Benevolent Policies: Bureaucratic Politics and the International Dimensions of Social Policy Expansion

Research on the welfare state has devoted considerable attention to social policy expansion. However, little is known about why governments expand social policies serving groups with limited power on issues with low visibility. I call these “benevolent policies.” This class of social policies improves population well-being but produces minimal political gains for the governments enacting them. Why do governments expand benevolent policies if political incentives for reform are weak? I investigate this question by focusing on government responses to malnutrition. Drawing on nine months of fieldwork, including 71 interviews, I argue that the origins of policy expansion can be found in the government bureaucracy. Bureaucrats with technical expertise—technocrats—can play a defining role, deploying international pressure to court executive support and orchestrate policy change. Their actions help explain the Indonesian government’s unexpected expansion of nutrition policies, which serve low-income women and children and address micronutrient malnutrition.

W hy do governments expand social policies? This question is central to our understanding of the welfare state. The existing literature offers two broad explanations for why governments act. First, governments are responding to groups with political power. Governments have incentives to introduce policies that appease groups which can threaten their political survival (Dion 2010;Esping-Andersen 1990;Garay 2016;Korpi 1978). Second, governments are addressing visible issues to garner political support or preempt political backlash (Diaz-Cayeros, Estevez, and Magaloni 2016;Haggard and Kaufman 2008;Hicken and Simmons 2008;Niedzwiecki 2018;Wong 2006;Zucco 2013).
Left unknown is why governments expand social policies for groups with limited political power on issues with low visibility. I call these "benevolent policies," identifying them as an overlooked class of social policies. Groups with limited power are less able to exert pressure on governments and elicit favorable reform. Issues with low visibility are less likely to be recognized by the public, reducing the likelihood that governments can use them for short-term political gains.
Conventional wisdom suggests that governments do not have strong incentives to expand benevolent policies and are unlikely to do so-yet empirical evidence, which I will discuss, shows there are conditions under which this occurs. I focus on explaining this puzzle.
Emblematic of benevolent policies are nutrition policies. 1 For women, malnutrition 2 prior to and during pregnancy is associated with restricted fetal development. In children, evidence links malnutrition to an estimated 3.1 million deaths every year, approximately 45% of global child deaths, primarily in low and middle-income countries (Black et al. 2013). Yet nutrition policies do not adequately respond to these challenges (WHO 2013). Government inaction, I contend, is a product of weak incentives. The recommended nutrition policies primarily serve a group with limited political power (low-income women and children) and address an issue with low visibility (micronutrient malnutrition), reducing political incentives for policy reform. 3 Certain governments, however, have defied expectations. This article focuses on Indonesia. After years of ignoring pervasive levels of malnutrition, the Indonesian government suddenly sprang into action and expanded its nutrition policies in 2012.
What explains the expansion of benevolent policies? Existing theories do not offer persuasive explanations. The social policies studied in the welfare state canonserving groups with political power (e.g., the upper class, unions) or addressing issues with high visibility (e.g., cash transfers, health insurance)-lend themselves to overt, and often contested, forms of redistribution. From this vantage point, policy expansion is a product of conflict among domestic groups, organized to protect or advance their interests, or of a government's pursuit of political gains. Government bureaucracies are viewed as mere arbiters of competing interests, bureaucrats within them as passive agents.
By contrast, I argue that the origins of benevolent policy reform can be traced to the government bureaucracy. In this article, I show that bureaucrats who have technical expertise in a subject-technocrats-can play a defining role. Technocrats harness the international pressure created by international organizations (IOs) to broaden their political connections within the bureaucracy. They use these connections to build vertical networks and elevate the desired policy onto the executive agenda, then leverage horizontal networks to coordinate policy change.

Contributions
This article makes three contributions. First, I conceptualize a new class of policies-benevolent policieswhich the comparative welfare state scholarship has overlooked and undertheorized. While the empirical focus of this article is nutrition policies, the concept travels to any social welfare and public goods policy that serves a group with low power on an issue with low visibility. Examples include policies that serve migrants, refugees, low-income and racialized groups, or policies that address non-communicable diseases, sanitation, long-haul COVID, or water fluoridation, as I will discuss. This article opens up analytical space for studying evidence-based policies that dramatically improve welfare but provide governments with limited short-term political gains.
Second, I draw attention to technocrats as an important catalyst of change. For challenging issues like malnutrition, we tend to think of powerful social movements or vote-seeking politicians as the drivers of change. But technocrats, I argue, and their strategic deployment of international pressure within the bureaucracy, can also act as a source of policy change. While the comparative politics scholarship has "tended to ignore the bureaucracy itself" (Pepinsky, Pierskalla, and Sacks 2017, 250), recent research is advancing knowledge on bureaucratic politics (Bersch 2019;Bozcaga 2020;Centeno, Kohli, and Yashar 2017;Ding 2020;Kim 2020;Toral 2021;Utama 2014). I join this work by drawing on studies which bring the oftdivorced comparative politics and international relations literatures into conversation (Dion 2008;Keck and Sikkink 1998;Martínez Franzoni and Sánchez-Ancochea 2014;Orenstein 2013;White 2011), continuing efforts to dissolve boundaries that have traditionally separated subfields.
Third, this work intervenes in a long-standing debate on democracy and welfare. Influential studies have produced contradictory results. Some find that democracy reduces infant mortality (McGuire 2010), whereas others do not (Ross 2006). The concept of benevolent policies helps explain this disagreement. I show that democracy's effects on social policies are not uniform. Although electoral competition contributes to the expansion of policies addressing issues with high visibility, it sidelines issues with low visibility. Studies have primarily documented democracy's effect on policies for high-visibility issues like cash transfers or health insurance. But the sidelining of low-visibility issues such as micronutrient malnutrition, which contributes to 45% of global child deaths, can explain why democracy produces mixed results.
This article begins by discussing the concept of benevolent policies. These policies improve social welfare but produce minimal political gains for governments enacting them. While governments do not have strong incentives to expand these policies, some have done so. In 2012, the Indonesian government suddenly prioritized nutrition, marshaling resources toward the formulation and introduction of new nutrition policies. This was a break from past behavior. For decades, malnutrition rates remained "stubbornly high" (Rokx, Subandoro, and Gallagher 2018, 19). Despite the availability of simple, low-cost policy solutions, the government did not appear to pay serious attentionuntil 2012 when it unexpectedly introduced a raft of new policies.
To explain policy expansion, I develop an analytical framework that shines a bright light on the inner workings of the government bureaucracy. After describing my methods, rationale for case selection, and the data collected, I illustrate the analytical framework using the Indonesia case. I conclude by explaining the relevance of the benevolent policies concept for other social policies, then discuss the theoretical contributions of this article and an agenda for future research.

BENEVOLENT POLICIES
In this article, I introduce the concept of "benevolent policies" to explain why governments fail to tackle serious societal problems. I use the term, "benevolent," to refer to acts that improve welfare but produce limited short-term political gains. The term is slippery, however, given the myriad of ways in which it can be used. In certain contexts, it conjures up images of a paternalistic state, guiding the supposedly misguided. In other settings, it is bound up with histories of racism and colonialism, reminiscent of the imperial impulse and so-called civilizing missions that occurred under the benevolent Western gaze.
I use the term, benevolent, as a reference to the benevolent dictator who acts for the benefit of her people rather than her own personal betterment. Benevolent policies advance citizen well-being, not just a government's short-term political interests. Reflecting this, scholars have used the concept of a "benevolent state" to refer to governments that "act out of a primary concern for the welfare of all its citizens" and use "policies … to find solutions to recognized social problems" (Marcuse 1978, 21).

Good Policies, Bad Politics
Benevolent policies have two core characteristics: they serve groups with limited power and address issues with low visibility. These policies improve welfare and well-being. But in comparison with policies that serve groups with high power or address issues with high visibility, benevolent policies produce limited short-term political gains for the government. These policies cannot easily be used to win votes, reward supporters, anchor coalitions, enhance government legitimacy, or ease popular discontent. The gains accrue to beneficiaries, a group with limited power, not the government enacting them. Moreover, welfare improvements take years to become apparent, limiting the government's ability to claim credit for positive outcomes. These are good policies-but bad politics.
As there are an infinite number of groups to serve and issues to address but a finite amount of resources, governments must prioritize. Given their desire to continue governing, governments have strong incentives to serve groups with high power and address issues with high visibility, as I will explain.

The Group: Political Power and Policy Expansion
The "group" refers to two or more individuals who share a common characteristic. The group's political power is shaped by its access to resources. These resources, whether they be funds, capital assets, or political connections, provide the group with greater influence over the government's political survival.
Governments have strong incentives to respond to groups with significant resources (Ramesh 2000), often governing on behalf of the powerful rather than the people. In the United States, the government has accommodated the demands of the upper class, resulting in a welfare state that "showers its benefits far more generously on the haves than on the have-nots" (Mettler 2011, 25). Similarly, in Uganda, the government has channeled favors to groups whose support they wish to secure (Mwenda and Tangri 2005).
A group's political power can be augmented through collective action and social mobilization (Díez 2013). When "marginalized groups are incorporated into the political arena and possess mobilizing power, there is greater pressure on the state to respond to their needs and expand social protection policy" (Pribble 2011, 194). In Sweden, the government responded to bluecollar workers who, organizing through labor unions, used strikes and protests to exert their political power (Korpi 1978). In Argentina, the government expanded social policies for unemployed and informal workers when they organized roadblocks and demonstrations (Garay 2007). The government is more likely to attend to groups that have mobilized and can threaten its survival.
The most "pressing problems" are those that are "forced on the policymakers through pressure from injured or interested outside parties" (Hirschman 1981, 146). These outside parties, however, vary in power, and social policies reflect the relative strength of different interest groups (Mesa-Lago 1978). It is domestic societal pressures emanating from groups with high power-that is, groups with access to substantial resources or the ability to coordinate largescale collective action-that demand government attention and policy action. When a group has high political power, there is a need to appease it or, at the very least, avoid provoking its ire. For this reason, governments have strong incentives to prioritize groups with power.

The Issue: Visibility and Policy Expansion
The "issue" is the social problem that exists in society. The issue's visibility is reflected by the public's awareness of the issue or the public's awareness of the government response. 4 The more immediate the consequences of the issue or the government response, the higher the public's level of awareness. For example, crises such as earthquakes and nuclear power plant accidents, which have immediate consequences, often precipitate policy action (Birkland 1998). The public is very aware of the issue and the government's role in remedying it.
Governments have strong incentives to address issues with high visibility. For social policies, an example is economic poverty. People know when they lack access to money and that the consequences can quickly result in deprivations such as hunger or homelessness. When governments expand social policies, such as conditional cash transfers (CCTs), the public sees the benefits of government action right away. For this reason, CCTs "with clear attribution of responsibility have provided electoral advantages to incumbent presidents" in Latin America (Niedzwiecki 2018, 12). Indeed, "the immediate electoral rewards to incumbents who implement and expand programs can entice rulers around the world to adopt CCTs" (Zucco 2013, 811).
In addition to cash transfers, governments have used other social policies to garner political support (Diaz-Cayeros, Estevez, and Magaloni 2016). Following democratization, governments in Korea and Taiwan had strong incentives to extend health insurance to the broader citizenry, as everyday citizens had gained the ability to vote (Wong 2006). Similarly, in Malawi, the government garnered political support by addressing food insecurity and expanding agricultural subsidies (Dionne and Horowitz 2016). In Tanzania, the government improved political approval because of its efforts to address malaria through bed net distribution (Croke 2021).
Governments have also used social policies strategically, introducing reforms to preempt popular discontent. For instance, "social insurance in the early Latin American adopters" was "initiated by political elites who pre-empted the pressures for reform with policies aimed at diverting labor from radical demands" (Malloy 1989, 319). 4 My conceptualization of low visibility shares parallels with work by Jacob Hacker and Suzanne Mettler. Hacker examines social policies affected by "subterranean politics, only occasionally involving a broad circle of participants and resisting the scrutiny that public programs typically invite" (2002,83). My definition, by contrast, does not specify the participants. For Mettler, submerged policies have low visibility as the government's role is obscured (Mettler 2011). By contrast, I define low visibility as low public awareness of the issue or government response.
Of its arsenal of social policies, governments have incentives to prioritize those that address high, rather than low, visibility issues. Like clientelistic exchanges, where politicians rely on the distribution of particularistic benefits rather than public goods to win office (Mares and Young 2018), social policies are politically useful when the recipient is aware of the issue addressed, or the government response. Therefore, governments are better positioned to garner political support or preempt discontent when they expand policies addressing high-visibility issues.

The Puzzle of Nutrition Policies
Unlike policies studied in the welfare state canonwhich serve groups with high political power or address issues with high visibility-benevolent policies are a separate class of social policies. They serve groups with limited power on issues with low visibility. To be sure, a group's level of power and an issue's visibility are not set in stone. A group can augment its power if it gains access additional resources or effectively mobilizes. Similarly, an issue's visibility can change. Advocates can launch information campaigns to increase awareness of an issue. An incumbent government can draw attention to its role in introducing policies the public may not previously have been aware of.
However, benevolent policies serve groups and address issues before political action is taken-when the group has low power and the issue has low visibility. Nutrition policies 5 are a classic example. Chronic malnutrition affects 149.2 million children worldwide (UNICEF/WHO/World Bank 2021). To address the issue, nutrition experts 6 have drawn upon scientific evidence and recommended that governments introduce nutrition policies which serve low-income women and children (a group with low political power) and address micronutrient malnutrition (an issue with low visibility. For the poor, the hardship of economic adversity, including the sheer mental and physical exhaustion accompanying life at the margins, constrains their ability to demand government action (Thawnghmung 2019). Concerned with daily survival, they rely on piecemeal responses rather than on organized approaches. Indeed, a history of inadequate access to social services can give way to "diminished expectations" reducing the likelihood that the poor organize (Holland 2018).
At the same time, many low-income women and children work in the informal sector in Asia, Africa, Latin America, and the Middle East, taking jobs as domestic helpers or street vendors, for example.
Informal workers are occupationally diverse and geographically dispersed, reducing their opportunities to organize and gain political power.
While there have been instances of women and informal workers mobilizing (Garay 2007), this has occurred around visible issues. By contrast, lowvisibility issues, such as micronutrient malnutrition, do not serve as effective rallying points. Micronutrients are essential vitamins and minerals such as iodine, Vitamin A, iron, and folate. Micronutrient malnutrition-the inadequate intake of essential vitamins and minerals-is commonly called "hidden hunger" or an "invisible" issue (Haddad 2013;Semba 2012).
Often people do not know their diet is insufficient. They might eat starchy staples, such as rice or potatoes, which provide the calories to stave off hunger but not the essential vitamins and minerals they need. As micronutrient malnutrition only becomes apparent when the problem is severe, accumulating over long periods, deficiencies can cause impairments but remain undetected. As a result, low-income women and children are unlikely to mobilize and demand improved nutrition policies.
Nutrition experts have developed effective interventions (Zlotkin et al. 2005) and identified optimal delivery methods (Baxter et al. 2014). However, governments will not necessarily adopt evidence-based recommendations. This is, in part, because the public is not typically aware of the government's role in improving micronutrient intake. For example, governments can introduce an iodized salt policy, adding iodine to pregnant women's diets, thus preventing irreversible physical and cognitive impairments in their children. However, the public does not often link government policies like salt iodization to their well-being. As it can take years to reap the benefits of nutrition interventions, governments are less able to claim credit for improving outcomes. In effect, the "costs of reforms are prompt and clear" but the "political rewards are delayed and uncertain" (Nelson 2004, 31).
Benevolent policies, such as nutrition policies, might be categorized as "easy" redistribution because they are fiscally cheap and politically uncontested (Holland and Schneider 2017). But expanding these policies is, in practice, hard as governments lack strong incentives to pursue reform. The public does not mobilize to demand expanded policies, and governments cannot use these policies to garner short-term political gains. If public policy is "whatever governments choose to do or not do" (Dye 1972, 18) then benevolent policies shed light on government inaction-the less studied aspect of public policy.
Yet certain governments have expanded benevolent policies, despite the absence of powerful groups agitating for reform or strong political incentives to pursue change. What explains this?

EXPLAINING POLICY EXPANSION
Policy expansion refers to intentional effort on behalf of the government to redress past omissions or address emerging problems. The government places the problem on the policy agenda, formulates new policy, and then issues the new or revised policy. For benevolent policies, where domestic societal pressures are weak, there is a need to "bring the state back in" to understand policy reform (Skocpol 1985). My work builds on classic scholarship on bureaucratic politics to carry out this analysis (Grindle and Thomas 1991;Heclo 1974;Orloff and Skocpol 1984;Rothstein 1998).
Government bureaucracies are increasingly staffed by technocrats-bureaucrats who have expertise in specific subjects. Technocrats have training in areas such as nutrition, computer science, economics, and education. Their professional training influences their ideologies. These ideas are powerful, shaping what they perceive as problems and appropriate solutions. Their ideologies become a guiding force behind their actions, motivating them to pursue specific policies. Their professional training also shapes their networks, connecting them with knowledge-based experts in IOs, as well as universities and nongovernmental organizations (Haas 1992).
In Brazil, computer science professionals were supported by fellowships and grants for overseas graduate training in the 1970s. After returning home, they joined the government and "for the sake of their ideas and ideology, elected to act as political and ideological 'guerrillas' within public institutions" (Adler 1986, 677). They sought to place their ideas on high-level agendas, ultimately leading Brazil to develop its domestic computer policy. In Chile, economics students were offered the opportunity to take courses at the University of Chicago in the 1950s and 60s. They returned to Chile, convinced that a competitive free market economy was the solution to developmental problems. The ideologically driven Chicago Boys, a group of neoliberal technocrats, designed economic policy during the Pinochet era (Silva 1991).
Technocrats, such as the guerillas in Brazil or Chicago Boys in Chile, fight for their ideas. These ideas, however, "do not float freely" (Risse 1994). Technocrats' ability to influence government policy comes not from the merits of their proposed policies or the strength of their convictions but, rather, from their access to executive decision makers, such as the head of state. Policy making is inherently political and what becomes policy must be examined in the context of power.
From the lower echelons of government, to mid-level management, to cabinet level positions, technocrats can be found throughout the bureaucracy. It is the latter, with their access to executive decision makers, who are best positioned to influence policy. In Brazil, the guerillas succeeded in influencing policy when their commission was placed in the Ministry of Planning, which had direct links to the president (Adler 1986). In Chile, the Chicago Boys were able to shape policy because they were senior members of government who enjoyed the support of the president (Silva 1991).
Technocrats may believe policy expansion is needed, but in the absence of political connections, it is unlikely that they can turn their ideas into policy. Executive support is essential for policy reform (Martínez Franzoni and Sánchez-Ancochea 2014;Rosenberg 1979). This is especially true for benevolent policies. And like social policies more broadly, benevolent policies address "wicked problems," which "require substantially greater coordination efforts than do relatively tame problems that fall neatly into the domain of a single government organization" (Peters 2018, 2). Education policies, for example, work best if students can access transportation to school and job opportunities after graduation. Similarly, the Ministry of Health must harmonize nutrition policies with agricultural and education efforts. Therefore, to pursue policy expansion, technocrats must convince actors in the executive office, as well as relevant ministries, to mobilize for reform.
I argue that once in their positions, technocrats continue to fight for their ideas, even in the face of opposition. Indeed technocratic expertise in the bureaucracy has long policy legacies (Pierson 2000). Technocrats will continue to advocate and train new generations to push for policy change. In the absence of direct access to executive decision makers, they will devise alternative methods to gain the ear of executive decision makers.
One strategy, which I highlight in this article, is the use of international pressure to construct a broader coalition that incorporates well-connected government officials into policy efforts. The coalition, with both expert knowledge and political connections, builds vertical networks to court leadership support. Executive support augments the coalition's power within the bureaucracy. The coalition uses this power, along with its horizontal networks, to pressure other ministries into cooperation. I show how this process of internal constituency building 7 facilitates policy expansion by introducing an analytical framework that looks outside the state, then inside the bureaucracy.

Looking Outside the State: International Organizations
Constructivist research shows how IOs use social processes to shape domestic policies (Avant, Finnemore, and Sell 2010;Dobbin, Simmons, and Garrett 2007;Risse, Ropp, and Sikkink 1999). For instance, IOs organize international meetings such as conferences, workshops, and forums. At these meetings, IOs put forward a core set of ideas that emphasize the need for governments to prioritize specific issues and policy solutions (Dion 2008;Weyland 2004). IOs assess governments on their fidelity to the recommended policies. Governments that adopt the policies receive praise, while those that do not are shamed, stoking reputational concerns (Kelley and Simmons 2015).
These processes of socialization can lead to policy change. For example, the United Nations Educational, Scientific, and Cultural Organization's efforts led to a wave of governments adopting new science policies, despite the absence of domestic demand for reform. The IO "took up science policy as a cause and… 'taught' states the value and utility of science policy organizations" (Finnemore 1993, 566). Similarly, I draw attention to the role of socialization.
However, I am not making the case for IOs as teachers of norms. Rather, I argue that IOs, through their efforts at socialization, serve as "coalition magnets" (Béland and Cox 2016) when there is "congruence" between IOs' efforts and technocrats' agendas (Acharya 2004).
The term, "coalition," typically refers to coordination between different groups to achieve a common goal (Kwon 2007;Nam 2015). For example, transnational advocacy networks have built new links between actors in civil society and IOs (Keck and Sikkink 1998). Similarly, human rights norms have facilitated the mobilization of both domestic and international groups (Black 1999). However, I modify the definition of coalition to refer to coordination between individuals of the same group. The coalition I describe in this article is primarily composed of actors within the government bureaucracy.
IOs draw attention to a range of policies. Their efforts are especially important for benevolent policies, where domestic pressures for policy expansion are weak. But IOs do not drive the domestic agenda. Instead, IOs' efforts at socialization serve as the focal point around which technocrats coalesce and mobilize like-minded individuals within the government bureaucracy. These individuals believe that policy expansion is needed, but their experiences have taught them that technical knowledge alone cannot facilitate reform. 8 They view international pressure as an opportunity to advance their agenda and use it to integrate knowledge-based experts from their formal and informal networks into their efforts.
Meanwhile, IOs are inviting high-ranking government officials to international meetings, with the purpose of socializing these individuals. Technocrats follow up with government officials and bring them into their network. I call this group "the coalition."

Looking Inside the Bureaucracy: The Coalition
In this scenario, technocrats are no longer acting alone. IOs have articulated similar ideas, granting legitimacy to their calls for policy expansion. Technocrats strategically deploy international pressure to facilitate the construction of a coalition, broadening the network of actors within the government bureaucracy involved in their cause. I emphasize the centrality of these networks, which encompass both formal and informal linkages, to explain policy expansion (Lee and Goodman 2001). These include vertical networks, which connect technocrats to executive decision makers, and horizontal networks, which cut across ministries.
The coalition leverages international pressure, in conjunction with its expanded networks, to gain the support of executive decision makers. Studies have documented the use of this strategy by activists outside the government. For example, the Council of Europe (CoE) sought to push governments to change their views of citizenship and minority rights. German activists, who shared the CoE's ambitions, used CoE efforts as a "tool for generating pressure on political elites in Bonn" (Checkel 1999, 105).
Similarly, the coalition uses IO efforts to place pressure on executive decision makers within their own government. As social policy issues "in health and education can rarely capitalize on a sense of urgency" (Kaufman and Nelson 2004, 13), international pressure becomes a tool for insisting on immediate action. The coalition draws upon its political connections to set up meetings with executive decision makers. Citing international scrutiny and reputational concerns, the goal is to get the head of state's support in policy expansion.
Executive support signals that policy expansion is a priority (Grindle 2004). The coalition then uses executive decision makers' commitment to policy expansion, in conjunction with international pressure, to create a sense of urgency for reform and demand interministerial cooperation. This is a dynamic process, shaped by interactions over time. IOs play a role in structuring the international environment, but their efforts alone do not decisively create new domestic policies. Their attempts at socialization may draw attention to the importance of certain policies, but the extent to which this leads to policy change is constrained by the domestic context. It is technocrats, who work inside the bureaucracy and understand the complexities of the policy-making process, who play a crucial role in facilitating coordination within the bureaucracy (Figure 1).
Here, I lay out the scope conditions for the analytical framework, specifying the conditions under which it has more, or less, explanatory power. For benevolent policies, policy expansion is likely to have its origins in the bureaucracy, given that domestic societal pressures are weak. Technocrats who believe there is a need for reform play a critical role. They are better positioned to facilitate policy expansion when they have access to executive decision makers. In the absence of direct access to leadership, technocrats can devise alternative methods for obtaining leadership support, such as constructing a coalition with politically connected members. This is true for all welfare states, truncated or not.
However, technocrats are more likely to emerge and wield influence in certain contexts. When a bureaucracy is staffed by highly trained technocrats with longterm careers, they are more likely to emerge to champion reform. Their higher level of training provides them with the skills needed to identify and address societal problems while also socializing them into a certain way of thinking. There is an increased chance that they belong to the same knowledge-based communities as their IO counterparts. This overlap increases the likelihood that technocrats share similar ideologies and view IO efforts as "congruent" with their own agendas. With extended time horizons, these technocrats see the value of benevolent policies, which pay off in the long run. These technocrats are better able to use IO pressure to garner executive attention if governments are, or recently have been, foreign aid recipients. Governments that have accepted aid are enmeshed in complex relationships with IOs and thus more susceptible to external influence.

METHODS, CASE SELECTION, AND DATA
This article uses the case study method to explain policy expansion. While large-n studies have the advantage of generalizability, case studies are "particularly useful in explaining cases that do not fit an existing theory, in order to explain why the case violates theoretical predictions and to refine or replace an existing hypothesis" (Levy 2008, 5). As benevolent policies have been overlooked in the literature, hiding in plain sight, I use the intensive study of a single case to carry out theory building.
The Indonesian government, and its sudden introduction of new nutrition policies, is the ideal case to study. Conventional explanations-including theories of industrialism, democracy, and the power resource theory-do not offer persuasive accounts for the policy expansion that took place in 2012.
Studies that fall under the broad umbrella of industrialism suggest that economic growth occurs in tandem with the expansion of social policies (Rimlinger 1971;Wilensky 1974). But while the Indonesian government experienced rapid economic growth from 2000 onward, nutrition policies remained inadequate. For more than a decade, the government had the resources to address malnutrition but did not take significant policy action. This suggests that economic growth was not responsible for reform.
Another prominent strand of literature links democracy to the expansion of the welfare state (Haggard and Kaufman 2008;McGuire 2010;Wong 2006). Electoral competition creates incentives for policy makers to respond to low-income citizens' needs, facilitating the expansion of redistributive policies. However, democracy does not change a government's incentive to address issues with low visibility. In the 2000s, Indonesia's democratization was accompanied by the expansion of pro-poor policies-but also an intensified focus on visible issues. The government prioritized policies with electoral returns such as cash transfers. Although these could indirectly improve nutrition, the focus on visible issues sidelined micronutrient malnutrition. Therefore, theories of democracy cannot explain policy expansion.
Another body of research, organized around the power resource theory, links partisan ideology to policy expansion (Esping-Andersen 1990;Korpi 1978). In the 2000s, the "left turn" in Latin America was accompanied by the expansion of social policies, suggesting that left-leaning parties had a decisive influence on the introduction of redistributive policies (Huber and Stephens 2012). However, the left-right divide is largely imperceptible among Indonesian political parties (Fossati et al. 2020). Another element of the theory links union mobilization to reform. In Indonesia, however, poor women and children often work in the informal sector and are unlikely to unionize, reducing the power resource theory's analytical purchase.
To investigate alternative drivers of policy change, I carry out inductive process tracing, using empirical observations to derive propositions and formulate sequences (Hall 2013). This approach is "often used for the purpose of theory development through the identification of key events and through the specification of hypotheses about how these events connect together" (Falleti and Mahoney 2015, 229). This article's primary data draw upon nine months of fieldwork conducted during 2014-16 as part of a broader research project on the politics of welfare state expansion. 9 I carried out 71 in-depth, semistructured interviews in Jakarta, Indonesia (2015) as well as Bangkok, Thailand (2014), Manila, Philippines (2015), Phnom Penh, Cambodia (2016), and Vientiane, Laos (2016). This process involved speaking with policy makers in governments, nutrition specialists affiliated with IOs, and representatives from civil society on the nutrition policy-making process. I interviewed individuals from groups with varying mandates to verify claims. In addition, I carried out informal observations in the nutrition section of an IO over three months (2014) and at three closed-door regional nutrition meetings in Southeast Asia (2014-15), observing interactions between representatives from governments, IOs, and civil society. In the next section, I draw upon these data to show how technocrats facilitate policy expansion.

NUTRITION POLICIES IN INDONESIA
In the late 1800s, a mysterious illness was ravaging the Asia Pacific region. Beriberi, which affects the cardiovascular and nervous system, was initially thought to be an infectious disease. Fearing threats to its colony, the Dutch government sent scientists to Indonesia to investigate. They discovered that beriberi was the result of a vitamin deficiency, not a pathogen. 10 This finding propelled Indonesia's program of research in nutrition forward. The Special Institute for Nutrition Research was established in Indonesia, providing advice to the government. The Institute carried out the nutrition program for the Ministry of Health in the 1950s and later became a division of the national government. Around this time, Indonesian universities began adding nutrition sciences to their curriculum, increasing the country's technical capacity.
More broadly, opportunities for advanced training would continue to grow in Indonesia. The Ford Foundation, for example, began sending students overseas and many studied economics at the University of California at Berkeley. By the early 1960s, these students returned to Indonesia. They took up academic positions and later, positions in the government where they would become known as the Berkeley Mafia.
In 1967, after a bloody power struggle, General Suharto became president of Indonesia, ushering in a period of authoritarian rule. To modernize the bureaucracy and respond to development challenges, technocrats came to the fore of his New Order regime. To be sure, this did not turn the bureaucracy into an apolitical institution. Nor did it result in a policy-making process free from corruption. But as part of "a wish to 'catch up' with the West," the government sought to develop technical expertise and was supported by donors such as the World Bank and International Monetary Fund (Wicaksono 2018, 631).
In the authoritarian New Order government, the Berkeley Mafia held key ministerial positions, heavily influencing policy. Widjojo Nitisastro, the most prominent member of the Berkeley Mafia and one of Suharto's most trusted advisors, was appointed head of Bappenas (the National Planning Agency). 11 Bappenas was a powerful agency, holding unassailable authority over development planning and budgeting. There, Nitisastro established an intergenerational network of technocrats. Indonesian students, including those specializing in nutrition, continued to study abroad, and upon their return they joined universities and the government.

Policy Expansion in the 1970s
The expansion of nutrition policies in the 1970s illustrates how Nitisastro, who supported nutrition policies and had access to the president, shaped policy reform. While Nitisastro was a market-oriented economist, he nevertheless advocated for government intervention to achieve development goals (Chalmers 1997). In his keynote speech at the National Institute of Sciences in 1970, he declared "we must not operate under the illusion that economic development will in itself cause development in other sectors" and indicated there was a need for the government to expand social policies (Soekirman 2008).
At Nitisastro's instruction, nutrition was included in national development planning in 1978 (Soekirman 2008). In 1979, the government expanded its nutrition policies, introducing the Family Nutrition Improvement Program (Usaha Perbaikan Gizi Keluarga or UPGK). Many of the program's interventions, such as vitamin A supplementation, iron-folate tablets, and nutrition education and counseling, produced limited short-term political gains.
Policy expansion, then, can be traced to the executive support. 12 Technocrats in Bappenas advocated for the introduction of nutrition programs (Heaver 2005). And these technocrats, especially "top decision makers from Bappenas," were crucial for establishing "an interministerial national program in nutrition" (Rohde and Hendrata 1982, 227). Reflecting on this, a technocrat working in the bureaucracy in the 1970s noted the importance of Nitisastro's connections to the president for elevating nutrition on the government agenda (Table 1). 13 The Family Nutrition Improvement Program continued until the 1990s.

Policy Retrenchment in the 1990s and 2000s
In 1993, Suharto announced the sixth development cabinet. The leadership of Bappenas 14 as well as several ministerial posts, previously held by Nitisastro's protégés, were taken over by individuals who sympathized 10 Christiaan Eijkman, the Dutch director of the lab, discovered that beriberi was caused by a deficiency in Vitamin B1 (thiamine). He won the Nobel Prize for his discovery. 11 Badan Perencanaan Pembangunan, commonly shortened to Bappenas.
12 A similar dynamic can be seen in the expansion of family planning policies, which had the support of a technocrat with access to the president. Widjojo Nitisastro completed his PhD at Berkeley between 1957 and 1961. In his dissertation, he calls "for 'a massive development effort' and an enhanced family planning programme to provide employment and reduce population pressures" (Chalmers 1997, 55). In 1967, Nitisastro was appointed head of Bappenas. In 1968, the government rapidly expanded family planning policies. Technocrats attributed policy expansion to the cultivation of presidential commitment and their ability to coordinate supporting ministries' efforts (Shiffman 2004). 13 Personal Interview, October 8, 2015, Jakarta, Indonesia. 14 Ginandjar Kartasasmita took over leadership of Bappenas.
with Bacharuddin Jusuf Habibie's ideas on development. Habibie had held a number of key bureaucratic positions in the Indonesian government and had a close personal relationship with the president (Amir 2008).
In contrast to Nitisastro, Habibie was an aeronautic engineer by training and placed industrial sectors at the core of development. But "concentrating on hightechnology development inevitably entails marginalizing the development of other sectors more beneficial for the populace as a whole" and this "failed to address social issues such as poverty and unemployment" (Amir 2008, 320).
In 1997, Indonesia was hit by a massive currency depreciation. As the economy plummeted, the government had no choice but to accept IO loans and the accompanying policies, including social safety nets. The government introduced the JPS program (Jaring Pengaman Sosial), a social safety net program, with the World Bank, Asian Development Bank, and bilateral donors providing budget and program support (Sumarto, Suryahadi, and Widyanti 2002). The JPS program was intended to mitigate the adverse effects of the crisis including poverty, unemployment, and malnutrition. However, nutrition efforts largely ignored micronutrient intake. Indeed, the "inclusion of too many other health issues may have diluted the nutrition interventions" (WHO 2012).
Amidst mass protests, Suharto resigned as Indonesia's president in 1998, handing over office to Habibie. This marked the country's transition to a democratic system of governance. By this time, the executive decision makers who had supported nutrition policies had been replaced or were losing power. Nitisastro's last formal position with the government's Economics Assistance Team ended in 2001 (Table 1).
At the same time, democracy, long hailed for its potential to facilitate the expansion of the welfare state, was not pushing nutrition policies to the forefront of Indonesia's policy agenda but was instead sidelining it. While the extension of the voting franchise increased the power of low-income groups, it sidelined lowvisibility issues. Politicians had a better chance at securing office if they addressed high-visibility issues, and they relied on policies like JPS and cash transfers. JPS was "a key plank" in Habibie's attempt "to survive politically … at a time when the first post-Suharto elections were looming" (Aspinall 2014, 812). The following two administrations 15 continued to reach out to poor voters in the same way, as low-income groups favored parties that demonstrated concern for social welfare (Mujani and Liddle 2010).
In the 2009 elections, Susilo Bambang Yudhoyono, the then president, campaigned for reelection on a pro-poor platform. It was "the introduction of massive cash programs for the poor that triggered Yudhoyono's meteoric rise from electoral underdog to almost unassailable frontrunner" in 2009 (Mietzner 2009, 4). Even as the economy rebounded and began to skyrocket, social policies focused on high-visibility issues.
The Indonesian economy rumbled along on high for years. Yet amidst significant economic growth and poverty reduction, malnutrition persisted. According to the World Bank (2021), 42.4% of children in the newly democratized country were affected by chronic malnutrition in 2000, as measured by rates of stunting (low height-for-age; Figure 2). A decade later, the numbers had hardly improved and 39.2% of children were malnourished in 2010.
As nutrition policies had not reemerged as a political priority, observers remained concerned. "Progress on reducing stunting rates has not been as good as on other social development indicators in Indonesia," they lamented (Rokx, Subandoro, and Gallagher 2018, 19). In the scramble to introduce policies that would secure political office, nutrition policies had been left behind.
The UNICEF office in Indonesia observed that malnutrition remained a stubborn issue, in part because "stunting and maternal nutrition are not easily visible. People are generally not aware that malnutrition is a problem" (UNICEF Indonesia 2012, 2). This made it difficult to get the issue onto the national agenda. In interviews conducted in 2015, interviewees agreed, attributing the difficulty of getting nutrition policies prioritized to low visibility. As a technocrat explained, Nutrition problems is unseen. If someone suffers from anemia, we cannot see it right? This is why [the] nutrition problem is so difficult… . We are struggling to convince everyone that nutrition is very important. 16 As emphasis shifted to visible health policies, nutrition policies fell to the wayside. This, I contend, contributed to the dismantling of Indonesia's Family Nutrition programming, with its low visibility, was slowly dismantled and replaced by more visible health services. And because technocrats advocating for nutrition no longer had the ear of the president, malnutrition continued to persist on a large scale. Government attention never reached the same levels as in the 1970s. Commenting on this, an employee of a bilateral donor in Indonesia observed that since democratization, "everything has stalled." 19

Policy Expansion in 2012
Against this backdrop, technocrats trained in nutrition continued to advocate for policy expansion. 20 But nutrition did not become a priority on the government agenda, even as the Indonesian economy recovered from the economic crisis and began to grow rapidly. It was not until IOs began concerted efforts to draw attention to malnutrition that the Indonesian government again expanded its nutrition policies. In 2010, the World Bank introduced the Scaling Up Nutrition Framework for Action at its spring meetings in Washington, DC. This laid the foundation for the "Scaling Up Nutrition" movement (SUN). IOs, including the World Bank, UNICEF, WHO, and Food and Agriculture Organization (FAO), renewed their efforts to pressure governments to expand nutrition policies.
Operating under the auspices of the UN, SUN began recruiting governments to join the movement and commit to "developing or revising national policies, strategies and plans to scale up nutrition" (SUN 2018, para. 3). Indonesian technocrats saw this as an opportunity to advance their agenda. They had been pushing for the prioritization of nutrition for years, yet it continued to be sidelined. SUN does not have legal status, nor does it provide financial resources to governments that commit to policy reform. Instead, SUN places pressure on governments by drawing attention to the need for expanded policies that serve poor women and children and address micronutrient malnutrition. In 2011, SUN organized a meeting at the United Nations General Assembly in the United States, emphasizing the prevalence of malnutrition and the need for government 17 The data were smoothed by removing an outlier in . 18 Personal Interview, October 20, 2015, Jakarta, Indonesia. 19 Personal Interview, October 28, 2015, Jakarta, Indonesia. 20 Personal Interview, October 8, 2015 action (SUN 2011). Over the next few years, their efforts would build to a crescendo. Indonesian technocrats coalesced around SUN and began reaching out to individuals sympathetic to their cause within the bureaucracy. They formed a coalition, calling themselves "the Sunners." 21 Over time, the coalition began to gain attention. Within the bureaucracy, Indonesian technocrats working on social policies, but not specifically on nutrition, acknowledged the emergence of the coalition. 22 Outside the bureaucracy, representatives working for an IO in Thailand,23 in addition to nutrition consultants based in the US, observed how the coalition had formed in response to SUN. 24 A Sunner, whose early calls for nutrition reform had gone unheeded said, "nutrition in a country is influenced by the world and its perceptions of the issue," and "now, we have started moving, because of the international." 25 Central to the Sunners' plan for policy reform was convincing the president to join SUN. However, the Sunners did not have direct access to him. To reach the president, a Sunner drew on her connection to the Minister of Health, Endang Rahayu Sedyaningsih, who had attended the 2011 SUN meeting. As the Sunner explained, A couple of months after that [the 2011 SUN meeting], I met her [the Minister of Health] and I asked her whether we will participate in the SUN movement or not. She said, "yes, we will." But at that time she was not sure how we could send the interest to the SUN Global Secretariat… . So we waited for a while and then we met again. And again, I asked her whether she is willing to write a letter to the UN. 26 The Sunner kept pressing the Minister of Health. Eventually, the Minister of Health agreed and obtained approval from the President, who sent "the letter to Sec-Gen Ban Ki Moon, saying that Indonesia was willing to be part of the movement." 27 After the letter was sent to the Secretary-General, a Sunner said, "I initiated a meeting of the senior nutritionists in Indonesia and I said to them we need to establish a policy framework. We cannot only involve in the movement without having a policy framework in place." She used Indonesia's commitment to SUN to divert resources to the formulation of new nutrition policies. As the Sunners gained momentum, the coalition grew to include Indonesians based at universities and nongovernmental organizations.
The Sunners were keen to demonstrate progress to IOs. This was especially true for policy reform, as SUN evaluates governments on their national policies and then compares them with other governments. Government officials attended meetings every six weeks by telephone to report progress updates. A Sunner said, "We were monitored by the [SUN] Secretariat through the teleconference, and the teleconference was done every six weeks. So it was very intensive." 28 As such, they diverted resources to formulating policies.
The Sunners used this pressure to keep policy efforts on track. This facilitated the rapid drafting and introduction of nutrition policies. By this point, the Sunners had generated momentum by obtaining commitment from the president and developed policies that adapted international recommendations to the Indonesian context. While these policies created basic guidelines, the Sunners needed buy-in from different ministries to set up policy implementation (Sardjunani and Achadi 2016). Describing past attempts to introduce policy change, an Indonesian national working for an IO in Jakarta said, "there is a Coordinating Minister but anything that needed coordination in the past did not work." 29 The Sunners knew this. But this time, they could use the international pressures to insist on the necessary coordination. Because the president had committed to SUN and international attention remained focused on nutrition policies, they could push recalcitrant ministries to cooperate. At the same time, they were drawing on their informal networks and reaching out to government officials with whom they had existing relationships.
The Sunners invoked the Indonesian government's commitment to SUN to convene high-level meetings and presented internationally recommended policies as "best practices" to avoid internal conflict. 30 They brought drafts of the newly formulated policies to show that the core work had been completed, and they emphasized the need to introduce new policy, given the external attention on Indonesia's high levels of malnutrition.
In 2013, the president launched Presidential Decree 42, the National Movement to Accelerate Improvements in Nutrition (Peraturan Presiden Republik Indonesia Nomor 42 Tahun 2013 Tentang Gerakan Nasional Percepatan Perbaikan Gizi). The presidential decree created a taskforce within the government spanning several ministries, which would coordinate efforts to improve nutrition. The interministerial taskforce, while slow to get started, reported to the president and convened 13 government ministries and UN agencies on malnutrition.
Later, Indonesia's 2015-19 National Medium-Term Development Plan (Rencana Pembangunan Jangka Menengah Nasional or RPJMN) recognized nutrition as requiring the cooperation of different ministries and incorporated the nutrition policy framework. New indicators, introduced through these policy changes, were harmonized with the 2011-15 National Food and Nutrition Action Plan (2011-15 Rencana Aksi Nasional Pangan Dan Gizi or RAN-PG).
The expansion of nutrition policies demonstrates how technocrats used IO pressure to augment their political power within the bureaucracy then obtained executive support and mobilized relevant government officials for policy expansion (Table 1). By 2015, there was a buzz of excitement-that this might be the beginning of improved nutrition in Indonesia. 31

CONCLUSION
At the heart of all policy making is a tension between technical expertise, on one hand, and politics, on the other. Evidence-based policy solutions may exist and have the potential to dramatically improve lives but fail to become policy if politics do not work in their favor. This is exemplified by benevolent policies. The concept makes legible a host of policies affected by government inaction. For instance, refugees and migrants often lack state recognition, making it difficult to accumulate resources or organize. Recently the WHO drew attention to the "invisible" epidemic of noncommunicable diseases (NCDs) in Europe. Refugees and migrants often have higher rates of certain NCDs, yet governments have been slow to introduce relevant policies. Similarly, sanitation policies in Indonesia have been plagued by inaction. "Citizen demand is insufficient to catalyze government action" and governments "are not poised to take credit" (Winters, Karim, and Martawardaya 2014, 33). The benevolent policies framework explains why this has resulted in inadequate sanitation policies.
The benevolent polices framework can also help us anticipate government inaction. COVID-19, for example, has predominantly affected low-income and racialized individuals in the US. Many have recovered, but an estimated 15 million people will be affected by long-haul COVID, experiencing a range of debilitating symptoms months after being infected (Phillips and Williams 2021). Like other postinfection syndromes, such as myalgic encephalomyelitis/chronic fatigue syndrome, the medical community has been ambivalent about recognizing the illness, lowering public awareness. The benevolent policies framework suggests that the government will be slow to introduce relevant policies.
Under certain conditions, however, governments do act. For example, in Canada, low-income citizens are almost twice as likely to suffer from poor oral health compared with high-income citizens. While these individuals rarely organize to demand action on oral health, the government introduced water fluoridation policies that have reduced tooth decay in children.
How do these "good policies" overcome "bad politics?" In this article, I focus on nutrition policies in Indonesia. I foreground the bureaucracy, an explanatory variable that often slips into the background. Technocrats, I argue, can be an important catalyst for policy change. Even after policies are hollowed out, technocrats remain within the bureaucracy and can serve as the spark behind policy reform. I show that technical expertise has long policy legacies. Technocrats continue to fight for their ideas, even in the face of policy retrenchment, and leverage alternative strategies to orchestrate policy reform.
I draw on both the comparative politics and international relations scholarship, weaving together research from the two subfields, to make my argument. Technocrats deploy IO pressure to build a coalition and broaden their political connections within the bureaucracy. Technocrats, by harnessing international pressure, shape the channels for information flow, building vertical networks to elevate the policy onto the executive agenda and horizontal networks to coordinate different ministries for policy change.
Therefore, I show how policies get initiated, formulated, and introduced-even when domestic pressure for reform is weak. In doing so, I am adding analytical precision to the welfare state literature, which has paid limited attention to the government bureaucracy.
When I spoke with interview participants, there was a sense of optimism about the new policies that were introduced in 2012. However, these high hopes must be tempered by a broader view of the Indonesian state. As a Sunner told me, "policy implementation is our next big challenge." 32 Indonesia faces governance challenges, including the capture of state institutions by 30 Personal Interview, October 28, 2015, Jakarta, Indonesia. 31 Personal Interview, October 8, 2015Personal Interview, October 24, 2015, Jakarta, Indonesia;Personal Interview, October 28, 2015, Jakarta, Indonesia. 32 Personal Interview, October 26, 2015, Jakarta, Indonesia. various elites. Despite efforts around democratization, corruption remains the norm rather than the exception and administrative malfeasance is a perennial issue. Nevertheless, the country appears to be headed in a better direction. After expanding nutrition policies in 2012, chronic malnutrition declined, with stunting rates dropping to 36.4% in 2013, then declining to 30.8% in 2018 ( Figure 2). Research indicates that in Indonesia elite capture is not inevitable (Alatas et al. 2012). However, the decentralized nature of the Indonesian state suggests there is likely to be variation in policy implementation (Shair-Rosenfield, Marks, and Hooghe 2014). Further work is needed to explore the influence of nonstate actors, such as religious groups (Cammett and MacLean 2014;Cammett and Şaşmaz 2021), shared identity (Singh 2015), and social network structures (Cruz, Labonne, and Querubín 2020) on nutrition policy implementation and subnational variation.
Future research on benevolent policies can illuminate the conditions under which governments act, when clear, short-term political gains are limited. As I show in this article, work in this area will complement the rich and dynamic scholarship on the factors that frustrateor facilitate-welfare state expansion.

SUPPLEMENTARY MATERIALS
To view supplementary material for this article, please visit http://doi.org/10.1017/S0003055421000927.

DATA AVAILABILITY STATEMENT
Research documentation and data that support the findings of this study are available at the American Political Science Review Dataverse: https://doi. org/10.7910/DVN/MEWHPE.