Anda di halaman 1dari 31

State Involution: A Study of Local Finances in North China, 1911-1935

Author(s): Prasenjit Duara

Source: Comparative Studies in Society and History, Vol. 29, No. 1 (Jan., 1987), pp. 132-161
Published by: Cambridge University Press
Stable URL:
Accessed: 10/11/2010 22:38

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact

Cambridge University Press is collaborating with JSTOR to digitize, preserve and extend access to
Comparative Studies in Society and History.
State Involution:A Study of Local
Finances in North China, 1911-1935

George Mason University


Beginning around the turn of the twentieth century, the Chinese state
launchedonto a course of developmentthatseemed to resemblethe process in
early moder Europe that Charles Tilly and others have called state making
(Tilly 1975). The phenomenon of an expanding state structurepenetrating
levels of society untouchedbefore, subordinating,co-opting, or destroying
the relativelyautonomousauthoritystructuresof local communitiesin a bid to
increaseits commandof local resources, appearedto be repeatingitself in late
imperial and republicanChina. The similarities include the impulse toward
centralization,bureaucratization,and rationalization;the insatiable drive to
increase revenues for both military and civilian purposes; the violent re-
sistance of local communities to this inexorable process of intrusion and
extraction;and the formationof alliances between the state and local elites to
consolidate their power (Duara 1983).
Nonetheless, a curious paradoxis to be found in the Chinese experience of
state making throughthe republicanperiod (1911-49): the expansion of the
power of the Chinese state occurs concomitantly with growing anarchy in
local society. In other words, given thatthe ability of the state to control local
society does not measureup to its ability to extractfrom it, state making is a
concept thatcan only partiallyexplain these developments.We need to gener-
ate a concept capable of explaining the paradoxof simultaneoussuccess and
failure, growth and disintegration, within the same structure. Here I will
address this paradox by examining local finances in north China and will
develop a construct which I call state involution to explain this particular
patternof state expansion.
State involution is a variation of the state-makingprocess wherein the
formalstructuresof the state grow simultaneouslywith informalstructures-
such as tax farmersand mercenarysoldiers. While the formal state is depen-
dent upon the informal structuresto carry out many of its functions, it is

I wish to thankThomas Bernstein, Helen Chauncey, Bryna Goodman, Sudipto Mundle, Philip
Oldenburg,Mary Rankin, Evelyn Rawski, Thomas Rawski, and William Rowe for their com-
ments on this article.
0010-4175/87/1792-1334 $5.00 ? 1987 Society for ComparativeStudy of Society and History


unable to extend its control over the latter. As the state grows in the involu-
tionary mode, the informal groups become an uncontrollablepower in local
society, replacinga host of traditionalarrangementsof local governance. The
concept of state involution may be not only applicable to the problems of
states in the developing world, but, as we shall see, may also drawattentionto
the limitationsof historicalmodels takenfrom "successfully made" states for
the study of developing states.
This study is also part of a burgeoningtrend in Chinese social history that
seeks to articulatethe complex relationshipsbetween state and society in late
imperialand republicanChina. As Philip Kuhnand Susan MannJones (1979)
have suggested, a succession of paradigmshas dominatedthe historiography
of state-societyrelations, beginning with Marx and Weber. These writersand
the first generationof historians, impressedby the immensityof the imperial
Chinese state, viewed the forms of local society and local elites largely as
outgrowthsof this overwhelming state structure.The authoritystructuresof
local communities seemed to be entirely controlled by the imperial state
throughthe examination system, the bureaucracy,and official ideology.
This paradigmcame to be supersededin the 1960s by the "gentry-society"
paradigmthat saw the literatias mediatorsbetween state and society, possess-
ing a dual identityas state servantson the one hand, and local magnateson the
other. Historically sophisticated, this paradigm saw the gentry playing an
equilibratingrole, balancing the interestsof state and society duringperiods
of dynastic strength, but noted the dominance of gentry and local interests
duringperiods of dynasticdecline. Nonetheless, as local studies mushroomed
duringthe next decade and a half, even the broaderexplanatorypower of this
paradigmappearedto fall short in many respects.
It seems to me thatthe currenttrendthatseeks to critiquethe gentry-society
paradigmembodies two inchoate tendencies. Although these tendencies are
not always distinguishable,they have identifiablydifferentinterpretivefocii.
One, which we will call the elite-centeredview, emphasises a more complex
understandingof the gentry and local society part of the old paradigm.The
other, the state-centeredview, urges a more thoroughgoingunderstandingof
the structuresof the state.
Let us review the elite-centeredview first. Studies of local society indicate
that the gentry-society paradigm did not sufficiently differentiatethe elite
eitherspatiallyor socially. In certainregions and especially in core economic
areas,the state may have had to concede much more autonomyto elites thanit
was preparedto do among elites in peripheralregions (Skinner 1977). Fur-
ther, the state may have been much more susceptible to the interests and
pressuresof the higher levels of the gentryelite thanto those at the lower end.
Indeed, there were profoundsplits within the gentry in termsof their interests
and the extent of their identificationwith the imperial state itself. The most
significant revision of the gentry-society paradigm comes out of research

which suggests that there were long-term secular developments among the
elite that would defy the balancing and cyclical characterof state-gentry
The state-centeredcritique of the gentry-societyparadigmfocusses on the
natureof the late imperialand republicanstate. In the first place, the state did
not relate to local communitiesjust throughthe gentry. Indeed, it often dealt
directlywith a varietyof nongentryand nonelite groups:merchants,peasants,
and marginalgroups. Nor were gentry institutionsthe only channels through
which the state controlled and communicatedwith these groups. It utilized
corporategroups, networks, and all mannerof symbolic resourcesembedded
in popularcultureand kinship in often ingenious ways. Recently, a numberof
scholarshave developed various models to characterizestate-societyrelations
thatdo not involve gentryor even necessarilyelites as the centerpieceof their
formulations.These include the autonomist,the pluralist,the corporatist,the
liturgical, and the brokeragemodel. All of these models are part of a trend
thatseeks to supplement,replace, or subsumethe gentry-societyparadigmby
highlightinga multiplicityof modes throughwhich the state dealt with local
communities (Eastman 1984; Geisert 1984; Fewsmith 1984; Mann 1984;
Duara 1983).
A second aspect to this state-centeredrevisionism is that it tends to stress
the strengthening,at least in some importantrespects, of the state throughthe
late nineteenthand twentieth centuries. It is this aspect of the state-centered
argumentsthat brings out a latent conflict with its twin, the elite-centered
view, which until a few years ago representedthe orthodoxposition in argu-
ing that throughthe late imperialand republicanperiod, the state had lost its
power to local elites (Wakeman 1975). Crudely sketched, the state-centered
revisionism locates the beginnings of state strengtheningprecisely where the
other view sees the beginning of its decline relative to local elites-in the
aftermathof the great rebellions of the mid-nineteenthcentury (Liu 1978;
Kuhn 1979). From that point onward, proponentsof the state-centeredview
see growth (though not necessarily uninterrupted)in state power throughthe
nationalistperiod, while the otherview sees a continuousgrowthin the power
of elites and other parochialgroups.
Thus, what we have is a split image. The paradoxicalsituationwe spoke of
earlier-an expansion of state strengthwith a simultaneousincrease in anar-
chy-is an expressionof this same split image. Whatis needed is a lens thatis
able to align the two halves. The concept of state involution, althoughbelong-
ing empirically to the state-centeredperspective, represents a preliminary
effort to serve as such a lens. It holds that a key element to bringingthe two
sides back togetherlies within the structureof the state itself, especially at its
boundarieswhere it meets local society.
The expansion and penetrationof the state, and the enlargementof its
domainin Chinatook place in an ideological climate thatwas self-consciously

modernizing. Various practices and techniques that can be subsumed under

the rubricof rationalization.wereimplementedand utilized, not only during
the Nationalist period but also during the militant warlord period. Yet, ra-
tionalizationwas only partlyresponsiblefor the expansion of state resources.
By the mid-1930s, it was clear that the contradictionsand pressuresof state
organizationsoperating in an environmentof what was perhaps a stagnant
per-capitaincome (Yeh 1979:104) were to lead to the overwhelmingprepon-
derance of modes of expansion that were essentially traditionaland, as we
shall see, involutionary. As such, these modes of expansion may well have
been suitedto certainculturalcontexts where the state sought a restrictedrole,
as in traditionalChina, and where efficiency was not in itself an absolute
value. However, in the context of a modernizing ideology and an immense
increase in the role of the state, the reproductionof these traditionalpatterns
could have very serious social and political consequences.


I adaptthe concept of involutionfrom CliffordGeertz, who originallyapplied

it to his study of Javanese wet-field agriculture(1963). According to Geertz,
involutionis a process wherebya social or culturalpatternpersists and fails to
transformitself into a new patterneven afterit has reacheddefinitive form. In
Java, expansion of agriculturethroughthe colonial and postcolonial period,
which takes place through more or less constant returnsto scale, manifests
itself in a baroque elaborationof the existing institutionalpattern. While it
does not lead to a serious fall in per-capitaincome, it nonethelessinhibitsthe
possibility of economic growth (in terms of per-capitayield).
What I call state involution is a process, like state making, whereby the
state expands its role in society in moderntimes. What differentiatesit from
state making is the mannerin which this expansion takes place. We may say
that state making occurs where expansion takes place on an efficient basis,
whereasin state involution, expansionoccurs accordingto a logical procedure
similarto the one describedby Geertz. I will define efficiency in the system of
tax administrationas follows: the system is efficient when theformal struc-
ture of the state acquires an increasingproportionof the resources collected,
by whateveragencies, at the sourcefrom the taxpayer. Conversely, a system
is inefficient when the formalstate structureis unableto increaseits share(not
the absolute value) of the total resources extractedfrom the taxpayer.
State involution occurs when state organizationsexpand not through the
increasinglyefficient use of existing or new inputs, which in this case refersto
personnel and other administrativeresources, but through the replication,
extension, and elaborationof an inheritedpatternof state-society relations.
For China, I have characterizedthis patternas the brokeragemodel (Duara
1983). The most readily understandableform of the state broker is the tax
farmer, who obtains a right to collect a commission on the taxes he collects

for the state. Although the rate of the commission may sometimes be regu-
lated, the structuressupervising their collection are typically weakly devel-
oped, and the tax farmer will collect what he can through custom and his
bargainingpower. In imperial China, this patternof state interventionwas
applied not just to tax farmingbut to virtuallyall forms of state functions in
local society. Salaries of clerks, runners,and other functionarieswere nomi-
nal, if any, and therewas tacit understandingthatthey would extracta certain
"fee" from the public on every occasion thatthey performeda state function.
Thus, these personnelformed a class of entrepreneurialagents who brokered
governmentservices and whose fees were not regulatedby the state but by
their power in society.
The basic principles of the brokeragemodel exist in various parts of the
world, particularlybut not exclusively in the developing world. In India, for
instance, the ordinaryperson often needs to use the services of an intermedi-
ary in order to approacha bureaucrator to get some official business done.
Frequently,the intermediaryis a professional brokerwhose investment and
source of livelihood are his connections in "the office," which he nurtures
with great care. In societies where such brokersare dominant, the state not
only loses potential revenue to the broker, but precisely because the broker
needs to buy his access to the bureaucracy(through bribes, commission
sharing,etcetera)the state also foregoes control over a significantproportion
of the income of its bureaucrats.The more serious political consequence of
this phenomenon is that, as bureaucratsturn their attention increasingly to
these alternativesources of livelihood, they become quasi-brokersthemselves
and lose their identificationwith the goals of the state. State involution cli-
maxes at a level where the proliferation of brokerage across the society
becomes a barrierto the rationalizationof state organizations,which are now
doomed to modes of expansion that are oppressive and disruptive of local
It is not hard to imagine why this particularpatternof expansion became
generalized in twentieth-centuryChina. It was the easy option historically
availableto the state in a hurryto expandits resources.Even in the nineteenth
centurythereis reasonto believe thatit was throughexpandingbrokeragethat
the state was able to cope with a rapidlygrowing populationwhile maintain-
ing the formal administrativeapparatusat an unchanging scale. When it
resortedto the brokeragepatternto fulfill its goals in the twentiethcentury,
the republicanstate not only swelled the strata of brokers encrusted on its
edges numerically, but created a deepening effect as entrepreneurialbro-
keragepenetratedlevels of social organizationuntouchedpreviously, such as
the village.
In the perfect version of state involution in state finances, we should see a
situation in which every unit increase in the revenues of the formal state
structureis accompaniedby a commensurateincrease in the income of the

informalbrokerageformationsover which the state has little control. In other

words, in a situation of a stagnant economy, the involutive state can not
develop systems of bureaucraticresponsibility at a rate faster than the en-
trenchmentof entrepreneurialstate brokers-a process which is itself induced
by the increasedextraction and interventionof the state in local society.'

The normative goal of the Qing imperial state (1644-1911) was to avoid
extensive interventionin society. This, together with its growing inability to
manage the increasingly complex scale of its problems from the late eigh-
teenth century, caused the central governmentto suffer an enormousdecline
(by almost two thirds)in the real value of the revenuescollected from the land
tax and its surcharges,or direct taxes, in the 150 or so years from 1750 until
the early 1900s (Wang 1973:113). Prices almost consistently outranthe land
revenuesof the imperialstate until around1905. Since the greaterpartof local
government(province and county) income came out of the land tax and its
surchargesduring much of the period, the declining real values of collected

I In an historian'sdreamworld of complete information,the difference between state making

(sm) and state involution (si) may be expressed in a precise and quantifiablemanner. In both
cases, the ratioof the rate of growth of revenue (R) (or the formal income of the state) to the rate
of growth of national income (Y) should be rising. Let the rate of growth be expressed by the
superscripted., then
sm and si => (R/Y) > 0.
The difference between the two is expressed with referenceto the total paymentsextractedfrom
the tax-paying populace, or surplus, where S is composed of revenue (R) and the income of
brokers (B):
S = R + B.
RIS + BIS = I = S/S.
State making is defined as a situation existing when the rate of growth of revenue to national
income is rising and when the ratio of revenue to surplus is growing faster than the ratio of
brokerageincome to surplus:
sm => (R/Y) > 0 and (R/S) > (BIS).
State involution is defined as a situationexisting when the rate of growth of revenue to national
income is rising and when the ratio of brokerageincome to surplusis either growing fasterthan,
or at a rate equal to, the ratio of revenue to surplus:
si => (RIY) > 0 and (BIS) > (RIS).
The purpose of engaging in the formal exercise is the expectation that the model may be
rigorouslyutilized to study state involutionin a society where some estimateof brokerageincome
may be made available(as in India, where vigorous efforts are being made to estimatethe volume
of "black money" in circulationin the economy). Thanks are due to Sudipto Mundle, Chandra
Kant, and Adhip Chaudhurifor helping me to constructthe model. Note also thatstate involution
should be distinguished from S. J. Tambiah's ratherdifferent but inspiring concept of admin-
istrative involution (Tambiah 1985:252-86).

taxes affected local governmentas well (Wang 1973:121, 125-26). However,

the story with commercial or indirect taxes was very different. The rate of
growth of these taxes began to accelerate in the latter half of the nineteenth
centuryin response to the great crises of the mid century at both the central
and local levels (Wang 1973:81). Indeed, there is evidence that these re-
sources providedthe basis as well as the agency for state strengtheningin the
latter half of the nineteenth century, at least at the regional level (Mann
1984:ch. 4; Rowe 1983).
Then, in the first decade of the twentiethcentury,the modernizationefforts
of the government suddenly began to create a much greater role for state
organizationsin local society, from both an organizationaland an extractive
point of view. For example, in Ding County, Hebei (then Zhili), prices rose
by 7.5 percentfrom 1903 to 1911, while in the same period the burdenof the
land tax and its surchargesgrew by 35 percent (Wang 1973:119, 126; Feng
and Li 1936:487). The deep and far-reachingimpactof these tax increasesand
other organizationalchanges has been documented in detail (Duara 1983).
What we should note is that from this time onward, while there were some
reversals in individual years, the secular trend is toward an increase in the
budgets of local governments.
In this section on provincial finances in Hebei and Shandong, I will first
attemptto estimate the increase in provincial income and expense over the
period 1913-35, that is, for most of the republicanperiodbefore the Japanese
invasion. I will comparethe rate of growthof provincialincome with the rate
of growth of gross domestic product in order to see if the role of local
governmentin society was expanding. I will also look at the changing com-
position of provincial income and, finally, at the efforts that were made
toward rationalizationof tax administration.
Before plunging into local finances, however, it would be useful to clarify
the role played by the central governmentduringthis period, especially as it
affected provincial finances. Notwithstanding the reform efforts of the
Yongzheng emperor (1722-36) (Zelin 1985), the most striking feature of
central-localfiscal relationsduring the Qing was the absence of an authentic
separationof sources of revenue between different levels of the administra-
tion. Beginning with the fiscal reformsof 1909 and throughsuccessive legis-
lation until 1916, the governmentdid clarify distinct sources of centralgov-
ernment and provincial government revenues. However, the sources of
revenue for levels below the province remained undifferentiated,thereby
giving the province enormouspowers over the lower levels (Li 1922:70-83).
Fiscal relationsbetween the center and the province were relatively stable
duringthe years of Yuan Shikai's rule (1912-16), but duringthe succeeding
period under the warlords (ca. 1916-27), the military governors of the
provinces often kept a good proportion of the land revenue, which the-
oretically still belonged to the center (Li 1922:83). Under the Nationalists

(1927-37), the land revenue and its surchargesbecame a provincialitem, but

we should not draw hasty generalizationsabout the financial weakness of the
centralgovernmentin north China as a result of this. For instance, when we
look at the distributionof the tax income among different agencies of the
government in 1929 (Gamble 1968:183) and then in the mid 1930s, from
Ding County, Hebei (NFWH 1934:39) the revenue income of the central
governmentincreased by over 150 percent during these five or six years, a
rate of growth that is greaterthan that of central governmentincome for the
nation as a whole (Young 1970:102).
Hebei and Shandongwere importantprovinces from a fiscal point of view.
In 1931-32, Hebei had the second highest provincial expenditures in the
nation, and Shandong, the seventh (Chang 1934:234). Table 1 traces the
growthof real income and expenditurein the two provinces and also indicates
the percentagechange over the preceding recordedyear in the table.2
The most striking feature about the table is the manner in which the in-
creases in provincial revenues take place. There is a clear cascading pattern
thatoccurs over threedistinctperiods in both provinces, which correspondsto
three differentpolitical regimes in republicanChina:the early republicunder
Yuan Shikai, representedby the years 1913, 1914, and 1916; the warlord
period, including 1919 and 1925; and the period of Nationalistrule, including
the remainingyears. Within this configurationit becomes particularlyclear
that, while within each period there was a rather slow rate of growth of
provincialincome with periodic declines, there were ratherstartlingincreases
in provincial income from one period to the other.
The average annual rate of growth of provincial revenues in Hebei from
1913-16 to 1931-34 was 42 percent. In Shandong, it was 56 percentover the
same period. These rates are dramatically higher than the annual rate of
growthof gross domestic product, which was only 1.08 percentfor the period
from 1914-18 to 1931-36 (Yeh 1979:105), and satisfy a necessarycondition
of state making and state involution at the provincial level. The enormous

2 The
provincial-levelestimatesof income and expendituretakenfrom ZhangYifan are in turn
derivedmainly from the yearbookof the Shanghai-basednewspaper,Shenbao(ShenbaoNianjian
1966:G-36). The compilers of the yearbookreportthat the data for the period priorto 1925 are
based on the annualprovincialbudgets;those for the period after 1925 are takenprincipallyfrom
the Chief Statistical Office (Jujichu) of the Nationalist government. These have been supple-
mented by relevant publicationsof the provincial government as well as materialsfrom news-
papers and journals. Although we have no foolproof method of testing the accuracy of these
formal estimates (I talk about informal values in the text), I have found them to be broadly
consistent with other Chinese sources for periods for which we have such estimates (see Zhang
1936:174, 180, 196). County-level data are more consistently reliable than provincialestimates.
This is because they are based on an intensive investigationof the recordsand practicesof a few
counties conductedby researchersfrom Nankai University in Tianjin, specifically, Feng Huade
and Li Ling. As for the qualitative materials, their general tone, which speaks of a growing
burden upon the rural populace in north China, is certainly consistent with the researches of
scholars from other parts of China.

Real Income and Expenditure and Yearly Percentage Change, Hebei and Shandong

Hebei Province

Percentage Real Percentage Real P

Real income change expend. change income

1913 1,625,484 8,255,433 1,877,047

1914 5,776,648 255 3,220,212 -60 2,658,095
1916 2,644,561 -54 2,131,015 -34 2,321,067
1919 10,899,222 312 19,438,756 343 11,953,867
1925 9,603,207 -12 11,267,475 19 10,696,957
1931 331,131,782 224 31,132,598 176 20,053,039
1932 220,576,769 -34 20,576,769 -34 21,733,853
1933 225,517,821 24 25,517,821 24 23,341,584
1934 116,056,000 -37 15,735,023 -38 25,610,000
1935 23,689,666

SOURCE:Zhang Yifan (1935a:6; 1935b:5).

fromnominalfiguresby mebyutilizinganindexof purch
Note: Allrealvaluesinthisarticlehavebeenconverted
Universityteamandbasedon wholesalepricesin Tianjin(INCP1937).

expansion of provincial income suggests the increasing control of the prov-

ince over the resourcesof society and, throughincreases in expenditures,the
weightierrole it was playing in society as well. This is a point of considerable
importancein its own right. The bureaucraticpower of the central state had
become parcelized, but in the process the fiscal foundationsof the provincial
units had been strengthened,a fact that would enable them to play a more
importantrole than heretofore.
Although I have emphasized that the growth of provincialfinancial power
is significant in itself, it is nonetheless importantto see to what extent this
growth representedthe transferof revenues from the central governmentto
the provinces. Or, in other words, what was the magnitudeof the new reve-
nues that the province was able to generate?It is well known that the land tax
became the de jure revenue of the province duringthe warlordperiod and the
de facto revenue during the Nationalist period. Indeed, furtherinvestigation
revealed that for the first three years shown in Table 1 (1913, 1914, and
1916), the income values in Hebei do not include revenues from the land tax.
This is because until the death of PresidentYuan Shikai in 1916, the center
continued to receive the bulk of this revenue (although the province was
collecting a surchargeon the land tax that usually exceeded the 30 percent
permittedby the Yuan Shikai government) (Sun 1935:155, 238). Of these
three years, we have estimates for the land tax collected in Hebei only for
1916. But this datum will adequatelyserve as a baseline. When we add the
value of the real land revenue collected in Hebei in 1916 to the othervalues in
Table 1, we get the result outlined below.
Table 2 presents data on new revenues generatedby the province, and the
picture here is far less dramaticthan that in Table 1. Income from the land
revenue shows a decline in 1925 from not only 1919, but even from 1916.
Some caveats, however, are in order. These figures of provincial income in
the warlordperiod may conceal more than they reveal because, althoughthe
provinceswere retainingmuch of the land tax in practice, they did remit some
amountsto the center, and especially Hebei, which was usually controlledby
Beijing. In other words, Hebei province was probablygeneratingmore reve-
nues than the figures tell us, but exactly how much more is a mysterybecause
we do not know what proportionof the land revenueit passed on to the center.
Moreover, the militarygovernors of the northernprovinces, including Hebei
and Shandong, often levied taxes and provisions which did not appear in
formalbudgets, and concerning which it would be impossible even to hazard
an estimate. On the other hand, revenues from sources otherthan the land tax
were increasing by as much as 40 percent duringeven this period. Thus the
growth of the state in the period 1916-25 proceeded on a partial basis,
increasingrevenues in certain sectors while stagnatingor even declining in
others. The annualrate of growth of new revenues of the province duringthe

Real Income from Land Revenue and Other Sources and Yearly Percentage
Change, Hebei Province, 1916-1934 (in yuan)

Land revenue Other revenue Total revenue

Income Change Income Change Income Change

1916 8,181,696 2,644,561 10,826,257

1919 7,165,412 -12 3,733,810 41 10,899,222 0.6
1925 5,971,422 -16 3,631,785 -3 9,603,207 -12
1931 3,971,440 -33 27,160,342 648 31,131,782 224
1932 5,632,143 42 14,944,626 -45 20,576,769 -34
1933 6,306,931 12 19,210,890 28 25,517,821 24
1934 5,446,864 -13 10,609,136 -45 16,056,000 -37

SOURCE: Shenbao Nianjian (1966:G-41). See also NFWH (1934. 14-19:20).

Note: All the figures have been deflated by the Nankai index (INCP 1937).

warlord period (1916 to 1919-25) grew at 2.7 percent, thus still keeping
ahead of the gross domestic product.
However, there continues to be a leap in provincial government income
from the middle and early period to the Nationalistperiod, by over 10 percent
per annumfrom 1916 to 1931-34. This is so even thoughthe land revenue is
itself eitherdeclining or increasingvery marginally.The criticalelement here
is revenue from sources otherthan the land tax. The initial increasein yields,
by a factorof seven and a half from these sources between 1925 and 1931, is
moderatedconsiderablyover the next few years as the depressionbegins. But
even after the highly successful tax cuts implementedby the provincial gov-
ernment in 1934 following the Second National Financial Conference, the
increase in this category over 1916 is 300 percent and over 1919 is 184
percent.This is in keeping with the view that the Hebei government'sexperi-
ments with the system of commercial taxation that began in the mid-1920s
were highly successful (Mann 1984:ch. 6).
For Shandong, we unfortunatelydo not have specific land revenue values
for either 1919 or 1925, but the trendfrom 1916 until the republicanperiod is
much the same as in Hebei althoughthe increasesare far less dramatic.From
1916 to 1934, land tax revenues in Shandong increasedby only 28 percent,
whereasother revenues increasedby 246 percent. Total new revenues in the
period from 1916 to 1931-34 increasedby about 50 percent, or just under4
percent per annum. Incidentally, we also see the much higher proportion
formed by the land tax in Shandongthan in Hebei (Zhang 1935b:6).
Thus the bulk of the increase in provincial income, especially after 1919,
came from outside the land tax, a situationunlike that in Jiangsu, where the

weight of the land tax including the provincial surcharge almost doubled
between 1915 and 1933 (Sun 1935:292-95). The principalcomponentof the
taxes outside of the land tax was the juanshui, the indirect taxes, which
included the deed tax, professional taxes, and excise taxes. In Hebei, the
juanshui formed an annualaverage of 31 percentof provincialincome for the
six years between 1919 and 1925; and formed 23 percent of the provincial
income of Shandong for the three years between 1931 and 1934 (Zhang
1935a; 1935b:7,8). It is to be noted that these taxes affected the peasantryas
much as any other group. This was true even for the professional taxes on
merchantsbecause merchantspassed on any additionaltaxes to the consumers
(Sun 1936:36-37).
We now turnto an analysis of provincialexpenditures.Detailed figures on
these are not available, but it is nonetheless possible to form some general
impressionsof the characterof the interventionof provincial governmentin
local society.
Military spending, reflecting conflicts between competing state organiza-
tions, was an importantitem of expenditure in both provinces, especially
duringthe periodfrom 1919 to 1930. From 1920 until 1925, militaryexpendi-
ture in Hebei constitutedapproximately50 percentof total provincialexpen-
diture, and was closer to 60 percentin Shandong. Spendingon warfareduring
these years created major deficits in the Hebei budget in 1920, 1924, and
1925, and in 1919, 1922, and 1925 in Shandong (Ding 1931:419). The
northernwarlordsfought the AnZhi war in 1920, the FengZhi war in 1922,
and the second FengZhi war in 1924. Deficits from these adventuresled to
growing public debts (forced loans from banks and merchants), the interest
paymentson which ensured ballooning provincial expenditures.
Militaryspending is not necessarily contradictoryto the expansion of state
power during the early stages of state making. As Rudolf Braun and others
point out, in eighteenth-centuryEurope, military costs were not only by far
the largest item in the state budget, they generally exceeded the sum of all
otherpublic expenses. Braunevaluates warfareas "the moving motor of the
whole developmentof public finance" (1975:135). The huge deficits created
by such militaryexpendituresled not only to the evolution of a tax system, but
also to the raising of public loans and the funding of public debts. The
successful states were able to control the strainsproducedby these methods
by creatingan institutionalframeworkin which fiscal obligations came to be
regardedas part of the public duty. Having attaineda certain stable level of
extractionfrom society, the managersof these states were then able to put
their finances on a sound institutionalbasis (Braun 1975:316).
What was the extent of financial rationalizationin north China before the
Nationalist takeover in 1928? From 1911 until 1928, several efforts were
made to streamline the administrationof provincial finances and increase
availablerevenues. First, underthe Yuan Shikai government,the measureof

the levy changed from the silver tael, which had a fluctuatingsilver content,
to the more standardizeddollar, or yuan. In 1914, the arbitraryrates of
conversion of the grain levies were eliminated in favor of a single currency
system in which all taxes were levied. Finally, the first steps were taken to
convertthe Qing militarybannerand othertax-freelands into taxablelands, to
ascertainand consolidatethe many surcharges,and to carryout a land investi-
gation (Sun 1935:371).
Implementationof many of these proposalsandplans began only duringthe
period 1917-25. In 1917, in Hebei, the counties established a collection
office underthe county finance department,supervisedby the magistrateand
financeofficer. The office had a staff of approximatelythirty-five,intendedto
replace the activities of the traditionalbroker. In 1922, specialized depart-
ments were created to collect tax arrears,make new revenue registers, and
increasethe amountof lands subject to tax. All of these measures, however,
did not amountto very much in the sphereof land revenue. This is evident not
only from the marginalincreases in these revenues, but also in the continued
presence of brokers and their entourage through the entire period (Li
1938:998-1000; Duara 1983:336-51). The organizationalchanges designed
to increasecommercial revenues, in particularthe creationof the short-lived
Committeeto OrderZhili ProvincialFinances underGovernorYang Jiangde
in 1924-25, yielded more positive results (Chen 1934:13-16; Ding
1931:409-13; Zhang 1935a:3).
Under the Nationalists (1927-37), all military expenses belonged the-
oreticallyto the center. However, provincialmilitaryexpenses did not disap-
pear. Provinces were expected to make contributionsto regionally based
national armies, the precise amounts of which fluctuated according to the
needs of these armies. For instance, Hebei and the other northernprovinces
made substantialcontributionsto the regional army underZhang Xueliang's
vice command in 1931 and 1932 (HSZ 1936:126-27). In addition, the
provincemaintaineda local securityforce for which a tax, the gonganfei, was
levied. In Hebei, military expendituresbased on these two items averaged
about 25 percent of provincial expendituresbetween 1931 and 1934.
Data for nonmilitaryexpendituresin Shandongare available for the years
1921 and 1931, as shown in Table 3, and these enable us to observe the
changingpatternof civilian expenditure.The increase between the two peri-
ods amountsto an annual growth rate greaterthan 50 percent. This kind of
growthis impressiveby any standards,and clearly reflects the expandingrole
of the provincial state. In the first place, the provincial government had
increased its capacity to intervene in society by strengtheningits admin-
istrativeand police apparatus.Even more significant, it was developing the
infrastructureof the economy and society by the almost ten-fold increase in
the categoryentitled Reconstruction,Communication,and IndustrialRegula-
tion and by new expenditureson governmentbusiness enterprises. The in-


Percentage Distribution of Provincial Expenditurein 1921 and 1931, and

Percentage Increase in Real Expenditurefrom 1921 to 1931, Shandong Province

Percentage of
total Percentage increase
in real expenditure
Function 1921 1931 from 1921 to 1931

Party affairs - 4
General admin. 34 16 177
Judicial admin. 6 9 757
Financial admin. 8 8 448
Police 12 16 686
Education, culture 21 13 254
indus. regulation 8 14 880
Debt charges 7 3 115
Govt. business enterprises - 8
Miscellaneous 4 9
Total 100 100 486

SOURCE:C. M. Chang (1934:238).

Note: All realvaluesaredeflatedby Nankaiindex1926=100(INCP1937).

crease in educational expenditure also points to this build-up of the social

infrastructure. All of these developments certainly seem to indicate a trend
towards state making during the period under Nationalist rule. However, only
when we look at how these increases were generated at the lower levels such
as the county can we understand that rationalization of fiscal administration
was only a small part of what was taking place. A complex involutionary
process at the base of this expansion was to undermine much of the growth,
releasing powerful effects that reached beyond the county seat and into the
countryside itself.


Although there was some clarification between the central and local sources
of revenue in the early republic, provincial and county revenues continued to
remain undifferentiated. This led not only to domination by the province in
the system of local finances, but to general confusion in county finances.
Thus, for the republican period as for the Qing, it is important to keep in
mind the difference between formal and informal sources of county income.
The most important of the formal revenues of the county were the surcharges
levied upon taxes. The late Qing practice of granting localities the right to
impose surcharges for projects at their own level-levies like the haoxian,

and later the police and educational surcharge (Wang 1973:52)-became

broadenedafter 1914 when localities were officially permittedto collect a
surchargeon many taxes, including the land tax, the deed tax, and other
miscellaneoustaxes, as long as the surchargedid not exceed 30 percentof the
main tax. However, the province arrogatedthe right to these surchargesfor
itself. During the warlordperiod the county negotiated its share of the sur-
charge with the province (Peng 1945:1-2,131-32; Sun 1935:155).
After 1928, when the land tax and its surchargesbecame the revenues of
the province, the province permittedcounties to levy surchargesat a variable
percentageof the main tax for their own projects. In Hebei, this was approx-
imately 56 percent of the main tax (Peng 1945:3; Feng 1935:702-3). On
commercialtaxes, which were mostly farmedout, the county surchargewas
not supposedto exceed 50 percentof the provincialtax in each county (HNK
1939:60). However, the surchargerates frequently increased beyond these
limits and may have been about three or four times the land tax in the 1920s
and 1930s (Sun 1935:167;NFWH 1934.3:6,8). This is a significant increase
over the approximately5.15 percentin Qing times and 30 percentin the early
republic. This increase in the surchargerate beyond the legally permitted
limit, then, constituteda category of informal income.
Anothersource of formal income of the county was called the mouchuan,
better known among the peasants as tankuan. Originallythis was a levy, or
chaiyao, in the northernprovinces imposed to supply the emperor, officials,
or armies in transit with provisions. Although it became more and more
common throughthe nineteenthcentury, each county's share was still quite
limited duringthis period. In 1911, the share of each county was formalized,
converted into cash, and made its supplementaryincome for the funding of
modernizationprojects. The uniqueness of tankuan lay in the manner in
which it was levied-on the village communityas a whole ratherthan on the
individual.The difference between the mouchuanor tankuanand the permit-
ted surchargelay in the fact that the formerhad no statutorybasis, and it was
expected that the county authorities would confer with the local elite in
determininga reasonablelevy. Thus, althoughwe will call it a formalrevenue
source, it might be more appropriateto think of it as a conventionalsource of
revenue (Xu 1936:6; HNK 1939:58).
As the tankuanbecame regularized,however, there appearedanotherform
of it-the infamousbaidi or provisionaltankuan.Provisionaltankuan,which
rarelyappearedon county budgets, was imposed wheneverlocal authorities-
the county, ward, the military, or even the province-needed funds on a
temporarybasis. It began its increase during the warlordperiod as a conse-
quence of the wars thatravagedthe region, and became partof informallocal
finances through the Nationalist period because of the increase in county
activities and resultantshortagesin local budgets (HSZ 1936:62). Provisional
tankuancame to be distinguishedfrom the regulartankuanthat was levied on

a seasonal basis and appearedon county budgets. However, like the regular
tankuanit was not levied on the individual, but on the village or some other
collective unit, which devised its own means of sharing the tax burden(Xu
The provisional tankuan quickly became one of the most arbitraryand
vexatious of levies, and was, of course, almost impossible to supervise or
measure.Yet it probablycontributedmore to governmentrevenuesthanmany
othercategories (Zhang 1935a:12; NFWH 1934.5:163) and, more significant,
it constituted an importantsource of livelihood for government and quasi-
government personnel who inserted themselves in the collection process.
Precisely because they were outside the regular budgetaryprocess and not
easily susceptible to public regulation, tankuan levies tended to become
swollen as they made their way up the administrativehierarchy.State brokers
at all levels of this hierarchysuperimposedtheir own demandson the original
levy, and it was extremely hard to distinguish whether this share was to be
used for public or private purposes.
Whatposition does the county occupy in the financial pictureof the nation
as a whole? In 1934 the total income of the county from surchargeson the
land tax of all counties in Hebei was 5 million yuan comparedto the 6 million
yuan in provincialincome from land revenue. In Shandongin the same year,
it was 11 million yuan compared to the 15 million yuan provincial income
from land revenue (NFWH 1934.7-12:20,21). Thus, the publicized income
of the counties was only slightly less than the land revenue. After the Na-
tionalisttakeover, county income grew rapidly-at the rate of approximately
17 percent per year from 1928 to 1933-even throughthe depression years
(Feng 1935:712).
Nonetheless, the share of revenue the county was permittedto retain was
consideredto be grossly unfairby many contemporaryobserversof the fiscal
system. They believed thatthe county was deprivedof far too much of the tax
income and that much more should come back to the localities than actually
did. According to the figures for 1934 presented to the Second National
FinancialConferenceby Li Jinghan,the counties collected only 75 percentof
what the provinces collected, and only 40 percent of what was collected by
the central government from a single county. Of the total governmentalin-
come in the mid-1930s, the center received 52 percent, the provinces 27.5
percent, and the counties only 20 percent. Since Li believed that very little of
the revenuecollected by the higher levels of the governmentwas spent for the
development of local communities, he advocated fiscal reforms that would
grantthe counties 50 percent of the total tax income (NFWH 1934:39; Feng
Looking at the distributionof sources of income at the county level shown
in Table 4, the most strikingfeatureis the contrastto the sources of provincial
income shown for Hebei in Table 2. Whereasincome from nonlandedsources


Percentage Distribution of Sources of Income, Qinghai and Ding Counties,


Land tax Deed tax County Mouchuan Yashui

surcharge surcharge property (Tankuan) surcharge Total"

Qinghai County
1928 57 12 15 - 5 89
1929 30 7 8 47 3 95
1930 31 14 7 40 4 96
1931 45 19 14 3 9 90
1932 36 17 11 17 9 90
1933 39 11 7 30 5 92
Ding County
1929 26 4 - 17 22 69

SOURCE: Feng Huade (1935:700).

"The remaining4 to 11 percent is made up of a multitudeof miscellaneous duties. For Ding
County, see Gamble (1968:183).

at the provinciallevel rangedfrom 70 to 87 percentof total revenuesbetween

1931 and 1934, at the county level, income from the land tax, countyproperty
tax, and mouchuan-all sources connected with land-represented 62, 64,
and 76 percentsfor 1931, 1932, and 1933. It is hardto tell how generalized
this phenomenonwas. Ding Countydoes recordhigh incomes fromyashui, or
commercialtaxes, in 1929, and Sidney Gamble also shows a relatively high
proportion(28 percent)of that county's income deriving from special excise
duties-the animal and cotton, and peanut and wood taxes. These taxes
perhaps enabled the county to apply lower mouchuan assessments than in
Qinghai (Gamble 1968:183).
A nationwidestudy of the sources of county income in thirteenprovinces
shows a picturecloser to Qinghaithanto Ding County. Surchargeson the land
tax occupied over 60 percentof the income, and the yashui less than 1 percent
(Sun 1936:37). While this piece of evidence tends to confirm the image that
county income derived mainly from direct taxes, it also reaffirmsthe general
impressionof the success that the Hebei governmenthad with indirecttaxes
from as early as the 1920s and continuingthroughthe 1930s. Nonetheless, we
should keep in mind that even here the provincial governmentgained much
more from this source of income than did the lower levels.
What is the patternof county expenditures?The following discussion ana-
lyzes the resultsof C. M. Chang's study of the distributionof county expendi-
turesfor 130 counties in Hebei in 1931, and Sun Shaocun's averagepercent-
age distribution figures for samples from thirteen provinces in the early
1930s, both shown in Table 5.


Percentage Distribution of County Expenditure,Hebei Province and Nationwide

Hebei, 1931 Nationwide,

early 1930s
Function (in yuan) Percentage Percentage

Party affairs 58,787 0.71 1.27

Financialadmin. 237,088 2.85 3.31
Police, anddefense 3,235,450 38.92 19.58
Education 3,167,711 38.10 31.77
Self-govt. 902,556 10.85 4.81
Modernization 542,667 6.53 5.80
Charities, relief 54,270 0.65 4.02
Generaladmin. 21,116 0.25 15.44
Other 94,595 1.14 14.00
Total 8,314,240 100.00 100.00

For Hebei, C. M. Chang (1934:236-37); nationwide, Sun Shaocun (1936:39).

The chief difference between Hebei and the nationwide average occurs in
the item for police and defense. The Hebei counties spent almost 20 percent
more on this item than the average county in the nation. Police and defense
costs were proportionatelyhigher in the northernprovinces, reflecting, per-
haps, a higher degree of disorderand militaryviolence in these provinces. In
Shandong, for instance, these items occupied 32.43 percent of the county
expenses, while in Jiangsuthey amountedto only 18.65 percent. The general
patternof distributionin Hebei is repeatedin the individualcase of Qinghai
County, for which we can also construct a time series, shown in Table 6.
Several points should be noted in connection with the data shown in Table
6. Education was clearly an extremely importantfunction of local govern-
ment, and the bulk of funds for education was concentratedin the develop-
ment of primaryschools (Feng 1935:713). It is interestingthat a greatershare
of the budgetwas spent on educationduringthe warlordperiodthan underthe
Nationalists,although,of course, the absoluteamountspent on educationwas
probably greater during the Nationalist era. Local security was the other
priority, and during 1929 costs of police and security constitutedalmost 75
percentof the entire expenses of the county. Both of these items conform to
the generalpicturewe have for Hebei as a whole. Both also tended to dimin-
ish in relative importanceas other modernizationprojectsbegan to mushroom
throughthe 1930s. The expenses of the ward (a subcounty institutionof the
twentiethcentury) basically included the maintenanceof a few functionaries
and a police force. The gradual growth of this item suggests the effort to
strengthenlocal government at this level (Feng 1938b:1048).

Percentage Distribution of Expenditure, Qinghai County, Hebei Pr

1919-23 1924-28
(Annual (Annual
Function avg.) avg.) 1928 1929 193

Education 41.19 43.36 47.80 23.69 22.3

Police 48.75 36.72 49.67 27.32 19.3
Security - - - 47.13 39.6
Modernizationprojects 7.29 3.14 2.20 1.73 9.7
Ward govt. - 12.34 - - 8.8
Other 2.77 4.44 0.33 0.13 -
Total 100 100 100 100 100

SOURCES:Feng Huade (1935:708; 1938c:1041-42).


One featureof county finances that came underheavy criticism from con-
temporaryobserverswas the shortageof resourcesavailable for each govern-
mentalactivity undertaken,a phenomenonthat was seen to be largely created
by the governmentitself. lnsteadof utilizing the increasedrevenuesto consol-
idate or improve previously existing functions and departments,the county
was undertremendouspressureduringthe 1920s and 1930s from the province
to createdepartmentsand bureauswith additional, "modernizing," functions
such as departmentsof land registration,sanitation,roads and bridges, party
training, and many others. These departmentswere allocated such paltry
amountsthat for a serious administratorthe situation was a veritable night-
mare. For instance, the departmentof water control and forest conservation
was grantedthe niggardlysum of 480 yuan, an amountthat would barelypay
the wages of two clerks (Feng 1938b:1051).
Indeed, so great was the pressureto develop the facade of moderngovern-
ment that administrativeredundancy,exacerbatingthe shortage of available
resources, was the result. Modern-style fire brigades coexisted with tradi-
tional fire-prevention organizations. There were police organizations at
county, township, and ward levels-the last of which also supervised the
militia-with overlappingjurisdictions and little co-ordination. Critics be-
lieved that these security arrangementscould have been more economically
and efficiently organized under a single administrative center (Feng
1938b:1045;HNK 1939:58).
All of this expansion, of course, meant a vast increase in the numberof
personnel,which in turnrestrictedfunds availablefor equipmentandprojects.
In a study of four counties, including Qinghai, Feng Huade discovered that
wages alone claimed 70 to 80 percent of total expenditurefor all categories,
but particularlyfor the top three items-police, education, and wardexpenses
(Feng 1934:510-512). At the level of the ward as well, 90 percent of the
expenses were absorbedby wages. Moreover, from 1929 to 1931, the real
wage bill actuallyincreasedas more and more bureausand departmentsbegan
to be added (Feng 1935:739,713).
The high wage-to-equipmentratio suggests, among other things, that the
largenumbersof governmentpersonnelwere unableto achieve a high level of
efficiency. As an example, Feng cites the case of Qinghai, where there were
150 policemen but only a very few modernweapons. He believed that under
these circumstances the level of police efficiency was so low that many
officers could well have been dismissed from the force withoutmucheffect on
the performanceof the unit as a whole (Feng 1938b:1045). Since most of the
county revenues went to pay wages and there was little left over to implement
projects, the system of local finances ended up, to a significant degree, as a
system supportingthe livelihood of an increasing number of state brokers.
Local governmentbecame what one writer has called an "employment con-
traption" (Geisert 1984). Feng sums up the situation succinctly: "This dis-

proportionatedistributionof expenditures (on salaries), especially for such

items as industrialdevelopment (or water control), which by its very nature
requiresgreateroutlays on equipmentthan on salaries, results in the creation
of a nominal staff whose function is actually not to function. Under such
circumstances, money spent is equivalent to money wasted, although the
burdenon the peasant is not in the least alleviated" (Feng 1934:512).
Responsibilityfor this situation lay more with the provincial government
than with the county, since it was by provincialorderthat the modernization
measureswere pushedthrough.At the same time, the provincesoughtto limit
severely the funds available to the county or, rather, to make certain that
provincial domination over local revenues would not be compromised. It
soughtto do this in 1928 by establishingfinancialbureausat the county level,
orientedtowardthe province and with considerablepower of supervisionand
control over the county administration.Despite their limited success, these
bureausexercised such great financial pressuresupon the counties that many
of them were reduced to insolvency (Kuhn 1979:121-25).
The variouspressuresfrom the province combinedto encouragetrendsthat
were in fact contraryto provincialgoals. First, in orderfor local governments
to fulfill just their minimal obligations, the unreportedtax income of county
governmentsbegan to rise rapidly, the result of higher surchargerates (Li
Ling:1935) and more frequent, or higher, provisional tankuanlevies (Chen
1934:13-15; HNK 1939:62-63; Sun 1935:355; Feng 1938a:1119).
Second, the additionalgovernmentpersonnel, hired as governmentdepart-
mentsexpanded,were, as a resultof these pressures,employed undercircum-
stancesthatwere similarto Qing times. The fact thatthe wage bill was spread
so thinly over an increasingnumberof personnelin the proliferatingorganiza-
tions of local governmentmeantthatthese workerswould be forced to depend
on sources otherthan their formal salariesfor sustenance.The greatestoppor-
tunity for state brokersto pursue their pecuniaryinterestswas to be found in
the provisionaltankuan. There was income to be had here all the way down
the hierarchyof collection. Personnelat each level addedtheir shareonto this
levy, which could emanate at various points and which was the most elusive
to supervise.
Tankuanalso providedthe channel throughwhich the most significant and
novel aspect of the involutionaryprocess took place in republicantimes-the
extension of the brokeragepatterndownward to the ward and the village.
Brokersat these levels addedtheir shareonto the downwardmovementof the
levies and subtractedthem in their upwardmovement of taxes. This was true
for both military and civilian governmental demands (HNK 1939:58; Sun
1935:355; Xu 1936:54, 57-60; Feng 1938a:1119). Official reports of the
AdministrativeYuan in Nanjing, elaboratingupon the severityof the tankuan
demand, described as "the greatest cause of the suffering of the people,"
repeatedlydemonstratethis swelling or piggy-back effect. For instance, in

Xingtai County, ShanchuanWard, the reportsstate that the forty villages and
towns of the ward had already spent several times the budgeted amount for
security expenses by the middle of the fiscal year in 1933. The increases
apparentlyoriginated at the ward level and were absorbed by the militia
(NFWH 1934.5:164). In anotherexample, villages paid dues that were five
times the formal tankuanlevy. The difference seems to have been absorbed
by police and self-defense personnel. There are many such cases to be found
in the literature, especially at the village and ward levels, indicating the
burdenof this levy. They also illustratethe mannerin which the levy became
an importantfoundationsustainingthe entire structureof formal and informal
government by means that were essentially unsupervised and arbitrary
(NFWH 1934.5:167, 168; Xu 1936:122-27; Duara 1983:220-26).
By 1934 the gravity of the situation had become clear to provincial au-
thorities.The expansionof revenues and growthof the formalstate apparatus,
so assiduously cultivated at the provincial and county levels, turned out to
have been raisedon the uncertainsoil of traditionalbrokerage.In part,provin-
cial efforts to squeeze and control the lower levels had led to the uncontrolla-
ble proliferationof subadministrativepersonnel and practices. The profuse
undergrowthof state brokeragehad also spreaddeeperinto local society as the
state sought to penetrate down to the village, creating new strata of en-
trepreneurialbrokers in the wards and villages. Consequently, in order to
sustainthe apparatusof proliferatingbrokerageas much as to maintainlegiti-
mate governmentfunctions, the rates of informal tax collection had risen to
levels that were quantitativelyunknown but politically dangerous.
In 1935, the provinces embarkedon a programto diminish the severity of
provincial controls by prohibiting tax farming and enhancing the financial
powers of county government(Kuhn 1979:126-29; Peng 1945:134). In Hebei
these measures were intended to dovetail with the provincial drive to reduce
the tax burdenand rationalizeadministrationthat had begun in 1934. Hebei's
much-publicizedreductionof the tax burdenby 550,000 yuan in 1934 (which
was regardedas among the most successful in the nation) was expected to go
further,to 1.5 million yuan, for the next year as a result of the new reforms.
Shandong'scontributionunderGovernorHan Fuchun was considerablyless,
but it, too, was expecting to make strides in reducing expendituresthrough
these reforms (NFWH 1934.7-12:10149-50; Zhang 1935b:2).
The Japaneseinvasion of northChina, which followed soon after, made it
difficult for the reforms to rectify the existing system, and contemporary
observers were less sanguine about the potential of the reforms than the
provincialauthorities.The critics believed thatthe new tax collectors were not
necessarilybetterthanthe old ones, and were certainlynot more efficient than
tax farmers(Zhang 1935a:3). The crucial issue, it appearsfrom the foregoing
analysis, is the extent to which a county's control of its own finances would
enable it to transformthe strataof brokersthat underlaythe formal govern-

ment system into a stratum of committed officials. For whatever reason,

because of pressuresfrom the communistsor the Japanese,or the absence of
managerialability, when the Japanesearrivedon the scene they found a state
structurecontinuing to expand througha process of involution.


What was the impact of state involution on the peasant masses who con-
stitutedthe bulk, and the most vulnerablesection, of the Chinese population?
There are two areas in which one can investigate this impact: first, in the
increase in the tax burden and, second, in the transformationof the institu-
tional frameworkof ruralpolitics. It would be foolhardyto claim to be able to
estimatewith any degree of precision the impactof state involutionon the tax
burdenof the peasants, and the great variationfrom region to region would in
any event render such efforts worthless. Nonetheless, there are sufficient
historicaldata to permitdelineationof broadtrendsin this burden.The quan-
titative and qualitative material available suggests that there was a secular
increase in the total tax burdenof the northChinese peasantthroughthe first
half of the twentieth century, and at times the increase was staggering.
An exhaustiveestimateof the total tax burdenon the peasantwould have to
consider (1) the formal taxes of the different levels of governmentincluding
the village and the military, and (2) the informal charges of each of these
levels. Moreover, it would have to adjustfor (3) the rate of inflation, (4) the
exchange rate between silver and copper, and (5) the terms of tradebetween
agriculturalcommoditiesand manufacturedcommodities. While the natureof
the sources is such that we are unableto ascribeprecise or total values to each
of these factors, I believe that it is nonetheless possible to demonstratethat
they each worked to increase the tax burdenof the peasantryright up to the
period of the Japanese occupation (when scarcely anyone doubts the enor-
mous increase in this burden).
In the precedingsections of this articleI have triedto estimatethe growthof
formalrevenuesof the differentlevels of governmentand have adjustedthese
increases for the rate of inflation. We know that provincial real income,
representingnew revenues, increased at a rate greater than 10 percent per
annumfrom 1916 until 1931-34 in Hebei (Table 2) and that the formal real
income of the county government doubled between the end of the warlord
period and 1932 (Feng 1935c:712). The increase in central governmentin-
come is also relevantbecause the salt tax and various commercialtaxes also
affectedthe ruralpopulace. The income of the centralgovernmentfrom Ding
Countyincreasedmore than three-foldfrom 1929 until the mid-1930s (Gam-
ble 1968:183;NFWH 1934:39).
Next we need to add the informalor unbudgetedincome of local govern-
ment and the levies, commissions, and fees chargedby state brokers.If there
is one thing that we know aboutthese values it is thatthey were increasing,if

for no other reason than that the state functions and numberof statutoryand
nonstatutorystaff were increasingat all levels of government,especially those
below the county. According to C. M. Chang, "the actualexpendituresof all
units of local government, including wastes and leakages, would be no less
thantwice" the values indicatedby the publishedrecords(Chang 1934:236).
These estimates do not include levels of government like the village, the
ward, and the military, for which we do not have formal records.
Sidney Gamble estimatedthe increasein village expenses solely for village
purposesfor Hsin Chuangvillage in centralHebei, over the period 1907-32.
These figures exclude all taxes and levies made by the county or the ward or
by their functionaries,but they do include military assessments. In terms of
the increasein real expenses (deflatedby Nankaiprice indices, 1926 = 100),
village expenses duringthe early republic, excluding the militarycomponent,
were approximately373 yuan, whereas duringthe early 1930s they were 670
yuan, an increaseof 180 percent. If we include militaryexpenses, the increase
is approximately315 percent. It should also be kept in mind that the increase
in village expenses from around 1905 or so was probablyeven more dramatic
because it was aroundthen that the modernizationmeasureswere first under-
taken, and before that there was little by way of public expense in the village.
In addition, given that, until at least the mid-1920s, transactionsamong the
Hebei peasant population were ordinarily conducted in copper while taxes
were generally collected in silver, the silver-copperexchange ratio is a rele-
vant factor in determiningthe real tax burdenof the peasant. From 1914 until
1923, the index of the exchange rate between silver and copperrose from 100
to 152, a 52 percent increase in less than ten years (SPC 1935:84).
Finally, in orderto assess the termsof tradebetween the ruraland the urban
sectors, I use the indices of commodity prices at wholesale in north China
classified by the Nankai University group (INCP 1937). According to these
figures, the period from 1913 through 1935 saw a situationin which (except
for 1925, 1926, and 1927, when values were almost on par) ruralhouseholds
consistently paid more for what goods they bought than they received for
those they sold.
The qualitative literaturerecording the oppressive characterof this tax
burden is voluminous (Eastman 1975:181-244) and even includes official
reports. In the Nationalist period, statementsof the National Financial Con-
ferences declaredthat abolishing the various forms of excessive taxationwas
the only means of reviving the villages (NFWH 1933.6:36-37; 1934.3:6-8).
Just as significant as the weight of this burdenwas the sheer unpredictability
of the provisionaltankuan,which made productionplans extremely uncertain
for peasant households (Feng 1938a:1105).
The high tax rate and its unpredictabilitymade taxationan importantobject
of ruralunrest. Studies of peasant violence in China throughthe 1920s and
1930s by Lucien Bianco, TanakaTadao, and others emphasize the fact that

antitaxprotests, and not class violence, formed the majorproportionof peas-

ant protest (Bianco 1975:315; Tanaka 1955:379,429). Even communist
cadres in Shandong "tended to believe that rent and interestreductionwere
not importantrevolutionarytasks, since they could neitheractivatethe major-
ity of peasants nor weaken the forces of feudalism in the countryside. ...
Indeed some cadres had found that alleviating the tax burden was the most
pressing demand of the masses" (Pepper 1978:262).
In the villages the involutionaryprocess was expressedas a vicious circle in
which the increaseddemandsof the state led to the proliferationof brokerage,
which in turnled to yet higher demands. Underthese conditions, arbitrariness
in the exercise of power was beginning to insinuateitself at the very heartof
the political process in the village. During the republicanperiod, entrepre-
neurial state brokers began increasingly to replace the various political ar-
rangementsthrough which peasants had traditionallygoverned themselves
and their relationshipwith the state. Elsewhere, I have tried to show how
communityleadership, articulatedthroughreligious, lineage, patronage,and
cooperativenetworks, sufferedconsiderabledisintegrationas it was unableto
meet the increasingdemandsof the state (and other factors)upon community
resources. More and more, the traditionalcommunityleadershipcame to be
replacedby political entrepreneursand state brokerswho began to be known,
in the vocabularyof the countryside,simply as "local bullies" (Duara 1983).
The local bully (tuhao, wulai, eba) was a ubiquitous character whose
notorietywas so great in the 1920s and 1930s that the Nationalistsconducted
a campaigndirected mainly against him. He has been describedas someone
who, in the late nineteenthcentury, possessed enormousphysical power and
was frequentlytied to the network of county underlingsand brokers(Smith
1970:158-69). Althoughhe often used his powers destructively,in those days
he customarilyoperatedon the marginsof the political process in the village.
However, with the involutionaryexpansion of entrepreneurialopportunities
that public office providedduringthe republic, the local bully came to domi-
nate political office in the village. What distinguishedthe bully from other
political leaders in the village was his approachto power; he pursuedoffice
for the entrepreneurialgains, and did so at the expense of the interestsof the
communityhe supposedlyled. In this sense the local bully is to be defined not
as a social category, but as a political type with a distinctive approachto
power. And it is this definition that makes him virtually indistinguishable
from the state broker, who had operatedout of the county only in imperial
times. Small wonder then that contemporaries,such as the reformerLiang
Shuming, speak of the startling visibility of local bullies among revenue
farmers,tax collectors, ward and village headmen, and brokersof every hue
throughoutthe period (Gu 1933:7; Liang 1971:198; Duara 1983:221-26).
This, then, bringsus back to our originalproject,thatof realigningthe split
image of state and society in ruralcommunities, the growth in the roles both

of the state and of social groupscompeting with the state, and the paradoxical
expansion of state power with a simultaneous growth of anarchy. A key
element in this projectis the creationof a core groupof political entrepreneurs
to staff the lower ordersof the expandingstate. An importantcharacteristicof
this group is that, in identity and interest, it is exclusively of neitherstate nor
society, but belongs simultaneouslyto both. These brokersare linked to the
statebecause it is throughthe state that they derive theirpower, yet they form
an interest group in competition with the state that eludes state control and
focusses its entrepreneurialgoals on society. The circumstancesof unprece-
dented state penetrationduring the republic, together with the weakening of
traditionalchannels of influence such as the old networksamong the gentry,
allowed the state brokers to become a dominant predatorypower in rural
society below the county level. Thus, the concept of state involution, which
accountsfor the growth of the state as well as that of the brokers,returnsthe
state to a central position not only in Chinese social history, but also in the
communist revolution.
The prominence of state brokers in the communist revolution is again
revealed in Suzanne Pepper's study of northChina, where we see that, apart
from taxation, local bullies and corruptionwere the most importantlocal
issues in the programof revolutionarymobilization. In JunanCounty, Shan-
dong, local bullies and corruptionwere at the top of targetsof attackby the
communists, while interestrates and rent reductionran a poor fourth. So also
in Laidong County. Baba Takeshi's recent work shows the extraordinary
importanceof the reform of fiscal administrationin the establishment of
communist power in Shandong (Pepper 1978:268-69; Baba 1984:45-48).
If furtherresearch continues to provide evidence for direct causal links
between state involutionin Chinaand the communistrevolution, we will need
to modify several of the arguments offered to explain the revolution. For
instance, a common assumptionaboutthe role of the state in revolutionis that
the state must weaken before a revolutionarycondition can develop. Theda
Skocpol, who otherwise has a complex understandingof the state, can be said
to hold such a view. State involution suggests a more differentiatedperspec-
tive on the state where growth in certainspherescould itself unleasha process
of self-destructionand revolutionarytransformation(Skocpol 1979:27-29).
In anotherexample, ChalmersJohnson's view that the Japaneseinvasion of
China permittedthe communists to mobilize the peasantryby appealing to
their nationalist sentiments would have to account for the persistence of
bullies, corruption,and heavy taxationas mobilizing issues throughthe revo-
lution, issues generated to a significant extent by state involution (Johnson
1962). Certainly all three were burning local issues through the 1920s and
1930s, and elsewhere I have tried to show that Japanese rule often simply
intensified these problems by driving the involutionaryprocess even harder
(Duara 1983:ch.6). This is not to overlook the monstrous violence that the

Japanesearmy perpetratedin the countryside, but to suggest that in certain

equally importantrespects, the effects of the Japaneseregime in northChina
should not be too sharplydifferentiatedfrom those of its native predecessors.
State involutionwas the common mode of expansion runningthroughall the
Chinese republicanregimes.
Finally, to returnto the comparativeproblem of the state, Tilly and his
colleagues focus on European states that had been successfully made and
survived into the twentieth century (Tilly 1975). Tilly himself notes that the
survivorswere only a handfulfrom among some five hundredstateformations
that were in existence in Europe around 1500. He enumeratesseveral useful
historicalconditions that distinguishthe survivorsfrom the bulk of cases that
went under. However, because he focusses on the survivorsratherthan the
failures, he is unableto specify what it was thatthe failureslacked to succeed
in the transition.
Does our study of state involution, of imperfectstate making, suggest any
answers?The first difference that would strike anybody looking at the ques-
tion comparativelyis that the Chinese economy, unlike the successful Euro-
pean ones, was not growing appreciablyduringthe twentiethcentury. Given
that per-capitaincome was more or less stagnantand that state income was
rising throughthe period, this could mean thatthe tax burdenon the populace
was greater in China than in the successful cases. That situation, however,
need not necessarilyhave been the case. The Chinese state could have gener-
ated a good deal of its income by trenchinginto the unreportedincome of the
brokersinsteadof cutting into the consumptionlevels of the ruralpopulation.
The point is that it was unable to do so.3 In the end, the question that the
Chinese experience poses to the Europeancases, as does the experience of
developing states generally, is: How are traditionalstates historically trans-
formedinto moder ones when the very agencies and channelsof transforma-
tion are resistantto change and obstruct change by controlling the flow of
resourcesand information?This is the questionthatmust now be asked about
the successful state formationsif they are to provide a meaningfulframe of
reference for the others.

BabaTakeshi.1984. "SantoKonichiKonkyochini okeruZaiseiMondai"[Problems
in financialadministration baseareas].Shikan,110
in the Shandonganti-Japanese

3 An importantfactor behind the Chinese communists' ability to generatecritical increasesin

revenueduringthe early period was their eliminationof state brokers, which they accomplished
by building political organizationslinked to state structuresfrom the grass-rootslevel upward.
Incidentally,with the gradual weakening of collective structuressince the agrarianreforms of
1978, we may presentlybe witnessing a recrudescenceof entrepreneurialbrokeragemediationsin
local society.

Bianco, Lucien. 1975. "Peasants and Revolution." Journal of Peasant Studies, 3:2
(April), 313-35.
Braun, Rudolf. 1975. "Taxation, Sociopolitical Structure,and State Building: Great
Britainand Brandenburg-Prussia"in The Formation of National States in Western
Europe, Charles Tilly, ed., 244-327. Princeton.
Chen Juren. 1934. "Dingxian Tianfu Xianzhuang" [The presentconditionof the land
tax in Ding County]. Minjian Yuekan, 1:4 (June).
Chang,C. M. 1934. "Local GovernmentExpendituresin China." MonthlyBulletinof
Economic China, 7:6 (June), 233-46.
Ding Da. 1931. ChagokuN6son Keizai no Hokai [The disintegrationof the Chinese
ruraleconomy] (translatedfrom the Chinese). Dalien.
Duara, Prasenjit. 1983. "Power in Rural Society: North China Villages, 1900-
1940." Ph.D. diss. HarvardUniversity.
Eastman,Lloyd. 1975. TheAbortiveRevolution:China underNationalist Rule 1927-
1937. Cambridge, Mass.
. 1984. "New Insights into the Nature of the Nationalist Regime." Republican
China, 9:2 (February),8-18.
Feng Huade. 1934. "Local GovernmentExpenses in Hopei." Monthly Bulletin on
Economic China, 7:12 (December), 507-13.
. 1935. "Xiandifang Xingzhengzhi Caizheng Jichu" [The financial basis of
local administrationin the county]. ZhengzhiJingji Xuebao, 3:5 (July).
. 1938a [Written in 1932]. "Hebeisheng Gaoyangxiande Xiangcun Caizheng"
[Village finances in Gaoyang County, Hebei], in Zhongguo Jingji Yanjiu, Fang
Xianding, ed. Changsha.
. 1938b [Written in 1934]. "Hebeisheng XianCaizheng Zhichuzhi Fenxi" [An
analysis of county expenditures in Hebei Province], in Zhongguo Jingji Yanjiu,
Fang Xianding, ed. Changsha.
. 1938c [Written in 1935]. "Wuguo Xianshouru Zhiduzhi Tezheng" [Special
features of the system of county revenues in our country], in Zhongguo Jingji
Yanjiu, Fang Xianding, ed. Changsha.
Feng Huade and Li Ling. 1936. "Hebeisheng Dingxianzhi Tianfu" [The land tax of
Ding County]. ZhengzhiJingji Xuebao, 4:3 (April).
Fewsmith, Joseph. 1984. "Response" to Lloyd Eastman. Republican China, 9:2
(February), 19-27.
Gamble, Sidney D. 1944. "Hsin Chuang: A Study of Chinese Village Finance."
HarvardJournal of Asiatic Studies, 8:1 (March), 1-33.
. 1968. Ting Hsien: A North China Rural Community. Stanford.
Geertz, Clifford. 1963. AgriculturalInvolution:The Process of Ecological Change in
Indonesia. Berkeley.
Geisert, Bradley. 1984. "Response" to Lloyd Eastman.RepublicanChina, 9:2 (Feb-
ruary), 28-39.
Gu Meng. 1933. "Benggui Guozhengzhongzhi Hebei Nongcun" [Hebei villages in
disintegration].Zhongguo Jingji, 1:4/5 (August).
HNK. 1939. Hokushi no Nogyo Keizai [The agriculturaleconomy of north China].
HSZ. 1936. Hokushi Shijo Zokan [A broad look at conditions in north China], Nan-
ManshuTetsud6 KabushikiGaisha, comp. Dalien.
INCP. 1937. "Index Numbers of Commodity Prices at Wholesale in North China
Classified by Stages of Production(1926 = 100)." Nankai Social and Economic
Quarterly, 9:4 (January), 1085-87.
Johnson, Chalmers A. 1962. Peasant Nationalism and Communist Power: The
Emergence of RevolutionaryChina, 1937-1945. Stanford.

Kuhn, Philip A. 1979. "Local Taxationand Finances in RepublicanChina," in Select

Papers from the Centerfor Far Eastern Studies, No. 3. 1978-79, Susan Mann
Jones, ed., 100-36. Chicago.
Kuhn, PhilipA., and Jones, Susan Mann. 1979. "Introduction"to Select Papersfrom
the Centerfor Far Eastern Studies, No. 3. 1978-79, Susan Mann Jones, ed., v-
xix. Chicago.
Li, Chuan-shih. 1922. Central and Local Finance in China. New York.
Li Ling. 1935. "Hebeisheng QinghaixianzhiTianfu ji chi Zhengshou Zhidu" [The
land tax and its system of collection in Qinghai County]. TianjinDagongbao (27
. 1938. "Hebeisheng Qinghaixianzhi Tianfu ji chi Zhengshou Zhidu" [The
land tax and its system of collection in Qinghai County] in ZhongguoJingji Yanjiu,
Fang Xianding, ed. Changsha.
Liang Shuming. 1971. "Gangao JinzhiyanDifang Zizhizhe" [Warningto people who
talk of self-governmenttoday], in ZhongguoMinzuZiqiu YundongzhiZuihouJuewu
[The final awakeningof the Chinese people's movementof self-salvation]. Taipei.
Liu, K C. 1978. "The Ch'ing Restoration," in Late Ch'ing, 1800-1911, Part 1, John
K. Fairbank,ed., Vol. X of The CambridgeHistory of China, 409-20. Cambridge.
Mann, Susan. 1984. "State Building and Market Taxes in Presocialist China."
NFWH 1933-35. Nongcun Fuxing WeiyuanhuiHuibao [Reportsof the Committeeto
Revive the Villages], AdministrativeYuan, Nongcun Fuxing Weiyuanhui, ed.
Peng Yuxin. 1945. Xian Difang Caizheng [Local finances at the county]. Zhongjing.
Pepper, Suzanne. 1978. Civil War in China: The Political Struggle, 1945-1949.
Rowe, William T. 1983. "Hu Lin-i's Reform of the GrainTributeSystem in Hupeh,
1855-1858." Ch'ing-shih Wen-t'i. December, 33-86.
ShenbaoNianjian [Shenbao yearbook]. 1966 [1935]. Shanghai;rpt. Taibei.
Skinner, G. W. 1977. The City in Late Imperial China. Stanford.
Skocpol, Theda. 1979. States and Social Revolution: A ComparativeAnalysis of
France, Russia, and China. Cambridge.
Smith, ArthurH. 1970 [1899]. Village Life in China. New York.
SPC. 1935. Silver and Prices in China, Committeefor the Study of Silver Values and
CommodityPrices, Ministry of Industries,comp. Shanghai.
Sun Shaocun. 1936. "Difang Caizheng duiyu Nongcun Jingjide Yinxiang" [The
impact of local finances on the village economy]. Zhongguo Nongcun, 2:9
Sun Zoji. 1935. Zhongguo TianfuWenti[Problemsof land taxes in China]. Shanghai.
Tambiah, Stanley Jeyaraja. 1985. Culture, Thought, and Social Action: An An-
thropological Perspective. Cambridge, Mass.
TanakaTadao. 1955. KindaiShinaNoson no Hokai to Nomin T6os [The disintegration
of the village and peasant struggles in modern China]. Tokyo.
Tilly, Charles, ed. 1975. The Formation of National States in Western Europe.
Wakeman,Frederic. 1975. The Fall of Imperial China. New York.
Wang, Yeh-chien. 1973. Land Taxation in Imperial China, 1750-1911. Cambridge,
Xu Zhengxue. 1936. "Nongcun Wenti-Zhongguo Nongcun Benggui Yuanyinde
Yanjiu" [Village problems-Research into the causes of the collapse of the Chinese
village]. Nanjing.
Yeh, K. C. 1979. "China's National Income, 1931-1936," in Modern Chinese
Economic History, Chi-ming Hou and Tzong-shian Yu, eds., 95-128. Taipei.

Young, Arthur. 1970. "China's Fiscal Transformation,1927-1937," in The Stren-

uous Decade: China's Nation-BuildingEfforts, 1927-1937, Paul T. K. Sih, ed.,
83-121. New York.
Zelin, Madeleine. 1985. The Magistrate's Tael: RationalizingFiscal Administration
in Eighteenth-CenturyCh'ing China. Berkeley.
Zhang Sen. 1936. "Tianfu yu Difang Caizheng" [Land taxes and local finances].
Dizheng Yuekan,4:2-3 (March), 192-203.
Zhang Yifan. 1935a. "Hebeisheng Difang CaizhengzhiJiantao" [An investigationof
local finances in Hebei Province]. Fuxing Yuekan,4:4 (December).
. 1935b. "Shandongsheng Difang Caizhengzhi Jiantao" [An investigation of
local finances in Shandong Province]. Fuxing Yuekan,4:4 (December).