Two
State Power and Economic Activism during the New Policies, 1068-1085' The Tea and Horse Trade and the "Green Sprouts" Loan Policy

Paul J. Smith

Modern historians of China have identified two long-term movements whose intersection helped shape the economic statecraft of the New Policies. On the one hand, as Twitchett, Skinner, and Feuerwerker among others have observed, from about the eighth century "there was a clear tendency for public and private [sectors of the economy] to be more clearly differentiated, and for the private sector to be enlarged at the expense of the public."^[1] The chief symptom of this trend, which continues into the twentieth century and stands as a defining characteristic of China's late imperial era, was the steady withdrawal of government from direct involvement in the economy, as market activity expanded beyond the regulatory capacity of the premodern state.^[2] On the other hand, for many historians of the middle and late imperial eras the Northern Sung, in particular the statist reform program of the New Policies, marks the high point of direct state involvement in the economy and society, and the culmination of a three-century cycle of economic activism that also began in the eighth century.

How are the long-term disengagement of the state from the economy and

[1] Albert Feuerwerker, "The State and the Economy in Late Imperial China" (unpublished manuscript, 1983), p. 12. See also Denis Twitchett, "Merchant, Trade, and Government in Late T'ang," Asia Major , n.s. 14, no. 1 (1968): 63-95, and G. William Skinner, "Urban Development in Imperial China," in G. Wm. Skinner, ed., The City in Late Imperial China (Stanford: Stanford University Press, 1977), pp. 3-32.

[2] For an analysis of the organizational constraints on an expansion of the state commensurate with the growth of population, see Skinner, "Urban Development," pp. 19-24. As Skinner points out, problems of size, complexity, and span of control restricted government involvement in all areas of local administration, not just commerce.

the three-century cycle of economic activism related? One solution is to view the activist cycle that culminated in the New Policies as a final attempt by a uniquely professionalized bureaucratic elite to keep the burgeoning economy under the control of the state. Here I attempt to support such a view by contrasting two measures that represent the New Policies approach to economic mobilization, and that highlight the ways in which the activist state intervened in the market economy. These are the Szechwan Tea and Horse Trade, which monopolized Szechwanese tea to pay for the purchase of cavalry horses from Tsinghai, and the "Green Sprouts" farming-loan act, which sought to turn the state into the primary provider of rural credit. Comparison of the Green Sprouts and Tea and Horse measures is recommended both by their similarities and by their differences. As I will discuss later in the chapter, both policies reflect a typical characteristic of Sung economic regulation: the creation of regional state enterprises and agencies to participate directly in the market economy; and both measures were generated by a common activist blueprint for reform that authorized field agents to employ state power to compete with private economic actors for control over the returns to commercial activity. But similar policy impulses take different organizational shape depending on the nature of the specific task and the environment within which it is to be performed. In later sections I analyze the organizational structure of the Superintendancy for Tea and Horses (tu-ta t'i-chü ch'a-ma-ssu , hereafter termed the Tea and Horse Agency, or THA) and the circuit-level Ever-Normal Granary intendancies (t'i-chü ch'ang-p'ing-ssu , or ENGs), which administered the Green Sprouts loans, and attempt to explain why the same action blueprint produced a radically decentralized enterprise based on the autonomous, entrepreneurial use of power in the first instance, but a tightly centralized network of dependent agencies in the second. Finally, in my conclusion I speculate about why post-Sung administrations gradually abandoned direct state involvement in the market economy, as shifts in the social basis of official service promoted the view that the state might regulate but could not dominate society.

The Sources of Activist Statecraft

Since the following pages will be devoted to a discussion of the nature and institutional shape of economic activist statecraft during the New Policies, I begin with a working definition of economic activism. By economic activism I mean, first, a propensity to participate in the commercial economy both directly, through monopolies and government enterprises, and indirectly, through commercial taxation; and second, a commitment to the use of state power both to promote economic activity and to increase the government's share of the economy's total resources. With this definition as a starting

point, if we accept as useful the hypothesis that from mid-T'ang on the Chinese economy began to expand more rapidly than the regulatory capacity of the state, how can we explain the synchronous cycle of economic activism that peaked some three centuries later in the New Policies? Many factors converged to produce the activist statecraft of Northern Sung, but no convincing explanation should exclude the interplay between the rise of a bureaucratic elite that benefited from the expansion of government activity and the emergence of strong frontier states that forced China into a perpetual search for new resources with which to subsidize defense.

As Denis Twitchett has shown, even as burgeoning market expansion convinced T'ang finance officials that commerce could no longer be suppressed or adequately controlled, the rebellion of An Lu-shan in 755 left the T'ang court with few sources of revenues apart from commerce. Once the state had lost control over the flow of direct taxes from agriculture, a newly emergent group of specialized financial administrators began systematically to exploit the rapidly growing commercial economy by imposing state monopolies on the sale of salt, tea, and wine, and ad valorem taxes on the transit and sale of commercial goods. By the rise of Northern Sung in the tenth and eleventh centuries, as Robert Hartwell has demonstrated, the T'ang nucleus of financial specialists had developed into a functionally integrated financial service, disproportionately staffed by members and associates of a professional bureaucratic elite that specialized in and drew a large part of its income from government service, and that was oriented to the use of state power to regulate and tax commerce and industry.^[3]

Through its encouragement of trade and industry, expansion of the money supply, dissemination of new knowledge and techniques, and pursuit of predictable and consistent economic policy, the Sung financial administration initially helped to foster China's spectacular economic transformation. At the same time, Sung specialists, building on the foundation laid by their T'ang predecessors, created the elaborate bureaucratic apparatus needed to tax and intervene in expanding commercial activities. In the course of the eleventh century they built up a field administration of some 2,000 commercial-tax stations and another 3,000 government economic in-

[3] For Twitchett's account, in addition to "Merchant, Trade, and Government," see also "The Salt Administrators after the Rebellion of An Lu-shan," Asia Major , n.s. 4, no. 1 (1954): 60-89; Financial Administration under the T'ang Dynasty (Cambridge: Cambridge University Press, 1963), pp. 49-65, 97-123; and "The Composition of the T'ang Ruling Class: New Evidence From Tunhuang," in Denis Twitchett and Arthur Wright, eds., Perspectives on the T'ang (New Haven: Yale University Press, 1973), pp. 47-86. For Hartwell's analysis of the composition of the Sung elite and its relationship to policy formation, see "Demographic, Political, and Social Transformations of China, 750-1550," Harvard Journal of Asiatic Studies 42, no. 2 (1982): 405-26. On the rise and nature of the Northern Sung professional financial service, see Robert M. Hartwell, "Financial Expertise, Examinations, and the Formulation of Economic Policy in Northern Sung China," Journal of Asian Studies 30, no. 2 (1971): 281-314.

stallations (ch'ang-wu ) to collect the empire's taxes on trade and manage its salt, wine, and tea monopolies and other fiscal enterprises.^[4]

Because of its widespread fiscal apparatus, in Northern Sung the government was able to draw more of its revenues from the dynamic commercial sector of the economy and to draw a larger share of the economy's total output than at any other time through the late nineteenth century. According to the financial expert Chang Fang-p'ing, for example, combined receipts from the commercial sector amounted, at 15 million strings of cash, to about 23 percent of the government's total income of 65.6 million units at the turn of the eleventh century, rising to 36 percent of its total income of 126 million units in the late 1040s. By 1076 combined commercial receipts came to 50 million strings, roughly equal to the value of the direct-tax (liang-shui ) quota of 52 million units, and so just under 50 percent of the government's total income. After the Sung, commercial taxes did not again loom so large in the revenues of the Chinese state until the early twentieth century, when Ch'ing finances came under foreign supervision. At the same time, according to estimates by Albert Feuerwerker and others, it appears that the Sung state was able to take in almost twice as much of its economy's total output as either the Ming or the Ch'ing state.^[5]

As long as the state was able to use its tax revenues to provide services that stimulated investment, Sung economic activism facilitated economic

[4] Peter Golas provides a clear description of Sung fiscal organizations in his chapter "The Sung Financial Administration," esp. p. 14, in The Cambridge History of China , vol. 5, in preparation. For an analysis of the sources and distribution of income and expenditure during the Sung, see Robert M. Hartwell, "Government Finance and the Regional Economies of China ca. 750-1200," unpublished paper prepared for the PARSS Seminar on Historical Data and Theories of Rational Choice, 1987. Studies of specific government contributions to various features of economic development during the Sung abound. For a synthesis of the chief Chinese and Japanese literature as of the early 1970s (drawing particularly on the work of the Tokyo school), see Mark Elvin, The Pattern of the Chinese Past (Stanford: Stanford University Press, 1973).

[5] Chang Fang-p'ing, Lo-ch'uan chi (SKCS ed.), 25/25b-26a; Ma Tuan-lin, Wen-hsien t'ung-k'ao (Shih-t'ung ed.), 4/59b (hereafter WHTK ). During the Ming and early and middle Ch'ing the land tax accounted for about 75 percent of total government revenues. See Ray Huang, Taxation and Government Finance in Sixteenth-Century Ming China (Cambridge: Cambridge University Press, 1974), p. 46; and Yeh-chien Wang, Land Taxation in Imperial China , 1750-1911 (Cambridge: Harvard University East Asian Series, 1973), table 4:2, p. 72, and table 4:4, p. 80. Feuerwerker estimates Northern Sung taxation at about 13 percent of total economic production, in contrast to between 6 and 8 percent for the Ming and Ch'ing (through the late nineteenth century), speculating that the difference may reflect the greater ease of taxation in the context of relatively higher urbanization rates during the Northern Sung. See "The State and the Economy," table 1, p. 4, and p. 7. Feuerwerker's figures are quite speculative, as he himself suggests, but they suggest an order of magnitude that is accepted by, inter alia, Marianne Bastid, "The Structure of the Financial Institutions of the State in the Late Qing," in Stuart Schram, ed., The Scope of State Power in China (New York: St. Martin's Press, 1985), pp. 74-75.

growth. But the impact of the activist state on a growing economy was mediated by a second critical component of Sung China's political environment, the rise from the seventh century on of powerful and united steppe empires. Encircled by stable, literate, and often hostile states, the Sung was unquestionably the most beleaguered of the major dynasties, and the only one forced to acknowledge equal players in a multistate Asian system. By 1127 Sung would lose half of its territory to the Jurchen Chin, but the dynasty had been threatened by rising frontier states from its very inception. At the founding of Sung in 960 the Khitan Liao were already established well inside the Great Wall frontier to the north, and shortly thereafter the Tangut Hsi-Hsia began an expansion outward from the Ordos that by 1036 gave them control of all the oasis cities of the Kansu corridor. Sung policymakers were constantly aware of the power of their northern neighbors, and of their own vulnerability. In many ways the competitive nature of the Asian geopolitical system between the tenth and thirteenth centuries helped foster the innovations that have come to characterize the Sung economic revolution. Sung's precocious development of mining, metallurgy, armament, and naval technology all stemmed in part from the need to repel sophisticated steppe cavalries, while the state's unusually active promotion of international trade was aimed at supplying much-needed revenues and such essential foreign commodities as horses. Later dynasties paid a price for their relative freedom from competitive pressures: following the collapse of the Yuan dynasty in the mid-fourteenth century the steppe no longer presented a formidable threat to China, and as Elvin has suggested for the late imperial era as a whole, and Wakeman for the Ch'ing, the absence of strong international competitors fostered a sense of self-sufficiency and isolationism in the last two dynasties that contributed to their political and technological stagnation.^[6]

[6] Sung encirclement is discussed in greater detail in Paul J. Smith, Taxing Heaven's Storehouse: Horses, Bureaucrats, and the Destruction of the Sichuan Tea Industry, 1074-1224 (Cambridge: Harvard University, Council on East Asian Studies, 1991), chapter 1. See also the essays collected in Morris Rossabi, ed., China Among Equals: The Middle Kingdom and Its Neighbors, 10th-14th Centuries (Berkeley: University of California Press, 1983). William McNeill, The Pursuit of Power: Technology, Armed Force, and Society since A.D. 1000 (Chicago: University of Chicago Press, 1982), chap. 2, provides an extremely clear synthesis of Western scholarship on the interrelationship between war, markets, and technology in Sung China. For the retarding effects of the absence of competitors, see Elvin, The Pattern of the Chinese Past , pp. 215-25; Frederic Wake-man, "China and the Seventeenth-Century Crisis," Late Imperial China 7, no. 1 (June 1986): 20-23. In excluding the Manchus themselves as a threat from the steppe I follow the position that the Ming fell under the weight of its own systemic sclerosis, which rendered any solution to the externally generated crises of climate, famine and pestilence, bullion-flow, and the rise of the Manchus impossible. See Ray Huang, 1587: A Year of No Significance (New Haven: Yale University Press, 1981), and William Atwell, "The Seventeenth-Century Crisis in China and Japan," Journal of Asian Studies 45, no. 2 (Feb. 1986): 223-44.

Chou Po-ti has suggested that the fiscal innovativeness of Sung stemmed directly from its military weakness, which compelled financial administrators to invent new methods of generating revenues to pay for expensive technological and logistical solutions to military vulnerability. This points to the malign side of the relationship between the rise of the steppe and the Sung economy. The need to maintain a constant defensive readiness against actual or implied foreign threats was translated into a growing fiscal burden that transformed the Sung fisc into what Shizuo has referred to as a perpetual "wartime economy." By the middle of the eleventh century, for example, the regular army numbered 1.25 million men, for whom grain and other provisions had to be imported from the southeast at great cost to the government. As of 1065 defense expenditures alone consumed 83 percent of the government's cash and 43 percent of its total yearly income, surpassing by 35 percent the entire Ming budget of 1502. Escalating defense expenditures forced the activist state to dig ever deeper into the burgeoning economy, raising the danger that overtaxation would encounter diminishing returns and bring investment and economic growth to a halt. Indeed, by 1059 revenues and production had already succumbed to overexploitation in the southeastern tea industry, forcing the government to suspend its eighty-year monopoly.^[7]

Beginning in the late T'ang, then, the rise of a professional financial service converged with the growing costs of geopolitical insecurity to produce an activist orientation to economic statecraft. When a sense of crisis overtook the Sung leadership in the mid-eleventh century, this activism provided the foundation for the radical state-directed experiments of the New Policies. The proximate origins of the New Policies can be traced back to the early 1040s, when the Sung lost every important battle in a three-year war with the Tanguts that underscored the impotence of China's huge but aging army and drove the economy into two decades of inflation. The Tan-gut wars initiated a quarter century of political reexamination, but neither the Ch'ing-li reform movement under Fan Chung-yen nor later reform initiatives by such leaders as Ssu-ma Kuang and Ou-yang Hsiu in the 1050s and 1060s successfully addressed the chronic problems of physical security and financial solvency, and military and economic crises again converged in the mid-1060s: the Tanguts stood poised to invade the Tibetan buffer zone of Tsinghai, threatening Sung's only horse supplier and imperiling its very

[7] Chou Po-ti, Chung-kuo ts'ai-cheng shih (Shanghai, 1981), pp. 259-60; Sogabe Shizuo, Sodai zaisei shi (Tokyo, 1966), p. 3; Shiba Yoshinobu, "Sodai shiteki seido no enkaku," in Sodai shogygoshi ronso (Tokyo, 1974), p. 128. The Ming figure is extrapolated from Ray Huang, Taxation and Government Finance , p. 46. On the southeastern tea industry, see Paul J. Smith, "Interest Groups, Ideology, and Economic Policy-making: The Northern Sung Debates over the Southeastern Tea Monopoly," unpublished paper presented at the Regional Conference, Association for Asian Studies on the Pacific Coast, Eugene, Oregon, 1977.

borders, and for the first time government expenditures exceeded receipts.^[8] Emperor Shen-tsung's accession in 1067 brought an even greater sense of urgency and provided the opportunity for a new approach, and the new emperor issued a broad appeal for fundamental reforms^[9] Within a year the emperor had thrown his support behind Wang An-shih, who proposed meeting the crises of the time by thrusting the power of the bureaucratic state still further into society and the economy in order to mobilize the resources of the nation for defense and national renewal. In conception and execution the resulting New Policies represent the culmination of the trend of economic activism begun three centuries earlier in the mid-T'ang.

The New Policies and Economic Activism

Many eleventh-century political thinkers saw an intensifying conflict between public (kung ) and private (ssu ) interests, but they offered widely different solutions. As Peter Bol shows in his comparison of the political visions of Wang An-shih and Ssu-ma Kuang, Ssu-ma accepted the proliferation of private interest groups as irreversible and sought only to prevent private interests from capturing the public institutions and functions of the state, and thus to ensure that as the sole public and impartial agency in society the state could mediate between private interests to promote social and political stability. Wang An-shih, on the other hand, came close to rejecting a legitimate place for private interests in the polity (without forgoing the manipulation of private interests as a means of inducing public-minded behavior), and from early on dedicated his career to recreating in the present an imagined ancient unity that would obliterate barriers between public and private spheres and collapse distinctions between the state and society.^[10]

Wang's idealized commitment to reunifying state and society informed a variety of practical policy proposals, such as the attempt to merge clerks and officials through reforms of the subbureaucracy, or to unite farmers and soldiers in his pao-chia local policing system.^[11] But the goal was most

[8] Ch'eng Min-sheng argues that the deficit, traditionally put at 4.2 million units, was a statistical fiction that reflected the inability of Sung financial officials to keep track of the national accounts by 1065. See his "Lun Pei-Sung ts'ai-cheng ti t'e-tien yü ch'eng-p'in ti chiahsiang," Chung-kuo-shih yen-chiu , 1984, no. 3:27-40.

[9] Sung-hui-yao: chih-kao 60/3a-4a (hereafter cited as SHY:CK ).

[10] See Peter Bol, "Government, Society, and State: The Political Visions of Ssu-ma Kuang and Wang An-shih," chap. 3 in this volume.

[11] In 1072 Wang told emperor Shen-tsung that "to unite clerks and officials and farmers and soldiers as one is the first obligation of imperial governance" (li yü shih, ping yü nung, ho wei i, ts'u wang-cheng chih hsien-wu yeh ). Li T'ao, Hsü tzu-chih t'ung-chien ch'ang-pien (Shih-chieh shu-chü ed.), 237/8a (hereafter cited as HCP ). For an extended discussion of this issue, see Miyazaki Ichisada, "O Anseki no rishi goitsu saku—Soho o chushin to shire," in his Ajia-shi kenkyu , vol. 1 (Kyoto, 1957), pp. 311-64; anti James Liu, Rearm in Sung China (Cambridge: Harvard University Press, 1959), p. 82.

sharply enunciated in his theories of political economy. Wang put public finance at the very center of public affairs, even investing finance with a sense of moral mission: "governing is for managing resources, and managing resources is what is meant by moral duty" (cheng-shih so-i li-ts'ai, li-ts'ai nai so-wei i yeh ). Public finance could assume a moral dimension for Wang precisely because he recognized no rightful demarcation between the public and private sectors of the economy: household, public fisc, and the resources of the natural economy were interlinked, so that if any one sector was to be enriched they must all be enriched. As Wang proclaimed in his Myriad Word Memorial of 1058, the proper object of public finance (chih-ts'ai , governing resources) had always been "to utilize the energy of all under heaven to produce wealth for all under heaven, and to use the wealth of all under heaven to meet the needs of all under heaven."^[12]

Wang's normative model of state and society generated an unconventional, although by no means unprecedented, diagnosis of the mid-century fiscal crisis. The conventional view was represented to the emperor by Ssu-ma Kuang, after he was appointed to investigate the deficit crisis in mid-1068. Ssu-ma reported that "the reasons for the current budgetary shortage lie in wasteful administrative expenditures, unrestrained bestowal of emoluments and rewards, an overpopulated imperial household, a bloated bureaucracy, and an inefficient military."^[13] Shen-tsung dutifully instituted a campaign for fiscal responsibility in the imperial household and established an Office of Expenditure Reductions (ts'ai-chien chü ), but by early 1069 the economizing campaign appears to have run out of steam, and the emperor turned instead to Wang An-shih.

Wang had consistently refused to see excessive expenditures or inadequate resources as the primary reasons for budgetary crises. On the broadest level, the state was straitened because "there was no Way (tao ) in the governance of resources," and its officials had "lost the Way of creating wealth."^[14] By the "Way of creating wealth" Wang meant the institutions, techniques, and authority through which the state was meant to govern the exchange values (ch'ing-chung ) and collection and disbursement (lien-san ) of goods and money in the economy as a whole. Though today we

[12] Wang An-shih, Wang Lin-ch'uan ch'üan-chi (Shih-chieh-shu>-chü ed.), 73/464, "Ta Tseng Kung-li shu"; 75/479, "Yü Ma Yun-p'an shu"; 39/222-23, "Shang Jen-tsung Huang-ti yen-shih shu" (hereafter cited as WLC ).

[13] HCP:Shih-p'u 3A/14b-15a; WHTK 24/232c.

[14] WLC 39/222; 75/479.

may see the eleventh century as the high point of Chinese economic expertise, from Wang's perspective it was the very erosion of economic skills that was responsible for the financial problems of the state. For example, in early 1069, when Shen-tsung sought Wang's advice on how to manipulate the value of Shen-hsi's currency to improve frontier grain supply, Wang replied that first it was necessary to revive the basic techniques of economic regulation (literally the "technique of opening and closing [the flow of money] and collecting and disbursing"—k'ai-ho lien-san chih fa ) as once practiced by the Chou dynasty Treasury Officer (ch'üan-fu ):

The office of Treasury Officer was the means by which the Former Kings controlled and regulated would-be engrossers (chien-ping ), measured and equalized the differences between rich and poor, transformed and circulated the wealth of all under heaven, and caused all benefits to flow from a single source.^[15]

In short, for Wang the appropriate way to enrich state and society was to stimulate the economy, which meant suppressing private monopolies, re-distributing wealth, and facilitating the smooth flow of resources throughout the empire. But as Wang and the emperor agreed, the establishment of necessary regulatory authorities was hampered by the small number of officials competent in the techniques of economic management (li-ts'ai ). And as Wang argued in his "Request for a Finance Planning Commission" (chih-chih san-ssu t'iao-li ssu ), through incompetent economic management the state had forfeited its fiscal prerogatives to private economic actors:

Most commodities used by the court are levied from places where they are not produced, or demanded before their season. Wealthy merchants and great traders have taken advantage of the crises. This allows both public and private interests to usurp control over exchange values and the collection and disbursement of money and goods.^[16]

[15] Yang Chung-liang, Tzu-chih t'ung-chien ch'ang-pien chi-shih pen-too (Sung-shih tzu-liao ts'ui-pien ed.), 66/2095 (hereafter CPPM ); HCP:Shih-pu 4/4b-6a. The Treasury Officer (see Chou-li , ch. 4) was often invoked by Wang as the model regulatory office, particularly in regard to the farming loans. Historically the Treasury Officer seems to have served as the ideal model for activist intervention in the market economy. In 10 A.D. Liu Hsin recommended it to Wang Mang, describing it as a Chou official "[for control of money in public markets], whose office received [by purchase for public storage] that which was unsold, and [when there was a scarcity] gave [out for sale] to those who desired to obtain." See Nancy Swarm, Food and Money in Ancient China (Princeton: Princeton University Press, 1950), p. 335.

[16] SHY:Shih-pu 4/5b; CPPM 66/2096-97. A summons was issued to all officials in the capital and the field with any competence in economic matters to memorialize their views, emphasizing that "individuals of all sorts (chu-ssu jen ) will be permitted to state their views before this commission." (CPPM 66/2097). For the Sung shih authors it was this creation of the Finance Planning Commission that launched the New Policies. See T'o T'o et al., Sung shih (Beijing ed.), ch. 14 (hereafter SS ).

In the second month of 1069 Shen-tsung authorized the creation of the Finance Planning Commission to "promote the arts of economic management," to coordinate economic planning, and to provide Wang An-shih with an agency through which to "recapture authority over exchange values and collection and disbursement and return it to the public domain (kung )." This move not only formally inaugurated the New Policies, it also launched a government campaign against "engrossers" (chien-ping ) in the commercial economy. The term "engrossers" had been used at least as early as Hsun Tzu (fl. 298-238 B.C ) to mean individuals who used extraordinary wealth and power to expropriate others. Since at least the Han the label had been used to categorize classes of economic actors, such as great landlords, merchants, and mining industrialists, whose wealth and power the state saw as destabilizing or coveted for itself.^[17]

Wang An-shih inveighed against "wicked engrossers," who had shattered the ancient unity between public and private property, in an early poem entitled "The Engrossers." Under Wang's aegis the term was deployed in the same way that class labels have been used in the People's Republic of China, to justify new uses of state power and to identify potential or actual targets of state campaigns. Wang himself emphasized the suppression (ch'ueh-che ) of engrossers as an essential component of economic statecraft, as we have seen; and in fact the need to suppress engrossers provided the explicit justification for two of the three policies (State Trade and Green Sprouts) most closely associated with the Finance Planning Commission. The 1072 edict that established a branch of the State Trade Agency (shih-i wu ) in the capital, for example, lamented that "traveling merchants from all over the empire who bring goods to the capital are put in great distress by the 'engrossing houses,' and many must sell at a loss and go out of business." That same year Wang An-shih described to the emperor the prac-

[17] WLC 70/445, "Ch'i chih-chih san-ssu t'iao-li." The memorial is undated. Ch'i Hsia, not wholly convincingly, ascribes it to Lü Hui-ch'ing, handpicked by Wang An-shih to administer the Commission. Su Ch'e was selected to coadminister with Lü, but in his first memorial denouncing "superfluous clerks, superfluous soldiers, superfluous officials" he took his stand with the proponents of "fiscal responsibility" rather than economic activism. See HCP:Shih-p'u 4/7a-14b; CPPM 66/2096; Ch'i Hsia, Wang An-shih pien-fa (2d ed., Shanghai, 1979), p. 270. For examples of the term chien-ping , see Chung-wen ta tz'u-tien , entries 1511.40-41; Daikanwa jiten , entries 1483.115-117. T'ung-tsu Ch'ü writes that the "term ping-chien or chien-ping suggests a person whose extraordinary wealth and power enabled him to encroach upon the people. Literally, it means to swallow up, to encroach, and thus to grab the property of others." See his Han Social Structure (Seattle: University of Washington Press, 1972), p. 394, n. 7. For an example of the use of the term in Han, see Wu-ti's currency reform of 120 B.C. , which was intended "to provide [money] for governmental expenses, and to curb unsettled, unscrupulous monopolists (Swann's translation for chien-ping ) [and their followers]." Swann, Food and Money , p. 266, translation of Han Shu 24B/1163.

tices of "engrosser" comb wholesalers who tried to depress the price of raw materials brought to the capital by traveling merchants, and "engrosser" tea-guild magnates who tried to raise the price of tea sold wholesale to less powerful guild members. In each case the State Trade Agency intervened to assume the function of primary wholesale distributor, inserting itself between the smaller merchants and the great metropolitan trading houses, in order to neutralize the monopolist powers of the largest guild merchants and gain a wholesaler's mark-up for itself. The campaign against "engrossers" was not confined to the great guild merchants of the capital, and the New Policies administration was equally willing to deploy state agents to compete with regional merchant groups, as in the Tribute Transport and Distribution Act (chün-shufa ), or with the kingpins of the rural economy, as we will see below in connection with the Green Sprouts policy.^[18]

Wang's commitment to pitting state agents against private "engrossers" directly contradicted conventional fiscal wisdom, for it required expanding rather than reducing the size of government and its expenditures. But Wang An-shih openly encouraged bureaucratic expansion and the increased expenditures this entailed, as a means of stimulating economic activity and generating greater revenues. In response to calls for the abolition of a waterworks post, for example, Wang wrote that "only with many officials can [essential] tasks be accomplished. So long as these tasks are accomplished there is nothing wrong with great [official] activity. And large expenditures will stimulate increasing prosperity. So long as they stimulate prosperity, what is the harm in great expenditures?"^[19] Shen-tsung was never as sanguine as his prime minister about the wisdom-of bureaucratic expansion, however, and the Hsu tzu-chih t'ung-chien ch'ang-pien records an attempt by Wang to instruct Shen-tsung in the economic benefits of

[18] For Wang's poem, which Bol dates to 1053, see WLC 5/22. The 1072 edict is in Sung hui-yao: shih-huo 37/14a (hereafter SHY:SH ). On State Trade Bureau intervention, see HCP 236/11b-13a, discussed at length in Kato Shigeru, "On the Hang or the Associations of Merchants in China" (in English), Memoirs of the Research Department of the Toyo Bunko 8 (1936): 68-69. Intervention should have equalized prices to the advantage of the smaller commercial interests. In the case of the comb market, the traveling merchants had themselves taken their case to the State Trade Bureau, which bought up their goods. According to Wang An-shih's report, "the 'engrossers' then wanted to utilize the new [state trade] laws to buy the entire lot (chan-mai ), but the [state trade administrator] instead distributed it to the local comb shops" (HCP 236/11b). In the case of the tea industry, Wang reported that "since we have now established a State Trade [Bureau] these ten or so [greatest guild] houses—the 'engrossing houses'—will have to buy and sell at the same price as the poor merchant houses, which is why these ten houses dislike the new policy and slander it. [I] recently obtained this information from a report by a member of the tea guild, but the same situation is to be found in all other guilds." HCP 236/31a; Kato, "On the Hang," p. 68. On class labeling in the People's Republic, see Richard Curt Kraus, Class Conflict in Chinese Socialism (New York: Columbia University Press, 1981), especially chap. 2.

[19] WLC 62/391, "K'an-hsiang tsa-i"; James Liu, Reform , p. 48.

appointing additional officials to administer the local Farming Loan and Service Exemption (mien-i ) funds:

The number of [county-level] officials in charge of distributing and collecting the Service Exemption and Ever-Normal funds does not exceed 500, and salaries for these 500 are not more than 100,000 strings [annually]. This year profits [from the two accounts] reached 3 million strings, at a [total administrative] cost of but 300,000 strings. Establishing new offices does not lead to unnecessary expenses.^[20]

Although Shen-tsung remained apprehensive about adding extra bureaucrats, under Wang An-shih's direction the number of qualified (not necessarily actively employed) officials registered the greatest increase of the entire dynasty, jumping 41 percent from 24,000 in 1067 to over 34,000 men in 1080.^[21]

But it was not enough to make government larger; if the state was to compete successfully with private interests in the expanding marketplace, then officials had to think and act not only like bureaucrats but also like entrepreneurs. It was precisely this vision of a society transformed by bureaucratic entrepreneurs that Wang An-shih put forth in his Myriad Word Memorial of 1058. There Wang blamed the economic, military, and moral ills of society on the deficiencies of the state, which was staffed mostly by men selected by the examination system for their strong memories and literary skills, rather than for the practical experience essential for good government. In chapter 3 in this volume Peter Bol addresses in detail the four stages—instruction, nurture, selection, and employment—Wang held necessary for producing an appropriate moral and political leadership over the long term. It is the second two stages that provide the core of Wang's theory of mobilizing talent for immediate reform.

In brief, Wang argued that men with the skills and abilities needed to meet the needs of the time had to be mustered from all levels of society, tested in actual government affairs, and the truly able ones appointed for long terms of office to posts that suited their qualifications.

In this way, intelligent, able, and energetic gentlemen [shih ] will be able to put all their intelligence into the pursuit of achievement, without having to worry that their projects will go unfinished and their accomplishments be left incomplete. . . . Once appointed, they must be given exclusive authority, not hampered or bound by this or that regulation, but permitted to carry out their ideas [chuan yen er pu i-erh i fa shu-fu chih, erh shih chih te hsing ch'i i ].^[22]

[20] HCP 250/14a-b (1072.2).

[21] See John Chaffee, The Thorny Gates of Learning in Sung China (Cambridge: Cambridge University Press, 1985), p. 27, table 4.

[22] WLC 39/220; see also p. 224, "she kuan ta-ti crier tang chiu ch'i jen . . . hou ke-yi tse ch'i yu wei."

Once a man has been appointed to office, Wang continued, he should be trusted to "select men of like character and put them to the test for a period of time, then appraise their abilities and make recommendations to the ruler, at which point they will be granted salaries and ranks." Above all, Wang stressed the need for full discretionary authority, writing that "there has never been a single case in history that has shown it possible to obtain good government even with the right man in power if he is bound by one regulation or another so that he cannot carry out his ideas."^[23]

The mobilization of bureaucratic entrepreneurship then incorporated five key elements: (1) men with the requisite practical and managerial skills were to be recruited from all levels of society; (2) the most able were to be assigned strategic posts or tasks on the basis of their proven skills, and not their formal credentials; (3) these action-oriented cadres were to remain in their respective positions for as long as it took to accomplish their assigned goals; (4) they were to be permitted to choose their own subordinates; and (5) they were to be given exclusive authority to experiment in the achievement of their tasks.^[24]

Wang An-shih's action strategy allowed a reform-minded leadership to define critical political and economic tasks, assign these tasks to aggressive and innovative officials in the field, and delegate extraordinary power to the men who were most successful in meeting reform goals. Wang had put his theory into practice once in 1060, when as member of an ad hoc committee to review horse-procurement policy he recommended that Hsueh Hsiang be appointed to the concurrent posts of Shensi fiscal intendant, regulator of Chieh-chou salt, and intendant for horse purchases and pastures, and given a broad mandate to reorganize the region's finances.^[25] Then from 1069 to 1076, as head of a government charged by the emperor to initiate wide-scale reforms, Wang had an opportunity to put his theory fully to the test. The following sections will focus on the Green Sprouts farming loans and the Szechwan tea and horse trade to assess the ways in which Wang's theories of economic activism and bureaucratic entrepreneurship animated the creation and enactment of New Policies economic reforms.

[23] WLC 39/224.

[24] My interpretation of Wang's Myriad Word Memorial as a theory of bureaucratic entrepreneurship is informed by Joseph Schumpeter, The Theory of Economic Development (New York: Oxford University Press, 1961), especially chap. 2; and James D. Thompson, Organizations in Action (New York: McGraw-Hill, 1967). Eugene Lewis has put Thompson's analysis of organizational power to similar use in his Public Entrepreneurship: Toward a Theory of Bureaucratic Political Power (Bloomington: University of Indiana Press, 1980).

[25] HCP 192/4a-b, 7a-9b; WLC 42/243-4, "Hsiang-tu mu-ma-so chü Hsueh Hsiang chatzu,"

Economic Activism and Policy Formation

I have ascribed to Wang An-shih a fiscal theory that sought to assert the regulatory authority and developmental initiative of the state in the market economy, and an action strategy that would mobilize bureaucratic entrepreneurs to enlarge the regulatory capacity of the state. In the realm of reform economic policy, fiscal theory and action strategy converged in the creation of new public agencies empowered to participate directly in the market economy, administered by officials chosen for their entrepreneurial characteristics. The Szechwan Tea and Horse Exchange and the Green Sprouts farming loans exemplify both the theory and the strategy. Both policies aimed at displacing private economic actors from key sectors of the economy; both policies came to emphasize revenue maximization out of an initially broader spectrum of objectives; and both policies were administered through new regional-level agencies—the Tea and Horse Agency (THA) and the Ever-Normal Granary intendancies (ENGs)—run by officials selected according to entrepreneurial criteria. But in the actual exercise of power the similarities between the THA and the ENGs break down. As I will show, the THA intendants were encouraged to create an autonomous public enterprise, and given a broad mandate to wield unrestrained public authority in the marketplace. The Tea and Horse Agency represents Wang An-shih's model of decentralized bureaucratic entrepreneurship in its purest form. The Ever-Normal Granary intendancies, on the other hand, were much more tightly restricted in their use of public power and became an example of what I describe as centralized managerial agencies. I discuss the reasons for these differences in a later section.

Displacing Private Interests

The Szechwan Tea and Horse Exchange and the Green Sprouts farming-loans act both replaced key private economic actors with agents of the state. The Green Sprouts policy took aim at the cycle of rural debt and propertylessness that seemed to be intensifying throughout the eleventh century. As Peter Golas argues in his useful survey of the literature on rural China during the Sung, by mid-century roughly 80 percent of rural landowning households (that is, "owner households," chu-hu ) were ranked in the lowest two grades (4 and 5) of a hierarchy consisting of five grades of commoners and a single grade of officials. Building on the work of Yanagida Setsuko, Golas estimates that as a class the grade 4 and 5 households (numbering about 10.5 million households) possessed only about 22.5 percent of the empire's cultivated lands, with average holdings of just fifteen mou (about two acres), roughly three mou less than was needed to meet the minimal food requirements of a family of five. On the whole, households in the

lowest two grades lived at the margins of economic independence and were often forced to supplement their incomes through hired labor or as tenants on other people's lands. As Colas puts it, "many of them were chronically in debt, and mounting debts led all too often to forced sale or foreclosure of their land."^[26]

All Chinese dynasties faced the problem of keeping poor peasants solvent and accessible to the state, in order to ensure minimal tax receipts and manpower for the military and public works. In Six Dynasties and early T'ang the principal lever used by the state to protect peasant solvency and independence was periodic land redistribution. Ever since the collapse of the equitable field (chü-t'ien ) system during the mid-T'ang, however, the state had ceased to intervene directly in land distribution, and private transactions in land had become routine.^[27] Having lost its capacity to limit the size of landholdings by command, the state was forced to devise other ways to keep at least some land in the hands of the smallest peasant landowners. As land and the entire rural economy became increasingly commercialized the peasant's ability to hold onto his plot depended more and more on his access to money and credit, and for New Policies reformers rural credit came to seem the fulcrum on which state power could best protect peasant solvency. In an important memorial in the ninth month of 1069 the Finance Planning Commission blamed peasant distress on the monopoly over rural credit wielded by "engrossing households": "The reason people are burdened by deficits is that in the seasonal gap separating the old harvest from the new, engrossing households take advantage of the crisis to demand interest rates of 100 percent. Consequently, would-be borrowers are often denied the funds they need."^[28]

The Commission proposed a solution that married a model of state-supplied rural credit that had been worked out independently by local officials to the enormous reserves of the Ever-Normal Granary system. A prototypical rural credit policy had been devised in the 1040s by the finance official Li Ts'an, who issued cash loans called "green sprouts cash" (ch'ing-miao ch'ien ) for the planting of dry-land crops in the spring, to be paid back in grain at harvesttime in the fall. Wang An-shih himself instituted a similar practice during his term as magistrate of Yin county in Chekiang in

[26] Peter Golas, "Rural China in the Song," Journal of Asian Studies 39, no. 2 (1980): 302-4. Chu-hu constituted two-thirds of the total rural population of 1050; k'e-hu , guest households, were officially regarded as migrants, with no land of their own. Naturally these categories raise a number of descriptive and analytical problems and have spawned a considerable literature, which Colas surveys on pp. 305-9.

[27] Twitchett, Financial Administration , pp. 1-23; Golas, "Rural China," p. 299.

[28] SHY:SH 4/16a. This key text, which comprises the basic legislation for the ch'ing-miao policy, is analyzed by Sudo Yoshiyuki, "O Anseki no seibyoho no shiko katei," Tokyo daigaku daigakuin kiyo 8 (1972): 172-74.

about 1049.^[29] Wang revived the notion in 1069 in the now-lost Documents on the Green Sprouts (Ch'ing-miao shu ), but temporarily bowed to Su Ch'e's protest that the combination of easy loan money, 20 percent interest, and local-government distribution and collection would corrupt otherwise honest residents and hand over irresistible coercive powers to already corrupt clerks. But at about this time Wang Kuang-lien, a managing supervisor in the Ho-pei Fiscal Intendancy, sent up a request for several thousand Buddhist ordination certificates to capitalize the local enactment of a Green Sprouts loan policy in Shensi, and Wang An-shih was able to push through adoption of Wang Kuang-lien's "Shensi Green Sprouts loan act" on a regional experimental basis.^[30]

The sources are muted on the details of Wang Kuang-lien's plan, when it was promulgated nationwide in the ninth month of 1069, capitalization for the loan fund was generated by liquidating the grain reserves in the Ever-Normal (ch'ang-p'ing ) and Universal Charity (kuang-hui ) granaries. Traditionally, the Ever-Normal Granaries were authorized to buy up grains when prices were cheap for resale when prices were dear, or in times of natural disaster. By the onset of Shen-tsung's reign, however, reformers and reform opponents agreed that the mechanism had ceased to work effectively: as a 1069 memorial from the Finance Planning Commission complained, 15 million piculs of grain and strings of cash sat idly in the Ever-Normal and Universal Charity granaries while relief aid was doled out piecemeal from the sheng-ts'ang granaries used to store grain for officials' salaries.^[31]

Naturally, conservatives and reformers disagreed about the cause of the Ever-Normal system's failure. Ssu-ma Kuang, for example, saw the issue as a problem of personnel, not of institutions. In a retrospective assessment in 1086, Ssu-ma agreed that the system was slow to respond to price changes. Before a grain purchase could be made, he wrote, the purchase request had to travel up the administrative hierarchy of county, prefecture, and judicial intendant to the Court of Agricultural Supervision; by the time a purchase authorization traveled back down the prices had all changed and the request was rendered irrelevant. But Ssu-ma insisted in 1086, as he had in 1070, that the problem with the Ever-Normal granaries lay, not in the system, but in the poor selection of administrators.^[32]

[29] Chou Po-ti, Chung-kuo ts'ai-cheng ssu-hsiang shih (Fukien, 1984), pp. 232-33; SS 330/ 10619; SS 327/10541. Wang's policy included interest charges; both measures were reported to be well received.

[30] HCP:Shih-pu [SP ] 5/20b-21a; SHY:Shih-huo [hereafter SHY:SH ] 4/16a-17b; SS 339/ 10822.

[31] On the granary systems, see SS 176/4275-42. The 1069 memorial is in SHY:SH 4/16a; 18a.

[32] HCP 384/1b-2a. Ssu-ma Kuang, Ssu-ma wen-cheng chi (Ssu-pu pei-yao ed.), 7/4a, "Ch'i t'iao-li-ssu ch'ang-p'ing-shi shu" (hereafter cited as SMWCC ).

The reformers, however, condemned the system itself and sought to install a new one. Although in principle the granary system was intended to provide a mechanism for the seasonal adjustment of prices, the Finance Planning Commission asserted that granary stocks were only sold off in years of truly bad harvests, and then only to the advantage of "shiftless idlers in the cities (ch'eng-shih yu-shou chih jen )."^[33] It is possible, however, that the Commissioners were moved chiefly by the capitalizing potential represented by the granary reserves, and in its directive of the ninth month of 1069 the Commission proposed a simultaneous solution to the problems of price equalization, disaster relief, and seasonal credit crises that transformed the nation's granary stocks into a "Green Sprouts" fund for rural investment. Control of the granaries was transferred from the circuit judicial intendants to the fiscal intendants, and almost immediately thereafter to the newly created Intendants for Ever-Normal Granaries (ENG), who were empowered to convert the reserves into a liquid loan fund.^[34]

Actual administration of the loan fund was assigned to the county magistrates and local village officers. The new policy was advertised locally through public notices. Potential borrowers were organized into mutual-guarantee groups (pao ) of five (later ten) or more households under the personal supervision of the county magistrates, with the assistance of village elders and household chiefs, and each household was permitted to contract for a loan of currency or grain proportional to its household wealth.^[35] Guest households (k'e-hu ) could borrow only by entering a group with an owner household (chu-hu ); and urban and suburban households could borrow in groups of five households if they put up collateral, but not until all interested villagers had obtained their loans. The loans were to be made in the spring and repaid in two installments after taxes in the summer and fall. A series of rules were established to protect borrowers against unfair manipulation by officials. For example, loans could be repaid in either currency or grain, with exchange rates set to ensure only that the government lost none of its basic capitalization (pen-ch'ien ), and forced loans (i-p'ei ) were expressly prohibited. Opponents of the Green Sprouts policy, as we shall see, charged that officials routinely disregarded these provisions.

Nothing less than a complete rejuvenation of the agrarian economy was expected from the new measure, which was seen as a means of rationalizing

[33] SS 176/4276; SHY:SH 4/16b. According to Han Ch'i, who vigorously opposed the new policy, under the old Ever-Normal system lower-grade villagers were issued vouchers that entitled them to buy up to 3 tan of grain at reduced prices at the granary in the county (seat?); poor urban residents and vagrant families were permitted to buy smaller amounts on a daily basis, and their access was tightly controlled. Upper-grade residents were, according to Han, excluded. See SHY:SH 4/28b.

[34] SHY:SH 4/17b; HCP:Shih-pu 6/13a.

[35] SHY:SH 4/19a; HCP 252/27ab.

the distribution and collection of grain and money to equalize commodity prices, minimize the unequal geographic distribution of resources, and ensure adequate supplies of relief grain. Most important, by supplanting private landlords and moneylenders as the principal source of rural credit, the state could "enable the peasants to hasten to the management of their affairs so that engrosser households will be unable to take advantage of their crises." With these basic problems solved, a galvanized officialdom could "exhort and induce [the peasants to greater productivity] and promote the benefits of irrigation, so that everywhere agriculture will be improved."^[36]

State control of the Szechwan tea industry was more directly tied to financial imperatives than was the rural credit policy. The tea and horse trade grew out of Wang Shao's state trade agency, which constituted a direct assault on mercantile profits. As military affairs commissioner for Ch'in-feng circuit, charged with prosecuting his own plan to recover the Tibetan territories of Hsi-ning and the T'ao River Valley that formed the centerpiece of New Policies military strategy, Wang Shao was ever alert for new revenues to support his campaign. In 1070 he requested permission to open a State Trade Bureau (shih-i wu ) in Ch'in-chou to capture the profits of frontier trade:

I do not know how many hundreds, thousands, and ten-thousands of strings worth of goods flow from all four corners of the Fan (Tibetan) lands to us every year, yet the profits of the merchant travelers all return to the people. I wish to set up a State Trade Bureau in this circuit and to borrow government funds to use as capital [with which to buy commodities to trade to the foreign merchants in order to] capture [for the state] the profits [that otherwise flow] to merchants and traders.^[37]

Wang An-shih actively promoted Wang Shao's recommendation, even reminding him in 1072, just after Wang Shao had captured Hsi-chou, to open a state trade bureau as soon as the town was walled, "so that Fan and Han, official and private interests will all benefit."^[38] That same year the state trade policy was enacted in K'ai-feng as well; in 1074 it was expanded into the Superintendency of State Trade, charged with breaking the monopoly of great guild merchants, speeding the circulation of goods, and lending out government goods and money at 20 percent interest.

Wang Shao's state trade policy led directly to the tea and horse trade. From the start Wang Shao needed to keep his frontier trading posts stocked

[36] SHY:SH 4/16a-17a.

[37] SHY:SH 37/14a; Smith, Taxing Heaven's Storehouse , pp. 44-47. The classic article on Wang Shao's frontier campaign is Enoki Kazuo's "O Sei no Kasei keiryaku ni tsuite," Moko gakuho 1 (1940): 87-168.

[38] WLC 73/464-65, "Yü Wang Tzu-ch'un shu."

with Szechwanese commodities, principally silk and tea, and from the capital Wang An-shih pushed through measures to ensure that he got an adequate supply. In late 1073 officials were sent to Szechwan both to buy tea for Wang Shao and to explore the possibility of opening a State Trade Bureau in Ch'eng-tu itself. But in mid-1074 Wang Shao changed the direction of events, memorializing that "The Westerners have begun to bring good horses to the frontier. All they desire is tea, but we do not have enough to trade to them. I request that the tea-purchase office buy tea in all haste."^[39]

The planning investigation for the Ch'eng-tu bureau was abruptly canceled, and in the eleventh month of 1074 two of the investigators, Li Ch'i and the Szechwanese P'u Tsung-min, were named the first intendant and co-intendant for tea purchasing for Western Szechwan and Han-chung and charged with buying tea in Szechwan and transporting it to the horse markets of Ch'in-feng and the newly opened Hsi-ho circuits.^[40]

The procurement of horses, a strategic good, was traditionally, and in Sung eyes naturally, managed by the state; the horse side of the THA enterprise added nothing new to the arena of state-managed trade. But on the tea side the THA intendants, driven by the need to create a self-financing trade, expanded the scope of state intervention in producer and consumer markets to an unprecedented degree. The state's relationship to the Szechwanese tea industry mirrored the region's turbulent response to Sung unification. The early Sung state's confiscatory policies in Szechwan at the end of the tenth century had produced the shattering rebellion led by Li Shun and Wang Hsiao-po, and even at the time of the New Policies opponents of economic activism in Szechwan invoked the cautionary message of that uprising. Political turmoil had made Szechwan hard to tax, and from 980 to 1059 a legislative wall was erected that left the Szechwanese tea industry internally unregulated but prohibited tea exports out of Szechwan itself, in order to preserve northwestern markets for the state's monopolized southeastern tea. When the southeastern tea economy collapsed in 1059, all restrictions were lifted, and for the first time in almost a century Szechwanese tea could circulate freely. Deregulation stimulated a booming triangular trade composed of Shensi salt, Szechwanese tea, and Inner Asian animal and mineral products, all under the control of private Shensi guest merchants (k'e-shang ). It was with these merchants that the new tea-purchase intendants had to compete, without destabilizing the industry or fomenting unrest.

[39] See HCP 217/9b-10a; HCP 219/8b; SHY:SH 37/18a; SS 186/4549-50. Wang Shao's memorial is in SHY:CK 43/47b.

[40] Much of the following discussion is taken from Smith, Taxing Heaven's Storehouse , chap. 4. The main records for the tea and horse trade during the New Policies are in SHY:CK 43 and SHY:SH 30.

The tea and horse intendants soon crafted an ambitious solution that paired a monopsony on tea production in Szechwan with a monopoly on tea sales in Shensi. On the tea-purchase side, the Tea and Horse Agency quickly laid claim to the tea industries of Szechwan and the Han-chung Basin, inserting itself between producer and distributor as the sole legal purchaser of the tea crop. Whereas Szechwanese tea households (ch'a-hu ) in the unregulated market had sold to traveling merchants, shop owners, local consumers, and in barter markets along the Szechwan-Tibetan frontier, under the monopsony regulations that were in full force by 1077 they could sell only to government purchasing officials in the forty-seven markets operated by the state. Although the government paid cultivators less than had private merchants, because of new demand opened up by the horse trade they could sell much more than before. Cultivators shifted from small-scale production of fancy tea for Chinese consumers to bulk production of low-quality tea for the state export trade; by 1085 the Szechwanese tea industry was producing, at thirty-eight million pounds yearly, more tea than the entire southeastern industry produced at any time during the Northern or Southern Sung, and a level of output not reached in any subsequent period in Szechwan until the twentieth century.^[41]

The fifteen thousand horses bought yearly from the Tibetans took only one-eighth to one-sixth of the Agency's tea, and the intendants concentrated on expanding the market in which to sell the rest. Because southeastern tea was still unregulated, no government agency had a proprietary interest in it, and the Tea and Horse Agency intendants took advantage of this administrative vacuum to get all of northwest China, from the frontier markets of Kansu in the west to the Chinese heartland prefectures within the Wei and Yellow Rivers to the east, designated an exclusive Szechwan tea-marketing zone. Moreover, this long-distance trade was operated entirely by Agency officials, who ran 332 tea-sales markets throughout northwest China, at every level of the administrative hierarchy from prefectural city to frontier stockade. Tea merchants were encouraged to buy tea from the Agency, but they could sell it only in the intraregional markets of Szechwan.^[42]

[41] Smith, Taxing Heaven's Storehouse , table 5, p. 218.

[42] For the enabling acts, see HCP 334/13a; SHY:SH 30/18b, 23a. The scope of THA control over the Szechwan tea industry is conveyed by the first 2 of 38 Revised Articles of the Tea Market Agency promulgated in 1083. The first article defined the monopsony operation in Szechwan: "All the tea-producing districts of Ch'eng-tu-fu and Li-chou Circuits and Chin-chou have established markets to buy out the tea of the cultivator households. Guest-(merchants) are permitted to buy tea in the government markets (on payment of a 10 percent license fee) and to sell it in the four circuits of Ch'uan-hsia (i.e., Szechwan) and on the borders of Chin-chou. All those who illicitly and heedlessly buy and sell, barter and trade in, or enter the territories of Shen-hsi may be seized and informed on by the general populace, and will be dealt with according to the articles on violation of the [Fukien camphorated] La-tea Act" (SHY:SH 30/18b).

A second article delimited the government's marketing region: "[Shan-fu West (i.e., Shen-hsi), Hsi-ho and Lan-hui Circuits and the two prefectures of Wen and Lung] constitute a Government Tea Restricted Region (kuan-ch'a ti-fen ). For all merchants of any circuit selling Szechwanese tea, southern tea, la-tea, or various teas without licenses, permit commoners to inform and seize. All violators will be dealt with according to the regulations on violation of the La-tea Act" (SHY:SH 30/93a).

The Szechwan tea and horse trade relegated merchants to minor partnership in a huge state-run enterprise. Public interest was occasionally claimed as justification for the government monopoly, but in general the THA intendants simply cited evidence of high merchant interest and profits as sufficient warrant for annexing a given market and prohibiting further private trade. Because of the overriding need to provision and finance frontier defense the THA intendants could simply disregard private mercantile interests, without having to justify their expansion into private markets.^[43]

Though the Green Sprouts program and the Szechwan tea monopoly both set state agents against private economic interests, their initial goals were quite different: the rural credit policy was enacted as an economic welfare measure, and the tea monopoly was established to provide exchange commodities for a publicly conducted foreign trade. But in a policy-making environment driven by the desire to eliminate deficits and generate new revenues, which most scholars agree Shen-tsung's court was, both policies were soon transformed into predominantly revenue-producing operations.

In its 1069 directive establishing the Green Sprouts loans, the Finance Planning Commission explicitly disavowed any diversion of the loan fund to the public treasury: "The loan policy is for the sake of the people; the government (kung-chia ) will claim no benefit from its receipts." Nor were provisions made to subsidize administrative costs of the loan policy, but within a few months interest rates of 20 to 30 percent were established in order to finance it. These interest charges were among the most controversial features of the Green Sprouts measure. Wang An-shih claimed that the Chou li provided a philosophical justification for interest charges, but opponents dismissed Wang's reading of the text and rejected outright the propriety of government usury. Han Ch'i, for example, in one of the most interesting rejoinders, argued that the people's taxes and services to the state should stand in lieu of interest charges. In fact Wang agreed with his critics that

[43] See SHY:SH 30/12b on the Ming-shah tea trade; SHY:CK 43/48a on Hsing-yuan-fu and Yang-chou in the Hanchung Basin; and SHY:SH 30/13a on Liu Tso's intent to monopolize the Shensi-Szechwan tea and salt trade. In 1086 Investigator Huang Lien successfully invoked the financial imperatives of frontier defense as a basis for retaining the THA monopoly even in the midst of Restoration attempts to dismantle the Tea and Horse Agency. HCP 381/ 22b-23a.

charging interest on government loans was by no means ideal, but he argued that in order to remain solvent the loan fund had to be rationally managed: salaries, transport costs, emergency supplements, and spoilage all had to be provided for. For this reason, Wang concluded, a 20 percent interest rate was the very pillar of rural credit reform.^[44]

Thus the Green Sprouts loan policy was intended to be self-financing, but not profit-making. The revenue potential of the loan fund was enormous, however, and given the New Policies' financial goals it was inevitable that the fund would be thoroughly exploited. In mid-1070 jurisdiction over the Ever-Normal Granary system was transferred from the Finance Planning Commission to the Court of Agricultural Supervision (ssu-nung-ssu ), newly reconstituted under the leadership of Lü Hui-ch'ing. Soon the growing loan fund joined service exemption revenues as the chief source of money for Ssu-nung-ssu projects. From 1070 to 1085 Ever-Normal Granary funds were routinely used to finance flood and famine relief, irrigation projects, and border provisioning and defense. By 1072, as was noted above, Wang An-shih had begun to praise the loan policy at least as much for its profitability as for its redistributive justice. And in 1082, two years after Shen-tsung built the Yuan-feng Treasury to hold the Court of Agricultural Supervision's growing riches, eight million strings of cash in surplus farming-loan profits from around the country were deposited in its vaults.^[45]

As critics such as Han Ch'i and Ssu-ma Kuang had predicted all along, then, revenue production soon superseded rural credit relief as the primary operational goal of the farming-loan policy. Although social welfare ideals constrained the pursuit of profit, Ever-Normal Granary officials were judged, not by the efficiency of their disaster-relief operations, as Ssu-ma Kuang recommended, but by the volume of loans issued and collected. Overall, the farming-loan system operated under a minimum quota of 13.9 million strings collected on loans of 11 million strings, for a net profit quota of 26.5 percent. Individual officials were in turn judged for promotion or demotion primarily (though not exclusively) on the basis of their success in meeting local loan targets, which were not abolished until 1086.^[46] Social welfare ideals could not be ignored in an agency ostensibly dedicated to agrarian relief, and the contradictions between social and fiscal goals had a marked impact on the power structure of the Ever-Normal Granary system, as I discuss more fully below. But it is clear that for the duration of the

[44] SHY:SH 4/16b; Han Ch'i's Chia-chuan , quoted in HCP:Shih-pu 7/26a; WLC 73/464, "Ta Tseng Kung-li shu." See also WLC 41/239, "Shang wu-shih cha-tzu."

[45] HCP 211/9a-b; 330/12a-b. This source contains extensive material on farming-loan subventions to other agencies.

[46] HCP 374/11b-13a; 226/6b; 287/5a-b; 330/12a-b; 358/2a. Han Ch'i's and Ssu-ma Kuang's criticisms are taken up in greater detail below.

New Policies making money became the chief objective of the farming-loans measure, while economic welfare was relegated to a secondary place.

The Tea and Horse Agency made the transition to a predominantly revenue-producing agency by a less circuitous route. When appointed intendant and co-intendant in 1074, Li Ch'i and P'u Tsung-min were charged simply with buying tea in Szechwan and transporting it to the horse markets of Ch'in-feng and the newly opened Hsi-ho circuits. Implicitly, they were expected to make the operation self-financing, but even that charge was never specified.^[47] As with the Green Sprouts loan fund, however, the revenue potential of the tea not needed for the horse trade was enormous, and the THA intendants very quickly learned how to realize that potential and to transform profits into bureaucratic power. In earlier work I have already demonstrated the relationship between Agency profits and the bureaucratic power cycle; here I simply note that the THA chiefs were able to translate every increase in their net-profit quotas into greater organizational power, as measured by increases in their formal status and that of their Agency and by rising influence and autonomy in their relationship to other agencies and internal resources and personnel.^[48]

The chief source of Agency profits was the monopoly in Shensi. According to figures provided by the Szechwanese opponent of the tea monopoly, Lü T'ao, the Agency could realize a net profit of 200 to 300 percent on sales of Szechwanese tea in the northwest.^[49] (Private tea merchants, on the other hand, earned only about 25 percent on their sales of tea in the intraregional markets of Szechwan.) From tea sold directly to foreign traders and domestic consumers in the northwest and to merchants in Szechwan the tea and horse enterprise not only became self-financing but also generated substantial surpluses that could be allocated to defense agencies in Shensi, or deposited in court-controlled discretionary funds (feng-chuang ) in Shensi and Szechwan. Yearly surpluses amounted to 400,000 strings of cash in 1074, 600,000 strings by 1076, 1 million strings (the new quota) by 1083, 1.6 million strings in 1084, and 2 million strings—4 percent of the imperial cash income of 48.5 million strings—in 1085.^[50] And of course the Agency paid for the horse trade.

As with the Green Sprouts farming loans, the transformation of the tea and horse trade into an enterprise devoted essentially to revenue production is revealed by the Agency's incentive system. And without the conflicting claims of social welfare, the Tea and Horse Agency intendants could devote

[47] I impute the implicit expectation that the tea and horse trade be self-financing on the basis of my analysis of the history of Sung horse procurement policies. For the detailed argument, see Smith, Taxing Heaven's Storehouse , pp. 38-44.

[48] Ibid., chap. 5.

[49] Lü T'ao, Ching-te chi (Ssu-k'u ch'üan-shu ed.) 3/9a.

[50] Smith, Taxing Heaven's Storehouse , Appendix D, pp. 332-34.

themselves to making money with unalloyed zeal. In order to translate the fiscal interests of the state into the private interests of the officials and local participants in the Tea and Horse Agency, the intendants devised a unique achievement-oriented incentive system that dispensed liberal promotions and material rewards for the production of surplus revenues.

The most striking component of this system split a percentage of the extra-quota profits between the tea-market inspectors (chien-kuan ), service personnel, and key local-government officials. In the monopsony markets of Szechwan and Han-chung, which sold tea to shops and local traders, the inspectors shared 5 percent of all earnings above their market's net-profit quota with their commoner stock controllers, warehousers, scale operators, and bookkeepers; another 1 percent went to the local-government official, usually the county magistrate, who served as nominal auditor for the market. A similar system applied to the monopoly markets of the northwest. And in government markets where tea had been traded for silver, textiles, grains, and assorted other commodities, officials and clerks shared 10 percent of all net profits realized in converting the trade goods into cash, providing the liquidation took place within six months. Indeed, the constant and profitable turnover of tea and other goods for cash took on the status of an Agency obsession, testified to in notices sent to market officials to "expedite rapid purchases and sales, prevent the stagnation of capital" (wu-ling mai-mai t'ung-k'uai, wu-chih fang-chih ch'ien-pen ).^[51]

The Structure of Economic Activism: Entrepreneurial and Managerlal Intendancies

It should be clear from the discussion above that as policies that mobilized state agencies to compete with private economic interests and sought to increase the state's share of total economic resources, Green Sprouts and the Szechwan tea monopoly exemplified New Policies economic activism. Both policies also shared a basic feature of the New Policies administrative model: the proliferation of new circuit intendancies to fulfill reform tasks at the level of the economic, administrative, or military region.

As Winston Lo has shown in his important analysis of circuit intendancies and Sung territorial administration, Wang An-shih and his colleagues, rather than working through existing machinery and entrenched personnel, preferred to create new institutions to carry out reform tasks. At the political center, the reformers breached the three-part division of financial, mill-

[51] Though the Agency's incentive system was enacted after Wang An-shih had retired from office, the congruence between it and Wang's theory of uniting officials and clerks is clear. For the incentive regulations, see Lu Shih-min's 38-article Revised Regulations of the Tea Market Agency of 1083, sixth month, in SHY:SH 30/18b-21a. On expediting purchase and sale, see Lü T'ao, Ching-te chi 1/12a.

tary, and executive authority by creating new fiscal agencies (such as the Finance Planning Commission) or reviving old ones (such as the Court of Agricultural Supervision), and by routing orders around the Directorate of Military Affairs, to centralize general policy-making and command functions in the executive branch of government dominated by Wang An-shih and his followers. Specific reform initiatives were then delegated to such new single and multicircuit intendancies as the intendants for state trade, the pao-chia and militia intendants, the Office for the Management of Frontier Finances, and of course the Ever-Normal Granary and Tea and Horse intendants. And the reformers increasingly empowered circuit intendants to bypass conservative prefectural administrators in exercising administrative control over the counties, by giving them greater powers over the appointment and promotion of county magistrates.^[52]

The circuit intendancies, then, served as the central link in a chain of command joining reformist policymakers, region, and locality. The Intendancy for Tea and Horses projected state power into the tea production and sales markets of Szechwan and Shensi through an executive triad of super-intendant, co-intendant, and vice-intendant. Similarly, the Ever-Normal Granary intendancies linked K'ai-feng to agrarian producers and borrowers all over the empire, through intendants and co-intendants in the capital region and, at one time or another during the New Policies era, twenty-two of the empire's twenty-three circuits.^[53]

To summarize the argument to this point, then, the Green Sprouts and Tea and Horse policies embodied a common policy impulse—restoring the state as the dominant actor in the commercial economy—and were embedded in homologous institutions—regional and subregional state agencies, or intendancies. But when we focus on the ways in which the THA and ENG intendants were allowed to fulfill their assigned tasks—that is, on the actual structure and exercise of power—it becomes clear that power was decentralized to the THA and ENG intendancies to strikingly different degrees. Reform policymakers granted the THA intendants considerable autonomy over critical inputs (whom to tax, where and from whom to buy), disposi-

[52] Winston Lo, "Circuits and Circuit Intendants in the Territorial Administration of Sung China," Monumenta Serica 31 (1974-75): 89. For Lo's model of the Sung system of dual control over county magistrates, see pp. 92-95. In 1086 Ssu-ma Kuang retrospectively denounced Wang An-shih's practice of creating new intendants, assistant intendants, and staff supervisors for each reform policy, and staffing them with young, low-ranked, and frivolous men with seniority status no greater than prefectural vice-administrator, county magistrate, or market inspector. (HCP 368/23b-24a.)

[53] For the initial deployment of intendants and co-intendants, see CPPM 68/2162. The only circuit for which I have not identified an intendant is Huai-nan-hsi, but the absence may be simply circumstantial. The circuit's granary system was no doubt administered from neighboring Huai-nan-tung.

tion of output (how and to whom to distribute its goods and services), and internal structure, most importantly its personnel. In the case of the Ever-Normal Granary intendants, these same decisions were made at the center, primarily by the Court of Agricultural Supervision. The THA intendants obtained for themselves precisely that entrepreneurial autonomy prescribed by Wang An-shih, while the Granary intendants served as managers of an administrative apparatus closely controlled from the center.

Wang An-shih's entrepreneurial strategy, it will be recalled, entailed assigning proven innovators to long terms of office in strategic posts, regardless of their formal credentials, and allowing them to staff their own organizations and experiment in the fulfillment of their tasks unfettered by bureaucratic red tape. The first set of characteristics describes the selection of entrepreneurial types; the second set, the exercise of entrepreneurial power. The men selected as Tea and Horse and Ever-Normal Granary intendants shared certain "entrepreneurial" traits that made them, on taking up their posts, relatively indistinguishable. But once they were in position, as we will see, the similarities diminished.

Attributes of the Intendants

The first two components of Wang's entrepreneurial strategy, selecting action-oriented and practical men regardless of formal credentials, can be specified by mode of entry and formal civil service rank. In the case of the THA, six of the seven men appointed to one of the triad of intendant's positions between 1074 and 1085 rose to office, not just as fiscal experts, but as fiscal experts who had displayed their aggressiveness, ingenuity, and loyalty to the reform cause in previous New Policies posts or tasks. To cite but three examples, the first intendant, Li Ch'i, had been commended for his management of currency and silk reserves in Shensi before being sent to Ch'eng-tu. His successor, the Szechwanese Li Chi, had worked for Lü Hui-ch'ing and served the reform cause as fiscal staff supervisor (chuan-yun p'an-kuan ) in Hopei before getting the THA intendant's post. And Lu Shih-min, the most expansionist of the THA empire-builders, served as Li Chi's managing supervisor (kan-tang kung-shih ) before his appointment as co-intendant in 1080, then superintendant in 1083, where he expanded and consolidated the dominant position of the THA in two macroregions (the Upper Yangtze and the Northwest) and three industries (tea, salt, and the local commodities trade).^[54]

In accordance with Wang's prescription for giving men power on the basis of their success and not their rank—which Wang's critics lambasted as the employing of mean and petty men—the stipendiary ranks of the early

[54] For sources on the careers of the THA intendants, see Smith, Taxing Heaven's Storehouse , Appendix C, pp. 327-31.

intendants at appointment were almost uniformly low: between 8b and 7b, with one exceptional 6b (Liu Tso). More important, the court acknowledged that the intendants stood at the bottom of the civil service hierarchy, and took steps to counterbalance their low status. In 1080, for example, the court granted a new assistant intendant a titular commission so that despite his low rank he could "awe the prefects and county magistrates he would encounter." And in 1083 the emperor himself ordered that "although the ranks of the (THA intendants) are low, they must be permitted to wield authority as they see fit."^[55]

The Ever-Normal Granary intendancies were filled by much the same sort of men, and indeed two (Yen Ling and Ch'eng Chih-shao) went on to serve as THA intendants in the postreform era. I have identified seventy-one men appointed intendant or co-intendant in one of the twenty-two circuits or the capital from between 1069 and early 1086. Obviously, so relatively large a sample will show much greater variation than the seven intendants of the THA, but a preliminary analysis of the seventy-one cases reveals largely the same pattern of technical skills, New Policies affiliation, and low rank. One of the first twelve men sent out to implement the farming-loans directive of the fourth day of the ninth month of 1069, for example, was Wang Kuang-lien, whose recommendation as managing supervisor of the Hopei fiscal intendancy had brought the farm-credit idea to fruition. Wang was made Ever-Normal Granary intendant for Hopei that very month, and a year later was given the concurrent post of judicial intendant as well.^[56] Ch'en Shih-hsiu, sent by emergency relay to manage the Ching-hsi-Huai-nan water conservancy in 1070, was promoted in place to Ching-hsi Ever-Normal Granary intendant around 1072, where he continued to advise on irrigation.^[57] And Wang Ku, who as registrar of the Court of Agricultural Supervision was despatched to investigate natural disasters in Liang-che Circuit in 1075, was subsequently kept on as granary co-intendant, then transferred to the intendant's post in K'ai-feng the following year.^[58]

If anything, the reformers reached further down the administrative hierarchy for their Ever-Normal intendants than for the THA. At least four men were promoted to intendancies from posts as county magistrate, including Court of Agricultural Supervision chief Lü Hui-ch'ing's younger brother Lü Wen-ch'ing. Moreover, at least three men were appointed to the

[55] SHY:CK 43/65a; HCP 341/2a.

[56] SHY:SH 4/17b; HCP 214/7a. At least four other men held the two posts concurrently.

[57] HCP 215/12a; HCP 233/18a. The full set of Ever-Normal responsibilities included, in addition to the granaries, management of service-exemption accounts, water conservancy, ferries, and in some areas state trade. (SS 167/3968.)

[58] In 1081 Wang Ku was fired as Ever-Normal Granary intendant of Ching-hsi circuit for criticizing the granary policy. HCP 269/22a; 275/10b; 313/7b-8a.

position of intendant from executory-class posts as civil aides. In almost every case, however, the men who were jumped from county or local government posts had distinguished themselves in endeavors related to granaries or farm loans. Finally, the stipendiary ranks of the forty-one appointees for whom the datum is available were even lower than those of their THA counterparts: one at 9a, ten at 8b, twelve at 8a, nine at 7b, six at 7a, and the three still in executory class.^[59]

At entry, then, there was little to distinguish the men appointed to the Ever-Normal and THA intendancies. But in the ways they were used in office the differences were more pronounced. The divergence begins with length of term. Wang An-shih prescribed long terms, to give incumbents full opportunity to see their projects through to completion, since in his view office holding was a mandate for experimentation. The THA intendants exemplified his theory fully: until reform opponents set a maximum term of thirty months for Agency chiefs in 1086, there was no statutory limit on the length of time an intendant served. The early intendants were kept in office as long as they kept innovating, their ambitions fed with imperial awards, greater authority, and promotions in rank. Thus the average term of office for the New Policies THA intendants was a remarkable forty-nine months, a figure that is only slightly skewed by P'u Tsung-min's presence in Ch'eng-tu for the entire eleven years. Moreover, THA intendants rarely went on to other posts. One of the first seven intendants was transferred out in an unexceptional manner, and a second was impeached when his new salt-monopoly venture went awry. The others served until death, illness, or the Restorationist policy reversal intervened; and of these the only man to return to office in the postreform era—Lu Shih-min—added another sixty-five months as THA intendant to his original sixty-six when Emperor Che-tsung revived the entire apparatus on reaching his majority in 1094.

Long tenures promoted the development of the local and intra-agency networks indispensable to decentralized bureaucratic entrepreneurship. But they also posed the risk that power might be abused, or that longtime specialists might monopolize the performance of critical tasks. My information on the tenures of the seventy-one Ever-Normal Granary intendants is not precise enough to yield a representative average, but it does not appear that they were unusually long. There were some exceptions: Fang Tse, an affinal relation of Lü Hui-ch'ing's, was in Chiang-hsi for sixty months; Chao Hsien

[59] HCP 246/1b; 267/13a; 271/13a; 305/6b; 307/4a; 347/1a. Family connections clearly played a role in some of the intendancy appointments, and at least fourteen intendants had family connections to more prominent faction members—though in Han Cheng-yen's case his uncle, Han Ch'i, was one of the most vehement antireformers. Three of the seven early THA intendants were also related to prominent political partisans, and in two of these cases (Lu Shih-min and Fan Ch'un-ts'ui) their relations were also against the reforms.

was in Ho-tung for forty-nine months; and Liu I had the same tenure in Chiang-nan. But most of the identifiable terms vary between fifteen and twenty-five months. And in contrast to the powerful but futureless THA intendancy positions, a term as Ever-Normal Granary intendant often heralded the start of a typical specialized career in finance. The ENG intendancy was inserted into the bottom of the rotation of circuit posts and was followed by promotion to fiscal staff supervisor (twelve cases), vice fiscal intendant (four cases), judicial intendant (four cases,) fiscal intendant (one case), or one of the new reform-era circuit posts (four cases).^[60]

But Wang An-shih himself had recommended a significant variation of long tenure in 1071, when he defined it as promotion to higher intendant positions within the same circuit. Wang specifically cited the case of Ch'en Chih-chien, who served serially in Ching-hsi as Ever-Normal Granary intendant, provisional fiscal staff supervisor, and provisional vice fiscal intendant in the early 1070s. Court of Agricultural Supervision chief Ts'ai Ch'ueh endorsed Wang's plan in 1078, recommending that successful granary intendants be promoted to judicial intendant of the same circuit. At least four granary intendants in addition to Ch'en fit this new pattern of regional as opposed to functional specialization, which was to become typical of the THA intendants in the Southern Sung, when recentralization and decreasing productivity led to severe curtailment of the Agency's autonomy. In both cases this variant of long tenure allowed the court to benefit from a man's knowledge of a region but retain centralized control over his performance, by preventing him from exploiting a particular office for personal ends yet not sacrificing his close familiarity with the problems and resources of his assigned region.^[61]

Rights of Personnel Selection

Differences between the THA and the Ever-Normal Granaries become starker still when we move from the characteristics of the intendants to the powers they were allowed. One of the most critical prerogatives for any bureaucratic agency, stressed by Wang An-shih in his Myriad Word

[60] HCP 267/9b-10a; 285/5b; 286/6b; 307/4b; 329/22b; 324/1a-6b; 372/15a-b. See Hart-well, "Financial Expertise," pp. 286-89.

[61] HCP 212/5b; 229/7a; 236/6b. As Robert Hartwell has shown, intraregional specialization in diverse careers characterized Southern Sung office holding generally. See "Transformations of China," p. 400. Where the government sought to maintain tight surveillance over circuit officials, it could array one official against another through a system of checks and balances that Winston Lo has referred to as functional interchangeability and overlapping jurisdiction (Lo, "Circuits and Circuit Intendants," pp. 86-91). At the same time, by rotating individual officials through the entire complement of circuit offices in each region the government could capitalize on each man's familiarity with local conditions and feed a variety of informed opinion into the process of regional decision making. For further discussion, see Smith, Taxing Heaven's Storehouse , pp. 212-16.

Memorial, is the right to appoint its own personnel. In order to further Agency success in the procurement of tea, horses, and above all revenues, the THA intendants were granted selection privileges that earned for the Agency unique status as "an organization permitted to appoint its own subordinates." And because the Agency's revenues kept growing, the intendants retained their selection rights over some forty-five field administrators and one hundred to two hundred petty officials even after the personnel privileges of virtually all other government agencies were rescinded in 1081.^[62]

The autonomy of the THA over its own personnel selection was enhanced by extraordinary access to the special "right of appointment in contravention of regular procedure" (pu i ch'ang-chih chü-pi ), and the privilege of coselecting with the fiscal intendant the magistrates of key tea-production and transport counties. The right of irregular appointment empowered administrators to place men with special skills and experience in posts for which they lacked the specified civil service rank, and to fill critical vacancies or shuffle personnel without awaiting court approval. Agency intendants successfully argued that irregular appointment was crucial for preserving the profitability of the tea and horse enterprise, and from 1077 on, when all but the most limited use of the privilege was prohibited elsewhere in the government, it became standard practice in the THA.^[63] The right to coselect the magistrates for at first four and in later periods up to ten key production and transport counties turned tea procurement and transport into a central concern of local government, considerably extending the effective authority of the Agency. And finally, THA selection powers were enhanced by Wang An-shih's "law for appointments and transfers to distant offices" (yuan-kuan chiu-i chih fa ), which allowed Szechwanese to request multiple terms of office in prefectures close to their family residences, and offered ambitious and still unsettled Szechwanese the quid pro quo of home service and bureaucratic advancement in return for taxing their own region on behalf of the state.^[64] Although only two of the first seven intendants were Szechwanese, at least one of them was always in office, ensuring a continual Szechwanese presence in Ch'eng-tu. At the staff level, despite a complaint from a later period that the intendants appointed only Szechwanese to subordinate positions, the only formal restriction imposed on them was that in the case of dual offices—for example, a single market with two inspectors—only one appointee could be Szechwanese.^[65]

Unlike the Tea and Horse Agency, which developed into an autonomous, fully staffed enterprise, the individual Ever-Normal Granary in-

[62] SS 167/3969. On the recentralization of appointments, see WHTK 38/361b-c.

[63] For a sample of decisions, see SHY:hsuan-chü 28/10b-12a.

[64] HCP 214/21a-22b; SS 159/3721-24; Lo, "Circuits," pp. 77-79; Smith, Taxing , pp. 105-7.

[65] SHY:CK 43/102b-103a.

tendants served as a new level of management set between center and locale to coordinate procedures actually performed by regular government functionaries. At the prefectural level one official, usually the vice-administrator, or two (including a civil aide) in prefectures of ten counties or more, supervised overall granary and farm-loan operations. In the subordinate counties, either the magistrate or the county finance official among the civil aides directed the distribution and collection procedure. The county was the key link in the system, and a county magistrate could bring the operation to a halt, as did Ch'en Shun-yü in Shan-yang county (Huainan) in 1070, when he blocked distribution of the Green Sprouts loan funds and then impeached himself in protest against the policy. Prefectural and county officials in turn presided over a clerical staff specially assigned to granary functions that was not to exceed three per prefecture and two per county.^[66]

To help them motivate their subordinates the ENG intendants were given relatively liberal sponsorship privileges: as of 1077 each intendant could recommend nine men yearly for promotion from the executory to the administrative class. But their direct appointment rights were quite limited. On the whole, ENG intendants were merely allowed to recommend, in conjunction with the fiscal intendant, men already in a prefectural or county-level post for assignment in place to a granary-related task.^[67] The THA intendants in contrast routinely created new posts to which they could recommend men currently out of office, giving them substantial powers of patronage.

When the Granary intendants, and even their director in the Court of Agricultural Supervision, were allowed to make new appointments, the central government maintained a close watch. For the rural credit policy authorized officials to hand out a substantial portion of the national treasure in the form of loans, and the government was always concerned about getting its treasure back. Nor was it easy to keep track: in 1075 officials in K'ai-feng county neglected to record loans of nonglutinous rice worth 60,000 strings of cash and overlooked the collection of another 40,000 strings of cash, all because county officials did not know how to keep the Green Sprouts' books.^[68] In such emergencies the Court of Agricultural Supervision despatched its own officials to teach registration procedures. But for new appointments of court executive assistants and ENG granary custo-

[66] SHY:SH 4/16b; HCP 250/4b; 289/15a; 212/7a-b; 249/6b. The same number of clerks again were deployed for the agricultural irrigation policy.

[67] SHY:SH 4/16b; HCP 250/4b; 282/2a.

[68] HCP 267/13b-14a; 269/2b. The Court of Agricultural Supervision was under orders to keep track of the volume of cash loans in circulation and collected; the volume of grains collected and in arrears; the number of defaulters; the volume of grain purchased and sold, and at what unit prices; the volume of relief aid; and the total interest earned on loans. HCP 214/28b.

dians the court director and the circuit granary intendants were ordered to choose men who were qualified in the first place. In particular, only officials who had already served as county magistrate and had passed their annual merit rating with the very highest honors, or in the absence of such luminaries men whose comportment was exceptional and seniority high, were to be considered for the two job categories.^[69] It is obvious that when it came to distributing the state's reserves, the government's chief decision makers wanted stolid, upright managers, not entrepreneurs.

Operational Autonomy

The central design of Wang An-shih's action strategy was to let bureaucratic entrepreneurs wield power freely in crafting innovative solutions to New Policy tasks. In its hunger for tea monopoly revenues, the New Policies court granted the THA intendants uncommon authority to disarm their rivals, protect their output in tea and funds from arbitrary expropriation by other agencies, engage in ancillary businesses, and above all to expand their marketing region.

In the course of the New Policies at least six prominent Szechwanese, nine regional financial administrators, and four central-government officials were fined, demoted, or cashiered for questioning Agency practices; in 1078 a blanket order prohibiting other government bureaus from interfering with the Agency was issued. With its critics neutralized, the Agency was free to consolidate and expand its operations. In order to protect its profitability and to ensure that its profits were all counted toward the promotional incentives and cash bonuses that propelled the enterprise along, organizations receiving THA cash or tea had to follow strict accounting and payback regulations established by the intendants, with all extraquota subventions registered as net sales.^[70] And to put its surplus funds to the most profitable use, the intendants were permitted to enter a variety of sideline businesses. In the northwest, Agency officials bought and sold Inner Asian valuables, traded in grain, and exported Szechwanese silks, while in Szechwan they sold Shensi salt and dealt in local textiles, paper, and medicinal herbs and made loans to tea cultivators and wineshop operators at 20 percent interest. The THA brought under its own umbrella operations that in other regions were shared by the salt monopolies, Ever-Normal granaries, and bureaus of state trade.

But the most important component of Agency entrepreneurship was its freedom to create new markets for Szechwanese tea. The geographic expansion noted earlier turned 7.5 million new domestic consumers and much of Inner Asia into drinkers of Szechwanese tea, giving the Szechwanese indus-

[69] HCP 280/21b; 283/13b.

[70] See Smith, Taxing Heaven's Storehouse , pp. 268-77.

try the largest marketing region it was ever to have. The market possessed all the instability of monopoly markets, of course, and when chief minister Ts'ai Ching sought out new markets for his remonopolized southeastern tea in 1102, bureaucratic competition dispossessed Szechwan of two-thirds of the extraregional tea drinkers that bureaucratic entrepreneurship had first obtained. But during the high tide of activism under the New Policies, the THA was given a free hand.

The THA was organizationally autonomous. It answered to the Secretariat-Chancellery, and after 1080 to the Ministry of Finance, but it was directly subordinate to neither. Though located at the same level in the administrative hierarchy, the Ever-Normal Granary intendants were all subordinate links in an apparatus directed centrally by the Court of Agricultural Supervision. Consequently they were much more tightly controlled.

The farming-loans policy prompted an obstructionist movement led by such elder statesmen as Fu Pi and Ou-yang Hsiu, who from the prefectural posts to which they were rusticated for opposition to the reforms openly and systematically encouraged county officials to disobey the loan directives. But even after the obstructionist movement was silenced in 1071, Shentsung was so shaken by the insurrection of his most prestigious advisers that the farming-loans policy was kept in constant view, and the granary intendants were under continual surveillance.^[71]

The Ever-Normal Granary intendants were subject to both horizontal and vertical control. As one of a cluster of circuit-level officials responsible for regional administration, the granary intendant was bound by a system of checks and balances that Winston Lo has described as functional interchangeability and overlapping jurisdiction. For example, both the fiscal and the judicial intendants had oversight responsibility for the credit policy, primarily to prevent forced loans; and in 1073 all fiscal intendants, fiscal staff supervisors, and judicial intendants were ordered to comanage the Ever-Normal Granary system. Conversely, the granary intendant also served to keep his colleagues in check—for example, by reporting on the antireform activities of fiscal intendant Fan Ch'un-jen in Ch'eng-tu in 1071, or investigating local finances and the salt administration in Ching-hu in 1084. But unlike their peers in the THA, granary intendants were discouraged from expanding into the activities of other circuit agencies on their own initiative, and at least three Ever-Normal Granary intendants were fined or summarily dismissed for exceeding their proper authority (yueh chih ).^[72]

[71] The obstructionist movement and the debates it inspired are documented in detail in CPPM 68, passim.

[72] Lo, "Circuits," pp. 86-91; SHY:SH 4/18b-19a; HCP 220/22b; 244/5a; 279/13b; and HCP 224/21b-22a; 348/15a. Mutual surveillance also meant mutual responsibility. In 1084 the granary intendant, vice fiscal intendant, and a prefect, vice-prefect, and managing supervisor in Kuang-nan East were summarily dismissed and demoted for a scandal and cover-up in the salt administration. HCP 345/5a. The ENG intendants fined or dismissed were Chang Shang-ying and Chang Wan in 1079, and Li Hsiao-po in 1080. HCP 287/2a; 299/21a; 307/4a. Court of Agricultural Supervision director Ts'ai Ch'ueh launched a move in 1078 to free his granary intendants from outside interference, and in particular sought to protect ch'ang-p'ing funds from encroachment by the fiscal intendant. But Ts'ai was promoted to the Council of State in 1079, and under his successor Li Ting the campaign died down. HCP 292/2b; 293/8b; 298/8b-9a.

Finally, because they were responsible for so wide a range of critical and delicate activities, the Ever-Normal Granary intendants were also subject to ad hoc surveillance from above. At least seven circuit granary operations in all parts of the empire were put under outside investigation between 1071 and 1076, for problems that included illegal collections, negligence, ineffective disaster relief, and inadequate water-conservancy projects.^[73] Constrained both laterally and from the top, Ever-Normal Granary intendants were subjected to a degree of centralized control that was not extended to the Tea and Horse Agency until the middle of the following century.

Choosing Centralization Strategies

Both the Green Sprouts farm loans and the Szechwan tea monopoly grew out of an activist effort to replace private economic actors with public agencies in order to revitalize the economy and generate new revenues for the state. And both were administered regionally by men chosen according to the same entrepreneurial criteria. But while the THA intendants were encouraged to use bureaucratic power in an experimental, entrepreneurial way, their counterparts in the Ever-Normal Granary system were tightly controlled. In the case of the Green Sprouts loan policy the entrepreneurial mandate was located much higher up the administrative hierarchy, in the Court of Agricultural Supervision. The farming loans act promoted state economic control through increased centralization; the Szechwan tea monopoly pursued the same goal through decentralization. How can we understand the choice of two such different strategies for the same goal?

To begin with decentralization, what might be called the frontier factor stands out. Distance alone would recommend for frontier regions a degree of decentralization that would be unnecessary closer to the capital. Transport distance from K'ai-feng to Szechwan rendered all transfers of information, men, and goods costly and time-consuming, making delegation of broad discretionary authority to regional-based agencies the most effective way to govern. And since the tea and horse trade yoked the Shensi economy so strongly to Szechwan's, once fiscal authority had been decentralized in

[73] Investigators included Shen Kua, Chang Tun, Liü Sheng-ch'ing, P'u Tsung-meng, Hsiung Pen, and Hsu Hsi. See HCP 227/11a; 236/2b; 237/14a; 245/2a-b; 247/3a; 248/4b; 271/ 18a-b; 272/6b.

Szechwan, it became advantageous to delegate it to the same agency in Shensi as well. Hence, from 1079 to 1082 Li Chi not only headed the THA but served concurrently as Shensi fiscal intendant.^[74] Closer to the capital, in the center of the political network, so great a delegation of power would have been both imprudent and unnecessary.

Szechwan's political history also commended decentralization as a useful strategy for enhancing state control. Because it was culturally and geographically isolated from the political heartland, Szechwan had proved relatively impervious to centralized taxation policies since the mid-T'ang. But the rebellions of Li Shun and Wang Hsiao-po had fractured multigenerational patron-client bonds, forcing Szechwan's magnate classes to turn to the state for protection and opening them to political recruitment. From the early eleventh century on increasing numbers of Szechwanese adopted bureaucratic careers as their primary mobility strategy, and bureaucratization of the Szechwanese elite gradually fostered integration of the region into the national polity. But even after the rebellions Sung administrators were wary of laying too heavy a hand on the Szechwanese economy: for most of the eleventh century, for example, Szechwan provided only 5 percent of the empire's twice-yearly tax quota, or about half of its percentage of the national population.

With the fiscal and military crises of the 1060s, however, the state felt compelled to increase its claims on the Szechwanese economy. Wang An-shih's "law for appointments to distant offices" enabled the state to maximize fiscal control in Szechwan by recruiting the growing pool of technically skilled and ambitious Szechwanese of low rank, whose chief hope for mobility lay in successful pursuit of a bureaucratic career, into serving in their own region as agents of the state. By grafting native recruitment onto the strategy of bureaucratic entrepreneurship, reform policymakers decentralized operational and personnel authority to the men who had most to gain from state economic activism and thus acquired for the state unprecedented access to Szechwan's surplus product.

But the efficacy of enhancing state revenues by decentralizing entrepreneurial power in distant regional agencies does not explain why, among all other possibilities, power flowed to the Tea and Horse Agency. One part of the explanation has to do with the degree to which the THA satisfied goals viewed as critical by the center. The sociologist James Thompson, in his useful and insightful book Organizations in Action , has supplied an axiom that links an organization's power to its operational success. Thompson writes that "an organization has power, relative to an element of its task environment [those parts of the environment that are relevant to goal setting

[74] At the same time the superintendant's headquarters was transferred from Ch'eng-tu to Ch'in-chou. HCP 297/16b; 299/12b; 330/9a-b.

and attainment], to the extent that the organization has capacity to satisfy needs of that element [in our case state policymakers] and to the extent that the organization monopolizes that capacity." During the New Policies era the state's greatest needs in Szechwan and the Northwest frontier were for tea, horses, and revenues, and the THA was a generous (and in the case of tea and horses a monopolistic) supplier of all three. In order to keep tea, horses, and revenues flowing, policymakers granted THA intendants extraordinary power and autonomy. But however much the THA might resemble a private monopolistic enterprise, ultimately its power flowed from the state. And to the degree that changes in the task environment—particularly changes in policy objectives and marketing opportunities—undercut THA effectiveness, the state withdrew its support. In 1102, for example, Ts'ai Ching's reclamation of eastern Shensi for southeastern tea in 1102 created a surplus of Szechwanese tea and a doubling of the price ratio between tea and horses; and the Jurchen conquest of North China in 1127 deprived the THA of not only the rest of its extraregional markets, but most of its horse suppliers as well. Each shock progressively undermined the THA's capacity to supply the revenues and horses on which its power had been built, prompting successive administrations to curtail that power and reimpose a higher level of control. By the middle of the twelfth century, although Szechwanese regional administration was even more decentralized than under the New Policies, the THA itself had been reduced to but one component of a functionally overlapping fiscal network subordinated to the new post of viceroy (hsuan-fu shih ).^[75]

Organizational success, in short, was rewarded with organizational power, and the organization that ceased to be successful found its powers increasingly diminished. From this perspective we can understand why the ENG intendants were never granted the degree of autonomy enjoyed by their THA peers: for the Ever-Normal Granary operation was charged with a set of mutually contradictory goals that drew it in conflicting directions and made it impossible for the policy and its administrators to achieve the unambiguous success registered by the Tea and Horse Agency.

On the one hand, the ENG intendants were under great pressure to generate revenues through rapid circulation of the loan fund. Circulation was measured by quotas for distribution and collections, inspiring such abuses as forcing loans on the basis of household registration; issuing new loans to households with unpaid prior debts; and overzealous collection of back debts. On the other hand, the ENG intendants presided at the same time over the state's primary mechanism for relieving a peasantry that both reformers and reform opponents agreed was stretched to the limit by nat-

[75] Thompson, Organizations in Action , pp. 30-31. For the decline in THA powers, see Smith, Taxing Heaven's Storehouse , pp. 191-217.

ural disaster and (for the reformers at least) inequalities of wealth. Thus even as it pushed the ENG intendants to deliver profits from the farming loans, the central government had to monitor them carefully to ensure that grain reserves were adequate to meet emergency needs, that relief grain was in fact distributed, and that loan collections did not dispossess poor peasants.^[76] The inconsistencies that stemmed from this tangle of conflicting goals are illustrated by the debate over loan eligibility and the problem of deferred repayments (i-ko ).

The question of loan eligibility placed the reformers in a vulnerable ideological position. As a social welfare measure, the Green Sprouts policy was intended to "suppress engrossers and relieve the poor and weak." But as a means of generating revenues, the Green Sprouts loan fund had to be circulated widely and to households that could be counted on (or forced) to repay. Consequently, in early 1070 the Finance Planning Commission pegged loan limits to the household grading system and extended eligibility to virtually all resident (as opposed to vagrant, fou-lang ) families. The new order entitled grade 5 and guest households to 1,500 cash; grade 4 to 3,000 cash; grade 3 to 6,000 cash; grade 2 to 10,000 cash; and grade 1 to 15,000 cash. The order then added that "if there are remaining funds, county officials are authorized to assess the situation and offer additional loans to households of grade 3 and above; if there are still surpluses, then interested households from the urban wards and suburbs with property and businesses to offer as collateral will be allowed to form into five-family guarantee groups and obtain loans under the Green Sprouts statutes."^[77]

Han Ch'i immediately attacked the new measure for providing public loans to the very same interests—the engrossers—it was purportedly seeking to suppress:

It is the rural households of grade 3 and above and the propertied urban and suburban households that have heretofore been "engrossing families." Now they are all given loans and charged 1,300 cash for every 1,000. The government is simply chasing interest payments. The measure absolutely contradicts the stated intent of the policy to suppress engrossers and aid those in need.^[78]

[76] HCP:SP 7/la; CPPM 68/2162; HCP 252/27a-b; 254/1b-2a. On new loans to already indebted households, see Ou-yang Hsiu's charge in HCP 211/12a-16a. In the spring of 1071 Han Ch'i complained from his post as vice-administrator of T'ai-ning-fu that the Hopei ENG intendant refused to issue grain from the loan fund for the hungry poor, insisting that they be fed from the sheng-ts'ang instead. In response Shen-tsung warned the fiscal, judicial, and ENG intendants against withholding poor relief and causing vagrancy. HCP 221/la-b. In 1074, Shen-tsung worried that although loan revenues were substantial, over 70 percent of the system's resources were in circulation at a time when widespread disaster made adequate relief-grain stores essential. The granaries were ordered to keep half their resources in reserve. HCP 256/15a-b; 272/7a-b.

[77] Wang An-shih to Shen-tsung, 1070/2, in SHY:SH 4/20b; and SHY:SH 4/19a, as quoted by Han Ch'i in 1070.2.1.

[78] SHY:SH 4/19b.

Han's charge apparently caught the reformers off guard and forced them to defend their position publicly and before the emperor. Shortly thereafter Wang An-shih added his personal response to an official rejoinder by the Finance Planning Commission:

Rural households of grade 3 and higher as well as wealthy urban and suburban households also encounter periods of financial trouble and then have to incur private debts. How can they all be "engrosser families"? At the present time, after making loans to poor families the surplus is loaned to these (wealthier) grades, so they too are not forced to contract loans at 100 percent from private houses. This is precisely within the intent of the original imperial decree to repress engrossers.^[79]

But Han Ch'i utterly rejected the official response:

Not only does this official know that [these grade 3 and wealthy urban and suburban households] are "engrossing families"; everyone under heaven knows it! The only reason the Finance Planning Commission denies that they are engrossing households is that it wants to push even more Green Sprouts funds on them to get back still more interest.^[80]

We need to know more than we know now about Sung rural society to judge whether Han Ch'i's condemnation of all upper-grade households as "engrossers" is accurate. But as a political tactic Han's attack was brilliant, for it highlighted the essential contradictions between the social-welfare and revenue-generating components of the Green Sprouts policy. Though Han did not succeed in reversing the Green Sprouts measure, he did force Wang An-shih and the Finance Planning Commission into lengthy public justification of their policy of universal eligibility.^[81]

The problem of deferred payments cut directly to the heart of the conflict between social and fiscal objectives. The Green Sprouts directive of 1069 specified that repayments be made after the summer and fall harvests of the year of loan issue, but allowed extensions of one harvest period in the case of disaster or poor crops. Subsequent orders extended the repayment period up to two years, but even with the more liberal provisions there were always borrowers who could not meet their obligations. Debtors and their guarantors were subject to confiscation, first of their collateral and then of their immovable property. As early as 1071 the Court of Agricultural Supervision obtained permission to funnel the returns from debt sales of property back into the ENG capital fund.^[82]

But even perfectly legal debt collections and property seizures could intensify rural disintegration, and in periods of endemic disaster the Court of

[79] SHY:SH 4/23a.

[80] SHY:SH 4/27b; also in Han's Chia-chuan , cited in HCP:Shih-pu 7/26a-27b.

[81] See, for example, SHY:SH 4/22b-24a, 24b-25b.

[82] SHY:SH 4/16b; HCP 228/7b; 279/23a-b; 294/8b-9a.

Agricultural Supervision was pressured to grant payment deferrals. Yet deferrals created their own set of problems. First, a debtor under deferred status was presumed to be a bad risk and was legally prohibited from receiving further loans. But denying a loan amounted to withholding essential agrarian relief: in 1074 the administrator of Chi-chou, in Hopei, predicted that the denial of Green Sprouts loans to deferred planters of spring wheat would diminish the following season's harvest as well as force farmers to go hungry. The court immediately ordered the ENG intendants of five northern circuits to grant interest-free loans of 1,000 cash to all spring-wheat farmers below household-grade 3, even if they had deferred debts. Two years later, however, the emperor censured granary officials for issuing loans to deferred debtors, accusing them of pursuing profits and good merit ratings at the expense of prudence.^[83]

In addition, collecting from deferred debtors raised delicate political as well as practical problems. In 1075, for example, the Court of Agricultural Supervision got permission to set the summer of 1076 as the deadline for repaying loans issued in Liang-che, Huai-nan, and Chiang-nan two years earlier, agreeing to accept labor service in exchange for money or goods. By the summer of 1077, however, collections had still not progressed, and the Secretariat-Chancellery this time sent its own agents to press for repayment following good harvests in Liang-che and Huai-nan. The Secretariat's collections were halted again the following spring for the two lowest household grades; and in 1079 the Court of Agricultural Supervision was in turn forced to accept another extension for the poorest households in Liang-che. Predictably, the deferrals, politically essential though they may have been, were financially disastrous: loan collections were 13 percent off quota in 1080 and fell 1.8 million strings into the red a year later. Moreover, the politically unavoidable deficits spurred even tighter control, and in 1083 the Finance Ministry was ordered to lead an investigation of the ENG intendancies. In contrast with the THA, whose prodigious fulfillment of a consistent set of policy expectations protected it from outside interference, the inability of the Ever-Normal Granary system to satisfy an intrinsically contradictory set of social and fiscal objectives provoked suspicion and careful scrutiny from its superiors as well as its opponents.^[84]

Finally, the Green Sprouts policy was centralized at a higher level and watched more closely than the Szechwan tea monopoly because direct state intervention in the agrarian economy, and efforts to get more out of the peasantry than custom and direct taxation warranted, kindled greater controversy than state control of commerce. In contrast with the Szechwan tea monopoly, critics of the Green Sprouts policy rejected the fundamental

[83] HCP 258/17a; 272/2b.

[84] HCP 268/18a-b; 283/10a; 288/11b; 289/13a; 292/3b; 297/1b; 332/10a-b.

principles and goals of the policy and occupied positions throughout the government, including the Ever-Normal Granary administration itself.

Among the many criticisms leveled at the policy, three in particular shed light on conservative fiscal views. Opponents rejected the theory of poverty on which the policy was justified; they criticized the dangerous reliance on money that the farming loans promoted; and they argued that the state was a far harsher creditor than private lenders could be.

As we have seen, the architects of the Green Sprouts policy blamed rural poverty on the powerful "engrossers," who could turn the endemic crises of agrarian production to their own advantage. But critics of the policy blamed rural poverty, not on exploitation, but on the spendthrift habits of the poor. In his polemic against the Green Sprouts policy and the Finance Planning Commission of 1070, for example, Ssu-ma Kuang outlined a model of rural society that portrayed wealthy families as a safety net for the poor:

The reason there are poor people and rich people is difference in talent, nature, and intelligence. The rich are more intelligent, their concerns deeper and their thinking more long-range, and they gladly work their muscles and toil their bones, dress plainly and eat sparingly. And to the end they do not want to borrow from others. Thus their households normally enjoy a surplus, and they do not go to the wolves [become hopelessly dependent]. Now poor people are timid and weak and grasp whatever they can rather than working industriously; they don't think of the long term, and if in one drunken day they get rich, soon there will be nothing left; when they are hard pressed they borrow from other people, and as the debts accumulate they are unable to repay them, until they reach the point that they must sell their wives and children. . . . Thus it is that the rich often make loans to the poor in order to enrich themselves, and the poor often borrow from the wealthy in order to survive. Though bitterness and happiness are not equally distributed, at least [rich and poor] mutually aid one another, in order to guarantee their livelihoods.^[85]

Anticipating Tung Wei's market-oriented theory of famine relief documented in chapter 7 in this volume by Robert Hymes, Ssu-ma Kuang identified wealthy rural families as the foundation of a stable order, and profit as the lure that moved the wealthy to aid the poor. For Ssu-ma Kuang, any policy that sought to displace the rural rich threatened to undermine society and the state.^[86]

[85] SMWKWC 7/3a, "Ch'i pa T'iao-li-ssu ch'ang-p'ing-shih shu."

[86] For Tung Wei's views, see chapter 7. Ssu-ma Kuang's projection of the impact of the Green Sprouts policy on the rich is worth quoting at length: "Now county officials [are themselves] issued interest-funds to lend to the people in spring and fall. None of the wealthy people want the loans, but the poor people do want them. Because the Intending Officials want to distribute [loans] on a wide scale in order to accumulate merit, they do not inquire into people's wealth or poverty, but just force loans on them according to their household grade. The wealthy are assigned relatively large debts, the poor somewhat smaller. Large loans go up to 15 strings, small ones not less than 1,000 cash. Prefectural and county officials and clerks fear [getting stuck with] the responsibility of absconders and defaulters, so order the poor and rich to array themselves together in pao-chia guarantee groups. The wealthy are made the chiefs (k'uei-shou ), the poor obtain money, and in no time at all the money is gone. In the future if the millet or wheat [harvests] are small or don't come up, they cannot even pay their twice yearly tax, let alone the interest payments. Since they cannot repay, the fears of the clerks are spread to all four quarters. The wealthy cannot avoid it, and must themselves repay the debts of the many households" (SMWCC 7/3b).

Ssu-ma then outlines a scenario for universal immiseration and disaster: "Once the poor are completely exhausted, the wealthy will also become poor; your minister fears that after ten years time there will not be many wealthy. And once the wealthy are completely exhausted, if the nation (kuo-chia ) should have the misfortune of a border emergency, and have to raise many troops, then from whom will the monies for grain, cloth, and military provisions be raised?" (ibid., 3b).

Prominent critics also charged that the farming-loans policy changed the calculus of rural poverty by first pumping unnecessary money into the countryside, then siphoning it out again through monetized collections of fees and taxes. In 1074, for example, Feng Ching, second privy councillor, asserted that when the people of Hsiang-fu county, K'ai-feng, hear that the government is distributing loan funds they put up whatever they own as collateral to get money that they do not need: "They just see that the government is giving out money, and there are none who do not want some; so they pile up debts, and when the time comes they cannot repay."^[87] Cash loans were seen as an inducement to debt and extravagance. As Su Shih wrote in a retrospective assessment of the policy:

Peasant households balance expeditures against income and economize in clothing and food, so that even if poor they still meet their basic needs. But when peasants can get more money than they need their expenses naturally increase, and there is nothing they will not do [to get still more].^[88]

Having habituated the peasants to money, however, the government made money increasingly hard to get by drawing it out of circulation through loan collections, the service-exemption fees, and inequitable exchange rates for the payment of taxes. Ssu-ma Kuang called attention to the problem in 1074, alleging that monetized payments forced the peasantry to sell everything they had to obtain the one thing they lacked—cash—in order to pay the state. The currency shortage undermined the farming loans' original objective as a rural credit mechanism. For, as Wang Yen-sou argued in 1086, when it came time to repay their loans poor peasants once

[87] HCP 252/8a.

[88] HCP 384/10b. Cf. Shang-kuan Chün's assessment, HCP 378/18b.

again had to borrow from the "rich families and great clans" at inflated interest rates to obtain the cash they needed to pay off the state. Thus the Green Sprouts policy had come full circle: "established in the name of suppressing the engrosser households, it wound up giving them great assistance."^[89]

Moreover, the state had far greater power to press its debtors than private lenders had. In 1069 Ssu-ma Kuang had predicted that although the rich could merely "nibble away" at the poor, county clerks and officials would come down on the peasantry with the irresistible legal and police powers of the state. According to Shang-kuan Chün's appraisal in 1086, although private lenders had charged nominally high interest rates, they had at least been flexible and relatively benign in their capacity to exact repayment. By the time a peasant finished paying off the ruthlessly efficient agents of the state, on the other hand, the price of a loan had risen to 50 or 100 percent. Wang Yen-sou graphically illustrated the monetary and physical cost of a government loan: first, in order to register for a loan, the guard chief, guarantee-group head, and notary for household registration all had to be paid off. Then, if at collection time there was not enough money to pay, when the government clerk appeared at the door he had to be bought off with food and drink. But the clerks could not be bribed endlessly, and for tens of thousands of debtors the inevitable day arrived when they were dragged off to be beaten and their property was seized. Nor did Wang Yen-sou assume that the Green Sprouts policy was intentionally malicious. The problems were simply the unavoidable consequence of government interference in the rural economy. "To strive intentionally for the benefit [of the peasantry]," he concluded, "is not as good as bringing them benefit by leaving them alone."^[90]

It should be evident from the discussion to this point that it was not merely a small circle of conservatives who rejected the rural credit policy; Green Sprouts critics were located throughout the government, constituting a vocal and persistent base of opposition to the measure at all levels. At the court level, complaints by such elder statesmen as Ssu-ma Kuang, Han Ch'i, Ch'en Sheng-chih, and Tseng Kung-liang kept Shen-tsung uncertain of the policy's merits and sensitive to its potential excesses and embroiled Wang An-shih and the Green Sprouts measure in continual and divisive court controversy. In 1070 Ssu-ma Kuang, against the emperor's wishes,

[89] SMWCC 7/10a-b, "Ying-chao yen ch'ao-cheng ch'ueh-shih chuang." Ssu-ma pointed to the low price of grains despite widespread famine as proof of the efficiency with which the government siphoned cash out of the countryside. For Chang Fang-p'ing's confirmations, see HCP 269/2b. On government manipulation of the exchange rate between money and goods in its own favor, see Ssu-ma's 1071 memorial "Chun t'i-chu Shen-hsi ch'ang-p'ing—kuang-hui-ts'ang ssu tieh," SMWCC 7/6a-7a. Wang Yen-sou's argument is in HCP 376/17a.

[90] SHY:SH 4/18a-b; HCP 378/18a-b; 376/17a-b.

resigned in protest against the policy, and Ch'en and Tseng used illness as an excuse to withdraw from the fray. But continued attacks against Green Sprouts forced Wang An-shih temporarily out of office in 1074, and debate over the program compelled the court to confront fundamental questions of factionalism and the proper limits of remonstrance. Outside the court, Fu Pi and Ou-yang Hsiu used their posts as prefectural administrators to spearhead an obstructionist movement against distributing the Green Sprouts loans in Ching-tung East and Huai-nan East circuits in 1070 and 1071; and they were joined by a fiscal intendant in Hopei and a vice fiscal intendant in Shensi, as well as county administrators in K'ai-feng and Liang-che. Moreover, even the administrators of the policy turned against it: in 1070 Chang Tz'u-shan used his refusal of an ENG intendancy to lambaste the New Policies generally, while the Ho-tung ENG intendant Liang Tuan recommended that the Green Sprouts be abolished; in 1077 the Chiang-nan West ENG intendant Fang Tse counseled that the associated hired-service system be ended; and in 1081 Wang Ku memorialized against the rural credit policy after administering it in Liang-che, K'ai-feng, and Ching-hsi.^[91]

Opposition to the Szechwan tea monopoly, in contrast, was far more limited in political and ideological scope. The first wave of opposition to the monopoly, in 1077, was confined entirely to native Szechwanese in Szechwanese prefectural posts (Lü T'ao, Chou Piao-ch'üan, and Wu Shih-meng) and the censorate (Chou Yin). These four gained the dismissal of one intendant (Liu Tso), but were themselves all neutralized through transfers, dismissals, and a "muzzling order" that prohibited outside officials from interfering in THA affairs.^[92] Opposition to the THA was effectively silenced until the Restoration movement of 1086, when Lü T'ao and his fellow Szechwanese Su Ch'e were joined in their efforts to dismantle the THA enterprise by the elder statesman Liu Chih.

Ideologically, it is clear that state regulation of subsidiary crops and industries, including textiles and salt in addition to tea, provoked less fundamental resistance than official intervention in the staple grain economy. In the particular case of the Szechwan tea monopoly, even Lü T'ao and Su Ch'e were more opposed to the THA as an entrepreneurial enterprise than

[91] On the resignations of Ssu-ma et al., see CPPM 68/2191; SHY:SH 4/24b. Court debates over the policy, all involving direct participation by a confused and deeply troubled emperor, are richly documented in CPPM 68 and 69. See especially pp. 2164, 2169, 2171, 2173, 2176, 2180, 2185-89, 2221. For Ou-yang Hsiu, see HCP 211/12a-16a; SHY:SH 5/7a-b. For Fu Pi, see HCP 220/6b-7a; 222/6a-7b; 224/14b; and SHY:SH 5/8b-9a. For others, see SHY:SH 4/22a-b; SHY:SH 5/6b; HCP 212/7a-b; and for ENG administrators themselves, see HCP 210/10b; HCP 212/14a; HCP 285/5b-6a; HCP 313/7b-8a.

[92] HCP 283/9b; 284/15a-16a; SHY:CK 43/50b. For Lü T'ao's three critical memorials of 1077, see Ching-te chi 1/4a-23b.

to the idea of a state-run tea and horse trade. In particular, they condemned THA autonomy and its control over county government; the unrestrained profit seeking associated with THA involvement in nontea commodities, the profit-sharing-incentive system, and the forced sale of tea to Shensi residents; and the exploitative burden the tea monopoly placed on tea producers and Szechwanese conscripted into tea-transport service. But both accepted the need to trade tea for horses and merely urged that instead of imposing a region-wide monopsony, the state purchase the tea it required in the open market. Similarly, both grudgingly agreed that the costs of frontier defense entitled the state to claim a share of the profits from the tea trade, but they pressed the government to obtain its revenues from the collection of sales and transit taxes and licensing fees from an otherwise free trade (t'ung-shang , literally "via merchants") instead of from monopolization.^[93]

Conclusion

Despite the anti-New Policies atmosphere of the Yuan-yu era (1086-1094), even the relatively modest recommendations of Lü T'ao and Su Ch'e were disregarded. In the absence of a critical mass of opposition, the Tea and Horse enterprise was too valuable, and its value too greatly tied to monopoly control, for even the Restoration government to tolerate more than partial deregulation and an agreement to end abuses. Even the late-eleventh-century Szechwanese critic of the monopoly, Yang T'ien-hui, observed that with a financial surplus of two million strings each year and a quota of 15,000 to 18,000 horses, "it is not that there is no desire to change [the intendancy operation], but that in truth it cannot be changed." The Green Sprouts policy, on the other hand, was gradually but irreversibly dismantled during the first quarter of 1086, and the experiment in state control of rural credit was brought to a decisive conclusion. Can we conclude from this that state control of the tea and horse trade was successful, but state intervention in rural credit markets was not? Unfortunately, we cannot. So politicized was the Green Sprouts debate that there is simply too much rhetoric and too little concrete information in surviving records for us to evaluate the overall impact of the policy with any confidence. Seasonal credit crises were an endemic component of agrarian life, and certainly some sectors of the poor peasantry in some regions of the empire must have benefited from the establishment of public credit institutions. But no observer has yet surfaced to provide the same locally specific and richly textured description of the impact of the Green Sprouts policy on the peasantry

[93] Lü T'ao, Ching-te chi 3/5b-7a; 8b-9a; Su Ch'e, Luan-ch'eng chi (Ssu-pu pei-yao ed.) 36/9a-b.

that Lü T'ao has given of the Szechwan tea monopoly and the tea producers of P'eng-chou. Thus, although we can outline the probable best and worst results for the peasantry, we cannot yet assess its overall impact.^[94]

I have argued that both the tea and horse trade and the Green Sprouts measure emerged from a common blueprint for reform that sought to mobilize bureaucratic entrepreneurs to reassert the regulatory authority of the state in the market economy, but that only the Tea and Horse Agency, which consistently surpassed revenue expectations through its control of a cash crop in a frontier region, was allowed to develop into an autonomous entrepreneurial enterprise. Although the THA outlasted the New Policies, its entrepreneurial character proved more transitory. The promotion of entrepreneurship as a policy preference seems to have been unique to the New Policies leadership, and later administrations only reluctantly tolerated the radical decentralization of authority that lay at the heart of the entrepreneurial model.^[95] Moreover, shifting central policy goals merged with a changing geopolitical environment to erode the productivity of the Tea and Horse Agency over time, compromising the Agency's claim to extraordinary powers.

On the revenue side, the ability of the THA to provide surplus funds was undermined by the progressive loss of its extraregional markets to the revived Southeastern Tea Monopoly in 1103, and the Jurchen conquest of North China in 1127. By the 1130s the THA was able to subsidize only a small fraction of Szechwan's thirty-six-million-string defense bill; from mid-century on THA revenues routinely fell below quota; and by the end of the century the THA had forfeited its role as a revenue-producing agency. Jurchen possession of the Hsi-ho horse-marketing region, in conjunction with the Tangut occupation of Hsi-ning, had an even more chaotic impact on THA horse procurement. Total output of the Southern Sung THA averaged only 10,000 horses annually, less than two-thirds of the Northern Sung quota. More critically, usable northern "war horses" (chan-ma ) constituted only half of the annual quota; the rest were made up of weak and undersized Szechwanese "halter and bridle" ponies (chi-mi-ma ), largely unsuitable

[94] For the investigative report of Huang Lien, who was himself named intendant in 1086/ 6, see HCP 381/22a-23b. Yang T'ien-hui is quoted in Fu Tseng-hsiang, Sung-tai Shu-wen chits'un (1943 ed.), 26/10b-11b. On the ending of the Green Sprouts, HCP 368/26a; 374/4b-5a; 375/4b-5a. I discuss Lü T'ao's material in Taxing Heaven's Storehouse , pp. 63-68 and 220-27.

[95] Although the question requires greater systematic study, it appears that late Northern Sung and Southern Sung administrators viewed the decentralization of critical authority to autonomous regional agencies, especially military agencies, as a necessary evil rather than a valuable policy option. The best general discussion in English is still Lo, "Circuits," pp. 83-103. Hartwell, perhaps prematurely, sees the growth of powerful regional administrators during the Southern Sung, particularly the viceroys (hsuan-fu-shih ) and general supply commissioners (tsung-ling ), as a usurpation of central-government functions ("Transformations of China," pp. 397-98).

for riding. As horse numbers, quality, and convoy survival rates plummeted during the Southern Sung, the THA and ancillary purchasers could meet only about 70 percent of the government's annual equine needs.^[96]

The government's response to declining THA productivity was predictable: over the course of the twelfth century the size, operational scope, and organizational prerogatives of the THA were steadily trimmed. By mid-century the THA had been transformed from an autonomous superregional enterprise, answerable only to the central leadership, into one link in a functionally integrated regional administration, tied through a system of collective responsibility to the circuit fiscal intendants and vice-intendants, and the regional supply master (tsung-ling-so ) and pacification commissioner (an-fu chih-chih shih ), all of whom were subordinate to the regional viceroy (hsuan-fu shih ).^[97] At several times toward the end of the twelfth and the beginning of the thirteenth century, production shortfalls even forced the government to transfer direct control of the THA or its resources to other regional, extraregional, or central-government agencies.^[98] As organizational output dwindled, then, control over the THA was correspondingly centralized, until by the 1150s it had slipped into the same externally regulated managerial configuration that characterized the Ever-Normal Granary system under the New Policies.

There remained one crucial distinction between the New Policies ENGs and the Southern Sung THA. In the case of ENG administration of the Green Sprouts measure, centralized surveillance was at least partly intended to restrain official exploitation of the staple-growing peasantry. But cash-crop tea cultivators were less ideologically sheltered. Because the Southern Sung court was continually pressed for funds for which there were few new sources, it was very slow to reduce revenue quotas set when the market for tea was robust. Since revenue quotas determined the bureaucratic influence of the Agency and the material and career expectations of its personnel, even after the loss of North China and an empire-wide tea glut undercut the market for all tea, including Szechwan's, the THA was put under enormous pressure to replace entrepreneurial profits generated by the sale of tea to merchants and consumers with increased and inescapable imposts levied directly on tea producers. Until well into the thirteenth

[96] The trajectory of the tea and horse trade over time is discussed in Smith, Taxing Heaven's Storehouse , chap 7.

[97] See ibid., pp, 207-17.

[98] In 1186 the Pacification Commissioner was ordered to take over joint control of the THA's accounts because of its financial debts; in 1192 the THA owed horses as well, and the court transferred its capital funds to the Hu-kuang Supply Master to buy horses locally; in 1203 the THA was split into two branches, one in Ch'eng-tu and one in Hsing-yuan-fu (Hanchung), and both were put under the control of the Bureau of Military Affairs. See Smith, Taxing Heaven's Storehouse , pp. 216-17.

century, when the records run out, tea market officials were driven by the rules of the Agency to force tea households into debt and to confiscate their plantations in order to meet long-outdated quotas, as bureaucratic entrepreneurship degenerated into confiscatory taxation. And in contrast with the rural credit measure, there is no evidence that THA functionaries ever refused to administer the tea monopoly.^[99]

But were the growers of staple grains, in contrast, well served by the disestablishment of the rural credit measure? As Richard yon Glahn and Robert Hymes document, famine and credit relief problems became even more pressing during the twelfth and thirteenth centuries. Yet as Schirokauer, Walton, and de Bary among others in this volume attest, many of the most prominent Southern Sung political thinkers indicted excessive bureaucratic centralization as the cause of contemporary social, political, and economic problems, to which they sought solutions, not in the bureaucratic entrepreneurship of state agents, but in the moral entrepreneurship of community-minded gentlemen. The community granary (she-ts'ang ) model popularized in the late twelfth century by Chu Hsi illustrates this shift in ideological orientation from central-institutional to local-voluntaristic solutions quite clearly. As yon Glahn demonstrates, Chu Hsi embraced the goals of Wang An-shih's Green Sprouts measure, but attempted to correct its flaws by severing the link between rural credit assistance and state revenue production, privatizing administration of the community granaries to "local literati gentlemen" (hsiang-jen shih chün-tzu ), and demonetizing the loans exclusively to grain rather than cash.^[100] Whereas Wang An-shih took aim at the rural credit crisis by despatching state agents to displace the rural rich, Chu Hsi attempted to circumvent the inadequacies of state policy by exhorting some among the rural rich to manage credit relief for the poor. But community granaries were seriously underfinanced and highly susceptible to bankruptcy, and von Glahn shows that over time they were likely to be absorbed by the state and turned into vehicles for episodic charity or local price stabilization, rather than the mechanism for perennial credit assistance that Chu Hsi sought.

Chu Hsi's community granary model represents a paradigmatic shift from the centralized bureaucratic state to the local voluntaristic community as the focus of political action. But Chu Hsi and his Tao-hsueh associates had little impact on Southern Sung policymakers, who perpetuated in however attenuated a form the basic statist policies of the Northern Sung. During the early years of the Ming, however, literati suspicion of the cen-

[99] The rising exploitative burden is reflected in changes in the smuggling laws, surtaxes, registered quotas, and the laws on debt liability. See ibid., pp. 227-42.

[100] For Chu Hsi's assessment of the Green Sprouts measure, see his "Wu-chou Chin-huahsien she-ts'ang chi," in Chu wen-kung wen-chi (Ssu-pu ts'ung-k'an ed.) 79/18a. Chu Hsi's community granary model is analyzed by Richard von Glahn in his chapter in this volume.

tralized bureaucracy converged with Hung-wu's mistrust of self-serving bureaucrats to yield a state structure that, though highly centralized at the top, relinquished almost all capacity to intervene in and shape the economy. Ray Huang has pointed up the sharp contrasts between Sung and Ming financial administration. Whereas the Sung state (and here we restrict our view to Northern Sung) was willing to exercise its financial powers to reshape and tap into an expanding economy in order to increase revenues without overburdening the taxpayers, the Ming founder shackled the development of his financial apparatus in order to centralize fiscal authority in the position of the emperor and maximize political stability. The result was a financial administration that was remarkably self-denying, in that it reduced its own operational capacity to the minimum, neglected to develop revenues from industrial and commercial sources, and refused to consider the possibility of seeking assistance from private quarters. The whole tone of the administration was regression rather than progress. Huang demonstrates that ultimately the underfinanced Ming government was unable to provide basic public services, which retarded the progress of technology and the economy as a whole and forced administrators to shift the mounting costs of government to those portions of the taxable population who were least able to resist extraschedule taxes.^[101]

For Huang the Ming system represents a significant break in Chinese fiscal history, and he concludes that "from this time onwards, the main aim of government finance was to maintain the political status quo, and it ceased to exhibit any dynamic features." Historians of the Ch'ing have largely accepted Huang's interpretation. In a recent study of Ch'ing fiscal administration, for example, Marianne Bastid finds little difference between the basic financial institutions of the Ming and the Ch'ing, and points out that these institutions were employed "as instruments of regulation and balance rather than of compulsion and unification." In direct contrast with even the least entrepreneurial Sung fiscal practices and institutions, Bastid notes that during the Ch'ing "the principle of personal direct responsibility to the emperor of all officials, together with a rapid turnover except in the highest and. lowest grades, limited the possibility of developing a body of regulations and conventions specific to financial institutions which would allow such institutions to protect themselves and expand while still assuring their own internal policing."^[102]

It is now recognized that effective state intervention plays an integral role in economic transformations, and that active state participation was as

[101] On the political frustrations of Tao-hsueh adherents during the Southern Sung, see the chapters by John Chaffee and James Liu in this volume. On the Ming situation, see Huang, Taxation and Governmental Finance , pp. 315-16 and 321.

[102] Huang, p. 323. Marianne Bastid, "Financial Institutions of the State in the Late Qing," pp. 67-68, 75-76.

central to the early industrial revolution as it has been to later industrial change.^[103] Historians of China as well as comparative historical sociologists acknowledge that economic development in late imperial China was seriously impeded by what Ray Huang calls the "limited handling capacity" of Ming and Ch'ing financial institutions, which were unable to accumulate or distribute capital in ways that would promote economic growth. In her provocative comparison of economic development in China and Japan in the seventeenth to twentieth centuries, the sociologist Frances Moulder (following Eli Heckscher) has employed the term "provisioning" to characterize the very weak interaction between the late imperial Chinese state and the economy. For Moulder, a political economy oriented to pro-visioning aims

above all to acquire revenues and supplies for a static or relatively slowly expanding state civil and military apparatus and to ensure an adequate supply of food and other necessities to the rural and urban populations in order to forestall riots and rebellions against the upper classes. The state that has a provisioning policy plays a relatively passive role in the economy. Provisioning policies generally neither suppress commerce, industry, and private capital accumulation nor especially encourage their development.^[104]

Moulder contrasts the provisioning policies of Ch'ing China (and, she argues, Tokugawa Japan) with the mercantilist orientation of the state-building European regimes of the eighteenth century on and the industrial capitalist states of the present day. "The key element of mercantilism," she writes, "is the conception of the economy as a national economy whose wealth and strength can be increased in part through government action"; among these actions she includes the promotion of capital accumulation, expansion, technology diffusion, and the encouragement of a national marketing infrastructure through, for example, currency unification and the improvement of internal transportation.^[105]

[103] For a useful and provocative interpretation of the state and economic development, see Dietrich Rueschemeyer and Peter B. Evans, "The State and Economic Transformation: Toward an Analysis of the Conditions Underlying Effective Intervention," in Peter B. Evans, Dietrich Rueschemeyer, and Theda Skocpol, eds., Bringing the State Back In (Cambridge: Cambridge University Press, 1985), pp. 44-77.

[104] Frances V. Moulder, Japan, China and the Modern World Economy (Cambridge: Cambridge University Press, 1977), p. 48. Moulder's interpretation is generally consistent with that of Dwight Perkins in "Government as an Obstacle to Industrialization: The Case of Nineteenth-Century China," Journal of Economic History 27, no. 4 (1967): 478-92. Even during the High Ch'ing in the eighteenth century the state impinged only very lightly on the economy. As Feuerwerker argues, "the traditional Chinese economy did not depend upon any direct economic role performed by the central government, or usually even by local officials, for its continuing operation and for such prosperity as it achieved." See State and Society in Eighteenth-Century China: The Ch'ing Empire in Its Glory (Ann Arbor: University of Michigan Center for Chinese Studies: 1976), p. 89.

[105] Moulder, pp. 49-50.

I think that the mercantilist category Moulder reserves for the centralizing states of Europe and the industrial states of today can be usefully extended to the activist economic orientation of the Northern Sung, and especially to the statist experiments of the New Policies as represented by the Tea and Horse Agency and the rural-credit measure. But this returns us to the question raised at the outset of this chapter: for if provisioning policies are a measure of the disengagement of the state from the economy, how can we explain the transition from a mercantilist to a provisioning orientation in the post-Sung era?

The question is a complex one, and any complete explanation would have to consider changing international contexts and the question of technical capacity, particularly the limits on formal organizational capacity imposed by premodern technology. But in conclusion I would point to changes in the relationship between the state and the dominant elite as an underlying factor in the transformation of the Chinese political economy. Following Twitchett and Hartwell, I argued at the outset that economic activism from the late T'ang through the Northern Sung was promoted by the rise of a professional bureaucratic elite that depended on government service for its power and income and benefited therefore from the expansion of government activities. As Hartwell has demonstrated, long-term social, economic, and demographic changes in the structure of elite power from the eighth century on were exacerbated by the intense factional power struggles of the New Policies and its aftermath, which nullified the formal organizational tactics used by professional elite families to guarantee government careers for their offspring. With the intensification of factionalism and political purges, separate families of the national hereditary elite increasingly abandoned the endogamous marriage practices and career specialization on which their power had rested, and adopted the same strategies of career diversification and local social, economic, and political consolidation as the local gentry. The late eleventh and the early twelfth century, Hartwell concludes, saw "the disappearance of the professional elite as a cohesive status group made up of families who specialized in government service and the coming to the fore of local gentry lineages who encouraged a division of labor among their progeny with government service as only one possible career choice."^[106]

Hartwell has identified a persuasive correlation between the disintegration of the national bureaucratic elite and the rise of a new parochialism in Southern Sung policy-making, as local careers and interests replaced the national ambitions and intellectual horizons of officials in the Northern Sung. Moreover, it now seems clear that the rise of the local gentry in the twelfth century marked a watershed in the evolution of the Chinese elite that ultimately transformed the foundation of state power and economic

[106] Hartwell, "Transformations of China," pp. 416, 421-22.

policy-making. As Hymes and Ebrey have documented for the Sung, and Elvin, Beattie, and Wiens for the Ming and Ch'ing, social, institutional, and demographic factors all converged from the twelfth century on to diminish the role of government service in the lives of the dominant gentry elite. Even the examination system drove a wedge between the elite and the state, for the increasing ratio of degree holders to government positions made entry into the civil service progressively remote and made the examination system as much a mechanism for acquiring status and connections as a ladder to official success.^[107]

Although official position probably remained the quickest route to power and fortune throughout the late imperial era, for any given family the downward slide out of the civil service was far more predictable than entry into it. Under the circumstances, elite mobility strategies gradually clustered around the control of land, commerce, and credit, supplemented by the widespread pursuit of the social and cultural rather than the official rewards of education. It is true that starting in the early Ch'ing large numbers of lower-degree holders began to specialize in local managerial functions, such as tax farming, pettifogging, school supervision, and irrigation and bursary management. Though in some instances illegal or semilegal, these managerial roles became an intrinsic component of the local government apparatus. But rather than extending the power of the central government into local society, they tended to transform the empire's administrative machinery into a weapon in the private economic struggles of the upper gentry and its managerial functionaries. In consequence, as Philip Kuhn and Susan Mann Jones have observed, "as the monarchy lost its capacity to defend its realm against the assertion of private interests, the role of the central government itself in dominating and defining the sphere of public interest was being irreparably damaged."^[108]

These changes in the basis of elite power are well known, but I suggest that they lie behind the long-term shift from a mercantilist to a provisioning political economy. Recent political theory has stressed the ways in which

[107] Hartwell, "Transformations of China," pp. 400, 424; Robert Hymes, Statesmen and Gentlemen: The Elite of Fu-chou, Chiang-hsi, in Northern and Southern Sung (Cambridge: Cambridge University Press, 1986); Patricia Ebrey, Family and Property in Sung China: Yuan Ts'ai's Precepts for Social Life (Princeton: Princeton University Press, 1984); Mark Elvin, "The Last Thousand Years of Chinese History: Changing Patterns in Land Tenure," Modern Asian Studies 4, no. 2 (1970): 97-114; Hilary J. Beattie, Land and Lineage in China: A Study of T'ung-ch'eng County, Anhwei, in the Ming and Ch'ing Dynasties (Cambridge: Cambridge University Press, 1979); Mi Chu Wiens, "Lord and Peasant: The Sixteenth to Eighteenth Centuries," Modern China 6, no. 1 (1980): 3-40; Chaffee, Thorny Gates , p. 27, table 5, Ho Ping-ti, The Ladder of Success in Imperial China (New York: John Wiley, 1964), chap. 3.

[108] See Ho, chap. 4; Philip Kuhn and Susan Mann Jones, "Dynastic Decline and the Roots of Rebellion," in John K. Fairbank, ed. The Cambridge History of China , vol. 10:1 (Cambridge: Cambridge University Press, 1978), pp. 110-16 and 162.

states are more or less autonomous from, and often in direct conflict with, the dominant social strata that supply their personnel. Nonetheless, I think there are at least two broad but testable generalizations that can be made about the ways in which long-term social and politico-economic changes were intertwined. First, the growing localism and independence from government service of the gentry as a class fostered a preference for minimalist, noninterventionist, provisioning economic policies on the part of gentrymen as members of the government. And second, when some participants in the late imperial Chinese state did make sporadic attempts to enhance the power of government in society, an independent dominant class that also controlled the institutions of local administration was easily able to repel these efforts.^[109] In the centuries following the collapse of the national bureaucratic elite and the rise of the local gentry, the Chinese imperial state lost both the will and the capacity to expand the arena of public power.

[109] For a discussion of the literature on the autonomous state, see Theda Skocpol, States and Social Revolutions (Cambridge: Cambridge University Press, 1979), pp. 24-33, and "Bringing the State Back In: Strategies of Analysis in Current Research," in Evans, Rueschemeyer, and Skocpol, Bringing the State Back In , pp. 3-43. On the situation in late imperial China, see, for example, Ray Huang, 1587 , and Madeleine Zelin, The Magistrate's Tael (Berkeley: University of California Press, 1984), as well as the comparative analyses in Skocpol, States and Social Revolutions , and Barrington Moore, Social Origins of Dictatorship and Democracy (Boston: Beacon Press, 1966).