Organized crime remains a subject of intense debate within the sphere of social sciences, primarily due to its profound implications for individuals, commercial entities, and entire nations. This form of criminal activity, which encompasses violence, corruption, extortion, and drug trafficking, has been empirically linked to disruptions in economic development, the political process and global prosperity (Acemoglu et al. 2013; Alesina et al. 2018; Battisti et al. 2020; Dal Bó et al. 2006; Daniele and Geys 2015; Pinotti 2015; De Feo and De Luca 2017; Baraldi et al. 2022; Bernardo et al. 2021; Balletta and Lavezzi 2023). A prime example of such a criminal group is the Sicilian mafia, an organization with roots dating back to the early 1800 s, which continues to exercise significant territorial influence. As argued by Gambetta (1993), the Sicilian mafia operates much like a business enterprise, generating, fostering, and vending protection to the private sector. Skaperdas (2001) generalizes this point, describing the different instances in which the demand for protection, or security, from a criminal organization, can emerge. In the case of Sicily, various historical accounts affirm that the genesis of the Sicilian Mafia was triggered by the need to safeguard the territory from predatory threats, especially during an era marked by weak governmental security and widespread banditry in the region (Varese 2014).

Starting with the assumption that the mafia proliferated in regions where institutions lacked strength, the economic literature has endeavoured to elucidate the genesis of the Sicilian mafia in the 19th century by scrutinizing various hypotheses. Bandiera (2003) provided an initial interpretation by demonstrating a correlation between land fragmentation and the activity of criminal organizations.

Differently, recent contributions have shed light on the relationship between natural resources and the prevalence of such groups. Specifically, they associate the existence of criminal syndicates with the necessity to safeguard areas rich in agricultural produce and natural resources. Notably, Buonanno et al. (2015) present evidence indicating that the mafia initially established a stronghold in regions abundant with sulfur mines. In related work, Ciccarelli et al. (2023) observed that mafia activity spread more easily in communities with small to medium-scale sulphur mining, while it had far less impact in towns with higher rates of mineral extraction. In a parallel vein, Dimico et al. (2017) postulate similar assertions, accentuating a significant correlation between the existence of criminal syndicates and citrus orchards.

In contrast to origin-focused studies, Acemoglu et al. (2020) emphasize the fluctuation in the presence of the mafia between 1885 and 1900. They propose that drought-induced wage demands from peasants incited a ‘demand’ for protection among landowners, who sought the Mafia’s assistance in quelling peasant uprisings.

This research aims to enrich the debate by examining how the Sicilian Mafia thrived in a context characterized by weak institutions and pronounced socio-economic disparities in land ownership. These disparities, manifested in the widespread diffusion of large-scale farming estates known as latifundia, fostered the socio-economic conditions for organized crime. Concentrated land ownership, particularly pronounced in western Sicily, exacerbated economic inequality and weakened institutional oversight, conditions historically associated with the rise of organized crime. This historical variation, as documented by Tocqueville (1821) and Sylos Labini (2014), forms the foundation for our analysis.

In this context, Pazzona (2010) offers valuable insights by exploring the relationship between land ownership and Mafia presence, emphasizing the role of resource competition and emerging social classes following the abolition of feudalism. However, while valuable, Pazzona’s analysis does not address potential endogeneity between land inequality and Mafia activity.^Footnote1 Nevertheless, our study refines and extends Pazzona’s work by employing a larger dataset of 265 municipalities, allowing for greater granularity in examining land concentration. Crucially, we introduce a continuous measure of latifundia, constructed from municipal-level data on crop diversification (Mortillaro 1854), a source also used by Acemoglu et al. (2020), to more accurately capture land inequality. Additionally, we propose a novel hypothesis: natural disasters, particularly earthquakes, acted as critical exogenous shocks, reshaping land distribution and fostering the socio-economic conditions for Mafia emergence. Using earthquake intensity as an instrumental variable, our analysis addresses endogeneity concerns and isolates the causal pathway between land inequality and organized crime.

The differential impact of earthquakes on land distribution provides a natural experiment to understand how external shocks shape socio-economic landscapes. Affected areas experienced localized disruptions to landownership structures due to property destruction, population displacement, and shifts in land use. In contrast, unaffected areas retained their pre-existing concentrated patterns of ownership. To support this hypothesis, we also consider other historical reforms–such as the abolition of feudalism and the sale of church properties–that could have interacted with seismic events in shaping land ownership. Historical evidence, however, suggests that these broader reforms did not substantially alter landownership structures. For example, Renda (1997) and La Manna (2019) argue that the abolition of feudalism primarily served the interests of the aristocracy and the emerging bourgeoisie, leaving the rural poor largely disenfranchised. Similarly, the redistribution of church lands, intended to democratize land ownership, often enabled wealthy elites to acquire these properties, further entrenching land concentration rather than alleviating it (Barbera Cardillo 1988; Falzone 1984). See also (Pazzona 2010, pp. 6-7), and (Carraro 1976, p. 94). In contrast, earthquakes had geographically localized effects, creating unique conditions that likely disrupted existing land distribution patterns in the affected areas.

Our analysis reveals a robust positive relationship between the prevalence of latifundia in the 19th century and the emergence of Mafia activity at the beginning of the 20th century. By employing an instrumental variable strategy, we demonstrate that earthquakes significantly influenced land ownership patterns, creating conditions conducive to the rise of organized crime. These findings contribute to the broader literature on the socio-economic consequences of natural disasters (Barone and Mocetti 2014; Belloc et al. 2016; Buonanno et al. 2023) and deepen our understanding of the historical roots of organized crime (Acemoglu et al. 2020; Buonanno et al. 2015; Dimico et al. 2017). By highlighting the critical role of exogenous shocks in shaping land distribution, this study underscores their enduring impact on institutional and economic trajectories.

The rest of the paper is organized as follows. Section 2 describes the historical connections between latifundia, earthquakes, and the emergence of the Sicilian Mafia and provides details on our identification strategy. Section 3 introduces the methodology for the empirical analysis. Section 4 discusses our dataset. Section 5 presents our results, and Sect. 6 concludes.

In this section we describe the salient characteristics of the relationship between the latifundia and the emergence of the Mafia, and provide the details of our identification strategy.

2.1 Latifundia and mafia

Latifundia originated with the dissolution of Feudalism in 1812.^Footnote2 Land was not part of the fief anymore but, far from being allocated equally among private owners, it widely became concentrated in large estates: “between 1812 and 1860, the number of large landowners in Sicily rocketed from 2,000 to 20,000” (Gambetta 1993, p. 91, quoting Mack Smith’s History of Sicily). In particular, the end of Feudalism implied that peasants changed their status from a social servitude (i.e., they had a right for gleaning and pasturage under Feudalism) to workers on the job market, but ended up in economic and political dependency as the avid bourgeoisie expropriated them (Blok 1969, p. 95). As such:

the latifundium generated the forces for its own destruction: the peasant movement. But the large Sicilian estate possessed its proper control and defense mechanism […] mafia was the most crucial equipment of latifundium (Blok 1969, p. 95).

Latifundia owners were mostly absent from the countryside. The administration of the latifundia was secured by leaseholders (gabellotti), who held the lands under a six-year lease and paid a fixed fee in advance to the owner, or by stewards (amministratori) “who managed the estates on account of the owner” (Blok 1969, p. 97). The cultivation of the latifundia was based on a three-year cycle, in which the first year the land was kept idle (fallow), the second year wheat was cultivated, and the third year was devoted to pasture.

The gabellotti sub-leased small strips of land to peasants on short-term contracts. Preparation of the fallow was organized by gabellotti with their servants in May. In September, the land was sub-leased to peasant-sharecroppers (mezzadri), who took care of subsequent ploughing, sowing, weeding, harvesting, etc., until the threshing of the next August. The gabellotto advanced seeds and loans, to be repaid with interest at the time of the harvest. The living conditions of the mezzadri were meager, as they were permanently in debt, making it difficult for them to repay the loans (soccorsi) that the gabellotti made them (e.g., in the form of seeds) (Blok 1969, p. 98). Among other outlays, mezzadri had to pay gifts that were used to pay the campieri, who acted as guards of the fields. In other words, peasants were paying for protection (Blok 1969, p. 98).

The standard approach of gabellotti, soprastanti (e.g., high-rank supervisors), and campieri toward the peasant-sharecroppers was violent and coercive. Gabellotti appointed campieri and soprastanti, while campieri filled the rank-and-file positions, mostly as favors granted on the basis of personal relationships in the nearby rural towns, where most of the population was living, given that the countryside was largely an unfriendly environment (Blok 1969, p. 97). This form of latifundia featured little investment in capital and/or technological improvements. As such, it generated a condition of permanently low standards of living for the peasant-sharecroppers. In other parts of Italy, by contrast, the mezzadria system implied that the sharecroppers were better endowed and could therefore live in better economic conditions (Blok 1969, p. 101).

The Sicilian mafia emerged in these circumstances. Most of the leaders of local mafia units, denoted as cosche (“local sodalit[ies] of one or more powerful men and their followers”), were gabellotti (Blok 1969, pp. 105-106).^Footnote3 From this description, a particular role for the mafioso appears, that is:

The mafioso was a mediator whose advice, support, and benefits were sought by peasants on the one hand and landowners on the other. He rendered services to both categories, at the same time using them as resources. (Blok 1969, p. 106)

Skaperdas (2001), p. 185 also stresses this aspect:

In rural Sicily the mafioso used to perform straightforward intermediary and judicial functions that were conferring legitimacy from the community.

The estate owners used mafiosi for their ends, but mafiosi should be understood as a rather autonomous group. They were not at all comparable to comparable to “feudal henchmen” that were completely dependent on the landlords (Blok 1969, p. 106).

The role of this early mafia, however, is more nuanced than some works portrait, in particular, once its role of autonomous mediator between the landlord and the peasants is understood. For example, Acemoglu et al. (2020) stress the necessity that Sicilian landlords had to protect themselves, through the mafia, against peasants’ rebellion, at the time in which it became better organized through the fasci siciliani of socialist inspiration. The historical reconstruction above, however, suggests that a mafia emerged at a time in which peasants’ rebellion was far from being organized, as the mafia appeared as an autonomous provider of protection (or security, (Skaperdas 2001, p. 174)) and other “mediation services” to both the landlords and the peasants.

To sum up, this historical reconstruction makes it clear that the Sicilian Mafia emerged as a response to weak institutions, offering protection and mediation services within a context of concentrated land ownership and socio-economic dependency, as discussed by Falzone (2019) and Dickie (2004).^Footnote4 It also suggests that a mafia emerged at a time in which peasants’ rebellion was still not organized, as subsequently happened through the fasci siciliani analyzed by Acemoglu et al. (2020). However, Sicily’s socio-economic landscape was far from uniform. While these dynamics explain the Mafia’s emergence in Western Sicily, external shocks such as seismic events in Eastern Sicily created a markedly different trajectory. The widespread destruction caused by these earthquakes triggered significant shifts in land ownership, governance, and economic practices, providing a natural experiment to understand the historical roots of Sicily’s socio-economic dualism. These seismic events offer a unique lens to analyze the divergence between regions dominated by large estates and those characterized by land fragmentation. The following subsection examines into how these natural disasters reshaped Eastern Sicily, altering patterns of land tenure and socio-economic outcomes.

2.2 Exclusion restriction: earthquakes and latifundia

The validity of the exclusion restriction central to our empirical approach requires to demonstrate that seismic events introduced variation in land distribution patterns that is independent of other contemporaneous reforms. The following discussion explores the geographically localized impacts of earthquakes, contrasts these effects with broader island-wide reforms, and addresses alternative channels that could influence the prevalence of latifundia.

2.2.1 Seismic events and reconstruction policies in eastern sicily

Between 1600 and 1850, eastern Sicily experienced 21 seismic events ranging from medium to very high intensity, while western Sicily recorded only three of medium intensity.^Footnote5 Among these, the 1693 Val di Noto earthquake and the 1783 Messina-Reggio Calabria earthquake were particularly consequential, reshaping the socio-economic landscape of southeastern Sicily. Post-disaster reconstruction efforts–encompassing wealth redistribution policies, infrastructural development, and institutional reforms–provided unprecedented opportunities for upward mobility among lower-income households.

The 1693 Val di Noto earthquake stands as one of the most devastating natural disasters in Sicilian history. It claimed around 60,000 lives, destroyed over 45 towns and villages, and affected more than 5,600 square kilometers (Puleo 2010). A subsequent tsunami compounded the destruction along the Ionian coast from Capo Passero to Messina (Piatanesi and Tinti 1998). In response, Sicily’s Viceroy, Juan Francisco Pacheco (Duke of Uzeda), and his lieutenant, Giuseppe Lanza (Duke of Camastra), orchestrated extensive rebuilding efforts under the central government’s auspices. They prioritized the reconstruction of churches and urban centers, symbolic acts of resilience that also facilitated the development of the Sicilian Baroque style–an architectural tradition distinguished by its intricate ornamentation and harmonious urban planning (Tedesco et al. 2007; Puleo 2014). At the same time, the central government introduced policies aimed at stimulating recovery and private investment. Key measures included facilitating property transfers, adjusting financial obligations, granting tax exemptions, and encouraging land-use mechanisms such as emphyteusis (Ligresti 1993; Gianrizzo 1997; Condorelli 2006).

A key policy of the reconstruction strategy was the widespread application of emphyteusis – an arrangement granting the holder (emphyteuta) the right to use and improve another’s land in exchange for an annual fee.^Footnote6 Emphyteusis facilitated the transfer of unused or damaged plots and fostered stable economic conditions for rebuilding. As documented by Condorelli (2006), the municipal senate of Catania actively encouraged the allocation of unreconstructed properties to individuals committed to restoring them. Complementary measures included reducing property prices by one-third, eliminating mortgages on sold buildings, and standardizing urban land values. Records from the Lettere Viceregie in the State Archives of Catania and Palermo enumerate at least 115 emphyteutic agreements and related transactions, underscoring the pivotal role this policy played in reviving the urban landscape. The revenues generated were frequently reinvested into private construction projects, further bolstering post-disaster recovery efforts.

Religious institutions also employed emphyteusis to advance reconstruction. For instance, the Convent of San Michele Arcangelo, lacking funds to meet the Duke of Camastra’s rebuilding mandate in 1695, granted some lands in perpetual emphyteusis to raise the necessary capital (Rizzo Pavone and Iozzia 1995). Likewise, the Convent of San Francesco d’Assisi in Catania leveraged its substantial real estate holdings by issuing emphyteutic and redeemable leases, generating stable revenues for reconstruction (Casagrande 1906). A particularly innovative use of emphyteusis emerged in Militello, where the Benedictine Monastery initiated a pioneering social housing program in 1708. Monastery lands were allocated for the construction of modest homes (case terrane) and multi-story residences (case solerate), with financial support provided through low-interest loans and emphyteusis agreements allowing eventual ownership. As discussed by Abbotto (2008), this project not only addressed immediate housing needs but also offered a regulated, sustainable model of social housing that blended architectural integrity with economic sustainability.

Beyond land reforms, the central government granted temporary tax exemptions to assist disaster-stricken municipalities. Duties and royal levies were reduced or fully waived, except for those essential to sustain religious services and local administration. This fiscal relief offered breathing room to communities, allowing them to channel resources more effectively into reconstruction (Consiglio Patrimoniale 1693).

Inheritance laws were also adapted to the new circumstances. Traditionally, baronial lands given as dowries would revert to the original family upon the daughter’s death, preserving the unity of large estates. After the catastrophic loss of life during the earthquake, an exception permitted lands to pass to surviving husbands if their wives were victims of the disaster, reflecting the urgent need to address unprecedented demographic pressures (Gianrizzo 1997).

Similarly, the 1783 Reggio Calabria and Messina earthquake prompted the government, under Vicar General Francesco Pignatelli, to pursue additional reconstruction efforts financed through agrarian reforms, including the expropriation and resale of ecclesiastical properties (Placanica 1970). These measures aimed to mitigate social unrest and incentivize residents to remain in the affected areas.

These localized effects underscore the plausibility of earthquakes as an exogenous source of variation in land distribution patterns, independent of broader historical reforms

2.2.2 Contrasting the experiences of eastern and western sicily

Emerging research (Abdullah et al. 2016; Mendoza and Jara 2022; Keerthiratne and Tol 2018) suggests that natural disasters, while destructive, can foster socio-economic mobility by opening new opportunities and redistributing resources. Although direct income data for pre- and post-disaster Sicily are unavailable, it is plausible that the institutional reforms implemented during reconstruction enhanced social mobility in Eastern Sicily. More equitable access to land, the liberalization of property transactions, and incentives to rebuild likely enabled disadvantaged households to acquire new assets. Over time, these changes encouraged a shift from monoculture toward more diversified, intensive farming practices, ultimately promoting a more balanced and resilient economy (Vinciguerra 1999).

In contrast, Western Sicily, having experienced far fewer seismic shocks, underwent no comparable socio-economic transformation. It transitioned from feudal arrangements into a landscape dominated by large estates (latifundia), characterized by concentrated landownership and extensive farming methods. This entrenchment of inequality and the absence of institutional innovation created an environment conducive to the emergence of the Mafia, which thrived on weak governance and persistent economic disparities (Renda 1997; Sylos Labini 2014). The stark divergence between Eastern and Western Sicily highlights how exogenous shocks, rather than uniformly applied reforms, can significantly reshape land distribution and institutional trajectories.

The divergence between these regions supports the argument that seismic shocks, rather than uniformly applied reforms, drove the distinct socio-economic trajectories observed across Sicily.

2.2.3 Other possible channels

While earthquakes had geographically localized impacts, reforms like the abolition of feudalism were uniformly applied across Sicily. By examining these alternative channels, we argue that their limited effectiveness in disrupting entrenched power structures further supports the exclusion restriction.

The abolition of feudalism in 1812, for instance, formally dismantled the feudal order by granting full ownership of estates to aristocrats. However, rather than democratizing land access, it effectively reinforced elite dominance. Communal rights such as gleaning and pasturage, which had previously provided subsistence for peasants, were abolished, deepening their dependency on landlords and consolidating existing power hierarchies (Fiume 1984; Renda 1997; La Manna 2019). Similarly, attempts to redistribute church properties after unification were frequently co-opted by elites, who exploited corruption and forged contracts to acquire large tracts of land, perpetuating the latifundia system (Barbera Cardillo 1988; Falzone 1984; Marino 2007).

Although these island-wide reforms were significant, they neither fundamentally altered the entrenched socio-economic structure nor created broad-based opportunities for upward mobility. By contrast, the localized impacts of seismic shocks in Eastern Sicily fostered genuine land fragmentation, economic diversification, and social advancement. Reconstruction efforts, including emphyteusis agreements and liberalized land sales, enabled lower-income groups to climb the economic ladder. These changes stood in stark contrast to the largely ineffective, uniformly applied reforms elsewhere on the island (Ligresti 1993; Puleo 2014).

These distinct, geographically specific changes induced by earthquakes provide a more convincing explanation for the divergent outcomes observed across Sicily. Unlike uniform reforms that often reinforced existing inequalities, seismic events introduced region-specific conditions that disrupted entrenched power structures. This localized variation underpins the exclusion restriction to our empirical strategy, ensuring that the observed differences in socio-economic trajectories between Eastern and Western Sicily can be plausibly attributed to the exogenous shocks imposed by earthquakes rather than to other, more pervasive historical reforms.

We start our analysis by providing reduced form evidence on the associations of Earthquake Intensity with Latifundia and Mafia presence in 1900. The first-stage regression, examining the impact of Earthquake Intensity () on Latifundia (), is represented by the following linear regression:

where  denotes the regression error.  is a – dimensional vector of controls that includes other possible determinants of Mafia, geographical controls, and Valley fixed effects. Furthermore, to explore the association of Earthquake Intensity with the Mafia’s presence (), we employ a linear regression that analyzes the connection between the Mafia and Earthquake Intensity, treating the latter as an exogenous variable.

where  denotes the regression error.

Our main econometric model takes the form of a linear instrumental variables (IV) model allowing for the endogeneity of Latifundia using Earthquake Intensity as an instrumental variable (IV). For each municipality , the Mafia is assumed to be described by the following equation:

where the coefficient of interest is , which measures the effect of the extent of Latifundia on the intensity of the Mafia presence. The error term  is assumed to be conditionally mean independent of Earthquake Intensity, . This assumption implies that including control variables is critical to ensure the exogeneity of the IV, , as . Put differently, provided that the control variables are effective, the unobserved factors should not systematically vary with the instrument, Earthquake Intensity in this case. It is worth noting that this assumption focuses our analysis on the latifundia channel, while other potential influences on Mafia presence–such as land division, agriculture, or natural resource management–are controlled for but not directly examined. This approach, informed by literature like the effects of sulfur mining (Buonanno et al. 2015), citrus industry (Dimico et al. 2017), and peasant revolts (Acemoglu et al. 2020), allows us to concentrate on Latifundia’s unique impact.

As discussed in Section 2, we argue that the core exclusion restriction of our IV approach, which posits that earthquake intensity influences Mafia presence solely through its impact on latifundia, is plausible. This is consistent with Bisin and Moro (2021), who suggest that latifundia possess strong persistence due to their deep historical roots in Sicily from the 1600 s to the 1700 s. In our setting, the lagged instrument initially affected land division, and through the persistent nature of land division, it influenced latifundia prevalence by the 1850 s. Consequently, this impacts Mafia presence in the mid-1800 s. The substantial time gap between the earthquakes and the Mafia’s emergence, along with controls for other historical events and economic conditions, supports the validity of our exclusion restriction. Estimation of model (3) relies on two-stage least squares and inference is based on standard errors clustered at the district level.^Footnote7 We test for a weak instrumental variable using the tests proposed by Stock and Yogo (2005) and the robust methodology of Kleibergen and Paap (2006) to account for heteroskedasticity.

Finally, as a robustness exercise in our empirical analysis, we also employ the IV spatial Durbin model with autoregressive errors. This approach allows us to account for spatial dependencies that might exist in the dependent variable, endogenous and exogenous regressors, as well as in the spatial errors, further validating the reliability of our main result. In this model, we define  and . The model is then given by

where , with  being a zero mean projection error. The weight matrix , which is the (i, j)th element of an  fixed spatial contiguity weight matrix, is considered in both row-normalized and non-normalized forms. This model is estimated using the Generalized Spatial Two-Stage Least Squares (GS2SLS) method, which accounts for the potential endogeneity of  using  as an IV. The inclusion of this robustness check serves to reinforce the integrity and comprehensiveness of our analysis.

In this section, we present the variables used in our empirical model, followed by a discussion of additional controls related to alternative channels for Mafia presence and variables associated with the prevalence of latifundia, which serve as alternative mechanisms distinct from the earthquake hypothesis. Our approach accounts for these alternative channels, ensuring that the primary channel remains robust to the inclusion of these controls. Table 1 presents summary statistics, and detailed variable descriptions are provided in Table 6 in the Appendix.

4.1 Main variables: earthquake, latifundia, and mafia presence

The dependent variable, intensity of Mafia presence, is measured as an ordinal variable ranging from 0 to 3, where 0 indicates no Mafia presence and 3 denotes high Mafia intensity. This measure is derived from Cutrera (1900), who compiled municipal-level data on Mafia activity by integrating information from law enforcement agencies and newspapers, covering 285 Sicilian municipalities during the late 19th century. This dataset has served as a benchmark for measuring Mafia presence in influential studies, including Acemoglu et al. (2020), Buonanno et al. (2015), Ciccarelli et al. (2023), and Dimico et al. (2017), underscoring its significance and reliability in the literature.

The key independent variable, Latifundia, refers to large areas of land used extensively, characterized by a rotation of cereals, pastures, and fallow land, with minimal capital and labor input (Lorenzoni 1910). Using a municipal-level dataset from Mortillaro (1854),^Footnote8 we measure latifundia as the share of arable land devoted to non-irrigated cereals and grassland. This measure captures the extent of latifundia in each municipality. The combined dataset from Cutrera (1900) and Mortillaro (1854) provides 265 observations at the municipal level.

To address potential endogeneity concerns, we construct an instrument defined by the maximum Earthquakes Intensity recorded in Sicilian municipalities between 1600 and 1900. Additionally, we use an alternative version of the instrument focusing on the intensity of the 1693 Val di Noto earthquake (Earthquake Intensity 1693), measured on the Mercalli scale (from 4 to 11). Figures 1a–c illustrate the geography of Mafia presence, earthquake intensity, and latifundia prevalence. The comparison highlights that Mafia presence is less prevalent in parts of Sicily that experienced higher seismic activity.

4.2 Control variables

The following sections provide details on the control variables used in the empirical analysis.

4.2.1 Alternative channels of mafia presence

To ensure that the relationship between latifundia and Mafia presence captured in the second stage of the IV regression is not confounded by other factors, we include controls for alternative mechanisms potentially driving Mafia presence.

First, we examine the role of agricultural profitability and natural resources. The production of valuable commodities has been identified as a significant driver of Mafia activity (Buonanno et al. 2015; Colajanni 1885; Lupo 2009). In this context, we include the presence of sulfur Caves, which were central to Sicily’s economy during the Industrial Revolution (Cunha 2019). Furthermore, as highlighted by Lupo (1988), Dimico et al. (2017), and Bandiera (2003), we account for the cultivation of high-value crops, including Citrus, Olives, and Vineyards, which significantly influenced the local economic activity.

Second, we account for economic activity and population density, as Mafia activity was historically concentrated in densely populated coastal areas around Palermo, where latifundia were largely absent (Lupo 1988). To capture this dynamic, we include controls for Population Density and a dummy variable, Urban, which equals 1 for municipalities located within 10 kms of major cities. These variables reflect the impact of urbanization and concentrated economic opportunities on the emergence of organized crime.

Finally, we include geographical features that may influence Mafia activity. Rugged terrain and higher altitudes can hinder law enforcement and provide strategic advantages for organized crime, as noted by Nunn and Puga (2012) and Lupo (2004). To account for these factors, we control for Ruggedness and Altitude, which capture variations in the physical landscape that could facilitate criminal activity or inhibit governance.

4.2.2 Alternative channels of latifundia presence

The abolition of feudalism and the redistribution of church lands were transformative events in Sicilian history, yet their effects were uniformly distributed across the island. This lack of geographical variation prevents their direct inclusion in our analysis. However, these historical factors remain critical to understanding the broader context of land redistribution and its consequences. To establish the validity of our instrument, which relies on earthquake-induced land redistribution as a credible source of exogenous variation, it is essential to demonstrate that its effects are not confounded by these contemporaneous reforms. As argued in Section 2.2, we must ensure that the variation in land ownership patterns we attribute to earthquakes is not simply a reflection of broader, island-wide historical changes that might independently influence Mafia activity.

To address these concerns, our analysis incorporates four key controls that account for alternative mechanisms influencing the prevalence of latifundia. These controls allow us to robustly isolate the unique variation introduced by seismic events and assess their role in shaping land distribution patterns.

Water availability emerges as a pivotal determinant of agricultural productivity and land use patterns, particularly in the context of western Sicily, where resource scarcity has historically supported the predominance of latifundia and monoculture (Blasi 1968; Vinciguerra 1999; Sylos Labini 2014). To capture this dynamic, we include a measure of Drinking Water availability in our analysis. Furthermore, inland water Scarcity serves as a secondary instrument, allowing us to account for variations in water resources that might independently influence land distribution patterns while ensuring the robustness of our findings.

Land suitability also plays a central role in shaping the prevalence of latifundia. Fertile soils favored the cultivation of high-value crops such as Citrus, Olives, and Vineyards, whereas less productive soils were often dedicated to cereal farming. To address this factor, we incorporate the share of hectares allocated to high-value crops alongside land suitability indices for Citrus, Cereal, and Olive cultivation, derived from agro-ecological data provided by the IIASA-FAO Global Agro-Ecological Zones (GAEZ) project (Buonanno et al. 2015).

Another critical dimension is the variation in infrastructure and connectivity, which reflects the contrasting economic demands of extensive latifundia systems and intensive agricultural practices. The former required minimal transport networks, while the latter relied on well-developed infrastructure to optimize production and distribution. To account for these differences, our model includes measures for the presence of Post Roads, proximity to Ports, and distance to major Rivers.

Finally, we consider the impact of malaria prevalence, as areas dominated by latifundia were often characterized by bio-geographical conditions conducive to the spread of malaria (Cammareri Scurti 1907). Including a control for malaria prevalence enables us to account for its potential effects on rural labor dynamics and agricultural practices, ensuring a comprehensive evaluation of the factors influencing land distribution patterns.

4.2.3 Fixed effects

Finally, to account for unobserved heterogeneity across different geographical areas in our analysis, we incorporate Valleys fixed effects. During the Muslim domination, starting in 948, Sicily underwent a territorial reorganization that divided the island into three large regions known as Valleys. This division, referenced in the Norman document of 1094 titled Antiquae Divisiones Saracenorum, remained in effect until the Bourbon administrative reform of 1812. The Valleys, or “Valli,” correspond to specific sub-regions: Val di Mazara in the west, Val di Noto in the southeast, and Val Demone in the northeast (see Fig. 2). As noted by Epstein (1992), these regions exhibit differences in cultural legacies, production systems, and urban structures. By including Valleys fixed effects, we aim to control for these region-specific unobserved influences that may affect both land ownership patterns and the historical prevalence of organized crime.