7.1 Introduction
Without an active secretariat, decisions under international treaties would often be ill-prepared, and an informed negotiation process would be much more difficult to achieve. Formally, secretariats are supposed to be neutral technocrats and not meant to influence democratic decision-making processes. In reality, however, things are usually different. In fact, it is almost impossible to provide “impartial information,” since even the volume of the information provided and the way it is prepared and introduced into the debate generally have some political impact. This relates to what Barnett and Finnemore (Reference Barnett and Finnemore1999: 708) have identified as the “irony of depoliticized appearance.” An active secretariat has to act behind the scenes, “indeed in the corridors and hotel bars of conference venues” (Bauer Reference Bauer2006: 34), and this hidden and informal action may be a key determinant of any progress to be achieved.
At the same time, the secretariat’s influence necessarily constrains the role of elected decision-making bodies. Influence thereby relates to both the design of policy outputs and the control of decision-making processes. Thus, secretariats need to strike a “delicate balance between the activism that is needed to make a difference and the risk of being perceived as questioning or even challenging specific interests of individual parties to the treaty,” that is, objectionable political interference (Andresen and Skjærseth Reference Andresen and Skjærseth1999: 7; Bauer Reference Bauer2006: 34). From a normative perspective, the role the secretariats should assume in this context depends on a number of context variables. These include the complexity of the problem that calls for the knowledge of specialized experts and the diversity of political preferences that call for a clear predominance of the democratic decision-making bodies and a less active role of the secretariat (see, e.g., Alesina and Tabellini Reference Alesina and Tabellini2007, Reference Alesina and Tabellini2008; Hawkins et al. Reference Hawkins, Lake, Nielson, Tierney, Hawkins, Lake, Nielson and Tierney2006b). A number of studies exist that compare the influence of different secretariats along these lines. Biermann and Siebenhüner (Reference Biermann and Siebenhüner2009) have provided a comprehensive discussion of different international environmental agreements.
This academic literature reflects a recent trend in the international relations literature to shift the focus to international bureaucracies as relevant independent actors, and not just acting on behalf of their member states (Barnett and Finnemore Reference Barnett and Finnemore1999, Reference Barnett and Finnemore2004; Hawkins et al. Reference Hawkins, Lake, Nielson and Tierney2006a; Johnson Reference Johnson2013a, Reference Johnsonb; Johnson and Urpelainen Reference Johnson and Urpelainen2014; Michaelowa, Reinsberg, and Schneider Reference Michaelowa, Reinsberg and Schneider2018), and a simultaneous trend within the economic theory of bureaucracy to consider a more realistic objective function for civil servants that significantly departs from the simple resource maximization perspective introduced by Niskanen (Reference Niskanen1971) (see, e.g., Alesina and Tabellini Reference Alesina and Tabellini2007: 173; Dewatripont, Jewitt, and Tirole Reference Dewatripont, Jewitt and Tirole1999a, Reference Dewatripont, Jewitt and Tiroleb).
However, most of these scholars seem to concentrate on the process of initial delegation: How much autonomy should states delegate to an international bureaucracy for a specific task, what safeguards should they impose to limit agency slack (both shirking and slippage), and how does the empirical difference observed between the responsibilities delegated to different international organizations reflect the theoretical (functionalistic) expectations about the extent of delegation?
In contrast, our analysis focuses on the behavior of international bureaucracies once they are established. Such studies are typically carried out from a sociological or anthropological perspective, which has now also found its way into political science (see Barnett and Finnemore Reference Barnett and Finnemore2004 for a general discussion; for an in-depth study of individual organizations see, e.g., Weaver Reference Weaver2008 for the World Bank). In contrast, Hawkins and Jacoby (Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006) combine the detailed observation of international bureaucracies’ activities with the principal–agent approach generally used within the economic theory of bureaucracy. By doing so, they adopt the assumption of international civil servants rationally following their own, independent, objectives. Hawkins and Jacoby illustrate their theoretical approach with the example of how the two main institutions of the European Convention of Human Rights (ECHR) changed the way of accepting cases and their court decisions in response to enlarged country membership and thus went clearly beyond their original mandate. The timing of events allows the authors to argue convincingly that it was the purposefully designed strategy of the international bureaucracy, rather than the state-led initial institutional design, that resulted in the considerable autonomy of these institutions.
Yet further empirical illustrations of the suitability of the principal–agent approach for the explanation of the behavior of international bureaucracies are rare. In the context of a case study on the International Monetary Fund, Gould (Reference Gould, Hawkins, Lake, Nielson and Tierney2006: 306 ff.) argues that the principal–agent approach is useful to predict the behavior of the principals but much less so to predict the behavior of the secretariat. She concludes that exclusively relying on principal–agent theory may well explain why an international bureaucracy is endowed with a certain level of autonomy but not in which way it will actually make use of this autonomy. Regarding environmental bureaucracies, Manulak (Reference Manulak2017) uses a principal–agent framework to study the entire process from initial delegation to subsequent attempts by states to informally control the secretariat of the United Nations Environment Programme. Other work on environmental bureaucracies sometimes reflects upon dynamics within bureaucracies but does not create a systematic and comprehensive link to principal–agent theory.
In this chapter, we attempt to overcome these problems by combining the rational choice approach of the principal–agent framework with some of the more constructivist ideas of principal–agent interactions that may eventually lead to revisions of the principals’ initial objectives. This goes beyond conventional principal–agent theory, which assumes a static set of (mutually conflicting) preferences for the agent and the principal. This new framework allows us to capture the interesting process of bureaucracies reinterpreting and redefining their rules – a process highlighted both in Barnett and Finnemore (Reference Barnett and Finnemore2004) and in Hawkins and Jacoby (Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006).
We believe that over and above a more realistic and detailed definition of the international bureaucracies’ objectives (as compared to pure budget maximization), it is these dynamics in the interaction between the agent and the principal that should allow us to derive more precise predictions about concrete activities of international civil servants within the general rational choice approach of the principal–agent model. Moreover, rather than studying bureaucratic behavior in general, we focus on the analysis of bureaucratic strategies triggered by resource growth – a rather typical situation for many international organizations.
Empirically, we illustrate our arguments with the example of the secretariat of the United Nations Framework Convention on Climate Change (UNFCCC) (henceforth simply called “the Secretariat”) and notably its relatively technical branch responsible for international market mechanisms, especially the Clean Development Mechanism (CDM). In such an area that is politically much less contested than, for example, emission reduction commitments, we expect the greatest chance to observe the development of the independent role of an international bureaucracy. In addition, this particular case study enables us to combine both econometric analysis based on quantitative data on resource growth, the range of delegated activities, and actual policy decisions (see Michaelowa and Michaelowa Reference Michaelowa and Michaelowa2017) and more in-depth qualitative analysis based on document analysis and interviews, on which we focus here. Our data are unique in that they allow us to measure a resource increase exogenous to deliberate decisions by the principal. This is crucial for our empirical identification strategy.
In addition, the initial role of the Secretariat has been relatively well researched, providing us with a sound basis for our analysis. In particular, Depledge (Reference Depledge2005, Reference Depledge2007) and Yamin and Depledge (Reference Yamin and Depledge2004) provide a detailed account and discussion of the Secretariat’s tasks and activities. Moreover, Busch (Reference Busch, Biermann and Siebenhüner2009) analyzes the UNFCCC Secretariat within the comparative theoretical framework for different environmental treaty secretariats provided by Biermann and Siebenhüner (Reference Biermann and Siebenhüner2009).
In this chapter, which complements and expands upon the more quantitative analysis undertaken in Michaelowa and Michaelowa (Reference Michaelowa and Michaelowa2017), we analyze whether the Secretariat has become more powerful over time, whether and how it started to directly influence policy processes, and how these developments are linked to the growth in financial resources. To do so, we first provide a more detailed and general theoretical framework in Section 7.2. Next, in Section 7.3 we provide an introduction to the specific UNFCCC case study and notably a description of the unexpected resource flow to a certain area of the Secretariat due to the CDM. In Section 7.4 we analyze two cases where the Secretariat increased its influence on rule-setting. In Section 7.5 we discuss how the drying up of the CDM market has led to an even stronger tendency for “top down” rule-setting as well as “Parkinson’s law”-style responses by the Secretariat, while in Section 7.6 we provide our conclusion.
7.2 Principal–Agent Interactions When International Secretariats Grow: A Theoretical Framework
As pointed out by Alesina and Tabellini (Reference Alesina and Tabellini2007: 173), there is no established standard model of bureaucratic behavior so far. In this context, a central problem appears to be the appropriate specification of the arguments in the bureaucratic objective function. Despite considerable discussion leading to a variety of different models in the rational-choice literature, consensus appears to emerge on at least some issues.
First, it is generally agreed that the sole consideration of the general budget (Niskanen Reference Niskanen1971) or the discretionary budget (resources minus cost) (Migué and Bélanger Reference Migué and Bélanger1974; Niskanen Reference Niskanen1975), that is, the consideration of resources as the only argument bureaucracies really care about, does not provide us with a sufficiently realistic theoretical framework to predict the specific development of diverse international bureaucracies. Barnett and Finnemore (Reference Barnett and Finnemore1999: 706) mention the opposition of North Atlantic Treaty Organization staff against political expansion plans in the late 1990s as one example that calls for other, supplementary, arguments.
Second, while budgetary concerns should not be the only ones considered, this does not mean that they are unimportant. Despite the obvious oversimplification, the focus on resource expansion objectives within the rational-choice framework has been able to provide useful explanations to a variety of phenomena observed within international organizations (see, e.g., Vaubel Reference Vaubel, Vaubel and Willet1991, Reference Vaubel2006). Even Barnett and Finnemore (Reference Barnett and Finnemore1999: 706) acknowledge that “there is good reason to assume that organizations care about their resource base and turf.” They insist, however, that resources should not be the only motivation taken into account as bureaucrats will usually weight them against other objectives. If other objectives are seriously taken into account, their striving for resources may also lose some of its negative connotation. In fact, at least to some extent, resources may then simply represent a means for the bureaucracy to reach other, socially more highly valued, ends.
Third, the existing literature has already brought up good candidates to complement the list of variables in the bureaucratic objective function. According to many authors (e.g., Alesina and Tabellini Reference Alesina and Tabellini2007, Reference Alesina and Tabellini2008; Barnett and Finnemore Reference Barnett and Finnemore1999, Reference Barnett and Finnemore2004) bureaucrats want to show technical excellence in their field of expertise. This may be related to the goal of acceptance and prestige within their professional community (Alesina and Tabellini Reference Alesina and Tabellini2007), to related career concerns (Dewatripont, Jewitt, and Tirole Reference Dewatripont, Jewitt and Tirole1999a, b; Holmstrom Reference Holmstrom1982), or simply to their normative commitment to the services (i.e., eventually, the public good) delivered by their organization (see, e.g., Hawkins and Jacoby Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006: 223). Indeed, since bureaucrats are not hired at random, but from a community of people who self-selected into this specific field of activity in the first place, they should also be expected to be more dedicated to this field than the average citizen (Häfliger and Hug Reference Häfliger and Hug2012). In fact, many of the functionalist explanations for the very existence of international organizations focus on the states’ willingness to delegate tasks to an organization with particularly strong substance-related preferences, so that it would help them to overcome problems of credible commitment, notably under potentially time-inconsistent preferences (Hawkins et al. Reference Hawkins, Lake, Nielson, Tierney, Hawkins, Lake, Nielson and Tierney2006b: 18–19).
Another typical objective bureaucracies are expected to value is autonomy or power (Barnett and Finnemore Reference Barnett and Finnemore2004; Hawkins et al. Reference Hawkins, Lake, Nielson, Tierney, Hawkins, Lake, Nielson and Tierney2006b; Vaubel Reference Vaubel, Vaubel and Willet1991). We follow Hawkins et al. (Reference Hawkins, Lake, Nielson, Tierney, Hawkins, Lake, Nielson and Tierney2006b: 8) in defining autonomy more broadly than discretion, a term that is used to refer to autonomy within the restricted area of explicitly delegated activities. Autonomy implies some freedom for independent action and thereby, eventually, some influence on actual policy outcomes.
To some extent, it is a logical implication of the principal–agent model itself that bureaucrats should value autonomy. The reason is that autonomy allows them to follow their own objectives as opposed to those of the principals. From this perspective, however, autonomy is purely instrumental, while it has been frequently considered as an objective in its own right (see, e.g., Barnett and Finnemore Reference Barnett and Finnemore2004; Hawkins and Jacoby Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006). Indeed while one might conceive autonomy as instrumental to, for example, the objective to show the bureaucracy’s technical competence or, alternatively, to the objective to enlarge the bureaucracy’s resource base, one might also conceive these two other objectives as instrumental to autonomy. As Barnett and Finnemore (Reference Barnett and Finnemore2004: 21) put it:
Bureaucracy is powerful and commands deference, not in its own right, but because of the values it claims to embody and the people it claims to serve. IOs [International organizations] cannot simply say, “we are bureaucracies, do what we say.” To be authoritative, ergo powerful, they must be seen to serve some valued and legitimate social purpose, and, further, they must be seen to serve that purpose in an impartial and technocratic way using their impersonal rules.
All in all, we have thus identified three widely agreed objectives that may all be final goals but also means to achieve some of the other goals. Bureaucratic utility (UB) could thus be specified as a function of these three terms: UB = f (autonomy, excellence, resources), whereby the exact functional form, that is, the weighting of the different arguments and the way they interact, remains to be specified depending on the more specific context within which the analysis takes place. For instance, in a politically contested and nationally salient context, striving for autonomy should be less relevant because further delegation may a priori be unrealistic. The opposite should be true in a context in which policy decisions depend primarily on technical assessments. Moreover, the dynamics between the different objectives can be expected to depend on context too. If, as in the empirical example discussed later, we observe an exogenous (and substantial) increase in financial resources, the expected dynamics would be that these resources will be used to promote the other two objectives (and then, only in a second step, further resource growth).
While the list of objectives considered in the bureaucratic utility function may omit some variables that could reasonably be added (again depending on context), it should generally provide a sound basis for a reasonable prediction of bureaucratic behavior within international organizations. The above specification also shows that the rational choice framework with an explicitly defined utility function does not preclude altruistic behavior, or bureaucratic behavior that effectively serves the internationally agreed goals of the organization (be it based on altruistic motivations or not). Thus, while we need to make some assumptions about the bureaucratic objective function, we do not necessarily need to make normative judgments in order to predict the developments of international bureaucracies within the framework of a principal–agent model.
By setting up a more specific bureaucratic objective function we respond to one problem identified in the literature with respect to the application of the principal–agent approach to the identification of bureaucratic behavior (Barnett and Finnemore Reference Barnett and Finnemore1999; Gould Reference Gould, Hawkins, Lake, Nielson and Tierney2006). However, this leads to yet another problem: Integrating autonomy in the objective function (and all the related dynamics suggested earlier) is at odds with the static formulation of the traditional principal–agent framework. In such a static framework, striving for autonomy cannot lead anywhere because the rules are defined once and for all. As Hawkins and Jacoby (Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006) convincingly argue, de facto, many international bureaucracies do succeed in obtaining greater autonomy over time. They do so both by reinterpreting existing rules (and gradually changing accepted practices) and by convincing the principals that rules might have to be changed (i.e., by convincing them to formally delegate more autonomy). Barnett and Finnemore (Reference Barnett and Finnemore2004) also make a strong point that the rules themselves may be endogenous to bureaucratic activity.
While principal–agent models have been adjusted to include multiple principals and several hierarchical levels of principals and agents (e.g., citizens delegating a task to their respective governments who in turn delegate it to an international bureaucracy), they typically ignore that the principals’ interests and priorities may change over time (Stone Reference Stone2011: 26). The adjustment of the principals’ beliefs about how much authority they should delegate would typically rather be discussed in a constructivist framework. Yet the two can be fruitfully married here, since the adjustment of beliefs may well be a very rational choice by principals, notably in the context of imperfect information that the principal–agent model supposes anyway, and thus in line with the general assumptions of this model.
We believe that it is important to highlight the break with the static version of the principal agent–model because these dynamics are essential to explain bureaucratic behavior. While Hawkins and Jacoby (Reference Hawkins, Jacoby, Hawkins, Lake, Nielson and Tierney2006) do not explicitly mark this theoretical break, the actual importance of these dynamics in their work is omnipresent, notably in their empirical analysis.
Delegation of autonomy is not a once and for all decision but is subject to constant adjustments either through the reinterpretation or through the formal revision of existing rules. Principals decide (and redecide) based on an ongoing optimization between the reduction of their own workload and improved outcomes due to the use of bureaucratic expertise on the one hand and the cost of engaging the bureaucracy on the other hand. The consideration of cost includes not only the direct financial cost of maintaining an international civil service but also the political cost of potentially undesirable bureaucratic decisions. Principals are induced to accept or even actively promote greater bureaucratic autonomy (i) if they receive relevant information (related, for instance, to a large international crisis or a change in public awareness), (ii) if new external resources (e.g., from the private sector) become available to cover some of the cost, or (iii) if change is obfuscated, for example, if decisions about relevant procedural rules are hidden in the midst of complicated technicalities, or simply, if change is gradually creeping in.
This third channel is driven by the bureaucracy itself, notably by showing its excellent technical capabilities or by generating trust by giving itself a very technocratic and apolitical appearance. Both external factors (i) and (ii) can complement and facilitate the bureaucracy’s strategy in this respect. For instance, the bureaucracy can make use of new information by interpreting it in a way that makes its expertise more desirable, or it can use new resources to enhance its capacity and the actual and/or perceived quality of its services. For instance, by hiring competent staff, the principal may be more easily convinced to leave relevant tasks in the responsibility of the secretariat. In addition, if staff are sufficiently large and experienced, it may be more easily able to convince the principal of the predominance of technical and process-related issues even regarding decisions that are, de facto, less technical than political. Finally, it may take the time to carefully draft propositions in a way to increase its own procedural rights without anyone noticing, or it may simply overwhelm the principals with so many issues to decide upon that they cannot help but delegate some of these decisions (or their preparation) back to the secretariat. Generally speaking, with increasing resources, a bureaucracy disposes of greater means to increase pressure for more autonomy and greater means to show competence. This should enable the bureaucracy to extend the compelling offer to reduce the workload of principals, thereby enhancing its own freedom of action and its impact on actual decision-making.
Apart from the dynamics we introduce into the model, there is one more way in which we wish to deviate from the traditional principal–agent framework. The principal–agent model usually adopts the normative perspective of the principals. In our context, this does not necessarily make sense because the principal–agent relationship we observe is only a subset of a wider hierarchical principal–agent framework, and the interests of national delegates at international organizations may themselves largely deviate from the interests of the ultimate principal, namely the population in the member countries. Thus, we could observe situations in which the national delegates willingly delegate more authority to the secretariat to reduce their own workload, while the general public would have preferred these issues to be decided at a political level. In other situations, political positions of national delegates may be driven by narrow vested interests, in which case the general public would prefer an international bureaucracy dedicated to the delivery of the global public good (i.e., to the central goal of the organization) to take over responsibility.
In brief, this implies that unless we take the whole picture into account, a normative judgment cannot be made. In the following, we thus concentrate on a positive analysis and only hint to the potential normative implications here and there without the intention to be conclusive in this respect.
7.3 The UNFCCC as an Empirical Case Study
The UNFCCC was agreed upon by the governments participating at the United Nations Conference on Environment and Development (Rio Conference) in 1992 and entered into force in February 1994. So there is now thirty years of experience with the work of the Secretariat. Our study covers the time period including 2016, that is, until the entry into force of the Paris Agreement. The dynamics after 2016 have changed considerably compared to the preceding period and would warrant a separate assessment. For example, from 2021 onward the CDM, which has provided a significant share of the Secretariat’s funding (see discussion later), has been replaced by new international carbon markets under Article 6 of the Paris Agreement (see Ahonen et al. Reference Ahonen, Kessler, Michaelowa, Espelage and Hoch2022).
Initial Delegation to the UNFCCC Secretariat and Prospects for Further Development
The existing literature on the Secretariat refers to its initial set-up and the first few years of its existence. According to Depledge (Reference Depledge2005: 70 ff.) the Secretariat’s activities were purposefully constrained by the member states to ensure a minimum of technical functionality while avoiding any kind of political interference such as experienced with other environmental treaties. The Secretariat’s activities include the provision of relevant logistics, procedural management, advice to the relevant presiding officers, technical advice in general, drafting text, and the facilitation of informal discussions. Depledge (Reference Depledge2005: 73) underscores that on the basis of this mandate, the Secretariat indeed chose a strictly apolitical “behind the scenes approach” as opposed to the approaches of other treaty secretariats such as the early ozone regime. Similarly, according to Busch (Reference Busch, Biermann and Siebenhüner2009: 251),
the climate secretariat is a ‘technocratic bureaucracy’ that has not had any autonomous political influence…. It has not promoted its own agenda or pursued specific approaches but has responded to requests of parties. It has functioned as an important and valuable but passive information hub in the climate regime that does not autonomously interfere with any political, scientific, or public discourses.
In his study, which refers to the early to mid-1990s, he concludes that the UNFCCC Secretariat is one of the least powerful among nine environmental treaty secretariats under comparison (see also Bauer, Busch, and Siebenhüner Reference Bauer, Busch, Siebenhüner, Biermann, Siebenhüner and Schreyögg2009).
Biermann and Siebenhüner (Reference Biermann and Siebenhüner2009) provide a broader study, in which these nine environmental treaty secretariats are compared. The analysis is based on a comprehensive theoretical framework distinguishing between the cognitive, normative, and executive influence of these secretariats (Biermann et al. Reference Biermann, Siebenhüner, Bauer, Biermann and Siebenhüner2009). Within this theoretical framework, Busch (Reference Busch, Biermann and Siebenhüner2009) identifies problem structure as the central argument for the strong constraints imposed on the UNFCCC Secretariat. Climate change is a politically complex issue on which scientific results continue to evolve and which cannot be solved by technical or administrative means. The salience of the problem as well as the cost of public action is high, and national interests are widely diverging.
When political positions diverge and when the issues are nationally very salient, and compliance to adverse decisions very costly, member countries will not give decisions out of hand and instead keep them directly within political decision-making arenas (Biermann et al. Reference Biermann, Siebenhüner, Bauer, Biermann and Siebenhüner2009). Stone (Reference Stone2011: 23) and more recently Manulak (Reference Manulak2017) provide a complementary rationale for this behavior. They argue that member countries know that autonomy delegated to international bureaucracies can always be used by powerful countries to exert informal influence via these bureaucracies. Thus, granting autonomy to bureaucracies may effectively mean granting more power to some members relative to others. When issues are politically highly contentious, member countries may thus try to avoid such delegation in the first place. According to Stone (Reference Stone2011), a similar argument applies when there is little international consensus about the fundamental purposes of the organization.
In contrast, issues of high technical complexity call for stronger delegation to an international bureaucracy, because such tasks require considerable time and expertise, and autonomous decisions of political committees as well as close monitoring of the bureaucracy become very expensive.
While overall the political element clearly dominates the technical element in international climate policy, it should be noted that the general field covered by the climate negotiations hides a lot of specific, and indeed sometimes also quite technical, issues. Correspondingly, the Secretariat is no monolithic block, and within the Secretariat, some areas dealing with more technical issues may more easily gain autonomy than others – which may lead to quite imbalanced developments within the bureaucratic structure.
However, if at all we expect any change over time, this would require that bureaucrats have somehow been able to convince the principals that more delegation is advantageous for them. According to our theoretical framework, the bureaucracy may try to gradually enhance its autonomy by exerting some influence on the principals. Until recently, the literature suggested, however, that in the case of the UNFCCC, the “straitjacket” imposed on the Secretariat, which rules out any proactive or independent role, also influences its culture in a way to make such developments rather improbable (Bauer, Busch, and Siebenhüner Reference Bauer, Busch, Siebenhüner, Biermann, Siebenhüner and Schreyögg2009: 178; Busch Reference Busch, Biermann and Siebenhüner2009: 261). Depledge (Reference Depledge2005: 73) even concluded that any influence the Secretariat may have critically depends on its invisibility. None of these authors seemed to believe that the Secretariat’s role would see significant changes in the future and that the Secretariat would itself even try to make it change. However, more recently van Asselt and Zelli (Reference van Asselt, Zelli, Jordan, Huitema, Asselt and Forster2018), Hickmann et al. (Chapter 3), and Well et al. (Chapter 4) assert that the Secretariat has become more proactive.
On the basis of our theoretical framework we empirically assess how the inflow of resources and information contributed to the Secretariat’s more proactive role.
A relevant flow of information that would challenge the balance of interest leading to the initial delegation decision cannot be observed. While scientific outcomes have led to a stronger international consensus on the reality of anthropogenic climate change, views on the implications widely diverge, and there is no consensus whatsoever on the responsibilities and commitments to be taken over by individual members (Gupta Reference Gupta2012). The strong political differences between countries are further illustrated by the failure of the Copenhagen conference in 2009 and the postponement of any decision about further steps to 2015. From this perspective, it can thus not be expected that the Secretariat should have become more powerful.
However, we do observe a significant externally induced growth in resources that was, in addition, fully unexpected by member countries and the Secretariat alike. This will allow us, in the following sections, to specifically analyze the impact of resource growth on the dynamics of our model.
The Development of UNFCCC Resources
Let us start by considering some descriptive statistics regarding the overall development of UNFCCC resources, along with some initial interpretations. Following the UNFCCC’s entry into force in February 1994, it took two years to establish the Secretariat. The first budget, available for the biennium 1996/1997, shows expenditures of about USD 4.5 million per year. Until 2015, it increased twentyfold, with particularly strong absolute increases in 2007–2010 (see Figure 7.1). The general growth in resources is mirrored by the growth in staff, which increased from 34 (20 professional and 14 administrative positions in 1995; Depledge Reference Depledge2005: 63) to 558.5 (346 professional-level and 212.5 administrative posts) in 2015. In 2016, both budget and staff numbers fell significantly.
Development of financial and human resources of the UNFCCC Secretariat
Note: The Secretariat budget includes the core budget and all trust funds except the Trust Fund for Participation in the UNFCCC process, which includes only transitory positions. When figures are reported other than for a full calendar year, they are annualized assuming an even spread of expenditures over the year.
The initial major growth phase seems to be related to the preparation of the Marrakech Accords (November 2001) that provided the detailed specifications for the implementation of the Kyoto Protocol agreed upon in 1997. The second major growth phase could then be related to the Kyoto Protocol’s actual entry into force in February 2005, which implied, in particular, the regular assessment of the parties’ greenhouse gas emissions and the evaluation of methodologies and projects submitted in the context of market mechanisms (trade in emission reduction certificates). Finally, the specific rise in 2009 could be related to the expected tasks in the context of the Copenhagen conference in late 2009, which was supposed to bring about an agreement on the follow-up to the Kyoto Protocol after the end of its first commitment period. These interpretations are only partially plausible, however, when looking at the more specific distribution of funds within the Secretariat.
In the period studied, the most important trust fund was the fund for the CDM. The CDM was a market mechanism under the Kyoto Protocol that allowed generation of emission credits (Certified Emission Reductions, CERs) from mitigation projects in developing countries. CERs could be used by industrialized countries to fulfil their Kyoto emission targets. CDM projects had to be formally registered by the CDM Executive Board (EB), which is supported by Secretariat staff in its decision-making. For this service, a fee was charged by the Secretariat both for registration of CDM projects and for CER issuance. The inflow of money through this source was much higher than originally predicted (see Michaelowa and Michaelowa Reference Michaelowa and Michaelowa2017: 251). This led to the accumulation of a surplus, which developed over time as shown in Figure 7.2, as reallocation of funds to other areas of Secretariat activities was impossible (Michaelowa and Michaelowa Reference Michaelowa and Michaelowa2017: 252).
CDM revenues for the secretariat and the development of the accumulated surplus over time (million USD)
Much of the steep rise in total Secretariat staff and expenditure after 2007 can be explained by the resources for the CDM. Note in particular that the strong increase in total expenditure between 2007 and 2010 observed in Figure 7.1 and not explained so far coincides exactly with the increase of CDM revenues. Thus, the Copenhagen meeting in November 2010 may be one explanation for the latter, but the market-driven rise of the CDM along with the related resource growth appears to be the predominant one. Obviously, the two explanations are not mutually exclusive as there may be a happy cooccurrence of needs and available means. However, as the political focus of this major international conference on climate change was on renewed emission reduction commitments, rather than CDM-related activities, there is some doubt about how CDM-related financial resources could have been used in this context.
Owing to the failure of the Copenhagen conference, the willingness of key countries to engage in acquisition of emission credits fell significantly in the early 2010s. A key example is the European Union, which prohibited the import of certain kinds of CERs and also introduced a maximum quota, which was attained around 2013. The CER price fell from EUR 12 in early 2011 to EUR 0.4 in early 2013 and stayed below EUR 1 until the end of the period. After a peak in CDM registration in late 2012, which was due to an EU deadline for credit eligibility, registration slowed to a trickle (see Michaelowa, Shishlov, and Brescia Reference Michaelowa, Shishlov and Brescia2019).
As resources were restricted to being used in the context of the CDM, the revenue increase until 2010 directly translated into staff increases in the Secretariat’s CDM department (see Figure 7.3). Afterward, staffing was kept constant while the revenue surplus increased further until 2012 (see Figure 7.2). This was due to problems in hiring sufficiently knowledgeable staff. During that period the Secretariat outsourced substantial work to external consultants. When the CDM market crisis hit (see discussion in Section 7.5 and Michaelowa, Shishlov, and Brescia Reference Michaelowa, Shishlov and Brescia2019), initially consultancy assignments were scaled back while the Secretariat’s effective staff started to decrease slowly. Only in late 2015, approved staff levels were slashed massively.
CDM versus total secretariat staff
Summing up the developments between the entry into force of the UNFCCC in 1994 and of the Paris Agreement in 2016, the most striking features are thus not only the strong increase in overall financial and human resources but also the increasing dominance of the CDM part within these resources, which is institutionally protected by the international agreement not to use income from the CDM fees for any other purposes. As shown in Figure 7.2, in 2012 the CDM revenues exceeded the entire budget of the Secretariat. Only after 2014, the role of the CDM in the Secretariat’s work declined, long after the market had come to a standstill.
The theoretical framework derived in Section 7.2 suggests that bureaucrats would make use of an important external resource flow like the CDM revenues to influence the principals in order to obtain more autonomy. The following subsection will apply the general theoretical framework to the specific context of the UNFCCC to lay the groundwork for the case studies in Section 7.4.
The Secretariat as Rule-Setter in the Case of a Resource Glut
Our theoretical framework summarized in Sections 7.1 and 7.2 suggests that external resource flows will indeed provide an opportunity for the international bureaucracy to influence its principals more effectively. In addition, the external resources allocated to the Secretariat make the international bureaucracy weigh less heavily on the member countries’ budget, which directly influences the principals’ balance between the cost and benefits of delegation. For both reasons, principals can be expected to rely more heavily on the bureaucracy, to delegate more responsibilities, and to accept the related increase in resources (if relevant). With increased autonomy and resources, which can be used to hire additional skilled professionals, the international bureaucracy can then even further influence the principals, so that we would expect a self-reinforcing effect of the initial external shock whose dynamics might fade out only after some time, when a new equilibrium is reached. Such changes in the influence of the Secretariat could also influence its own self-perception, its organizational culture, and its confidence in pushing for even further autonomy. This would be a specific example for the self-reinforcing dynamics of the initial resource flow. We will not be able to precisely disentangle the individual pathways driving these developments (e.g., whether it is the impact on organizational culture, on the quality of bureaucratic services, or on staff’s self-confidence that is primarily responsible for these self-reinforcing dynamics). In the following section, we discuss two cases where the Secretariat expanded its influence on rule-setting. For a quantitative analysis of the early increase in CDM resources and the influence on rule-setting, especially in the context of the checking of quality of project documentation, see Michaelowa and Michaelowa (Reference Michaelowa and Michaelowa2017).
7.4 Changes in CDM Rules and Regulations Concerning the Secretariat’s Freedom of Action
The stronger role of the Secretariat can be best exemplified in two specific areas: (i) decisions about baseline and monitoring methodologies for potential future CDM projects and (ii) the issue of a standardization of baseline methodologies (standardized baselines). In both cases the stronger role was made possible by the increased availability of expert manpower at the Secretariat, which enabled the Secretariat to argue that it would provide faster and more high-quality methodology-related work than what the “bottom-up” external expert review process could provide. A third case on rules for request for review of problematic project proposals was discussed in depth in Michaelowa and Michaelowa (Reference Michaelowa and Michaelowa2017: 255–256).
Case 1: Baseline and Monitoring Methodologies
Baseline and monitoring methodologies were key to determine the amount of emission credits of a CDM project. They thus directly influenced the amount of money a country received for the export of emission credits and were thus commercially important. Project developers could submit methodology proposals,Footnote 1 which were evaluated by the Methodologies Panel (Meth Panel) and then submitted to the CDM EB, which normally followed the Meth Panel’s recommendation. Traditionally, methodology submissions were evaluated by independent desk reviewers chosen by the Meth Panel. According to information from the EB, the increasing role of the Secretariat is due to the EB’s assessment that the Meth Panel could not handle the increasing number of methodologies. In June 2007, a preassessment of proposals by the Secretariat was introduced (see decision EB 32, Annex 13). While one Meth Panel member selected by the Secretariat would check this (para 7), the Secretariat would develop a draft recommendation (para 14). From February 2010, the Secretariat could skip the independent desk review (see decision EB 52, Annex 9, para 18) if supported by two members of the Meth Panel chosen by the Secretariat itself. It is likely that the Secretariat did not choose overly critical Meth Panel members if it wanted to push a methodology. Moreover, from 2010 onward, the Secretariat started to engage in methodology development on its own initiative, an area previously reserved to external developers. From late 2012 the Secretariat could officially propose methods, while at the same time being the only institution systematically involved in the quality assessment of the methodologies (see Figure 7.4, and summary of the rulemaking changes in UNFCCC 2013a). As no private CDM project developer was willing to invest into methodology development after the CDM price crash, all methodologies submitted since 2012 were developed by the Secretariat or its Regional Collaboration Centres (see Section 7.5). The shift from independent review to secretariat-led rulemaking likely reduced the overall stringency and conservativeness of methodologies.
Changes in procedures for approval of baseline and monitoring methodologies over time
Note: The final decision on a proposed methodology is always taken by the EB
As experience with methodology application accumulated, flaws became visible and project developers were able to ask for a revision of approved methodologies. Usually, they aimed to reduce the transaction costs linked to the use of a methodology as well as its stringency. Traditionally, the Meth Panel prepared the recommendation whether to engage on a revision of a methodology while the Secretariat did the completeness check of the revision request. This was similar to the traditional division of labor between the Secretariat and the Registration and Issuance Team (RIT) for CDM projects. In October 2007, EB 35 (Annex 13, para 8) introduced drafting of the recommendation by the Secretariat. From November 2010 onward (decision EB 54, Annex 2), the Secretariat was able to initiate methodology revisions on its own initiative (para 7). It could then hire outside consultants for preparation of the draft recommendation but needed the approval of the Meth Panel chair (para 14), before selecting one or two Meth Panel members for review. Subsequently, the Secretariat could call for public comments and change the methodology draft if it found the comments relevant. Here again, the approval of the Meth Panel chair was required (para 21d). The Secretariat could also trigger “editorial amendments,” which just needed to be approved by the Meth Panel chair and entered into force automatically unless an EB member objected (paras 29–30). This meant that the Secretariat was able to control the whole revision process if supported by the Meth Panel chair. As noted by one of our interviewees, this would have made sense to speed up processes if the Meth Panel as a whole had difficulties to find a consensus. Yet, it clearly implied an increase in responsibilities for the Secretariat. Here, the shift from the project developer-led revisions to secretariat-led revisions likely increased the overall stringency and conservativeness of methodologies.
Case 2: Standardized Baselines
A strong influence of the Secretariat was also visible in the process of standardizing baselines (i.e., scenarios against which emission reductions by CDM projects had to be assessed). Initially, the EB had asked the Secretariat to develop proposals in consultation with the Meth Panel. However, the rules agreed in September 2011 (decision EB 63, Annex 28) effectively allowed the Secretariat to bypass the Meth Panel with regard to evaluating submissions of standardized baselines (EB 68, Annex 32, paras 15, 16, and 22) and to send the baseline directly to the EB (para 27). While there was still an obligation to include two Meth Panel members assessing the Secretariat’s proposal (para 23), this safeguard could be easily weakened by selecting members disposed favorably to the Secretariat’s proposal.
In November 2011, the EB approved guidelines for standardized baselines and related performance benchmarks developed by the Secretariat (EB 65, Annex 23). A dispute between the Meth Panel and the Secretariat arose regarding the appropriateness of these guidelines. After a long debate, this dispute resulted in an official “information note” sent by the Meth Panel to the EB in November 2012 (UNFCCC 2012d). Thereby, for the first time, an official committee of the UNFCCC questioned the quality of the Secretariat’s work.
In a similar way, a dispute broke out in the context of the determination of the benchmarks used within the standardized baselines. The contested numerical values were hidden in a document innocuously named “Work programme on standardized baselines” (EB 65, Annex 22, para 10). Again, the Meth Panel openly criticized the Secretariat’s approach as documented in another official “information note” for the EB (UNFCCC 2012c).
Since 2013, the standardization drive of methodologies by the Secretariat accelerated. Within three years, standardized shares of nonrenewable biomass – relevant for efficient cookstove projects – were calculated for thirty-four countries. That this was not done according to market demand is shown by the fact that these parameters were used only by forty-one CDM activities – thus on average by just one activity (UNFCCC 2016a). Moreover, forty standardized baselines were developed on a country level for various sectors with the support of Regional Collaboration Centres until the end of 2018, as shown in Table 7.1.
| Year | Power | Waste | Cookstoves | Agriculture | Energy efficiency | Forestry |
|---|---|---|---|---|---|---|
| 2013 | 2 | 1 | ||||
| 2014 | 2 | |||||
| 2015 | 3 | 5 | 2 | 1 | ||
| 2016 | 5 | 5 | 1 | 2 | 1 | |
| 2017 | 1 | 4 | ||||
| 2018 | 2 | 1 |
One of the national delegates to the UNFCCC articulated concern about the strong concentration of resources in the CDM-related part of the Secretariat. During the interview, he underscored that, in general, the Secretariat’s inputs and advice have been extremely useful: “There have been a few cases where the Secretariat put things on the agenda, which created a lengthy process. But in other cases, if the Secretariat had been followed, a lot of time could have been saved. Overall, the negative cases are infinitesimal as compared to the positive side.” Despite this highly positive overall appreciation of the Secretariat’s work, he asserted that the accumulation of resources in the CDM part of the Secretariat clearly required restructuring.
This perspective on resources was challenged by other respondents. They believed that it makes sense for the Secretariat to concentrate resources on an area that is more technical and less politically contested. It was mentioned that a certain financial buffer for the CDM was actually intended to overcome “bad times.” At the same time, another respondent pointed at the experience from Joint Implementation (JI), where the number of projects and thus income from fees had not risen in the same way as for the CDM (see Table 7.A1 for JI staffing, revenue, and budget). In this area, the Secretariat had been much less active in promoting new rules and processes and in proposing increases in its own responsibilities to the corresponding political committee.
Overall, the two cases show how over time and in line with the increase in staff resources the Secretariat took over significant responsibilities in the development and revision of baseline and monitoring methodologies. As such methodologies were critical in determining the monetary revenues of activities under the CDM, the Secretariat was now able to influence which types of activities would be able to generate emissions credits under which circumstances.
7.5 Secretariat Reactions to the CDM Market Crisis
Even after the CER market price had crashed, the Secretariat still kept a large staff force active in the CDM department. Given the reduced inflow of projects, initially Secretariat-led development of rules was intensified as discussed in Section 7.4. A number of baseline and monitoring methodologies were developed by Secretariat staff, and a large number of methodologies were revised. Moreover, outreach to various stakeholders was undertaken to increase the attractiveness of the CDM, for example, setting up a web platform for voluntary cancellation of CERs, which was launched in September 2015 (UNFCCC 2016c). Before the emergence of national-level policy instruments such as the Korean emissions trading scheme and the Colombian carbon tax that accepted CER cancellation certificates in 2017, the platform was used only to a small extent. After the emergence of international climate finance institutions such as the Green Climate Fund, the Secretariat had attempted, in vain, to market CERs to these institutions.
Moreover, the regulatory documents defining the project cycles and other key regulatory steps were aggregated in overall documents, a work that may have pleased legal practitioners but did not have any immediate impact on the use of the mechanism.
Last but not least, a loan scheme to support project developers was launched in April 2012. By late June 2016, 191 applications had been received, with 78 loan agreements signed totaling USD 6.21 million (UNFCCC 2016c: 15). The scheme was closed at the end of 2018. UNFCCC (2018a) states that while 63 loans with a total volume of USD 3.7 million had actually been paid out, only about half were expected to be fully repaid while more than 20 loans were likely to be written off completely, with the remaining loans likely to be partially repaid. Already in 2012 it would have been clear to any observer that this money could never be paid back in the difficult market situation with CER prices close to zero.
The most visible activity was the setting up of Regional Collaboration Centres in all major world regions. This started immediately after the market crisis: The center in Lomé, Togo, started in January 2013, followed by Kampala, Uganda (May 2013), St. George’s, Grenada (July 2013), and Bogota, Colombia (August 2013). Another one in Bangkok followed in September 2015. While initially the focus of the centers was on developing CDM project pipelines, they did become general capacity-building entities, supporting knowledge transfer in the context of the Paris Agreement and its new concepts such as nationally determined contributions (NDCs), as shown directly in the statement “Set up to spread the benefits of the clean development mechanism (CDM), the RCCs have broadened their role since adoption of the Paris Agreement on climate change in December 2015, supporting the development and implementation of countries’ nationally determined contributions to climate action under that agreement” (UNFCCC 2017a: 1). As per this new mandate, the key activities of Regional Collaboration Centres in 2018 were to support the development of measurement, reporting, and verification systems and the elaboration of studies on domestic carbon pricing policy instruments and NDC partnership plans (UNFCCC 2018b).
7.6 Conclusions
The UNFCCC Secretariat was able to mobilize an unexpected volume of revenues from the CDM, a market mechanism under the Kyoto Protocol that proved surprisingly attractive to the private sector between 2005 and 2011. The inflow of over USD 350 million within a decade led to a rapid expansion of staffing at the Secretariat and a tendency to take over rule-setting under the mechanism. We provide evidence through case studies on rules for development of baseline and monitoring methodologies as well as standardization of such methodologies. This complements evidence found through regression analysis by Michaelowa and Michaelowa (Reference Michaelowa and Michaelowa2017). Our assessments confirm theoretical considerations that an international bureaucracy tends to take over tasks from its member governments as soon as its resources increase.
The collapse of the CDM market from 2012 onward initially led to a “hibernation” attitude of the Secretariat, which only slowly laid off staff and even increased Secretariat-led rule-making, despite lack of activity on the market. Standardization of methodologies was undertaken and the rulebook streamlined significantly. However, this can be partially seen as a manifestation of Parkinson’s Law, as a number of activities were undertaken that clearly did not have significant benefits, such as loans to project developers and brokerage activities to find buyers for emission credits. Only after five years of crisis was a serious downscaling of staff undertaken. At the same time, activities of remaining staff were tacitly reoriented to support negotiations under the Paris Agreement and its implementation, for example in the context of Regional Collaboration Centres that could no longer support identification of projects.
Acknowledgments
We thank Liliana Andonova, Carsten Hefeker, and Tana Johnson for their very thoughtful comments that have helped us to substantially improve this chapter. Moreover, we thank all those that shared their precious time with us during our interviews, and without whom this chapter would have been much less rich in concrete examples.
Open access