Policymaking through opposing ideas? Framing conditionality and unconditionality in Finnish parliamentary discourse

In the ideational branch of institutionalist literature, policy outcomes are often explained in terms of a single idea or a set of ideas prevailing over others. This article explores a unique case where the Finnish government simultaneously introduced policies grounded in the ideas of unconditionality – in the form of a basic income experiment – and conditionality, reinforced through reforms concerning unemployment benefits. The study focuses on the ways in which political actors, in their framing strategies, connected their arguments to cognitive paradigms and normative public sentiments located in the background of policy debates. Through a thematic analysis of documentary data, the study finds that the seemingly contradictory unconditional and conditional measures were made compatible by framing them in connection to the prevailing cognitive activation paradigm. However, the framing of stricter conditionality additionally made use of normative arguments, connecting to public sentiments on what may be considered appropriate.


Introduction
A growing branch of institutionalist literature has turned to ideas as an explanatory factor in policy analyses. The ideational approach views institutional stability and change as resulting from sentient agents' actions, which in turn are influenced by their values and causal beliefs. Consequently, policy formulation and political decision-making can be portrayed as a competition between ideas or a struggle of interpretation of policy goals and policy alternatives, as policy actors attempt to construct the need to reform and 'sell' their policy proposals to the public. (Béland, 2005;Béland, 2016;Cox, 2001.) Prior literature on ideas has emphasised that ideas change and paradigms shift over long periods of time (Hall, 1993). Moreover, empirical literature on ideational stability and change has often explained the outcomes of policy processes in terms of a single idea or a set of ideas prevailing over others. (Blyth, 2001;Cox, 2001Cox, , 2004Lieberman, 2002;Niemelä & Saarinen, 2012;Kuivalainen & Niemelä, 2010;Perkiö, 2019.) In contrast, this case study provides a unique case in which social policy has been developed simultaneously on the basis of two opposing ideas.
Seeking to contribute to the literature on institutional change and ideational processes, this article examines the processes in Finnish social policymaking that have simultaneously advanced the ideas of unconditionality -in the form of the basic income (BI) experiment conducted in 2017-2018 -and conditionality, specifically in the case of unemployment benefits, where conditions and sanctions were reinforced through reforms introduced in 2016-2017. Several authors have noted the apparent contradiction of these policies (Halmetoja et al., 2018;Perkiö, 2019). The objective of this article is to examine and explain how ideas and ideational frames were used by political actors in these controversial processes.
The ideational approach emphasises that ideas that actors communicate through their discourse 'inform not only their belief in what they want but what they deem to be appropriate, legitimate, and proper.' (Béland & Cox, 2011, 3). Policymakers' success in legitimising policy decisions is largely dependent on their ability to communicate their ideas in a politically and culturally acceptable manner (Béland, 2005). Framing may create imaginary realities and open or close windows of opportunity for policy alternatives. Therefore, those who can define the interpretative frame for a given policy have a great deal of power. The understanding of the recent developments in Finnish social policies can be advanced through a close analysis of the ideas and discourse in which the contradictory approaches, i.e., more conditional and entirely unconditional benefits, were framed and 'sold' to the public as a coherent set of policies. The contribution of this article is to make this process visible, by showing how policymakers utilised cognitive and normative framing to subsume both approaches under the meta-frame of increasing employment.
As a further contribution to literature on ideas, this study addresses the need, highlighted by Campbell (2002, 27), to explore processes by which 'frames are constructed, tested, transformed, and fit to the prevailing normative frameworks and cognitive paradigms residing in the background of policy debates'. Rather than focusing on a single aspect of political decision-making, the analysis is directed both to government bills and to parliamentary debates, which allows exploring both the factual and the more symbolic aspects of ideational framing. Engaging in a form of 'thick description' (Geertz, 1973) of relevant policy debates, the study seeks to provide information on the causal mechanisms that may facilitate or inhibit change (Campbell, 2004, 55-56). Finally, this study also expands the literature on Finnish political discourse on BI (Koistinen & Perkiö, 2014;Perkiö, 2019), by connecting it to the broader discourse on the unemployment benefits that BI has been proposed to replace.

Ideas and framing
Ideas can be understood as the changing causal beliefs, values and perceptions that provide individual and collective actors with 'guides for action' (Béland & Cox, 2011). The turn to ideas in institutional literature has expanded our understanding of policy development, demonstrating that institutional stability or change can often be explained through the interplay of ideational processes relating to institutional constraints, interests and political power. (Béland & Cox, 2011;Carstensen & Schmidt, 2016;Schmidt, 2010.) In theoretical literature, ideas have been identified, among others, as paradigms, broad political ideologies, widely shared cultural understandings and public philosophies, as well as policy programmes, problem definitions and policy solutions. (Béland, 2016;Daigneault, 2014;Hall, 1993;Mehta, 2011.) Campbell's (2004) typology distinguishes between four types of ideas: policy paradigms and public sentiments, located in the background of policy debates, and programmes and frames, located in the foreground of policy debates. Paradigms are understood as structured, cognitive background assumptions that constrain the range of alternatives available to decision-makers (see also Hall, 1993;Daigneault, 2014). Public sentiments, on the other hand, refer to the normative assumptions and shared understandings that limit what decision-makers perceive as acceptable and legitimate to their constituents and to themselves. In the foreground of policy debates, programmes are prescriptions about how institutions and instruments should be used in specific situations, which helping to provide courses for policy action. Finally, frames are the symbols and concepts that policy actors use to legitimise policy proposals and make certain pro-grammes more appealing than others.
Framing is first and foremost a strategic tool, most apparent in political debates where political actors define problems -constructing the 'need to reform' -and prescribe policy solutions, seeking to justify their ideas to other actors and to the public. (Béland, 2005;Campbell, 2004;Cox, 2001;Mehta, 2011.) Framing is, essentially, a struggle of interpretation (Pfau-Effinger, 2005) in which actors use their ideas to influence discourse and, by highlighting certain features of a political issue over others, encourage it to be viewed in a particular light. The way problems and solutions are framed can considerably contribute to certain policies becoming implemented over others. (Hiilamo & Kangas, 2009;Kangas et al., 2014.) The analysis of political discourse is one method for understanding how such ideational processes play out and how they affect policy outcomes (Schmidt, 2011).
Ideas, including frames, can be cognitive or normative. Cognitive ideas are used to provide guidelines for political action and to legitimise policies and programs by 'speaking to the interest-based logic and necessity' (Schmidt, 2008, 306), presenting them as the most robust solutions to specific problems. Normative ideas, on the other hand, attach values to political action, speaking to how policies resonate with the principles and norms accepted among the general public, thus basing their value on their appropriateness. Literature on ideational framing has emphasised the latter type, arguing that attaching frames to culturally accepted and shared norms likely renders them more acceptable not only to other policy actors, but also and especially to the public. (Béland, 2009;Cox, 2004;Hiilamo & Kangas, 2009;Kangas et al., 2014.)

The basic income experiment and unemployment protection
Since the rise of unemployment in the early 1990s, Finnish labour market and social policies underwent a shift which has been called paradigmatic (Kantola & Kananen, 2013). The earlier focus on universalism and redistribution was replaced by the ideal of national competitiveness, and social policy turned to more targeted and selectivist measures aimed to alleviate poverty (Kananen, 2012;Kuivalainen & Niemelä, 2010). Employment was encouraged by obliging participation in the labour market on both economic and ethical grounds, and by identifying and fixing systemic issues related to work incentives. (Björklund, 2008;Keskitalo, 2007.) In the case of unemployment benefits, from 1990s to 2010s most governments, regardless of composition, 1 have either reinforced the conditions for benefit eligibility or increased sanctions for refusing offers to work or to participate in employment services. (Varjonen et al., 2019.) The social legitimacy and public support of benefit schemes largely depends on the public perceptions of deservingness. Across Europe, the public has been shown to consider the unemployed and immigrants as less deserving of social protection than the elderly, the sick and the disabled. As activity and individual responsibility for one's welfare are highly stressed at the political level, it is likely that in the public view, the unemployed assume more blame for their situation than the other categories, among whom there can be less doubt about the involuntariness of their neediness. (van Oorschot, 2006;van Oorschot & Roosma, 2017.) However, attitudes towards the welfare state in general vary to a great extent between supporters of different political parties. In Finland, voters of centre-right parties have been shown to be more critical towards welfare measures and likely to expect that generous social security risks making its recipients lazy. (Saari, 2017, 89-92.) The high priority on activation and individual responsibility at the level of both policy-making and political discourse has affected discourse on BI as well. One of BI's core principles is that the right to income is not conditional on any behavioural requirements (van Parijs, 1995), which is in conflict with the prevailing activation paradigm (Wieshaupt, 2011). However, BI proposals in Finland have largely been framed to accommodate the policy paradigm of their time. Where some of the proposals in the 1980s embraced the ideas of personal freedom and self-development, the BI models of the 1990s and 2000s have been presented as systemic modifications that solve pragmatic problems, rather than as radical reforms that advance a universalist concept of social justice (Koistinen & Perkiö, 2014). More recently, BI advocates have argued that the unconditionality of BI would encourage individual activity: as BI would not be conditional on behavioural requirements and it would not be reduced by other types of income, it would reduce welfare bureaucracy and incentive traps that are considered to discourage jobseekers from taking on short-term and part-time jobs. The argumentation has therefore progressed from emphasising normative ideas to utilising primarily cognitive framing. By highlighting the way that BI would enable everyone to enter the job market and reward all activity, this framing also aims to undermine normative objections to BI, which have largely been grounded in the principle that income support only belongs to those who need it or are willing to reciprocate in return. (De Wispelaere, 2015, 95;Perkiö, 2019.) In late 2014, the Centre Party (CP), then the largest party in the opposition, suggested that Finland should conduct regional experiments on BI in order to facilitate reconciling social security with short-term and part-time work. After the 2015 parliamentary election, the CP formed a new government together with the Finns Party 2 and the National Coalition Party (NCP). In its platform, the government promised to improve work incentives and shorten periods of unemployment, and seek to improve the consolidation of benefits with earned income. The key measure would be an unemployment benefit reform, which would include stricter obligations to participate in employment and activation services. The government would also promote 'a culture of experimentation', to develop new solutions to societal problems. One of these experiments would be an experiment on BI. (Prime Minister's Office, 2015.) The task of planning the experiment was given to a research consortium led by the Social Insurance Institution of Finland, which proposed a broad experiment with several sample groups with different levels of BI and income tax rates (Kangas & Pulkka, 2016). However, the government's preparation of the experiment legislation was influenced by budgetary constraints, legislative issues and time limitations. A government proposal for the Act on the BI experiment was drafted quickly and presented to the Parliament in the autumn of 2016 (HE 215/2016). The stated aim was to study what effects an unconditional BI would have on the target group's labour market behaviour. Referring to budgetary limitations, the government argued that target group should be limited to those whose behaviour BI could be expected to influence the most. Therefore, the group would consist of jobseekers whose basic-level unemployment benefit would be replaced with a BI of 560 € (equivalent to the unemployment benefit after taxes) for a period of two years. Their conditional unemployment benefit would be made temporarily unconditional. They would no longer be required to register as unemployed jobseekers at the public employment services, and they would receive no sanctions for refusing services or job offers. They would also be able to study or work in paid employment or as an entrepreneur without their BI being affected. The results would be measured on the terms of how the unconditional benefit would affect the participants' likelihood of finding employment or participating in activation measures.
Alongside the BI experiment, the government implemented several reforms to unemployment benefits, primarily concerning the conditions of conduct, i.e., the behavioural requirements that benefit claimants are expected to fulfil to maintain eligibility (Clasen & Clegg, 2007) (see Table 1). First, the government increased waiting days and shortened the maximum duration of the earnings-related allowance (HE 113/2016). Then, it made the obligations as well as the sanctions related to accepting job and training offers stricter (HE 210/2016). In addition, it made periodical jobseeker interviews, which had formerly been conducted at the discretion of the employment services, obligatory after every three months of unemployment (HE 209/2016). In the final major reform, jobseekers were obligated to fulfil an 'activity condition' by working in salaried employment or participating in employment training for a specific time for every three-month period of unemployment. Failure would result in the benefit being cut by 4,65 % for the following three-month period. (HE 124/2017.) A common factor to these reforms was the strengthening of conditionality: jobseekers were gradually asked to put more effort into either looking for work -even if only for short term -or participating in services that might aid them in finding employment.
The BI experiment and the unemployment benefit reforms differ in that the former was a two-year experiment with a relatively narrow target group, whereas the latter constituted legislation that would be in effect indefinitely. However, the decision on when and how to conduct policy experiments is just as much a political decision as is passing other legislation. Furthermore, when viewing the policy processes as a whole, both types of measures underwent the same stages of the policy cycle, i.e., agenda-setting, formulation, decision-making, implementation and evaluation (Howlett et al., 2009). The BI experiment also required new legislation, whose drafting was preceded by extensive analyses concerning the way BI relates to existing national and EU legislation (Kalliomaa-Puha et al., 2016;Kangas & Pulkka, 2016). There was an apparent contradiction in the government's approach to unemployment benefits: the aim of the BI experiment was to test the behavioural effects of making unemployment benefits unconditional, while at the same time the same benefits were gradually being made more conditional for other benefit recipients. As unconditionality was at the core of the entire BI experiment, this raises the question of how it was made compatible with policies that modified the unemployment benefit system in the opposite way.

Aim, data and methods
In light of the theoretical and empirical literature on ideas and framing, as well as the literature on BI and unemployment benefit conditionality in Finland, this study analyses the way that the two opposing ideas of unconditionality and conditionality were reconciled in political discourse and how the framing of policy solutions was connected to previously identified background ideas (Campbell, 2004). Furthermore, rather than studying BI discourse in isolation, this study places it in a broader policy context concerning benefit conditionality and jobseekers' status.
The main data consist of policy documents concerning the BI experiment and the unemployment benefit reforms implemented during the centre-right government's term. This documentation includes government bills, committee reports and parliamentary discussions concerning five legislative changes  Table 1). To supplement these with other relevant parliamentary discussions, keyword searches on perustulo (basic income) and perustulokokeilu (basic income experiment) were conducted using the Finnish Parliament's database. The government bills and committee reports were included in the data in order to obtain a more complete picture of the 'factual' justifications for the proposed reforms, and to see if and how they were used in the framing of the issues in the parliamentary debates. In total, the assembled data consisted of approximately 60 documents, of which the parliamentary discussions deemed relevant for the study totalled over 400 pages (see Appendix for a list of referenced documents). The data were analysed through thematic analysis, which is concerned especially with the identification and analysis of repeated patterns of meanings in a set of data (Herzog et al., 2018). The objectives were to identify key actors and their proposals; the problems the proposed policies were designed to address; what policy alternatives or programmes were discussed; and how they were ideationally framed. The primary focus was on discourses concerning benefit conditionality, unconditionality, and activation. The aim was to see how different parliamentary actors framed their ideas in political discourse. The analysis followed an inductive, data-driven approach, in which the themes were identified in the data, rather than attempting to fit them into a pre-existing coding frame. Whereas a deductive approach is useful when pre-existing categories are established and can be utilised, induction allows for a categorisation where the identified themes are more closely linked to the data. Braun and Clarke (2006) have distinguished six phases of thematic analysis: familiarising oneself with the data, generating initial codes, searching for themes, reviewing themes, defining and naming themes, and producing the report. Following this process, the data was first read and re-read and then initially coded into ten categories. These were then further analysed for themes and eventually collated into three major themes of analysis: incentive traps, the activating effect of employment services, and prevention of social exclusion.
From a validity standpoint, focusing on discourse risks overlooking institutional and interest-based explanations for policy development. However, the starting assumption of this study is that policy documents, including parliamentary proceedings and speeches, provide direct evidence of the ideas of policy actors (Daigneault, 2014). Furthermore, a case study involving a 'thick description' (Geertz, 1973) of policy processes, exploring for and interpreting signification in the data, may allow for a more in-depth analysis of the mechanisms that facilitate of inhibit policymaking. (Campbell, 2004, 55-56.)

Incentive traps
From the early plans regarding the BI experiment, the government made it clear that its goal was to reform social security to encourage participation and improve work incentives. The government's proposal itself (HE 215/2016), in a straightforward manner, specified key problems of the social security system. First, in some instances, short-term employment could have little or no impact on an individual's income, due to taxation and to the withdrawal of means-tested benefits. Second, accepting short-term work could result in delays and interruptions in benefit delivery. Third, the system was considered too complex and difficult to comprehend from the claimant's point of view, which made the extent of benefit withdrawal in the event of low-income work difficult to predict. The BI experiment's stated purpose was to examine what effects the removal of these unemployment traps and bureaucracy traps would have on the target group, i.e., recipients of basic-level unemployment benefits. The goal was to find whether this would increase the target group's -primarily long-term jobseekers and labour market entrants -employment, or their participation in employment services. The government's proposal thus framed the experiment of an unconditional BI almost exclusively as a way to study incentive traps and the effects of their removal. Hence, the idea of unconditionality, inherent in BI, was enveloped in the broader policy goal of increasing employment.
The Green League and the Left Alliance heavily criticised the proposal for its limited target group and the lack of changes to taxation. The experiment would provide no information on BI's effects on, for example, self-employed persons or low-income temporary workers, and in any case, the model would be too expensive to scale up for the entire population. The Greens, who marketed BI as 'reforming social security to serve the needs of the changing working life', made a separate legislative initiative proposing a more comprehensive experiment (LA 103/2016). The opposition parties also criticised the fact that BI would have no effect on supplementary means-tested benefits (primarily housing benefit and social assistance) that many jobseekers received. As these individuals would continue to suffer from this 'bureaucracy trap of all bureaucracy traps' (Left Alliance member, PTK 106/2016), the experiment would produce trivial results. Thus, the main point of criticism was that the experiment was harnessed to serve goals that have little to do with a true universal BI.
In the parliamentary debates, the CP and the Finns Party were the most active defenders of the proposal, emphasising that the difficulty of reconciling short-term work with social security benefits was a topical issue that everyone wanted to solve. The changing labour market required innovative solutions to simplify the social security system and to improve work incentives. Building on the argumentation established in the government's proposal, they argued that the experiment would provide information on how incentive traps and bureaucracy discourage employment and on how improving incentives affects jobseeker behaviour. As the Minister of Social Affairs and Health (then a member of the Finns Party) put it, 'as we all know, our current social security system is quite the patchwork, and there are very challenging incentive traps for many groups of people' (PTK 30/2016). Rather than lead to an implementation of BI, the information that would be gathered could later be used in developing the benefit system and designing a simpler social security model. Thus, the experiment could become a 'bridge to the new social security of the 2020s' (Minister of Local Government and Public Reforms, CP member), which adapts better to shortterm work and the reality of individuals having multiple sources of income. The experiment would also advance the government's efforts of introducing an 'culture of experimentation' to policy development. The purported lack of incentives was also the main reason for increasing the conditions and sanctions in the unemployment benefit system. Each of the government's proposals on unemployment benefit reforms stated that the goal was to increase employment by incentivising jobseekers to be active and to find work as quickly as possible. Shortening benefit periods, increasing jobseekers' personal contacts with employment services, and making sanctions for refusing job offers or services more demanding were all expected to have positive behavioural effects. The measures should make jobseekers more willing to look for and accept offers for the kinds of short-term and part-time work that they would otherwise not take up. (HE 113/2016;HE 209/2016;HE 210/2016;HE 124/2017.) In the parliamentary debates, the left-wing opposition parties argued that unemployment was predominantly a demand-side problem and that increasing conditionality and sanctions for not finding work would be unjust if there simply were not enough jobs available. However, the government, and especially the NCP members, framed unemployment as a structural issue. They argued that there were jobs in the hidden job market, which could be discovered only if jobseekers were committed and active. Increasing conditionality was seen as the best way to achieve this goal. The mismatch between job openings and jobseekers was a key part of the problem definition: there were even labour shortages in some areas, but the benefit system did not encourage mobility or retraining. For the government, reforms that would increase labour supply would help with these issues. The inherent logic, common to the efforts to increase conditionality, was that 'the human is a rational being, which reacts to incentives and acts according to them' (NCP member, PTK 144/2017), meaning that the unemployed could be made more active by appealing to their calculative self-interest (see Kildal, 2011). The government, and especially NCP members, brought up research on the behavioural effects of labour market policies, which had shown that the prospect of sanctions and of losing eligibility for benefits can increase the probability of a jobseeker to find work (PTK 134/2016). In particular, the impending time limit of earnings-related benefits was said to motivate jobseekers: 'it is astonishing how many find a job just a little before the benefit ceases to be paid' (NCP member, PTK 128/2016). If the opposition agreed that jobseekers would be better motivated to find work if there was a higher monetary reward for becoming employed (as would be the case with BI), by the same logic, the same result could be achieved by the threat of cutting benefits. In both scenarios, 'a larger gap would be created between the economic situations of a working and a nonworking individual' (NCP member, PTK 134/2017). While the government parties emphasised that they did not consider the unemployed lazy, they implied that due to incentive and bureaucracy traps, the current system encouraged at least a portion of the labour force to remain unemployed. Increasing conditions and sanctions could therefore 'push' jobseekers to seek and accept potentially unappealing short-term jobs that were further from their homes, that did not necessarily match their skills and that paid perhaps less than their benefits. 'In modern society, it cannot be assumed that in every situation a jobseeker finds the kind of work that he/she personally wants to do' (NCP member, PTK 128/2016).
The Social Democratic Party (SDP), the Green League and the Left Alliance did not challenge the idea of incentive traps as the main problem. The key difference was that in the opposition's view, the reforms would push those jobseekers who could not meet the tougher conditions to rely on social assistance. As this last-resort benefit involves heavier means-testing, it would only exacerbate the incentive traps and bureaucracy these individuals faced. (StVM 22/2017;PTK 135/2017;PTK 144/2017.).
The government parties' framing of both the BI experiment and the unemployment benefit reforms were strongly tied to the activation paradigm that has since the 1990s dominated the Finnish discourse on labour market policies. Incentive traps and bureaucracy traps were recalled as the systemic problems that cause structural unemployment. Solving them was therefore in the interest of every party, whether they were in the government or in the opposition. It is notable that the opposition parties largely accepted the idea of incentive traps as a key problem, and the concept of jobseekers as rational actors who respond to incentives. This highlights the effectiveness of connecting the framing of a policy solution to a prevailing policy paradigm. Further, when framed in such similar ways, the simultaneous implementation of policies grounded in the ideas of unconditionality and conditionality seems a little less incongruous. In the government parties' framing, the goal of the BI experimentation was not the implementation of BI, but the gathering of information in order to increase jobseeker activity, to reduce incentive traps and to further develop the social security system to everyone's benefit.

The activating effect of employment services
Another argument for active labour market policies has been that money should be spent to support 'active' participation instead of 'passive' distribution (van Berkel & Møller, 2002). This type of framing also played a role in the unemployment benefit reforms, prompted especially by measures that made it possible to use funds appropriated for unemployment benefits as pay subsidies. More broadly, the government's proposals emphasised the active role of employment services and especially employment plans, in which the services and job-search goals for individual jobseekers are outlined. The government argued that mandatory jobseeker interviews would allow more frequent contacts with the services and more regular updating of these plans. This, in turn, would make employment services more efficient, increase jobseekers' activity and participation in services, and consequently improve their likelihood of finding work. (HE 209/2016.) In the parliamentary debates, many government parties' MPs also emphasised the role of public employment services in promoting employment. Jobseekers' situations would be assessed individually to provide them with 'personalised help and support' (NCP member, PTK 38/2018) and to determine what types of actions and services would best help them find work. For a jobseeker, meeting employment services professionals personally was a way to build a 'path' back to working life. A frequently cited example was Denmark, where jobseeker interviews, coercive measures and stricter sanctions were said to produce a higher employment rate and a more equal society. (PTK 135/2017.) It is notable that the importance of employment services was largely accepted also among the opposi-tion parties. Especially the SDP hoped for more resources for activation policies that 'encourage people to participate in rehabilitative work, training and employment' (SDP member, PTK 134/2017). This opinion was most evident in discussions concerning the activation model, which would cut benefits for jobseekers who did not work or participate in services for a specified time every three months. This was a particularly difficult question for the left-wing parties, because it meant that jobseekers could be penalised regardless of their efforts, if no services were offered. (StVM 22/2017; TyVL 10/2017.) SDP members made frequent references to the statement of the Constitutional Law Committee (PeVL 45/2017), which had also viewed this aspect as problematic. While some MPs from the Left Alliance criticised increasing sanctions in general, for most SDP members, the core issue was the inequality in access to the services that could spare jobseekers from sanctions. A possible solution that was repeatedly brought up by the SDP and the Greens, and one that the Committee also had proposed, was to either give jobseekers a subjective right to these services or to expand the range of acceptable activities so that everyone willing would be able to fill the activity requirement. 'The state must have an obligation to offer these activation services, and if the jobseeker refuses, it is acceptable that this results in some sort of sanctions' (SDP member, PTK 48/2015) (PTK 134/2017; PTK 135/2017; TyVL 10/2017.) The government parties' MPs' argumentation highlighted the contrast between 'active' measures versus the 'passive' role of unemployment benefits. The contrast was made all the more evident by naming the most controversial reform 'the activation model'. In part, allocating resources to employment services and pay subsidies was presented as simply a smarter and more efficient use of public resources. This argument was also extended to lament the 'old' labour market policy, which was purportedly only concerned with the provision of passive income transfers, after which 'there is zero contact with the unemployed person' (NCP member, PTK 135/2017). Jobseekers themselves would be required to change their mind-set, in that they should not take income support for granted but should always associate it with the obligation to be active and look for work.
Some NCP and CP members viewed passive benefits as having an almost paralysing effect, because they essentially pay people to do nothing: 'to leave a person to passively stay at home, looking at the walls, is not caring... it is abandonment'. It was the state's responsibility to provide unemployed individuals with meaningful services and 'help them help themselves rise out of poverty' (CP member, PTK 135/2017). Stricter obligations to participate in activation services were framed as a compassionate act. In turn, those fit to work had the responsibility to 'primarily strive to take care of their own livelihood' (NCP member, PTK 106/2016). If jobseekers did not find work that matched their skills or location, they were required to accept whatever work was available, because nobody should have a 'subjective right to enjoy income support without the obligation to work' (NCP member, PTK 136/2016). The benefit system was thus blamed for rewarding inactivity and allowing those reluctant to work to stay at home at other people's expense. (PTK 136/2016;PTK 134/2017;PTK 135/2017;PTK 38/2018.) Interestingly, the BI experiment, and BI in general, also drew criticism from some NCP and CP members on these same grounds (PTK 48/2015;PTK 38/2016;PTK 136/2016).
The way that the government parties framed conditionality in terms of employment services and activity thus drew from both cognitive and normative sets of ideas. Referring to the resources and improved effectiveness of employment services is a useful strategy, as it connects directly to the prevailing activation paradigm. Again, this can be very effective, as there is relatively little disagreement across the political spectrum on the role of services in promoting activity and employment. Ultimately, the question becomes not whether sanctions are right or wrong, but whether there are enough services for everyone. On the other hand, the contrast created in the government parties' MPs' discourse between passivity -both of the benefits and of the benefit claimants -and the ideal of activity touched directly on the normative appropriateness of receiving benefits despite being able to work. As Campbell (2004, 99) notes, framing often works through discursive oppositions of this nature.

Prevention of social exclusion
The Greens and the left-wing parties opposed the unemployment benefit reforms also on the grounds that stricter sanctions and benefit cuts would force jobseekers to rely on social assistance and eventually cause some of them to leave the labour market entirely, leading to poverty and long-term unemployment. In the government's proposals, however, stricter conditionality was marketed as a way to combat long-term unemployment: the reforms encouraged accepting any type of short-term jobs, which could in time turn into longer contracts or, at any rate, make jobseekers better qualified in the eyes of potential employers. (HE 210/2016;HE 124/2017;StVM 23/2016.) In the parliamentary debates, the CP and the NCP extended this argument by bringing up the association between long-term unemployment and social exclusion: 'According to some researchers, it is only a question of a few months [of unemployment], after which a person begins to become marginalised in society' (CP member, PTK 135/2017). Long-term unemployment was presented as a downward spiral that an intervention from the employment services could help solve or even prevent. The obligation for jobseekers to remain in more frequent contact with employment services would provide tools to take better care of 'those who, perhaps due to a longer period of unemployment, have become side-tracked from the labour market' (NCP member, PTK 136/2016). As different groups of jobseekers require different kinds of services, frequent contacts would also enable better assessment of individual situations and allow the jobseeker and employment service personnel to look for solutions together. Furthermore, coercing jobseekers to accept any kind of short-term job, instead of allowing them to passively wait for work that matched their skillset, would shorten their periods of unemployment; this could subsequently increase the likelihood of them finding jobs that match their skills in the longer term. (PTK 107/2016;PTK 126/2018;PTK 131/2016.) The government parties emphasised that young jobseekers had found personal contacts at the employment services particularly useful. They argued that employment services professionals often identified issues that could lead to marginalisation and necessitated a variety of professional assistance, which could then be provided. As a stricter obligation to participate in services would help prevent social exclusion, it should not be seen as a punishment but as a form of caring. (PTK 107/2016.) Coercive measures were necessary in themselves, because they made young jobseekers step outside and actually 'face the question of what they plan to do with their lives' (CP member, PTK 38/2018). Again, the current system was presented as problematic in this respect, especially in the context of intergenerational unemployment: 'at the moment, our system is too passive…we really have third-generation unemployed families, where [the youth] have no model of leaving their homes... if we do not encourage them to find work, we're doing a really immoral thing, especially towards our youth' (CP member, PTK 135/2017).
Framing conditionality as a tool to prevent social exclusion, especially that of the youth, is another strategy that reaches out to normative ideas of what is considered appropriate. While not directly blaming long-term jobseekers for their situation, this framing portrays them as passive subjects in need of guidance. Paying a benefit that is sufficiently high to enable subsistence, without enough reciprocity, is conversely framed as a disservice to its recipients, especially young jobseekers who may not themselves even know what is best for them. In other words, stricter conditionality is presented as serving the interest of the individual just as much as that of the society. What makes this type of framing different from the others is that the ultimate goal is not simply reaching a higher employment rate but also increasing the well-being of the society's citizens and, particularly, its youth.

Discussion
The objective of this article was to explore the process by which, in the course of three years, the Finnish government implemented an experiment on an unconditional BI, while simultaneously increasing the conditionality of unemployment benefits. The aim was to explain how the potential ideational conflict was resolved in the government's proposals and parliamentary debates, through an analysis of the processes by which frames are constructed and fit to the prevailing ideas in a society (Campbell, 2002).
The findings of this study highlight two contributions to the theoretical literature on ideational and institutional stability and change. First, while ideational literature has often portrayed policymaking as the outcome of a single idea dominating over others, the Finnish case demonstrates that it is possible to implement policies grounded in ideas of an opposing nature. The outcome of such a process is likely dependent on a careful balancing of cognitive and normative framing of the issues in question.
The framing of the government's proposals was limited to cognitive ideas (Schmidt, 2010). They placed significant emphasis on defining the problems (see Mehta, 2011) that needed solving, and subsequently establishing a causal relationship between the problems and the proposed solutions. The solutions were primarily framed in terms of the cognitive activation paradigm, arguing for ways in which the proposed policies would influence jobseeker behaviour and encourage the take-up of short-term work. In large part, the government parties' MPs expanded on this framing in the parliamentary debates, but their strategies also made use of normative ideas.
In terms of ideational framing, the conflict between unconditionality and conditionality was resolved by portraying the BI experiment simply as a way to procure information about incentive traps and about the behavioural effects that increasing incentives to work and temporarily removing sanctions would have on employment and on participation in services. The justifications for the BI experiment therefore directly connected to the dominant activation paradigm. Ultimately, the reconciliation of contradictory policies was facilitated by the contradictions of the BI experiment itself vis-à-vis the more traditional concept of universal BI. First, limiting the target group to unemployed jobseekers already distanced the experiment from universal BI. Second, while the key aim of the experiment was to test benefit unconditionality, the concept of unconditionality in itself was not featured at all in the parliamentary discourse, and normative arguments such as human freedom, emancipation, dignity etc. that BI advocates have usually presented (see e.g., Standing, 2005Standing, , 2017van Parijs, 1995;van Parijs & Vanderborght, 2017) were equally lacking.
Thus, BI was exclusively presented as an instrument to increase employment. This, to a large extent, is also in line with the way BI's political proponents have recently been framing the issue (Perkiö, 2019). The argumentation for the unemployment benefit reforms was also largely tied to the same paradigm, but the cognitive framing related to incentives, behavioural effects, and the benefits of active employment services was interlocked with normative framing that highlighted the appropriateness of activity versus the dangers of encouraging passivity and failing to prevent social exclusion. This framing emphasised everyone's responsibility to contribute either by working or, in the case of unemployment, by doing everything they can to re-join the labour market (Björklund, 2008). Framing welfare retrenchment measures by referring to the deservingness of benefit claimants is an effective way not only to build support for policy proposals, but also to influence public opinion (Slothuus, 2007). Rather than portraying the proposals as solutions to specific problems, this type of argumentation attempts to fit the frames to the normative frameworks that specify what the public would deem appropriate. Therefore, the choices policymakers make in terms of cognitive and normative framing plays a crucial role. The process of implementing potentially ideationally conflicting policies is greatly facilitated if the policies can, on a cognitive level, be framed in terms of the same paradigm and as helping solve issues that nearly all political parties can identify as shared societal problems. In this case, there was no attempt to accommodate normative framing of unconditionality and conditionality, but it is likely that this would be more difficult.
The second theoretical contribution concerns the association between ideas and institutional change. Some ideational scholars have pointed out that framing policy proposals in ways that connect with established background ideas -paradigms and public sentiments -is usually more likely to reinforce institutional stability rather than to expedite change (Campbell, 2004). In this case study, the introduced policies -with the exception of the BI experiment -can be seen as a part of a long process of incremental institutional change that has increased demands placed on unemployed individuals, particularly long-term jobseekers (Varjonen et al., 2019). As far as policy implications concerning BI can be construed, framing BI in ways that conform to the hegemonic (Ferree, 2003) incentives discourse is understandable in the light of BI's political feasibility (De Wispelaere, 2015;Perkiö, 2019) and can be effective in proposing experiments. However, overlooking normative framing, which has been shown to be effective in garnering public support, may not be the most effective strategy to advance its implementation.