Talking about trees: the territorial classification of native forests in the Argentinian Chaco

SESs refer to those systems which include a human socio-economic component, interacting with an ecological or environmental component (Berkes et al 2000, Ostrom 2007, 2009, Preiser et al 2018). The SES framework stresses the embeddedness of the socio-economic component into the ecological component (Daly and Farley 2011, Daly 2015) (figure 1).

Zoom In Zoom Out Reset image size

At times the representation of SES further distinguishes between the social and the economic component, noting that the latter is embedded in the former (Folke et al 2016, Olsson et al 2017). The SES and related frameworks have enormous value from a heuristic perspective. However, its operationalisation has notoriously remained a challenge, particularly to capture the formation of institutions (Haller et al 2018). We suggest a novel way in which the SES approach can be operationalised.

HM is a method based on dialectics, which perceives reality in terms of interconnected processes (Engels 1892, Ollman 1993). Initially developed by Marx and Engels (Marx and Engels 1970), HM posits that history proceeds from material conditions rather than from ideas. The connections between HM and SES stems from the fact that a critical aspect of HM concerns the 'metabolic relationship' between human society and nature (Foster 2000, Saito 2017). According to HM, this relationship is trans-historical (i.e. it is a constant feature of human societies). However, its specific form at a certain point in time is the result of the historical (here intended as inherently social) and material (including necessarily nature) conditions under which society operates. These conditions include both matter/nature and social/historical relationships, in a way which resonates with the modern definition of SES (Preiser et al 2018). HM studies the historical processes leading to the emergence of a particular socio-economic configuration out of the interaction of its various parts, while noting that the 'whole' system constraints and selects the parts.

We use HM to open up the 'black-box' of the socio-economic subsystem within a SES. To this end we begin with an important passage from Marx's Capital, where, while discussing the role of technology, he notes 'Technology reveals the active relation of man to nature, the direct process of the production of the social relations of his life, and of the mental conceptions that flow from those relations (Marx 1990, p 493)'. The incipit of the passage clearly points at the fact that the socio-economic system is embedded in the biosphere. However, it is the rest of the passage that gives us an indication of what the constituent elements of the socio-economic subsystem could be. Technology affects social practices, which include not only production per se, but also, social reproduction and the way of thinking. A further elaboration (Harvey 2017) allows us to identify six 'moments' forming the socio-economic subsystem. They are technological processes, social relations, material production processes, daily life and social reproduction (i.e. material/economic processes), development of mental conceptions and institutional processes (i.e. cultural/institutional processes). It is important to stress how HM sees cultural/institutional processes emerging out of material/economic ones. However, this does not imply technological or economic determinism. In an important passage in the preface to A contribution to the critique of political economy Marx refers to the 'legal, political, religious, artistic or philosophic—in short, ideological forms in which men become conscious (p 12)' of the material economic processes (Marx 2018). Hence, any experience of the material/economic processes is influenced by the cultural and institutional ones. Bringing the two pieces together we can decompose the socio-economic subsystem into a set of material/economic processes, and a set of cultural/institutional processes. This allows us to operationalise the SES framework and study the formation of institutions.

We illustrate our framework in figure 2.

Zoom In Zoom Out Reset image size

The material/economic sphere engenders not only 'dominant' visions/cultures that reinforce the status quo, but also 'critical' ones. These contrasting visions reflect the way in which the material/economic practices are perceived by different activists, particularly in respect to issues of distribution of assets and output (Wood 1995). Both dominant and critical visions/cultures fight to become hegemonic and establish what Gramsci called 'common sense' (Gramsci 2014, Crehan 2016). Although a thorough discussion of Gramsci's contribution is beyond the scope of this article, it is still worth remarking how 'culture' and 'common sense', by their effect on institutions and on the material/economic sphere (and ultimately on the biosphere) play an important role in history. The effect of cultures/visions on institutions is not direct but mediated. Cultures/visions provide the 'core beliefs' behind which activists can organize and form coalitions capable of shaping institutions and social behaviour (Crehan 2016, Weible et al 2020, Winkler 2020).

We use fsQCA, a particular case of QCA. QCA is a set-theoretical method, specifically designed to deal with situations in which a small or moderate number of cases is available to the researcher (Rihoux and Ragin 2009, Kahwati and Kane 2018). A brief description of the method is provided in the supplementary information. Here we would just like to point out that this method, while preserving the richness of information typical of qualitative approaches, allows for the systematic identification of patterns via the use of set theory and Boolean algebra.

We are interested in explaining the proportion of the native forest in each province that has been assigned to the category of low conservation value (GREEN). From our theoretical framework we know that the explanatory factors include the environmental context, the material/economic structure and the existing visions/cultures about deforestation. With respect to the environmental context, we consider the extent of native forests in each province, at the time when the provincial regulations establishing the OTBN were approved (i.e. 2009–2013 as indicated in supplementary table 1), expressed as a percentage of the total province area (NATFOR). The underlying rationale is that provinces with larger forest areas are more likely to have a higher proportion of green in the OTBN maps, since they feel there is a relative abundance of natural resources which can be sacrificed to promote development (Gutiérrez 2017). With respect to the material/economic structure, our purpose is to capture those elements that, reflecting technology, production and reproduction processes and social relationships, relate positively to the outcome of interest. For technology and production, we consider the proportion of the province area occupied by agricultural units with clear delimitations (AUD). This type of agricultural units denotes capital-intensive production, as opposed to small-scale peasants who almost exclusively operate on lands without clear demarcation of limits. Regarding reproduction, we note how the diffusion of capital-intensive agriculture is also associated with declines in rural populations, as small-scale peasants and indigenous peoples are expulsed from the land (Mioni et al 2015). We therefore consider the proportion of the population in each province being characterised as rural (RURPOP). Concerning social relationships, the diffusion of capital-intensive agriculture brings with it the dissolution of patron-client or family-based work relationships, in favour of wage labour. We use the proportion of non-family workers (NONFW) in agriculture to capture the type of social relationships reflecting a certain material/economic configuration positively associated with the outcome of interest. The combination of these three factors allows us to determine to what extent the material/economic configuration is favourable to capital-intensive agriculture and thus likely to yield a high proportion of green in the OTBN maps.

Our theoretical framework suggests that a particular material/economic configuration will engender both visions/cultures that are organic to it and antagonistic to it. In the context of the Chaco ecoregion, it has been noted how two distinct visions/cultures with respect to the issue of the expansion of capital-intensive agriculture and deforestation exist: one, associated with large-scale producers, sees agricultural expansion in terms of development and poverty eradication; the other, held mainly by peasants and indigenous peoples, sees it mainly in terms of loss of cultural identities and lack of access to institutions (Zepharovich et al 2020). Recent research points to the existence of two opposed state-society coalitions, one pro and the other anti-deforestation, in the formulation of the OTBN (Gutiérrez 2017, Figueroa and Gutierrez 2018). Rather than accounting directly for the impact visions/cultures have on institutional configurations, we consider the presence and the strength of state-society coalitions that express the contrasting cultures/visions. Drawing on (Fernández Milmanda and Garay 2019a, 2019b), we assume the strength of the pro-deforestation (PRODEF) coalition to be directly related to the presence of large landowners. We therefore consider the share of the total agricultural area belonging to extremely large (i.e. larger than 20 000 ha) agricultural units. To determine the strength of the anti-deforestation coalition (ANTIDEF) we carry out an extensive literature search in five different databases in English and in Spanish (see supplementary information for details). We interrogate the databases to determine whether for each province: (a) anti-deforestation organizations existed at the time the regulation for the OTBN were discussed; (b) whether such organizations had a weak local base or (c) whether they developed a strong local base; (d) whether they had established strong links with local (provincial) institutions and/or (e) with national institutions. We expect anti-deforestation coalitions to be most effective at shaping institutions when they emerge out of a diffuse/popular base and when they display strong ties with both the national and provincial state apparatus. On this basis, we assign different scores to the ANTIDEF variable, in the following way. If no anti-deforestation existed (case (a)), then the ANTIDEF condition would take the value of 1. One additional point is assigned to the value of ANTIDEF for each additional criterion met (with the exception of criterion (c), which would give two additional points, thus assuming that having a strong base 'includes' having a weak base). As a result, the ANTIDEF condition can take a value ranging from 1 (if no anti-deforestation organization existed) to 5 (if one or more organizations existed, with a strong local base, connected to both local and national institutions). The literature search also allowed us to identify nine local experts who worked on the elaboration of the OTBN (see supplementary information). They were interviewed to validate both the PRODEF and ANTIDEF indicators for each province. The conditions and the outcomes used in the analysis are summarised in table 1, alongside the data sources.

We note that there are some differences in the timeframes of some variables (i.e. some variables refer to 2009–2013, while other to 2001 or 2002). The chosen data sources, however, represent the best available information. Moreover, the current analysis is not dynamic but relies on the variability across the different provinces. We believe that over the period 2002–2013 the interprovincial heterogeneity has not substantially changed. Such a belief is supported by the fact that, while there has been a substantial expansion of the agricultural sector across all the provinces, particularly over the period 2004–2009, the relative number of active agricultural enterprises in each province hardly changed between 2002 and 2013 (see supplementary information and supplementary table 6).

Before specifying the models and performing the analysis, the raw data are calibrated and transformed into fuzzy-set data. For space reasons, the calibration strategy is presented in the supplementary information.

After calibration, we construct a new variable representing the aggregate material/economic conditions (ECON) conductive to GREEN, by taking the average of the variables AUD, NONFW and non-rural population, denoted as ∼RURPOP (i.e. the ∼ denotes the negation or complement of a variable). This procedure has the purpose of reducing the number of possible logical remainders (i.e. theoretically possible combination of conditions that are not observed in practice) and mitigate the issue of limited diversity (Pahl-Wostl and Knieper 2014). We analyse the following model:

$$\begin{align}{\text{GREEN}} = f\left( {{\text{NATFOR,ECON,PRODEF,ANTIDEF}}} \right).\end{align} \tag{ 1 }$$

We do test whether any of the factors on the r.h.s. of equation (1) qualify as necessary conditions (see supplementary information for details). Necessary conditions should be dropped from the fsQCA analysis, which focuses on sufficient conditions (Rihoux and Ragin 2009). In line with standard practice, we also look at the conditions that can explain the complement of GREEN (i.e. non-GREEN written as ∼GREEN). Formally this is equivalent to analysing equation (2) below. All the analyses are performed with the fsQCA software:

$$\begin{align}\sim {\text{GREEN}} = & \,\,f\left( {{\text{NATFOR,ECON,PRODEF,}}}\right. \nonumber \\ & \left. {{\text{ANTIDEF}}} \right){\text{.}}\end{align} \tag{ 2 }$$

The truth table for equation (1) is presented in table 2. For the analysis of sufficiency conditions, due to space reasons, we only display the intermediate solutions (table 3) while the parsimonious and complex solutions can be found in the supplementary information (supplementary tables 3 and 4).

The solution paths in table 3 indicate that the large extent of native forests (NATFOR) in combination with either the absence of strong anti-deforestation coalitions (∼ANTIDEF) or with an economic base oriented towards capital-intensive agriculture and a strong pro-deforestation coalition (ECON*PRODEF) is conductive to the outcome GREEN. The large extent of native forests (NATFOR) is a core condition, as it appears also in the simple solution (see supplementary table 3). This is in line with previous research indicating that provinces with large forest areas perceive forests as abundant and not in need of protection (Figueroa and Gutierrez 2018).

The first solution path in table 3 (NATFOR*∼ANTIDEF), shows how the outcome GREEN occurs when NATFOR is combined with the absence of strong anti-deforestation movements (∼ANTIDEF). Formosa and Tucuman, among others, well exemplify this case. In Formosa native forests account for around 61% of the provincial area, the largest proportion in the whole Chaco ecoregion. The formulation of the OTBN in Formosa has not received much scholarly attention. However, it has been noted how the peasant movement—relevant anti-deforestation activists in Formosa—are 'weak and co-opted' (Fernández Milmanda and Garay 2019b). Our interviews with local experts confirm this. As a result, Formosa turns out to be the province with the largest share of native forest area declared as of low conservation value (GREEN), namely 74.25%. Some important caveats apply. The provincial regulation identifies two categories of GREEN. The first one, defined as a corridor where up to 20% of the area can be deforested (Torrella et al 2018) and a non-corridor area where deforestation can reach up to 60% (Vallejos et al 2021). Indeed, Formosa represents an exceptional case in that the OTBN has been formulated mainly by a team of ecologists based at the University of Buenos Aires (personal communication of one interviewee). The province of Tucuman has experienced significant deforestation in the past with most of the remaining forests falling in the Yunga ecosystem, situated on land too steep and/or in areas with too high precipitation to be of interest to agriculture (personal communication of one expert). Despite the past deforestation, Tucuman still has relatively high forest cover, which stands at about 40% of the provincial area. With respect to the presence of anti-deforestation activists, in Tucuman only two organizations have been mentioned: Pro-Yunga and Pro-Eco. The former is centred mainly around its founder, while the second is perceived as too radical. In both cases, the organizations do not emerge out of a diffuse base, although Pro-Yunga is reported to have connections to local institutions (Langbehn 2013, Langbehn et al 2016). Overall, the first solution path shows a good coverage (both overall and unique). Notice that Tucuman also appears in the second solution path (NATFOR*ECON*PRODEF). This path posits that alongside the core factor NATFOR, it is the prevalence of capital-intensive agriculture (ECON) and strong pro-deforestation coalition (PRODEF) that are conductive to a high proportion of GREEN. Agricultural units with clear delimitations account for a relatively high share of the provincial area, about 22%; non-family workers represent more than 24% of the agricultural workforce; about 80% of the population lives in urban areas. Overall, these factors indicate an economic structure in line with capital-intensive agriculture. Moreover, agricultural land ownership is quite concentrated. Units larger than 20 000 ha (representing 0.05% of total agricultural units) account for over 13% of the agricultural area, while units larger than 1000 ha (representing less than 2% of the total agricultural units) account for about 64% of the total agricultural area and units smaller than 1000 ha (representing about 98% of the total agricultural units) account for about 36% of the agricultural area. This suggests that the pro-deforestation coalition is likely to be quite powerful. San Luis shows a similar pattern, with about 23% of the forests classified as GREEN. The area of native forests is large, about 41% of the provincial area, the economic structure is oriented towards capital-intensive agriculture and land ownership is quite concentrated, with units larger than 20 000 ha (representing 0.7% of farms) accounting for over 20% of the agricultural area (Minaverry 2018). The second solution path also shows a good solution coverage, with Salta being the only province that is uniquely covered by this solution. Salta has 19.23% of its forest classified as GREEN. While this seems low compared to the 74.25% of Formosa, it is still above two thirds of the provinces. Salta has the third highest extent of native forest (NATFOR), covering about 53% of the provincial surface. Scholars describe the organized pro-deforestation activists in Salta (e.g. La Rural, ProGrano) as having strong ties to government and administration (Langbehn 2013). Land concentration captures this aspect quite well, with Salta scoring the highest among the Chaco provinces. Large farms (above 20 000 ha) represent about 0.5% of the agricultural units and account for over 32% of the planted area. The economic configuration (ECON) shows that capital-intensive agriculture is quite prominent in Salta. At the same time, Salta is peculiar as it is also characterised by the presence of a strong anti-deforestation movement. This aspect is discussed in the next subsection.

The truth table associated with equation (2) is presented in table 4. The parsimonious and the intermediate solutions, for the sufficiency conditions associated to equation (2), coincide, and are presented in table 5. The complex solutions are displayed in the supplementary information (supplementary table 5).

The solution paths in table 5 indicate that strong anti-deforestation coalitions (ANTIDEF) or the small extent of native forests in combination with an economic structure based on extensive agriculture (∼NATFOR*∼ECON) can explain ∼GREEN. The province of Cordoba, with no GREEN area in its OTBN, exemplifies the first condition well. Cordoba has relatively strong anti-deforestation activists, including both local and international ones (e.g. Greenpeace), (Gutiérrez 2017). The Cordoba peasant's movement (MOCACO) and the representative of the Centre for Human Rights and the Environment (CEDEHA) were important players in the provincial committee for the elaboration of the OTBN (COTBN) and achieved a strong mobilization in favour of the protection of native forests (Cáceres 2015). Through their involvement in COTBN, these activists had strong links with the state apparatus. It is worth noting that while Cordoba has no GREEN areas, anti-deforestation activists heavily criticized the OTBN law 9814/2010 for allowing activities that could lead to deforestation in red and yellow zones, contravening to the requirements of the NFL (Silvetti 2013, Gutiérrez 2017). The province of Salta also appears in the first solution path. However, while Salta is characterised by the strong presence of anti-deforestation activists, the share of GREEN in the OTBN map is relatively high (and not low, as one would expect from the strong presence of anti-deforestation coalitions). The elaboration of the OTBN in Salta has been highly conflictual (Seghezzo et al 2017), due to the significant presence of peasants and indigenous peoples, alongside capital-intensive agricultural units (Gabay and Alam 2017). This has resulted in the province having a considerable area of yellow (medium conservation)—also termed 'social yellow' (Barboza et al 2020).

The second solution path (∼NATFOR*∼ECON) implies that provinces with small forest cover (∼NATFOR) and an economic structure oriented towards extensive agriculture (∼ECON) will probably approve OTBN with low proportion of GREEN. The provinces of Catamarca and La Rioja, where forest cover stands at about 24% and 11% of the total area respectively, exemplify this path well. The proportion of native forests classified as GREEN is relatively low in both provinces, standing at around 12% in Catamarca and less than 4% in La Rioja. Both provinces are characterised by an economic structure prevalently composed of small producers.

As already mentioned, the current analysis rests on the assumption that the heterogeneity across the provinces has not changed over the period 2002–2013. This assumption is well supported for the factors reflecting ECON and PRODEF, as the relative importance of the agricultural sector in the various provinces hardly changed over this period. However, we also note that an expansion of the agricultural sector took place in the Chaco ecoregion. We expect that such a development makes the case for the realization of the outcome GREEN sharper.

Summing up, HM is based on dialectics and as such it understands that the 'truth is the whole': a system emerges out of the interactions of its parts (Levins 2008). When applied to the study of a SES, HM provides important insights. First, it sheds lights on what the different parts of the system are and how their interaction does matter. Institutions should not be studied in isolation but in conjunction with the larger environmental, material/economic and cultural context. Secondly, HM posits that material/economic processes engender cultures/visions and institutions. It is not a mere chance that in the provinces with an economic structure oriented towards capital-intensive agriculture, like Tucuman for example, we also find strong pro-deforestation coalitions. Thirdly, HM emphasises the existence of contradictions, reflecting distributional issues, which can provide leverage points to transform the system. Lastly, HM explicitly recognises that cultures/visions can shape institutions and in turn impact on the material/economic sphere. Salta, again, is a good example. Here most of the native forests have been classified as of medium conservation value (yellow). Although this result is often criticised as being insufficient, one must appreciate how it follows from the mass mobilisation of peasants, indigenous peoples and civil society in a province characterised by a strong capital-intensive agricultural sector.