Enabling Factors for Better Multiplier Effects of the LEADER Programme: Lessons from Romania

: LEADER is an EU development method that aims to stimulate local actors to cooperate and co-produce ideas and projects that otherwise would not be possible. Therefore, the Local Action Groups (LAGs) should not only focus on implementing the Local Development Strategies but also to actively contribute to the development of their territory. The aim of the present paper is to underline the most important tangible indirect multiplier effects produced by the LAGs in Romania in the 2014–2020 Programming Period and to identify the enabling characteristics and conditions for maximizing such effects in future LEADER actions. The study was conducted using the structured interview as a primary method for collecting data. The results were analyzed using the Principal Component Analysis and Hierarchical Cluster Analysis. The most important multiplier effects were the amount of non-LEADER grants that LAGs managed to attract and the innovation level of the projects supported from LEADER funding. The results show that the performance of LAGs is linked to the size of their team, their experience, and the involvement of their partners. However, not all LAGs managed to generate signiﬁcant multiplier effects, suggesting that they still lack the experience necessary to successfully implement the method in their territories.


Introduction
LEADER, the acronym from the French rural development action 'Liaison Entre Actions de Développement de l'Économie Rurale' was firstly introduced in the European Union (EU) public policies framework (Common Agricultural Policy, CAP) in 1991, as a measure for better tackling local rural development needs [1]. The first edition of LEADER was implemented between 1991 and 1994 to promote and test a new rural development model, based on bottom-up and inter-sectorial cooperation. During its second edition (LEADER II. between 1994 and 1999), the program was embraced in almost half of the rural areas of EU member states, while in its third stage (LEADER+, 2000(LEADER+, -2006 the long-term needs of the territory became the main focus of the method [2,3]. Starting with 2007-2013 CAP Programming Period, LEADER became the 4th Axis of the Rural Development Program and was financed through the European Agricultural Fund for Rural Development (EAFRD) [3].
In its most recent edition (2014-2020), the program was extended under the broader concept of Community-Led Local Development (CLLD) [4]. CLLD promotes links and cooperation between sectors and local actors that should create multiplier effects by coproducing ideas and projects that otherwise would not be possible [5]. The method was created to serve as a practical laboratory for innovations multiplication and adoption [6,7], and it is considered one of the most innovative EU rural development actions [8]. The program is implemented at local level by multi-actor partnerships, formed from public, private, and social actors, called Local Action Groups (LAGs). They are responsible for creating and implementing area-based local development strategies (LDSs) funded through having a pluridimensional nature, emphasizing the importance of social innovation. They argue that a basic LEADER task is to revitalize and engage the local population by applying the core principles of the bottom-up approach and by fostering innovation. Cooperation and networking are other core LEADER principles that can bring additional multiplication effects. Ballesteros and Hernandez [30] underlined that they favor good-case experience exchanges between LAGs that can engage in jointly initiative and projects which can trigger higher impact for the rural communities. In general, cooperation is approached in three distinct ways. Firstly, it is seen as a form of learning from more experienced LAGs. It is also perceived as a way to engage in innovative projects and as an opportunity to develop the institutional capacity and to attract a critical mass of actors in projects development [3]. Wojewodzka-Wiewiorska [27] points out that the very nature of LEADER enables the development of social capital, and thus, it can lead to an increase in the level of trust and joint initiatives of the local stakeholders. These are highly important aspects especially for Eastern-European EU MS countries, such as Poland and Romania, where the low level of trust between people inherited due to their former communist history still represents a great local drawback [31,32]. By creating and promoting forums and networking opportunities for the local stakeholders, the program also generates other indirect multiplier effects such as the improvement of their capacity building and know-how transfer [33]. Svobodová [34] argues that one of the priorities of the LAGs should be to increase the quality and impact of animation activities and to build social networks in order to increase the participation and local initiative as a premise for better multiplication effect. By stimulating training and employment, the LAGs can also indirectly have an additional multiplier effect by slowing down depopulation and attracting new inhabitants [35].
In Romania, the LEADER approach continues in its second implementation period. The first LAGs were selected in 2011, less than a decade ago. In the 2014-2020 programming period, 239 Local Action Groups were funded, with a total budget of EUR 563 million [36]. The weak administrative capacity, the strong local negative political influences, and the high level of mistrust between actors inherited from the socialist era mostly explained the low LEADER success [31]. It has been noted that the policy instruments were inadequately implemented to local realities [37]. In addition, the quality of the LAG's Development Strategies was strongly influenced by the socio-economic data availability, proving to have low direct impacts especially in supporting the shift toward a low-carbon and climateresilient economy, and knowledge transfer and innovation [26]. Responding to the existing academic literature gaps on the LEADER approach, this paper aims to: (1) underline the most important tangible direct (non-LEADER) and indirect multiplier effects of the LAGs in Romania; (2) identify the enabling LAGs characteristics that can better maximize such multiplier effects in future rural development public policies; and (3) establish the conditions for maximizing the multiplication effects in future LEADER actions based on the awareness level of the LAGs managers. Thus, three hypotheses were formulated: H1 better multiplication results are obtained when the private and NGO partners are better engaged in the decision-making process of the Local Development Strategy; H2 LAGs with a larger and more experienced team are able to generate more consistent multiplication effects, by creating new projects supported from other non-LEADER funding; H3 more-experienced LAGs, with a greater capacity to involve partners, are generating better multiplier effects. The results can offer important insights and lessons regarding the implementation of Community-Led Local Development in countries with a weak participatory culture. LAGs should not act as one-off project but as an integral part in the sustainable development of European rural areas [5]. Therefore, the results should also indicate LAGs real level of sustainability. Are they able to continue their activity and to generate development in rural areas without LEADER funding?

Materials and Methods
The present research was conducted in the North-West Development Region of Romania that corresponds to the second level of the European Nomenclature of Territorial Units for Statistics (NUTS 2). The size of the investigated area is similar to other previous LEADER studies [38,39]. In the 2014-2020 LEADER programming period, there were 31 active LAGs in the region with a total budget of EUR 72 million.
On average, they have an area of 881.23 km 2 and a population of almost 40,000 persons (Table 1), with an old-age-dependency ratio higher when compared to the national (23.7%) and European (27.4%) average levels ( Table 1). They have a relatively large ethnic and cultural diversity, with most of the population living in rural areas. Structured interviews were conducted with the LAGs' managers to: (1) identify the most important tangible direct (non-LEADER) and indirect multiplier effects generated by the Romanian LAGs; (2) point out the LAGs characteristics that can better explain such multiplier effects; and (3) determine actions needed for maximizing the multiplication effects in future LEADER actions and to improve LAGs sustainability. Ballesteros and Hernandez [30] previously argued that the LAGs managers are best suited to be engaged in such studies, giving their key position in the organization and the fact that they represent a link between the funding authorities, the board of directors, and the local population. The interview methodology was previously used in other similar research [29,[41][42][43]. Structured interviews are a controlled way to obtain information from interviewees, where questions are clearly established in advance. Thus, the responses become comparable by reducing the variation among responses [44].
The interviews were implemented between July and August 2020, and included 10 questions, divided into two sections ( Table 2). The interview guide was designed after an online focus group with six Romanian experts of the LEADER program (advisors, researchers, and LAG managers, May-June 2020). Due to the sanitary SARS-CoV-2 virus pandemic constraints, the research was implemented by telephone interviews. The method limits the interviewer's ability to obtain certain contextual information and the length of the interview, but it is less time consuming and less expensive [45].
The questions were both open-ended and closed-ended, using the Five-Points Likert Scale for the later. This method is used to evaluate the attitude of the respondents regarding a set of statements, using scales with common categories such as "strongly agree", "agree", "neutral", "disagree", and "strongly disagree" [46]. The first four questions are designed to identify the characteristics that might affect the indirect impact of the LAGs in their territories such as the number of employees, their professional qualification and experience, and the partnership experience. The next five questions (from Q5 to Q9) are used to selfevaluate the level of involvement of the local actors in implementing the local strategy; the quality of the cooperation activities developed with other LAGs; the innovation level of the implemented projects using LEADER funding; and the direct and indirect multipliers effects. The last open question aimed to find out the awareness about how LAGs can maximize multiplier effects in the future LEADER actions (Q10).

Q1. Employees number;
Open-ended; Q2. Expertise and education related to the employee's job description; Open-ended; For each employee; Determined based on their academic degree/studies and previous experience. Q3. Expertise and education related to the manager's job description; Open-ended; Determined based on their academic degree and previous managerial experience Q4. Partnership history; Open-ended.

Local Development Strategy (LDS) implementation
Q5. Partnership involvement in the LDS settlement (private, public and NGOs stakeholders); Closed-Ended; Using a five-level Likert scale format, where 1 meant 'not at all' and excellent'; Q6. National and international cooperation actions; Closed-Ended; Using a five-level Likert scale format, where 1 meant 'strongly disagree' and 5 strongly agree'; Q7. Innovation level of the projects supported through LEADER funding; Closed-Ended; Using a five-level Likert scale format, where 1 meant 'not innovative' and 5 very innovative'; Q8. Direct multiplier effects (Non-LEADER projects); Open-Ended; Multiple responses are allowed; Value of each project (in EUR); Q9. Indirect multiplier effects (LAG capacity to engage local stakeholders in developing projects) to add intangible results Closed-Ended; Using a five-level Likert scale format, where 1 meant 'not at all' and excellent'; Q10. Enabling conditions for maximizing multiplication effects in future LEADER actions; Open-ended.
Source: Own data.
In total, 21 responses were recorded, representing 67% of the 31 regional LAGs. Although this represents a satisfactory level, as shown in other studies [29], it may introduce a bias, since the successful LAGs could be more likely to respond to an interview. The lack of response of some managers could also suggest a lack of interest in other additional activities not covered under the LDS. Data from the open questions was manually coded. This method allows for a better data understanding and transparency [47]. Responses that expressed a similar concept were assigned the same code in order to establish their frequency [48].
The variables used in the analysis were in line with those proposed by other previous studies [49]. Variables such as the history of LAGs and the partners' involvement in creating and implementing the strategies represent the background of the organizations, while their capacity to successfully implement the LEADER methodology is given by the number of the employees and their level of expertise. The direct multiplier effects that LAGs had outside of their LDS are determined by using the value of other non-LEADER projects implemented by LAGs, while the indirect effects are represented by the scores for cooperation activities and innovation level ( Table 3). The variables were validated using an exploratory Principal Component Analysis (PCA) with varimax rotation (Table 4). This approach is one of the most often used in other LEADER-related studies [25,50].
Hierarchical cluster analysis was used to classify similar LAGs into homogenous groups. This method allows one to empirically identify the most important characteristics in maximizing the multiplier effects. Therefore, the LAGs from a certain cluster should be as similar as possible to each other but as distinct as possible from the LAGs belonging to other clusters [51]. The method minimizes variability within clusters and maximizes variability between them [49]. The Squared Euclidean Distance and Ward's method were used to generate similar sized clusters, with a similar degree of tightness. The hierarchical clustering does not require a priori decision regarding the final number of clusters [52]. However, the results are strongly influenced by outliers; therefore, they were firstly identified and removed (LAG 16) from the data set using the Single Linkage method [51]. Cluster analysis was previously used by other authors in studies related to LAGs impact [25,49,50,53]. A possible limitation of the method could be the small number of the cases. Although the literature does not impose any constraints related to the minimum number of elements that are required to successfully apply the method, it is generally accepted that having more cases provides stronger results [51]. The analyses were performed using the Statistical Package for the Social Sciences (SPSS 20.0).

Interview Results
The main results about the structure and profile of the LAGs are (Table 3) Most of the LAGs were active in the previous CAP programming period (2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014), and only five are at their first experience, with a total average of 8.2 years of local partnership experience (Q4). Public organizations (3.9) were the most active partners in defining and adopting the LDS, followed by the private (3.6) and the NGO actors (3.0).
LAGs scored average in the national cooperation activities (2.85) but have reported poor results for international cooperation (2.50) (Q6). The projects supported by the LDS were innovative (3.4), but the responses have high deviations (Q7).
While some organizations reported to have funded projects from non-LEADER actions, the majority of the LAGs reported no such direct multiplier effects (Q8). The average value of the direct multiplier project was EUR 105 300 (4% in average from their RD budget).
Overall, the respondents believe that the LAGs managed to appropriately engage and motivate the local actors and thus generating good indirect multiplier effects (Q9). No respondents considered that LAGs had zero indirect multiplier effects.
More than half of the LAGs managers (11) believe that "knowledge transfer" is the most important enabling factor for better multiplier effects in future LEADER projects (Q10). It was followed by reinforced "networking and cooperation activities" (nine mentions), "the Non-LEADER grants", and the "Understanding the local needs and opportunities" (both with 8 mentions) (Figure 1).

Principal Component Analysis Results
Four principal components (PC) were extracted using the computed eigenvalues (Table 4). Together, they explain 75.6% of the variance, a satisfactory level as shown in other studies [25].
The variables with high loadings are essential in identifying the dimensions captured in each component. The first Principal Component (PC1) shows that, in general, the LAGs with better-qualified employees have better results in cooperation activities, and reports more multiplication effects. The second PC shows that the LAGs that have strong previous experience and were managed by experienced managers have the capacity to properly engage the local partners in the decision-making process. The third PC shows that better actor involvement in the decision-making process of the LDS, together with strong national cooperation actions, increases the level of innovations in the projects financed at the local level. The fourth PC reveals the important links existing between the direct multiplier effects and the LAGs team size. However, a negative correlation could be observed with the influence importance of the public partners in designing and implementation the LDS. All these variables have loadings above 0.5; therefore, they were all selected in the subsequent Cluster Analysis [25].

Cluster Analysis Results
Cluster Analysis reveals four cluster types in the region (Table 5 and Figure 2).  Cluster 1-LAGs with high multiplication performance The first cluster, and the smallest (2 LAGs), contains the groups with the highest performance regarding both the indirect and direct multiplier effects. On average, they managed to implement non-LEADER projects with a total value of almost a third of their initial budget, while the projects supported from LEADER funding were self-assessed as being highly innovative (4.5). They are characterized by higher number of employees, experienced management activity, and consistent involvement from private and NGO partners in the LDS decision-making process.
Cluster 2-LAGs with good multiplication performance Most of the LAGs from the second cluster report moderate success in attracting non-LEADER grants (on average 4.4% from the initial budget) as direct multipliers LEADER effects. However, they implemented, in average, more non-LEADER projects and had a good indirect multiplier effects, mostly using strong international cooperation experiences and an above-average number of qualified employees that had the capacity to proper valorize an increased engagement of public and private actors in the LDS implementation.
Cluster 3-LAGs with low multiplication performance The third group contains the LAGs with low multiplication capacity. None of these LAGs were able to attract other non-LEADER funding sources, while their indirect multiplier effects score was around average levels. They presented low scores regarding international cooperation. Their actors were highly involved in the LAGs decision-making process and have a longer program experience, compared with Cluster 4, which led to more innovative projects and better results in national cooperation.
Cluster 4-Unexperienced LAGs The LAGs from the fourth cluster are the youngest ones, being active only in the 2014-2020 CAP Programming Period. They had reported poor results in the case of the cooperation activities, the level of innovation, and the involvement of their partners. The levels of managerial experience together with the level of professional expertise were below the average.

Factors That Enable LAGs Success
The Guidance on Community-Led Local Development [5] portraits LAGs as active presences in the territories that create links between the local actors in ways that should generate multiplier effects in their territories and promote a sustainable development of the rural area. However, they firstly have to build the capacity and resources that their communities need in order to take initiative [34]. The results from the current study are in line with the above-mentioned arguments. In general, more-experienced LAGs had managed to better engage their partners in the local decision-making process and obtained better multiplication results, thus verifying H3. There is a strong contrast between experienced (Cluster 1 and 2) and non-experienced partnerships (Cluster 4), which have reported poor results in all assessed categories. Partnership experience and the human capital quality proved to be important characteristics that enable LAGs to maximize their multiplier effects. LAGs need to mobilize the potential of the human resources in the process of local development. The lack of human capacity can be a serious matter when dealing with such a complex program [54]. On average, the LAGs had five employees, a similar number as those reported in other studies from Italy and Czech Republic [55,56]. This is probably the optimal level given the budgets allocation from LEADER funding. However, the results show that the groups with the best direct multiplier effects (Cluster 1) have, on average, more employees. This aspect, the number of employees with expertise, and the engagement of the private and NGO partners seem to emerge as the key aspects that differentiate between lower and higher performance LAGs, verifying the assumptions formulated in H1 and H3. However, the results are pointing out that even in the best cases, the LAGs lack the capacity to engage in both cooperation and non-LEADER projects. The lack of human resources is also perceived as a negative aspect by the respondents. This is not only a Romanian reality. Ballesteros and Hernandez [30] argue that LAGs can only cover their core activities and cannot hire technical staff to better respond to the needs and opportunities of the territory or to engage in truly innovative projects. This limits the LAG initiatives as well their ability to serve as examples and to stimulate local population [57].
Complying with the principles of the bottom-up approach is another important factor that influences LAGs success. The results show a negative effect between the involvement of the public partners and the direct multiplier effects (PC4). Respondent number 3 argued that the LAGs cannot generate such effects "unless the bottom-up approach will function without interferences from the local public authorities. Only then we could support and engage in truly innovative projects". The results are in line with those presented by Navarro et al. [7]. They point out that lack of flexibility and the top-down control of the regional and national agencies limit the level of innovation.

Tangible Multiplier Effects of the LEADER Program in Romania
The most important variable used to differentiate between high-and low-performance LAGs was the total amount of non-LEADER grants that LAGs managed to attract. This indicator measures the direct impact that the groups had outside their LDSs. This allowed LAGs to create and promote their own initiatives and projects and to create additional value in their territories. Most of the LAGs have not obtained grants outside the LEADER funding, suggesting that they have not fully embraced their role as an active generator of development in the rural community, and raises serious questions regarding their level of sustainability. However, research dealing with these kinds of direct multiplication effects is mostly missing from the academic literature.
With the exception of Cluster 1, the results also reflect a similar situation when compared to the results of LDS implementation. In general, the LAGs with the best direct multiplier effects outside of the LDS have also reported the best results within the frame of the strategies. Cluster 2 LAGs managed to create more than double the amount of jobs initially planned (208%). They were followed by the LAGs from Cluster 3 (174%) and Cluster 4 (140%) [58].
Cooperation activities, innovation, knowledge transfer, and the capacity to engage the local stakeholders in the development of their rural area are other important multiplication effects that resulted from LEADER implementation in Romania. Volk and Bojneck [18] see cooperation as an opportunity to introduce new good practices in the rural area and as a way to improve the performance capacity of LAGs. The best results in this category were obtained by LAGs from Cluster 2, which were especially interested in transnational cooperation, while LAGs from the other clusters have reported poor results regarding cooperation activities. Almost none of the latter have engaged in international cooperation, focusing on local activities. This may be explained by a lack of initiative and experience. That is in line with the results reported by Bumbalova et al. [54], who showed that some LAGs lack the capacity to create important links with those from other countries. The low level of trust between stakeholders involved in the LAGs process and the administrative barriers [43] could be other factors that lead to such poor results. These are important findings that should be considered when preparing for the next CAP Programming Period. Thus, when selecting the future LAGs, it should incorporate international cooperation as a mandatory requirement in the evaluation process of their LDS, in line with previous studies and suggestions [43].
Cooperation and innovation are among the key features in creating strong indirect multiplier effects. Innovation is especially relevant in the context of the post 2020 CAP regulations that seek the modernization of the agriculture and rural areas through the promotion of knowledge, innovation, and digitalization [10]. The LAGs generally reported either very good results (Clusters 1 and 3) or very bad ones (Cluster 4). They are strongly linked with a greater involvement of private and NGO actors. A similar situation was reported by Navarro et al. [7], who show that the level of innovation is dependent on the existence of a network of stakeholders that are highly involved in the territory. The poor results could be a consequence of the bureaucratic barriers established from the central level [59]. Ballesteros and Hernandez [30] show that those administrative requirements are limiting the capacity of LAGs to engage in innovative and more complex projects. In this context, it is necessary for the LAGs to be granted more freedom in their work and management in the future Programming Period as also pointed by Alonso and Masot [28]. To strengthen to concept of innovation, some authors are suggesting that LAGs should be more involved in defining this core principle [7]. However, the results seemed to indicate a lack of agreement regarding the characteristics of innovation. Some of the respondents consider innovation as "something new in the territory" (respondent number 18), while others are arguing that it rather represents a "new kind of value that is brought or created in the territory" (respondent number 10). The results are in line with those seen in previous research [7,57].
The existence of the LAGs can facilitate the creation of local partnerships in order to develop the territory [7]. These are in line with the responds offered by LAG managers. In this regard, respondent number 10 said, "The LAG should act as a vector for local development through which a better capacity for collaboration at the local level can be created [...]. Also the LAG, through its projects and its own staff, can be a source of knowledge transfer to the local community and key actors in the territory". A similar vision was shared by respondent number 9: "The existence of the LAG stimulates the local community through the promotion and information activities it undertakes, through the involvement of the beneficiaries in the network of information multipliers, created by the LAG, and through the fact that decisions regarding future development directions will be taken using the bottom-up approach". The transfer of knowledge and ideas can be facilitated in many ways. The animation process is deemed as crucial. This allows the local actors to learn about the opportunities that are present in the territory and stimulate them to engage in the decision making process and investment [2]. The results in this category seem to confirm the EU hypotheses mentioned in the document The Guidance on Community-Led Local Development [5] and other findings such as De Luca et al. [33] and Svobodová [34]. A big advantage for the future CAP programming period is the fact that they can build on the already-existing capacity.

Conclusions
As previously shown, most of the LEADER-related research is generally concerning the direct outputs of the LDSs implemented by LAGs. However, by focusing on the direct and indirect multiplier effects generated outside of the strategies, the present work aims to fill these important gaps identified in the literature. The results of the PCA and cluster analysis are verifying the three hypotheses postulated in the first section, showing that these multiplier effects are strongly linked to the previously existing partnership experience, the size and level of expertise of LAGs team, and the strong involvement of the non-public partners. However, not all LAGs managed to achieve high multiplier effects. Although most of them were active at a local and more immediate level, few of them managed to fully integrate the core LEADER principles, such as cooperation and networking in their activity. Even fewer managed to obtain non-LEADER grants and to develop their own projects. This highlights the strong challenges faced by the unexperienced LAGs. The results suggest that even after two programming periods, some LAGs have not achieved the maturity needed to successfully implement the Community-Led Local Development program in Romania. The findings are also indicating a low level of sustainability in the case of Romanian LAGs, as most of them would not be able to continue their activity in the absence of LEADER funding. In the next CAP programming period, better results can be obtained by incorporating in the LAGs' selection criteria other important elements that can maximize future multiplier effects like number and experience of LAGs employees, the quality of the animation process and national and international cooperation activities. Such measures should also lead to an increase in their capacity to implement high innovative projects, enabling them to have a more active and consistent role in the sustainable development of the rural areas.
However, the study presents some limitations. The structured interview used in data collection reduces the variation among responses due to its strict format [44]. Conclusions could be limited by the fact that the study investigates the subjective opinions of the LAG managers. In addition, only 21 out of the 31 LAGs from the North-West Development Region of Romania wanted to participate in the study. Although this is an acceptable proportion, the lack of response of some LAGs managers could be itself interpreted as a lack of interest in other additional activities than the ones established by the LDS. As in other previous research about LEADER [38,39], due to its limitation to a NUTS 2 development region, results cannot be generalized to the entire EU area. Nevertheless, the region is representative for Members States that joined EU after 2004 and especially for Romania. Overall, the results can help in building more-efficient Community-Led Local Development actions in the future CAP programming period. After the current sanitary crisis, additional local actors and regions can be incorporated in the study by using direct interview techniques.