Rural Districts and Business Agglomerations in Low-Density Business Environments. The Case of Extremadura (Spain)

Abstract

The strategy of the institutionalization and development of business agglomerations, in any of its analytical aspects (industrial district, local production system, cluster, etc.), has not had great results in Spanish regions with low business-density, probably due to the difficulty of finding an adequate implementation framework in administrative, geographic, and institutional terms. Based on the limitations presented by the identification methodologies of business agglomerations in low business-density territories, in this work we propose some methodological corrections that allow for reconciling these economic realities with the institutional and geographical framework offered by the local action groups (LAGs). This reconciliation is a useful tool to take advantage of the economies of agglomeration and, consequently, to explore the possibilities of endogenous development in rural areas, so that it can be a factor to take into account when planning and executing the public strategy of local and rural development. Finally, the results obtained for the specific case of Extremadura, the only Spanish region listed as a less developed one in European rural development policies, are presented.

1. Introduction

The local action groups (hereinafter LAG) have become the main tool of the European Union for structuring the local and rural development strategy [1,2], this being the reason why industrial or rural development policies in areas with low business-density, or rural areas, must consider them. In a way, they exemplify the open participation of the main economic agents with a presence in each territory at the county level, bringing community decisions on rural development closer to the rural territories of the member states. Conceived as a strategic tool, LAGs emerged with a dual function: on the one hand, they should be in charge of planning and channeling funds for the European rural development strategy in the territories, and on the other, they must contribute to the dynamism of the socioeconomic fabric of rural regions, directly attacking structural problems that affect them, such as depopulation and inequalities in living standards with respect to the urban environment [3] (p. 596), [4]. One way to face such challenges is by enhancing and optimizing the region’s endogenous resources [5] (p. 230), [6], wherein the correct definition of productive specialization seems crucial to us. In this sense, the tools offered by the theory of business agglomerations for shaping the LAG strategy cannot be ignored; as such, we consider its adaptation to the rural environment necessary.

“Business agglomerations” is a generic way of referring to the different terminologies that have been defined by the literature to define the grouping of firms around a certain territory (industrial districts, cluster, local productive systems, rural districts, quality agri-food districts,…). Each of these concepts presents its nuances, although they all start from the same premise: the concentration of companies that are dedicated to the same product or productive chain in a given territory. Broadly speaking, the industrial districts (IDs) [7,8] and, more generically, the local productive systems (hereinafter LPSs) [9,10], are socioeconomic realities that are based on taking advantage of the endogenous industrial growth capacity that certain geographical enclaves have, which constitutes an attraction factor that favors the location of companies and, consequently, the formation of specialized business agglomerations in a certain product or branch of activity, in rural areas that have set up a so-called rural district [11]. These realities generate a series of competitive advantages, allowing small and medium-sized companies, which by themselves would not have the financial capacity to invest in technology or to execute an internationalization strategy [12,13,14], to do so, being able to balance, through cooperation and agglomeration, the scale economies associated with large companies, in Chandlerian terminology [15]. Undoubtedly, this favors the generation of employment and income, allowing local and rural development [16,17,18,19,20,21,22,23,24,25,26] and, even if it is only for an arithmetic effect, regional development too [27,28,29,30,31,32]. Thus, the aforementioned concepts of ID and LPS have evolved towards newer and more recent theoretical notions such as the rural district (RD) or the quality agri-food district (AFD), more appropriate to the nature and characteristics of the rural regions and environments [33] (Legislative Decree No. 228 (18 May 2001) relative to the Italian normative), or even as “bio” districts [34,35].

Given the above, the main objective of this work is to evaluate the theoretical and practical lessons of business agglomerations and to facilitate their incorporation into the rural development strategy by LAGs, particularly with regard to the detection and identification of the endogenous productive capacities of the territories to which they are circumscribed, so that they can enhance the comparative advantages associated with them, and may also prioritize investments, allowing a better use of resources to achieve the objectives of income and employment generation and fixation of the rural territories. In summary, we seek to find the tool that allows one to localize business agglomerations into the LAG regions without giving up the postulates of the economies of agglomeration; that is, to locate municipal or supramunicipal business agglomerations with a capacity for generating incomes and employment and with influence and significance throughout the LAG region, so that they can be used as an economic engine for it, as well as being a focus for the attraction of new investment. To meet this objective, the text has been structured into four sections, in addition to this introduction. In the first, the reasons that in our opinion explain the poor practical development of theories of the Italian school of industrial districts in Spain, or at least their lesser degree of consideration compared to the Italian case when articulating rural development, are analyzed. In the second section, we reflect on various ID or LPS identification methodologies, and in particular, on their advantages and limitations when used in the LAG development strategy. In the third section, we propose some methodological adaptations that would facilitate, in our opinion, such use. Finally, in the fourth section, we outline the main conclusions of the investigation.

2. From Theory to Practice, from the Industrial District to the Rural District

This article arises from the authors’ conviction that in Spain, the enormous scientific and theoretical efforts that many regional researchers have made in the last two decades in the field of business agglomeration analysis are not translating into applied results in regions with low business-density (district effect [36,37,38,39,40,41], i-district effect [42,43], social capital [44,45,46,47,48]). As an example, and unlike what has happened in other nations, Italy is, without a doubt, a reference in this field, not only for the remarkable development of the existing research in this regard [49,50,51,52,53], but for the broad regulatory development that the industrial districts have had in this country, which are already a relevant element in industrial policy planning [54,55]. The creation of the National Observatory of Industrial Districts (http://www.osservatoriodistretti.org/ (accessed on 1 May 2020)) is clear proof of this, which denotes the institutional commitment decided by the promotion of this type of economic reality. There has not been a regulatory or institutional development in Spain that efficiently explores the potential of these agglomerations, and this has been the case even in the regions where the greater historical development of such industrial agglomerations has been evidenced, which have also been those on which scholars have focused most of the research efforts in this regard, namely, the Valencian region [39,56,57,58], Catalonia [59,60,61,62], and the Basque Country [59,63,64,65,66].

The previous reflection, which seems clear despite the fact that the elements that should serve as the basis for the inclusion of industrial agglomerations in the country’s industrialization strategy are known with some precision, is even more true if we refer to the agrarian field, where the whole path, including the scientific one, has yet to be covered. In this sense, at least three aspects seem relevant to us, which, if given their full value, would contribute to the better planning of productive activities in rural areas. The first one is the adaptation of the concept of business agglomeration to the reality that we find in agricultural environments. This aspect has already been partially resolved by the Italian school of industrial districts, having coined the concept of the rural district, whose theoretical specifications are assimilable to the rural agglomerations that we find in Spain and other Mediterranean countries [67,68]. In our opinion, this is crucial, since it determines, for example, the methodology to be applied for the identification and detection of these rural agglomerations, as well as in defining the tools to be used in their empirical analysis and in developing other not-yet-studied concepts, such as the so-called quality agri-food districts, which are also linked to a greater extent to the agrarian environment.

The second aspect to take into account is the absence of specific legislation that protects and develops these realities in rural areas. It should be noted that Spain has been applying legislation for a number of years to promote industrial districts under the name of innovative business groups ((hereinafter IBGs, http://www.minetad.gob.es/PortalAyudas/AgrupacionesEmpresariales/Paginas/Index.aspx (accessed on 1 May 2020)). These realities, which have already been analyzed in the context of Spanish industrial policy by Trullén and Callejón [69], bring together different forms of agglomeration, namely, industrial districts, value chains, knowledge-intensive activities and ICT-intensive activities, and tourism [70] (p. 380). In our opinion, this legislation, in its current formulation, is not adequate to link economic activity to the territory, something that should be a priority in the rural development strategy [71,72]. In fact, the need to have a sufficient critical mass to access the financing lines included in the regulations has led to the association of companies from different provinces and regions, so that the IBGs have ended up being institutions without a clear link to a certain locality or region [73,74]. The correlation between the detected business agglomerations and the IBGs existing in the Extremadura region is presented in

Table 1. Business agglomerations vs. innovative business groups (IBGs) listed in Extremadura, 2013.

Agglomeration (Sector and County)	LPS and ID 1	IBG 2
Agri-food in Don Benito	Yes	No
Agri-food in Jaraíz de la Vera	Yes	No
Agri-food in Montijo	Yes	No
Agri-food in Valle del Jerte	Yes	No
Meat in Fregenal de la Siera	Yes	No
Meat in Higuera la Real	Yes	No
Cork in San Vicente de Alcántara 3	Yes	No
Packaging in Mérida	No	Yes
Energy in Badajoz	No	Yes
Metal in Badajoz and Jerez de los Caballeros 4	Yes	Yes
Health in Cáceres	No	Yes
ICT in Cáceres	No	Yes
Tourism in Cáceres	No	Yes
Various in Navalmoral de la Mata	Yes	No

¹ Local Productive Systems (LPS) and Industrial Districts (ID). ² In the public funds program for IBGs in 2008, the Extremaduran Federation of Furniture and Wood Entrepreneurs (http://www.fedexmadera.com/es/.html (accessed on 1 May 2020); consultation May 2020), which is currently not recognized as an IBG, appeared among the beneficiary institutions. The same occurred with the Extremadura Construction Materials Cluster, which in 2008 also received funds due to its status as an IBG, which it no longer has. ³ The Extremadurian Cork Cluster, based in San Vicente de Alcántara, was a beneficiary of the IBG funds program in 2007. However, it has subsequently lost the status of IBG. It is important to mention that this IBG term contains companies from all over the country, which minimizes the agglomeration effect in competitive terms. ⁴ The metal IBG is based in Badajoz. It is, however, the industrial agglomeration of metal located by Boix and Galletto [75] in Jerez de los Caballeros. Source: Own elaboration.

.

Finally, the third aspect has a methodological nature, and refers to the fact that the ID or LPS identification, detection, and analysis methodologies usually take the so-called local workforce systems (LWS) [76] as spatial reference. This term fits, more or less, with the municipal term, and in no case adheres to the region or the LAG territory-of-influence. This factor must be corrected if what is intended is to incorporate the theory of business agglomerations into the strategic planning of LAGs. In addition, its correction is also desirable to assess the regional relevance of the agglomeration, its impact on the economic and social development of the region, and its supramunicipal area of influence; in short, to evaluate and measure the agglomeration effect of the region.

To sum up, when looking for a methodology for the detection of business agglomerations in the rural world, the most appropriate type of agglomeration is the so-called rural district. In this sense, Castillo and García [67] suggested that the basic territorial unit that best adheres to the theoretical definition of this type of agglomeration is the local action group.

3. Methodological Limitations for Regional Analysis of Rural Agglomerations

Starting from the existing methodologies for the identification and detection of business agglomerations, in

Table 2. Industrial districts (IDs) and/or local productive systems (LPS) identification methodologies and their adaptation to the local action group (LAG) geographical area.

Methodology 1	Strong Points for Its Application to LAG	Weaknesses for Its Application to LAG
Courlet and Pecqueur 2 [21,78]	It uses the municipality. It only detects the industrial branch but is easily integrated into a value chain analysis	It uses variables for specialization and a minimum requirement of establishments, leaves variable outputs out of analysis Comparative based on the national total.
Sforzi-ISTAT 3 [75,76]	Institutional recognition. Academic recognition. Wide use and notoriety.	LWS as territorial unit. Only focused in industrial sector. Its prevalence index rules out polyspecialized districts. It rules out protodistricts. Based only in employment.
Sforzi-ISTAT for big business systems 4 [79,80]	Same as in the previous case.	Same as in the previous case. The predominant type of company in rural areas is the SME.
Laine [81] 5 [77,82]	Identifies LWS and ID. Greater flexibility than previous methodologies. Even starting from a criterion such as LWS, its noninclusion allows localized realities to maintain the characteristics of an LPS.	It omits the economic importance of the dominant activity of the LPS. It does not include an international competence criterion of the LPS. Quite restrictive methodology, excluding other forms of agglomeration such as protodistricts.
Hernández, Fontrodona, and Pezzi [83]	Greater flexibility than previous methodologies. It includes identification criteria based on internationalization and economic importance. It does not consider the LWS as a territorial unit. It detects all types of business agglomerations.	It does not take employment into account, giving too much importance to the variable number of companies. It does not clearly define a scale of the types of companies or specialities. It lacks criteria that distinguish between large companies and SMEs.
Integrative methodology Puig, Plá, and Linares [84]	Identifies LWS and ID. Greater flexibility than previous methodologies.	Uses variable occupation leaving out variables from other interesting studies.
Italian experimental methodology 6	It uses quality variables normally associated with regions (protected designations of origin or protected geographical indications)	It is based on variables relative to the land factor. Almost exclusively linked to agriculture and livestock. Low weight of variables such as employment in industry.

¹ We leave out of the analysis of the methodologies used [85,86,87,88,89,90,91,92] as they have already been improved, in our opinion, by more recent methodologies. ² Adapted by Climent for the study of La Rioja. ³ The Sforzi-ISTAT methodology, although it has undergone several updates, is considered here in its ISTAT version [93,94]. This is one of the most contrasted methodologies in the existing literature, whose results have served as the basis for other research. ⁴ Sforzi-ISTAT methodology, but changing the criterion related to the size of the dominant industry from SMEs to large companies. ⁵ Corrected by [77,81]. ⁶ We use the version provided by Legislative Decree No. 228 (18 May 2001). In Spain, and specifically in the case of Castilla-La Mancha [67,68], it has been used to analyze rural districts, but is very focused on population movements, and not on productive specialization and business concentration. We have ignored it in this analysis.

we have tried to synthesize the advantages and disadvantages that these present for their adaptation to the territorial analytical framework proposed here; that is, the areas of influence of the current LAGs. Broadly speaking, if we do an overall analysis, we find four major methodological limitations for the analysis of business agglomerations at the county level or within the geographic demarcation associated with LAGs. The first of these is the delimitation of the productive specialization of the territory. In this sense, the existing methodologies usually start from the search for a productive specialization in a smaller geographical area of the region, usually municipal or close to it, when taking the LWS as a functional administrative (and geographical) unit [77].

The second major limitation that these methodologies present as regards being useful in the LAG strategy is their industrial orientation. That is, these methodologies usually ignore the fact that productive specialization is not necessarily limited to the industrial field, and may be found in activities in the agricultural or service sector. Furthermore, they do not contemplate the existence of branches or value chains that include agricultural, industrial, and tertiary activities (from the production of raw materials to the commercialization of manufactures), despite the fact that one of the main lessons of the theory of business agglomerations is the promotion of the vertical integration of processes, or the integration of the value chain of products. This aspect is key to the identification of the comparative advantages in rural areas, which are usually found in the availability of a certain raw material or natural resource, regardless of whether its industrial transformation has developed in the region.

The third limitation is the use of the national context as a frame of comparison when determining the productive specialization of the territory in a given productive activity. In our opinion, this prevents the detection of business agglomerations that show some relevance in rural areas but appear less significant in the national context. As an example, an agglomeration of 20 companies that generate 200 jobs will be significant and should be considered in a hypothetical regional development strategy if it is located in a certain rural region, but it will probably be diluted if it is located in the metropolitan area of a large city. Failure to take this aspect into account supposes the exclusion of business agglomerations from rural development policies, which, although not very relevant at the national level, constitute or may constitute an economic engine for some rural areas.

The last limitation has to do with the restrictive nature of the businesses that make up the agglomerations being studied. Normally, the existing methodologies adopt criteria oriented towards the identification of agglomerations of small and medium-sized companies, without prejudice toward the existence of works that have been concerned with the locations of large company districts [66,67]. In our case, we understand that this “SME vs. large company” approach is unhelpful, since the existence of an agglomeration of SMEs is as relevant to the development of a rural environment as the location of an agglomeration led by one or more large companies. Thus, the methodology to be used should be flexible enough to include both realities.

Source: Expanded from [26] (p.129).

4. Methodological Adaptation to Regions with Low Manufacturing Density

The exercise carried out in the previous section leads us to conclude that the methodology most easily adaptable to the geographical area of the LAG is that designed by Lainé [81], with the improvements that have been introduced by other authors [33,77,82]. The resulting methodology can be applied to geographical areas wider than that delimited by the LWS, without the detected agglomerations losing the theoretical characteristics of LPSs—those that empower them to achieve competitive advantages. However, this methodology continues to be quite restrictive, since it does not identify realities such as protodistricts [95,96,97,98] nor does it allow the detection of extended value chains, since it focuses solely on industrial activity. Furthermore, it requires a high business-density for the location of the agglomeration, which makes it difficult to apply it to the regions with the highest rurality and depopulation index, as is the case of Extremadura [99,100,101]. It is difficult to identify LPSs based on this methodology in regions with little or no industrialization, such as Extremadura in Spain [102,103,104,105], not only for the reason of industrial arithmetic (scarcity of industries, low active population in the secondary sector, etc.), but also due to the scarcity of sources available on a regional scale. For this reason, we consider a methodological adjustment that emerges from Hernández, Fontrodona, and Pezzi [83] to be appropriate, which is useful when we work with regions with a low manufacturing density, such as Extremadura.

In this section, we make a methodological proposal that allows for a better adjustment to the reality of the least economically developed regions, allowing the identification of LPSs in more ruralized and not strictly industrialized environments. This proposal does not invalidate the aforementioned methodologies, but it is based on them, particularly the one used by Hernández, Fontrodona and Pezzi [83] for Catalonia. Furthermore, it seems to us a more flexible proposal, since it does not predetermine either the territorial unit of reference for the analysis or the codification of the activities with which to work. In this sense, it allows for by-county and regional analyses and exercises to identify LPSs of the value chain and polyspecialized ones, thus not adhering to the mere detection of manufacturing LPSs (it would, in fact, allow for the identification of rural LPSs specialized in the agriculture, livestock, or extractive industry).

In accordance with the above, a previous step to adapt the methodology is to choose the geographic level to which it will be applied. As we have seen, the way to integrate LPSs into the European regional development strategy is to use the LAG’s territory-of-influence as a geographical unit. For the analysis of the productive specialization of the possible LPS identified, we understand that it is better to use an aggregated classification of the branches of activity, since, although it lacks specificity, it facilitates the identification of value chain LPSs, that is, agglomerations, that work in different parts of the production chain of a specific branch. In this sense, it seems appropriate to use the sectoral grouping of activities proposed by the CNAE 2009 (

Table 3. Sectoral classification of the CNAE 2009 activities proposed.

Classification	CNAE 2009
Agri-food industry	01. Agriculture, livestock, hunting, and related services (Except 0116. Plant cultivation for textile fibers and 0128. Cultivation of spices, aromatic, medicinal and pharmaceutical plants) 03. Fishing and aquaculture 10. Food industry 11. Manufacture of beverages 12. Tobacco industry 462. Wholesale trade of agricultural raw materials and live animals (4624. Wholesale trade of leather and skins) 463. Wholesale trade of food products, beverages and tobacco
Forestry and forest products	02. Silviculture and forest exploitation 16. Wood and cork industry, except furniture; basketry and plaiting
Chemical, plastic, and petrochemical industries	05. Extraction of anthracite, coal, and lignite 06. Extraction of crude oil and natural gas 091. Support activities for the extraction of oil and natural gas 19. Coke ovens and oil refining 20. Chemical industry 22. Manufacture of rubber and plastic products 4671. Wholesale trade of solid, liquid, and gaseous fuels, and similar products 4675. Wholesale trade of chemical products
Metallurgical industry	07. Extraction of metallic minerals 24. Metallurgy; manufacture of iron, steel, and ferroalloy products 25. Manufacture of metal products, except machinery and equipment 4672. Wholesale trade of metals and metal ores 4677. Wholesale trade of scrap metal and waste products
Nonmetallic mineral product industries	08. Other extractive industries 099. Support activities for other extractive industries 23. Manufacture of nonmetallic mineral products
Textile and clothing	0116. Plant cultivation for textile fibers 13. Textile industry 14. Manufacture of clothing 4641. Wholesale trade of textiles 4642. Wholesale trade of clothing and footwear
Leather and footwear	15. Leather and footwear industry 4624. Wholesale trade of leather and skins
Paper, publishing, and graphic arts	17. Paper industry 18. Graphic arts and reproduction of screen-printed media
Pharmaceutical manufacturing	0128. Cultivation of spices, aromatic, medicinal, and pharmaceutical plants 21. Manufacture of pharmaceutical products 4645. Wholesale trade of perfumery and cosmetic products 4646. Wholesale trade of pharmaceutical products
Manufacture of computer and communications products	26. Manufacture of computer, electronic, and optical products 27. Manufacture of electrical material and equipment 465. Wholesale trade of equipment for information and communication technologies
Machinery manufacturing	28. Manufacture of machinery and equipment. Not included elsewhere 33. Repair and installation of machinery and equipment 466. Wholesale trade of other machinery, equipment, and supplies
Automotive industry	29. Manufacture of motor vehicles, trailers, and semitrailers 30. Manufacture of other transport material
Products for domestic use	31. Manufacture of furniture 4643. Wholesale trade of household appliances 4644. Wholesale trade of porcelain, glassware, and cleaning articles 4647. Wholesale trade of furniture, rugs, and lighting appliances 4648. Wholesale trade of clocks and jewelery 4649. Wholesale trade of other articles for domestic use
Other types of industries	32. Other manufacturing industries
Supplies and waste management	35. Supply of electrical energy, gas, steam, and air conditioning 36. Collection, purification, and distribution of water 37. Collection and treatment of wastewater 38. Collection, treatment, and disposal of waste; valorization 39. Decontamination activities and other waste management services

Source: Own elaboration from Galetto and Boix (2006: 8) and from the table of equivalences between CNAE 93 Rev. and CNAE 2009 Rev. of the Spanish National Institut os Statistics.

), which would distinguish 16 major productive branches with various activities, each representing the vertical integration that exists within them.

Once the statistical information has been compiled according to the regional territorial demarcation (LAG) and the proposed classification of activities (

Table 4. Description of indicators.

Indicator	Description	Period
Number of employees	Extremadura companies included in SABI	Average data: 2012–2014 Lifecycle: 1993–2018
Number of companies	Businesses and establishments (SABI)	Average data: 2012–2014 Lifecycle: 1993–2018
Income	Operating income from SABI-listed Extremadura companies	Average data: 2012–2014 Lifecycle: 1993–2018
Internationalization	International company SABI indicator	Indicator without temporary referece
Social Capital	Formal relations between companies (participated, shareholder, etc.) reflected in SABI	Indicator without temporary referece

Source: Own elaboration.

), our proposal suggests the following three steps: (1) look for the productive specialization of the LAG territories and verify the relative importance of this (that is, the LPS that is identified) in the economy at the regional or sectoral level; (2) once the previous one has been verified, look for formal (or informal) signs of collaboration or cooperation between the companies that make up the LPS; and (3) verify the international character of the LPS companies, that is, their exporting vocation (this has to happen at least for some of the companies that make up the agglomeration).

The first of the steps suggests slightly modifying the specialization index set forth in the criteria used in other methodologies, so that it is sensitive to the size of the companies. This is achieved by calculating the index based on the number of companies and the number of employees, and not only using the number of firms that work in the productive activity considered; that is, converting the equation of criterion 6 into the following two equations.

Equation (1): Depending on the number of firms

$$\mathrm{L}{1}_{\mathrm{ij}}=\frac{\frac{{\mathrm{E}}_{\mathrm{ij}}}{{\mathrm{E}}_{\mathrm{j}}}}{\frac{{\mathrm{E}}_{\mathrm{i}}}{\mathrm{E}}}$$

(1)

where:

L1_ij is the specialization index in territory i and in sector j measured in terms of the number of companies (“territory i” being understood as the LWS or LAG territory considered, and “sector j” as the productive activity on which we apply the methodology);

E_ij is the number of firms of sector j in territory i;

E_j is the total number of firms in sector j in the geographical area that we are going to take as a reference (we consider it convenient to take the region or autonomous community, and not the nation, as the top territorial reference unit, in order to make the agglomerations’ detection process more flexible);

E_i is the total number of firms in territory i (of all the productive sectors);

E is the total number of firms in the territory taken as a reference (that is, the number of firms in all sectors in the region, which serves as reference).

Equation (2): Depending on the number of employees

$$L{2}_{ij}= \frac{\frac{L_{ij}}{L_j}}{\frac{L_i}{L}}$$

(2)

where:

L2_ij is the specialization index in territory i and in sector j measured in terms of number of employees (“territory i” being understood as the LWS or LAG territory considered, and “sector j” as the productive activity on which we apply the methodology);

L_ij is the number of employees of sector j in territory i;

L_j is the total number of employees in sector j in the geographical area that we are going to take as a reference (we consider it convenient to take the region or autonomous community, and not the nation, as the top territorial reference unit, in order to make the agglomerations detection process more flexible);

L_i is the total number of employees in territory i (of all the productive sectors);

L is the total number of employees in the territory taken as a reference (that is, the number of employees in all sectors in the region, which serves as reference).

Specialization will be verified when the specialization index in both cases is greater than 1, as this would indicate that, in terms of both firms and employment, the LAG territory considered presents a degree of specialization higher than the upper reference territory (in this case, region). This step must also verify, as Hernández, Fontrodona, and Pezzi [83] maintain, that the LPS has an important influence at the regional and/or sectoral level. These authors propose that the relative weight of the productive branch in the LPS should be greater than 15% of the productive branch in the reference space (region), or what is the same, that the production of the main productive branch in the LAG territory represents more than 15% of the total of the same productive branch at the regional level. The relative importance of the LPS would also be verified if it represents more than 0.1% of the set of productive activities in the region under study. Both seem adequate to us, so we endorse them.

To correct the limitations presented by any methodology relative to the specialization index, two criteria enunciated by Laine are proposed [81].

Criterion Number of employees. This criterion is complementary to the previous one, since it serves to verify the productive specialization of the territory through the active population. In addition, following Giner, Santa María, and Fuster [65], and taking again their more restrictive criterion, we consider that a LAG contains an LPS if it has at least 200 employees directly dedicated to a specific branch of activity, in which the territory would be specialized.

Criterion Business density. This criterion tries to verify the presence of a high geographic concentration of businesses that are dedicated to the production of the same product or to the same branch of activity in the analyzed LAG territory. For this to be verified, the number of firms dedicated to the same productive activity per km² in the LAG territory under consideration must be higher than the average of the same indicator in the geographical area chosen to establish the comparison (the region or autonomous community, preferably, in this case).

Finally, following Hernandez, Fontrodona and Pezzi [83], the existence in the LPS of social capital is desirable for a better result, ensuring for it, for instance, business cooperation. Such a factor has a positive impact on competent performance of firms at the international level, and it can be verified with the following two criteria.

Criterion Internationalization. One of the main characteristics of LPSs is that they provide a competitive advantage that allows companies, even if they are small, to compete in the international market. That is why the acceptance of this criterion requires the verification of the existence of companies belonging to the agglomeration that compete in the international market, that is, that export all or part of their production.

Criterion Business cooperation. Another characteristic that the theory of industrial districts assumes is the existence of business cooperation between the companies that make up the agglomeration, in such a way that the existence of business cooperation must be a sine qua non condition to identify an LPS. The measurement of business cooperation can be verified formally and informally, although we understand that a simple way to do it is verifying the existence of agreements between companies or the participation of some of them in the capital of others.

In short, this methodology allows us to identify business agglomerations of a local or regional nature with contrasting importance in terms of employees, number of companies and income generated, and with a significant influence at the LAGlevel and a high level of business cooperation and presence in international markets.

As the intention is to enable the construction of local development strategies covered by the rural development strategy at the European level, it is necessary to understand at what stage of its life cycle the agglomeration is, that is, whether it is in an incipient development stage or in a mature or decline stage. This is important because the actions to be implemented in each case are different due to what the LPS and the companies inside it really need from an institutional point of view [106,107,108,109,110,111,112,113]. In this sense, to identify this we will use the methodology described by Branco and Lopes [106], and Rangel [113], which uses the indicators of employees, number of companies, and income generated to catalogue each of the agglomerations previously detected.

5. Result for the Extremadura Case

The local sources available for deriving the indicators that we have been describing are difficult to find. As such, we use the database built by Rangel [26], which is described in the following table.

The use of this methodology shows us up to 22 productive specializations in Extremadura with a root at the local or regional level, considering their relative importance in terms of number of companies, employment, and level of generated income. These 22 LPSs are located in 13 LAG territories, which implies that there is polyspecialization in some of them. Mostly, we find that the LPSs that start from an advantage in agriculture and livestock (rural districts) are very relevant, as reflected in

Table 5. Rural districts (LPS) in Extremadura.

Specialization—Location	Companies	Companies Specialization Index	Employment	Employment Specialization Index	Income Generated (EUR)	Sector Weight (%)1	Extremadura Weight (%)2
Campiña Sur—Meat	139	1.80	454	1.64	125,466,662.82	3.90	0.92
Campo Arañuelo—Agri-food	136	1.24	1124	1.60	198,770,797.48	4.92	1.45
Campo Arañuelo—Metallurgical	20	1.21	293	2.18	59,028,949.83	4.57	0.43
Jerte—Agri-food	34	1.39	333	2.52	80,197,505.10	1.98	0.59
La Serena—Meat	146	1.33	473	1.07	145,678,973.43	3.60	1.06
La Serena—Granite	76	6.16	413	7.31	34,735,251.07	11.63	0.25
La Siberia—Meat	80	1.59	266	1.34	74,187,954.75	1.84	0.54
La Vera—Agri-food	111	1.40	434	1.59	98,016,876.79	2.43	0.72
Lácara—Agri-food	202	1.49	1174	1.64	153,603,726.12	3.80	1.12
Miajadas-Trujillo—Agri-food	109	1.18	657	1.49	157,331,112.82	3.89	1.15
Sierra Grande-Tierra de Barros—Agri-Food	305	1.16	1461	1.21	373,502,535.69	9.24	2.73
Sierra San Pedro-Los Baldíos—Cork	68	15.55	466	20.73	113,633,209.54	58.80	0.83
Sierra Suroeste—Meat	154	1.65	718	1.00	149,821,763.41	3.71	1.09
Sierra Suroeste—Jewelry	10	1.18	326	16.76	80,706,681.30	68.67	0.51
Sierra Suroeste—Metallurgical	28	2.01	892	6.48	948,623,622.58	73.41	6.93
Tentudía—Meat	119	2.01	354	1.86	55,735,863.99	1.38	0.41
Tierra de Barros—Metallurgical	80	2.01	421	1.83	86,487,143.35	6.69	0.63
Vegas Altas—Agricultural Machinery	50	1.48	203	1.63	25,288,543.43	15.04	0.18
Vegas Altas—Agri-food	357	1.32	3193	1.69	746,653,970.96	18.47	5.46
Vegas Altas—Chemical produtcs	36	1.46	228	1.82	107,941,686.10	32.32	0.79
Zafra-Río Bodión—Agri-food	80	1.24	266	1.49	74,187,954.75	8.14	2.40
Zafra-Río Bodión—Metallurgical	31	1.35	363	2.12	47,373,551.59	3.67	0.35
Total	2371		14,512		3,936,974,336.90

¹ Percentage of the total revenue generated by LPS in the Extremadura sector to which the specialization belongs. ² Percentage of total revenue generated by LPS in the total Extremadura economy. Source: Own elaboration.

.

Finally, we present the life cycle status results for each of the LPSs, represented in

Table 6. Life cycle of Extremaduran rural districts.

Agri-food and meat districts
Growing	Maturity	Decline
Vegas Altas—Agri-food Zafra-Río Bodión—Agri-food	Campo Arañuelo—Agri-food	Lácara—Agri-food Miajadas-Trujillo—Agri-food
Agri-food and meat quality district
Growing	Maturity	Decline
La Vera—Agri-food Tentudía—Meat Sierra Grande-Tierra de Barros—Agri-Food Jerte—Agri-food	Sierra Suroeste—Meat Campiña Sur—Meat La Serena—Meat	La Siberia—Meat
Other districts
Growing	Maturity	Decline
Campo Arañuelo—Metallurgical Vegas Altas—Agricultural Machinery Vegas Altas—Chemical products	Sierra Suroeste—Jewelry	Sierra San Pedro-Los Baldíos—Cork Zafra-Río Bodión—Metallurgical La Serena—Granite Sierra Suroeste—Metallurgical Tierra de Barros—Metallurgical

Source: Own elaboration.

. This analysis is based on the parent trend by income, employment and business indicators from 1993 to 2018.

In our study, the intention is not to analyze the impact of the LEADER program through the LAG territories, because this fact is already perfectly well described in the research developed by Nieto and Cárdenas for the case of Extremadura [3,4,5,114,115,116]; nor is our intention to define the location of Extremadura’s industry [117], but it is instead to check whether the methodology described allows us to identify and detect productive specializations and business agglomerations at the local or regional level whose economic influence is significant at the LAG level, so that this specialization can be enhanced in the rural development strategy.

In the Extremadura case, unlike regions with high business-density, we find that the business agglomerations and productive specializations that have been detected have a moderate level of employment and generated income, as shown in Table 5. However, some cases, particularly those with a special link to agri-industry, have a clear growing trend. In them, taking into account economic theory, it can be understood that they have a certain competitive advantage that favors companies and projects linked to productive specialization. Following the Italian example described by Toccaceli [118], these territories considered to be rural districts fit into policies developed through LEADER projects (LAG) or in the Common Agricultural Policy.

6. Discussion and Conclusions

The local productive systems identified in Extremadura by the methodology proposed in this paper are characterized by their modest contribution to the regional level in terms of employment and number of firms, this being much lower than the contribution evidenced by the industrial districts identified in other studies at the national level [15,75,76,78]. In this sense, what is verified is that these agglomerations have a great impact in terms of income and employment when the analytical and comparative territorial framework is local, and even regional [33], as evidenced, for example, in the business agglomeration dedicated to the cork manufacturing found in the Sierra de San Pedro-Los Baldíos, located in the west of the Extremadura region. Its importance has led to the fact that, facing the decline stage of the agglomeration, all the agents that comprise it (employers, workers, institutions, research units, etc.) have worked in unison to reverse the situation [119], in what can be classified as an effect of the social capital that the agglomeration possesses.

Among the productive specializations that have been identified for Extremadura, we find a common nexus, namely, they are all based on the possession and use of natural resources, which gives the territory a uniqueness in the form of a comparative advantage that can be used in international trade. We observe this fact in other similar studies carried out at the national and international levels, and in particular in studies carried out in regions such as Andalusia [21] or Castilla-La Mancha [120], where the same phenomenon happens. Even in Italy, a paradigm of the economic literature on agglomeration economies, we already refer to agri-food industrial districts or rural districts [118], and there is also a similar pattern that links business agglomeration with the exploitation of endogenous natural resources, especially in regions with a high incidence of rural areas, such as Sardinia [121].

Another aspect that should be highlighted from the results obtained is the verification of polyspecialization in several of the Extremadura territories. Indeed, the existence of more than one productive specialization has been found in several Extremadura regions, which has positive effects on their economic development, perfectly described by Ruíz [122], as observed in the greater dynamism that regions such as Vegas Altas del Guadiana (one of those in which polyspecialization has been more clearly evidenced) present [12].

In line with the foregoing, empirical evidence shows that the agglomeration industry when organized in the form of agglomeration obtains better results in competitive terms than when it is achieved in a dispersed (non-agglomerated) way [123]. In this sense, we find that the agglomeration of activity identified in Extremadura around a product or branch of activity permits a capacity for the integration of the value chain, ranging from primary activities to wholesale trade, and in some cases passing for the complete transformation of the products. This fact, which can be presented as a common behavior pattern in border regions [124], invites us to think that the clusters detected exhibit the behavior described by industrial ecosystems, in accordance with green and circular economy policies.

All of the above contributes to the design of a bottom-up development strategy for Extremadura, since the methodology allows for locating local productive systems in rural areas with a significant influence on employment, number of firms and income generated at the local and regional level (or in the territories of influence of the local action group), based on the unique production and resource endowment that some Extremadura territories have, and with the possibility of developing primary, secondary and even tertiary branch activities around these products or resources. In short, transforming natural resources into value-added products makes possible the development of services linked to these productive specializations, in particular of a touristic nature, a fact that would lead to a full use of LEADER development strategies, which have been put into practice in Extremadura as regards rural tourism as well [125].