Three
The Chinese Political System and the Political Strategy of Economic Reform

Susan L. Shirk

The greatest challenge of China's economic reform is the political one. The national leaders who seek to improve the functioning of the economy by introducing market reforms must formulate and implement these reforms through the Communist political system. Their political strategy of economic reform, if it is to succeed, must reflect the actual power relationships operating in the political realm. The political system is not static, however; economic reforms in communist states usually are accompanied by some reform of the communist political system. Changes in the political institutions and rules of the game modify the context in which bargaining over economic reform policies occurs. From the standpoint of communist political elites, political reforms, such as modification of the relationship between the Communist Party and the government, are instruments for furthering economic reforms, important elements of their political strategy of economic reform.

Political bargaining over policy proposals for economic reform is very intense. A transformation of the economic structure involves redistributing authority and rewards among sectors, bureaucratic agencies, and regions. Those who benefit from market reforms naturally support them and strive to get the best possible deal under the new rules. The groups who were favored and protected by the old command economy and who feel threatened by changes in the economic system resist the reforms or fight to retain as much of their original privileges as they can. As Vice-Premier Tian Jiyun observed in a 1986 speech: "The overall reform of the economic structure is, in a sense, a readjustment of power and interest, in which a large amount of contradictions exist. Among them, there are contradictions between the central authorities and the localities; between the state, the collective, and the individual; between

one department and another; between one locality and another; between departments and localities; and so on and so forth" (Tian 1986).

Theoretically, these changes in the economic system, by increasing efficiency, should benefit everyone. But as economic theorists have observed, the redistributive effects of changes in the rules of the economic game are bound to create group conflict (Pratt and Zeckhauser 1985). Only if there is an institutional framework for resolving these conflicts through bargaining will economic restructuring occur.

In this chapter I examine the formal institutional relationships and rules of the game governing national economic policy-making as they have evolved during the 1980s in China. I argue that the political strategy of economic reform devised by Deng Xiaoping and Zhao Ziyang has used effectively the institutions of the Chinese political system to bargain with various groups and build support for reforms; but by compromising and postponing the most contentious issues, the strategy created new economic problems, which put political obstacles in the path of reform.

The Relationship Between The Communist Party And The Government: Principal And Agent

The relationship between the Communist Party and the government is at the core of any Leninist political system. The Party is "the organized expression of the will of society" (Schurmann 1968, 110). It leads the work of the government (usually called by Chinese the "state," guojia ), but remains distinct from the government.

Although the Chinese Communist Party (CCP) is intertwined with the state and with the society, it is a distinct political institution.^[1] The CCP is an elite membership organization with 47.75 million members, less than 5 percent of China's total population (Xinhua , 17 June 1988). Most Party members are employed as workers, farmers, professionals, and so forth and only participate in Party activities part-time. Yet they are required to obey the direction of Party committees within their work unit and at higher levels. When I talk about the "Party" I am referring to it as a formal bureaucratic institution with over 1 million professional, full-time staff called Party cadres and a permanent structure at all levels of the political system. At the apex of the CCP is the national Party organization headed by the Standing Committee of the Politburo, the Politburo, and

[1] This view of the Communist Party as an institution distinct from the government challenges Lieberthal and Oksenberg's perspective of a merged institutional structure (1988). While I agree with Lieberthal and Oksenberg that there is a unified chain of command at the apex of the Chinese political system, I see this authority structure as a manifestation of the delegation relationship between Party and government, not as evidence of the lack of boundaries between Party and government.

the Central Committee, which are served by a Secretariat and other departments. At provincial, municipal, and county levels there are similar leadership bodies (CCP committees) and staff offices. Every agency of the national, provincial, and county governments has within it a Party committee and, until recently, a Party leading group. Although some members of these Party oversight bodies play two roles, as both government and Party officials, these bodies are clearly defined as part of the Communist Party.

The language of Western institutional economics is useful for conceptualizing the formal authority relationship between Communist Party and government: It is an "agency relationship" in which the Communist Party is the "principal" and the government is the "agent." The Party has formal political authority over the government, which does the actual work of administering the country. The Party's authority over the government is based primarily on its authority to appoint and promote government officials (nomenklatura ) (Burns 1987).^[2] The Party also sets the general policy line (luxian ), which the government implements, and oversees the work of the government. Finally, it is responsible for the ideological remolding of government cadres and of all other members of society.^[3]

The relationship is analogous to the relationship between the ruling party and the government bureaucracy in a democratic system. (In a parliamentary democracy, the majority party in Parliament appoints the cabinet and directs the work of the bureaucracy. In a presidential democracy, two principals, the president and Congress, lead the bureaucracy.)^[4] The Party politicians oversee the work of the government bureaucracy. The bureaucrats may appear to work autonomously, because the politicians intervene infrequently. But the bureaucrats know that they cannot stray too far from the preferences of their principals. If they do, they will be publicly criticized and fired by their political masters, a lesson that Ann Burford (head of the U.S. Environmental Protection Agency) and Yang Zhong (PRC Minister of Forests) learned the hard way.

The crucial difference between communist and democratic systems is

[2] According to the Constitution, top positions in the state bureaucracy are filled by the National People's Congress, but, as Yan Jiaqi says, "the political reality is that the candidates nominated or recommended by the CCP are always elected" (Yan Jiaqi 1988).

[3] "In authoritarian systems without active legislatures, the principal form of control over the bureaucracy has been the mass party, particularly its Leninist variant. The Leninist party exercises control over the bureaucracy by managing appointments and promotions, monopolizing leading posts, supervising the indoctrination of state officials, disciplining Party members who hold government positions, and setting the bureaucracy's principal policy guidelines" (Harding 1981).

[4] For an interesting comparison of Party-government relations in democratic and socialist systems, see Yan Jiaqi (1988).

the political accountability of the principals. In communist systems the Communist Party is not formally accountable to the citizenry; it claims to reflect the will of the people by leading them toward a communist future, but there is no institutional mechanism making them accountable to citizens' present preferences. In a democracy, the politicans are elected and therefore are the agents of their constituents. If they enact policies or allow their bureaucratic agents to take actions that violate their constituents' perceived interests, they will be voted out at the next election.

Although the Communist Party has the ultimate authority in a communist polity, it cannot administer the country on its own. Like any principal in a large organization, it has limited information. The only way for a principal to solve this problem of limited information is to delegate authority to agents. The Communist Party allows the government bureaucracy to make and implement policies because the bureaucrats have better information than Party leaders can possibly have. The bureaucrats have specialized information, while the Party leaders must know about everything; the bureaucrats are close to the problem, whereas the Party leaders are remote.

Once a principal has delegated authority to its agents, the problem of control arises. How does the Communist Party know whether government bureaucrats are carrying out policies that conform with the Party's preferences? Bureaucratic agents naturally distort the information they pass up to their political masters, in order to place themselves in a good light. In democracies, elected politicians manage the agency problems of "hidden action" and "hidden information" through various mechanisms (Arrow 1985). They require regular submission of reports, hold public hearings, or enfranchise interest groups to scrutinize administrative actions. By empowering concerned groups to oversee bureaucratic agencies, they reduce the costs of constant monitoring while guaranteeing that their influential constituents are satisfied. As elected politicians, they do not need to know about all actions of appointed officials, only those actions that might displease their constituents. As Mathew McCubbins and Tom Schwartz put it, democratic politicians prefer a "fire alarm" approach to the control of bureaucratic agents to the much more costly "police patrol" approach (McCubbins and Schwartz 1984). In communist political systems, however, the Communist Party is reluctant to enfranchise constituent groups to oversee government operations; the Party's political monopoly depends on the political demobilization of the population. Without the help of citizen groups, Party leaders have a particularly difficult time acquiring information about government actions. Despite their two-thousand-year tradition of bureaucratic authoritarianism, the

Chinese are no better than any other communist regime at monitoring bureaucratic behavior.^[5]

Faced with this structural problem (and the immediate practical problem of controlling bureaucrats who had served the pre-Communist regime), Soviet Communist Party leaders in 1919 developed a method that we now call "parallel rule." The Chinese system of Party control over the government is essentially identical to the Soviet system on which it was modeled. Communist Party members are appointed to the top positions in government agencies, and in each agency all Party members are organized under a Party committee (subordinate to the Party committee at the next level). The hierarchy of government organs is overlaid by a parallel hierarchy of Party committees that enables Party leaders to supervise Party members in the government and lead the work of the government.

The Communist parties in both the Soviet Union and China established specialized functional departments at central and provincial levels to oversee the work of government economic agencies. Staffed by specialists, these Party organs for agriculture, industry, finance, and so forth gradually took over the policy-making functions from the government, especially at the provincial level.^[6] The state administrative apparatus became redundant, a "duplicate administrative structure," as Lin Biao described it in 1969 (Harding 1981, 284). The Communist Party became the locus of bargaining over economic policies. A 1953 regulation, "The Decision on the Central Authorities' Leadership Over the Work of the Government," formalized this fact: "All the major and important general and specific government policies, and all major questions concerning the government's work, must first be submitted to the central authorities for examination and approval. And only after the relevant discussions are carried out by the central authorities, and the relevant decisions or approval given by the central authorities, can major and important general and specific government policies begin to be implemented" (Zhou 1987).

[5] This problem is not remedied by letters to the editor, denunciations to higher-level officials, open meetings with government officials, or any of the other practices vaunted by the Chinese press as signs of the new "democracy" in China. Public opinion polls are a partial remedy to the Party's information problem (and from the perspective of an authoritarian elite are preferable to elections because they provide information without imposing accountability), but they do not provide sufficiently specific information on the actions of particular agencies.

[6] "In the past all levels in the localities set up certain organs as counterparts to the government, and also assigned some industrial secretaries, agricultural secretaries, standing committee members in charge of finance and trade or culture and education, and so on. They monopolized government work to a greater extent than at the central level, hence the amount of work involved in separating party from government there is greater than at the central level" (Zhao 1987b).

The Chinese Communist Party leaders, for reasons not yet entirely clear, took the method of parallel rule even further than the Soviets. As former Premier Zhao Ziyang said, "China is one of the socialist countries most seriously afflicted by lack of separation of party and government" (Zhao 1987c).^[7] The Chinese established in all government agencies Party "leading groups" or "fractions" (dang zu ) responsible for actually administering the work of the agency. Although these institutions originated in the Soviet Union, they came to play a more significant role in China.^[8] The Party leading group within a government agency is much more powerful than the Party committee (author's interviews). The Party committee only supervises the lives and thoughts of the Party members within that agency. But the Party leading group has authority over the non-Party bureaucrats as well as those who are Party members. The Party leading group plays a decision-making role, setting policy for the entire sector (xitong ), not just for the agency.^[9] For example, the Party leading group in the Ministry of Metallurgy leads the work, not only of the ministry, but also of subordinate provincial and municipal bureaus and even of steel mills run by the ministry and bureaus. It has the authority of appointment, removal, and transfer of officials for the entire sector (Burns 1987). The Party leading group is much smaller than a Party committee, consisting only of three to five people (usually the Party secretary, the minister, and several vice-ministers). The Party leading group essentially appropriated the authority of the official heads of the government agency, creating a confusing system of dual administrative leadership.^[10]

At the enterprise level, the Chinese Communist Party also went beyond the Soviet Party in playing an administrative role (Ma 1987; Zhao 1987c). Soviet Party committees have always limited their role to supervision of Party members in the enterprise while allowing the manager full responsibility for production. The Chinese, since the Eighth Party Congress (1956), have put the manager under the leadership of the Party

[7] Franz Schurmann (1968) made the same observation twenty years ago.

[8] Schurmann (1968) described the function of the Party fraction as more limited than that of the Party committee. Basing his analysis only on Party rules and handbooks, he said that the fraction only supervised Party members within the agency, while the committee had decision-making powers within the agency (160). Barnett (1967), whose findings were based on interview data, described the Party fraction as "the real center of power and authority in the ministry, more important even than the Party committee" (24).

[9] The Guangdong Party handbook quoted in Schurmann (1968, 160) implied the broader authority of the leading group when it stated, "Party fractions, in cases where it is necessary to carry out their tasks, may direct the work of the basic-level Party organizations in the unit concerned."

[10] "In the government setup, the CCP bypasses the government executives at all levels by leading all government departments directly through the 'leading party groups' in government institutions" (Yan 1988).

committee. As Zhao Ziyang himself said in 1987, the issue of Party control over factory management became a "yardstick for supporting or opposing party leadership. ... Every time we undertook a campaign, this setup was strengthened, to the extent that the Party committees monopolized many administrative matters" (Zhao 1987c).

Under this system the Chinese Communist Party not only controlled the government with a tight, constant "police patrol," it actually substituted itself for the government. The organizational lines between the Party and government blurred, and the delegation relationship disappeared almost entirely. For example, the head of the CCP committee overseeing all the Party leading groups and committees within the agencies of the central government sat (and still sits) in the government's State Council (author's interview). The Finance and Economics Leadership Small Group, five to seven top leaders who make most of the important economic policy decisions, although formally a Party unit, was located in the government section (Northern) of Zhongnanhai when Zhao Ziyang, as premier, was in charge of economic policy; when Zhao was transferred to Party secretary, the Finance and Economics Leading Group moved to the Party section (Southern) of Zhongnanhai (author's interview).^[11] Carol Hamrin in chapter 4 in this volume describes the extent to which Party and government functions merged in all of the leadership groups that make decisions in specialized policy domains.

Such tight control could only be achieved at a tremendous price. The informational advantages of bureaucratic delegation, that is, expertise and specialization, were lost. Poorly educated veteran Party officials made policy decisions on the basis of political instincts rather than technical knowledge. Party members serving in government agencies were promoted more for political loyalty than for professional accomplishment (Harding 1981).

No longer willing to pay the price of poor-quality decisions and inefficiency, the Party leaders moved to transform the relationship of Party and government in the 1980s. The Party delegated more responsibility to the government bureaucracy, especially in economic policy-making. The Standing Committee of the State Council, meeting twice a week, took charge of the economy (although it still received recommendations from the Finance and Economics Leadership Small Group). The Politburo, having earlier abolished its specialized economic units, limited itself to setting the overall political line of economic reform and ratifying

[11] Lieberthal and Oksenberg (1988) cite as evidence of organizational integration the fact that top Party and government organs are both housed in the Zhongnanhai compound. I cite as evidence of organizational boundaries the fact that within Zhongnanhai, Party organs are grouped together at the southern end and government organs are clustered at the northern.

important economic policies made by the government. At the provincial level, specialized Party departments overlapping their counterpart government departments were abolished (Zhao 1987b). In enterprises, the Party's role was reduced, and administrative responsibility was restored to managers. The CCP Constitution was revised to gradually eliminate the Party leading groups within government agencies (Renmin Ribao , 2 November 1987).^[12] Civil service reforms were proposed to establish a dual structure within the bureaucracy, a cadre of professional civil servants, selected by meritocratic examination and promoted on professional criteria, alongside the administrative officials appointed and promoted by the Party organization departments (Zhao 1987b; Yan 1988; Xinhua , 18 November 1987; Burns 1988).

The expanded discretion of the government bureaucracy is illustrated by a recent example of conflict over agenda setting in the State Council. (Agenda setting is a prerogative of the bureaucracy typical of most political systems.) In February 1988, Bo Yibo, vice-chairman of the CCP Central Advisory Commission, came to a regular State Council meeting with a commission proposal on transportation safety, probably designed to embarrass the government bureaucracy for recent accidents. Wan Li, the vice-premier chairing the State Council meeting (the premier, Li Peng, had left to participate in a Politburo Standing Committee meeting) refused to alter the agenda to allow discussion of the issue. According to the unofficial report of the incident, the State Council leaders had learned in advance about Bo's intention and had obtained Deng Xiaoping's support for sticking to the original agenda. Bureaucratic authority has widened, but at this stage it still is fragile and dependent on the support of the preeminent leader, Deng Xiaoping (Lo 1988).

The widening of government's discretionary authority did not destroy the principal-agent relationship between Party and government. Zhao Ziyang emphasized that the CCP Central Committee should retain its leadership over the government "in political principles and orientation and in major policy decisions," and continue to appoint leading cadres for central state organs; and that the provincial Party committees also should retain political leadership and personnel appointment powers.

[12] Zhao Ziyang in his work report to the Thirteenth CCP Congress said, "Party leading groups in government departments should gradually be abolished, since the practice of making these groups responsible to the Party committees of the next higher level which have approved their establishment is not conducive to unity and efficiency in government work" (Zhao 1987b). At a Party meeting on Party work in central state organs in March 1988, Politburo Standing Committee member and premier, Li Peng, explained that while Party leading groups in government ministries and commissions would be abolished "step by step," Party committees would be retained and become "even more important" (Li 1988).

The 1984 reform of the nomenklatura system decentralized the authority to appoint, remove, and promote government personnel to lower-level Party organizations but still left it in Party hands. Chinese leaders promise that in the future the Party will supervise the work of the government but not substitute itself for the government. Separation of Party and government will actually strengthen Party leadership of the government. "Leaders must keep very cool; they must stand high and see far, consider things carefully, and avoid getting entangled in a pile of routine affairs. They cannot truly play a leading role if they are entangled in trivia all day long" (Zhao 1987c). Another commentator pointed out that while state organs "can choose either to accept or not accept the Party's policies," if a state organ refuses to follow the Party's policies, which are "absolutely correct," "then Party organizations and Party members within the state organs should supervise and ensure the correct implementation of the Party lines and the general and specific policies, by giving play to the exemplary role of the Party members within the state organs" (Zhou 1987). In other words, if the government does not agree with the preferences of the Party, then the Party should use its authority over Party members in the bureaucracy to impose these preferences on the government.

The Party's control over the government has made the economic reform drive possible. Party leaders have the power to propose new directions for reform and prod the bureaucracy to action. The government bureaucracy would never have taken such bold initiatives on its own. Deng Xiaoping and Zhao Ziyang were able to dominate a conservative minority within the CCP to promulgate a Party line of economic reform. By making support for economic reform the current political-ideological line, the Party has made it impossible for anyone, inside or outside the government, to publicly oppose the reform drive (although people's definitions of reform vary widely). And by replacing thousands of incompetent or conservative government officials at central and local levels, the Party has empowered a new cohort of officials, eager and able to promote economic reforms.

The delegation of greater discretion to the government bureaucracy also offered substantial advantages to a Party leadership intent on economic reform. Enhancing the independence of the bureaucracy gives government officials a greater incentive to be efficient and to take actions that are economically rational but politically risky. As Tang Tsou has observed, modernization requires that the Party grant all professionals, including government bureaucrats, greater autonomy (Tsou 1983). By shifting the locus of economic policy-making from the Party to the government, reformers made it more difficult for Party conservatives

(who until recently retained their power in the Politburo) to sabotage the reform drive.^[13] And by delegating policy-making to government commissions and ministries, they devised a reform package acceptable to key economic groups.

On the other hand, Party conservatives were obviously unhappy with moves to restrict the Party's powers. Party resistance to limiting its role was even stronger at the provincial and local levels, where Party officials continued to meddle in economic affairs. Some of the formal organizational reforms, such as eliminating Party leading groups within ministries or closing down specialized Party departments in provinces, were implemented only halfheartedly and were reversed by the conservatives after 1989. Nonetheless, the locus of bargaining over national economic-reform policies was moved over to the government side, where it remained even after 1989.

The Policy-Making Process: Management By Exception

Under the revised delegation relationship between the Chinese Communist Party and the government, the locus of bargaining over economic policy has shifted to the government arena. The Communist Party sets the general line for the economic reforms, but it delegates to government the formulation and implementation of specific policies. The Standing Committee of the State Council has become the main policy-making arena, while the Standing Committee of the Communist Party Politburo is much less active.^[14] The State Council is advised by a network of recently established policy research centers, described by Nina Halpern in chapter 5.

The Chinese government bureaucracy makes policy according to the decision rules characteristic of "management by exception" (Lawler 1976). At each level of the organizational hierarchy, agency representatives make decisions by a rule of consensus. If they all agree, the decision is automatically ratified by the higher level. If the bureaucrats cannot reach consensus, then the decision is referred to the higher levels, and if the higher levels cannot agree, then either nothing happens or the ultimate principal, the Communist Party, intervenes to impose a solution.

The advantages of management-by-exception, from the standpoint of the principals (the leaders of the government and the Communist Party) are that (1) it exploits the superior information of subordinates; (2) it

[13] Ironically, it was China's proto-legislature, the National People's Congress, under the leadership of Party conservative Peng Zhen, that was responsible for sabotaging some of the government's reform proposals, such as the bankruptcy law.

[14] According to Barnett (1985), the same shift from Party to government can be observed in the making of foreign policy.

relieves the principals of the costs of constant intervention in the policy process; (3) it gives all the agents who will implement a policy a voice in the formulation of that policy; and (4) it gives agents an incentive to resolve their differences and come to agreement.

Most corporations and many political systems delegate policy-making to subordinate agencies and use management-by-exception as the most efficient way to manage this delegation relationship. Parliamentary systems like Japan or France may appear to be dominated by their bureaucracies, but they are actually ruled by parliamentary party majorities who have effectively used management-by-exception to allow their bureaucratic agents to work out policy packages. If the politician principals seem invisible and rarely intervene to overrule a bureaucratic decision, it is because they have structured the bargaining game to produce outcomes satisfactory to them (and to their most influential constituencies).^[15]

The structure of the government bureaucracy reflects the Communist Party's notion of which groups should be represented in policy deliberations. Establishing a particular set of bureaucratic agencies, and organizing collective choice so that individuals represent the preferences of their agencies, enfranchises certain groups but not others (it is striking, for example, that the national labor organization is represented in key economic meetings in Hungary but not in China [Comisso, personal communication, 1984]). Just as a ruling party in a democracy structures decision-making processes so that its most important constituents are satisfied, the leaders of the Chinese Communist Party structure decision-making processes so that their most important constituents are satisfied.

The Chinese bureaucracy is organized by function (education, culture, public security) and by economic sector (agriculture, coal, machinery). Ministries are expected to articulate the interests of their particular sector. When ministers or vice-ministers are called together to discuss a policy proposal, they are expected to represent the perspective of their

[15] Management-by-exception is not the prevailing mode of decision making in the American political system. Bureaucratic agencies make their recommendations directly to the president or Congress, and the locus of bargaining is the congressional committee, not the bureaucracy. One possible reason for this difference between presidential and parliamentary systems is the separation of powers characteristic of presidential systems. Congress and the president, as dual principals, share authority over the bureaucracy; Congress is reluctant to shift the decision-making process to the bureaucracy for fear that the president might strengthen its influence over that process. In parliamentary systems, and in the Chinese communist system at the present time, the ruling party has no rivals for control of the bureaucracy and therefore can delegate to it decision making using management-by-exception. We may speculate that if the National People's Congress becomes a genuine legislature and exercises its authority over the bureaucracy, the Party might retract its generous grant of decision-making authority from the bureaucracy (unless the Party is able to completely dominate the NPC like a ruling parliamentary party in a democracy). These speculations are based on my conversations with Mathew McCubbins.

particular ministries. Press articles criticizing "selfish departmentalism" (e.g., Renmin Ribao , 21 January 1983) seem futile and naive when we recognize that Chinese government institutions are structured to encourage expressions of departmental points of view.^[16]

At higher levels, the commissions (State Planning Commission, State Economic Commission, State Science and Technology Commission) and the State Council promote the aggregation of departmental interests, but they also reinforce the articulation of sectorial interests. Within the State Economic Commission and the State Planning Commission bureaus are divided by economic sector, so that the head of the energy bureau, for example, argues for resource allocations to the coal, petroleum, electricity, and nuclear power industries. Within the State Council the degree of aggregation is greater. Each vice-premier and state councillor is assigned responsibility for a sector or a function, such as agriculture, industry, or finance. While such divisions of responsibility encourage specialized expertise in policy-making, they also guarantee that sector-based bargaining continues right up to the top of the government hierarchy (Chen 1987).^[17] The comprehensive agencies have their own organizational viewpoints as well: The State Planning Commission represents the macroeconomy, the State Economic Commission represents the enterprises, and the Ministry of Finance represents the central state (author's interviews).

When the Chinese CCP leaders set up their national economic bureaucracy in 1953, they gave industry a stronger voice than agriculture. There was only one ministry representing agriculture, but over ten ministries representing various industries. Because the fiscal system was set up to obtain revenue almost entirely from the earnings of state factories, central-government officials had a financial interest in keeping agricultural prices low and industrial profits high. This bias toward industry reflected the policy of Soviet-style industrialization, which prevailed at the time. Once institutionalized in the government structure, the bias

[16] Ellen Comisso (1986) argues that in communist regimes, the control of the Communist Party over bureaucratic careers makes it impossible for managers and government bureaucrats to articulate the particularistic interests of their units. Comisso has identified the key principal-agent relationship in communist regimes, but fails to recognize the advantages to the principal of delegation and management-by-exception. The Communist Party expects officials from subordinate government agencies and enterprises to articulate particularistic interests, and benefits from the bargaining among these interests.

[17] Vice-heads of provinces, cities, and counties are also assigned sectorial jurisdictions and are expected to bargain to get more benefits for their sector. One country cited as an example in the press had one head and thirteen deputy heads: "At the official work meetings they attended the deputy county heads asked the county head for power, material resources, and manpower in the interest of the parts they represented, thus bothering the county head instead of helping him carry out his work. For this reason, it is not good to have too many deputies" (Xinhua , 5 July 1987).

was perpetuated, as the continuing underinvestment in agriculture illustrates (Lardy 1983).

The bureaucratic reorganizations that have been a frequent occurrence since the 1950s are efforts to change the structure of interest articulation and aggregation, as well as to improve efficiency. Whenever ministries are merged or divided, raised to commission level or demoted to bureau level, the voices of various sectors are strengthened or weakened. The elevation of education from ministry to commission status in 1987, for example, was designed to give education more clout in the contest for government resources, in addition to enhancing coordination of educational activities under different ministries. Lynn Paine, in chapter 7 in this volume, describes the disadvantages the education sector suffers because of its organizational weakness.

Management-by-exception requires a bureaucratic hierarchy with formal equality among the units at any one level. If one ministry could impose its preferences on the entire group, then the advantages of management-by-exception would be lost. The Chinese government, by putting all ministers (as well as all bureau heads, all vice-ministers, etc.) at equal rank, satisfies this requirement. (An important exception is the Ministry of Finance, which by virtue of its revenue function, which is critically important to the survival of the state, ranks at the commission level [Lieberthal and Oksenberg 1988]). Although each ministry's actual influence varies with the prestige of its minister (and State Council overseer), its function, its control over subordinate enterprises, its financial contribution, and so forth (Oksenberg 1982; Lieberthal and Oksenberg 1988), the formal status of each ministry is the same. It is also crucial that the ministries are subordinate to organizations at the next level of the hierarchy. The State Planning Commission and the State Economic Commission (and to a lesser extent the State Science and Technology Commission) have formal leadership relations (Lieberthal and Oksenberg 1988) over the ministries and can resolve issues that ministries have been unable to agree on. At the apex of the government sits the State Council, which has leadership authority over all its subordinate agencies and can impose a solution when neither the ministries nor the commissions can find one.

The obvious fact that bureaucracies are organized by sector and function, not by geographic region as legislatures are, is worth emphasizing. Regional concerns are brought into government policy-making in a number of ways. The State Economic Commission and the State Planning Commission have regional as well as sectorial bureaus that voice the demands of the Northeast, the Southwest, and so forth. And whenever economic policy proposals are being debated, the heads of the relevant provincial (and some municipal) bureaus are invited to participate; pro-

vincial governors or vice-governors attend the most important policy work conferences. Each province sends a delegation to the annual national planning and budgeting meetings to lobby in its own behalf. Provinces are equal in rank to ministries, and disputes between provinces and ministries often filter up to the commission or State Council level. Yet, provinces do not have permanent formal representation in the bureaucratic arena where most economic policy-making now occurs.

The lack of formal regional representation in government policy-making is an institutional anomaly if we consider that since 1957, and especially since 1980, the provinces have received an increasingly large share of resources and authority. Moreover, the two most important constituencies for top Communist Party leaders currently are the central-government bureaucracy and the provincial officials. The institutional "payoffs" to the first group are quite clear, that is, extending bureaucratic autonomy and expanding the size of the bureaucracy. Yet the resources and authority of the provinces have not been recognized by institutional changes. (Provincial demands for an institutionalized voice at the Center could result in a strengthening of the national legislature, the National People's Congress. Or provincial officials might advocate shifting economic policy-making from the State Council, where they are not represented, back to the Politburo, where they currently are the largest bloc of votes.)

The Chinese government has been making economic policies according to management-by-exception at least since the early 1980s.^[18] Research by Lieberthal and Oksenberg (1988) and by Lampton (1987) found that consensus was the rule governing economic decision making in the central government, and that any participant could veto proposals.^[19] In contrast to a legislative setting in which majorities rule, minorities have the power

[18] A serious deficiency in our understanding of Chinese politics is the lack of a "base line" with which to compare current political practices. The analysis here, as well as previous studies of Chinese policy-making based on documents (e.g., Teiwes 1979, 1984; MacFarquhar 1974, 1983; Oksenberg 1971), suggests that during Mao Zedong's reign the locus of economic policy-making was the Communist Party instead of the government. Little discretion was delegated to government institutions. It would be valuable to confirm this conclusion with information from retrospective interviewing in China.

[19] A consensus-decision rule does not necessarily require unanimity. A bureaucrat from a weak agency may not have an effective veto. If eveyone knows that an agency is weak and that its position will not be sustained by the higher level, it may be unable to block a measure or force a compromise, whereas a bureaucrat from a strong agency could succeed in doing so. The writings of Lampton (1987), Oksenberg (1982), Lieberthal and Oksenberg (1988), and Shirk (1989) have suggested that agencies that generate revenues, have control over a large number of enterprises, and have ties to top Party leaders have more clout than agencies without these resources, but more case studies of policy-making are needed before we can speak with any certainty about the differential ability of agencies to veto measures under management-by-exception in China.

to obstruct action in bureaucracies run by management-by-exception. Whenever an agency refuses to compromise and consensus cannot be reached, the issue is "tabled" or is referred to a higher level for resolution. As the secretary-general of the State Council complained, "many problems remain unsolved for a long time simply because of the objection from a minority. ... Too many people exercise veto in our public organs" (Chen 1987).

Much of the time of the government bodies above the ministries is taken up by what the Chinese call "coordination" (xietiao ) work, which means resolving conflicts among subordinate units. The State Economic Commission (SEC) and (to a lesser extent) the State Planning Commission are supposed to provide final resolution to interagency (as well as interregional and agency-region) disputes.^[20] But because the internal organization of these bodies reinforces sectorial and regional divisions, disputes often are passed up for resolution at an even higher level. The Standing Committee of the State Council spends much of its twice-weekly meetings resolving interministerial (or interprovincial or ministerial-provincial) differences over relatively minor issues (author's interviews). A temporary institution called the "Adjustment Office" (tiaojie bangongshi ) was established under the premier during the mid-1980s (Lieberthal and Oksenberg 1988) to settle conflicts that were obstructing the progress of reforms and take some of the load off the Standing Committee of the State Council.

Bargaining among bureaucratic agencies is structured by management-by-exception. Every agency representative must decide whether to sign on to a lower-level decision that does not entirely satisfy its preferences, or hold out and force the intervention of the higher levels. To make this decision the agency representative must seek information about the preferences of the higher-level bureaucrats to anticipate what their decision would be. Unless the agency representative has reason to believe that the higher levels would make a decision more favorable to their agency's interests than the lower-level decision, the representative will compromise to reach agreement at the lower level. After all, a bureaucrat or an agency that consistently refuses to compromise will certainly not be popular with upper-level leaders. Agency officials are the

[20] The Economic Reform Bureau of the SEC spent most of its time resolving disputes among subordinate units sparked by changes in the economic structure, such as disputes between ministries and Wuhan over control of factories after Wuhan became an independent planning entity (author's interview). After the SEC and the SPC were merged in June 1988, the State Council was so overloaded with disputes to resolve that in 1990 they set up a new commission responsible for xietiao work, the Production Commission, consisting of the former SEC Production dispatch Bureau (shengchan diaodu zu ) and Enterprise Management Bureau (qiye guanli zu ), to relieve their burden (author's interviews).

employees of the State Council and are appointed by the Communist Party. A reputation for rigidity may be punished by demotion or other measures.^[21] Therefore, in most circumstances an agency will settle for a less-than-optimal decision at the lower level instead of gambling on a better decision at a higher level. From the standpoint of the principal, management-by-exception works well when it allows key groups to articulate their interests but creates incentives for them to compromise their differences without forcing the principal to intervene.

By this standard, Chinese management-by-exception has not worked well during the late 1980s. From the perspective of the top leaders of the government (articulated by Chen Junsheng, then secretary-general of the State Council), the work of the government has been impeded by constant arguing and "the escalation of coordination." Problems that once were solved at lower levels are now being pushed up to the State Council (Chen 1987). There are several reasons why the management-by-exception system is working poorly. First, divisions within the Party leadership encourage intransigence among subordinates. Conflicts among principals increase uncertainty for agents and lead them to gamble on the higher-level resolution of the conflict. Even without specific information (usually acquired through factional relationships) that someone at the top will support their position, subordinates expect more favorable treatment when there is leadership conflict. Central and provincial officials currently are the two major constituencies within the CCP Central Committee.^[22] Top Party leaders compete for power by building up support among these two key groups. They welcome opportunities to appeal to particular officials by pressing their interests when policies are decided. Yet even while top leaders are glad to have opportunities to build support by helping ministers or provincial officials, they complain that incessant bureaucratic wrangling has overloaded the system.

A second reason for the growing burden of coordination work is that participation in policy deliberations has been expanded to include representatives of more groups. Each policy is debated in a series of meetings to which a large number of affected units, both local and central, as well

[21] There is no evidence that uncooperative behavior is punished by cutting an agency's budget, which often is the case in democracies run by management-by-exception.

[22] In this sense the relationship between Party and government in China is not a pure agency relationship. There is reciprocity between principal and agents in all agency relationships; the principal generally believes that the organization will run more smoothly if the agents are kept satisfied. But Party leaders in China are politically dependent on government officials to an even greater degree because these officials belong to the CCP and participate in choosing the leaders of the CCP. Top provincial and ministry officials are constituents as well as agents of the Party leaders. This fact may help explain why Communist parties, even when they use tight "police patrol" control of the government, seem to do such a poor job of oversight.

as technical experts, are invited. The composition of these meetings is not set by any rules (in contrast to policy deliberations in Congress or other democratic legislatures), allowing the top government leaders who are orchestrating the process a high degree of flexibility. The widening of consultation is designed to prevent overconcentration of power in the hands of a few individuals in the Party. Deng Xiaoping, in his important 1980 speech "On the Reform of the System of Party and State Leadership," argued that it was this overconcentration of power that had produced the Cultural Revolution and other excesses of Mao Zedong's rule (Deng 1980). By expanding participation in the policy process, Deng and other top Party leaders hope to improve the quality of decisions and build support for them among key constituent groups. But inviting more groups to sit at the bargaining table also complicates the process of building consensus. Instead of a small, stable group of participants who are willing to trade votes on one issue because they trust others to pay them back on the next issue, there is a larger, ad hoc group that finds agreement much more problematical.

Finally, economic reform, by creating new financial interests among bureaucratic and regional agencies, has made bargaining among them more intense. Policies designed to improve economic incentives by allowing ministries, provinces, local industrial bureaus, and so forth to retain a share of their renminbi (Chinese currency) and hard-currency earnings have made bureaucratic agencies more profit-conscious. They fight more fiercely to obtain preferred policies because the stakes are higher. For example, the dispute about which ministry would take charge of the production of refrigerators and washing machines, which took five years to resolve in the era of economic reform (Renmin Ribao , 17 July 1984), would have been settled more quickly and easily under the old system. The creation of new financial incentives sometimes makes it easier to work out a compromise by giving all affected parties a share in the deal. There are many examples, especially at the local level, of potentially profitable joint ventures and other projects that were approved only because every agency with approval authority was given a percentage of the profits (Lieberthal and Oksenberg 1988). But, on the other hand, this new mode of resolving differences by dividing the profits encourages agencies to press their demands in the bargaining process. They recognize that assertiveness can pay off in several percentage points of a profitable venture. As the secretary-general of the State Council notes, "A profitable undertaking invariably draws the intervention of many departments, with every department demanding a slice of the cake, and none will make any concession. Sometimes this ends in a confrontation" (Chen 1987).

Management-by-exception tends to bog down in the Chinese govern-

ment bureaucracy because the bureaucracy is so poorly institutionalized. There are few rules about who participates in which decisions. The sequence of decision gates is ambiguous in all policies except those related to the planning and budgeting cycles. Policy decisions are not permanent, because a leader who does not like the outcome of one work meeting can call for another with a different set of participants.^[23] When Chinese officials talk, they often differentiate between "hard" and "soft" decisions (author's interviews). Moreover, all the conflicts among agencies, among provinces, and between agencies and provinces, must be resolved by higher-level bureaucratic organs. There is no administrative law or independent judiciary to settle jurisdictional disputes among them. The absence of judicial institutions puts the entire load of conflict resolution on the central government and the Party.

Policy Outcomes Under Management-By-Exception

The policy outcomes produced by the Chinese system reflect the structure of the government bureaucracy and the process of management-by-exception. Scholars of communist political systems have noted that policies in such systems tend to be incremental (Hough 1977; Bunce 1981). Drastic changes in policy direction or massive shifts in the allocation of resources are inhibited by the requirement that every agency must agree to them. The consensus rule produces policies under which everyone benefits, or at least no one loses too much. As one high-level policy adviser in China put it, "We must use all our policies to 'coordinate' interests among agencies and localities" (author's interview). Radical change is rejected, delayed, or watered down in the process of building consensus. Party principals occasionally can impose policies on their bureaucratic agents if the issue is of overwhelming urgency. But such instances are rare because the cost of dictating policy to a reluctant agency usually is failure of implementation.

An example of a major policy change delayed by lack of consensus is the restoration of ranks in the Chinese military, first proposed in 1980 but not implemented until 1988. (The military-ranks issue is discussed by Jonathan Pollack in chapter 6 in this volume). In 1984 the People's Liberation Army received new uniforms with epaulets in preparation for

[23] Governments with a higher degree of institutionalization solve the problem of instability in institutional choice by implicit contracts agreeing not to reconsider issues already decided by the group (Schwartz 1988). China has not yet established this institutional rule. Therefore, as Chen Junsheng writes, "A decision is formally made, a correct one at that, but just because it involves the interests of a certain department, unit, or locality, some people will refuse to put it into effect under all kinds of pretexts, and the result is a lot of arguing back and forth" (Chen 1987).

the restoration of ranks. The final decision to restore ranks was delayed until October 1988 because of conflict between retired officers and active officers over whether the retired officers should receive formal ranks. As a Hong Kong newspaper account described the situation, "These two opinions were locked in stalemate for a long time, and a consensus of opinion could not be achieved. This slowed down the process of restoring military ranks." The stalemate was broken by a compromise that restricts formal ranks to active officers but "soothes" the retired officers with special medals and an affirmation of their previous ranks (Fu Meihua 1988).

In China the normative reflection of consensus decision making in economic policy-making is an ideology one might call "balancism" (pinghengzhuyi ). According to this ideology, the function of the state is to balance out inequities among units created by arbitrary policies, particularly administratively set prices. Fairness (gongping ) requires that no unit lose too much because administrative prices make them less profitable than other units or because past decisions (such as investments in fixed assets) work against them under current formulas. All government bodies are supposed to adjust their policies to prevent large disparities in benefits among units. The expression "ku le bu zhun ," meaning "disparity between sadness and happiness," is frequently used by government bureaucrats to explain that a particular policy modification was necessary to guarantee equity. Representatives of ministries, localities, or enterprises who are in a strong competitive position to benefit from economic reforms (i.e., their products are in demand in the market, the irrationalities of the price system work for them, and they are more efficient and productive in their operations) often complain that government organs worry too much about taking care (zhaogu ) of the weak and not enough about promoting the strong (author's interviews). (In chapter 11 in this volume Andrew Walder analyzes this redistributive tendency in economic policy-making at the municipal level.)

The conservative bias of management-by-exception makes the political challenge of economic reform formidable. Introducing a market through a bureaucracy, especially one operating under management-by-exception, is extremely difficult. Under the consensus rule even a minority of ministries or provinces who prefer the status quo to the proposed changes could obstruct progress in reform. The Party principals retain the authority to take significant policy initiatives—the reform drive has been sustained despite short-term economic problems and group conflicts because Deng continues to promote reform ideas—but they must obtain the agreement of at least most of the bureaucratic institutions that will implement the policies.

Zhao Ziyang's political strategy of economic reform was designed to

meet this challenge. First of all, it was a strategy of gradualism. Rather than rushing ahead with a comprehensive, radical transformation of the entire system, which would threaten the vested interests of many, Zhao was extremely cautious. Zhao recognized that there were many political risks inherent in economic reform and that therefore they were "required to act carefully in reforms, like wading across a river by holding on to the rocks in it" (Gongren Ribao , 13 March 1985). He understood that "the reform of the economic structure is actually a process in which various interests and relations are readjusted and redistributed" and that if the interests of a particular department or locality were harmed by a particular reform, it would oppose the reform (Hongqi , March 1987). Gradual reform might be a long-term process, requiring as long as several generations, but no other strategy had a chance of success in the context of China's political system.

To minimize the threat to central economic agencies, Zhao Ziyang made the crucial decision to create a dual-track system, gradually expanding the market sector while maintaining the plan sector, instead of replacing plan with market at one shot. The policy allowing enterprises to sell their above-quota output on their own at market prices created an economic incentive for managers to press for more market opportunities and allowed central planners to save their functions and their face. Stimulated by the new incentives, the economy grew rapidly, especially in the market sector. This strategy of "letting the economy outgrow the plan" (Barry Naughton, personal communication, 1984) created a transitional dual economy, with numerous accompanying economic problems, but it was politically very successful.^[24]

The sequencing of reforms also reflects the political realities of the policy-making process. Zhao Ziyang decided to postpone the most redistributive policies, which would create the most intense conflict within the bureaucracy, namely, the "hard budget constraint" (i.e., bankruptcy) and price reform. Forcing enterprises to take sole responsibility for their own profits and losses (zi fu yingkuei ) is an essential component of market rationality, but it is politically infeasible under current conditions. From the perspective of "balancism" it would be unfair to punish enterprises that cannot make profits because of external, "objective" (keguan ) causes (i.e., prices, demands of planners, fixed assets, etc.). The burden would fall mainly on a few sectors (coal, steel, heavy machinery) and the inland provinces where these sectors are concentrated (author's interviews).

Zhao Ziyang began talking about price reform in 1984. He set up a

[24] I have no evidence that this policy was produced by bargaining within the government bureaucracy, but it was clearly designed to satisfy the interests of the bureaucracy.

group to prepare price reform policies and announced in early 1985 that the State Council would soon take action on price readjustment in industry as well as agriculture (Xinhua , 1 January 1985). The fact that a thorough adjustment of industrial prices (i.e., raising the prices of raw materials such as coal, iron, and steel) was continually postponed and has yet to be introduced,^[25] whereas food prices have been liberalized, suggests that it is the potential for intense bureaucratic conflict, as well as the public reaction to inflation, that stopped Zhao from pushing ahead with urban price reform.

Participants in economic policy-making often explain the delay in these two key dimensions of reform as a reflection of Zhao Ziyang's caution on tackling issues that would be opposed by the most powerful agencies within the government bureaucracy, namely, the State Planning Commission, the Ministry of Finance, and the heavy-industry ministries. Zhao pushed up against the stone wall of the state bureaucracy, they say, and he went through only where he found loose stones; he did not waste time pushing against stones that would not move (author's interviews).

The policies that have emerged from the bureaucracy since 1980 suggest that the progress of economic reform was sustained by appeasing the powerful heavy-industry ministries. When reform proposals were first introduced, heavy-industry interests expressed open opposition to them (Solinger 1982). As the prime beneficiaries of the Soviet-style command economy, they viewed economic reform as a direct threat. By replacing some ministers and making reform the ideological line, the Party discouraged outright opposition from the heavy-industry ministries. And over time, the bureaucrats in these ministries came to recognize that they could do better by demanding a larger share of the benefits of reform than by opposing all reforms (author's interviews).

In exchange for their support of various reform policies, the heavy-industry ministries received valuable side-payments. The clearest example of such a side-payment is the "departmental contracts" (bumen chengbao ). The ministries in charge of fuels, raw materials, and transportation perceived that they were falling further and further behind under the dual economic system. Sectors with many enterprises operating under low (or no) plan quotas could sell on the market and reap high profits from the higher market prices, while their sectors had to continue to produce almost entirely for the plan. The petroleum, coal, metallurgy, railroads, and communication ministries, and the airline, petrochemical, and nonferrous metals corporations demanded and received special "departmental

[25] The prices of specific materials were adjusted during this period, but a comprehensive industrial price reform was not undertaken. The issue of price reform was raised once again during the summer of 1988 but abandoned in the fall when inflation rose to an alarmingly high rate.

contracts." The ministries contracted with the State Planning Commission to receive a certain amount of investment and inexpensive plan inputs in exchange for delivering a certain amount of output to the plan for the next five years. Any production above this amount the ministries and their subordinate enterprises were free to sell on the market at higher prices (Xinhua , 10 February 1985; Shijie Jingji Daobao , 25 May 1987). The contracts guaranteed access to the market and higher profits to these particular agencies, which allocated the plan burdens and market opportunities to the enterprises under them. The departmental contracts were very popular with the State Planning Commission, because they helped guarantee plan procurement (the lure of the market sector had made this task increasingly difficult), and naturally with the ministries, which benefited from them (author's interviews). The departmental contracts violated the reform principle of increasing enterprise autonomy from administrative control, but they were politically expedient.

The need to obtain the support of the influential heavy-industry ministries for the package of economic reforms probably also explains the striking stability of the shares of central budgetary investment allocated to economic sectors during the period of economic reform. Heavy industry, which claimed the lion's share of state investment under the old system, found its investment allocation cut drastically in 1979–80 when top leaders sought to improve proportional balance by shifting resources to light industry (Solinger 1988). The blow to heavy industry was so severe, and its representatives complained so persuasively, that heavy industry's share of investment was increased again after 1982 (State Statistical Bureau 1985). Since that time heavy industry has continued to receive favored treatment by the Center, and light industry has increased its share of total investment only because most investment is now controlled at the local, not the central, level. Meanwhile, agriculture, which is weakly represented in the central government, has seen its share of central investment reduced despite a 1979 promise that it would be increased (Xinhua , 5 October 1979).

Many reform policies bear the mark of consensus decision making. One example is the decision to create a new tax system for industrial enterprises with a policy called "substituting tax for profits" (li gai shui ). The goal of this policy was to place the financial relations between enterprises and the central government on a stable, institutionalized basis. Instead of each enterprise bargaining for a particular rate of profit retention, all enterprises producing the same product would have to pay at a uniform tax rate. When this policy was originally proposed in 1983 the Ministry of Finance recommended a moderately high tax rate. At a series of meetings called to discuss the proposal, representatives of

heavy-industry ministries and inland provinces complained that under such a heavy tax burden their many unprofitable enterprises would be forced to close. To achieve consensus it was necessary to revise the policy. The tax rate was reduced, but then the Ministry of Finance came in with simulations predicting that the policy would reduce state revenues. The revenue gap was filled by tacking on a so-called adjustment tax (tiaojie shui ), which was applied only to the most profitable medium-sized and large state enterprises (most of them in coastal cities like Shanghai). The adjustment tax was set on an individual-enterprise basis, thereby violating a fundamental principle of reform but creating the compromise necessary to win approval of the policy (author's interviews).

The constraint of consensus building also requires that every policy change leave no one substantially worse off than before. The li gai shui tax regulations guaranteed enterprises that they would pay in taxes no more than the profit they remitted the year before. Whenever the li gai shui policy resulted in a loss of revenues to a particular province, the central government made compensation by reducing the share of total revenue the province was required to remit to the Center (Lieberthal and Oksenberg 1988). Officials agree that the only politically feasible approach to price reform is to compensate those who are hurt by it, with wage subsidies to urban consumers in the case of food prices, and with tax exemptions for manufacturing enterprises in the case of raw-materials prices. One reason that industrial price readjustment has become more elusive over the course of reform is that as the central treasury becomes increasingly strained (because local governments and enterprises are allowed to retain a larger share of financial revenues), it is more difficult for the Center to come up with the financial side-payments necessary to make the package broadly acceptable (author's interview).

Gradualism, postponement of the most divisive issues, side-payments, and compromise—all are the marks of a market reform filtered through a political system characterized by management-by-exception. Zhao Ziyang's strategy recognized the necessity of building consensus of support for economic reform. As he himself put it, "When adopting a reform measure, we must do our best to benefit all quarters concerned so that our reform will always have the support of the broad masses of the people and its success will be guaranteed" (Zhao 1987a).^[26]

[26] A more critical view of Zhao's reform strategy, expressed recently in a People's Daily article advocating price reform, is that the strategy involved the implementation of only those reforms that were "shallow and easy" and "of immediate benefit to the masses" without altering the fundamental character of economic relations. "Reform Involves Risks and Difficulties, Pass the Barrier Through Arduous Efforts," Renmin Ribao, 9 June 1988, FBIS , 9 June 1988, 37.

Center And Locality: Federalism In A Unitary State

The People's Republic of China is a unitary state, with formal authority constitutionally held by the central government. Yet in reality, China, even before the introduction of economic reforms, was much more decentralized than the Soviet Union (Schurmann 1968). Beginning in 1957, the Center shared with the provinces the authority to approve projects, control industrial enterprises, plan production, allocate materials, and collect fiscal revenues. Despite periodic attempts at recentralization, the trend since 1957 has been progressive decentralization to the provincial level (Naughton 1985, 1987; Wong 1985, 1986).

Why China took the path of administrative decentralization is a highly significant but as yet unanswered question. Schurmann suggests that decentralization was made possible by the strength of the Communist Party at the provincial level; in contrast, the Stalinist purges decimated the Soviet party at the regional level (Schurmann 1968). Building on Schurmann's point, we might speculate that because of the strong Party base in the provinces, the CCP leaders created a Central Committee in which provincial representatives played a major role. The leadership enfranchised three major blocs within the Central Committee, officials from the (government and Party) Center, officials from the provinces, and People's Liberation Army officers.^[27] Leaders competed for power by building support among these key constituencies. Whenever a Party leader perceived that rival leaders were blocking his policy initiatives by their control over the central bureaucracy, he attempted to build support for his initiatives by "playing to the provinces." According to this analysis, Mao Zedong stressed administrative decentralization to win provincial support for policies promoting revolutionary transformation in 1957, 1964, and 1967-70, and Deng Xiaoping used the same strategy to win provincial support for economic reform policies in 1980.

Administrative decentralization offers economic as well as political advantages to the central leadership. Provincial officials are the agents of the central Party and government. Delegation of authority, whether to government ministries or to provincial departments, improves efficiency by exploiting the superior information of agents. And in a state-run economy, a profit-sharing rule, like the one included in the fiscal decentralization policies promulgated in 1980, improves the incentives for agents to be more efficient.

The prior decentralization of the Chinese system has had profound consequences for the course of the post–1978 economic reforms. The

[27] Although this analysis concentrates on only two blocs within the Party, the central bureaucracy and the provinces, the third bloc, the military, played a very important role during certain periods, such as the 1970s.

cumulative effect of the progressive devolution of authority and resources from 1957 through the Cultural Revolution was to create a political system in which a substantial share of the planning decisions, management of factories, control of raw materials, and receipt of fiscal revenues was in the hands of local officials. Provincial Party secretaries also had an important voice within the CCP Central Committee. When Deng Xiaoping looked around for a group that could become the core of a reform coalition and that could counter the vested interests of the central economic bureaucracy in the command economy, he soon identified provincial officials. These local officials would play a key role in the implementation of reform policies because they controlled most of the enterprises in the country. And as the largest bloc within the Central Committee (36.8 percent in the twelfth and 39.4 percent in the thirteenth [South China Morning Post 1987]) they could provide critical political support within the Party if and when conservative leaders tried to challenge the reforms.

To win the support of provincial officials for the reform drive, Deng Xiaoping introduced a radical fiscal decentralization in 1980. This policy, officially called "apportioning revenues and expenditures between the central and local authorities, while holding the latter responsible for their own profit and loss" (Caizheng 1980) allowed provinces to fix for five years the amount of revenues they must remit to the Center and keep a proportion or all of the revenues over that amount. Provinces were assigned to five different categories of treatment, ranging from Guangdong and Fujian, who retained 100 percent of their above-quota revenues, to the three municipalities of Beijing, Shanghai, and Tianjin, who retained none of their above-quota revenues. In addition, provinces and the lower levels were permitted to keep all the profits from the enterprises controlled by them (Donnithorne 1981; Fujimoto 1980; Han 1982). The main advantage of this policy, colloquially called "eating from separate kitchens" (fen zhao chifan ), was the strong incentive for local authorities to expand their revenue base by developing their local economies. True, the growth motivated by new fiscal incentives produced some undesirable consequences: uncontrollable local investment, most of it going into profitable but reduplicative and wasteful processing plants, leading to supply shortages, inflation, and budget deficits; local protectionism; and excessive local imports, causing national imbalances in foreign exchange.^[28] The results of fiscal decentralization were dis-

[28] "In recent years, the power of the localities to act autonomously has been expanded, and the system of 'eating in one's own kitchen' has been practiced in finance: in order to increase their own financial revenues, the various places have neglected the question whether it is in conformity with the needs of the whole nation, and blindly developed the processing industry which reaps the greatest profits, resulting in duplicative construction; they even go so far as to refuse the sales of high quality productions from other places in local commercial departments, with a view to guaranteeing a market for their own productions, with the effect of carving up the regions" (Xue 1983). One fact that vividly illustrates the phenomenon of local expansion is that as of 1988, Zhejiang, a province with only 70 counties and cities, had 115 beer breweries (Gu and Li 1988).

torted by the irrational price structure and the lack of hard budget constraints, rather than by the fiscal policies themselves. The entrepreneurial energy sparked by fiscal decentralization was impressive nonetheless. And as a political strategy to win provincial support for economic reform, it was extremely successful.

Many of the other reform policies introduced in the 1980s also reflect the strategy of "playing to the provinces." Under the leadership of Zhao Ziyang, whose entire career was based in the provinces, administrative decentralization became a key element in the political strategy of economic reform. Central ministries were told to send down (xiafang ) control of their enterprises to provincial or municipal authorities. Local authorities were authorized to retain enterprise depreciation funds and expand other sources of extrabudgetary revenues. The financial freedom of provincial officials was further enhanced by transforming capital-construction funds into bank loans and by granting more autonomy to provincial bank branches. In the realm of foreign trade, provinces were permitted to set up their own trade corporations and delegated authority to approve imports and joint ventures. And after the implementation of the 1988 foreign trade responsibility system, provinces contracted with the Center to share the foreign exchange revenues from trade (as well as the local currency profits and losses from trade), much as they do local fiscal revenues (Yao 1988). The one reform policy that harmed the interests of provincial authorities, the 1983 li gai shui tax system, which took the financial profits of local enterprises away from provinces, was scrapped after four years; in 1987 it was replaced with profit-sharing contracts (chengbao ), which restored the provinces' claim to local-enterprise profits (author's interview). Policies granting special financial and planning authority (called jihua danlie ) to eleven cities (fourteen, as of February 1989), special foreign trade and investment authority to four special economic zones and fourteen coastal cities, and full provincial status to one region (Hainan Island) enhanced the economic power of China's most flourishing cities and brought these cities into the reform coalition along with the provinces.

The result of all these reform policies has been to shift the center of gravity in economic administration from central agencies (tiao ) to local government (kuai ).^[29] This shift is more valuable for provincial and munici-

[29] Although the national reformist leaders have bent over backward to appeal to provincial interests, the provinces still are under the sway of the Center. True, the Center lost some of its leverage when it dispersed control over revenues, material resources, and rents. Provincial officials bent on making profits often evade central regulations (Shirk 1988). Yet the Center remains the principal. It retains formal authority over the provinces by virtue of its power to appoint provincial governors and Party secretaries and to set the contracts (fiscal, production plan, and foreign trade) for each province. The formal authority of the Center over the provinces fundamentally constrains the bargaining between provincial and central officials. If a provincial Party secretary or governor steps out of line, the central Party leaders can punish him or her by demotion or transfer. The competition among provinces for the best revenue-sharing contract with the Center gives the Center another way to reward or punish the behavior of provincial officials. The provincial agents may have been delegated more authority than before, but they are still agents. If central officials do not punish provincial officials for failing to implement central directives, it is because the central officials choose not to—because they need the support of provincial officials in the Central Committee—not because they are unable to. This principal-agent perspective emphasizes the dependence of the provinces on the Center, in contrast to Lieberthal and Oksenberg's depiction of central-provincial relations as an interdependent relationship with the balance generally in favor of the Center (Lieberthal and Oksenberg 1988).

pal authorities than previous decentralizations because it occurs in the context of an increasingly marketized economy. With the market offering everyone new opportunities for making money, whoever controls access to the market has the opportunity to collect "rents" (Krueger 1974). Reform policies have both expanded market exchange and decentralized from Center to locality the rents collected by administrative regulation of the market. Having been delegated the authority to approve construction projects and imports, and to set license fees and other local commercial taxes, provincial and municipal authorities can, in effect, sell tickets to the market (Xinhua , 5 May 1988). If they charge money and use it to develop local infrastructure, as the Tianjin mayor, Li Ruihuan, was famous for doing (author's interviews), they are called statesmen. If they charge money and put it in their own pockets, they are called corrupt criminals. And if they take their payment in the currency of political support, they are called political bosses. The rent-decentralizing implications of many reform policies explain why provincial and municipal authorities are such enthusiastic supporters of economic reform.

At least some of these local rents were spent on expanding the size of local government. Top Party leaders, eager to win the support of local officials (and appease central ones), have tolerated a dramatic buildup of the national administrative apparatus. The average annual increase in the number of government cadres reached 330,000 per year; before 1980 the average increase was 110,000 per year. By the end of 1986 the total staff of government offices and organizations was 7.34 million, 78.2 percent higher than 1978 (Tang 1987). The increase in administrative expenditures that accompanied this growth in the size of government (a 250 percent increase over 1978) can be seen as a side-payment to keep local-and central-government officials satisfied during the period of reform.

The prior dispersal of resources and authority not only laid the foundation for Chinese leaders to make local officials the core of its reform coalition, it also had the surprising effect of motivating central economic bodies like the State Planning Commission and the Ministry of Finance to support important reform policies. These central organs were frustrated after years of trying to sustain central management of the economy in an environment characterized by dispersed material inputs and revenues. Planning and financial officials saw their actual control over the economy slipping away over the years. From their perspective, a formal sharing rule, dividing the functions, resources, and revenues between Center and province, was preferable to a continuing erosion of de facto control. At least the sharing rule would prevent further deterioration of their position and permit them to retain their current degree of economic control. This perspective of central comprehensive agencies explains why the Ministry of Finance proposed the 1980 fiscal decentralization, the State Planning Commission proposed the 1984 planning reform, and the Ministry of Foreign Economic Relations and Trade proposed the 1988 reform in foreign trade contracting (author's interviews).

China's form of decentralized communism has helped the Party leadership build political support for economic reform. Not only did it make provincial and municipal officials a natural constituency for reform, it also gave central economic officials a reason to support reforms that would preserve the sharing of control between Center and locality. One hypothesis that stems from this analysis is that the Soviet Union, which begins reform as a much more centralized system, will find the political challenge of economic reform even more difficult than China's.

Conclusion: Evaluating The Political Record Of Economic Reform

The political challenge of economic reform in communist states is to devise policies that build a coalition of support for the reform drive while also improving economic efficiency. The reformist leaders at the top of the Communist Party, Deng Xiaoping and Zhao Ziyang, appeared to have done a masterful job at designing a political strategy of reform that suited the Chinese political institutions. While maintaining the CCP's leadership over the government, they delegated to the government bureaucracy increased discretion. This modification of the principal-agent relationship between Party and government seemed to dilute the influence of conservative Party leaders who might have sabotaged the reform drive, at the same time that it improved the quality of policy decisions. Yet, as the ultimate principals, the top Party leaders retained the power to take policy initiatives, set the ideological line, and replace government

cadres. The management-by-exception method of decision making employed under this revised delegation relationship between Party and government provided the framework for bargaining out reform policies that could be supported by key bureaucratic agencies. Although the widening of consultation within the government bureaucracy made the task of building consensus more difficult, it meant that the policies that emerged would be satisfactory (or at least tolerable) to all groups. Zhao Ziyang's strategy of gradual reform, playing to the provinces while postponing the most redistributive measures and working out compromises by giving side-payments, was effective at sustaining the momentum of reform until 1988.

The political choices made during 1980–88 built a coalition of support for economic reform but made the political challenge of future reforms much more difficult. The path of reform reflected more a political logic than an economic logic. By postponing industrial price reform and the introduction of a hard budget constraint, the reformist leaders stimulated rapid economic growth accompanied by material shortages, budget and foreign exchange deficits, inflation, and corruption. These new economic problems backfired on the reform drive at both the elite and the mass level. They provided ammunition for Party conservatives who opposed radical reforms and made a play to reclaim dominance within the Party. At the mass level, the social mobilizational effects of market reforms and public dissatisfaction with inflation and corruption sparked social protest, beginning in late 1986 and building to the massive demonstrations in eighty-four cities during spring 1989. The conservatives pointed to protests as evidence that reform was leading to social chaos (luan ) and that new Party leadership was needed to restore stability. Despite the achievements of Deng and Zhao at brokering reform policies through the Chinese bureaucratic system, the combination of elite power struggles and social unrest destroyed the momentum of reform in 1988–89. As the principal-agent framework would predict, once the conservatives strengthened their hold on the Communist Party leadership organs (with the firing from the Party secretary position of Hu Yaobang in 1987 and Zhao Ziyang in 1989), the policies emerging from the government changed from reform to retrenchment. Whether the reform drive will recover its momentum in the future depends on the outcome of the competition for Party leadership after Deng Xiaoping's death.

References

Arrow, Kenneth J. 1985. "The Economics of Agency." In Principals and Agents: The Structure of Business , edited by John W. Pratt and Richard J. Zeckhauser. Boston: Harvard Business School Press.

Barnett, A. Doak. 1967. Cadres, Bureaucracy, and Political Power in Communist China. New York: Columbia University Press.

———. 1985. The Making of Foreign Policy in China: Structure and Process. Boulder: Westview Press.

Bunce, Valerie. 1981. Do New Leaders Make a Difference? Executive Succession and Public Policy Under Capitalism and Socialism. Princeton: Princeton University Press.

Burns, John P. 1987. "China's Nomenklatura System." Problems of Communism 36 (September–October): 36–51.

———. 1988. "Reform of Contemporary China's Civil Service System: Proposals of the 13th Party Congress." Paper presented to the Annual Meeting of the Association for Asian Studies, March 1988.

Caizheng. 1980. "The Temporary Provisions for Carrying Out a Financial Management System of 'Apportioning Revenues and Expenditures Between the Central and Local Authorities, While Holding the Latter Responsible for Their Own Profit and Loss.'" Caizheng , no. 12 (December); Joint Publications Research Service (JPRS) , no. 77592, 16 March 1981, 35–39.

Chen Junsheng. 1987. "Increase the Work Efficiency of Public Organs." Renmin Ribao , 19 March, Foreign Broadcast Information Service (FBIS) , 1 April, K33–39.

Comisso, Ellen. 1986. "Introduction: State Structures, Political Processes, and Collective Choice in CMEA States." In Power, Purpose, and Collective Choice, Economic Strategy in Socialist States , edited by Ellen Comisso and Laura D'Andrea Tyson. Ithaca: Cornell University Press.

Deng Xiaoping. 1980. "On the Reform of the System of Party and State Leadership," 18 August. Selected Works of Deng Xiaoping, 1975–1982. Beijing: Foreign Language Press, 1984.

Donnithorne, Audrey. 1981. Centre-Provincial Economic Relations in China. Canberra: Australian National University, Contemporary China Papers, no. 16.

Fu Meihua. 1988. "Inside Story on the Restoration of Military Ranks in China." Kuang Chiao Ching (Hong Kong), 16 May, FBIS , 20 May, 20–24.

Fujimoto, Akira. 1980. "The Reform of China's Financial Administration System." Japan External Trade Research Organization, China Newsletter , March, 2–9.

Gongren Ribao. 1985. "Reforms Must be Carried Out Step by Step." Gongren Ribao , 13 March, FBIS , 25 March, K26–27.

Gu Wanming and Li Li. 1988. "Expanded Use of Extra-Budgetary Funds." Xinhua , 17 January, FBIS , 28 January, 13–14.

Han Guochun. 1982. "A Brief Introduction to the System of 'Apportioning Revenues and Expenses Between the Central and Local Authorities, While Holding the Latter Responsible for Their Own Profit and Loss in Financial Management.'" Caizheng , no. 7 (5 July), JPRS , no. 82018, 19 October, 16–19.

Harding, Harry. 1981. Organizing China: The Problem of Bureaucracy, 1949–1976. Stanford: Stanford University Press.

Hongqi. 1987. "Fully Understand the Long-Term Character and Difficulty of the Reform." Hongqi , March, FBIS , 16 March, K13–15.

Hough, Jerry. 1977. The Soviet Union and Social Science Theory. Cambridge: Harvard University Press.

Krueger, Anne. 1974. "The Political Economy of the Rent-Seeking Society." American Economic Review 64: 291–302.

Lampton, David M. 1987. "Water: Challenge to a Fragmented Political System." In Policy Implementation in Post-Mao China , edited by David M. Lampton. Berkeley and Los Angeles: University of California Press.

Lardy, Nicholas. 1983. Agriculture in China's Modern Economic Development. Cambridge: Cambridge University Press.

Lawler, Edward E. III. 1976. "Control Systems in Organizations." In Handbook of Industrial and Organizational Psychology , edited by Marvin Dunnette. Chicago: Rand McNally.

Li Peng. 1988. "Li Peng Addresses Party Work Meeting." Xinhua , 22 March, FBIS , 23 March, 15–16.

Lieberthal, Kenneth, and Michel Oksenberg. 1988. Policy Making in China: Leaders, Structures, and Processes. Princeton: Princeton University Press.

Lo Ping. 1988. "Notes on the Northern Journey." Cheng Ming (Hong Kong), 1 April, FBIS , 1 April, 42–46.

McCubbins, Mathew, and Thomas Schwartz. 1984. "Congressional Oversight Overlooked: Police Patrols vs. Fire Alarms." American Journal of Political Science.

MacFarquhar, Roderick. 1974. The Origins of the Cultural Revolution. Vol. 1, Contradictions among the People, 1956–57. New York: Columbia University Press.

———. 1983. The Origins of the Cultural Revolution. Vol. 2, The Great Leap Forward, 1958–1960. New York: Columbia University Press.

Ma Hong. 1987. "Ma Hong, General Director of State Council Economic and Social Development Research Center, Answers Reporter's Questions." Guangming Ribao , 29 October, FBIS , 13 November, 37–38.

Naughton, Barry. 1985. "False Starts and Second Winds: Financial Reforms in China's Industrial System." In The Political Economy of Reform in Post-Mao China , edited by Elizabeth J. Perry and Christine P. W. Wong. Cambridge: Harvard University Press.

———. 1987. "The Decline of Central Control over Investment in Post-Mao China." In Policy Implementation in Post-Mao China , edited by David M. Lampton. Berkeley and Los Angeles: University of California Press.

Oksenberg, Michel. 1971. "Policy Making Under Mao, 1949–68: An Overview." In China: Management of a Revolutionary Society , edited by John M. H. Lindbeck. Seattle: University of Washington Press.

———. 1982. "Economic Policy-Making in China: Summer 1981." China Quarterly , no. 90 (1982): 165–94.

Pratt, John W., and Richard J. Zeckhauser, eds. 1985. Principals and Agents: The Structure of Business. Boston: Harvard Business School Press.

Renmin Ribao. 1983. "Resolutely Stop Acts of Wrangling Detrimental to the Overall Interests," 21 January, FBIS , 24 January, K5–6.

———. 1984. "Breaking with Departmental Bias, On the Importance of Strengthening Sectoral Management during Industrial Reform," 17 July, 1.

———. 1987. "Draft Revision of Some Articles of the Constitution of the Chinese Communist Party," 2 November.

———. 1988. "Reform Involves Risks and Difficulties, Pass the Barrier Through Arduous Efforts," 9 June, FBIS , 9 June, 37–38.

Schurmann, Franz. 1968. Ideology and Organization in Communist China. 2d ed. Berkeley and Los Angeles: University of California Press.

Schwartz, Thomas. 1988. "The Meaning of Instability." Paper presented to the Department of Political Science, University of California, San Diego, June.

Shijie Jingji Daobao. 1987. "Different Types of Contracted Managerial Responsibility Systems Used by Enterprises," 25 May, 7.

Shirk, Susan L. 1988. "The Acquisition of Foreign Technology in China: The Bargaining Game." Paper presented to the 17th Pacific Trade and Development Conference, Bali, Indonesia, July.

———. 1989. "The Political Economy of Chinese Industrial Reform." In Remaking the Economic Institutions of Socialism: China and Eastern Europe , edited by Victor Nee and David Stark. Stanford: Stanford University Press.

Solinger, Dorothy J. 1982. "The Fifth National People's Congress and the Process of Policymaking: Reform, Readjustment, and the Opposition." Issues and Studies 18, no. 8 (August 1982): 63–106.

———. 1988. "Disinvestment and the Politics of Pleading." Paper presented to the Annual Meeting of the Association for Asian Studies.

South China Morning Post. 1987. "Paper Analyzes New Central Committee." Saturday Review , 7 November, FBIS , 9 November, 18–19.

State Statistical Bureau. 1985. China: A Statistical Survey in 1985. Beijing: New World Press.

Tang Tian. 1987. "The Readjustment of the Cadre Distribution Structure Is Imperative." Liaowang , no. 36, 7 September, FBIS , 18 September, 18–19.

Teiwes, Frederick C. 1979. Politics and Purges in China: Rectification and the Decline of Party Norms, 1950–1965. White Plains, N.Y.: M.E. Sharpe.

———. 1984. Leadership, Legitimacy, and Conflict in China. Armonk, N.Y.: M. E. Sharpe.

Tian Jiyun. 1986. "Speech at a Conference of Central Organs." 6 January. Xinhua , 11 January, FBIS , 13 January, K21.

Tsou, Tang. 1983. "Back from the Brink of Revolutionary-'Feudal' Totalitarianism." In State and Society in Contemporary China , edited by David Mozingo and Victor Nee. Ithaca: Cornell University Press.

Wong, Christine P. W. 1985. "Material Allocation and Decentralization: Impact of the Local Sector on Industrial Reform." In The Political Economy of Reform in Post-Mao China , edited by Elizabeth J. Perry and Christine P. W. Wong. Cambridge: Harvard University Press.

———. 1986. "Ownership and Control in Chinese Industry: The Maoist Legacy and Prospects for the 1980's." In U.S. Congress, Joint Economic Committee, China's Economy Looks Toward the Year 2000 , vol. 1. Washington, D.C.: U.S. Government Printing Office.

Xinhua. 1979. "Decisions on Some Questions Concerning the Acceleration of Agricultural Development," 5 October. In The People's Republic of China, 1979–1984: A Documentary Survey , edited by Harold C. Hinton, vol. 2. Wilmington: Scholarly Resources, Inc.

———. 1985. "Zhao Comments," 1 January, FBIS , 2 January, K9–11. "Metallurgical Conference's New System Approved," 10 February, FBIS , 12 February, K16–17.

———. 1987. "Forum Discusses Deng's Speech," 5 July, FBIS , 6 July, K5–8. "New Management Systems to Categorize Cadres," 18 November, FBIS , 18 November, 17–18.

———. 1988. "Provisional Regulations Banning Unjustifiable Impositions on Enterprises" (promulgated by the State Council on 28 April 1988), 5 May, FBIS , 9 May, 31–33. "Central Committee Reports on CPC Membership," 17 June, FBIS , 17 June, 17.

Xue Muqiao. 1983. "China's Current Economic Situation, Analysis and Prospects." Renmin Ribao , 3 June, FBIS , 13 June, K34–41.

Yan Jiaqi. 1988. "'Separation of Party and Government Work' in China as Viewed From a Comparative Angle." Wen Wei Po (Hong Kong), 23 March, FBIS , 14 April, 42–45.

Yao Yilin. 1988. "Report on the Draft National Economic and Social Development Plan for 1988." Paper presented to the First Session of the Seventh National People's Congress, 26 March. Renmin Ribao , 17 April, FBIS , 20 April, 18–27.

Zhao Ziyang. 1987a. "Report on the Work of the Government." Paper presented to the Fifth Session of the Sixth National People's Congress. Beijing Domestic Service, 25 March, FBIS , 26 March, K1–25.

———. 1987b. "Advance Along the Road of Socialism with Chinese Characteristics." Work Report to Thirteenth National CCP Congress. Xinhua , 25 October, FBIS , 26 October, 10–34.

———. 1987c. "On the Separation of Party and Government." Speech to the Preparatory Meeting for the Seventh Plenary Session of the Twelfth Central Committee on 14 October 1987. Xinhua , 26 November, FBIS , 27 November, 13–16.

Zhou Yi. 1987. "The Question of Party Leadership During Political Structure Reform." Lilun Yuekan , no. 7 (25 July), FBIS , 1 September, 18–23.