One
Introduction: The "Fragmented Authoritarianism" Model and Its Limitations

Kenneth G. Lieberthal

The reforms that began at the end of the 1970s created opportunities for unprecedented scholarly access to government agencies in the People's Republic of China (PRC), thus vastly increasing our potential for understanding bureaucratic relationships and policy processes there. The resulting research quickly began to yield a small harvest of publications,^[1] which moved our understanding beyond the limitations of the documentary sources and refugee interviewing on which scholars had relied for the most part until then.^[2] It became increasingly possible to probe perceptions, to trace in detail specific policy decisions, to become familiar with the recommendations made by various staff organs, to match formal bureaucratic relationships against actual behavior of officials, and so forth. This volume demonstrates some of the returns from this unprecedented access.^[3]

A mapping of China's major bureaucratic sectors shows, however, that scholarly penetration of China's far-flung bureaucratic leviathan proved highly uneven. The decision by leaders in Beijing to intensify the country's economic relations with the outside world afforded particularly extensive contact with economic organs. Scholars could interview pertinent Chinese officials, staff in the country's enterprises, and numerous foreigners who conducted business with the Chinese. Beijing's dealings with the World Bank provided additional information to those concerned with economic decision making.^[4] From this access, a burgeoning literature began to flesh out a model for understanding economic decision making in the Chinese polity that is reviewed below.^[5] This model might be called "fragmented authoritarianism." Its key features are discussed below.

China, of course, consists of more than a set of economic bureaucracies. In broad terms, the country's national political organizations are perhaps best grouped into six functionally defined clusters.^[6] Each cluster has a number of different bureaucratic units in it, and each is also assigned a distinctive set of tasks. These clusters and their core tasks are:

Economic Bureaucracies. These seek to make the economy grow in order to satisfy the material needs of the country. A wide array of specific bureaucratic hierarchies fall within this cluster, and there are many horizontal and vertical cleavages in it. But priority commitment to economic growth generally enhances the resources and prestige of the organs in this cluster, albeit specific strategies of growth inevitably favor some units (e.g., the banking system) over others (e.g., the planning system).

Propaganda and Education Bureaucracies. These have responsibility for shaping the values and knowledge of China's citizens. They encompass the formal propaganda organs (including the mass media), the educational system, and most research units.

Organization and Personnel Bureaucracies. These run the system of personal dossiers and provide critical staff work for individual career assignments. While these bureaucracies do not themselves actually make personnel decisions, they can strongly influence those decisions through their power to collect and utilize data on individuals.

Civilian Coercive Bureaucracies. These provide the civilian fist that is used as necessary to protect the communist system and to implement policies. They include the public security system, the judicial system, the prison and forced labor administration, and intelligence/counterintelligence units. As with the economic bureaucracies, there have been sharp internal tensions within and among these bureaucracies, but, in general, periods of high coercion provide bureaucratic benefits to many of the units in this group.

Military System. The People's Liberation Army (PLA) is virtually a separate state within the Chinese system.^[7] It is directly subordinate to the Party leadership, and its leading body, the Military Affairs Committee of the Party, is coequal in rank with the government State Council. In addition to providing security for the country against external threats, the military has in the past been assigned extensive domestic political roles. This bureaucratic cluster also includes the military-related science and technology system and the extensive production, transport, and service facilities run by the General Logistics Department (houqin bu ) of the PLA.

Communist Party Territorial Committees. There is, in addition to the above, a core political function that is exercised on a territorial basis by Communist Party secretaries. Each of these individuals attempts to coordinate and prioritize activities within his or her geographical domain and to represent the interests of that domain in dealings with both higher and lower levels.

China has other important bureaucratic systems, such as foreign affairs, united-front work, and mass-organization work, that are not included in the above listing. But the above six clusters of bureaucracies identify the core organs that have nationwide hierarchies and that exercise strong executive power. This "cluster" approach downplays the distinction between Communist Party and government organs—a distinction that, as Susan Shirk explains in chapter 3 in this volume, can be politically very important. But this "cluster" approach has the virtue of grouping bureaucratic systems in much the same way that the Chinese leaders organize themselves to run their system.^[8] It has the additional

virtue of enabling us to identify reasonably clearly the contours of the access foreign scholars have gained to China's bureaucratic labyrinth.

Other than the economic bureaucracies—for which foreigners have enjoyed extensive contacts at all levels of the national hierarchy and in many regions of the country—foreign access has been limited and selective. Of the remaining five areas, perhaps the greatest access has been to the propaganda and education cluster. The education portion of this cluster became quite open to foreign contacts during the 1980s.^[9] Within the propaganda portion, access to Chinese writers and to people in the mass media proved particularly extensive.^[10] The core propaganda apparatus in the Communist Party itself (that is, the propaganda departments under various Party committees) remained more opaque.

Very few scholars gained extensive access to the personnel, civilian or military security, and Party territorial clusters for intensive interviewing. Only two authors have published work based on direct interviewing of personnel officials,^[11] and no one has published a book-length study based on direct access to either the security apparatuses or the territorial Party leaders. Even in the most open periods of reform, moreover, Chinese officials have considered direct questions about either the security clusters or the core Party bureaucratic operations to be inappropriate and overly intrusive.^[12]

A second bias in the interviewing done to date involves the bureaucratic levels that have been accessible to foreign scholars. Some scholars have been able to undertake systematic interviewing up to the vice-ministerial level, with occasional meetings with individuals above that rank. But no scholars have enjoyed systematic access to individuals with the active rank of minister or higher.^[13] The functional "cluster" unevenness in

interviewing is thus further complicated by incompleteness in the hierarchical levels of China's political system to which foreigners have gained access.

The result is that the "fragmented authoritarianism" model contained in the literature is based overwhelmingly on intensive research on only one of the six "clusters" and on access only to certain parts of the multitiered Chinese bureaucratic hierarchy. The editors of this volume have contributed to the slim literature that has explicated a "fragmented authoritarianism" model for China's policy process, and we convened the 1988 conference, whose papers make up this book, as a vehicle for both elaborating upon and testing the limits of the validity of our findings as they apply to other clusters and bureaucratic levels. In organizing this conference we sought to include scholars who had gained direct access to each of the "clusters" noted above, as well as individuals who had focused on every major level of the national bureaucratic hierarchy.

As the chapters in this volume indicate, we succeeded in obtaining contributions from scholars who had interviewed in four of our six "clusters." (No scholar, to our knowledge, has enjoyed sufficient access to the civilian security cluster or to territorial Party secretaries to provide major insights into these bureaucratic clusters.) In addition, while we have included chapters that focus on each national bureaucratic level of power (Center, province, major municipality, county, and township) and on urban and rural localities, we have left a number of empty "boxes" in a hypothetical matrix of clusters and bureaucratic levels. This volume does not, to put it differently, treat every cluster on each level of the national bureaucratic hierarchy.

Nevertheless, this volume does, for the first time, focus the attention of scholars who have gained direct, sustained access to Chinese officialdom in various clusters on the task of explaining the characteristics of the policy-making process in these different clusters. Special chapters by Carol Lee Hamrin and Nina Halpern focus on the Center above the level of ministries and commissions to understand better the interplay of personal and organizational elements in leading China's vast bureaucratic apparatus at the very top. The result of this array of rich presentations is a more textured, contoured understanding of the structures, processes, and types of relationships characteristic of various Chinese clusters and bureaucratic levels than has previously been available. The result enables us to identify both similarities and differences among the various parts of the Chinese system.

To appreciate the distinctive, broad-gauged contributions these chapters make to our understanding of the Chinese bureaucratic system, a brief review of the "fragmented authoritarianism" model provides a helpful starting point. This model has been embodied in several works

that the editors authored or coauthored,^[14] as well as in other works.^[15] The chapters in this volume add new richness to our understanding of "fragmented authoritarianism," provide an understanding of the limitations to which this model is subject, and explain the reasons for those limitations. They also provide a basis for anticipating the systemic consequences of various changes in China's political priorities and the degree of malleability in the Chinese bureaucratic system. But first, the essentials of the "fragmented authoritarianism" model can be described as follows.

The Starting Point: The "Fragmented Authoritarianism" Model

Just as the reform era itself created opportunities for intensive research on Chinese bureaucratic practice, the content of the reforms influenced the specific research agenda in important ways. Decentralization of decision-making authority was a key reform initiative, and much research sought to understand better the dynamics of a system in which Beijing seeks to work with lower levels rather than to dictate to them.

There are, in broad terms, three dimensions to the study of decentralization and centralization: value integration; structural distribution of resources and authority; and processes of decision making and policy implementation.^[16] These three dimensions are not mutually exclusive, but they do highlight different aspects of the system under scrutiny. The "fragmented authoritarianism," explained below, focuses especially on the latter two dimensions.

Values can provide a strong basis for integration of policy making and implementation. In China under Mao, for example, the top leaders used massive doses of ideological indoctrination as a vehicle for achieving greater fidelity to the goals of the leaders even in the absence of informational and other resources adequate to assure the desired level of compliance. In contemporary Japan there is often sharp conflict in policy before a consensus has been reached, but the various parties to the decision

can then act with confidence in the knowledge that all concerned will obey the "consensus" decision once it has been made.^[17] Shared values, in short, can substantially affect the operations of a political system. Value consensus can basically reduce the need of the political leadership to develop additional resources to assure fidelity to their priorities and compliance with their policies.

Political structure, by contrast, examines the formal allocation of decision-making authority. Various political systems have, of course, adopted a very wide array of allocations of power, and significant structural variations may appear even in one political system over time. In China, for example, the government ministries exercised considerable power over the economy during the First Five-Year Plan of the mid 1950s, the Communist Party expanded its relative influence over the economy during the Great Leap Forward of 1958–61, and the military became a key decision-maker regarding the economy during the late 1960s. The reforms of the 1980s brought a significant devolution of budget-making authority (a process that had actually begun in earnest in 1971), with Party and government officials in major municipalities gaining greatly increased influence over economic decision making. These variations amounted to structural changes in the distribution of basic decision-making power in China's economic sphere.

In general, the structure of centralization and decentralization is highly complex. These terms encompass, for example, considerations about the span of control of various organs, the levels at which particular decisions are nested, the bases of resources of various players in the system, the degree of specificity that characterizes decisions, the amount of flexibility formally allowed in policy implementation, and so forth.^[18] The clarity with which organizational boundaries and decision-making procedures are specified affects structural centralization.

Structural dimensions of centralization and decentralization by themselves, of course, provide at best an inexact guide to the real operations of a political system. The focus of much scholarly analysis is the actual policy process that characterizes decision making and implementation. Forces that influence that process and help to shape it include, inter alia, the values held by participants in this process, the structural distribution of authority and resources, and the structure of rewards.

The "fragmented authoritarianism" literature touches on all three

dimensions of analysis: value integration, structural elements, and policy process. To a significant degree, however, the literature has emphasized the latter two. This fact probably reflects three problems. First, the study of values is inherently very difficult, and this is especially true in a society like China that does not permit the types of survey research and participant observation that are best suited to reaching well-grounded conclusions about value integration. Second, the huge changes in value hierarchies that the reforms demanded as compared with the late Maoist era produced a generally unstated consensus among Western scholars of China that the level of value integration among PRC officials during the late 1980s was low. And third, as was noted above, foreign scholars have gained the greatest access to the economic bureaucracies;^[19] the relative ease of observing bargaining over control of tangible resources in this sphere quite naturally focused the attention of researchers on the material (instead of the value) dimension of the Chinese policy process. Because of these factors, the fragmented authoritarianism model has devoted most of its attention to the structural allocation of authority and the behavior of officials related to policy process, especially those behaviors that have characterized the reforms.

The fragmented authoritarianism model argues that authority below the very peak of the Chinese political system is fragmented and disjointed. The fragmentation is structurally based and has been enhanced by reform policies regarding procedures. The fragmentation, moreover, grew increasingly pronounced under the reforms beginning in the late 1970s, as the following brief examples illustrate.^[20]

Structurally, China's bureaucratic ranking system combines with the functional division of authority among various bureaucracies to produce a situation in which it is often necessary to achieve agreement among an array of bodies, where no single body has authority over the others. In addition, the reforms' decentralization of budgetary authority enabled many locales and bureaucratic units to acquire funds outside of those allocated through the central budget, which they could use to pursue their own policy preferences. This cushion of "extrabudgetary" funds in turn permitted many locales to become less sensitive to the policy demands from higher levels.

Procedurally, several major changes designed to produce more effective information and incentive systems had the additional effect, according to the fragmented authoritarianism model, of contributing to the fragmentation of authority of China's economic decision making. The leaders reduced the use of coercion—purges, labeling, demotions—against those who propose ideas that eventually are rejected, thus emboldening participants to argue forcefully for their proposals. The stress laid on serious feasibility studies during the 1980s also, de facto, encouraged various units to marshal information to support their own project preferences, often in competition with others. The encouragement given to many organs to become increasingly self-supporting through bureaucratic entrepreneurship also strengthened the tendency of bureaucratic units to work vigorously to promote and protect their own interests in the policy-making process. The general decline in the use of ideology as an instrument of control increased the "looseness" of the system, and decentralization in personnel management permitted many bureaucratic units additional initiative. All of these changes thus combined to reduce the extent to which organs respond in disciplined fashion to instructions from higher levels.

The resulting situation proved to be quite complex, and no simple characterization captures even its essential features. To an extent, the above developments seem to have produced increased bargaining in the Chinese bureaucratic system. "Bargaining" involves negotiations over resources among units that effectively have mutual veto power. Fragmentation of authority encouraged a search for consensus among various organs in order to initiate and develop major projects. This consensus, in turn, required extensive and often elaborate deals to be struck through various types of bargaining stratagems. Perhaps the fact that the issues under study generally concerned the allocation of real resources made it more likely that visible bargaining would characterize much of the decision making in the economic cluster. In any case, the fragmented authoritarianism model focused attention on the importance of bureaucratic bargaining, and one major concern of this volume is to determine the conditions under which such bargaining does and does not occur.^[21]

This system still retained some important elements of coherence. Research found that bureaucracies retained a real sense of mission and purpose. It also revealed striking instances of "policy communities" that formed around particular projects and issues and that cut across formal bureaucratic lines. This research, moreover, did not conclude that the leaders at the top of the system are helpless. Rather, it portrayed a system

in which the information available to the leaders, the strategies of the leaders themselves, and the actual policy implementation are affected in significant fashion by the structural and procedural aspects of the bureaucracies, at least in the economic cluster.

The fragmented authoritarianism model thus did not present the Center as helpless, the bureaucracies as unable to cooperate, or the locales as all-powerful. But it did seek to identify the causes of fragmentation of authority among various bureaucratic units, the types of resources and strategies that provide leverage in the bargaining that evidently characterizes much decision making, and the incentives of key individuals in various units,^[22] in order to gain a better grasp on the ways in which bureaucratic structure and process affect Chinese policy formulation, decision making, and policy implementation. On balance, as was noted above, this model lays considerable stress on bargaining relationships.

The fragmented authoritarianism model echoes some of the literature on bureaucratic politics, such as Graham Allison's Essence of Decision .^[23] Allison argues that organizational processes (what he terms "standard operating procedures") and bureaucratically shaped politics influence decisions, and thus that rational-actor assumptions about decision making (in his case, regarding national security) do not fully capture the forces that shape policy-making and implementation. Similarly, the fragmented authoritarianism model regarding China under the reforms does not argue that rational problem solving at the top does not occur. It rather details other dimensions of the system that are not adequately captured in a straightforward application of a rationality model.

The fragmented authoritarianism model thus seeks to put into better perspective two well-developed groups of literature concerning policy-making in post-1949 China.^[24] The first group posits essentially a rationality model for understanding Chinese political outcomes. This literature implicitly assumes both that top-level leaders can exercise enormous leverage over the political system as a whole and that these same leaders make decisions by identifying problems and then adopting responses to those issues that derive logically from their own value preferences^[25] —or power needs in intraelite struggles.^[26] Books in this literature, whether

they focus on elite politics per se,^[27] or on the politics of various functional areas, such as public health,^[28] agriculture,^[29] or education,^[30] explain policy decisions almost exclusively in terms of leadership perceptions and preferences.

The second group stresses the extent to which China is what has been termed a "cellular" society. This literature focuses on the consequences of various policies over the years that have decentralized decisions and produced bureaucratic devices that permit localities effectively to block upward flows of information and to blunt higher-level initiatives that cascade down on local leaders. While this literature initially focused primarily on economic linkages,^[31] in more recent versions it has been extended to the political system, too.^[32] The experiences of businessmen who dealt with China during the 1980s added much anecdotal evidence to support the cellular model of the Chinese system, as localities repeatedly demonstrated an ability to frustrate the policies of the upper-level authorities.^[33]

To repeat, the fragmented authoritarianism model acknowledges the great insights offered from elite-oriented rational-actor approaches and from a cellular conception of the system. However, it adds a third necessary ingredient to the equations: the structure of bureaucratic authority and the realities of bureaucratic practice that affect both the elite and the basic building blocks of the system.

The structures that link the top and the bottom of the system—in order to function—require negotiations, bargaining, exchange, and consensus building. Fragmented authoritarianism finds wanting both strictly "top down" and "bottom up" views of Chinese politics. This model, rather, focuses attention on the effects of the interactive processes among the constitutent elements of the Chinese polity. In so doing, it finds that the system is somewhat but not totally fragmented. The fragmentation has not reached the point where its constituent parts have the legitimate autonomy characteristic of a pluralist system.

Fragmented Authoritarianism In Broader Perspective

The authors of the major works that have contributed to the fragmented authoritarianism model recognize that this model reflects research primarily on economic decision making. An important contribution of this volume is its exploration of the extent to which this model adequately conveys the dynamics of decision making beyond the economic cluster of bureaucracies (and within that cluster, beyond decision making on major new investment projects). Every chapter makes a substantial contribution to this effort. The following discussion culls out some of the important factors that emerge from this study and that modify to a greater or lesser extent the fragmented authoritarianism model as an adequate description of the entire Chinese bureaucratic system.

Like most of the literature on bureaucratic bargaining in China, Barry Naughton's contribution in chapter 9 focuses on decision making concerning major investment projects. Naughton notes the irony that Chinese leaders intended the reforms to reduce the amount of bargaining in the management of the economy by substituting market forces for bureaucratic management of the decision-making process. Instead, the reforms increased the role of bargaining in this sector, and they changed the types of resources that the various parties brought into the bargaining arena.

The two key resources that structure bargaining positions regarding proposed major investment projects, according to Naughton, are control over information and skills and control over resources. The bureaucratic bargaining literature to date has emphasized the latter almost to the exclusion of the former. As is often noted elsewhere, the reforms have, on balance, shifted control over resources—especially over budgetary resources—to lower bureaucratic levels in the system. But this same reform effort has effectively created greater central-level control over economic information and skills.

Enhanced availability of information for the Center commenced with the rehabilitation and expansion of the State Statistical Bureau, which

had been devastated during the Cultural Revolution. It then proceeded further via two channels: considerable relaxation of the rules governing release of economic data;^[34] and the establishment of various new policy-research organs at the central level to enhance directly and indirectly the ability of the leaders to produce more highly coordinated economic policy based on greater availability of information. In chapter 5 Nina Halpern analyzes the important influence on policy process that the new policy-research institutes exerted. As a result of these measures, the Center could undertake more sophisticated economic projections and planning.^[35]

This shift in several types of resource did not reduce bargaining in the system. The Center no longer retained the direct economic levers necessary to obtain its wishes by simple command. Nevertheless, it did command information, expertise, and some critical resources that the localities would need in order to pursue economic development effectively, especially in terms of infrastructure development. As a consequence, the Center used its superior information and skills to bargain with the localities to obtain cooperation in the pursuit of the Center's priorities.

As the Center's priorities became better informed, therefore, its bargaining position vis-à-vis the localities was correspondingly enhanced. In contrast, the greater transparency of the Chinese system under the reforms created the impression that the prereform planned economy was characterized by an extremely strong Center. In reality, lack of information and skills at the Center kept this level of the system considerably weaker than outward appearances suggested.

As Naughton's focus on information and skills highlights, the reforms produced very complex changes in the distribution of pertinent resources in the Chinese political system. Focusing on decision making in the economic cluster, Naughton finds, not surprisingly, that the net result is an enriched and highly complex set of bargaining relationships. Other chapters in this volume provide information both on the sectors of the system in which bargaining plays less of a role and on the conditions under which bargaining flourishes in the Chinese polity.

Carol Lee Hamrin, in chapter 4, focuses on a key layer of bodies—comprising "leading groups"—that links the top few leaders with the country's far-flung bureaucratic network. This arrangement groups most of the major bureaucracies into broad functional clusters and assigns a small group, typically led by a member of the Politburo or its

Standing Committee, to manage that policy cluster and to act as liaison between it and the top decision-makers. These "leading groups" do not appear on organizational charts, and their membership is not announced. But they serve vital functions in the Chinese system: they work up materials for the consideration of the top leaders, initiate pertinent policy research, resolve some issues that cannot be handled at a lower level, and coordinate activities among the various bureaucracies in each bureaucratic cluster.

Hamrin provides far more detail on the origins, development, functions, and politics of these leading groups than has ever been presented before. Noting that these groups grew out of a clustering of the bureaucracies that originally took place in order to facilitate personnel assignments, Hamrin stresses that these bodies have become an extraordinarily concentrated expression of centralized power in China. At this level, the system is highly personal. The heads of these various "leading groups" typically have known each other for many decades, and their relationships cannot fruitfully be analyzed in terms of a fragmented authoritarianism model. There is, rather, a great deal of informal contact and maneuvering, with control over the various "leading groups" an important prize in the political jockeying in the capital.

In chapter 5 Nina Halpern focuses the analysis just one step below the "leading groups." She deals with the research institutes that were established—primarily at Zhao Ziyang's initiative—to develop policy options for the top leaders. While he was premier of the State Council, Zhao devoted considerable attention to creating an extensive "in-house" policy-research capability. The research institutes he established lacked line functions, but as staff organs they nevertheless influenced in important ways the information flows within the government and enhanced the ability of the leaders to develop and implement coordinated policies.

The institutes greatly increased the leadership's information on policy externalities, which effectively enhanced the leaders' ability to formulate coordinated policies. These institutes also gave the top leaders greater leverage vis-à-vis their own line organs by reducing their dependence on the line organs for vital information. And the research institutes forced many of the line bureaucracies to consider more fully the externalities of their policy proposals during the drafting process, as the bureaucrats knew that their proposals would have to compete with those of the pertinent research institutes. The Maoist system did not encourage this broader vision.

Nina Halpern's analysis emphasizes that at this level of the system bargaining (which, again, focuses on exchange and mutual veto power) does not capture the essence of the relationship between the research institutes and the top decision-makers. The institutes, after all, did not

seek to trade off resources as much as they sought to influence policy thinking. A command relationship, in which the institutes faithfully provided the top leaders with the information they sought, would have drastically reduced the real benefits of the institute structure for the leaders. Halpern characterizes the resulting relationship as one of "competitive persuasion."

This relationship mixed both formal bureaucratic and highly personal elements. In essence, the institutes competed with line bureaucratic units (and with each other) to persuade Zhao Ziyang and other top officials of the wisdom of the institutes' policy proposals. Their effectiveness depended in part on the personal ties between the institute leaders and the top political leaders, as is shown by the changes (generally, for the worse) in the stature and effectiveness of various institutes since Zhao Ziyang's ouster in mid-1989. But the fact that the institutes demonstrably had the ear of powerful officials also made line bureaucratic units take the institutes seriously and try to adjust their own policy advocacy to a system in which these institutes provided the top leaders with serious information and analytical work. The net result contributed to the enhanced availability of information to the top leaders that is noted in chapter 9.

Below the level of the top leaders and the research institutes, the system is dominated by bureaucratic bodies, most with line functions. Key issues about these bodies concern their responsiveness to political leaders, the extent to which their actions are governed by formal rules and regulations, and the ways in which they deal with each other. This is an extremely complex system, but the various chapters in this volume permit some important generalizations to be made.

As Melanie Manion illustrates in chapter 8, on the cadre retirement policy in the personnel system, the ability of the top leaders to reach clear and detailed decisions seriously affects the functioning of the system as a whole. Bureaucratic middlemen tend to regard policies that are flexible and ambivalent as the equivalent of having no policies at all. They are more likely to implement policies to the extent that the policymakers reduce the middlemen's risks, constrain their actions, and coordinate the various demands made on them. But these suggest that the crucial elements are matters of policy content and coordination, not of bureaucratic structure and rules. Where policies lack the necessary specificity, clarity, and "fit" with other related activities, the middlemen studied by Manion would simply fail to implement them. This was thus not a bargaining situation so much as it was a matter of policy compliance. And policy compliance depended crucially on the ability of top leaders to produce well-coordinated sets of clear, detailed policies.

Indeed, the importance of policy content versus bureaucratic institutionalization of the Chinese system is evident time and again in these

chapters. For example, even in the mid and late 1980s, nobody seems to have seriously questioned the right of the top leaders to reach any decisions (and enforce any redistribution of power) that they saw fit. This ability of higher-level leaders throughout the system to impose their will on lower levels without serious institutional (versus self-imposed policy) constraints shows through repeatedly, as in Paul Schroeder's treatment of the relations between Wuhan and Hubei and in David Zweig's analysis of the approach taken by county governments to subordinate administrative towns and townships. Bureaucratic behavior inevitably limits the information available to the top leaders and may produce various distortions in policy implementation. But no institutional regulations in themselves seriously constrain the options available to the top leaders as a group.

This is a system, then, where the balance between policy and institutional integrity tips very heavily in favor of the former, even under the reforms. In chapter 3 Susan Shirk stresses, in this regard, that the reform agenda required that the top political leaders adopt reform as the official political line. Only within this context could government officials at lower levels implement the reforms. The government alone, Shirk argues, could not have generated the reform policies against the wishes of the Party apparatus. The fact that the top leaders adopted "reform" as the political line, however, enabled lower-level government officials to implement many reform policies even in the face of obstructionism by the Party bureaucrats who wished to slow down this process. Shirk does not argue that the adoption of a political strategy by the top leaders is by itself adequate to produce results in China. She asserts, rather, that the adopted political strategy creates the framework within which subsequent bargaining generally occurs. The bargaining, in turn, reflected several facets of the reform effort: its focus on economic issues; disagreements among the leadership about specific reform policies; and lack of consistency and specificity that characterized reform decisions.

In sum, the picture above suggests that the top leadership in China remains very powerful, despite the reforms. At the top, bureaucratic boundaries fade even as leaders compete for control over bureaucratic resources. Personalities and personal relationships assume tremendous importance at the pinnacle, and then through "leading groups" and other devices the small coterie of very top officials ties into the huge bureaucratic clusters through which they govern China. While the reforms have decentralized administrative control over many resources, they have also in various ways (especially through the enhancement of the availability of information) increased the potential leverage of top leaders vis-à-vis their own bureaucracies. While no changes have occurred that effectively limit the right of the highest leaders to change the

rules of the system itself, the leaders' ability to elicit lower-level compliance is reduced when there is evidence of leadership tentativeness or disunity.

Several chapters in this volume deal specifically with bureaucratic clusters that do not concern the economy. In these other clusters, there appears to be far less evidence of bargaining relationships than has generally been found in decision making and policy implementation in the economic cluster. This may, of course, in part reflect the possibility that it is simply more difficult to identify bargaining behavior in bureaucratic sectors that deal more with intangibles such as propaganda and education, but the pertinent chapters suggest that the differences are real and are important.

Lynn Paine's examination of the educational system casts light on a bureaucratic cluster that produces, in her words, a "mushy" product—that is, it is very difficult in the short term to determine the payoff of specific investments of resources in education. Overall, moreover, the educational system is bureaucratically weak at every level. It consumes, rather than generates, resources, and it has repeatedly been associated with producing values and behaviors that the top leaders have regarded as anathema. Paine finds that this mushy, resource-poor bureaucratic system is not characterized by extensive bargaining among its various units. Paine reveals, not bargaining to protect and enhance resources, but a process of "groping," where the Center sets vague standards, the locales implement the policies quite flexibly, and then there are iterations of interactive adjustment as units with very limited organizational resources seek to find ways to accomplish their tasks. In the process, these units adopt what Paine describes as "loosely jointed coping mechanisms" to get things done. The Chinese Communist Party remains extremely powerful in defining the boundaries of the acceptable in the educational sphere, but the policy dynamics within this sphere bear little relationship to those regarding major economic decision making.

Jonathan Pollack's examination of the military system intrudes into an area where very little was known previously about policy process. Pollack portrays the People's Liberation Army as a virtual internal empire within the PRC, and he sees it as having early on reached a compact with the Party in which the military would remain loyal so long as it could protect its vast and varied domain. Within this system, Pollack argues, personal ties rather than institutional positions determine the power of people at and near the top of the system. The closer to the apex of the military system, in short, the less command derives from specific rules and norms and the more it is personalistic.

Pollack examines Deng Xiaoping's far-reaching reforms in the military in terms of their stress on budgetary restraint, effectiveness in producing

new technology, and progress in creating more resilient institutions. While Pollack suggests that Deng made very substantial progress along all of these lines, he argues that during the 1989 crisis the military system reverted very much to its more personalistic, prereform character.

Although Pollack is not able to address directly the issue of bargaining relationships in the military, the thrust of his analysis suggests that the overall effect of the reforms had been to strengthen the role of central organs within the military. Greater need for technological improvements to prepare for modern warfare effectively enhanced central coordination in all phases of the weapons research, development, testing, and evaluation system. He does not present evidence that would suggest that the reforms in the military, like those in the civilian economic sphere, led to policy processes in which bargaining became increasingly important to policy outcomes. Possibly the pertinent policies—stress on high technology, on professionalism, on institutionalization, and on budgetary restraint—created an environment less conducive to wide-ranging bargaining. Because there remain sharp limitations on our information about this sphere—despite the fact that Jonathan Pollack breaks considerable new ground in chapter 6—the answers to many questions about policy process in the military must at this point remain tentative. This is especially true since, in the wake of June 4, 1989, many of Deng's efforts to professionalize the military have yielded to renewed emphasis on politicization of the armed forces.

The chapters that focus, therefore, on personnel (Manion), education (Paine), and coercion (Pollack) find that the reforms of the past decade do not appear to have nurtured the pervasive bargaining relationships that researchers have found in the economic sphere. To be sure, some bargaining characterizes the dynamics of every functional system, but the differences of degree suggested here are substantial and important. Bargaining behavior evidently requires at least that tangible resources be at stake and that substantive policies permit leeway in implementation. These conditions are not met in significant measure for much of China's bureaucratic behavior.

Three chapters focus specifically on distinctive subnational bureaucratic levels in China: Paul Schroeder examines the provincial level (relations between Hubei and Wuhan when the latter was granted provincial budgetary status); Andrew Walder analyzes fiscal and budgetary dynamics at the municipal level; and David Zweig illuminates the political dynamics of rural urbanization at the subcounty level. All three focus substantial attention on economic decision making at these three levels.

Schroeder, in chapter 10, stresses the failure of the Chinese system, even under the reforms, to develop political or legal formulas that would permit the development of stable definitions of authority. The result,

which he details in his examination of the changing formal status of Wuhan municipality vis-à-vis Hubei province, is fluidity, competition, incessant bargaining, and a resulting muddle. Where jurisdictional issues arise (as they do constantly, given the many changes initiated by the reformers), the legal system is not mature enough to settle the issues and to provide case-law precedents to guide the behavior of others. Vague policies from the Center essentially produce bargaining among lowerlevel units, where the parameters of the situation remain uncertain. This chapter illustrates well the interconnections among the formal administrative system, the legal system, the economic system, and substantive policy decisions in shaping the results of reform initiatives. Schroeder finds that the actual relationships among these various elements are fluid and that the system as a whole, therefore, lacks the institutional regularity necessary to permit reform decisions to have their desired effect.

Andrew Walder, in his analysis in chapter 11 of the processes that determine the financial flows between city budgets and enterprises, argues that bargaining activity does not necessarily stem from fragmentation of authority, as the fragmented authoritarianism model has stressed. Walder notes, indeed, that the key characteristic of the municipal fiscal environment is that of concentration of power rather than of its fragmentation. But this concentration of budgetary power and resources occurs in a larger environment of unreformed prices, unequal capital endowments, and unclear rules and regulations. He finds plenty of bargaining behavior in the process of determining financial flows between the city budgets and enterprises he examines, but these negotiations are intended to reduce individual responsibilities, avoid risks, and assure broad equity (or "fairness") in the distribution of and accounting for financial resources. Walder concludes that it is more important to understand the characteristics of the environment that structure bargaining than to concentrate attention on the dynamics of the resulting bargaining itself.^[36]

Looking at rural urbanization, David Zweig examines the sphere where the reforms since the late 1970s reportedly had gone the farthest toward reducing bureaucratic authority in favor of having decisions driven by market forces. Small-scale rural enterprises have flourished with the dissolution of the communes and the policy decision to permit peasants to leave farming in favor of working in small towns. These

changes suggest that rural towns are relatively freewheeling places, but Zweig's research paints a distinctly different picture. While the production and marketing of agricultural produce are no longer monopolized by local officials, the county's control over development assistance and investment for expanding urban infrastructure has assured a continuing—and in many cases an expanding—role for the county government in county-towns and townships. Bureaucratic administrative control over rural town development, as a consequence, remains very powerful, and the characteristics of that control—such as lack of institutional constraints on the authority of upper levels—will sound very familiar to those who have read the other chapters.

The Schroeder, Walder, and Zweig contributions draw remarkably similar pictures in their examination of various subnational levels in China. They all agree that the reforms significantly decentralized control over economic resources, that subnational bureaucracies became more important in decision making on economic issues, that the system lacks clear rules and a stable distribution of authority, and that these circumstances have combined with vague or inconsistent policies from above to produce and contour widespread bargaining behavior. Very considerable budgetary power, moreover, now appears to be concentrated at the level of the municipality and of the rural county (at least, in the latter case, vis-à-vis the rural townships). This concentration of power has not reduced bargaining—it has, rather, reshaped it.

These explorations into four bureaucratic clusters and of the top and bottom of the system challenge and modify the fragmented authoritarian model in several important ways. They may be briefly summarized as follows:

The reforms have not produced straightforward decentralization of China's bureaucratic system. Rather, they have had complex effects. Reform policies have consciously decentralized decision making in the economic sphere and have given lower bureaucratic levels more control over fiscal resources. The reduced role of ideology has reinforced fragmentation of the system. But countervailing trends, such as enhancement of the Center's authority to acquire and analyze information, have also been nurtured by the reforms. The resulting system structures new bargaining relationships and dynamics of decision making, but the Center continues to hold serious cards in this economic game.

Fragmentation of the bureaucratic system is most severe in the domain from the ministries through the provinces. Above the ministries and below the provinces, this is a political system characterized by extraordinary concentrations of power. The key to whether bar-

gaining occurs at the bottom of the system is not simply the extent to which power is fragmented.

The reforms since the late 1970s have produced only very limited progress toward institutionalizing the political system. While top leaders no longer launched disruptive political campaigns to shake up the bureaucratic institutions, the system as a whole failed to develop institutional ways of allocating authority on a stable basis. The legal system does not function in a fashion that enables it to adjudicate key issues and establish stable precedents. Law and regulation combined do not pose effective bars to the adoption of policies that redistribute power and violate past commitments.

While formal rules do not constrain the top leaders in reshaping the system, three other factors do limit considerably the effective leverage of those leaders. First, officials at lower levels tend to ignore or circumvent top-level decisions that are vague or inconsistent. Second, substantive policy goals, especially those of the reformers, demand that top leaders allow their subordinates considerable leeway, as any other approach chokes off information, reduces enthusiasm and creativity, and precludes China's developing the dynamic society envisioned by the reformers. Third, top leaders generally recognize that policy implementation, especially regarding major economic projects, can be slowed down and made more difficult by unenthusiastic provincial and lower-level officials, and thus they often seek to bring those officials on board rather than to coerce them into compliance.

For many parts of the political system, processes other than bargaining tend to play very important roles in determining how bureaucratic units deal with each other. These may include, inter alia, the personal maneuvering at the apex suggested by Carol Hamrin, the competitive persuasion noted by Nina Halpern, or the coping mechanisms detailed by Lynn Paine. Bargaining is more likely to occur where tangible resources are at stake, both parties need each other, and the rules that govern decisions are not fixed and clear. Quite often, one or both of the first two of these conditions is not present.

Another way to put the findings in this volume into perspective is to look at the Chinese political system as being in transition from a traditional hierarchical system toward a more modern, market-oriented system. In the former, activities are guided primarily by traditional vertical relationships within the bureaucratic apparatus, while in the latter a wider range of activities is shaped by pure rule-guided and especially market relationships. This is a basic set of changes sought by the post-1978 reforms. Figures 1.1 and 1.2, which build upon, modify, and am-

plify a presentation David Zweig made at the Tucson conference, provide a framework for an understanding of the resulting situation.

In the traditional system, relations are shaped primarily by informal criteria, such as personal connections and actual control over resources. The more modern variant posits a greater role for formal institutional boundaries, accepted rules, and laws. In short, in the more modern polity, the role of personal elements is reduced and that of more formal criteria is enhanced.

As is indicated in these charts, for each of these two types of polity there are three basic structural relationships: those where both parties

are inside the bureaucracy, those where one is in and one is out, and those where both parties are outside the bureaucracy. Within each of these, moreover, the charts distinguish relationships among unequal parties (vertical/asymmetrical relationships) from relationships among basically equal parties (horizontal/symmetrical relationships). The resulting ranges of types of behavior for each structure of relationship in the traditional and modern polities are indicated.

In the more modern version of vertical/asymmetrical relationships, pleading is replaced by persuasion and patron-client ties are replaced by rule-guided behavior (box 1), corruption (including rent-seeking) is supplanted by service according to regulations (boxes 3 and 4), and assistance is rendered on the basis of merit (box 5). The modern analogue of horizontal/symmetrical relationships substitutes rule-guided behavior for bargaining and corruption/rent-seeking and has pure market relationships replace guanxi (boxes 2, 4, and 6).

To repeat, an objective of the 1980s reforms in China was to shift from a situation where activities are guided primarily by traditional vertical

relations within the bureaucratic apparatus to one where a wider range of activities is shaped by pure rule-guided and especially market relationships.^[37] Thus, the reformers sought to move the polity from a bureaucratic leviathan that operates in traditional fashion toward a polity that acts in more modern fashion and in which the bureaucratic apparatus plays a narrower role.

As the chapters in this volume confirm, the reality of the reforms fell far short of the ultimate goals of shifting to more modern forms of behavior and of sharply curtailing the scope of bureaucratic activity. Rather, the major changes in the 1980s moved an increasing array of decisions from the "Both In" vertical cell (box 1) in the Traditional Polity chart to the "Both In" horizontal cell (box 2) and the "One In, One Out" vertical and horizontal cells (boxes 3 and 4) in that same chart. The actual scope of bureaucratic activity remains surprisingly pervasive,^[38] and these changes reflect the fact that the reforms in many ways flattened China's bureaucratic hierarchy and increased the importance of nonbureaucratic sectors without changing the nature of relationships to those of a more modern polity. Only within the decision-making apparatus developed directly under Zhao Ziyang's aegis in the Zhongnanhai, as is detailed by Nina Halpern, did behavior move in very substantial measure toward the appropriate box (box 1) in the chart Modern Polity.

One important result of these shifts that is not highlighted in this volume is that they have nurtured corruption and rent-seeking behavior both within the bureaucracies and between officials and the population. These behaviors reflect in part the combination of the rapid development of nongovernmental efforts, the commercialization of many governmental activities, continued bureaucratic power to intervene in the market, and confusion over norms. Note that these forms of behavior predominate even as the unofficial party acquires resources that create a relatively symmetrical relationship with the official party. Thus, the very process of reform itself has created conditions that nurture corruption by relaxing former rules and habitual practices, diffusing authority and resources, and creating confusion over guiding norms.

The charts and the discussion of the fragmented authoritarianism model reflect a preliminary way of looking at the trends in the kinds of

relationships—and in the resulting behavior—that are occurring under the reforms in the PRC. The particular data in the various chapters flesh out this framework in some detail. No polity operates in a purely "modern" fashion (nor should it), but the differences between low levels of institutionalization combined with expansion of the role of the market to produce widespread rent-seeking behavior and corruption, versus far greater use of law and regulation to achieve market-driven outcomes, is serious and important.

The Volume In Perspective

The chapters in this volume are far richer than the above comments can convey. Each provides analytical insights, a "feel" for the bureaucratic arena being addressed, and, typically, illustrative material that conveys a nuanced appreciation of the forces that shape political and bureaucratic outcomes in China. Overall, however, they do not provide substantial comment on the extent to which the system has changed over time, on the role of the Chinese Communist Party in the polity, on the impact of culture on policy process, and on relations between state and society. These are all important issues for putting into perspective the content and coverage of chapters that follow, and this "Introduction" thus concludes with a few comments on each of these matters.

How Constant Has China's Bureaucratic Practice Remained?

The 1980s reform era created the impression of major change in the way China was governed. In part, those perceived changes were substantive and important. As was noted above, for example, the reforms significantly redistributed flows of information in the system and greatly reduced the role of ideology as a factor in structuring policy formation and implementation. But there is a danger of exaggerating the changes that the reforms produced in China's bureaucratic practice. As figures 1.1 and 1.2 highlight, the system has moved only very partially in the directions sought by the reformers.

Clearly, there have been important continuities as well as changes. The fundamental structure of the Chinese bureaucratic system that was established in the 1950s, for example, remains in place to the present and continues to exert tremendous influence on policy process. Mao Zedong himself altered the scope of authority of the various bureaucratic clusters over time in his quest to keep the political system responsive to his desires. Deng Xiaoping and his reform-minded colleagues continued this practice by considerably enhancing the resources and authority of the economic cluster at considerable cost to the organization/personnel, propaganda/education, coercive, and Party territorial clusters. At the beginning of the

1990s it appears that hard-liners may attempt to increase the resources of the five "losing" clusters of the reform era as part of their strategy for building support for turning back some reform initiatives. While all of these efforts have redistributed authority and resources, none has fundamentally changed the nature of the system. The "losing" clusters during each period remain in the wings as potential resources for political contenders who seek a change of course.

A number of additional factors make it difficult to judge the extent to which bureaucratic practice under Mao differed from the findings presented in this volume. For the Maoist era, we obviously lack the kind of detailed studies based on direct access that are contained here. In addition, the Chinese media were far less informative about this earlier period than they became during the 1980s. And many Chinese interviewees, acting in the best of faith, nevertheless tend to recall past situations in conformity with the current official views concerning those previous periods. Thus, the 1980s demonology concerning the Maoist era has affected recapitulations of decision making during that era.

Even the extent of pressures for change effected by the 1980s reformers is not unprecedented in the PRC. The Chinese reforms beginning in the late 1970s sought changes in important areas: bureaucratic organization, the scope of responsibility and definition of tasks of key bureaucracies, the distribution of bureaucratic resources, and the nature of the process by which decisions are made. In broad terms, the reform leadership of the country tried to make the system less personalized, less ideological, less centralized, and more sensitive to economic rewards for greater efficiency and dynamism.^[39]

But the bureaucracies under Mao also had to adapt to very different environments in terms of the intensity of ideological pressure,^[40] the openness to the outside world (i.e., to the USSR and Eastern Europe during the 1950s),^[41] the decision-making models they should follow,^[42] fiscal and budgetary environments,^[43] and so forth. In short, pressures for change of this magnitude have been a recurrent feature of China's

bureaucratic world.^[44] As is detailed in this volume, moreover, the 1980s reforms achieved only part of their ambitious agenda for change.

The question concerning the extent to which the bureaucratic world detailed in this volume also characterized the Maoist period must, therefore, remain unanswered. Significant changes have undoubtedly occurred in some aspects of policy process, especially since the role of ideology in the system has diminished greatly and the reforms appear to have significantly flattened bureaucratic hierarchies. But there may be more continuity than we assume, and researchers should keep this possibility in mind as they undertake further studies on the Maoist era.

The system as described and analyzed in this volume probably will remain more constant in the future than changes in elite-level political rhetoric might suggest. It is made clear, for example, that, while there is no constitutional or even normative bar to the central leadership's radically altering the distribution of authority crafted by its predecessors, this would require either that one leader emerge as a new strongman or that strong agreement be reached among all the top leaders that the system should move in this new direction. Without such agreement, policy decisions will lack the clarity, consistency, and detail that are necessary to bring a high probability of lower-level compliance. Without such agreement at the very top, to put it differently, there is apt to be widespread sabotage of national directives by officials at each subnational level. If the top-level initiatives move the system toward a more centralized system producing less information, moreover, then lower levels will be in a position quietly to achieve greater degrees of freedom through manipulation of information that goes to the leadership. In short, by focusing on fundamental structure and process, we examine here factors that indicate that the Chinese system is not nearly as malleable as are the dynamics of political contention at the apex of the system.

At the time of this writing there is a possibility that China will experience dramatic political change in the coming years. The factors under scrutiny here are of such a fundamental nature, however, that to some extent any successor system is likely to embody many of the features and dynamics that are discussed in these pages. Major changes would, of course, be evident, but the issues explored would very likely retain significant salience even if leaders who reject communism were to govern China.

What About the Communist Party?

The Chinese constitution recognizes the Chinese Communist Party as the "sole leader" of the Chinese system. Even the major reform docu-

ments, such as Zhao Ziyang's speech to the Thirteenth Party Congress,^[45] reserve for the Party the responsibility of making all key decisions, of checking on their implementation, and of retaining the power of appointment of individuals to every important post. Yet the actual state of the Party—its organization, work methods, and to some extent even its bureaucratic identity—remain matters on which there is little information and much confusion among foreign scholars.

Discussion of the Party is bedeviled, of course, by the difficulty of gaining access to information. The Chinese press during the 1980s became relatively frank about problems in the government, but it treated the Party gingerly. Broad comments about difficulties in the Party tended to be supported only with flimsy anecdotal data. Solid research—if any has been done—remains unavailable to foreign analysts. Foreigners themselves understandably have felt that probing the details of Party organization and functioning would stretch too far the hospitality they have been accorded by their Chinese hosts. Yet, as Susan Shirk indicates in chapter 3, it is terribly important to clarify the role and dynamics of the Party in the Chinese political system.

More attention must be paid to career mobility within the Party, personal characteristics of Party members, personal ranks in the Party, and other factors that affect relative authority and that therefore contour behavior of Party officials. In addition, much more effort must then be made to understand the procedures by which issues are considered within the Party and the views of various Party bodies. It is, of course, quite possible that the extent of the Party's role varies considerably by sector—probably being more important in propaganda, organization, public security, and rural work, and less important in the urban economy and the military (although even this may vary considerably over time). Only empirical research can determine this. Until this research is considerably more advanced than at present, however, all comments about the functioning and evolution of the Chinese system—indeed, about the structure of authority within that system—must be made with some caution.

In sum, the fact that even a volume such as this, which investigates in detail the decision-making process in various bureaucratic arenas, basically is forced to give little consideration to the Chinese Communist Party should be put high on the agenda of research in the China field, although it is difficult to be optimistic about how much can be learned about the Party in the environment of the early 1990s. Given the fact that in most government units the top Party officials are also the top govern-

ment officeholders, it may be that lack of direct consideration of the Party itself does not seriously distort our understanding of policy process. But future research should seek to provide the empirical base for determining whether (or to what extent) this is the case. This volume does not satisfactorily accomplish this task.

How About Chinese Culture and Society?

The analyses provided in these chapters typically make little or no reference to Chinese cultural characteristics. It has been argued that Chinese culture is crucial for understanding political and bureaucratic behavior in the PRC, as cultural approaches explain the nature of political alliances, expectations of political behavior, attitudes toward authority relations, and even the fundamental strength of political organizations.^[46] These authors probably would not argue against the idea that cultural factors affect the style and some other aspects of political and bureaucratic behavior. For example, the weakness of formal legal authority, the legitimacy of virtually unbridled concentration of power, and a proclivity toward negotiation may all be in part attributable to cultural influences. The authors nevertheless find that they do not have to utilize cultural variables to explain the major behavior patterns they identify.

There is also little consideration of the relations of state and society. The various chapters include careful analysis of the relationships among top leaders, key staff organs, and various bureaucratic units. The chapters by Naughton, Walder, and Zweig also address the issue of the relations between local government units and various enterprises. The focus here is explicitly on the political system and its internal dynamics, however, not on the relations between this system and the larger Chinese population.

Organization Of The Volume

The volume begins with analyses that are national in scope. David M. Lampton considers systematically the conditions that structure bargaining behavior, and Susan Shirk analyzes overall reform strategy and implementation. In their focus on the Center, Carol Lee Hamrin examines the evolution, structure, and politics of the "leading groups" headed by the top leaders, and Nina Halpern explicates the impact on policy process of the development of think tanks at the Center.

Four authors then take up various bureaucratic functional clusters:

Jonathan Pollack on the military; Lynn Paine on the educational system; Melanie Manion on the personnel system; and Barry Naughton on the economic system. Finally, three authors focus specifically on subnational levels of the system: Paul Schroeder on the provincial level; Andrew Walder on the municipal level; and David Zweig on county level and below in the countryside.

Many of the "functional" chapters in fact deal with both national and local levels, and each of the last three examines the politics and policy process of localities in their relationship to other levels of the political system. Thus, although the chapters are arranged to march "down" the national governing hierarchy, the actual relationships among these chapters are more complex than may appear on the surface. In this tension between orderly appearance and complex reality, this volume imitates the bureaucratic system about which it is written.