Two
A Plum for a Peach:Bargaining, Interest, and Bureaucratic Politics in China

David M. Lampton

China is a centrally planned society without central planning. China is a centralized polity without centralized authority. These aspects are in dialectical relationship to each other.

—INTERVIEW WITH CHINESE CITIZEN, FEBRUARY 1988, IN NEW YORK

In Article 16 of the constitution adopted by the Thirteenth Party Congress we find the following: "In case of controversy over major issues in which supporters of the two opposing views are nearly equal in number, except in emergencies where action must be taken in accordance with the majority view, the decision should be put off to allow for further investigation, study and exchange of opinions followed by another vote." In the previous constitution, the section was the same except that the last word was "discussion."

Within connection networks, the law of exchange generally works. ... I will give you a plum in return for a peach. ... The practice of exchanging power for goods, exchanging goods for power, and exchanging power for power and goods for goods is extensively pursued and injures the public interest to profit the private interest, and lines one's pockets with public funds.

—PEOPLE'S DAILY, 10 DECEMBER 1986, P. 1

China's economic reforms have changed the monopoly of public ownership and the pattern of equal distribution. As a result different interest groups are emerging in the society. These groups, representing various economic forces, make different demands on the reforms. Conflicts arise as they evaluate the social changes from their own perspectives. ... Not only do interest gaps exist between different groups, but views of people within groups are divergent.

—CHINA DAILY, 4 APRIL 1988, P. 4

Some preliminary elements of this chapter appeared as "Chinese Politics: The Bargaining Treadmill," Issues and Studies 23, no. 3 (March 1987): 11–41. In making subsequent revisions, the suggestions of James Reardon-Anderson, A. Doak Barnett, Thomas P. Bernstein, and Harry Harding were most helpful. Finally, I thank Kenneth Lieberthal and all the participants in the conference on "The Structure of Authority and Bureaucracy in China" for their trenchant comments.

How are we to understand the structure of authority and the political process in China, particularly in the era of reform? How substantially have these structures and processes changed in the course of reform and how are they evolving?

Are the quotations with which this chapter opened simply the rhetorical fig leaves covering the knuckles of an iron-fisted and highly capable centralized regime? The violence of Tiananmen Square in mid-1989 superficially supports such a conclusion. Or, on the contrary, is there sown within these disjointed vignettes a serious understanding of the Chinese system of authority and the bureaucratic apparatus?

In this chapter I argue that bargaining is one of several forms of authority relationship in China , that it has been of central importance in the Chinese policy process throughout the Communist era, and that it became increasingly important in the first decade of reform in the post-Mao era. Even in the wake of the violence of Tiananmen, cooked up by an aging cabal of Party elders holding few (or no) formal positions of authority, bargaining remains a key feature of the system. By "bargaining" I mean an authority relationship of "reciprocal control ... among representatives of hierarchies."^[1] Bargaining has been conspicuous in technical and economic decisions, though it is by no means limited to this domain. Other forms of authority relationships are documented elsewhere in this volume: hierarchy and command, market relations, patron-client ties, pleading, and rent-seeking or corruption.

One of the principal conceptual contributions of this entire volume is to point out that there are many forms of political and authority relationships in China and that the kind of authority relationship depends on where in the social and bureaucratic hierarchies the respective parties are located, who the various parties to the authority relationship are, and what resources they possess. Bargaining occurs among proximate leaders, persons of equal rank, or among immediate superiors and subordinates. Bargaining is most in evidence when one is dealing with two or more bureaucracies of approximately equal resources, none of which can carry out an undertaking without the cooperation of the other(s), but which cannot compel the cooperation of the other(s) and cannot persuade a senior authoritative leader or institution to compel the other(s) to cooperate. Senior authoritative leaders may not intervene, because they lack the knowledge to decide, they do not care, their resources are insufficient to enforce a decree, or the leadership is itself divided.

To be more specific, the circumstances that favor a bargaining process are situations in which there is collective leadership, disagreement

among authoritative elites, parties of about equal bureaucratic rank, decisions of high complexity with multiple trade-offs, and decisions in which interdependencies are complex and extensive. Issues that must be resolved in such circumstances frequently are addressed through an intensive process of consensus building in which leadership, at all levels, is hesitant to act until there is a consensus among subordinates and among competing bureaucracies.

The Chinese system is distinctive, not because bargaining occurs (which is a generic feature of politics, per se), but because frequently so many individuals and organizations must agree or acquiesce before one gets action. Americans sometimes see themselves as uniquely hamstrung by a "checks and balances system"; the Chinese decision system often is hamstrung by a complex bargaining process and the need to build a consensus.

There are several reasons that bargaining became more prevalent in the decade of reform (1978–1988) following Mao Zedong's death: leadership became more collective, and political agendas were dominated by complex economic and development issues that reflected great interdependencies, many trade-offs, and high complexity. Perhaps even more important, the very notion of partial social interests articulating their needs was legitimized in the wake of the Chairman's death, and extreme collectivist ideology was delegitimized. Finally, the unrelenting growth in the size of both state and Party bureaucracies, combined with the decentralization of economic power, proliferated the number of organizational bases with clout.

Nonetheless, even under Mao there was extensive bargaining among localities and functional bureaucracies, reflecting the need to resolve conflicts among interdependent localities and bureaucracies when the rigidities of the vertical command system obstructed timely and appropriate decisions. The potential of the marketplace to make decisions was almost entirely untapped.

If one examines the literature on Chinese foreign and domestic politics, there is a curious and conspicuous disjuncture. Analysts have long realized that bargaining and negotiating processes are central to interactions with foreigners; the Chinese, we have been told, are master bargainers.^[2] However, when we look at domestic political processes, we see that

there has been insufficient appreciation of the role that bargaining and negotiation play in the consensus-building and decision-making process. Richard Solomon's observation about negotiations with foreigners applies with nearly equal force to internal politics. "Chinese officials sometimes give the impression that agreements are never quite final. They will seek modifications of understandings when it serves their purposes, and the conclusion of one agreement is only the occasion for pressing an interlocutor for new concessions."^[3]

This chapter will address these questions: What role does bargaining play in the repertoire of authority relations? What resources provide the greatest leverage in what types of bargaining situations? In what policy domains is bargaining most pronounced, in what areas is it least in evidence, and why? What do political participants bargain over, and who is entitled to enter the process? What tactics do participants employ to enhance their positions? What factors account for the behavior of particular bargainers? What consequences does bargaining have for the political system and its policy outputs and outcomes?

A Comparative Framework For Viewing Bargaining

Political leadership in any society has two inescapable tasks: the calculative function (to identify problems, to assess the options for solution, and to make choices) and the control or coordination function (to assure that multiple actors comply with policy and coordinate their actions when necessary). As Robert Dahl and Charles Lindblom explained long ago, in discharging these tasks, leaders must choose from a limited number of basic means: hierarchy or "command" structures, markets (or price mechanisms), voting and preference counting systems, and bargaining.^[4] Each tool provides both a means by which leaders and followers gain information and make choices ("calculation") and a means by which leaders control (or coordinate) subsequent behavior. Every society employs a combination of each tool, and every polity is therefore a "mixed" system; what distinguishes one polity from another is the mix , the issues with respect to which various tools are employed, the levels of the society at which various mixes prevail, and the stability of the mix.^[5]

Looking at the PRC developmentally, in the first decade of reform in the post-Mao era we see that there was a relative decline in the use of hierarchical command (and within many, but not all, hierarchies there was a delegation of authority downward). Within the elite there was more

collective leadership. There was an increase in the use of market mechanisms, but it proceeded only far enough to create opportunities for corruption and other rent-seeking behavior by those in the hierarchy who were best positioned to reap benefits from manipulations of the disjunctures between the administrative hierarchy and the marketplace. There was a limited increase in the use of voting and preference counting systems (particularly in county and below governance and in the use of public opinion surveys). There was a significant increase in the use of bargaining, because the agenda was dominated by complex economic and technical issues, with multiple trade-offs, in the context of a collective leadership and bureaucracies of growing size and technical complexity.

What do we mean by "bargaining" and what are its practical and theoretical consequences? Dahl and Lindblom provide a departure point. "Bargaining is a form of reciprocal control among leaders . ... Leaders bargain because they disagree and expect that further agreement is possible and will be profitable. ... Bargaining commonly means reciprocity among representatives of hierarchies " (emphasis added).^[6] Bargaining, therefore, is a process of reciprocal accommodation among the leaders of territorial and functional hierarchies. Bargaining occurs because these leaders believe that the gains to be made by mutual accommodation exceed those to be made by unilateral action (if that were possible) or by forgoing agreement altogether.

A bargaining perspective does not mean that China is on the road to free enterprise—indeed, bargaining can be one means by which established hierarchies endeavor to prevent the further erosion of their power, a way to avoid the increased use of market forces. Moreover, bargaining does not mean that established hierarchies are being dismantled; on the contrary, bargaining is one reflection of the fact that there are large, competitive bureaucracies and territorial administrations that are absolutely central to the functioning of both the society and the polity. Bargaining does not necessarily result in more coherent policies or more efficient governance. Bargaining is more akin to protracted guerrilla warfare within and between large-scale organizations.

Consequently, bargaining has sown within itself a number of pathologies among which are blocked leadership, minority veto, control by the organized (which in China generally means powerful bureaucracies, to include localities), and an inability to assure that the policy to emerge from one bargaining process is consistent with another, for coherence among policies is difficult to assure. Bargaining is a process characteristic of what Jerry Hough calls bureaucratic (and I would add localist) plural-

ism, though its cultural roots in China run much deeper than just the post-1949 bureaucratic structure.^[7]

The Systematic Causes Of Bargaining Activity

Bargaining results from several structural factors, many of which existed prior to the demise of Mao Zedong, and several of which made bargaining substantially more pronounced in the era of reform after his death. The enduring factors are several. First is the existence of massive, parallel, and interdependent (but inadequately coordinated) bureaucracies and territorial administrations, which must coordinate policy but which lack established mechanisms to do so without negotiations and protracted consensus-building efforts.

Second, divergent societal interests become embedded in the various state and Party bureaucracies, experts and technocrats come to play an increasingly potent role, and bureaucracies develop their own distinctive ideologies and cultures. Central, territorial, and bureaucratic leaders cope with the resulting complexity and conflict through accommodation and arduous consensus-building efforts.

Third, though the scope of state planning underwent important changes between 1978 and 1988 (with guidance planning assuming a more prominent role and mandatory planning playing a declining role), the process of moving resources in society was still a political and bureaucratic decision in considerable measure. Without a market to match supply and demand, and voting and preference counting systems practically nonexistent, elaborate bargaining mechanisms developed to fill the void.

Fourth, work by Vivienne Shue goes some distance in further explaining why bargaining activity has been an enduring feature of the Chinese polity under communism.^[8] To simplify Shue's rich argument: the Chinese polity is a "honeycomb" structure in which localist and familial values are dominant. The "Center's" capacity to impose its objectives and norms generally is quite limited; the alternative to compulsion has generally been negotiation. This honeycomb structure is the origin of what Naughton called "the implementation bias," that is, the situation in which every central initiative will be distorted in favor of the organization or locality responsible for implementation.^[9]

Fifth, territorial administrations and vertical functional organizations (the kuai and the tiao ) embody a variety of interests,^[10] and the minister (buzhang ) of a ministry has the same rank in the system as a provincial governor (shengzhang ). With respect to any given issue, specific ministries find that their interests and policy preferences correspond with, or diverge from, those of a complex array of other ministries and territorial units. Local leaders not only are advocates for their own territory's interests, they also are mediators of disputes among the autonomous, but interdependent, vertical hierarchies that intersect in their localities. Local leadership engages in this mediation with its own agenda. Central leaders at the commission, vice-premier, and Politburo levels become arbiters among provinces and ministries. One can view the Chinese hierarchy as a top-down command system or, sometimes more accurately, as an inverted sieve in which issues that cannot be resolved at lower levels are kicked up to the next higher level able to negotiate a resolution. As Susan Shirk recounts in chapter 3, there is a desire to resolve conflicts at the lowest possible level. The problem is, however, that deadlocks among entities at lower levels assure that higher levels remain overwhelmed by squabbling subordinates.

Finally, another enduring aspect of the Chinese political milieu is the deeply shared value among both superiors and subordinates that "fairness" exists when there has been "consultation" and when the outcome of "consultation" is not to leave an individual, family, locality, or organization without adequate wherewithal to subsist and accomplish its assigned duties, unless there is a self-evident and overriding social interest that can be demonstrated. Even then, fairness requires that there be "just" compensation. What constitutes "just compensation" can become the subject of protracted negotiation.

In addition to those factors that have accounted for bargaining relationships throughout the Communist era, the first decade of reform created circumstances that favored even more bargaining behavior. Most notably, during the period 1978–88 China was in the awkward transition stage of reform. A small, growing, and dynamic market sector uneasily coexisted with the rigidities of the still dominant administered economy. The scarcities and rigidities of the dominant system provided opportunities (and problems) for the decontrolled sector, and vice versa; bargaining behavior (legal, illegal, and quasi-legal) of all sorts was the result.

Moreover, the pursuit of individual, local, and organizational interest was legitimate in a way in which it never was under Mao. And, the

pursuit of interest became absolutely essential in China's greatly expanded intercourse with the outside world after Mao's death.

Further, the very nature of the post-Mao agenda—modernization and economic growth—lent itself to bargaining, for a number of reasons. The agenda was dominated by economic and technical choices in which the trade-offs among issues were complex, and powerful domestic bureaucracies frequently found themselves at loggerheads without the benefit of market, voting, or command systems adequate to resolve issues.

And finally, structurally, prior to June 1989, there was a carefully balanced Politburo Standing Committee of Zhao Ziyang, Hu Qili, Li Peng, Yao Yilin, and Qiao Shi. The cleavages evident among the bureaucracies and localities were mirrored in the leadership. An elite that could not agree among itself found it difficult to enforce its will on recalcitrant and deadlocked subordinates.

What Do Leaders Bargain Over?

I shall first indicate the breadth of issues over which bargaining occurs and then examine two specific cases that reveal the process more clearly. While the issues and arenas recounted below emphasize national organizations and national issues, these processes are mirrored throughout the bureaucracy and units at all levels. The range of policy issues that generate bargaining is broad. One bargains over what is scarce: in the PRC, financial resources, power and position in the hierarchy, high-quality goods and services, and access to the international system and to highly skilled personnel are among those things most sought.

Bargaining is intense in the budgetary process. Budgetary resources are allocated among the various functional "systems" (xitong ).^[11] Each "system's" prior share of the budget, particularly the "operating" portion of the budget, is the base from which marginal changes are negotiated for the next year—"the fixed sum system."^[12] Within "systems" and individual bureaucracies, resources (particularly "operating" resources) are treated as a "lump" (kuai ) to be "carved up" (qie ) among subordinate entities according to the percentages previously applicable.^[13]

Bargaining, therefore, occurs around the edges of the budget. Which units will suffer marginal cuts and reap modest gains, and which units will have their budgets charged for investments that benefit other systems or organizations, all become important questions.^[14] One of the recurring

budgetary issues in building the Gezhouba Dam, for instance, was how much of the budget of the then Ministry of Water Conservancy and Electric Power (MWCEP) should be spent on increasing the lock capacity of the dam to meet the needs of the Ministry of Communications, the agency responsible for inland shipping. The Ministry of Communications could make demands, knowing that the cost of meeting its desires would not come from its budget, because all project expenditures are charged against the budget of the lead agency, in this case the MWCEP.^[15]

In speaking of how investment allocation decisions are made, Barry Naughton explains, "We can speculate that ... actual allocation decisions are determined largely by the influence that different Beijing-based bureaucracies can bring to bear, and by various ad hoc sharing arrangements. ... Ministries struggle to protect their power bases and keep subordinates busy; in order to succeed in this struggle, they must insure that at least their share of the total investment is not too drastically reduced."^[16]

Revenue raising also involves negotiation. In the early 1980s, when Beijing was experimenting with a system in which enterprises were being permitted to keep a portion of profits ("profit retention")—which effectively reduced central revenues in the short run—"enterprises negotiated long and hard for the best possible retention rates."^[17] As was explained in a 1982 interview in a Beijing ministry, many factories'

[NEX]

retention rates are decided by the local bureau and province. The State Planning Commission has a "general principle" that the profit retention rate should not be higher than 12–13 percent, though he [the interviewee] noted it has changed every year and the situation has been "chaotic" [hen luan ]. The average is about 10 percent, though in some cases it is as low as 5–6 percent and in some cases it is higher. [Critical to determining where in the permitted range the allowed profit retention will fall is an assessment of the degree to which the enterprise is disadvantaged by the price system.] ... For example, the price for agricultural machinery is low and in favor of the peasants. So, they [the agricultural machinery enterprises] cannot change the sales price, but they can try to get to retain a higher percentage of the profit.^[18]

Similarly, with the experimental implementation of a tax system in the 1980s that created the prospect of enterprises keeping more money and

then remitting taxes directly to the Ministry of Finance (thereby bypassing counties that had previously taken a slice of the financial pie), some counties began to discriminate against enterprises in which the new revenue system was first being introduced, "and may have demanded kickbacks in exchange for supplying the trial enterprises with the desired commodities."^[19] In summarizing the politics of the process of moving from a profit remission to a tax-based revenue system, Bachman says:

In an effort to win approval for ligaishui [the substitution of taxes for profit remission], central leaders apparently compromised on contentious issues. The Center made two fateful agreements that have checked the more revolutionary implications [of change]. ... Beijing stated that enterprises would retain about the same amount of money under [the new system] ... as they had retained under profit retention. It also announced that there would be no change in the distribution of central-local finances. In other words, to overcome the (potential) resistance of key local interests (factory managers and local officials), the Center agreed that the redistributive dimensions of the ligaishui would be minimal.^[20]

Lieberthal and Oksenberg note the complex revenue-sharing deals that were worked out in Shanxi province. In speaking of fourteen large coal mines, they observe:

In the past, these large mines transferred all their profit to the provincial government . Under the new revenue system [ligaishui ], the mines no longer remitted all profits to the province. Instead, they [the mines] retained their profits and paid a tax [to the Center]. ... The Center then devised a way to compensate Shanxi for this loss in revenue, namely, to reduce the amount which the Center collected in revenue from Shanxi (emphasis added).^[21]

Localities and bureaucratic organizations also bargain over new revenue sources and subsidy levels. Authorities in one county explained to me how the Center and localities had negotiated an arrangement whereby localities would build small-scale hydroelectric plants (which the Center wanted) in exchange for the localities' being able to dispose of the resulting revenue as they wished (yusuanwai ). Moreover, any excess electrical power beyond local needs would be purchased by the centrally managed power grids, at a high price, even if the grid did not need (and could not use) the energy at the time the locality wished to sell it.^[22]

In one Beijing interview, the rationale for electrical price subsidies was explained. "I then asked why they subsidize small electrical power plants with higher purchase prices and why they sell power in the Gansu highlands at a way lower price than elsewhere? He said in explaining subsidies, 'political factors are key.' 'Sometimes cost benefit analysis counts for nothing.' ... [He said] that provinces are important political powers and you can't ignore them."^[23] Provinces are potent bargainers because they often are represented in the Central Committee and have complex inter-personal and other ties with those in the elite; because they have the principal power to appoint (with central approval) those occupying key positions in both territorial and functional units within the province; because they often are major sources of central revenue; and, in the end, because it is they who must implement policy.

In late 1987 a group of economic officials was in the United States. On the West Coast they became involved in a discussion about how the government in Washington, D.C., decides where to build various construction projects. They were told by the American respondent that political connections played a decisive role, to which one member of the Chinese group responded, "Why, that's not very different from the way we do things in China."^[24]

In the more consumer- and profit-oriented environment of 1978–88, the desire of firms to enter new, growing, and more lucrative markets, and the desire of ministries and enterprises previously in those markets to protect their shares, gave rise to competition and an intense bargaining process. Take washing machines, for instance. They are comparatively simple to make, previously they were manufactured by the light-industry ministry, and they are in high demand, with good profit margins. Predict-ably, the Ministry of Light Industry did not want the Ministry of Machine Building to begin making washing machines. But, the State Planning Commission (SPC) approved this, and many conferences were held as a result.^[25]

Indeed, conferences to bargain over market shares and product lines appear to be a way of life. For example, in 1981 there was a conference on machines for civil use, attended by the SPC, the Ministry of Machine Building, and other related ministries and localities. The conference covered bikes, radios, fans, TVs, watches, clocks, electric meters, refrigerators, sewing machines, and washing machines; it fixed production levels and determined which factories would make what. The conference

lasted ten days, and the meeting was preceded by discussions that occurred over a year.^[26]

Ministries (and other units) also clash continually over their respective jurisdictions. Oksenberg recounts the difficulties that Bo Yibo encountered in efforts to reduce the overlapping and duplicative organizational structure in the shipbuilding industry:

The solution which Bo and other top leaders embraced ... was to group factories in a single industry into a single, independent corporation operating directly under the Machine Building Commission. A pilot project in the shipbuilding industry was to group the major shipyards in Shanghai into a single corporation. Previously, the shipyards were under several jurisdictions: the Sixth Ministry of Machine Building, the Ministry of Communications, and several municipal departments. However, neither the Sixth Ministry nor Communications wished to lose their shipyards. ... Bo Yibo, with the staff of the Machine Building Commission behind him, nevertheless had to conduct the extensive negotiations for the formation of the new corporation personally. Several trips to Shanghai were necessary. Even then, with all of his prestige, the result was a hybrid organization. The ministries concurred only when it was decided the head of the new corporation would be one of the vice-ministers of the Sixth Ministry and the head of its board of directors would be a vice-minister of communications.^[27]

Concisely, therefore, bargains are struck over revenue sources, budgets, personnel, organizational jurisdictions, market shares, production rights, subsidy levels, investment allocations, and jobs. Anything that is scarce and is sought by organizations can provide the raison d'être for bargaining.

Water Projects: The Bargaining Process Up Close

Because water is a scarce resource with multiple uses (with one use frequently precluding another), and because water traverses administrative boundaries (with upstream users affecting the interests of those downstream), analyzing how decisions are made in this field clearly reveals the bargaining dimensions of Chinese politics. The controversies swirling around the planning, construction, and management of the Danjiangkou and Three Gorges dam projects reveal both the complex constellation of interests that must be accommodated and the broader character of the political system, at least with respect to economic, organizational, and technical decisions. Further, both of these projects involve ongoing decision processes that straddle the Maoist and post-Mao eras;

they thereby reveal the continual role of bargaining in the Chinese polity, as well as its even more prominent role in the era following the Chairman's death.

The Danjiangkou Dam

The middle and lower reaches of the Han River in Hubei province flood often; it is a constant menace to the major metropolis of Wuhan at the confluence of the Han and Yangtze rivers. In 1955 China's government began a flood-diversion project along the Han (Dujiatai), which was part of a more comprehensive development plan for the entire Han River Basin. In 1958 the State Council approved construction of the Danjiangkou Dam as stage two of that development effort.

During its construction a number of problems slowed the project, as is detailed elsewhere.^[28] Here the focus is on several issues that became the foci of protracted bargaining during the planning, construction, and management phases: the dam's height; the priorities to be assigned irrigation, flood control, and electrical power generation; the problem of relocating displaced persons; and the issue of local opposition.

When construction started on Danjiangkou, the dam's initially planned height was 175 meters, with the water level to be maintained at 170 meters (above sea level).^[29] However, throughout the entire construction phase (which ended in 1974), there was intense controversy over what the dam's height should, in fact, be. During construction, the bargaining process produced agreed-upon dam heights of 140 meters, 152 meters, and 162 meters, as well as the initial height of 175. Throughout much of the construction period, no one knew how high the dam would be; the foundation that was built could support a much higher dam. Further, even in 1989, debate was still going on about whether to heighten the dam—sixteen years after construction was "finished."

In 1965 the planners (who wished to minimize construction and displaced-person costs) accepted a dam height of 152 meters, but Hubei and Henan provinces (the two provinces that shared the resulting reservoir) wanted a higher dam (so they would get more electrical power and irrigation water). In 1966 the dam was approved for 162 meters (which Hubei and Henan wanted), but the water level was to stay at the 145-meter level (as in the 152-meter dam) until the problem of what to do with displaced persons could be solved. The negotiating process produced a perfectly predictable outcome. The difference between the two sides was split down the middle, with one side getting an acceptable dam height and the other avoiding the immediate problem of displaced per-

sons and attendant costs.^[30] One interviewee explained: "In 1965 ... the Center approved a revised dam height of 152 meters with a water level of 145 meters. Because Hubei and Henan were unhappy with this ... the Center agreed in 1966 to a 162-meter dam, with an initial water level of 145 meters to gradually be raised to 157 meters as the relocation problem was solved."^[31]

The essence of the problem was that as the dam's height rose the opposition of counties and special districts that would be inundated became more intense; the number of persons who would be displaced would grow, and the number of cadres who would lose jobs would escalate, and all of this would greatly magnify political conflict and increase expenditures. The kernel of the bargaining outcome was that nobody got all of what they wanted, when they wanted it; few persons, organizations, or localities lost everything (at least immediately); and the number of dislocated persons and financial expenditures rose more gradually than would have been the case had they proceeded with the initial plan.

Though we cannot see the entire bargaining process, aspects of it are clear. One of the most contentious issues concerned the fact that Hubei province would get most of the flood control and electric power benefits and Henan province would get excessive numbers of refugees (in proportion to its benefits). Intense negotiations between the two provinces were conducted to redress this imbalance.

The leadership of the two provinces got together and Hubei agreed to take 80,000 of Henan's displaced persons (yimin ). According to central figures, there were a total of 356,000 refugees, with 130,000 in Henan and 226,000 in Hubei province. ... They said the local figures put the total number of displaced persons at 390,000 [obviously higher than the central estimate]. I asked why Hubei agreed to assume the burden and was told plainly that Henan was poor, their displaced persons' plight was worse, and "Henan's benefits from the reservoir were not as great."^[32]

Perhaps because the Center pays local authorities for each displaced person, the central authorities had an estimate for displaced persons well below that of the local authorities, whose receipts went up with the number of refugees. In this kind of bargaining situation, proponents tend to minimize costs imposed on others and puff up the benefits to localities; the localities whine about the damage and mafan ("bother") and deprecate their benefits.

These negotiations resulted in a written agreement between Hubei and Henan, an agreement that specified the distribution of electric

power, water, and displaced persons.^[33] Henan, though poorer and weaker than Hubei along many dimensions, did quite well in the bargaining process. Why? First, desperation can be a valuable political resource. Henan's provincial leaders could continually argue that if the refugee problem was not adequately solved, they would be unable to effectively remove people from areas to be inundated. In all political systems, people who are highly motivated to resist can exact large tolls. Second, the delays that Henan could impose on "closing" a deal were, in themselves, very expensive for Hubei. Every year that Wuhan's industries were starved for power and floods caused destruction all along the Han River proved enormously costly to Hubei. Just by being able to slow progress, Henan could inflict costs on Hubei higher than the concessions Hubei would have to provide Henan to close a deal. In short, the weak can be strong!

A second ongoing issue has been whether the principal use of Danjiangkou should be electrical-power generation, flood control, or irrigation. To manage the dam to maximize one objective is to diminish the extent to which the other purposes can be fully realized. To use water for irrigation means that less water flows through the turbines to generate power. "The contradiction is sharp," I was told in one interview.^[34] To maximize power output, the water level needs to be kept high; to provide insurance against flood, the water level should be kept low. To irrigate fields, the water level should be high, but water should not run through turbines and thereafter "uselessly" (from the vantage point of the local farmer) flow downstream.

Each of these purposes is not only of more importance to some mass constituencies than others (e.g., electrical power for industry and water for peasants, to greatly oversimplify), each purpose is also organizationally embodied in a particular ministry or set of bureaus within ministries. Agricultural concerns are the responsibility of the Ministry of Agriculture, Forestry, and Animal Husbandry. Electrical power has its bureaucratic proponents in the Electrical Power Ministry (or in the Electrical Power Bureau when there is a consolidated Ministry of Water Resources and Electrical Power), heavy industrial organizations, and big urban areas with great concentrations of population and industry. As for flood control, the Ministry of Water Conservancy historically has seen itself as the defender of the peasantry against flood. Territorial actors also have different interests. For instance, with respect to Danjiangkou, Henan province (upstream) was concerned primarily about irrigation water (although Zhengzhou did want more electrical power), and Hubei's provin-

cial government (downstream) was more concerned with flood control and Wuhan's electrically starved industry (e.g., Wuhan Iron and Steel). One sees, therefore, a Byzantine process of negotiation within provinces (between the province and "its" special districts and municipalities and the many vertical functional organizations that intersect in its domain) and between and among provinces.

In the case of Danjiangkou, complex and literally never-ending discussions among ministries and localities occurred over the dam's utilization priorities. Initially, flood prevention was the first priority, followed by electrical power, irrigation, navigation, and aquaculture, in that order.^[35] Since then each ministry keeps raising the issue of the priority of "its" use, continually trying to reopen the case in order to better its position.^[36] Issues frequently are never resolved; they just ebb and flow over time.

Another contentious issue that resulted in the dam's not being built to its designed height, an issue that still weighed against raising it in the late 1980s, is the opposition of localities along the reservoir's edge. My interview notes recount:

Another reason the 175-meter dam was not built as planned ... was because of the displaced persons problem. The fact is, I was told, that Henan province, even though it had no particular need for the flood control aspects of this dam, would agree to the higher dam, but it was Henan's Nanyang Special District which adamantly opposed it then, and does today, for the obvious reason that it would be one of the areas to be inundated heavily.^[37]

In Hubei province, it is Jun Xian [county], with a long and illustrious history dating from the Tang, and Yun Xian. In Henan province it is Zhechuan Xian. These three old county towns would go under water and Xiangyang Special District in Hubei province and Nanyang Special District in Henan opposed it. According to ... [my interviewee], the provincial government in Henan would go along with this [raising the dam] because of the benefits to Zhengzhou [the provincial capital] and irrigation, but they can't persuade the affected localities.^[38]

The Three Gorges Project

The Three Gorges, which sits astride the Hubei and Sichuan border, is a strategic choke-point at which floods that originate in Sichuan province (and devastate Hubei, Hunan, and other localities downstream) could theoretically be contained. Moreover, the reservoir that would be formed by a dam in the Three Gorges could drive turbines that would energize

much of central and eastern China's energy-starved industry. Finally, raising the water level in the gorges could improve navigation and increase the size of ships able to reach Chongqing, now China's largest city. So strategic is the Three Gorges that leaders from Sun Yatsen, through Mao Zedong, to Deng Xiaoping, Zhao Ziyang, and Li Peng have considered the project.

The undertaking has been approved at least twice "in principle" in the post-1949 era (once in 1958 at the Chengdu Conference and again in April 1984),^[39] only to have the start of construction aborted because approval to commence actual building had been made contingent upon the resolution of myriad, presumably minor, technical, financial, and political details, such as, How high should the dam be? What is to be done with the displaced persons, inundated factories, cadres out of work, grain production and tax revenues that would evaporate, and administrative centers? Who pays, how much, for all this relocation, and where do you put the people? What would be the effect of such a huge project on upstream fisheries and estuaries and harbors above and below the dam? What would be the useful life of the dam, given siltation? How will shipping across the dam be affected, and how much shipping growth is it prudent to plan for? Is flood control best achieved by one gigantic project, or several smaller ones? Who will receive the resulting electrical power? Can it be efficiently transmitted to distant locations where it is most urgently needed, and at what cost? Is a dam of this scale safe? Or would it be, as Mao and the Ministry of National Defense at times feared, a huge "bowl of water on our heads"?

Each of these questions has no obvious answer, and the various possibilities all have their popular and bureaucratic constituencies, which are not averse to making their case in protracted consultations. I shall not recount here the more than three decades of the project's tribulations; these have been well documented elsewhere.^[40] But I can advance a generalization that I believe will accurately predict both bureaucratic and local behavior most of the time—an "iron law of bargaining." The locality or bureaucracy will almost always exaggerate the costs that another unit's proposal will inflict, minimize the benefits received, exaggerate the mafan (bother) to itself, exaggerate the benefits others receive, understate its own resources, overstate the resources of others, and generate one-sided data supportive of its case.

As a result of such processes, in the case of the Three Gorges Project, promoters of the project face a bargaining dilemma: to weld a coalition big enough to win support for the dam, they must provide benefits to a

vast constellation of groups. But this requires a dam so enormous that the resulting costs and negative outcomes create intense opposition, high financial expenditures, and other risks that top decision-makers are loath to ignore.

The system has been unable to reach closure on this issue after almost three decades of wrangling. In May 1983 the State Planning Commission (SPC), then headed by Yao Yilin, convened a meeting to assess the Three Gorges feasibility study, which had been submitted to it by the MWCEP and the Yangtze River Valley Planning Office (YRVPO). This study assumed a water level of 150 meters, which was comparatively low and thereby would reduce both negative outcomes (displaced persons and inundated urban and rural land) and simultaneously reduce benefits (to shipping, flood control, and electrical-power generation). Yao declared:

For more than twenty years the debate over the Three Gorges water resources project has concentrated principally on the problem of the dam's height. That a high dam generates more electricity and that the flood control results are better is easy to see. However, the inundation is too much, the investment is too big, the masses upstream are unable to agree, the burdens on the state finance also cannot be borne. However, the relevant ministries and localities have not been reconciled to the low dam and, because of this we have debated for many years and still are unable to decide. If we continue to debate, I think this generation of ours will be unable to accomplish anything on this.^[41]

The battle continues without respite. In December 1988 the cover on Beijing Review proclaimed, "Three Gorges Project Given the Go-Ahead." Scarcely a month later Vice-Premier Yao told the fourth meeting of the Standing Committee of the Seventh Chinese People's Political Consultative Conference, "In the next five years, it is absolutely impossible to start the Sanxia Project, so people do not need to spend too much energy debating the issue for the time being."^[42]

Who Bargains, In What Arenas, And What Strategies Are Employed?

Who Bargains?

Generally, only proximate leaders bargain—that is, equals in the hierarchy, and entities (persons, organizations, factions, localities) one step above and one step below that level. These are the actors that can legitimately make demands. For instance, provinces bargain with one another and

with ministries as equals, with prefectures and counties immediately below, and with the primary central commissions (e.g., the SPC, the State Economic Commission, etc.) immediately above.

One's capacity to affect policy content (as distinct from implementation) diminishes greatly as one moves downward from the point of decision . In policy formulation, bargaining goes on between the principal bureaucracies and territorial administrations, except at the lower reaches of the hierarchy, where low-level territorial and bureaucratic actors must take account of the very real, albeit more diffuse, social forces that they seek to manage. Nonetheless, there is a representative quality to the process, inasmuch as each bureaucratic or territorial actor sees part of its job as being to reflect at least the minimal interests of subordinates, not so much out of democratic considerations as out of the realization that effective implementation requires the cooperation of subordinates and the realization that one's own interests are inextricably linked to one's organization or locality.

Depending on the policy issue, the implementation process not only involves intense bargaining among the territorial and functional bureaucracies, it also frequently involves officials dealing with a broad range of individuals and small social groups that have no formal standing or political role, yet who must be reckoned with if policy is to be effectively implemented, at least without coercion. As the modernization of China proceeds, the fragmentation of Chinese society proceeds apace, and as it does, the array of social forces affecting implementation will multiply. As this occurs traditional Chinese fears of chaos and immobilization are not unwarranted.

How to meet urgent problems, of a vast scale in a timely manner, in a bargaining system that seeks consensus amidst increasing pluralization is a genuine problem for the system. Obviously, there are alternative tools at hand, such as command and hierarchy or greater use of the market. Each of these also has liabilities; more command and hierarchy reduces flexibility and innovation; increased use of the market increases inequalities and accelerates pluralization.

In What Arenas Does Bargaining Occur?

There are several arenas (or types of arenas) that play central and recurrent roles: the SPC, other commissions, the Standing Committee of the State Council,^[43] the Politburo, state councillors, ad hoc interprovincial and interministerial committees, the National People's Congress, the national finance and planning conferences, materials-allocation confer-

ences, central work conferences, and specific policy-issue committees. These forums are replicated at all system levels . At this, we see only the tip of the iceberg of conflict-resolving arenas. Below I concentrate simply on national arenas.

The SPC (in 1989 consisting of the pre–April 1988 State Planning Commission and the old State Economic Commission) is one of the most important bargaining arenas. As a commission, the SPC stands as a buffer between individual ministries and the State Council. In an interview I was told that the way in which allocations among ministries are determined is "complicated" and that there is "lots of discussion." The interviewee went on to say that every province and municipality wants allocations that they consider small but that become a huge sum when aggregated. The SPC is the organ charged with mediating these conflicting claims and avoiding deficits and material bottlenecks.^[44] Because the SPC is divided into functional bureaus, conflict patterns within the organization tend to reflect those in the wider bureaucratic environment.^[45]

Another, and indeed higher, arena for bargaining among territorial and functional interests is the State Council, meaning either its Standing Committee (with a large permanent bureaucracy of its own, which needs a great deal more study) or the entire State Council. Conflicts that cannot be resolved by individual commissions or ministries are referred upward to this next higher level. For instance, to paraphrase one interviewee, "Say there is a high value investment to be made ... and say four provinces or municipalities all want it ... obviously this is a difficult task. ... The SPC can reach agreement with mayors, but if that is not possible, it then goes to the Standing Committee of the State Council."^[46]

It appears from my interviews, as well as from the work of Oksenberg, that one of the key systemic problems is that an excessive number of issues cannot be resolved by the ministries and localities themselves, even though, as Susan Shirk says in chapter 3, there is a desire to resolve issues at the lowest possible level. This can overwhelm the top echelons of the State Council (and Politburo) in a Niagara Falls of issues, a torrent whose volume is greatly increased by the fact that market mechanisms are not in place that elsewhere greatly reduce the number of items requiring conscious bureaucratic (political) decision in the first place.

In an effort to reduce the number of issues kicked to the next higher level, the Chinese have repeatedly tried to merge organizations that habitually conflict (and yet need to coordinate policy). The Water Conservancy and Electric Power ministries are just one of several such cases. In 1982, for instance, the Water Conservancy and Electric Power

ministries were merged. As one interviewee expressed the logic of the move, discussions would now be "in the family" rather than "between families."^[47] This union ended in divorce in April 1988, once again, with the re-creation of the Ministry of Water Resources and the merger of the Electric Power Ministry and other units into a new Ministry of Energy Resources.^[48] What is fascinating is that even when the water-conservancy people had been joined in a "shotgun union" with the electric-power people, individuals from each of the old ministries still referred to their new organization as the Ministry of Water Conservancy and Ministry of Electric Power, respectively.

Not only are conflicts resolved in the SPC and the State Council, there also is a rich repertoire of regularized and ad hoc procedures, conferences, and committees to resolve disputes in both the formulation and the implementation phases of the policy process at all system levels. Among the most important are the National Finance Conference and the National Planning Conference. One informant described the National Planning Conference as a Chinese market in which delegations from various provinces and ministries worked out deals. Lieberthal and Oksenberg explain that in the planning process the Ministry of Finance (MOF)

must go over the draft plan from the SPC with a view to its financial feasibility and implications. Not surprisingly, the perspectives of the three major participants in this process (the line ministries and provinces; the SPC; and the Ministry of Finance) often differ as to both the revenue and expenditure implications of their proposals. ... The key forums for resolving the differences among these various groups in the annual planning cycle were generally the National Planning Conference and the National Finance Conference. ... Since not all problems could be solved at these meetings, a central work conference that brought together provincial Party first secretaries as well as Politburo members and other key officials usually met to reach final decision on particularly contentious issues. The Politburo itself met separately to determine its position, where necessary.^[49]

When the Politburo finds itself deadlocked, the supreme leader, in this case Deng Xiaoping, and before him Mao Zedong, becomes the court of last resort. From a system-development perspective, the critical questions for the future are these: Will the system be able to reduce the number of issues bumped up to this level for resolution? And will individual leaders assume less importance as institutions achieve greater legitimacy?

Finally, one of the most interesting developments of the first decade of reform in the post-Mao era was the rise of the National, and local,

People's Congress (NPC) as an arena for bargaining. For instance, during the March 1989 session of the NPC, Xinhua carried the following report, quoting one NPC deputy named Yang Lieyu: "However, he noted, as most provincial governors and mayors of major cities are NPC deputies, who usually spoke a lot, asking for everything ranging from favorable policies to energy and raw materials, from cabinet members and ministers, deputies' group discussions sometimes became bargaining sessions between the central government and local authorities."^[50]

Strategies for Bargaining and the Necessary Resources

Strategies are contingent upon the actor under consideration, the policy issue in question, whether one is trying to promote, or frustrate, a specific initiative, the bundle of interrelated issues on the agenda at the same time, one's resources and position in the hierarchy, and the wider social-political-economic contexts. A few examples may serve to illustrate the rich and sophisticated diversity of strategies employed.

Foot-In-The-Door And "Fishing Projects."

In this gambit, the bargainer tries to secure a commitment that will permit work on a project to start. The idea is to obtain an initial commitment that is nonthreatening to potential or actual adversaries but that will not preclude the possibility of enlarging the project at some later date, thereby creating a situation in which each stage's sunk costs (combined with the presumed benefits of the next phase) become justification for taking the next step toward the initiator's ultimate objective.

At the Fourth Session of the Sixth National Chinese People's Political Consultative Conference, a speech by Qian Jiaju concisely explains the strategy: "In the past 30-odd years ... we had a fondness for big projects but ignored actual effects, and thus paid an enormous price ... from those "fishing projects" (which were said to need limited investment at the beginning according to the planners but later involved more and more additional funds after they were started). We should not follow this stupid practice any more."^[51]

For this strategy to work, several conditions must be met and several resources must be in the bargainer's possession. First, and most important, the decision under consideration must be separable into stages, and each stage must generate sufficient benefits to be defensible in its own terms. For instance, dams can be raised, and each level of dam can, at least possibly, provide benefits that justify each step in its own terms. Obviously, some decisions are all or nothing, and any benefits are contin-

gent on completion of the whole process. Such issues make this strategy difficult to employ.

The bargainer seeking to use this strategy must have the capacity to disguise "his" ultimate objective, or at least convince potential opponents that the first step will not offend their interests and that subsequent steps are not preordained. Some decisions cannot credibly meet this condition, and some bargainers simply are too identified with the ultimate objective to disassociate themselves from it. For instance, because the YRVPO has worked for a high Three Gorges Dam for decades, any attempt by it to promote a more modest project would be seen as the entering wedge for its more expansive ambitions.

Whipping Up Support and Faits Accomplis

In this strategy, local or ministerial leaders use the media to fan support for a project, begin work on it, and then point to their responsiveness to "public opinion" as the rationale for, in effect, having presented superiors with a fait accompli. "And, if one whips up opinion through newspapers and broadcasting stations and starts the so-called preliminary work before the feasibility study of a project is completed, so as to make the project an accomplished fact and force the central authorities to agree with the plan, one's practice, I think, is definitely a violation of the basic procedure."^[52]

For this strategy to be feasible, the bargainer must have resources that he or she can independently deploy without central approval. As Naughton points out in his study of the loss of central control over investment resources, the "localization" of control over capital was one of the most profound changes of the first decade of post-Mao reform.^[53] The combination of decentralized resources and decentralized decision authority have created a situation in which the Center has been presented with faits accomplis on an enormous scale in the capital construction area.

Another precondition for this strategy to be effective is that the sanctions for presenting higher authorities with faits accomplis cannot be too severe. The fact that so many local leaders have felt free to proceed in defiance of central preferences is one indication of the Center's diminished reach in the wake of 1980s reform.

Painting a Rosy or a Black Picture—Cooking the Books

In the same way that supporters of an initiative try to minimize perceived costs and uncertainties, opponents frequently exaggerate them. For instance, in one interview I asked about the price tag for the Three Gorges Project,

noting that I had seen cost estimates that varied by a factor of five. My respondent, who was an ardent supporter of the project, bluntly replied that those who advance extremely high estimates "oppose the Three Gorges Project."^[54]

For this strategy to be effective, the bargainer must possess seemingly credible information. It is no accident that each ministry and locality has its own statistical units that tend to produce data supportive of local or organizational goals. I noted above, for instance, that central and local estimates concerning the number of refugees that would be generated by the Danjiangkou Dam diverged in a predictable fashion, with each side of the dispute promoting data most consistent with its interests. Similarly, respondents at one ministry at which I interviewed assumed that the figures I quoted from another ministry had been distorted in a way consistent with that other ministry's interests. One implication of this process is, of course, that more information will not necessarily speed up the decision process, unless the data are collected using agreed-upon methodologies in the first place.

A Little Something for Everyone

This is a coalition-building process in which the scale of an undertaking is enlarged to provide benefits to all of the strategic groups that could obstruct agreement. In cases such as the Three Gorges Project, there may exist a situation in which any enlargement of the scale of the project not only attracts some wavering elements, it also scares away others by virtue of the project's very size. The costs of building a big enough coalition may simply be prohibitive, both economically and politically.

Getting to Key Decision-Makers, "Old Friends," and Relatives

The utilization of political networks cannot be overlooked.^[55] In the bargaining process, one's success may hinge on the "connections" (guanxi ) one possesses and the IOUs that can be collected. Personal networks, though not limited to organizations, are built into them. A leader, such as Yu Qiuli, who has extensive personal networks throughout the planning, energy, and military bureaucracies has great power and influence, not only because he directly controls powerful organizations, but also because he has been able to place loyal friends into a broad range of other organizations.^[56]

When asked how political actors promote their interests, one inter-

viewee placed particular emphasis on face-to-face meetings with decision-makers and contacts with old friends. Close personal ties with decision-makers, he asserted, could be decisive.^[57] Also not to be overlooked, in the "new China" as in the "old", marriage patterns establish lines of influence and obligation that can be decisive in certain bargaining settings (we need to know a great deal more about this).

The Implications Of A Bargaining System View

What are some of the practical consequences of bargaining for both the process of making policy and the outputs and outcomes of that process? I believe that five consequences are of particular importance:

Decisions generally are slow in coming; the process of consensus building and negotiation is protracted. The more geographic areas and functional systems that must be involved in the process the more laborious will be the process of negotiation.

It is difficult to definitively say when a decision really has been made. Frequently, decisions are made "in principle," with nettlesome details left for future resolution. The requisite resolution of these details may never occur. Decisions concerning the building of nuclear power plants and the Three Gorges Project are excellent examples of this phenomenon.^[58] There is an indeterminacy to outcomes. Issues seem to rise like Lazarus on the agenda; they never stay buried. This process is the political equivalent of protracted guerrilla warfare.

Even once a policy is formulated and adopted, the implementation process is characterized by negotiation among and between levels of the hierarchy, sometimes all the way down to the grass roots. Each level slightly deflects policy in a direction favorable to its interests; by the time one has moved through six, seven, or more layers of the system, the cumulative distortion (not to mention bureaucratic constipation) can be great. Almost invariably, unanticipated and unwelcome consequences are part of the implementation process from the Center's perspective.

One of the biggest mistakes that the Center can make is to set too many high-priority goals simultaneously. For both formulation and implementation to be effective, the elite must be united on the objective and willing to expend considerable political and economic resources as a seemingly endless bargaining process unfolds. There

is no substitute for elite attention and the focused use of resources in the bargaining process. However, the ability to focus elite attention and resources is often diluted by divergent priorities among elite members, their different support bases, and the shared desire of the entire leadership to produce rapid change. This shared desire to produce rapid change almost assures that the elite will bite off more than it can chew.

Because bargaining is extensive, the legal framework is poorly developed, and social norms and system legitimacy suffered egregious harm in the Cultural Revolution era, it is exceedingly difficult to separate legitimate and necessary bargaining activity from corruption. It is essential that the system create a legal framework and affirm widely shared values, procedures, and norms to govern this activity. The race is between the process of establishing these norms and the loss of system legitimacy. This may be the most important race in which Beijing's leaders are running.