Abstract
Clientelism, deeply interwoven with critical societal issues such as democratic erosion, rising inequality, and persistent poverty, has further been spotlighted for its incendiary role in regions such as Latin America and Africa, where it acts as a catalyst for violence. Recognizing the profound implications of this topic in both contemporary and future dialogues, our article offers an in-depth examination through a two-pronged approach: a meticulous bibliometric analysis and an encompassing literature review. By leveraging various data and scientific mapping techniques, we aim to analyze the evolution of clientelism research, highlighting pivotal contributors and shifting paradigms. Our analysis notably underscores the preeminent role of U.S.-based institutions and scholars, a reflection of the nation’s foundational contributions to clientelism theory. Further enriching our study, we employ the ‘tree of science’ metaphor to identify seminal works, emphasizing prominent research avenues such as democracy, corruption, Latin America, and Africa. We conclude by mapping out prospective avenues for future research in this domain.
Introduction
The wave of democratization experienced since the mid-1970s in much of southern Europe, Latin America, Africa, and Asia entailed challenges in the emergence and definition of systems and processes of political participation in the countries of these regions (Huntington, 1991; Scott, 1972). It is in this context that clientelism has emerged as a recurring characteristic in incipient democracies, where the mechanisms of political competition are usually flawed by deinstitutionalized relationships of interest between political elites and voters (Keefer, 2007; Keefer and Vlaicu, 2008; Thachil, 2014).
Thus, political clientelism is presented as the establishment of asymmetric transactional ties between political actors, with expectations of electoral access to positions of representation and their eventual voters (Eisenstadt and Roniger, 1980; Hicken, 2011). This relationship-building usually materializes through mechanisms ranging from strategies of favoritism using public or private resources (Kramon, 2016; Robinson and Verdier, 2013), as in the case of vote buying (Nichter, 2008), to forms of blackmail such as those grouped under the concept of armed clientelism (Eaton, 2006).
On the contrary, although clientelism is a category that has been widely analyzed from the perspective of political studies, its beginnings are mainly rooted in contributions from anthropology and sociology (Schmidt, 1977). This interdisciplinarity has been important because it has allowed the adoption of different methodological tools applicable to research on the phenomenon; this is the case of both ethnography, widely used in anthropology, and quantitative studies, increasingly present in political science (Aspinall and Berenschot, 2019).
Regarding the central axes of study, several seminal documents have explored the important adaptability of clientelism to all types of social and political contexts (Hicken, 2011). Therefore, part of the most relevant literature on the subject has established how clientelist relations promoted by political parties (political machines) have become not only a structural characteristic of less consolidated democracies (Desai et al., 2020; Gans-Morse et al., 2014; Stokes, 2005; Szwarcberg, 2015) but also promoters of poverty and inequality traps (Fergusson et al., 2022; Nichter, 2010; Stokes, 2021). Another analytical dimension that has drawn the attention of researchers has been the role of brokers as mediators in the relationship between patrons and clients, in terms of their nature, typologies, characteristics, and action strategies (Camp, 2017; Hicken et al., 2022; Stokes et al., 2013).
Of course, much of this work has been linked to specific case studies or comparative exercises (Kitschelt, 2000, 2007) where Latin America (Garay et al., 2020; Levitsky, 2007; Ocantos et al., 2014; Roniger, 2019), Africa (Dyzenhaus, 2021; Ochieng’ Opalo, 2022; Van de Walle, 2007; Vicente and Wantchekon, 2009; Weipert-Fenner et al., 2018), and Asia (Aspinall, 2014; Das and Maiorano, 2019; Jamal, 2022; Yıldırım, 2020) remain natural laboratories for the analysis of the interrelationships between clientelism, democracy, and development.
Despite the relevance of this topic, almost no research has been conducted to date that establishes the evolution and trends of research on clientelism at a global level. So far, only one article has been identified: Pellicer et al. (2022) carried out a meta-analysis of 40 documents which, on the basis of a cluster analysis, generated a typology of clientelism based on two dimensions (the equal–unequal dimension and the individual–universal dimension).
Accordingly, this article intends to contribute to this area, by undertaking a bibliometric analysis that incorporates performance analysis, scientific mapping, and network analysis, to establish the origins, evolution, and trends of worldwide research on clientelism. To do this, the Web of Science (WoS) and Scopus databases are used, and through bibliometric tools (R-studio, Gephi, and Tree of Science) and network analysis procedures, the most relevant authors, countries, journals, and documents in this area are identified; the latter are classified using the metaphorical scheme of the tree (roots, trunk, and branches). Finally, through network analysis, the main clusters in which research on clientelism is currently framed are established.
Specializing in the statistical analysis of bibliographic data, bibliometric studies stand as an advanced category of systematic literature reviews (Lim et al., 2022). Their rising popularity is attributed to their adeptness in handling large data sets, thereby facilitating the detection of emerging trends, collaboration patterns, and key research foci (Mukherjee et al., 2022). Such analyses are instrumental in mapping out the intellectual landscape of a specific field based on existing literature (Verma and Gustafsson, 2020). When executed rigorously, bibliometric studies lay a solid foundation for fostering innovative and impactful advancements in a given discipline (Donthu et al., 2021). Moreover, their versatility and scope surpass those of other review methods, making them particularly useful tools for enriching both theory and practice in a field of study (Fan et al., 2022; Mukherjee et al., 2022). For these reasons, the present article adopts this research approach.
The first section of this article is the theoretical approach, describing the origin of and main orientations of research on clientelism. The second section, materials and methods, explains the procedure undertaken to obtain, process, and analyze the information. The third is the performance analysis, presenting descriptive data regarding worldwide scientific production on clientelism. The fourth is the scientific mapping, analyzing the citations and co-citations of authors and documents. The fifth is the network analysis, where the most relevant works in the area are identified and classified, using the tree metaphor (roots, trunk, and branches), and the clusters identified are developed and elaborated on. Finally, the conclusions and limitations of the article are presented, along with an agenda for future research.
Theoretical approach
The study of clientelism as a political phenomenon began, marginally, in the 1950s and 1960s (Eisenstadt and Roniger, 1980). However, a deeper approach to this conceptual category began to gain relevance in the 1970s, when it was shown to be a characteristic of political regimes that were not full democracies in regions such as Latin America, Africa, Southeast Asia, and the less developed areas of Europe (Landé, 1983; Schmidt et al., 1979; Scott, 1972). Since then, the analysis of clientelism has focused on establishing the dynamics of the client–patron relationship within the framework of electoral competition, where the political parties that participate in it assume a strong de-ideologization as a consequence (Eisenstadt and Roniger, 1984; Gherghina and Nemčok, 2021; Graziano, 1976; Hicken, 2011; Hilgers, 2011; Pellicer et al., 2022). Thus, political clientelism is initially conceived as the establishment of affective, asymmetric, reciprocal, and transactional (clientelist) relationships between actors who have expectations of access – or effective access – to the administration of scarce resources (patron-political party-state) that are of interest for others who do not possess them (client-voter) (Legg, 1972).
At the beginning of the 1980s, the study of political-clientelist phenomena flourished, and its dynamic character was noted (Eisenstadt and Lemarchand, 1981). In this sense, analyses of the particularities of political systems, mechanisms of resource distribution within them, and stages of socioeconomic development determined research on clientelism and its transformations (Landé, 1983). Similarly, at this stage, research began to address the effect of clientelism and its strong structures and dyadic interactions on the stagnation or promotion of a variety of processes: political and economic modernization, social integration, independence from international actors, and the renewal of elites and even revolutionary movements (Eisenstadt and Roniger, 1981; Poggi, 1983).
The final decades of the 20th century were characterized by strong processes of transition toward democracy in countries that were heirs to authoritarian regimes of a civil or military nature. Latin America, Africa, Southeast Asia, and ex-communist European and Mediterranean countries emerged as examples of complex democratizing projects (Ekiert, 1991; Huntington, 1991, 1997; Joseph, 1997; Lee, 2002; Mainwaring and Viola, 1985; Remmer, 1992). Hence, the relationship between the use of clientelist practices within these transition processes and the consolidation of democracy in these regions significantly marked the research agenda during this period. In this way, studies on the instrumentalization of clientelism by political parties in southern Europe showed that these actors became authentic political machines (Mavrogordatos, 1997; Papadopoulos, 1997). Likewise, works such as that of Birch (1997) pointed to the challenges imposed by clientelism on the consolidation of democracy in settings like post-Soviet Ukraine. For his part, Auyero (1999), based on the Argentine case, proposed scenarios similar to those observed in previous cases, establishing relationships between the presence of political clientelism and the limitations of democracies in Latin America.
On the contrary, in this same period, studies such as those by Fox (1994, 1996) explored the multidimensionality of democracy far beyond the universal exercise of the vote. Studying the Mexican case, Fox investigated how it was possible to find clientelist elements in different hierarchical power relations that went beyond the simple scenario of electoral competition. Unlike the theoretical approaches developed in previous decades, this period led to a great interest in establishing what concerned the field of clientelist practices and what concerned other types of negotiations between hierarchically asymmetric actors involving an exchange of political rights for social benefits (Abers, 1998).
In the first decade of the 21st century, the expansion and deepening of studies on clientelism implied important methodological advances. An example of this tendency is the analysis of Auyero (2000) regarding the Argentine case, where he applied ethnographic instruments directly in the field in order to establish the logics of the transaction networks of essential goods and services between citizens and political intermediaries. Along the same lines is the work of Wantchekon (2003), who carried out an extensive field study in Benin, West Africa, to investigate the incidence of clientelism in electoral behavior during the country’s presidential elections, managing to differentiate the impact of clientelist messages by region and gender.
Likewise, Hallin and Papathanassopoulos (2002), with a rigorous application of comparative methodologies applied to a sample of several countries in Latin America and Mediterranean Europe, established that contexts with high political instrumentalization of the media, limited autonomy of the professional practice of journalism, and the politicization of regulatory bodies are fertile for the rise of clientelist practices. Regarding this interrelation between the role of the media and the phenomenon of clientelism, we also find the important contribution, from the discipline of anthropology, of Roudakova (2008). Within the framework of comparative studies, Clark’s (2004) research in this period also stands out, in which he dealt with the analysis of patron–client relationships between moderate middle-class Islamic groups in Egypt, Jordan, and Yemen.
Although the literature on clientelism has largely focused on case studies, it was not until the first decade of this century that research on Africa and the Middle East achieved an increasingly important position (Chanie, 2007; Clark, 2004; Lindberg, 2010; Lust, 2009; Van de Walle, 2007; Vicente and Wantchekon, 2009). Regarding Latin America, some of the social particularities of the region began to determine more and more the study of the evolution of clientelism (Szwarcberg, 2012). In this sense, dynamics such as armed conflict, criminal structures, and the illicit drug trade began to be identified as relevant variables in the evolution of clientelism in countries such as Colombia (Eaton, 2006) and Brazil (Arias, 2006).
At this stage, because of the democratization processes launched in various parts of the world in the preceding decades, several investigations began to establish systematic differences in performance between the youngest democracies and the oldest and most consolidated ones. Studies such as those of Keefer (2007), Keefer and Vlaicu (2008), and Manzetti and Wilson (2007) showed that, in general, young democracies are characterized by weak rule of law and corruption in their public administrations, which causes the inability of political competitors to make credible promises to citizens, at the same time that it empowers clientelism as a mechanism of access to and permanence in power. Similarly, populism, another of the most characteristic dysfunctions in fragile democracies, began to be widely studied during this period in terms of its collusion with clientelism (Alamdari, 2005; Levitsky, 2007; Penfold-Becerra, 2007; Thompson, 2010).
In recent years, the multidimensionality of the phenomenon of clientelism has made it necessary to explore new nuclei of analysis, as is the case of the dynamics around the construction of loyalties. For example, Stokes et al. (2013), from a comparative perspective, investigate how, through different redistributive strategies, loyalties are created between clients and patrons that go beyond electoral situations to become long-term relationships. In addition, the literature has also been interested in the scope of the relationships between patrons and brokers (Aspinall, 2014; Auerbach and Thachil, 2018; Holland and Palmer-Rubin, 2015; Novaes, 2018), demonstrating that the loyalty ties between them are relative and can be easily broken based on the individual or organizational interests of the latter (Aspinall, 2014; Holland and Palmer-Rubin, 2015; Novaes, 2018). Likewise, recent research has indicated that, thanks to the complexity of the social networks in which political brokers are immersed, the construction of loyalties that feed clientelism is increasingly viable without the need for strict supervision systems (Aspinall and Hicken, 2020; Hicken et al., 2022; Hicken and Nathan, 2020; Ravanilla et al., 2021).
Another approach that has strongly marked clientelism studies in recent years is the inclusion of different dimensions linked to development (Fergusson et al., 2022; Shchukin and Arbatli, 2022; Yıldırım and Kitschelt, 2020). Thus, the classical approach to the phenomenon from the clientelism–democracy binomial has been expanding to include other variables, such as poverty and inequality. Studies like those by Hicken and Weitz-Shapiro are pioneers in this line of analysis (Hicken, 2011; Weitz-Shapiro, 2012). The most recent research has examined how clientelism becomes more effective in the midst of conditions of poverty and inequality, causing harmful effects on development results due to the creation of a poverty trap that leads governing parties to lack incentives to promote redistributive policies (Lo Bue et al., 2021; Stokes, 2021).
Materials and methods
Bibliometric research, regarded as a specialized form of systematic literature review, involves the application of quantitative and statistical techniques to bibliographic data, such as publications and citations (Donthu et al., 2021; Fan et al., 2022). Its primary purpose is to quantify scientific output, assessing its quality and impact (Cobo et al., 2011). Beyond these metrics, bibliometrics aims to understand the social, intellectual, and conceptual structure of scientific knowledge by exploring bibliographic networks, like those generated by co-words, co-citations, or co-authors (Batagelj and Cerinšek, 2013). As a result, such studies have proven invaluable, offering unique opportunities to enhance both theory and practice in the scientific realm (Mukherjee et al., 2022).
Within bibliometrics, scientific mapping emerges as a subdiscipline aiming to visually represent entire scientific fields. These maps, be they two-dimensional or three-dimensional, depict the landscape of science, highlighting themes and specific topics of the field under study (Noyons, 2001). This process allows for the identification of interconnections between disciplines, fields, specialties, documents, and authors (Small, 1999). In this light, the specific scientific knowledge of a field is perceived as a system, employing a network structure to represent interactions between various scientific actors (Boccaletti et al., 2006).
To conduct this study, we adopted a bibliometric procedure supported by scientometric tools and indicators to decipher the intellectual structure and discern patterns and trends in a knowledge area (Chen, 2017; Leydesdorff, 1987; Noyons, 2001). We carried out various analyses, including author co-citation analysis, document co-citation analysis, and word co-occurrence analysis. In addition, we used techniques for network visualization (Herman et al., 2000), alongside indicators and metrics that assess the impact of publications and authors, such as citation count and the h-index.
Below, we elaborate on the indicators and analyses used:
Author co-citation analysis (White and Griffith, 1981). Co-citation analysis, a method for scientific mapping, posits that authors cited together frequently share thematic similarities (Chen, 2017; Hjørland, 2013). The advantage of this analysis is that it unveils seminal publications and authors in a field (Donthu et al., 2021).
Document co-citation analysis (Small, 1973). This technique, used in scientific mapping, is rooted in the notion that if two publications share references, they likely cover similar contents (Donthu et al., 2021; Weinberg, 1974). It aids in illuminating central themes shaping the intellectual structure and its evolution over time within a specific research field (Zupic and Čater, 2015).
Word co-occurrence analysis (Callon et al., 1983). Word co-occurrence analysis applies text mining techniques to document titles, abstracts, and keywords. By examining the frequency of these word repetitions, dominant themes in the field can be identified (Van Eck and Waltman, 2014).
Citation count (Garfield, 1955). This metric refers to the total number of times an author or academic article has been cited by others. It is among the most employed factors in gauging relevance and ranking authors/documents based on how often their works have been cited (Rovira et al., 2019).
h-index (Hirsch, 2005). This metric measures both the productivity and bibliometric impact of publications. It’s applied to researchers, institutions, journals, countries, or research groups, becoming a widely accepted indicator already incorporated by major databases (Bornmann et al., 2011).
JIF (Journal Impact Factor). It is an indicator to determine the impact of journals, generated from the data indexed in the WoS Core Collection (Garfield, 1955).
SJR (SCImago Journal Rank) (González-Pereira et al., 2010). Calculated using Scopus database information, this metric gauges a journal’s scientific influence. It considers both the number of citations a journal receives and the importance or prestige of the journals from which these citations originate (Scimago Journal & Country Rank, 2022).
Data collection
The data necessary to undertake the scientific mapping of worldwide research on clientelism were obtained through the combination of publications registered in WoS and Scopus, currently considered the principal databases (Bar-Ilan, 2008; Martín-Martín et al., 2018; Mongeon and Paul-Hus, 2016; Pranckutė, 2021), and using them jointly allowed a broader and more complete vision of the knowledge in a specific area (Echchakoui, 2020; Zhu and Liu, 2020). For the search, the terms ‘clientelism’ OR ‘clientelismo’ OR ‘clientélisme’ were used, and as parameter title, abstract, and keywords. This search generated 1739 records in Scopus and 1486 in WoS (consultation date 10/01/2023), which, when contrasted, established 614 duplicates which were then eliminated, consolidating a total of 2611 publications.
Visualization and analysis
For the analysis and visualization of the data, the Bibliometrix (Aria and Cuccurullo, 2017) and Gephi (Mathieu et al., 2009) tools were used. Bibliometrix (version 3.1) is a free-to-use tool that has several functions for bibliometric studies, including author co-citation analysis, collaboration network analysis, document co-citation, and co-word analysis, and also allows one to work with several databases (Aria et al., 2020). It has been used in various studies of this type (Alhawari et al., 2021; Derviş, 2020; Di Vaio et al., 2021; Duque et al., 2021; Homolak et al., 2020; Queiroz et al., 2022), giving it credibility as a tool for scientific mapping.
Gephi (version 0.9.4) is an application used for network analysis, allowing one to know in detail the way in which its various components and their respective indicators interact. In this case, it was used to analyze the network of co-citations, and it has also been used by previous investigations which have generated important results (Donato, 2017; Donthu et al., 2020; Ferguson, 2012; Jacomy et al., 2014; Meier, 2020).
Impact and relationship metrics are calculated for the documents that make up the citation network, such as Indegree (the number of times it has been referenced), Outdegree (the number of times it has referenced others) (Wallis, 2007), and Betweenness Centrality (the degree of intermediation) (Freeman, 1977), with this latter metric showing that a document is both referenced and references others (Zhang and Luo, 2017). These indicators are used to classify documents using the tree of science metaphor (Robledo et al., 2022; Valencia-Hernandez et al., 2020; Zuluaga et al., 2022). Based on this analogy, three categories are created with which to classify the documents. The roots (high indegree) are the classical documents, or those considered hegemonic, based on their high level of citation (Wallis, 2007). The trunk (high betweenness) includes the documents that cite the roots but which, at the same time, are cited by the branches (Zhang and Luo, 2017), making them structural since they connect the theoretical foundation that originates in the classics with the most recent research. Finally, the branches (high outdegree) are the documents based on the classical and structural ones (Wallis, 2007); these works show the main tendencies within which research in the area is currently framed.
Subsequently, to identify the main clusters in which the global research on clientelism is framed, the clustering algorithm was designed by Blondel et al. (2008). This technique allows, through a co-citation analysis (applied to the documents that are part of the network created from the records coming from the databases), us to classify the documents by groups (given common citations), thus establishing the main research fronts in the area (Chen, 2017). Furthermore, PageRank (Page et al., 1999) is calculated, allowing the most representative documents of each group to be established, taking into account the citations received by other works that are highly cited within the network (Ding et al., 2009; Leydesdorff, 2009; Yan et al., 2010). Finally, through text mining programmed in R-studio, specifically with the WordCloud package (Ohri, 2012), the keywords and titles of each publication in each group are identified and used to generate word clouds, thus facilitating the identification of main themes. This review technique has been used successfully in recent studies (Duque et al., 2021; Grisales Aguirre et al., 2023; María et al., 2023; Robledo et al., 2023). This methodological procedure follows the suggested structure by Donthu et al. (2021), which is described below (Figure 1).
Research methodology.
Results
Performance analysis and science mapping
Figure 2 shows the behavior of the publications since the year 2000, sorted by database and calculating the total based on the complete records from WoS and Scopus, minus the eliminated duplicates. The trend shows that publications in the field of clientelism have increased steadily over the past two decades. The effect of this is reflected in the average number of publications between the years 2000 and 2023, which reached 113 publications per year, although the growth has been faster since 2015 as, between 2015 and 2022, the average was 216 publications. In addition, approximately one in two works registered in the databases was published between 2018 and 2022. These data reflect the fact that the topic is current and that the scientific community is interested in expanding research in this important area.
Publication trends.
Figure 3 lists the journals with the largest number of publications in the field of clientelism. Because WoS and Scopus are contrasted, and their impact indicators are not comparable, it is necessary to present them separately for each journal in order to have a more complete and comprehensive look at their impact (López-Illescas et al., 2008). The impact indicators of the journals are associated, such as the JIF generated from the data indexed in the WoS Core Collection (Garfield, 1955), the SJR calculated using the information contained in the Scopus database (González-Pereira et al., 2010), as well as the quartile in which they are classified in each database.
Publications by journals.
The journal with the highest number of publications on clientelism is the North American journal Comparative Political Studies with 61 records in total. In addition to leading the list, it is classified in quartile 1 of the JIF and SJR, making it a high-impact editorial. It is followed by Democratization (a specialized political science journal) with 43 publications. The eight journals that complete the list are renowned editorials: 9 of the 10 are classified in quartile 1 of the SJR, making them leaders in terms of global research on clientelism.
Table 1 presents the countries that contribute the most to the field of clientelism. Researchers from the United States provide 756 documents, representing close to 30% of global production in the area, followed by the United Kingdom with 325. These two countries are the main powers in the subject. Analyzing the 10 most relevant journals in the field (Figure 3), 4 are from the United States and 6 are from the United Kingdom. This list shows that the contributions to global research on clientelism come mainly from these countries. Likewise, there are other European nations (Germany, Spain, the Netherlands, France, and Italy) and North American ones (Canada and Mexico) that also contribute, but to a lesser degree.
Publications by country.
An analysis of the collaboration network between researchers from different countries shows that scientists from the United States and the United Kingdom are the most active in producing articles on clientelism. In fact, 4 out of 10 articles published on the topic are by authors from these countries. The strong collaboration and integration of researchers from these regions on issues related to clientelism explains this trend. However, researchers from the United States stand out for having a broader research network, reflected in the greater number of co-authors from other countries in their articles. This fact raises the scientific production of American researchers on this subject.
Table 2 shows an analysis of the most productive authors in the field of clientelism. Their h-index and number of citations in the WoS and Scopus databases have been evaluated. The most published author in the area is Edward Aspinall, associated with the Australian National University, with a total of 15 articles in these databases. His h-index in WoS is 19, while in Scopus it is 22. However, the author with the highest impact indicators on the list is James Robinson, a researcher at the University of Chicago. Robinson registers an h-index of 45 in WoS and a total of 15,881 citations, while in Scopus he has an h-index of 47 and 22,425 citations. It is interesting to note that, in Scopus alone, his 5 registered articles have received 959 citations, equivalent to an average of over 191 citations per article.
Publications by author.
NR: not recorded.
It is important to mention that 13 of the 20 authors included on the list are associated with institutions in the United States. This finding is consistent with previous research and suggests that the United States is an important country for research in the field of clientelism. In general, this table provides an overview of the productivity and impact of authors in the area. These data may be useful for future researchers seeking to identify and collaborate with leading researchers. In addition, it could also provide a basis to identify critical areas where further research and the development of future lines are needed.
Figure 4 shows the citation and collaboration networks between authors in the field of clientelism. The first network represents the influence of an author within the scientific community, through the importance of their research and the level of references to it (White and Griffith, 1981). In this case, the five most influential authors, those considered references or key authors in the field of clientelism, are Susan C. Stokes (University of Chicago), Herbert Kitschelt (Duke University), Javier Auyero (University of Texas), James C. Scott (Yale University), and Allen Hicken (University of Michigan). The second network shows the groups of researchers with the highest number of publications as co-authors, and this network is closely related to Table 2. For example, the group formed by Edward Aspinall and Allen Hicken (the first- and ninth-ranked authors on the list of the largest number of publications on the subject) is the most important, since these researchers are co-authors of three articles registered in the databases. In addition, Edward Aspinall and Ward Berenschot (the first- and second-ranked authors) are co-authors of two publications, showing how collaborative work can lead to greater productivity, as in the case of Aspinall.
Authors networks.
Network analysis
To present the network, the tree scheme is used (Figure 5), in which, metaphorically, the documents are classified into three categories, allowing one to describe the origin, evolution, and trends of research in this area. At the roots are 10 seminal documents, also known as classical or hegemonic, which present the theoretical bases of the subject. In the trunk are the 10 documents known as structural, those which connect the classics with the most recent, but especially those that begin to define trends in the area. Finally, the branches are the four main clusters identified, those which determine sub-areas or currents that frame research trends in the field of clientelism.
Clientelism tree.
In the following section, we analyze the documents that make up each part of the tree.
The tree of science metaphor
Seminal documents (roots)
The roots of the tree present a characterization of the central axes of the seminal documents that have determined the bases of the study of clientelism as a phenomenon object of analysis from the social sciences.
Clientelism as a conceptual category has been addressed predominantly in political studies; however, anthropologists and sociologists were the first to use this concept to describe certain hierarchical power relationships that tended to be present in rural societies (Scott, 1972). More precisely, the approach to clientelism from different disciplinary areas has provided the same plurality of methodological approaches for its analysis, ranging from ethnography, widely used in anthropology and sociology, to quantitative analysis, increasingly common in political science (Auyero, 2000; Scott, 1972; Stokes, 2005).
Since their beginnings, studies on clientelism have focused significantly on the analysis of the incidence that, in scenarios of transition and democratic consolidation, the processes of political competition acquire, which, at the same time, are distorted by a strong protagonism of non-institutionalized relations between patrons and clients (Kitschelt, 2000; Scott, 1972). Clientelism has shown an important capacity to adapt to all types of social and political contexts, whether democracies or autocracies (Hicken, 2011). However, it is within the framework of incipient and incomplete democracies that particularistic and corrupt practices like clientelism tend to structurally define the electoral competition between political machines that function as political parties (Stokes, 2005). In this sense, Latin America, Africa, Southeast Asia, and some regions of southern Europe have been natural laboratories to analyze the relationship between clientelism and democratic quality, due to the institutional deficit that their political and electoral regimes present or have presented (Kitschelt and Wilkinson, 2007b; Piattoni, 2001).
The literature shows that the multifaceted nature and diversity of scenarios in which clientelism occurs make it difficult to take a unified position regarding its conceptual definition (Hicken, 2011). Nevertheless, it is possible to find common elements that identify the phenomenon regardless of the concept adopted: dyadic relationships of mutual instrumentalization (patron–client), exchange of benefits (goods and services-political support), asymmetry in the position of power (the patron with the ability to offer, and the client with the need to receive), and iteration in the relationship (expectation of building ‘loyalties’ in the medium and long term) (Hicken, 2011; Nichter, 2008; Stokes et al., 2013).
Structural documents (trunk)
Since the establishment of clientelism as a focal point in social science research, scholarly endeavors have largely centered on examining the interactions between patrons and clients. Various studies have deduced that these relationships can hinge on factors such as the origin of the proffered resources (whether public or private) and whether the recipients are individuals or groups (Gherghina and Volintiru, 2017). For instance, there are individual clients who, in return for their allegiance, procure public benefits like job positions or contracts (Heath and Tillin, 2017). Conversely, there are those who merely trade their votes for private financial gains. Contrary to traditional literature’s perspective, subsequent research has illuminated that classic vote buying doesn’t always culminate in a full-fledged successful transaction. Supporting this viewpoint, Guardado and Wantchekon (2018) illustrated through their study across five African nations that the act of directly paying individuals for their votes bears a significantly diminished influence on electoral turnout and political endorsement than what political entities anticipate. Similarly, in the milieu of Mexico’s 2012 presidential elections, Nichter and Palmer-Rubin’s (2015) research highlighted that clients who overtly display their political affiliation – say, by showcasing political banners in their residences – reap considerably more rewards compared with those who remain reserved.
Conversely, we can discern groups that exchange their allegiance for public advantages, often materialized as ‘pork barrel projects’ or targeted subsidies (Penfold-Becerra, 2007). In addition, there are instances where groups pledge electoral backing in return for private contributions to community associations, potentially of religious or athletic character (Gherghina and Volintiru, 2017). Continuing on this trajectory of a centralized clientelist approach surpassing conventional brokers, we encounter the concept of associative clientelism (Holland and Palmer-Rubin, 2015). Here, political entities attempt to delegate to interest groups or ‘organizational brokers’, where prospective clients already have entrenched ties of solidarity and loyalty. Examples can be seen in certain farmer organizations in Mexico and street vendor collectives in Colombia.
Crucially, studies like the one by Schaffer and Baker (2015), employing a comparative methodology in Latin America, reveal that pinpointing opinion frontrunners in social dialogue networks offers better dividends for political factions. In essence, clientelist outcomes are amplified when focused on the influential circle surrounding a social figurehead, who generally garners the lion’s share of the compensation.
As time has advanced, the exploration of clientelism has transitioned from debates around its conceptual framework and pinpointing its inherent components to a more profound scrutiny of the intricate dynamics among patrons, clients, and middlemen or brokers (Cruz, 2019). These brokers, with their pivotal role as mediators, have garnered immense prominence in the discourse on clientelist ties. Their range of involvement spans from straightforward vote buying to intricate voter engagement tactics (Khemani, 2015). In this vein, a broker’s intimate connection to the clients elevates their significance in this mechanism. It is chiefly due to this proximity that one can ascertain voter inclinations, their propensity to participate in voting, the favors they solicit, and the means to oversee their voting patterns (Cruz, 2019).
Cluster 1. Clientelism and democracy
One of the main focuses of analysis that has remained valid since the appearance of clientelism as a conceptual category is its relationship with the quality of democracy and political institutions. Much of the focus of research in the area has been aimed at establishing how, in societies with incipient processes of democratic consolidation, clientelism appears as an element of distortion of the classic vision of party competition, which should foresee a relationship between voters and political elites based on ideological identities, programmatic proposals, and the results of political management (Kitschelt, 2000; Stokes et al., 2013). Thus, the establishment of clientelist relationships by offering non-institutionalized and selective material incentives in exchange for political support vitiates the expected transparency of democratic competition (Gans-Morse et al., 2014; Kitschelt, 2000; Schaffer, 2007).
This cluster addresses the causal relationship between the modernization of democracy, state capacity, and the presence of clientelism as a mechanism of atypical political relations between elites and voters (Keefer, 2007; Keefer and Vlaicu, 2008; Kitschelt and Wilkinson, 2007b; Kuo, 2018). Studies like that of Bardhan (2022) posit that clientelism turns clients into supplicants for favors, eroding the concept of citizenship, a vital element of democracy. In contrast, works like that of Nichter and Peress (2017) aim to reconsider the general view that describes the client as an actor with minimal autonomy facing the patron. It has been possible to establish that citizens can approach the political machines on their own initiative to formulate their requests and, also, can punish those same machines by voting for their competitors when their demands are not granted.
Another dimension addressed is that the personal nature of the clientelist relationship weakens the state’s capacity to strengthen itself as a provider of public goods since clients are prone to value the patron’s favors more when they are delivered through private benefits than through programs of public health, education, or infrastructure (Bardhan and Mookherjee, 2020). This dynamic stimulates a decrease in the capacity of the state as a consequence of low tax collection, since by lowering the expectations of citizens in terms of obtaining public goods, the incentive to comply with their tax obligations decreases (Fergusson et al., 2022; Kyriacou, 2022; Pittaki, 2021).
In the same way, it is possible to identify that the analysis of the harmful incidence of clientelism in the processes of the democratic construction and consolidation of the state as a guarantor of rights remains valid in research on the phenomenon. It can be affirmed that clientelism promotes the creation of ‘poverty traps’, where the patron–client dependency relationship is systematically nourished by the resignation of citizens to the effective exercise of their democratic rights in a context of convenient deinstitutionalization of the state (Durán and Raciborski, 2008; Stokes, 2021; Wood, 2018).
Cluster 2. Clientelism and corruption
Another important axis of analysis studies the direct relationship between corruption and clientelism as a recurring characteristic in countries with high levels of poverty and inequality, incipient democratic processes, and weak institutional structures (Kawata, 2017; Kitschelt and Wilkinson, 2007b). Even so, although it is difficult to determine a unidirectional causal relationship between the two phenomena, much of the literature suggests that clientelism can help explain why some political systems tend to be more susceptible to corruption than others (Keefer, 2007; Keefer and Vlaicu, 2008). Research shows that countries with a high presence of clientelist practices are more corrupt (Lindberg et al., 2022) than those where the main characteristic of political competition is the open debate of electoral proposals of a programmatic nature (Bulut, 2020; Kuo, 2018).
The connection between clientelism and corruption considers different dimensions (Briquet, 2020). First, common clientelist practices like vote buying are, in themselves, illegal behaviors and, therefore, punishable (Hicken, 2011). Likewise, the need for resources to consolidate patron–client relationships increases the incentives to access funds through illicit means, not only from public sources but also private ones (Gherghina and Volintiru, 2017). On the contrary, studies based on experimental data reaffirm that clientelism reduces the ability of citizens to control corruption via political means (Leight et al., 2020). The selective benefits that political elites offer to voters stimulate the renunciation of the effective exercise of control and accountability mechanisms against the corrupt, generating institutional capture through the culture of impunity (Manzetti and Wilson, 2007; You, 2015). Other findings show that the strong clientelist structures that link traditional patrons and clients prevent the participation of new political actors and, with it, the renewal of elites (Engler, 2016). On the contrary, studies that have explored the instrumentalization of public employment for clientelist purposes (Bøttkjær and Justesen, 2021; Robinson and Verdier, 2013) are important. In this case, it has been identified that clients can receive not only posts in public office as compensation for their support, but also the right to earn their own corrupt income through the use of mechanisms like bribes and the blackmail of citizens, which replaces the classical ‘extraction of the state by extraction through the state’ (Trantidis and Tsagkroni, 2017).
Cluster 3. Clientelism and Latin America
Latin America, being one of the scenarios par excellence that reveal the scope of the relationship between clientelism and democratic construction processes, has remained a recurring object of cutting-edge research on this phenomenon from a diversity of approaches and perspectives (Hilgers, 2012; Magaloni et al., 2007; Nichter, 2010). From a methodological dimension, several investigations in the region have adopted ethnographic approaches to investigate the logics of social and political relations arising from clientelist transactions (Auyero, 2000; Hilgers, 2009; Szwarcberg, 2015). Although clientelism is a recurring characteristic in Latin American political systems, it is still a stigmatized practice, which is why this type of methodology has allowed the establishment of bonds of trust between researchers and actors involved in the phenomenon, allowing the development of studies with significant scientific validity (Gonzalez-Ocantos and Oliveros, 2019).
Some studies that have marked the evolution of clientelism in Latin America have investigated the characteristics of the voters whom the political machines aim to recruit as clients. Calvo and Murillo (2004), Stokes (2005), and Szwarcberg (2012) have shown that, analyzing the Argentine case, the efforts of the political parties are prevailingly directed toward capturing undecided voters. However, Nichter (2008), through empirical research, moves away from this hypothesis, pointing out that, on the contrary, political machines in the Argentine case show a preference for voters with whom there is an established empathy, but who have a propensity for electoral absenteeism.
On the contrary, Latin America has been the subject of important research that links the triad of clientelism, violence, and criminality (Gay, 2012; Hilgers et al., 2017; Sarkar and Sinha, 2022). Along these lines, the Colombian case stands out because the processes of decentralization of the political system have promoted the transfer to the territories of political and economic resources that become the spoils of war for armed groups, generating clientelist relationships between them and citizens that are mediated by violence (Eaton, 2006; Moncada, 2016; Osterling and Sanin, 2020). Similarly, the presence of drug trafficking in the region has promoted strong cooperative links between the cartels and subnational politics, generating fluid clientelist relationships between political authorities, traffickers, and citizens, in a framework of profound deinstitutionalization (Darién, 2021; Rebolledo and Villegas, 2010; Singh, 2019; Velasco et al., 2018).
Another interesting dimension in Latin America is its relationship with rural dynamics. The structural abandonment of the countryside in the region has promoted strong dependency relationships between patrons and clients, where the former tend to view the latter as easily corruptible actors due to their vulnerable situation, as Ansell (2018) and Sabourin (2020) show in the case of Brazil, and Landini (2013) in Argentina. On the contrary, findings such as those of Diamond (2021) in Colombia reveal that in some rural areas of the country, it is the clients who donate small resources to the campaigns of their patrons, thus establishing a reciprocal relationship based on expectations of the retribution of favors once the electoral objective has been achieved. Finally, analyzing the Mexican case, Garay et al. (2020) offer interesting contributions regarding the growing preference of citizens for community associativity as a mechanism to access social programs, to the detriment of clientelism in both rural and urban settings.
Cluster 4. Clientelism and Africa
In addition to the previous axis, the African case appears in the center of gravity within studies on clientelism. The 1990s are considered the quintessential period of transition toward democracy in Africa, specifically in the countries of the sub-Saharan region (Lynch and Crawford, 2011). However, clientelism as a common practice in most of the countries in the region did not arise with the wave of democratization; on the contrary, it was already a widespread characteristic in the authoritarian regimes that preceded this process (Pitcher et al., 2009; Van de Walle, 2007). Another distinctive element of the African case is a visibly more precarious level of economic and institutional development than in other regions of the world where clientelism is present, which determines a differentiated nature and evolution of the phenomenon (Kitschelt and Wilkinson, 2007a; Van de Walle, 2007; Van Ham, 2015).
Another factor that has drawn attention during decades of research on clientelism in Africa is its link to ethnicity (Berman, 1998; Bratton et al., 2012; Lemarchand, 1972; Yıldırım and Kitschelt, 2020). Although ethnic conditions appear, at first sight, to be a determining factor in the construction of political loyalties in the region, cases such as Ghana (Harding, 2015; Ichino and Nathan, 2013; Lindberg, 2013; Lindberg and Morrison, 2008) and Senegal (Koter, 2013) show that, surprisingly, this element does not significantly define voter support. However, the plurality of existing contexts in Africa reveals that the relationship between ethnicity and clientelism becomes more complex according to the variables analyzed. Along these lines, Isaksson and Bigsten (2017) revealed in a study of 25 African countries that the predominant ethnic composition in the area of residence of voters is more relevant than the ethnic affiliation of individuals when establishing clientelist relations. Likewise, Kramon (2019), in a study that collected samples in 15 countries, identified that in the presence of hierarchical and centralized leaderships of ethnic groups, non-clientelist structures of voter mobilization tend to be generated, but in their absence, clientelism emerges from the need to establish links through individualized approaches.
Studies like that of Driscoll (2018) in Ghana, question the hypothesis that, in the presence of a competitive party system, the rate of democracy increases while clientelism tends to decrease. In this case, we analyze how volunteers see their ability to demand rewards increase in the face of the vulnerability that their parties experience in close elections. On the contrary, research like that of Sigman (2022) in Ghana and Benin explores the role of offering public positions beyond its function as an element of electoral mobilization, instead seeing it as a source of subsequent political financing through the control and appropriation by bureaucrats (clients) of public resources destined for elites (patrons).
It is also important to highlight recent works such as that of Jöst and Lust (2022), who, through the analysis of Malawi, Kenya, and Zambia, conclude that, the greater the social density, the greater the preference for electoral handouts, while expectations of receiving social services in the future decrease. Likewise, this study showed that, in the communities analyzed, the non-offering of gifts can be perceived as an announcement of the candidate’s incapacity to offer social services. Finally, studies such as those by Kao et al. (2022) in Malawi, and Blattman et al. (2019) in Uganda, prove that voters increasingly support candidates who base their campaigns on proposals for the offer of public goods, to the detriment of those with a clientelist profile. These studies even show that voters have learned to take clientelist handouts without this implying an effective change in their political preferences and their vote.
Future research agenda
Below, we present the research agenda that identifies the topics which, according to the cluster analysis, should be addressed in the future:
Conclusions
This work contributes to global research on clientelism by conducting an analysis of the origins, evolution, and trends of this area of knowledge. Through bibliometric tools and techniques, the mapping of the scientific production associated with the topic was shown. It was also possible to classify the most relevant documents using the metaphor of the tree of science and, finally, to identify the main approaches in which their research is currently framed, and in the future agenda.
The scientific mapping shows that 83% of the literature available in the WoS and Scopus databases is in English, 9% in Spanish, 3% in French, and the remaining 5% in other languages. This indicates that when conducting searches on clientelism, it is prudent to articulate these languages to not exclude research that may be relevant. Likewise, the predominance of American academia on the subject is clear, with the United States being the country that presents the highest levels of production. In this way, its researchers, institutions, and journals lead the list of publications, in both volume and reputation.
According to the metaphor of the tree of science, from ‘the roots’ the approach to clientelism as a concept has shown a predominant interest in political studies. However, a review of the methodological evolution in the study of clientelism leads us to the contributions made by anthropology and sociology at the beginning of the construction of this conceptual category. Thus, from ethnography, very close to anthropology, to the new quantitative trends adopted by political science, the methodologies applied to the study of clientelism have stood out for their interdisciplinarity, reflecting a constant dialogue between qualitative and quantitative approaches. In addition, comparative methodologies and case studies have also played an important role in the study of the phenomenon, as demonstrated by the numerous studies presented in this article about clientelism in Latin America, Africa, and Southern Europe.
On the contrary, the seminal documents show clientelism as a phenomenon almost exclusive to incipient, incomplete, and weak democracies and present mainly in Latin America, Africa, and Asia.
The subsequent evolution in the study of clientelism, as demonstrated by ‘the trunk’ of our analysis, has focused its attention on the dynamics of non-institutionalized relationships between political elites (patrons) and voters (clients), as well as on the role-played by intermediaries (brokers). It was identified that the mechanisms and resources, public or private, used to build loyalties between the actors are widely addressed in these studies: vote buying, offers of contracts, pork barrel projects, public jobs, and coercion, among others.
As a result of the scientific mapping, four main clusters were identified, to which a large part of the research on the subject belongs. The first ‘clientelism and democracy’ reveals important studies that link the relationship of clientelism to the quality of democracy and its institutions. Along these lines, it is important to note the growing attention given by research to the effects of clientelism on the role of the state as a guarantor of social well-being, being a direct cause of the generation of poverty and inequality traps.
In the second cluster, ‘clientelism and corruption’, it is possible to establish that, although it is difficult to establish an unequivocal relationship between the two phenomena, the presence of clientelism does allow one to infer why certain political systems are more vulnerable than others to the endemic affectation of corruption. In this cluster, it is important to highlight a growing interest in the private sector as an increasingly common actor in the promotion of clientelism.
The prevalent presence of clientelism in certain types of democracies has prompted a large amount of research based on case and area studies. Thus, the third and fourth clusters show a clear interest in the dynamics of clientelism in Latin America and Africa. In the case of Latin America, the third cluster shows numerous works on Argentina, Brazil, Colombia, and Mexico. Similarly, the links of clientelism with elements like rurality, drug trafficking, and the presence of solid criminal structures in this region have been an important part of the new research agenda.
In the fourth cluster, the literature on Africa shows that clientelism does not necessarily go hand in hand with democracy. On the contrary, it was already a practice present in authoritarian regimes prior to the democratizing wave in the 1990s. Another distinctive element in the research on this region is the scope of the ethnic dimension in the instrumentalization of clientelism.
Finally, it is important to mention relevant studies that have focused on the analysis of clientelism in southern Europe, the former socialist republics, and, especially, the countries of Southeast Asia.
These conclusions should be considered in the light of the limitations of our study. First, the main use of WoS and Scopus for the initial tracking of information may be a limitation because, although they are today considered the main databases globally, it is possible to find important journals and publications on the subject that are not incorporated into their indices, thus omitting them from this mapping. Therefore, other works in the same area may consider using alternative sources and databases in their tracking exercise to compare the results presented here. In addition, the use of other techniques, procedures, and bibliometric tools may be useful to compare the findings obtained in this work and, thus, to achieve a greater generalization of the results.
Footnotes
Authors’ note
In the case of the first co-author, this article is the result of his academic work as a full professor at the Universidad Militar Nueva Granada.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
