Introduction
Popular explanations of poverty – or so-called poverty attributions – have been researched extensively, as they are part of the cultural values that structure the design of social security systems and determine the legitimacy of welfare programs for the poor (Lepianka et al., Reference Lepianka, van Oorschot and Gelissen2009). The longstanding literature on poverty attributions traditionally focusses on a three-tier model, distinguishing between structural, individualistic, and fatalistic explanations (Feagin, Reference Feagin1972). Many studies examining popular views on poverty have sought to confirm this dominant model across Western societies. However, there has been a growing recognition of more nuanced perspectives on poverty attributions. In particular, scholars point to the relevance of alternative explanations (Lepianka et al., Reference Lepianka, van Oorschot and Gelissen2009) and to important heterogeneity across countries and regions (van Oorschot and Halman, Reference van Oorschot and Halman2000; Da Costa and Dia, Reference Da Costa and Dia2015). As it has become clear that a one-size-fits-all approach to poverty attribution is far from optimal, there is an increasing focus on uncovering which types of explanations are salient in specific contexts.
This paper investigates poverty attributions in Guyana, which is an English-speaking developing country in the mainland of South America. This is relevant to broadening the scope of explanations of poverty to contexts with different socio-economic realities and cultural orientations compared with those usually obtained in Western countries. Guyana is particularly interesting, as the country faced a number of economic reforms, a deep economic and social crisis, and relatively high and rising poverty rates in the last decades of the 20th century (Gafar, Reference Gafar1998, Reference Gafar2004). Guyana went from a market economy in the 1960s to a state-controlled economy with nationalised industry in the 1970s and 1980s, and back to a market-oriented economy in the face of the economic crisis in 1988 (Gafar, Reference Gafar1998). In 1980, the poverty rate in Guyana was estimated to be 26% (Gafar, Reference Gafar1998). However, the Guyana Poverty Reduction Strategy Paper indicates that the poverty rate was 43.2%, 36.3%, and 36.1% in 1992, 1999, and 2006, respectively (Government of Guyana, 2011).
Like other Caribbean countries, Guyana’s poverty is predominantly concentrated in rural areas, particularly in the interior regions inhabited by Indigenous (Amerindian) peoples (Bowen, Reference Bowen2007). Among its poor, self-employment in the agricultural or labour-intensive work is the primary source of income. The level of education is quite low, with less than 15% of household heads having completed secondary education (Government of Guyana, 2011). Guyana is a diverse nation with six main ethnicities: including Afro-Guyanese (29%), Amerindians (11%), Indo-Guyanese (40%), Chinese (0.2%), Portuguese (0.3%), and White (0.01%) (Bureau of Statistics of Guyana, 2016). Historical ethnic tensions, especially between the two largest groups (Afro-Guyanese and Indo-Guyanese), have significantly impacted the nation’s political, economic structure, and social structure. This has resulted in uneven wealth distribution, often contingent on the ruling party (Edwards, Reference Edwards2017). Over time, the country has also experienced a rise in income inequality, with Indo-Guyanese making progress in this regard at the expense of other groups since the implementation of structural adjustment policies (Constantine, Reference Constantine2017).
Guyana was recategorised as an upper middle-income country in 2015. If the new category is used to determine the poverty head count, it would mean that approximately 61% of the population lived in poverty in 2006 and by 2019, this number stood at 48% (World Bank, 2022). Poverty reduction has thus been an important issue in public affairs in Guyana for many decades. This has most likely left its mark on public opinion and increases the relevance of popular explanations to better understand the perceived legitimacy of past and future poverty reduction policies (van Oorschot and Halman, Reference van Oorschot and Halman2000). To study poverty attributions in a non-Western context, this paper employs a factor analytic framework allowing identification of potentially relevant alternative attributions of poverty. In contrast to previous studies, there is no forced-choice between types of poverty attributions, which could potentially reveal a wider range of relevant explanations beyond the three-tier model. Moreover, our study aims to identify key variables with explanatory capability for attributions of poverty and test whether previously identified and popular explanatory mechanisms operate equivalently in a different cultural and economic context. This approach addresses two shortcomings in the existing literature. First, previous research identified several explanatory frameworks, such as socio-demographics, awareness of poverty, experience of poverty, values and beliefs, and national context (Lepianka, Reference Lepianka2007), but rarely tested them simultaneously. Second, most studies focus on Western societies, raising the question of whether and to what extent their explanatory utility is context-specific.
This paper takes into account three particular sets of determinants: socio-economic and demographic characteristics, feelings of resentment, and ideological values. As the most common factor, several studies have shown that the well-off, for instance, tend to harbour more individualistic beliefs, while low-status groups tend to blame society for their hardship (Bucca, Reference Bucca2016). However, self-interest has limited explanatory capability for the poverty explanations individuals adopt (Kluegel and Smith, Reference Kluegel and Smith1986; Cozzarelli et al., Reference Cozzarelli, Wilkinson and Tagler2001; Nasser et al., Reference Nasser, Singhal and Abouchedid2005; Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010). Alternatively, subjective evaluations and feelings of disadvantage play a potentially crucial role (Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010; Hunt and Bullock, Reference Hunt, Bullock, Brady and Burton2016). Nonetheless, with a few exceptions, the influence of experience of insecurity and injustice on poverty attributions has not been systematically investigated. To explore the role of social experiences, we focus on three interrelated feelings of resentment, that is, economic insecurity, relative deprivation, and powerlessness, to explain attributions (cf. Abts, Reference Abts2012; Van Hootegem et al., Reference Van Hootegem, Abts and Meuleman2021; Abts and Baute, Reference Abts and Baute2022). Finally, we investigate the impact of values and ideas, as attributions are not neutral, but embedded in broader belief systems that give meaning to social events (Feather, Reference Feather1985; Lepianka, Reference Lepianka2007). To consider the multidimensionality of these ideas, the role of egalitarianism, left–right self-placement, and beliefs on social inequality are considered. Analysing survey data of the Values and Poverty Study in Guyana (VAPO Guyana), we test which explanations of poverty can be retrieved in a non-Western context and how self-interest, resentment, and values affect these different types of poverty attributions.
Theoretical framework
Expanding the traditional poverty attributions
The traditional three-tier model of poverty attributions distinguishes among individualistic, structural, and fatalistic explanations: (1) individualistic attributions assign responsibility to the poor themselves and link destitution to dispositions and behaviours of the impoverished, like lack of effort and loose morals; (2) structural explanations blame social and economic forces external to the individual assigning responsibility to structural inequalities in society, low wages, lack of jobs, absence of schools to educate the population, prejudice, and discrimination; and (3) fatalistic explanations assume that poverty is beyond the control of both individual and society as, the poor have bad luck and lack the ability to influence their own destiny (Feagin, Reference Feagin1972). The validity of this three-dimensional model was upheld by many researchers over time. Several studies confirm the individualistic, structural (e.g., Furnham, Reference Furnham1982; van Oorschot and Halman, Reference van Oorschot and Halman2000; Cozzarelli et al., Reference Cozzarelli, Wilkinson and Tagler2001; Nasser et al., Reference Nasser, Abouchedid and Khashan2002; Sales, Reference Sales2002; Bullock, Reference Bullock2004), and fatalistic dimension (Feather, Reference Feather1974; Morçöl, Reference Morçöl1997; Nasser et al., Reference Nasser, Abouchedid and Khashan2002), while some have made use of the framework without assessing it (e.g., Bastias et al., Reference Bastias, Cañadas, Figueroa, Sosa and Moya2019; Özpinar and Akdede, Reference Özpinar and Akdede2022). However, some authors argue that the three-tier model is not always entirely supported by the data and might not be exhaustive as it fails to reflect the ambiguity of public perceptions of poverty. For this reason, they propose to recognise more complex and extensive views on poverty attributions (e.g., van Oorschot and Halman, Reference van Oorschot and Halman2000; Lepianka, Reference Lepianka2007; Lepianka et al., Reference Lepianka, van Oorschot and Gelissen2009).
Specifically, the three-tier model has been improved and complemented in two distinct ways. First, an important refinement entails the recognition of (additional) subdimensions of the three traditional types of poverty explanations. For instance, Morçöl (Reference Morçöl1997) identifies two individualistic attributions, two structural explanations, and one fatalistic dimension. The two individualistic dimensions differentiate between perceived personality flaws and target behaviours, while the structural attributions distinguish between abstract factors related to the socio-economic system and income distribution, and tangible factors concerning employment and lack of education, which focus more on concrete life experiences. Similarly, Bolitho et al. (Reference Bolitho, Carr and Fletcher2007) also reveal five dimensions that encompass one individualistic, one fatalistic, and three types of structural attributions. Specifically, the structural dimensions differentiate among poverty due to conflict, third world governments, and international exploitation. This illustrates that rather than the existence of a single fatalistic, structural, and individualistic dimension, the three types of attributions can be disaggregated into several meaningful subdimensions.
Second, the model is improved by identifying more complex and hybrid explanations of poverty that draw from more than one of the traditional explanations (Hunt and Bullock, Reference Hunt, Bullock, Brady and Burton2016). There are, for instance, hybrid poverty attributions that not only capture the presence or absence of agency (blame vs. non-fate) on one axis, but also the nature of the responsibility (individual vs. collective/social) on a second axis (Lepianka, Reference Lepianka2007). Combining the resulting quadrants, some authors make distinctions between perspectives of individual blame and societal blame on one axis, and individual fate and social fate attributions on the other (van Oorschot, Reference van Oorschot2000; van Oorschot and Halman, Reference van Oorschot and Halman2000). In this context, van Oorschot and Halman (Reference van Oorschot and Halman2000) highlight that the individual-social (structural) axis was already present in the pioneering work on poverty attributions, and that individual blame versus individual fate distinctions were also discernible. However, these dimensions were seldom empirically combined to yield four quadrants of poverty attributions that combine and go beyond the traditional classifications. For instance, the individual fate perspective (which perceives poverty to be a result of unfortunate circumstances) focusses on the individual but goes beyond individual’s control, which differs from the conventional explanation attributing poverty to a lack of thrift. Similarly, the social fate perspective, considering poverty as being an unavoidable facet of contemporary life, assigns responsibility to macro processes beyond individual control, thus combining elements of traditional structural and fatalistic explanations (see van Oorschot and Halman, Reference van Oorschot and Halman2000). In addition, some studies identify cultural explanations of poverty, which draw on more than one of the typical dimensions. Cultural explanations put the responsibility, for example, on the breakdown of the nuclear family and on being born into poverty (Cozzarelli et al., Reference Cozzarelli, Wilkinson and Tagler2001; Lepianka et al., Reference Lepianka, van Oorschot and Gelissen2009; Terol-Cantero et al., Reference Terol-Cantero, Gelabert, Costa-López and Vázquez-Rodríguez2023). As such, this explanation holds poverty as self-inflicted, but also as caused by the social surroundings of the individual and the values being passed on from generation to generation. These more diverse attributions that build upon more than one of the traditional explanations of poverty demonstrate the importance of going beyond the ideal types of the three-tier model.
Building on these two aforementioned contributions to the original three-dimensional model, we expect that six dimensions of attributions might be supported. These include individual blame, social structure, economic structure, social fate, individual fate, and family. Individual blame is prominent in most poverty attribution research. Social and economic structure are refinements of structural explanations, drawing upon the differentiation between tangible (economic structure) and abstract (social structure) explanations, as elucidated by Morçöl (Reference Morçöl1997). The conceptualisation of dimensions related to social fate and individual fate takes inspiration from the work of van Oorschot and Halman (Reference van Oorschot and Halman2000). Social explanations that incorporate a reference to fatalism are classified under the category of social fate, whereas those devoid of such fatalism fall under the umbrella of social structure. The inclusion of the family dimension is informed by the relevance of cultural explanations that tap into multiple traditional facets of poverty. However, it is worth noting that this choice to include a single dimension for culture and family is influenced by data limitations in this study. A cultural dimension could potentially subsume indicators pertaining to family alongside other factors (e.g., access to quality education and types of jobs available), as observed in the study by Cozzarelli et al. (Reference Cozzarelli, Wilkinson and Tagler2001). Such indicators could, potentially, be subdivided into separate subdimensions to measure distinct elements of culture. These expected dimensions of poverty attributions, expand the three-tier model by including multiple forms of individualistic (i.e., individual blame and individual fate), structural (i.e., economic and social structures), and fatalistic (i.e., individual fate and social fate) explanations and by identifying hybrid forms of attributions that combine multiple dimensions (i.e., family). However, this six-dimensional model stems from theoretical expectations (and data limitations in the case of family) and is not yet guaranteed to be empirically relevant in a non-Western context.
The forced-choice design that requires respondents to select from particular poverty explanations and which is regularly employed in research, is suboptimal when there is uncertainty in preferences. It could lead to discomfort among respondents and to the selection of options that do not necessarily reflect true views (Dhar and Simonson, Reference Dhar and Simonson2003; Lepianka et al., Reference Lepianka, van Oorschot and Gelissen2009). As an alternative intended to take this preference uncertainty into account, a factor analytic approach is adopted that attempts to capture underlying dimensions of individuals’ ratings of various explanations of poverty.
Regional and cultural heterogeneity in poverty attributions
An extensive body of research on poverty attributions illustrates that levels of support for various explanations diverge across Western societies. Support for individualistic explanations of poverty, for instance, appears to be ubiquitous in the United States to the extent that individual blame is considered essential of American ideology (Kluegel and Smith, Reference Kluegel and Smith1986). In addition, individual attribution is salient among the British (Furnham, Reference Furnham1983), Irish, New Zealanders (Hirshberg and Ford, Reference Hirshberg and Ford2001), and Luxembourgers whereas Canadians and Australians place less emphasis on blaming the poor (Reutter et al., Reference Reutter, Stewart, Veenstra, Love, Raphael and Makwarimba2009). System blame, in contrast, characterises West-European and Scandinavian countries: poverty is mainly considered the result of social injustice leading to greater support for social explanations (van Oorschot and Halman, Reference van Oorschot and Halman2000; Niemelä, Reference Niemelä2008; Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010).
Apart from the substantial regional variation across Western countries, there is important cultural diversification among developing countries. In addition to the presence of individualistic and fatalistic explanations, there seems to be especially strong support for the social blame perspective in developing countries (Payne and Furnham, Reference Payne and Furnham1985; Hunt, Reference Hunt1996; Hine and Montiel, Reference Hine and Montiel1999; Davids and Gouws, Reference Davids and Gouws2013). One reason why citizen’s attributions might be different in developing countries is that they often face harsher socio-economic realities (Feather, Reference Feather1974). However, apart from the economic conditions, it might also be related to variability in values, attitudes, and modes of causal attribution that are developed through diverging socialisation processes (Feather, Reference Feather1974; Furnham, Reference Furnham1982; Kreidl, Reference Kreidl2000). In addition, it might be connected to the presence of policies that especially protect the interest of the wealthy and suppressed anger from minorities or lower social classes towards the ruling regime, government institutions, and the structures that reproduce inequality (Kluegel and Smith, Reference Kluegel and Smith1986). It follows from the existing cultural heterogeneity that not all recognised and theoretically distinguished poverty attributions are relevant in all contexts and there is need for caution when generalising explanations across countries and especially between Western and non-Western contexts. As a result, we adopt a more exploratory approach aimed at determining whether these theoretically relevant attributions of poverty can also be retrieved in the Guyanese context, which is not inherently self-evident. Hence, we refrain from formulating explicit hypotheses, as it is very well plausible that some attributions might not effectively encapsulate people’s views on poverty in Guyana.
Explanations of poverty attributions
Poverty attributions have been related to a series of explanatory frameworks. Although they have been rarely tested simultaneously in a non-Western context, we link the relevant poverty attributions to three different theoretical frameworks, that is, self-interest, feelings of resentment, and ideological beliefs. Each of these frameworks could help clarify why certain groups are more likely to adhere to particular attributions of poverty.
Early work has considered self-interest as a main driver for individual’s positions towards poverty (Furnham, Reference Furnham1982, Reference Furnham1983; Payne and Furnham, Reference Payne and Furnham1985; Kluegel and Smith, Reference Kluegel and Smith1986; Hayati and Karami, Reference Hayati and Karami2005). Economically disadvantaged groups (women, minorities, working class, lower educated, and lower incomes) are expected to be more likely to ascribe poverty to structural inequalities instead of individual characteristics whereas economically advantaged groups (men, whites, middle/higher classes, higher educated, and higher incomes) should explain poverty in more individualistic terms (Hunt, Reference Hunt1996; Cozzarelli et al., Reference Cozzarelli, Wilkinson and Tagler2001; Nasser et al., Reference Nasser, Abouchedid and Khashan2002; Svallfors, Reference Svallfors, Rothstein and Steinmo2002; Yúdica et al., Reference Yúdica, Bastias and Etchezahar2021). These relations have mostly been explained by self-interest theory (Form and Hanson, Reference Form and Hanson1985; Hasenfeld and Rafferty, Reference Hasenfeld and Rafferty1989; Hunt, Reference Hunt1996), which holds that individuals will favour explanations that legitimise their welfare state (in)dependence. As such, the self-interest hypothesis suggests that the greater support for structural explanations by groups that are disproportionately poor stems from their interest in a type of attribution that justifies their larger dependence on social provisions. However, these relations could also be explained by a greater awareness of and exposure to poverty among certain groups, which might lead to adoption of more structural and external explanations (Lepianka, Reference Lepianka2007).
Whereas some studies report significant effects of socio-structural and demographic variables and lend support to self-interest theory, others report low or no influence at all and question the relevance of personal interest in itself as an explanatory framework (Kluegel and Smith, Reference Kluegel and Smith1986; van Oorschot, Reference van Oorschot2000; Cozzarelli et al., Reference Cozzarelli, Wilkinson and Tagler2001; Nasser et al., Reference Nasser, Singhal and Abouchedid2005; Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010). Some authors point to the relevance of “subjective” experiences of disadvantage to understand how welfare state attitudes and beliefs about the causes of poverty take shape (e.g., Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010; Abts, Reference Abts2012; Hunt and Bullock, Reference Hunt, Bullock, Brady and Burton2016). Generally, the role of these experiences can be framed from the “underdog perspective,” whereby the more individuals feel deprived and disadvantaged, the more inclined they are to adopt the perspective of the worst-off in society and to adhere to structural poverty explanations (Nilson, Reference Nilson1981; Lepianka, Reference Lepianka2007).
We focus on these experiences of disadvantage, specifically by exploring the role of resentment arising from structural inequities as a lens to comprehend attributions of poverty. In contrast to self-interest perspectives, resentment is strongly group-based and emphasises broader subjective grievances concerning the functioning of society. It encompasses sentiments of loss and injustice due to experienced disadvantages, often directed towards the elite and the other (Hoggett et al., Reference Hoggett, Wilkinson and Beedell2013). Triggered by economic, ethno-cultural and political grievances, resentment is a moral reaction to feelings of insecurity, injury, and injustice, manifesting as deep-rooted feelings of discontent, anger and frustration when individuals find themselves unable to openly voicing their grievances, and consolidating in attributing blame externally. In this study, we conceptualise resentment as a “structure of feelings” anchored in three pivotal components: status insecurity, group relative deprivation, and powerlessness (Abts, Reference Abts2012; Hoggett et al., Reference Hoggett, Wilkinson and Beedell2013; Van Hootegem et al., Reference Van Hootegem, Abts and Meuleman2021; Abts and Baute, Reference Abts and Baute2022). Since “we resent what we judge unjust” (Ure, Reference Ure2015, p. 3), both status insecurity and relative deprivation can breed resentment. Status insecurity pertains to a person’s loss of standing within society, while relative deprivation focusses on the position of that person’s ingroup relative to relevant outgroups, entailing a sense of undeserved inferiority and perceived violations of justice norms: the ingroup feels neglected by others, while outgroups seemingly enjoy undeserved rewards (Barbalet, Reference Barbalet1992; Solomon, Reference Solomon and Schacht1994). They feel collectively and unfairly deficient, which may be linked to external blame attribution. However, the emergence of resentment is contingent upon chronic and persistent experiences of powerlessness (Demertzis, Reference Demertzis, Clarke, Hoggett and Thompson2006; Capelos and Demertzis, Reference Capelos and Demertzis2018). Beneath the surface lies a profound critique of the prevailing social system, including its economic structures. This critique is coupled with a yearning for transformative change and a call for accountability. Consequently, individuals consistently seek external targets to assign blame (Solomon, Reference Solomon and Schacht1994). Our argument posits that resentment, encompassing economic insecurity, relative deprivation, and feelings of powerlessness, fuels profound disillusionment with contemporary society and its establishments, holding them accountable for perceived injustices, including the manifestation of poverty (see also Rogenhofer et al, Reference Rogenhofer, Abts, Klein and Bertin2023).
Powerlessness, to begin with, includes a “learned, generalized expectation that outcomes of situations are determined by forces external to oneself” (Geis and Ross, Reference Geis and Ross1998, p. 233). Powerlessness coincides with an external locus of control, which encompasses the perception that actions, situations, and events are generally controlled by exogenous factors such as luck, fate, or the actions of powerful others (Rotter, Reference Rotter1966). This external orientation might spill over to the believed causes of poverty, as they make structural and fatalistic attributions that deny individual responsibility more likely (Heaven, Reference Heaven1989).
Group relative deprivation encompasses “a judgment that […] one’s ingroup is disadvantaged compared to a relevant referent” (Smith and Pettigrew, Reference Smith and Pettigrew2015, p. 2). The relatively deprived believe that other groups receive preferential treatment while their own group is portrayed as systematically and undeservingly disadvantaged. This structural feeling of injustice leads to blaming the established system, welfare state, and political elites (Abts, Reference Abts2012; Smith et al., Reference Smith, Pettigrew, Pippin and Bialosiewicz2012; Spruyt et al., Reference Spruyt, Keppens and Van Droogenbroeck2016; Rogenhofer et al., Reference Rogenhofer, Abts, Klein and Bertin2023). Consequently, the government could also be held responsible for the disadvantaged situation of those at the lower end of the societal ladder, thus being more likely to favour structural explanations.
Status insecurity is defined as the fear of losing the acquired living standard and the position within the economic hierarchy (Castel, Reference Castel2003). This insecurity is accompanied by a strong anti-establishment rhetoric, and a broader loss of hope and power (Hoggett et al., Reference Hoggett, Wilkinson and Beedell2013; Standing, Reference Standing2014). This experience could hence consolidate in system blame and fatalistic explanations of poverty. In general, we hypothesise that the more resentful in terms of powerlessness, relative deprivation, and status insecurity, the less inclined individuals are to attribute poverty to individual responsibilities and dispositional deficiencies, and the more likely, they see poverty as a consequence of external and structural determinants.
Beyond self-interest and feelings of resentment, poverty attributions do not arise in an ideological vacuum but are part of wider belief systems that provide an interpretative framework and give meaning to societal events (Feather, Reference Feather1985; Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010). To recognise the multidimensionality of these ideological dispositions, multiple beliefs and values are linked to poverty attributions. As a commonly used operationalisation of the ideology framework, the impact of left–right political orientation is investigated. It is long-established that right-wing or conservative individuals are generally more likely to adopt individualistic explanations and to blame the poor (Pandey et al., Reference Pandey, Sinha, Prakash and Tripathi1982; Lee et al., Reference Lee, Jones and Lewis1990; Özpinar and Akdede, Reference Özpinar and Akdede2022), as they adopt more conditional notions of solidarity, are more likely to accept social inequality and hold more negative attitudes towards the poor (van Oorschot, Reference van Oorschot2006; Weiner et al., Reference Weiner, Osborne and Rudolph2011). In addition to political ideology, egalitarianism is linked to the various poverty attributions. Those who prefer to make outcomes in society more equal are expected to be more likely to attribute poverty externally and to blame the system, because they are more supportive of targeting the poor as well as of redistributing vertically and thus see a collective responsibility in alleviating their fate (Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010; Gugushvili and van Oorschot, Reference Gugushvili and van Oorschot2020). As a final dimension, we focus not so much on ideas about the desired level of equality, but on beliefs about the wealth distribution in their society. People who perceive a higher degree of inequality are generally less tolerant towards inequality and perceive larger hurdles in obtaining social mobility, which should go hand in hand with more external or structural attributions (Davidai, Reference Davidai2018; García-Castro et al., Reference García-Castro, Rodríguez-Bailón and Willis2020).
Data and methods
Data
This study is based on data from the Values and Poverty Study in Guyana (VAPO Guyana), a cross-sectional, face-to-face survey conducted under the supervision of the University of Guyana and University of Ghent. Data collection was executed in two phases. Between May and June 2012, fieldwork was executed in the coastal regions, which account for approximately 90% of the population. The interior regions, which are less accessible, were surveyed between October and November 2013.
In the first phase, a two-step sampling procedure was used. Municipalities (or Neighbourhood Democratic Councils in the case of rural areas) were randomly selected with probability proportional to their size, and clusters of 12 households were identified within each selected area. A respondent of each household was then chosen following the birthday rule (based on the most recent birthday). This sampling procedure resulted 87 clusters within 51 municipalities, with a total of 1,048 individuals interviewed. In the second phase, which focussed on the interior regions, a similar sampling approach was employed. However, due to inaccessibility and sparse populations, communities instead of municipalities were used as sampling units. Seventeen communities were selected pragmatically. A total of 509 respondents were randomly selected and interviewed. To compensate for both oversampling of the interior regions and for gender nonresponse, the data are weighted using iterative proportional fitting.
Indicators
Dependent variables
The VAPO Guyana employed several five-point rating scale items (completely disagree–completely agree) to measure the poverty attributions to individual blame, social structure, economic structure, social fate, individual fate, and family. While the indicators of the individual blame explanation target specific behaviours of the poor, the proposed measurement of the social structure attributions includes two items that identify stigmatisation and exploitation as the social forces that cause poverty. The economic structure explanation focusses on the influences of levels of employment and wages, and the social benefits that might be available to individuals. The two fatalistic dimensions separate indicators that focus on the individual (individual fate) (e.g., lack of intelligence and talent, bad luck or disability, and punishment from God), from those that point to the social circumstances (social fate) (e.g., not having a voice, not getting the same chances, and poverty being an unavoidable part of modern life). The attribution to family concentrates on the breakdown of family and community life and on the inadequacy of familial support. The question wordings and percentages of respondents agreeing with each of these items are provided in Table A1 in the appendix.
Independent variables
Assets and education are included as explanatory variables of poverty attributions to operationalise the social structure. The assets variable was constructed from responses to seven items that focus on the presence of various assets in household. Respondents were asked whether a computer, washing machine, refrigerator, generator, bathtub, flush toilet, and a vehicle (or outboard motor) were owned by the household. The responses to each of these items were coded 1 = yes and 0 = no and the score for each individual was computed as the mean-centred sum over the items. This variable is used instead of household income to which there was much nonresponse. Education is divided into the following three classes: primary, secondary (reference category), and university. In addition, age, gender (0 = female; 1 = male), ethnicity (Indo, Afro, and mixed and others [reference category]), and location (rural, hinterland, and urban [reference category]) are included as explanatory variables.
Powerlessness, group relative deprivation, and economic insecurity are each measured by multiple five-point scale items (completely disagree–completely agree). The three indicators of powerlessness probe whether respondents feel incapable of influencing things that happen to them, whether they believe that they have no control on society, and whether they feel powerless. Group relative deprivation is measured by asking individuals whether they believe that their group is disadvantaged by the government, whether they feel systematically neglected, and whether they think the government does more for other ethnic groups. Economic insecurity is measured by three items which ask whether respondents think that their financial worries will increase, they will have difficulties maintaining their financial position, and their children will have more difficulties.
Values and beliefs are captured by three variables: political ideology, egalitarianism, and beliefs about inequality. Political ideology is measured as left–right orientation, which is measured by a single item on an eleven-point scale (0 = left; 10 = right). Egalitarianism is operationalised by three items that probe whether class differences should be smaller, income should be made more equal, and the government is responsible for reducing income differentials. A confirmatory factor analysis (CFA) on the three latent constructs of resentment and egalitarianism shows that the measurement model fits the data very well (χ2 = 90.177; df = 48; CFI = 0.998; TLI = 0.997; RMSEA = 0.024; WRMR = 0.728) and that despite some moderate indicator loadings for powerlessness, all factor loadings are sufficiently high (see Table A2 in the appendix for question wordings and factor loadings). For model simplification and to not include too many latent variables in the final model, factor scores of this measurement model are saved to be included as predictors in the structural equation models. To measure beliefs about inequality, respondents are presented with five diagrams, accompanied by verbal descriptions, of different “types of society.” This question asks which representation of how resources are distributed in Guyanese society best fit their opinion. The diagrams display: (A) a society with a small upper class and the majority at the bottom of society (reference category); (B) a society like a pyramid with a small upper class, more people in the middle and most at the bottom; (C) a society like a pyramid but with few people at the bottom; (D) a society with most people in the middle; and (E) a society with many people near the top and few near the bottom. For all variables used in our analysis, a summary of descriptive statistics is included in the appendix (see Table A3).
Statistical modelling
To analyse the data, we first employ CFAs to obtain an adequate measurement model for poverty attributions. All the models are estimated with the weighted least squares means and variances adjusted estimator (WLSMV) to take into account the ordered categorical nature of the rating scales used to collect the responses to the items. A six-factor measurement model for the poverty explanations is evaluated first and modifications are made as warranted. Following this, a structural equation model is estimated with all three explanatory frameworks included. The fit of this model (χ2 = 370.072; df = 142; CFI = 0.970; TLI = 0.947; RMSEA = 0.033; WRMR = 0.871) is adequate. To nevertheless determine the contribution of each group of variables to the explained variance, we also estimated stepwise models where the socio-demographic variables are included in the first model; the social experiences of resentment are added in the second model; and the ideological values are incorporated into a final model. However, only the coefficients of the final model are discussed in the results and the first two models are used to examine the incremental changes in the explained variance. All analyses are conducted using the Mplus software (version 8.3; Muthén and Muthén, Reference Muthén and Muthén2017).
Results
Salient poverty attributions
The initial six-factor poverty attributions model fits the data poorly (χ2 = 1,070.145; df = 89; CFI = 0.911; TLI = 0.879; RMSEA = 0.084; WRMR = 2.293; see Table A4 in the appendix for an overview of the fit of all measurement models). Large modification indices (greater than 100) for one indicator of individual blame (poor people are not motivated enough) and for one indicator of social fate (poverty is an unavoidable part of modern life) indicate that these items fail to discriminate among the included factors. In addition, the item of social fate has a low standardised loading (0.21), and thus, fails to adequately measure the latent construct. These items are dropped in subsequent models.
A second, revised, model with the two non-discriminating items excluded still fails to fit adequately with respect to the RMSEA (χ2 = 548.278; df = 62; CFI = 0.950; TLI = 0.926; RMSEA = 0.071; WRMR = 1.746). In this regard, it is noticed that the proposed individual fate factor is poorly recovered, while the item about bad luck and disability, and the one about punishment from God fail to discriminate adequately among the factors. Hence, ultimately, we drop the individual fate factor entirely from the model. This results in a good fitting five-factor model (χ2 = 191.123; df = 34; CFI = 0.981; TLI = 0.969; RMSEA = 0.054; WRMR = 1.106), which is deemed to be appropriate to measure poverty attributions in Guyana. Table 1, which displays the final measurement model illustrates that all factors have good convergent validity with average variance extracted exceeding 0.50. Furthermore, the factors also show adequate discriminant validity given that the square root of the average variance extracted (lowest value of 0.72) for each factor exceeds its correlation with each of the other factors. Apparently, Guyanese citizens do not prominently differentiate individual bad luck or misfortune as a key explanation for poverty, but only recognise individual blame as an option when individualising poverty. This observation does not mean that the individual fate dimension is irrelevant in all contexts or needs to be theoretically discarded. Instead, it suggests that it is not distinguishable or applicable within the context of Guyana.
Table 1. Factor loadings and correlations for the five-factor measurement model.
χ2 = 191.123; df = 34; CFI = 0.981; TLI = 0.969; RMSEA = 0.054; WRMR = 1.106.
From Table A1 in the appendix, which reports the percentage of respondents who agree or completely agree with each of the items, it is clear that especially structural and fatalistic explanations (values exceed 70%) are endorsed. Most statements of the social structure, economic structure, and social fate attributions are supported by almost 80% of respondents, which illustrates the overwhelming support for these explanations. This is in line with the strong support for the system blaming perspective in many of the previously investigated developing countries (Payne and Furnham, Reference Payne and Furnham1985; Hunt, Reference Hunt1996; Hine and Montiel, Reference Hine and Montiel1999; Davids and Gouws, Reference Davids and Gouws2013). In the case of Guyana, the popularity of social explanations might be related to the precarious economic situation and high poverty rate, as bad economic conditions seem to strengthen the belief that poverty is not self-inflicted. Indeed, the individual blaming perspective is adopted far less, as no more than about 30% agrees with the statements. The family attributions seem to hold a more moderate position, as it is supported by about half of all respondents. This intermediary position is in line with the family explanation building on elements of both structural and individual attributions.
Predicting poverty attributions
Table 2 presents the standardised coefficients for the structural equation models estimated. For the dummy variables, the coefficients are only standardised with regard to the dependent variables, as this makes the coefficients more meaningful. Furthermore, Table 2 presents only the results from the model that estimates the effects of all the predictor variables simultaneously, but we describe changes in the explained variance resulting from estimating models only with social structural variables and with experiences of resentment. This is done to determine the added explained variance of each explanatory framework.
Table 2. Structural equation model for the five types of poverty attributions (N = 1,503).
Fit final model: χ2 = 370.072; df = 142; CFI = 0.970; TLI = 0.947; RMSEA = 0.033; WRMR = 0.871.
Abbreviations: IB, individual blame; SS, social structure; ES, economic structure; SF, social fate; FA, family.
* p ≤ 0.05.
** p ≤ 0.01.
*** p ≤ 0.001.
Starting with the social structural variables that are used to measure the self-interest framework, the results indicate that older people are more likely to blame the social structure and to attribute poverty to social fate. This may be because they have been exposed longer to economic hardship, which makes them recognise the impact of these forces. Gender only has a significant influence on family attribution, whereby men are less likely to blame family breakdown for poverty. While primary education is associated with higher levels of individual blame and blame of the changing role of family, higher educational attainment is related to lower levels of these attributions. Despite contradicting the self-interest logic, it aligns with the enlightening effect of education (Kreidl, Reference Kreidl2000). The effect of socio-economic status in terms of the household assets confirms the self-interest theory, as possessing more household assets is associated with a stronger propensity for individual blame and with less support of social fate as an explanation for poverty (Svallfors, Reference Svallfors, Rothstein and Steinmo2002; Yúdica et al., Reference Yúdica, Bastias and Etchezahar2021; Özpinar and Akdede, Reference Özpinar and Akdede2022). Ethnicity is an important predictor: Indo-Guyanese are more likely to blame the individual and less likely to attribute poverty to the social structure. In contrast, Afro-Guyanese are less likely to blame the individual and are more likely to attribute poverty to social fate. As the Indo group was both the largest and the politically influential ethnic group during data collection, these results are consistent with self-interest theory (Form and Hanson, Reference Form and Hanson1985; Hasenfeld and Rafferty, Reference Hasenfeld and Rafferty1989; Hunt, Reference Hunt1996). Last, citizens living in rural regions are less prone to blaming economic and social structures, but are more likely to attribute poverty to family breakdown.
Examining experiences of resentment, the results reveal that powerlessness, to begin with, increases adherence to the economic structure and social fate frameworks. However, in contrast to theoretical expectations, it does not exert an influence on the individual blame attribution. These effects largely align with our expectations and provide support to the argument that those who are powerless adopt an external locus of control and, as such, relate social problems to external factors (Rotter, Reference Rotter1966; Heaven, Reference Heaven1989). On the other hand, group relative deprivation only fosters the fatalistic explanation. The perception that one’s group is systematically disadvantaged compared to other groups is associated with the belief that poverty is an inherent aspect of social fate. Those who feel like an underdog themselves are more inclined to adopt the perspective of the marginalised and hold overarching social processes accountable for the existence of poverty (cf. Nilson, Reference Nilson1981). Although this is in line with our expectation, the insignificant relation with all other attributions contradicts our theoretical predictions. In contrast to our expectations, economic insecurity has no significant relationship with any of the poverty attribution. This might be because economic insecurity constitutes predominantly an individual experience of disadvantage, one that might not necessarily inform judgements on society as a whole.
Shifting our focus to the impact of ideological beliefs, we observe that left–right placement only relates significantly to the economic structure attribution, whereby right-wing respondents are less likely to blame the economy. Surprisingly, the alignment does not predict any of the other attributions. It is, however, important to acknowledge that left–right placement in Guyana might not perfectly mirror its interpretation in Western countries, but could, for instance, only encompass differences in economic orientations. Egalitarianism emerges as a robust predictor, displaying significant associations with each of the poverty attribution types. In line with our predictions, individuals who strive for a more equal society are less likely to blame the poor themselves, but instead attribute poverty more strongly to social structure, economic structure, social fate, and family circumstances. Finally, with regard to the inequality beliefs, we observe that individuals who perceive the distribution of resources in society as pyramid-shaped are less likely to make attributions according to economic structure or social fate compared with those who believe that a sizable majority resides at the bottom of the societal hierarchy. This indicates that individuals who are generally more aware of poverty and perceive a large degree of inequality are less tolerant of it and perceive larger structural barriers to upward social mobility (Lepianka, Reference Lepianka2007; Davidai, Reference Davidai2018; García-Castro et al., Reference García-Castro, Rodríguez-Bailón and Willis2020).
As previously mentioned, to determine the incremental contribution of each of these frameworks to the explained variance of poverty attributions, we estimate two additional models: (1) a model exclusively including social structural variables, and (2) a model incorporating both social structural variables and experiences of resentment. The results are displayed in Table A5 in the appendix. The initial model, that includes only the demographic variables, explains 9.1%, 8.9%, 4.8%, 8.2%, and 6.3% of the variance in individual blame, social structure, economic structure, social fate, and family attributions. The other model, which adds experiences of resentment, yields a small increase (less than 3 percentage points) in the explained variance of social structure attributions. More substantial increases are observed for economic structure (an increase of 7.2 percentage points) and social fate (an increase of 9.9 percentage points). The final model (as displayed in Table 2 and discussed earlier) shows that the incorporation of ideological beliefs further increases the explained variances for social structure (by 7.2 percentage points), economic structure (by 12.5 percentage points), and social fate (by 7 percentage points). In total, 9.8%, 18.2%, 24.5%, 25.1%, and 9.4% of the variances in individual blame, social structure, economic structure, social fate, and family explanations, respectively, are explained.
From these explained variances, we can conclude that while the socio-demographic characteristics generally exhibit only a moderate explanatory power, they account for almost all of the explanatory power for individual blame and family structure attributions. Experiences of resentment are pivotal for explaining economic structure and social fate attributions, whereas ideological beliefs play a crucial role in explaining social structure, economic structure, and social fate poverty attributions – which are all external attributions. Overall, we can conclude that each of the three frameworks contributes to explaining poverty attributions, within a non-Western context. Furthermore, all three frameworks constitute relatively independent pathways for comprehending poverty attributions explanations. Notably, the values framework emerges as the strongest predictor among the three.
Conclusion
The conventional three-tier model of poverty attributions, which distinguishes individualistic, structural, and fatalistic explanations for poverty, has increasingly been recognised as overly restrictive, as it fails to capture more nuanced viewpoints that distinguish subdimensions or hybrid attributions. Moreover, this model is seldom scrutinised in a non-Western context. This study explores which poverty attributions are present in Guyana, a developing country in South America, and examines the factors underpinning these poverty attributions by linking them to socioeconomic and demographic variables, experiences of resentment, and values.
Our findings indicate that the traditional three-tier model inadequately represents the perspectives held by the Guyanese population. Instead, we unveil significant subdimensions within structural attributions, highlighting a differentiation between economic and social explanations. Furthermore, we identify a hybrid dimension that associates poverty with family breakdown, situated at the intersection of individual and social blame viewpoints. Additionally, we demonstrate that one type of attributions frequently found in Western countries, namely individual fate (van Oorschot and Halman, Reference van Oorschot and Halman2000; Lepianka et al., Reference Lepianka, Gelissen and van Oorschot2010), is absent in the Guyanese context. Thus, our study not only underscores the importance of extending the three-tier model beyond conventional explanations but also emphasises that distinct cultural contexts may give rise to alternative models of poverty attributions. Consequently, future research should consider alternative dimensions while also recognising the inherent cultural diversity in explanations of poverty.
Furthermore, we demonstrated that socio-economic status and ethnicity, for instance, exert influence on attributions in line with self-interest theory. Social experiences of powerlessness, group relative deprivation, and economic insecurity, which serve as indicators of experiences of resentment, also contribute to our understanding of poverty attributions. Specifically, powerlessness fostered several attributions, illustrating that most of the poverty explanations relevant to Guyanese society emanate from an external locus of control. Additionally, our study revealed that ideological values have the most substantial explanatory power for poverty attributions. However, we do observe a discrepancy between internal attributions (individual blame and family) and external attributions (social structure, economic structure, and social fate), with the latter being more effectively explained by the distinct frameworks. In sum, our analysis underscores that each of the frameworks plays an important role in explaining crucial aspects of these attributions, within a non-Western context, and these frameworks largely operate independently of each another.
Although there are evident structural constraints limiting the feasibility of certain types of poverty policies in Guyana, examining public opinions can offer insights into whether and which policies would be considered legitimate by the general population. Individuals who tend to attribute poverty primarily to individual behaviour for poverty will be less likely to support welfare interventions for the poor (Yúdica et al., Reference Yúdica, Bastias and Etchezahar2021), whereas those who emphasise social structures or fate are more likely to support state-driven interventions (Marquis and Rosset, Reference Marquis and Rosset2021). Furthermore, specific types of poverty policies may garner support depending on the attributed causes of poverty by individuals (e.g., Marquis and Rosset, Reference Marquis and Rosset2021; Özpinar and Akdede, Reference Özpinar and Akdede2022). For instance, a tendency to associate poverty with individual shortcomings might lead people to endorse policies focussed on job creation and skill development. Conversely, fatalistic attributions might foster support for direct assistance to impoverished individuals, while structural attributions could enhance support for policies targeting minimum income and free education (see Özpinar and Akdede, Reference Özpinar and Akdede2022). Considering the prevalent adherence to structural or external explanations in Guyana, there is likely a substantial support for the latter type of policies addressing these societal dynamics and hence aiming to alleviate poverty.
As education serves as an antidote against assigning individual blame to poverty, prioritising improvements in education could potentially result in a population that is even more open to policies aimed at assisting those in poverty. However, it is important to note that distinct perspectives are held by the largest ethnic groups in Guyana – Indo-Guyanese and Afro-Guyanese. These two groups have a history of ethnic tensions (see Edwards, Reference Edwards2017). This ethnic divergence appears to extend to their views on the root causes of poverty, thereby influencing their potential support for various poverty alleviation policies. Achieving consensus and garnering support from these major ethnic groups for the same anti-poverty measures would likely require a delicate balancing act by policymakers in Guyana. The findings regarding resentment also present a potential policy dilemma. On the one hand, experiences of resentment are frequently linked to heightened welfare state criticism (Abts, Reference Abts2012; Van Hootegem et al., Reference Van Hootegem, Abts and Meuleman2021), potentially hindering the effectiveness of poverty reduction policies. On the other hand, our results underscore that especially the feeling of powerlessness is associated with external attributions, which, in turn, correlate with greater support for government intervention (Marquis and Rosset, Reference Marquis and Rosset2021). Consequently, policymakers are confronted with a complex challenge: they must simultaneously work to mitigate experiences of resentment to empower individuals and enhance their internal efficacy, while also persuading them about the structural forces underlying poverty that necessitate effective poverty policies.
Although it offers valuable insights into poverty attributions and their primary determinants within a developing and non-Western society, this study is not without limitations. An important drawback, for instance, lies in the potential interpretation of two items (serving as indicators of social fate) as outcomes rather than causes of poverty, which could have affected the structure of the model. We therefore advance our findings about social fate as preliminary and encourage further research employing different items for a more robust exploration. Apart from this, dependence on the available items might have resulted in a model that overlooks other salient attributions. For example, we estimated a single factor for family attributions which might not fully encompass the potential cultural explanations of poverty. Consequently, the poverty attributions model developed may not provide the most comprehensive overview for Guyana. The challenge ahead lies in designing survey questionnaires that incorporate an even broader array of poverty attributions items, enabling a more comprehensive assessment of the relevance of various dimensions.
Competing interests
The authors declare none.
Appendix
Table A1. Question wording and percentage (completely) agreeing for each item of the poverty attributions.
Terms in grey are not included in the final measurement model.
Table A2. Questions wordings and factor loadings for items of the social experiences of resentment.
χ2 = 90.177; df = 48; CFI = 0.998; TLI = 0.997; RMSEA = 0.024; WRMR = 0.728.
Table A3. Summary of descriptive statistics of all variables included in analysis
Table A4. Model selection.
Table A5. Stepwise structural equation models for the five types of poverty attributions
Abbreviations: ES, economic structure; FA, family; IB, individual blame; SF, social fate; SS, social structure.
