I. INTRODUCTION
The relationship between democracy and bureaucracy has been figured as fraught for a long time. A century ago, Max Weber’s classic account concluded that although “[b]ureaucracy inevitably accompanies modern mass democracy,” the two also “inevitably … conflict”: bureaucracy’s tendency to empower “a closed status group of officials” who prefer to work in secret grates against democracy’s promotion of “equal rights” and reliance on “public opinion” (Reference Weber1978, 983–85, 48; see also Espeland Reference Espeland2000). A century before that, Alexis de Tocqueville’s famous study of a young America connected its democratic success to its lack of effective central administration (Ernst Reference Ernst2014, 1; see de Tocqueville Reference de Tocqueville, Mansfield and Winthrop2000). And in the present, the image of unelected bureaucrats obstructing the will of the people has become commonplace in a range of democracies, from members of the European Union to the United States (Hamburger Reference Hamburger2014).
At the same time, democratic aspirations do not simply implement themselves; any modern polity needs some administrative apparatus. That apparatus, moreover, is not an “unthinkable,” ungraspable abstraction (Bourdieu Reference Bourdieu, Fernbach, Cambridge and Press2014, 3). Although its workings may look murky, the administrative apparatus of the state is a social entity: an arena of “human activity” (Dewey Reference Dewey and Boydston1984, 241; see Novak Reference Novak2015). The administrators who populate that arena, who make the state run, are thus key players in the life of any modern democracy. Given how contested their role in democratic governance has been, it is striking how little bureaucrats themselves are consulted on the issue.
Using interviews and participant observation with bureaucrats, I do just that. Locating my inquiry in the government of Taiwan, I ask how administrators understand their place in a democratic government and what, in their view, renders their actions legitimate.Footnote 1 To be clear, I do not assess whether my subjects comport with some predefined notion of legitimacy. Rather, I seek local understandings of what constitutes legitimate political action, specifically asking how democratic legitimacy looks to those charged with producing it. Eliciting the understandings of those who populate the government, I see bureaucracy “from the native’s point of view” (Geertz Reference Geertz1974).
Taiwan is one of the few success stories among “third wave” democracies (Huntington Reference Huntington1991; Carothers Reference Carothers2002). With no history of democracy and no internationally recognized sovereign status, Taiwan moved quickly to achieve the hallmarks of democratic governance in the 1990s and has avoided the “illiberal” turn of a number of recently democratized countries (Zakaria Reference Zakaria1997). Freedom of press and association, free and fair contested elections, interparty transfers of power, a professionalized civil service, and civil liberties protections were established quickly and firmly.Footnote 2 Both exemplar and outlier, Taiwan illuminates how bureaucracy can adjust to democratic transition. And it can help us understand how democracy can flourish, and what it can look like, even in unlikely locations (Paley Reference Paley2008).
My research subjects—administrators at different levels of a range of central government and Taipei City agencies—root their own legitimacy in the porous quality of their bureaucracy. They present their work as an ongoing collaboration with others, especially legislators, scholars, and other members of the public. For my subjects, administrative accountability involves not just executing the legislature’s commands but producing them in the first place. And they present bureaucracy as a key site for political participation by people outside the government.
I do not address the efficacy of Taiwanese government or gauge the level of its populace’s satisfaction with it. Rather, I ask how those who make the government work understand it and conceive of its legitimacy. How governors understand governing may be a particularly crucial question for Taiwan, which cannot claim the mantle of recognized de jure sovereignty to obscure the contingencies of its de facto self-governance (Markell Reference Markell2003). Over the last few decades, the government of the People’s Republic of China (PRC), across the Taiwan Strait, has become increasingly adamant in claiming ownership over Taiwan, an island it has never ruled. Government administrators—like everybody else—know that Taiwan’s continued self-determination hangs on its continued recognition as autonomous, if not technically sovereign, by both domestic and international audiences.
Taiwan, then, is a polity that must continually reassert and justify its existence as a polity. That fact highlights in an ongoing way the existential ramification of political legitimation.Footnote 3 But my findings here have implications for less precarious places, too: in a democracy, after all, governance is always contingent on local legitimation. The Taiwanese administrators I study consistently emphasize the porous, interactive nature of their governing apparatus, in contrast to the kind of “state effect” that produces an illusory separation of state from society (Mitchell Reference Mitchell and Steinmetz1999). My interlocutors ground state legitimacy not in some imagined autonomy from society but in the opposite. Government legitimates itself insofar as it participates in the world as one social actor among others, in relationships of mutual influence, communication, and responsibility.
This vision encourages us to question whether bureaucracy and democracy are really as opposed as some make them out to be. I posit instead that the role bureaucracy plays in democratic governance will differ across cultures and contexts. Here, I explore how that role is understood within a democratic administration itself. My interlocutors’ celebration of the bureaucracy’s interdependence with other branches contrasts with some common assumptions about the proper separation of powers in presidential political systems (Ackerman Reference Ackerman2000). Their reliance on constant interaction with multiple private actors may be surprising in a government characterized for decades by the executive supremacy and tight links to large economic players typical of developmental states (Johnson Reference Johnson1982; Öniş Reference Öniş1991; Evans Reference Evans1995). And their conviction that the bureaucratic production of statutes legitimates both the bureaucracy and the resulting legislation problematizes assumptions about principals and agents in the administrative state (Daniels Reference Daniels2014).
II. INTERNAL ADMINISTRATIVE CULTURES OF ELEVATOR-LEVEL BUREAUCRATS
Scholarship about administrative agencies often focuses on how courts evaluate administrative action, as well as on how legislatures mandate and constitutions authorize it (Mashaw Reference Mashaw2005, Reference Mashaw2007; Rose-Ackerman and Lindseth Reference Rose-Ackerman and Lindseth2010; Boughey Reference Boughey2013; Bignami and Zaring Reference Bignami and Zaring2016). In this literature, administration thus gains legitimacy from other government institutions. In contrast, some have urged greater attention to the “internal law of administration” through which bureaucracies structure their own work (Mashaw Reference Mashaw2012; see Wyman Reference Wyman1903). Some have even argued that, given the size and complexity of modern states, this internal law of administration may present the most important constraint on administrative power (Katyal Reference Katyal2005). On this view, internal strictures are as legitimate as, and more effective than, judicial review for setting standards for the exercise of discretion and ensuring the use of expertise in decision making. Such “processes, guidelines, … polic[ies,] internal management structures,” and “[t]ransagency … oversight and coordination” mechanisms produce the everyday of administrative action—and its justification—more than the exceptional moments inscribed in sporadic legislative commands and judicial rulings (Metzger and Stack Reference Metzger and Stack2017, 1248). Other scholars have proposed greater attention to the larger and more differentiated sphere of influence that provides the context for administration. This view moves out of the bureaucracy to place administrators in multinodal “accountability networks” that include “elected officials, organized interests, the courts, and the general public” (Bignami Reference Bignami2011, 861). This work presents administrators as constrained by multiple nodes of influence, often in tension with one another.
These two approaches enrich our understanding of administration and the way it functions in its particular contexts. They present bureaucracy as both a social field with its own internal regularities and as an active participant in larger social worlds (Bernstein Reference Bernstein2017). Building on this research, I look both to the inner life of administration and to its outward place in larger social and historical contexts. And I view internal administrative law and external accountability networks as aspects of administrative culture (see Meidinger Reference Meidinger1987).
By culture, I mean a dynamic, historically situated foundation for the construction of meanings and values (Silverstein and Urban Reference Silverstein, Urban, Silverstein and Urban1996; Gal Reference Gal2005). I do not take culture to be something that can be captured in a text (Agha Reference Agha2003); nor does it produce unchanging values (Langbein Reference Langbein1997; Eppinger Reference Eppinger2013). Rather, it forms an evolving basis for the intuitions and evaluations that influence understandings of the world and provides the terms on which they change. Culture matters to understanding bureaucracy and democracy: cultural contexts can imbue even similar phenomena with different meanings and values for their participants. For instance, it may be quite common for people to avoid recourse to legal remedies. But that avoidance may sometimes mean that people prefer to “lump” rather than contest negative experiences (Engel Reference Engel1984, Reference Engel2005), while in another context it may instead show that people contest negative experience through channels other than law (Bernstein Reference Bernstein and Trautner2018; see also Chuang Reference Chuang2004).
Similarly, administrative action gains meaning not just through the structural relations of government but through participants’ own understandings of it. This is because administrative agencies are staffed not by offices or positions, but by people. In contrast to a tradition that viewed bureaucracy as an object of social commentary rather than as a social field itself (Herzfeld Reference Herzfeld1992), recent scholarship has explored how bureaucratic practices, identities, and values are forged through interactions with a range of people and processes (Hoag Reference Hoag2010; Latour Reference Latour2010; Hull Reference Hull2012; Cabot Reference Cabot2014; Thelen et al. Reference Thelen, Vetters and Benda-Beckman2014; Fassin et al. [2013] Reference Fassin, Brown and Fassin2015; Mathur Reference Mathur2016; Hoag and Hull Reference Hoag and Hull2017). Administrators’ relationships to their own positions and their own work forms a key vector of an administrative accountability network and a key aspect of internal administrative culture. Moreover, this is true not only of the “street level bureaucrats” (Lipsky Reference Lipsky1971) who form the usual object of cultural study, but also of the regulators who formulate the regulations and policies that implement democratic decisions—what we might call the elevator-level bureaucrats whom I present here. These administrators enact the power of the state not by deciding the fate of some particular supplicant but by setting general terms of conduct. How they understand the legitimacy of their own actions thus has implications for how governance is done.
III. CONSTITUTION, ELECTION, ADMINISTRATION—AND DEMOCRATIZATION
Taiwan’s government administration has undergone significant transformations within the lifetimes of most of my interlocutors. That historical trajectory, in turn, informs administrators’ understanding of bureaucracy and its place in democratic government. I focus here on three areas: the Constitution, popular elections, and the administrative apparatus itself. Key to this history, bureaucracy provided a key site for the development of democracy.
The Constitution
While the existence of a constitution can function as an indicator of democratic governance in some circumstances, Taiwan’s Constitution was for decades a means of entrenching dictatorship. When the Allies expelled Japan from its colonies in 1945, they gave authority over Taiwan to the Republic of China, then under the control of the KMT party (also known as the Chinese Nationalist Party, Kuomintang, and Guomindang), which was embroiled in a long civil war with communist forces. Although the two sides managed to produce a constitution in 1946, the KMT prevented its application on Taiwan (Kerr Reference Kerr1965; Wang Reference Wang2002; Phillips Reference Phillips2003). After losing the civil war in 1949, the KMT establishment fled to Taiwan, where it continued to claim legitimate authority over China, despite having no actual power there, for decades.
This claim justified a slew of other measures. It underlay the Temporary Provisions Effective during the Period of Communist Rebellion, which displaced constitutional provisions that would facilitate democratic rule (Ginsburg Reference Ginsburg2002b). It made it unthinkable to amend the Constitution to reflect its actual geographic reach, which would imply giving up the claim to China (Yeh Reference Yeh1990). “It was as if the Constitution were frozen,” Weng Yueh-Sheng told me, using a common phrasing.Footnote 4 Weng is widely considered the godfather of Taiwanese administrative law. “People couldn’t talk about it, and I was unwilling to teach constitutional law in Taiwan at that time.”Footnote 5 And it gave the KMT a reason to suspend national elections, since most of the Chinese electorate lived under communist rule (Yeh Reference Yeh1990). The Constitution would not go into effect for decades, and when it did, it bore the traces of its creation in a place it did not govern, by parties vying for dictatorial control—in Cheng-Yi Huang’s words, an “unpopular sovereignty” (Reference Huang2017, 117). For decades, then, the Constitution instantiated the dictatorship’s determination to prevent meaningful popular participation in electoral politics.
Elections
Similarly, elections were a key tool of dictatorial rule in recent memory; as a young scholar and activist told me, he had grown up knowing that “elections were fake.”Footnote 6 The KMT dictatorship ruled out national-level elections, but still sought to legitimize its “mobilizational authoritarianism” by offering an alternative to the Communist People’s Republic (Rigger Reference Rigger1999, 8). Lacking a social infrastructure on Taiwan, which had been a Japanese colony during the KMT’s formation and rise to power in China, the one-party state instituted elections for local and low-level offices in which local factions competed for central resources. Low-level elections bolstered the KMT’s democratic bona fides and helped it penetrate Taiwanese society: it could incorporate local elites into nationwide party structures while protecting the party-state from local influence (Bosco Reference Bosco1992). To do that, the KMT latched onto existing lines of alliance, using traditional forms of association as conduits for the new central government power (Wu Reference Wu1987). Elections entailed little ideological contestation; they were a tool for normalizing dictatorship.
Democracy activists started infiltrating this process and imbuing elections with more ideological significance starting in the 1970s, helping to push the state to legalize opposition parties and institute elections at all levels of the political system (Rigger Reference Rigger1999). This transformation, like the KMT’s elections, built on the status quo, incorporating aspects of existing systems into new developments (Wu Reference Wu1987; Chun Reference Chun and Unger1996). Contemporary voting, in turn, retains much of the personalistic tenor developed under martial law. Legislators—even those in the central government—are closely tied to individual constituents (Lin Reference Lin1996; Chin Reference Chin2003; see Lerman Reference Lerman1978), for whom they are expected to furnish personalized help. People often turn directly to their national legislators for help: the legislator is like a service desk, a “window (chuangkou),” expected to help constituents solve problems either directly or by channeling them to others.Footnote 7 As Wu Nai-teh (Reference Wu1987, 265) put it, “[a]s the entrepreneurs collect money, the politicians collect friends.”
Such highly particularized interest group politics stand somewhat in tension with a widespread belief in an identifiable public good that government is supposed to serve, a conviction nurtured in Taiwan’s dictatorial developmental state (see also Johnson Reference Johnson1982). The freedom democracy offers to pursue divergent political goals makes possible a disarray that throws the definition of the public good into question. In contrast, government administration remains peopled with long-term career professionals hired through a highly competitive examination system—a steady and expert population that can claim some power to entrench a rule of law. Interlocutors in the Taipei city government, for instance, often figured legislators, with their frequent requests to bend the rules for particular constituents, as a challenge to the rule of law (Bernstein Reference Bernstein2008). In the central government, interviewees grouped bureaucrats and judges as “officials (guan)” hired through civil service exams and promoted on regular tracks, while legislators—embroiled in relationships of aid and obligation with individual constituents—were something of a wild card, less predictable and subject to institutional strictures than the professionalized civil service.
Administration and Administrative Law
Constitutions and elections are seen to be central characteristics of a republican form of government (Pettit Reference Pettit1997). But like any cultural universal, they can have different valences in different contexts (Sahlins Reference Sahlins1976). Taiwan’s democratization history—also the history of my research subjects—indicates how its particular trajectory can lead administration to play a central role in an ideology of democracy. The civil service formed a key site in which the infrastructure of the democratic transition developed. The KMT’s 1949 retreat to Taiwan brought roughly two million migrants, most linked to the party, from China to Taiwan. Through education, career pathways, and violence, the KMT instituted an ethnicized regime that distinguished local Taiwanese (bensheng ren) from these Chinese migrants (waisheng ren).Footnote 8 To alleviate some of the conflicts that ethnic division caused, the party began “localizing (bentuhua),” opening up government posts to local Taiwanese, in the 1970s (Chun Reference Chun2000). As the rest of the world withdrew recognition of the KMT party-state on Taiwan in favor of diplomatic relations with the People’s Republic of China, localization gave the KMT some claim to providing a more democratic alternative to China. The liberalizing movement that started in the 1970s—which led up to democratization—was phrased largely in the terminology of ethnonational awakening and empowerment for local Taiwanese. By changing the methods of administrative staffing in ways that dramatically altered the composition of the staff, meanwhile, localization made governance available to the governed. It successfully integrated the government apparatus, giving the disenfranchised majority a voice in policy decisions—even while that population still largely lacked a voice in meaningful elections.
In the wake of the democratic transition, Taiwan began instituting legal restrictions on administrative power. Administrative law had become a focus for reform discussions already under martial law because it was considered safer and “less connected to politics” than the Constitution, as Weng Yueh-Sheng put it when explaining why he changed course from writing about constitutional law to writing about administrative law after returning from his studies in Germany.Footnote 9 Weng, for instance, proposed revising the administrative appeals process to provide easier access to independent courts, an approach instituted in 1998, a decade after martial law ended. At a time when the Constitution served dictatorial purposes and “suggesting constitutional revision [was a] political taboo” (Yeh Reference Yeh1990, 85–87), administrative law provided a safer, and potentially more legitimate, avenue to express democratic impulses.
Early in the democratization process, the prominent legal scholar Jiunn-Rong Yeh criticized the absence of channels for private parties to express their views: “Owing to institutional deficiencies in incorporating citizen input, citizen participation in Taiwan has been accomplished mainly through protests or civil disobedience” (Reference Yeh1990, 94). In 2000, Taiwan passed its Administrative Procedure Act (APA). The APA created standard channels for private party participation in governmental decisions, imposing for the first time a slew of requirements for agencies promulgating regulations (Baum Reference Baum2011). Agencies are required to announce and accept comments on proposed regulations (Taiwan APA art. 154) and hold hearings for adjudications that impair the interests of private parties (Taiwan APA art. 102).
Each stage of the historical trajectory described here built on and incorporated what existed before it: the dictatorial use of the Constitution has not been forgotten, and elections have retained their personalistic quality, even as they have come to structure a newly democratic government. National government agencies, meanwhile, remain staffed with professionals broadly representative of the people they govern.
IV. LEGISLATION AS AGENCY ACCOUNTABILITY
It was the end of a long interview with several administrators in the National Health Insurance Administration. This agency implements Taiwan’s universal health insurance system, instituted in 1995. My interviewees spoke of their workplace in highly affective terms (see Martin Reference Martin2019). “We got married in the department,” one of them said; “we had our children here.”Footnote 10 When they boasted of the agency’s broad popularity, they sounded a bit like parents proud of their own child. What was it, I asked, that lent the agency’s actions legitimacy?
“Fundamentally,” responded Ms. Xian, the most senior administrator in the group, “these child laws (zifa) all come from mother laws (mufa).”Footnote 11 The term mufa, “mother law,” refers to a statute; zifa, “child law,” refers to the regulations born of that mother law. These family-based terms are the standard way administrators refer to laws and regulations. The terms zifa and mufa instantiate a familial idiom that crops up in many Taiwanese terms for social relationships.Footnote 12 The prevalence of family as metaphor likely draws on a traditional Confucian emphasis on hierarchical family relations as a model for social life (Greenhalgh Reference Greenhalgh1984; Fei Reference Fei, Hamilton and Zheng1992; Fingarette [1972] Reference Fingarette1998). But the words also express a guiding principle of German administrative law, a major influence on Taiwan, which insists that agency action must be based in a specific statutory delegation of authority (Singh Reference Singh2001; Taiwan APA art. 10). Thus, the first basis for the agency’s legitimacy was the way statutes governing national health insurance delegated rulemaking authority to the agency. This formulation clearly places administrative acts in a subordinate position to legislative ones.
Yet, as in many familial relationships, the situation is not quite so simple. Under Taiwan’s Constitution, executive agencies have the authority to independently propose bills to the legislature (ti’an quan, the right to raise an issue) (Constitution of the Republic of China (Taiwan), art. 58).Footnote 13 Administrative departments have components responsible for proposing legislation (fagui weiyuanhui or falü shi). Often, a substantive component will draft proposed legislation, which the legislative drafting component edits and revises before proposing the bill to the legislature.
In addition to this independent legislative drafting authority, agencies also play an advisory and supporting role to legislators. A legislator interested in putting forward legislation usually turns to an agency for advice on an issue within its sphere of expertise; most legislators then have agency personnel draft the statutory text. Thus, agencies are routinely tasked with producing differing, even conflicting, versions of proposed statutes for different legislators. Individual administrators also often play an informal advisory role through phone calls or side conversations with legislators or aides. Representatives from relevant agencies are expected to attend legislative sessions where bills are read into the record, so that they can answer questions that come up during discussion and debate.
Administrative involvement in legislation often follows these different routes simultaneously. Administrators reported that the legislature often considers a number of different drafts (banben) of bills on the same topic, which may include competing versions commissioned by different legislators along with the administration’s independent proposal, all drafted by the same agency. Administrators from diverse agencies told me that it was not uncommon for their agency to advise legislators on drafting a bill to say what the legislator wanted, while also warning the legislator of the problems inherent in his or her approach and separately drafting a different bill for the agency to propose independently. Administrators also participate in producing the statement of purpose, the lifa liyou (“law-enacting reason”), that normally accompanies each statutory provision and clarifies what agencies are expected to accomplish.
Agency involvement in legislative production was so thorough that administrators sometimes went out of their way to emphasize that legislators had the authority to propose legislation on their own. That agencies wrote the statutes that legislators voted on was unremarkable to my interlocutors. But that legislators sometimes took it upon themselves to produce statutory text independently was considered sufficiently noteworthy that a number of interviewees took pains to ensure I understood that this was a possibility.
Although every zifa comes from a mufa, administrators do not present the relationship as one of unilateral command by the legislature, nor of mechanical implementation by the agency. Rather, administrators describe the process of producing regulations for statutes as one of ongoing conversation and negotiation by different parts of the agency—substantive component, legislative drafting component, individual bureaucrats—with different parts of the legislature—the whole body, specific committees, individual legislators.
In my fieldwork, bureaucrats thus described themselves as playing a large role not just in implementing the laws of the democratically elected legislature, but also in producing them. Moreover, administrators presented their deep involvement in statutory production not just as part of their job, but as part of what justified their job. When I mentioned that some observers might see the high level of agency involvement in statutory drafting as a challenge to the proper separation of powers, interviewees generally expressed confusion and disbelief. As one high-ranking administrator put it, having an agency draft legislation “is called accountability. I [the agency] am responsible for these things. So when I think that the way we do these things needs to be modified, well, I ought to express our suggestions. So for administrative agencies to propose legislation is extremely legitimate and reasonable.”Footnote 14 Another mid-level administrator was clearly shocked by my suggestion that some people might find it inappropriate for agencies to write law. Letting her arm fall to the table and leaning in across it, she burst out, “That is really strange! But it—the agency—it bears the responsibility for its policies” implementing the statute.Footnote 15
This view echoed what Taipei City administrators described. Taipei’s Department of Legal Affairs (fawuju), for instance, is charged with evaluating whether proposed city regulations conform to applicable laws (see Asimow et al. Reference Asimow, Bocksang Hola, Cirotteau, Dotan and Perroud2019). To determine the meaning of the law, administrators in that department looked primarily to the documents and records produced in the writing of the legal provision, and even sought out the people who produced it to discuss the issues in person. Specifically, they would contact the administrators who had worked on a given law, even when those administrators had moved on to posts in the central government or gone into private employment. As in the central government, my Legal Affairs interviewee explained, it is usually the administration that proposes new laws in the city government. It is therefore the bureaucrats who best know what a law was meant to accomplish.Footnote 16
Legislatures form a primary vector in the accountability network of a democratic bureaucracy. My interlocutors routinely emphasized the importance of legislative delegation (shouquan) to the administrative branch. They commented that “our democracy has become quite mature.”Footnote 17 They praised the rapid development of what one judge called government actors’ “separation of powers consciousness (fenquan yishi).”Footnote 18 But recognizing the legislature’s part in the administrative accountability network without including a cultural analysis only reveals so much. Legislatures can play different practical and ideological roles depending on what meanings people ascribe to the interaction of executive and legislature.
Some might imagine administration as an “agent” to a legislative “principal.” In this image, accountability flows in one direction: bureaucrats who fulfill legislative mandates are accountable to legislatures that issue them. This principal-agent model “assume[s] that at the center of any given political system there is a single actor that can be characterized as sovereign at a particular point in time” (Ginsburg Reference Ginsburg2002a, 248).Footnote 19 Yet my administrative interviewees did not present themselves in this way. Although they clearly saw themselves as beholden to the statute they implemented, that statute did not arrive at their door a stranger, a unilateral command from the legislature. Rather, it took shape largely in administrators’ own hands, through an interactive process in which agency actors exercised both expertise and authority.
This vision of separated powers consciousness in a mature democracy did not involve political branches jealously guarding their work from one another’s influence. Rather, administrative agencies remained “accountable” to the legislature and the people by participating actively and openly in the production, as well as the implementation, of legislation. Taking responsibility for the regulations they promulgated included being responsible for the statute that authorized those regulations. This formulation retains some aspects of the typical developmental-state structure, where the executive leads on economic policies that the legislature more or less rubber stamps (Johnson Reference Johnson1982; Amsden Reference Amsden, Peter, Rueschemeyer and Skocpol1985; Gold Reference Gold1986; Bishai Reference Bishai1991; Ginsburg Reference Ginsburg2001). But it incorporates a new emphasis on dynamic interaction, consultation, and multiparty deliberation: administrators did not present themselves as properly determining the content of the laws for others to pass. Rather, they figured themselves as legitimated participants in a broad-based legislative process with contested proposals and unpredictable results. In the view of my Taiwanese interlocutors, extensive administrative involvement in statutory production is not only permissible, but necessary. And it is not just necessary for making laws efficacious; it is also necessary for making them legitimate.
V. PUBLIC PARTICIPATION AS POLITICAL ETHICS
Taiwan’s APA mandates a notice-and-comment process for rulemaking, but when I asked administrators how the APA affected their work, most had little to say about it. This is not to say that they were not aware of the APA, nor that they did not follow it. But their own conception of their work clearly did not rest on the requirements laid down by this law. For instance, when I asked a long-term, mid-level administrator in the Ministry of the Interior how the APA had affected her work, she said it had not: the law merely codified conceptions and practices that were already prevalent in her department.Footnote 20 This combination of awareness of the APA with a consistent waving-off of its importance characterized many of my conversations with government administrators.
Instead of describing their work as structured by the APA’s notice-and-comment requirements, administrators largely echoed Ms. Xian. After noting that every child law came from a mother law, Ms. Xian went on to explain that her agency’s actions had legitimacy because “[t]here’s a lot that we have to consult about with everyone.” Standards for private conduct, agency spending, rate setting, every decision has to “go through meetings.” Ms. Xian said the word “meetings (kaihui)” slowly, with emphatic enunciation, hands bobbing up and down to emphasize each syllable. Agency personnel meet “with everyone: with consumers, with doctors—we meet. It’s not just that we ‘put up information on a website.’ Our meetings have a representative character” that allows the agency to develop “consensus.”Footnote 21 As I have written elsewhere, reaching consensus is frequently the stated goal of meetings regarding government regulations, including those at which administrators seek to ensure that private parties adhere to the regulations that govern them (Bernstein Reference Bernstein2008).
Scholarly Participation as Bureaucratic Action
As Ms. Xian’s comment indicates, Taiwanese administrators consult not only with legislators but with a range of parties outside the government. For one thing, agencies drafting legislation and regulations rarely do so on their own. They typically consult with scholars and experts in the relevant field. Taiwan’s academic world is highly integrated with the government. Numerous Taiwanese interlocutors have told me, with some satisfaction, that Taiwan’s government has the world’s highest proportion of PhDs.Footnote 22 It is quite typical, moreover, for a successful scholar to be appointed to a government post; a number of people described a revolving door between government and academia. Holding government posts, however, is just the tip of the iceberg in academic involvement in producing laws and regulations.
Agencies routinely convene panels of scholarly advisors. By law, agency appeals panels (suyuan hui)—the intra-agency bodies that review private party appeals from administrative acts that affect private rights—must include at least 50 percent scholars or other experts (Chang Reference Chang2014, 266). Professors also often sit on commissions that determine private party eligibility for government benefits. For instance, a Taiwanese anthropologist explained that when indigenous groups apply for government funding to support the preservation of their cultural heritage, scholars constitute part of the evaluation committee. As a participant in this kind of evaluative panel, she said, she felt she played a role of epistemological mediation. As the panel looked through indigenous group applications, consulted local governments, and spoke with indigenous group representatives, scholars in the group had to explain to administrators both what the indigenous practices were and how they should value such cultural heritage. On the other hand, scholars on the panel also often found themselves explaining to applicants what exactly was subject to funding and how they should formulate their applications.
Scholarly involvement does not end at evaluating case-specific decisions, though. It extends to drafting the regulations that implement statutory mandates, and even the statutory language that agencies propose to the legislature. For instance, the anthropologist was involved not just in panels evaluating indigenous group funding applications but also in revising the law governing cultural heritage funding itself. Agencies will regularly hire scholars and experts to do this work, often through multiple rounds: one committee of scholars will produce a first draft of legal text, which the agency will pass on to a second committee of other scholars for commentary.
This kind of scholarly input, moreover, occurs not only in the arenas of cultural recognition but all over the regulatory spectrum. For instance, a mid-level administrator in the National Development Council, which has a broad economic development mandate, described how his agency created regulations implementing part of a recent public land use law (Guotu Jihuafa). The agency outsourced the production of standards for determining public land conservation areas to a committee of scholars, whose work was then reviewed by an interagency committee of government employees. Both administrators and scholars treat scholarly input into regulations and laws as highly influential; many used the word “entrust (weituo),” often used for government outsourcing, to describe the relationship.Footnote 23
Scholars’ participation is not only influential; it is also spread broadly across academia. One environmental studies professor—herself a vocal activist for government reform—complained that agency advisory panels took up a big part of her working hours. I asked if she was particularly in demand. “It’s a small pool,” she responded (in English); scholars in general seemed to be in demand. Another professor agreed, noting the challenge of participating in such committees when she viewed her own scholarship as an enterprise in critique. Yet, even coming from a critical perspective, she received so many invitations to sit on government commissions that she had to turn most of them down to leave time for her own research.
In the accountability network of Taiwanese administrators, then, scholars constitute a special node of mutual influence—not just a population to which administrators are accountable but one on which they depend to do their work. Like administrators, who participate in lawmaking by advising legislators and writing legislative text, scholars write regulations to be reviewed by other scholars and administrators, and produce legislative text proposals to be presented to the legislature by agencies as mediators and collaborators. The structures through which regulations are produced, then, support administrators’ presentation of their work as highly interactional, embedded in social networks that exceed the agency, and dependent on perspectives from outside the government.
Bureaucracy as a Social Actor
Scholars, of course, are not the only people administrators encounter. Administrators I interviewed from a range of agencies emphasized the centrality of public input and public opinion. The kinds of input they referred to took the form of direct communication between agencies and the populations under their purview. These included meetings agencies held with constituents, constituent phone calls, and personal conversations with affected parties, as well as Internet-based comment submissions, often organized as a group effort with a number of signatories. Recall that Taiwan’s Administrative Procedure Act requires agencies to solicit comments on proposed rules. When I raised these requirements myself, administrators recognized them and agreed that they bound agencies. But no administrator I interviewed themselves raised these provisions as a source of public participation. In other words, although the law requires some amount of public consultation, the administrators I studied did not refer to the law as the reason for the consultation they did. Instead, administrators presented their interactions with their publics as the ethical basis for their claim to recognize and serve the public good.
This lack of interest in the law was echoed in discussions of the possibility that an agency’s decisions might embroil it in litigation with private parties. Modeled on the German system, Taiwanese administrative litigation allows private parties to challenge decisions that directly affect their interests through an agency-internal review (suyuan) that can be appealed to an independent Administrative Court system for the judicial review of agency action (susong). As in the German system, parties may litigate only decisions that directly affect their own interests; there is no facial challenge for regulations as there is under, for instance, the American APA. Litigation risk for agency action is thus lower in Taiwan than in some (though not most) other systems. Still, private parties do utilize the administrative appeal and Administrative Court system. Indeed, the Administrative Court system has been developing continuously for twenty years, recently expanding to include a new level for challenges involving limited monetary sums.Footnote 24 And a challenge to a specific application of a regulation may end by invalidating the rule itself.Footnote 25 The administrators I talked to, though, never brought up litigation risk as a consideration in regulation drafting. When asked about it, one administrator responded that his agency dealt with a court’s invalidation of agency action when it happened; it was not something worth planning for.Footnote 26
Administrators thus discounted the significance of legal routes through which private parties may seek to change or challenge agency decisions. Instead, they referred frequently to more direct communications between affected populations and administrators, unmediated by courts or formal legal process. And they phrased this interaction less as a legal requirement than as a species of political ethics, often described in personal, affective terms. A number of administrators contrasted—sometimes ruefully—the power that bureaucrats had held in the era of dictatorship with the training they received now to “serve the people (wei renmin fuwu).” What this slogan meant, one administrator explained to me, was that when private parties called up and yelled at you, you stayed on the phone.Footnote 27
This sense that government administrators played an important role in public life was echoed in the Taipei city government. For instance, a high-level Taipei urban planning administrator, discussing a conflict about zoning in a popular commercial and residential area, explained that the government’s role is less to tell people what to do and more to mediate among their competing interests, to make obeying government policy in their interest.Footnote 28 Indeed, a number of administrators described their work as influenced by the fear of public opprobrium, of being “scolded” or “cursed” (ma) by the public. When I probed what the negative consequences of being ma would be—Could an administrator lose his job? Be transferred to a less desirable post?—my interviewees focused on the fact of public disapproval. The negative consequence of being scolded, as they described it, was the scolding itself.Footnote 29 These administrators thus expressed a sense that public confidence or trust was a condition for successful agency work.
Not only administrators but those with whom they tussled routinely adverted to bureaucracy’s central role in effectuating democratic impulses. In a typical example, Dujun, an activist student, described her work in an urban neighborhood whose residents had gathered data to demonstrate that the air quality was worse than the Environmental Protection Agency (EPA) had claimed. The EPA, Dujun explained, takes the average air quality of larger areas, missing differences among localities within an urban area. The EPA had taken steps to mitigate the problems the residents had found, but residents and activists were dissatisfied that they were the ones gathering the data: “They feel that this is something the government ought to do.”Footnote 30
VI. LEGALITY AND PUBLIC PARTICIPATION
The expectation—the requirement—that government agencies be not only responsive to, but out ahead of, the demands of civil society groups was a common theme for both the activists I have talked to over the years and the administrators who form the target for their demands. Both administrators and activists often present citizen involvement as a crucial aspect of government legitimacy. Like the activists I talked to, administrators described bureaucracy as the proper place for public participation.
Some activists have tried to mobilize the power of law to push government agencies into greater and more proactive public consultation (Tu Reference Tu2015). For instance, Taiwan’s APA does not require an agency to publicly respond to comments, and some activists complained to me that this could render the comment process a sham that required agencies to collect comments that they could then simply ignore.Footnote 31 In the limited area of hospital reimbursement through the National Health Insurance Administration, this push has recently borne some fruit: a Constitutional Court opinion recently ruled that the agency must not only hold hearings on reimbursement rates, but must also publish its rationales for adopting and rejecting participant comments.Footnote 32 This requirement has not yet been extended to other regulatory areas, but this judicial move echoes the development of notice-and-comment rulemaking in the US context, where judicial rulings, rather than the language of the Administrative Procedure Act itself, drove the compilation of rulemaking records and the publication of responses to comments.Footnote 33
When I asked National Health Insurance Administration interviewees how the ruling would affect their work, however, Ms. Xian waved her hand dismissively. Administrators developing a regulation, she replied, already “evaluate whether the comments [they receive] make good points.” When they reject a proposal in the comments, administrators already write up explanations of that decision in internal decision-making documents. The new judicially imposed requirement “actually … just means making our [existing] process public, is all.”Footnote 34 The internal culture of this administration thus presented private party input as part of its expertise and a basis for its legitimacy; whether it was required by law was presented as a side issue.
VII. LEGAL AMALGAMS
Taiwanese law, government, and political culture have incorporated a host of ideas and influences from its Chinese imperial rulers (1644–1945), Japanese colonizers (1895–1945), and KMT dictatorship (Dirlik Reference Dirlik1975; Kirby Reference Kirby1984, Reference Kirby1990; Lee Reference Lee2004; Huang Reference Huang2008). These powers in turn brought others to bear. Germany, Japan, and America have provided lasting influence, and others have played parts as well. The KMT had Soviet advisors for a time (Kirby Reference Kirby1990; Lee Reference Lee2004). The drafters of the APA took inspiration from German and American procedural laws, and consulted others from Australia, Japan, and South Korea (Baum Reference Baum2005).
The organization of courts shows a similar pluralistic dynamism. While the Administrative Court system is modeled on that of Germany, it has evolved, recently introducing a new level with the addition of District Court Administrative Tribunals (difang xingzheng susongting) to adjudicate smaller claims.Footnote 35 Justices on Taiwan’s Supreme Administrative Court told me that the judicial branch was considering a number of further changes, including subsuming the Administrative Court system into the civil-criminal system altogether, and introducing jury trials on the American model for at least some civil-criminal cases.Footnote 36
Taiwanese students, professors, and civil servants continue to study law and policy abroad, primarily in Germany, Japan, and the United States, bringing ideas from those countries back home. Acquiring an advanced degree from one of those countries is an important career-enhancing move for a government administrator. Legal scholars told me that journal editors choose referees with the same study-abroad (liuxue) background as authors, and that a retiring academic would presumptively be replaced by someone who had studied abroad in the same country. Legal thought from these three countries thus exerts a continuing influence, but not in a static way: the ongoing interaction of legal world views means that any imported approach will be transformed and localized through interaction with others.
Taiwan’s continuously fraught relations with China, meanwhile, have exerted another kind of influence. The KMT’s continued assertion of authority over China after 1949, despite lacking political control there, bent both elections and the Constitution to dictatorial ends, making administration an attractive target for those interested in democratic reform. Cross-Strait tensions also helped entrench the ethnicized stratification (mainlander versus local Taiwanese) that localization (bentuhua) addressed, contributing to the democratic bona fides of the administration. And China’s continued assertion of authority over Taiwan, despite lacking political control there, pushes administrators to demonstrate Taiwan’s viability as a nation on the world stage (Bernstein Reference Bernstein2008).
This constant, built-in interaction of different legal and political world views may contribute to a hallmark of Taiwan’s democratic transition: creativity and debate in its legal and political arenas. Administrative structures, personnel, and laws have undergone fairly radical changes over the decades around democratization and continue to develop rapidly as the administrative apparatus adjusts to developments in its changing political context.
In other words, despite—or perhaps because of—the importance of foreign influence, Taiwan is not a land of legal transplants that retain their original characteristics (Watson [1974] Reference Watson1993). Instead, the country is characterized by legal amalgams: intersecting influences that Taiwanese people continually incorporate, change, and localize. Understanding its political culture, therefore, entails attending not just to the historical roots and family resemblances of its legal or political systems (Zweigert and Kötz Reference Zweigert, Kötz and Weir1998) but to the particular historical trajectories, social circumstances, and local values in which they evolve (Legrand Reference Legrand1997).
In Taiwan, this has meant incorporating a range of outside influences while developing a distinct local culture of vocal democratic demand mixed with a reliance on—indeed a demand for—a highly involved government. This in itself does not, of course, make Taiwan unique. The widespread understanding of government bureaucracy as obligated to articulate and pursue a general public good, for instance, resonates with understandings of government in French political discourses throughout the twentieth century (Jones Reference Jones2002). As another example, American administrative agencies are also heavily involved in producing statutory text (Parrillo Reference Parrillo2013; Shobe Reference Shobe2017).
Yet these characteristics take on particular forms and meanings in Taiwan’s particular context. As Jeeyang Rhee Baum (Reference Baum2011) has argued, for example, the seemingly similar APAs produced under seemingly similar circumstances in Taiwan and Korea emerged from quite different internal political struggles and served different purposes for those who passed them. And despite its similarities, American bureaucrats’ role in legislation is not well publicized or discussed in administrative law textbooks, and even appears to be in some tension with common discussions of separation of powers in the American context (Ackerman Reference Ackerman2000). Taiwanese administrators, in contrast, present participation in legislation as integral to their accountability to both government and populace. Taking the legislative mantle, they play the role of both principal and agent in the governmental structure, suggesting that principal-agent models may be of limited utility in the complex power distributions that characterize governments.
Relatedly, Taiwan’s bureaucracy has built on, rather than simply rejected, its role in the country’s dictatorial developmental-state period. Retaining its centrality in a modified way, the bureaucracy has presented itself as a key site for public participation in creating the common good. The freedom to pursue the divergent political goals that democracy offers also presents the possibility of a disarray that throws the definition of the public good into question. During my fieldwork in the early 2000s, members of just about every social group with whom I came in contact told me something along the lines of, “Taiwan is too democratic” (Bernstein Reference Bernstein2017).Footnote 37 Even some activists I studied—who would have been sanctioned for their activism under martial law (Lewis and Cohen Reference Lewis and Jerome2013)—said so. The explanation most gave had to do with the loss of consensus (gongshi) on questions of the public good. Administration, in contrast, can be framed as identifying and pushing for a general public interest in the face of the miscellany of private interests pursued by legislators beholden to individual constituents.
I do not mean to suggest that there is something unusual about interest group politics, which are surely a feature of any polity. What I emphasize here is the way my research subjects figured legislators’ responsiveness to private interests as a challenge to public policy and even the rule of law, values that administrators presented themselves as furthering. Staffed with long-term career professionals representative of the population they govern and hired through an independent, examination-based process, bureaucracy can be both a foundation for democracy and a bulwark against its potentially dangerous excesses.
Although here I focus on the views of administrators, my experience in Taiwan—both studying activists and interacting with people from a range of other social groups—suggests that this attitude may be widely shared. This impression has support in a survey in which only 24.7 percent of Taiwanese respondents agreed that “the legislature should check the executive” (Chu et al. Reference Chu, Diamond, Nathan and Shin2010, 32). From one perspective, this may look like an authoritarian hangover: the population of a country recently emerging from a dictatorship in which the executive monopolized public discourse retains the ideological predilections of that dictatorial propaganda. It may indeed owe something to that history. But it builds also on a more recent history in which administrators have come to occupy a newly participatory administrative environment, and to treat that openness as central to their own legitimation, while legislatures retain their habits of personalistic deal-making. At least in some eyes, then, bureaucracy can be a fulcrum for popular participation and democratic legitimacy.
VIII. CONCLUSION: ADMINISTRATIVE CULTURE AS DEMOCRATIC INFRASTRUCTURE
I have argued here that studying the cultural aspects of bureaucracy—the bases on which people ascribe meaning to structures and practices—is key to understanding the legitimation of the administrative state. We should take the world views of bureaucrats as seriously as we take those of the people they regulate. Viewing administration as a cultural arena can provide insight into how governments justify their actions. And it opens up avenues to consider the variety of bases that even successful democracies can turn to for legitimacy.
Given Taiwan’s remarkable success at democratic transition—all the more unlikely given its lack of recognized de jure sovereignty—it is worth considering the role of administrative culture in democratic viability more broadly. My findings here suggest that administrative culture provides a foundation for more visible indicators of democracy, like elections and constitutions, to take effect, even as it provides a bulwark against the potentially destabilizing effects of democratic transition. While scholarship and public discourse often counterpose administration with democracy, my interlocutors suggest a different way to understand the relationship. Bureaucracy, in their presentation, does not compete with democracy but provides its condition of possibility.
Administrators’ descriptions of government present a particular ideal of democratic process. In this image, rather than being primarily bound by legal strictures, state actors work in ongoing, direct dialogue with the populace under their purview and the legislators who guide their discretion. Interested private parties speak directly to bureaucrats through meetings and comments. Scholars, another kind of private party, participate in writing laws, producing regulations, and making adjudicatory decisions.
Government employees, in turn, are the agents not just of the premier who is their boss, nor of the legislature that enacts their statutes, but also of the public itself (see Kahn and Brennan-Marquez Reference Kahn and Brennan-Marquez2014). In this vision, the bureaucracy represents the general will and furthers the public good at least as much as—perhaps more than—the democratically elected legislature. Ideologically, the administrative state plays a central role not just in implementing policy but in creating and legitimating it. It connects the people to their government. It even connects the legislature to its own legislation.
I do not mean to suggest that this ethical commitment renders Taiwan’s government inherently legitimate, nor that it effectively ensures that all interests are considered in administrative decisions. Taiwanese citizens are not uniformly enamored of their government, and struggles about the distribution of power and the role of law constitute as much of the political life of this polity as of any other. My point, rather, is to elucidate how those citizens who populate the government present the grounds for its legitimacy.
Qualitative research on administrators does not generally see bureaucracy as a site for popular participation. Yet that is how the bureaucrats I study present their workplace. They imagine government not as a superordinate decider but as an embedded social actor, implicated in normal social processes and subject to their strictures, conventions, and values. Key to this role is both the administration’s ongoing cooperation with the legislature and its relationships of creatively evolving accountability and obligation with the public. On this view, democracy ties the populace not just to its elected representatives but to the administrative state itself.
This bureaucratically oriented version of direct democracy—perhaps I should say direct bureaucracy—helps constitute the legitimacy of administrative action for those who practice it in my field site. This local ideology illuminates how bureaucracy, rather than sullying or competing with democracy, can be seen as co-constitutive with it. For my subjects, a key aspect of this legitimately democratic bureaucracy is its ongoing entanglement with things outside itself. Bureaucrats describe a highly porous administration shot through with multiple influences from other institutions, groups, and individuals. And they describe that as how it should be.
