Introduction

Political struggle is about the negotiation of status quo versus change. Be it a discussion about pension reforms, the fight against racism or a vote of no confidence about a government: It all boils down to the same question. Is the status quo acceptable, or do we need change?

In situations where the question of status quo versus change gets onto the political agenda, groups invested in challenging or defending the status quo need to convince the public and gather parliamentarian majorities. It is known from social psychology research that individuals do not always decide on rational grounds when it comes to potential losses or gains. Therefore, groups trying to persuade others of their preferred option need to consider carefully how they present potential gains or losses that result from either implementing a policy change or maintaining the status quo. While social psychology research—more specifically, prospect theory (Kahneman & Tversky, 1979)—allows for the deduction of some practical advice for effective political communication (Brie & Dufresne, 2020; Hameleers, 2021; Soroka & McAdams, 2015), this advice does not include how precisely arguments need to be told. Yet, argumentative strategies are required in political debates to cater to the needs of different audiences.

The Narrative Policy Framework (NPF) is an upcoming approach in policy process research that analyses the use of narratives in political communication (Shanahan et al., 2018). The NPF emanates itself from insights in the field of cognitive psychology, which highlight the relevance of narratives for human thinking (Berinsky & Kinder, 2006; Jones & Song, 2014). Narratives are a very popular heuristic to both memorize and relate information (Jones & McBeth, 2010). With narratives, we assemble complex information in a captivating form, simultaneously imbuing this information with normative evaluations and establishing connections between otherwise isolated aspects (Stone, 2012). Thus, narratives are an adequate and often-used tool to formulate political arguments (Jones & McBeth, 2020; Schlaufer, 2016). As narratives may be populated with a multitude of different characters, such as heroes, villains, or victims, and as they may be equipped with further elements such as a setting, a plot, or a moral of the story, narratives allow for the crafting of many different arguments centering around a given political preference.

The question of status quo (SQ) versus policy change (PC) has already been addressed in recent NPF studies (Schlaufer et al., 2023). However, although these studies show that this distinction is important for political communication, they remain unclear about which mechanisms are at play to understand different usages of narratives. Hence, we contend that both the NPF and prospect theory may benefit from their mutual integration. In this article, we propose such an integration to find out how exactly SQ and PC are recounted in public debates.

In the next section, we introduce both the NPF and prospect theory, and we derive theoretical conjectures for an empirical application. We present the case and research design for this application in Sects. "Farm animal welfare: a contested policy domain" and "Research design", namely the examination of farm animal-welfare policy debates in France and Germany in 2021. Animal-welfare policy is a well-suited case for our research endeavor, as the policy field features well-established and clearly demarcated coalitions supporting either the SQ or PC. The two coalitions stand for different conceptions of agricultural policy, known as “exceptionalism” and “post-exceptionalism” (see Sect. "Farm animal welfare: a contested policy domain"). Animal-welfare policy can thus be considered to be a typical case of a policy conflict, which is an appropriate case selection to test the empirical validity of a new theoretical argument (Gerring, 2008, pp. 648–650). By extending the examination to two countries, we render the analysis more robust and allow for empirical nuancing. France and Germany share many similarities but differ regarding the amount and restrictiveness of animal welfare policies. The data consist of newspaper articles in four major French and German newspapers. The articles were manually coded for narratives according to a standardized NPF procedure and subsequently differentiated according to whether they refer to the SQ or a PC.

To test our conjectures, we rely on descriptive statistics supported by qualitative interpretation. The findings are presented in Sect. "Findings". We find that the way SQ and PC are recounted in narratives depends on narrators’ preference for either the SQ or PC. For example, the animal welfare coalition in Germany generally depicts policy change as a means to improve the life of farm animals, thus transforming them from victims to beneficiaries of the agricultural system. By contrast, the agricultural coalition in France denigrates visions of policy change by stating that it will generate many costs for farmers who are already suffering from an economically harsh environment. In Sect. "Discussion", we relate our findings to existing research and formulate suggestions for future studies. The article ends with a conclusion (Sect. "Conclusion").

Theory

Foundations of the Narrative Policy Framework (NPF)

The Narrative Policy Framework (NPF) is a theoretical lens that has received growing attention within policy process research in the past years (Husmann, 2015; Ruff et al., 2022; Terlizzi, 2021). Deviating from discourse analysis and interpretivist approaches to narratives (Fischer & Forester, The Argumentative turn in policy analysis and planning 1993; Stone, 1988) that often conceptualize narratives as unique, NPF proponents claim that narratives are measurable and comparable. Trying to render a framework on policy narratives “clear enough to be wrong” (Jones & McBeth, 2010, p. 331), the NPF contains clear-cut definitions of narrative components and features a set of hypotheses that are to be valid across case-specific contexts (Jones et al., 2023). Being convinced that such an approach towards narratives is fruitful, our analysis builds upon the NPF. The NPF starts from the basic assumption that policy narratives are central to the policymaking process at different stages, from the agenda setting to the termination phase. For example, a convincing narrative may be essential for policy entrepreneurs to put a policy on the agenda (McBeth & Lybecker, 2018), or the existence of powerful narratives can make the difference whether a policy gets adopted or rejected in a popular vote (McMorris et al., 2018).

Though policy narratives have been studied from different disciplinary perspectives already for a long time, the NPF aims to provide a systematic approach for the empirical study of these narratives. Policy narratives are an expression of different social constructions of reality that influence public policymaking. These constructions are bounded by belief systems, social norms, or political ideology (Shanahan et al., 2017). Actors strategically use such narratives to voice their policy preferences and influence the policy process (Jones & McBeth, 2010). In the definition of the NPF, a policy narrative includes four core elements (the “narrative form”) that also guide the empirical analysis:

  1. 1.

    Setting: Situates the narrative within a specific policy problem situation and context.

  2. 2.

    Narrative character: Narratives contain at least one character. Originally, these included victims, villains, and heroes to describe who is likely to be harmed or suffering from a problem, who is to blame and who is able to solve the problem (Shanahan et al., 2018). Recently, further characters have been introduced into NPF research, such as beneficiaries, allies, or opponents (McBeth et al., 2005; Merry, 2016; Weible et al., 2016).

  3. 3.

    Plot: Used to allocate the characters and their interactions in time and space.

  4. 4.

    Moral of the story: Equals the suggested policy solution, thus, what should be done to solve the problem narrated and who should do this.

The NPF additionally defines “narrative content”, namely by specifying that narratives contain expressions of policy beliefs and that narrators may use different narrative strategies to convince audiences of their opinion (Shanahan et al., 2018). Policy beliefs can be operationalized via ideology (McBeth et al., 2022) or cultural theory (McMorris et al., 2018). An important contribution regarding the characterization of policy beliefs within policy process research has been developed within the Advocacy Coalition Framework (ACF). The ACF proposes a three-tiered belief system which shapes policy actors: deep core beliefs, policy core beliefs and secondary beliefs. Deep core beliefs are not policy-specific but related to ideology and culture. One way to conceptualize these in the ACF is also by incorporating insights from cultural theory (Jenkins-Smith et al., 2017). In contrast, policy core beliefs are specific to the policy subsystem (Jenkins-Smith et al., 2017; Sabatier & Weible, 2007). The ACF has also introduced the so-called devil shift in policy process research, which has later been included in the NPF as a narrative strategy. The devil shift borrows a key finding from prospect theory, namely the observation that people remember losses better than gains, which also influences policy actors (Jenkins-Smith et al., 2017).

Frequently employed narrative strategies are the devil/angel shift, causal mechanisms, and scope of conflict (Shanahan et al., 2017). A recent contribution by Kuenzler and Stauffer (2023) adds the component of “policy dimension” to the narrative content. In the policy dimension, narrative elements can be differentiated according to whether they refer to a policy’s substance (i.e., the policy’s implementation model) or to a policy’s process (i.e., the stage a policy is situated in like agenda setting, decision-making etc.; Kuenzler & Stauffer, 2023, pp. 16–19). We return to the policy dimension below.

A further important assumption of the NPF is that policy narratives operate at different levels simultaneously, namely the micro, meso, and macro level. The majority of NPF analyses is situated at the meso level, asking how policy actors develop narratives and form coalitions (Jones & McBeth, 2010). The micro and the macro level, on the contrary, remain less explored. The micro level looks at how individuals inform and are informed by narratives (Flores et al., 2023), and the macro level concerns the embedment of narratives in the cultural and institutional context (Stauffer, 2023).

The past years have seen an extremely dynamic evolution of the NPF, including the theoretical and methodological refinement and the application to different countries and political levels (Schlaufer et al., 2022; Stauffer & Kuenzler, 2021; Tosun & Schaub, 2021). Though NPF analyses have started at the policy subsystem level (Bandelow & Hornung, 2019; Heikkila et al., 2014; Leary et al., 2017), there is an increasing interest of NPF scholars in conducting comparative analyses, also to test larger theoretical propositions (Dunlop et al., 2021; Smith-Walter & Jones, 2020). By comparing policy narratives in France and Germany, we add to this research agenda.

Integrating the distinction between status quo and policy change into the NPF

The distinction between status quo (SQ) and policy change (PC) has received increasing attention in recent NPF research (Chang & Koebele, 2020; Gottlieb et al., 2018; Schlaufer et al., 2023). Scholars have referred to SQ versus PC as an alternative explanation to the winning/losing partition. According to the winning/losing partition, policy actors will use systematically different narrative strategies depending on whether they perceive themselves to be winning or losing in a policy debate (Shanahan et al., 2017, p. 193). For example, some studies found that coalitions perceiving themselves as winning are more prone to use an angel shift, i.e., to emphasize their “ability and/or commitment to solving a problem, while de-emphasizing the villain” (Shanahan et al., 2013, p. 459). In a similar vein, losing coalitions were identified as prone to use a devil shift, thus “exaggerating the power of an opponent while understating the power of the narrating […] coalition” (Shanahan et al., 2013, p. 459). Although in some cases, this distinction between winning and losing was deemed to be a successful explanation for the use of narrative strategies (Schlaufer, 2016), these results could not be replicated across all contexts: In some studies, winning versus losing did not result in a statistically significant difference (Heikkila et al., 2014); in others, coalitions were shown to use strategies contrary to the ones expected (Leong, 2015). This is in line with the theoretical argument that winning and losing may also be evaluated on other criteria than the prevailing of SQ or PC (Marsh & McConnell, 2010). For instance, even if a coalition formally loses in a policy debate, it might still “win” in the sense that thanks to participating in the debate, it gained publicity and was able to attract supporters for future struggles.

Following the mixed findings on the winning/losing partition, Gottlieb et al. (2018) noted that „the side of the issue embraced by the narrator” (Gottlieb et al., 2018, p. 817) might be a better predictor for the use of different narrative strategies. In their analysis of school choice policy in Nevada, Chang and Koebele (2020) built on this proposition by venturing in the discussion of their findings that “[s]tatus-quo and reform-focused policy positions may influence the use of the devil-angel shift strategy” (Chang & Koebele, 2020, p. 640). An examination of narrative strategy usage by government and opposition coalitions in three urban policy debates in Moscow (Schlaufer et al., 2023) found empirical support for this assumption. Governmental actors consistently used the angel shift to promote their proposed reforms, whereas oppositional actors used the devil shift to problematize the same reforms (Schlaufer et al., 2023, p. 80).

While these recent developments have pushed researchers towards including the distinction between status quo and policy change as a relevant explanatory factor for actors’ usage of narratives, we think that an important piece of the puzzle is still missing. In our view, it does not only matter which side of an issue an actor stands on, but what they actually talk about in their narratives. For example, consider an animal welfare group fighting for the introduction of a meat-free veggie day in public facilities. To convince others of their opinion, they might both argue for the policy change they consider necessary, or they might argue against the status quo they deem unacceptable. Both arguments can be delivered via narratives, and these narratives may be equipped with different narrative elements to deliver the animal welfare group’s preferred message. Simultaneously, an opponent to a public veggie day might argue in favor of the status quo or against the policy change to disapprove its introduction. Table 1 provides a fictive illustration.

Table 1 Selection of possible narratives for an animal welfare group fighting for and an opponent group fighting against the introduction of a veggie day (fictive illustration)
Full size table

Thus, by relying on the assumption formulated by recent NPF research, namely that PC proponents will prefer the angel shift, whereas SQ proponents will prefer the devil shift, we only look at one side of the coin. To convince others of the necessity of PC, the animal welfare group can also use other narrative elements than heroes—most notably villains, provided they are depicted as representatives of the SQ. The current assumption regarding SQ versus PC does not account for this possibility. Beyond that, for the animal welfare group it might be even smarter to rely on SQ villains than on PC heroes, as we argue in the following section.

Therefore, we propose to complement the NPF with the distinction of SQ versus PC for narrative elements. As this distinction, similarly to the distinction between substance and process, relates to the question of whether a narrative element supports or opposes a policy, we consider a conceptualization of SQ versus PC as an additional component of the policy dimension adequate, as proposed by Kuenzler and Stauffer (2023). Figure 1 provides an overview of the NPF’s existing conceptualization of narrative content (Jones et al., 2023), together with the newly proposed policy dimension and its two aspects “substance and process” and “status quo and policy change”. In Sect. “Data and Method”) we propose a systematic operationalization of SQ versus PC in narratives that may be applied independently of the specific policy domain investigated.

Fig. 1
figure 1

Notes: Grey background: existing conceptualization of narrative content according to Jones et al. (2023); white background: newly proposed policy dimension

Integration of the distinction between status quo and policy change in the NPF’s narrative content.

Full size image

In the two first conjectures, we abstract from the above example to denote which narrative elements are basically useful for proponents of the SQ or PC, while taking into consideration the new distinction of SQ versus PC narrative elements:

Conjecture 1. Status quo proponents will rely mostly on status quo heroes, beneficiaries, and morals, as well as policy change villains, victims, and problems.

Conjecture 2. Policy change proponents will rely mostly on status quo villains, victims, and problems, as well as policy change heroes, beneficiaries, and morals.

But which narrative strategies are expected to be more useful for SQ and PC proponents to convince others, and why should they be more effective? To answer these questions on the underlying mechanism, we rely on the social psychological prospect theory.

Integrating insights from prospect theory into the NPF

While the integration of the distinction between SQ and PC is relatively new to policy narrative research, the differentiation is already well established in other social science disciplines. Prospect theory, which was established by Kahneman and Tversky (1979), is the most influential behavioral theory to explain decision making under risk and uncertainty. Since the SQ is certain and expectations about the future are uncertain, individuals do not always decide for outcomes that increase their expected utilities. In risk situations, when there is a question whether the SQ is acceptable or whether change is needed, individuals perceive gains and losses according to their reference point (Kahneman & Tversky, 2000). Thus, individuals tend to overvalue deteriorations of the reference point in comparison to analogous improvements of the reference point. As a result, they are more likely to take a risk with losses than a risk with gains. This psychological bias is referred to as endownment effect, which posits that the loss of a good is perceived as more negative than the analogous gain of a good is perceived as positive (Kahneman & Tversky, 2000). Another core cognitive bias theorized by prospect theory is the certainty effect, which means that individuals that are faced with gains overweight small certain gains in contrast to large probable gains. Therefore, they are taking more risks to avoid losses viewed as certain than to secure gains which are probable (Masters, 2004).

To sum up, according to prospect theory, the reference point strongly influences how an uncertain situation is perceived. The reference point is subject to complex dynamics, depending on how it is embedded in a communicative context. For instance, the reference point need not necessarily equal an individual’s own SQ: Say that a UK animal welfare group praises Israel’s veggie day (Naveh, 2014) with the aim to criticize the UK diet in public facilities, the reference point is not the SQ in the UK but the SQ in Israel. Furthermore, a gain can be perceived as a loss if the previously set reference point would have received higher benefits. For example, if a UK policy enables public facilities to offer a veggie day, it can be perceived as loss if the UK policy proposal originally included even a compulsory introduction of a veggie day (reference point = policy change promised in the past). The promotion of certain reference points can therefore influence how policy preferences evolve (Maoz, 1990).

While many studies support the empirical relevance of prospect theory (Camerer, 2011), some studies point out limitations (Kühberger, 1998; O’Keefe & Jensen, 2006), mostly by emphasizing the need for additional moderating or mediating factors (Latimer et al., 2007; Shen & Bigsby, 2013). Levy criticizes that prospect theory “is a reference-dependent theory without a theory of the reference point” (Levy, 1997, p. 100). This is why research has incorporated different factors to determine an actor’s position, among them heuristics and emotions (Mercer, 2005). Heuristics and emotions can influence how individuals perceive potential losses or gains (Welch, 1995). Similar to this, narratives about e.g., heroes or villains trigger certain emotions and serve as heuristic for a losing or gaining reference point.

Another limitation frequently mentioned in the context of prospect theory concerns the fact that it is mainly tested in laboratory experiments. As these settings are artificially generated by researchers, the practical relevance of prospect theory might be limited (Barberis, 2013). Thus, studies such as the one at hand add to knowledge about prospect theory by applying its insights to real-world contexts.

While prospect theory primarily focuses on the effect of differently formulated statements on individuals, its insights bear important implications for actors who intend to maximize the effect of their statements on audiences. Due to humans’ inclination to prefer a certain present over an uncertain future (certainty effect), and due to their inclination to prioritize potential losses over gains (endowment effect), it makes sense for SQ proponents to focus on denigrating a proposed PC, for example by stating that implementing the change will be too expensive, will cause unintended side-effects, or will lead to a disastrous deterioration from the SQ. PC proponents, by contrast, should focus on a negative depiction of the SQ: The SQ is untenable because it makes a lot of people suffer, it leads directly into an abyss, we cannot go on like this etc.

We integrate these prospect theory insights into the NPF by formulating conjectures for the devil/angel shift. We chose the devil/angel shift as a first avenue for this integration because it is frequently used in NPF studies (Pattison et al., 2022; Schaub, 2021) and because its calculation resulting in a numeric valueFootnote 1 allows for a precise empirical evaluation of theoretical conjectures. Due to the combined certainty and endowment effects proposed by prospect theory, we formulate the following conjectures for the devil shift:

Conjecture 3.

Status quo proponents will rely mostly on a policy change-related devil shift.

Conjecture 4.

Policy change proponents will rely mostly on a status-quo related devil shift.

The next section introduces the topic of farm animal welfare.

Farm animal welfare: a contested policy domain

To investigate the way status quo (SQ) and policy change (PC) are embedded in policy narratives, we choose the policy domain of farm animal welfare. In many Western democracies, debates about farm animal welfare feature a strong polarization of actor coalitions due to two diverging conceptions of agricultural policymaking, namely exceptionalism and post-exceptionalism. While exceptionalism generally is considered to dominate agricultural policymaking and thus represents the SQ, post-exceptionalism comprises a host of challenges to the SQ, therefore representing PC. In the following, we introduce these two conceptions and discuss their manifestation in farm animal welfare policies.

Agricultural policy presents a unique policy area that has traditionally received special treatment by policymakers. Hence, public policy scholars have characterized agricultural policy as “exceptional” (Grant, 1995; Skogstad, 1998). Policy exceptionalism is defined as a set of complementary ideas, institutions, interests and policy instruments that have consolidated in the agricultural sector over decades and that have resulted in a high compartmentalization of the sector. Exceptionalism ascribes a unique role to the agricultural sector, which justifies high state intervention and a general aversion against market liberalization to guarantee food security and stable income for farmers (Daugbjerg & Feindt, 2017). Other key characteristics of exceptionalism are well-organized lobbyists that together with complementary institutions and policy instruments have led to the exclusion of actors with diverging interests and to a high degree of compartmentalization of agricultural politics. The EU’s Common Agricultural Policy has for long followed this exceptional path, diverting government funds to support the agricultural sector in the member states and to ensure stable incomes for farmers and high production coupled with low prices for consumers. However, already in the 1980s, this exceptional approach to agricultural policymaking was challenged by rising demands for market deregulation. In the 1990s, the increasing importance of sustainability further added to the contestation of this original set of ideas (Feindt, 2010; Grant, 1995). Both, the demands to open up agricultural policy to the market and to integrate environmental concerns, have contributed to a partial transformation of ideas in the sector and to the rise of so-called post-exceptionalism. Post-exceptionalism describes the intrusion of new ideas and interests in this historically closed agricultural policy area. These new ideas include a less critical stance towards market orientation and internationalization as well as rising importance of sustainability standards and changing consumer concerns. The latter comprise the importance of environmental friendly but also animal welfare friendly production (Daugbjerg & Feindt, 2017).

These ideational transformations also manifest in the specific area of farm animal welfare. Following the ideational framework of exceptionalism, in the past decades, animal production has been trimmed for cost efficiency and the maximization of production output—often at the expense of animal welfare (Grethe, 2017). However, in the past years, sustainability concerns have increasingly been discussed also with regard to animal production, including the environmental and climate impacts of growing and fattening animals for human consumption. With rising consumer concerns about animal welfare, legislation to protect the welfare of farmed animals has been passed at the EU level as well as at the national level in many member states (Horgan & Gavinelli, 2006; Lundmark et al., 2014). Next to state-level regulations, there is also a considerable expansion of private and industry-driven animal welfare labels and initiatives (European Commission, 2009; Parker et al., 2017; Sandøe et al., 2022). The expansion of regulatory as well as private policies in animal welfare both can be accredited to new actors entering the previously closed sector, an indication for a decompartmentalization of the field. Since animal welfare serves sustainability goals, e.g. by increasing stable spaces which in turn reduces pollution of the soils, it is therefore an essential element of the post-exceptional turn in agricultural policymaking. Despite the rising importance of animal welfare, improving welfare standards in animal production remains a costly and therefore contested issue. Previous empirical studies have found a high degree of polarization when it comes to farm animal policies. This polarization is also reflected in party political competition, be it at the EU level in the European Parliament or at the national level (Vogeler, 2019b, 2022).

Next, we provide information about the study’s research design.

Research design

In this section, we introduce the two cases of France and Germany, followed by a presentation of our data and methods approaches.

Farm animal welfare in France and Germany

To conduct a first empirical application of the status quo (SQ) versus policy change (PC) distinction in policy narratives, we focus on the two countries France and Germany. We decided for these two cases because they share cultural, institutional and industrial similarities: First, France and Germany are neighboring countries, which means that they share a long common history with comparable and common values, beliefs and opinions among citizens in both countries. This set of shared values and beliefs might be reflected in similar public discussions on social and environmental issues. In fact, citizens of France and Germany both have a comparably high level of public concern for the welfare of farmed animals, as seen in Eurobarometer results from 2016 (Directorate-General for Communication, 2016).

Second, both countries are members of the European Union (EU) and are therefore subject to the same regulations and policy goals set at the EU level (for more information on EU regulations, see Appendix 1). As European citizens, the French and the Germans are involved in European initiatives such as “End the Cage Age”, a European Citizen Initiative that called for the prohibition of all cage systems in animal husbandry.Footnote 2 This should be mirrored in similar public debates.

Third, France and Germany are among the biggest producers of animal products, and both have a large agricultural sector (European Commission, 2019). Animal production in France consists largely of milk production (13.1% of total agricultural output), cattle (10.0%), and pigs (4.4%; European Commission, 2023a). Germany’s animal production focuses largely on milk (23.3% of total agricultural output) and is equipped with significant shares of pig (11.5%), cattle (6.9%), and poultry production (4.8%; European Commission, 2023b).

Despite their many similarities, France and Germany also exhibit some differences. Most notably for the analysis at hand, the amount and restrictiveness of existing policies that promote farm animal welfare vary considerably. With the term “policy”, we refer to any substantial political decision taken by authorized institutions (Schmidt, 2010, p. 601), thus including, for instance, both laws and ordinances as well as regulative and incentivizing measures. While Germany is among the countries with the highest legal standards worldwide, France is rather in the lower midfield in the EU, with policies rarely surpassing the basic requirements of the EU (Vogeler, 2019b; for more information on animal welfare policies in France and Germany, see Appendix 1).

While we do not expect the amount and restrictiveness of existing policies to influence the narrative usage of SQ and PC in a country, we still want to account for such a possibility in our research design. The analysis of two cases with different policy contexts might provide us information in this regard. To conclude, examining both France and Germany allows for a first robustness test of the conjectures formulated in this article’s theory section.

Data and method

To examine the public debates in France and Germany, we collected newspaper articles from two daily, national newspapers per country for the period of January 2020 until December 2021. This time period is not only chosen because of its recency, but also because animal welfare has received increasing public visibility in the last years due to supranational initiatives (e.g. “End of the Cage Age”) and changing consumption habits towards less meat- and more plant-based diets. We decided for newspaper articles instead of other forms of mass communication due to their high degree of comparability, their ability to represent the broad public and their depth of content, which allows to identify narratives (see also Bandelow & Hornung, 2019). For Germany, we chose the Frankfurter Allgemeine Zeitung (FAZ, rather conservative; Burkhardt, 2012) and the Süddeutsche Zeitung (Süddeutsche, leftist-liberal; Hachmeister, 2012). Both newspapers have a high circulation but differ in their ideological outline, allowing us to identify opposing narratives, e.g. the view of farmers’ unions compared to the one of non-governmental organizations. For France, we also chose two weekly national newspapers with a high circulation and differing political orientation: Le Figaro (conservative and economically liberal; Wagener & Berg, 2012) and Le Monde (leftist-liberal; Wagener & Berg, 2012). We extracted the data via Nexis Uni, Factiva, Europresse, and the FAZ’s newspaper archive. The searches are documented in appendix 2. They included the following combination of keywords, translated into French and German and checked by native speakers: ““Farm animal*” and (*farming or *slaughter* or *transport* or *welfare or *torture* or “dairy cattle” or grazing* or “species appropriate” or *breed* or outdoor* or *meat* or cage* or anaesthesia)”. This broad combination of search terms allows for increasing the probability of retrieving all articles of interest. Besides, we were interested in articles about farm animals, therefore we deleted in a next step all articles about pets and wild animals if they were not used agriculturally (e.g., hunting). In sum, we extracted 29 articles covering the debate in France (8 articles in Le Monde, 21 articles in Le Figaro) and 55 articles in Germany (41 from the FAZ, 14 from Süddeutsche), cf. Appendix 3. Thus, solely based on the distribution of the newspaper articles, we firstly find that there is more public discussion about farm animals in Germany than in France.

In a next step, we checked the single articles and coded narratives qualitatively based on a codebook that was generated beforehand (see Appendix 4). We used the Discourse Network Analyzer (DNA) by Leifeld et al. (2019) for coding the statements.Footnote 3 Thus, we annotated statements of actors—also called concepts—in the selected media documents based on our previously defined and theoretically informed codebook. We used the concepts “hero”, “villain”, “victim”, “beneficiary”, “problem” and “moral” to code the statements of different actors/organizations. The concepts were concretized, meaning that we specified for example who exactly is the villain or what is the problem. The characters based on the NPF can also be animals, e.g. “chickens suffering in the cage” would be coded as “victim: chickens”, or “cows allowed to graze in the pasture” would be coded as “beneficiary: cows”. As Jones et al. (2014) illustrate, characters do not need to be individual humans and can also be non-humans as broad categories or animals. We coded both indirect and direct statements of persons and/or organizations. If no specific actor was named (e.g. “some”/ “many”), we did not code the statement. We differentiated the following organization types: agriculture, animal, business, citizen, court, environment, media, party, science, state and other. Descriptions of these categories can be found in the codebook in Appendix 4. Subsequently, the problem and moral codes were inductively aggregated into broader categories, cf. Appendix 5.

Two additional steps led to the final dataset used for the analysis at hand. On the one hand, we categorized narrators in the dataset according to their belonging to either the coalition of SQ proponents (in the following: “agricultural coalition”) or the coalition of PC proponents (in the following: “animal welfare coalition”).Footnote 4 The categorization was based on fundamental case knowledge, which we derived from document analysis, semi-structured interviews with policy actors and experts, and previously conducted studies. Appendix 6 provides an overview of the two coalitions in both France and Germany.

On the other hand, we categorized all narrative elements according to status quo or policy change. We operationalized this distinction by determining whether a narrative element refers to a past or present state (status quo) or to a future state (policy change). This constitutes a proxy, which in our case attained a high validity, as a qualitative in-depth check of elements categorized with this approach, based on substantive knowledge about the policies at hand, revealed. While such in-depth checks should always be conducted to ensure validity, we propose this approach as a practical solution also for future research endeavours, because it allows for coders to apply the categorization without requiring in-depth case knowledge.

Our results are based on descriptive statistics and on the calculation of the devil-angel-shift. The formula for the devil/angel shift, where the number (N) of each category are related to each other, reads as follows (Stauffer, 2023):

$$\frac{{\left( {N^{Heroes} - N^{Villains} } \right)}}{{\left( {N^{Heroes} + N^{Villains} } \right)}}$$

The formula returns values ranging between −1 and + 1, where negative values indicate a devil shift and positive values an angel shift.

Findings

In the following, we present our empirical findings. Please refer to Appendix 7 for an overview of our dataset, both in terms of organizations participating in the debates in France and Germany between 2020 and 2021, and in terms of narrative elements appearing in the debates.

Table 2 summarizes the number of narrative elements uttered by the agricultural coalition. While as predicted, the agricultural coalitions in France and Germany exhibit a multitude of negative policy change (PC) depictions, with PC related villains, victims, and problems, unexpectedly, they also denigrate the status quo (SQ). Therefore, conjecture 1 is partially corroborated by the public debate on animal welfare in both countries. An in-depth investigation of these SQ depictions reveals that in France, farmers complain about society unfairly accusing them of neglecting their animals and of polluting the environment—a criticism widely known in France as “agribashing”. In Germany, the agricultural coalition mainly draws attention to allegedly high costs and low rentability of animal production, a condition that in their view necessitates PC. This farmer (SQ proponent) depicting the regional government in Bavaria as a villain constitutes an example:

Table 2 Evaluation of conjecture 1, agricultural coalition (SQ proponents)
Full size table

So instead of a second conventional stable, [farmer] Naß built four completely new stables that even meet organic standards. He spent one million euros on this, financed entirely by credit. “But the Bavarian government did not support the conversion to this alternative form of animal farming in any way," says Naß. (FAZ, July 29 2020).


The German agricultural coalition’s general dissatisfaction with the SQ also explains the surprisingly high amount of PC depictions with narrative elements such as heroes and morals: Their proponents suggest PC measures such as grandfathering for farmers who invest in stable upgrading to improve farmers’ wellbeing.

As displayed in Table 3, the data of both France and Germany corroborate conjecture 2 (“PC proponents will rely mostly on SQ villains, victims, and problems, as well as PC heroes, beneficiaries, and morals”). The two investigated animal welfare coalitions focus on negative depictions of the SQ and positive depictions of policy change, thus promoting post-exceptionalism vis-à-vis exceptionalism. The following quotes serve as illustration, with the first quote constituting a SQ-related victim depiction from the French coalition and the second quote a PC-related moral from the German coalition:

Table 3 Evaluation of conjecture 2, animal welfare coalition (PC proponents)
Full size table

The life of a chicken is a life of misery in large sheds, without ever seeing the light of day, herded together with tens of thousands of animals. (Animal welfare NGO in Le Monde, September 10 2020, own translation).


The so-called competence network for livestock farming, led by Borchert, had already in February argued for an elevation of prices for meat, sausages, milk, eggs, and other animal products, in the form of an animal welfare tax. (Public commission for the promotion of animal welfare in FAZ, August 28 2020, own translation).


The French animal-welfare coalition also features a small amount of positive SQ depictions. These depictions are praises of advances for animal welfare that have already been achieved. Similarly, the German animal-welfare coalition exhibits a minority of 5 negative PC depictions (opposed to 26 positive policy change depictions), denouncing policy measures that in the narrators’ view are not sufficient for animal welfare. Such a negative PC depiction by the German animal welfare coalition is displayed in the following quote:

The non-governmental organization Foodwatch […] criticized that [a commitment by the food discounter Aldi to stop selling meat with the lowest category of an animal welfare label by 2025] would fool customers into believing that this would change something about the situation of animals. No animal welfare labelling can help against the diseases and misery of millions of animals," said Foodwatch. (Animal welfare NGO in FAZ, June 26 2021, own translation).


We now turn to the devil/angel-shift calculations. As may be concluded from Table 4, conjecture 3 (“SQ proponents will rely mostly on a PC-related devil-shift”) is corroborated for France and partially corroborated for Germany. The agricultural coalition in France features an unequivocal devil shift in their PC related communication (− 1.00, see column 3). The German agricultural coalition too exhibits a devil shift in their PC related communication, although it is less clear-cut (− 0.33, see column 6). Again, this finding may be ascribed to the fact that the German agricultural coalition also proposes their own PC measures (cf. above, conjecture 1).

Table 4 Evaluation of conjectures 3 and 4, devil/angel shifts for both coalitions
Full size table

Conjecture 4 (“PC proponents will rely mostly on a SQ related devil-shift.”) is corroborated by the data from Germany and rejected for France. Whereas the German animal welfare coalition exhibits a strong devil shift (− 1.00, see column 6) in their SQ related communication, the French animal welfare coalition’s SQ communication is characterized by a moderate angel shift (0.33, see column 3). The following quote of a PC proponent serves as an example for France:

It's thanks to [consumer information on egg packaging concerning farming standards] that we've reduced the suffering of laying hens, who used to have only an A4 sheet of paper to move on," recalls Agathe Gignoux, Head of Public Affairs at Compassion in World Farming France. (Animal welfare NGO in Figaro, January 28 2020, own translation).


Finally, we take a moment to compare the devil-angel shift values calculated both with the conventional approach and with the distinction proposed in this article between status quo and policy change (see column 2 and 5 in Table 4). As Table 4 shows, the conventional devil/angel shift values are always situated somewhere in between the values calculated for SQ and PC related narrative elements. This is logic, given the fact that the latter are calculated on data subsets of the former. Thus, the conventional calculation provides us with a more moderate picture than what we actually see in the data if we distinguish between narrative elements depicting the status quo or policy change. For three of four devil/angel shift calculations for coalitions in Germany and France (cf. Table 4), the substantial message remains the same, that is, if the conventional approach leads to the conclusion that a coalition uses a devil shift, then so do the two calculations for SQ and PC, respectively. However, the case of the German animal welfare coalition proves that the conventional approach may also lead to false and oversimplified conclusions: While the conventional calculation signals a moderate devil shift (− 0.17), the distinction between status quo and policy change shows that actually, the coalition uses a differentiated and polarized communication strategy with simultaneously a strong denunciation of the status quo (devil shift of − 1.00) and a moderate appraisal of possible policy change (angel shift of 0.43). Thus, we caution researchers to be aware of the potential pitfalls the conventional devil/angel shift calculation entails.

Discussion

Table 5 provides a summary of the conjectures’ empirical evaluation in the French and the German cases. All four conjectures find partial or full corroboration in at least one of the two investigated cases. Finding empirical evidence for the theoretical differentiation between status quo (SQ) and policy change (PC) highlights the benefit of integrating this new theoretical aspect into the NPF: It allows for a more precise analysis of strategic arguments by different coalitions.

Table 5 Summary of conjectures’ empirical evaluation
Full size table

While recent NPF studies (Chang & Koebele, 2020; Gottlieb et al., 2018; Schlaufer et al., 2023) had already put forth the relevance of SQ and PC as an explanatory factor for narratives, they applied the distinction exclusively to narrators, hypothesizing that PC proponents will use the angel shift, and SQ proponents will use the devil shift. Thereby, they only captured part of coalitional communication that makes sense from a strategic point of view.

Departing from the same premise as Chang and Koebele (2020), Gottlieb et al. (2018), and Schlaufer et al. (2023), we demonstrated that for a comprehensive understanding, it is necessary to investigate what narrators actually refer to in their communication: SQ or PC. Both may be strategically interesting for them, and as our investigation shows, both are actually referred to and depicted with differentiated narrative strategies. Furthermore, we extend the argument by not only including heroes and villains as done previously with the devil/angel shift hypotheses, but also victims, beneficiaries, problems and morals.

Additionally, our analysis highlights pitfalls that may occur with the conventional calculation of the devil/angel shift, an often-investigated narrative strategy in NPF research (Pattison et al., 2022; Schaub, 2021). Whereas in most instances, the conventional calculation led to more moderate values than what we actually perceived empirically, in one case the conventional calculation even provided a distorted impression.

Future research building on the distinction between SQ and PC in the NPF might extend the analysis towards other narrative strategies such as scope of conflict or causal mechanisms. As first findings by Schlaufer et al. (2023) indicate, differentiating between SQ and PC might also matter for these additional narrative strategies.

For researchers working with prospect theory, this study adds to the understanding of the reference point by introducing narratives, more precisely narrative strategies, as an important heuristic used by policy actors to elaborate on their reference point. For instance, by portraying PC proponents as villains that cause harm, SQ proponents can evoke the endowment effect for their audience. Simultaneously, PC proponents may portray SQ actors as villains to emphasize the potential losses that would result from maintaining the SQ. Thus, the study demonstrates that the reference point is also influenced by narrative elements. In addition, this study adds methodically to prospect theory. While previous prospect theory research is mainly based on laboratory experiments, looking at public debates and analyzing them quantitatively and qualitatively adds to the practical relevance of prospect theory, which is often criticized for its artificial examination setting (Barberis, 2013).

In a last step, we leverage three empirical instances in which the distinction between SQ and PC did not perform as expected to reflect on possible causes and propose amendments for future research.

An unexpected behavior of the investigated coalitions that occurred repeatedly concerns the policies they referred to in their narratives. For example, the agricultural coalition in Germany did not only denounce PC measures aimed at promoting animal welfare, but also propose their own measures to promote farmers’ wellbeing. Similarly, the French animal welfare coalition not only denigrated the SQ for a lack of measures promoting animal welfare, but also highlighted already achieved advancements they consider important. We think that these unexpected findings are at least in part a consequence of our research design. Our selection of newspaper articles held the issue area “farm animals” constant, but not the policy itself. So, while all narratives centered around farm animals, narrators referred to a multitude of different possibilities of PC versus SQ in this area, with their pro/contra positions shifting depending on the specific topic at hand. We argue that focusing on a single policy decision might lead to a stronger corroboration of the formulated conjectures, as this would keep the topic context more constant. This is a first proposition we would like to put forward for future research. For example, a debate preceding a popular or parliamentary vote might be an interesting avenue to put this argument to the test.

Second, the French and German agricultural coalitions’ communication warrants attention. Both depict the SQ mainly with villains, victims, and problems, a finding running contrarily to our first conjecture. This criticism of the SQ even by those actors who supposedly benefit from it may be interpreted as a sign of general discontent with the current agricultural system in the European Union. Recent research supports this assumption for German farmers (Bethge, 2023). This also points to the fact that in some policy debates, the central question is not whether to have PC or remain with the SQ, but which PC is needed and how far it should go. Such debates are often prevalent in climate policy (Gjerstad, 2017; Malmborg, 2023a, 2023b, 2023c), and our study indicates that farm animal welfare policy might be at a similar point. Future research could conduct in-depth examinations of debates where the status quo is deemed unacceptable by all participants. It might be worthwhile to investigate how this impacts the narrative strategies employed.

A third interesting finding concerns differences between the French and the German case. While, as expected, the distinction between SQ and PC performs mostly similar in the two countries, especially conjecture 4 shows that only in Germany, the animal welfare coalition used a devil shift to depict the SQ. With the French animal welfare coalition exhibiting an angel shift in the same area, the question arises what led to this difference. As described in Sect. "Farm animal welfare in France and Germany" , France is lagging behind with animal welfare compared to other European countries. Thus, the amount and restrictiveness of existing policies might play a certain role after all in that the French animal welfare coalition could feel the need to signal willingness to compromise. By praising the few achievements that have been made in this area, they might show that PC is actually possible and that there is a noticeable effect for animals. Future research investigating cases with PC coalitions suggesting especially pronounced deviations from the SQ could provide further insights here.

Harking back to exceptionalism and post-exceptionalism introduced in Sect. "Farm animal welfare: a contested policy domain" of this article, our analysis allows for an assessment regarding the conflict between these two diverging conceptions of agricultural policy in the area of farm animal welfare. The fact that both agricultural coalitions in France and Germany denigrate the status quo seems to indicate that exceptionalism, with its one-sided focus on maximizing production and lowering costs, is on the downgrade in Western Europe. However, this does not mean that the conception of post-exceptionalism is fully embraced by all actors: While in France, agricultural actors mostly refrain from proposing any solutions, German agricultural actors state that they depend on financial support to improve farm animal welfare. This difference between the two countries again might be rooted in the varying states of legislation. In Germany, several animal welfare policies have already been implemented, so farmers have the opportunity to evaluate both costs and benefits of such policies first-hand. France, by contrast, is lagging behind with animal welfare legislation (see also Sect. "Farm animal welfare in France and Germany" and Appendix 1). Hence, improvements towards more farm animal welfare might appear costlier or riskier for French farmers, thus sparking more resistance.

Conclusion

This article aimed to answer the following research question: How are status quo (SQ) and policy change (PC) recounted in public debates? By integrating insights from prospect theory into the Narrative Policy Framework (NPF), we investigated narratives about SQ and PC in newspaper articles on the topic of farm animals. Comparing public debates in France and Germany between 2020 and 2021, we found that the way SQ and PC are recounted in narratives depends on narrators’ preference for either the SQ or PC. For example, the animal welfare coalition in Germany generally depicts PC as a means to improve the life of farm animals, thus transforming them from victims to beneficiaries of the agricultural system. By contrast, the agricultural coalition in France denigrates visions of policy change by stating that it will generate many costs for farmers who are already suffering from an economically harsh environment. The distinction between SQ and PC also matters regarding the use of the “devil/angel shift” narrative strategy: Our analysis shows that the conventional approach often leads to a more moderate and sometimes even distorted picture compared to what we see when applying the distinction between SQ and PC. While not all conjectures were fully corroborated by the analysis, we are confident that the distinction between SQ and PC constitutes a useful tool for future NPF studies.

The study at hand is also beneficial for researchers working with prospect theory. On the one hand, narratives are introduced as a possibility to depict an actor’s reference point. On the other hand, prospect theory insights were applied to a real-world setting, thus contributing to knowledge about the practical relevance of the approach.

The analysis is subject to some limitations. First, our period of investigation from 2020 to 2021 largely overlaps with an extraordinary time both socially and economically, and not only in France and Germany, but worldwide: The COVID-19 pandemic. We chose this period with the intention to conduct a timely analysis, and to examine a period with increased public visibility for animal welfare and with changing consumption habits towards more plant-based diets. However, it is possible that the health crisis might have impacted our research object—after all, animal production hit the spotlight several times during the pandemic, as evidenced for example by the scandal of massive mink culling in Denmark to prevent virus mutations. Although we cannot fully account for the possibility of such an effect, a comparison of newspaper article frequencies both before and during the years of the COVID-19 pandemic shows that at least in quantitative terms, no crowding-out effect of farm animal topics occurred (cf. appendix 8). Furthermore, the qualitative investigation of the analyzed articles did not reveal any anomalies when compared to previous years.

Second, the analysis of newspaper articles bears the limitation that we did not investigate narrators’ direct communication, but narratives mediated via journalists and editorial teams (Shanahan et al., 2008). Hence, we cannot guarantee that narrators’ communication was related appropriately, and the balance of actors found in the newspaper articles might not equal the balance in the general public debates. We tried to amend for the first hazard by selecting multiple newspapers renowned as high-quality outlets. Additionally, the narratives identified in these newspapers correspond well with what we know about the narrators and the policy field more generally from existing research (Hårstad, 2023; Vogeler, 2019b). Concerning the balance, the selection of multiple newspaper sources also served as a remedy. Furthermore, as our research interest is not concerned with an accurately balanced representation of actors in the investigated debates but focuses on narrative content, we refrained from corrective measures such as weighing actors’ statements.

Third, the amount of data used for our analysis constitutes a limitation. Farm animal welfare is a topic of low to medium salience in public debates, as is evidenced by the fact that in Germany, 55 articles appeared in two distinguished newspapers within two years, and in France only 29. Thus, the evaluation of our conjectures should be considered to be preliminary, and we encourage fellow researchers to extend the analysis to other debates with higher salience.

Our study shows that the NPF’s analytical strength is enhanced by integrating the distinction between SQ and PC in narrative elements. This distinction enables further empirical nuancing of actors’ narrative communication, and it allows in combination with insights from prospect theory for new conjectures about actors’ use of narrative strategies such as the devil/angel shift. Simultaneously, the study demonstrates that narratives are important heuristics to define the reference point. Thus, prospect theory research might benefit from looking at narrative characters as heroes or villains to better understand strategic argumentations based on uncertainty and risk.

This article also sheds light on debates surrounding farm animal welfare in Western Europe. Our analysis shows that not only the coalitions promoting animal welfare are unsatisfied with the SQ, but also the agricultural coalitions. The latter complain about costs and lacking support from politics and society. Whereas the German animal welfare coalition also denounces the current system, the French animal welfare coalition seeks compromises by praising what has already been achieved for farm animals. Overall, a multitude of policy proposals are prevalent in both the French and the German debate. Thus, the findings generated in this study may also serve as an optimistic sign for future improvements of animal welfare.