Informally Governing International Development: G7 Coordination and Orchestration in Aid

Informal groupings like the G7 aim to address global development challenges but lack the administrative and budgetary capacity to drive change directly. Instead, the G7 seeks to catalyze international action that reﬂects its priorities. For example, the G7 attempts to set the international development agenda by publishing annual communiqués with actionable commitments designed to inﬂuence the behavior of G7 donor countries, non-G7 donor countries, and international organizations. But questions about the G7’s ultimate impact persist, as critics contend the informal G7 can do little more than pay lip service to development challenges. We provide empirical evidence that the G7 shapes international development in two ways. First, when the G7 emphasizes a policy area in its annual communiqués, donors allocate more aid to that policy area. Second, when the G7 highlights a policy area in its annual communiqués, donors establish more trust funds in that policy area. This suggests the G7 serves simultaneous coordination and orchestration roles in international development: it coordinates its member states’ aid and orchestrates non-G7 bilateral and multilateral aid. The study’s theory, approach, and ﬁndings can inform further research on whether and how informal organizations ultimately affect states, formal international organizations, international cooperation, and global governance.


Introduction
Do informal international organizations influence policy?Informal governance entities are increasingly central to global governance ( Vabulas and Snidal 2013 ;Westerwinter, Abbott, and Biersteker 2021 ;Roger and Rowan 2022 ), yet their impact is often ambiguous.Any effects must manifest through the behavior of other actors, including intermediaries like formal intergovernmental organizations ( Abbott et al. 2015b ;Downie 2022 ) and member states themselves ( Kirton 2006 ;Morin et al. 2019 ).Moreover, ambiguity may be intentional insofar as soft law governance approaches serve state interests by reducing obligation through strategic imprecision ( Abbott and Snidal 2000 ).Such imprecision can make informal groupings "only loosely related to the structure of particular problems … and not especially effective," leading to questions about the conditions under which they ultimately affect policy ( Roger 2020 , 206-7).
This study considers the case of informal governance in international development by investigating the G7's influence on aid donors.The G7, which publishes annual communiqués that include dozens or hundreds of policy commitments, has, since its inception in 1975, placed economic and social development at the center of its agenda.Yet the impact of the G7's development rhetoric has been unclear at best ( Kirton 2006 , 461).A common critique is that G7 policy commitments are inevitably mere lip service: "how effective could the G7 be without a permanent secretariat … how could the opacity, proliferation, and relative informality … contribute to the building of a coherent system of global economic governance?" ( Pauly 2021 , 2).Meanwhile, others see the G7's informality as an essential feature of any potential contributions to global governance, including international development ( Morin et al. 2019 ;Roger 2021 ;Cooper 2022 ).
Previous studies have tested for G7 effects by identifying whether or not member states comply with G7 communiqué pronouncements by implementing corresponding policies at home ( Kirton 2006 ;Kirton and Larionova 2018 ).This study considers a more general and wider-ranging pathway through which the G7 may affect policy across not only member states but also non-G7 states and formal international organizations: informal governance through coordination and orchestration .
We hypothesize that the G7 helps informally set the international development agenda through simultaneous coordination and orchestration roles.First, we argue that the G7 coordinates the development efforts of its members, as seen in other policy areas such as security and domestic economic management ( Payne 2008 , 522, 525).Second and simultaneously, we argue the G7 orchestrates the development efforts of non-G7 bilateral and multilateral donors ( Abbott et al. 2015b ) by providing an ideational agenda and, in turn, material support for specific development goals through its annual communiqués ( Gstöhl 2007 ).
We quantitatively test for G7 coordination and orchestration in international development with two observable propositions.The first is in donor spending: when the G7 emphasizes a policy area in its annual communiqués, we ex-pect donors to spend more on that policy area.The second is in institutional creation: when the G7 highlights a policy area in its annual communiqués, we expect donors to establish more trust funds-institutions that direct aid to specific policy areas-in that policy area.We find evidence in favor of both propositions.G7 member states, non-G7 states, and international organizations dedicate more money and trust fund capacity to the G7's policy priorities and less money and trust fund capacity to the G7's non-priorities.
The findings suggest the G7 is a source of informal governance in international development.Specifically, a strong G7 relationship to aid flows and trust fund establishment indicates informal governance of donor behavior, with implications for understanding the politics of aid.Furthermore, to our knowledge, this study is among the first to quantitatively test for informal organization effects in global governance (joining Westerwinter, Abbott, and Biersteker 2021 ).The findings thus have broader implications for the literature on global governance and international organizations, providing evidence that informal groupings like the G7 can affect states, formal international organizations, and international cooperation more generally.
The next section details the G7 and its relationship to international development.The third section reviews the literature on informal governance, hypothesizing that the G7's annual summit communiqués are a coordination and orchestration tool through which the G7 can informally shape donor behaviors.The fourth section outlines the empirical strategy for testing the hypothesis.The fifth section presents results and notes limitations.The conclusion discusses the implications.

The G7 and International Development
The G7 is an informal grouping of seven of the world's largest economies.While the G7's original purpose was to coordinate domestic economic policies among these economies, the G7 has, from its inception, also sought to address global development issues ( Fischer 2020 , 5).This scope manifests in the G7's annual summit communiqués, which include commitments that reach beyond G7 domestic policy to include commitments about contributions to international development.G7 development commitments cover a range of economic and social policy areas that ultimately aim to enhance peace, human security, economic growth, governance, health access, climate resilience, social equality, and political rights outside of the rich world ( Fratianni, Savona, and Kirton 2003 , 1).
If the G7 affects international development in practice, its summit commitments must drive resources to these purposes.To this end, the G7 explicitly targets multilateral and bilateral aid.1 Efforts to affect donor flows from the major international financial institutions (IFIs), regional multilaterals, G7 member states, and non-G7 donor countries are typically the "centerpiece" of G7 development commitments ( Fratianni, Savona, and Kirton 2003 , 1).
Anecdotally, G7 members and partners claim G7 meetings substantively influence international development by steering donor resources.Member states see the G7 as a coordination forum where they can "mobilize" funds for various social and economic development purposes ( United States Government 2022 ).In addition, non-G7 donors such as the World Bank report that many funding initiatives and innovative developmental financing instruments are "launched in the context of G7" meetings ( World Bank 2010 , 23).For example, at the 2009 L'Aquila summit, the G7 "pledge[d] to advance by the end of 2009-consistent with our other actions aimed at an improved global governance for food security-the implementation of the Global Partnership for Agriculture and Food Security."This commitment was designed to be "an important means through which [G7] national authorities, supported by [the World Food Program and] other specialized Agencies, Funds, and Programmes, together with non-governmental organizations, can provide help to people facing acute hunger" ( G7 2009 ).G7 countries followed through by establishing a corresponding food security trust fund.Non-G7 donors also subsequently contributed funds, and the World Bank, which endorsed the initiative, has been running the fund since ( World Bank 2021 ).The G7 effectively set the agenda for food and agriculture aid, both coordinating member state aid and orchestrating other donors.
Debt relief provides another oft-repeated example of the G7 effectively setting the agenda for international development.1987 Venice summit communiqué statements explicitly targeted Paris Club donors and "urged [them] to work out necessary technicalities to ensure [debt rescheduling in poor countries] by the end of this year."Later that year, the Paris Club offered substantial debt relief for the first time ( Gstöhl 2007 , 13-14).The 1992 Munich communiqué targeted IMF loan expansion, and the IMF subsequently increased lending.The 2000 Okinawa summit communiqué urged HIPC debt cancellation for at least 20 countries, and creditors ultimately canceled debt for 22 countries ( Gstöhl 2007 , 15).Over time, multilaterals and other donors "have shown a remarkable degree of compliance" with G7 debt relief agendas expressed in annual summit communiqués ( Gstöhl 2007 , 20).
In practice, such effects are plausible because each G7 summit communiqué is the product of a series of meetings between representatives from member-state governments, partner governments, partner organizations, and civil society.Important meetings in the build-up to each summit include, among others:2 r meetings of ministers from governments' executive bu- reaucratic units; r meetings of public and private sector experts and special- ists that attend thematic G7 "working group" sessions; r meetings of civil society actors participating in thematic G7 "engagement group" sessions.
Meanwhile, "sherpas" from each member state meet regularly throughout the year to negotiate and prepare final G7 annual communiqué commitments based on feedback and input from the ministerial, working group, and engagement group meetings.As the Council on Foreign Relations phrases it, these envoys "hammer out policy initiatives at meetings that precede the gathering of national leaders" to decide the policy commitments to be included in the final summit communiqué ( Council on Foreign Relations 2023 ).
We highlight this process to emphasize that G7 summit communiqués are the product of strategic negotiations and discussions among member states.These preliminary efforts are explicitly designed to identify jointly backed initiatives that annual communiqués then promote.In this sense, members use the G7 informal forum as a strategic opportunity to articulate certain goals and thus set the agenda for the international community as well as themselves.
Yet skepticism remains about the degree to which the G7 process and summit communiqués substantively affect global policy-including international development efforts.Evidence that the G7 follows through on its own development commitments is mixed at best ( Kirton 2006 , 461).Scholars have trouble accounting for important aspects of the G7's informal nature (see discussion in Cooper 2022 ).Popular commentary often critiques underwhelming commitments and lack of action on climate, food, health, debt relief, or other developmental issues ( Harvey 2021 ;Woodcock and Merrick 2021 ;Mathiesen and Herszenhorn 2022 ).
This debate leaves the G7's impact on international development unclear.Critics focus on the lack of outcomesbut the informal G7 cannot directly affect outcomes on the ground.Proponents emphasize the G7's role as a coordination forum-but this role and any effect are often difficult to identify or disprove beyond anecdotal evidence.

The G7 As Development Coordinator and Orchestrator
This study accordingly scopes analysis to a plausible and quantifiable question: does the G7, as an informal international organization, have an empirically observable relationship to donor behaviors?Specifically, we ask whether the G7 helps inform what multilateral and bilateral donors prioritize in their aid efforts, particularly regarding (1) direct spending decisions and (2) trust fund establishment. 3This section reviews the informal governance concepts of coordination and orchestration, applies them to the G7, and identifies the observable implications of such G7 roles in international development.

Coor dination, Or chestration, and the G7
Informal governance entities like the G7 do not have the capacity, expertise, or legal standing to direct resources and staff toward desired ends.Instead, informal entities affect global governance by, among other functions, coordinating the activities of their own members and orchestrating the activities of non-members.Through these dual informal functions, the G7 can affect the behavior of its member states as well as other actors ( Gstöhl 2007 , 3).
We highlight these dual informal functions to theorize why the G7 is not merely reducible to the behavior of members alone.Insofar as G7 member states pool resources and agenda-setting efforts through G7 coordination, this yields the potential for an additional external-facing global governance role despite the informal grouping's lack of bureaucracy ( Gehring and Urbanski 2023 ).In other words, the G7 may orchestrate the behavior of non-member actors since the G7 communiqués serve as a focal point that elucidates how others can align with G7 priorities.In the development context, such actors include non-G7 bilateral and multilateral donors.

G7 COORDINATION
First, coordination involves serving as a forum for solving collective action problems.Member states establish and participate in formal and informal international organizations to collaborate for mutual gains ( Martin and Simmons 1998 ).As the G7 is a small group of powerful states that share core characteristics and interests, it is reasonable to suspect that a functionalist logic explains the G7's emergence and persistence.Coordination across a limited but important set of policy areas benefits participating members (see Koremenos, Lipson, and Snidal 2001 ).Over time, the G7's agenda has expanded "from the original economic, monetary and financial issues to matters of development, environment, foreign and security policy, and now covers almost all important global problems" ( Gstöhl 2007 , 2).Across time and agenda expansions, the G7 remains a policy coordination forum, manifest in annual communiqués that include actionable commitments agreed-to by member states ( Payne 2008 ).In this sense, the G7 performs an informal governance role by providing a collaboration opportunity where member states may coordinate policies.The degree of implementation, and thus the effect of these coordination efforts and the communiqués they inform, varies and is often criticized on various grounds ( Putnam and Bayne 1984 ;Kirton 2006 ;Bayne 2018 ;Kirton and Larionova 2018 ).However, the coordinating logic underpinning the G7 remains, and this study takes a new approach to test for its presence in the context of international development policy.

G7 ORCHESTRATION
Second, as the food and debt relief anecdotes above highlight, the G7 often seeks to affect the behavior of other actors.But the G7 cannot directly govern them.An alternative informal governance strategy is "orchestration:" mobilizing and facilitating the voluntary cooperation of other actors.These other actors are "intermediaries."Orchestrators enlist intermediaries by articulating "ideational and/or material inducements to create, integrate and maintain a multi-actor system of soft and indirect governance, geared toward shared goals that neither orchestrator nor intermediaries could achieve on their own" ( Abbott et al. 2015b , 4).In other words, orchestration is a strategy by which actors may guide third parties ( Abbott et al. 2016 , 2) over which they lack "hard control" ( Abbott et al. 2015b , 4). 4nformal groupings like the G7 are a prime example of orchestration efforts in global governance, created by members "for the express purpose of orchestrating" ( Abbott et al. 2015a , 378).Their agendas articulate "guidance as to pressing governance issues and plausible policy solutions.This influences intermediaries' priorities and helps them define their strategies" ( Abbott et al. 2015b , 15).For example, the G20, a descendant of the G7 ( Viola 2020 ), has been shown to orchestrate the policies and behaviors of several multilateral organizations and non-G20 states ( Viola 2015 ;Downie 2022 ).Similarly, the G7 (orchestrator) uses annual meetings and communiqués to signal shared ideational and material goals on which non-member actors (intermediaries) may act.
Orchestrators ultimately enlist intermediaries in the pursuit of governing certain "targets."Targets often are states or private entities, but not necessarily so.Targets may also be "governance goals" broadly conceived, including the provision of public goods ( Abbott et al. 2015b , 5-6).For example, the World Health Organization enlists intermediary health organizations to "improve health services" in places it operates ( Abbott et al. 2015b , 5-6; see also Hanrieder 2015 ).In that case, then, the target is the work on the ground in locations where the health orchestrator and intermediaries jointly operate.We follow this example to define the target of our application.In the context of this study, the target of G7 orchestration is aid allocation to developing countries.The G7 (orchestrator) seeks to govern international development by inducing other donors (intermediaries) to align their aid allocations and efforts (the target) with the G7 agenda.
Following the orchestration model outlined in Abbott et al. (2015b , 4), and with our nuanced target as defined above, a G7 informal orchestration role in international development can be conceived as in Figure 1 .
Orchestration is likely central to any influence the informal G7 has in international development.Because the G7 does not have a hierarchical relationship with non-G7 bilateral and multilateral donors, they can only enlist such other actors voluntarily through material and ideational inducements.Accordingly, if the G7 affects international development in ways other than merely coordinating its own member states' development efforts, it does so by inducing other donors to prioritize the G7's goals ( Gstöhl 2007 ).Communiqués include explicit ideational inducements and corresponding (if indirect, as detailed below) material incentives for other donors to follow.Such an agenda-setting mechanism is a tool by which the informal G7 can orchestrate intermediaries to act on behalf of its agenda.In this sense, G7 communiqués are comparable to the Sustainable Development Goals, which inform international development financing efforts insofar as they provide "guidance [on] pressing governance issues and plausible policy solutions" ( Abbott et al. 2015b , 15).
Orchestration follows a functionalist logic in that orchestrators further their goals by "tapping into the capabilities of third-parties" ( Abbott et al. 2016 , 2).As described earlier, the G7 strategically explicates goals for the international community.As an informal grouping without an autonomous bureaucracy or set of resources, the G7 has no agency itself.It is, however, greater than the sum of its member-state parts because it is the choice to act in concert that provides the platform to orchestrate other intermediary actors to behave in ways that serve the groups' governance goals.This reflects a functionalist logic insofar as these goals aim to induce non-G7 actors (intermediaries) to contribute their resources to G7 priorities.The power of networks may also underpin why intermediaries would be orchestrated by the agenda explicated in G7 communiqués (see Mattli and Seddon 2015 ).
Yet questions persist about the impact of this informal grouping.In international development specifically, whether and the extent to which the G7 orchestrates intermediary donors remains unclear.

Theorizing G7 Informal Governance in International Development
If the G7 coordinates and orchestrates international development, this would yield at least two observable implications.G7 development policy priorities, reflected in calls to action expressed in annual summit communiqués, would af- To reiterate, such an informal governance role would simultaneously affect aid behavior across both G7 member states themselves (through the coordination function) and non-G7 bilateral and multilateral intermediaries (through the orchestration function).Below, we theorize why we expect this to be the case in both donors' funding allocations and trust fund establishment.First, the G7's annual summit communiqué summarizes the agreed-to positions of G7 member states, articulating the outcome of annual coordination efforts.As seen across policy areas, to varying degrees, G7 governments often act on the policy commitments included in the communiqué ( Kirton 2006 ).If the G7 has a coordination effect on member-state development contributions, we should expect subsequent aid allocations to reflect the policy priorities expressed in the G7 communiqué.Moreover, bilateral aid agencies based in G7 countries should gain support and resources at home should they contribute to G7 priorities (on the domestic politics of aid support, see Milner and Tingley 2010 ; Dietrich 2021 ).
Simultaneously, the communiqué clarifies the G7's annual developmental priorities for intermediary donors.These non-G7 bilaterals and multilaterals gain clarity about what the G7's development agenda is and thus guidance on what to spend their money on should they wish to align with the G7 development agenda.Even if the language of a particular communiqué commitment does not only refer to developing countries, it indicates the issues at the heart of the G7's agenda in a given year.Donors can then act on this information.
Such explicit ideational agenda-setting creates a further material inducement for intermediaries to pursue the G7's goals.To be sure, these material inducements are indirect; it is not the case that the G7 provides money directly to the donors that follow the G7's priorities.Instead, G7 communiqués indicate the policy areas that G7 states are prioritizing, signaling potential material gains for intermediary donors that align their aid contributions to those G7 priorities.
These material benefits are multiple.For multilaterals seeking approval and resources from their primary G7 principals, communiqués indicate how they can obtain material financial and procedural efficiency gains by adhering to the G7's agenda ( Stone 2011( Stone , 2013 ; ;Clark and Dolan 2021 ).For bilateral aid agencies in non-G7 countries, communiqués clarify how they can increase alignment with the agenda of major donors.Insofar as effectiveness is a primary concern ( Dietrich 2013 ), these donors would perceive impact gains by following the G7's priorities.Even if it leads to ineffective herding and redundancy in practice ( Frot and Santiso 2011 ), following G7 priorities yields visibility for donors seeking to "fly their flag" as relevant aid contributors ( Canavire-Bacarreza, Neumayer, and Nunnenkamp 2015 ).Following G7 priorities may also provide legitimacy.When the G7 supplies information about the development priorities of the most important donors, other donors learn what is considered legitimate in international development, and those policies subsequently diffuse across actors (for exam-ples of such effects in international development, see Swiss 2012 ; Baccini, Heinzel, and Koenig-Archibugi 2022 ).
Multilateral and bilateral intermediary donors thus have both ideational and material reasons to have their spending orchestrated by G7 development priorities.These multiple logics underpin why "donors often use IGO agendas as reference points when determining spending priorities, helping intermediaries" make aid allocation decisions ( Abbott et al. 2015b , 15).
Second, and for many of the same reasons, we theorize that new donor institutional arrangements are affected by G7 priorities.The policy areas to which donors dedicate new trust funds are likely to be influenced by the G7's agenda.Trust funds are ad hoc institutions established by donors to direct aid to specific development policy areas, which are then given to and administered by multilaterals ( Dietrich 2013 , 701;Reinsberg 2017 ;Reinsberg, Michaelowa, and Knack 2017 ;Dietrich, Reinsberg, and Steinwand 2022 ).The ad hoc nature of trust funds means donors establishing new trust funds are responsive to topical, salient development priorities.Indeed, G7 members have expressly established trust funds following G7 policy coordination efforts.Recall the example of the 2009 L'Aquila summit discussed above, where the G7 served a clear coordination role in the establishment of a large food trust fund ( World Bank 2021 ).
In addition, and again simultaneously, the G7's annual summit communiqués provide explicit ideational inducements for new trust funds established by non-G7 donors.As with spending, the communiqué indicates priorities that would likely have reputational and legitimacy benefits should these intermediaries dedicate their new funds to them.Establishing new trust funds that match G7 priorities yields visibility for donors interested in being seen as major contributors to development.Also, as with spending, there may be material efficiency and impact gains since following G7 priorities means a donor country is coordinating with the Western development agenda.Insofar as trust funds are centralizing aid within the administrative capacity of larger donors like the World Bank, this may have a lead donor effect that addresses challenges associated with aid proliferation and fragmentation ( Steinwand 2015 ;Gehring et al. 2017 ).There is anecdotal evidence that this logic informs donor behaviors.Referring once again to the food aid example, non-G7 countries such as Norway and Korea followed the G7's prompt to prioritize food security in developing economies at the 2009 L'Aquila Summit by contributing subsequent trust fund resources to the GAFSP ( World Bank 2021 ).
Concerns have been raised about the potential for such behavior to yield visible but ineffective donor use of trust funds.The World Bank highlights that too many small trust funds reduce impact by creating excessive overlap ( World Bank 2019 , 117), and donor government agencies themselves raise similar concerns ( Norway Office of the Auditor General 2021 ).But our focus here is donor behavior at the time of trust fund establishment, not the impact of the trust funds on aid performance over time (see Heinzel, Cormier, and Reinsberg 2023 ).G7 orchestration of donors may or may not be ultimately developmental.But regardless of the implications for long-term effectiveness, the G7 communiqués do provide clear guidance on what it would take to align with the G7's agenda at a given point in time.

Hypotheses
G7 summit communiqués provide an ideational agendasetting tool that anecdotal evidence suggests can informally coordinate and orchestrate donors' direct spending outlays and trust fund efforts.Implicit is an indirect material inducement: the G7's annual communiqués signal ways in which donors may realize efficiency gains, possible effectiveness gains, and reputational or visibility benefits by aligning their cash allocations and trust funds to G7 priorities.We hypothesize that the G7 does coordinate and orchestrate donors, with two observable implications and thus two testable hypotheses: H1: When the G7 prioritizes a policy area, donors will allocate more aid to that policy area.H2: When the G7 prioritizes a policy area, donors will establish more trust funds in that policy area.

Empirical Strategy: Data and Model Specifications
Do G7 priorities influence aid donors?We test for G7 coordination and orchestration using G7 summit communiqués as an observable agenda-setting tool.Data on G7 commitments, aid flows, and trust funds allows us to construct a panel dataset from 1975 to 2021 organized by policy topic and year. 5We begin the discussion with the explanatory variable because it establishes the policy area units used to set the panel structure.We then discuss dependent variables, control variables, and model choices.

Explanatory Variable
Our explanatory variable is the annual change in the G7's emphasis on a policy area in its yearly communiqués.G7 communiqués are divided into brief, enumerated sections of text.Since 1975, the G7 Research Group (G7RG) 6 has coded the enumerated sections of text that represent policy commitments ( G7 Research Group 2019 , chap.2). 7In the G7RG codebook, policy commitments are defined as "discrete, specific … publicly expressed, collectively agreed-to statements of intent" in the G7's annual communiqués that indicate a forward-looking "promise or undertaking by summit members" ( G7 Research Group 2019 , 4).Commitments are thus similar enough in terms of specificity that they are "actionable," allowing G7RG coders to track over the coming year whether member states comply with these commitments.
The G7RG then groups each communiqué commitment into one of the 30 policy areas listed in Table 1 .In essence, 5 Results are robust to using a panel limited to the years 1990-2021.See Supplementary Appendix C . 6Based at the University of Toronto (http://www.g7.utoronto.ca/about/index.html,last accessed December 7, 2023). 7The full G7RG dataset codes commitments across 34 policy areas.We drop three policy areas that are unlikely to affect the specific topics emphasized or de-emphasized in aid policy each year: East-West Relations, IFI Reform, and International Cooperation.We also omit the overly broad "Developing Country" policy category, as this catch-all term does not relate to a specific policy area by which we could observe more or less aid going into a specific policy area each year (see Supplementary Appendix F for a detailed discussion).We do, however, use the "Developing Country" category as a control in robustness checks to account for relative emphasis on non-rich countries in the annual communiqué rhetoric (see Supplementary Appendix F ). the G7RG codes the annual rhetorical agenda of the G7 by policy area.This coding effort yields a dataset that identifies the rise and fall in the relative importance of these policy topics in the G7's communiqué over time.As noted above, even if the language of a particular commitment does not only refer to developing countries, it indicates the policy areas central to the G7's agenda each year.The G7RG uses this data to track whether G7 members comply with commitments after summits.However, are interested in the commitment rhetoric as an observable agenda-setting device for the international development community.To our knowledge, it is the best available and most succinct measure of the agenda that the G7 sets for itself and others each year.
Our explanatory variable is based on the G7RG's count of commitments addressing the 30 policy areas each year.We generate a dummy variable indicating whether there were more commitments in a policy area than last year: G7Commit = 1 if the G7's communiqué includes more commitments in a policy area than the previous year, and G7Commit = 0 if the communiqué includes the same number or fewer commitments in a policy area than last year.
We expect that if the G7 places more emphasis on a policy than it did the previous year ( G7Commit = 1), this will be associated with donors dedicating more cash and trust fund resources toward that area.In contrast, if the G7 does not prioritize a policy area as much as it did the previous year ( G7Commit = 0), this will be associated with donors allocating less cash and trust fund resources toward that area.
Since communiqués do not include the same number of commitments each year, we probe the possibility that the G7 is merely an "agenda-expander" that increases the number of commitments across all policy areas over time.Simple expansion would undermine the assumption that the G7 shifts priorities each year, and thus our premise that this variable can capture variation in G7 priorities across competing or alternative policy areas over time.
Figure 2 confirms that mere agenda expansion is not a theoretical or statistical concern.No simple upward trend exists in the number of commitments within any policy area.There is, instead, significant annual variation within each policy area.Significant drops in the subsequent year follow many increases in the number of commitments in a policy area.Indeed, there are nearly three times as many observations where G7Commit = 0 (954) as observations where G7Commit = 1 (342). 8A decrease in a policy area's relative 8 Author calculations.

Annual Variation in G7 Commitments
importance to the G7 from year to year is nearly three times as likely as an increase.This makes G7Commit a conservative indicator of when the G7 places a new or renewed degree of emphasis on a policy each year.The upshot is that there is significant variation in the variable G7Commit across policy areas over time.

Dependent Variables
We hypothesize that the policy areas emphasized or deemphasized each year by the G7 serve an informal governance role in international development, with two observable implications.The first is in donor aid commitments (H1).The second is in the policy focus of new donor trust funds (H2).We include a dependent variable for each hypothesis.
The aid dependent variable is Spend : the annual change in direct multilateral and bilateral aid commitments within each policy area.We group multilateral and bilateral aid because we expect simultaneous coordination of G7 cash allocations and orchestration of non-G7 target actor cash allocations.9Data are from the OECD Creditor Reporting System (CRS).We map CRS purpose codes, which identify the policy area for which funds are allocated, to corresponding G7 policy areas.See Supplementary Appendix B for the G7 policy area and CRS purpose code mapping.
We also test three alternative spend variables.First, we create a non-G7 variable of multilateral and non-G7 bilateral spending ( NonG7Spend ) to confirm that these relationships are not merely driven by G7 aid flows (in other words, that there is evidence of orchestration of non-G7 donors, independent of G7 member-state aid coordination).Second, we create a bilateral-only variable ( SpendBilatOnly ) to check that multilateral flows do not drive these relationships.10Third, we create a percentage variable measuring the percent of total aid allocated to that policy area that year ( PerTotAidSpend ).Results are robust to using these alternative dependent variables.See Supplementary Appendices D and G for results and further discussion.
The trust fund dependent variable is NewTFs : the number of new trust funds dedicated to a policy topic each year.Data are from an updated version of the multi-bi-aid dataset ( Eichenauer and Reinsberg 2017 ), which tracks the trust funds established by donor countries each year.That dataset codes the policy area of trust funds using CRS purpose codes, allowing us to map trust funds to G7 policy areas using the same CRS codes as the spending dependent variable (see Supplementary Appendix B ).
Augmented Dickey-Fuller tests indicate stationarity of Spend (as well as the level of Spend ) and NewTFs .

Control Variables
Each model includes unit (policy area) and time (year) fixed effects.As explained below, we use multiple leads and lags of G7Commit as placebo tests in half of the models.We also include a lagged dependent variable in all specifications, accounting for inertia in donor spending and trust fund establishment ( Carey 2007 ). 11These controls account for unobservable contextual effects in global governance and development aid.We lastly control for the cumulative number of trust funds ever targeted toward a policy area ( TFsEver ), as historically important topics may be more likely to attract donor cash and trust fund emphasis.

Models
Our first set of estimations model Spend with fixed effect OLS models: An important concern is the extent to which G7Commit and Spend are jointly determined by "fads" or general, if difficult to empirically identify, trends in global governance priorities.Short of obvious instruments, we probe this with a placebo test using multiple lags and leads of G7Commit.If spending is more likely a function of long-term trends in global priorities than G7 coordination and orchestration, then spending would be "sticky"-global priorities identified this year would also be associated with shifts in spending in the past or the future.In our application, if multiple lags or multiple leads of G7 commitments are significantly associated with aid allocation, this would be evidence that the G7 and aid spending are less likely linked through coordination and orchestration and more likely linked through global norms and trends in aid.Moreover, if including these deep lead and lag placebos erased the significant associations found in simpler models, this would also be evidence against the hypotheses.We use such placebos in a second set of models: We then estimate NewTF using unit and time-fixed effect Poisson models, due to the count nature of NewTF : We finally use another series of placebo tests in Poisson models to probe whether the relationship between G7Commit and NewTFs more likely reflects faddish aid trends or our hypothesis about more immediate G7 priorities.If multiple lags or multiple leads of G7 commitments are significantly associated with trust fund establishment, this would be evidence that the G7 and trust fund establishment are linked not through coordination and orchestration but 11 Supplementary Appendix E drops lagged dependent variables to remove potential Nickell Bias.through contextual factors such as global norms and trends in international development.These models take the form:

Results
Table 2 presents three estimations of the first model above, building to the full specification.Model 1 is a simple correlation.Model 2 includes lags of spending and trust fund emphasis.Model 3 adds the final control for the historical importance of this policy area to donors.All estimates provide evidence in favor of H1.An increase in G7Commit is associated with increased spending of between $30 and $34 million USD on that policy topic each year.
We then use placebo models to test for evidence that this relationship is more likely a function of shifts in global priorities than an immediate link between G7 communiqué commitments and aid spending.Model 4 uses one year of leads and lags.Model 5 adds two years of leads and lags.Model 6 adds three years of leads and lags.If these deep lead and lag placebos are significant, or if the G7Commit relationships from Table 2 become insignificant, this would be evidence that the G7 and aid spending are linked less through coordination and orchestration than through general aid norms and trends.
Table 3 shows that only commitments this year are significantly correlated to shifts in aid spending in a policy area.Figure 3 visualizes this in plot form.While not a causal test, this is evidence that the association we find between G7 commitments and aid allocations is less likely a function of broader shifts in global aid priorities than it is a function of our theory of G7 coordination and orchestration.The coefficients also remain consistent with Table 2 , estimating an increase of between $27 and $34 million USD when the G7 increases its emphasis on a policy area in its communiqués.
Table 4 then reports estimates of Poisson models, where NewTF is the dependent variable.Model 7 is a simple correlation and Model 8 includes lags of spending as well as trust funds.Model 9 controls for the relative importance of this policy area to donors historically.All models suggest G7Commit is significantly associated with how many trust funds donors establish in a given policy area each year.All   3 G7Commit lead/lag placebo tests else equal, G7 emphasis on a policy area is estimated to correspond to a 10 percent-11 percent shift in the expected log count of trust funds newly established in a policy area each year.
Again, short of a plausible instrument, we seek to address endogeneity concerns due to global development fads.We isolate the timing of the relationship, checking whether multi-year lags and leads of G7Commit are associated with NewTF .If the significant relationship identified in Table 4 is merely a function of longer-term global development policy trends, we would find deep lags or leads of G7 commitments to be associated with trust fund establishment or find the G7Commit relationships in Table 4 become insignificant.
Table 5 shows only G7 commitments in the same year and last year correlate to the establishment of new trust funds in a policy area.Future realizations of G7 commitments are never significant.This finding suggests that the association we find is less likely due to broad aid trends than G7 coordination and orchestration.It is also inconsistent with reverse causality.Given the reported pattern, donor trust fund activ- ity is unlikely to drive future G7 rhetoric (i.e., the G7 does not retroactively justify prior behaviors).Nor are the second and third lags of G7 rhetoric significant.This also suggests that the long-run salience of a policy area is less likely responsible for the identified relationship than a short-term G7 agenda-setting role in aid. Figure 4 visualizes these nonrelationships across G7Commit leads and lags.

Robustness Tests
The appendix reports more information on data, including descriptive statistics ( Supplementary Appendix A ) and the purpose code mapping by which we link CRS data to G7RG data ( Supplementary Appendix B ).The appendix also includes additional tests and robustness checks.As noted above, results are robust to limiting the 1975-2021 sample timeframe to a 1990-2021 sample timeframe ( Supplementary Appendix C ).As also noted above, we model a non-G7 spend variable to check that the G7 relationship between G7 commitments and aid spending is not merely a function of G7 member-state spending.We find evidence of G7 orchestration of intermediary actors when removing G7 flows ( Supplementary Appendix D ).We model a bilateral-only dependent variable, removing multilateral aid from our spending commitment measure.The results suggest that bilateral donors may respond more quickly to G7 orchestration through some modalities (multi-bi aid) rather than others (direct bilateral aid), and evidence of orchestration of non-G7 bilateral direct aid persists with a single lag ( Supplementary Appendix D ).We report models that drop lagged dependent variables from models in Tables 2 and 4 to ensure Nickell Bias is not responsible for re-  5 G7Commit lead/lag placebo tests sults in models with lagged dependent variables and unit effects ( Supplementary Appendix E ).We replace year effects with other controls that apply across sectors each year ( Supplementary Appendix F ).This appendix includes models controlling for the year-to-year change in the total number of G7 commitments to account for general reductions and expansions in G7 agendas.It also includes models controlling for the year-to-year change in less-specific "developing country" G7 communiqué commitments coded by the G7RG, checking that general mentions of developing countries in communiqués are not responsible for the associations between commitments and aid allocations.Findings are robust across these alternative specifications.
We then test alternative dependent and explanatory variables to check that the results are robust to using percentage measures of spending, trust fund establishment, and commitments.These measures present statistical problems but allow us to check that the evidence found above is not dependent on a specific transformation of variables.See Supplementary Appendix G for more details and results.
Lastly, we account for rhetoric across IOs using the Intergovernmental Policy Output Dataset (IPOD) ( Lundgren et al. 2023 ).IPOD codes 37,000 IO policy pronouncements by policy area, many of which link in some intuitive way to the G7 commitment policy areas used in this study.We map each G7 category to its closest IPOD category, then control for the number of outputs in that IPOD category that year.This estimation helps isolate G7 commitments from other IO pronouncements, providing a strong control for unobservable "context" insofar as IPOD variables are a proxy for the general salience of a policy area across the international community each year.Not only do G7 commitment results persist, but IPOD variables are negatively associated with aid spending and trust fund establishment. 12See Supplementary Appendix H for details and results.
12 Because inclusion of IPOD variables significantly reduces sample size, we present related models only as a robustness test.

Limitations and Instrumentation Strategies
There are limitations to this empirical strategy, which stem from the challenge of identifying causal evidence of informal governance effects using observational data.The placebo tests provide evidence that spending is not merely a sticky function of fads in global development priorities.But placebos, of course, do not eliminate the possibility that the relationship between G7 commitments and spending or trust funds is a function of common omitted factors.Instrumentation would be necessary for establishing causality.We cannot identify any plausible instruments for models of trust fund establishment.However, we use a potentially excludable instrument based on the rotational nature of summit hosts, who have control over the final wording of annual communiqués ( Tallberg 2010 ), to test the robustness of the aid commitment models to using a 2SLS estimator.These models provide further evidence supporting H1 (further discussion and results in Supplementary Appendix I ).Together, much empirical evidence suggests that the G7 influences international development through coordination and orchestration.

Conclusion
This study theorizes that the G7 has an informal governance role in aid.It tests whether the G7's annual summit communiqués help coordinate and orchestrate donors, serving as an agenda-setting tool through which the G7 can inform international development efforts.Empirical tests provide evidence consistent with this two-pronged informal G7 role in international development.When the G7 prioritizes a policy area, multilateral and bilateral donors direct more cash and trust fund resources toward those priorities.
To our knowledge, we provide some of the first large-N evidence of the influence of informal organizations, considering their role as coordinators and orchestrators in global governance.A quantitative approach to studying informal governance has limits, in that causality is inherently difficult to establish in a statistical study of informal governance ef-fects.Unobservable contextual factors loom as a major limitation.Yet the preponderance of statistical evidence supports anecdotal reports that the G7 serves an informal governance role in international development.While more research is needed, this study has implications for the international political economy of development and the study of informal organizations in global governance.
On the former, the study contributes to our understanding of donor behaviors in international development.Informal groupings like the G7 can coordinate their own members' activities and orchestrate those of non-member intermediaries.G7 coordination and orchestration may or may not give rise to developmental impact, and we are agnostic on this point.As discussed earlier, a G7 governance role in international development need not be effective, as overlap and herding may be as likely to emerge as efficiency gains.While this study does not probe whether informal G7 governance in international development is effective on the ground over time, it does identify evidence of its presence.In addition, identifying G7 orchestration in international development gives rise to questions about its reach.For example, some major sources of official finance strategically ignore some aspects the Western development agenda ( Cormier 2023 ).Unpacking the conditions under which actors are more or less influenced by G7 and other informal orchestration efforts would advance not only research on international development but also international cooperation more generally.
On the latter, the study provides an example of a specific policy area (international development) where informal organizations can influence a plurality of actors and institutions, and thus global governance.By providing ideational and material inducements through its communiqué, the G7 informs international development policy despite the lack of formal staff, permanent resources, expertise, or legal standing.The G7 is not the only informal organization that seeks to coordinate and orchestrate policies ( Roger 2023 ), and informal intergovernmental organizations and transnational public-private partnerships are proliferating ( Andonova, Hale, and Roger 2017 ;Reinsberg and Westerwinter 2021 ;Vabulas and Snidal 2021 ;Roger 2023 ).This raises questions about which institutions are able to produce outputs that match the ever-increasing demand for global problem-solving and which ones get to determine the agenda in an increasingly crowded governance space.To address such questions, informal organizations' communication tools, like G7 communiqués, may provide useful material for further investigation of the politics and impacts of informality across different areas of global governance.

Table 1 .
The 30 policy areas in the G7RG Communiqué dataset

Table 2 .
Estimating the G7 relationship to aid spending

Table 4 .
Estimating the G7 relationship to trust funds.