Pathways towards power shifts: State-society synergy

Policy reformers often make bold promises to improve government responsiveness to citizen demands. Yet such proclaimed openings from above often fall short, get diverted, or are blocked. This study uses the state-society synergy approach to analyze exceptional cases when reformers within the state managed to deliver openings for citizen action that tangibly empowered otherwise excluded or marginalized groups. What happens when these reform strategies are attempted? We used process tracing, combined with qualitative comparative analysis, to identify patterns across 19 cases in the global South where state actors created a more enabling environment for citizens ’ collective action. The study compares the triggers and scope of enabling state actions, the breadth and intensity of collective action, roadblocks within the state, and whether or not these interactive processes led to substantive power shifts in favor of the excluded. We find that half of these openings led to shifts towards greater power for either citizens or reformist actors within public institutions, in spite of both structural obstacles and governmental roadblocks. Notably, power shifts occurred where reformers ’ initiatives to enable collective action were themselves most intensive (often but not always backed by political change). Windows of opportunity were often open only briefly, until reformers lost power, and the pathways that led to power shifts combined collaborative and adversarial relationships. The power shifts identified were all incremental and uneven, and many were limited to subnational arenas. Though some later stalled or were partly rolled back, from the point of view of socially and politically excluded groups they represented tangible improvements in the balance of power. While tangible openings from above are rare and conventional theory would expect little institutional change, the state-society synergy framework shows how state actions to reduce the risks or costs of collective action can enable pathways to power shifts.


Introduction
Both national and international policy reformers often make bold promises to improve government responsiveness to citizen voice.Time and again, such proclaimed openings from above fall short, get diverted or blockedas most theories of bureaucracy, political parties, collective action and protest would lead one to expect.Yet in exceptional cases reformers within the state manage to deliver tangible openings for citizen action, which sometimes do lead to power shifts for the socially excluded.Such 'cracks in the system' may look small from afar -but may loom large when seen from below, in context.Lessons from these outlier cases are relevant for informing both more nuanced theories and practical reform strategies.This study asks: how do openings from above enable the socially excluded to engage in collective action?When does such collective action in turn manage to lead to power shifts?
This study seeks to identify patterns of mutually-reinforcing interaction between reformist government actors and socially excluded citizens with an analytical framework inspired by a least likely case from Mexico.In 1979, under Mexico's then-authoritarian regime, officials concerned with poverty alleviation launched a national network of thousands of village food stores.Unexpectedly, this program convened participatory citizen councils to co-manage and oversee food delivery, the first-ever free spaces for autonomous, regional-level community organizing under a harsh regime of boss rule.For some officials leading this opening from above, citizen oversight through these councils was instrumental to block leakage and elite capture by vested interests, while other officials had the more ambitious goal of encouraging participatory development.One third of these 300 regional oversight councils became an autonomous social movement, leading to both bureaucratic backlash and spillover effects that bolstered a web of smallholder producer organizations.Just over a decade later, the councils' advocacy networks in congress retained enough clout to block a Finance Ministry proposal to

Analytical frameworks
Openings from above that lead to inclusionary power shifts do not fit well with conventional social science theories of institutional change.These theories focus primarily on either the state or the social actors involved.Some frameworks focus on the interests and organization of the state and/or political systems to explain institutional change, while others emphasize the influence of economic forces or social actors on the state.For example, political scientists focus on laws, official policies and governance institutions while social movement theorists across disciplines focus on extra-institutional protestwhich does indeed drive most power shifts that consolidate rights.Yet not all interactions between collective action and government reform fit the classic 'pressureresponse' model.
While state-or society-centric theories of change recognize actors on the 'other' side, each tends to treat those actors as external to their frameworks.Indeed, the governance literature acknowledges that protest or advocacy can motivate policy change but does not focus on interactive dynamics that enable reformers both in state and society, while the social movement literature stresses how perceived political opportunities can trigger cycles of protest.Yet openings from above can also enable cycles of other forms of collective actionas many of the cases here show. 4 Historical studies of the drivers of effective, inclusionary governance stress long-term, virtuous circles of state-society interactionthough they do not spell out the causal mechanisms (e.g., Mungiu-Pippidi & Johnston, 2017;Putnam, 1993).Analysts of governance reforms also find positive feedback loops that can drive institutional change in the shorter term (e.g., Rothstein, 2011).This study addresses a specific genre of openings from above that may enable such virtuous circles.While these windows of opportunity for reform have great potential, at the same time they may close quickly after elections, be bounded to specific sectors or subnational territories, and may unfold below the radar of national politics and policies.Such openings face daunting obstacles, appear to be rare, and are not well-understood.To ground these questions in a broader analytical framework, the state-society synergy approach offers an alternative lens to the conventional bifurcated analysis of state and social actors.Frameworks, in contrast to theories or models, identify relevant pieces of a puzzle in terms of key concepts that can inform further researchincluding multiple possible configurations (Emerson & Nabatchi, 2015).
The state-society synergy framework goes beyond the identification of complementarities and co-production between state, society and market to focus on embeddedness: relationships that bridge the statesociety divide. 5Because of this study's concern with power shifts, it extends past the concept of embeddedness to recognize more explicitly that collaborative, pro-reform state-society coalitions may face competing state-society coalitions that seek to block inclusionary institutional change.While such cross-sectoral ties can enable rent-seeking, they can also contribute to more effective development institutionsoften at the sectoral and or subnational level (e.g., Tendler, 1997).The state-society synergy approach seeks to identify where and how agents of change in state and society mutually enable one another, sometimes with a specific focus on the empowerment of the socially and politically excluded (e.g., Fox, 1996).Viewed in the context of the longstanding theoretical discussion of structure and agency, the goal here is to identify patterns of reciprocal interaction between actors within both the state and society (agency) that can lead to at least incremental power lective action distinguishes such cases from authoritarian "state-mobilized movements" (Ekiert, Perry, & Xiaojun, 2020).Diverse state actors combine carrots and sticks to prod citizens to mobilize in order to "rule by other means."This study's case selection excludes coercive approaches to induce mobilization because its goal is to identify pathways towards pro-citizen power shifts.
3 Historic cases of mutual empowerment between state reformers and mobilized constituencies that drove large scale structural reforms include Mexico's peak period of agrarian reform in the 1930s, Kerala's iterative series of social democratic reforms through the last third of the 20th century, Chile's brief radical reform government (1970)(1971)(1972)(1973), and the period of most extensive agrarian reform in the Philippines in the 1990s (Borras, 1999;Fox, 1992;Hamilton, 1982;Heller, 1996;Heller, Harilal, & Chaudhuri, 2007).
4 Meanwhile, analysts of non-governmental organizations are increasingly recognizing blurred boundaries between state and society (Brass, 2016). 5For foundational work on state-society synergy, see Evans (1996), Houtzager andMoore (2003), Joshi and Moore (2004), Migdal (2001), Migdal, Kohli andShue (1994), andOstrom (1996), among others.The focus of the related "polity" approach is on "how societal and state actors are constituted, how they develop a differential capacity to act and form alliances, and how they cooperate and compete across the public-private divide" (Houtzager, 2003: 2).Studies of ultra-local governance institutions find they "straddle" state and society (Read & Pekkanen, 2009).Applied policy analysis also addresses exclusively collaborative approaches to state-society synergy, though without focusing on power shifts for the excluded.See, for example, Guerzovich and Poli (2020) on collaborative social accountability, Ansell and Gash (2007) and Emerson and Nabatchi (2015) on collaborative governance, and Sidel and Faustino (2019) on "coalitions for change" among policy insiders in both state and society.
shifts (alteration of structural constraints). 6 The focus here is on one specific genre of state-society synergythe set of openings from above that deploy tangible actions to enable mobilization from below.This framework involves a non-zero sum understanding of power shifts.That is, the political or bureaucratic initiatives in this set of cases do not involve elites 'giving up' power to the socially excluded.In contrast, these elites seek to bolster potential external allies in order to increase their own capacity in relation to the rest of the state.In other words, the analysis seeks to identify cases where change agents within state and society were mutually empowering.This framework takes into account both collaboration and conflict across the state-society dividein contrast to literatures that address one or the other. 7 This genre of enabling openings from above can be called "sandwich strategies" (Fox, 1992(Fox, , 2015))."Strategy" emphasizes the role of agency from within the state, in contrast to a model based purely on external pressure leading to government response.The "sandwich" metaphoras both noun and verbunderscores the process of convergence between actions both from above and below that can contain or weaken vested interests.For these reasons, that metaphor originally framed this study.Yet the comparative analysis revealed two major limitations of the metaphor.First, the obstacles to change may not be located in the "middle."While resistance to inclusionary reform in some of the cases was indeed at meso levels, in other cases the obstacles were lateral forces, such as rival agencies and/or vested interests outside the state.Second, when international actors are also involved, they introduce a multi-level dimension that eludes a simple sandwich. 8 Enabling actions by actors within the state can be found in both democratic and undemocratic regimes, as well as in both strong and institutionally fragile settings (as some cases addressed here show). 9One of the most relevant literatures for understanding state-society synergy strategies focuses on the origins and institutional design of national participatory institutions in relatively high-capacity states governed by democratic regimes. 10These institutions of "empowered participatory governance" involve semi-autonomous state-society interfaces (Fung & Wright, 2003).These official channels for public engagement are also known as "invited spaces," though in many contexts such official channels for public engagement are weak, socially exclusionary or politically confined (e.g., Cornwall & Coehlo, 2007).In practice, national participatory institutions also vary widely across subnational territories and sectors.The more dynamic variants can be considered "state-sponsored activism" (Rich 2019).Moreover, the character of invited spaces can also be contested and change over time.In other words, sometimes invited spaces created from above can be claimed and transformed from belowbut when, and how?
Particularly relevant to this study is the concept of political opportunities, from political process theory within the social movement literature.This concept emphasizes that changes in the political environment can enable collective action, such as shifts in the broad degree of political openness, the availability of influential allies, divisions among elites, or the threat of reprisals (Tarrow, 1994: 77-80). 11Yet using the concept of political opportunity to explain institutional change can be so broad as to come close to the generic "context matters."In response, scholars call for the identification of how specific dimensions of context are directly relevant to specific actors (Meyer & Minkoff 2004: 1464).This study contributes to this literature on political opportunities by honing in on one specific element of context with direct implications for social action: tangible actions by reformers that reduce the risk or cost of collective action for specific social actors.
This study focuses on what happens when reformists deliver openings from above that encourage mobilization from below.Both conventional state-or-society frameworks and the state-society synergy perspective would expect that such initiatives would be rare, and that even those outlier efforts would often be blocked or diverted.Research strategies that seek to identify outlier cases are relevant here.The literature on policy reform and public services, especially in low-income settings, has identified "pockets of effectiveness" or "islands of integrity" within otherwise low-functioning systems of governance (e.g., McDonnell, 2020;Roll, 2014).Finding these positive outliers requires getting inside the black box of the state, which is a crucial step towards classifying pathways to institutional change (Andrews, 2015;Peiffer & Armytage, 2019;Tendler, 1997).Yet this literature tends to stress institutional insulation rather than embeddedness, to protect insider reformers from capture or diversion.Moreover, the dynamics of diffusion -how such pockets spread and manage to offset obstaclesare still not well understood, especially in more patrimonial or fragile settings. 12 This study seeks to identify and learn from those outliers by comparing 19 diverse cases of state-society synergy initiated by state actor efforts to enable collective action from throughout the global South, drawing on cases that cut across sectors and scale.Across all these cases, state actors went beyond discursive promises of change or new policies only on paper to take measures that tangibly changed the enabling environment for collective action.The specific motives of these state actors varied, but they shared an interest in strengthening their own capacity in relation to the rest of the state by strengthening the voice and action of excluded stakeholders.Although some of these openings from above lasted just a few years, the focus here is on identifying and drawing lessons from their dynamics and the pathways they initiated.The case studies provide the evidence needed to assess whether state-6 Mutually-reinforcing strategic interaction between agents of change in state and society is also key to a classic analysis of regime transitions in the 1980s (O'Donnell, Schmitter, & Whitehead, 1986).A key difference in this study is its focus on power shifts involving the socially excluded.A recent turn in the study of social movements also emphasizes strategic interaction (Jasper et al., 2022), though without addressing the possible relevance of pro-reform initiatives from within the state. 7The literature on participatory governance and social accountability addresses state-society collaboration, while the literature on social movements focuses on conflict.Yet in practice, social actors and strategic policy reformers may draw on both repertoires of action, as many of the cases in this study show. 8To apply the metaphor of a triple-decker sandwich would be a bridge too far, accentuating the sandwich term's Western bias.Yet the idea has been appropriated and transformed into other contexts, notably in the case of the "bibingka strategy" in the Philippines (Borras, 1999(Borras, , 2001). 9For a recent body of research that seeks to identify patterns of empowerment and accountability in fragile and conflict-affected settings, see Anderson et al. (2022). 10These national processes are especially well-documented in Brazil (e.g., Abers & Keck, 2013;Mayka, 2019;Rich, 2019;Wampler, 2015;Wampler & Goldfrank, 2022).This literature on Brazil is also notable for its focus on the mobilization of public servants as advocates of policy reform, particularly in the health and environmental sectors.
11 Social movement theory defines political opportunities for collective action in terms of "changes in opportunities that lower the costs of collective action, reveal potential allies and show where elites and authorities are vulnerable" (Tarrow, 1994: 18).Scholars subsequently called for more precise operationalization of this broad umbrella concept (e.g., Goodwin & Jasper, 2012;Meyer & Minkoff, 2004).The analytical framework here emphasizes and operationalizes just one dimension of "political opportunity" as specific state actions that tangibly enable collective action of the excluded.In contrast to the social movement literature, this agenda for identifying potentially power-shifting repertoires of collective action is not limited exclusively to contentious mass protest. 12The literatures on diffusion reflect the classic theoretical assumptions about state-society boundaries mentioned above.The study of policy diffusion focuses on the replication of state actions, while the social movement literature addresses the horizontal diffusion of protest.Indeed, both can spread in waves.In contrast, the state-society synergy framework would suggest unpacking those waves to identify possible mechanisms of replication of virtuous circles of interaction between pro-reform actors in state and society.
society interaction led to sustained power shifts in favor of the socially and politically excluded, even if those shifts were limited to specific sectors, subnational territories or were eventually contained or rolled back.
The comparative analysis involves analyzing each case through the lens of a set of conditions that, through an interactive process, may follow a pathway of change towards power shift.The definition of these conditions, each comprised of multiple factors, is informed here by the theoretical literatures on state-society relations and on social movements, as well as inductively drawing on the case analysis.These conditions include contextual triggers, the nature of each opening, the character of the collective action in response, and the emergence of roadblocks from within the state and elsewhere in society, which can ultimately lead to power shift (see Figure 1).Each condition is constituted by 4-6 distinct possible dimensions.
The various possible dimensions of the triggers reflect different theories of change.For example, state-centered policy analysis and principal-agent theory suggest that changes in political leadership are necessary conditions for significant action by government agencies.In contrast, social movement theory indicates that pressure from below would be the primary driver of openings from above.One might also expect exogenous shocks would be needed to create the space for significant new initiatives, or that international aid agencies might pressure state actors to create openings from above.Our conceptualization of the dimensions of openings from above, collective action, and roadblocks draws both from a priori understandings of these conditions and from inductive analysis of cases.Finally, we define power shift as occurring when under-represented social actors gained some degree of increased leverage, related to the opening and sustained over at least several years, even if that leverage was subsequently rolled back.This assessment is qualitative, relational and context-specific, which means that power shifts are identified in relation to power relationships that existed prior to openings (rather than meeting the same minimum criteria applied across diverse cases).Power shifts can thus be incremental and limited to specific government programs and/or subnational territories.
In sum, the analytical framework described in Figure 1 identifies pathways of change in which a trigger makes possible an opening from above that in turn enables collective action from below.The opening and/or the collective action may be met with roadblocks.The extent and nature of the opening and the collective action will influence their ability to overcome roadblocks in order to potentially generate power shift.
Building on case-by-case process tracing with qualitative comparative analysis (QCA), we find that the presence of bureaucratic reformers and an intensive opening from above were both necessary conditions for power shifts, almost always also accompanied by intensive collective action.Intensive openings are defined as having broad, tangible reach across people and territory.Intensive collective action is defined similarly in terms of scalebreadth of involvement by the socially excluded.The combination of intensive openings from above triggered by bureaucratic reformers and intensive collective action from below managed to overcome roadblocks.Where reforms were blocked at the national level, sometimes reformers were able to protect subnational enclaves of inclusion in hybrid, state-society institutions.But roadblocks occurred in all cases where power shift did not occur.

Case selection & data
This study's approach to case selection is akin to a positive deviance strategy, initially developed in nutrition studies (Pascale, Sternin, & Sternin, 2010) but then expanded across other sectors, in which researchers try to understand positive outcomes in the face of contrary odds.Explanations of anomalies can shed light on pathways towards change.Our approach differs from positive deviance analysis in that case selection did not rest on a positive outcome, but instead on evidence that state actors took measures to tangibly reduce the risks or costs of collective action. 13Since there is no existing data set that gathers such cases of openings from above, we identified as many cases as possible through an extensive process of literature review and dialogue with experts across the governance and development fields as well as specific sectors (health, education, agrarian reform, etc.), and through social media crowdsourcing.We asked experts if they knew of tangible initiatives by state actors to create space for societal action for proaccountability or institutional change and combed both academic and practitioner literature that addressed state-society interaction.To maximize the number of potential cases, we set few inclusion boundaries beyond the existence of an opening from above.These cases of openings did not have to lead to substantial collective action, could occur at national or subnational levels, could come from any sector under any kind of regime, and could be motivated for any reason. 14Cases were limited to the global South in large part to avoid potential North-South differences explaining observed outcomes.We drew from recent history (the late 20th century to the present) in an effort to reach a comparable threshold of amount of information available on each case.Assembling this set of outlier cases makes visible an under-recognized pathway of inclusionary change across widely varying contexts and is designed to investigate how that particular pathway unfolds (in contrast to assessing it in comparison to other pathways to inclusionary power shifts).
We explored approximately 75 cases, quickly rejecting about a third, and then ruling out another third after extensive investigation based on primary and secondary literature, as well as consultation with experts.Most cases of apparent openings did not involve substantial, tangible actions to reduce the risk or cost of collective action.To maximize diversity of context in the comparative analysis and to prevent the overrepresentation of countries with more extensive track records of institutional and civic innovation (e.g., Brazil, India), the number of cases per country was limited to two. 15This methodological choice involved tradeoffs, constraining the potential size of the pool of known cases in order to emphasize the unexpected presence of this outlier process across diverse contexts (including more fragile and undemocratic settings).
The final set of 19 cases draws from Latin America (8), South Asia (5), East Asia (4), and sub-Saharan Africa (2). 16Table 1 summarizes the key elements of each case, including the opening from above and the outcomes.The state-society interactions in some cases have a much longer duration than othersranging from a year or two to decades.Approximately half of the cases involve national-level government actors, while the other half take place at a state or provincial level.Almost 13 Note that this very bounded case selection criterion does not include key features of the social movement literature's approach to political opportunities, which emphasize discursive promises from sympathetic elites or protestors' perceptions that elites will respond favorably to their claims (e.g., Goodwin, 2012: 292, 299). 14In contrast to studies of social movements that focus on explaining mass protest, this case selection process was driven by the presence of the independent variablethe tangible opening.Openings were coded in terms of whether or not they were considered intensive (a context-dependent assessment).Collective action here is a possible intermediate outcome of varying intensityand can take forms other than protestwith power shifts as the ultimate outcome. 15For countries with more than two cases identified, case selection emphasized those where state action to enable collective action was most robust, where implementation dynamics were most well-documented, and where the scale involved national and/or subnational reformers (not purely municipal level openings).The Nigeria cases were identified in collaboration with our partner organization, the Centre for Democracy and Development, as part of a parallel comparative study of five cases in that country. 16Analysis of African cases included a total of five Nigerian cases that met the criteria for opening, as well as consideration of reform experiences where the scholarly literature ultimately did not identify evidence of tangible measures that reduced the risk/cost of collective action (Tanzania, Ghana, Burkina Faso and South Africa).all cases had some involvement by international donors or aid agencies, although in no cases were donors or aid agencies the primary drivers.This project commissioned 15 cases by academic and practitioner experts, many of whom had already analyzed them extensively.Four other well-documented cases drew exclusively on secondary sources.(See Table 1.).
Dominant state-centric and society-centric accounts in the literature may have hidden some cases.It is also possible that because many enabling reforms were either short-lived or deliberately low-profile, they were not documented.As the universe of cases of such attempts is unknown, the representativeness of this sample of cases is also unknown.The cases analyzed may over-represent those with aid agency involvement because of their greater evaluation resources and visibility to scholars.

Comparative method
The comparative analysis presented below is grounded in the process tracing of the individual cases (Beach & Pedersen, 2013).The cases were then analyzed according to factors relevant to the reform strategy: context; state actor characteristics and actions; social actor characteristics and actions; roadblocks; and outcomes.This informed the case coding in terms of five phases in the analytic framework shown in Figure 1: the proximate trigger, the nature of the opening, patterns of collective action, any roadblocks, and whether power shifts occurred.The co-authors coded the cases independently, revisiting the case evidence and consulting with case authors to resolve differences and to ensure consistency.
The comparative analysis involved two key components.The first involved assessing the relative frequency with which particular dimensions were present within each phase of the process, as well as their degree of intensity (involving both breadth and depth).The second step applied QCA in order to understand the combinations of conditions associated with power shifts. 17 Specifically, the first component of the analysis involved identifying the presence of four to six possible dimensions (defined below) of: each opening's proximate triggers, the nature of each opening, the character of collective action in response, and patterns of roadblocks from within government (either resistance or reformists' loss of power).The key question about the outcome of each process involved an assessment of whether these dynamics led to power shifts of some kind for the socially and politically excluded.Because most of the cases featured donors or aid agencies, the analysis also considered their roles in the different phases of the process.
Following this descriptive classification, we then assessed the intensity of each condition.This judgement of intensity relied on contextspecific interpretations of whether one or more of these dimensions had significant, strong, or extensive breadth and/or depth.Intensive triggers included major political transitions, the presence of bureaucratic reformers recognized by insiders and often even outsiders as entrepreneurial innovators, and/or prior mass mobilization directly targeting the issue at hand.Intensive openings included the provision of substantial government resources to social organizations or large numbers of people, momentous policy or legal changes, and/or granting citizens meaningful oversight of government programs.Intensive collective action included protest, deep citizen participation in resource allocation, and/or citizens exercising meaningful oversight of government programs.Intensive roadblocks threw up substantial barriers to either the opening from above or the collective action, including the loss of power by political and bureaucratic allies, complete clientelist takeover of openings, and/or threats of violence against social actors supported by government in the openings. 18 We conceptualized power shifts as occurring in three distinct arenas.First, within society, the consolidation of representative organizations of the excluded and/or sustained pro-poor entitlements to resources constitute evidence of power shift.Second, within the state, power shift may involve the substantive (not pro forma) implementation of institutional changes that favor the excluded, including policy continuity across changes of government and the enforcement of rights (even if uneven and incomplete).Third, the creation and persistence of hybrid government institutions that bridge state and society by sharing power over authoritative decisions are also evidence of power shift.
In order to confirm pathways leading to power shift (or its absence) across this "medium N" set of cases, we turned to the QCA method.QCA is both an approach and a set of techniques that are especially appropriate for comparative analysis where there is great within-case complexity and where the researcher anticipates "multiple conjunctural causation" (Rihoux & Lobe, 2009).This methodological concept reaches back to John Stuart Mill and addresses "situations in 17 To situate QCA in the broader methodological context of different logics of comparison, see della Porta (2008).This comparison across diverse cases is also informed by "abductive analysis" and its recognition of "intersituational variation" (Tavory & Timmermans, 2014: 78). 18For discussion of different kinds of roadblocks in the context of the one of the casesnotably the difference between active opposition and a refusal to implement laws -see Sultan and Mahpara (2023).which a given outcome may follow from several different combinations of causal conditionsfrom different 'causal recipes'" (Ragin, 2008: 124).The focus is on identifying both necessary and sufficient conditions.Because of QCA's emphasis on interdependence and path dependence, it treats conditions as distinct, but does not assume or require them to be independent of one another.QCA method uses non-linear mathematics to identify pathways of change that involve multiple moving parts, in contrast to statistical methods that seek to isolate the relative weights of different independent variables.In addition to its appropriateness for the size of this sample and anticipation of conjunctural causation, QCA also encourages an iterative analytical process, which includes reassessing coding decisions and even case inclusion based on what is learned from the analysis and closely mirrors our overall approach to the analysis described above.
The pathway framework in Figure 1 informed the definition of the conditions included in the QCA.We used crisp-set techniques, coding conditions as "1{\Prime} if intensive as described above, or 0 otherwise.We chose this approach, over fuzzy-set QCA, because the diversity of cases and outcomes made it difficult to systematically assign each case to multiple levels of any given condition.QCA generates a truth table, which shows all possible combinations of conditions included in the model.A model with four conditions and one outcome like ours has 2 4 , or 16, rows.QCA then uses Boolean algebra to produce solutions, or recipes of conditions necessary and/or sufficient to produce the outcome in question. 19The QCA method's focus on "causal complexity," with its emphasis on identifying both necessary and sufficient conditions, is designed to recognize mutually-reinforcing interaction between conditions.
Like all models, including standard frequentist models (e.g., regression), with QCA there is always a risk of omitting a condition important to explaining the outcome, which is most significant if that condition has separate influences on two other conditions (endogeneity).We have taken the best available route to protect against such risks in any

Pakistan
A coalition between an autonomous governmental national commission & a CSO led to partial reform of anti-women laws, but lack of government support ultimately weakened the commission.

National Commission on Status of Women Peru (Puno)
A district-level health services monitoring partnership between the government ombuds agency & indigenous health rights defenders led to a national commitment to expand monitoring, but electoral change prevented implementation.

Triggers
Triggers refer here to contextual events or processes that directly enabled the opening from above. 20The presence of bureaucratic reforms was the most widespread dimension, triggering openings in 16 of the 19 cases.Political transitions preceded 13 openings and usually involved the election or assumption of office by government leader(s) promising reform, as in the case of campaign promises in Sri Lanka or a major legal change, such as Colombia's new constitution, which made possible collective ethnic land titling.Notably, three cases of bureaucratic initiative occurred even in the absence of political change, with each followed by intensive openings.Indeed, in some cases reformist officials tried to stay away from politics, choosing to remain low profile to avoid provoking backlash (as with Peru's provincial level public defender's office and in Mexico's national community food councils).Even in the many cases where political transitions occurred, bureaucratic reformers were not merely subordinate agents.These innovators exercised their own agency, with motivations including professional norms-as in the cases of Brazilian health officials or Sri Lankan housing planners-or ideology, as in the cases of feminist officials in Pakistan and Bangladesh, insider advocates of participation in Mexico, and Afrodescendant rights defenders making their way up within the ranks of the Colombian state.International aid agencies (multilateral, bilateral and private) often provided diplomatic and financial support to government reform efforts before the opening -brokering coalitions rather than pressuring with conditionality.Aid agency roles that went beyond supporting ongoing in-country opening initiatives were rare (in contrast to their widelyrecognized influence on other kinds of policies). 21 Another possible explanation for triggering openings involves political leaders' specifically electoral motivations.Surprisingly, the process tracing found only two cases where electoral motivations directly accounted for enabling actionsin Bolivia's decentralization reform and the Sri Lankan Prime Minister's ambitious campaign promise to build a million houses.He then turned to an innovative urban planner who launched participatory sites and services provision to reach scale.In Brazil's rural education case, electoral change was necessary but not sufficient to account for enabling actions, insofar as post-electoral protest was needed to push the government to put a campaign promise into action.
In contrast to what classic social movement theory would predict, in only three cases was protest a clear proximate cause of the opening from above: Brazil's rural landless movement, Brazil's HIV advocacy movement, and the second People Power civic mobilization in the Philippines.Nine other cases involved modest degrees of preexisting mobilization of constituencies, likely contributing indirectly to the opening, but without being a driving force.Also contrary to expectations, just three other openings were iterations of previous cycles of reform, including the cases of community forestry in Mexico, India's official social audits, and participatory housing policy in Sri Lanka.Moreover, few openings were triggered by shocks -whether crises, disasters or perceived direct threats to specific social groups (as in the two China cases).Instead, the presence of bureaucratic reformers was the most frequent enabling condition present.

Openings
As defined here, the openings from above involved actions by state actors that tangibly reduced the risks or costs of collective action, especially for the socially excluded.The most intensive openings involved both breadthwith clear reach in terms of people and territory -and depth, meaning more substantive tangible measures by state actors.
In 13 cases, government agencies directly contributed to grassroots collective action with tangible resources, including field organizers, direct support for membership organizations, training of grassroots leadership or relevant service provision.For example, in five cases government agencies directly formed or reinforced representative mass membership organizations of the excluded (e.g., HIV advocacy organizations in Brazil, Afro-Colombian community land councils, Colombia's National Association of Peasant Service Users, Mexico's community forestry enterprises and Mexico's food councils). 22In Indian states, hybrid state-society agencies fielded frontline organizers who convened safe spaces for public hearings where the excluded could give testimony about government performance (activist community health workers in one state and social auditors in two other states).In the Philippines, the national social welfare department deployed field organizers nationwide to convene village leaders to lead participatory project decisionmaking and provided training for mass membership civic organizations to oversee textbook distribution to the last mile.In Bangladesh, financial support to CSOs enabled them to broaden access to stakeholder consultations around the country.In Brazil, the education ministry provided intensive trainings to rural schoolteachers and convened local and state officials to encourage co-governance of schools with organized agrarian reform communities.In the cases where the opening from above involved tangible resources that reached the grassroots, national or state agencies enabled either the consolidation of mass organizations or the creation of safe public spaces for citizen voice.
In 12 cases, reforms either promoted co-governance of programs or recognized citizens' right to oversight via participatory policy monitoring.Official participatory programs shared targeted information about the services that agencies were supposed to provide, which allowed social actors to make informed claims and hold specific government actors accountable for their commitments.In Bolivia, rural municipalities co-governed with "territorial base organizations" and public oversight committees.In Brazilian states where the landless movement was strong enough and politicians agreed, their organizations co-governed rural schools in agrarian reform settlements.Hybrid state-society regional councils in Colombia managed, monitored and enabled a wave of collective titling of ethnic territories.In Mexico, the food distribution agency created the first nationwide social accountability program, allowing regional councils of community representatives to oversee staple food deliveries from warehouses to village stores.Reform-minded federal and state policymakers in Nigeria gave official permission to donor-funded CSOs to do independent monitoring of national social programs.In Pakistan, the government created a commission to monitor and promote compliance with international and constitutional women's rights commitments.A regional office of the 20 Indeed, the metaphor of "trigger" has the disadvantage here of possibly implying a determinative rather than an enabling role. 21An example of a case where donors were present but did not trigger the opening, consider Bolivia's Popular Participation Law.Aid officials were focused on other simultaneous policy reforms.Donors such as USAID paid salaries of government consultants, yet the promulgation of the law caught major donors by surprise, as former World Bank country staff have reported (Faguet, 2012: 6-7).For comparative analysis of five outlier cases where an aid agency played more determinative roles in driving openings from above, see Fox (2020). 22On the contributions of mass membership organizations of the poor to development, see Chen, Jhabvala, Kanbur and Richards (2007) and Esman and Uphoff (1984), among others.
Peruvian public ombuds agency provided community health rights defenders with credentials and legal support to validate their right to oversee public clinics and advocate for patients.The education ministry in the Philippines recognized and supported CSO oversight of procurement of textbooks and trained large national civic organizations to be field monitors of textbook distribution to the districts.
In approximately half of cases (10), government created or recognized spaces for active consultation with CSOs.These consultative bodies operated at different levels in different contexts, with some more focused on policy implementation and others more involved in policy/ law design.For example, in Bangladesh senior government reformers convened CSOs to design and advocate for domestic violence legislative proposals.Brazil's pioneering HIV policy created numerous hybrid consultative bodies.Colombia's brief but large-scale wave of ethnic collective land titling included regional multistakeholder councils to encourage policy implementation and address conflicting land claims.In Mexico, government consultations with community forestry organizations informed legal reforms.Nigerian state officials engaged with CSOs to listen to their findings from monitoring social programs.In Pakistan, the government created an autonomous commission to mediate its relationship with civil society on women's rights.
In 13 cases, changes in laws or policies established new rights or entitlements that enabled social action.These changes varied widely, both in terms of their degree of formal institutionalization and in terms of the degree to which official commitments were carried out in practice.Laws mandated openings for public action in six cases.Bolivia's Popular Participation law recognized and empowered territorial base organizations and municipal oversight committees.Brazil's HIV law guaranteed AIDS treatment, which legitimated claims.Implementing legislation for Colombia's new constitution enabled Afrodescendant land councils to solicit collective titles to ancestral domain.India's Rural Employment Guarantee law mandated state governments to carry out social audits.Mexico's forest law recognized and supported self-managed peasant and indigenous community forest enterprises.The Philippines' law on public procurement enabled CSOs to monitor bidding and contracting for textbook provision.In seven other cases, national policies or programs created openings.Brazil's rural education policyitself a response to movement advocacycreated spaces for collaborative implementation.In the community-driven development program in the Philippines, municipal forums of elected village leaders selected projects.In Sri Lanka, then-innovative housing policy allowed urban and rural community development councils to co-manage local planning, including entitlements to self-built improvements and recognition of addresses.
To sum up, openings from above were diverse, varying in their intensity and institutionalization.The most intense openings included multiple dimensions and were likely to include resources targeted towards the support of frontline field staff and mass membership organizations, in contrast to weaker openings limited to consultations.

Collective action
The most widespread response to openings involved the reinforcement of social organizations, especially those with grassroots membership (16 cases).The decentralization law in Bolivia gave grassroots territorial organizations standing to engage with local governments.Resources from the Brazilian National AIDS Program sustained existing civic associations, helped create new ones and enabled a national alliance.In Colombia, government support launched 300 municipal level peasant organizations, which formed a nationwide organization.The state government's community health workers in Chhattisgarh, India monitored health programs, supported other rights struggles, and tried to bargain collectively.Forest policy provided resources and technical support for agrarian communities in Mexico to form self-managed timber enterprises.Also in Mexico, the village food store program formed local and regional oversight committees, some of which gained autonomy.Pakistan's national commission brought together women's organizations in a national conference and provided cover that allowed them to petition, protest, and advocate.The community-driven development program in the Philippines encouraged formation of new grassroots rural social organizations.Sri Lanka's participatory housing policy strengthened urban community councils and associations of village organizations.
Many of the social organizations supported by the opening from above also monitored policy implementation, making it one of the most widespread dimensions of collective action (14 cases).In Bangladesh, the coalition against gender violence monitored implementation of the anti-domestic violence law.In Bolivia, oversight committees created under the decentralization law monitored municipal governments.Brazil's AIDS movement monitored policy implementation by state and municipal health services.Community health workers in India were active in official local committees to monitor government social programs.In India's states, citizens participated in thousands of public hearings to debate the findings of government-supported, independent validation of social program implementation.In Mexico's rural food program, regional councils met in the warehouses to monitor operations and deliveries to village stores.In Nigeria, school management committees and parent-teacher associations as well as CSOs organized to monitor the school feeding program.In the Philippines, a broad national civic coalition monitored the textbook supply chain and reached 80% of school districts at peak.
Collective action involving civil society policy advocacy also occurred quite frequentlyreferring to campaigns that addressed government policy rather than one-off problems, targeting different levels of government (13 cases).In Bangladesh, gender violence campaigners advocated for legal reforms, including broadening the legal definition of the household.In Brazil, the AIDS movement called for funding for medications, organized legislative caucuses and engaged in participatory policy consultation bodies.Also in Brazil, the landless movement advocated for a national rural education policy and engaged with state governments to carry it out.In China, the national leader of the disabled people's federation advocated to allow continued use of three-wheeled vehicles by disabled drivers, which had been banned.In Colombia, the national peasant organization advocated for stronger land reform laws, first within the system and then from the opposition after the government changed.In Mexico, community councils mobilized congressional allies to defend the national rural food store program from threatened elimination.In the Nigerian state of Ogun, CSOs engaged with state officials in multistakeholder meetings, calling for increased portions for meager school lunches.In Pakistan, the national commission worked with CSOs to advocate for reforms to gender-biased laws.In Peru, the national health rights coalition allied with community defenders to advocate for health ministry guidelines to recognize citizen monitoring nationwide.In the Philippines, the textbook monitoring coalition advocated for dedicated budgets to fund last-mile delivery to schools.
Protest was less common across the cases and was concentrated in a handful of countries (Brazil, China, Colombia, Mexico, and Pakistan).In Brazil, after an allied political party won the presidency, the landless movement organized a large march to push the government to begin implementation of the rural education policy.In China, once insiders leaked information about government plans to build a toxic industrial plant, a sizeable citywide street protest challenged the threat.Also in China, disabled drivers protested government efforts to ban their vehicles, including direct action and 'hidden resistance.'After a change in the Colombian government ended hope for land reform, one wing of the new national peasant association radicalized and led a wave of militant land invasions.In Mexico, when regional food council concerns about poor service delivery went unheeded, council members carried out warehouse takeovers.In Pakistan, a women's rights coalition petitioned and protested against traditional men-only councils and impunity for sexual violence.Most of these cases of protest were enabled, directly or indirectly, by insider support from government officials (albeit low profile and often transitory).
Direct CSO or citizen participation in resource allocation was relatively infrequent, found in only four cases.In Bolivia's decentralization, territorial social organizations gained the right to participate in allocating municipal funds.In Brazil, federal backing encouraged the state governments most open to collaboration with the landless movement to share decision-making over the siting of new schools.In the Philippines, elected village leaders met in municipal forums to choose among local project proposals.And in Sri Lanka, government-backed community associations allocated loans for housing improvements, coordinated community infrastructure investments and oversaw community contractors building local public works.
In sum, patterns of collective action primarily involved the formation or reinforcement of social organizations, coalitions or spinoffs, CSO or citizen monitoring of policy implementation, or policy advocacy.

Roadblocks
Roadblocks were intensive in 12 cases and almost all (18) of the cases encountered some degree of resistance.Sometimes these responses limited the opening in the first place and other times they attempted to block the collective action that followed the opening.The most common form of roadblock involved active resistance by other actors within government to carrying out enabling reforms (13 cases).
In Bangladesh, law ministry officials, parliamentarians and Islamist groups tried to block the passage of anti-domestic violence laws, and then officials limited implementation.Conservative religious legislators in Brazil resisted reform to support a stigmatized disease (HIV), the health minister resisted increasing the national budget for HIV medication, and subnational politicians/bureaucrats failed to support the policy.Also in Brazil, officials in half the states -where the landless movement lacked clout -ignored the federal government's rural education policy reform.In China, national economic growth policies incentivized local governments to overlook concerns about polluting industries raised by citizens, scientists and environmental policymakers.In India, even where state government agencies effectively carried out social audits, anti-corruption agencies took few followup actions in response to findings of malfeasance.Even though Mexico's government food distribution agency created the food councils, much of the agency staff resisted their autonomy, rejected their claims and purged proparticipation field staff.Suspicion among education officials, food vendors and head teachers in Ogun State, Nigeria significantly slowed rollout of monitoring of the school feeding program.Also in Nigeria, federal cabinet ministers resisted third party monitoring of anti-poverty programs.In Pakistan, the national government resisted funding its own commission on the status of women.Local Peruvian health professionals initially resisted citizen monitoring of service delivery, out of concern that citizens lacked medical training.
Another common roadblock to both openings from above and ensuing collective action occurred when pro-reform policymakers lost power (12 cases), often because of electoral change (seven cases).In Brazil, national elections displaced federal allies of the rural education policythough it survived in two states.In China, the main national environmental ministry ally of the anti-toxics movements lost power to senior rivals, reducing responsiveness to citizen concerns.In Colombia, a national election brought in pro-landowner elites who opposed agrarian reform and the national peasant organization.Later on, another election brought in conservative elites, ending most new Afro-Colombian community land titling.In Mexico, after the food program's first decade and a half, new policymakers weakened government support for food councils.In Nigeria, national elections changed the balance of power within the presidency, leading to the reassignment of social programs to another ministry and the end of government support for CSO-led monitoring.After the same election, change in the party controlling the Ogun state government led to loss of a key state champion and turnover in local education officials, thus limiting the capacity of parentteacher associations and school management committees to monitor the school feeding program.In Pakistan, lack of governmental support for its commission on the status of women blocked the naming of a new director.In Peru, electoral change ended incipient health ministry support for citizen monitoring.In the Philippines, a high-level corruption scandal led senior allies in the education ministry to leave the government, blocking the CSO textbook monitoring campaign.In Sri Lanka, after electoral change the new government moved housing policy away from a sites and services approach and ended support for community councils.
Partisan or clientelistic political intervention diverted elements of reform efforts to enable collective action in almost half of cases (nine).In some Bolivian municipalities, radical decentralization created openings for partisan clientelism.Local political operatives, rather than autonomous social movements, led the collective land titling processes in some Colombian communities.In some regions of rural Mexico, government food agency officials and local elites controlled the oversight councils.A newly-elected government in Nigeria's Ogun State purged the school cooks to create patronage opportunities, disrupting the school feeding program monitored by CSOs.In some areas of the Philippines, local elites captured the community-driven development program.And local elected officials influenced the rural councils managing Sri Lanka's housing program.
Finally, in a relatively small number of cases (five), roadblocks took the form of threats or acts of violence enabled or tolerated by government actors.In China, disabled tricycle drivers who defied or protested the ban faced police violence.In Colombia's ethnic territories, attacks by paramilitaries backed by the government displaced newly-titled communities.In Pakistan, the head of the national commission on the status of women received threats because of her work.
In response to roadblocks, reformists frequently had to carry out direct advocacy with other agencies or levels of government (12 cases).For example, in Brazil, the director of the HIV program advocated for provision of medication with the skeptical health minister.In China, a member of the political elite who also led the federation of disabled persons lobbied local officials and local branches of the federation to allow disabled people to preserve livelihoods and to defend those who protested.In the Indian state that led the process of launching a hybrid social audit agency to limit corruption in the rural employment program, the chief minister told local politicos to seek rents elsewhere, and to leave the rural employment program alone.In Nigeria, the vice president had to convince a skeptical cabinet to allow CSO monitoring of social programs for the first time.
Overall, roadblocks came primarily from vested interests within government, but also from religious, political, or economic elites who felt they stood to lose from reforms.Political and bureaucratic transitions also closed windows of opportunity for enabling collective action.

Power shifts
Twelve of the 19 cases of the enabling strategy led to substantial relative power shifts.Power shifts unfolded in multiple arenassome more within society, others more within the state.Table 2 synthesizes the nature of power shifts in each case.Power shifts took the form of greater capacity of state institutions to include and respond to the socially excluded in some cases, more consolidated social organizations and capacity for representation in others, or both.Power shifts in all but one of the cases involved a strengthening of voice and entitlements, either through the consolidation and recognition of representative social organizations and networks, or through sustained pro-poor access to resources.In 10 of the 12 cases, power shifted as reforms were implemented and institutionalized, either through new laws or policies that lasted, or through the enforcement, at least to some degree, of rights promised by new laws or policies.Eight of these legacies of reforms were embedded in some kind of state-society power-sharing institution that consolidated and survived changes in national political context.
The power shifts identified were all incremental and uneven, and many were limited to subnational arenas.Though some later stalled or were partly rolled back, from the point of view of socially and politically excluded groups they represented tangible improvements in the balance of power.As Table 2 shows, power shift was multi-dimensional in all but one case, involving changes to both elements of the state-society dynamic.
To sum up patterns of power shifts, they unfolded across a wide range of contexts and sectors.They almost all began with the presence or empowerment of bureaucratic reformers, often (but not always) put in place through political transition.Most often, collective action in response involved the consolidation of broad-based, representative social organizationssuch as Afro-Colombian community land councilsas well as to institutional channels for the excluded to exercise voice, as with social audits' public hearings in India.Not surprisingly, the openings from above that did not lead to power shifts also tended to be relatively brief.Often, elected and politically appointed reformers had to spend much of their first term in office just to begin to open the window a crack (as in Nigerian cases), leaving little time for openings to generate collective action.That said, some of the openings did manage to sink institutional roots and survive for more than a decade, in spite of changes in the party in powerespecially where they were grounded in large subnational governments (as in India and Brazil).The QCA analysis that follows confirms the combinations of conditions associated with power shift, and lack thereof.

QCA results confirm a primary pathway to power shift
The QCA model directly reflects the pathway framework outlined in Figure 1, incorporating as conditions each of the four main phases of the process: trigger, opening from above, collective action, and roadblock.Table 3 presents the case by case assessments that went into the truth table analysis for the QCA.Because all of the cases involved an intensive trigger of some kind, this step in the QCA analysis focused on the two dimensions of the trigger most theoretically relevant to our framework, the occurrence of political transition and the presence of a bureaucratic reformer.These two dimensions were also the most frequent. 23All other conditions were coded as intensive or not.
The QCA analysis finds two necessary conditions for power shift: the presence of a bureaucratic reformer, and an intensive opening.Furthermore, all of the cases with intensive collective action involved intensive openingsand all led to power shifts.In only two cases did power shift occur without intensive collective action.Furthermore, in five cases, intensive openings from aboveusually together with intensive collective actionled to power shifts in spite of substantial roadblocks, suggesting that collaborative efforts of state and society actors can overcome roadblocks.In contrast, three cases without power shift lacked a bureaucratic reformer, and all cases without power shift lacked an intensive opening and intensive collective action.All cases without power shift also featured substantial roadblocks.
The results of the formal QCA analysis in Table 4 both confirm the conclusions drawn from Table 3 as well as augment them.The results show two pathways to power shift.The first, which covers the majority Table 2 Types of power shifts observed. of cases with power shift (10), shows mathematically that a bureaucratic reformer, an intensive opening, and intensive collective action lead to power shift, regardless of any roadblocks that occurred.The second, less prominent pathway, which covers slightly less than half of cases with power shift (5), includes political transition, a bureaucratic reformer, an intensive opening, and no intensive roadblock and leads to power shift.This second solution indicates the importance of political transition to triggering the state-society synergy process.It also suggests that in the absence of an intensive roadblock, intensive collective action may not be necessary to achieve power shift.The existence of two pathways to power shift is consistent with the QCA method's capacity to reveal multiple pathways to the same outcome.
The QCA analysis reveals two pathways to the absence of power shift.The first, which covers slightly less than half (3) of cases without power shift, indicates that the absence of a bureaucratic reformer, the absence of an intensive opening, the absence of intensive collective action, and the presence of a roadblock prevents power shift from occurring.The second, which also covers three of the cases without power shift, indicates that the absence of political transition, the absence of an intensive opening, the absence of intensive collective action, and the presence of a roadblock result in the absence of power shift.This set of two solutions suggests that the absence of a bureaucratic reformer or of a political transition is equally damaging to jumpstarting the state-society synergy process.Notes: ~ indicates the absence of a condition.Both sets of solutions have values of 1 for solution consistency; solution coverage equals 0.92 for the power shift model, and 0.57 for the ~ power shift model.Consistency refers to the proportion of cases with those conditions that experience the outcome while coverage refers to the proportion of cases with the outcome that experience those conditions.

Role of international aid agencies
Most openings from above involved some support from international aid agencies, though they were not primary drivers in any of the cases.Specifically, aid agencies were involved in 12 out of 19 cases, contributing either to the trigger or to the opening from aboveoften as convenors and brokers as well as funderssometimes of advocacy coalitions as well as government agencies.In eight of these cases, aid agencies played some role throughout the process.Aid agencies were not, however, a necessary condition for the occurrence or absence of power shift: they were involved in eight out of 12 cases with power shifts (Brazil-HIV, Colombia-Afrodescendant land titling, both India cases, both Philippines cases, and Sri Lanka), and in four out of seven cases without power shifts (Bangladesh, both Nigeria cases, and Peru).The two cases with an intensive opening but without intensive collective action (India-Mitanin and Sri Lanka) both had aid agency involvement, suggesting that aid agencies could perhaps bolster some stages of the process sufficiently to overcome gaps in others.They showed a capacity to play modest but catalytic roles (as in the Indian cases) or to provide national policymakers with additional room for maneuver in the face of either inertia or opposition from elsewhere within the state (as in Brazil, Mexico and Pakistan). 24These roles played by international aid agencies were very consistent with the bolder wing of the community of practice known as "Thinking and Working Politically," which encourages aid agencies to respond to national windows of opportunity with nimble support for reform coalitions that may include excluded groups (e.g., Teskey, 2022: 7).

Conclusion
When windows of opportunity allow state actors to attempt to enable collective action by the socially and politically excluded, what pathways do they follow and how far do they get?Rarely were these openings direct responses to social movement or electoral pressures.These openings were primarily initiatives led by bureaucratic innovators with diverse motives, often (but not always) enabled by political leadership change.This comparative analysis identified a substantial number of cases that led to shifts towards greater power for either citizens or reformist actors within public institutions, in spite of both structural obstacles and governmental roadblocks.These power shifts are both effects and causes of the difficult construction of countervailing powerdefined as "a variety of mechanisms that reduce… the power advantages of ordinarily powerful actors" (Fung & Wright, 2003: 260).
This comparative analysis of a diverse array of cases found that more intensive government enabling measures are associated with more intensive patterns of collective action. 25In spite of unfavorable contexts, the combination of intensive openings and collective action can drive power shifts and can sometimes partly overcome roadblocks.Yet in contrast to most experiences with official channels for participation, these cases of openings can be considered outliers -tangible state measures to enable collective action are rare.The cases involving more intensive openings are outliers within outliers.
When it comes to recognizing and explaining the trajectories of statesociety coalitions for change, academic theory still lags behind the actual practices of advocates of inclusionary reform.This study finds that the conventional dichotomy in the development and governance literatures between either collaborative or adversarial state-society relationships misses a key pathway towards inclusionary institutional change.When state reformers take tangible actions that enable the socially excluded to exercise collective voice, those social actors may both collaborate with and confront the same state, just as the reformers themselves may need to challenge unresponsive counterparts within their own government.This three-dimensional dynamic of collaboration plus conflict can drive virtuous circles of reciprocal interaction that in turn enable insiders to pursue further reforms.More often, however, those windows of opportunity are open only briefly, until reformers lose power.Though reformers' intentions may vary widely, their main constraint is their political weakness vis-à-vis the rest of the statewhich is one reason why they needed to enable collective action from below to pursue their goals in the first place.That leaves the question: what is left after openings close?
Shifting political winds at the national level often displace proparticipation political leaders and state managers.Several of the cases here suggest that the survival of openings is enhanced when pro-reform innovations are embedded in subnational institutions that involve statesociety power-sharing, as in cases of state government agencies in India (social audits and community health workers) and Brazil (HIV policy councils).In contrast to the widespread pattern of aid agency investments whose commitment to citizen engagement is primarily discursive, multiple cases here also demonstrate that resources from international aid agencies can bolster governmental openings from above that enable collective action.
To sum up, this study has identified a pathway to change that involves building countervailing power through reciprocal interaction across the state-society divide.The state-society synergy framework makes visible the mutual empowerment between insiders and outsidersa weapon of the weak that can drive inclusionary institutional change.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.attempted to strengthen the countervailing power of the excluded, see Fox (2020). 25This finding is consistent with Mansuri and Rao's comprehensive metaanalysis of official efforts to induce participation, which concludes that successful community-led development at scale requires robust support from governmental accountability institutions (2013: 287).These distinctions underscore the relevance of strategic approaches to governance reform to enable citizen action, in contrast to the much less intensive, tactical interventions that are the main focus of field experiments (Fox 2015).

Table 1
State-society synergy strategy cases & outcomes.Bolivia Law granted mainly indigenous rural citizens right to elect municipal authorities & gave grassroots organizations municipal oversight & decision-making powers, consolidating some autonomous indigenous organizations.Popular participation law Brazil HIV policy reformers created hybrid participatory councils & CSO umbrella networks at multiple levels, which sustained rights to nondiscrimination & the provision of antiretroviral therapy.HIV response Brazil (Ceará) A coalition between the Landless Movement and federal education officials designed & implemented a national rural education reform, opening state level power-sharing over curriculum, hiring, school governance & siting of new schools.Rural education Colombia Government reformers supported the launch of the first nationwide peasant association, though the next government weakened it.A legacy of consolidated mass organizations survives in some regions.Hybrid state agencies convened tens of thousands of social audits, including public hearings that led to problem-solving, recovered funds & discipline for corrupt officials.Social audits Mexico Policymakers formed regional councils to oversee a large-scale village food store network; some councils gained autonomy & survived at least two decades, helping to fend off program elimination.Community food councils Mexico Forest policymakers' support for community rights to resource management led to the consolidation of the largest self-managed community forestry sector in the world.Community forestry Philippines Government recognized grassroots organizations' involvement in village processes to propose & manage projects through Kalahi program.Councils of village representatives made project funding decisions for local public goods.Participatory development program Philippines Education ministry officials encouraged nation-wide participatory civil society oversight throughout the textbook supply chain, reducing leakage and increasing efficiency.Textbook monitoring Sri Lanka Government reformers bolstered both urban & rural community associations, resulting in improvements to housing quality and local infrastructure & regularization of tenancy.BangladeshPolicymakers passed a gender violence law that recognized new rights for women & the state's obligation to protect them in their homes, but the law was never fully implemented & few cases were filed.
Anti-domestic violence law China (Xiamen)Environmental policymakers leaked information about planned toxic plant to press.One citywide protest led to relocation of plant, but others did not & senior policy ally lost power.EnvironmentChinaNational leader of the Disability Rights Federation supported local petitions & protest against a ban on tricycles that slowed its implementation, but local government support for ban persisted and police repressed protestors.Disability rights Nigeria (Ogun)Independent oversight of the school feeding program by CSOs, parent-teacher associations & school management committees contributed to improved service delivery, but elections disrupted the program.School feedingNigeriaFederal reformers convened CSOs to carry out third party, donor-backed monitoring of large-scale National Social Investment programs, but national elections reconfigured program management and undermined monitoring.

Table 3
Conditions Used in QCA of Cases.

Table 4
Pathways to Power Shift among Cases.