig-
nificant effect of monastic income exposure on urban population growth fifty years before the
Revolution; if any, the effect seems to be small and negative.
Table 3: Monastic Income Exposure and Trends before the French Revolution
Dep. var.: Urb. Pop. Growth 1750-1800 1700-1800 1600-1800 1700-1750 1600-1750
(1) (2) (3) (4) (5)
log(Monastic Income Exposure) -0.03 -0.00 0.00 0.02 0.07
(0.043) (0.067) (0.104) (0.054) (0.094)
Controls Yes Yes Yes Yes Yes
Region fixed effects Yes Yes Yes Yes Yes
Observations 229 229 228 229 228
Adjusted R20.12 0.15 0.10 0.06 0.04
Notes: This table presents OLS estimates of the effect of monastic income exposure in 1768 on urban population growth at the arrondissement
level. The dependent variables are percent changes in urban population at the date specified in the column header. Each column controls
for the caloric suitability of the land, ruggedness, initial urban population levels and region fixed effects. Standard errors clustered at the
département level are in parentheses. * p<0.1, ** p<0.05, *** p<0.01.
28
5 Mechanisms
In the previous section, I showed that the reallocation of monastic land improved agricultural
productivity. In this section, I investigate plausible mechanisms linking land reallocation to
productivity gains in agriculture. First, I examine how the reallocation of monastic land affected
land inequality and land consolidation. In a second exercise, I study how this reallocation
affected physical capital investments and labor force organization.
5.1 Land Inequality and Land Consolidation
As explained in Section 2.2, the confiscation and auctioning of monastic lands triggered by the
Vente des Biens Nationaux opened the possibility for rich peasants and bourgeois to increase the
size of their landholdings, driving land inequality. There is no consistent data measuring land
inequality and land fragmentation in the Enquête Agricole of 1852. The first consistent data on
land inequality is available from the Enquête Agricole of 1862, and at the département rather
than arrondissement level. From this, I compute the average farm size to gauge land inequality
in the mid-19th century. As an additional variable to measure land inequality, I calculate the
share of large landowners in each arrondissement using data from the Enquête Agricole of 1852.
I measure land fragmentation using data from Legoyt (1843) to compute the average number
of parcels per owner at the département level. Even at the département level, the data can be
useful for detecting fragmentation and persistent differences in landholdings patterns.
Table 4reports regressions of monastic income exposure, proxying land reallocation, on
the percentage of large landowners (columns 1-3), average farm size (columns 4-6) and the
number of parcels per owner (columns 7-9). The results show that, on average, arrondissements
with higher land reallocation had a larger proportion of large landowners and larger farms.
In particular, column 3 shows that a doubling in monastic income exposure leads to a 6pp
increase in the share of large landowners (one-third of a standard deviation), conditional to
29
my main controls and region fixed effects. Looking at the effect of land reallocation on the
average farm size in column 6, I find statistically insignificant results. One possibility is that
when region fixed effects are included, there is insufficient within-region variation remaining
to estimate the effect. This is likely as farm size data are available at the département level,
one NUTS level higher than arrondissements.29 Reassuringly, the effect of land reallocation
on farm size is positive and highly significant when I remove region fixed effects in column 5.
The estimated effect is economically important; column 5 predicts that a doubling in monastic
income exposure is associated with a 2.65-hectare increase in the average farm size in 1862
(46% of a standard deviation), conditional on caloric suitability of the land, ruggedness and
urban population levels in 1750.
The results also indicate that arrondissements with higher land reallocation had less frag-
mented agricultural exploitations, as measured by the number of parcels per owner in 1843
(columns 7-9). Specifically, column 9 indicates that land reallocation has an economically large
and statistically significant negative effect on land fragmentation. Indeed, a doubling in monastic
income exposure lowers the number of parcels per owner by 1.45 (42% of a standard deviation).
Overall, I identify a consistent pattern indicating that the reallocation of monastic land
following the French Revolution triggered an increase in land inequality and a decrease in land
fragmentation.
5.2 Capital Investment
As established in the previous section, the Vente des Biens Nationaux initiated an increase in
land inequality, allowing rich peasants and the bourgeois to purchase Church land and create
large and less fragmented farms. Galor and Zeira (1993) and Galor and Moav (2004) argue that
inequality is conducive to economic development when the prime engine of growth is physical
capital accumulation. The logic is that, at early stages of development, inequality channels
29In my sample, I have farm size data for 85 départements.
30
Table 4: The Effect of Monastic Land Reallocation on Land Inequality and Land Fragmentation
Dependent variable: % Large landowners Farm size Parcels per owner
(1) (2) (3) (4) (5) (6) (7) (8) (9)
log(Monastic Income Exposure) 0.16 0.12 0.06 3.38 2.65 -0.39 -1.78 -1.77 -1.45
(0.015)*** (0.018)*** (0.027)** (0.756)*** (0.901)*** (0.905) (0.534)*** (0.527)*** (0.599)**
[0.017]*** [0.019]*** [0.025]** [0.623]*** [0.759]*** [0.720] [0.490]*** [0.493]*** [0.548]**
Caloric suitability No Yes Yes No Yes Yes No Yes Yes
Ruggedness No Yes Yes No Yes Yes No Yes Yes
Urban population in 1750 No Yes Yes No Yes Yes No Yes Yes
Region fixed effects No No Yes No No Yes No No Yes
Observations 354 354 354 354 354 354 354 354 354
Adjusted R20.40 0.52 0.60 0.22 0.30 0.68 0.13 0.15 0.56
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, on land inequality and land fragmentation in the mid-19th
century at the arrondissement level. I use the share of large landowners in 1852 (columns 1-3), the average size of an agricultural exploitation in 1862 (columns 4-6) and the number of parcels per
owner in 1843 (columns 7-9) as dependent variables. See Section Bof the Appendix for more details on the variables used. For each dependent variable, I first display the bivariate relationship
in the first column, then I include my main set of controls (caloric suitability of the land, ruggedness and urban population levels in 1750) and finally I add region fixed effects. Standard errors
clustered at the département level are in parentheses and Conley (1999) standard errors, with a Bartlett kernel and a cut-off distance of 100km, in brackets. * p<0.1, ** p<0.05, *** p<0.01.
31
resources towards individuals with a higher propensity to save, fostering investment and capital
accumulation. Therefore, a plausible mechanism by which agricultural productivity could have
improved is higher investment in physical capital and, in particular, mechanization.
To measure investments in physical capital in the agricultural sector, I take the number of
scarifiers and extirpators reported in the Enquête Agricole of 1852. These plowing machines
were used to lift, mix, clean and divide the earth before and after the harvest to facilitate the
work of the plow and increase yields. They existed during the 18th century but diffused more
broadly only after the Revolution: “The use of the extirpator in France is not very old, and its
use is far from being as widespread as it should be” (Bixio,1844, author’s translation p. 200).
A possible explanation for the slow diffusion of these machines was their high price.30 Only
landowners that were sufficiently large could acquire such expensive physical capital.
In Table 5, I examine the relationship between land reallocation and investment in physical
capital. As column 1 reveals, there is a positive and highly significant unconditional relationship
between land reallocation and physical capital, as measured by the number of scarifiers and
extripartors in 1852. This effect is robust across the different specifications. In particular,
column 3 reveals a sizeable effect. The point estimates suggests that a doubling in monastic
income exposure leads to a 76% increase in the number of scarifiers and extirpators.
The opportunity to increase productivity through mechanization was dependent on crop
types. I investigate the importance of investment in physical capital, and therefore mechaniza-
tion, to increasing agricultural productivity through a placebo test using vineyard yields. As
wine production requires relatively less intensive use of physical capital than producing cereals,
I expect that productivity gains were less marked for vineyards than wheat fields.
Table 6compares the effect of land reallocation on vineyards yields (columns 1-3) and wheat
yields (columns 4-6). Columns 1 and 2 reveal a positive effect of land reallocation on vineyard
30About one hundred francs (Bixio,1844). In comparison, the average French agricultural laborer was earning
about 175 frances a year (based on the average daily wage and average working days of male agricultural laborers
reported in the Enquête Agricole of 1852.
32
Table 5: The Effect of Monastic Land Reallocation on Mechanization
Dependent variable: log(Scarifiers and extirpators)
(1) (2) (3)
log(Monastic Income Exposure) 0.98 0.77 0.76
(0.184)*** (0.206)*** (0.364)**
[0.201]*** [0.228]*** [0.295]**
Caloric suitability No Yes Yes
Ruggedness No Yes Yes
Urban population in 1750 No Yes Yes
Region fixed effects No No Yes
Observations 354 354 354
Adjusted R20.09 0.12 0.21
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, on
mechanization in 1852 at the arrondissement level. I use the number of scarifiers and extirpators as dependent variable. See Section Bof the
Appendix for more details on the variables used. I first display the bivariate relationship in the first column, then I include my main set of
controls (caloric suitability of the land, ruggedness and urban population levels in 1750) and finally I add region fixed effects. Standard errors
clustered at the département level are in parentheses and Conley (1999) standard errors, with a Bartlett kernel and a cut-off distance of 100km,
in brackets. * p<0.1, ** p<0.05, *** p<0.01.
yields. However, this becomes insignificant when I add region fixed effects in column 3. By
contrast, the effect of land reallocation on wheat yields is positive and highly significant across
all specifications (columns 4-6). This suggests that part of the effect of land reallocation on
agricultural productivity is indeed the results of investment in physical capital.
5.3 Family Labor
As observed by Allen (1988), the rise of English labor productivity in the 18th century was
partly the result of the substitution of family labor with the hiring of specialized labor. In
particular, Allen (1988) notes that the per acre employment of women and children declined
faster along farm size than that of men. To capture the gradual replacement of family labor
by more specialized hired male workers hired on larger farms, I calculate the share of labor by
women and children required to farm one hectare of wheat from the Enquête Agricole of 1852.
Table 7shows that the effect of land reallocation on family labor is negative. This relation-
ship is robust across all specifications. The results suggest that part of the positive effect of land
33
Table 6: The Effect of Monastic Land Reallocation on Productivity: Wine vs. Wheat
Dependent variable: log(Vineyards yields) log(Wheat yields)
(1) (2) (3) (4) (5) (6)
log(Monastic Income Exposure) 0.36 0.36 0.15 0.27 0.27 0.14
(0.048)*** (0.058)*** (0.101) (0.035)*** (0.033)*** (0.051)***
[0.058]*** [0.061]*** [0.106] [0.043]*** [0.038]*** [0.047]***
Caloric suitability No Yes Yes No Yes Yes
Ruggedness No Yes Yes No Yes Yes
Urban population in 1750 No Yes Yes No Yes Yes
Region fixed effects No No Yes No No Yes
Observations 279 279 279 279 279 279
Adjusted R20.21 0.22 0.30 0.37 0.41 0.53
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, on wine
and wheat yields in 1852 at the arrondissement level. I use two different dependent variables: wine yields (columns 1-3) and wheat yields
(columns 4-6). See Section Bof the Appendix for more details on the variables used. For each dependent variable, I first display the bivariate
relationship in the first column, then I include my main set of controls (caloric suitability of the land, ruggedness and urban population levels
in 1750) and finally I add region fixed effects. Standard errors clustered at the département level are in parentheses and Conley (1999) standard
errors, with a Bartlett kernel and a cut-off distance of 100km, in brackets. * p<0.1, ** p<0.05, *** p<0.01.
reallocation on agricultural productivity was channelled through a decrease in family labor, as
observed by Allen (1988) for 18th-century England.
Table 7: The Effect of Monastic Land Reallocation on Family Labor
Dependent variable: % Female and child labor per hectare of wheat
(1) (2) (3)
log(Monastic Income Exposure) -0.05 -0.05 -0.04
(0.015)*** (0.016)*** (0.018)**
[0.014]*** [0.015]*** [0.017]**
Caloric suitability No Yes Yes
Ruggedness No Yes Yes
Urban population in 1750 No Yes Yes
Region fixed effects No No Yes
Observations 354 354 354
Adjusted R20.10 0.12 0.48
Notes: This table presents OLS estimates of the effect of monastic land reallocation, proxied by monastic income exposure in 1768, on family
labor in agriculture in 1852 at the arrondissement level. I use the share of female and child work required to farm one hectare of wheat as
dependent variable. See Section Bof the Appendix for more details on the variables used. I first display the bivariate relationship in the first
column, then I include my main set of controls (caloric suitability of the land, ruggedness and urban population levels in 1750) and finally I
add region fixed effects. Standard errors clustered at the département level are in parentheses and Conley (1999) standard errors, with a Bartlett
kernel and a cut-off distance of 100km, in brackets. * p<0.1, ** p<0.05, *** p<0.01.
34
6 Conclusion
This paper uses a historical setting to explore how a market-based land reallocation can affect
agricultural productivity, and through which mechanisms. Focusing on French agriculture in
the first-half of the 19th century, I analyze the consequences of the major land reform triggered
by the French Revolution, known as the Vente des Biens Nationaux. Through the Vente des Biens
Nationaux, and in the span of five years, 6% of French land was confiscated and auctioned to
secular owners. According to Lecarpentier (1908, author’s translation p. 4), this was “the
most important event of the Revolution”. Specifically, I focus on the reallocation of monastic
land, which represented a substantial part of Church land (Bodinier and Teyssier,2000). The
Vente des Biens Nationaux favored the emergence of capitalist farmers with large farms and
mechanized production techniques. In contrast, areas with less monastic land were unable to
establish large agricultural domains suited for mechanization.
Using data collected from primary and secondary sources, I proxy variations in monastic
landholdings across French arrondissements before the Revolution using the income and location
of each monastery. I show that areas with higher levels of monastic land reallocation, proxied
by monastic income exposure in 1768, had higher levels of agricultural productivity in the
first-half of the 19th century. I shed light on the mechanism, focusing on the changes in farm
size and land fragmentation introduced by the Vente des Biens Nationaux. I find that areas with
higher levels of monastic land reallocation had larger and less fragmented farms in the mid-
19th century. Consistent with Galor and Zeira (1993) and Galor and Moav (2004), I show
that the land reallocation produced both an increase in inequality of land ownership and an
increase in physical capital and agricultural productivity. Finally, consistent with Allen (1988),
I provide evidence that land reallocation induced a substitution of family labor with the hiring
of specialized male workers in agriculture.
The dissolution of French monasteries was part of a larger historical phenomenon, where
secular powers throughout Europe were gradually attempting to control, or temper, the eco-
35
nomic importance of the Church and monasteries. These attempts, increasingly frequent since
the Protestant Reformation, were both cause and consequence of the gradual modernization of
European societies that continued with the Industrial Revolution. I view the investigation of
the social and economic consequences of other historical episodes of dissolution and reallocation
of monastic properties as a fruitful area for future research.
36
References
Adamopoulos, T. and Restuccia, D. (2014). The Size Distribution of Farms and International
Productivity Differences. American Economic Review, 104(6):1667–1697.
Adamopoulos, T. and Restuccia, D. (2020). Land Reform and Productivity: A Quantitative
Analysis with Micro Data. American Economic Journal: Macroeconomics, 12(3):1–39.
Allen, R. C. (1988). The Growth of Labor Productivity in Early Modern English Agriculture.
Explorations in Economic History, 25(2):117–146.
Allen, R. C. (2000). Economic Structure and Agricultural Productivity in Europe, 1300–1800.
European Review of Economic History, 4(1):1–25.
Anselin, L. (2001). Spatial Econometrics. A Companion to Theoretical Econometrics. Hobo-
ken NJ: Blackwell Publishing Ltd.
Bairoch, P. (1988). Dix-huit Décennies de Développement Agricole Français dans une Per-
spective Internationale. Économie Rurale, 184(1):13–23.
Bignon, V. and García-Peñalosa, C. (2021). The Toll of Tariffs: Protectionism, Education and
Fertility in Late 19th century France. CEPR Discussion Paper No. DP16069.
Bixio, A. (1844). Maison Rustique du XIXe siècle, volume 1. Paris, Librairie Agricole.
Bodinier, B. (1988). Les Biens Nationaux dans le Département de l’Eure de 1789 à 1827. PhD
thesis, Paris 1.
Bodinier, B. and Teyssier, É. (2000). L’Événement le plus Important de la Révolution: La Vente
des Biens Nationaux (1789-1867) en France et dans les Territoires Annexés, volume 53. Société
des Études Robespierristes.
Brondel, N. (2008). L’Almanach royal, National, Impérial: Quelle Vérité, Quelle Trans-
parence ? (1699-1840). Bibliothèque de l’École des Chartes, pages 15–87.
Buringh, E. (2021). The Population of European Cities from 700 to 2000: Social and Economic
History. Research Data Journal for the Humanities and Social Sciences, 1(aop):1–18.
Cantoni, D., Dittmar, J., and Yuchtman, N. (2018). Religious Competition and Realloca-
tion: The Political Economy of Secularization in the Protestant Reformation. The Quarterly
Journal of Economics, 133(4):2037–2096.
Carré de Busserolle, J.-X. (1882). Dictionnaire Géographique, Historique et Biographique d’Indre-
37
et-Loire et de l’Ancienne Province de Touraine, volume IV. Tours, impr. de Rouillé-Ladevèze.
Cassini and EHESS (2021). Des villages de Cassini aux communes d’aujourd’hui.
http://cassini.ehess.fr/fr/html/index.htm.
Cinnirella, F. and Hornung, E. (2016). Landownership Concentration and the Expansion of
Education. Journal of Development Economics, 121:135–152.
Conley, T. G. (1999). GMM Estimation with Cross Sectional Dependence. Journal of Econo-
metrics, 92(1):1–45.
Cornia, G. A. (1985). Farm Size, Land Yields and the Agricultural Production Function: An
Analysis for Fifteen Developing Countries. World Development, 13(4):513–534.
Daudin, G. (2010). Domestic Trade and Market Size in Late-Eighteenth-Century France. The
Journal of Economic History, 70(3):716–743.
de la Croix, D. and Perrin, F. (2018). How Far Can Economic Incentives Explain the French
Fertility and Education Transition? European Economic Review, 108:221–245.
Diebolt, C., Menard, A.-R., and Perrin, F. (2017). Behind the Fertility–Education Nexus:
What Triggered the French Development Process? European Review of Economic History,
21(4):357–392.
Finley, T., Franck, R., and Johnson, N. D. (2021). The Effects of Land Redistribution: Evi-
dence from the French Revolution. The Journal of Law and Economics, 64(2):233–267.
Foster, A. D. and Rosenzweig, M. R. (2011). Are Indian Farms Too Small? Mechanization,
Agency Costs, and Farm Efficiency. Unpublished Manuscript, Brown University and Yale
University.
Foster, A. D. and Rosenzweig, M. R. (2022). Are There Too Many Farms in the World?
Labor Market Transaction Costs, Machine Capacities, and Optimal Farm Size. Journal of
Political Economy, 130(3):636–680.
Franck, R. and Galor, O. (2021). Flowers of Evil? Industrialization and Long Run Develop-
ment. Journal of Monetary Economics, 117:108–128.
Franck, R. and Galor, O. (2022). Technolo-Skill Complementarity in Early Phases of In-
dustrialisation. The Economic Journal, 132(642):618–643.
Franck, R. and Michalopoulos, S. (2017). Emigration During the French Revolution: Con-
sequences in the Short and Longue Durée. Technical report, National Bureau of Economic
38
Research.
Furet, F. and Ozouf, J. (1977). Lire et Écrire: l’Alphabétisation des Français de Calvin à Jules
Ferry. Paris, Éd. de Minuit.
Galor, O. and Moav, O. (2004). From Physical to Human Capital Accumulation: Inequality
and the Process of Development. The Review of Economic Studies, 71(4):1001–1026.
Galor, O. and Moav, O. (2006). Das Human-Kapital: A Theory of the Demise of the Class
Structure. The Review of Economic Studies, 73(1):85–117.
Galor, O., Moav, O., and Vollrath, D. (2009). Inequality in Landownership, the Emergence of
Human-Capital Promoting Institutions, and the Great Divergence. The Review of Economic
Studies, 76(1):143–179.
Galor, O. and Özak, Ö. (2016). The Agricultural Origins of Time Preference. American
Economic Review, 106(10):3064–3103.
Galor, O. and Zeira, J. (1993). Income Distribution and Macroeconomics. The Review of
Economic Studies, 60(1):35–52.
Ghatak, M. and Roy, S. (2007). Land Reform and Agricultural Productivity in India: A
Review of the Evidence. Oxford Review of Economic Policy, 23(2):251–269.
Gollin, D., Parente, S., and Rogerson, R. (2002). The Role of Agriculture in Development.
American Economic Review, 92(2):160–164.
Goñi, M. (2022). Landed Elites and Education Provision in England: Evidence from School
Boards, 1871-99. Journal of Economic Growth, pages 1–47.
Goudot, G. (2006). Monachisme Clunisien et Vie Rurale sous l’Ancien Régime. Histoire
Societes Rurales, 25(1):9–35.
Grantham, G. W. (1993). Divisions of Labour: Agricultural Productivity and Occupational
Specialization in Pre-Industrial France. The Economic History Review, 46(3):478–502.
Greer, D. (1951). The Incidence of the Emigration during the French Revolution. Harvard
Historical Monographs, 24.
Heldring, L., Robinson, J. A., and Vollmer, S. (2021). The Long-Run Impact of the Dissolu-
tion of the English Monasteries. The Quarterly Journal of Economics, 136(4):2093–2145.
Hoffman, P. T. (2000). Growth in a Traditional Society: The French Countryside, 1450-1815.
39
Princeton University Press.
Hoffman, P. T., Postel-Vinay, G., and Rosenthal, J.-L. (2019). Dark Matter Credit. In Dark
Matter Credit. Princeton University Press.
Jarvis, A., Reuter, H. I., Nelson, A., Guevara, E., et al. (2008). Hole-Filled SRTM for the
Globe Version 4. available from the CGIAR-CSI SRTM 90m Database (http://srtm. csi. cgiar.
org), 15(25-54):5.
Jessenne, J.-P. (1987). Pouvoir au Village et Révolution: Artois 1760-1848, volume 6. Presses
Univ. Septentrion.
Lecarpentier, G. (1908). La Vente des Biens Ecclésiastiques pendant la Révolution Française. F.
Alcan.
Lecestre, L. (1902). Abbayes, Prieurés et Couvents d’Hommes en France: Liste Générale d’après
les Papiers de la Commission des Réguliers en 1768. Picard.
Lefebvre, G. (1972). Les Paysans du Nord pendant la Révolution Française. Paris, Colin.
Legoyt, A. (1843). La France Statistique. Paris, L. Curmer.
Lewis, W. A. (1955). The Theory of Economic Growth. London, Allen & Unwin.
Marin, B. and Marraud, M. (2011). L’Enquête Agricole de 1852. L’Atelier du Centre de
Recherches Historiques. Revue Électronique du CRH.
Mendola, M. and Simtowe, F. (2015). The Welfare Impact of Land Redistribution: Evidence
from a Quasi-Experimental Initiative in Malawi. World Development, 72:53–69.
Moriceau, J.-M. (2002). Terres Mouvantes: Les Campagnes Françaises du Féodalisme à la Mondi-
alisation, 1150-1850: Essai Historique. Fayard.
Newell, W. H. (1973). The Agricultural Revolution in Nineteenth-Century France. The
Journal of Economic History, 33(4):697–731.
Polanyi, K. (2001). The Great Transformation. Boston, Beacon Press.
Restuccia, D. and Rogerson, R. (2013). Misallocation and Productivity. Review of Economic
Dynamics, 16(1):1–10.
Restuccia, D. and Rogerson, R. (2017). The Causes and Costs of Misallocation. Journal of
Economic Perspectives, 31(3):151–74.
40
Rostow, W. W. (1990). The Stages of Economic Growth: A Non-Communist Manifesto. Cam-
bridge University Press.
Sée, H. (1927). La Vie Économique de la France sous la Monarchie Censitaire (1815-1848). Li-
brairie F. Alcan.
Squicciarini, M. P. (2020). Devotion and Development: Religiosity, Education, and Economic
Progress in Nineteenth-Century France. American Economic Review, 110(11):3454–3491.
Squicciarini, M. P. and Voigtländer, N. (2015). Human Capital and Industrialization: Evidence
from the Age of Enlightenment. The Quarterly Journal of Economics, 130(4):1825–1883.
Sée, H. (1925). La France Économique et Sociale au XVIIIeme siècle. Armand Colin.
Tackett, T. (1986). Religion, Revolution, and Regional Culture in Eighteenth-Century France:
The Ecclesiastical Oath of 1791. Princeton University Press.
Tocqueville, A. d. (1967). L’Ancien Régime et la Révolution. Paris, Gallimard.
Toutain, J.-C. (1961). Le Produit de l’Agriculture Française de 1700 à 1958. 2: La Croissance.
Paris, ISEA.
Vigneron, S. (2008). Les Mécanismes du Marché Foncier dans les Campagnes du Nord de la
France au XVIIIe siècle. L’Exemple du Cambrésis et de la Région Lilloise. Revue du Nord,
(2):391–428.
Wilkin, A. (2011). Communautés Religieuses Bénédictines et Environnement Économique,
IXe-XIIe siècles: Réflexions sur les Tendances Historiographiques de l’Analyse du Temporel
Monastique. Ecclesia in Medio Nationis–actes de la rencontre de Conventus, pages 101–150.
Young, A. (1792). Travels in France during the Years 1787, 1788 & 1789. George Bell and
Sons.
41
Appendix
A Additional Figures
42
Figure A-2: Correlation between the percentage of Church Land Redistributed during the
Vente des Biens Nationaux and Monastic Income Exposure – with controls
-10 0 10 20
Percentage of Church Land Redistributed in 1789
-1 -.5 0 .5 1
Log(Monastic Income Exposure)
coef = 4.8466882, (robust) se = .91761884, t = 5.28
Notes: This figure plots the relationship between the percentage of Church land reallocated through the Vente des
Biens Nationaux and log monastic income exposure in 1768. Residuals and coefficient estimates from Table D-4,
column 6.
44
B Variable Definitions and Sources
Variable Definition and Source
Dependent Variables
Wheat Yields Average yield of wheat per hectare in an arrondissement in 1852, as
reported by Marin and Marraud (2011) from the Enquête Agricole of
1852.
Days per hectare of Wheat Total number of days required to farm one hectare of wheat in an
arrondissement in 1852, calculated using data reported by Marin and
Marraud (2011) from the Enquête Agricole of 1852. In particular, it
includes the time needed to perform all operations, including plough-
ing, sowing and harvesting and all types of labor force, including days
from men, women, children and animals.
Agricultural Wage Average daily wage of agricultural laborers in francs in an arrondisse-
ment in 1852, as reported by Marin and Marraud (2011) from the
Enquête Agricole of 1852.
Share of Large Landowners Number of landowners “owning property in the arrondissement with-
out residing there” and landowners “residing in the arrondissement but
not cultivating themselves” over the total number of landowners in an
arrondissement in 1852, calculated using data as reported by Marin and
Marraud (2011) from the Enquête Agricole of 1852.
Farm Size Average size of a farm in hectares in an département in 1862, calculated
using data reported by the Enquête Agricole of 1862.
Parcels per Owner Average number of parcels per owner in a départment in 1843, calcu-
lated using data reported by Legoyt (1843).
Scarifiers and Extirpators Total number of scarifiers and extirpators in an arrondissement in 1852,
as reported by Marin and Marraud (2011) from the Enquête Agricole
of 1852.
Vineyards Yields Average product per hectare in hectolitres in an arrondissement in
1852, as reported by Marin and Marraud (2011) from the Enquête
Agricole of 1852.
Share of Female and Child Labor Number of women’s and children’s days of labor required to farm one
hectare of wheat over the total number of days required to farm one
hectare of wheat in an arrondissement in 1852, calculated using data
reported by Marin and Marraud (2011) from the Enquête Agricole of
1852.
(continued on next page)
45
Variable Definition and Source
Explanatory Variables
Monastic Income Exposure Distance weighted sum of monastic incomes in livres tournois in an
arrondissement in 1768, calculated using (1) and a distance cutoff of
100km. Data on monastic incomes and location comes from the France
Ecclésiastique, the Almanach Royal and Lecestre (1902). For more de-
tails on the sources, see Section Cof the Appendix.
Caloric Suitability Average caloric yields given the set of crops that are suitable for culti-
vation before 1500 in an arrondissement, calculated using data reported
by Galor and Özak (2016) at a 5-degree resolution level.
Ruggedness Average ruggedness index in an arrondissement, calculated using eleva-
tion data reported by Jarvis et al. (2008) at a 5-degree resolution level.
Urban Population levels Total urban population in an arrondissement in a given year, calculated
using data reported by Buringh (2021) at the city level.
Market Potential in 1794 Distance-weighted sum of 1794 population levels in an arrondissement,
calculated using data reported by Cassini and EHESS (2021) at the city
level. In particiular, I consider as a city all municipalities with 1,000
or more inhabitants in 1794 and use the following formula: M Pa=
[j1/dac ·P opc], where P opcis the population of city cin 1794 and
dac is the kilometric distance between the centroid of arrondissement
aand city c.
Market Integration in 1790s Total external suppliers of an arrondissement in the 1790s, calculated
using data reported by Daudin (2010).
Share of Emigrés Number of Ancien Régime supporters who fled France during the
French Revolution (émigrés) over total population in a département
in the 1790s, as reported by Greer (1951).
Literacy in 1786 Share of grooms who signed their wedding licenses with their names
in a département over the 1786-1790 period (as opposed to those who
marked it with a cross), as reported by Furet and Ozouf (1977).
Subscriber Density Average density of Encyclopédie subscibers in an arrondissement in the
1750s, calculated using data reported by Squicciarini and Voigtländer
(2015) at the city level.
Share of Refractory Priests Average share of priests who refused to swear the oath of allegiance in
1791 in an arrondissement, calculated using data reported by Squiccia-
rini (2020) at the district level from Tackett (1986).
(continued on next page)
46
Variable Definition and Source
Banks in 1850 Total number of banks in activity in an arrondissement between 1800
and 1850, calculated using data reported by Hoffman et al. (2019).
Distance to Paris The distance in kilometers from Paris to the centroid of an arrondisse-
ment. Author’s calculations.
Distance to Bishoprics in 1789 The distance in kilometers from bishoprics and archbishoprics in 1789
to the centroid of an arrondissement, calculated using data reported on
Wikipédia.
Dependent Variables used in Appendix
Share of Church Land in 1789 Hectares of land owned by a Church-related entity in 1789 (monas-
teries, bishoprics, etc.) over the total hectares, calculated using data
reported by Finley et al. (2021) at the district level from Bodinier and
Teyssier (2000).
Explanatory Variables used in Appendix
Distance to London The distance in kilometers from London to the centroid of an ar-
rondissement. Author’s calculations.
Distance to Fresnes-sur-Escaut The distance in kilometers from Fresnes-sur-Escaut to the centroid of
an arrondissement. A