In Chile, Law 20.390 of October 16th, 2009, incorporated a constitutional reform on Regional Government and Administration. However, few changes have occurred in practice: i) the creation of autonomous regional councillors (CORE) in 2013; ii) the Constitutional Organic Law on the Transfer of Competences (approved in 2014) iii) the creation of three new divisions in the Regional Governments (GORE), one of them being the Productive Development and Industries division (2016) and iv) the implementation of decentralisation in the allocation of funds to promote innovation and productive development (2015) with pilot experiments in three regions of Chile (Antofagasta, Biobío, and Los Ríos). As part of this, the Regional Productive Development Committee (CDRP) was created, a territorial development unit with powers to allocate funds for productive development with autonomy. The latter is the largest experience of decentralization in Chile, which is why it will be the subject of comparison in this research. The other thirteen regions, including Coquimbo, continued with the centralised system for allocating public funds. This allocation of funds for innovation and productive development to companies is made through deconcentrated institutions with offices in each of the regions named CORFO and SERCOTEC (deconcentrated entities). These elements will make the main difference between the cases to be considered: Coquimbo and Antofagasta.

There is in this experience an interest in decentralization and institutional change (creation of the CDPR), i.e., in fostering participation of local actors for promoting local and regional development. Decentralization and promotion of local and regional actors is linked to a certain kind of development: endogenous development. Given all this, we are interested in understanding better how endogenous development has been promoted in two case studies (Coquimbo’s and Antofagasta’s). We will use a qualitative approach evaluate in which of these cases endogenous development has been more achieved.

There are diverse approaches for local and regional development (Pike et al. 2017), from top-town to bottom-up approaches, including combinations of them. The first one suggests that local and regional development should be directed from the national level, while the second one emphasizes the need for the participation and strengthening of local actors. Pike et al. (2017) argue that no one of these is always better. Given the national interest shown by the above Chilean policies aiming at promoting regional development by means of strengthening local actors, we evaluate (via simulation) how such experiences have fostered endogenous development, a kind of bottom-up development.

On the other hand, the creation of the CDPR in the Antofagasta’s case represents a change of the governance, i.e., an institutional variation, as part of the effort for strengthening local participation. Though the effects of institutional variation in local governance for local and regional development is ambiguous and not well understood (not always effective as expected), as pointed out by Rodríguez-Poce (2020), well directed institutional changes towards decentralization are a promising strategy, and thus suggested, from the perspective of endogenous local and regional development.

Though no form of local and regional development is always the best, there are good experiences with endogenous development. An adequate institutional framework is necessary to implement public policies for innovation and technological development (Angelelli et al., 2006). In this context, it is important to emphasise regional development and decentralisation through resources such as the capabilities and potential of all the actors in each locality (Ochoa & Pilioneta, 2006). It is precisely the endogenous development approach that seeks to promote development based on regional actors and productive factors (Tödtling, 2020).

In the above backdrop, the main objective of this study is to compare the relationships between the actors linked to the regional productive development system of Coquimbo (non-decentralised) and Antofagasta (pilot region of decentralisation) to determine and measure the differences in the level of local cooperation in the allocation of public funds for innovation and productive development in two regions with different degrees of decentralisation. Though Coquimbo has a deconcentrated model: we will call it “centralised model”.

We study the behavior of the system by carrying out social simulations. (Axelrod, 1997; Squazzoni, 2012). We use the SocLab framework which has been used to the study of concrete and ideal forms of organization (see e.g. Sibertin-Blanc et al., 2006; El-Gemayel et al., 2011; Sibertin-Blanc et al., 2013a; Terán et al. 2015) on the basis of sociological inquiries. Pike et al. (2017) and Rodríguez-Pose (2020) suggest to learn from case studies in order to understand better the complexities of local and regional development. Social simulation is useful to comprehend better real situations by virtual experimentation of case studies, complementing real case studies. The present paper aims at generating additional knowledge about two case studies by virtual experimentation.

The results of the simulation allow validating that the pilot decentralisation model (Antofagasta) has better conditions of articulation and cooperation between local actors, that is, promotes endogenous development, by allowing a high degree of influence of local actors and presenting a better state of critical resources for this purpose.

The research presents evidence regarding the relevance of the decentralisation process in decision-making to promote and strengthen the endogenous development of the regions, through greater autonomy and coordination in the allocation of public funds for innovation and productive development, to promote the development of the regions with their own particularities by their local actors who best know the potentialities and shortcomings of their region. Additionally, this is pioneering research in terms of the method used, which was through simulation based on the theory of organised action systems.

The rest of this paper is organised as follows. The second section extends the theoretical framework of the paper given in the introduction. Section three describes the case studies. Section four examines the methodology, i.e., the SocLab simulation platform. Section five describes the model. The next section presents the results and their implications for regional endogenous development. The last section offers some conclusions.

2. Theoretical framework

Pike et al. (2017) examine diverse approaches for local and regional development, from top-town to bottom-up approaches, including a combination of them. The first one suggests that local and regional development should be directed from the national level in a centralized form, while the second one emphasizes the need for the participation and strengthening of local actors. We use a qualitative approach to evaluate the simulation models in terms of their contribution toward favouring endogenous development, a form of bottom-up development. Below, we will review Pike et al. (2017)’s and the endogenous perspectives of local and regional endogenous development, briefly comparing them. Additionally, we will refer to the qualitative and quantitative levels of approaching local and regional development.

3. The cases of study: Coquimbo and Antofagasta

The regions of Coquimbo and Antofagasta have been chosen to carry out this study, they are territories located in the northern macro-zone of Chile. The Coquimbo Region is characterized by having a diversified economy, standing out for its contribution to GDP, the sectors associated with mining (26.7%), personal services (12.5%), financial and business services (11.2%), construction and housing and real estate services (17.4%), also highlighting agriculture, fishing, tourism and commerce (CORFO, 2022). Antofagasta, for its part, is mainly based on mining activity linked to copper and lithium (52.6%), a sector that is not only the economic engine of the region, but also a fundamental pillar in the country's economy. Following in proportion to regional GDP are the financial and business services sector and the construction sector, both with a 10.3% share of the region's economy (Yáñez, 2019).

The main regional authorities in each region are three: i) intendants, ii) Regional Council (CORE) and iii) Regional Government (GORE). The Intendants, a senior administrative figure in the regional governments, must answer to the President of the Republic. This position is in charge of allocating the budget to the regional governments, as well as managing the investment funds. This is a position of trust of the president, therefore, if he does not comply with the task entrusted to him, he is removed from his position, generating uncertainty in the management of the region. The Regional Council (CORE) is a public collegiate body with powers over regional government and administration. They oversee effective citizen participation in the regional government. Since 2014, the councillors are elected by universal suffrage, in direct voting, for periods of four years, and can be re-elected. The councillors, together with the intendants, comprise the regional government (GORE).

This paper contrasts two approaches to regional development in Chile. On the one hand, the traditional centralist model (Coquimbo case). On the other hand, a form of decentralisation oriented towards regional endogenous development, creating a territorial development unit with certain attributions (CDRP in Antofagasta case).

In the case of Antofagasta, the approach to decentralisation of productive development was achieved in two ways: i) the legislative way that emphasised the creation of regional development services, and ii) the administrative way that proposed a model of decentralisation of national services or agencies through programmes. Under this structure, a hybrid of legislative and administrative approaches was created. This change gives political responsibility for productive development to the GORE (Regional Government) and leaves the execution of regional policies to specialised executing agencies called Regional Productive Development Committees (CDPR). This institution should define priorities for productive development at the regional level and the budget allocations necessary to implement regional policies in this area and play an articulating role in regional and national policies.

There are currently two models for the allocation of public funds for innovation and productive development in Chile: i) one in which resource allocation decisions are made by public institutions at the central level (CORFO and SERCOTEC), which, for this research, will be called the centralised model, represented by Coquimbo region; and ii) the regions where a CDPR has replaced the functions of CORFO and SERCOTEC in fund allocation decisions, which for this study, is the case of the decentralised model of Antofagasta.

CORFO and SERCOTEC are deconcentrated public services dependent on the Chilean Ministry of Economy, whose mission is to support the development of smaller companies through subsidies and technical assistance. In Coquimbo’s case, the administrative organization of the two main entities that support productive development in the territories depends on the Central Government and, as pointed out by Ferreiro et al. (2019, p. 79) , this reality implies that their deconcentrated public services with national coverage respond directly to the national directors and ministers located in the centre. On the other hand, in Antofagasta, while the CDPR (replacing local CORFO and SERCOTEC dependencies) is decentralized.

The criteria for decision-making of officials designated by the central government: CORFO and SERCOTEC in Coquimbo, as well as the Intendant (head of GORE) both in Coquimbo and Antofagasta, is importantly aligned with that of central government. Members of the regional deconcentrated units of CORFO and SERCOTEC’s decision making is very similar to that of members of these entities in Santiago - officials of both are designated by the central government -. On the other side, as the central government is supported by, and comes from, a certain political party or by a certain political coalition that usually includes several political parties, all these somewhat aligned functionaries could have certain political bias in relation to these political entities. This possible bias was implicitly captured or included in our models when identifying the relations and actors, as well as when defining the effect functions of the relations on the actors, the interest of the actors on the relations, and the solidarities between actors.

4. Methodology

The research methodology consists of virtual experimentation with a model developed on the SocLab platform. The SocLab framework (Sibertin-Blanc et al. 2006, 2013a, 2013b) has been developed into a formalisation of the Sociology of Organized Action (SOC) of Crozier and Friedberg (1980) . In SOC, the authors studied how social organisations are regularised, as a result of counterbalancing processes involving power relationships between social actors. Social organisations may include community interaction similar to that occurring in systems of regional development, as studied here.

In SOC, power relationships are based on the actors’ mastering of certain resources needed by others. This theory recognises that actors have bounded rationality, and that their environment is complex and shapes their behaviour (Simon, 1982). Consequently, the behaviour of each actor is strategic in the social game defined by their interaction with others. As a meta-objective, every actor adjusts his/her behaviour to obtain a satisfying level of capability to reach his/her goals, which is a balance of the aims assigned to his/her role within the organisation or social interaction, and of his/her particular objectives. The end of a social game is to reach a stationary state, i.e., a configuration where no actor has reasons to change his behaviour, i.e., where for all actors, the satisfaction of the actor is larger than his/her aspiration.

The management of a resource determines how it can be used by others so that it is the base of a relationship between the manager and the users of the resource. The dependency of an actor on a resource is expressed by its stake and effect function, as shown in Table 1 for our base model. The stake of an actor on a resource expresses how much it depends on this resource, on a scale null = 0, negligible = 1,… , significant = 5,… , critical = 10. Each actor has 10 stake points to distribute on the set of all resources.

The effect function of a resource on an actor expresses to what extent the management of the resource, by the actor controlling this resource, impedes or supports the actor in the achievement of its goals.

The impact of a resource r having the management m upon an actor a, i.e., the capability it provides to this actor to achieve its goals, is the value of the effect _{r, a()} function applied to m, weighted by the stake of a on r. Consequently, an actor gets some satisfaction, as an aggregation of the impacts it receives from the resources it depends on, weighted by the solidarities it gives to other actors, and, conversely, it exerts some influence, as an aggregation of the capabilities it grants to actors who depend on the resources it controls (Sibertin-Blanc et al. 2006, 2013a, 2013b). Formally, it can be said that, when a system of organised action is in a configuration m = (mr)_{r 𝜖 R}, the capability received by an actor a is defined as:

From this, we can define the satisfaction of a as (calling A the set of all actors):

Conversely, the influence exerted by actor a over actor b is defined as the sum of the capabilities it supplies to the actors, towards which b has solidarity by the management of the resource(s) it controls, that is:

Finally, the influence of a is the sum of the influence of a on all actors:

Actors seek to maximise their satisfaction. The linear relationships between the variables assumed in Equations (1)-(4) are explained in (Sibertin-Blanc et al., 2006). Firstly, the introduction of the linearity is in line with the Sociology of Organized Action (SOC) of Crozier and Friedberg (1980) : In all equations we added the value due to each involved actor or relation; for instance, in equation (1), the capability given to actor a per each relation r the actor manages is added (this particular capability is proportional to the stake of a on r and the effect of relation r on actor a), while in equation (2) the satisfaction that actor a receives from each actor b he has solidarity with is aggregated (the satisfaction a receives is proportional to the solidarity a has towards b, and the capability b achieves). Secondly, the operation of the model requires to normalise diverse variables, e.g., the sum of the actors’ stakes and then to attribute to each one the same amount of stake points to be distributed on the relations he participates. This normalization grants the same investment to each actor, i.e., the same possibility of personal implication in the game. It makes possible to quantify several concepts of the SOA by numerical values belonging to the same scale of values, and thus to compare them. For instance, the relevance of a resource could be estimated as the sum of the stakes placed by the whole population of actors on the relation supported by this resource, as those stakes reflect the importance of these relations for the actors (in section 5.2 the relevance of the resources in our model is given). Similarly, by calculating other variables (equations (1)-(4)) using linear relations, facilitates maintaining in a comparative scale these other variables of the model (satisfaction, etc.). The SocLab decision-making algorithm is described in Sibertin-Blanc et al. (2013b), and El-Gemayel et al. (2011) .

The output variables, e.g., influence and satisfaction, indicating the situation of the actors can help us in evaluating the degree of endogeneity of local and regional development in the simulated cases (Coquimbo and Antofagasta). Particularly, uneven values of influence: low values for some actors and high levels for other actors imply that the first actors have week power, low capacity to participate with relevant impact in the game, and likely low potential for autonomy, since the game depends on powerful actors, what means that the game is unfair (some actors are too week, while other are too strong). This is an unfavourable situation for endogenous development, given that it requires good participation and autonomy of all actors. Additionally, sustainability, a property of local and regional development in accordance to Pike et al. (2017) , and by the theory of endogenous development, requires effective participation of all actors. Thus, the distribution of influence of actors gives us a qualitative approach and criteria (following Pike et al. 2017) measuring and comparing endogeneity of the models representing two case study. Other similar criteria could be defined, e.g., based on the distribution of satisfaction of among actors.

5. Simulation Model

The information to develop the model was obtained through interviews conducted virtually through the Zoom platform, from May to June of 2021, that is, before the newly elected authorities (Regional Governors) took office and replaced their predecessors, the Regional Intendants. Since mid-July the highest authority in both regions analysed is the Governor. To include the vision of diverse actors (SERCOTEC, Intendant, etc.) the interviewed met the following requirements: i) work in CORFO, SERCOTEC or another government institution, ii) has held public office and iii) knows well the public system.

Four interviews with thematic script of guided open questions were conducted with two people in Antofagasta and two in Coquimbo. It was necessary to conduct the interviews in four moments or stages, each one having the following characteristic: i) the objective of the paper and the elements of the simulation model were explained.; ii) open questions were asked, and meaning and examples of key concepts, e.g., solidarities were given; iii) information was verified; and, iv) answers were validated with other external experts. Regarding point iii), validation, it was done through expert judgement (interview with two experts) with extensive experience in the public sector. Our way of collecting information was a variant of that presented in Sibertin-Blanc et al. (2013a) .

A model with four actors was elaborated. The first actor is in charge of the analysis and allocation of public funds for innovation and productive development. In the case of the Coquimbo Region, there are two autonomous agencies: CORFO and SERCOTEC, which we will call CorfoSercReg. For the Antofagasta Region, this role is fulfilled by the Regional Productive Development Committee -CDPR, created for this purpose within the framework of the decentralisation process previously described. The second actor is the Regional Council (CORE). The third actor included in the analysis is the Regional Government, (GORE). The fourth actor is the Central Government (CentralGob), which includes all the authorities that issue guidelines from the city of Santiago, where the government authorities are located. Each actor controls the following resources.

CORFO/Sercotec (Coquimbo’s case) - CDPR (Antofagasta’s case):

CORE:

GORE:

CentralGov: