Nine
Hierarchy and the Bargaining Economy: Government and Enterprise in the Reform Process

Barry Naughton

Before economic reform China ran a command economy, and economic decisions were made and evaluated within a hierarchical bureaucracy. Because decision-makers were not directly subject to competitive pressures or external review, decisions generally emerged from a process of negotiation and bargaining within the bureaucracy. Bargaining in the upper reaches of the bureaucracy determined the choice of investment projects, and bargaining between enterprises and their superiors over planned targets determined current production levels. This system permitted inefficient decisions to be made repeatedly because the costs of wrong choices were not borne by any single individual or work unit, but were instead diffused through the economy as a whole. From the beginning, one of the hopes of reform was that it would reduce the scope for decision making by bargaining, substituting instead an objective "discipline of the market."

In fact, economic reforms have transformed the nature of bargaining relations in China's industrial economy, changing the bargaining positions of superiors and subordinates, transferring control over large blocks of resources, and introducing elements of market competition into the decision-making process. But the hope that reform would somehow diminish the overall importance of bargaining within the bureaucracy has been sorely disappointed. Instead, reforms have caused bargaining relations to become more complex and even more pervasive. Within the state sector the maintenance of the administrative hierarchy has preserved the basic precondition for the bargaining relationship, while the increasingly complex and diversified economic environment has enriched the content of the bargains that can be struck. In this new bargaining environment the roles of the central government and state

enterprises have been recast, without, however, solving the question of what those roles should be.

The new bargaining relations are the result of changes in the forces that shape the bargaining environment. Changes in two areas are particularly striking. First, the relative strength of different parties in the hierarchical bargaining process—determined primarily by their control over resources and information—has shifted. The central government has been weakened by a decline in the volume of resources under its direct control, but strengthened by an increase in information and skills and by a broader range of instruments at its disposal. In certain respects this combination has led to an unexpected increase in the strength of the central government. Second, the coexistence of plan and market sectors has led to bargains with increasingly complex and diverse contents. There are simply more economic variables to be bargained over than there were in the past. From the enterprise standpoint, this aspect of change has been crucial. While enterprises have more resources at their disposal, they also face a vastly more complex bargaining environment, and bargaining has shifted from predominantly plan bargaining to a complicated mixture of plan, exchange, and redistributive bargaining. The combination of ambiguous shifts in power relations and increasing complexity in the economic system has resulted in an increased prominence for bargaining overall in economic decision making. This is ironic, for we might have supposed that it would be precisely in the economic realm that market relations would replace the give-and-take of bureaucratic politics at an early stage of the reform process.

In the first section of this chapter I introduce the conceptual framework by describing a pure "command-bureaucratic" economy. Stress is laid on the closed, monopolistic nature of the system and on the degraded character of information flows within the bureaucracy. The prereform Chinese economic system is described with reference to these characteristics. The second section begins the discussion of the contemporary Chinese environment by examining the role of the central government in Beijing in the investment process, stressing changes in bargaining strength. The third section examines the changing bargaining environment at the enterprise level. This necessarily involves a discussion of the "two-track" strategy for economic reform and of the nature of enterprise response to the new risks and opportunities facing them. In the fourth section, I examine the basic structural conditions that serve to maintain and reproduce the bargaining relationship at the enterprise level: I stress the continuing bilateral monopoly that prevails between enterprises and their superiors. Some of the implications of this way of looking at the Chinese reform process are presented in a brief concluding section.

The Command-Bureaucratic Economy

The economic system China operated before reform is often called a "command economy," and this term accurately captures the nature of most economic decision making in such a system. Information is collected from production and consumption units; the information flows upward through bureaucratic channels; decisions are made on the basis of this information; and commands are issued down through the same bureaucratic channels to determine production decisions. Vertical flows of information and command are thus the basis of the system. The label "command economy" applies to the economic system as a whole because the "commands" issued by superiors are the central features that organize the system. Just as the flow of energy through an organism or a machine determines the physical form of that system, so the flow of commands through the bureaucracy determines the characteristic forms of the command economy. Individual incomes are determined by the extent to which commands are carried out (degree of plan fulfillment); input purchases and output sales are planned to enable production commands to be fulfilled; and financial flows are set to accommodate those planned tasks. Moreover, various planning exercises are expected to mesh into a single integrated plan that expresses the will of those who command. Thus, the command economy is "monomorphic," or uniform: all economic decision making is organized in such a way that it replicates and is subordinate to the basic "command" relationship. Similarly, economic organizations generally have the same internal structure regardless of their rank in the hierarchy, because all organizations serve the same functions of relaying commands and information. This monomorphism characterizes the command economy regardless of whether it is highly centralized or relatively decentralized.

The command economy is also monolithic. That is, there are no significant organizations outside the planned economy that compete with units inside. Because of the lack of competition and markets, prices do not carry much information useful in economic decision making. Instead, important information flows mainly through a few narrow channels connecting lower and higher levels. The importance and scarcity of official information channels means that they become the focus of interest of many different individuals. Those at lower levels, for example, have an interest in retaining information so that they can use this scarce resource to advance their own careers. The same superior-subordinate relations are used to structure information gathering and to issue commands, so the incentives to manipulate and distort those relations are very great. Bargaining within the bureaucracy is

concentrated on the level of commands coming down the bureaucratic chain and the type of information going up: it is predominantly "plan bargaining."

In ordinary times this bargaining takes the form of "hiding reserves." Lower-level units wish to conceal capacity from their superiors in order to obtain plans that are easy to fulfill. In that way, they can be assured of a quiet life and an adequate income. Thus, the economic system as a whole tends to sink into a low-level equilibrium of low productivity and low effort. The situation is neatly captured by an epigram from Eastern Europe: "They pretend to pay us, and we pretend to work." In this respect, the command economy resembles any bureaucratic system, which may fall into this low-level trap when morale is low, tasks routinized, and external checks weak. The basic problem is that the narrow channels connecting subordinates to superiors become clogged with pseudoinformation, which is often intentionally distorted. While the system continues to report thousands of "bits" of data, the actual information content is quite limited. Production data are abundant, but these reflect merely an institutional consensus about appropriate levels of effort, rather than actual information about attainable levels of output. Uniformly organized production units all report more or less adequate performances, and it becomes impossible to know, for instance, the potential savings in energy usage that could be achieved by a drastically reshaped enterprise. Because of this impoverished information flow, it becomes difficult for leaders to get the kind of response from the "command economy" that they desire. The "command economy"—a model of subordination to the leader's will— becomes instead the "bureaucratic economy"—a model of unresponsiveness. It would be best to characterize these systems as "command-bureaucratic economies" to capture both the authoritarian flavor and the sense of unresponsiveness which the word bureaucratic has come to carry in daily language.^[1]

One of the curiosities of the command economy is the tendency of the system to generate an opposite, superresponsive kind of behavior during certain exceptional periods. During exceptional "forward leaps" a different response emerges, which we can call Stakhanovite after the Soviet coal miner who hewed 104 tons of coal—fourteen times his output quota—on one particularly good day in 1935. During these periods the incentive structure is altered to reward exceptional achievements, and production workers and units begin vying to overfulfill their plans by ever more astonishing margins. Suddenly, all the desire to conceal reserves is abandoned: in the context of a revivalist spirit, the worker-hero

[1] Charles Lindblom characterizes these systems as having "strong thumbs, no fingers." See Politics and Markets: The World's Political-Economic Systems (New York: Basic Books, 1977), 65–75.

shatters the stagnation of the bureaucratic system. Certainly the most spectacular example of this behavior shift was the Great Leap Forward in China, when it seemed for a period that all the laws of nature had been repealed. In 1958 cadres eager for recognition reported spectacular grain harvests, leading the central leadership to believe that China's total harvest had increased by a miraculous amount. The commands that followed included instructions to reduce the acreage sown to grain and increase deliveries of grain to the state, thus leading directly to disastrous famine.^[2] While Stakhanovite leaps forward seem to be the opposite of the bureaucratic economy, they are really just the flip side of the same phenomenon. In both cases, the tangling of the incentive system and information flows creates a distorted and degraded flow of information; the indeterminacy of the whole system, because of the absence of external checks, permits the most outrageous outcomes to appear acceptable for a period. The extremes of stagnation and Stakhanovite leaps forward are both more likely when central planners have a weak and uncertain grasp over concrete decision making in the economy.^[3]

What determines the level of effective control over the economy exerted by central planners? The size and complexity of the economy play a major role, and Chinese planners would face a formidable control problem under any conceivable system. More specifically, however, effective control basically depends on two factors. The first is the direct control over resources—particularly investment resources—exercised by planners, and the second is the quality of the information available to planners. In both these respects, Chinese planners were exceptionally weak before reforms. Even before reforms, financial control over one-third of investment had been decentralized, and only two-thirds of investment was disbursed directly through the government budget. Slightly less than half of state investment went for projects that were included in the central-government investment plan. Moreover, central-government resources were tied up in the misguided "Third Front" development strategy, leaving central planners with little freedom to shift resources in

[2] The process of misinformation can explain policy mistakes through late 1958, but it cannot, of course, explain the criminal continuance of the Great Leap Forward after such misinformation was recognized, probably by the beginning of 1959, and certainly after the Lushan Plenum in summer 1959.

[3] In a recent analysis, the Stakhanovite movement in the Soviet Union has been explicitly linked to a radical attempt to shake up a system sinking into bureaucratic stagnation. See J. Arch Getty, Origins of the Great Purges: The Soviet Communist Party Reconsidered, 1933–1938 (New York: Cambridge University Press, 1985). As individuals, Stakhanovite workers received substantial material rewards and, on occasion, high political office. P. Krivonos, the leading Stakhanovite in the railroads, parlayed his achievement into a position on the Ukrainian Central Committee. John Armstrong. The Politics of Totalitarianism: The Communist Party of the Soviet Union from 1934 to the Present (New York: Random House, 1961), 91–93.

response to changing priorities.^[4] Even more striking was the decline in the quality of the information available to planners. During the Cultural Revolution the State Statistical Bureau was reduced sharply in size, and most of its functions were incorporated into the planning commissions at all levels. This meant that the government sacrificed a semiautonomous source of information and became completely dependent on information channeled directly through the industrial hierarchy. Moreover, the whole scope of statistics gathering changed. Previously, the Statistical Bureau had been charged with gathering data about the entire economy, but as control over significant blocks of resources was decentralized, the data-collection network shrank to cover only those areas directly under central control. Thus, when control over "technical transformation" investment was decentralized, the government ceased to collect information about this important component of investment. The government literally did not know how much total investment was taking place.^[5] Similarly, materials that were under local control and "balanced" by local authorities were not incorporated into the central-government balancing process at all. Of course, the government continued to collect output figures, but it made no effort to coordinate sources and uses of this important portion of total output. Finally, the decimation of the technically skilled economic bureaucracy meant that only relatively crude direction of resource flows could take place. Only a few hundred commodities were centrally planned, whereas in the Soviet Union several thousand such commodities are planned. With limited skills and limited information, the central government was unable to exercise detailed control over even those portions of the economy where it nominally possessed absolute authority.

China's economic system before reform was unquestionably a "command economy," but it is of only limited value to describe it as a centrally planned economy. While the Center had enormous formal authority, and while the ultimate centralization of the Communist Party and other aspects of the political system cannot be neglected, the Center was extraordinarily weak compared with other planned economies. As a result, the industrial economy was unresponsive to attempts to regulate daily decision making but was at the same time vulnerable to recurrent periods of

[4] For the central-government investment plan, see the Appendix. For the other points, see Barry Naughton, "The Decline of Central Control over Investment in Post-Mao China," in Policy Implementation in Post-Mao China , ed. David M. Lampton (Berkeley and Los Angeles: University of California Press, 1987), 51–80; and "The Third Front: Defense Industrialization in the Chinese Interior," China Quarterly , no. 115 (Autumn 1988).

[5] See Wang Yifu, ed., Xin Zhongguo Tongji Shigao (A Draft History of Statistics in New China) (Beijing: Zhongguo Tongji, 1986 ), 165–82; Planning Section, Fushun Colliery, "How We Unified Statistical Work for Technical Transformation," Tongji , 1981, no. 4:11–12.

leaping forward. One case that occurred on the eve of economic reform is symptomatic of the weakness of central control. In November 1977 the State Council approved a proposal to build at Baoshan in the Shanghai suburbs an advanced iron mill, to be imported from Japan and capable of supplying 5 million metric tons (MMT) of iron annually. Within six months of approval, the cost of the projected plant had approximately quadrupled, as it was expanded to a comprehensive steel mill producing 6 MMT of iron and 6 MMT of steel, plus continuous hot and cold steel rolling mills. Yet at this time no blueprints or construction plans had ever been submitted to the central Planning Commission. In July 1978 the Planning Commission finally obtained and approved blueprints, but the discovery that the designs submitted incorporated a further expansion of the project and still more cost increases led to escalating doubts. Subsequent rethinking led to the recognition that the project was deeply flawed, but by that time contracts had been signed with Japanese suppliers that basically locked the government into the proposal. Ultimately, the government proceeded with a project with a total cost of over 20 billion yuan (over $4 billion at today's exchange rates), in spite of the fact that planners had not possessed any detailed information about the project until it was already under way.^[6] Surely few decisions of this magnitude have ever been made on such a flimsy information base. We can speculate that cases like this, occurring after the door had been decisively closed on the Cultural Revolution era, forced China's leaders to recognize the weakness of their bureaucratic decision-making process and made them more receptive to the possibility of economic reform.

The Central Government And The Control Of Investment

The early stages of economic reform in China were dominated by a process of decentralization. But while decentralization was taking place, a countervailing movement that improved the skills, information, and control available to the central government can also be discerned. This accumulation of skills was not in any sense contrary to the ideals of reform; indeed, increased sophistication of central officials was a key objective of reform in China, as in other socialist countries. Improved government skills would be necessary to guide the economy through the complexities of an increasingly marketized economy open to the outside world. However, in the face of constant change and recurrent crises,

[6] Liu Jingtan and Hong Huiru, "Report on the Baoshan Steel Mill," Gongye Jingji Guanli Congkan , 1981, no. 2:29–32; Martin Weil, "Baoshan: A Symbol of Change in China's Industrial Policy," in China Under the Four Modernizations , ed. U.S. Congress Joint Economic Committee, 1:365–93.

central leaders have understandably used their newly available skills and information, and this has involved the central government in new activities despite the general decentralization process. Moreover, the two-track or "piecemeal" reform process in China has left a wide range of activities potentially open to central-government involvement. As a result, while the volume of resources directly under the control of the central government has declined, the Center's bargaining position has in other respects been enhanced by its greater access to information and the wider range of tools at its disposal. These contrasting trends can be seen most clearly through an investigation of the crucial focus of government activity in a command economy, the investment process.

By acceding to a dramatic reduction in its direct control over investment resources, the central government created the economic space that allowed reform to proceed in the late 1970s. Enterprises were given direct control over a substantial portion of profits and depreciation allowances that had previously been drawn into the state budget, and the government budget shrank as a proportion of the economy. The clearest indicator of direct government control over investment resources is the amount of investment that is funded directly through the government budget (shown in figure 9.1 as a proportion of national income [net material product].) Between 1978 and 1981 budgetary investment fell by half, from almost 15 percent of national income to slightly over 7 percent; moreover, the decline was persistent, with budgetary investment declining each successive year. This is an extremely unusual phenomenon, virtually unprecedented in the experience of centrally planned economies.^[7] Between 1981 and 1984 budgetary investment stabilized. While it rose slightly as a proportion of national income to above 8 percent, it continued to decline as a proportion of total investment, which was increasing rapidly. After 1984 budgetary investment again entered a declining phase and fell below 6 percent of national income in 1987.

In the early reform years the central leadership acceded to this diminution in its direct control and began rebuilding the institutions that would improve the quality and volume of its information about the economy. At this point central planners had little choice, since they quite literally had no central plan: the existing planning procedures, as of the late 1970s, had produced only the grandiose Ten-Year Plan, which was widely recognized as unrealistic and had been discarded. The rebuilding of information networks began with the rehabilitation and strengthening

[7] I discussed this phenomenon in 1983 in "The Decline of Central Control over Investment in Post-Mao China," in Lampton, ed., Policy Implementation . This section is an extension and revision of the analysis presented there.

Fig. 9.1.
Fixed Investment: Budgetary and Central Government

of the State Statistical Bureau. The Statistical Bureau then gradually expanded the coverage of its data-collection network, improving the reliability and meaningfulness of data. In the sphere of investment the Statistical Bureau gradually progressed from collecting information only about capital construction within the state economy to collecting information about all kinds of fixed investment economy-wide, including that carried out by individual households. Computerized information centers under the State Council now regularly collect data from all large industrial enterprises. Several new research institutes labor with increasing sophistication to interpret and analyze the available information.^[8] As the available information base improved planners began to engage in a series of long-range projections of the future of the economy, projections

[8] See chapter 5 in this volume, and "Fixed Investment Statistics," China Statistics Monthly (University of Illinois at Chicago), 1988, no. 2:1–2.

that became gradually more meaningful and specific.^[9] Even as direct central control over investable resources declined, a gradually more sophisticated process of economic strategizing led to increased demands that investment accord with central-government priorities.

As central planners gradually developed a coherent vision of future economic development, they began to engage in a continuous tug-of-war with local interests. Generally speaking, central planners tried to increase the flow of investment to key bottleneck sectors, particularly energy and transport. Local governments and enterprises, by contrast, tried to develop industries with high profit rates and prospects for rapid growth. Because China's distorted price system keeps the profitability of energy low, and projects in this sector require large-scale investments often beyond local capabilities, the energy sector was rarely targeted by localities. They gambled instead that bottlenecks in transportation and energy would ultimately be taken care of by the central government, and that localities with the most promising early development of profitable industries would be able to hold on to those assets.

From the beginning of the reform process, the central government repeatedly stressed the need to develop energy production, but energy investment nevertheless stagnated between 1978 and 1982. This stagnation came about in part because central planners simply did not have workable plans for energy development. They had not carried out the detailed work of selecting sites and projects, drawing up blueprints, and working out supply arrangements. For a period they hoped the foreign multinationals would solve some of the problems by developing offshore oil fields, but this hope was disappointed. Priority to energy and transportation remained a long-range goal rather than a concrete task for operational plans. In addition, planners felt they could temporarily survive by allowing the industrial structure to shift toward a lighter, less energy-intensive pattern. The new decentralized funding mechanisms that were devised were used predominantly for light-industry investment. For example, when new programs of bank lending for fixed investment were initiated, 13 billion yuan worth of fixed investment were funded by bank loans, and of this, 68 percent went for light industry (mid-1979 through mid-1982).^[10] During this period even central-government projects were frequently in light and consumption-goods industries. Ultimately, however, the growing economy would need more investment in energy and transport.

[9] Carol Hamrin, China and the Challenge of the Future (Boulder, Colo.: Westview, 1990). Barry Naughton, "China's Experience with Guidance Planning," Journal of Comparative Economics 14, no. 4 (December 1990): 743–67.

[10] Zhu Tianshun, "Medium Period Equipment Loans Make a Big Contribution," Zhongguo Jinrong 1982, no. 22:1–3, translated in JPRS Economic Affairs, no. 303, 4–6.

During the early 1980s the central government gradually created a menu of projects in the energy and transport sectors that it wished to carry out. The result was a steady revival in the importance of the central investment plan. Statistics are shown in figure 9.1 on investment spending on all projects included in the central-government investment plan. (For a description of the data, see the appendix to this chapter.) Between 1978 and 1981 spending on central-government projects declined substantially but not as much as budgetary investment. Spending on central projects declined from slightly over 10 percent of national income to just below 8 percent, a reduction of 2.5 percent of national income—a very substantial change, but quite a bit less than the decline in budgetary investment. In 1978 spending on central projects was significantly less than budgetary investment. If we think of the central government as simultaneously establishing an investment-funding mechanism and a program of investment projects, these two activities in conjunction yielded a substantial surplus in 1978, equal to slightly over 4 percent of national income. Thus, after spending on central projects was complete, budgetary funds were still available to fund projects planned by local governments.

From 1981 through 1984 budgetary investment and spending on central-government projects were roughly equal. This does not mean that all budgetary investment went directly to central-government projects: some central projects were funded in whole or in part through bank loans, and some through extrabudgetary retained funds; conversely, some budgetary investment went to local-level projects. Netting out these flows, central-government projects and budgetary investment were roughly in balance, with a slight deficit amounting to about 1 percent of national income. In order to fund central-government projects the central government would have to draw in extrabudgetary funds or bank loans equal to about 1 percent of national income, even if no budgetary funds went to local-level projects.^[11] During this period, household and enterprise saving was rising rapidly, so drawing on surplus funds was not difficult, and no serious economic problems arose.^[12] After 1984 the spending trend for central-government projects diverged from that of budgetary investment.

[11] Note that this deficit in the central-government investment program is entirely separate from, and in addition to, the government budget deficit. The budget deficit measures the borrowing necessary to fund actual government expenditures; the investment program deficit measures the borrowing necessary to complete the investment program after actual government expenditures have been used.

[12] See the important flow of funds data given in Macroeconomic Research Office, Economic System Reform Research Institute, "The Macroeconomy in the Process of Reform: Distribution and Use of National Income," Jingji Yanjiu (Economic Research), 1987, no. 8:16–28. The data from this article are reproduced and discussed in Barry Naughton, "Macroeconomic Management and System Reform in China," in From Crisis to Crisis: The Chinese State in the Era of Economic Reform , ed. Gordon White (London: Macmillan, 1990).

While budgetary investment dropped further, spending on central-government projects climbed steadily; by 1987 it had surpassed 10 percent of NMP (net material product), regaining the 1978 level. In order to fund this substantial investment program, the central government, by 1987, had to borrow funds equal to 4 percent of national income. Thus, while the central government surrendered direct control over economic resources, it did not reduce its aspirations in a corresponding fashion. By the second half of the 1980s a significant disparity had developed between what the central government wished to accomplish and the resources at its disposal. The central plan had been reborn, but whereas before reform the central government had directly disposed of the resources to carry out this plan, it now had to devise additional mechanisms to draw resources into planned projects.

Central-government planners evolved three complementary strategies to attain their investment objectives: they concentrated their own resources on priority sectors; they harnessed the financial resources of the banking system to those priorities; and they drew on local financial resources through a version of "matching funds." While the first of these strategies was merely a rationalized version of the old planned system, the other two involved the central government in bargaining exercises with local officials and enterprises. Each of these strategies will be examined in detail.

The most important single component of the central government's concentration of its own resources has been the creation of a special program of priority projects. A gradually increasing number of projects had been planned with "rational time-tables and guaranteed supply of materials." These projects have first claim on materials still under state control. This is a kind of plan within the plan, and its significance can be seen from the figures in table 9.1. The priority-investment program has grown steadily, both in absolute terms and as a proportion of national income. Moreover, all these projects have access to in-plan materials provided at subsidized prices, so that real resources are concentrated on the priority program to an even greater extent than the financial data indicate. In essence, the central government is subordinating the surviving elements of the material-allocation system to its development strategy. Before reforms the material-allocation system was charged with the nearly impossible task of delivering resources to all state-run factories for all production needs. But as reform has deepened and the economy has diversified, the allocation system, now much smaller relative to the economy as a whole, has increasingly been targeted to this investment program. The central government no longer has nominal control over all the materials in the economy; but the remaining control over materials is now used almost exclusively to carry out central-government priorities.

The priority-investment program also differs from the old central plan in the quality of the design and planning activity. Unlike the Baoshan steel mill, symbolic of the low quality of planning prior to reform, today's priority projects are carried out with reasonable preparatory work, feasibility studies, and complete sets of design documents. Indeed, a number of the projects utilize international funds, such as those from the Japanese Development Bank and the World Bank, and thus have to comply with international standards for project appraisal and implementation, including competitive bidding for some parts of the work. The central plan has thus been strengthened in a technocratic sense; with a better understanding of the economy as a whole and better utilization of trained manpower, the economic returns of these projects will generally be higher than those of projects during the 1970s before reform. This is true both because individual projects are better designed and because the program as a whole is targeted more effectively on bottleneck sectors. At the same time, priority status and access to materials serves to ensure the completion of the projects. In 1987 the capital-construction expenditure plan for these projects was 109 percent fulfilled, while the remainder of the plan was only 90.5 percent completed.^[13] Although this is an improved version of the central plan, relative to what China had before, it still suffers from the inherent liabilities of central planning, as was discussed in the preceding section.

In order to carry out the broader central-government plan (including the priority plan as one component), the government still needs access to additional financial resources. As central plans grew, bank lending was harnessed to the needs of this plan. By 1985, at the latest, this had worked a fundamental change in the composition of both central-government investment and the use of bank loans. Figures are presented in table 9.2 on the proportion of different types of capital-construction investment going to three bottleneck sectors—energy, transport, and heavy raw-materials

[13] Zhongguo Jiben Jianshe (China Capital Construction), 1988, no. 2:8–9.

industries—in 1985. Table 9.2 shows the striking difference between the composition of central and local-level investment and confirms the reluctance of local governments and enterprises to invest their own money in bottleneck sectors. According to the same source, 80 percent of electricity investment in the past few years has been on central-government projects. Even more striking is the extent to which bank lending conforms to central-government priorities. Originally introduced to permit flexible, decentralized financing of consumer goods industries, bank lending for fixed investment has increasingly been reshaped to serve central-government investment projects. While 68 percent of fixed-investment lending went to light industry in 1979–82, 62 percent went for a subcategory of heavy industry and transportation in 1985.^[14] Given the division of responsibility over different sectors, this means that the central government effectively preempts most fixed-investment lending, and local governments and enterprises have correspondingly fewer financial resources. The banking system is unable to serve as an independent, decentralized funding source, for it is squeezed between the demands of the central government and local officials. Banks are obligated first to fund central-government projects; subsequently, they are placed under enormous political pressure to put remaining resources in the service of projects favored by local governments.

The second source of additional financing is the retained funds of

[14] There are no direct data on the proportion of bank loans going to central projects. (Cross-tabulations of investment categories are not generally published.) Also, the comparison is not precise, since the 1979–82 figures include all fixed-investment lending, whereas the 1985 figure includes only capital construction, excluding technical-transformation investment. This biases the comparison somewhat, but I have been unable to locate additional data on the sectoral composition of technical-transformation lending.

localities and enterprises. Increasingly, the central government requires localities to contribute "matching funds" in order to see critical infrastructure projects in their regions. For example, while the projects in electricity generation are overwhelmingly central, the money for them comes to a significant degree from the localities. Through the end of 1984, twenty-six large-scale electricity-generation projects had been carried out in the central plan, but using the joint resources of central and local authorities; the localities had provided 3.5 billion yuan, or 51 percent of the total cost.^[15] Thus, the central government resolves the disparity between its limited financial resources and its responsibility for crucial sectors of the economy by negotiating for additional resources project by project. In this negotiation the central government is quite powerful: it possesses the design resources and seed money needed for large-scale energy development, and it possesses enormous leverage over the economic system as a whole, including the material-allocation and banking systems. Yet it also requires the cooperation and financial resources of local governments and enterprises. Thus, the Center and local governments are now involved in a classic bargaining situation: each has something the other needs. It is in the interests of both parties to get together, and it is in the interests of each to shape the resulting bargain to their own advantage.

One example of this bargaining process is presented by the province of Shandong.^[16] Shandong is a large, slightly above-average coastal province, which has grown rapidly in recent years and has also experienced significant energy shortages. It is estimated that energy supply is 20 percent below demand. Shandong is the site of 11 of the 190 central-government priority projects, and it accounted for 8.5 percent of 1986 national priority-investment expenditure. Shandong thus receives substantial central-government support, and central projects include two large power plants, a very large ethylene plant, and two ports. Shandong's development strategy is therefore inextricably bound up with its relations with the central government.

To obtain central-government investment projects, Shandong makes—and publicizes—major contributions to those projects. Shandong provides a significant portion of the money for these projects, and must also organize land requisition and purchase; supply of water, electricity, and transport; supply of local building materials (cement, bricks, stone, and sand); and design and construction services. Thus, a substantial part of

[15] 1985 Zhongguo Baike Nianjian (China Encyclopedic Yearbook), 226.

[16] The following discussion is based on the Sixth Five-Year Plan for Shandong, Dazhong Ribao , 2 May 1983, and a series of articles in Zhongguo Jiben Jianshe (China Capital Construction), 1986, no. 9:22; 1986, no. 12:35; 1987, no. 1:18–19; 1987, no. 2:10–11; 1987, no. 5:18–30; 1988, no. 1:12.

Shandong's investment policy consists of coordination and support for central-government investments. An important reason for this support is to demonstrate to the central government that the locality is making a contribution, thus ensuring a future flow of central resources to Shandong. A significant public relations effort goes on to reassure the central government that its money is well spent. Yet these contributions are also a substantial burden to Shandong. The central-government financial contribution to a project covers only a portion of total cost and is generally fixed at the beginning of the plan year, leaving the locality to deal with the frequent cost overruns, while quantitative controls on investment strictly limit the province's total investment. Moreover, the province has to pay taxes on its investment spending—even when that spending goes to central-government projects—whereas the central government itself is exempt from construction taxes.

The province's strategy is to publicize all contributions to central-government projects while seeking ways to minimize the actual burden of those projects. Shandong authorities argue for tax exemptions and additional investment authority in order to carry the burden of central-government projects; at the same time, their own projects have been further "decentralized," placed under the nominal control of rural collectives so that they disappear from the provincial investment quotas. Every investment decision is thus shaped by the desire to draw in the largest possible amount of central resources (including centrally approved bank loans) while simultaneously protecting local resources as much as possible. Compared with the prereform system, Shandong is much less a passive agent of central-government plans; yet its more active role is shaped by obligations as much as opportunities and is still dominated by the need to draw resources from the central government. Shandong follows a particular strategy of high-visibility support for the central government, which is consistent with its generous endowment of central projects. Other localities follow a low-profile strategy of quietly draining resources from central projects, hoping that, ultimately, the Center's commitment to those projects will insure their completion. In this way, local projects that have little protection in case of policy changes can be completed as quickly as possible.^[17] The bargaining relation between Center and locality is here in full flower.

Local governments find themselves squeezed between the Center and their enterprises. On the one hand, they must bargain with the Center to maximize central-government investments in their territory; on the other hand, they seek to retain as much as possible in the way of re-

[17] This strategy is described in Huang Kai, "Macro-control of the Scale of Fixed Investment," Zhongguo Jinrong (China Finance), 1986, no. 2:60.

sources and finances. They must promulgate their own development strategies, but they typically find themselves highly constrained by lack of expertise and experience. The obvious source of additional resources is the locality's own enterprises, which it can tap for a range of voluntary and involuntary contributions. But localities must strike a balance between drawing the resources they need from enterprises and allowing them sufficient resources for growth. This generally ends with local governments assuming a paternalistic, somewhat benevolent attitude toward all their enterprises, protecting them from the vicissitudes of the marketplace and encouraging their development, without regard for any particular development strategy. The relation between enterprises and their superiors will be discussed further below (see also chapter 11 in this volume), but it should be apparent that the difficult position in which localities find themselves will be reflected in their relations with their subordinate enterprises.

By following the three strategies described, the central government has managed to engineer a rebound in energy and transportation investment. Such investment was at a peak in 1978 before the initiation of reform. Energy investment as a proportion of total state fixed investment reached a low point in 1982 and since then has climbed to approximately the 1978 proportion (22–23 percent of investment); transportation investment has displayed a similar pattern, declining until 1981 and recovering since to 13–14 percent of investment. Measured both by the proportion of national income going to central-government projects and by the proportion of total investment going to central-priority sectors, central direction of investment appears as strong now as it was in 1978 before reforms began.

From the aggregate numbers, it would appear that the central government has simply been scrambling to get back to where it was in 1978. In fact, its economic role has been strengthened by access to three types of superior information. The Center's coordination of technical information has improved, so that individual projects are better. Much of the energy investment in 1978 was actually wasted; this is particularly clear in petroleum and is evident in other central-priority sectors as well, such as steel. Second, the central government has utilized better information about the economy as a whole, including long-range forecasting, to concentrate on areas where the case for central coordination is stronger. Generally leaving smaller-scale, consumer-oriented production to the localities and enterprises, the Center has focused on large-scale infrastructure projects, which often span provincial boundaries. These two factors together have caused a gradual improvement in the provision of energy and transport services; while these sectors have remained bottlenecks, industrial growth has nevertheless accelerated, and the shortages persist relative to a much

larger volume of output. Finally, the Center's increased information about local-government activities permits it to engage in specific bargains with localities about individual projects. It is this last type of information that ultimately has allowed the Center to subordinate the banking and material-allocation systems to its objectives, using the resources in those systems as bargaining chips to shape local behavior.

Nevertheless, these factors must always be seen in the context of the decline in direct central control over resources. Central planners are arguably stronger, and certainly more capable, than before reform, but they must deal with a vastly more complex economic environment in which many different agents have control over resources. This environment constantly threatens to overwhelm central-government actions. Although central actions have increased the investment share of priority sectors, the results still fall quite a bit short of central-government objectives. The Sixth Five-Year Plan, covering 1981–85 but drawn up only in 1982, called for the completion of 400 large and medium-sized investment projects, but in fact only 235 (59 percent) were completed by the end of 1985. In 1987, 63 out of 74 planned large projects (85 percent) were completed on schedule.^[18] The sustained central-government focus on energy and transportation has just barely offset the bias toward light-industry investment created by decentralization of resources.

Examination of the investment process thus shows a complex set of changes. Although the central government's direct control over investment resources has been unambiguously reduced, it has been able to use its authority over the economic system as a whole, in combination with substantially enhanced information about the economy, to increase its indirect control over investment. But the nature of the central government's "indirect control" instruments is still highly imperfect: rather than manipulating objective economic levers to control the market environment and thus shape lower-level decisions, central planners instead achieve indirect control by engaging with lower levels in a case-by-case bargaining process. In an immediate sense, this works: it has increased the flow of investment resources into bottleneck sectors. Yet this strategy is clearly a second-best alternative to more fundamental reforms, including increases in the relative price of energy and transport. The central government's expedient policies draw local governments into further complex bargaining relations with the central government, rather than confronting them with more rational costs and opportunities for their own investments. A modest recentralization of finances, bringing financial capabilities in line with central-government ambitions, combined

[18] Hu Changlin, "When Will the System of Allocating Blocks of Investment and Lending to Functional Departments Be Reformed?" Zhongguo Jiben Jianshe (China Capital Construction), 1986, no. 12:7–8; 1988, no. 3:41–42.

with price rationalization and greater autonomy for those controlling decentralized finances, would be preferable. In this way, energy shortages could be addressed without enmeshing all parties in an overly complex bargaining relationship.

More generally, the strategy that has been followed does not seem capable of resolving the fundamental problems of the command-bureaucratic economy. While the Center has better information about local activities and can bargain with localities about a relatively small number of large-scale projects, it still lacks detailed knowledge and control of the bulk of economic decisions. In a sense, the central government can achieve certain objectives precisely because those objectives are circumscribed in scope, while the economy as a whole is just as resistant to specific manipulation by administrative means as it always has been. The new bargaining relations can only compensate for some of the deterioration in authority relations. Moreover, as the central government must cover a deficit in its investment program equal to 4 percent of national income, it competes for savings and pressures the banking system to create credit at an excessive rate. The central government is powerful but needy, and this tends to destabilize the system.

Enterprises And The Two-Track System

At the opposite end of the hierarchy from the central government, enterprises are the building blocks of the industrial economy. A few basic principles have guided enterprise reforms from the beginning. Reformers held that enterprises should cease to be administrative subdivisions of the government bureaucracy and should instead become economic entities making independent decisions based on price and profitability. Enterprises would become "relatively autonomous commodity producers," responsible for their own profits and losses, and economic—rather than administrative—means would be used to accomplish economic objectives. Most reformers envisage a continuing activist role for government, but that role is carried out through the manipulation of "economic levers," such as interest rates and prices.^[19] Evidently, such principles require a reduction in direct government interference with enterprise decision making and imply a limited and indirect role for government that requires substantial sophistication.

These principles are too broad to serve as a specific blueprint for the reform process. The concrete strategy that has dominated the Chinese reform process has been that of allowing a gradual expansion of markets

[19] All these principles can be found in the basic urban reform charter, passed by the Central Committee in October 1984. "Zhonggong zhongyang guanyu jingji tizhi gaige de jueding," (20 October 1984) in Zhongguo Jingji Nianjian 1985 , I/1–I/13.

outside the confines of the planned economy. The planned economy survives, but its size has been held roughly constant, while the regular growth of the economy has steadily swelled the proportion of economic activity carried on outside the plan through the market or marketlike exchanges. The leadership hoped that this evolutionary process would result in the gradual marketization of the economy, while the maintenance of a planned sector would anchor the system during the period when leaders were learning to use "economic levers" and indirect market regulation. This reform strategy has often been referred to as a "two-track" system combining both traditional planning and market operations. One of the essential features of this strategy—and one that gives the Chinese system its great novelty—has been that the two-track system is applied not only to the economy as a whole but also to each individual enterprise. Thus, nearly every state-owned factory operates with a portion of its output planned by government and a portion produced according to market demand.

At its best, the two-track-system strategy of reform held the promise of introducing market forces into the state-run industrial system at a rapid pace. Rather than waiting years for a comprehensive rationalization of prices and taxes, enterprises would be confronted with the opportunity of operating in the marketplace immediately. The command economy would persist, but its scope would be strictly limited. As a result, the monomorphism that made all decision making subject to the command relationship would be broken, since enterprise decisions about the growing portion of the economy would be made on the basis of profitability considerations determined by market prices. At the same time, the growth of a market sector would shatter the monolithic state economy; state-run enterprises operating out-of-plan would compete with collective and private enterprises, and markets would provide information and competitive pressures that could be used to reshape the state economy. Alternately stated, if plan targets were frozen, enterprises would face market prices on the margin, and the enterprise's plan would serve as a lump-sum tax, having no effect on current operations. The administrative economy would thus be gradually dissolved by market forces, and the bargaining economy progressively replaced by the objective rule of the market.

In this framework it is essential that the planned "track" be frozen. If the size of the planned economy is not fixed, then enterprise decision making cannot be fully freed to depend on the market. If the enterprise expects that its market-oriented behavior will have repercussions on the level of planned targets or inputs, it will inevitably take those repercussions into account in its decision making. As a result, advocates of the two-track system have tended to oppose using the old command-

economy structure to solve problems, even to the extent of opposing price reforms within the administrative economy. This position has been bolstered by a perception of the bureaucratic economy as a clumsy and unresponsive system, evident in a recent article by advocates of the two-track system:

The debate over whether it is better to adjust planned prices or to simply decontrol stems from different assessments of the ability of the government. Those who advocate price adjustment believe strongly in the ability of the government to control the price reform process ... [while we believe that] there is a serious contradiction between highly centralized price adjustment and the decentralized structure of interest groups. The central government will never be able to calculate the impact of each individual price change as accurately as the localities and ... enterprises, and the ultimate burden of price adjustment will inevitably be shifted onto the state treasury.^[20]

In other words, advocates of the two-track system felt that the flows of information in the command economy were too crude and distorted to be useful in a rationalization of economic relations. A bureaucracy with weak information-gathering capabilities and correspondingly limited management abilities had no choice but to rely on the expansion of markets outside the system to realize its ideals of reformed enterprise behavior. The strategy of freezing a clumsy bureaucracy—rather than trying to rationalize it—implied that the financial and authority relations linking enterprises to their superiors would persist, though, one would hope, in fossilized form.

Unfortunately, this strategy has not yet been adequate to make the enterprise-reform principles into reality. It has been impossible, first of all, to freeze the planned "track" of the economy. The overall size of the command economy (the scale of central-government production and allocation plans) has indeed stayed relatively constant, declining steadily as a proportion of total economic activity as the economy grows. Yet from the perspective of an individual enterprise, plans are anything but stable. Some enterprises experience shrinking plans and are uncertain about the speed of shrinking. Other enterprises discover that their plan is increasing, and they may welcome this when it promises access to cheap inputs. For example, when the foreign joint-venture automobile companies, such as Beijing Jeep and Guangzhou Peugeot, ran into problems with raw-material supplies, the government provided them a special benefit: incorporation into the plan. The overall constancy of the plan thus conceals considerable variation in plan targets at the enterprise level.

[20] Hua Sheng, Zhang Xuejun, and Luo Xiaopeng, "Ten Years of Reform in China: Retrospective, Rethinking, and Prospects," Jingji Yanjiu , 1988, no. 9:16.

Moreover, in the traditional command economy, input supplies are subordinate to the production plan, so that only the level of the production plan itself is important. However, in China, as bureaucratic capabilities declined during the Cultural Revolution, and as the system became less regularized during the reform process, there began to be substantial discrepancies between production plans and the allocation of inputs. Enterprises today cannot necessarily get the full quota of inputs required to fulfill their compulsory plan. Thus, even for a given production plan, enterprises face differential supplies of low-price inputs, and ultimately supply depends upon the nature of enterprise relations with their superiors. Here, paradoxically, the deterioration of bureaucratic capabilities makes it harder to freeze the bureaucratic portion of the economy, because the bureaucratic system is incapable of generating a single parameter to represent the level at which the plan should be fixed. Only enterprises that have obtained administrative recognition of their priority status can hope to receive a full complement of subsidized inputs.^[21] The individual enterprise cannot regard the plan as being fixed in any meaningful sense. A persuasive case to a superior might always make a difference; "plan bargaining" persists in the two-track system.

Outside the plan, the environment in which enterprises operate is still not a pure market environment. Indeed, in general, enterprises do not sell their outside-plan output at market-clearing prices. In most cases, outside-plan output is sold at a higher price than planned output, but the price is kept low enough that shortages continue to arise and the enterprise must decide to whom to deliver its products.^[22] This is a very curious phenomenon that requires explanation: Why do enterprises not exploit their apparent ability to charge higher prices? There seem to be three mutually reinforcing causes of this type of enterprise behavior. First is the government effort to restrain price increases in the inflationary environment that has prevailed since 1984. The old bureaucratic networks persist and are used to pressure enterprises to refrain from raising prices; in this sense, the failure to charge market-clearing prices is simply a result of the failure to free the enterprise from the old authority relations. Such pressures vary greatly from region to region and from sector to sector, depending upon the bureaucracy and type of good involved—producers of certain consumer goods are subject to more

[21] This generalization is based on a sample of twenty-two enterprises visited in the summer of 1988. Individual enterprises reported receiving anywhere between 20 percent and 100 percent of the main inputs required for in-plan production.

[22] Of twenty-two enterprises visited (see n. 21), each was selling some outside-plan output at a higher price. However, only two were selling at a price that equated supply and demand, and all the others practiced some type of informal rationing.

pressure, as are factories in "conservative" provinces with active price-control efforts.

The second factor is that enterprises find it in their own interest not to charge high, market-clearing prices for their output. When enterprises sell their output at the highest possible price, they receive all the benefit from the sale in overt, monetary form: high profits. Profits are highly visible, and they are highly taxed. As a result, although the enterprise may reap greater paper profits from such a sale, it may find itself with little real benefit after delivering taxes and other revenues to its superiors. When the enterprise chooses to sell its output at a lower price, it forgoes some money income but receives something else in return. Most commonly, it receives access to other goods at a concessionary price. If these are consumption goods, the enterprise benefits directly: the goods can be sold to the enterprise's workers without any overt subsidization and without restrictions. It appears to be a pure market transaction and is thus subject to no taxes and no quantitative limits. (If the enterprise had taken the income in money form and then distributed it to its workers, that distribution would have been subject to limitations on bonuses.) Furthermore, the enterprise creates a hedge against future supply and price uncertainties by creating a long-term cooperative relationship with another enterprise. Current income forgone is "saved" by accumulating capital in the form of guanxi (connections). By building up guanxi the enterprise obtains greater flexibility and security: guanxi can be used to shift income from one period to another, or to guarantee supplies of materials when the need is critical. For the enterprise to forgo guanxi for some short-run financial advantage would be shortsighted indeed.

Finally, the benefits received by forgoing enterprise profits can accrue to individuals, particularly managerial and sales personnel. This can take the form of a slightly lavish dinner and a carton of cigarettes, or it can involve major bribery. The danger of corruption is, of course, endemic to a two-track system, because of the existence of more than one price for any given commodity. Overt bribery and illicit sale of in-plan goods, however, can be extremely dangerous and seem to be relatively infrequent. A far more common pattern is the creation of long chains of buyers and sellers, each of whom raises the price of a good by, say, 10 percent, just enough to cover "handling charges." The individual privileged enough to be one of the links in this chain earns a moderate cash reward and acquires the gratitude of the individual who is the next link. No unambiguously illegal activity has taken place, and everyone involved has profited, except the original enterprise. Only at the end of this long chain does the good sell for something like a true market price.

For these three reasons, enterprises and their superiors collaborate in

complex exchanges at other than market prices. On occasion, an enterprise may dispose of goods at market prices when it has no "connection" with the purchaser, and logically even the longest chain of buyers and sellers has a final link, so market prices exist. But it is extremely difficult to determine what those prices are, and difficult to object when transactions take place at other prices. Thus, there is a kind of "exchange bargaining" that goes on between enterprises and also affects enterprise relations with their superiors. Enterprise managers try to maximize a complex mixture of sales revenues and outside-sales benefits, and this leads them into elaborate bargaining arrangements.

When all the enterprise's supplies have been purchased, and all the output has been sold at various prices, the enterprise's ultimate profit is still far from determined. The final bargain that must be struck is that which determines the financial relations that link enterprises to their superiors. These have remained unregularized and subject to a bewildering array of inconsistent provisions. In part, this is due to the failure to successfully implement new fiscal systems, such as the "tax for profit" system and the consequent prevalence of the "contracting" (chengbao ) system. Under the contracting system, enterprise financial obligations to superiors are determined as part of a multiyear contract negotiated between the two sides. Such a system is designed to provide high-powered incentives to enterprises by allowing them to retain a high percentage of their incremental revenues. At the same time, the contracting system indicates the inherent difficulty of the two-track system: the multiyear contract is an attempt to "freeze" the financial tribute the enterprise pays annually to its superiors. The fact that such a freeze has to be specified in a negotiated contract between the enterprise and its superior simply demonstrates that the planned track of the economy was not frozen to begin with. Quite the contrary, the financial relations between enterprises and superiors have been subject to constant, virtually annual, changes between 1978 and 1988. Unavoidably, each time financial provisions are altered, every aspect of the enterprise's economic health, behavior, and relations with its superiors enters into the process of adjustment. The striking of a deal dividing up enterprise revenues, which Kornai has described in Hungary and labeled "redistributional bargaining,"^[23] is prevalent in China as well.

Redistributional bargaining in China is peculiarly unconstrained because of the absence of clear fiscal and financial regulations. Almost any parameter can be altered through negotiation between superiors and subordinates. For example, there is uncertainty as to whether bank loans

[23] Janos Kornai, "The Hungarian Reform Process," Journal of Economic Literature , 24 (December 1986): 1691–1701.

should be repaid before or after taxes, so the authority to repay before taxes thus becomes a benefit that superiors grant enterprises. Actually, the discretion available to superiors is virtually unbounded; at one plant visited in 1988 the superiors had decided that the enterprise was having legitimate difficulties in fulfilling its profit-remittance contract because it had such a large volume of loans to repay. The superiors simply allowed the enterprise to count 50 percent of the loans it repaid as remitted profits for the purpose of calculating the revenue split, so that the enterprise succeeded in fulfilling its contract. In this case, the contract fulfillment was entirely imaginary. Taxes are slightly more difficult to alter, since the central government has repeatedly insisted that taxes, particularly those on cigarettes and liquor, should not be forgiven. Nevertheless, tax forgiveness is very significant. In Jiangsu in 1986 local authorities forgave taxes equal to a remarkable 12 percent of total budgetary revenue.^[24] Every financial parameter is subject to negotiation and thus subject to constant change.

An additional aspect of redistributional bargaining links it to my earlier discussion of the central government and its investment plan. I noted above that, before reform, enterprises retained some funds for investment purposes, but that the central government collected no data on this investment and had no way of controlling it. Local governments and enterprises, in collaboration, were thus completely autonomous users of these funds. In the current period, enterprises control substantially more funds, but the central government no longer closes its eyes to their use. Instead, the government by the mid-1980s had promulgated a plan for enterprise-level investment in technical transformation. This plan, which includes projects for about half of all large enterprises, classifies enterprises and projects according to level of priority and establishes technologies and processes that are to be introduced.^[25] This is not a detailed central investment plan, like the priority capital-construction program described above. However, it is the basis for central-government involvement in redistributional bargaining. Enterprises with authorized and high-priority investments are supposed to retain more profit and have easier access to bank loans than other enterprises. Thus, the central government uses this planning exercise to reach down the administrative hierarchy and shape the bargain between the enterprise and its direct superior in ways that reflect central-government priorities. Clearly, enterprises can benefit by appealing to patrons at the central level.

[24] Tian Yuan, "Price Reform and the Transference of Property Rights," Jingji Yanjiu , 1988, no. 2:15.

[25] "Strengthen the Long-range Planned Categorization of Large Enterprise Technical Transformation," Zhongguo Jiben Jianshe (China Capital Construction), 1985, no. 7:4–7.

The enterprise is thus enmeshed in a complex bargaining relationship with its superiors. Plan bargaining, exchange bargaining, and redistributional bargaining are all taking place simultaneously; in fact, along with the face-to-face personal relationships involved, they are woven into a single inextricable chain of bargains—repeatedly struck and constantly reopened—between enterprises and their superiors. No wonder, since both enterprises and superiors face a vast realm of indeterminacy, in which everything—price, plan, supply, tax, credit—is subject to change and negotiation. All these parameters can be altered by the enterprise's superior, and only an extraordinarily bold, or extraordinarily foolish, enterprise manager would choose to operate "on the market" as if the wishes of his superior did not matter. Conversely, only an exceptionally obtuse manager would ignore the opportunity to improve his lot and protect himself against adverse outcomes that is possible by the appeal to his superiors. The astute enterprise manager will always keep open the channels to his superiors that permit him to reopen the bargaining relation at any time.

What is true at the enterprise level is also true one step higher in the bureaucratic hierarchy. Local government officials must always be aware of the possibilities involved in currying favor with their superiors. Indeed, since there is no final blueprint for reforms, localities must be given the freedom to experiment and must be supported in this experimentation process. Localities learn very quickly that it is in their interest to have some kind of reform pilot project.^[26] Localities receive special benefits to operate pilot projects, and if one is deemed successful, local officials will undoubtedly receive a career boost. As a result, reform "experiments" proliferate; at the end of 1986 there were seventy-two cities that were "comprehensive-reform pilot cities," sixteen "medium-size city administrative-reform pilots," twenty-seven banking-reform pilots, thirteen producers-goods-marketing pilots, fourteen housing-reform pilots, plus labor-reform pilots, coastal development zones, and so forth.^[27] In each of these "experimental cities" local authorities could hope to obtain benefits and preferential treatment from the central government by extending benefits and preferential treatment to their subordinate enterprises. Thus, just as local governments were squeezed by the central government in the investment-allocation process, so do they squeeze themselves into the system-reform process. The reform process, and the real distribution of benefits within the state sector, evolve from this constant contention and interplay of interests.

[26] An expression has arisen: "Whoever has a pilot project, benefits" (shei shidian, shei deyi ).

[27] Zhongguo Baike Nianjian 1987 , 201–2.

Bilateral Monopoly

By itself, the indeterminacy of so many economic parameters cannot explain the persistence of bargaining relations at the enterprise level. Indeed, given the rise of marketlike institutions in China during the past ten years, the question is why market forces do not in the long run eliminate the indeterminacy. Ultimately, the crucial factor is that state enterprises are in a relationship of bilateral monopoly with their superiors. The two-track reform strategy implies the retention of the administrative hierarchy, and because of the retention of the administrative hierarchy, enterprises cannot escape the influence of their superiors, while superiors cannot escape their ultimate reliance on the productive capacity of their enterprises. The importance of the retention of the hierarchy is thus not primarily that the enterprise remains subject to arbitrary commands from its superior (though this is sometimes a factor). Rather, the importance lies in the fact that the enterprise and its superior are forced to make a deal with each other. This is a special case of the general problem of small numbers exchange: one buyer confronts one seller and there are no competitive forces to drive the two into competitive equilibrium. The transaction takes place, given there are gains to be realized, but the benefits are divided between the two parties according to their relative bargaining power.

Within the state sector relations between the enterprise and its superior body are always situations of bilateral monopoly. Within any given urban area there is always a sector-specific bureaucracy that has responsibility for a certain type of production. For example, production of machine tools in Taiyuan will be subject to the Taiyuan machinery bureau, or to a similar provincial or national body. That bureau has the authority to permit or prevent production of machine tools by anybody in the city of Taiyuan, and this type of authority is universal and entirely formalized. It is impossible that "nobody" could be in charge of a certain kind of production in a certain locality; every locality and every type of production is automatically established as a monopoly.

Two particular aspects of the Chinese economic system affect the way these local-sectoral monopolies are managed, but they do not change the basic monopoly condition. First, it is a curiosity of the relatively decentralized Chinese system that industrial production in rural areas is controlled, not by sectoral bureaucracies of broad geographic scope, but by local bureaucracies with authority over a broad range of industrial sectors. Since there are so many local bureaucracies (more than 2,000 county-level governments), urban producers cannot be completely protected from competition arising in a nearby rural area. Indeed, markets

in which rural producers play a role are the most competitive markets that exist in China. In most cases, though, rural industries have limited technological capabilities and produce goods of poor quality. Although this is gradually changing, competition from rural factories does not yet significantly change the market position of most modern urban factories, particularly in an economic environment characterized by excess demand and inflationary pressures.^[28] Urban producers of a given product would be far more threatened by the threat that another urban factory would cross sectoral boundaries and produce a competing product, and it is precisely this danger from which they are protected by the monopoly power their superiors exercise over their own product lines.

Second, the enterprise may be subject to a number of different supervisory bodies that make its task more difficult. For instance, the municipal labor bureaus, tax offices, and banks may intrude from time to time on enterprise decision making. Even worse, during certain periods the enterprise may be subject to "dual leadership" in which authority is shared by local and national authorities. These peculiarities make the job of running a factory much more difficult, but they do not fundamentally alter the enterprise's position. Except for brief periods following on administrative reorganization, an enterprise always has an immediate superior (zhuguan bumen ), which is clearly identified. The fact that individual decisions must be made taking into consideration the opinions of various "related administrative agencies" (youguan bumen ) makes the deal-making process more complex and onerous. However, this does not change the basic situation, which is that the enterprise must procure the assent of its superior organ to every significant decision, while additional approvals may still be required for some of these decisions. Quite frequently the enterprise's superior organ negotiates with the related administrative agencies in place of the enterprise. In all cases the one permanent and inescapable relationship every enterprise has is the one with its superior organ.

Conversely, the enterprise possesses significant task-specific capital, which cannot be easily exploited by other parties. The superiors cannot simply dispense with the existing enterprise and its workers and hire somebody else to produce machine tools. Because of China's decentralized management system, most enterprises are subordinate to authorities in the city in which they are located; this means that superior authorities have under their control only a handful of enterprises, or

[28] Competition from rural factories depends on the ability of competitors to transport their products across administrative boundaries. That condition is not always met in China. In the final analysis, though, state enterprises are more often protected by the difficult technological requirements of certain goods, especially heavy industrial products, than by barriers to interregional flows.

even a single enterprise, capable of producing a given product. Because their span of control is narrow, it is difficult for the superior to monitor enterprise performance by comparing it with other enterprises, and difficult to apply severe sanctions to the enterprise because there are few alternate sources of supply. Ultimately, bureaucratic superiors must deal with the existing enterprises, which are their only viable sources of money and output. Bureaucratic superiors in China worry constantly about the possibility that subordinates will restrict output, substituting perfunctory performance (or nonperformance) for hard work and frustrating the objectives of superiors by slowdown and egregiously sloppy work. The relationship between superiors and enterprises is thus the same as that between management and workers within the Chinese enterprise, and for the same reason: top and bottom are locked into a relationship that neither can escape. The ultimate bargaining power of those at the bottom derives from the fact that they can frustrate the deal-making process so that both sides lose. Because those at the top are aware of this power, they strive to keep those on the bottom mollified and part of the bargaining process. This one-on-one bargaining process is thus the inevitable outcome of a basic condition of bilateral monopoly, in which neither side can do without the other.

This state of affairs is perpetuated by the financial and pricing policies that the central government carries out. By maintaining low prices on energy and basic foods; by subsidizing capital (interest rates) and social security and health benefits; and by protecting factories from import competition, the government ensures that the great majority of state factories turn a paper profit regardless of their economic efficiency. The large stream of accounting profits generated in industry—industry accounts for over 80 percent of budgetary revenues—ensures that there will always be a deal to be struck over enterprise revenues. Both superiors and subordinates have a major incentive to stay in the bargaining process so that they can reap a share of these accounting profits. The bilateral monopoly persists because neither party would be likely to gain by breaking up the relationship, even if that were possible. We might say that the monopoly power exists, not to keep the two sides in the bargaining relation, but rather to keep other parties out. Having between them the disposition of the surplus generated in industry, the two sides then bargain over the precise distribution of benefits.

Market forces cannot yet dissolve this monopoly relationship, because there is always some combination of concessions and benefits the superiors can grant that will enable the enterprises to survive even the fiercest competition. State enterprises have large buildings and extensive arrays of equipment for which they pay little; and they have access to substantial financial resources at very low cost. It is not difficult for even a very

inefficient enterprise to "out-compete" his potential rivals from rural industry. No superior will acquiesce in the collapse of a sector under his control, because that will shut him out of the stream of benefits generated by that sector. Thus, both parties remain locked into the hierarchical structure and the bilateral monopoly bargain that goes with it. One consequence is the maintenance of the monomorphic organizational form from the command economy. The internal organization of the enterprise is standardized to match that of its superior organs: functional departments within the enterprise report to—and bargain with—parallel departments in their superior organs, as well as management in their own enterprise. According to a survey of 170 managers of large factories in Liaoning province, the two greatest sources of discontent were the lack of control managers had over personnel decisions, and their inability to alter the organizational structure of their enterprises.^[29] Given this state of affairs, Chinese industry is unable to reap the efficiency gains that come from idiosyncratically specialized organizational forms, or those that come from diversified multiproduct corporations with significant economies of scale.

Finally, the maintenance of the administrative hierarchy and the underdevelopment of outside-of-plan markets has meant that the monolithic character of the state economy has not been fundamentally changed. In the discussion of the central-government role in investment, it was argued that the information available to central planners had been greatly increased. Planners had better technical information for project design and long-range projections, and more timely and complete information about the behavior of the economy as a whole. These generally can be understood as improvements and rationalizations of the information-gathering facilities that normally characterize command-bureaucratic systems. That is, even for a command-bureaucratic system, China had an exceptionally weak information-coordinating capability before reforms, and this weakness has been partially rectified. However, if market-oriented reforms are to work, whole new channels for the circulation of information must be created. These channels operate through the market mechanism itself to provide information to decentralized agents (not just to central planners) about economic opportunities. In this respect, the Chinese reform has not

[29] Liaoning Statistical Bureau, "The Opinions of Managers of Large and Medium Size Enterprises on Several Problems of Reform," Tongji, 1987, no. 3:32–33. These opinions are more striking in that the managers were not anxious to escape from the authority of their supervisors in other respects: 37 percent wanted a larger compulsory plan; 33 percent thought the plan was about right; only 30 percent wanted a smaller plan. On the organizational form of enterprises, see also the discussion in Max Boisot and Xing Guoliang, "The Nature of Managerial Work—Chinese Style," paper presented at the International Conference on Management in China Today, Leuven, Belgium, June 19–21, 1988.

really transformed the state-owned sector. Indeed, the two-track system has in some respects impeded the circulation of information in the economic system. Since enterprises do not regularly sell above-plan output at market prices, it is difficult to tell what market prices are. This—combined with rapidly changing market prices in an inflationary environment—has meant that market prices cannot really serve as readily available "shadow prices" that could be used to assess state enterprise performance. At the same time, each enterprise's capital stock and inputs are purchased at widely varying prices. As a result, an enterprise that purchased its fixed capital at low state-set prices will show a much higher profit rate than one that purchased machinery outside the plan; enterprises dependent on high-priced inputs may show losses while less well run factories in the same sector generate profits because of access to subsidized inputs. These factors are too complex to allow systematic calculations at the higher level to correct for them. For instance, who can say whether textile production is more efficiently carried out in small-scale rural factories or in state urban factories? Certainly the central government has no information at its disposal that would permit a ready answer to such a question.

Because of the weakness of markets, both central planners and local agents continue to base much of their decision making on pseudoinformation. Nominally profitable enterprises are propped up or expanded, without accurate information about the actual comparative productivity of those enterprises. Local monopolies are perpetuated because it is impossible for governments or enterprises to devise alternative strategies based on mutually beneficial specialization and trade. While technical information available to the Center has been enhanced, local information on economic choices has not been qualitatively improved. The two-track system has thus not been able to change the fundamentally hierarchical nature of the system, and this has perpetuated the basic bargaining relations of that system.

Conclusion

The reform strategy of a two-track system has opened up the monolithic command economy somewhat by allowing private and collective enterprises to play an increasingly important role. However, it has not altered the fundamental authority relationship that ties enterprises to their superiors and perpetuates the bilateral monopoly between them. As a result, the monolithic and monomorphic nature of the state-run economy has been preserved to an important extent. Enterprises and their superiors continue to be entangled in complex bargaining relationships. One of the most obvious and immediate consequences is that enterprises can never be

held fully accountable for their operations. Every aspect of enterprise performance is related in some way to the bargains the enterprise has struck; an enterprise therefore always has someone in the administrative hierarchy to share the blame for unfavorable outcomes. Conversely, any unfavorable outcome has a possible remedy somewhere in the bargaining process; some negotiated package of concessions and benefits is always potentially available. The lack of accountability naturally reproduces the "soft budget constraint" characteristic of bureaucratic economies. Pressures to raise efficiency are correspondingly reduced, and destabilizing types of behavior are encouraged.

At the same time, the central government has returned to the investment arena with a renewed program for development of priority areas, but without direct control over the resources needed to implement that program. As a result, the national government increasingly competes with enterprises and local governments for available investment resources. The central-government investment deficit combines with the "investment hunger" created by soft budget constraints at the enterprise level to generate significant macroeconomic pressures. In order to accommodate the demands of central and local governments and enterprises, the banking system is continuously prodded to provide more loans to finance investment. The result is a continuing expansion of total demand in the economy and—since the expansion in demand outpaces the expansion in supply—a serious inflation problem. By the middle of 1988 the economic system and reform strategy were in the midst of a major crisis brought about by steadily accelerating inflation.

In a general sense, this crisis can be attributed to the continuation of the bargaining economy. The central government uses the bureaucratic apparatus to reach for more than it could accomplish with its own resources. The enterprises rely on the bureaucratic apparatus to shield them from losses and to avoid stark pressures to economize. These two forces collide in an escalation of demands that the economy cannot accommodate. The continuation of the bureaucratic economy means that the cost of wrong decisions can still be shifted onto society as a whole, rather than being borne by individual decision-makers. The difference today is that such costs are increasingly visible. During the period of the command economy, the monolithic economic structure was able to hide the effects of poor economic choices; there were few markets to register disequilibria, and poor choices became manifest only very gradually with a recognition that real growth and improvement in living standards were occurring much more slowly than seemed warranted. Today, the disproportion between total demand and total supply shows up quickly with the emergence and acceleration of inflation. This could be an advantage if central planners were able to respond quickly to signs of imbalance. In

any case, though, a successful reform process demands a high degree of responsiveness in each individual market, as the system really starts to reward more efficient producers and cut down the flow of resources to the less efficient. The polymorphous bargaining relations that characterize China's industrial hierarchy today tend to deflect the impact of any positive or negative shocks, thus diffusing the pressure for better performance that those shocks would otherwise create. They also serve to disperse costs so that they are borne by the economy as a whole and show up as inflation. Although the system has been partially opened up, the prevalence of complex bargaining relations within the state-run industrial hierarchy is a symptom of the system's inability thus far to establish new principles of operation and decision making on which to base future development.

Appendix

The basic data sources are Zhongguo Guding Zichan Touzi Tongji Ziliao 1950–1985 (China Fixed Investment Statistical Materials), 59, 64, 218–19; 1988 Tongji Nianjian (Statistical Yearbook), 564, 566, 605; 1988 Tongji Zhaiyao (Statistical Abstract), 67. In the table below, technical transformation includes "other" investment. Column (5) is the sum of columns (1) and (2), divided by net material product , and column (6) is the sum of (3) and (4), again divided by NMP.

Data on budgetary and extrabudgetary investment . The capital-construction figures are consistent and readily available. However, I wish to include foreign borrowing disbursed through the central budget in budgetary investment. This is available through 1985 in the first source cited, and 1985–86 figures are available in

1987 Tongji Nianjian , 467n. The 1985 figure thus given is 4.103 billion, and the 1986 figure is 3.003. The latter figure can be confirmed with reference to 1988 Tongji Nianjian , 564. However, the 1987 figure is not available. Since total state investment funded by foreign capital increased 31.2 percent during 1987, and budgetary foreign borrowing of all kinds increased 36 percent, while foreign direct investment increased 33 percent in renminbi terms (1988 Tongji Nianjian , pp. 565, 732–33, 762), I have assumed that foreign-funded investment channeled through the budget increased by one-third to 4.0 billion.

Figures for technical renovation are more complicated because data on "other" investment are sometimes included and sometimes separated, but separate data are never presented with the breakdown we require. Data are good for technical transformation, not including "other," for 1981–85 in the first source cited. However, the 1980 data are incomplete and seem unreliable. Figures for 1978–84 on financing for the category of technical transformation-including-"other" are available in the earlier Statistical Yearbooks, e.g., 1983, English, 360, and 1985, 452. These are used for those years. These show that budgetary finance of "other" investment was very small but not zero. The composition of "other" investment financing 1985–87 thus is the only remaining financing problem. Overall, the composition of "other" investment changes considerably through 1983, but is quite stable in the 1984–87 period. In 1984 it can be calculated that 1.25 billion, or about 10 percent, of "other" investment was funded from the budget. This proportion has been applied for the years 1985–87. In the calculation of budgetary and nonbudgetary shares of investment some approximation is required, but the possible margin of error is very small.

Data on Investment Included in Central and Local Plans . Data for capital construction are clearly presented in the first source cited. It is less clear what proportion of technical transformation and other investment should be counted as central government. The figures from 1981 through 1987 for technical transformation-excluding-"other" are good, but there are no figures at all for "other." Unfortunately, there is no alternate source where the proportion of technical transformation-plus-"other" that is included in the central plan is given. I have therefore assumed that no "other" investment is central. An alternate assumption would be that petroleum-depletion allowances, which account for a little over 50 percent of "other" in 1984–87, are also "central." This alternate assumption would raise the proportion of total investment included in the central plan in every year, but would have virtually no impact on the trends that are the subject of the analysis.

The other problem is determining the proportions of technical transformation that are incorporated in the central-government plan during 1978–80 (there was no "other" investment during those years: thus, this is implicitly technical transformation-plus-"other"). I simply assumed central projects were equal to 27 percent of technical transformation in those years, as they were in 1981. This is plausible because the proportion of capital construction accounted for by central projects is quite stable during the 1978–80 period, and we know that technical transformation investment at this time was predominantly under local control. The maximum error involved in this procedure would be if we undercounted central-government technical transformation projects by 2–3 bil-

lion in 1978 through 1980, which would amount to at most 1 percent of NMP in 1978, or 1/2 percent of NMP in 1980 (a more likely error, since bank lending for technical transformation starts to become important at that time, and that might fund central-government projects). Thus, there is a margin for error in the 1978–80 figures larger than that involved in the calculation of budgetary shares. Nevertheless, the margin is not sufficient to materially alter the qualitative conclusions presented here.