The Politics of Incremental Progressivism. Группа авторов

to services but without impacting the wealth of the rich. The case of São Paulo suggests that easy distribution policies may be implemented under any government (although they are usually also first developed under left‐wing administrations), while hard distribution only happens during left‐wing governments.⁸

With this distinction in mind, we can return to policy production processes. Once decisions are made, policies must be produced and delivered, which brings bureaucrats, private contractors and policy community actors to center‐stage. These actors interact with politicians within policy‐specific governance patterns (Le Galès 2011) and are involved in the production and operation of all policies, irrespective of producing easy or hard redistribution.

Concerning median voter mechanisms, once easy redistribution policies are in place, they tend to continue regardless of changes to who controls the executive.⁹ The production of hard redistributive policies, however, shows a different trajectory. These policies are interrupted or sharply reduced during right‐wing administrations, but they do not die completely, showing different degrees of resilience. Instead, they enter a kind of latency period and may be reanimated later, after the next government swing. This process is not entirely compatible with current agenda‐setting theories, which are produced by advocacy coalitions (Sabatier and Jenkins‐Smith 1993) or by combinations of politics, problems and solution streams aligned by political leaders (Kingdom 1984). It is true that, as these theories predict, São Paulo's policies involved several political actors and groups strategically defending their interests and ideas, surrounded by institutions and historical legacies, as well as dealing with socially constructed problems and mobilizing existing solutions. However, traditional agenda theories suggest that policies that enter the agenda have “won” and tend to stay (like easy redistribution initiatives). In contrast, our cases show that many others (hard redistribution) follow a winding trajectory, shrinking or being discontinued, entering latency and being reanimated in the next friendly government.

These latencies and latter reanimations were made possible because the memory and organizational/operational capacities of these policies remained within the policy community, migrating inside and outside the State in the hands of bureaucrats, but also activists, academics and professionals who entered and left government. By policy communities, we mean the relational and issue‐based fields in which the actors of a policy sector interact (Marques 2003) beyond State and societal borders (Sellers 2010), not cohesive and unified actors amalgamated through collective action. In many cases, it is difficult to draw a hard line between State and society since actors circulate between many roles within these communities (Banaszak 2010; Abers 2019), reflecting the proximity between the urban reform movement and technicians in many municipalities. In the case of São Paulo, this has been the most critical influence/presence of civil society actors in policy production, participating in essential feedback mechanisms of policy change that connected State capacity building and policy production in non‐Weberian ways (Sellers 2010). The former process empowered actors central to the resilience of the latter, allowing policies to be reanimated from latency in subsequent governments.

      However, these processes worked differently according to governance patterns, multiplying the variations between policy sectors. Several policy‐specific elements help to explain different rhythms and resiliences that reinforce or hamper latency. The presence of actors and finance from higher levels – both federal and international – create more resilience and make the return from latency easier, along with policy institutionalization (in laws, administrative procedures and organizational structures). Highly capacitated bureaucracies and conditions of institutional insulation also contribute to resilience. The effect of policy instruments works along the same lines, as micro‐institutions that operationalize policies and depoliticize implementation once in progress, regardless of their intrinsically political character, sustaining the policy's logic even in the absence of concrete actors (Lascoumes and Le Galès 2007). One of the most critical elements to reinforce resilience and allow reanimation from latency, however, is the fit (Skocpol 1992) and embeddedness of policy actors in the bureaucracy and society, mainly in civil society organizations and academic circles. Finally, policies that hurt the interests of elite actors, as well as those of service providers, tend to be less resilient, as programs that produce hard redistribution. As we shall see in the following chapters, these processes operate differently by policy area, contributing to various degrees of resilience and different rhythms into and out of latency.

Although the chapters will develop a detailed account of policy changes, Table I.1 below summarizes the trajectories of the most relevant 30 urban redistributive programs.¹⁰ Each line represents a program within the studied policies, indicating their starting moments, latencies and reanimations through time, as well as the adopted intensities for their implementation (shown in grayscale). Columns represent governments, except for the last two, which classify the initiatives in terms of the types of redistribution (easy or hard) and trajectories (oscillated, came to stay or failed).

We can not only see that many redistributive programs were developed, but also that these became increasingly common as time passed. Among the policies, bus and waste collection services form the more significant proportions of recent changes, both areas consolidating changes that would remain in place for at least four administrations. Housing was the policy with the most significant number of innovations shifting back and forth from latency, while the same trajectory also characterizes traffic control and development regulation. Urban renewal presents increasing stability over recent years, associated with the learning process and the institutional consolidation of its main instrument. The subway, the only state‐level policy on the table, also presents changes, sometimes connected to municipal changes (and vice‐versa). Free passes, for example, were either produced by legislative initiatives (by the municipal council or the state legislature) or introduced following their establishment on municipal buses (and the other way around), as was the case of the integrated smart card. In any case, the main redistributive policy change in the subway was the construction of new lines to peripheries, which tended to be rare.

TABLE I.1 Trajectories of redistributive policy changes.

Although these patterns will undoubtedly become more evident throughout the chapters, the table provides a first glimpse of essential regularities. Some preliminary caution is necessary, however, to make the cell of the table more easily comparable. Although the table includes 30 programs, the comparison becomes more direct if we just consider municipal programs and disregard those that existed almost always (single bus fares and construction of new housing units) as well as the one that existed only in one government (the creation of a waste collection tax), reducing the universe to 24 programs. Since the municipal administration includes the same number of left‐wing, center‐right, and right‐wing municipal governments (three each), simple comparisons are already illustrative of the incidence of political ideology.¹¹ As predicted by partisan politics theories, most changes in municipal policies started and/or were reanimated in left‐wing administrations: 79% of the programs, while 15% of the programs started or restarted in centre‐right governments and just 7% of them in right‐wing governments, considering all intensities of implementation.¹²

However, when municipal programs were already underway, the difference between governments of different