Institutional Transplant and Cultural Proximity: Evidence

from Nineteenth-Century Prussia

Giampaolo Lecce and Laura Ogliari∗

This version: July 2015

Abstract

The effectiveness of institutional transplant is likely to depend on the underlying cultural envi-ronment of the receiving country. This paper provides the first evidence that the positive effect ofimporting good institutions is attenuated when the receiving territories are characterized by cul-tural traits in conflict with those embedded in the imported institutions. We test this assumptionusing county-level data from late nineteenth-century Prussia. This allows us to exploit both thequasi-natural experiment generated by the radical Napoleonic institutional reforms and the deeplyrooted cultural heterogeneity across Prussian counties. First, using religious affiliation as a proxyof cultural commonality, we find that the impact of the exported institutions is halved in Protes-tant areas. Then, using hand-collected data on pre-Napoleonic reigns we show that kingdoms withstronger ties to French culture exhibit a more effective transplant even when controlling for insti-tutional proximity. Our findings are consistent with the hypothesis that cultural incompatibilitybetween the country exporting the institution and the receiving areas is a significant determinantof a successful transplant.

∗Lecce: Department of Economics, Universitá Bocconi (email: giampaolo.lecce@phd.unibocconi.it); Ogliari: Depart-ment of Economics, Universitá Bocconi (email: laura.ogliari@phd.unibocconi.it). We thank Alberto Alesina, MaristellaBotticini and Guido Tabellini for their precious supervision ; Davide Cantoni, Michela Carlana, Italo Colantone, RosarioCrinó, Nicola Gennaioli, Francesco Giavazzi, Eliana La Ferrara, Tommaso Nannicini, Tommaso Orlando, Nicola Persico,Severine Toussaert and Ludger Woessmann for useful comments, suggestions and fruitful discussions as well as seminaraudience at Universitá Bocconi. We gratefully acknowledge financial support by Fondazione Cariplo and by UniversitáBocconi.

1

1 Introduction

Economists have long argued the importance of good institutions for economic growth.

Rule of law, better enforcement of contracts, security of property rights are usually

associated with better economic outcomes, such as higher investment in physical and

human capital, technological progress and industrial production. Some countries have

adopted legal institutions from foreign systems.1 However, the growth effect of trans-

ferring a well functioning institution from its original setting may, in principle, be

ambiguous. In particular the reception of the foreign institution crucially depends on

cultural compatibility and institutional proximity with the exporting country.2 3

In this paper we argue that the transfer of a well functioning and positive institution

can be ineffective (or less powerful) as a result of interacting with local culture and social

norms.4 Specifically, we show that the transplant of the Napoleonic Code, precisely

a legal institution, had a strong heterogenous effect on economic performance across

Prussian regions characterized by different cultural traits. Specifically, this paper asks

how a transplanted civil code may affect economic performance across different cultural

environments. To identify the effect of a change in legal institution, we exploit a

historical natural experiment represented by the Napoleonic invasion of Prussia in the

early 19th century. We combine this setup with heterogeneity in culture across regions.

Our novel finding is that Napoleonic Code transplant in modern Germany was more1It is the case of Japan, where the national Civil Code was first modeled on the draft of the German Code. More

recently Chinese antitrust law, the Antimonopoly Law, was shaped in the mold of Western antitrust laws.2The fortune of an institutional transplant depends also on other numerous conditions, as nicely illustrated by Fisher

(1903):

What are the conditions which assist fusion, what are the subtle causes that arrest it, what historicalmedium acts as a conductor, what medium again is non-conducting, what elements of national genius,taste, and character are capable of exportation, what incapable, and why some characteristics are moreeasily assimilated or imitated than others, these and a host of other subsidiary questions will alwaysarise when conquest, migration, or mere juxtaposition places two different races in a position where it isimpossible for either to remain unaffected by the characteristics and ideals of the other.

3Historically, we have examples of good laws that are heterogeneously adopted across countries (or within societies).Putnam and his colleagues, for example, exploit a natural experiment involving an institutional reform in the early1970s in Italy. Italian central governments established new regional governments that ideally would function identicallythroughout the country, but in practice they didn’t. The scholars attributed this discrepancy, that was mainly observedbetween center-north and the south, to different levels of social capital.

4See Alesina and Giuliano (2015) for a detailed survey on these two aspects.

2

effective in areas with stronger cultural linkages with France. Our first proxy of cultural

linkage is the Protestant share of population. Given that France was predominantly

Catholic at that time, it serves as an inverse proxy for cultural proximity. This measure

has several advantages, it is available for all 451 Prussian counties, it is well measured

and varies substantially across different areas. Clearly, however, Protestant share may

not capture all aspects of cultural commonality that maybe relevant for institutional

transplant, hence we also construct alternative proxies - based on hand collective data

- which captures ties between each reign and the Kingdom of France and the attitude

of each ruler toward either the French enlightened ideals or French habits and customs,

both measured before French revolution.

We perform several robustness checks to ensure that our results are not spuriously

driven by observable and unobservable characteristics of Prussian regions. First, we map

individual counties to pre-napoleonic reigns which allows us to condition the estimation

on a full set of reign fixed effects which control for any difference in pre-existing social

norms, historical facts and economic characteristics. Second, we explore a wide range of

alternative specifications using a large set of controls and different proxies for economic

performance and institutions. Third, we show that our effect is not contaminated by

past implementation of liberal reforms in some Prussian areas. Finally, we explore

a number of competing explanations such as: (i) human capital accumulation, (ii)

religious fractionalization within county and (iii) religious diversity with neighboring

counties. We find none of these stories to fully explain our results.

Our paper contributes to several streams of research broadly focusing on the ef-

fect of institutions and culture on economic performance. The first important stream

we connect to is the one investigating the importance of good institutions on economic

growth. Starting with the seminal work of Douglas North 5, many scholars have empha-

sized that institutions "matter". One of the first paper showing a positive correlation

between measure of institution and economic growth was the one by Knack and Keefer

(1997). They documented a positive relation between GDP and "protection against

expropriation risk". Significant the works by Acemoglu et al. (2001, 2002) in which5See, for example, North (1990).

3

the economic performance is explained by the persistence of institutions, in particu-

lar exploiting the heterogeneity in colonization across the different European colonies.

Similar results have been proposed using different proxies for institutions such as cor-

ruption, political instability and rule of law. The latter has been widely investigated by

a series of papers by Shleifer and his coauthors relating institution and the legal origin

of countries.6 In a more recent paper, extremely relevant for this work, Acemoglu et

al. (2011) exploited the variation in institutional reforms during Napoleonic campaigns

across 18th and 19th century. In particular the paper shows that the radical reforms

imposed by the French Army in the Western part of modern Germany had significant

effects on long-term economic performance.

The other important strand of literature we refer to is the one analyzing the effect of

culture on economic performance. Societies, through history, developed their own cul-

ture and these cultural traits are persistent over time and affect their economic growth

path.7 Closely related to our paper the works by Becker and Woessman (2009) and

Cantoni (2014).Using historical data on Germany they both test the Weberian hypoth-

esis8, which attributed the rising industrialization in Western Europe to Protestant’s

ethic, providing controversial results. Cantoni (2014) analizes the impact of Protes-

tantism on urbanization, used as a proxy for economic prosperity, starting from the

17th century. He finds no significant effect of Protestant ethic on economic development

even using different specifications (IV regression) and different subsamples (Calvinist).

Becker and Woessman (2009) use an argument similar to the one proposed by Bot-

ticini and Eckstein (2005, 2012) stressing the importance of human capital to explain

economic prosperity. They claim that a side effect of Martin Luther’s preachings was

to boost literacy rate. Using cross-county variation in Prussia during the 19th century,

they find evidence of higher level of human capital in Protestant areas, thus providing

an alternative channel to explain the higher prosperity of Protestant regions.

We provide the first empirical evidence of the heterogeneous growth effect generated6For example Glaeser and Shleifer (2002) and La Porta et al. (2008).7See Alesina and Giuliano (2010); Algan and Cahuc (2010), Bisin and Verdier (2000), Doepke and Zilibotti (2008);

Fernandez, Fogli and Olivetti (2004); Galor and Moav (2002); Greif (1993); Guiso, Sapienza and Zingales (2008),Nannicini et al. (2013), Nunn and Wantchekon (2011) and Tabellini (2008, 2010). Fernandez (2011) provides a detailedreview on this literature.

8See the seminal work by Max Weber (1930) "The Protestant Ethic and the Spirit of Capitalism".

4

by a law transplant in a variegated cultural environment, thereby marrying these two

important strands of literature. There is, however, an emerging theoretical literature

on the interplay between cultural traits and institutions. A seminal paper by Acemoglu

and Jackson (2014) models the interaction between law enforcement and social norms9.

An interesting prediction of their model is that laws in conflict with prevailing social

norms may not be fully accepted and thus may backfire. In the context of Napoleonic

invasions, this might imply that the reception of French institutions can be different

depending on underlying cultural environment of the several states receiving the Code.

In contrast to the existing contributions, in particular Acemoglu et al. (2010,2011)

and Buggle (2015), our analysis does not focus on the positive effect of French institu-

tions but on their heterogeneous reception across the different reigns and the impact

on economic performance. We are also aware that culture and institutions are endoge-

nous variables, hence we use the exogenous variation in reforms, Napoleonic campaigns

before the Congress of Vienna and the introduction of Napoleonic Civil Code in order

to quantify the effects of culture (mainly identified with religious beliefs and attitudes

towards French culture), good institutions and their interaction. We analyse the unique

environment provided by the institutional and cultural heterogeneity of the Prussian

Empire. 10

The rest of the article is organized as follows. Section 2 reviews the historical back-

ground, discussing French military campaigns with the consequent introduction of the

Civil Code and its abandonment after the Congress of Vienna. Section 3 presents

our data, provides some descriptive statistics and our empirical specification. Section

4 provides evidence on the significant effect of interaction between culture, measured

as religious beliefs or French ties and French institutions on economic performance.9Also Bisin and Verdier (2015) have a theoretical model on the interaction of culture and institutions.

10For example, Rowe (2003) notes that:

Cultural diversity was acceptable in the eighteenth-century Rhineland. It was unwelcome to revolution-ary state-builders. The French interior minister, visiting in 1798, was shocked by the numerous customsthat confronted him, and which he blamed for obstructing ‘the fusion of the various parts of these diversecountries into a completely republican single whole’. The old territorial divisions were so great as to utterlyconfuse revolutionary commissioners posted to the region. For them, the imposition of cultural and aboveall linguistic uniformity was not merely a question of administrative efficiency. Nor was it an example ofcultural imperialism, for the revolutionaries evinced little interest in attempting to impose French beyondthe Republic’s natural frontiers.

5

Section 5 tests also the robustness of these findings using a wide range of alternative

specifications. Section 6 discusses some alternative potential channels and Section 7 in-

vestigates whether the main results are driven by pre-Napoleonic institutional features.

Finally Section 8 concludes.

2 Historical Overview

2.1 The situation of the Holy Roman Empire before 1800

The territories of the Holy Roman Empire had always been characterized by a huge de-

gree of heterogeneity. Since its foundation in 962 AD the Holy Roman Empire featured

high fragmentation furthermore the several layers of political power became gradually

more autonomous as the Emperors focused their attention on their local kingdoms.

Pivotal in the progressive disintegration of the Empire was the Protestant Reforma-

tion. Starting as a protest against the corruption of the Roman Catholic Church the

Reformation quickly spread out throughout central Germany gaining the support of sev-

eral princes who wanted to stress their political and religious independence. In 1555,

after several years of war, the Emperor and the Protestant German princes signed a

peace treaty in Augsburg. The principle of cuius regio eius religio (“whose realm,

his religion”) was affirmed making Lutheranism an official religion of the Empire. The

ambitions of the Emperor to centralize power and rule over a unified empire was thus

permanently shattered. Religious and political wars continued to afflict the Holy Ro-

man Empire until a stable resolution was reached with the Peace of Westphalia. By

1648 the Empire was just a confederation of German princes who, in their own lands,

had the right to legislate, impose taxes, organize an army, mint and engage in foreign

policy.

The political fragmentation (see Figure 1) of the Empire translated into an institu-

tional and cultural diversity which was much deeper than the one due to the differences

in religious affiliation. During the eighteenth century, some rulers, inspired by enlight-

enment principles, enacted reforms in their states promoting literacy and simplifying

justice and administration (one of the most notable example of enlightened king is

6

Frederick the Great in Prussia). 11

[FIGURE 1 HERE]

2.2 The French Revolution and the Napoleonic Military Campaigns

The emergence of revolutionary France as an aggressive and strong military power at

the end of the eighteenth century marked the end of the Empire. The first crushing

victories by the French armies created a power vacuum in the area known as modern

Germany, and Napoleon did not waste time in transforming it into a "cordon sanitaire"

between France and its traditional Eastern enemies (see Figure 2). In his campaign to

expand the French Empire, Napoleon was driven by geo-political concerns more than

by the economic outlook of the region.12 With the treaty of Luneville in 1801, the

left bank of the Rhine was annexed to the French Empire.13 The effective end of the

Holy Roman Empire was determined when the Habsburg in 1805 ceded part of their

German territories to Napoleon allies. The following year, central Germany was unified

in a confederation of sixteen states (Confederation of the Rhine) whose protector was

Napoleon. The members of the Confederation promised to supply their "protector with

a military contingent of sixty-three thousand men" (Lefebvre 1969) and, in return for

their support, they were given higher statuses or additional territories. For example

the free cities of Augsburg and Nuremberg were annexed by Bavaria and Frankfurt

to Dalberg, Nassau became a duchy and Dalberg became the prince primate of the

Confederation of the Rhine.

[FIGURE 2 HERE]

2.3 The Rise of Institutions

It is not possible to determine the quality of institutions across the Holy Roman Empire

before French invasion, given the strong heterogeneity resulting from the territorial11For more details, see Arvind and Stinton (2010).12As it was also stressed by Acemoglu et al. (2010).13According to Fisher (1903), this treaty has also been called the "First Revolution of Germany" given that a "territory

of 150,000 square miles, peopled by 3,500,000 inhabitants, and amounting to about a seventh part of the populationand territory of the whole Empire" was transferred to the French Empire.

7

fragmentation, still some useful general features can be identified. The institutions

of the "ancien regime" were predominant across the territories and feudal privileges

were still deeply rooted throughout the Holy Roman Empire. Even in those areas

where serfdom had been abolished, peasants were subject to several restrictions and

burdened by a list of duties and services they should provide to their lords. Equality

before the law was still far away, some classes (aristocrats, clergy, military) benefited

particular preemptions, while other groups were discriminated (ex. Jews). At the urban

level there were some other privileged lobbies such as guilds, who controlled the major

sectors and in some cases were limiting the development and the growth of the industry

they controlled.14

Once the last vestiges of the Holy Roman Empire were abolished, the new ideals and

the spirit of the French Revolution pervaded the new territories. Napoleonic era was

characterized by the rise of a new elite in the place of the old aristocracy which could

participate in the political and economic life through representative councils, electoral

colleges and chambers of commerce. The German territories under the influence of

the French Empire, even the satellite states, benefited from a series of institutional

reforms imposed by Napoleon.15 The introduction of the Civil Code maybe was the

most important among the reforms.

Promulgated on 21 March 1804, the Code Napoleon regulated all social interac-

tion from family matters to economic contracts, importantly, it consecrated absolute

property rights to which the code dedicated a total of 1776 articles.16 The process of

codification continued with the promulgation of the Code of Civil Procedure (1806),

the Commercial Code (1807), the Criminal Code and the Code of Criminal Procedure

(1808) and the Penal Code (1810). All these codes were also adopted, sometimes with

some modifications or with different timings, by the satellites states of the Confed-

eration of Rhine.17 Interestingly, some states decided to retain the codes even after14A good example in the Rhenish area is provided by Kisch (1989). The author provides an example of the limitations

the guild imposed on the adoption of new technologies.15Acemoglu et al. (2010) remark that "The reform of the administrative and fiscal system, the enactment of writ-

ten legal codes, the restructuring of agricultural relations, the abolition of guilds, the emancipation of Jews and thesecularization of church lands are often cited among the beneficial reforms introduced by the French Rule in Germany".

16A huge amount when compared to 515, the number of articles regulating person. See Woolf (1991).17See Acemoglu et al. (2010) for details.

8

Napoleon’s fall.18

French rule implied also a more efficient model of administration compared to the

past. In line with the Revolutionary principles, it was hierarchical, centralized, and

based on the separation of powers. This new efficient bureaucratic system was coupled

with a fiscal reform that introduced budgeting and rationalization of public expendi-

tures. The impact of public reformation generated hostility among local population,but

it laid the foundations for a modern system of public finance.

Napoleon disseminated radical progressive reforms throughout Europe. Despite the

death toll in German territories, his invasion implied modern bureaucracy, meritocracy,

the end of privileges of aristocracy, the limitation of Church power, legal equality,

property rights and a better public finance system. He set in motion the transition

from the "ancien regime" to the modern era.

3 Data and Specification

Our main source is the ifo Prussian Economic History Database (iPEHD) built by

Becker et al. (2012). We use Prussian county level data collected from different censuses

over the period 1871-1892. In particular we choose these waves of data because they

guarantee the widest geographical coverage, the highest number of states (including

former members of the Confederation of the Rhine) and also a sufficient period for

the new institutions to be effective.19 We complement iPEHD database with unique

hand-collected data on pre-Napoleonic reigns and rulers.

We now describe our institutional and cultural variables (Section 3.1), dependent

and other relevant variables (Section 3.2), and then provide some descriptive statistics

(Section 3.3).

3.1 Institutional and Cultural Variables

Our main measure of institutional change is a binary variable, which takes value one if18A commission was set up to decide whether to extend Prussian law or keep the French one. The Commission ended

its work in 1818 and recommended the preservation of the French judicial system. French law remained in force inRheinland until 1900. But also local population and business community where at the forefront to retain the code. SeeRowe (2000) and Diefendorf (1980) for more details.

19Acemoglu et al. (2011) find the effect of institutions on economic growth especially after 1850.

9

the county is either in the provinces annexed to the French Empire (e.g. Rhineland) or

in a satellite state (e.g. Kingdom of Westphalia).20 This measure is meant to capture

the presence of Code Napoleon. The counties invaded at the beginning of the century

imported Napoleonic code and better legal institutions. Cultural traits are notably

difficult to measure. We start by using religious affiliation to capture cultural diversity.

Our main proxy is the share of Protestant in a county in 1871 (see Figure 3).

[FIGURE 3 HERE]

This measure has two main advantages. First, religious beliefs are suitable to capture

cultural commonality, in particular "[we] expect that two countries with the same religion

tend to have similar cultures" (Guiso et al. 2009). Second, Protestant share is an

easily measurable and transparent variable. We construct other measures to capture

two different cultural dimensions that may be crucial for reception of the Civil Code,

namely Pro-French and French Ties.

[FIGURE 4 HERE]

We hand-collect historical information on the reigns located on the territory of

modern-day Germany and Poland. We map all the counties in our database into their

relative eighteenth century reigns, this allows us to have a unique database and to

identify all the historical rulers in the relevant territories. Our classification is crucial

for the creation of pre-Napoleonic cultural measures. The first measure of historical

cultural commonality is based on the ties between the reign and the Kingdom of France

before the revolution. We expect that early interaction with French courts and customs

should facilitate the assimilation of French institutions. We investigate whether, during

the eighteenth century, one of the rulers had a French direct relative (mother/father),20In the robustness checks we use an alternative definition that captures the direct effect of French empire with respect

to satellite kingdoms. The First French Empire was defined by the Rhine, except in the Northern territories, so we usea smaller subsample with respect to the previous one, because we do not want to capture the effect of some counties inthe Confederation of Rhine. As we discussed above, the latter was a confederation of client states of the First FrenchEmpire but they were formally independent. A good example is provided by the Kingdom of Westphalia, which can beconsidered a “model”, since at the beginning of the XIX century different legislative reformation were introduced (equalrights, Jews equal rights, right for a jury trial, etc). This kingdom was formally ruled by Napoleon’s brother JeromeBonaparte and it was officially an independent state.

10

then we verify if the reign was previously occupied by the French Army for a period

longer than one year. We thus construct a dummy variable French Ties that equals

one if the ruler had a French relative or the reign had been occupied after the Peace

of Westphalia. The second variable we construct, Pro-French, captures the inclination

of the ruler toward either the French enlightened ideals or French habits and customs

according to historical sources. Even if this measure is highly subjective, it nicely cap-

tures attitudes towards the foreign culture. Finally, we construct a measure of historical

institutional proximity. In particular, we collect data on progressive reforms in the edu-

cational, judicial or administrative system implemented in each reign between 1700 and

1790. We classify as enlightened those rulers that implemented at least one progressive

reform and we create an index based on the fraction of years the progressive king was

in power. For example, Frederick the Great was king for 46 years and he implemented

educational reforms, hence he contributes to Brandenburg index by 0.51 (i.e. 46 over

90 years). Our institutional proximity is equal to one if all rulers, in that time span,

have been enlightened. This index is meant to capture similarity between Napoleonic

code, which is grounded in innovative and liberal ideals, and pre-existing institutions.

3.2 Other Variables

Our main measure of economic performance is the average annual incomes of male ele-

mentary school teachers from the 1886 Education Census. Since we do not have general

county-level income data this proxy is the only direct measure of income available for all

counties at the time.21 We use several historical controls in order to capture pre-existing

differences across counties. In particular we control for the presence of Hanseatic or

free imperial cities in sixteenth century, since they benefited from particular economic

and diplomatic privileges. In order to capture human capital and cultural differences

we also control for the presence of schools and monasteries - relevant educational cen-

ters - in 1517 and the presence of universities during the Holy Roman Empire period.

Finally, following the literature, we control for pre-napoleonic economic development21It has already been used as a proxy for income in general in other studies (see Lee, Galloway, and Hammel 1994,

Becker and Woessmann 2009).

11

using urban population in 1500.22

Additional geographic controls are the distance from Berlin (in km), the latitude (in

rad) and a dummy variable for polish-speaking provinces (mainly located in the East).

Contemporaneaous controls include demographic and social characteristics of the

population and industrial features in the county.

3.3 Descriptive Statistics

The descriptive statistics, reported in Table I, reveal that about half of the sample was

under the control of Napoleon between 1792 and 1814. Among these counties 20% were

formally annexed to the French Empire. Our main economic outcome variables seem

homogeneous on average across the sample while there are significant differences in

cultural proxies. First, areas controlled by Napoleon display a significantly lower share

of Protestants and Rhineland is mainly Catholic. Second, regions that were invaded by

Napoleon had already stronger ties with France. Finally, enlightened monarchs seem

to be more concentrated in the Eastern regions. The latter finding is mainly driven

by Fredrick The Great of Prussia whose liberal attitudes and reforms deeply pervaded

the Electorate of Brandenburg. Historical controls differ across treated and non treated

territories suggesting that western areas were richer even before the arrival of Napoleon.

Nineteenth century controls do not display significant differences in the treated areas.

[TABLE 1 HERE]

3.4 Identification strategy

In this section we present the empirical model testing our central hypothesis that the

heterogeneous reception of the Napoleonic Code depends on the pre-existing cultural

traits of the county. Specifically, in kingdoms culturally more similar to the French

Empire - sharing either enlightenment ideas, the same religion or having historical ties

with France - the institutional transfer was more effective.22See De Long and Shleifer (1993).

12

Our baseline model is as follows:

yi = α+ β1Culturei + β2Napoleoni + β3Culturei ×Napoleoni +Hiβ4 +Giβ5 +Eβ6 +Xiβ7 + εi,

where yi is the average income of male elementary school teachers in county i Culturei

denotes one of the cultural proxies we discussed above (defined at the county or pre-

Napoleonic reign level), Napoleoni is a binary variable for the adoption of the Napoleonic

code. Hi, Gi, Ei and Xi are respectively vectors of historical, geographical, educational

and economic controls while εi is a standard error term.

The key coefficient is the one on the interaction between the French Institution and

the cultural measures. We expect β3 to be significantly different from zero and, in

particular, that cultural similarity and institutions positively interact.

Our identification strategy relies on two main assumptions. First, the areas invaded

by Napoleon were not chosen because they were more prosperous, i.e. French occupa-

tion is exogenous. As extensively argued by Acemoglu et al. (2011), French military

invasions were not driven by economic reasons but mainly by geographic and historical

ones (ex. Rhine was historically considered the natural border of French empire). Sec-

ond, religious affiliation is persistent across centuries and thus the share of Protestants

in 1871 captures cultural traits that already existed before the arrival of Napoleon.

Previous empirical evidence supports this assumption, persistence in religious distribu-

tion is observed when comparing religious affiliation data after the Peace of Augsburg

(1555) with the more recent one. (see Figure 4).23

4 Institutional Transplant and Religious Beliefs

In this section we test the impact of religion on the effectiveness of the French insti-

tution. Table 2 shows a significant positive effect of Protestantism and of the legal

institution on teachers’ wage, in line with previous work. The most striking result

is the negative and significant coefficient for the interaction term across the different23Spenkuch (2010) provides detailed data and figures of Protestant and Catholic distributions in Germany after the

Peace of Augsburg.

13

specifications. This result suggests that Napoleonic institution has a weaker impact on

economic performance the higher the Protestant share in the county. In particular the

mean wage of a teacher is 10-11 % points lower in Protestant counties who received the

Code compared to Catholic counties who received the Code. Our estimates prove to

be robust to the inclusion of several control variables. In column 2, we control for a set

of geographical controls, in particular for the distance from Berlin, the Prussian capi-

tal, and from the district capital in order to take into account periphery. In columns

3-5 we add historical and current controls. The former allows us to control for pre-

existing differences while the latter allows to capture the economic and social outlook

of the county after the Congress of Vienna. We find that some of these variables have

an effect significantly different from zero but the significance of the interaction term

coefficient is barely affected.

[TABLE 2 HERE]

5 Robustness

This section investigates a series of robustness checks to validate our findings (Section

5.1). Then we implement a Diff-in-Diff specification using an alternative database.

5.1 Checks on the main specification

This section presents a battery of alternative specifications in order to investigate the

robustness of the baseline estimates, the results are reported in Table 3.

Clustered Standard Errors Panel a) clusters standard errors at the pre-Napoleon

reign (row 1) and 1871 Prussian political district level (row 2) to allow for an arbitrary

variance-covariance matrix capturing potential serial correlation in the residual error

term. The coefficients are always statistically significant as in the baseline estimates.

14

Fixed Effects In Panel b) we include pre-Napoleonic fixed effects. We first include

reign fixed effects (row 3) and then we include rulers’ fixed effects. In particular, the

latter are a smaller number since several kingdoms had the same ruler. Our results are

robust to the inclusion of these controls.

Additional Controls A potential concern could be that the result is induced by

differences in purchasing power across regions or by other drivers that influence the

demand for teachers, and hence their wages. In panel c), we show that our evidence

is preserved when adding supplementary controls addressing these issues. We first

include a price measure to capture potential differences in purchasing power across the

different counties. This proxy is constructed as the ratio between total expenditures

in new school buildings in 1886 over the total number of new school buildings. This

measure should capture variation in housing prices, a relevant component of CPI.24

Albeit not perfect, this is the only available measure of historical unit values at county

level. We then add a group of demographic variables in 1871 - including household size,

share of population born in the county, share of population of Prussian origin, share

of females and share of the population under 16, that might influence the demand of

teachers. The results always confirm the baseline estimates.

Outliers In panel d), we show that our evidence is not driven by influential obser-

vations. First, we trim (row 7) and winsorize (row 8), respectively, the extreme 1% of

observations of our dependent variable. We then compute a measure of influence (row

9), specifically how much an observation has effected the estimate of a regression coef-

ficient. In particular we computed the difference between the regression coefficient of

the interaction term (Protestant x Napoleonic Code) calculated for the entire dataset

and the regression coefficient calculated with the observation deleted, scaled by the

standard error calculated with the observation deleted.25

24See Moretti (2013)25The cut-off value we use for highly influential observation is : 2/sqrt(n). Even though, our results are robust to

different cut-off values.

15

Alternative Institutional Variables In panel e), we show that the results hold

when using different proxies for the French institution. Specifically we reduce our

treatment group depending on the exposure to Napoleonic code. We start by using a

more conservative definition of institutional variable, namely we define as invaded only

those territories formally annexed to the French Empire (row 10). In this subsample

the effect of the treatment might be reinforced since, not only the code was imposed,

but also the administrative structure and local governors were replaced by the French

ones. We then exclude the territories under the direct control of the French Empire

and we consider only those territories in the Confederation of the Rhine (i.e. Satellite

States) which adopted the code (row 11). Finally, we use years of French domination as

a proxy for intensity-to-treatment (row 12). Reassuringly, all measures yield the same

message as the preferred Napoleonic-invasion dummy.

Alternative Protestant Variables In panel f), we show that the results are robust

to using alternative measures of religious affiliation. First (row 13), we use a dummy

that equals one if the county has an absolute Protestant majority (share of Protestant

>50%). Crucial for our identification is the persistence of the religious affilation, hence

using a binary variable attenuates the potential concern that Protestant share may have

changed over time. We then construct a historical measure of Protestant majority in

seventeenth century using data from Cantoni (2012). The advantage of this measure

is that it is defined two centuries before the arrival of Napoleon however it has two

limitations: the historical religious affiliation is defined at reign level - not county

level- in 1600 and it is not available for the entire sample. Finally, in order to address

endogeneity concerns, we follow the literature instrumenting Protestant share with the

distance from Wittenberg. We identify the exogenous variation in Protestantism using

the concentric diffusion of Protestantism in Prussia around Luther’s city. Note that

the coefficients maintain the same sign as in the baseline specification, and remain

significant at conventional levels.

Alternative Outcomes Variables Panel g) shows that the baseline results hold

16

for alternative proxies of economic prosperity. In row (16) we use the only other direct

measure of income available, that is wage of daily laborer in 1892.26 Following Becker

and Woessmann (2009) we use income tax revenue per capita (row 17) and the size of

the non agricultural sector (row 18). The coefficient of the interaction term remains

consistently negative and statistically significant across the different specifications.

[TABLE 3 HERE]

5.2 Diff-in-Diff Specification

Our specification allows us to investigate several dimensions of institutions and culture

but most of our variables are measured after the Napoleonic invasion. We have many

historical controls and we introduce reign level fixed effects to capture pre–existing

differences. However, our specification may generate concerns about pre-trends, in

particular that we are not fully accounting for these.

We use data from Acemoglu et al. (2011) in order to implement a DID strategy.

They have 19 independent states (or provinces of larger states) in nineteenth-century

Germany and their outcome variable is the urbanization rate

The main drawback of this dataset is that it contains less information for our pur-

poses. In particular, data have a higher level aggregation and we are not able to

investigate potential channels explaining our results.

Our baseline specification is as follows

uit = δt + µi + α1Post+ α2Post×Napoleon+ α3Post×+α4Post×Napoleon×+εit

where Post is a dummy variable that equals one if we consider a time period after

Napoleonic invasion (i.e. second half of nineteenth century), Napoleon is a dummy

that captures French presence in the state and Prot is the share of Protestants around

1800. Column 1 shows that after the Napoleonic invasion, there is a negative inter-26The table displays the results for the male laborers in urban areas. The coefficients are virtually the same when

using wage of rural male daily laborer or wage of female daily laborer. Results are available upon request.

17

action between French institutions and Protestant share, consistent with our previous

results. These findings are robust also in columns 2-3, where respectively we implement

a weighted regression (weighted by total population in 1750) and an unweighted one

but controlling for the level of population in 1750. Columns 4 provides similar results

but using a different measure of French institutions, i.e. the number of years of French

presence, while column 5 shows consistent results using year dummies. The evidence

suggests that this "moderating effect" due to the negative interaction between culture

and the civil code is significantly affecting the urbanization rate since 1875, i.e. 60 years

after the Congress of Vienna.

[TABLE 4 HERE]

6 Possible Alternative Explanations

This section investigates alternative channels proposed by the literature that might

explain our findings. Our results are robust across the different alternative explanations

and these confirm the fact that culture plays a relevant role in law transplants. We

investigate three possible alternative channels: human capital accumulation, religious

fractionalization and diversity from neighborhood.

Human capital development is a significant driver of economic growth but it is also

closely related to the diffusion of Protestantism, as already stressed in the previous

literature.27 A potential concern is that some of the institutions brought by Napoleon

had a negative impact on the human capital accumulation of the regions. Narrative

evidence suggests that Napoleon’s educational policies were not aimed to boost literacy

but targeted higher education in order to breed well-prepared military and administra-

tive elites. Hence, technical schools were promoted (Polytechnic, Conservatory of art

and trades, etc.) and lycees introduced. Napoleon, in fact, paid very little attention to

primary education - and even less to the education of girls - which was mainly managed27For the effect of human capital on economic growth see, for example, Barro (2002) and Gennaioli et al. (2014).

18

at local level and left to religious institutions. These policies could have led to a rele-

vant role of Catholic parishes in the invaded areas weakening Protestant human capital

accumulation. We use literacy rate in 1871 and its interaction with Napoleonic Code

in order to test the validity of this potential channel. In column 1 of Table 4 we show

that including the percentage of literate significantly affects our dependent variable but

it does not affect the significance of the interaction term. Hence, human capital cannot

alternatively explain our result.

Most of the counties invaded by French troops were mainly Catholic. An alternative

explanation of the negative interaction between Protestantism and institutions may be

that those counties with a higher share of Protestants are also highly fractionalized. A

potential concern is that the interaction term reflects the effects of religious diversity.

Several works have previously investigated the cost and the benefits of diversity, whether

racial, ethnic, religious, or linguistic.28 Fragmented societies are often more prone to

poor policy management and pose more politico-economic challenges than homogenous

ones; however, a diverse cultural or ethnic mix also brings variety in abilities and

experiences that may be productive and lead to innovation and creativity. A highly

fractionalized area can be a better recipient of French institutions if the diversity fosters

open-mindedness and ability to adapt to changes. Furthermore if religious heterogeneity

is correlated with political instability, French domination might have both a sharper or a

weaker effect in those counties that are more heterogeneous. We construct a Herfindal

Index using the shares of three biggest religious groups (Protestants, Catholics and

Jews) and we use this measure and its interaction with the Napoleonic dummy to test

this alternative explanation. Column 2 shows that our results are robust even when we

include a measure of religious fragmentation.

Another concern may be that counties that display internal religious homogeneity

might be very diverse from their neighboring counties and thus economically and po-

litically disadvantaged. Since Protestant counties in the western part of the sample are

a minority among a Catholic majority, this issue could drive our main results. We cre-

ate a measure of religious distance with neighboring counties as the difference between28For example Alesina and La Ferrara (2005)

19

Protestant share of the county and the average Protestant share of the neighbouring

counties. Column 3 confirms that the negative interaction term is not determined by

religious distance from the Catholic neighborhood. Finally, column 4 includes all the al-

ternative explanations, we implement a horserace, and findings support our hypothesis

that cultural incompatibility drives the negative coefficient between Protestant share

and institutions.

[TABLE 5 HERE]

7 Discussion: Institutional or Cultural Proximity?

Having shown that our evidence is robust to possible measurement issues, we devote

this section to investigate whether Protestant areas are characterized by other cultural

features and different institutional qualities that can explain the negative interaction in

our findings. Territories characterized by more liberal rulers who enacted progressive

reforms in the eighteenth century, might have been more willing to accept Napoleonic

institutions. Moreover, French institutions might be better received in those areas with

historical higher exposure to French culture, through protracted French occupation or

family connections. Hence, we want to disentangle the effect of institutional proximity

from the one of cultural commonality on the effectiveness of the institutional transplant.

To this purpose, we construct a measure of institutional reforms based on judicial, ad-

ministrative and educational progressive reforms implemented in the eighteenth century

(see Section 3.1). Table 5 presents our results. Columns 1-2 show that the pre-existing

liberal institutions amplify the effect of Napoleonic Code but they do not delete the ef-

fect of cultural similarity. In columns 3-6 we show that other dimensions of culture have

a positive and significant effect on the transplant. We obtain similar results when we

use French Ties or the Pro-French dummy. This suggests that the cultural environment

is important for the effectiveness of French institutions and Catholic territories might

have a more favourable attitude towards French culture. When running our different

cultural and institutional measures against each other (columns 7-8) we find that the

20

effect is explained by both dimensions. This appears to support the hypothesis that

a reign that was historically more exposed to liberal and enlightened ideas was more

prone to accept and implement Napoleonic institutions but also culture commonality

plays a relevant role. Earlier connections and links with the exporter have a positive

impact in the reception and the effectiveness of the transferred institutions. Finally, the

interaction term between Napoleonic Code and Protestant share has a lower magnitude,

supporting the fact that religious beliefs are not the only cultural dimension important

for economic transplant.

[TABLE 6 HERE]

8 Concluding Remarks

This paper is the first attempt to gauge the heterogeneous effect of a law transplant on

economic performance due to cultural diversity across territories.

Nineteenth-century Prussia provides a very suitable context to test this hypothesis as

it allows us to exploit a quasi-natural experiment, the Napoleonic military campaign, in

an enviroment characterized by a rich variety of cultural traits. All our evidence points

to a moderating effect of the impact of French institutions on economic performance due

to the absence of cultural similarity. Indeed, the positive effect of radical institutional

reforms brought by the French empire in territories either sharing religious beliefs or

having experienced previous contact with France is almost double compared to the one

in culturally distant areas. Our results survive even when controlling for a measure of

institutional proximity between pre-existing institutions and the new imported ones.

Although we analyze a very specific historical environment and extrapolation to

other contexts might be hazardous, our findings call for extreme care when deciding to

export seemingly good institutions, for the transplant may fail when it conflicts with

local culture and pre-existing institutions.

21

References

[1] Acemoglu, Daron, Simon Johnson and James A. Robinson. "The colonial origins of

comparative development: An empirical investigation." American Economic Re-

view 91, no. 3 (2001): 1369-1401.

[2] Acemoglu, Daron, Simon Johnson and James A. Robinson. "Reversal of Fortune:

Geography and Institutions in the Making of Modern World Income Distribution."

Quarterly Journal of Economics 107, no. 4 (2002): 1231-1294.

[3] Acemoglu, Daron, Davide Cantoni, Simon Johnson and James A. Robinson. "From

Ancien Regime to Capitalism: The Spread of the French Revolution as a Natural

Experiment." in: J. Diamond and J. A. Robinson (Eds.), Natural Experiments of

History , Cambridge, Mass.: Harvard University Press, 2010.

[4] Acemoglu, Daron, Davide Cantoni, Simon Johnson and James A. Robinson. "The

consequences of Radical Reform: The French Revolution." American Economic

Review 101, no. 7 (2011): 3287-3307.

[5] Acemoglu, Daron and Matthew O. Jackson. “Social Norms and the Enforcement

of Laws.” NBER Working Papers (2014) No. 20369.

[6] Alesina, Alberto and Paola Giuliano. "The power of the family." Journal of Eco-

nomic Growth 15, no. 2 (2010): 93-125.

[7] Alesina, Alberto and Paola Giuliano. "Culture and Institutions."Journal of Eco-

nomic Literature (2015), forthcoming .

[8] Alesina, Alberto and Elian La Ferrara. "Preferences for Redistribution in the Land

of Opportunities." Journal of Public Economics 89, no. 5 (2005): 897-931.

[9] Algan, Yann and Pierre Cahuc. “Inherited Trust and Culture.” American Economic

Review 100, no. 5 (2010): 2060-2092.

[10] Arvind ,Thiruvallore T. and Lindsay Stirton (2010). “Explaining the Reception of

the Code Napoleon in Germany: A Fuzzy-Set Qualitative Comparative Analysis”.

Legal Studies 30, no.1 (2010):1–29.

22

[11] Barro, Robert J. "Human Capital and Growth." American Economic Review 91,

no. 2 (2001): 12-17.

[12] Botticini, Maristella and Zvi Eckstein. "Jewish Occupational Selection: Education,

Restrictions, or Minorities?" Journal of Economic History 65, no.4 (2005): 922-

48.

[13] Botticini, Maristella and Zvi Eckstein. The Chosen Few: How Education shaped

Jewish History 70–1492. Princeton, NJ: Princeton University Press, 2012.

[14] Becker, Sascha O. and Ludger Woessmann. "Was Weber Wrong? A Human Capital

Theory of Protestant Economic History." Quarterly Journal of Economics 124, no.

2 (2009): 531–596.

[15] Becker, Sascha. O., Francesco Cinnirella, Erik Hornung and Ludger Woessmann.

"iPEHD - The ifo Prussian Economic History Database." CESifo Working Paper

No. 3904 (2012).

[16] Bisin, Alberto and Thierry Verdier. “ ‘Beyond the Melting Pot’: Cultural Trans-

mission, Marriage, and the Evolution of Ethnic and Religious Traits.” Quarterly

Journal of Economics 115, no. 3 (2000): 955-988.

[17] Bisin, Alberto and Thierry Verdier. “ ‘On the joint evolution of culture and insti-

tutions.” Working paper (2015)

[18] Buggle, Johannes C. "Law and Social Capital - Evidence from the Code Napoleon

in Germany." Working paper (2015)

[19] Cantoni, Davide. "Adopting a New Religion: The Case of Protestantism in 16th

Century Germany " Economic Journal 122, no. 560 (2012): 502–531.

[20] Cantoni, Davide. "The Economic Effects of the Protestant Reformation: Testing

the Weber Hypothesis in the German Lands."Journal of the European Economic

Association (2014), forthcoming.

23

[21] De Long, Bradford J. and Shleifer, Andrei. "Princes and Merchants: City Growth

Before the Industrial Revolution." Journal of Law and Economics 36, no. 2 (1993):

671-702.

[22] Diefendorf, Jeffrey M. Businessmen and Politics in the Rhineland, 1789-1834.

Princeton, NJ: Princeton University Press, 1980.

[23] Doepke, Matthias and Fabrizio Zilibotti. "Occupational Choice and the Spirit of

Capitalism," Quarterly Journal of Economics 123, no. 2 (2008): 747-793.

[24] Fernandez, Raquel, Alessandra Fogli and Claudia Olivetti. "Mothers and Sons:

Preference Formation and Female Labor Force Dynamics" Quarterly Journal of

Economics 119, no. 4 (2004): 1249-1299.

[25] Fernandez, Raquel. "Does Culture Matter?" in J. Benhabib, M. O. Jackson and A.

Bisin (Eds.), Handbook of Social Economics , Vol. 1A, North Holland, 2011.

[26] Fisher, Herbert A. L. Studies in Napoleonic Statemanship:Germany . Oxford:

Clarendon Press, 1903.

[27] Galor, Oded and Omer Moav. "Natural Selection And The Origin Of Economic

Growth." Quarterly Journal of Economics 117, no. 4 (2002): 1133-1191.

[28] Gennaioli, Nicola, Rafael La Porta, Fernando Lopez-de-Silanes and Andrei Shleifer.

“Growth in Regions” Journal of Economic Growth 19, no. 3 (2014): 259–309.

[29] Glaeser, Edward L. and Andrei Shleifer. "Legal Origins." Quarterly Journal of

Economics 107, no. 4 (2002): 1193-1229.

[30] Greif, Avner. "Contract Enforceability and Economic Institutions in Early Trade:

the Maghribi Trader’s Coalition." American Economic Review 83, no. 3 (1993):

525-43.

[31] Guiso, Luigi, Paola Sapienza, and Luigi Zingales. “Long Term Persistence.” NBER

Working Papers (2008) No. 14278.

24

[32] Guiso, Luigi, Paola Sapienza, and Luigi Zingales. “Cultural Biases in Economic

Exchange?.” Quarterly Journal of Economics 124, no. 3 (2009): 1095-1131.

[33] Kisch, Herbert. From domestic manufacture to industrial revolution: The case of

the Rhineland textile districts . New York: Oxford University Press, 1989.

[34] Knack, Stephen and Philip Keefer. “Does Social Capital Have an Economic Payoff?

A Cross-Country Investigation.” Quarterly Journal of Economics 112, no.4 (1997):

1252–1288.

[35] La Porta, Rafael, Fernando Lopez-de-Silanes, and Andrei Shleifer. "The Economic

Consequences of Legal Origins." Journal of Economic Literature 46, no 2 (2008):

285-332.

[36] Lee, Ronald D., Patrick R. Galloway and Eugene A. Hammel. "Fertility Decline

inPrussia: Estimating Influences on Supply, Demand, and Degree of Control."

Demography 31 (1994): 347–373.

[37] Lefebvre, Georges. Napoleon. New York: Columbia University Press, 1969.

[38] Moretti, Enrico. "Real Wage Inequality." American Economic Journal: Applied

Economics 5, no. 1 (2013): 65-103.

[39] Nannicini, Tommaso, Andrea Stella, Guido Tabellini and Ugo Troiano. "Social

Capital and Political Accountability." American Economic Journal: Economic Pol-

icy 5, no. 2 (2013): 222-250.

[40] North, Douglass C. Institutions, Institutional Change, and Economic Performance.

Cambridge: Cambridge University Press, 1990.

[41] Nunn, Nathan and Leonard Wantchekon. "The Slave Trade and the Origins of

Mistrust in Africa." American Economic Review 101, no.7 (2011): 3221-52.

[42] Putnam, Robert D., Robert Leonardi and Raffaella Y. Nanetti. Making Democracy

Work. Princeton, NJ: Princeton University Press, 1993.

25

[43] Rowe, Michael. From Reich to State: The Rhineland in the Revolutionary Age,

1780-1830 . Cambridge: Cambridge University Press, 2003.

[44] Rowe, Michael. "The Napoleonic Legacy in the Rhineland and the Politics of Re-

form in Restoration Prussia." in D. Laven and L. Riall (Eds), Napoleon’s Legacy:

Problems of Government in Restoration Europe. Oxford and New York: Berg Pub-

lishers, 2000.

[45] Spenkuch, Jorg L. "The Protestant Ethic and work: Micro evidence from contem-

porary Germany." MPRA Paper (2010).

[46] Tabellini, Guido. "The Scope of Cooperation: Values and Incentives." Quarterly

Journal of Economics 123, no. 3 (2008): 905-950.

[47] Tabellini, Guido. "Culture and Institutions: Economic Development in the Regions

of Europe." Journal of the European Economic Association 8, no. 4 (2010): 677-

716.

[48] Weber, Maximilian K. E. The Protestant Ethic and the Spirit of Capitalism. Lon-

don: Allen and Unwin, 1930.

[49] Woolf, Stuart. Napoleon’s Integration of Europe. New York: Routledge, 1991.

26

Figure 1: Holy Roman Empire in 1789

Figure 2: Counties Under Napoleonic Influence

Figure 3: Share of Protestant 1871

Figure 4: Confessional Division 1560

Table 1: Summary Statistics

Not Invaded Invaded p-value

Mean (a) Obs. Mean (b) Obs (a)-(b)

French Empire 0 200 .22 251Satellite States 0 200 .78 251Av. income of male elementary school teacher 6.82 200 6.91 251 0.00Share of Protestant .77 200 .54 251 0.00Liberal Att. .9 199 .59 228 0.00French Ties .035 200 .29 251 0.00French Bias .89 200 .58 251 0.00County Area (log) 11.07 200 10.57 251 0.00Number of Farms 1882 (log) 8.66 200 8.9 251 0.00Year in which annexed by Prussia 1696 200 1796 251 0.00Total Population (log) 10.88 200 10.74 251 0.00Universities Holy Roman Empire .035 200 .0797 251 0.05Imperial city in 1517 .01 200 .0876 251 0.00Hanseatic city in 1517 .055 200 .0916 251 0.14City population in 1500 .0001 200 .0008 251 0.09

Table 2: Dep. Var. Log average wage, male elementary teacher 1886

(1) (2) (3) (4) (5)

Napoleon 0.0917*** 0.0822*** 0.102*** 0.0844*** 0.104**(0.0235) (0.0230) (0.0188) (0.0205) (0.0438)

Share of Protestant 0.115*** 0.134*** 0.178*** 0.174*** 0.143***(0.0291) (0.0294) (0.0239) (0.0266) (0.0275)

Napoleon ⇥ Share of Protestant -0.0971*** -0.112*** -0.180*** -0.155*** -0.0812*(0.0342) (0.0335) (0.0256) (0.0283) (0.0441)

Constant 8.504*** 8.674*** 6.765*** 6.146*** 6.033***(0.312) (0.355) (0.338) (0.396) (0.632)

Geographic Controls yes yes yes yes yesHistorical Controls no yes yes yes yesEconomic Controls no no yes yes yesEducation controls no no no yes yesReign FE (38 groups) no no no no yes

R2 0.400 0.441 0.648 0.659 0.800Obs. 447 447 447 447 447

Notes: Geographic Controls: latitude, distance from Berlin (km), area of the county (log), dis-tance from the district capital and polish speaking area. Historical controls: year of annexa-tion to Prussia, population in 1500, schools in 1517, universities in the Holy Roman Empire,Hanseatic or Imperial cities and monasteries per sqkm in 1517. Economic controls: total popu-lation size (log), percentage of county population in urban areas 1871, percentage of labor forcein mining 1882 and number of farms 1882 (log). Education controls: percentage of pupils withdistance to school over 3 km, total number of pupils 1886 (log), total number of teachers 1886(log) and number of free apartments for male teachers 1886. Robust standard errors in paren-thesis. *** p<0.01, ** p<0.05, * p<0.1

Table

3:Robustness

Check

s

Institution

sProtestan

tInstitution

s⇥

Protestan

t

Coe↵

.Std.Err.

Coe↵

.Std.Err.

Coe↵

.Std.Err.

Obs.

R2

a)C

lustere

dStd

.Err.

1)AtPre-N

apoleonic-R

eign

0.08

44**

(0.036

8)0.17

4***

(0.054

6)-0.155

**(0.060

5)44

70.65

92)

At18

71District

0.08

44**

(0.038

4)0.17

4***

(0.052

8)-0.155

**(0.064

9)44

70.65

9

b)F

ixed

E↵ects

3)Pre-N

apoleonic

Reign

FE

0.10

4**

(0.043

8)0.14

3***

(0.027

5)-0.081

2*(0.044

1)44

70.80

04)

Pre-N

apoleonic

RulerFE

0.14

1***

(0.029

9)0.13

1***

(0.026

4)-0.092

6***

(0.033

6)44

70.78

9

c)A

dditionalControls

5)Price

(Realestate

unit

values)

0.08

30**

(0.039

0)0.17

6***

(0.053

4)-0.151

**(0.067

8)44

10.65

96)

Price

andDem

oCon

trols

0.05

94*

(0.033

5)0.14

9***

(0.042

3)-0.106

*(0.056

8)44

10.69

6

d)O

utliers

7)Trimming

0.08

18**

(0.037

7)0.15

3***

(0.051

3)-0.159

**(0.062

1)43

90.65

48)

Winsorising

0.08

29**

(0.036

6)0.17

0***

(0.054

3)-0.154

**(0.061

5)44

70.66

29)

DfBeta

0.09

53**

*(0.026

7)0.17

1***

(0.050

5)-0.186

***

(0.055

9)42

90.72

9

e)A

ltern

ativeIn

stit.Var.

10)French

Empire

0.15

3***

(0.036

7)0.20

0***

(0.027

7)-0.241

***

(0.071

7)25

30.78

611

)Satellite

States

0.05

03**

(0.023

3)0.17

7***

(0.028

7)-0.114

***

(0.032

3)39

30.65

12)Yearof

French

Dom

ination

0.00

889*

**(0.001

35)

0.16

3***

(0.030

5)-0.012

***

(0.003

3744

70.66

6

f)Altern

ativePro

testantVar.

13)Prot.

Maj.

1871

0.06

11**

*(0.017

3)0.10

9***

(0.017

4)-0.122

***

(0.022

)44

70.65

414

)Can

toniProt.

Maj.

1600

0.00

807

(0.019

6)-

--0.049

0**

(0.020

4)43

80.64

615

)IV

(Distance

Wittemberg)

0.17

4***

(0.047

6)0.31

5**

(0.127

)-0.282

***

(0.062

5)44

60.64

4

g)A

ltern

ativeDependentVar.

16)Wag

eUrban

MaleLab

.18

920.27

9***

(0.034

6)0.12

5***

(0.039

6)-0.246

***

(0.041

9)43

00.68

917

)IncomeTax

Reven

uep.c.

0.00

253

(0.116

)0.25

9*(0.138

)-0.373

**(0.152

)42

10.36

518

)%

Pop

inMan

uf./S

ervices

0.00

6(0.019

1)0.06

07**

*(0.022

2)-0.038

8*(0.022

9)44

70.75

3

Notes:

Depen

den

tvariable:Log

averag

ewag

e,maleelem

entary

school

teacher

1886

unless

di↵erentlysp

ecified

inthetable.Rob

ust

stan

-darderrors

inparenthesis

unless

di↵erentlysp

ecified

inthetable.**

*p<0.01

,**

p<0.05

,*p<0.1

Table 4: Di↵erence in Di↵erence, Dep. Var. Urbanization Rate

(1) (2) (3) (4) (5)Baseline Weighted Initial Years of Napoleon

Urbanization French ⇥ Years

Post 1850 9.925*** 10.38*** 8.734*** 11.64***(2.442) (2.325) (2.861) (2.949)

Napoleon ⇥ Post 1850 21.78*** 21.23*** 23.29***(7.467) (4.889) (6.842)

Post 1850 ⇥ Nap. ⇥ Prot. Share -28.13*** -29.18*** -30.45***(8.458) (6.563) (8.340)

Years of French ⇥ Post 1850 1.454***(0.272)

Post 1850 ⇥ French Yrs. ⇥ Prot. Share -2.233***(0.751)

Napoleon ⇥ 1750 10.49(12.01)

Napoleon ⇥ 1800 20.68(14.45)

Napoleon ⇥ 1850 19.38(12.50)

Napoleon ⇥ 1875 37.67**(16.93)

Napoleon ⇥ 1900 50.93**(20.67)

Prot. Share ⇥ Nap. ⇥ 1750 -12.46(14.67)

Prot. Share ⇥ Nap. ⇥ 1800 -24.65(16.57)

Prot. Share ⇥ Nap. ⇥ 1850 -24.65(15.06)

Prot. Share ⇥ Nap. ⇥ 1875 -47.80**(18.84)

Prot. Share ⇥ Nap. ⇥ 1900 -62.55**(21.95)

Constant 8.526*** 9.012*** 8.505*** 8.524*** 5.531***(0.832) (0.712) (0.826) (0.851) (1.611)

Number of id 19 19 19 19 19R2 0.506 0.530 0.509 0.503 0.878Obs. 109 109 109 109 109

Notes: All regressions have territory and year fixed e↵ects. Robust standard errors clustered by territory.Weighted regressions are weighted by territories total population in 1750. *** p<0.01, ** p<0.05, * p<0.1

Table 5: Dep. Var. Log average wage, male elementary teacher 1886(1) (2) (3) (4)

Share of literate Rel. Frag. Rel. Distance Horserace

Napoleon 0.0550*** 0.0848*** 0.0789*** 0.0525**(0.0188) (0.0198) (0.0225) (0.0206)

Share of Protestant 0.0867*** 0.168*** 0.165*** 0.0801***(0.0267) (0.0260) (0.0283) (0.0276)

Napoleon ⇥ Share of Protestant -0.0925*** -0.152*** -0.145*** -0.0860***(0.0264) (0.0281) (0.0295) (0.0275)

% of literate 0.423*** 0.435***(0.0566) (0.0608)

Religious Fragmentation 0.0744** 0.00931(0.0323) (0.0364)

Rel. Dist. from Neighbours -0.0482 -0.0414(0.0408) (0.0418)

Constant 5.003*** 6.089*** 6.151*** 4.967***(0.434) (0.404) (0.413) (0.450)

Geographic Controls yes yes yes yesHistorical Controls yes yes yes yesEconomic Controls yes yes yes yesEducation controls yes yes yes yesR2 0.691 0.663 0.660 0.694Obs. 447 447 431 431

Notes: Geographic Controls: latitude, distance from Berlin (km), area of the county (log), dis-tance from the district capital and polish speaking area. Historical controls: year of annexa-tion to Prussia, population in 1500, schools in 1517, universities in the Holy Roman Empire,Hanseatic or Imperial cities and monasteries per sqkm in 1517. Economic controls: total pop-ulation size (log), percentage of county population in urban areas 1871, percentage of laborforce in mining 1882 and number of farms 1882 (log). Education controls: percentage of pupilswith distance to school over 3 km, total number of pupils 1886 (log), total number of teachers1886 (log) and number of free apartments for male teachers 1886. Robust standard errors inparenthesis. *** p<0.01, ** p<0.05, * p<0.1

Table

6:Dep.Var.

Logavera

gewage,male

elementa

ryschoolteach

er1886

Instit.Proxim

ity

Culture:Pro-French

Culture:French

Ties

Horserace

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Nap

oleon

-0.263

***

-0.193

***

-0.304

***

-0.240

***

-0.035

40.07

36*

-0.258

***

-0.249

***

(0.044

8)(0.058

0)(0.036

7)(0.050

5)(0.038

9)(0.041

6)(0.045

1)(0.043

8)Shareof

Protestan

t0.13

1**

0.12

8**

0.16

7***

0.11

5**

0.12

7**

(0.055

2)(0.049

1)(0.055

0)(0.052

2)(0.047

9)Nap

oleon⇥

Shareof

Protestan

t-0.077

5*-0.068

2-0.151

**-0.064

3-0.074

8*(0.045

2)(0.043

1)(0.067

0)(0.045

7)(0.041

4)Institution

alProxim

ity

-0.334

***

-0.308

***

-0.369

***

0.74

3***

(0.061

9)(0.058

9)(0.034

0)(0.162

)Nap

oleon⇥

Inst.Proxim

ity

0.35

8***

0.34

5***

0.41

0***

-0.746

***

(0.062

3)(0.059

8)(0.034

6)(0.174

)Pro-French

-0.514

***

-0.479

***

-1.689

***

(0.073

7)(0.059

1)(0.282

)Nap

oleon⇥

Pro-French

0.68

1***

0.65

4***

1.86

4***

(0.076

8)(0.068

1)(0.276

)French

Ties

-0.056

5-0.071

7-0.191

***

(0.043

4)(0.045

7)(0.026

6)Nap

oleon⇥

French

Ties

0.11

8**

0.12

2**

0.26

0***

(0.056

1)(0.056

4)(0.042

1)Con

stan

t7.14

8***

7.24

8***

7.08

5***

7.19

3***

5.53

0***

5.90

3***

7.04

4***

7.13

7***

(0.525

)(0.457

)(0.538

)(0.456

)(0.739

)(0.641

)(0.388

)(0.456

)

HistoricalCon

trols

yes

yes

yes

yes

yes

yes

yes

yes

Geograp

hic

Con

trols

yes

yes

yes

yes

yes

yes

yes

yes

Econom

icCon

trols

yes

yes

yes

yes

yes

yes

yes

yes

R2

0.69

60.70

80.70

00.71

40.64

60.66

70.73

10.72

0Obs.

447

447

426

426

447

447

447

426

Notes:

Geographic

Controls:latitude,

distance

from

Berlin(km),

area

ofthecounty

(log

),distance

from

thedistrictcapital

and

polishsp

eakingarea.Historicalcontrols:year

ofan

nexationto

Prussia,pop

ulation

in15

00,schoolsin

1517

,universities

inthe

HolyRom

anEmpire,

Han

seatic

orIm

perialcities

andmon

asteries

per

sqkm

in15

17.Economic

controls:totalpop

ulation

size

(log

),percentage

ofcounty

pop

ulation

inurban

areas18

71,percentage

oflabor

forcein

mining18

82an

dnumber

offarm

s18

82(log

).Educationcontrols:percentage

ofpupilswithdistance

toschool

over

3km,totalnumber

ofpupils18

86(log

),totalnumber

ofteachers18

86(log

)an

dnumber

offree

apartm

ents

formaleteachers18

86.

Standarderrors,clustered

atthepre-N

apoleon-reign

level,in

parenthesis.

***p<0.01

,**

p<0.05

,*p<0.1