Country Reporting on the Sustainable Development Goals—The Politics of Performance Review at the Global-National Nexus*

Abstract With the adoption of the Sustainable Development Goals (SDGs), debates on governance through goal-setting and performance review have gained momentum. In this article, we explore how the politics of performance review played out in public sustainable development reporting at the global-national nexus. By examining the case of Swedish reporting to the United Nations High Level Political Forum in 2017, we find policy translation, accountability preparation and identity formation to be key functions of SDG reporting. We draw attention to the performative and political features of these functions in the sustainable development realm. With a fast approaching deadline, policy translation of global indicators to the national context glossed over politically contentious issues. Reporting served to enable peer review among governments rather than hierarchical accountability of domestic politics. Moreover, the identity formation function of SDG reporting was strong on the international stage while domestically it was challenged by broader political struggles. In conclusion, our study bears witness of the formative power of public reporting for SDG governance. We call for comparative research to allow for further theory-building on the politics of public reporting on sustainable development.


Introduction
The adoption of the UN Sustainable Development Goals (SDGs) in 2015 marked the birth of a holistic global agenda that merges environmental and development concerns. In contrast to their predecessor, the eight Millennium Development Goals (MDGs) which targeted the Global South, the SDGs provide guidance for all countries worldwide until 2030. other countries. In methodological terms, the case is therefore an extreme rather than a representative case (Flyvbjerg 2006, 230). Moreover, choosing a high-income country enables us to study how such a country handles traditional divisions between international development cooperation and national sustainability politics when adapting the SDGs to the national context. Taken together, this allows us to explore functions of national public reporting as part of the broader trend of governance by goal-setting alluded to above. We expect the reporting process to matter beyond review and experience-sharing, the stated aims of the HLPF (cf. Hansen and Mühlen-Schulte 2012). Our research question is therefore: What functions does HLPF reporting fulfil in sustainable development politics at the global-national nexus?
The material for our case study consists primarily of 13 elite interviews with individuals who in different capacities participated in Swedish reporting to the HLPF in 2017: government officials of the Ministry for Foreign Affairs, statisticians at the government agency of Statistics Sweden, officials working for the government's independent committee for the 2030 Agenda, and representatives of large Swedish civil society organizations (CSOs) with both international and national focus. Interviews were conducted in Stockholm between December 2017 andFebruary 2018. For reasons of anonymity, we refer to interviewees in general terms, mentioning only their sectoral belonging (government official, statistician or civil society representative). Moreover, our empirical material comprises a rich stream of policy documents on the 2030 Agenda in Sweden and internationally, such as preparatory reports on which the official Swedish HLPF report was based. We also use news items and debate articles from Swedish newspapers as background material.
Our overarching contribution is to take the initial steps of theorizing the politics of public reporting on sustainable development by suggesting key functions of such reporting at the global-national nexus. While each of the functions is already the subject of different literatures, they have not before been considered side by side in light of goal-setting governance. Empirically there is still limited research on HLPF-reporting and on how the SDGs are adapted at national levels. This article, hence, also contributes a case study that can be used for comparative purposes in future studies of country reporting to the HLPF. Such comparison can nurture further theory-building on the politics of reporting. In short, our interest lies in how global governance goal-setting meets domestic policy-making-not in global governance itself. Below, we first review a selection of relevant research that offers theoretical underpinning for identifying functions of public reporting. We proceed by providing a brief overview of the HLPF and the Swedish reporting process. Thereafter, we explore three main functions of reporting in the case of Swedish HLPF-reporting in 2017: policy translation, accountability preparation and identity formation. In the conclusion, we call attention to how the global-national nexus created tensions related to those functions of reporting and we propose a future research agenda.

Public Sustainable Development Reporting at the Global-National Nexus
This section reviews a selection of research on public policy-making, governance by numbers and organizational theory in order to further elaborate our notion of public reporting on sustainable development at the global-national nexus. The section thereby provides theoretical ground for identifying possible functions of such reporting. To begin, our study concerns public reporting by governments which requires us to theoretically understand reporting as part of a process of public policy-making more generally. In theory, policy processes have often been described as continuous cycles of agenda-setting, policy formulation, decision-making, implementation and evaluation, even if most authors emphasise that these processes are by no means linear in practice (Sabatier and Weible 2014, 8). Of interest here is that the use of evaluations, such as rankings and performance reviews, has increased considerably in recent times (Van de Walle and Cornelissen 2014). This creates an 'evaluation society' where reporting practices become a central part of policymaking. Evaluations 'help to structure what we talk about, what we think are important problems, and what we think we will do about these problems' (Dahler-Larsen 2012, 4). This suggests that reporting practices impact all stages of the policy-making process and are formative for knowledge that underpins policy decisions.
While we define the reporting process in broader terms than measurable indicators, research on governance by numbers clearly offers theoretical leverage to our analysis. Such research emphasizes that the interpretation and utilization of numerical data is always dependent on someone to collect it, make sense of it, contextualize it and translate it (Leite and Mutlu 2017, 73). Statistical expertise plays a key role when quantitative data is used in order to justify policy decisions and ground them in an appeal to objectivity. Research also shows that statistics and other quantitative data might 'take on secondary lives' in the course of being transferred from one context to another. The significance, purpose and effects of quantified data may shift away from original intentions (Merry 2016;Beerli 2017, 57). For example, as 'raw information' collected by experts moves towards those making big decisions, it is often edited in a way that removes assumptions and ambiguity. This makes knowledge appear more robust than it actually is (Davis, Kingsbury, and Merry 2012, 8-9). Prior research has demonstrated that the power of numbers is great in the realms of development and sustainable development. Indeed, an earlier Special Issue of this Journal identifies several governance and knowledge effects of the MDGs. Intended to harmonize reporting on the Millennium Declaration, the MDGs eventually took on a life of their own and became definitional for development agendas. They biased performance evaluations and narrowed conceptualisations of development at the expense of human rights and capabilities perspectives (Fukuda-Parr 2014;Fukuda-Parr and Yamin 2015).
A key assumption in literature on public policy is that performance evaluation on public services enables accountability to citizens by making governments more transparent (Van de Walle and Cornelissen 2014). Political accountability is thus the ultimate justification for performance reviews. To be effective, accountability requires information on the basis of which citizens can hold governments accountable for their performance (or lack thereof) (Merry 2016, 3;Bexell and Jönsson 2017, 17). Interestingly, research on performance reporting shows that parliaments are not strengthened by the increased information flow on public services, contrary to expectations (Van de Walle and Cornelissen 2014, 6). This can partly be explained by the expertise required in the twin processes of data collection and data analysis, leaving performance evaluation and reporting to the domain of experts (Pollitt 2006, 52). At the global-national nexus, horizontal peer review among governments intersects with hierarchical accountability of domestic politics. What happens to chains of accountability relations at this nexus is an empirical question whose answer is likely to vary between political systems.
Broadening our focus beyond functions related to policy-making, literature of constructivist orientation calls attention to the formative effects of performance review on identities. Studies on organizational identity construction show that organizational communication through annual reports and the like, convey organizational identity claims to internal as well as external audiences (Schultz and Hernes 2013). Literature within the field of critical organization studies argues that sustainability reporting by private companies handles conflicting stakeholder demands by developing organizational facades (Cho et al. 2015). Research invoking a theatre metaphor further demonstrates that report presentation on different stages involves identity formation elements-actors strive to convince their audiences through their stage appearance. In this vein, Death (2011, 1-2) sees global summits as stages where 'political theatre' with symbolic and performative dimensions plays out in front of a global audience. Such theatre conveys that political elites are serious on global challenges, strengthening their self-image. Rigney (2001, 153) suggests a front-and backstage distinction highlighting that what happens on the frontstage is open to external audience sight while the backstage is only for fellow actors, hidden to the audience behind a curtain. Governments develop different rhetorical devices as stages and audiences vary and will perform according to particular 'normative scripts' of those stages (Halliday, Block-Lieb, and Carruthers 2010).
In conclusion, this theoretical discussion has allowed us to tentatively identify key functions of public sustainable development reporting at the global-national nexus. First, we should regard such reporting as part of a broader public policy-making process and our case study therefore serves to draw more precise conclusions on what policy functions reporting fulfilled in the early years of SDGs policy-making. Moreover, prior research makes clear that we need to view reporting as a manifestation of a broader trend of performance review for which accountability is the justificatory rationale. Our case study examines accountability functions of this particular case of reporting. Finally, research on organizational identity calls attention to reporting being an element of identity construction, depending on audiences' reactions.

Swedish Reporting to the UN High Level Political Forum
This section introduces the HLPF followed by a brief overview of the Swedish reporting process during 2017. The HLPF is mandated to review implementation of the 2030 Agenda, and to enable learning and best practice sharing. It is an intergovernmental forum for UN members, where nongovernmental actors are included through the Major Groups system. Annual HLPF week-long meetings are held under the auspices of the UN Economic and Social Council. During HLPF, countries present their Voluntary National Reviews and answer questions from Major Groups representatives and other countries. In 2017, 43 countries presented HLPF reports compared to 22 countries in 2016. Reporting guidelines are quite open-ended but increasingly formalized each year. The level of ambition and format of these reports therefore have varied greatly. The Voluntary National Reviews thus far revolve much around how countries organize work towards implementation of the SDGs, rather than actual fulfilment of indicators (UN DESA 2018; cf. Persson, Weitz, and Nilsson 2016). The HLPF is concluded by a three-day ministerial segment, ending with the adoption of a Ministerial Declaration negotiated by states' permanent representatives to the UN. Every fourth year, the HLPF convenes at the level of Heads of State and Government at the UN General Assembly.
The Swedish Government's 2017 Voluntary National Review (hereafter the HLPF report) built to a large extent on preparatory domestic reports. In early 2017, the government agency in charge of producing official statistics (Statistics Sweden), was tasked with writing a report preparing the ground for the HLPF report. As a result, Statistics Sweden provided the first outline of the interpretation of SDG indicators in a Swedish context (Statistics Sweden 2017a). 1 The main part of indicator-level adaptation to the Swedish context was carried out by a group of 10 statisticians formed for this purpose at Statistics Sweden. They were instructed to map the availability of data and suggest alternative SDG indicators suitable for the Swedish context. In addition, a report of the government's independent committee for the 2030 Agenda and memos from 86 Swedish government agencies and Swedish embassies across the world fed into the reporting process (Swedish Government 2017).
Country reporting on the sustainable development goals 407 The subsequent writing of the Swedish HLPF report led to some unusual procedures cutting across policy-making fields. At the time, responsibility for national SDG implementation resided with the Minister for Public Administration at the Ministry of Finance. This allocation of responsibility is due to the key role of public agencies and municipalities for domestic SDG attainment. Responsibility for Sweden's international SDG contributions resided with the Minister for International Development Cooperation and Climate at the Ministry for Foreign Affairs. The drafting of the report therefore saw an unusual attempt of joint ownership between the two ministries. The HLPF report was written by an experienced government official of the Ministry for Foreign Affairs supported by a steering group of representatives from the Ministry of Finance, the Ministry of Enterprise and Innovation and the Swedish Agenda 2030 Ambassador at the Ministry for Foreign Affairs. There was also a small reference group of CSOs (Concord Sweden, Social Forum, the Swedish Society for Nature Conservation and the Swedish Women's Lobby) with whom the HLPF-report author consulted. A broader group of stakeholders of about 120 people was invited twice by the Ministry for Foreign Affairs and there was a report-launch event of about 400 invited people. Nonstate actors were invited to provide best practice examples for Sweden and the 2030 Agenda-Collection of Examples, attached to the official HLPF report. In a unique process, the draft report was then handed over to the Ministry of Finance for ministry deliberations. The bulk of the final version of the report is devoted to a preliminary assessment of the 17 SDGs in Sweden (based on the Statistics Sweden report) and to an overview of on-going and future work in all sectors of society to implement the 2030 Agenda. As we will return to later, there is also a brief section that brings up examples of conflicts of objectives. That section mentions tensions between sustainable development and consumption as well as conflicts arising when Swedish companies do business in countries where environmental concerns or human rights are not respected. Finally, a large Swedish delegation was put together by the Ministry of Finance to represent Sweden at the HLPF in July 2017, and the Minister for Public Administration presented the Swedish report at the HLPF.

Policy Translation
We proceed by substantiating our theoretical discussion in empirical terms. Swedish HLPFreporting shows that one key function of reporting was to ignite rapid policy translation of global goals to the national context. We find that formative elements of policy translation were provided by procedural conditions of reporting and substantive pre-existing national policies. The procedural conditions were, in sum, the formal institutional processes related to report ownership and approval, the resources allocated to the reporting party (time, staff, and administrative support), demands on interaction with stakeholders, and not the least, the deadline provided and ensuing presentations of the report. Our interviews with employees of Statistics Sweden show that the procedural prerequisites for compiling the first Statistics Sweden report were tight. The short time frame provided for the task limited their work (interviews, December 2017). The group had about six weeks at its disposal and no extra funding was provided from Statistics Sweden centrally. Several interviewees conveyed they had to find time during coffee breaks, in between their ordinary tasks, to squeeze in work on the SDG indicators. Clearly, the HLPF deadline hastened the process. 'The reason for writing our report was that Ardalan Shekarabi [Minster for Public Administration] was to present at the HLPF in the summer. I don't think they realised how much work this would imply' (interview, December 2017).
One interviewee argued that Swedish SDG ambitions look great, but when government agencies are not provided with additional resources to perform the ground work for reporting, such ambitions appear symbolic (interview, December 2017). For their part, several Statistics Sweden interviewees felt they could not work through the mapping as thoroughly as needed. They tried to capture what was most important and saved more in-depth work for a second Statistics Sweden report due in the Autumn of 2017, after the HLPF had taken place. They agreed, however, that despite the short time frame they were satisfied with the end result. It was an unusual experience for Statistics Sweden employees to be provided first with indicators, then trying to decide what those indicators could shed light on in the Swedish case. Normally, statisticians would first identify a societal phenomenon in need of scrutiny, then they would consider indicators and data that might apply (interview, December 2017).
Once the report was to be written at the Ministry for Foreign Affairs, we discern tensions due to Swedish silo-based government organization. Tensions stemmed from the different mandates of the Ministry for Foreign Affairs and the Ministry of Finance who were supposed to cooperate on HLPF reporting. The former Ministry deals with broad global issues aiming to 'rock the boat', while the latter deals with public administration and safeguarding the national budget, making sure there is order in public governance and that 'the train does not gear off the tracks' (interview, December 2017). This was manifested in a tug of war around environmental and economic sustainability. One interviewee conveyed that '[d]uring the reporting process we had to deal with issues not sorted out before. Many late nights'. We also discerned frustration among representatives of the different governmentmandated bodies working in parallel on producing baseline reports on the 2030 Agenda (Statistics Sweden, the government's independent 2030 Agenda committee, and the two ministries). This was due to a perceived lack of coordination and mutual recognition of each other's reports: 'We were an underused resource. Hello, talk to us, we can help you! We had so much knowledge and they did not use us' (interview, December 2017).
Substantive conditions that shaped policy translation were the reporting mandate, preexisting policies, the format of the report, the expertise of the rapporteurs, the data on which the report was based and the interpretation of data. UN instructions on Voluntary National Reviews were quite open, leaving great freedom to the government to decide where to put the report's emphasis. Attempts to define how the globally agreed SDGs apply in the Swedish context involved choices and trade-offs shaped by pre-existing national policies and by professional norms of those compiling data. Statistics Sweden interviewees recall that a surprising amount of their work concerned trying to understand the intentions of Transforming Our World, asking themselves questions like: 'Why do the indicators bring up this particular aspect? Why are these vastly different indicators put under the same overarching goal?' (interviews, December 2017). According to one interviewee, the translation of SDG indicators to Swedish statistical thinking required pragmatism and a willingness to do things a bit differently than is usually the case (cf. Merry 2016, 7). The group mandated to consider indicators for the Swedish context needed to think about the quality of statistics in a different way as well. Its members collected material that statisticians do not normally work with such as reports, evaluations and research products, to get a sense of what was possible to do to with the SDG indicators. This resulted in an 'ad hoc feeling' (interview, December 2017).
Yet, several of the interviewees found certain SDGs easily applicable in a Swedish setting. This was for instance the case for indicators of SDG 3 on health and well-being where definitions existed and where Swedish statistics was available. In contrast, SDG 1 (on poverty) was considered more difficult. One interviewee explained that the measurement of poverty was a sensitive issue, very political and not up to statisticians to take a Country reporting on the sustainable development goals 409 stance on. Sweden does not have a national poverty measure of the multidimensional kind preferred by the UN. Poverty was portrayed as a sensitive issue also within the Statistics Sweden agency, where the term 'low income' was usually preferred to 'poverty', due to the political connotations of the latter term (Statistics Sweden 2017a). Overall, some of the statisticians mentioned that the SDG indicators do not provide a sufficient instrument in itself for review of sustainable development in Sweden, rather the SDGs provided an umbrella. They believed pre-existing Swedish goals systems such as the Swedish Environmental Objectives with measurable indicators would retain primacy (interviews, December 2017).
The extent to which the HLPF report should acknowledge goal conflicts was a matter of debate between CSO members of the reference group and the report author of the Ministry for Foreign Affairs (interviews, December 2017 and January 2018). A report by the government's independent committee for the 2030 Agenda (Agenda 2030 Delegationen 2017, e.g., 62, 95) pinpointed additional goal conflicts to the ones brought up in the HLPF report. Development-oriented CSOs (through Concord Sweden) raised two substantive issues: Sweden's international SDG work should not only revolve around development cooperation but engage with broader global affairs; and the HLPF-report should bring up challenges and goal conflicts faced by Sweden. CSO representatives of the reference group recall that there were changes in that direction during the reporting process. Yet, they would have wanted the Swedish government to go further in this regard and to discuss how to handle goal conflicts. In their opinion, report drafts paid too much attention to how successful Sweden was compared to most other countries. Representatives of two CSOs thought the Swedish HLPF report should have gone further in spelling out ambitions around international work on climate change and on sexual and reproductive health rights. This notwithstanding, being part of a reference group for the official HLPF report was considered quite unique, particularly in comparison with the situation for CSOs in many other countries (interviews, December 2017 and January 2018).
In sum, the reporting deadline spurred policy translation and provided its procedural and substantive prerequisites. The HLPF report conveys knowledge on sustainable development from the Swedish government's political perspective, invoking credibility from the prior Statistics Sweden report. Yet, the reporting process involved a tug-of-war on knowledge claims on two fronts, between two ministries with different policy mandates and between government and civil society agendas. This indicates that portraying the status of sustainable development is, despite references to indicators, politically sensitive even within a narrow circle of policy makers.

Accountability Preparation
The second function of reporting that we distil from our material is that of preparing for accountability. When reporting resides at the global-national nexus, accountability relations become more intricate than at the national or international level per se. Prior research shows that the use of quantitative indicators carries a promise of systematic accountability (Kelley and Simmons 2015, 57), while this promise is not necessarily fulfilled in practice. Our material shows that statistics rapidly took centre stage after the adoption of the 2030 Agenda as practical concerns on follow-up and review entered the political realm. Indeed, the main challenges highlighted by reporting countries at HLPF 2017 concerned a lack of disaggregated data and of data capacity and support (UN DESA 2018, 33). Two interviewees testify that after the adoption of the 2030 Agenda in 2015, there has been an explosion of interest in statistics globally. Before that, the attitude of politicians towards statistics was rather one of 'do not disturb us now, we are negotiating!' (interviews, December 2017). Tellingly, it was first hard for statisticians to make their voice heard at all, then suddenly everyone asked them 'where are the indicators?'. In the opinion of one interviewee, this lack of early involvement of statistical expertise makes some of the SDG targets impossible to follow-up in a statistical sense, though the targets are highly important in themselves.
Statisticians argued that statistics enable SDG review without political considerations, illustrated in the following quote: 'in the world of statistics, we measure things without considering what image this might provide of individual countries, we do not advocate certain policies, but are supposed to provide an account of how a country is doing' (interview, December 2017). High expectations on indicator-based review were held by government officials and civil society representatives alike. Representatives of Swedish CSOs among our interviewees agreed on the central role of reliable statistics for follow-up of the 2030 Agenda. They increasingly engage in quantitative review practices themselves. One telling example is the Swedish Federation for Lesbian, Gay, Bisexual, Transgender and Queer Rights (RFSL) that called attention to a gap in data on inclusion of LGBTQ people globally. The RFSL therefore together with UNDP and the World Bank, aimed to develop an LGBTQ inclusion index that will be part of CSO monitoring of states' SDG commitments (interview, January 2018).
Our material shows that, at least in the Swedish case, the HLPF in its present form is not an accountability forum where indicator-based reporting is scrutinized critically. Opinions on whether this was good or bad differed between our interviewees. Differences did not align with a simple division between government representatives and nongovernmental actors. Some government officials and CSOs saw the HLPF as a constructive learningforum that did not suffer from the usual UN conflicts. 'It is much easier to have a discussion about our issues at the HLPF and states do not react in a hostile way, which happens in the regular human rights mechanisms framework' (interview, January 2018). Others considered the lack of effective accountability features a weakness of HLPF. Illustratively, one interviewee said: [The HLPF] is not politically sensitive. It is not a problem to put a company representative on the Swedish chair in the HLPF plenary, because nothing politically sensitive will happen there. To me as a supporter of the 2030 Agenda, this is sad. It conveys that the HLPF is a nice forum where people meet to talk but it is not about accountability. You do not put a nonstate actor on the Swedish chair in a Security Council negotiation or in human rights compliance mechanisms (interview, December 2017).
In contrast, political contention on sustainable development characterised the parallel intergovernmental setting of negotiations on the Ministerial Declaration that concludes the HLPF each year. On the non-public backstage, the global politics of sustainable development agreements was played out in those negotiations, conducted by states' permanent representatives to the UN. Negotiations concerned the wording on politically contentious matters among states, such as foreign occupation, equality, sexual and reproductive health rights, and human rights. 2 Among our interviewees, both government officials and civil society representatives found those negotiations highly worrying, risking to weaken the political status of the 2030 Agenda. One interviewee felt that some states' attempts at renegotiating the 2030 Agenda 'exploded in our faces' (interview, December 2017). There was uncertainty on which version of the 2030 Agenda would be the valid one if the HLPF Ministerial Declaration of 2017 was weaker than Transforming Our World.
The Swedish HLPF report acknowledges both international and domestic follow-up systems, promising that 'Sweden will remain active in the international follow-up system Country reporting on the sustainable development goals 411 established for the 2030 Agenda and the global goals. Sweden will also continue to actively follow up, document and report on how the work to implement the Agenda is progressing' (Swedish Government 2017, 86). The report states that existing domestic follow-up systems will keep their relevance and that a forthcoming proposal from Statistics Sweden will form the basis for national SDG indicators. Thus far, reporting has not ignited political accountability in the Swedish political context. We note a lack of involvement of the Swedish Parliament in the reporting process of 2017 and in broader 2030 Agenda policy planning (Bexell and Jönsson 2016;Agenda 2030Delegationen 2018. Two interviewees emphasized that the organizational structure of Government Offices does not work well in relation to the cross-cutting nature of the 2030 Agenda (interviews, December 2017). The SDG umbrella lent itself well for reporting to an international audience of governments familiar with the global goals, but less well for Swedish domestic political structures.
In sum, reporting served as preparation for accountability by aligning with expectations on indicator-based performance review. Lacking parliamentary involvement and being largely unknown among the general public, the reporting process remained an elite project for an international audience, including of international CSOs. Reporting at best provided the groundwork for future domestic accountability. At the HLPF, contentious politics were played out on the backstage of negotiations on its Ministerial Declaration rather than in the HLPF plenary. Yet, those negotiations are equally important for the political status of the 2030 Agenda.

Identity Formation
The third and final function of reporting that our case study calls attention to is identity formation, invoking the theatre metaphor introduced above. Already at the adoption of the SDGs in 2015, the Swedish government (consisting of Social Democrat and Green Party ministers) declared its ambition to make Sweden an 'SDG champion state'. Notably, the Prime Minister created a High-Level Group of nine state leaders supposed to drive commitments to realize the SDGs-besides Sweden also Brazil, Colombia, Germany, Liberia, South Africa, Tanzania, Timor-Leste, and Tunis. The top ranking of Sweden in the widely cited index compiled by the Bertelsmann Stiftung (2016) that assesses countries' capacity to fulfil the goals, is often invoked by politicians to justify Sweden being a champion state. The Swedish reporting process reinforced this script through the composition of the delegation representing Sweden at the HLPF and through the report presentation at the HLPF. Below we discuss these two reporting elements in turn.
In our interpretation, the selection of participants to the official Swedish HLPF delegation had strong symbolic dimensions supporting identity formation. The government frequently referred to its inclusive approach to SDG work, involving cooperation with all sectors of society (Swedish Government 2017). The Ministry of Finance, in charge of selecting participants, demanded that all organizations invited should be represented at the highest level by its chair or secretary-general (interview, January 2018). Interviewees from all sectors commented on the selection of delegation members, portraying it as a last-minute decision involving much uncertainty on who was to be included. The HLPF delegation was led by the Minister for Public Administration and the Swedish Ambassador for the 2030 Agenda. It consisted in representatives from government agencies, municipalities, trade unions, CSOs, parliamentarians, private business and academia. Government officials pointed to the value of having a broad delegation. It would send a message to other countries that realization of the SDGs requires the involvement of all sectors of society (interviews, December 2017). Sweden received positive feedback on the broad composition of its HLPF-delegation. Moreover, delegation participants were encouraged to highlight the multi-sectorial nature of the delegation when talking to representatives of other countries (interviews, December 2017 andFebruary 2018).
Other interviewees, both among government officials and CSOs, were overall positive but also raised critical points. Some delegation members were not clear on what they were supposed to do during HLPF and did not fully understand expectations on their participation. The broad composition of the delegation was generally considered a strength but also involved tensions: Things can become a little messy when nonstate actors on the one hand are included in an official governmental delegation to represent Sweden in a UN SDG setting, on the other hand they are supposed to hold that same government accountable for its SDG work (interview, December 2017).
Outside of the official delegation, a couple of Swedish CSOs participated at the HLPF through the UN Major Groups system of nonstate actor representation. One civil society interviewee argued that '[w]e were all in New York because we came in our own capacity as organisations working at the UN level, but it sends a different message if organisations come in official capacity as part of a state delegation' (interview, January 2018). Certain CSOs outside of the official delegation were disappointed that several planned meetings with government officials were cancelled on short notice. In their view, the Ministry of Finance did not prioritize civil society contacts.
Moving on to report presentation, we regard the HLPF frontstage as one of identity formation for governments who presented reports. Reporting in intergovernmental fora might either reinforce or challenge the role identity communicated by governments. Governments were not challenged by other governments at the HLPF plenary. At side events, Sweden was asked to speak about its societal consultations on SDG implementation. Swedish government officials were also asked to participate in panels on Goal 17 (means of implementation) at both HLPF 2016 and HLPF 2017. This was taken as a confirmation of the Swedish 'champion' role (interview, December 2017). The 15-minutes presentation of the Swedish HLPF report was shared between the Minister for Public Administration, the chair-person of Malmö municipality, the chair of the Confederation of Swedish Enterprise and a youth CSO representative. A few other country presentations in plenary also included nongovernmental speakers (UN DESA 2018, 52). In terms of content, CSO representatives were pleased that the Swedish presentation raised Swedish societal challenges during the oral presentation, stating that '[t]he HLPF is not only supposed to be a PR-event' (interview, January 2018). This encouraged some other countries to mention their challenges as well. CSOs from other countries were impressed that Swedish CSOs dared to ask questions on Swedish weapons export to the Swedish government in plenary (interview, January 2018). The Minister for Public Administration answered in plenary by emphasising that civil society is extremely important for SDG implementation and that the government would continue its dialogue with civil society, including on gender issues. The Minister emphasized that sexual and reproductive health rights are an important part of Swedish feminist foreign policy but did not respond specifically to the question on Swedish weapons export. 3 We find the identity formation functions of SDG reporting to be strong on the international stage. On the domestic stage, in contrast, role identity is ascribed to the government rather than to the country as a whole. Domestic party politics invite more critique towards the government, even if the SDGs have not been a concern for broader national political debate. On the domestic political stage, critique stating that the government was too late in launching a Country reporting on the sustainable development goals 413 national action plan on the 2030 Agenda has been raised by CSOs and political opposition party representatives. For instance, CSOs argued that Sweden was not 'best in class' (Agenda 2030: "Sverige är inte bäst i klassen" 2017), prompting answer from responsible minsters who emphasised progress and Swedish leadership (Sverige ska vara ledande i arbetet med Agenda 2030 2017). The long-term nature of the 2030 Agenda led several interviewees to comment on the political vulnerability of Swedish SDG ambitions. Future Swedish governments might not want to enact the 'SDG champion state' role that the coalition government of the Social Democratic Party and Green Party (2014-2018) constructed.
In sum, identity-making occurred as the Swedish 'champion state' role was enacted on the international HLPF scene, reinforced by its audiences' approval. In contrast, on the domestic scene, the identity formation effects of reporting were weak due to pre-existing political critique according to which the government was slow in formulating its 2030 Agenda action plan. Studying reporting in terms of a performative act opens up for finding functions beyond those directly related to policy-making, but of equal importance for theorizing the politics of public sustainable development reporting.

Conclusion
Performance reviews are increasingly used throughout the policy-making process, underpinning knowledge claims and political priority-setting. This article has explored the functions of one performance review practice, that of public sustainable development reporting. Our case study on Swedish reporting to the HLPF shows that already during the early years of the SDGs, reporting practices influenced how countries approached the SDGs. The case confirms that public sustainable development reporting has multiple important functions and it points to challenges and tensions arising from reporting at the global-national nexus. We have identified three main functions that make public reporting on global goals a worthwhile study object, namely policy translation, accountability preparation and identity formation. While these functions can be found in other realms of self-evaluation as well, the Swedish HLPF-reporting process highlights their performative and political features in the sustainable development realm. Below we briefly discuss how the three functions played out at the global-national nexus, followed by suggestions for further research.
First, reporting kick-started policy translation driven by adaptation of global indicators to the national context. This global-national nexus provided challenges to the standard operating procedures of those involved. Statistics Sweden needed to compile data of less traditional kind in order to assess SDG indicator relevance and applicability in the Swedish context. Moreover, it was a backward order for Statistics Sweden to receive indicators from the global level first, then decide what Swedish societal concerns those could shed light on. The encompassing scope of SDG reporting also challenged issue-based policymaking structures of Swedish government offices. UN-reporting has traditionally resided with the Ministry of Foreign Affairs rather than the Ministry of Finance, and SDG policy translation brought looming tensions between ministry mandates forefront. A tension with particular bearing on knowledge production was how to portray goal conflicts. Thus, under the surface of policy translation lies the politics of performance review arising from contention on who is to decide how the broad 2030 Agenda applies in Sweden, beyond and within government.
Second, while accountability is the most obvious function of public sustainable development reporting in theory, our case shows that horizontal peer review among governments was more prominent than hierarchical accountability of domestic politics. While HLPF reporting processes are supposed to be inclusive at the national level, their ultimate audience remains other governments in a setting devoted to learning rather than to systematic scrutiny of indicators. At the same time, our interviews made clear that reporting served to at least prepare for future accountability by providing the government's view of problems and solutions in the sustainable development realm. This could be used for future assessment of how the government lives up to ambitions. Yet, in light of the long-term nature of the 2030 Agenda, a key concern raised by several interviewees was that future governments could not be held accountable on the basis of prior governments' SDG ambitions. This underscored the need of increasing the Swedish parliament's ownership of SDG policy translation, illuminating one of the many connections between functions of reporting.
A third reporting function was that of identity formation on the part of the Swedish government at the time. Understanding reporting as a performative act of identity construction allows for finding functions beyond those directly related to the policy-making process. With regard to identity formation, we found tensions between what played out on international and national stages. On the intergovernmental HLPF stage, the Swedish 'champion state' role was reinforced by its audiences' approval. On the domestic stage, in contrast, role identity was ascribed to the government rather than to the country of Sweden as a whole. Claims to role-model leadership by the government were questioned by political opposition parties and civil society representatives. Accordingly, at least in democratic societies, selfreporting conducted by governments becomes subject to broader political struggles around competing visions of the ends and means of sustainable development.
A research agenda for the future should be based on comparative studies of countries' reporting to the HLPF. Such comparisons can underpin more comprehensive theorizing of the politics of reporting on sustainable development than we have been able to provide in the present article. As reporting processes related to the 2030 Agenda grow in prominence, other functions will be observed as well and the relationship between the ones pinpointed here can be further explored. Research questions should be asked on whose voice matters for policymakers when review and follow-up schemes are constructed on the basis of SDG indicators. While we are still in the early days of the SDG era, research on the MDGs demonstrates that reporting measures might reinforce a gap between what is measurable through SDG indicators and the broader goals of Transforming Our World. Finally, future studies need to scrutinize the political features of review processes nationally and internationally. In particular, we believe that the role of national parliaments in SDG review should receive attention for the duration of the 2030 Agenda. Without doubt, goal conflicts related to the SDGs in national policy-making will nurture future research on the politics of reporting on sustainable development.

Disclosure statement
No potential conflict of interest was reported by the authors.