Reputation for legitimacy : independent fiscal councils : an enquiry into the components that shape the reputation of independent fiscal councils

Reputation for legitimacy: independent fiscal councils

Maarten Anton Johannes Kavelaars

Research project: The Political Economy of Financial Crises Supervisor: dr. Jasper Blom

Student number: 10533273 Amount of words: 26.315

A thesis submitted for the degree of Master of Science in Political Science, specialisation track Political Economy, at the University of Amsterdam

ACKNOWLEDGEMENTS

To my experience, the past five months have not only been about writing a thesis, nor about finishing a master's programme — first and foremost, they have demarcated the end of a four-year period as student in Amsterdam.

I am forever thankful to the people that made it such a success, both intellectually and personally. That includes several of my lecturers and professors that have challenged me over and over again to go beyond the boundaries of what used to be common and accepted to me. Intellectually, that also includes my friends — I believe I can say that at this point in time — from Rethinking, whose interests have led me to dive into worlds I did not know the existence of, but whose determination has simultaneously convinced me to always be directed by what you think yourself. Take nothing as given.

Personally, I want to show my gratitude to three people in particular. 'Pa' and 'mam', you both probably still do not know what my thesis is about. I am however forever thankful for all the opportunities you have given me and all the efforts you have put in supporting me. Do not be shy to be proud at yourselves too.

Giulia, I think this thesis has been a burden for you most of the time. Despite my uncontrolled propensity to pay attention to my laptop instead of to you, my love for you goes to a level indescribable in words. Although quite important to me, that level this thesis will never achieve.

To my grandmother, who has intellectually triggered me from a very young age on, but was not able to see her grandson graduating.

ABSTRACT

What components shape the reputation an independent fiscal council (IFC) enjoys? Although an independent fiscal council (IFC) needs a good reputation to gain popular support among an audience and to be successful in containing deficit biases, this question has not been addressed yet. This thesis aim to fill that gap. Based on a discourse analysis of 840 texts on the U.S. Congressional Budget Office (CBO) and Netherlands Bureau for Economic Policy Analysis (CPB) in three American and three Dutch newspapers, it is taken into account what elements are relevant in constituting the reputation IFCs enjoy, how these elements interact in doing so and, on the basis of a supplementary literature research, what factors explain why it is that the core constituting elements are as relevant as they are. By taking together the findings from the two cases, it is demonstrated that an IFC’s reputation is constituted by whether the quality of its work is perceived to be good or not by the public. As the way in which this audience takes this quality into account is rooted in and cannot be separated from an IFC’s historical and political context, an IFC’s reputation is shaped by all the components that together constitute how the quality of an IFC’s work is valued by its audience.

LIST OF CONTENTS

1. INTRODUCTION

...

7

2. THEORETICAL FRAMEWORK

...

14

2.2 Elements constituting reputation ...17

3. METHODOLOGY

...

24

3.1 Case selection strategy ...24

3.2 Gathering empirical material ...28

3.3 Operationalisations ...35

4. RESULTS

...

42

4.1 The Congressional Budget Office ...42

4.2 The Netherlands Bureau for Economic Policy Analysis ...55

5. CONCLUSION

...

72

List Of Newspaper Texts ...77

List Of References ...79

1. INTRODUCTION

Assuming that serving citizens as properly and fairly as possible is a goal societies should aim to achieve, societies share an interest in a political discussion in which (1) the participants have an appropriate understanding of the discussion material at hand and (2) the outcome of the discussion is translated into policies that reflect this outcome. Exchanging proper arguments on a fair basis is as vital for any other policy area as it is for the realm of policymaking that influences all others: public finance. Within the domain of the budget however, the democratic end to serve citizens to the best can substantially be eroded by the so-called phenomenon of ‘deficit biases’ (after Roubini and Sachs, 1989); situations of politically unconstrained decision-making that result in rising budgetary deficits and expanding public debt levels.

Deficit biases are rooted in various kind of mixes of; (1) insufficient understandings of the complex nature of budgetary policies which leads to short-term budgetary overoptimism or overconfidence (Rogoff and Bertelsmann, 2010); (2) rent-seeking behaviour by which politicians flaunt with policies to get re-elected, which materialises into higher debt- and deficit-levels (Von Hagen, 2010); (3) short-sightedness, in which the fruits of measures lie within an administration’s term instead of in the long run (Alesina and Tabellini, 1990); (4) time-inconsistency problems, in which policymakers decide to implement policies but later on renege on them as effects are different ex

ante and ex post (Rogoff and Bertelsmann, 2010); and (5) the ‘common pool problem’, in which the

benefits of fiscal expansion are enjoyed by a part of the group — e.g. a ministry, or interest groups — while the burden is carried by the whole group (e.g. Jacobs, 2007).

Notwithstanding the difficulties involved in demarcating what counts as ‘appropriate understanding’ and ‘translation into policies’, most of the deficit biases are neither a net-contributor to the first nor to the latter. Overconfidence and overoptimism hinder the formulation of budgetary policies that serve citizens in the long run; rent-seeking behaviour, short-sightedness and the common pool problem sideline the process by which the outcome of a proper and fair discussion is translated into policies; strategies behind time-inconsistency problems do not serve citizens fairly in general . 1

The time-inconsistency problem arises as policymakers announce a policy proposal in order to achieve behavioural change of actors

1

that will be generated by these announcements. When this behavioural change is achieved policymakers can decide to withdraw from their announcement. For example, a minister can announce to implement a tax reduction for companies that want to build factories, but increase tax rates for these companies once the factories are build. In some cases time-inconsistency problems might lead to rising deficit- and debt-levels. Within the context of this research’s argument, time-inconsistency problems are considered dubious as they mislead rather than fairly serve citizens.

Because they hinder the flourishing of a proper and fair budgetary discussion, deficit biases are problematic.

Although the diagnosis of the deficit bias problem is often rather different , its troublemaking nature 2 is recognised within academic circles (e.g. Alesina and Tabellini, 1990; Jonung and Larch, 2006; Krogstrup and Wyplosz, 2006; Beetsma et al, 2009) as well as policy realms (e.g. Hagemann, 2011; OECD, 2010). Consequently, wide-ranging efforts have been made by e.g. the European Commission and the OECD to embed budgetary policymaking in institutional configurations of fiscal governance — “the overall system of arrangements, procedures, institutions that underlie fiscal policymaking” (Ayuso-i-Casals et al, 2006, p. 652) — that are able to tackle the phenomena at the root of deficit biases (European Commission, 2017; Von Trapp et al, 2016).

Such configurations are constituted by three elements (see Ayuso-i-Casals et al, 2006). First, procedural rules can provide for a clear legal distribution of roles (who does what?) and power (who is allowed to do what?) to the different actors involved in the budgeting process, thereby limiting the ability of actors to materialise deficit biases. Second, numerical fiscal rules can provide for fixed targets and ceilings for budgetary aggregates, and set benchmarks for budgetary policies , thereby 3 legally limiting the ability for a deficit bias to expand.

Third, procedural and numerical rules can be complied with effectivity only when accompanied by institutions that (1) evaluate whether these rules are respected, (2) ascertain that the aggregated budgetary objectives are achieved, and (3) model how budgetary policies should be implemented to reach a goal (Debrun et al, 2012). Many countries have therefore founded an at least on paper 4 independent body that forecasts economic trends and analyses the progress of public budgeting (Magaro, 2013), so-called independent fiscal councils (IFCs).

The ‘neutral’ models with which IFCs forecast, monitor, and evaluate inhabit specific political-economic assumptions (e.g. Boumans and Davis, 2015, pp. 189-205; Van Staveren, 2013) —

Deficit-biases are widely considered problematic because they inflict rising debt- and deficit-levels, not because they sideline the

2

proper functioning of a democratic process. Whether it can objectively be evaluated that these rising levels are problematic, is however debatable (Blyth, 2013).

Far-going efforts have been made in the harmonisation of numerical rules. The EU not only implemented the European Stability

3

Mechanism (2012) and the European Semester (cycle of coordination of policies, implemented 2011), but also targeted fiscal governance more specifically, by defining standards members must observe in designing their budgets as part of the so-called ‘six- pack’, which inhabits the rules that annual deficit levels may not be higher than 3 per cent of GDP and public debt levels not higher than 60 per cent of GDP.

Among others Belgium, the Netherlands, Denmark, Austria, the United States, Mexico, Korea, Sweden, Canada, Hungary, Slovenia,

4

assumptions that produce non-neutral, distributive economic outcomes (Bader and Van den Berg, 1993; Hendriks, 2011). With these models, IFCs can determine for the political arena which budgetary options might be worth considering and which not. In a normativity-loaded fashion, IFCs can subsequently shape the terms upon which the political debate on the budget takes place. Citizens therefore share an interest in the presence of ‘input-legitimacy’ of IFCs; the political right to set the terms of a democratic debate in a normative way.

Simultaneously, empirical evidence demonstrates that IFCs can help to reduce biases in forecasting economic trends (Hagemann, 2011; Jonung and Larch, 2006) . Reducing this bias improves the 5 ability to anticipate which budgetary policies reach a political goal and which do not — e.g. when it comes to pro- or anti-cyclicality issues. As is indicated what decisions might serve citizens and what not, the fact that IFCs subsequently determine for the political arena which budgetary options can be considered is hence not only a normative intervention, but also a vehicle for hindering unconstrained decision making. Hereby, potentially arising deficit biases can be contained. Citizens therefore also have an interest in IFCs that, despite their normative models, possess output-legitimacy, which is achieved when the IFC does its job and effectively contains deficit biases.

As legitimacy can be defined as “the capacity of a political system to engender and maintain the belief that existing political institutions are the most appropriate or proper ones for the society” (Lipset, 1959, p. 86), the need for IFCs to bear the democratic right to obtain their activities implies that, because they lack a regular democratic mandate, IFCs should maintain broad-based support for their agency and activities. As this support exists only by the virtue of citizens that are convinced that the IFC is doing a good job, IFCs will bear this support if they have a good reputation among citizens (Calmfors, 2011; Debrun et al, 2012; Magaro, 2013). Such a reputation is at the same time a crucial driver behind an IFC’s output-legitimacy (Magaro, 2013). As IFCs have no formal decision-making responsibilities, they can only effectively contain deficit biases when politicians are incentivised to follow-up on their advices. With IFCs having a proper reputation, the political costs of ignoring its advises and analyses will become too high. Hence, politicians will feel incentivised to act in accordance with the advise of IFCs. Reputation is thus a source of soft power.

In short, IFCs that have a good reputation among citizens bear legitimacy because (1) this reputation enables them to contribute to containing deficit biases in an effective manner and (2) this

Reducing forecasting bias is often applauded as it enables to reduce deficit- and debt-levels. It however also prevents budgetary

5

reputation reflects that the agency enjoys popular support for its activities among citizens. The alternative of IFCs lacking a proper reputation is worse; although IFCs do not need input-legitimacy if their work lacks substantial influence, a poorer reputation disables IFCs to effectively contain deficit biases. The case of IFCs bearing a proper reputation is therefore one worth striving for.

A view on IFCs in advanced economies reveals that strong variation exists in the reputation enjoyed by different IFCs. Von Hagen (2010) observes that in Germany, the Council of the Wise Man, the Council of Academic Advisors to the Minister of Finance and the Joint Business Cycle Forecasting group “lack visibility and respect by the government and/or their professional peers, with the result that their reports and recommendations to not catch much public attention”. In Greece, the finance minister accused the Parliamentary Budget Office right after it published its first report of lacking “the expertise, experience and accountability” needed “to prepare, cross-check, and publish reports about the status and outlook of macroeconomic and fiscal figures” (EUBusiness, 2011). Simultaneously, Calmfors (2011, p. 22) notes that a proper reputation is attributed to the Economic Council in Denmark. Bjerkholt (2005) and Hendriks (2011, pp. 125-137) argue that the Norwegian and Dutch IFC enjoy similar reputations. Hence, these IFCs seem to enjoy broad public support for their activities.

An academic debate has emerged on the elements that constitute the reputation IFCs enjoy. Calmfors (2011) argues that “[i]n the long term, the influence of a fiscal watchdog [his label for IFCs] is determined by the quality of its work” (p. 22). Magaro (2013) argues that an IFC’s reputation is constituted by “independence, transparency and breadth of the analyses” instead of by quality alone (p. 75). In the end therefore, Magaro argues, the independent character of an IFC and its ability to resist political interference, is key in gaining a proper reputation. Kopits (2011) outlines that the degree to which an IFC is “home-grown and home-owned in every aspect” is key in the formation of an IFC’s reputation. Kopits therefore argues that in order to gain broad-based support an IFC must fit into a its institutional context in such a way that it conforms to a political culture and addresses local needs — a conclusion Debrun and Takahashi (2011) support. Hendriks (2011) elaborates on the institutional power of the Dutch IFC, the CPB. She demonstrates how the CPB is a facilitator of consensus-building within Dutch politics by providing politicians with a common language, and how the CPB is an enforcer of realistic demands when it comes to budgetary policies. At the same time, she demonstrates how the stakeholders of the CPB (e.g. governors) value the agency’s role of ‘outside referee’ that scores policies. This suggests that Hendriks regards that fitting properly into the institutional context is — at least within the case studied — a key constituting element of reputation.

Despite the potential validity of each of the arguments, a lack of systematic insight into what elements constitute the reputation IFCs have is afflicting the existing debate. This research addresses that issue. In doing so, it has a fourfold purpose. The first aim is to theoretically elaborate on all the elements that are potentially relevant in constituting the reputation of IFCs. The second aim builds on the first and embodies the purpose to identify what elements are relevant in constituting the reputation of IFCs. As it is not expected that one of the elements that will be identified is the sole constituent of an IFC’s reputation, this research thirdly aims to analyse which of the relevant element interact with each other and how they do so in constituting an IFC’s reputation. Based on this aim and all the previous ones, the research’s fourth and final aim is to provide for insight into why the core constituting elements are as relevant as they are.

As it is expected that the way the different elements interact with each other in constituting the reputation of IFCs cannot be separated from the underlying factors that constitute the way in which these elements are relevant, this research has the aim to comprehend the elements and factors together in what will become this research’s main assessment. To that end, the following research question has been formulated:

What components shape the reputation independent fiscal councils enjoy?

1.1 Contributions

This research aims to make a substantial contribution to the existing academic debate on IFCs in two ways. The first contribution stems from this research’s aim to not only identify the relevant elements that constitute an IFC’s reputation, but to also take into account how these elements interact with each other in doing so. The existing arguments on the elements that shape the reputation of IFCs do not take an elaboration on all the elements that potentially constitute this reputation as starting point. Consequently, these studies focus each in their own way on a rather narrow set of potentially constituting elements and do not take into account how distinguishable elements, such as quality and institutional context, interact with each other in constituting the reputation an IFC enjoys. As it is likely that none of the elements that will be identified is the sole sufficient constituent of an IFC’s reputation, taking into account their interaction is of primordial theoretical importance. By providing for a theoretical discussion that incorporates all the elements that potentially shape an IFCs reputation (first aim), this research is able to take into account the

interaction of these different factors in the empirical section (third aim). This is the research’s first academic contribution.

The second academic contribution stems from this research’s aim to enrich scholarly insights on what makes IFCs successful in their job. Although there is, roughly speaking, scholarly consensus that IFCs help to contain deficit biases, it is at the same time argued that the way in which IFCs do so is rather subtle and varies over context (Debrun and Takahashi, 2011; Fabrizio and Mody, 2006). As Debrun and Takahashi (2011) note, this subtlety complicates statistical attempts to identify what features of an IFC contribute to what extent to the IFC successfully doing its job. As soft power is a main driver behind an IFC being successful and reputation is the vehicle via which IFCs can gain soft power, this research inherently gives insight into the factors that shape an IFC’s successfulness. In contrast to existing research however, it does so with a qualitative approach that enables to make far more detailed within-case observations than statistical research could do. This allows one to track and trace the subtlety behind the successfulness of IFCs — subtlety has not been captured by statistical research yet. Providing for new insights on which features make IFCs successful and what not is therefore the second academic contribution this research aims to make.

Despite the importance of contributing academically, this research has been set out to make a policy-oriented contribution. As will be discussed later on, accounts into the history of some particularly successful IFCs demonstrate that the decisions taken by an IFC’s first director can thoroughly determine the organisation’s path later on (see e.g. Joyce (2011) and Den Butter (2010)). When the U.S. Congressional Budget Office (CBO) was founded, contested discussions for instance arose on who the CBO was to serve; budget committees of Congress preferred full control of its agenda. If Alice Rivlin, CBO’s first director, would not have rejected this proposal as firmly as she did, then, Joyce (2011) argues, the narrative on CBO’s nonpartisan status would never have been so dominant today. As will be discussed, the historical accounts on the IFCs (that contribute to this research’s fourth aim) demonstrate that the room for agency in successfully bringing about these elements, is significant. Taking into account what elements are relevant in constituting an IFC’s reputation therewith helps policymakers who serve IFCs, but also those in charge of other institutions of expertise that are for their legitimacy dependent on a good reputation, with an insight in where their agency should be directed to in order to enhance the success of these institutions.

This research proceeds as follows. In the next chapter, the theoretical fundament of this research will be laid out. From a discussion on the concept of reputation, the elements that potentially constitute an IFC’s reputation will be set out as different subhypotheses — this research’s first aim — and categorised under three hypotheses. The theoretical section is followed by the methodological chapter. Here, it will be motivated why this research has selected the Congressional Budget Office and the Netherlands Bureau for Economic Policy Analysis as cases. Subsequently, elaborated will be on the choice of this research to obtain a discourse analysis in three U.S. and three Dutch newspapers. The final part of the methodological section consists of an operationalisation of the elements as identified in the theoretical framework.

After the theoretical and methodological foundations of this research have been laid out, the results section will discuss the main findings of the discourse analysis and will answer the main question per selected IFC. This will be done in a way as to check for this research’s first aim, and to try and deliver on this research’s second, third and fourth aim. The conclusion will comprehend this research’s findings, answer the main question, discuss notable methodological limitations, set out a new agenda for academics, and give advise to policymakers where possible.

2. THEORETICAL FRAMEWORK

This chapter will lay the theoretical groundwork of this research. It starts with an elaboration on the concepts of IFC and reputation. From the discussion on reputation, a framework will be derived in which the identified elements that potentially constitute an IFC’s reputation will be embedded. Based on this identification — the first main aim of this research — three hypotheses will be proposed. As all the elements potentially relevant in constituting reputation will be discussed, a literature review will be integrated into the discussion underlying the hypotheses.

2.1 Concepts

Before elaborating on the concept of reputation, it will be defined what an independent fiscal council is. Principally speaking, any public institution with an (at least on paper) nonpartisan character that has the aim to inform and support the quality of the public debate on budgetary policymaking and the quality of budgetary decisions could be categorised as independent fiscal council (Debrun and Takahashi, 2011). Regularly, IFCs obtain these kind of activities by doing macro-economic forecasts, making cost-estimations of specific policies, calculating the costs and benefits of existing policies, assessing whether budgetary plans and processes are in line with procedural and numerical rules and, more generally, advising political actors on which policy interventions should be considered desirable and realisable and which not. The methodological section will provide for a specific operationalisation on what counts as IFC and what not.

2.1.1 Reputation as phenomenon

Reputation is a concept applied to a broad range of fields (for an overview, see Lang, 2015). To what field the concept is applied shapes how it is approached. Bourdieu (1992) sees reputation from a socio-cultural perspective, thereby defining the concept as a form of “symbolic capital validated against the perceptions of those in the know“; Craik (2009) applies reputation to social psychology, therefore highlighting how individuals develop reputations; management theorists approach reputation as being about “the arts of impression management” (after Goffmann, [1978] 2016).

As this research is about reputation of organisations, it follows the conceptualisation of Carpenter (2010, p. 33), that defines organisational reputation as the beliefs about an organisation’s intentions,

capacities, history and mission — beliefs that are embedded in a network of multiple audiences. Carpenter’s definition is built on Bailey (1971, p. 2), who considers reputation as not being about the intrinsic quality persons, organisations or institutions possess, but rather about the opinion others have of him/her/it.

The Carpenter and Bailey conceptualisations leave undiscussed how the reputation of organisations is formed, a gap filled by Craik (2009). He argues that often, reputation is not based on personal knowledge. Rather, an audience bases the reputation of an entity on what they hear others saying about it, until they accept a certain content of this reputation as ‘common knowledge’ or simply ‘true’. The relevant audience of an IFC is in this research demarcated as the group of people that is affected by the budgetary policymaking in which the IFC plays a role. The process of reputation formation, Craik argues, consists of three parts. A first necessary ingredient for earning reputation is visibility; the entity about which a reputation potentially exists should be known among the audience. The second ingredient is the reputation network; channels must exist through which information about a certain entity circulates. Based on these ingredients, and this is the third step, people select the part of that reputation that seems relevant to them, after sifting, correction and collation. Therefore, reputation should not be seen as static phenomenon; the process of forming certain reputations is a continuous one, and can change over time as people select new images to construct new common knowledges about the same entity.

2.1.2 Reputational categories

Above is described what reputation is and how it works as a phenomenon. This however leaves undiscussed a specification of the potential content of the reputation audiences attribute to organisations — a dimension endogenous to the concept as well. Such a specification enables to theoretically embed the elements that potentially constitute an IFC’s reputation into different categories. Carpenter’s elaboration on reputation (2010, pp. 46-47; with Krause, 2012) provides such a categorisation. He identifies four umbrellas that “comprise the structure of beliefs about an agency” (2010, p. 46). First, whatever the aim of the organisation, its performative reputation is about the audiences’ judgement of organisation’s capacity for effectively achieving the goals it has set. Second, an organisation’s moral reputation is about whether the IFC is considered to have ethically and morally defensible means — how it for instance demonstrates compassion for people less fortunate. Third, an organisation’s technical reputation reflects whether the organisation is seen as bearing the technical, analytical and methodological capacities to properly fulfil its tasks. Fourth, an organisation’s legal-procedural reputation is about “the justness of the process by which its

behaviour is generated” (p. 47); whatever its quality or performance, did the organisation follow the desired procedures when doing its job? As the four umbrella’s inherently overlap with each other (e.g. technical reputation provides for performative reputation), the aim of this research is to categorise the different subhypotheses under one umbrella if possible and more if needed.

Carpenter’s categorisation demonstrates that in theory, different characteristics of an organisation can be represented in the reputation an audience attributes to that organisation. The differences between categories can be subtle, but significant. For example, an audience can perceive the work of an IFC to be of good quality (technical) and value its work as effective (performative), but challenge the admissibility of that same IFC as it perceives the role that is assigned to the body (legal-procedural) as unfitting.

Although this research should leave open the possibility that in reality, a reputation attributed to an organisation is heterogenous and reflects an audience’s viewpoint on a variety of characteristics or dimensions of an organisation, an audience tends to attribute to an organisation a simplified frame. Why does an audience do so? As Carpenter and Krause (2012) argue, a trade-off will likely exist in the attention the audience can direct to the four dimensions of reputation. Although not intrinsically at conflict with each other, an audience’s focus on e.g. an organisation’s technical reputation might thus come at cost of an audience’s attention for the organisation’s moral or performative characteristics. Only in a simplified (or even unitary) way, the argument goes, a reputation is clear-cut enough to be preserved as common knowledge within an audience.

It is, as mentioned, not expected that one of the elements identified is the sole sufficient constituent of an IFC’s reputation. Rather, several relevant elements will likely be in play that, together with factors that explain the relevance of these elements, form the components that shape the reputation an audience attributes to an IFC. This should not imply that anything goes. Instead, as these elements interact with each other in a certain way, it is expected that the relevance of some of the elements is rooted in the relevance of others. Taking into account the interaction of elements and factors thus enables to identify those components that are crucial in constituting other relevant ones, and therewith in constituting an IFC’s reputation. For example, if an IFC is regarded as providing for political needs (a H2 subhypothesis) while it is seen as body of expertise (a H1 subhypothesis), then it can be logically deduced as well as empirically identified that the first is an effect of the latter. After taking into account this H2-H1 interaction, it can be demonstrated that the image of the IFC as body of expertise is, in the case of the example, the core constituting element of an IFC’s reputation.

The simplified framework an audience will likely attribute to an IFC is therefore expected to consist only of those components that are crucial in constituting an IFC’s reputation — ‘expertise’ is the simplified framework in the example. What reputational category constitutes the reputation of an IFC will thus be taken into account after it is identified how the relevant elements and factors interact, and what elements are essential in constituting reputation.

While it is likely that the reputation an audience will attribute to an IFC is a simplified characterisation, real simpleness is in fact absent — an organisation’s characteristics of expertise, performance, and morality are heterogenous and cannot simply be lumped together. “Nonetheless”, Carpenter (2010, p. 50) emphasises, “the fiction of a unitary organisation gives rise to certain realities of its own”. Reputations should therefore not only be seen as an interpretation or construction of a certain reality. As reputations shapes the input- and output-legitimacy with which IFCs contain deficit biases, they construct new realities as well.

As reputation has real consequences even if it is an oversimplified interpretation from reality, the perception of an audience is all that matters and all that should be studied when identifying the elements that constitute reputations. This research will therefore focus on perceptions of audiences.

2.2 Elements constituting reputation

In the following section, three hypotheses will be identified that each represent elements that potentially constitute an IFC’s reputation. From each of the three discussions, a hypothesis will be derived. The first two hypotheses are broad in scope; these consist of several elements (or subhypotheses, these are synonyms from now on) that in themselves can constitute an IFC’s reputation but can content-wise be grouped together. The hypotheses in which several elements are grouped together leave room for an empirical discussion on the way in which some parts under those hypotheses might be relevant while others might not.

Each of the elements discussed under the three hypotheses has a one-on-one relationship with one or more of the above discussed reputational categories. If it thus turns out that a relevant element plays a role as core constituting element — which is identified after taking into account its interaction with other elements —, then one can deduce that the category of reputation it is subjected to is present in a certain way. If, for instance, an audience perceives the quality of the

work of IFCs to be high (see H1), then this automatically implies that the IFC has a proper technical reputation — the latter measured by taking into account the first.

Subsequently, if one (or more) of Carpenter’s reputational categories is identified as covering the core constituting elements of an IFC’s reputation, then this reputational category inherently translates into the overall reputation an IFC enjoys — a one-on-one relationship as well. A good technical reputation implies that the organisation to which that proper quality is attributed enjoys a good reputation — when other categories do not inhabit core constituting elements and a trade-off as discussed indeed turns out to be present. Quality is then, within the technical reputation category, the constituting element. If empirical material indicates that, even after taking into account the interaction between different elements, such a trade-off is absent, then the reputation attributed to an IFC is constituted by multiple reputational categories.

The to be identified elements constitute the relevance of a reputational category in a one-on-one way. However, it is expected that they can do so via two mechanisms. First, elements can directly influence an audience’s viewpoint on that IFC and so the reputation that audience attributes to the agency. Second, elements can influence the viewpoints and acceptance of that IFC within political or other influential domains. This translates into more positive or negative messaging from these domains on the IFC’s legal-procedural (“we are delighted with its work”), performative (“it is doing a great job”), and/or technical (“these people are the real experts”) reputation , which is expected to 6 shape the view of an audience on that IFC and hence the reputation the audience attributes to it.

Some of Carpenter’s reputational categories (such as performative reputation) are in the coming section linked to more than one of the potentially relevant elements. Which of the elements play a critical role in constituting those reputational categories and which not will be identified by taking into account the interaction of these different elements.

2.2.1 Perceived quality of activities

The first potentially relevant element constituting reputation is the perceived quality of an IFCs work. As this element is about whether it is perceived that an IFC bears the technical and methodological capabilities to properly fulfil its tasks, the hypothesis that will be derived from the discussion can be categorised under Carpenter’s ‘technical reputation’. As will be discussed below,

Morality is not mentioned here as it is the reputational category under which no identified element will be categorised.

it is expected that an audience can perceive the technical reputation of an IFC to be of certain height either because it sees the IFC as expert, or because it sees the IFC as independent. From these two arguments the two subhypotheses under H1 will be derived. The relevance of these subhypotheses will be taken into account separately.

The complexity of society, and hence the difficulty of designing proper budgetary policies, has increased in the last decades. As societal tendencies seem to become more unpredictable, budgetary policymakers face an expanding number of uncertainties. Several scholars (e.g. Hall, 1989; Blyth, 2002; Hendriks, 2011) have argued that as consequence, policymakers (in budgeting but also in other areas) become more and more reliant on experts. As such, it is argued, the role of expertise in politics is to an increasing extent applauded.

Given the increased appreciation of expertise, the first hypothesis proposes that the perceived quality of work of IFCs might be key constituting element of the reputation of IFCs. Calmfors (2011) explicitly emphasises that quality is the key element. If this quality is perceived to be high, he argues, so will be the IFC’s reputation; if lower, so will be an IFC’s reputation.

However, academic dispute exists on what an audience actually takes into account when it values the technical capabilities of an IFC. On the one hand, Calmfors (2011) argues that the audience’s perception on an IFC’s quality depends on the actual quality of that IFC’s work. This argument bears two assumptions; (1) audiences indeed examine the quality of work when valuing its expertise, and (2) this audience is epistemologically competent to make such an assessment . 7 Evaluating the validity of the latter brings one into deep psychological terrain and is hence beyond the scope of this research. The research will therefore be limited to evaluating the first assumption and will take into account whether technical abilities are relevant in an audience’s attribution of certain quality to an IFC.

On the other hand, Hagemann (2011) argues that an audience does not assess an IFC’s technical abilities when it values the quality of that IFC’s work. Instead, he argues, audiences perceive the quality of work of an IFC to be high when they see IFCs as independent and politically neutral, and this quality to be lower if the independence of IFCs is disputed or considered absent . Neutrality 8

Assessing the quality of an IFC in a certain manner brings, as discussed, real consequences. These consequences however lie in the

7

legitimacy of IFCs, not in their quality — the latter does not change as consequence of changing perceptions about it.

As perceptions are the only phenomena to be studied, it is irrelevant for evaluating this subhypothesis whether IFCs are actually

8

should in this context be seen as indicator for the perceived ‘scientificness’ of an IFC’s work, which, if scientificness is related to quality, indicates how an IFC’s quality is considered.

If quality indeed turns out to play a role, the aim is to assess what explanatory role both subhypotheses discussed play in this. For that assessment to be made however, the following hypothesis first has to be confirmed:

H1. The reputation an IFC enjoys depends on how the relevant audience values the quality of the IFC’s work.

2.2.2 Reputation and institutional context

The second hypothesis proposes that the extent to which an IFC fits properly into its political-institutional context is the key element for an IFC’s reputation. Four ways in which an IFC is institutionally aligned are identified. As each of the specific subhypotheses differ substantially, these, instead of the overarching H2, will be linked to Carpenter’s reputation categorisation. It is therefore not only the aim to take into account whether an IFC is perceived to be institutionally aligned to its context; the relevance of each of the four elements will be taken into account separately as well.

Kopits (2011) and implicitly also Hendriks (2011) argue that if an IFC fits well into its context, it will bear a proper reputation; if not fitting well an IFC will not. If a match between an IFC and its institutional context is perceived to be missing, Kopits argues, the audience is likely to regard the IFC as an ‘externally imposed alien body’. A good reputation, he argues, can therefore not be imported from abroad — i.e. supranational players as the European Commission and OECD pushing for adjustments in institutional configurations. With centralising institutional alignment Kopits does not argue that quality is completely irrelevant; key however is the existence of societal consensus on the position of an IFC. Debrun and Takahashi (2011) support this argument.

Kopits lists three mechanisms that can constitute whether an IFC is properly fitting in its institutional context . First, an IFC fits institutionally properly if an audience perceives the body as 9

It is phrased as “properly fitting in its institutional context” instead of “perceived to properly fit…” because the extent to which an

9

IFC matches with its institutional context is intrinsically, by nature, an interpretative phenomenon. If an audience perceives this institutional match to be good, than that match is already good in itself. The quality of an institutional match is shaped by how an audience perceives that quality to be. This is different with elements as ‘quality of work’, as it is arguable that one’s analysis will not become better if an audience values its quality more.

competent in addressing and challenging a lack of transparency in the budgetary decision-making process. For this mechanism to be relevant, the lack of transparency must be an issue the audience bothers about (1) while the IFC must be seen as bearing (input- and/or output-) legitimacy in addressing and challenging this situation (2). As an audience will value the IFC’s capacity to effectively enhance transparency when both are present, the transparency-mechanism can be categorised under performative reputation.

Second, an IFC fits properly in an institutional context if it is perceived to enhance discipline within budgetary politics. Discipline should therefore be an issue the audience bothers about, and the IFC should be seen as legitimately enhancing it. As with transparency, an IFC will be valued positively when it bears the capacity to do so effectively. The discipline-mechanism can thus be categorised under performative reputation.

Third, an IFC fits properly institutionally if it provides for political needs — for insights that a budgetary debate would otherwise not have obtained. It is expected that when an IFC provides for insights considered of additional value, its work will be appraised; if its work is seen unnecessary or even otiose, the institutional fitting and, consequently, reputation of an IFC is hindered. If an IFC provides for necessary insights, the process that generates its tasks will be seen as justifiable. The (un)necessity mechanism will thus be categorised under an IFC’s legal-procedural reputation.

A fourth, not by Kopits listed element can be identified. As touched upon, Magaro (2013) argues that the independent status of an IFC and its ability to resist what is seen as political interference is key in gaining a proper reputation. An emphasis on independence and neutrality is in this context not related to and a mean for the perceived degree of scientificness of an IFC (translating into a certain technical reputation, H1); it is considered separately from that and is instead seen as end on its own. That a nonpartisan status is seen as end itself demonstrates that this status is seen as distinguishing IFCs from other institutions — after all, if all institutions would be considered independent it would not be worth mentioning anymore. The presence or absence of nonpartisanship is emphasised because it demonstrates that an IFC fits properly within a certain institutional context. The emphasis on neutrality is then a mechanism that justifies the role attributed to the IFC. Hence, this subhypothesis is categorised under the legal-procedural category.

H2. The reputation an IFC has depends on the degree to which that IFC is aligned with and connected to its institutional context.

2.2.3 Influence as driver

The third hypothesis is not yet identified in the literature as potential constituting element. It builds on the idea that reputation not only shapes influence, but that influence is instead also an element that constitutes a reputation — a reciprocal relationship. This hypotheses will not be split into several subhypotheses, and is, as will briefly be elaborated upon, part of the performative type of reputation as it reflects how an audience’s judgement on the capacity and effectiveness of an IFC constitutes reputation.

IFCs that bear a proper reputation are inherently influential. If influential, IFCs can, as argued, effectively do their job. This effectivity can in turn shape the judgements of that audience on the IFC. Although an audience might assume that an IFC is influential because of the proper technical work it does or might justify the legal role of an IFC because of its influence, it is most likely that an audience will value the effectivity of an IFC positively when taking into account its influence — this effectivity is, as opposed to its technical abilities, procedural role and moral standards, directly generated by an IFC’s influence. Likewise, a lack of influence might inflict that an audience assumes and perceives the effectivity of an IFC to be low — rhetorically wondering: “why would it otherwise have a low influence?”. Influence is thus not only rooted in an IFC’s reputation, it can re-shape that reputation as well. If so, the relationship between influence and reputation is a reciprocal one.

In short, the third hypothesis proposes that the (absent or present) influence of an IFC as sole element constitutes the judgements of an audience on how effective the IFC is. This subsequently shapes an IFC’s performative reputation. The hypothesis reads:

H3. The reputation of an IFC is rooted in whether an audience perceives that IFC to be influential.

It is expected that in the above section, all the elements potentially constituting reputation have been identified. In order to make sure that this research’s first aim is indeed achieved, it will be empirically checked whether this is indeed the case. How that is done is set out in section 3.3.5.

3. METHODOLOGY

This section will discuss the methodological underpinnings of this research. These underpinnings enable to systematically analyse in the empirical section what elements are relevant in constituting reputation, how these different elements interact with each other, and why the core constituting elements are as relevant as they are — this research’s second, third and fourth aim. This section starts with a discussion on the research’s case selection. The case-selection part will be followed by an elaboration on the methodological approach to gathering the within-case empirical material. Finally, attention will be paid to how then the different elements will be identified in the data gathering; the operationalisation of the relevant elements.

3.1 Case selection strategy

Section 3.1 consists of two parts. In section 3.1.1, it will be set out which IFCs are selected. The subsequent section will provide for a motivation for why these cases have been selected.

3.1.1 Which IFCs?

In order to select IFCs, the concept of IFC needs to be operationalised first. When can an organisation be categorised as IFC and when not?

In order to be categorised as IFC, an institution should fulfil a specific set of tasks. It should (1) obtain macro-economic forecasts, (2) forecast the revenues and expenditures of a government, (3) substantiate future expectations on budgetary deficit- and debt-levels, (4) be involved in independent analyses on fiscal policies, for instance by estimating short- and long-term effects of measures, and (5) evaluate whether budgetary plans and outcomes are in line with set fiscal rules.

Three extra conditions an organisation has to meet to be qualified as IFC will be proposed (the latter two based on Debrun et al, 2009). First, as private bodies cannot be seen as parties that officially underlie fiscal policymaking, private bodies cannot be counted as part of configurations of fiscal governance (see the Ayuso-i-Casals et al (2006) definition). Therefore, these will not be regarded as IFCs. Second, international agencies (e.g. OECD, that aims to improve budgetary debates) are excluded; their mandate is often too broad and has too little practical application in order to be

influential, while they lack a clear audience that constructs that organisation’s reputation either. Third, audit institutions are excluded. Their focus is often backward- rather than forward-looking, and they have a primordially legal instead of economic character.

There are a some dozens of organisations passing the above operationalisation (see e.g. Von Trapp et al, 2016). As identifying which elements constitute the reputation of IFCs in what way is a subtle operation, a detailed enquiry per IFC is required. Two IFCs will therefore be selected: the U.S. Congressional Budget Office (CBO), and the Netherlands Bureau of Economic Policy Analysis (Centraal Planbureau or CPB). As these IFCs play a role in American and Dutch budgeting, all citizens affected by the budgeting of each of the IFCs are considered the relevant audience of that IFC .10

The CBO (founded in 1974) is the independent analytical servant of the legislative branch of the U.S. federal government (Congress) when it comes to budgeting issues. In serving Congress, the CBO obtains (1) macro-economic forecasts, (2) forecasts on federal expenditures and revenues, and (3) analyses on long-term budgetary sustainability. Moreover, it (4) provides Congress with estimations of the (financial) costs and benefits of policy proposals (Joyce, 2011; Von Trapp et al, 2016, p. 19), and (5) monitors whether budgetary plans and outcomes are in line with set fiscal rules (Joyce, 2011, pp. 108-111).

The CPB (founded 1945, legal basis since 1947) is publicly mandated to engage in the same five activities as just listed for the CBO (Hagemann, 2011, p. 83). The CPB is, additionally, involved in a broader range of activities, such as analysing and evaluating the budget to close the budgetary year and estimating the costs and benefits of policy proposals in party programmes during national election time (Hagemann, 2011; Von Trapp et al, 2016). Officially, both the CBP and CBO refrain from presenting their own normative viewpoints in discussions (Hagemann, 2011; Joyce, 2011).

3.1.2 Why these IFCs?

The existing literature on IFCs suggests that the CPB and CBO both enjoy a very good reputation. Following a Dutch newspaper, Hendriks (2011) observes: “[t]he CPB not only provides the facts and figures on the basis of which actors might discuss their policy preferences, but these policy preferences themselves are also shaped by CPB analyses. Political actors are expected to talk in

That excludes Americans and Dutch people living in foreign countries, and includes citizens from other countries living in the U.S.

10

CPB language when the economic effects of their policy plans are concerned. Policy plans that are ‘CPB-approved’ are taken seriously, because these plans are viewed as more robust, economically sound, responsible, and are ‘therefore not unrealistic’” (p. 133). The CBO seems to enjoy a rather good reputation as well. In 2011, Joyce argued that the “CBO has become a powerful and influential arbiter of economic and budgetary effects of policies” (p. 1; cf. Smith, 1988, p. 290).

Following the logic that differences cannot explain similar reputations, selecting the CPB and CBO is expected to enable to control to some extent for general ‘background noise’; the relevance of (system and sub-system level) factors that might play a role in the agencies’ operations in which the CPB and CBO differ (e.g. that the CPB closes the budgetary year while the CBO does not — the list of differences will be long) can likely be eliminated in this research, as well as in answering the main question regardless of the systems within which observations are made (Przeworski and Teune, 1970).

There is, however, a more specific reason to select the CPB and CBO; the country-specific context the two agencies are operating in potentially influences the mechanism via which relevant elements constitute the relevance of reputational categories and hence reputation of IFCs. Selecting the CBO and CPB enables to control for that influence.

As discussed theoretically, it is expected that elements constitute reputation via direct and indirect mechanisms. The country-level political context in which an IFC operates is likely to influence the indirect mechanism. As Katzenstein (1985, pp. 88) argues, whether institutions of expertise are accepted and valued positively or not can crucially depend on the degree to which a country’s political habit is to seek multi-party consensus or instead augment political differences. Accounts on the role of expertise in Dutch politics seem to support the Katzenstein argument. Hendriks (2011) touches upon a widely shared viewpoint when she argues that “Dutch politics is characterised by substantial cooperation and coordination between political actors, predominantly between government, employers’ associations and unions” (p. 61). As “[i]t appears easier to come to an agreement, when policy wishes are stripped from their political or ideological charge […]”, Hendriks argues, “Dutch political actors across the political spectrum seem indeed to prefer a practical, utilitarian approach to ideological assumptions” (p. 314). The pragmatic approach central to the consensual game subsequently paves the way for a large role of expertise within the political domain. As Wilts (1997) argues, expert knowledge is regarded instrumental “in finding the necessary political consensus by transforming the problem of having to choose between ideologically tainted political options into practically manageable policy issues” (p. 201).

In short, experts are particularly appreciated within more depoliticised political contexts as they facilitate the consensual approach by providing for an understanding of what the common interest implies. It is less likely that experts as IFCs will enjoy the same appreciation within deeply ideological contexts. The degree of polarisation might thus influence the way politicians talk about IFCs, which is part of the indirect mechanism via which an audience attributes to an IFC a reputation. Controlling for the degree of (de)politicisation in a country’s political domain is therefore necessary.

On the political-institutional axis that places partisan extrapolation and polarised behaviour on the one end and (multi-party) consensus-seeking behaviour on the other, the United States and the Netherlands are polar opposites . As demonstrated, the Dutch political game seems to be played 11 according to consensual rules. When it comes to seeking consensus, the United States has a rather different political climate. There is large scholarly consensus that over the last decades, the American political climate has become increasingly polarised, as deep cleavages between the partisan lines of Democrats and Republicans have arisen . Political differences are therefore 12 extrapolated. Jacobson (2013) for instance notes that “[p]artisan disputes over matters large and small, personnel as well as policy, occur almost daily. Polarised parties, combined with divided government, have made legislative gridlock the normal state of affairs in Washington, overcome only when dire necessity compels short-term compromises to stave off such disasters as default on the national debt or a government shutdown” (p. 688).

The CBO and CPB seem to enjoy roughly the same reputation but differ on the system-level variable of (de)politicisation. This enables to control for a potentially important system-level variable. This research thereby follows the logic of the most different systems same outcome strategy (MDSO, see for a discussion Anckar, 2008), which can logically speaking be seen as part of the broader most different systems design approach (MDSD, see Przeworski and Teune, 1970).

No measure exists yet that scales the consensus- or polarisation-seeking behaviour of Dutch vis-à-vis American politics in a

11

consistent index. However, Engelen et al (2010) apply an index of corporatism to the Netherlands, U.S. and some other countries based on the methodology of Hicks and Kenworthy (1998). The index is comprised out of ten indicators, of which only two are somewhat indicative for the degree in consensus-seeking behaviour in the political domain: the extent to which wage bargaining is coordinated and the level of cooperation between the state and interest groups. Despite the limited indication this index gives, it is noteworthy that the Netherlands scored .578 on a scale from 0 (full polarisation) to 1 (full consensus) in 2005, while the U.S. scored 0.023 in the same year. Although in a limited way, this seems to support the argument that the U.S. and Netherlands differ substantially when it comes to consensus-seeking.

This is at least the case for the federal level, the relevant level of U.S. governance in the context of this research.

As for a major system-level variable is controlled, the case selection approach implies that explanatory variables residing at the sub-system level are taken as the analysis’ starting point (Przeworski and Teune, 1970). Since perceptions on expertise, institutional alignment and influence are all elements that can constitute an IFCs reputation regardless of the system within which these perceptions are present, the elements identified all reside at the sub-system level. The selection design thus enables to devote primordial attention to evaluating these elements’ relevance. That these elements reside on sub-system level moreover enables to test their relevance in other cases, regardless the specific features of these cases (assuming that relevant elements constituting reputation are indeed found in this research).

Even when the relevance of one or more of the identified elements is confirmed, the relevance of numerous other variables can not be ruled out; the CPB and CBO will anyway share a number of characteristics. Logically speaking, these (some not further specified) variables are potentially constituting the reputation of IFCs as well . Following Przeworski and Teune (1970), although 13 confirmation of one or more of the hypotheses strengthens confidence that these instead of other variables are relevant as explanatory force, no rigorous inference on this proposition can possibly be made. Independent of this research’s outcomes therefore, further research should be directed toward testing the influence of the identified element(s) in other contexts.

3.2 Gathering empirical material

This section will discuss what empirical material will be gathered with what method.

3.2.1 Discourse analysis

As discussed, meaning is given to what an IFC is via reputation. As the concept of discourse can be defined as an identifiable set of produced and reproduced practices that entail “an ensemble of ideas, concepts and categories through which meaning is given to phenomena.” (Hajer, 2002, p. 63), the meaning an audience attributes to an IFC (its reputation) should be seen as set of discursive practices. In order to identify the meaning-making that constructs reputation therefore, discourses should be studied.

As Von Trapp et al (2016) for instance demonstrate, although the CBO (246 fte in 2011) and CPB (117 fte in 2012) differ in their

13

staff size, they both have a staff of serious size. Although it is unlikely that having a sufficiently large staff is a sufficient explanatory factor for reputation, it might be a necessary one. Whether this is the case can however not be examined within this research’s selection strategy.

This can be done by (1) surveying this audience on the reasons they have to attribute to a certain IFC a reputation, and (2) obtaining a discourse analysis (DA). As the formation of reputation is a process that exists only by the virtue of being permanently in the making, this research should leave open the possibility that the discursive practices constituting reputations are subject to change over time. In order to accurately capture not only the discursive practices that constitute reputations today, but also those that constituted reputations in the near past, the bias in data that exposes these near-past discourses should be reduced to a minimum. This is guaranteed for only if the data of the near past is not influenced by today’s discursive practices — a bias that might arise when obtaining interviews. As this bias is not present in studying words that already have been uttered, this research will obtain a DA.

Over the years, DA has become an overarching term for a large variety of specific methods to study discursive practices (see e.g. Bacchi, 2009; Hajer, 1995, 2006; Billig, 1987). Without taking a previously obtained DA as precedent, the aim of this research is to follow a ‘context-acknowledging’ approach to studying discursive practices. This means that studying words is considered central to identifying the process by which an audience gives meaning to an IFC, while these words should simultaneously be studied within their relevant context. Below, it will be discussed what words should be studied, section 3.2.5 will elaborate how these words will be analysed within their context.

As discussed, reputation formation of an IFC takes place through channels beyond the ‘known acquaintances’. The way some of us speak thus shapes the view of the world (reputation in this case) a much larger group develops. Given that the work of IFCs and the issue of budgetary policymaking is rather complex by nature, it is expected that the way in which meaning is given to the work of IFCs in media has generally large authority over how a larger group perceives the reputation of that IFC to be . The assumption is that it works the other way around as well; with no 14 or hardly any media attention, the work of an IFC will soon be forgotten within an audience. Media are therefore the most important vehicle via which the speakings of some shape the perceptions of a larger audience. As newspaper texts enable to study discourses in a more systematic way than visual media would enable to do, texts from a selection of Dutch and U.S. newspapers will be analysed.

Studying newspaper texts enables to identify both the direct and indirect mechanism via which elements constitute reputations. In newspapers, audiences both receive information on the activities

In contrast with for instance the game of football or political figures like Donald Trump; issues about which everyone seems to

14

of an IFC (direct mechanism) and read how politicians and other relevant actors comment on these issues (indirect mechanism).

Some academic and non-academic sources distinguish different types of newspaper texts . As these 15 categorisations are either too minimal or beyond what suits this research, this research distinguishes five different types of newspaper texts independent from earlier work: (1) ordinary newspaper articles that report on an event, item or discussion, written by a newspaper’s editor, (2) an interview (often with a well-known person) hosted by the newspaper, (3) columns via which appointed authors comment regularly on issues, (4) opinion articles in which an expert, well-known individual or organisation, or otherwise engaged person argues for a certain viewpoint, and (5) submitted letters via which readers comment on certain news issues.

Generally, the content of newspaper texts has authority over the viewpoint of the reader. In cases where the authority of the author and/or participant is absent, it is expected that the content of newspaper texts has no significant influence on the viewpoint and perceptions of the reader. In analysing relevant newspaper texts therefore, only the types of newspaper texts that contain such authority will be analysed. As this authority is based on the perceived qualifications of the author and/or participant(s) in the text, this research will analyse the first four types of newspaper texts (following the conviction that authority is present here). Submitted letters are excluded as authority is likely to be absent in such texts.

3.2.2 Newspaper selection

This section will discuss which newspapers are selected in the U.S. and Netherlands. It will provide with a motivation for the procedure underlying this selection, and will set out criteria newspapers have to meet in order to be selected. The section will end with an elaboration on the procedure of the selection of texts within the selected newspapers.

Morton (1986) distinguishes seven types of local newspaper texts. This categorisation is however based on the content of the

15

newspaper articles (e.g. whether it serves consumers or whether coming or past events are described). The distinction does not fit for this research as no selection on topic is needed because the articles selected are on the CBO and CPB. Roberts (2016) distinguishes three types of news stories; straight or hard news, featured news (which is more subtle in tone) and editorials. This however is a categorisation of types of news instead of types of articles. The BBC (2014), a non-academic source, distinguishes three types of newspaper texts; news articles, feature articles and editorial and opinion articles. This categorisation is too minimal for the purposes of this research. Some other academic sources (e.g. S. Carpenter, 2010; Hamdy and Gomaa, 2012) briefly note on the type of newspaper texts they select (e.g. editorials, opinion articles, columns) but do not make an explicit distinction between the different types of newspaper texts available.

In both the United States and the Netherlands, the three largest daily newspapers will be subjected to a discourse analysis. Selecting the largest newspapers is based on three motivations. First, although not all citizens affected by an IFC’s work (the relevant audience) will read (the largest three) newspapers, selecting the largest three newspapers provides for a relatively generalisable picture of the reputation the audience as a whole attributes to that IFC. This is the case for two reasons. First, widely-read newspapers are simply more representative for the reputation the relevant audience attributes to IFC’s than newspapers that are read only within a smaller circle. Second, selecting the largest three newspapers enables to capture — if necessary — the differences in the reputation different parts of the population attribute to an IFC. As will be discussed, the three newspapers selected in each country differ considerably in the type of reader they focus on; some newspapers are seen as focusing on the elite or insiders, while others focus on the so-called ‘middle-market’ reader.

Second, as can be derived implicitly from the discussion of the hypotheses, it is not expected that political-ideological signature is a relevant element in constituting reputation. The signature of newspapers does therefore not function as selection criteria. Third, selecting three newspapers for each country allows to triangulate the findings in each newspaper with those of two other newspapers. If, for whatever specific reason, a newspaper is an outlier in its discursive practices compared to the other two of a country, then selecting three newspapers prevents one from generalising the discursive practices of that one newspaper as representative for practices across all newspapers in a country.

On the one hand, the selection of the three largest newspapers might not represent the content of texts of all newspapers, as more outstanding types of newspapers are not the ones with the largest circulation. On the other hand however, this bias is irrelevant, as the texts subject to the analysis should represent the perceptions of large groups in society, instead of reflect the content of a large number of newspapers (including less-read ones).

Before proceeding with selecting three daily newspapers per country following a list of the size of circulation, every newspaper that is eligible for selection must a priori meet two conditions. First, only newspapers read on a nation-wide scale will be selected — newspapers that serve a local public will not be included in the search. The content of national newspapers serves readers from all over the country and is therefore with more certainty representative to the reputation-formation of a country-wide audience than the content of local newspapers. Second, the newspapers selected must be available as source within the LexisNexis Academic database, for two reasons. First, it prevents