Differentiation as Affirmative Action: Transforming or Reinforcing Structural Inequality at the UNFCCC?

ABSTRACT Structural inequality is at the heart of the struggle to prevent dangerous climate change. This makes the global climate regime a particularly interesting case, when it comes to conceptualising and assessing the role of international institutions as sites for the reproduction and transformation of macro-level inequalities that structure the international system. This article uses these interlinkages to, first, assess, in how far the debates, conflicts and doubts regarding effectiveness and justifications of affirmative action at the domestic level, introduced as a reaction to domestic structural inequality, can teach us something about the actual potential of and the obstacles to the transformation of structural inequalities through differentiation internationally. Second, it assesses whether and how institutional mechanisms of categorisation and (re-)distribution within the UNFCCC have led and are likely to lead in the future to a reinforcement or a transformation of global structural inequalities.


Introduction
Structural inequality is at the heart of the struggle surrounding the global attempts to prevent dangerous climate change. It is the source of mistrust, of gridlock and even apathy to comply with commitments, even considering ever-escalating news on the state of the global climate. 1 It comes, for instance, into play when looking at the historical responsibility for human-induced climate change, but also in assessments about the uneven capacities to mitigate current greenhouse gas (GHG) emissions and to adapt to a future already impacted by global warming. 2 Indeed, the links between climate change and structural inequalities regarding both individuals and states are "locked in a vicious circle, whereby initial socioeconomic inequalities determine the disproportionate adverse effects [from climate change] which in turn results in greater inequality". 3 However, not only the impacts of climate change are intertwined with inequality, also both mitigation and adaptation policies "may reinforce or challenge such structures". 4 The exact implications of, for instance, any local mitigation actions for a 1.5°C pathway, are unclear; as, for instance, "ambitious mitigation scenarios suggest competition between land-based mitigation (e.g. biofuels and afforestation) and food crops in lower income regions". 5 Climate policies are also likely to create shifts in commodity prices and restructure labour markets-an issue that has risen to prominence during the Katowice COP in 2018 under the theme of "Just Transition". 6 At the same time, the expansion of renewable energy might re-distribute impact and power to previously less powerful actors and open up room for new agency within the oligopolistic energy market.
These close interlinkages with global stratification patterns make the global climate regime a particularly interesting case when it comes to conceptualising and assessing the role of international institutions as sites for the reproduction and transformation of macro-level inequalities in the international system. As Fehl and Freistein highlight in the introduction to this Special Issue, institutions entail intertwined processes of categorisation and distribution that may both reinvigorate but also disrupt structural inequalities through both material (such as voting, distribution of financial support) and symbolic (such as processes of inclusion and exclusion) practices. 7 Indeed, the global climate regime has since its inception relied on various differentiated categories of states (described in more detail below) to determine who has to carry what kind of burden but also who has access to, for instance finance and other forms of capacity-building. This kind of differentiation is deeply embedded into the global climate regime through the principle of "Common But Differentiated Responsibilities" (CBDR) and Respective Capabilities (RC) for instance in Article 4 of the United Nations Framework Convention on Climate Change (UNFCCC) of 1992. 8 CBDR requires across a range of issues within the climate regime (including mitigation, climate finance, capacity building and technology transfer) that, while all states need to contribute to the global effort, the actual burden to be shouldered differs. Therefore, CBDR is also one of the most contested issues within climate change governance and , p. 207. 8 CBDR is more frequently used to refer to the general principle of common but differentiated responsibilities as mentioned in the Rio Declaration in 1992. In the UNFCCC, CBDR was "expanded" by the "respective capabilities", but this formulation is not used consistently throughout the entire convention. The actual difference is surprisingly complicated, and, as not immediately relevant to the purposes of this article, I will generally use CBDR, unless the difference to CBDR-RC becomes important to the argument. For a detailed discussion, see Pieter Pauw, Clara Brandi, Carmen Richerzhagen, Steffen Bauer, and Hanna Schmole, "Different perspectives on differentiated responsibilities: a state of-the-art review of the notion of common but differentiated responsibilities in international negotiations", Discussion Paper, No. 6/2014 (DIE, Bonn, 2014), p. 9.
disagreements on the justification, interpretation and temporal scope of the norm is one of the key obstacles in negotiation progress over time. 9 Both the aspiration of privileging particular target groups as well as the conflicts associated with this show interesting analogies to affirmative action (AA) processes at the domestic level. 10 AA refers to policies and practices that aim at redressing "past racial or gender imbalance and injustices". Such measures often come in the form of preferential treatment of members of the target group, and hence also speaks to the categorisation and distribution mechanisms that are highlighted in this Special Issue overall. So while in general "affirmative action occurs whenever an organisation goes out of its way (i.e. takes action) to assure (i.e. to affirm) that its practices operate without disadvantaging either sex or any ethnic group", the crucial difference to equal opportunity policies, for example, is that it assumes that "we live in a world where decades or centuries of discrimination have created a slanted playing field". 11 The notion of a "slanted playing field" is a different way of describing structural inequalities. As a first goal in this article, we thus investigate in how far the debates, conflicts and doubts regarding effectiveness and justifications of AA at the domestic level can teach us something about the actual potential of and the obstacles to the transformation of structural inequalities through differentiation internationally. This will help us with the second research goal, which is to assess whether and how institutional mechanisms of categorisation and (re-)distribution within the UNFCCC have led and are likely to lead in the future to a reinforcement or a transformation of global structural inequalities. If we treat differentiation in analogy to AA, we thus argue that in order to affect structural inequality, the distinction between developing and developed countries needs to be integral to the UNFCCC and important internal distribution mechanisms should be structured around this division. In turn, this privileging would feedback into the better positioning a state within the international environment and may even help to restructure some of the fundamental underlying stratification patterns. We can identify the potential triggers of reinforcement or transformation of inequality with the help the heuristic model of this Special Issue, in particular with regard to its focus on three observation points for either transformation or reproduction at the intersection between institution and environment. 12 The first observation point exists at the "outside-in" intersection between the stratified environment and the international institution, i.e. how an IO responds to its environment, for instance by incorporating social categories, from the external environment. The second observation point exist at the intra-organisational level at which distribution mechanism for existent or newly created goods, benefits but also burdens, are created. At observation point three, the heuristic model suggests the existence of transformative or reproductive processes, such as learning between international institutions, or as particular intra-institutional distributive decisions have effects on the broader global status of a particular actor. The remainder of this article will proceed in three steps: First, we show how the approach to differentiation in the UNFCCC can be, more or less, be divided into two phases, the "Kyoto"-Phase with a top-down approach to differentiation and the "Parisphase" which describes the shift to a hybrid or bottom-up approach (more details below). Second, we show how differentiation in these two phases corresponds or not to domestic AA, its processes, justifications and problems and third whether or not the mechanisms of differentiation may have served to transform or reinforce global structural inequalities.

Differentiation as affirmative action from Rio to Paris
The UNFCCC concluded at the Earth Summit in Rio de Janeiro in 1992 was the first comprehensive legal agreement on addressing global climate change and it introduced mechanisms of classification and differentiation to alleviate the effects of structural inequality that were argued impact negatively on the goals set by the Convention. In the course of the following three decades, the approach to differentiation across different areas of climate change policy was continuously debated, re-interpreted and changed. 13 Nevertheless, two broad phases can be identified that also resonate with shifts in the AA debate at the domestic level.
The core norm in this regard is the notion of "Common But Differentiated Responsibilities" (CBDR), for instance in Article 3(1) and Article 4(1), that confers that all states have a shared obligation to address climate change but with unequal responsibilities. Beyond differentiated commitments to mitigate the emission of greenhouse gases (GHG), governments are, for instance, expected to contribute climate finance in line with their responsibility for the problem and their economic capacity to do so (and are entitled to receive climate finance in the case of vulnerable developing countries); but CBDR also evoked procedural mechanisms such as differentiated monitoring and reporting schemes that allow all parties to assess what their peer were doing and, if necessary, take corrective measures. Annexes to the Convention sorted states into three groups with to different commitments, using pre-existing categorizations from the international environment: Annex-I includes OECD (Organisation for Economic Co-operation and Development) members of 1992 in addition to countries with Economies in Transition (EIT). Annex-II includes only OECD-members of Annex-I. Only Annex-II parties are obligated to transfer funds and technology to developing countries to enable them to mitigate and adapt to climate change. Non-Annex-I parties are those states that are traditionally considered "developing countries". The Convention further recognises particularly vulnerable groups, such as Small Island Developing States (SIDS) and Least Developed Countries (LDCs). Overall, thus, many benefits and obligations within the regime were differentiated according to categories that the UNFCCC imported from its institutional environment.
The Kyoto Protocol (KP) negotiated in 1997 set a collective GHG reduction target of 5 percent below the emission level in the year 1990 during its first commitment period from 2008 to 2012. It differentiated among parties by establishing individual and legally binding emission reduction targets for Annex-I states only, the "Quantified Emissions Limitation and Reduction Objectives" (QELRO), while exempting developing countries (including emerging economies) from any reductions, even voluntary ones. It thus split the burden of keeping within the calculated remaining carbon budget unequally and by doing so, it "deviated from the conventional wisdom of environmental law-making [that] require taking on specific obligation by all parties upon ratification". 14 A number of other formal distributive decisions were made at Kyoto based on the categories of parties in the Annexes, for instance pertaining to procedural issues such as transparency in reporting but also with regard to climate finance. The Clean Development Mechanism, a marketbased mechanism, was set up to allow industrialised countries and businesses to finance mitigation projects in developing countries in order to achieve their emission reduction targets 15 Monitoring mechanisms and punitive consequences were established to ensure compliance. The governance mechanism of the KP thus can be described as a topdown differentiation that prescribed legally binding commitments to selected states.
The Bali Action Program (BAP) adopted at the 2007 Conference of Parties to the UNFCCC (COP) highlighted an ever-growing appeal of the argument about the need for (large) developing countries to take action in parallel to developed countries. For instance, the BAP introduced "nationally appropriate mitigation actions" (NAMAs) for developing country that were conditional upon the provision of finance, technology and capacity as a first step towards a softening the sharp divide of obligations between Annex-I and Non-Annex-I countries. These tendencies to move towards developing country mitigation obligations was picked up again in the compromise negotiated among the US and the BASIC countries (Brazil, South Africa, India, China) during the 2009 COP in Copenhagen which concluded that all parties were to self-select and list mitigation commitments and actions. This step introduced a new flexibility for all instead of top-down differentiation that privileged developing countries. 16 Negotiations resulted in a political agreementthe "Copenhagen Accord". Yet, many developing countries opposed these results due to the non-transparent and undemocratic process, leaving them "hostage to a geopolitical game between the major powers of the world". 17 Nevertheless, the COP agreed to "take note" of the Accord and, in the subsequent process, over 140 countries indicated their support and over 80 countries shared information on their national emission reduction targets. The following COPs in Cancun and Durban (2010 and 2011) managed to lead to a new negotiation process of a "legal instrument or an agreed The reliance on market-based mechanisms illustrations the strong linkages of the climate regime with the existing economic system and its inbuilt structural inequalities. Market-based mechanisms have been among the most contentious issues of the recent Summits in Katowice and Madrid and important decisions on the rules for a global carbon market system were once more postponed to COP 26 in Glasgow in 2020. See Carbon Brief, "COP25: Key Outcomes Agreed at the UN Climate Talks in Madrid", available at: <https://www.carbonbrief.org/cop25-key-outcomes-agreedat-the-un-climate-talks-in-madrid> (accessed 23 January 2020). outcome with legal force under the Convention applicable to all parties". 18 With regard to categorisations of states and their distributive consequences, the decisions in Cancun started a process in which the differentiation between developed and developing countries was no longer to be taken for granted. Instead, the decision text embraced the self-selection of mitigation targets and endorsed "differentiation for all". Bit by bit, the notion of individual obligation and mandatory emission reduction commitments gave way to an aspirational language of collective global targets. 19 Universal commitmentsand thus the dissolution of the established categorizations of Annex-I and Non-Annex-I stateswere on the table in a move towards more symmetrical commitments. At least in parts, developing countries agreed to this in exchange for a promise of a second commitment period of the KP ending in 2020. The Cancun Agreements "took a tentative step in this direction by requiring developing countries to aim at achieving a 'deviation in emissions relative to business-as-usual' in 2020" and Cancun took further the Copenhagen Accords in further moving towards "disturb[ing] the direct link between 'Monitoring, Reporting and Verification' (MRV) of financing and MRV of mitigation actions". 20 The final blow to differentiation as it was previously known, was dealt by the Durban Platform on Enhanced Action, negotiated a year later during the Durban COP in Durban and the negotiation of a new, universal agreement "under the Convention" were brought underway. This process was supposed to be concluded in 2015. In order to enhance climate finance that would enable developing countries to continue to take part in the process, the Durban negotiations further decided upon the establishment of the Green Climate Fund.
Subsequently, the key development at the 2013 COP in Warsaw was for parties to agree to the introduction of Intended Nationally Determined Contributions (INDCs) as the main mechanism according to which Parties communicate what kinds of actions they considered to be fair and appropriate. This introduced a strong form of bottom-up, self-differentiation. The INDCs also serve as centrepiece for the new Paris Agreement on Climate Change (PA) of 2015 where they have become the Nationally Determined Contributions (NDCs). The key features of the PA are set in Articles 2, 4 and 14. Article 2 describes the overall goal of the agreement as "holding the increase in the global average temperature to well below 2°C and to pursue efforts to limit the temperature increase to 1.5°C above pre-industrial levels". This makes parties collectively legally obligated to achieve this goal. Art. 4 includes the long-term goal of reaching "global peaking of greenhouse gas emissions as soon as possible". This goal needs to be achieved through the preparation, communication and maintenance of NDCs that reflect each parties' "highest levels of ambition". NDCs are neither legally binding and nor subject to any international review as such before they feed back into the socalled "global stocktake" to consider the collective progress in achieving the goal of the agreement (Art. 14) and to catalyse new climate action over time. 21  this, and also relating to the question of inequality, is how each successive NDC is supposed to be a progression beyond the current undertaking (Art. 4.3, PA). Article 14 requires for parties to ramp up the individual pledges (that are not legally binding), every five years in order to pursue the goals of Articles 2 and 4.A majority of parties submitted their NDCs ahead of the COP. Absent of any joint rules or drafting, they varied widely in terms of their actual content (ranging from pledges of absolute net reduction of emissions and reductions in emission intensity to the description of policy targets, such as shares of renewables or forest covers). Also NDCs vary widely in terms of the conditionality attached to them. 22 Whether an intended contribution is actualised thus may still depend on financial or other support to enable the action. Paris concluded the shift away from the dichotomy of between developing and developed countries' commitments and towards non-binding self-differentiation for all. This process of convergence of obligations was to be supported by, among other features, the introduction of an Enhanced Transparency Framework (ETF) applicable to all members that allowed for better monitoring and "shaming" should a party fail to comply with its commitments. Interestingly, in the ETF, the core procedural obligations related to mitigation (and transparency) are almost fully symmetrical for all parties (aside from SIDS and LDCs), which followed the interests of Annex I states. 23 Paris thus constituted the completion of major reform in the governance structures of the climate regime, from a rather strict top-down approach in the KP (stage 1) to a hybrid form governance combining (weak) elements of top-down (through the collective goal-setting) and strong, individualised and flexible bottom-up policies through the individual NDCs (stage 2). After the Katowice Summit of 2018, which had the mandate to collectively take stock of ambitions and to develop the process on when and how states are expected to submit new NDC, the 2019 COP in Madrid was supposed, first, step up ambition as well as climate finance, and to finalise outstanding rules in particular with regard to the global carbon market of Article 6 of the PA. After the general optimism after Paris, a more realistic view has set in, pointing to the fact that current commitments continue to be insufficient to reach the Agreement's goal of limiting global warming to well below 2°.
In sum, differentiation in the UNFCCC has shifted fundamentally over time, which is mainly an interest pursued by developed, Annex-I parties. 24 Flexibility and national determination have largely replaced differentiation and common responsibilities as core norms. As we will show in the next section, the underlying core conflicts about who is supposed to carry which burden align with domestic debates of AA and pave the way for a reinforcement rather than the desired transformation of structural inequalities through informal, often power-related mechanisms. Designing privileges is difficult and highly contested: what we can learn from affirmative action In the following section, I will show that the discussions about AA and preferential treatment of minority groups at the domestic level have important similarities to the debates about special responsibilities of developed countries towards developing countries at the global level and that, therefore, the notion of AA helps to better understand and potentially predict the challenges and conflicts arising around the legitimacy and durability of the differentiation mechanisms in the climate regime over time, especially by looking at the different underlying justifications for specific privileges of specific groupings. 25 The basis for AA is the existence of persistent differences in achievement between different groups which can be attributed to past and present discrimination, structural factors or other causes beyond the control of the disadvantaged group. 26 The literature discusses essentially three not necessarily compatible justifications for domestic AA across issue-areas and the related functions it is expected fulfil in an ideal setting. These justifications are typically compensatory, redistributive and utilitarian. 27 The first implies that a minority group deserves or needs preferential treatment to compensate of historic injustices and retribution of past discrimination. A central problem with this justification is that the setting of temporal limits to this form of positive discrimination is necessarily contested, as only when the "guilt" is compensated, equal treatment can follow. But how past guilt is quantifiable is difficult to assess. Also, there is the challenge that both those privileged and those burdened by AA are not equivalent to those having exposed to or having committed respectively the historic injustice. 28 A redistributive justification argues for a fairer distribution of resources and/or opportunities today. Here, a problem arises around the definition of "fair distribution" vs. the notion that opportunities should be based on merit alone. Further, critics argue that the respective recipients are often already those most privileged within the minority group. The third, utilitarian justification argues that AA is needed and justified if society benefits from it, for instance through an increase of social cohesion, broadened participation and lower social division. This seems to be the legally and publicly most accepted version of AA today, but again, the drawing of boundaries both regarding the duration of the privileged treatment as well as access for specific individuals are potentially problematic.
Independent of the specific justification, the targeting of separate groups creates several challenges. Carol Bacchi introduced the notion of category politics to "talk about the ways in which categories figure in affirmative action debates" and in particular how "concepts and categories used in political debates are constructed and deployed for political purposes". 29 The reference to specific categories, first, can contribute to essentialising particular identities independent of other changes going on around them. In other words, AA incentivises people (or other entities) to focus on one of their identities alone and thereby increases rather than decreases competition and the consciousness of difference and discrimination between groups. This might prevent effective solutions to policy problems that do not specifically play to these group identities and thus potentially perpetuates marginalisation. The disaggregation into different groups can also create new or enhance existing unequal power relations between and among groups and privilege those who define the scope and boundaries of the respective category of AA recipients. 30 Categories in this sense are a tool to define and legitimize particular policies. Indeed, categories and categorisation are important tools in political power play and the placement of a state or individual in a particular category is "not obvious or self-evident". 31 If we now turn to the UNFCCC and the KP in particular, AA (or differential treatment) was legitimised and practiced (mostly) through its compensatory and but also redistributive justifications. 32 The underlying interpretation of CBDR was that developing countries today suffered from the past overexploitation of the global environment by developed countries in multiple ways, resulting, for instance, in unequal economic capacity today. Historical responsibilities trumped present responsibility and, since industrialised countries have done most of the damage, it should be only fair that developing countries be given flexibility in the environmental control measures overall, and in the climate change regime in specific. While many international environmental agreements instead, however, state universal commitments for all, for example the Montreal Protocol, and then qualify the more substantive norms with regard to developing countries, the UNFCCC and in particular the Kyoto Protocol have chosen this compensatory path, which designated groups of states have not been required to reduce their GHG emissions. 33 In climate terms, this implies that "states [are given] … emission allowances for further developing their production and consumer potential", even if this potentially clashes with the overarching goal of combatting climate change. Macro-level inequality and multiple hierarchies (of vulnerability, historical responsibility and varying economic capacity) thus have been deeply enshrined into the climate change regime since its initiation. The fact that there is no temporal or other qualified limit speaks for a compensatory justification that has been acceptable to all in the early days of the climate regime. Yet, this also led to considerable "stickiness" of the categorisation of states preventing a clear understanding of how the Annexes were to be adapted to shifting geopolitical conditions and circumstances. 34  Clearly, there are also utilitarian aspects in the mix, as they are in any international agreement in which there are burdens to share and benefits to and hence become part of the negotiation agenda or bargaining chip. For the reasons described in the text, I find that the other two justifications prevail at this point. 33 Anita Halvorssen, Equality among Unequals in International Environmental Law. Differential Treatment for Developing Countries (Boulder, Westview Press, 1999), p. 76. 34 No shifts between Annex/Non-Annex I/Annex II states have been made aside from some marginal changes, and for external reasons only, such as the addition of Liechtenstein and Croatia to the Annex I/II list (other factors contributing to shifts were the dissolution of Czechoslovakia and Yugoslavia, for example). Turkey has over the years (and again at the most recent COP in Katowice) attempted to be excluded from Annex I and, instead, to be considered a Non-Annex I developing country; so far unsuccessfully. Andolu Agency, "Turkey Wants to Be Downgraded to Developing Country status in Climate Change Agreement List", Hurriyet Daily News, 29 November 2018, available: <http://www.hurriyetdailynews.com/turkeywants-to-be-downgraded-to-developing-country-status-in-climate-change-agreement-list-139311> (accessed 11 January, 2019).
Annex I and Non-Annex I states, despite shifting economic realities. 35 The differentiation between the Annexes with regard to mitigation obligations is of particular intricacy. The way in which mitigation burdens had been split has a very strong distributive and potentially zero-sum character: Limiting emissions through a global target creates a scarce resourcea "carbon budget"around which a distributional conflict emerged. 36 Talking about a carbon budget thus means that there are limitations on those who can use up this space and emit greenhouse gases unrestrictedly to continue or enhance their development, for example. 37 Here, we find clear analogies with the problems associated with compensatory (but also redistributive) AA at the domestic level: through the tie of privileges (here of not having to do something) to a particular identity (here: "developing country according to the definition of 1992)", even large and rising states with growing capacities (at least when compared to other states of this grouping) are forced to continue to present themselves with a strong identity as developing country, rejecting any binding action for themselves. These apparent discrepancies ultimately led to gridlock and, for instance, the US' rejection of ratifying the KP, just as many groupings disadvantages by AA domestically argue against discriminating measures. This explains how differentiation has become the "lightning rod for conflict because it underpins fundamental disagreements about historical and continuing responsibility and burden sharing for climate action". 39 This conflict spilled (and continues to spill) over onto almost any issue discussed within the global climate regime, often in analogy to the discussion on AA domestically and has a great impact on the extent to which the UNFCCC as an international institution transforms or reinforces structural inequality in the international system.
And indeed, ahead of the Paris COP of 2015, growing economic prowess of China but also India, Brazil and a few other "muddled" the clear dividing lines between the two camps of "Northern" developed and "Southern" developing countries, which, at least in the eyes of some, mainly Annex I states, further delegitimised redistributive and, above all, compensatory justifications of AA and they pushed for their phasing out of the political process. This is in analogy to many domestic situations in which proponents of AA, nowadays, are (often legally) required to use future-oriented justifications, for instance, that AA promotes diversity in educational institution as a societal value, rather than serving as a compensatory or redistributive measure. The equivalent development in the global climate regime comes in form of the proposition of a global goal to which all contribute and are enabled to contribute that has replaced top-down differentiated mitigation obligations as a more legitimate, moderate form of self-differentiation. Thus, the compensatory justification of AA has given way to a more redistributive and even more so utilitarian one. 40 The preamble of the PA, for instance, declares that climate change is an "urgent and potentially irreversible threat to human societies and the planet and thus requires the widest possible cooperation by all countries, and their participation in an effective and appropriate international response". The claim for sovereign decisionmaking on this important issue is thus strong and has likely contributed to a greater acceptance of its legitimacy and created hopes for greater compliance. This design of the PA (both structured and flexible at the same time) is undeniably one of the factors that led to its fast ratification by parties. For instance, the qualifier "in light of national circumstances'" that was added to CBDR-RC allows parties to participate broadly, according to their own assessments of responsibility and capacity, enhancing also equality in participation and decision-making. 41 This, again, compares well to the current state of debate at the domestic level, where, today, utilitarian, future-oriented justifications prevail. It is unclear, however, whether these measures are better suited to achieve their goals.

Transformation or reinforcement through CBDR
The introduction to this Special Issue explains how a focus on the "intertwined processes of categorisation and distribution" needs to be at the heart of the argument about how international organisations reproduce and/or transform global stratification patterns in their global environment. In the global climate regime, differentiation in the form of CBDR is the central "intertwining mechanisms". If we now accept, as has been previously argued that differentiation is a form of international AA, we should expect that differentiation has a transformative impact on inequality. While I do not attempt to engage with the underlying normative argument, one can also reasonably say that the majority of member states to the UNFCCC indeed demands such a transformative impact, both in terms of formal, material practices and rules that administer the distribution of goods and burdens, but that in symbolic practices through status ascriptions, stigmatisation and socialisation, for example. 42 Fehl and Freistein's heuristic model of three observations points (OP) at which the processes of categorisation and distribution can be analysed is used here to point at some of the central mechanisms in which differentiation has played out as AA. 43 40 It could be argued that the emergence of "loss and damage" in the form of the Warsaw International Mechanism contradicts this point of view. Yet, the decisions at the most recent COP in Madrid did not go beyond the encouragement of parties to "establish a loss and damage contact point", the request for the WIM's Executive Committee to "clarify how developing country parties may access funding from the Green Climate Fund" and the establishment of the Santiago Network to "catalyse the technical assistance" required by vulnerable countries. Yet, no decision for loss-and-damage finance was taken, which has been postponed to COP26. UNEP DTU Partnership, "COP25 -Loss and Damage Outcomes", available: <https://unepdtu.org/wp-content/uploads/2019/12/ld-cop25.pdf>, 17 December 2019, (last accessed 24 January 2020). Thus, while a need for discussion loss and damage has been widely recognized, and compensatory features theoretically are part of loss and damage, in the reality for the climate regime, no such compensation is foreseen so far. In the Kyoto-phase of differentiation, consisting of a strong interpretation of CBDR within the UNFCCC, its protocols and follow-on agreements, helped define and redefine the relevant social categories as well as the valued rights and benefits that are associated with these categories. This is where the UNFCCC might be distinct when compared to the other institutions studied in this Special Issue. Whereas in other organisations, positions of power are identified and valued, in the UNFCCC, the conventionally, historically, economically powerful are those with less entitlements and more commitments, at least within the first stage of differentiation were the compensatory justification, even if not necessarily intentionally, matches most closely the justification of this AA. 44 Differentiation considered through the lens of AA thus was installed as a tool to disrupt inequality patterns of the international environment that in turn were seen as a major obstacle for achieving the regime's overall goal of combatting dangerous climate change. The most obvious example is the KP with its distributive decisions built on the set categories of developing vs developed countries imported from the external environment (OP1). The KP's rigid interpretation of the CBDR had enormous distributive consequences within (OP 2) through intra-organisation processes (for instance the development of differentiated monitoring schemes). This also triggered shifts at OP3, at the inside-out level where internal distribution decision had an impact on the external environment by redefining the macro-structure as well as the status of individual actors at the macrolevel beyond climate politics. Scott describes this forcefully: The CBDR-RC principle […] dictated a radical solution that challenged […] the status quo. It would function to help rising powers at the expense of Annex-I States; indeed, if damage to the climate were linked to economic activity and the regime as a whole was to prevent dangerous change to the climate, the cost to Annex-I States would become exponentially higher as the economies of major developing countries continued to grow. 45 Along with the United States (see above), other Annex-I member states (including the EU, Australia and Canada) pushed for more legal commitments for emerging economies; some took it as far as opting out of agreeing upon new individual targets as part of the negotiation of a second commitment period of the Kyoto Protocol that was signed in 2012. 46 This is very clear example of the undermining of formal AA-like processes that were installed to provide a balance to structural inequalities, even if, in practice they did not necessarily function exactly to that point. 47 Dissatisfaction with the form of differentiation thus overall prevailed and, as described above soon enough this compensatory / redistribution form of differentiation gave way to a more future-oriented, utilitarian version of bottom-up self-differentiation to encourage for the greatest possible participation of each member state.
Self-differentiation particularly in the realm of mitigation has now become the norm. If we look at this method, we can find, first, at Observation Point 1 and dissolution of the 44 Achievements in combatting climate change and the provision of leadership are even seen as tools to enhance global status in the international system as described in this recent article: Sarang Shidore and Joshua Busby, "One More Try: The International Solar Alliance and India's Search for Geopolitical Influence", Energy Strategy Reviews, Vol. 26 (2019).

47
I do not intend to state here that, normatively speaking, a revision of the strict division among Annexes was and continuous to desirable or not. Thanks to an anonymous reviewer for highlighting the ambiguity in the argument. strong connection to external forms of categorisations. The self-differentiating method thus was expected facilitate greater ambition as "category politics" receded, and, some scholars argue, that self-declared commitments that are transparently communicated do much more than binding commitments in holding policy-makers accountable domestically and globally. 48 Yet, the absence of intended differentiation mechanisms within (i.e. at OP2) may have counterproductive effects at OP3 as aspects of these new institutional mechanisms are likely to reproduce inequalities in international society. First, the submitted NDCs are estimated to lead to emissions levels that are inconsistent with the PA goals: if ambitions are not increased for the period after 2030 a projected global average temperature of 2.6-3.1°C above pre-industrial levels is expected. This will affect most drastically the most vulnerable in international society. At this point and as said above, before the joint rules negotiated at the Katowice COP kick in, NDCs vary widely in terms of form and scope of mitigation commitments. 49 A key feature here is the conditionality attached to them; i.e. whether a contribution are actualised depends on financial or other support to enable the action. If thus insufficient climate finance or technological support is offered (which is likely given the continued struggle to raise sufficient climate finance), some of the NDCs will not be fulfilled. The sovereign determination of nationally distinct criteria of criteria of fairness, responsibility and capability then might serve as an incentive to backslide on commitments. There are further unintended consequences of the call for self-differentiation and/or parallelism between efforts of developed countries and (some) developing countries. In particular, given the strength of the American demand for parallelism between the mitigation commitments and actions taken by developed and (some) developing countries, and the absence of political conditions for strengthening the overall mitigation effort, symmetry has been achieved at the cost of ambition, and by levelling down the mitigation efforts required of developed countries. 50 A lack of top-down decision-making on limits to the carbon budget thus includes the willingness to take the risk that the poorest will have to suffer in particular, thereby reproducing substantial inequalities. 51 Procedural equality may also be affected, albeit more so within the UNFCCC rather than within the international system as a whole: The combined effect of the growing meaninglessness of a classification as "Non-Annex I" member and the lack of an effective negotiation coalition with the ongoing fragmentation of the G77, pushes the most marginalised states into a position even further on the sidelines. For instance, while the BASIC coalition is set to gain from a "everyone-for-its-own" approach, developing countries' interests overall would have a stronger clout if the BASIC states shared a collective motif with the rest of the G77. Thus, the bottom-up approach does not necessarily provide the necessary flexibility for all states to "voice their priorities" in a fair and just manner but instead is likely to "give way to a post-equity era with focus on 48 Falkner, op. cit. Clive, This Changes Nothing, op.cit., p. 929. This is not to say that, for instance, the KP with its top-down mechanisms could have produced more efficient results. History has clearly proven that (thanks to an anonymous reviewer for pointing this out).
voluntary self-determined national contributions regardless of historical responsibility". 52 The solution offered within the PA is one of enhanced transparency leading, in an ideal world to more social pressure, accountability and voluntary compliance in a virtuous cycle of escalating ambitions. Yet, as argued convincingly by Gupta and v. Asselt, practices and effects of UNFCCC transparency mirror first order conflicts over scope of accountability and burden-sharing, rather than transcending such conflict. 53 As an IR scholar I find none of this surprising. Indeed, developing countries feared that cross-pressures from the convergence of obligations in reporting and transparency would eventually lead to the "creation of a top-down regime for the establishment of subsequent NDCs or of creating de facto limitations on the extent to which Parties, particularly developing countries, may exercise national determination in shaping and communicating their NDCs". 54 Rather than transforming the structure of the current social and economic system, most of the PA thus appears to reinforce it, also beyond the more narrow aspects of differentiation. 55 There is a continued commitment to industrial growth and a reliance on future technology to do the "hard work". Its Article 2 requests, for example, that the 2°C goal should be "achieved in the context of sustainable development". Article 10 further points to the need to accelerate, encourage, and enable economic growth. In addition, there are no mentions of any particular GHG sources and "not a single comment on fossil fuel use, nothing about how to stop the expansion of fracking, shale oil or explorations for oil and gas in the Arctic and Antarctic". 56 The lacking consideration of economic and energy policies within the PA is likely to encourage governments to sign up to the agreement, while continuing their investments into fossil fuels production and consumption. 57 Rather than engaging in preventing dangerous climate change, the principle of risk management and a reliance on technology has taken ground, which means that the international community counts on the hope that some technological solution will be found, including carbon capture and storage but also more futuristic options that will cushion the need for changing production and consumption behaviours too much. Yet, none of these technologies so far have had any proven impact on the global CO2 balance or have any reliable prediction to do so in the near to mid-term future. But more important to the argument here, the reliance on potentially very costly technologies further reinforces the advantages of industrialised nations as they are generally those who provide the new technologies and those who will benefit from their eventual widespread usage, which, in the end, again leaves us to see mechanisms of distribution and categorisation at observation point 1. Pre-existing inequalities underneath the systems of production of fossil fuels and conventional models of development are indeed so deeply entrenched into the UNFCCC and its negotiations and thus continue to shape many aspects of the global climate regime. The importance of the fossil fuel industry within the climate regime has been reemphasised by the presence of many coal lobbyists and sponsors at the most recent COP in Katowice.

Conclusion
The goal of this article was to assess whether institutional mechanisms of categorisation and (re-)distribution within the global climate regime as represented by the UNFCCC have led and are likely to lead in the future to a reinforcement or a transformation of global structural inequalities. The concept of AA has further been used to showcase how differentiation in the global climate regime, similarly to the evolution of the discourse on AA domestically, compensatory/redistributive justifications of this form of positive discrimination have becomes less and less accepted and eventually made way for a more utilitarian one. 58 There are good reasons for this, for instance real world developments that led to very different development paths among the so-called Non-Annex I states. Yet, again, similarly to domestic AA, this shift was unable to prevent further reinforcements of structural inequality and increasing vulnerability of already marginalised actors. Also, it may, given most recent climate predictions simply be too late to create to level playing field, which is why calls for a comprehensive 'loss and damages' have increasingly been voiced. While COP25 has made some limited progress in this area, loss and damage continues to be one of the most contested issue of the climate regime and the outlook for a comprehensive, and truly compensatory regulation is not very promising.
The heuristic model that guides contributions in this Special Issues helped to point out some of the mechanisms and interfaces where differentiation has actually rather counteracted its purpose and, in addition, actually blocked any creation of a "level playing field", even if this was not necessarily the primary goal of differentiation. For instance, as I have described the form of static differentiation offered by the Kyoto Protocol served essentially led to the demise of the top down approach, as it functioned "to help rising powers at the expense of Annex-I states" even beyond the UNFCCC within the wider global economic system.
Yet, a few further caveats need to be considered when we assess the institutional impact of shifting norms of differentiation on structural inequality, also in light of the comparative angle of this Special Issue. First, the fragmentation of climate change governance in specific with its unusually dense institutional setting, the diversity of relevant actors and the intense issue-linkages between climate change and almost any other field of global governance challenges (health, food, migration, natural disasters, pandemics etc.) gives rise to layered hierarchies and variable structures of power and problems. 59 For research on inequality and its transformation or reproduction, this means that it is difficult to 58 There are, however, also important limits to this comparison. Above all, the assumption of developing countries as deserving recipient (for whatever reasons) implies a certain "black-ball" view of the state, while Chancel and Piketty, for example, have estimated that within country inequality explains 50% of global GHG emission inequalityLucas Chancel and Thomas Piketty, Carbon and Inequality: From Kyoto to Paris (Paris School of Economics, Paris, 2015), p. 33-34. assign winners and losers within its specific hierarchies, but marginalisation and inequality are also likely to reinforce themselves through cross-over disadvantages. 60 Roberts et al. for instance speak about the "triple inequality" in the climate regime; first, inequality about who caused climate change, second, about who is able to share the burden and, third, about who is going to be most affected by the consequences of dangerous climate change. 61 Frequently, economic and other geopolitical interests thus at least used to be treated as much more important as any climate goals; and bargaining power games within the UNFCCC obstructed meaningful negotiations of a comprehensive climate agreement, further aggravating the "triple inequality", as "poverty and powerlessness […] directly deprive[d] developing countries of the technical, financial, and administrative capacity to effectively negotiate". 62 Within our structural perspective on inequality, it is further important to note that, the same climate policy, the same climate-related burden or opportunity will affect actors differently because of pre-existing inequalities. This context is also where this article's discussion of differentiation as international AA was set.
Indeed, the past three decades of climate change negotiation have led to ongoing circles of transforming and reproducing structural inequalities. Institutional attempts to alleviate and balance out structural inequalities by different forms of AA were almost without fail followed by informal processes that undermined these attempts. Whether the newest ramification in this cycle, the hybrid governance mechanism established in the PA, will overcome cyclical gridlock and ineffectiveness, is at least questionable and the continued lack of ambition by some of the major players leaves little room for optimism. The PA has added to the already complex and multidimensional hierarchies/inequalities an additional layer of a "primacy of domestic politics". 63 Despite growing societal engagement with the problem of climate change, 64 the absence of strong political agents pushing for change has not allowed the PA to provide a clear path to a transformation of global structural inequalities. Keeping within the 1.5°C pathway that most scientist see as the only way to avoid the most dangerous effects of climate change will ultimately require a more radical shift in global patterns of production and consumption that goes beyond the impact of an international institution. 65 And while there has been much optimism about the Paris mechanisms to serve as a blueprint for other "shared-fate" issue areas, such as air pollution, financial crises, migration and global pandemics under the condition of a sustainable pressure for global management, the insights offered by AA into the underlying mechanisms of progress vs backlash and the considerable and growing opposition against AA measures in many states, makes us less optimistic about this. 66 This interplay of formal and informal modes of inequality reproduction is likely to be an important feature of other organisations' stratification patterns and should be considered in depth in future research on the subject of this Special Issue.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributor
Dr. Miriam Prys-Hansen is Lead Research Fellow at the GIGA German Institute of Global and Area Studies and Academic Director of GIGA's Doctoral Programme. She received her D.Phil in International Relations from the University of Oxford. Her research covers issues of global and regional governance in Asia, in particular in the issue-area of climate change, and the impact of changing power constellations in the international system.