394 MODERN CHINA/JULY 1991 late Qing,especially after 1900,points to the substantial role of gentry in fee-taking.The funding of such bureaus,which were authorized by local officials and managed by the gentry,often necessitated the handling of tax funds by gentry managers.In some cases,tax collec- tion was the primary function of a gentry bureau.In Sichuan,for example,gentry bureaus were in charge of collection of land-tax surcharges known as jintie and juanshu,whose quotas by 1909 totalled nearly seven times the land tax itself.The gentry managers who supervised collection of these surcharges became an informal exten- sion of the bureaucracy,sharing their fees with the local magistrate and the yamen clerks(Lu Zijian,1984:305-383,790;Wei Yingtao and Zhao Qing,1982,2.661-666). Although many of the details of fee-taking in the late Qing are well known and show a continuity in bureaucratic practices that reaches back to the eighteenth century and earlier,such details require our attention because they can provide estimates of the scale of fee-taking that will be needed for evaluating salary reform.Direct evidence of the scale of fee-taking is scarce and mostly concerns local magistrates. Chang Chung-li(1962:29-31),based on sources covering three coun- ties located in Hunan,Jiangxi and Zhili,estimated the annual income of magistrates at 30,000 taels.2 The writings of Liu Yu,a magistrate in Sichuan in the 1860s,suggest that a lower estimate is in order, claiming that magistrates in Sichuan received customary fees of at least 3000 or 4000 taels per year,and quite commonly 10,000 taels or more (Lu Zijian,1984:549). This scanty evidence can be supplemented with evidence concern- ing the routine expenses of county magistrates,from which it is possible to infer the scale of fee-taking.One of the largest of these expenses was the salaries paid to private secretaries and advisors (muliao),who numbered at least five or six in each magistrate's yamen (Ch'u,1962:97).According to a memorial by a censor,Chen Hengqing (1902:146),each secretary received 500 to 1000 taels annually. Magistrates also faced incessant demands for money from their supe- riors at all levels of the bureaucracy.According to Liu Yu,ritual gifts demanded by superiors cost 10,000 to 15,000 taels,and in some cases even 20,000 taels,per year(Lu Zijian,1984:524,548).Similar figures
394 MODERN CHINA / JULY 1991 late Qing, especially after 1900, points to the substantial role of gentry in fee-taking. The funding of such bureaus, which were authorized by local officials and managed by the gentry, often necessitated the handling of tax funds by gentry managers. In some cases, tax collection was the primary function of a gentry bureau. In Sichuan, for example, gentry bureaus were in charge of collection of land-tax surcharges known asjintie andjuanshu, whose quotas by 1909 totalled nearly seven times the land tax itself. The gentry managers who supervised collection of these surcharges became an informal extension of the bureaucracy, sharing their fees with the local magistrate and the yamen clerks (Lu Zijian, 1984: 305-383,790; Wei Yingtao and Zhao Qing, 1982, 2.661-666). Although many of the details of fee-taking in the late Qing are well known and show a continuity in bureaucratic practices that reaches back to the eighteenth century and earlier, such details require our attention because they can provide estimates of the scale of fee-taking that will be needed for evaluating salary reform. Direct evidence of the scale of fee-taking is scarce and mostly concerns local magistrates. Chang Chung-li (1962: 29-31), based on sources covering three counties located in Hunan, Jiangxi and Zhili, estimated the annual income of magistrates at 30,000 taels.2 The writings of Liu Yii, a magistrate in Sichuan in the 1860s, suggest that a lower estimate is in order, claiming that magistrates in Sichuan received customary fees of at least 3000 or 4000 taels per year, and quite commonly 10,000 taels or more (Lu Zijian, 1984: 549). This scanty evidence can be supplemented with evidence concerning the routine expenses of county magistrates, from which it is possible to infer the scale of fee-taking. One of the largest of these expenses was the salaries paid to private secretaries and advisors (muliao), who numbered at least five or six in each magistrate's yamen (Ch'ii, 1962: 97). According to a memorial by a censor, Chen Hengqing (1902: 146), each secretary received 500 to 1000 taels annually. Magistrates also faced incessant demands for money from their superiors at all levels of the bureaucracy. According to Liu Yti, ritual gifts demanded by superiors cost 10,000 to 15,000 taels, and in some cases even 20,000 taels, per year (Lu Zijian, 1984:524,548). Similar figures
Hickey FEE-TAKING AND SALARY REFORM 395 appear in an 1854 memorial by the Sichuan provincial director of education,He Shaoji (Lu Zijian,1984:531-32). The expenses of local officials also generally included repayment of debts accumulated during a period of waiting for an appointment prior to assuming office.According to procedures described in gov- ernment statutes(and outlined in Watt,1972:45-55),the great major- ity of local magistrates were to be appointed directly by the Board of Civil Appointments(libu)in Beijing.However,over the course of the dynasty there appears to have been a steady rise in the percentage of appointments that were acting,rather than substantive,appointments. Local gazetteer evidence from Sichuan,for example,shows that, by the last decade of the Qing,seventy to eighty percent of local magistrates held their posts in an acting capacity and usually served for only about a year before being transferred to a new post(Zhongjiang xian zhi,1930:5/20a-b;Zizhong xian xuxiu Zizhou zhi,1929:6/ 13b-14a,20a;Daxian zhi,1933:6/34;Nanchuan xian zhi,1926, 3/8a-b;Yunyang xian zhi,1935:7/15a-b;Jintang xian xuzhi,1921: 6/3b-4a;Dazhu xian zhi,1928:7/6b-7a;Fushun xian zhi,1931:9/ 55a-57a;Qijiang xian xuzhi,1938:3/2b-3b;Suining xinzhi,1929: 6/46a-b;Shifang xian zhi,1929:3/6b).Unlike substantive appoint- ments,nominees for acting appointments were generally chosen at the provincial level.Thus,in the late Qing,many expectant magistrates waited for appointments in the provincial capitals.Liu Yu stated that this waiting period often lasted five or six years and required 1,000 to 2,000 taels per year in expenses,much of which was spent on gifts and banquets to win the favor of provincial-level officials who played crucial roles in the appointment process (Lu Zijian,1984:521,548). In the end,outright bribery of governors and provincial treasurers was usually necessary to secure an appointment.3 A 1905 impeachment of the Shanxi provincial treasurer indicates that magistracies in that province cost 3,000 taels;for the much wealthier province of Jiangsu, the fiction of Wu Woyao shows the cost ranged from 7,000 to 20,000 or even 30,000 taels (Yuzhe huicun,12 May 1905:2a;Wu Wo-yao, 1975:18-19). The minimum level of fee-taking practiced by late Qing magistrates can be estimated on the basis of the foregoing list of expenses.Since
Hickey / FEE-TAKING AND SALARYREFORM 395 appear in an 1854 memorial by the Sichuan provincial director of education, He Shaoji (Lu Zijian, 1984: 531-32). The expenses of local officials also generally included repayment of debts accumulated during a period of waiting for an appointment prior to assuming office. According to procedures described in government statutes (and outlined in Watt, 1972: 45-55), the great majority of local magistrates were to be appointed directly by the Board of Civil Appointments (libu) in Beijing. However, over the course of the dynasty there appears to have been a steady rise in the percentage of appointments that were acting, rather than substantive, appointments. Local gazetteer evidence from Sichuan, for example, shows that, by the last decade of the Qing, seventy to eighty percent of local magistrates held their posts in an acting capacity and usually served for only about a year before being transferred to a new post (Zhongjiang xian zhi, 1930: 5/20a-b; Zizhong xian xuxiu Zizhou zhi, 1929: 6/ 13b-14a, 20a; Daxian zhi, 1933: 6/34; Nanchuan xian zhi, 1926, 3/8a-b; Yunyang xian zhi, 1935: 7/15a-b; Jintang xian xuzhi, 1921: 6/3b-4a; Dazhu xian zhi, 1928: 7/6b-7a; Fushun xian zhi, 1931: 9/ 55a-57a; Qijiang xian xuzhi, 1938: 3/2b-3b; Suining xinzhi, 1929: 6/46a-b; Shifang xian zhi, 1929: 3/6b). Unlike substantive appointments, nominees for acting appointments were generally chosen at the provincial level. Thus, in the late Qing, many expectant magistrates waited for appointments in the provincial capitals. Liu Yu stated that this waiting period often lasted five or six years and required 1,000 to 2,000 taels per year in expenses, much of which was spent on gifts and banquets to win the favor of provincial-level officials who played crucial roles in the appointment process (Lu Zijian, 1984: 521, 548). In the end, outright bribery of governors and provincial treasurers was usually necessary to secure an appointment.3 A 1905 impeachment of the Shanxi provincial treasurer indicates that magistracies in that province cost 3,000 taels; for the much wealthier province of Jiangsu, the fiction of Wu Woyao shows the cost ranged from 7,000 to 20,000 or even 30,000 taels (Yuzhe huicun, 12 May 1905: 2a; Wu Wo-yao, 1975: 18-19). The minimum level of fee-taking practiced by late Qing magistrates can be estimated on the basis of the foregoing list of expenses. Since
396 MODERN CHINA /JULY 1991 it is apparent that formal salaries could cover only a small fraction of the costs of obtaining and holding office,such costs must have been met almost entirely by taking fees.Given the scale of expenses documented here,it is reasonable to estimate that,by the last decade of the Qing magistrates in posts with relatively few responsibilities must have collected at least 10,000 taels in fees per year,while in more burdensome localities the minimum figure was probably more like 20,000 taels.These estimates reflect the costs that even the most frugal magistrates could not avoid.Presumably,most local officials collected enough funds to both enrich themselves and meet the expenses of holding office.Thus,my estimates may be low by a wide margin. Nonetheless,such minimal estimates will assist in the analysis of salary reform below. The uncertainty in estimating fee-taking above the county level is even greater than for local magistrates,but there can be little doubt that the scale of fee-taking by individual officials increased progres- sively with rank in the provincial bureaucracy.This correlation be- tween bureaucratic rank and fee-taking income is borne out by pub- lished schedules of the financial contributions demanded by lineages from their office-holding members.These schedules,which assumed that the resources of provincial officials increased with rank,are the basis for the following estimates by Chang Chung-li(1962:31-32,40) of the annual income (in taels)accruing to various posts:prefect, 52,500;circuit intendant,75,000;provincial treasurer,150,000;and governor,180,000.This correlation between rank and income is confirmed by Liu Yu's estimates of fee-taking in Sichuan,where the governor-general was said to have received over 100,000 taels annu- ally in customary fees,the provincial treasurer 40,000-50,000,and intendants and prefects on the order of 10,000 taels each(Lu Zijian, 1984:547). The volume of fee-taking in the last decade of the Qing was probably greater than some of the above figures from the mid-nineteenth century indicate,for fee-taking probably increased as the costs of holding office rose over the last several decades of Qing rule.Officials surely passed on to subordinates and the people the increased expenses that arose from a price inflation that caused commodity prices to more than double between 1875 and 1910(Wang,1972:361).The last half
396 MODERN CHINA IJULY 1991 it is apparent that formal salaries could cover only a small fraction of the costs of obtaining and holding office, such costs must have been met almost entirely by taking fees. Given the scale of expenses documented here, it is reasonable to estimate that, by the last decade of the Qing magistrates in posts with relatively few responsibilities must have collected at least 10,000 taels in fees per year, while in more burdensome localities the minimum figure was probably more like 20,000 taels. These estimates reflect the costs that even the most frugal magistrates could not avoid. Presumably, most local officials collected enough funds to both enrich themselves and meet the expenses of holding office. Thus, my estimates may be low by a wide margin. Nonetheless, such minimal estimates will assist in the analysis of salary reform below. The uncertainty in estimating fee-taking above the county level is even greater than for local magistrates, but there can be little doubt that the scale of fee-taking by individual officials increased progressively with rank in the provincial bureaucracy. This correlation between bureaucratic rank and fee-taking income is borne out by published schedules of the financial contributions demanded by lineages from their office-holding members. These schedules, which assumed that the resources of provincial officials increased with rank, are the basis for the following estimates by Chang Chung-li (1962:31-32, 40) of the annual income (in taels) accruing to various posts: prefect, 52,500; circuit intendant, 75,000; provincial treasurer, 150,000; and governor, 180,000. This correlation between rank and income is confirmed by Liu Yii's estimates of fee-taking in Sichuan, where the governor-general was said to have received over 100,000 taels annually in customary fees, the provincial treasurer 40,000-50,000, and intendants and prefects on the order of 10,000 taels each (Lu Zijian, 1984: 547). The volume of fee-taking in the last decade of the Qing was probably greater than some of the above figures from the mid-nineteenth century indicate, for fee-taking probably increased as the costs of holding office rose over the last several decades of Qing rule. Officials surely passed on to subordinates and the people the increased expenses that arose from a price inflation that caused commodity prices to more than double between 1875 and 1910 (Wang, 1972: 361). The last half
Hickey FEE-TAKING AND SALARY REFORM 397 of the nineteenth century also saw a large influx into the bureaucracy of men whom the Qing state had rewarded with rank and office for their military and financial contributions to the suppression of the mid-nineteenth century rebellions.As a result,expectant officials faced increased competition for appointments to office and experi- enced longer waiting periods before receiving an appointment.Liu Yu,for example,observed,in 1873,that there were over one thousand expectant officials in Chengdu,rather than the less than one hundred who had resided there in the past (Lu Zijian,1984:524).With more expectant officials vying for an unchanging number of official posts, the cost of winning the favor of key provincial-level officials must have been driven up,thereby enlarging the debt that had to be paid off with the proceeds from holding office and fueling the propensity to collect fees. That inflation and a growth in the number of expectant officials probably contributed to a growth in fee-taking suggests that late Qing bureaucratic behavior cannot be explained solely in terms of moral decline and"corruption."Bureaucratic demoralization in the nine- teenth century may have increased the willingness of officials to enrich themselves at the expense of the state and the people,as some would claim,but the point remains to be proven.If there was such a decline in the moral standards of officials,that decline is,at most,only a partial explanation of this expansion in fee-taking. Precise measurements of changes in the scale of fee-taking and the exact mix of causes underlying such changes lie beyond our current knowledge,but it is clear that,viewed from the perspective of the finances of individual officials,state finance throughout at least the last several decades of the Qing dynasty was dominated by an informal sector of customary fees and gifts.The dominance of this fee-taking sector of state finance had two major consequences.First,with signif- icant portions of the finances of individual officials and their yamens escaping the Finance Ministry's accounting procedures,the private finances of officials and the finances surrounding government func- tions became inseparably entangled.Fees both funded public business and met the financial needs and demands of officials and their families and friends.Lack of state supervision over fee-taking meant that the size of fees was generally left to the discretion of officials on a
Hickey / FEE-TAKING AND SALARY REFORM 397 of the nineteenth century also saw a large influx into the bureaucracy of men whom the Qing state had rewarded with rank and office for their military and financial contributions to the suppression of the mid-nineteenth century rebellions. As a result, expectant officials faced increased competition for appointments to office and experienced longer waiting periods before receiving an appointment. Liu Yu, for example, observed, in 1873, that there were over one thousand expectant officials in Chengdu, rather than the less than one hundred who had resided there in the past (Lu Zijian, 1984: 524). With more expectant officials vying for an unchanging number of official posts, the cost of winning the favor of key provincial-level officials must have been driven up, thereby enlarging the debt that had to be paid off with the proceeds from holding office and fueling the propensity to collect fees. That inflation and a growth in the number of expectant officials probably contributed to a growth in fee-taking suggests that late Qing bureaucratic behavior cannot be explained solely in terms of moral decline and "corruption." Bureaucratic demoralization in the nineteenth century may have increased the willingness of officials to enrich themselves at the expense of the state and the people, as some would claim, but the point remains to be proven. If there was such a decline in the moral standards of officials, that decline is, at most, only a partial explanation of this expansion in fee-taking. Precise measurements of changes in the scale of fee-taking and the exact mix of causes underlying such changes lie beyond our current knowledge, but it is clear that, viewed from the perspective of the finances of individual officials, state finance throughout at least the last several decades of the Qing dynasty was dominated by an informal sector of customary fees and gifts. The dominance of this fee-taking sector of state finance had two major consequences. First, with significant portions of the finances of individual officials and their yamens escaping the Finance Ministry's accounting procedures, the private finances of officials and the finances surrounding government functions became inseparably entangled. Fees both funded public business and met the financial needs and demands of officials and their families and friends. Lack of state supervision over fee-taking meant that the size of fees was generally left to the discretion of officials on a
398 MODERN CHINA/JULY 1991 case-by-case basis.Uncontrolled,fee-taking opened the door to cor- ruption and blurred the line between legitimate funding of public administration and unjustifiable pursuit of profit at public expense. Moreover,fee-taking caused the private financial interests of officials to assume inordinate importance in the implementation of state finan- cial policy. A second consequence of the ubiquitousness of fee-taking was the development of patronage ties that linked together all levels of the bureaucracy through a network of intra-bureaucratic payments.Drawn into this patronage network,too,were local gentry and yamen func- tionaries who assisted with (and shared the proceeds of)the fee collection process.Patronage ties accompanied fee-taking because the acceptance of fees and gifts from a subordinate entailed reciprocal obligations-even in cases where the name and size of the payment were well established by custom.As late Qing critics of fee-taking were quick to point out,these obligations hindered superiors in exer- cising the kind of supervision over subordinates that was required by statute.Having taken fees from their subordinates,officials often faced heavy pressure from subordinates to overlook evidence of negligence, incompetence,or corruption.More often than not,officials avoided trouble and gave in to these pressures(Wang Shu'nan,1937:gongdu, 10/27b-29b;Zhu Shoupeng,1909:5/4927-4929).The result was that collusion in pursuit of private interests became the common pattern of interaction between different levels of the bureaucracy-more com- mon than conflict or cooperation for public purposes. Fee-taking,by virtue of its massive scale and its capacity to under- mine the authority of the central state,was the most important of those characteristics of Qing bureaucracy that made the system of public finance decentralized.Certainly,there were other institutional ar- rangements and historical developments that were conducive to de- centralization,such as the long-term decline in the administrative powers of provincial treasurers,who,in theory,were in charge of provincial finances and provided a check on the powers of the gover- nors(Fu Zongmao,1963:176).But the impact of these other decen- tralizing tendencies pales in comparison with the effects of fee-taking. Thus,a major challenge facing financial centralizers in the last years of the Qing was to eliminate fee-taking,create a new salary system
398 MODERN CHINA I JULY 1991 case-by-case basis. Uncontrolled, fee-taking opened the door to corruption and blurred the line between legitimate funding of public administration and unjustifiable pursuit of profit at public expense. Moreover, fee-taking caused the private financial interests of officials to assume inordinate importance in the implementation of state financial policy. A second consequence of the ubiquitousness of fee-taking was the development of patronage ties that linked together all levels of the bureaucracy through a network of intra-bureaucratic payments. Drawn into this patronage network, too, were local gentry and yamen functionaries who assisted with (and shared the proceeds of) the fee collection process. Patronage ties accompanied fee-taking because the acceptance of fees and gifts from a subordinate entailed reciprocal obligations -even in cases where the name and size of the payment were well established by custom. As late Qing critics of fee-taking were quick to point out, these obligations hindered superiors in exercising the kind of supervision over subordinates that was required by statute. Having taken fees from their subordinates, officials often faced heavy pressure from subordinates to overlook evidence of negligence, incompetence, or corruption. More often than not, officials avoided trouble and gave in to these pressures (Wang Shu'nan, 1937: gongdu, 10/27b-29b; Zhu Shoupeng, 1909: 5/4927-4929). The result was that collusion in pursuit of private interests became the common pattern of interaction between different levels of the bureaucracy-more common than conflict or cooperation for public purposes. Fee-taking, by virtue of its massive scale and its capacity to undermine the authority of the central state, was the most important of those characteristics of Qing bureaucracy that made the system of public finance decentralized. Certainly, there were other institutional arrangements and historical developments that were conducive to decentralization, such as the long-term decline in the administrative powers of provincial treasurers, who, in theory, were in charge of provincial finances and provided a check on the powers of the governors (Fu Zongmao, 1963: 176). But the impact of these other decentralizing tendencies pales in comparison with the effects of fee-taking. Thus, a major challenge facing financial centralizers in the last years of the Qing was to eliminate fee-taking, create a new salary system