Here we go again! Repetition and the Politics of Inclusive Institutional Reform

ABSTRACT This article introduces repetition as a sui generis temporal pattern that is central to the politics of inclusive institutional reform. Pushed by crises or leadership changes, inclusive reform initiatives spring up in moments of high salience only to die back again and be relaunched later, in a seemingly endless cycle. Drawing from public policy, historical and feminist institutionalist scholarships and the in-depth case studies of Birmingham and Turin city governments’ equalities offices, this article identifies (a) four drivers of repetition, to do with the nature of the problem, the nature of the institution, the nature of the change agents, and repetition’s self-reinforcing nature, and (b) two sets of effects (sedimentation and erosion) which follow concurrent logics of increasing and decreasing returns. Understanding repetition as a specific, non-linear pattern of change eschews teleologies of hope or doom, advancing theories of institutional change and serving as a useful guide for action for equity-driven reformers. It invites other scholars to engage with a new research agenda to systematically identify alternative types of temporal sequence and study their distinct logics and effects.


Introduction
The debate about institutional change typically focuses on two alternative temporal patters: punctuated and gradual change. 1 This article introduces repetition as an alternative, and overlooked, pattern.Based on two city governments' attempts over forty years to embed ethnic and racial inclusion in their work, I show how institutional reform agendas that require sustained political and administrative commitment are liable to go through repetitive cycles.A reform initiative is launched, then dissipates through resistance or neglect, only for a new reform initiative to emerge and dissipate in turn.This article proposes that repetition is a sui generis temporal pattern with independent effects on the institutions that engage in it.This is particularly pronounced in equity-driven reforms, but repetitive patterns also emerge in other policy areas.Therefore, understanding the reasons for and effects of repetition is central to understanding the politics of institutional reform.
Based on two in-depth case studies (the equalities offices of Birmingham and Turin city governments), I identify four sets of explanations for repetition and two concurring sets of effects.The prevalence of repetition in processes of inclusive institutional reform is driven by the stubborn nature of the problem they seek to address, the stubborn nature of the institution whose change is sought, the nature of the change actors driving the reform agenda, and the self-reinforcing nature of repetition.Once set in motion, repetition engenders two overlapping logics of change: an increasing returns logic, driving sedimentation effects, and a decreasing returns logic, driving erosion effects.Through repetition, new practices, expectations, and bureaucratic tools are gradually sedimented.This creates forms of path dependence that facilitate future cycles of inclusive innovation by easing internal resistance and creating easily re-deployable administrative practices.At the same time, repetition also erodes inclusive initiatives through fatigue, disillusionment, and the reinforcement of box-ticking and self-serving bureaucratic cultures.The repeated attempts in Birmingham and Turin to mainstream ethnic and racial inclusion and to embed stakeholder participation in local policymaking offer an illustration of erosion and sedimentation's combined effects.
In the rest of this article, I first reconsider the scholarship on institutional change to propose that repetition is an important and so-far neglected change pattern, which we are particularly likely to observe in cases of inclusive reform initiatives.Then, I introduce Birmingham and Turin's equalities offices and leverage these cases to demonstrate the prevalence of repetition in inclusive reform processes.The third section proposes four explanations for this prevalence, while the fourth discusses its effects.I conclude with a discussion of repetition's implications for the prospects of inclusive reform and its applicability beyond city-level inclusion policies.The article significantly advances our understanding of institutional change in theory and practice.It introduces a new pattern of change (repetition), demonstrating its analytical usefulness, and theorizing its drivers and effects.It encourages other scholars not only to apply the concept of repetition beyond the inclusivity policy area but also to identify other temporal sequences and consider how policy and institutional change initiatives are shaped by the type of sequence they are part of.Finally, it provides a new, practice-relevant understanding of the challenges of making institutions more inclusive, useful for both scholars and practitioners to navigate what is otherwise an often dispiriting process.

Institutional change and the nature of inclusive reform
There are two broad approaches to the study of how institutions change.Theories of 'disjunctive' (or punctuated) change see institutions as fundamentally stable and change as exogenous and rapid, resulting from critical junctures.Theories of 'evolutionary' change see institutions as malleable and change as endogenous and gradual, resulting from internal power shifts (Bernhard, 2015).For historical institutionalists that see change as disjunctive, the logic of increasing returns makes institutions path dependent (Pierson, 2000).Change occurs when exogenous events, typically a crisis of some sort, break this path dependence, creating permissive conditions for change agents to implement reform.Then a new path-dependent stability settles in, until the next exogenous shock.A similar logic is at the basis of Punctuated Equilibrium Theory (PET) in public policy, whereby long periods of incremental, small-scale change are interrupted by rarer moments of crisis-driven 'drastic change' (Adam et al., 2022).Other scholars have pointed to the limits of punctuated equilibrium, and to the many ways in which substantial institutional change can happen without a crisis (Streeck & Thelen, 2005).Even when institutions seem stable, actors engage in ongoing contestation and renegotiation, which reshape them gradually and radically (Mahoney & Thelen, 2010). 2  While there is general recognition that both endogenous and exogenous change patterns exist empirically, there are long-standing public policy debates about their relative prevalence, the magnitude of their respective change, and their mutual relationship (Baumgartner et al., 2009;Wegrich, 2021;Fernández-i-Marín et al., 2022;Knill & Steinebach, 2022a).At core, these are the questions that animate scholarships on, for example, policy accumulation (Knill & Steinebach, 2022a), policy learning (Dunlop, 2020;Assche et al., 2022;Leong & Howlett, 2022), and mechanisms of institutional continuity (Carstensen & Röper, 2022).Even with its internal differences, this scholarship points to broadly agreed understandings of institutional and policy change as complex, with different patterns likely to prevail in different contexts and policy areas, and change processes shaped not just by individual events but by their sequencing.
Historical institutionalists also note that, because institutional settlements are the outcome of power-distributional struggles, questions of power underwrite all processes of change.The relative power of change agents vis-à-vis continuity agents informs the outcomes of both disjunctive and evolutionary change (Pierson, 2015;Mandelkern & Koreh, 2018).However, power does not shift easily in a perfect pluralist balance, where today's losers can easily become tomorrow's winners.Rather, change processes are nested in selfreinforcing, though not immutable, power asymmetries that cement the winners' power and constrain the losers' capacity for effectively renegotiating institutions (Pierson, 2004;2015).This logic of self-reinforcing power means that, in unequal societies, all change is not equal.Change that challenges existing structuresparticularly deep-seated gendered, racialised, ethnocentric or ableist norms and practicesis going to encounter more obstacles than change that consolidates them.The latter (what we might call exclusive change) is in line with the observation that power begets power and institutions are shaped by the 'subterranean' path-dependant processes by which institutional winners gradually consolidate their control over resources, practices, and discourses (Pierson, 2015).Inclusive change is not simply the reverse process.As it implies a redistribution of power from the winners to the losers (empowerment), it cannot rely on subterranean, path-dependant shifts, but requires an explicit change agenda by equity-minded reformers.
Feminist institutionalism provides important tools to think about inclusive reform in the context of persistent, multiple, and intersecting power imbalances (Krook & Mackay, 2011).This body of work finds that even new institutions created with explicit inclusive agendas are always nested within 'old' informal norms and practices that tend to reproduce exclusionary forms (Chappell, 2011;Krook & Mackay, 2011;Mackay, 2014).This 'nested newness' means that 'no institutionhowever new or radically reformedis a blank slate' (Mackay & Murtagh, 2019, p. 11).This is in line with understandings among other historical institutionalists that change is 'bounded' (Pierson, 2004, p. 52) and past legacies shape new institutions' design and outcomes (Goodin, 1996).Scholars of race and minority politicsalthough typically not using institutionalist languagemake similar points about how formal democratic institutions are nested within persistent exclusionary norms and practices (Snoussi & Mompelat, 2019) and how those norms are self-reproducing and constitutive of institutions rather than glitches (Frymer et al., 2006).
Because it is nested within routinized exclusive practices and exclusive social, political, and economic contexts, and it is continually challenged, undermined, and hollowed out by the resurfacing of 'old' practices and norms, inclusive reform is unlikely to 'stick' and engender a strong path dependence.Part of the explanation rests in the 'liability of newness' (Mackay, 2014): new institutions are created with much enthusiasm and fanfare but their very newness makes them vulnerable to reversal, oblivion, and side-lining as they struggle to impose their legitimacy vis-à-vis established norms and practices.The reform potential of the 'new' is diluted or outright frustrated by a combination of open hostility, wilful ignorance, and the reformers' tendency to fall back on tried-and-tested 'old' practices and norms to establish the legitimacy of their fledgling initiatives.
Therefore, while it applies most clearly to racial and gender inclusion, the distinction between exclusive and inclusive change can be formulated widely to encompass all politics of reform.Reform that goes against the grain of existing institutional or societal power structures (inclusive reform) will not follow the same logics and temporal patterns as reform that goes with the grain of existing power structures, closing off even more avenues for the current losers to renegotiate existing arrangements (exclusive reform).Repetition is more likely to play out in the former.So, while this article deals with inclusive reform in the stricter sense of reform initiatives that explicitly and primarily aim at promoting ethnic and racial inclusion, its findings apply to a wider set of inclusive reforms in other contexts and policy areas.
Notwithstanding the abundance of empirical instances of repetition, theories of institutional change have not been alert to the implications of this temporal sequence.Despite the implied iterative nature of policy feedback loops and widespread interest in sequencing (Sewerin et al., 2020;Sewerin et al., 2022), the focus has typically remained on components, drivers, and effects of individual loops, rather than the drivers and effects of the iteration itself.A partial exception is Bernhard's (2015) treatment of 'non-critical junctures'.These are crises that, unlike critical junctures, do not bring about a new path-dependent institutional settlement.In explaining chronic regime instability as the result of a 'string' of such non-critical junctures, Bernhard prefigures the possibility of thinking of repetition (the string) as having an independent effect (chronic instability) that cannot be ascribed to any one single episode in the string.In a similar vein, inclusive reform initiatives can be understood as a string of missed or only partially captured opportunities for institutional reform.If we take each individual missed chance, we might conclude that the path dependence of old structures is insurmountable, bar rare occasions of radical institutional redesign.And even then, formal 'newness' will likely be diminished by 'old' informal norms and practices.While this is certainly part of the story of inclusive reform, it also misses another fundamental part: the fact of the string.

The prevalence of repetition in inclusive institutional reform
In a recent interview, the race inequity scholar David Gillborn commented that 'Working on educational inequity can feel like a soul-destroying task, especially when battles that I've fought decades before just keep on returning for a new generation'. 3Indeed, much to the frustration of reformers, initiatives to make institutions more inclusive of marginalized and minoritised people have a certain quality of being repeated (cf.Ahmed, 2012).Forty years of equalities in Birmingham and Turin's city governments (BCG and TCG) provide a useful vantage point on such repetition.City governments' equalities offices are ideal windows into the politics of inclusive institution-making: they engage explicitly with institutional reform initiatives, do so over extended periods of time, andas central governments become more exclusivecities are expected to be more committed to promoting equality and inclusion, making them most likely cases for successful inclusive reform (Cianetti, 2020).While Turin and Birmingham have been selected as they provide some of the longest histories of institutionalized inclusive reform in Europe, the author's engagement with Europe-wide urban policy networks confirms that they illustrate a pattern of repetition that resonates well beyond these two cases. 4 Over the past two decades many cities in Europe and worldwide institutionalized equality and inclusion agendas into dedicated administrative units.While I refer to them here as 'equalities offices', they can vary significantly in name (foregrounding, for example, equality, inclusion, or integration) and position within the wider administrative structure (they can be departments, sub-divisions, teams, or even a single officer).Variation is not only between cities: within the same local government, different such administrative units emerge over time.BCG and TCG both instituted their first equalitiesdedicated administrative units in the early 1980s.The empirical material for this paper has been collected by process-tracing their forty-year histories, using a combination of internal documents, stakeholders' reports, interviews, and participant observation. 5The equalities offices' timelines summarized below are the result of triangulation of these multiple sources.
Birmingham and Turin are mid-size post-industrial cities, struggling to revive their local economies after the collapse of their industrial base.They have higher-than-national-average unemployment and social exclusion and higher-than-national-average minority populations, 6 which tend to be at the sharper end of socio-economic vulnerability (O'Farrell, 2020;Rapporto Rota, 2017).This was exacerbated by a decade of austerity, which hit both city governments heavily.Beyond their similarities, Birmingham and Turin are embedded in significantly different national contexts regarding ethnic and racial inclusion policies (proactive in the UK, nearly absent in Italy), and have different histories of minority representation in city councils (significant in Birmingham, very limited in Turin). 7Yet, both cities have been pioneers of equalities-focused governance in their respective countries.Building on longer traditions of inclusive planning, they established their first dedicated equalities offices in the early 1980s: TCG's Office for Foreigners and Nomads in 1982, and BCG's Race Relations Unit in 1984.Over the next forty years, these early structures underwent significant change.
In Birmingham, the 1981 riots prompted the creation of the Race Relations Unit, tasked with advising the council on how to address institutional discrimination and improve equitable access to local services (Mano & Joly, 1994, p. 86).New riots in 1985 resulted in the Unit being placed under a higher-level official (an Assistant Chief Executive), the only such unit in Britain at the time being led by such a high-profile administrator (brap, 2016).A decade later, the UK-wide Macpherson Inquiry into the murder of Stephen Lawrence and subsequent biased police investigation brought the issue of institutional racism to the forefront of the national debate.BCG launched its own special Commission, which decried BCG's entrenched racist practices (2001 report) and insufficient steps to reform them (2002 report).A nationally-mandated new Race Equality Scheme was launched as a response.Following new riots in 2001, the 9/11 terrorist attacks, and the 2005 London bombings, 'community cohesion'with its subtext of dealing primarily with Muslim communitiesbecame the new mantra (Shukra et al., 2004) and 'named public sector officers and councillors suddenly had "cohesion" as part of their job title' (brap, 2016, p. 20).This resulted in repivoting equalities work towards faith-based organizations.At the same time, political uncertainty (in 2004-2012 no party had overall control of BCG) and internal conflict sapped the reform drive away from the Equality Directorate.External partnerships with civil society actors, like the Social Inclusion Process headed by the Bishop of Birmingham, partially filled this gap by conducting stakeholders consultations and making reform recommendations.The Black Lives Matter protests of 2020 pushed racial justice back into the agenda.This coincided with a reshuffle that brought a councillor with a strong equalities commitment back into the Cabinet and created the conditions for revamping BCG's equalities team and establishing a new Directorate of Strategy, Equalities and Partnerships at the end of 2021.
In Turin, riots and protests of the kind that marked Birmingham's recent history are rare.The timeline of the equalities office was instead driven by leadership changes and the fluctuating salience of the public debate on immigration.The election of the first communist mayor in over 20 years in the mid-70s and his collaboration with the equity-driven 'worker priest' Don Fredo Olivero resulted in the establishment of the Foreigners and Nomads Office, the first of its kind in Italy.The rise of the political temperature around the 'immigration emergency' in the 1990s prompted a new slew of initiatives by the progressive city leadership to deal with Turin's growing foreign-born population.This included attempts to promote migrant residents' political representation through an elective consultative body (the Consulta Stranieri) in 1994, which involved a significant investment but petered out by the end of the decade. 8This period also saw the creation of an Intercultural Centre in 1996 to support civil society organizations working on 'civic coexistence and the valorization of cultures', 9 a targeted urban development programme for disadvantaged neighbourhoods, and the proliferation of diversity-related offices across the local government. 10 A leadership change in 2006, together with the post-9/11 salience of 'community cohesion', like in Birmingham, brought a stronger focus on collaboration with faith organizations and inter-faith initiatives.A new Department for Integration was created 'with the mandate of defining a coherent intercultural policy' (Caponio & Donatiello, 2017).The new cabinet member overseeing the equalities portfolio in 2006-16 brought to the job her considerable Europe-wide contacts in intercultural policy networks and her expertise managing urban redevelopment projects.She oversaw the creation of a network of 'Neighbourhood Houses', independently managed spaces promoting inclusion and intercultural encounters.In 2016, as the culmination of ten years of work (TCG3), and not least as a response to new Islamist terrorist attacks in Europe and a growth in Islamophobia, a Patto di Condivisione (Pact of Sharing) was signed between the City and representatives of the Islamic centres of Turin.This created a new 'permanent coordination table' between TCG and the city's Islamic centres, and the flagship intercultural event 'Open Mosques'.Another leadership change in 2016 triggered another administrative reorganization: the Office for Rights was created, which works alongside an Interculture Office and an Equal Opportunities Office, with an explicit intersectional framework and a pivot to antiracism (TCG2a, TCG2b).New elections in 2021 resulted in a further reorganization of policy portfolios and responsibilities.
In both cities, once institutionalized, equalities remained part of the local agenda.However, periods of low salience and austerity negatively affected the ambition, internal appetite, and capacity for far-reaching reform.In Birmingham, in 1989 the Race Relations Unit 'had 31 staff in post, including race relations advisers in housing, social services, and education' (brap, 2016, p. 14) and used an elaborate consultative structure to include minorities more systematically in decision-making. 11However, in 1997, like many other similar units across the UK (Garbaye, 2005, p. 58), it was scaled down and merged with the Disability and Women's Units into a single Equalities Division.Its consultative structure was also dismantled and replaced with the Birmingham Race Action Partnership, which was hailed as an innovative approach to 'joined-up-working within and between institutions and communities' (Abbas & Anwar 2005, p. 64-5, emphasis in original), but that also soon lost momentum. 12When in the 2010s austerity hit the council with dramatic reductions in resources and personnel, balancing the budget became the main priority, pushing equalities on the back burner (BCG2a, BCG2b).The Equalities Division went through a series of reshuffles, mergers and scaling backs and was eventually dismantled.By the end of the decade, BCG was left with only one equalities officer.This was moved under different directorates and eventually ended up under Public Health, all the while fighting repeatedly to position the equalities function 'at the centre of the organization' and not as a sub-unit of any specific directorate where it risked to 'get lost' (BCG3).
Throughout these changes, equality mainstreaming and stakeholders inclusion in policymaking were repeatedly reiterated as key aims of inclusive reform.In the '80s and '90s they were at the core of the Race Relations Unit.The 2001 Race Equality Scheme also aimed at 'mainstreaming equality in [the council's] service delivery and employment activities', through regular monitoring, equality impact assessments, and stakeholder consultations.Similar aims motivated the 2018 Community Cohesion Strategy, the new 2021 equalities strategy Everyone's Battle, Everyone's Business, and the newly established Directorate (BCG2c).
In Turin, after big investments on inclusive urban regeneration in the '80s and '90s, the inclusive agenda remained but was marginal from the point of view of the actual budget (Belligni & Ravazzi, 2012).The new (in 2006) policy portfolio on 'integration of new Turin residents' did not have an independent budget but relied on funding from external sources.Later, austerity cuts resulted in the halving of TCG personnel, creating serious obstacles to the kinds of ambitious reform initiatives that had characterized Turin in the previous decades (TCG5, TCS2a).This, however, did not stop new mainstreaming and stakeholder participation initiatives.The need to mainstream and coordinate equalities work was particularly felt, especially as different offices had proliferated, working alongside each other on aspects of equalities with varying levels of coordination.The Department for Integration in 2006 was one such attempt at coordination.Most recently, in 2017, the Office for Rights led on the adoption of TCG-wide Guidelines for the Coordination of Intercultural and Participation Policies.These came with an informal coordination mechanism, bringing all relevant city officers together for a monthly meeting.After the 2021 elections, this informal mechanism got lost in the leadership change, although the need to do something about the splintered nature of equalities work was once again raised (TCG6; TCG9).
In the meantime, the Consulta experiment of the 1990s gave way to multiple initiatives to promote stakeholder participation.In a 2010 resolution, TCG committed to overcoming ad hoc collaborations and creating a council-wide strategy to empower community organizations. 13Similar priorities were restated a decade later by the new Office for Rights.The Rete Torino Plurale (Plural Turin Network) was established in 2020 to recognise and coordinate community organizations' Covid-19 relief work and reimburse their expenses, 14 as a step towards making them an integral part of city governance (TCG2b).The Office for Rights also launched a new citizens' engagement initiative for co-designing an Urban Pact on Antiracism as a Common Good. 15This was a flagship initiative, meant to create a new institutionalized network of civil society actors working closely with TCG to embed antiracism in all aspects of city governance.As of 2022, a new leadership is now planning to keep this network but with significant innovations in the way it will work going forward (TCG6; TCG8).It is too early to gauge the operational consequences of this promised further innovation.
These brief timelines show that, in both cities, attempts to make local government more inclusive followed a cyclical pattern.As one of my respondents put it, there is 'a constant scenario where the equality estate in the local authority is grown and shrunk, and as soon as it goes out of the limelight and the spotlight fades, it falls off the agenda' (BCS2).City council officials and activists described their experiences in a language replete with metaphors of repetition.They talked about 'coming full circle' and 'starting all over again' (BCG3), being in a 'perpetual loop' (BCS2), having 'been around long enough' to have seen similar initiatives before (B_Field, T_Field), and remembering other inclusive practices that 'are born and then get lost' (T_Field).As a jaded community activist put it, new inclusion initiatives spring up 'every five years or so … there are cycles of it in the city' (BCS1).
As schematically illustrated in Figure 1, each cycle starts with a leadership change or a crisis, which creates momentum for inclusive reform.While in Turin and Birmingham leadership changes were mostly political (following elections or reshuffles), the appointment of new administrative managers can be equally consequential, as each new manager is typically keen to deliver on a new change vision. 16Each crisis or leadership change results in a new reform initiative.Its ambition and the level of reach beyond the equalities office depends on the breadth and depth of institutional commitment.Reform initiatives' content, however, is recurrent: they typically aim to make inclusion a priority of the entire local government machine (mainstreaming) and/or to involve marginalized and minoritised communities in local governance (stakeholder participation).After a time, when the salience of equalities decreases, reform momentum is lost, resources are redirected elsewhere, and the initiative stalls or peters out, until the next moment of high salience.This ebb and flow results in a recurrent pattern of growing, shrinking, and moving of equalities offices within the wider city council machine, and a repetition of initiatives with similar aims.

Why repetition?
This section proposes four contributing explanations as to why repetition is such an inescapable part of the politics of making inclusive institutions, summarized in Table 1.
Inclusive initiatives tackle problems of systemic exclusion that are, by their very nature, stubborn.Officers are 'constantly battling' with stubborn exclusion issues (BCG4), confronting 'the same problems of a few months ago, and of last year, and the year before' (T_Field).The stubbornness of the problem is compounded by the 'cyclical dynamics of issue salience' (Capoccia, 2016, p. 20).External shocks (a riot, a terrorist attack, a protest) can increase the salience of exclusion and trigger the need for local government action.But the effects of 'triggering events' on an institution's reform drive are famously short-lived (Capoccia, 2016, p. 18).Other shocks (like austerity cuts) have the opposite effect of diverting attention and resources away. 17As the shift of emphasis from social inclusion to community cohesion shows, the definition of the stubborn problem also fluctuates over time.As a TCG officer put it: 'in the '90s [urban segregation] was the phenomenon and so, when it is the phenomenon, that draws attention.Later the phenomenon became immigration and so that shifted the citizens' and local government's attention a bit' (TCG1).Therefore, the combination of problem stubbornness, issue salience cyclicity, and periodic problem redefinition drives repeated inclusive reform attempts.
The problems that equalities offices want to solve are stubborn, but so is the institution they want to reform.Inclusive reform initiatives grind against an administrative machine whose explicit commitment to equality is not always guaranteed and, when it is, is not typically matched by actual appetite for radical change.In this context of 'status quo bias' (Pierson, 2000), reform coalitions are weak and volatile, and repetition itself can be a way of delaying or diluting reform.Even when highly-motivated equalities officers are involved, the sincerity of the reform drive at the level of the larger institution is not a given.Coming up with new flagship initiatives in response to the renewed salience of exclusion, with little evaluation of previous initiatives and little critical appraisal of the reach of the new proposed interventions, can become a goal in itself, a way of being seen to be doing something.If doing something is the goal, this can result in 'a pattern of replicating the same interventions again and again and again, even though we know that they don't make an impact' (BCS3), or launching the umpteenth stakeholder consultation process and 'hope that this time something comes out of it' (TCS2a). 18More or less wilful institutional amnesia results in the 'recycling' of forgotten past initiatives (Stark & Head, 2019).
Even when more radical reform processes are set in motion, commitment to reform is not evenly spread within the organization and, following the logic of 'nested newness' (Mackay, 2014), old norms and practices tend to persist or resurface.The reach, ambition and life cycle of equalities offices' initiatives are dependent on the level of support and commitment they get by the political and administrative leadership and by the officers of other, 'fatter' (TCG2a) departments.A significant part of equalities officers' time is therefore spent in internal advocacy to mitigate against this constant risk of 'reverting back' (BCG2), and to painstakingly build internal reform coalitions.This, however, means that changes in leadership undo this work and can and often do result in 'starting all over again' (BCG4).As leadership changes and 'old' bureaucratic practices empty new initiatives of their reform potential, newer initiatives will be needed, starting a new repetitive cycle.
The nature of change agents also encourages repetition.Equalities offices have an ambiguous role as internal brokers of institutional reform, which makes them particularly fruitful vantage points to study institutional change (Ahmed, 2012).It also makes them unstable.There is a basic tension between the very existence of a dedicated equalities office and the need to make equality a core concern of the whole institution.Having a dedicated office that 'does diversity' can mean that other units do not see equalities as their responsibility; however, not having a dedicated unit because equalities are the responsibility of everyone 'translates quickly in "no one"' (Ahmed, 2012, p. 136).This mainstreaming dilemma results in a recurrent 'problem of location' for the equalities function (brap, 2016).This was clearly articulated by a BCG official, reflecting on the re-establishment of the equalities team: 'I think that the next challenge is … avoiding the risk that comes from having a team of the rest of the organization then thinking: "Oh well, equalities is dealt with by the Equalities Team"' (BCG2c).
As the equalities function's appropriate location within the administrative machine is debated, 'the internal power politics of where [it sits] has been a constant issue going backwards and forwards' (BCS2).This fluid and highly politicized status means that equalities offices' structure, responsibilities, and priorities are always liable to be tinkered with, redesigned, or even completely overhauled following a change in leadership or responding to external events.In both cities, equalities responsibilities have been passed around under different departments, occasionally being placed under the mayor's office, and have been split and merged depending on their prioritizing and deprioritising in the wider council strategy (cf.Mano & Joly, 1994, p. 126).This also makes such departments always precarious vis-à-vis larger, more established departments, and always having to prove their legitimacy (cf.Mackay, 2014), in a way that shapes their working practices (T_Field).The equalities office's position vis-à-vis the wider institution at any given time affects the ambition and durability of their reform initiatives.Initiatives launched by smaller, more side-lined departments are likely to engender shorter reform cycles and leave fewer durable sediments.
Moreover, while both city governments affirm that equality should be at the core of their work, the work of equalities offices is not a statutory obligation for the city.Therefore, especially in times of cuts (BCG6) or if a new local government with less interest in these themes is elected (TCG1b), they are at high risk of being defunded and marginalized.One way for committed equalities officers to overcome endemic underfunding is to rely on external funding.While this means that some inclusive projects get done, projects have their own time-limited logic that also reinforces repetition at the expense of long-term inclusive reform (TCG6, T_Field).Moreover, as a lot rests on the drive, capacity, and internal and external network-building of individual highly-motivated officers, equality agendas become easily personalized.The personalization of equalities agendas also drives repetition, as 'the second [these highly-motivated individuals] disappear or are not there or move on, all that goes away.And then you're left at the position of starting again' (B_Field), sometimes with a 'new batch [of whom some] have got a political awareness of what's actually gone on before, and some haven't' (BCS2).
Finally, repetition has a self-reinforcing logic.As discussed in the next section, repetition breeds distrust towards the local government's commitment to equalities.Therefore, there is an institutional incentive to present inclusive initiatives as breaking with the past.As a respondent put it, 'we need to make something fundamentally different if we're going to achieve the big change that we need' (BCG5).Newness must be brandished to demonstrate to those that have seen similar initiatives before that this time it is for real.New equalities officers are keen to affirm that things have 'changed a lot now' (BCG7) and 'this time is different' (B_Field), and that even when working 'in continuity' with the previous leadership they are introducing 'fundamental novelties' and deploying similar tools but 'based on a new logic' (TCG2; also TCG8).In fact, one difficulty of studying repetition is that there are clear institutional paper trails (policy documents, council debates minutes, press releases, news items, social media posts) of when new initiatives come up, but the trail soon goes cold, so that it is difficult to determine when a 'new' initiative peters out and why.The difficulty of following the trail of the new is not only the result of researchers' preference for looking at reformers' strategies and neglecting the strategies of the statusquo actors (Capoccia, 2016, p. 23).It is also a material effect of the recurrent newness of reform initiatives: yesterday's new initiative is overshadowed by today's, which in turn will be overshadowed by tomorrow's.

What does repetition do?
This section turns to repetition's effects, summarized in Table 2.I argue that repetition's sedimentation and erosion effects, driven by opposing logics of increasing and decreasing returns, are central to the politics of making inclusive institutions.
Increasing returns are 'self-reinforcing or positive-feedback processes' (Pierson, 2000, p. 251), which mean that decisions taken early on in a sequence are likely to be locked in.Once an institutional form or practice is set in motion, positive feedback loops make it easier for it to stick than to be changed (Pierson attributes this to the 'cost of switching').In other words, the more something is done, the more it is likely to be done again.This institutional inertia can be frustrating for change agents, but it can also be a source of gradual progress: as inclusive initiatives get repeated, new norms and administrative practices get normalized and routinizedthey sediment.Change agents can then build on those sediments to push for further reform.Equality impact assessments (EIAs) in Birmingham are a case in point.They were established in a period in which the equalities workforce was being cut down to one.Implementation was pushed by this one officer and was resented in parts of the organization as another box to tick (BCG4, BCG5).Although the quality of engagement with EIAs varies, officers that were originally sceptical report that over time 'equalities has really started to bed-in into the way we work' (BCG5).The newly re-established equalities team plans to seize on this routinized practice (sediment) to make it a meaningful tool of equality mainstreaming, by strengthening EIA support and monitoring structures (BCG3, BCG7).Similarly, TCG officers talked about the practice of reusing ('copy-pasting') bureaucratic documents that were done ad hoc and against internal resistance for a specific project: 'this way we can transform a [one off] project idea into a [permanent] administrative practice' (T_Field).This routinization of equalities work gradually turns it from controversial to 'an element of normality for the city' (TCG3).It also makes it administratively easier.New initiatives are more likely to be accepted when they are based on models that are familiar to the broader organization, and routinized ways of working on thorny equalities problems create that familiarity (B_Field, T_Field).Tried and tested tools can be reached for, like interdepartmental and public-civic 'tables' in Turin, or stakeholder consultation schemes and public-civic partnerships in Birmingham.These are relatively easy to organize because they have been done before and everyone already knows who needs to do or sign what to make them happen, easing administrative resistance.
Even names can be useful sediments.As a BCG respondent suggested: 'we're now called the Inclusive Growth Directorate.I think that name had been in place for a while before anybody really knew what it meant, to be honest with you.And it's only over the last couple of years that we've really started to actually move in that direction' (BCG5).So, once equality enters the organization as a (at least rhetorically) key aim and commitment, bits of it sediment through repetition and remain available to change actors to mobilize when the opportunity arises (cf.Ahmed, 2012).Sedimentation through repetition, therefore, lowers the 'start up costs' (Pierson, 2000, p. 254) of trying again next time.This chimes with established evidence on the value of task repetition in cognition and language learning studies, which also show how each iteration of the same task is different as it builds on previous iterations (Bygate, 2018). 19 Sedimentation is not just an accidental by-product of repetition.Equalities officers are well aware, at least at an intuitive level, of the cyclical nature of inclusive reform, which means that they can intentionally plan for sediments.
In launching new initiatives that they know will face resistance and reversals, equalities officers try to, e.g., establish 'sticky' obligations for the administration, broker lasting relationships between other offices and community stakeholders, or produce 'copy-pastable' administrative practices that can sediment even if the initiative itself peters out (TCG2a). 20 Sedimentation, however, is accompanied by erosion, as repetition also triggers self-defeating dynamics.In economics, decreasing returns are the basic negative feedback mechanism that maintains 'optimal equilibrium' (the status quo).This logic also applies to political processes (Gains et al., 2005;Deeg, 2001).As the more something is repeated the more it is devalued, repetition erodes inclusive reform.Confronted with yet another inclusive initiative, actors within and outside the organization respond with fatigue, distrust that this time is really going to be different, and foot-dragging.In this sense, repetition is an additional, and so far ignored, type of self-undermining process (Jacobs & Kent Weaver, 2015).
Equalities officers recognise this distrust and fatigue.Upon relaunching the equalities team, BCG equalities officers reported feeling 'really optimistic about the direction of travel after such a long time' but recognizing that 'there's a lot of mistrust in the organization and a lot of apathy' (BCG4): 'there's also, and I absolutely get this, a scepticism because people have heard a lot of these similar sounding things in the past' (BCG3).This is also true for stakeholders, that have grown 'understandable disaffection' (BCG7).As one local activist put it, 'I've been working [on these issues] for about 20 years and I've seen a similar statement in different forms must be more than 20 times from the Council … That's great that they're saying that institutional racism exists.But they've been saying that for 20 years' (BCS3).
While it eases internal resistance, routinization of equalities work is also self-undermining.As inclusive practices are turned into bureaucratic routines, these are increasingly perceived as hollowed ('box-ticking') by both those inside and outside the organization.Repeated attempts at engaging stakeholders in 'another consultation' (B_Field) or 'another project' (T_Field) are received with fatigue and distrust.They are seen as serving internal administrative logics rather than providing a genuine avenue for empowerment.Moreover, out of habit and for administrative expediency, they often involve the 'usual suspects'that is, stakeholders that have collaborated with the council before, can navigate its bureaucratic systems, have the competencies to write bids and to respond to policy consultations, and 'speak progettese (project-writing jargon)' (T_Field).As one respondent put it: the way that local bureaucracies are sort of set up, the work of doing racial justice is often quite technocratic.It's about feeding the administration of the state, really.It's like, well, how many people have we engaged with this year?Were they from diverse backgrounds?You know, there's a sort of feeding the beast [which] can make it much harder to go beyond the performative.(BCS1b).
As the two cases show, sedimentation and erosion effects do not happen separately (sedimentation under certain conditions and erosion under others), but repetition triggers both at the same time. 21This explains the non-linearity of repetition patterns, whereby changes are not gradually moving in one clear direction, but self-reinforcing and self-defeating dynamics act together in shaping the opportunities and limits of each subsequent attempt at institutional reform.

Conclusions
I have argued that repetition is a sui generis temporal pattern, which shapes the politics of inclusive reform.This insight advances existing scholarship on institutional change in at least three ways.First, it overcomes a widespread tendency in the literature to focus on individual reform initiatives, and points instead to the longer temporality of reform processes.In so doing, it contributes to a recent strand in public policy research that looks at the 'forest' (aggregate patterns of change) rather than only individual 'trees' (single instances of policy change) (Knill & Steinebach, 2022a, p. 603).A focus on the sequence of reform initiatives shows that the outcomes of any new initiative do not depend solely on its ambition, design, implementation, the breadth and depth of the coalition that supports it, and the level of institutional resistance it encounters.Rather, outcomes are mediated by the repetitive sequence it is part of.Each new initiative builds on sediments of past initiatives, leaves its own sediments, and is undermined by the fact of being yet another initiative in a repetitive string.This combination of erosion and sedimentation results in an institution that is not completely stuck and unchanged but is also not changed in a clear linear directionone of steady if slow progress towards more inclusivity.The contradictions of such non-linear change are well summarized by one of my respondents, who said of BCG's most recent cycle of inclusive reform that it 'feels slightly different, but very similar, but different' to past cycles (BCG4).Attending to repetition, therefore, eschew teleologies of hope or doom to allow for complex and contradictory change.
Second, in taking on board feminists' insights on 'newness', this article also pushes them further by showing that newness can be recurrent.New inclusive initiativesby virtue of their very newnessare liable to be frustrated or only partially realized.Opportunities for reform interruption (Mandelkern & Koreh, 2018) and institutional amnesia (Stark & Head, 2019) abound, which make change less likely to stick.However, the problems they are meant to solve are persistent, so newer new initiatives eventually spring up.The politics of inclusive reform is thus trapped in a politics of repetition.The article's suggested distinction between inclusive and exclusive reform makes this insightand the feminist literature it builds onusable beyond gender and racial inclusivity policies to other instances of transformative institutional reform.
Third, the article advances our understanding of sequencing and its implications.It demonstrates the advantages of thinking in terms of different sequence types, overcoming rigid divisions between punctuated and gradual change.The repetitive sequence discussed here comes closer to gradual understandings of change.Yet, it is a sui generis gradual pattern.It is cyclical rather than simply cumulative, and it is punctuated by permissive opportunities for reform that, following Bernhard (2015), we could usefully understand as a string of non-critical junctures.Early events in the stringsuch as the decision to set up an equalities officeare likely to have strong path-dependent effects.Indeed, albeit in changeable shapes and sometimes in very diminished form, such administrative units once established are likely to persist.However, and contra Pierson (2000), this does not mean that events later in the string have necessarily less effect than earlier ones: each new initiative contributes to lengthening the string and making its effects (both positive and negative) stronger.Thus, taking repetition seriously is a way of taking time and sequencing seriously in the study of institutions (Pierson, 2004;Grzymala-Busse, 2011).
However, repetition is not the only possible type of sequence; it is a prominent oneand thus important in its own rightbut also a starting point for a new research agenda that systematically maps sequence types and their effects.If it is true, as Tilly suggested, that 'when things happen in a sequence affects how they happen ' (1984, p. 14), we must pay attention to the nature of that sequence.This article has argued that a repetitive sequence has specific characteristics that are most clearly felt in inclusive institutional reform.Different types of sequences are likely to interact with the timing of events differently, determining different patterns as to 'how things happen' and with what effects.Future research should not only investigate when and with what effects repetition applies to other policy areas, but also identify other sequence types, under what conditions they are most likely to emerge, and how they shape patterns of institutional change.Such an agenda can have deep implications for long-standing debates on, for instance: policy failure and success (for example, opening up questions about the effects of a string of failures), incremental vs crisis-driven change (whose mutual relationship and respective outcomes are likely determined by the kind of sequence they are embedded within), and processes of institutional learning (that are likely to evolve differently within different sequences).
While there is often a normative component to debates on incrementalist and gradual change (Adam et al., 2022), this article is not making a normative point regarding repetition but an analytic one: repetition is not an a priori good or bad policymaking strategy for the goal of building inclusive institutions; it is what happens empirically, and as such it cannot be ignored, not least because it has practical implications for equity-minded practitioners.In this sense, this article contributes to an emerging 'forward-looking' debate about how to design policies that bring about positive change (Sewerin et al., 2022).A clear understanding of why repetition happens and how it produces positive as well as negative effects is essential for reformers who want to design 'policies that intentionally stick' (Jordan & Matt, 2014).While it will not eliminate the erosive side of repetition, an awareness of how repetition works can take us closer to making more inclusive institutions.

Notes
1.For recent critical summaries in comparative politics see Gerschewski (2021), in public policy see Fernández-i-Marín et al. (2022).2. In what is perhaps a case of 'convergent evolution', historical institutionalist and public policy scholarships developed their own separate concept of and debate on 'punctuated equilibrium', with virtually no cross-pollination.While their emphases differ (focused on drivers of path dependency among historical institutionalists and on attention and agenda setting among public policy scholars), both point to broadly the same type of temporal sequence.More cross-field engagement would be fruitful.3. British Academy newsletter, 27 January 2022.4. The author discussed findings with local policymakers and equality activists at UNESCO's 2021 Human Rights Go Local academy, and the European Network Against Racism (ENAR)'s 2022 'Regions Against Racism' workshop.5.I conducted 24 interviews with civil servants and activists in Turin and Birmingham.Respondents are anonymised and identified by a code, indicating the city (B for Birmingham, T for Turin), their role (CG for city government, CS for civil society), and a number.I conducted interviews in 2019-2022, speaking with some respondents multiple times: in these cases, the respondent's number is followed by a letter identifying the interview number.I also attended webinars on BCG's new equality agenda and conducted participant observation on a 2020-21 TCG stakeholder-inclusion initiative.I followed the entire process, which involved TCG officials and 39 civil society organisations and took place mostly online due to Covid-19 restrictions.I attended all virtual plenary meetings and selected sub-group meetings, held regular debriefings with the TCG officer in charge of the scheme, and five one-to-one debriefings with civil society participants, resulting in a co-written feedback document that was shared with TCG officers.Notes from this fieldwork are part of this article's material and are referred to as B_Field and T_Field.The anonymised list of respondents is in Appendix 1. 6.According to the 2021 census, over 51% of Birmingham residents are from an ethnic minority (www.ons.gov.uk).Turin, the city of car manufacturer Fiat, attracted subsequent waves of immigration from the surrounding rural areas, Southern Italy, and abroad (Cingolani, 2016).As of 2021, about 15% of its population are foreign nationals, compared to less than 9% nationally (www.dati.istat.it).Since Italian censuses do not include questions on ethnicity or race, this is an underestimation of minoritised and racialised residents.
Collaborazione-per-Torino-Antirazzista.pdf.A description of the process by a city officer is available at: https://nuovenarrazioni. medium.com/pact-for-anti-racism-in-torino-the-co-designphase-b53dea60b34d.The English translation of the feedback reflection co-written by the author and five civil society participants is available at: https://nuovenarrazioni.medium.com/pactfor-an-anti-racist-turin-a-reflection-7b7feb6a161a.

Table 1 .
Explaining the prevalence of repetition.

Table 2 .
Effects of repetition.