International economic policy: a fuzzy set qualitative comparative analysis on think tanks in the press

Abstract

This study examines the conditions that explain leading think tanks’ successful strategies for the growth of world media representation of the European Union’s (EU) international economic policy (N = 19). A fuzzy set qualitative comparative analysis (fsQCA) was developed throughout 2009–2018, which begins with the euro crisis. The causal conditions used are: the years of experience of each think tank; the attributes of its media representation (number of languages in which they spread their messages in the media, thematic concentration, and media concentration); and the macroeconomic characteristics of the countries from which they carry out their main activity (competitiveness and belonging to the euro). This data has been obtained through Factiva® (Dow Jones & Company©), the Global Competitiveness Index (World Economic Forum), and the European Central Bank. The core conditions analysis shows the importance of the think tank experience and the diversity of languages for their successful strategies. Likewise, the different pathways suggest that the thematic concentration and membership of the euro are sufficient conditions to explain the increased influence of the EU think tanks on the international public agenda.

Keywords:

• Think tanks
• economic policy
• international economics
• fsQCA
• causal recipes
• media representation

SUBJECT CLASSIFICATION CODES:

• F53
• H00
• Z18

1. Introduction

Think tanks are organisations whose main objective is to generate ideas and conduct analyses and research to subsequently disseminate them, advise the political elite on the formulation of public policy, and promote debates on the public agenda (McGann & Weaver, 2000; Misztal, 2012; Montobbio, 2013; Stone & Denham, 2004). Moreover, research centres (Castillo-Esparcia et al., 2017), factory of ideas (Abelson, 2009), or defence coalitions (Sherrington, 2000) form a set of data, knowledge, and influence networks, competing for notoriety, recognition, and credibility (Pérez-Yruela, 2018).

Current research on think tanks highlights that diverse actors, such as experts, academics, political parties, companies, and the media, among others, participate in the generation of public policies (Plehwe et al., 2018). Several authors posit that the main strategy of think tanks to influence political decisions is to increase their visibility in the media (Lalueza & Girona, 2016), since it is the fundamental instrument for disseminating activities, research, and proposals from these entities (González-Enríquez, 2018). In this sense, there is a need for studies on think tanks to examine their instrumentality, expressed as ‘knowledge marketing’ (Stone, 2007).

The objective of this research is to analyse the combinations of conditions that explain the successful strategies followed by European Union (EU) think tanks to increase their global representation in the media. Thus, the necessary and sufficient conditions that lead to increased media visibility are identified. For this, a qualitative comparative analysis (QCA) of 19 European think tanks has been carried out and included in the international economic policy category of the 2018 Global Go To Think Tank Index Report (McGann, 2019).

With QCA it is possible to bring some of the spirit and logic of case-oriented investigation to small and medium sample studies. In this sense, unlike traditional statistical techniques, QCA allows conclusions to be drawn from particular cases (Ragin, 2008). This methodology allows researchers to overcome the limitations of conventional statistical techniques (Woodside, 2013). Various investigations that use QCA resort to small samples and provide relevant results (Cezar, 2020; Li & Ma, 2019; McLevey, 2014).

The sample has been built with EU think tanks that exceed 100 mentions in the media during the period under study (2009–2018). This 10-year stage begins with the euro crisis and includes the different bailouts to Greece, Portugal, and Ireland, as well as the request for financial assistance from Spain. In the same way, the stages of adjustment and austerity have been contemplated. The longitudinal analysis incorporates the overcoming of the euro crisis and reaches the recovery phase and subsequent economic growth.

The novelty of this research lies in using QCA to identify the main factors that allow the design of the different strategies aimed at increasing the media representation of think tanks, since it is a determining factor to participate in the topics of public discussion and influence political decision-making (Castillo-Esparcia et al., 2017). This methodology offers a valid option to understand the complexity of the factors that influence the increased presence in the media. It is focused on relations with sets, so the set-theoretical logic of QCA allows the examination of multiple equifinal alternatives that lead to the same result and the identification of the conditions that may be necessary or sufficient for said result.

2. Theoretical framework

This research examines the different strategies that explain the increase in media representation of think tanks. The causal conditions that explain the success of these strategies are: the experience of think tanks, their communicative context evaluated through the number of languages, thematic and media concentration, and the macroeconomic characteristics of the countries in which they operate, evaluated according to the competitiveness and belonging to the euro zone of each country.

The QCA methodology is applied in political science studies (Thomann, 2015) and economics (Nieto‐Alemán et al., 2019). In the specific case of think tanks, QCA is used to identify their different financing models (McLevey, 2014), assess the role of these institutions in environmental policies, and evaluate the membership of a think tank as a necessary or sufficient condition to access the governmental political elite (González-Bustamante, 2013; Olivares et al., 2014), among other studies.

2.1. Think tanks: Conceptualisation and characteristics

Given the nature and functions they perform, there are different types of think tanks, which has given rise to numerous definitions to conceptualise this phenomenon (Abelson, 2013; Chuliá, 2018; Medvetz, 2008). In the North American context, think tanks are independent or private policy research organisations present in increasing numbers around the world. More often than not, think tanks are established as non-profit organisations. When they operate internationally, they are usually categorised as non-state actors in global and regional politics (Stone, 2005). Requejo-Coll et al. (2000) add the European perspective and define think tanks as autonomous non-profit foundations while contemplating the possibility of support from the public administration and affiliation to professional groups, academic institutions, or political parties.

Precisely, this heterogeneity favours the categorisation of national ‘traditions’ referred to in these entities. In this way, an Anglo-American tradition is observed in which think tanks are constituted as independent organisations that collaborate in the analysis of public policies outside the government and political parties. In the Asian tradition, think tanks are integrated into the corporations themselves, are semi-independent, and usually have links with the government or with individual political figures (Baier & Bakvis, 2001; Stone, 2005; Stone & Denham, 2004). For their part, European think tanks are controlled by the government to a greater or lesser extent (Kelstrup, 2017) or are linked to already established public, semi-public, or private institutions (Chuliá, 2018). More specifically, and depending on the relationships between think tanks and political parties, the European pattern offers significant dependence and cooperation with the parties (Baier & Bakvis, 2001). In any case, and in general, Arshed (2017) points out that think tanks depend on some institutions. Medvetz (2012) underlines the subordination of these entities to political actors to influence public debate, to economic actors to obtain financing, and to the media to acquire a presence in the media agenda. The objective of think tanks to disseminate their research materialised in an increase in communication strategies since the 1970s (Castillo-Esparcia, 2009).

2.2. Communicative context

The media constitutes an essential element of politics (Bennett & Entman, 2001). The media representation of think tanks is the primary strategy to influence political decisions (Lalueza & Girona, 2016; Rich & Weaver, 2000) since ‘a think tank is more likely to be perceived as influential if it features frequently in newspapers and other mass media’ (Kelstrup, 2017, p. 132).

A media presence is a necessity for think tanks and they adapt their strategies to that objective, producing the type of information that the media demands (González-Enríquez, 2018). Media professionals frequently go to think tanks to request information on current aspects. Besides, journalists tend to use experts who regularly participate in their media (Graber, 1993).

Think tanks increase their media impact by specialising in a specific topic to become a reference source on that topic (Rich & Weaver, 2000). Thus, it is appreciated that thematic concentration is an essential tool in the design of their strategies.

The concentration of the messages around a selection of media can contribute to the fixation of speeches of think tanks. The cohesion of the opinions of such entities makes it possible for the media to disseminate information, which favours the creation of opinion states in the public debate (Davis, 2012). In this way, the cohesion of the opinions shows the importance that the media concentration represents in the configuration of think tank strategies.

The novel and in-depth analyses that think tanks display allow them to act as disseminators that summarise and present the most relevant information concerning topics that interest the public and the media (González-Enríquez, 2018).

The specific conditions of idiomatic diversity that characterises the EU (Gómez González-Jover, 2002) act as contingent factors within the thematic and media concentration strategies of European think tanks. The increase in transnational networks of think tanks that contribute to the research and formulation of public policies, both within the same country and across its borders (Kelstrup, 2016), supports the importance of analysing the number of languages in which think tanks spread their messages in the media.

2.3. Macroeconomic context

The last global economic crisis that began in the US in 2007 caused solvency problems for banks in advanced countries (Otero-Iglesias & Steinberg, 2019). To cope with the crisis, the governments of the EU adopted stimulus plans for a short period, specifically between 2008 and 2009. However, they quickly reversed this position and decided to implement structural reforms and strict austerity measures (Plehwe et al., 2018). The argument supporting these policies was to consider that fiscal discipline and labour market flexibility would improve the economic situation (Matthijs, 2014), secure the euro, and prevent the disintegration of the EU (Schmidt, 2016).

The sovereign debt crisis of some of the member states questioned the credibility of the euro as a currency (Otero-Iglesias & Steinberg, 2019). In that context, markets demanded austerity positions that were unpopular and politicians justified their decisions with economics experts (Coman, 2019).

On the contrary, although austerity measures have been found to be inefficient from a financial and social perspective (Blanchard & Leigh, 2013; Schui, 2014), various think tanks of the centre-right political spectrum scattered throughout the EU encouraged these ideas (Plehwe et al., 2018). Despite the apparent failure of austerity policies, they have been immune from criticism (Blyth, 2013).

In this context, think tanks played a fundamental role in the creation of a consensus aimed at favouring the restructuring of the financial system and in operating with austerity policies (Parrilla et al., 2016).

Besides, during the eurozone crisis, they managed to significantly increase their budgets and financial resources, contact networks, and visibility in the media (Coman, 2019). The evolution of the media representation of think tanks over time is not affected by the phase of the economic cycle. Thus, they generated debates about the design of public policies and became essential agents for the formulation of political recommendations (Castelló-Sirvent et al., 2020).

The analysis of competitiveness includes multiple dimensions of microeconomic origin. However, a detailed study of its foundations exceeds the scope of this work. Therefore, the Global Competitiveness Index (GCI) of the World Economic Forum (WEF) was used, an index commonly used in applied economics studies (Athari et al., 2020; Radulescu et al., 2019; Rusu & Roman, 2018).

The GCI constitutes secondary data to build country rankings and defines competitiveness as the set of institutions, policies, and factors that determine the level of productivity of a country (Schwab, 2010). Likewise, the leading countries in global competitiveness are characterised by also being the countries with a more significant innovative component, which implies a more sustainable economic development (Schilling, 2008).

Following Kelstrup (2016), the transnational linkage of think tanks with common interests suggests that the study focused on the euro zone as a relevant condition to explain their strategic behaviour. The structural imbalances of the economies of the different member states suggest that the countries of the EU that have adopted the euro as their official currency are likely to condition the behaviour and interests of think tanks that have been created and promoted in the respective country.

European think tanks act as intermediaries between experts/academics and states to disseminate political solutions that influence decision-making (Sherrington, 2000; Stone, 2007). Therefore, it is especially important to carry out a rigorous analysis of the role that competitiveness and belonging to the euro present for the growth of the media representation of European think tanks.

The doubts about the stability of the eurozone that had begun with the various sovereign debt crises articulated a public debate based on the risks that the high levels of indebtedness of the countries could represent for long-term growth (Reinhart & Rogoff, 2010). Subsequently, Krugman (2013) questioned the rationale behind these arguments by rejecting the data on which this research was based, among other aspects, due to the severe and obvious calculation errors it presented.

Given that think tanks depend on academic actors to endorse the objectivity of the research (Wiarda, 2008), the public debate during the period analysed was conditioned by the political and financial interests that displaced academic rigour. This circumstance was identified in Parrilla et al. (2016).

3. Methodology

The international economic policy category of the 2018 Global Go To Think Tank Index Report (McGann, 2019) has been considered, which includes 87 think tanks. Following Kelstrup (2017), the selection has been limited to think tanks in the European Union that exceed 100 mentions in the media during the period 2009 to 2018. Thus, the sample is composed of 19 think tanks (Table 1).

Table 1. European Union top international economics policy think tanks.

Rank	Name	Growth	Country
2	Bruegel	11,19%	Belgium
4	Vienna Institute for International Economic Studies	14,56%	Austria
7	Adam Smith Institute	8,42%	United Kingdom
10	Chatham House	6,48%	United Kingdom
18	Centre for European Policy Studies	0,91%	Belgium
22	Kiel Institute for the World Economy	1,66%	Germany
27	European Centre for International Political Economy	9,55%	Belgium
33	Institute for International Economic Studies	25,93%	Sweden
38	Centre d'Etudes Prospectives et d'Informations Internationales	20,11%	France
57	French Institute of International Relations	5,62%	France
61	Austrian Institute of Economic Research	−3,62%	Austria
72	Institute for International Political Studies	23,26%	Italy
76	Macroeconomic Policy Institute	10,36%	Germany
77	National Institute of Economic and Social Research	−0,80%	United Kingdom
79	TARKI Social Research Institute	−4,89%	Hungary
82	World Institute of Development Economics Research	43,98%	Finland
84	Organisation for Economic Co-operation and Development (OECD)	1,30%	France
86	Institut Montaigne	18,65%	France
87	Mercator Research Institute on Global Commons and Climate Change	4,74%	Germany

Source: Own elaboration.

The Factiva® tool, an information database owned by Dow Jones & Company © that provides access to more than 33,000 news sources, has been used to search content in media related to the analyzed think tanks.

The search was carried out about title, summary or news of all the media included in Factiva® in any language, for the ten years analyzed. The full name of the think tank has been used with quotes, regardless of its acronyms, to avoid false positives, since these acronyms, to a large extent, represent other identifying terms of different types of entities or elements randomly included in the contents.

We investigate using the QCA methodology (Ragin, 2008, 1987) based on the use of Boolean algebra. QCA is an asymmetric methodology that, in social research, allows a more precise analysis than other quantitative techniques posed from a symmetric perspective. Causal connections or other comprehensive connections between social phenomena can be very strong despite relatively weak correlations. On the contrary, correlation arguments are totally symmetric, while set-theoretical arguments are almost always asymmetric (Ragin, 2008).

As already mentioned, the QCA methodology is suitable for investigations with small samples (Cezar, 2020; Li & Ma, 2019; McLevey, 2014) and overcomes the methodological limitations of other statistical and inferential techniques (Woodside, 2013). In addition, it allows one to develop and profile the theory scheme (Redding & Viterna, 1999). The technique is based on the configuration of two groups of factors: the outcome or phenomenon that must be explained and the causal conditions or explanatory factors of the result. According to Fiss (2011), the procedure that must be followed with the QCA method involves carrying out four steps sequentially: identifying the property space, generating joint group measures, assessing the consistency that characterises the general relationships, and making a logical reduction of the resulting data.

In this study, we use the fuzzy set qualitative comparative analysis (fsQCA), a variant of QCA widely used in recent years (Roig-Tierno et al., 2017) and accepted in the advancement of multilevel theory (Lacey & Fiss, 2009), to identify how conditions that generate the outcome analysed are combined.

According to Ragin (2008), it is possible ‘to deconstruct a single symmetrical analysis into two asymmetric set-theoretic analyses, one focussing on sufficiency, the other on necessity’ (p. 7). The use of the fsQCA methodology in this study makes it possible to evaluate the necessary and sufficient conditions that configure successful and unsuccessful strategies in the growth of the media representation of the main think tanks of international economic policy in the EU.

The evaluation of the same result achieved by think tanks through more than one combination of causal conditions allows capturing the ideas of asymmetry (Fiss, 2011) and equifinality (Fiss, 2007). The analysis is performed for the presence and absence of each of the proposed attributes, following a standard practice of this methodology that allows identifying the asymmetry between the results of the analysis of the causes of a result and the analysis of the causes of its denial (Ragin, 2008). This is because ‘with fuzzy sets, no mathematical reason exists to expect consistency scores calculated for the negation of an outcome to be perfectly negatively correlated with consistency scores for the original outcome’ (Ragin, 2008, p. 137).

The research design process has avoided developing a mechanical approach and has followed diverse iterations within a case-oriented approach (Ragin, 2009a; Rihoux, 2013; Rihoux & Lobe, 2009), thereby guaranteeing methodological control (Rihoux et al., 2017). Consequently, we propose the following model: $$\text{GROWTH}=\text{f}\left(\text{EXP},\text{LAN},\text{SUBJECT},\text{MEDIA},\text{COMP},\text{EUR}\right)$$

Table 2 shows the nomenclature and details of the attributes, the secondary sources used and their calculation method.

Table 2. Conditions.

ATTRIBUTE		CONCEPT
Tipo	Condition	Indicator	Source
Outcome	GROWTH	Growth of world media representation. Average interannual variation rate of the media representation of each think tank for the period 2009–2018 according to the total number of contents published worldwide.	Factiva®
Think tank condition	EXP	Think tank experience. Number of years of experience of each think tank, taking as reference the year of its foundation.	Factiva®
Media conditions	LAN	Languages. Number of languages in which each think tank published its contents throughout the 2009–2018 period.	Factiva®
	SUBJECT	Subject concentration. Percentage of thematic concentration achieved in international media, calculating the relative importance of the contents of the 10 main topics discussed by each think tank throughout the 2009–2018 period.	Factiva®
	MEDIA	Media concentration. Percentage of concentration in international media, calculating the relative importance of the contents published by the 10 main media topics for each think tank throughout the 2009–2018 period.	Factiva®
Economic conditions	COMP*	Country competitiveness. Average competitiveness for the 2009–2018 period of the country in which each think tank develops its main activity. Calculation from the Global Competitiveness Index (WEF) with a base index number 100 constructed from a scale of 1–7.	World Economic Forum (WEF)
	EUR	Country with the single currency (with the Euro as currency). It takes values 0, 1 and registers the countries that have the euro as legal tender, from the countries in which each of the think tanks develops their main activity.	European Central Bank (ECB)

* The mean of the 10 years under study has been considered, since in all the countries studied the trend variation is positive and the differences in the mean interannual variation rates are not statistically relevant.

Source: Own elaboration.

The outcome (GROWTH) measures the average interannual variation rate of the media representation for the period 2009–2018. The causal conditions that help to explain the strategies of think tanks are: (1) experience (EXP) of think tanks, expressed as the number of years of experience of each think tank, taking as reference the year of its foundation; (2) number of languages (LAN) in which each think tank published its contents throughout the 2009–2018 period; (3) subject concentration (SUBJECT), expressed as a percentage of thematic concentration achieved in international media; (4) media concentration (MEDIA), expressed as a percentage of the concentration achieved in international media; (5) the country competitiveness (COMP), expressed as average competitiveness for the period 2009–2018 of the countries in which each think tank was created; and (6) with respect to these countries, if they adopted the euro as their currency (EUR).

The data matrix was calibrated in established group values that can vary between 0 and 1. Under the established criteria, three delimitation points were defined: totally inside, completely outside at the ends, and the crossing point (0.5) that identifies the anchor for a qualitative distinction is ‘inside’ or ‘outside’ of a set because the values of most cases do not meet the ideal types. This assignment has been possible from a direct method of attribution capable of designating the location of established theoretical anchors based on the knowledge and experience of researchers (Verkuilen, 2005). fsQCA performed uses the numerical values of each attribute of the model, for each think tank (Appendix). Since these values are not fuzzy values, it is necessary to calibrate the data matrix to convert them into simultaneously binary and metric values (Kent et al., 2009) that allow capturing variations in terms of sets (Ragin, 2009b).

The methodology was based on an in-depth prior knowledge of the think tanks analysed, as well as the most relevant academic theory, to allow researchers to make optimal decisions on the combination of conditions and about the expected directionality of the causal conditions. In this sense, the presence of EXP, MEDIA, and COMP was taken as a reference. For the other attributes, presence and absence were considered.

Table 3 shows the three established anchor points and descriptive statistics for the different attributes used by the model. For the calibration of the outcome (GROWTH), the imposition of preset reference values for other types of research was avoided (Schneider & Wagemann, 2012) and an average interannual variation rate of the media representation of 0 per cent, maximum ambiguity, was set as fully outside the average and fully inside average plus the standard deviation. For the COMP and LAN attributes, the median was set as fully outside and the average as fully inside. For the EXP, MEDIA, and SUBJECT attributes, an approximation to the median was taken as fully inside and the maximum ambiguity and fully outside points were set paying special attention to the knowledge of the cases (Rihoux et al., 2017). The EUR attribute is defined as dichotomous and, consequently, 1 means fully inside and 0 fully outside.

Table 3. Calibration and descriptive statistics.^a

CONDITION	CALIBRATION			STATISTICS
Concept	Fully inside	Maximum ambiguity	Fully outside	Max	Min	Average (standard deviation)	Median
GROWTH	22.40	10.39	0	43.98	−4.89	10.39 (12.00)	8.42
EXP	40	20	10	104	6	46.21 (31.77)	40
LAN	15	10	5	27	4	14.89 (6.99)	15
SUBJECT	60	50	40	68.61	39.02	56.01 (8.24)	56.06
MEDIA	60	50	40	87.45	39.95	58.63 (13.80)	57.47
COMP	95	85	80	100	78.15	94.50 (5.94)	93.80
EUR	1	–	0	–	–	–	–

^a See Appendix.

Source: Own elaboration.

To perform the analysis, the fs/QCA 3.0 software was used, which executes the Quine–McCluskey algorithm and provides a logarithmic function that allows calibrating the original values in logarithmic group scores according to values from 0 to 1 (Ragin et al., 2006).

Following Ragin and Pennings (2005), we know that ‘fuzzy sets retain almost all the essential mathematical properties of crisp sets and thus enable researchers to model complex and diverse constellations of case aspects and to assess set-theoretic relations’ (p. 425).

Consequently, fsQCA allows the creation of models based on the concept of cyclical causality (Fiss, 2011), which allows us to understand how the combination of several causal conditions can influence the success of think tank strategies.

4. Results

This document proposes the following research question: what combinations of necessary and sufficient conditions can lead to successful strategies to increase the global media representation of the main think tanks of the EU’s international economic policy?

The total media representation of the analysed think tanks offers a positive long-term regression (Figure 1). The outcome (GROWTH) shows a total asymmetric behaviour for the period 2009–2018 (Figure 2), characterised by high heterogeneity between think tanks (Figure 3).

Figure 1. Regression of world media representation of the sample (19 think tanks).

Source: Own elaboration

Figure 2. Growth of world media representation per years of the sample (19 think tanks).

Source: Own elaboration

Figure 3. Growth of world media representation, per years and per think tanks.

Source: Own elaboration

An analysis of necessary conditions for successful and unsuccessful strategies, taking an evaluation of positive and negative cases, is carried out (Ragin, 2000). A condition is considered necessary if it must be present for a given outcome to occur (Ragin, 1987; Rihoux & Ragin, 2009).

Ragin (2008) accepts that a condition is necessary if it exceeds the threshold of 0.9. Schneider et al. (2010) establish that a certain condition that achieves this level of consistency is ‘almost always necessary’, while, in a more restrictive interpretation, Dul et al. (2010) indicated that 5 per cent of cases might be counter-examples of the necessary condition, which would mean 95 per cent accuracy.

Following the most conservative criteria proposed by Dul et al. (2010), we analyse if any of the six causal conditions and their denials exceed the threshold of 0.95. The need analysis (Table 4) shows that there is no necessary condition for either the success or the failure of think tank strategies since it cannot be demonstrated that instances of the outcome constitute a subset of instances of a causal condition (Ragin, 2008).

Table 4. Analysis of necessary conditions.

CONDITIONS	SUCCESSFUL STRATEGIES		UNSUCCESSFUL STRATEGIES
	Consistency	Coverage	Consistency	Coverage
EXP	0.7634	0.4684	0.7256	0.5846
∼EXP	0.3230	0.4727	0.3402	0.6537
LAN	0.6016	0.3584	0.8919	0.6976
∼LAN	0.4923	0.7763	0.1796	0.3718
SUBJECT	0.7961	0.4778	0.8329	0.6563
∼SUBJECT	0.4272	0.6607	0.3372	0.6847
MEDIA	0.8985	0.5740	0.7049	0.5913
∼MEDIA	0.3602	0.4818	0.4921	0.8642
COMP	0.8908	0.4472	0.9053	0.5967
∼COMP	0.1967	0.6128	0.1614	0.6599
EUR	0.8009	0.4700	0.6880	0.5300
∼EUR	0.1990	0.3270	0.3120	0.6730

∼ means absence of an attribute.

Source: Own elaboration.

Although there are no necessary conditions, some conditions may be relevant in the sufficiency analysis (Schneider & Wagemann, 2012). An argument of sufficiency without necessity permits multiple paths (Ragin, 2008).

The sufficiency analysis of the proposed model (Table 5) shows the different paths followed by think tanks in the design and execution of their communication strategies. The existence of different routes and initial conditions for success and failure strategies connects with the concepts of equifinality (Fiss, 2011; Katz & Kahn, 1978) and causal complexity (Ragin, 1987). Pathways 1a, 1 b, and 1c show the paths that think tanks followed for the success of their strategies. Pathways 2a, 2 b, 2c, 2d, and 2e describe different causal combinations that help understand how the failure of their global growth strategies in the media occurred.

Table 5. Analysis of sufficiency for the outcome growth of world media representation^(b).

	ANTECEDENT CONDITIONS (intermediate solution and core conditions)^(c)
	Successful strategies			Unsuccessful strategies
Condition / Pathways	1a	1b	1c	2a	2b	2c	2d	2e
EXP	●	●				•	○	•
LAN	○	○	•	●	●	•	○	•
SUBJECT	•	•	○	•			•	•
MEDIA		•	●	•	○	•	•	º
COMP	•		•		•	•	•	•
EUR	º	•	●	○	•	º	•
Think tanks explained	1	2	2	2	5	2	3	2
Raw coverage	0.1204	0.2975	0.3328	0.2175	0.4120	0.1829	0.1539	0.3001
Unique coverage	0.1204	0.2146	0.1375	0.0781	0.1629	0.0435	0.0988	0.0304
Consistency	0.9958	0.9951	0.9053	0.9973	0.9696	0.8551	0.7910	0.9980
Intermediate solution
Coverage	0.7473			0.8021
Consistency	0.8495			0.9046
Parsimonious solution
Coverage	0.7475			0.8120
Consistency	0.8388			0.8638
Cutoff
Frequency	1			1
Consistency	0.8756			0.7910
Directional expectations	(1, - , - , 1, 1, - )			(1, - , - , 1, 1, - )

^(b) More detailed results are available at request to the authors.

^(c) Black circles indicate the presence of a condition, and white circles indicate its absence. Large circles indicate core conditions; small ones, peripheral conditions. Blank spaces indicate “don’t care”.

Source: Compiled by the authors using the software program fsQCA 3.0.

The coverage and consistency of intermediate and parsimonious solutions are suitable, both for recipes for success and for failure. According to Ragin (2008), the cut-off value indicates the existence of substantial consistency (≥ 0.75).

Table 5 presents the pathways of the intermediate solution and follows the proposal of Fiss (2011) on the nucleus and the causal periphery. The black circles show the presence of a condition and the white circles signify its absence. The size of the circles provides information about belonging to the core of the solution (large circles) or to the periphery (small circles). The distinction between causal nucleus and periphery introduces the notion of neutral permutations. According to Fiss (2011), ‘core conditions are those that are part of both parsimonious and intermediate solutions, and peripheral conditions are those that are eliminated in the parsimonious solution and thus only appear in the intermediate solution. Accordingly, this approach defines causal coreness in terms of the strength of the evidence relative to the outcome, not connectedness to other configurational elements’ (p. 403).

The consistency and coverage of intermediate and parsimonious solutions are satisfactory, in all cases greater than 0.8, a more demanding criterion than that proposed by Ragin (2008).

The different strategies followed by think tanks show very high consistency values. The raw coverage is suitable for the different pathways and, considering the type of research, totally acceptable. However, regarding unique coverage, it should be remembered that ‘the coverage gauges only empirical importance, not theoretical importance. A sufficient relation may be quite “rare” from an empirical point of view (and, thus, exhibit low coverage), but it still could be centrally relevant to theory. For example, the sufficient relation might be proof that a path that was thought to be empirically impossible, at least from the perspective of theory, in fact is not’ (Ragin, 2008, p. 55).

The configuration of solutions 1a, 1 b, and 1c helps to understand the successful strategies that they followed to achieve growth in the representation of the global media. Experienced think tanks (1a, 1 b) specialised in a few thematic niches disseminated their ideas in few languages, regardless of whether the think tank's country belonged to the eurozone. According to Fiss (2011), configurations 1a/1b are ‘neutral permutations’ of each other, respectively, in the sense that they share the same type of central conditions and only differ in their contribution conditions.

The think tanks of highly competitive eurozone countries that applied successful strategies (1c) expanded their subject areas of study, spreading their messages extensively through several media outlets in a wide variety of languages.

The analysis of pathways 2a, 2b, 2c, 2d, and 2e helps to understand the causal configurations followed by think tanks that implemented unsuccessful strategies.

In pathways 2a and 2b, diversity of languages and the absence of media concentration are basic conditions, regardless of whether the think tank country belongs to the eurozone. In highly competitive countries, pathways 2c and 2e suggest that the experience of think tanks and the dissemination of their messages in many languages are basic conditions. The think tanks that follow the 2d pathway belong to highly competitive countries in the eurozone, have little experience, spread their messages in few languages, and show a high thematic and media concentration.

Tables 6 and 7 show details of the think tanks that follow each of the strategies analysed in the model.

Table 6. Successful strategies^(c).

Pathway	Cases
1a	Institute for International Economic Studies
1 b	Institute for International Political Studies World Institute of Development Economics Research
1c	Centre d'Etudes Prospectives et d'Informations Internationales Institut Montaigne

(c) Cases with greater than 0.5 membership in term.

Source: Own elaboration.

Table 7. Unsuccessful strategies^(d).

Pathway	Cases
2a	National Institute of Economic and Social Research TARKI Social Research Institute
2 b	Organization for Economic Co-operation and Development (OECD) Centre for European Policy Studies French Institute of International Relations Bruegel European Centre for International Political Economy
2c	National Institute of Economic and Social Research Adam Smith Institute
2d	Mercator Research Institute on Global Commons and Climate Change Macroeconomic Policy Institute
2e	French Institute of International Relations Chatham House

(d) Cases with greater than 0.5 membership in term.

Source: Own elaboration.

5. Conclusions

The analysis of the presence of think tanks on the public agenda allows us to assess their ability to influence society. Think tanks show a generalised growth in their global media representation, although it is not uniform for all during the 2009–2018 period. This research studies the different successful and unsuccessful strategies of think tanks, based on various combinations of their experience, the number of languages in which they spread their messages, and thematic and media concentration. This analysis takes into consideration the competitiveness and monetary area conditions of the country in which each think tank operates.

The think tanks that managed to develop successful strategies relied on their experience, spread their messages in a small number of languages, and specialised in a thematic niche. With this strategy, they become reference sources for information professionals, who turn to them every time they need to delve into these specific topics. Conversely, various think tanks in the eurozone focused their discourse on a few media channels while they amplified their specialised topics and their messages in many languages. By covering different areas, they establish proposals and solutions in different contexts that resulted in a high presence in the media.

The empirical evidence suggests that, to achieve greater media representation, the heads of think tanks with little experience should design strategies aimed at specialising in a few topics and increasing the number of media and languages in which they disseminate them.

A limitation of this work is the absence of a content study or semantic analysis that explores the thematic interrelations that exist between the speeches of the different think tanks. This will make it possible to know in-depth the underlying forces that drive the logical configurations that have been detected in this study.

Future research should analyse the necessary and sufficient conditions that contribute to explaining the diversity of languages with which European think tanks spread their messages, depending on the degree of trade openness and economic influence of the countries from which the think tanks operate.

So far, little research has delved into the combinations of factors that determine the communicative capacity of think tanks. The contribution of this study involves examining the strategies followed by think tanks to increase their global media representation. Future research should explore the strategic behaviour of think tanks from the European tradition compared to think tanks from the Anglo-American culture.

Disclosure statement

No potential conflict of interest was reported by the authors.

Appendix.

Outcome and conditions

Table

THINK TANK	OUTCOME	CONDITIONS
	GROWTH	EXP	LAN	SUBJECT	MEDIA	COMP	EUR
Bruegel	11.19	14	14	49.92	48.79	93.80	1
Vienna Institute for International Economic Studies	14.56	45	15	56.06	62.66	93.58	1
Adam Smith Institute	8.42	41	21	46.64	64.76	97.55	0
Chatham House	6.48	98	27	53.67	47.73	97.55	0
Centre for European Policy Studies	0.91	35	23	47.13	45.06	93.80	1
Kiel Institute for the World Economy	1.66	104	25	53.91	50.40	99.30	1
European Centre for International Political Economy	9.55	12	15	55.04	49.49	93.80	1
Institute for International Economic Studies	25.93	56	4	66.28	47.27	100	0
Centre d'Etudes Prospectives et d'Informations Internationales	20.11	40	14	39.02	62.01	93.65	1
French Institute of International Relations	5.62	39	17	58.36	45.57	93.65	1
Austrian Institute of Economic Research	−3.62	91	20	65.69	77.48	94.24	1
Institute for International Political Studies	23.26	84	4	66.77	68.36	80.82	1
Macroeconomic Policy Institute	10.36	13	8	66.42	51.81	100	1
National Institute of Economic and Social Research	−0.80	80	20	56.16	61.55	98.24	0
TARKI Social Research Institute	−4.89	11	13	68.61	87.38	78.15	0
World Institute of Development Economics Research	43.98	34	6	57.92	87.45	100	1
Organisation for Economic Co-operation and Development (OECD)	01.30	57	19	49.39	39.95	93.65	1
Institut Montaigne	18.65	18	12	47.64	57.47	93.65	1
Mercator Research Institute on Global Commons and Climate Change	4.74	6	6	59.57	58.69	100	1

Source: Own elaboration.