UvA-DARE (Digital Academic Repository)
Extractive states: the case of the Italian unification
De Oliveira, G.; Guerriero, C.
Link to publication
Citation for published version (APA):
De Oliveira, G., & Guerriero, C. (2014). Extractive states: the case of the Italian unification. (ACLE Working
Papers; No. 2014-03). Amsterdam: ACLE, University of Amsterdam.
General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),
other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).
Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating
your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask
the Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,
The Netherlands. You will be contacted as soon as possible.
UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)
Download date: 16 Jun 2017
Extractive States:
The Case of the Italian Unification.∗
Guilherme de Oliveira and Carmine Guerriero
ACLE, University of Amsterdam
December 10, 2014
Abstract
One of the most puzzling issues in economics is the huge divide between the North
and South of Italy despite more than 150 years of common formal institutions. We document that the opening of the present-day gap in culture and economic outcomes is the
consequence of the region-specific tax policies selected between 1861 and 1911 by the
elite of the Kingdom of Sardinia, which annexed the rest of Italy in 1861. Regional preUnification tax revenues indeed are mainly explained by the contemporaneous farming
productivity but not by the region’s political relevance for the Kingdom of Sardinia’s
elite, whereas the opposite is true for the post-Unification period. Moreover, tax distortions, tax-collection costs, and each region’s political relevance significantly shaped
the increasing post-Unification differences in income per capita, electoral turnout, and
human capital between the North and the South. This evidence, which is consistent
with a two-region, two-social class model, sheds new light on the incentives of a dominating group in a political or economic union, e.g., the European Union.
Keywords: Extractive States; Political Union; Culture; Taxation.
JEL classification: H20; H70; N4; Z10.
∗
We would like to thank Marco Casari, Giuseppe Dari-Mattiacci, Enrico Perotti, Laura Rondi, Avraham
Tabbach, Guido Tabellini, and Paola Valbonesi and seminar participants at UvA for the insightful comments, and Raffaella Paduano for the help in gathering and arranging the data. We are also indebted to
Mark Dincecco, Giovanni Federico, Luigi Guiso, Andy Hanssen, Giovanni Iuzzolino, Paolo Malanima, Luca
Pennacchio, Paolo Pinotti, SVIMEZ, Giovanni Vecchi, and Andrea Vindigni for the data provided and to
the staff of the Biblioteca Nazionale in Roma for the guidance. Financial support by the Fundação para
a Ciência e Tecnologia through Grant SFRH/BD/76122/2011 is gratefully acknowledged. Corresponding
author: Carmine Guerriero. Address: Plantage Muidergracht 12, 1018 TV, Amsterdam, The Netherlands.
Phone: +31 (0)205254162. Fax: +31 (0)205255318. E-mail: [email protected]
1
Introduction
The huge cultural and economic differences among its regions despite 150 years of common
political and legal institutions have made Italy one of the most frequented laboratory in
economics. In 2008 indeed Southern Italy displayed a 12 percent lower share of respectful
people and a 40 percent lower income per capita than Northern Italy (Iuzzolino et al., 2011).1
A well-known literature has traced back this gap to the diverse political trajectories of these
regions during the Middle Ages (Putnam et al., 1993). In particular, the experience of a more
inclusive political system—i.e., the communes—would have helped Northern Italy develop
a stronger culture of cooperation easing property rights protection, contract enforcement,
and the functioning of formal institutions (Tabellini, 2010; Guiso et al., 2014). Yet, recent
papers have raised several doubts on this slant.2 First, Boranbay and Guerriero (2013) show
that the main driver of present-day culture in Europe is the need of sharing climate-driven
consumption risk during the Middle Ages rather than past political institutions, and that
up to the 17th century the two parts of Italy did not display significant differences. Second,
a growing body of research reveals that, in the aftermath of the Unification, the two blocks
fared similarly well in terms of development (Pescosolido, 2011; Ciccarelli and Fenoaltea,
2013). Building on these contributions, we document that the opening of the present-day
gap in culture and economic outcomes is the consequence of the region-specific tax policies
selected between 1861 and 1911 by the elite of the Kingdom of Sardinia, which annexed the
rest of Italy in 1861. These policies discriminated more the regions farther away from the
Northern frontiers and so less politically salient for the unified government.
These patterns are consistent with a two-region, two-social class model that we construct
to inform our empirical test. Formally, we consider a simple and yet general society split in
two mass one citizenries, each living in a region and consuming a private and a public good.
The Northern region represents the Kingdom of Sardinia and the Southern one can be seen
as any of the absolutists polities that were annexed at Unification. The production of the
private good is in the hands of the citizenry and its technology is linear in a region-specific
1
The first statistics represents the share of answers mentioning “tolerance and respect for other people” as
important qualities children should be encouraged to learn to the 2008 European Value Study.
2
Another literature claims that genetic differences are the crucial driver of the North-South imbalances (Lynn,
2010). Yet, this odd argument has been proven completely unfounded by Felice and Giugliano (2011).
2
efficiency parameter and the citizenry’s investment in a culture of cooperation. The citizenry
consumes the untaxed supply of private good, whereas a part of the taxed one contributes
to the production of the region-specific public good, and the remainder accrues to a mass
zero “elite.” Under autarky, each regional elite selects a tax rate maximizing the sum of
her rents and the welfare of the region’s citizenry net of linear tax-collection costs. Since
these costs are sizable because of limited state capacity, the equilibrium tax revenues fall
with the extent of technological efficiency. Under political union instead, both region-specific
tax policies are selected by the Northern elite who bears lower marginal tax-collection costs
than under autarky, devotes a share of the Southern tax revenues to public good production
lower than that committed by the Southern elite under autarky, and delivers a less useful
public good to the Southern population. This set of assumptions is consistent with the fact
that the Kingdom of Italy had a state capacity higher than those of the pre-unitary states
and was initially politically dominated by the elite of the former Kingdom of Sardinia. This
liberal bourgeoisie was mainly interested in getting the Northern industry off the ground.
Thus, extraction is cheaper under political union and escalates with the factors that boost its
returns and, in particular, the extent of technological efficiency. Provided that the Southern
technology is backward and so the disincentive from taxation is limited, tax revenues in
the South under political union are also higher than under autarky and decrease with the
South’s political relevance for the North, i.e., the share of tax revenues invested by the
Northern elite into the production of the Southern public good. Finally, Unification can
depress cultural accumulation and welfare in the South since both rise with the marginal
tax-collection costs and the South’s political relevance, which in turn curb extraction and so
investment distortion. The gist of the model survives when we allow the South to specialize
in a traditional sector. In this case, the risk of losing rents can discourage extractive policies.
To test these predictions, we study over the 1801-1911 period those regions that were
absorbed by the unitary state but were not part of the Kingdom of Sardinia before Unification. We proxy the post-Unification political relevance of each of these regions with the
average distance of its pre-Unification main city from the capital of the foreign power most
adverse to the Kingdom of Sardinia. Turning to tax policies, we exploit the fact that the
Italian economy was essentially agrarian and thus we focus on the revenues from land prop3
erty taxes. These hit the profitability from land property and, in particular, the landowner’s
effort and investment in technology. To capture the efficiency of the farming technologies
most widespread at the time, we look at the geographic determinants of the profitability
of sericulture and the production of wheat, citrus, and olives. Finally, we gage marginal
tax-collection costs with a proxy for fiscal capacity, i.e., military personnel per 1000 citizens.
Consistent with our model, pre-Unification tax revenues are mainly explained by the
contemporaneous farming productivity but not by the region’s political relevance for the
Kingdom of Sardinia’s elite, whereas the opposite is true for the post-Unification period.
Moreover, each region’s political relevance, the unified state’s marginal tax-collection cost,
and tax distortions significantly shaped the increasing differences in income per capita, electoral turnout, and human capital between the North and the South over the period in which
tax policies remained region-specific, i.e., 1861-1911. We measure tax distortions with the
difference between the tax revenues observed in the region and those that would have prevailed without Unification according to our positive taxation model. To determine whether
the correlations we document are, in fact, causal, we follow a two-step strategy. First, we
control not only for fixed region effects, but also for other relevant observable factors, like
wages, income inequality, energy endowment, and population density. Including these controls has little effect on the gist of our results. Second, we use insights from Altonji et al.
(2005) to evaluate how much greater the influence of unobservables would need to be, relative to observables, to completely explain away the relationships we uncover in the data. We
find that this influence would have to be on average 7 times greater than that of observables.
Hence, it is quite unlikely that our estimates can be attributed to unobserved heterogeneity.
Albeit a long literature has related the present-day North-South divide to post-Unification
policies (Aprile, 2010), no other work has built and tested a model clarifying how these
policies solved the trade-off between extraction-related losses—i.e., investment distortions,
marginal extraction cost, and military weakness—and the rent-seeking gains faced by the
unitary government.3 In this respect, our theory of endogenous extractive policies represents a natural complement to those contributions contrasting “extractive” and “inclusive”
3
Felice (2014) points at the extractive habits of Southern elite to explain the present-day divide. However,
this mechanism is completely irrelevant to explain post-Unification economic outcomes (see section 5.4).
4
institutions (North et al., 2009; Acemoglu and Robinson, 2012). Differently from them,
we focus on the determinants of the conflict arising between a dominating geographic—and
in general cultural—group and the rest of society given a certain level of the inclusiveness
of the prevailing political institutions. The German government’s opposition to the rescue
packages demanded by Greece just before the 2011 European sovereign debt crisis (Guiso et
al., 2013) and the extremely costly political tensions between the Basque Country and the
central Spanish government (Abadie and Gardeazabal, 2003) and between Northern Ireland
and the UK (Besley and Mueller, 2012) constitute a case in point of this type of conflicts.
The paper proceeds as follows. Section 2 reviews the key facts about the political organization, economy, and tax system of the Italian regions over the 1811-1911 period to motivate
the model. Section 3 illustrates the set up of the model and documents the robustness of its
message to alternative assumptions. Section 4 states the model implications that are tested
in section 5. Section 6 concludes. The appendix gathers both tables and figures.
2
Italy Before and After the Unification: A Primer
Next, we describe the political and economic contexts of the Italian regions over the
1811-1911 period, detailing at the same time the main innovations in tax policies.
2.1
The Era of Risorgimento
Italy emerged from the Congress of Vienna as a thoroughly designed strategic equilibrium
among eight absolutists states: the Northwestern Kingdom of Sardinia ruled by the Piedmontese House of Savoys; the Northeastern Kingdom of Lombardy-Venetia, which was under
the direct control of Austria; the Grand Duchy of Tuscany ruled by branches of the Austrian
Habsburg dynasty; the Southern Kingdom of the Two Sicilies in the hands of the Bourbons;
the Central Papal State, Duchies of Modena and Parma controlled by the Habsburgs, and
Duchy of Lucca then absorbed by Tuscany in 1847.4 Such a status quo reflected the territorial divisions established by the treaty of Campo Formio (1797) before the Napoleonic
conquests and served two key purposes. First, it deprived the Bourbons of any interest in
waging war being their only neighbor the Pope, who in turn was constrained by his religious
4
Our historical account is based on Hearder (1983) and Galasso (2011).
5
role. Second, it kept in check Austria and France by leaving the Kingdom of Sardinia as a
buffer state between the two powers. Yet, exactly this threat-based balance fed the ambitions of the Savoys making the Kingdom of Sardinia the champion of that part of the Italian
elites, inspired by the eighteenth-century Enlightenment and known as “patrioti,” longing
to establish a liberal unified state. Inspired by this aim and supported by urban workers
and lower military ranks, the patrioti organized a series of subversive acts in the wake of
the European unrests of 1820, 1830, and 1848. Even if none of these turmoils overthrowed
a pre-unitary regime, they forced the absolutist rulers to implement some of the liberal laws
and regulations brought by the Napoleonic armies and inspired by the French revolution.
This institutional discontinuity hit the whole peninsula but was particularly relevant for
the Southern population, who could finally see realized the reforms introduced during the
French seizure, i.e., the privatization of one third of the clerical and common lands initially
implemented in 1792 and the end of the Feudal system declared in 1806. The liberalization
surge did not release the Italian peasants from their destitution but allowed a rising class of
bourgeoisie to acquire part of the nester nobility’s domains and prioritize market-oriented
activities like arboriculture and sericulture over the classic subsistence farming, i.e., wheat
(Pescosolido, 2011). Both enterprises were capital intensive. Silk is obtained from the fibers
extracted from the cocoons of the larvae of the mulberry silkworm dissolved in boiling water.
The fibers are spun through reels, which at the time were powered by watermills (Britannica,
2014). The basic need of water favored the concentration of sericulture in the irrigated Po
valley. Citrus and olive trees instead cannot tolerate a temperatures below 4 degree Celsius
(Britannica, 2014). This explains their almost exclusive presence in the South (Dimico et
al., 2012). Farming productivity quickly increased in both the sharecropping-based Northern
farms and the Southern latifundia, which contrary to the conventional wisdom maintained
a net technological primacy until the Unification (Federico, 2007). Indeed, thanks to the
training, job security, and credit line assured to its permanent employees, the “latifondisti”
could invest in a setting characterized by a low level of conflict (Petrusewicz, 1996).
After two centuries and half of economic decline (Malanima, 2011), the Italian population
started to grow again doubling over the 1800-1860 period whereas the GDP per capita after
a first rise stagnated until the 1870 in the absence of any significant industrial revolution
6
(Pescosolido, 2011).5 This pattern is clearly depicted in the left graph of figure 1 where we
look at the regions that we will analyze in our empirical exercise—i.e., those outside the
Kingdom of Sardinia and absorbed by the Kingdom of Italy in 1861—differentiating them
between “North”—i.e., Emilia Romagna, Lombardy, Tuscany, and Veneto, “Center”—i.e.,
Latium, Marches, and Umbria, and “South”—i.e., Abruzzi, Apulia, Basilicata, Calabria,
Campania, and Sicily. The development of the pre-Unification Italian economy shows only
limited regional differences with a slight primacy of the South and a general condition of
backwardness. To elaborate on this last point, at Unification still 68 (only 3) percent of the
active population was employed (worked exclusively) in the agricultural (industrial) sectors,
and the production of iron (number of spindles) was less than 1 percent of the English one
(Pescosolido, 2011). The causes of this gap were the scarcity of coal and the paucity of
human and real capital.6 This second aspect manifested important regional dissimilarities.
While the Northern credit markets were slowly starting to become diversified with both
small institutes—i.e., “casse di risparmio”—and medium sized commercial banks ready to
serve the newborn industry, those of the South were still characterized by the plethora of
micro-credit institutions appeared from the 15th century on, i.e., “Monti” (Felice, 2014).7
Equally underdeveloped but even more polarized was the educational structure. The share
of illiterates aged over six in the total population ranged in 1861 from 53.7 percent in
Lombardy to 88.6 percent in Sicily and was significantly higher in the South (see table 1
for the definition, sources, and statics of the variables we use, and table 2 for further details
on the North-South divide). Obviously, in the absence of developed private markets, the
asymmetry in human and real capital accumulation could only reflect disparities in the fiscal
capacity and, in turn, the political objectives of the pre-unitary states (Dincecco et al., 2011).
The 1848 revolutionary wave was by far the most violent. The patrioti tried to convince
their rulers to participate in a war of independence against Austria led by King Carlo Alberto
5
The values of GDP-L for the pre-Unification South are inferred from the estimates of the GDP in Italy and
in the Center-North developed by Malanima (2006, 2011) but criticized from an ideological point of view
(Daniele and Malanima, 2014) by Felice (2014). We use them for descriptive purposes only.
6
In 1861 the total production of coal in Italy was 34,000 tonnes versus the 85 millions quarried in England,
the 18.7 millions extracted in Germany, and the 9.4 millions mined in France (Pescosolido, 2011).
7
Not surprisingly, over the 1861-1911 period, the deposits in “casse di risparmio” in 1861 lire per capita were
significantly larger in the North—i.e., 64.11 versus 4.49 (SVIMEZ, 2011), whereas the number of credit firms
per thousands of people was significantly higher in the South, i.e., 0.04 versus 0.03 (Missiaia, 2012).
7
of Piedmont. Yet, the lack of fervor of these monarchs first burked the liberal constitutions
granted after the revolts and then left Piedmont at the mercy of Austria. The inevitable
defeat pushed Carlo Alberto to abdicate in favor of his successor Vittorio Emanuele II, who
uphold the liberal constitution to calm down the internal uprisings and gain support for
his untamed territorial ambitions. This move allowed the Genoese merchant elite to obtain
public investments in transportation infrastructure and the king to raise the taxes necessary
to finance military spending. Military personnel as a portion of total population and the
related expenditures in the Kingdom of Sardinia became almost twice as big as in any other
pre-unitary state so that its public debt skyrocketed to a level that, in 1860, was half of the
debts of the remaining pre-unitary states (Dincecco et al., 2011). The lack of an expansionist
foreign policy implied lower investments in fiscal capacities elsewhere against a background
of regional differences. Because of fiercer domestic unrests, the average military spending per
capita in the Papal State and Two Sicilies were 50 (80) percent higher than those in Tuscany
(Modena). Vienna instead kept taxation low in Lombardy-Venetia to appease local elites,
who threatened to defect in favor of the Kingdom of Sardinia. Thus, pre-Unification revenues
from land property taxes in 1861 lire per capita remained slightly larger in the South than
in the rest of our sample (see table 2 and figure 1). This was the most conspicuous among
the direct taxes envisaged by the pre-unitary tax systems and, being based on the generated
rent, varied across areas (Marongiu, 1995). The larger costs of internal conflicts in the Two
Sicilies and the Papal State on the one hand and the subsidy allowed to Lombardy-Venetia
by Vienna on the other hand justify the above-mentioned North’s advantage in terms of
human and real capital. Nevertheless, this lead was minimal and not determinant from a
development point of view as also revealed by the average pre-unitary railway network, which
was not significantly different between North and South—i.e., respectively 0.028 and 0.022
km per capita (Romani, 1982; SVIMEZ, 2011), and by the tonnes of merchant navy per
capita in 1860, which were higher in the South, i.e., 0.035 versus 0.027 (Pescosolido, 2011).
2.2
Italian Unification and the Rise of the North-South Divide
During the 1850s the Piedmontese parliament’s power relative to the king grew steadily,
and its dominant figure became count Camillo of Cavour. Appointed prime minister in 1852,
8
he recognized that the Kingdom of Sardinia alone could not match Austria militarily. Hence,
he participated in the Crimean War (1853-1856) to win the favor of France and England.
Cavour’s strategy was a success, and in 1858 he signed a secret pact with Napoleon III to
wage battle against Austria. In the summer of 1859, the two allies defeated the Austrian
military in Lombardy. This victory triggered insurrections in Tuscany, the military leader
Giuseppe Garibaldi’s conquest of the South, and the successful invasion of the Papal State.
A new unified Kingdom of Italy was proclaimed on March 17th, 1861.
Over and above the annexation of Veneto (1866) and Rome (1870), the desire of getting
the Northern industry off the ground was the crucial objective of the first unitary governments, which were dominated up to the 1873 by the Piedmontese faction of the right wing
party (see figure 2). Expression of the rising liberal elite of the ex-Kingdom of Sardinia,8
they foisted Piedmontese legislation and, in particular, its liberal policies on the rest of the
country, favored the Northern industry in the allocation of public contracts, and increased
the taxes levied on the annexed reigns to pay back the mounting public debt and finance
new infrastructures:9 albeit optimal from the policymakers’ point of view, these policies soon
produced unplanned and dramatic consequences for the whole peninsula.
The Piedmontese excise, one of the lowest in Europe, was immediately dictated to the
whole state so that the regions part of the ex-Kingdom of Two Sicilies experienced a 80
percent fall in their customs. As a result, while the exports of silk exploded, those of olive
oil and citrus soon crashed when cheaper substitutes arrived from the USA. The 1887 rise in
customs did not change this trend since it affected mainly the North-based wheat farming and
production of iron, which was boosting thanks to the public subsidization of the Genoese
steamboat producers and the legal exclusion of Neapolitan firms from coastal navigation
(Pescosolido, 2011). A similar procurement policy was followed in the railways sector (see
figure 2). To elaborate, the relative spending in 1861 lire per capita remained for a long time
higher in the Northern part of Italy, which consequently ended up displaying a significantly
larger network, i.e., 0.1 versus 0.06 Km per capita over the 1861-1911 period (SVIMEZ,
2011). Agrarian policies were similarly discriminatory: just 4 percent of the pre-First World
8
The opposing faction lumped together the annexed reigns’ elites and championed through its leader, Marco
Minghetti, the idea of a modern decentralized state concerned with each region’s economic interests.
9
In 1861 the public debt amounted to the striking level of 40 percent of the total GDP (Pescosolido, 2011).
9
War spending in land reclamation was allocated to the South (Iuzzolino et al., 2011).
More important, the impressive infrastructural program supporting the Northern industry was financed out of highly unbalanced tax rises. In the first decade of the unified state,
the average total revenues from land property taxes skyrocketed from 1.78 to 7.22—1861—
lire per capita but the burden was borne chiefly by the South (Ministero delle Finanze,
1872). The 1864 tax reform indeed first established a sum to be raised—i.e., 125 millions,
and then allocated it among nine districts resembling the pre-unitary states—i.e., Piedmont,
Lombardy, Parma and Piacenza, Ex-Duchy of Modena, Tuscany, Ex-Papal State, Neapolitan
provinces, Sicily, and Sardinia (Marongiu, 1995).10 The regions outside the ex-Kingdom of
Sardinia provided 80 percent of the total revenues, which in turn were paid for the 50 percent
by the regions part of the ex-Kingdom of the Two Sicilies, 14 percent by those part of the
ex-Papal State, and only 36 percent by the other districts (see tables 1 and 2 and figure 1).
The most severe phase of these extractive policies came to an end with the achievement of
the balanced budget in 1876, but by then the divergence of the two economies was irreversible
(Nitti, 1905). By the end of the 19th century, the steel, clothing, and mechanical industries
were wiped out from the South, while a competitive manufacturing and heavy equipment
industry was established in the North (see table 2 and Iuzzolino et al., [2011]). Furthermore,
the economic impoverishment had deteriorated the government-citizenry relationship and the
cooperation in the South. On the one hand, a sizable part of the population experienced the
Unification as an undue occupation by the Piedmontese armies. This unchained a ferocious
series of crashes, which goes under the name of “brigandage.” Between 1861 and 1864,
this civil war brought around 20,000 Southern victims, imposed a de facto militarization
of the area,11 and opened the way to the massive emigration from the countryside (Felice,
2014). Accordingly, the 1890s ratio of gross emigration to total population accelerated in
the South surpassing by far the one in the Center-North (Iuzzolino et al., 2011). On the
other hand, the population started to display an increasingly lower culture of cooperation
as documented by the fall of the turnout rate (see the two rightmost graphs in figure 3),
which constitutes an outcomes-based measure of social capital (Guiso and Pinotti, 2013).
10
11
This mechanism endured until the introduction of the provincial cadastres in 1922 (Marongiu, 1995).
In 1870 there was half of the Piedmontese army—120,000 units (Del Monte and Pennacchio, 2012).
10
Contrary to the conclusions reached by a long literature on permanent institutions (Guiso
et al., 2014), the 1911 snapshot and not the 1861 one should be considered an outlier. To
see this, the leftmost graph in figure 3 compares the cumulated discounted numbers of years
of activity over the 1000-1600 period of Cistercian and Franciscan houses in the North and
South of Italy. Boranbay and Guerriero (2013) document that in the Middle Ages these
monks met the population’s demand for insurance against consumption shocks in exchange
for the acceptance of a culture of cooperation so that, at the European regional level, there is
a strong correlation between their expansion and present-day self-reported norms of respect
and trust. Thus, a glance to figure 3 suggests that post-Unification policies seem to be also
a powerful account of the present-day divide in culture across Italian regions.
Later on, the fascist regime’s aversion to migration and its rush to arming managed
through a series of investments in the Northern heavy industry have stretched even more the
North-South divide, which has been only scratched by the 1960s economic booms and state
aids.12 Despite these more recent events however, it should be clear at this point that the
present-day North-South gap opened because of the extractive—especially taxation—policies
implemented by the Kingdom of Italy’s government in the aftermath of the Unification.
3
Theory
Consider a territory divided in two regions r ∈ {N, S} each inhabited by a mass one of
equal citizens consuming a public good gr and a private good, whose demand and supply are
Dr and Xr . Building on section 2, the N region is the Kingdom of Sardinia, whereas S can be
seen as any of the other regions part of the Kingdom of Italy. Each citizenry has full property
rights on the Xr production technology, which is linear in the region-specific technological
parameter Ar > 0 and her investment in a culture of cooperation Cr , i.e., Xr = Ar Cr . This
assumption is consistent with the huge evidence according to which a culture of cooperation
is pivotal to curb transaction costs, expand anonymous market exchange, and facilitate the
division of labor (Tabellini, 2010; Guerriero, 2014). In terms of our historical case, Xr is the
product of a farming technology requiring an increasingly larger quantity of intermediate
inputs, a higher division of labor, and a more sophisticated technology, i.e., silk or citrus.
12
Iuzzolino et al. (2011) emphasize these channels, which are fully consistent with our model.
11
At time zero, the citizenry of region r fixes a level of culture to instill in its members
through a linear cultural accumulation technology (see Boranbay and Guerriero, [2013]).
This assumption incorporates into the model the key insights of evolutionary psychology
(Barkow, Cosmides, and Tooby, 1992) and Malthusian growth theories (Clark and Hamilton,
2006): group-specific cultural values result from a process that instills, via natural selection
or cross-punishment, norms maximizing the fitness of the group’s members. Thus, a citizenry
expecting higher returns from cooperation ends up showing a stronger Cr . At time one and
under autarky (political union), region r (N ) elite, who has mass zero, selects the rate(s)
tr (tUN and tUS ) at which the production of the private good is then taxed. At time two
and under autarky (political union), a share α < 1 (α in the North and αU < α in the
South) of the regional revenues is used by the elite (Northern elite) to produce the regionspecific public good gr , whereas the remainder winds up in region r (N ) elite’s pocket. The
public good production function is linear, and each tax rate maximizes the elite’s rents plus
the citizenry’s welfare net of linear tax-collection costs. Thus, αU < α implies that under
political union the Northern elite attaches a higher weight on the welfare of the citizenry of
her own region than on the welfare of the Southern citizenry or, similarly, that αU increases
with the political relevance of the South for the Northern elite. This set up captures the
post-Unification evolution of political institutions. At time three, each citizenry obtains a
concave utility from consuming the untaxed output and the region-specific public good.
3.1
Autarky
Under autarky, an r citizenry selects a Cr maximizing the objective function
q
p
p
Dr + γgr − Cr =
1 − t̂r Xr + αγ t̂r Xr − Cr = t̃r Ar Cr − Cr ,
(1)
where hats label equilibrium quantities, t̃r ≡ 1 − t̂r (1 − αγ), and γ gages the citizenry’s
relative preference for gr vis-a-vis Dr .13 We maintain that αγ > 1 so that region N (S)
citizenry prefers public to private good production because of, for instance, the economic
development achievable only through new infrastructures as railways (harbor maintenance
and land reclamation). The unique and global equilibrium levels of culture and consumption
13
The gist of the model will be similar should the citizenry’s subutility from consumption be Cobb-Douglas.
12
are Ĉr = t̃r A4r and D̂r = 1 − t̂r Ar Ĉr so that the citizenry’s and social welfare is Vr = t̃r A4r .
Taking into account the citizenry’s choice, an r elite selects a tax policy maximizing
(1 − α − K) tr Ar Ĉr + Vr ,
(2)
n
o
γ−1
where the marginal tax-collection cost is such that K > max 1 − α, γ . This last assumption is consistent with the fact that, before the warfare that lead to the unified state,
the taxation power of the Italian reigns was limited (Dincecco et al., 2011). The unique and
global solution to problem (2) is t̂r =
1
2Ar (K+α−1)
1
− 2(αγ−1)
, which falls with K and decreases
with Ar since at the margin the taxation gains in terms of social welfare equal 1−αγ whereas
the taxation costs in terms of collection expenses net of rents are increasing in the technological efficiency. Finally, t̂r has an uncertain link with α, which both depresses the elite’s rents
and augments the incentive to invest via a higher payoff from public good consumption. The
equilibrium levels of culture and welfare equal
Ar
8
αγ−1
+ 8(K+α−1)
, which obviously rises with Ar ,
α, and γ and decreases with K that can be seen as inverse measure of the feasibility of public
good production. Finally, tax revenues t̂r Ar Ĉr =
αγ−1
16(K+α−1)2
2
Ar
− 16(αγ−1)
fall with Ar , decrease
(increase) with K (γ) being the product of two functions falling (rising) with K (γ), and
have an uncertain relationship with α when Ar is sufficiently small. As discussed in section
2, this is the scenario fitting the most the economy object of our empirical exercise.
3.2
Political Union
Building on the insights of section 2, we maintain that under political union the Northern
elite: 1. extracts all the rents and devotes a share of the Southern tax revenues to public
good production lower than that committed by the Southern elite under autarky because
of its political primacy; 2. bears lower marginal tax-collection costs in the South
γ U −1
γU
<
K U < 1 − αU because of the militarization of the area; 3. faces a technological constraint in
producing gS —i.e., γ U < γ so that αU γ U < 1—because of the differences between the public
goods most conducive to development in each region, e.g., railways and harbor maintenance
(Pescosolido, 2011). Since the regional tax revenues not appropriated by the Northern elite
produce the region-specific public goods, the citizenry’s problem is the same as in autarky
so that ĈNU = VNU = t̃UN A4N = t̃N A4N = VN = ĈN , and ĈSU = VSU = t̃US A4S with t̃US ≡
13
1 − t̂US 1 − αU γ . Therefore, the Northern elite selects a pair tUN , tUS maximzing
(1 − α − K) tN AN ĈNU + 1 − αU − K U tS AS ĈSU + VNU + VSU .
As a consequence, t̂UN = t̂N and t̂US =
regions. Tax revenues in the South
1
− 2AS (1−α1 U −K U ) since there is
2(1−αU γ U )
A2
1−αU γ U
t̂US AS ĈSU = 16(1−αSU γ U ) − 16(1−α
U −K U )2
(3)
no trade between
decrease with K
and increase with both γ U and AS . This last comparative statics is different with respect
the autarchy equilibrium because now rent-extraction incentives are stronger than marginal
tax-collection costs and thus an increase in technological efficiency and thus investment calls
for more taxation. Finally, the tax revenues raised in the South falls with αU (are higher
than under autarky) whenever AS is sufficiently small (and γ U is not too different from γ).
Intuitively, a fall in the South’s political relevance fosters the rent-seeking motives of the
Northern elite but curbs at the same time the Southern population’s incentive to invest. If
technological efficiency is limited, the first effect must prevail. The following proposition
summarizes the equilibrium of the endogenous taxation game under the two regimes:
Proposition 1: Under autarky (political union), the tax revenues raised in the South
fall (rise) with technological efficiency AS , decrease with the marginal tax-collection cost K
(K U ), and increase with the public good quality γ. Whenever AS is sufficiently small (and
γ U is not too different from γ), they fall with the political relevance of the South for the
Northern elite αU (are higher under political union than under autarky).14
Next, we take stock of the results obtained so far to analyze the impact on culture and
economic outcomes of an exogenous shock turning autarky into a political union.
3.3
The Rise of the North-South Divide
Under political union, cultural accumulation and welfare ĈSU = VSU =
AS
8
+
1−αU γ U
8(1−αU −K U )
increase with both the marginal tax-collection cost K U and the South’s political relevance for
the Northern elite αU , which in turn curb extraction and so investment distortion. The mix
14
2
(αγ−1)
If −∆K < 0, −∆α < 0, and −∆γ < 0 capture the impact of the Unification, t̂U
S > t̂S if AS (K+α−1)2 >
U
16(1−αU −K U )(αU γ U −1−K U γ U +γ U −1)
dt̂S AS ĈS
γ∆α+α∆γ
< 0 for AS → 0 if
<
∆K+∆α , which is true for AS → 0. Also,
dαU
(1−αU −K U )4
U
U
0, which is the case since αU γ U < 1 and γ γ U−1 < K U < 1 − αU . Similarly, t̂U
S AS ĈS > t̂S AS ĈS for AS → 0
U U
U
U
(1−α γ )
dt̂ A Ĉ
if −∆α SdαSU S + ∆K 8(1−αU −K U )3 > ∆γ 16(1−ααU −K U )2 , which is true for ∆γ → 0.
14
of lower marginal tax-collection costs and inefficient public good provision thus implies that
Unification can depress both cultural investment and social welfare in the South provided
that K U falls below
γ−1 15
.
γU
In terms of our historical experiment, given a level of K U , regions
to which a relatively high αU is assigned will gain from Unification, whereas those for which
αU tends to 0 will irreparably lose. The following proposition restates this series of remarks:
Proposition 2: Under political union, both culture ĈSU and social welfare VSU in the
South rise with the marginal tax-collection cost K U and the political relevance αU , and they
can be lower than the corresponding levels under autocracy for K U and αU sufficiently small.
Our model provides a theory of endogenous extractive policies in a political or economic
union dominated by one of its components. One key extension to our analysis is to endogenize the parameter α in order to consider the passage towards more democratic political
processes giving the citizenry the chance of constraining the elite’s choices (North et al.,
2009; Acemoglu and Robinson, 2012; Boranbay and Guerriero, 2013). Similarly, the elite
might be induced to extract less surplus from the South if worried that the Southern citizenry
could opt out of the modern sector producing XS and specialize in a sector demanding no
investments. We leave the first robustness to future research being less related to the case
analyzed in our empirical exercise, but we study the second one in the following.
3.4
General Equilibrium Disincentives to Extraction
The Southern citizenry can consume an alternative good TS —i.e., wheat—produced
through a “traditional” technology, which is linear in the productivity parameter NS with
√
1−αU γ U
NS < A4S − 4(1−α
Being the traditional sector technology indeU −K U ) , i.e., TS = NS .
pendent of investment activities, the indirect utility of citizenry S producing TS equals
p
p
(1 − τ̂SU ) TS + αU γ U τ̂SU TS = τ̃SU NS , where τSU is the tax rate levied on the traditional
good in the South and τ̃SU ≡ 1 − τ̂SU 1 − αU γ U . A bit of algebra shows that τSU =
1
1−αU γ U
−
1−αU γ U
4NS (1−αU −K U )2
and VSU =
1−αU γ U
2(1−αU −K U )
the traditional sector whenever αU <
3−AS (
3γ U −AS
1−K U
so that the Southern citizenry will select
)
≡ αU , even if the Northern elite will al-
ways prefer otherwise.16 Thus, she is now willing to extract a weakly lower surplus by acting
dĈS
dK
dĈS
dα
dĈ U
U
dĈS
dαU
15
Since
16
The Northern elite prefers to tax XS
<0∧
> 0 ( dKSU > 0 ∧
U
U
> 0), ĈS > ĈSU for α = 1, αU = 0, and K U = γ γ U−1 if γ−1
K >γ .
√
√
2
U U
U
U
U U
U
U √
[2(1−α γ )+4 NS (1−α −K )] −16 NS (1−α γ )(1−α −K ) NS
if
<
16(1−αU γ U )(1−αU −K U )
15
as the political relevance of the South was at least αU to levy taxes on a more productive
activity. The Northern elite will face a similar incentive should the most taxed citizenry have
the chance of selling some of its production to the other group. In this case, extraction is
curbed by the prospect of cheap imports. Since regional trades were limited over our sample
(Iuzzolino et al., 2011), we leave this extension to the basic setup to future research.
4
Empirical Implications
Our model produces two sets of implications regarding the thirteen regions outside the
Kingdom of Sardinia and completely annexed by the Kingdom of Italy in 1861.17 The first one
concerns the differences between the pre and post-Unification tax policies. The second one
deals with the post-Unification distortions in cultural accumulation and outcomes produced
by the higher state capacity of the unified state and the lower weight on the South’s welfare
put by the central authority. These implications can be restated as testable predictions as:
Predictions: Pre-Unification regional tax revenues will fall with both technological efficiency and the marginal tax-collection cost but will be independent of the region’s political
relevance for the Kingdom of Sardinia’s elite. Post-Unification tax revenues instead will increase with technological efficiency and decrease with the marginal tax-collection cost and the
region’s political relevance. Finally, post-Unification distortions in outcomes and culture will
be higher the lower are the marginal tax-collection cost and the region’s political relevance.
5
Evidence
To test our predictions we need, first and foremost, meaningful proxies for the revenues
from the taxes on the most relevant production activities in our sample, the efficiency of the
related technologies, the cost of raising these taxes for each reign’s government and for the
unified state, and each region’s political relevance for the Kingdom of Sardinia’s elite.
√
1−αU γ U
, which is the case under our restriction on NS , i.e., if NS < A4S − 4(1−α
U −K U ) .
17
They are Lombardy, Veneto, Emilia, Tuscany, Marche, Umbria, Lazio, Abruzzi, Campania, Puglia, Basilicata, Calabria, and Sicily. Piedmont, Liguria, and Sardinia were part of the Kingdom of Sardinia, whereas
Trentino-Alto Adige (Friuli-Venezia Giulia) was annexed (in its entirety) to the Kingdom of Italy in 1919
(1918 even if Friuli was already part of it from 1866). Finally, Molise was separated from Abruzzi in 1963.
2
[1−αU γ U +AS (1−αU −K U )]
16(1−αU γ U )(1−αU −K U )
16
5.1
Measuring Taxation and Its Determinants
Consistent with the extant literature on the North-South divide (Ciccarelli and Fenoaltea,
2013), we base our analysis on 10-year benchmarks; moreover, we focus on the period, before
and after 1861, over which tax rates remained region-specific and our regions kept a stable
political and territorial identity, i.e., 1801-1911 (Galasso, 2011).18 We employ present-day
regional boundaries but our results are similar when we switch to the historical ones.
Since regional tax rates are unavailable, we focus on the revenues raised through the
land property tax—i.e., “imposta sul valore fondiario”—in 1861 lire per capita, Land-Taxes,
which represented the great part of total revenues from direct taxes—i.e., 43 percent over our
sample—and hit the average profitability from land property. Before Unification, Land-Taxes
ranges from zero in the Papal State to 3.26 in the Kingdom of the Two Sicilies. This pattern
conceals a huge variation in tax rates and income, which persisted after the Unification.
To show that our results are not driven by unobserved differences in land availability and
suitability, we document in the Internet Appendix that they remain similar if we divide total
revenues by the arable area in square kilometer collected from SVIMEZ (2011) or control for
the land quality estimated by the Atlas of the Biosphere interacted with time effects.
To gage the efficiency of the most-widespread farming activities, we look at the geographic determinants of the technologies producing silk, citrus, olives, and wheat. While
the first three represent the most profitable activities discussed in the basic model, the last
one embodies the traditional sector we analyze in section 3.4. Building on section 2, we
use as proxy for the efficiency of the sericulture technology the normalized first principal
component extracted from the share of the region’s area covered by large lakes and rivers
and the average growing season precipitation in the previous decade in ml, Sericulture.19
Turning to the efficiency of wheat (arboriculture) breeding, we consider the normalized first
principal component extracted from the land suitability(ies) for wheat (citrus and olive trees)
and the average growing season precipitation in the previous decade in ml (temperature in
18
19
This is roughly the period elapsing between the treaty of Campo Formio (1797) and the First World War.
The former (latter) variable is in grid format, covers the entire World (Europe) at a 0.5 degrees spatial
resolution (for the period 1400-1900), and is collected from the GLWD dataset (Pauling et al., 2006). To
the best of our knowledge, there is no information on the land suitabilities for mulberry and grapevine in
our sample. Characterizing the viticulture technology is an important avenue for future research since the
export of wine was an important source of income in 19th century Italy (Pescosolido, 2011).
17
Celsius), Wheat (Arboriculture).20 The choice is driven by the fact that the efficiency of
these technologies weakly grows with the average land suitability, precipitation, and temperature (Britannica, 2014), but only the first and second (first and third) geographic traits are
positively correlated with Wheat (Arboriculture). To cross validate our proxies, first we document their high correlation with the production of each item between 1871 and 1911, and
then we show that our conclusions are similar when we use either one of the land suitabilities
for citrus and olive trees to evaluate each technology (see the Internet appendix).
Turning to the marginal tax-collection costs, we follow Dincecco et al. (2011) and elect as
a proxy the military personnel per 1000 citizens, Army. A vast body of research has shown
that two key determinants of a state capacity to raise revenue are the risk of external or internal conflicts and the degree of political instability (Besley and Persson, 2009). Accordingly,
higher levels of Army should correspond to a lower marginal tax-collection cost (Dincecco et
al., 2011). Our results will be similar should we consider instead the share of the previous
decade in which the state participated in external wars (see the Internet appendix).
Finally, we employ as an inverse metrics for a region’s political relevance for the Kingdom
of Sardinia’s elite the average distance in kilometers of its pre-Unification main city from
the capital of the foreign power most adverse to the Kingdom of Sardinia, i.e., Vienna
over the 1801-1813, 1848-1881, and 1901-1914 periods, and Paris otherwise, Distance-toEnemies. Our choice can be explained as follows. Fallen under Napoleon in 1796, the
Kingdom of Sardinia together with the former Republic of Genoa became after the 1814
Congress of Vienna a buffer state against France. Yet, the declaration of the First Italian
War of Independence on Austria following the 1848 Spring of Nations initiated an enmity that
flowed into the Second and Third Wars of Independence. France took this opportunity to
draw up an alliance with the Kingdom of Sardinia with the twofold aim of gaining back Nice
and the Savoy and erecting a wall against Austria. This strict link with the Second French
Republic was suddenly interrupted in 1881 when French forces established a protectorate in
Tunisia (Galasso, 2011). Unable to materialize its important investments there, the Kingdom
of Italy secured in 1882 the Triple Alliance with Austria and Germany by committing to
20
The former (latter) variable is in grid format, covers the entire World (Europe) at a 0.5 degrees spatial
resolution (for the period 1400-2004), and is collected from the GAEZ dataset (Luterbacher et al., 2004).
18
mutual support in the event of a French attack. Yet, the deterioration of the relationship
between England and the Triple Alliance, because of Otto von Bismark’s “realpolitik” and
the conflicting interests in Northern Africa, promoted a rapprochement of France and Italy,
which spurred the 1902 colonial agreements (Galasso, 2011). The revived French-Italian axis
paved the way to the end of the Triple Alliance and the blast of the First World War.
5.2
Endogenous Taxation
Building on our model, we estimate by OLS tax policy equations of the following type
2
2
2
LTr,t = αr + β00 Ar,t + β10 Ar,t + γ0 Sr,t + γ1 Sr,t
+ δ0 Pr,t + δ1 Pr,t
+ r,t ,
(4)
where LTr,t is Land-Taxes in region r and year t, the vector Ar,t gathers Sericulture, Wheat,
and Arboriculture, Sr,t labels Army, Pr,t indicates Distance-to-Enemies, whereas αr account
for time-independent differences across regions like unobserved geography shaping the efficiency of the different farming technologies. Considering the squared terms into equation
(4) greatly increases the fit of the model and allows us to capture relevant nonlinearities in
the data, which are plausible given the functional forms of the equilibrium tax revenues (see
section 3.1). The key implications to be tested are that β0 is negative in the pre-Unification
sample and positive otherwise, γ0 > 0 since Sr,t is an inverse measure of tax-collection costs,
and δ0 is insignificant (positive and significant) in the pre(post)-Unification sample.
In judging the adequacy of the empirical strategy, it is worth to discuss the possible
endogeneity of the regressors in equation (4). First, the controls encapsulated in Ar,t are
clearly exogenous since they are all driven by either climate shocks or features of the region’s
terrain independent of human effort. Second, Distance-to-Enemies was shaped by events
outside the control of the Italian elites as the Vienna Congress, the Spring of Nations, the
French expansion in Tunisia, and the German realpolitik. Furthermore, since δ0 is identified
by the time variation in Distance-to-Enemies, it cannot simply capture the distance from
international markets. Similarly, as stressed in section 2, Army was mainly driven by the
boundary lines designed by the Vienna Congress and the resulting internal and external
dangers to the survival of each state (Galasso, 2011). Finally, the correlation between Army
and Distance-to-Enemies is limited—i.e., - 27 (1) percent in the pre(post)-Unification sample.
19
This excludes that the results could simply be driven by multicollinearity among the controls.
Columns (1) to (3) of table 3 display the estimates for three alternative specifications, the
first excluding both Army and Distance-to-Enemies, the second excluding only Distance-toEnemies, and the third one including all controls. Columns (4) to (6) of table 3 have
the same structure but are based on the post-Unification sample. To control for arbitrary
correlation within groups we allow for clustering by region. For the most part, the results
are consistent with the model predictions, and the implied effects are large. In particular,
the estimates reveal a negative and significant relationship between regional tax revenues
and proxies for efficiency of sericulture and arboriculture in the pre-Unification sample,
whereas the two variables lose significance in the post-Unification sample. Furthermore,
the more efficient is the outside option represented by the subsistence wheat breeding, the
weaker was the Piedmontese incentive to curb extraction after 1861. Distance-to-Enemies
instead is irrelevant to explain pre-Unification regional taxation but represents the strongest
predictor of post-Unification tax policies. In particular, a one-standard-deviation rise in the
distance from the relevant Kingdom of Sardinia’s enemies—i.e., 256 km—is associated with
a 0.7-standard-deviation increase in Land-Taxes, 3.57 lire per capita. Finally, Army is never
significant. This lack of statistical significance could be simply driven by the preferences of
the elite to resort to indirect taxes (Dincecco et al., 2011). All in all, it is fair to summarize
our results stating that pre-Unification policies were mainly driven by efficiency concerns,
whereas post-Unification taxation was the product of a rational extraction process by the
Kingdom of Sardinia’s elite whose impact on the other regions we measure as follows.
5.3
The Rise of the North-South Divide
Ignoring nonlinearities,21 we estimate performance equations of the type
Yr,t = αr + γ2 Sr,t + δ2 Pr,t + ζDr,t + νr,t ,
(5)
where Yr,t is the ratio of one among seven development indicators to its 1861 value, i.e.,
GDP, Turnout, Illiterates, VA F, VA T, VA C, VA M (see table 1).22 These indicators
21
22
2
2
2
The message of our analysis will be the same should we add to the specification Sr,t
, Pr,t
, and LT Dr,t
.
The message of the empirical exercise is similar when we look at the gross tradable annual farming production
20
are respectively the income in 1861 lire per capita, the electoral turnout, the share of the
population aged over six that was illiterate, and the value added in the foodstuff sector, in
the textiles sector, in the clothing sector, and in the manufacturing sector all expressed in
millions of 1861 lire. Turnout (Illiterates) is positively (negatively) correlated to a stronger
culture of cooperation (human capital) (Guiso and Pinotti, 2013).23 When needed, we
imputed a missing observation with the following data point: this approximation strategy
does not affect the gist of our empirical results. Indeed, we document that our results will
be very similar should we focus on the subsample 1871-1911 for which we observe almost
all the variables (see the Internet appendix).24 αr control for time-independent differences
across regions like the inclusiveness of the political institutions prevailing before Unification.
Dr,t is the difference between the observed Land-Taxes and the land property tax revenues
per capita forecasted using the specification illustrated in column (1) of table 3, LTD.25 LTD
assumes larger values when tax distortions are wider and incorporates unobserved shifters
of the tax-collection cost and of the regional political relevance. A caveat to our approach
can be that the Unification dramatically changed the structures of the regional economies
making our positive taxation model inadequate for the 1861-1911 sample. Yet, over this
period, the percentage of the active population employed in the industrial sectors in the
North (Center-South) raised (felt) by only two (one) percent—i.e., from 12 to 14 (13 to 12),
and the gist of our results is the same should we consider time dummies (see the Internet
appendix). These control for price shocks, epidemics, and changes in policies and technology
affecting the whole country. We expect that the estimates of γ2 , δ2 , and ζ are all significant
and negative (positive when the dependent variable is Illiterates).
A glance to figures 1, 2, and 3 confirms the model predictions. Figure 1 shows the opposite
evolutions of the level of income per capita and land property taxes in post-Unification Italy:
the regions less politically relevant for the Kingdom of Sardinia’s elite experienced the most
hindering tax policies and the weakest growth in well-being, whereas the opposite is true
per employee, the life expectancy in years, the average height of conscripted workers, the above-described
value added per capita, and the value added for the mining and wood sectors (see the Internet appendix).
23
Felice and Vasta (2012) provide a different but less reliable proxy for culture (Daniele and Malanima, 2014).
24
Of course, in this case, we define the proxies for Yr,t as the ratio of each indicator to its 1871 value.
25
We don’t use either the specification reported in column (2) or that illustrated in column (3) of table 3 since
post-Unification values of Sr,t and Pr,t should be irrelevant in the counterfactual autarky regime.
21
for the regions nearest to the French and Austrian borders. This pattern is even more clear
when the post-Unification distortions in land property taxes are compared with the change
in income per capita, turnout, and illiterates levels over the same period (see figure 2). Of
course, this evidence is inconsistent with a persistence of cultural-divide explanation of the
present-day gap: as already discussed, figure 3 further confirms this conclusion. Although
graphical comparisons are instructive, multivariate analysis is more convincing.
Panels A to D of table 4 list the estimates of equation 5 for three specifications: a first
one including only Army and Distance-to-Enemies, a second one incorporating only LTD,
and a last one considering all controls. The key observation is that the state capacity of the
unified state and the region-specific political relevance are strong predictors of the rise of
the North-South gap in culture and outcomes. A one-standard-deviation rise in Army—i.e.,
2.34—is associated with an almost 11 percent fall in income per capita, a 2.6 percent decrease
in the turnout, and a slightly higher than 6 percent rise in the share of population aged over
six that were illiterates with respect to their 1861 levels (see columns (1) to (3) of panel C).
Similarly, a one-standard-deviation rise in Distance-to-Enemies—i.e., 256 km—is associated
with a 7.7 percent fall in GDP, whereas a one-standard-deviation increase in LTD—i.e.,
5.2—implies a 5.8 percent fall in GDP and a 0.5 percent rise in Illiterates (see panel C).
5.4
Identifying Causal Relationships
The negative link between tax distortions and welfare discussed in the previous section is
consistent with the testable predictions produced by our model. Yet, this correlation could
also be explained by omitted variables. We pursue a two-step strategy to assess whether the
correlations documented so far are indeed causal. First we control for relevant observable
factors, and then we use selection on observables to assess the bias from unobservables.
5.4.1
Controlling for Observables
We consider four observables possibly confounding the effect of tax distortions on economic outcomes. First, we incorporate the ratio of estimated regional wage based on non
tradable goods to the Italian average, Wages (Felice and Vasta, 2012). This should exclude
that either the post-Unification selection of the conversion rates between regional currencies and the lire or the wage rigidities in the South drive our results. Second, to contrast
22
our model with the idea that fragmented land ownership and income created an extractive
bourgeoisie class in the South (Felice, 2014), we include into the analysis the Gini coefficient,
Gini (Vecchi, 2011). Along the same line, we experimented with using Illiterates as a control
variable rather than as a dependent variable, and we obtained qualitatively similar estimates.
Third, to take care of the technological drivers of the Industrial Revolution in Italy, we look
at the horsepower units of hydroelectric and water power production in millions of 1861
lire, HW-Power (Missiaia, 2012). Water represented the single and most important energy
endowment in the post-Unification period and became increasingly relevant for the rising
textile, clothing, and manufacturing sectors (Missiaia, 2012). Fourth, we tackle the possibility that demographic differences captured by the population per square kilometer—i.e.,
Population-D—determine outcomes by shaping the incidence of land property taxes.
The estimates listed in panels A to D of table 5 reveal that our results remain almost
intact once we consider alternative explanations of the North-South divide. Consistent with
the historical analysis of section 2, while exchange rate policies and the fragmentation of
economic power do not make an economically relevant dent in explaining the post-Unification
divergence in the welfare of the two blocks, both the initial energy endowment and the
population density help shed new light on the 1861-1911 events.
5.4.2
Using Selection on Observables to Assess the Bias from Unobservables
Despite our attempts to control for relevant observables, our estimates may still be biased by unobservable factors correlated with technological efficiency, state capacity, and the
political relevance of each region for the Kingdom of Sardinia’s elite. Next, we use selection
on observables to assess the potential bias from unobservables. We build on Altonji et al.
(2005) who provide an index gaging how much stronger selection on unobservables, relative
to selection on observables, must be to explain away the full estimated effect.26
To see how the index is calculated in table 6, consider two regressions: one with a restricted set of control variables and one with a full set of controls. Denote the estimated
coefficient for the variable of interest from the first regression γ R , where R stands for “restricted,” and the estimated coefficient from the second regression γ F , where F stands for
“full.” Then, the index equals the absolute value of the ratio γ F /(γ R − γ F ). The intuition
26
We use the version developed by Bellows and Miguel (2009) for possibly endogenous continuous variables.
23
behind the formula is as follows. First, consider why the ratio is decreasing in (γ R − γ F ).
The smaller is the absolute value of the difference between γ R and γ F , the less the estimate
is affected by selection on observables, and the stronger selection on unobservables needs to
be to explain away the entire effect. Next, note that the larger the absolute value of γ F is,
the greater is the effect that needs to be explained away by selection on unobservables, and
thus the higher is the ratio. We focus on the variables testing the key model predictions.
In particular, we look at the proxies for technological efficiency in the endogenous taxation
model run on the pre-Unification sample (see column (1)), the measures of state capacity
and the political relevance in the endogenous taxation model run on the post-Unification
sample (see column (2)), and at Army, Distance-to-Enemies, and LTD in the performance
equations (see columns (3) to (9)). These are also the regressors whose coefficients are more
statistically and economically significant in tables 3 to 5. The covariates added to the full set
are those listed in the last three rows of table 6. The median and the average of the indexes
in columns (1) and (2) (columns (3) to (9)) are respectively 4.36 and 11.35 (3.75 and 4.49).
Thus, to attribute the entire OLS estimates to selection effects, selection on unobservables
would have to be on average 7 times greater than selection on observables. In our view,
these results make it quite unlikely that our estimates are simply driven by unobservables.27
6
Conclusions
One of the most puzzling issues in economics is the huge divide between the North and
South of Italy despite more than 150 years of common formal institutions. We document
that the opening of the present-day gap in culture and economic outcomes is the consequence
of the region-specific tax policies selected between 1861 and 1911 by the elite of the Kingdom
of Sardinia, which annexed the rest of Italy in 1861. Regional pre-Unification tax revenues
indeed are mainly explained by the contemporaneous farming productivity but not by the
region’s political relevance for the Kingdom of Sardinia’s elite, whereas the opposite is true
for the post-Unification period. Moreover, tax distortions, tax-collection costs, and each
region’s political relevance significantly shaped the increasing post-Unification differences in
27
We will reach a qualitatively similar conclusion should we consider as extra control in the full set of covariates
one among Gini, HW-Power, and Population-D instead of the variable Wages.
24
income per capita, electoral turnout, and human capital between the North and the South.
We measure tax distortions with the difference between the tax revenues observed in the
region and those that would have prevailed without Unification according to the two-region,
two-social class model that we construct to inform our empirical exercise. To determine
whether the correlations we document are, in fact, causal, we follow a two-step strategy.
First, we control not only for fixed region effects, but also for other relevant observable
factors, like wages, income inequality, energy endowment, and population density. Including
these controls has little effect on our results. Second, we use insights from Altonji et al. (2005)
to evaluate how much greater the influence of unobservables would need to be, relative to
observables, to completely explain away the relationships we uncover in the data. We find
that this influence would have to be on average 7 times greater than that of observable factors.
Hence, it is quite unlikely that our estimates can be attributed to unobserved heterogeneity.
Of course, our results characterize the crack of the dawn of the North-South divide and
other more recent factors may have aggravated the disparities arose before the First World
War (Iuzzolino et al., 2011). Accordingly, evaluating more recent mechanisms that have
affected the diverging performances of the two blocks is an important avenue for further
research. Similarly, it is worth the effort applying our framework to understand the deep
conflicts that are shaking the European Union nowadays (see also Guiso et al., [2013]).
25
References
Abadie, Alberto, and Javier Gardeazabal. 2003. “The Economic Costs of Conflict: A Case
Study of the Basque Country.” American Economic Review, 93: 113-132.
Acemoglu, Daron, and James Robinson. 2012. Why Nations Fail: The Origins of Power,
Prosperity, and Poverty. New York, NY: Crown Publishers.
Altonji, Joseph G., Todd E. Elder, and Christopher R. Taber. 2005. “Selection on Observed
and Unobserved Variables: Assessing the Effectiveness of Catholic Schools.” Journal of
Political Economy, 113: 151-184.
Aprile, Pino. 2010. Terroni. Tutto Quello che è Stato Fatto Perché gli Italiani del Sud
Diventassero Meridionali. Milano, Italy: Piemme.
Barkow, Jerome H., Leda Cosmides, and John Tooby. 1992. The Adapted Mind: Evolutionary Psychology and the Generation of Culture. New York, NY: Oxford University
Press
Bellows, John, and Edward Miguel. 2009. “War and Local Collective Action in Sierra
Leone.” Journal of Public Economics, 93: 1144-1157.
Besley, Timothy, and Torsten Persson. 2009. “The Origins of State Capacity: Property
Rights, Taxation, and Politics.” American Economic Review, 99: 1218-1244.
Besley, Timothy, and Hannes Mueller. 2012. “Estimating the Peace Dividend: The Impact
of Violence on House Prices in Northern Ireland.” American Economic Review, 102:
810-833.
Boranbay, Serra, and Carmine Guerriero. 2013. “Endogenous (In)Formal Institutions.”
Unpublished.
Britannica (Encyclopedia). 2014. Encyclopedia Britannica. http://www.britannica.com
26
Ciccarelli, Carlo, and Stefano Fenoaltea. 2013. “Through the Magnifying Glass: Provincial
Aspects of Industrial Growth in post-Unification Italy.” The Economic History Review,
66: 57-85.
Clark, Gregory, and Gillian Hamilton. 2006. “Survival of the Richest: The Malthusian
Mechanism in Pre-Industrial England.” Journal of Economic History, 66: 707-736.
Corbetta, Piergiorgio, and M. Serena Piretti. 2008. Atlante Storico-elettorale d’Italia (18612006). Zanichelli: Bologna, Italy.
Daniele, Vittorio, and Paolo Malanima. 2014. “Perché il Sud è Rimasto Indietro: Il Mezzogiorno fra Storia e Pubblicistica.” Rivista di Storia Economica, 1: 3-36.
Del Monte, Alfredo, and Luca Pennacchio. 2012. “Agricultural Productivity, Banditry, and
Criminal Organisations in Post-Unification Italy.” Unpublished.
Dimico, Arcangelo, Alessia Isopi, and Ola Olsson. 2012. “Origins of the Sicilian Mafia: The
Market for Lemons.” Unpublished.
Dincecco, Mark, Giovanni Federico, and Andrea Vindigni. 2011. “Warfare, Taxation, and
Political Change: Evidence from the Italian Risorgimento.” Journal of Economic History,
71: 887-914.
Federico, Giovanni. 2007. “Ma l’Agricoltura Meridionale Era Davvero Arretrata?” Rivista
di Politica Economica, 97: 317-340.
Felice, Emanuele. 2014. Perché il Sud è Rimasto Indietro. Bologna: Il Mulino.
Felice, Emanuele, and Ferdinando Giugliano. 2011. “Myth and Reality: A Response to
Lynn on the Determinants of Italy’s North-South Imbalances.” Intelligence, 39: 1-6.
Felice, Emanuele, and Michelangelo Vasta. 2012. “Passive Modernization? The New Human
Development Index and Its Components in Italys Regions (1871-2007).” Unpublished.
Galasso, Giuseppe. 2011. Storia d’Italia. Torino: UTET.
27
Guerriero, Carmine. 2014. “Endogenous Institutions and Economic Outcomes.” Unpublished.
Guiso, Luigi, and Paolo Pinotti. 2013. “Democratization and Civic Capital.” In The Oxford
Handbook of Italian Economy since Unification, ed., Giovanni Toniolo. New York, NY:
Oxford University Press.
Guiso, Luigi, Helios Herrera, and Massimo Morelli. 2013. “A Cultural Clash View of the
EU Crisis.” Unpublished.
Guiso, Luigi, Paola Sapienza, and Luigi Zingales. 2014. “Long-Term Persistence.” Unpublished.
Hearder, Harry. 1983. Italy in the Age of the Risorgimento, 1790-1870. London: Longman.
Iuzzolino, Giovanni, Guido Pellegrini, and Gianfranco Viesti. 2011. “Convergence Among
Italian Regions, 1861-2011.” Bank of Italy Economic History Working Papers, 22.
Lynn, Richard. 2010. “In Italy, North-South Differences in IQ Predict Differences in Income,
Education, Infant Mortality, Stature, and Literacy.” Intelligence, 38: 93-100.
Luterbacher, Jürg, Daniel Dietrich, Elena Xoplaki, Martin Grosjean, and Heinz Wanner.
2004. “European Seasonal and Annual Temperature Variability, Trends, and Extremes
Since 1500.” Science, 303: 1499-1503.
Malanima, Paolo. 2006. “An Age of Decline. Product and Income in Eighteenth-Nineteenth
Century Italy.” Rivista di Storia Economica, 21: 91-133.
Malanima, Paolo. 2011. “The Long Decline of a Leading Economy: GDP in Central and
Northern Italy, 1300-1913.” European Review of Economic History, 15: 169-219.
Marongiu, Gianni. 1995. Storia del Fisco in Italia, volume I. Torino: Einaudi.
Ministero delle Finanze del Regno d’Italia. 1872, 1882, 1888. Annuario del Ministero delle
Finanze del Regno d’Italia. Torino, Italy: Stamperia Reale.
28
Ministero delle Finanze del Regno d’Italia. 1903, 1911. Bollettino di Statistica e di Legislazione Comparata. Roma: Tipografia Elzeviriana.
Missiaia, Anna. 2012. “Market vs. Endowment: Explaining Early Industrial Location in
Italy (1871-1911).” Unpublished.
Nitti, Francesco Saverio. 1905. La Ricchezza dell’Italia. Torino, Italy: Roux and Viarengo.
North, Douglass C., John J. Wallis, Barry R. Weingast. 2009. Violence and Social Orders: A
Conceptual Framework for Interpreting Recorded Human History. Cambridge University
Press: Cambridge, UK.
Pauling, Andreas, Jürg Luterbacher, Carlo Casty, and Heinz Wanner. 2006. “Five Hundred
Years of Gridded High Resolution Precipitation Reconstructions Over Europe and the
Connection to Large-scale Circulation.” Climate Dynamics, 26: 387-405.
Pescosolido, Guido. 2011. “La costruzione dell’Economia Unitaria.” In Enciclopedia Italiana, ed., Roma: Treccani.
Petrusewicz, Marta. 1996. Latifundium: Moral Economy and Material Life in a European
Periphery. Ann Arbor: University of Michigan Press.
Putnam, Robert, Robert Leonardi, and Raffaella Nanetti. 1993. Making Democracy Work.
New York: Simon & Schuster.
Ragioneria Generale dello Stato. 1909. Il Bilancio del Regno d’Italia. Roma: Tipografia
Nazionale di Bertero.
Romani, Mario. 1982. Storia economica d’Italia nel secolo XIX, 1815-82. Bologna, Italy: Il
Mulino.
SVIMEZ. 2011. 150 Anni di Statistiche Italiane: Nord e Sud: 1861-2011. Bologna, Italy: Il
Mulino.
Tabellini, Guido. 2010. “Culture and Institutions: Economic Development in the Regions
of Europe.” Journal of European Economic Association, 8: 677-716.
29
Vecchi, Giovanni. 2011. In Ricchezza e in Povertà. Il Benessere Degli Italiani dall’Unità a
Oggi. Bologna, Italy: Il Mulino.
Tables and Figures
Table 1: Summary of Variables
Definition and Sources
GDP:
Ratio of income in 1861 lire per capita to its 1861 value. Sources:
Malanima (2006, 2011); Felice and Vasta (2012).
Ratio of the number of voters relative to the number of electors
to its 1861 value. Source: Corbetta and Piretti (2008).
Ratio of percentage of illiterates in the total population over the
age of six to its 1861 value. Source: SVIMEZ (2011).
Ratio of value added in the foodstuff sector in millions of 1861 lire
to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Ratio of value added in the textiles sector in millions of 1861 lire
to its 1861 value. Source: Ciccarelli and Fenoaltea (2013).
Ratio of value added in the clothing sector in millions of 1861 lire
to its 1861 value. Source: Ciccarelli and Fenoaltea (2013).
Ratio of value added in the manufacturing sector in millions of 1861
lire to its 1861 value. Source: Ciccarelli and Fenoaltea (2013).
Taxes on land property in 1861 lire per capita. Sources: Dincecco et al.
(2011); Ministero delle Finanze (1872, 1882, 1888, 1903, 1911).
Difference between the observed Land-Taxes and those forecasted
using our positive taxation model.
First principal component extracted from the share of the region’s
area covered by major lakes and rivers and the average growing
season precipitation in ml in the previous decade and normalized
to range between 0 and 1. Sources: GLWD dataset
(http://worldwildlife.org/); Pauling et al. (2006).
First principal component extracted from the land suitability for
wheat ranging between 0 and 100 and the average growing
season precipitation in ml in the previous decade and normalized
to range between 0 and 1. Sources: GAEZ dataset
(http://www.gaez.iiasa.ac.at); Pauling et al. (2006).
First principal component extracted from the land suitabilities for
citrus and olive trees, which range between 0 and 100, and the average
growing season temperature in Celsius in the previous decade. The
index is normalized to range between 0 and 1. Sources: GAEZ
dataset (http://www.gaez.iiasa.ac.at); Luterbacher et al. (2004).
Military personnel per 1000 citizens. Source: Correlates of War
Project (http://www.correlatesofwar.org/).
Distance in km from Vienna between 1801 and 1812, from Paris
between 1813 and 1847, from Vienna between 1848 and 1881,
from Paris between 1882 and 1900, and from Vienna between
1901 and 1914. Source: Galasso (2011).
Ratio of estimated regional wage based on non tradable goods
to the Italian average. Source: Felice and Vasta (2012).
Gini coefficient. Source: Vecchi (2011).
Turnout:
Illiterates:
Economic
outcomes:
VA F :
VA T :
VA C :
VA M :
Land-Taxes:
Taxation:
LTD:
Sericulture:
Inputs:
Wheat:
Arboriculture:
Army:
Other
drivers of
taxation:
Distanceto-Enemies:
Wages:
Other
Controls:
Notes:
(1)
18011851
Variable
2.147
(1.323)
(2)
18611911
1.131
(0.342)
1.063
(0.142)
0.783
(0.191)
1.243
(0.307)
1.518
(0.880)
1.601
(0.715)
1.591
(0.651)
5.418
(5.139)
3.164
(5.207)
(3)
Difference
- 3.271
(0.601)***
0.306
(0.239)
0.323
(0.243)
- 0.018
(0.039)
0.281
(0.199)
0.298
(0.201)
- 0.017
(0.032)
0.547
(0.192)
0.532
(0.192)
0.014
(0.031)
5.134
(2.310)
8.828
(2.343)
- 3.694
(0.373)***
934.539
(311.239)
803.462
(255.570)
131.077
(45.600)***
0.965
(0.107)
44.268
(3.160)
HW-Power :
Horsepower unit of hydroelectric and water power production in
2.984
millions of 1861 lire. Source: Missiaia (2012).
(3.972)
Population-D:
Population per square kilometer. Source: SVIMEZ (2011).
153.091
(146.276)
1.
Columns (1) and (2) report the mean value and, in parentheses, the standard deviation of each variable, which are computed
respectively for the sample used to produce table 2 (3). Column (3) lists the difference between the means of each variable over
the samples used to produce respectively columns (1) and (2) and, in parentheses, its standard error.
2.
*, **, and *** denote a significant difference at respectively the 10 %, 5 %, and 1 % based on a t-test with unequal variances.
Gini:
30
Table 2: The North-South Divide
(1)
(2)
(3)
1801-1851
GDP-L:
North
341.690
(27.779)***
South
374.500
(52.116)
2.087
(0.501)***
0.431
(0.292)***
0.495
(0.205)***
0.355
(0.101)***
6.314
(3.433)*
684
(150.856)***
3.260
(0.203)
0.160
(0.135)
0.133
(0.064)
0.681
(0.167)
4.994
(1.179)
1142.417
(291.607)
Turnout-L:
Illiterates-L:
VA F-L:
VA T-L:
VA C-L:
VA M-L:
Land-Taxes:
Sericulture:
Wheat:
Arboriculture:
Army:
Distance-to-Enemy:
Wages:
Gini:
HW-Power :
Population-D:
(4)
1861
North
315.84
(0)
0.500
(0.023)***
67.500
(10.686)**
43.755
(21.830)
14.648
(14.139)
10.422
(3.716)**
126.623
(58.219)*
5.043
(1.278)*
0.445
(0.328)
0.508
(0.229)**
0.338
(0.116)***
8.337
(0)
565.250
(91.263)***
0.983
(0.055)
47.7
(0)
2.181
(0.844)*
175.999
(174.249)
(5)
(6)
1861-1911
South
320
(0)
0.644
(0.062)
85.783
(3.131)
22.439
(15.156)
5.544
(6.107)
3.873
(2.442)
60.498
(43.550)
13.908
(9.053)
0.168
(0.152)
0.140
(0.074)
0.658
(0.178)
8.337
(0)
893.833
(165.028)
0.977
(0.070)
43.8
(0)
1.102
(0.950)
125.788
(85.784)
North
375.321
(100.771)**
0.542
(0.062)***
45.129
(18.138)***
57.506
(31.726)***
32.744
(44.670)***
16.506
(8.715)***
224.636
(171.362)***
3.949
(1.548)***
0.448
(0.290)***
0.513
(0.200)***
0.342
(0.100)***
8.828
(2.378)
604.833
(123.294)***
1.008
(0.096)***
45.533
(1.530)***
3.825
(4.524)**
222.219
(205.120)
South
322.069
(76.961)
0.651
(0.062)
74.628
(9.287)
28.979
(21.507)
6.786
(7.191)
6.133
(4.498)
93.964
(77.001)
7.299
(7.015)
0.172
(0.140)
0.144
(0.069)
0.667
(0.167)
8.828
(2.361)
976.694
(240.897)
0.923
(0.082)
43.183
(3.773)
1.681
(1.944)
154.292
(104.478)
144.906
102.425
(134.891)
(64.316)
Number of
Observations
24
36
4
6
24
36
Notes:
1.
GDP-L, Turnout-L, Illiterates-L, VA F-L, VA T-L, VA C-L, and VA M-L label respectively the income in 1861 lire per capita, the
ratio of the number of voters relative to the number of electors, the percentage of illiterates in the total population, the value
added in the foodstuff, textiles, clothing, and manufacturing sector all in millions of 1861 lire (see table 1 for the sources of each
variable). The income per capita for the pre-Unification period is calculated building on the data from Malanima (2006, 2011).
2.
Each column reports the mean value and, in parentheses, the standard deviation of each variable.
3.
*, **, and *** denote a significant difference between the means in the North and South samples at respectively the 10 %, 5 %,
and 1 % based on a two-sample t-test with unequal variances. The North (South) sample includes Emilia Romagna, Lombardy,
Tuscany (Abruzzi, Apulia, Basilicata, Calabria, Campania), and Veneto (Sicily).
Figure 1: Income and Land Taxes Per Capita
Note:
1.
See tables 1 and 2 for the definition of each variable. The N ( S) group comprehends Emilia Romagna, Lombardy, Tuscany
(Abruzzi, Apulia, Basilicata, Calabria, Campania), and Veneto (Sicily). The C block includes Latium, Marches, and Umbria.
31
Figure 2: Political Power and Railway Spending Per Capita
Note:
1.
“Political-Power” is the share of the executive born in the region collected from Corbetta and Piretti (2008), “Railway Spending” is
the regional spending on railways in 1861 lire per capita collected from Ragioneria Generale dello Stato (1909). The North ( CS)
group comprehends Emilia Romagna, Liguria, Lombardy, Piedmont, Sardinia, Tuscany (Abruzzi, Apulia, Basilicata, Calabria,
Campania, Latium, Marches, and Umbria), and Veneto (Sicily).
Figure 3: Culture
Note:
1.
“Monasteries” is the cumulated discounted number of years of activity between 1000 and 1600 of Cistercian and Franciscan houses
(see for sources and construction Boranbay and Guerriero, [2013]). The series ending in N ( S) average the values for Emilia
Romagna, Lombardy, Tuscany, and Veneto (Abruzzi, Apulia, Basilicata, Calabria, Campania, and Sicily).
32
Table 3: Endogenous Taxation
(1)
(2)
(3)
(4)
(5)
(6)
Pre-Unification sample
Post-Unification sample
The dependent variable is Land-Taxes
- 8.397
- 8.088
- 8.900
30.810
0.269
- 4.946
Sericulture
(4.394)*
(4.110)*
(4.527)*
(210.542)
(192.291)
(194.059)
- 2.331
- 2.693
- 2.613
- 6.214
- 17.949
- 11.107
Sericulture 2
(1.697)
(1.688)
(1.729)
(48.056)
(43.572)
(47.114)
5.798
5.895
5.562
- 139.774
- 70.517
- 88.055
Wheat
(3.676)
(3.215)*
(2.802)*
(179.600)
(163.893)
(164.878)
7.859
6.901
8.042
78.698
121.700
130.358
Wheat 2
(2.773)**
(3.070)**
(3.466)**
(38.824)*
(45.708)**
(49.736)**
- 10.305
- 8.585
- 8.026
19.120
- 94.710
- 13.539
Arboriculture
(3.122)***
(2.381)***
(2.901)**
(49.237)
(69.334)
(74.976)
15.515
14.565
12.429
70.915
- 37.620
- 31.483
Arboriculture 2
(3.093)***
(2.845)***
(3.026)***
(49.077)
(65.677)
(75.819)
- 0.040
- 0.040
- 2.787
- 0.779
Army
(0.052)
(0.044)
(1.694)
(1.637)
0.005
0.004
0.190
0.078
Army 2
(0.004)
(0.004)
(0.107)
(0.094)
0.0003
0.014
Distance-to-Enemies
(0.0008)
(0.006)**
−7
- 2.15E
- 8.37E−6
Distance-to-Enemies 2
−7
(3.25E
)
(3.35E−6 )**
Estimation
Fixed Region Effects OLS
R2
0.48
0.49
0.52
0.11
0.20
0.22
Number of observations
78
78
78
78
78
78
Notes:
1.
Robust standard errors allowing for clustering by region in parentheses.
2.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
3.
All specifications include a constant term.
Figure 4: The Rise of the North-South Divide
Note:
1.
“GDP-N,” “Turnout-N,” and “Illiterates-N” are respectively GDP-L, Turnout-L, and Illiterates-L normalized so that their mean
for Italy in 1861 is 1 (see table 2). The N ( S) group comprehends Emilia Romagna, Lombardy, Tuscany (Abruzzi, Apulia,
Basilicata, Calabria, Campania), and Veneto (Sicily). The C block includes Latium, Marches, and Umbria.
33
Table 4: The Rise of the North-South Divide
Army
Distance-to-Enemies
Estimation
R2
Number of
observations
LTD
Estimation
R2
Number of observations
(1)
(2)
GDP
- 0.055
(0.005)***
- 0.0002
(0.0001)***
Turnout
- 0.012
(0.004)**
0.0000
(0.0000)
0.24
78
(3)
(6)
(7)
VA C
- 0.070
(0.012)***
- 0.0001
(0.0001)
VA M
- 0.058
(0.006)***
- 0.0001
(0.0000)**
0.08
(4)
(5)
Panel A. The dependent variable is:
Illiterates
VA F
VA T
0.034
- 0.022
- 0.042
(0.004)***
(0.005)***
(0.017)**
- 0.0001
- 0.0001
0.0000
(0.0000)***
(0.0000)**
(0.0001)
Fixed Region Effects OLS
0.26
0.04
0.02
0.06
0.05
78
78
78
78
(1)
(2)
(3)
(6)
(7)
GDP
- 0.024
(0.006)***
Turnout
- 0.004
(0.001)**
VA C
- 0.071
(0.012)***
VA M
- 0.072
(0.016)***
0.14
78
0.02
78
(4)
(5)
Panel B. The dependent variable is:
VA T
Illiterates
VA F
0.017
- 0.030
- 0.048
(0.004)***
(0.009)***
(0.020)**
Fixed Region Effects OLS
0.18
0.20
0.07
78
78
78
0.20
78
0.22
78
(1)
(2)
(3)
(6)
(7)
VA C
- 0.026
(0.018)
- 0.0003
(0.0001)*
- 0.067
(0.010)***
VA M
- 0.011
(0.012)
- 0.0004
(0.0001)**
- 0.071
(0.014)***
0.21
78
0.23
78
78
78
(4)
(5)
Panel C. The dependent variable is:
GDP
Turnout
Illiterates
VA F
VA T
- 0.045
- 0.011
0.027
- 0.001
- 0.012
Army
(0.007)***
(0.004)**
(0.005)***
(0.006)
(0.013)
- 0.0003
0.0000
- 0.0001
- 0.0002
- 0.0001
Distance-to-Enemies
(0.0001)***
(0.0000)
(0.0000)**
(0.0001)**
(0.0002)
- 0.016
- 0.001
0.011
- 0.031
- 0.046
LTD
(0.005)***
(0.001)
(0.003)***
(0.008)***
(0.018)**
Estimation
Fixed Region Effects OLS
R2
0.29
0.09
0.32
0.21
0.08
Number of observations
78
78
78
78
78
Notes:
1.
Robust standard errors allowing for clustering by region in parentheses.
2.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
3.
All specifications include a constant term.
34
Table 5: The Rise of the North-South Divide — Robustness
Army
Distance-to-Enemies
LTD
Wages
Estimation
R2
Number of observations
Army
Distance-to-Enemies
LTD
Gini
Estimation
R2
Number of observations
Army
Distance-to-Enemies
LTD
HW-Power
Estimation
R2
Number of observations
(1)
(2)
GDP
- 0.041
(0.006)***
- 0.0003
(0.0001)***
- 0.023
(0.006)***
1.175
(0.728)
Turnout
- 0.010
(0.005)**
0.0000
(0.0000)
- 0.004
(0.002)**
0.493
(0.208)**
(3)
0.36
78
0.17
78
(4)
(5)
Panel A. The dependent variable is:
Illiterates
VA F
VA T
0.025
0.002
- 0.005
(0.005)***
(0.007)
(0.017)
- 0.0001
- 0.0002
- 0.0001
(0.0000)**
(0.0001)**
(0.0002)
0.015
- 0.037
- 0.058
(0.004)***
(0.010)***
(0.030)*
- 0.675
0.978
2.055
(0.457)
(0.992)
(3.408)
Fixed Region Effects OLS
0.38
0.25
0.10
78
78
78
(1)
(2)
(3)
GDP
- 0.046
(0.007)***
0.00001
(0.0001)
- 0.013
(0.006)*
0.035
(0.009)***
Turnout
- 0.011
(0.005)**
- 0.0001
(0.0000)*
- 0.003
(0.001)*
- 0.015
(0.003)***
0.41
78
0.25
78
(4)
(5)
Panel B. The dependent variable is:
VA T
Illiterates
VA F
0.027
- 0.001
- 0.011
(0.005)***
(0.006)
(0.014)
- 0.0001
- 0.0001
- 0.0003
(0.0001)*
(0.0001)
(0.0003)
0.010
- 0.030
- 0.049
(0.003)***
(0.009)***
(0.020)**
- 0.001
0.016
- 0.027
(0.006)
(0.012)
(0.028)
Fixed Region Effects OLS
0.32
0.23
0.09
78
78
78
(1)
(2)
(3)
GDP
- 0.044
(0.007)***
- 0.0003
(0.0001)***
- 0.012
(0.005)**
0.034
(0.009)***
Turnout
- 0.011
(0.004)**
0.0000
(0.0000)
- 0.0002
(0.0014)
0.007
(0.004)*
0.44
78
0.13
78
(4)
(5)
Panel C. The dependent variable is:
VA T
Illiterates
VA F
0.026
0.0003
- 0.007
(0.005)***
(0.006)
(0.013)
- 0.0001
- 0.0002
- 0.0001
(0.0000)**
(0.0001)**
(0.0002)
0.008
- 0.027
- 0.026
(0.002)***
(0.006)***
(0.010)**
- 0.022
0.035
0.145
(0.005)***
(0.011)***
(0.033)***
Fixed Region Effects OLS
0.49
0.36
0.46
78
78
78
(1)
(2)
(3)
GDP
- 0.043
(0.006)***
- 0.0004
(0.0001)***
- 0.007
(0.006)
0.004
(0.002)**
Turnout
- 0.012
(0.004)**
0.0001
(0.0000)
- 0.003
(0.002)
- 0.0006
(0.0004)
(4)
(5)
Panel D. The dependent variable is:
VA T
Illiterates
VA F
0.025
0.002
- 0.005
Army
(0.004)***
(0.003)
(0.013)
- 0.0000
- 0.0003
- 0.0004
Distance-to-Enemies
(0.0000)
(0.0001)***
(0.0003)
0.003
- 0.019
- 0.009
LTD
(0.002)
(0.007)**
(0.015)
- 0.003
0.005
0.015
Population-D
(0.001)***
(0.002)**
(0.006)**
Estimation
Fixed Region Effects OLS
R2
0.43
0.12
0.58
0.46
0.39
Number of observations
78
78
78
78
78
Notes:
1.
Robust standard errors allowing for clustering by region in parentheses.
2.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
3.
All specifications include a constant term.
35
(6)
(7)
VA C
- 0.022
(0.020)
- 0.0003
(0.0002)
- 0.075
(0.014)***
1.133
(1.494)
VA M
- 0.005
(0.014)
- 0.0003
(0.0002)*
- 0.085
(0.020)***
2.282
(2.080)
0.23
78
0.28
78
(6)
(7)
VA C
- 0.028
(0.019)
0.0001
(0.0004)
- 0.062
(0.013)***
0.050
(0.039)
VA M
- 0.013
(0.011)
- 0.0001
(0.0003)
- 0.068
(0.017)***
0.034
(0.029)
0.25
78
0.25
78
(6)
(7)
VA C
- 0.023
(0.018)
- 0.0003
(0.0001)*
- 0.055
(0.007)***
0.088
(0.026)***
VA M
- 0.008
(0.011)
- 0.0003
(0.0001)**
- 0.054
(0.008)***
0.121
(0.023)***
0.38
78
0.60
78
(6)
(7)
VA C
- 0.021
(0.013)
- 0.0005
(0.0002)***
- 0.041
(0.012)***
0.010
(0.005)*
VA M
- 0.006
(0.009)
- 0.0006
(0.0002)***
- 0.044
(0.012)***
0.011
(0.004)**
0.40
78
0.47
78
Table 6: Using Selection on Observables to Assess the Bias from Unobservables
(1)
(2)
Land-Taxes
(3)
(7)
(8)
(9)
GDP
(4)
(5)
(6)
The dependent variable is:
Turnout
Illiterates VA F
VA T
VA C
VA M
10.25
10
12.5
0.67
0.71
5.5
0.83
4
1
2
3
2
4
3
1.030
1.030
1.030
1.030
1.030
1.030
1.030
The ratio is calculated for the variable:
Sericulture
Sericulture
17.694
2
9.266
Wheat
Wheat
23.568
2
43.945
Arboriculture
Arboriculture
3.522
2
4.028
Army
Army
1.030
2
1.765
Distance-to-Enemies
4
Distance-to-Enemies 2
4.702
LTD
The extra controls in the full set are
{Sericulture, Sericulture 2 , Wheat,
Wheat 2 , Arboriculture, Arboriculture 2 },
{Army, Army 2 , Distance-to-Enemies,
Distance-to-Enemies 2 },
{Wages}.
Note:
1.
2.
NO
YES
NO
NO
NO
NO
NO
NO
NO
YES
NO
NO
NO
NO
NO
NO
NO
NO
NO
NO
YES
YES
YES
YES
YES
YES
YES
The results listed in the first column are obtained using the pre-Unification sample, wherease the remainder are obtained building
on the post-Unification sample.
Each cell reports a ratio based on the coefficient for the relevant variable from two regressions, one with a restricted set of
covariates and another with a “full set” of covariates. The covariates included in the former are those for which the ratios are
calculated, the covariates included in the latter are those listed in the last three rows. The sample size of both regressions is the
same. Each reported ratio is calculated as the absolute value of γ F /(γ R − γ F ).
36
APPENDIX (FOR ONLINE PUBLICATION)
Supplementary Tables
Table I: Summary of Variables
Variable
Agriculture:
Life:
Economic
outcomes:
Height:
VAP F :
VAP T :
VAP C :
VAP M :
VA MI :
VA W :
Land-Taxes-K :
Taxation:
LTD-K :
Citrus:
Inputs:
Olive:
Land-Q:
Other
drivers of
taxation:
Note:
1.
2.
War :
Definition and Sources
Ratio of gross saleable annual farming production per employee to
its 1861 value with Italy 1861=100. Source: Esposto (1997).
Ratio of life expectancy in years to its 1861 value. Source:
Felice and Vasta (2012).
Ratio of average height of conscripted workers to its 1861
value. Source: Vecchi (2011).
Value added in the foodstuff sector in millions of 1861 lire per
capita to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Value added in the textiles sector in millions of 1861 lire per
capita to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Value added in the clothing sector in millions of 1861 lire per
capita to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Value added in the manufacturing sector in millions of 1861 lire per
capita to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Ratio of value added in the mining sector in millions of 1861 lire
to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Ratio of value added in the wood sector in millions of 1861 lire
to its 1861 value. Source: Ciccarelli and Fenoaltea (2009).
Taxes on land in 1861 lire per square km of arable land. Sources:
Ministero delle Finanze (1872, 1882, 1888, 1903, 1911).
Difference between the observed Land-Taxes-K and
those forecasted using our positive taxation model.
First principal component extracted from the land suitability for
citrus ranging between 0 and 100, and the average growing
season temperature in Celsius in the previous decade and
normalized to range between 0 and 1. Sources: GAEZ dataset
(http://www.gaez.iiasa.ac.at); Luterbacher et al. (2004).
First principal component extracted from the land suitability for
olives ranging between 0 and 100, and the average growing
season temperature in Celsius in the previous decade and
normalized to range between 0 and 1. Sources: GAEZ dataset
(http://www.gaez.iiasa.ac.at); Luterbacher et al. (2004).
Average land quality. Source: Atlas of the Biosphere available at
http://www.sage.wisc.edu/iamdata/
Share of previous decade in which the state participated in
external wars. Source: Authors’ codification.
(1)
1801-1851
299.910
(225.969)
(2)
1861-1911
1.123
(0.278)
1.167
(0.176)
1.007
(0.007)
1.016
(0.170)
1.213
(0.526)
1.277
(0.387)
1.273
(0.354)
4.191
(9.897)
1.321
(0.562)
1066.487
(1078.427)
0.0004
(0.0010)
(3)
Difference
- 766.578
(124.760)***
0.678
(0.200)
0.654
(0.200)
0.024
(0.032)
0.453
(0.221)
0.432
(0.221)
0.021
(0.035)
0.778
(0.132)
0.056
(0.097)
0.778
(0.132)
0.050
(0.077)
0.778
(0.132)
0.006
(0.014)
Columns (1) and (2) report the mean value and, in parentheses, the standard deviation of each variable.
Column (3) lists the difference between the means in two samples and, in parentheses, its standard error. *, **, and *** denotes
a significant difference at respectively the 10 %, 5 %, and 1 % based on a two-sample t-test with unequal variances.
37
Table II: The North-South Divide — Further Evidence
(1)
(2)
(3)
(4)
1801-1851
North
South
358.870
(98.502)**
410.558
(211.484)
0.463
(0.144)***
0.245
(0.110)***
0.033
(0.048)
0.814
(0.126)
0.606
(0.205)
0.067
(0.112)
North
68.300
(18.719)
32.525
(1.343)
164.019
(0.844)***
0.018
(0.005)
0.006
(0.004)
0.004
(0.001)*
0.052
(0.012)*
1.882
(1.596)
16.389
(6.413)*
1137.350
(478.011)
298.036
(281.340)*
0.430
(0.162)**
0.217
(0.122)**
4
(0)
Agriculture-P-L:
Life-L:
Height-L:
VAP F-L:
VAP T-L:
VAP C-L:
VAP M-L:
VA Mi-L:
VA W-L:
Land-Taxes-K :
LTD-K :
Citrus:
Olive:
War :
(5)
1861
(6)
1861-1911
South
87.617
(25.512)
31.767
(2.569)
160.030
(0.687)
0.014
(0.003)
0.003
(0.002)
0.002
(0.0005)
0.036
(0.008)
2.978
(5.604)
7.335
(5.487)
2304.568
(2069.236)
2150.091
(2034.097)
0.778
(0.137)
0.575
(0.222)
6
(0)
North
78.083
(18.390)***
38.396
(5.688)*
165.059
(1.094)***
0.019
(0.005)***
0.009
(0.009)***
0.006
(0.002)***
0.070
(0.031)***
4.682
(5.370)
22.242
(14.124)***
1071.017
(496.916)
234.813
(327.785)***
0.438
(0.143)**
0.223
(0.108)**
0.050
(0.078)
South
91.972
(28.414)
35.761
(4.659)
161.347
(1.047)
0.014
(0.004)
0.003
(0.002)
0.003
(0.001)
0.045
(0.015)
6.741
(13.181)
10.055
(8.416)
1325.722
(1474.626)
1176.382
(1447.718)
0.790
(0.125)
0.586
(0.205)
0.050
(0.077)
Number of
Observations
24
36
4
6
24
36
Notes:
1.
Agriculture-P-L, Life-L, Height-L, VAP F-L, VAP T-L, VAP C-L, VAP M-L, VA Mi-L, VA W-L represent respectively the gross
saleable annual production per employee, the life expectancy in years, the average height of conscripted workers, the value added
in the foodstuff, textiles, clothing, and manufacturing sector in millions of 1861 lire per capita, and the value added in the mining
and wood sector in millions of 1861 lire.
2.
Each column reports the mean value and, in parentheses, the standard deviation of each variable.
3.
*, **, and *** denote a significant difference between the means in the North and South samples at respectively the 10 %, 5 %,
and 1 % based on a two-sample t-test with unequal variances. The North (South) sample includes Emilia Romagna, Lombardy,
Tuscany (Abruzzi, Apulia, Basilicata, Calabria, Campania), and Veneto (Sicily).
Table III: Employing Land Taxes per Square Kilometer of Arable Land
(1)
Sericulture
Sericulture 2
Wheat
Wheat 2
Arboriculture
Arboriculture 2
Army
Army
2
Distanceto-Enemies
Distanceto-Enemies 2
(2)
Land-Taxes-K
- 630.092
- 28756.78
(903.736)
(52912.67)
- 705.273
7657.921
(543.404)
(11537)
450.252
4895.285
(701.056)
(41958.15)
1210.665
19778.82
(619.895)*
(11727.44)
- 669.343
746.576
(585.395)
(12479.47)
1071.386
- 9445.85
(505.432)*
(14476.75)
- 19.718
- 39.882
(13.233)
(268.360)
3.758
9.416
(0.838)***
(14.518)
- 0.008
2.250
(0.152)
(1.210)*
0.00003
- 0.001
(0.00006)
(0.0006)**
LTD-K
Estimation
R2
Number of
observations
Notes:
1.
2.
3.
4.
0.75
0.22
(3)
(7)
(8)
(9)
GDP
(5)
(6)
The dependent variable is:
Turnout
Illiterates
VA F
VA T
VA C
VA M
- 0.047
(0.007)***
- 0.011
(0.005)**
0.029
(0.005)***
- 0.017
(0.014)
- 0.037
(0.019)*
- 0.019
(0.012)
- 0.0002
(0.0001)***
0.00004
(0.00004)
- 0.0001
- 0.0002
(0.00004)*** (0.0001)**
- 4.66E−6
(0.0001)
- 0.0001
(0.0001)
- 0.0002
(0.0001)*
- 0.0001
(0.00003)**
- 5.05E−6
0.00004
(6.58E−6 )
(0.00002)**
Fixed Region Effects
0.08
0.29
- 0.0003
(0.0001)***
- 0.0003
(0.0001)***
0.14
0.17
0.27
(4)
- 0.002
(0.007)
- 0.0002
- 0.0002
(0.00004)*** (0.0001)
OLS
0.18
0.06
78
78
78
78
78
78
78
78
While the first column is based on pre-Unification data, the other columns are built on post-Unification data.
Robust standard errors allowing for clustering by region in parentheses.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
All specifications include a constant term.
38
78
Table IV: Cross-Validating the Proxies for Technological Efficiency
Sericulture
Wheat
Citrus
Olive
Arboriculture
Notes:
1.
0.66
0.17
0.42
0.69
0.90
0.99
Silk-P
Wheat-P
Citrus-P
Olive-P
Silk-P (Wheat-P) represents the production of silk (wheat) in kg (hectoliters). Citrus-Sales (Olive-Sales) is the production of
citrus (olives) in hundreds (hectoliters). All these variables are collected for the years 1871, 1881, 1891, 1901, 1911 from Direzione
Generale della Statistica e Lavoro (1886, 1890, 1900, 1911). Thus, each correlation is calculated for a sample of 65 observations.
Table V: Controlling For Time Effects and the Average Land Quality
Army
Distance-to-Enemies
LTD
Estimation
R2
Number of observations
Army
Distance-to-Enemies
LTD
(1)
(2)
GDP
- 0.141
(0.018)***
0.0001
(0.0001)
0.009
(0.007)
Turnout
- 0.008
(0.014)
0.0003
(0.0001)***
0.003
(0.003)
(3)
0.82
78
0.28
78
(4)
(5)
Panel A. The dependent variable is:
Illiterates
VA F
VA T
0.082
- 0.124
- 0.234
(0.009)***
(0.025)***
(0.106)**
0.0001
- 0.0001
0.0004
(0.0001)**
(0.0001)
(0.0003)
- 0.008
- 0.003
0.014
(0.003)**
(0.005)
(0.017)
Fixed Region and Time Effects OLS
0.88
0.75
0.38
78
78
78
(1)
(2)
(3)
GDP
- 0.159
(0.078)*
- 0.0001
(0.0002)
0.010
(0.009)
Turnout
0.042
(0.075)
0.0001
(0.0001)
0.005
(0.003)
(4)
(5)
Panel B. The dependent variable is:
Illiterates
VA F
VA T
0.147
- 0.337
- 1.163
(0.041)***
(0.096)***
(0.586)*
0.0001
- 0.0002
0.0001
(0.0001)*
(0.0002)
(0.0005)
- 0.006
- 0.007
- 0.002
(0.003)*
(0.006)
(0.015)
p-value for Land-Q
× 1861-1901 dummies
[0.02]
[0.06]
[0.47]
[0.27]
[0.33]
Estimation
Fixed Region and Time Effects OLS
2
R
0.82
0.37
0.91
0.81
0.54
Number of observations
78
78
78
78
78
Notes:
1.
Robust standard errors allowing for clustering by region in parentheses.
2.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
3.
All specifications include a constant term.
(6)
(7)
VA C
- 0.309
(0.062)***
0.0001
(0.0002)
0.001
(0.014)
VA M
- 0.309
(0.042)***
0.0003
(0.0002)
- 0.002
(0.007)
0.75
78
0.88
78
(6)
(7)
VA C
- 0.114
(0.218)
0.0003
(0.0004)
0.003
(0.019)
VA M
- 0.575
(0.226)**
0.0002
(0.0002)
- 0.005
(0.009)
[0.58]
[0.05]
0.76
78
0.89
78
Table VI: Alternative Proxies for Technological Efficiency
(1)
(2)
(3)
(4)
Pre-Unification sample
Post-Unification sample
The dependent variable is Land-Taxes
-10.546
- 10.400
- 5.128
2.689
Sericulture
(4.740)**
(3.785)**
(191.645)
(188.084)
- 1.819
- 1.479
- 3.749
- 46.480
Sericulture 2
(1.863)
(2.013)
(46.444)
(53.201)
6.409
6.333
- 74.879
- 64.323
Wheat
(2.901)**
(3.238)*
(163.881)
(158.547)
8.012
7.608
117.212
112.410
Wheat 2
(3.632)**
(3.903)*
(45.239)**
(58.752)*
- 5.080
- 46.361
Citrus
(2.092)**
(89.700)
5.985
- 41.458
Citrus 2
(1.748)***
(45.447)
- 3.810
- 229.442
Olive
(1.913)*
(176.828)
7.420
76.069
Olive 2
(2.452)**
(53.424)
- 0.030
- 0.038
- 3.232
- 5.285
Army
(0.053)
(0.055)
(3.881)
(4.378)
0.003
0.004
0.219
0.326
Army 2
(0.004)
(0.004)
(0.234)
(0.256)
0.00003
2.36E−6
0.011
0.013
Distance-to-Enemies
(0.0008)
(0.0007)
(0.006)*
(0.005)**
- 1.09E−7
- 1.00E−7
- 6.15E−6
- 6.90E−6
Distance-to-Enemies 2
−7
−7
−6
(3.06E
)
(2.92E
)
(2.70E
)**
(1.99E−6 )***
Estimation
Fixed Region Effects OLS
R2
0.50
0.51
0.23
0.26
Number of observations
78
78
78
78
Notes:
1.
Robust standard errors allowing for clustering by region in parentheses.
2.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
3.
All specifications include a constant term.
39
Table VII: An Alternative Proxy for the Marginal Tax-collection Cost
(1)
Sericulture
Sericulture 2
Wheat
Wheat 2
Arboriculture
Arboriculture 2
War
War 2
Distanceto-Enemies
Distanceto-Enemies 2
(2)
Land-Taxes
- 9.305
4.649
(5.062)*
(194.254)
- 2.315
- 9.994
(1.634)
(47.669)
5.571
- 95.639
(2.971)*
(165.490)
9.577
125.435
(3.518)**
(47.174)**
- 6.975
- 24.270
(5.043)
(74.433)
9.814
- 33.941
(4.427)**
(76.912)
- 1.653
- 34.406
(1.176)
(22.577)
5.681
302.718
(3.860)
(169.069)*
0.0001
0.015
(0.001)
(0.007)**
−7
- 1.86E
- 8.38E−6
(4.12E−7 )
(3.41E−6 )**
LTD
Estimation
R2
Number of
observations
Notes:
1.
2.
3.
4.
0.53
0.21
(4)
(5)
(6)
The dependent variable is:
Turnout
Illiterates
VA F
(7)
(8)
(9)
GDP
(3)
VA T
VA C
VA M
- 2.334
(0.155)***
- 0.043
(0.172)
0.626
(0.101)***
- 1.321
(0.252)***
- 2.511
(1.145)**
- 3.145
(0.689)***
- 3.101
(0.445)***
- 0.001
(0.0001)***
0.0000
(0.0000)
- 0.00002
(0.00003)
- 0.0004
(0.0001)***
- 0.0004
(0.0003)
- 0.0007
(0.0002)***
- 0.0007
(0.0002)***
- 0.015
(0.004)***
- 0.003
0.013
(0.001)**
(0.004)***
Fixed Region Effects
0.03
0.27
- 0.025
(0.007)***
OLS
0.32
- 0.036
(0.015)**
- 0.057
(0.008)***
- 0.058
(0.012)***
0.13
0.32
0.53
78
78
78
78
78
78
78
78
While the first column is based on pre-Unification data, the other columns are built on post-Unification data.
Robust standard errors allowing for clustering by region in parentheses.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
All specifications include a constant term.
0.36
78
Table VIII: Other Measures of Economic Development
Army
Distanceto-Enemies
LTD
Estimation
R2
Number of
observations
Notes:
1.
2.
3.
(1)
(2)
(3)
Agriculture
0.0004
(0.0075)
- 0.0003
(0.00005)***
- 0.024
(0.005)***
Life
- 0.011
(0.005)**
- 0.0001
(0.00004)
- 0.014
(0.006)**
Height
- 0.0007
(0.0002)***
- 2E−6
(2E−6 )**
- 0.0007
(0.0002)***
0.23
0.19
0.35
(4)
(5)
The dependent variable is:
VAP F
VAP T
0.00003
- 0.007
(0.002)
(0.010)
- 0.0003
- 0.0002
(0.0001)*** (0.0001)
- 0.008
- 0.017
(0.006)
(0.011)
Fixed Region Effects OLS
0.14
0.03
78
78
78
78
78
Robust standard errors allowing for clustering by region in parentheses.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
All specifications include a constant term.
40
(6)
(7)
(8)
(9)
VAP C
- 0.022
(0.011)*
- 0.0003
(0.0001)***
- 0.032
(0.007)***
VAP M
- 0.009
(0.006)
- 0.0004
(0.0001)***
- 0.035
(0.011)***
VA MI
- 0.114
(0.131)
- 0.005
(0.003)*
- 0.334
(0.204)
VA W
- 0.004
(0.010)
- 0.001
(0.0001)***
- 0.068
(0.011)***
0.20
0.22
0.03
0.28
78
78
78
78
Table IX: Focusing on the 1871-1911 Sub-sample
Sericulture
Sericulture 2
Wheat
Wheat 2
Arboriculture
Arboriculture 2
Army
Army
2
Distanceto-Enemies
Distanceto-Enemies 2
(1)
(2)
(3)
Land-Taxes
- 9.095
(107.503)
- 6.539
(47.358)
- 7.941
(99.091)
75.175
(52.964)
- 36.015
(71.637)
107.192
(98.862)
- 1.305
(1.238)
0.102
(0.066)
0.015
(0.005)***
- 6.36E−6
(2.35E−6 )**
GDP
LTD
Estimation
R2
Number of
observations
Notes:
1.
2.
3.
0.43
(6)
(7)
(8)
Turnout
(4)
(5)
The dependent variable is:
Illiterates
VA F
VA T
VA C
VA M
- 0.055
(0.010)***
- 0.016
(0.005)***
0.038
(0.006)***
- 0.003
(0.007)
- 0.018
(0.013)
- 0.034
(0.020)
- 0.016
(0.012)
- 0.0004
(0.0001)***
0.00001
(0.00005)
- 0.00003
(0.00002)
- 0.0003
(0.0001)***
- 0.0002
(0.0002)
- 0.0004
(0.0001)***
- 0.0004
(0.0001)***
- 0.027
(0.008)***
0.003
(0.002)
- 0.038
(0.018)*
- 0.060
(0.011)***
- 0.065
(0.015)***
0.39
0.15
0.003
- 0.031
(0.002)
(0.008)***
Fixed Region Effects OLS
0.53
0.20
0.07
0.21
0.22
65
65
65
65
65
65
65
65
Robust standard errors allowing for clustering by region in parentheses.
*** denotes significant at the 1 % confidence level; **, 5 %; *, 10 %.
All specifications include a constant term.
41