There’s a place for us? The Digital Agenda Committee and internet policy in the German Bundestag

1. Introduction

Almost a decade ago, internet policy entrepreneurs declared the “birth of internet policy” (“big bang der Netzpolitik”) in Germany (Spielkamp and Wragge, 2012). They considered several events – for example, the electoral success of the Pirate Party at the state level in Berlin and the massive mobilisation against the Anti-Counterfeiting Trade Agreement (ACTA) – as indications that the transformative power of the internet on society had gained an important and permanent spot on the political agenda. Therefore, there was a push from both outside and inside of the political system to create dedicated institutions for internet policy to secure the freedom of the internet and civil rights. In this way, internet policy would receive permanent attention and necessary expertise could be brought into the political system.

The Bundestag (German federal parliament) established a commission of inquiry (a so-called Enquete Commission) on Internet and Digital Society in 2010. This Enquete Commission discussed internet policy via a collection of issues, such as copyright, data protection, open access, freedom of the internet, and net neutrality. The commission recommended the creation of a permanent committee (see Bundestag, 2013, pp. 25, 41). After the following general election, the Bundestag indeed established a new permanent committee, the Digital Agenda Committee (DAC, Ausschuss Digitale Agenda), in February 2014. In hindsight, it is clear that public attention around internet policy in Germany peaked in 2011 (Hösl, 2019, pp. 285f.), and steadily declined after this point, probably because it shifted to migration policy and the rise of populism from 2015 onwards. Nevertheless, the DAC was established as a permanent committee, which is still present in the running legislative session. While further demands to create a ministry for internet policy were not fulfilled, the role of Minister of State for Digital Affairs, located in the Chancellor’s office, was created in 2017.

In this article, we analyse the DAC in its first legislative session (2013-2017) and examine the case of data retention legislation to determine if the DAC had the effect desired by internet policy entrepreneurs of the time. The focus of this research is relevant in three primary regards. First, the DAC not only marked a remarkable institutional change in Germany, but also represented the first body of its kind among all EU member states. The Tweede Kamer of the Netherlands is currently discussing the introduction of a digital committee in the Dutch parliament (see de Jong et al., 2020). Secondly, this work contributes to understanding how internet policy issues became institutionalised in Germany. Hösl and Kniep (2019) recently analysed institutional changes regarding internet policy in the Ministries of Internal Affairs and Economic Affairs. This research complements their work on the executive branch with an examination of the Bundestag. Thirdly, this research helps readers to understand the theoretical question of how new policies emerge.

The remainder of this paper is structured as follows: section 2 discusses the existing literature on the emergence of policy fields (policy subsystems) and on internet politics, and develops the theoretical foundation for this research. section 3 then describes the methods and material utilised for the empirical analysis that is detailed in 4 and 5. These sections analyse the DAC (4) and examine data retention as a case of internet-related law-making (5). Finally, section 6 discusses the results of the analysis and relates them to the broader literature (6).

2. Theoretical background

In recent years, there has been rising interest in the way new policies and policy subsystems emerge. A small number of previous studies have empirically focussed on different policies, including healthcare (Harrison et al., 2002), homeland security (LaPira, 2014), European science (Edler and James, 2015), climate adaption (Massey and Huitema, 2016), and fracking (Ingold et al., 2017). All such studies refer to theories of the policy process (see Weible and Sabatier, 2018) to conceptualise how new policies emerge.

On the specific topic of internet policy in Germany, the literature is divided into studies which apply theories of the policy process and those that adapt a field-sociological perspective. In the first group, the majority of authors have adapted the advocacy coalition framework (ACF) (see Fritz, 2013; Scheffel, 2016; and implicitly Greef, 2017). In the second group, authors have proposed an original theoretical perspective (see Hösl and Reiberg, 2016; Pohle et al., 2016; Reiberg, 2017; Hösl and Kniep, 2019). They introduced the sociological theory of strategic action fields (Fligstein and McAdam, 2011) to policy process research and combined it with discursive institutionalism (Schmidt, 2010). These authors have argued that internet policy is foremost “a field of struggle”, as prominently stated in the title of the work by Pohle et al. (2016). Although they do not refer to field theory, Haunss and Hofmann (2015) also emphasised power struggles when they applied the concepts of (de-)politicisation and subsumption. Therefore, they can be included in the second group in the literature.

A closer reading of Pohle et al. (2016) suggests that there may be a greater difference in wording than in substance between the two approaches in the literature, despite their very different starting points. Nevertheless, the field-theoretical perspective remains problematic in its theoretical micro-foundation (see Coleman, 1986). In contrast, well-established theories of the policy process clearly define bounded rationality as their micro-foundation. Thus, it is this notion that establishes the theoretical background for the analysis in this research (see Jakobi, 2019 for a longer discussion of theoretical frameworks for the analysis of internet policy in Germany).

In general, the emergence of an issue or a dedicated policy subsystem is a matter of policy change. Electoral politics and party competition can act as drivers of such change. However, over which issues do parties actually compete, and what explains policy change over long time periods? A multitude of different issues and social conditions constantly demand the attention of political actors, and not all issues and conditions are addressed. Indeed, some issues are ignored for a long time. On the one hand, public policy is problem solving, although it is structured by power politics. On the other hand, the rationality of political actors and organisations and their interactions are not comprehensive. According to the model of bounded rationality (Simon, 1985) actors have preferences and act intentionally with respect to them. However, this rationality is limited; in other words, it is bounded by many factors, such as cognitive abilities and biases, attention, emotions, and values. Different theories of the policy process that build upon this micro-foundation stress different limitations and different aspects of the policy process. This research – like the majority of the above-referenced policy process studies – refers to punctuated equilibrium theory (PET) (for a general overview, see True et al., 2018) to conceptualise how internet policy may have emerged in parliament and what to expect from the DAC in particular. In short, PET proposes that long periods of incremental change or near-stability alternate with short episodes of fundamental change in public policy. Because it accounts for both stability and change, this research applies PET instead of the ACF, which tends to emphasise stability.

Punctuated equilibrium theory focusses on limited attention as an important bound of rationality. According to the theory, neither individuals nor organisations can attend to everything at the same time; instead, they must prioritise issues. This limitation is a significant problem for decision-making, because decisions must be made one after another in serial mode. It is less problematic for discussing problems and proposing solutions. Individuals can become experts on a limited number of issues, and organisations can be established for specific issues. The permanent interactions between such experts and organisations create a specific policy subsystem. A number of policy subsystems can process multiple issues, problems, and proposals for solutions in parallel (see Baumgartner and Jones, 2009, p. 19 f.). However, while this parallel process benefits those issues already recognised by political actors, it makes it more difficult for new issues to attract attention. Actors with vested interest in such issues have to overcome not only potential political opposition, but also cognitive/organisational and institutional friction.

An important cause of cognitive and organisational friction is the fact that social conditions do not automatically translate into public policy problems. Only if a social condition is perceived as something that a government can and should change does it constitute a public policy problem and attract attention. Empirical information – what kind of information is (or is not) relevant, how that information is interpreted, what policy instruments are appropriate, what policy outcomes are to be expected – and evaluative information – what is (or is not) acceptable in regard to what values – together constitute a policy image or frame. Baumgartner and Jones (2009) do not discuss frames as concepts, but policy images and frames are defined in the same way. Daviter (2007, p. 655 f.), for example, uses the concept of frames and refers to almost the exact same literature as Baumgartner and Jones (2009). 1 Policy images not only guide the attention of individuals, but can also guide the attention of organisations. Although experts in a specialised organisation may disagree about policy solutions, they tend to share a policy image after interacting for a significant time. Political actors that are willing to invest their time and effort in pushing new issues, their favourite proposals or frames, are called policy entrepreneurs. When they want to promote new issues, they often try to change an existing policy image. Mobilisation is one strategy to attract attention around a new policy image. Different policy images often compete with each other, and political actors constantly try to manipulate policy images in favour of their interests. Environmentalists, for example, have mobilised against the policy image of nuclear energy as a modern, cheap, and safe energy source and promoted an alternative policy image of a dangerous and expensive technology with high environmental costs (Baumgartner and Jones, 2009, pp. 256-264).

Policy images first interact with organisations in a given policy subsystem. Organisations may be more receptive to some policy images than to others. For example, an agricultural committee will be more receptive to the policy image of pesticides as an efficient method to improve farmers’ profits than a public health or environmental committee, which will be more receptive to the image of pesticides as dangers to health or the environment. Secondly, policy images interact with venues, which are the institutional locations where authorities decide certain issues (Baumgartner and Jones, 2009, p. 32). Therefore, venue change is a second strategy to promote a given policy image. Policy entrepreneurs, for example, bring their issues to courts if they do not succeed in parliament or to the local level if they do not succeed at the federal level (Baumgartner and Jones, 2009, pp. 34 f.). Thirdly, policy entrepreneurs mobilise for their issues and try to change venues, but their success is also critically dependent on media attention (Baumgartner and Jones, 2009, pp. 103ff.). All of these interactions result in either positive (stabilising) or negative (destabilising) feedback loops.

When we apply this general perspective to parliaments, a necessary, but only first, condition to establishing a new policy issue is that either experts on a new issue are elected as MPs or already-elected MPs adopt the issue in parliament. An important second condition to permanently establish a new issue is that there is an organisational location that attracts attention to the issue, convenes experts on it, and helps to create and stabilise a shared policy image. Dedicated subcommittees, committees, or even committees with jurisdiction over a given issue reflect grades of increasing attention and serve to stabilise the respective policy. However, an existing committee system with accepted policy images creates friction that must be overcome by mobilisation. This research therefore expects to find evidence of competition between established and new policy images in the Bundestag, for example, in plenary debates. The stability of policy images and the successful transformation of policy images both critically depend on media attention. Moreover, there was a change in how the new issue of internet policy was called in the period of study in Germany. While “Netzpolitik” (internet policy) dominated the media discourse between 2006 and 2013, another term “digital policy” became prominent after 2013 (Hösl and Kniep, 2019, p. 16). Whereas internet policy was strongly associated with freedom, civil rights, and anti-surveillance (Hösl, 2019, pp. 285f.), digital policy rather referred to economic dimensions like the roll-out of ultrafast broadband to create economic opportunities.

The concepts of PET were developed in regard to the political system of the United States. In order to apply these concepts to the German context, it is necessary to account for the special structure of the Bundestag. The Bundestag is a collective body (parliament of factions; Schüttemeyer, 1992), in which the role of political parties is emphasised and factional discipline prevails. As a mixed-centre-type parliament, the Bundestag is located between the two prototypes of transformative parliaments (e.g., the US Congress) and arena parliaments (e.g., the UK House of Commons; see Polsby, 1975). Committee work (transformation) in the German parliament is at least as necessary as public debates (arena). In effect, decisions by the Bundestag are usually prepared and predetermined in standing committees before they are the subject of plenary debates. Although the latter is also true for the US Congress, the German committee business is distinct (see Siefken, 2018, p. 781; Martin, 2014). Committee jurisdiction in Germany is less exclusive and less public than in the US Congress. In addition to one leading committee, a case-dependent number of other committees are also involved in the Bundestag in the preparation of the final vote in plenary. They deal intensively with the law and support the leading committee by forwarding their voting results on the bill. The leading committee takes note of the information. After the vote of the leading committee, the decisive recommendation is given back to the plenary for the final vote on the bill. Therefore, not only the leading committee, but all involved committees, act as policy venues.

3. Material and methods

To answer the questions posed in the introduction, this research proceeds with the analysis in two steps. Recognising that policy images interact with venues, we first analyse the DAC as a potential venue (see section 4). We examine its origins, competencies, composition, and limits and the policy image(s) of its members in the period of study. In addition, we note how journalists and internet policy entrepreneurs outside of parliament perceived the DAC. As there is still no academic literature on the DAC, we include the official websites of the Bundestag and media reports in this analysis. Moreover, we conducted in-depth interviews with committee members in 2016, 2 which are evaluated qualitatively. 3

Secondly, the analysis examines whether there has been a change in policy images regarding internet issues (see section 5). To do so, we selected an internet-related issue that was discussed in the Bundestag before and after the creation of the DAC. Given this condition, a number of internet-related issues that were discussed in Germany could not be selected. Legislation on access blocking (Netzsperren) was a significant issue in the 17^th legislative session (2009–2013), but not in the 18^th (2013–2017). The Network Enforcement Act (“Netzwerkdurchsetzungsgesetz”) was a prominent issue in the 18^th legislative session, but not before. Moreover, net neutrality (“Netzneutralität”) is an international issue that will be largely decided at the level of the European Union. In contrast, the selected issue of data retention (“Vorratsdatenspeicherung”) was an important issue inside parliament and also for internet policy entrepreneurs outside of parliament (Ganz, 2015, p. 36; Wendelin and Löblich, 2012, p. 903) during the period of study. Indeed, internet and data protection were especially salient issues in public debate around 2015 (Hösl, 2019, p. 285). Laws regarding data retention were passed in 2007 (before the introduction of the DAC) and 2015 (when the DAC had been established).

Because legislation on data retention is a complicated matter, the related analysis proceeds in three steps (5.1-5.3). First, the origins of the laws are explained in a brief historical outline (5.1). A discursive network analysis is subsequently performed ( dna, 5.2) to determine whether and how the plenary debate has changed. This research uses dna because it is not only interested in potential changes in policy images, but also in who attempts to produce such changes. The data set for the discourse network analysis consists of the minutes of the plenary sessions for all three readings of the data retention bills of 2007 and 2015. Finally, the analysis examines if and how members of the DAC attempted to mobilise for policy image change and what other strategies they employed (5.3). The material utilised in this analysis consists of parliamentary documentation (5.1), plenary documents (5.2), and transcripts from in-depth interviews with committee members (5.3). 4

4. The Digital Agenda Committee: a venue for internet policy?

The DAC established in 2014 represents a remarkable institutional change and is the first committee of its kind in the European Union. The DAC was created upon the recommendation of the Enquete Commission on Internet and Digital Society (orig. Internet und digitale Gesellschaft). The internet and related issues had been previously discussed in the Bundestag. In the 14th legislative session (1998-2002), internet-related issues were granted limited attention in a subcommittee (New Media, “Neue Medien”) under the responsibility of a standing committee (Culture and Media, “Kultur und Medien”; see Schwanholz, 2019). Thus, the issues were discussed in the context of a media policy image. In contrast, the Enquete Commission, in its 2013 final report, recommended the creation of a new standing committee for internet policy and digitalisation in the Bundestag (see Bundestag, 2013, p. 41). Parties in the Bundestag followed up on this recommendation with a joint proposal and voted for the creation of the DAC in 2014. Following the recommendation of the Enquete Commission, the primary task of the committee was to address the growing importance of internet policy issues. In addition, the DAC was assigned to monitor the Digital Agenda of the federal government (Bundestag 2014). The Digital Agenda was a package of measures to realise digitalisation and identified seven priority areas for action (see Digitale Agenda, 2014-2017). Overall, the Digital Agenda had a strong economic focus, which this research refers to as the digital policy image, in contrast to the internet policy image described above.

In the period of study, the website of the Bundestag made no explicit reference to either the internet policy image or the Digital Agenda. In fact, for the committee chairman Jens Koeppen, digitalisation was broadly “a breathtaking transformation for society as a whole, a profound technical, social and cultural process of change that has affected all areas of our lives” (https://www.bundestag.de/webarchiv/Ausschuesse/ausschuesse18/a23). In the committee, the various aspects of internet policy, digitalisation, and network technologies were to be discussed in an interdisciplinary manner. Members of the committee did not see internet policy as a “niche topic”, and the committee described itself as an important source of inspiration for parliamentary work.

During this period of study, members of the DAC could raise issues on their own authority, invite ministers for oral governmental reports or expert testimonies on all issues of the Digital Agenda, and build networks with interest group actors outside of parliament or even parliamentarians of other nations (see Bundestag, 2016d). Nevertheless, the DAC was established only as an advisory committee and could not lead law-making (see Bundestag, 2016a). This limitation is explained by the fact that the DAC did not mirror any single governmental department (in contrast to all other committees), as there was and still is no ministry for internet policy or digitalisation in Germany.

The DAC consisted of 16 ordinary and 16 deputy members. All became members at their own request and many were former members of the Enquete Commission (see Bundesregierung, 2014). Individual motives for becoming a member of the DAC ranged from general interest in digital topics to professional backgrounds in the members’ former working lives. In this research, not all interviewees were aware of the actual objective of the DAC. For instance, one member interviewee planned to focus on the roll-out of ultrafast broadband (see Interview 1), which in fact falls under the responsibility of the Committee on Transport and Digital Infrastructure. As cross-sectional topics, internet issues and digitalisation extend over a number of governmental domains. For example, security issues pertain to the Ministry of Interior, whereas digital economy issues pertain to the Ministry of Economic Affairs, and broadband expansion pertains to the Ministry of Transport.

In the Bundestag, DAC members did not share a common policy image, but instead mirrored the ambiguous goals of the DAC. The MPs themselves reflected two distinct groups. On the one hand, several MPs highlighted individual (privacy) rights to be protected against the government, promoting an internet policy image. On the other hand, several MPs wanted to expand universal access to broadband internet, promote new digital business models, and emphasise opportunities over risks, reflecting a digital policy image (see Interviews 2 and 3).

Although the members of the DAC had insider knowledge and held specific functions (e.g., speaker, chairperson, responsible representative), they did not advertise their status as DAC members in public. They engaged with a number of interest groups and other relevant stakeholders outside of parliament, such as think tanks, but were not generally recognised as experts. Instead, they were either excluded from important topical events, such as panel discussions and conferences, or did not take part of their own volition. In short, DAC members competed with policy experts from other committees, who were recognised to have more in-depth knowledge (see Interviews 1 and 3). Social networks could have provided an alternative way for DAC members to become more visible. However, with respect to social media use, Schwanholz and colleagues concluded that members of the DAC made “much ado about nothing” (Schwanholz et al., 2018). The DAC had no institutional social media account, and committee members did not use their individual accounts to communicate their committee work. The analysis of the Twitter accounts of all DAC members in 2015 revealed that they used Twitter for self-promotion and comments on daily news, but not as a channel to discuss internet policy. Although this pattern is similar to other Bundestag committees, one might expect further discussion from members of the DAC, who were particularly interested in technology and digital innovations.

Moreover, early evaluations of the committee’s work rated it as poor: media reports and internet blogs lambasted the DAC for the wide gulf between its ambition and actual activity (see Schnoor, 2015; Voß, 2015). Although the committee tried innovative methods to better connect with citizens outside of parliament, committee members seemingly preferred to sit behind closed doors. The number of public hearings remained at a relatively low level disproportionate to non-public hearings, with only 22 public hearings and 70 non-public meetings (see Bundestag, 2017). In addition, the introduction of an online participation tool failed due to a lack of public interest.

Overall, members of the DAC did not become the most important contacts for internet policy issues in the Bundestag. The committee was in a phase of self-discovery and had to cope with a competing, imposed digital policy image by the federal government. This influence amplified ambiguity regarding the objectives sought by the committee. As such, expectations for the DAC from outside internet policy entrepreneurs and journalists may simply have been too high.

With its limited competencies and competing policy images, it is an open question if the DAC made a difference at all. To tackle this question, this research examined the case of data retention.

5. Data retention law-making in the Bundestag in 2007 and 2015

5.1 Short history of the data retention conflict

Since the 1990s, data retention has repeatedly entered the German political agenda. Data retention is defined as the obligation by service providers to store personal and traffic data arising from telecommunication and internet use and to make these data available for security services when necessary. When telecommunication was liberalised in 1996, data retention was discussed for the first time in competing policy images of data protection and security. In the late 1990s and early 2000s, the opposition Christian Democratic Party (CDU) proposed several laws that were rejected by the government coalition of the Social Democratic Party (SPD) and Green Party (die Grünen; see Bug, 2016, p. 677; Gausling, 2010, p. 47). The terrorist bombings in New York, Madrid, and London (2001/2004 /2005) raised worldwide awareness of new threats to national and internal security and increased the number of new national standards to fight terrorism. To ensure that all member states imposed data retention in the European Union, an EU data retention directive (No. 2006/24/EG) was adopted in 2006. The directive described categories and types of data and listed requirements for data access and duration of storage. The political process of the EU data retention directive was accompanied by large-scale protests. Not only citizens, but also judicial scholars doubted the directive’s effectiveness for preventing serious crime and its compatibility with the right to privacy (see Feiler, 2010, p. 19).

Germany passed its first data retention act in 2007 (see Szuba, 2011). The act encompassed a number of changes and reforms to German telecommunication law, namely the introduction of preventive data retention over a period of six months. In 2010, the German Federal Constitutional Court declared the data retention act to be incompatible with German constitutional law, and the act was repealed. Moreover, in 2014, the European Court of Justice repealed the EU directive. The court ruled that the directive was not compliant with the Charter of Fundamental EU Rights and the Treaty of the European Union.

This decision by the European Court could have been the end of the data retention conflict in Germany. Without the EU directive, there was no longer an obligation to incorporate data retention in national law. Furthermore, in Germany, the amount of public protest against data retention acts was remarkable (for details on political mobilisation, see Hornung, 2012, pp. 384–386). Notwithstanding these concerns, the Grand Coalition (i.e., the coalition of the CDU and the SPD) proposed another bill in 2015 after devoting a great deal of time to discussing different scenarios (e.g., the quick-freeze solution 5). While CDU officials pushed the issue, SPD members were disunited. After the terror attacks in January 2015 in Paris, the discussion accelerated and ultimately culminated in an agreement between the party leaders of the CDU and SPD to adopt another data retention act. However, this agreement was not simply accepted in the SPD. Thus, the question was decided in a SPD party convention (with a majority of 60% in favour of data retention). There was conflict on this policy issue for some time (Gathmann, 2015) because there was strong competition between the data protection and the security policy images. While the data protection policy image referred to the control of own personal data and civil rights, the security image emphasised the need for (limited) surveillance to prevent crimes. In late 2015, the German Bundestag adopted the second Data Retention Act. Again, interest groups outside of the parliament as well as the opposition Green Party in the Bundestag pushed hard against the law. In 2017, only a few days before the Data Retention Act of 2015 was set to come into force, the German federal network agency 6 declared that it would not enforce the act before the final court decision on its general application (see Bundesnetzagentur, 2017).

5.2 Organisation of the law-making process and plenary debates

Data retention was selected for this research as an important issue for internet policy entrepreneurs before and after the establishment of the DAC. In the period of study, the issue of data retention was highly controversial, and outside internet policy entrepreneurs strongly mobilised against it and used venue change strategies (going to court) to promote the data protection policy image. However, what can be expected from the law-making process in the Bundestag? Section 4 concluded that the DAC can only advise, and during the period of study, there were competing policy images in the DAC. Moreover, parliamentary decisions are processed in serial mode, which means that party competition has a strong effect. As there was the same coalition government, with no change in party preferences, in 2007 and 2015, one possibility is that there was no change at all. Indeed, there was no change regarding the final decision on the data retention bills. However, given the theoretical background of this study, an expectation of incremental change is more appropriate. Therefore, this study first examines the difference of involved committees and explores potential strategies of DAC members to compensate for the weak institutional position of the DAC. The study then turns to the plenary debates in order to identify traces of an internet policy image.

The leading committee in the data retention law-making processes in both 2007 and 2015 was the Committee of Legal Affairs. The DAC was among other committees advising the process (see Table 1). There was frequent information exchange between members of the DAC and all relevant federal ministries (Ministry of Economy, Energy, Transport & Digital Infrastructure; Ministry of Legal Affairs and Consumer Protection; and Ministry of Transportation). In addition, political groups regularly met in parliament to discuss digital issues and coordinate actions, and staff members met occasionally in formal and informal fora (see Interviews 1, 2, and 5). Moreover, some members of the DAC were also (deputy) members of other related committees (e.g., internal or legal affairs) and thereby had access to established venues for internet issues. As such, the DAC members, as members of other concerned committees, could have introduced internet policy arguments that were subsequently adopted in the debates. If this were the case, the analysis should find differences in the discourse networks of the data retention acts.

For the plenary debates, this research posed two competing possibilities. First, the analysis could find no change in the debate between 2007 and 2015, as there were already well-developed, legal arguments against data retention with an emphasis on freedom going back to the 1970s. The call to prevent an Orwellian state of surveillance, for example, was paramount for internet policy entrepreneurs, but it had been developed by data protection entrepreneurs since the late 1970s. This overlap is not surprising, as the early hacker community of the 1980s was (and still is) an important actor in the German data protection movement. Therefore, internet policy entrepreneurs could have simply decided to reinforce a data protection policy image in 2015.

The second possibility is that there was an incremental change through which the data protection policy image was expanded by internet-specific and more technically informed arguments in 2015. Internet policy entrepreneurs argued that their expertise was lacking in parliament. Hence, we seek to find any (or an increase of) internet-specific arguments over time that can also be attributed to members of the DAC.

As the analysis aimed to identify potential changes, we created codes for the analysis inductively. The analysis revealed many legal and security claims as expected, but also a small number of more technical arguments similar to those made by internet policy entrepreneurs outside of parliament like the Arbeitskreis Vorratsdatenspeicherung (http://www.vorratsdatenspeicherung.de) and especially Kurz and Rieger (2009) 7.For example, the arguments against data retention that all data can be hacked and abused or that the principle of data minimisation should apply were coded as internet policy claims. 8

Figures 1 and 2 illustrate the relevant affiliation networks for the data retention acts of 2007 and 2015. 9For the first data retention act, 21 general claims were used, with 126 references for the plenary debates in 2007. For the second data retention act, the number of general claims increased slightly to 25, but with a remarkable increase in references to 306 for the plenary debates in 2015. This finding can be interpreted in several manners.

First, there were clear discourse coalitions in favour and against the data retention bills in both plenary debates, that conform with the logic of the parties in government and in opposition. This finding is notable for the parties in opposition, as the Liberal Party (FDP), the Left Party (die LINKE), and the Greens (BÜNDNIS 90/DIE GRÜNEN) exhibited differences in many other policies, but agreed on this issue.

Secondly, the discourse network in 2015 was much denser than in 2007, as predicted from the increased number of statements. However, the higher density is not evenly distributed, in large part due to the increased number of positive statements. Thus, the pro-data-retention coalition clearly dominated the discourse in 2015 (comparison of negative statements: 55 in 2007 and 58 in 2015; comparison of positive statements: 67 in 2007 and 226 in 2015).

In Figures 1 and 2, the most prominent and contested 10 claims are placed between the discourse coalitions. Figures 3 and 4 depict all claims in 2007 and 2015 measured by their indegree centrality. Four findings are notable. First, there were nine references to two internet claims in 2007 compared to 41 references to all four internet claims in 2015. This increase in both claims and references to them indicates that the debate became slightly more internet-specific. Secondly, the top four claims were identical in 2007 and in 2015. Thirdly, the top four claims consist of three legal claimsand one security claim. This finding conforms with previous studies that found a discursive shift towards a preventive state (Bukow, 2011, pp. 35-38), but also indicates that this move was highly contested in parliament. 11 Fourthly, among the top six most prominent claims, there were mostly legal claims in 2007. In contrast, there was one internet claim (data can be protected against hacking and abuse) 12 in 2015 which was also present, but neglected in 2007. The gain in prominence of this internet claim can be interpreted as a post-Snowden effect, as proponents of the data retention bill argued that data are safe because they must be stored on file servers in Germany.

On the individual level, 15 MPs spoke in the plenary debates of 2015, of which four were members of the DAC. Three of those were also members of other important concerned (internal affairs) or even responsible committees (legal affairs, and consumer protection). There is no systematic relationship between being a member of the DAC and referring (negatively) to an internet claim in our data. 13 This finding corroborates with the fact that there was no dominant internet policy image in the DAC. Internet experts, in fact, mostly referred to legal claims, which suggests that they did not feel references to internet claims would resonate as strongly with the majority of MPs. Two additions to the analysis at the level of claims suggest that discourses have shifted slightly towards the direction of internet policy. In 2015, there were some references to experts in the internet community outside of parliament which became acknowledged only after 2010, when the Enquete Committee raised their general credibility. There were also relatively technical remarks in the plenary debates in 2015, whereas internet activists outside of parliament complained in 2007 that MPs lacked technical expertise.

In contrast to Scheffel (2016, p. 102), who found no increase in the density of the debate and negated that new arguments were introduced over time, our results indicate that the discourse did change over time. However, the clear dominance of mostly legal and (only) then security claims was striking in 2007 as well as in 2015.

5.3 Internal mobilisation

The discursive network analysis in section 5.2 stated that it was not only the density of the discourse that increased in 2015 compared to 2007, but also the number of positive statements on data retention law-making that were made by members of coalition parties. This finding is in contrast to the voting results that are analysed in this section. Table 2 presents the results of the votes on data retention laws in 2007 and 2015.

Between 2007 and 2015, the number of dissidents (i.e., the sum of negative votes and abstentions only from coalition parties) increased drastically from 13 to 50 MPs. In 2007, most MPs in the group of dissidents abstained from the vote, whereas in 2015, most MPs in that group voted against the bill. All dissidents came from the SPD, whereas the CDU voted unanimously in favour of the bill. Of all final bill votes in 2015, this bill received the highest number of opposing votes.

As illustrated in section 4, the SPD never agreed on data retention legislation. The only agreement was among members of the DAC: Social Democrats in the committee were always unified against data retention (see Interviews 4 and 3). However, committee standing orders do not allow individuals to jeopardise the vote. Thus, all members of the DAC were forced to vote for the bill. To bend this rule, they asked their deputy members for representation. In this way, the dissidents paved the path to vote against data retention in the final vote in plenum.

In short, Social Democrats abstained from the vote in the committee in order to vote against the law in the plenary. In fact, they could have attracted significant attention with a scandal vote in the committee that broke the standing orders of the committee. However, they expected disapproval by the party leader in parliament and isolation in their own political party, as a small group of MPs cannot usually embarrass the whole coalition on a single policy issue (here, data retention). Instead, Social Democrats were able to mobilise others by convincing them of their view (see Interviews 2 and 4), which explains the many opposing personal votes in the 2015 vote. As a downside to their covert internal mobilisation, the Social Democrats of the DAC were not acknowledged as internet experts inside or outside of parliament (see Interviews 4 and 2). In fact, neither individual nor collective efforts could have changed the final vote. Data retention was a top-down decision from the start (see section 4). Party leaders were in favour of the legislation and therefore pushed it through anyway. Nonetheless, the high number of 50 dissidents in 2015 attracted some attention inside and outside of parliament.

6. Discussion and conclusion

This research analysed the DAC as a potential venue (see section 4) and examined data retention (see section 5) as a relevant issue to contribute to the broader question of how new policies emerge. This research concluded that the DAC was not the venue for internet policy that was sought by internet policy entrepreneurs of the time. First and foremost, the DAC could only advise other committees and could not lead in law-making. In addition, the DAC did not promote a consistent internet policy image with an emphasis on freedom of the internet and civil rights in the Bundestag. From the start, there was a competing digital policy image with an emphasis on economic opportunities. Members of the DAC adopted differing policy images and failed to develop a shared policy image. As a group, they did not present themselves as internet policy experts and were not acknowledged as such from inside or outside of parliament. This is not to say that individual members did not acquire expertise, but rather, that they could choose to become either internet or digital experts.

In the case of data retention law-making, on the one hand, there was incremental change in the plenary debate and mobilisation efforts from members of the DAC to compensate for the weak institutional position of the committee. Social Democratic members of the DAC even found a way to vote against the data retention bill in 2015. On the other hand, the change in the debate, which became slightly more internet-specific, was an expansion of the data protection policy image rather than a change in principle. Regarding mobilisation, Christian-Democratic members of the DAC did not join the opposition against the bill, but stayed within the security policy image.

These findings do not mean that the creation of the DAC did not have any effect at all (or that it will have no effect in the future). As a permanent committee, the DAC would eventually become a venue by making its members experts on certain issues. In fact, the DAC strongly drifted to become a venue for digital policy. The evidence in this article of such a shift is supported by the theoretical framework. Punctuated equilibrium theory suggests that shifts in media and public attention are critical for the success of policy images and the mobilisation strategies of interested actors. In fact, digitalisation and economic topics associated with the internet rose in prominence in the media and public attention shortly before the DAC was created (Hösl, 2019, p. 286). Beginning in 2015, migration policy and populism finally dominated the public agenda, which detracted attention from internet and digital policy.

What do our results mean more generally? First, the results in this research cannot be generalised to internet or digital policy in other parliaments. Venues where MPs become accustomed to a policy image remain important. If this is not the case, policy images can easily shift, because parliaments experience regular and rather frequent changes in membership. However, parliaments are organised in different ways. Thus, the interaction of venues and policy images and mobilisation strategies are different. Nevertheless, the theoretical foundation adopted in this research can be adapted to further cases, such as the Dutch or European Parliament.

Second, the results regarding the Bundestag in this research complement the previous work of Hösl and Kniep (2019) on the executive. They found an overall semantic change from information society in the 1990s, to internet policy, to digital and cyber policy in the 2000s. More specifically, the Ministries of Internal Affairs and Economic Affairs both neglected the discourse on the internet in society, but incorporated internet and digital issues in established security and economic discourses. The result was different ministerial discourses until the Digital Agenda of the Federal Government fostered a uniform discourse in 2014. The contribution of Hösl and Kniep (2019) helps to explain how, translated into PET language, the digital policy image was reinforced and the internet policy image was destabilised in the DAC via its contacts to these ministries. We deliberately used our theoretical concepts in the last sentence since their theoretical perspective can easily be translated into the theoretical language of PET. The primary claim of Hösl (2019) about the salience of issues and attention is identical with that of Baumgartner and Jones (2005, 2009, 2012).

Third, the contribution of this research directly relates to the international policy process literature. The DAC and internet policy are of substantial interest for internet policy entrepreneurs and academics in related fields. In general, this research contributes to the still small number of recent articles that have employed qualitative methods and PET to understand policy dynamics (see section 2). A next step to progress this field would be to better integrate longitudinal data on changing media and legislative, executive, and party agendas from the Comparative Agendas Project (https://www.comparativeagendas.net/) into studies with a small number of cases. The author’s opinion about the shift in public attention in 2015 could, for example, be substantiated with data from The German Policy Agendas Project (https://gpa.uni-konstanz.de) once its data extends past 2013 to include internet or digitalisation as topics.