How to best maintain public goods in the realm of natural resources has long been a thorny issue for comparative governance. Such difficult and ultimately political questions as “how to best evaluate benefits to then apportion to individuals and collectivities” are complicated, as “the benefits are difficult to measure, value, or capture; occur in the long run, and are global, national and local. The costs, however, tend to be local and immediate” (World Bank 2000). Even advanced industrial democracies wealthy enough to “afford” conservation have a historically mixed record in preserving natural resources. Much of the planet's remaining wealth of natural resources—from forests to oil to minerals—is now to be found in the developing world, which faces a range of even more serious challenges in evaluating, administering, and preserving these public goods, as states there are subject to conditions that compound the difficulties of balancing short-term, individual interests and long-term, collective ones. The state's extractive and redistributive institutions are typically weak (and so it is much harder to redistribute a portion of the surplus from wealthy urban areas to poor rural ones), natural resource concessions tend to be captured by small and unrepresentative economic elites or foreign companies through special monopolies, population to arable land ratios are high, and remote areas where natural resources need the most protection are exactly those areas where the pressure on natural resources is likely to be the most severe—as there are in many of these areas relatively few other opportunities for earning a living. Officials entrusted with implementing locally unpopular conservation initiatives are often thinly stretched and face an unpleasant set of choices: finding themselves at odds either with local communities or with higher levels of the state as they deflect these initiatives. Environmental issues in general, and forestry management in particular, cut straight to the heart of how to administer the preservation of public goods under conditions of chronic administrative and resource shortages: balancing individual and collective interests, negotiating a consensus between short-term incentives and long-term sustainability, and, most critically, establishing sufficient government institutional capacity to regulate both (Dauvergne Reference Dauvergne1997; Doornbos, Saith, and White Reference Doornbos, Saith and White2000; Peluso Reference Peluso1992; Poffenberger Reference Poffenberger1990, 7–26).

This article focuses on a set of administrative reforms undertaken in China's forestry sector at the end of the last century, and places them in a wider historical and institutional trajectory of state transformation. There are three reasons for the focus on forestry. First, as environmental concerns, particularly carbon dioxide emissions and questions of biodiversity, climb higher up the agenda of international concerns, China's own position as a “carbon sink” and reservoir of plant biodiversity has become an important global issue in its own right. Second, forestry is an important sector within a substantial and growing literature in China area studies that considers China's chronic problems in implementing environmental regulations. These works focus largely on the political and bureaucratic incentives that push local cadres to privilege development over the environment and on the unresolved problem of unclear property rights (Harkness Reference Harkness1998; Ho Reference Ho2001; Jahiel Reference Jahiel1997; Smil Reference Smil1993). And third, because forests cover large and varied geographic areas and forestry policy making includes an unusually diverse set of core agendas and programs, forestry in microcosm offers an arena in which to make a preliminary assessment of policy making and implementation in the contemporary Chinese state, particularly at more local levels of governance. It is this final concern that is the subject of this article.

In the past twenty years, a vast literature on policy implementation and local governance in China has emerged. Broadly speaking, there are four overlapping perspectives that seek to make sense of the Chinese state and its ongoing gaps in policy implementation: (1) state weakness and retreat (Shambaugh Reference Shambaugh and Shambaugh2000), (2) state predation and corruption (Bernstein Reference Bernstein, Goldman and MacFarquhar1999; Lü Reference Lü Xiaobo2000), (3) bureaucratic bargaining between different layers of the state (Lampton Reference Lampton, Lieberthal and Lampton1992; Lieberthal Reference Lieberthal, Lieberthal and Lampton1992), and (4) the particulars of bureaucratic incentives for on-the-ground officials (O'Brien and Li Reference O'Brien and Li1999). Much of the empirical evidence presented here confirms elements of each of these lines of analysis: the state is retreating from complete ownership and heavy-handed direct management of all aspects of natural resources, there is clear evidence of corruption and predation in local forestry practice, local forestry officials do bargain with higher-level forestry officials to obtain better deals, and bureaucratic, financial, and market-based incentives do matter for decision making and policy implementation.

I begin with a slightly different assumption than is typical of the other four perspectives: that China's forestry reforms at the end of the twentieth century are part of a wider, more systemic process of state transformation in which the core question is how the state, particularly at more local levels, attempts to augment and transform its own institutional capacity to deliver new policies when the structural conditions for so doing are problematic. China is still very much a state in transition, one in which the institutions established to fulfill one set of tasks (managing a planned economy, indoctrinating the population, and extending social insurance through the institution of the work unit) are now in the process of adapting to a range of more complex agendas. The flip side of the “retreat” from the state-run economy and high levels of political indoctrination is, in fact, the state's attempted expansion of capacity in service delivery and regulation of emerging markets and social actors. This is visible in such diverse policy areas as tax reform (Yep Reference Yep2004); the creation of pilot programs to provide social insurance such as pensions, health care, and unemployment counseling to the citizenry at large (Duckett Reference Duckett2004; Jackson et al. Reference Jackson, Sleigh, Peng and Liu2005; Lee and Warner Reference Lee and Malcolm2003; Liu Reference Liu2004; Shue and Wong Reference Shue and Wong2007); and, of course, environmental initiatives. All of these new areas of state activity have to be implemented by local levels of government when the fiscal basis for so doing is shaky and the supervisory capacity of higher levels of the state is extremely uneven. How higher levels of the state are able to vertically integrate this process, make meaningful rules and gather relevant information, dampen corruption and rent seeking, monitor success and failure in this process of transition, and, perhaps most importantly, garner the compliance of local state agents remain open questions.

When compared to either forestry regimes elsewhere in the developing world or to new realms of augmented state activity in China (tax, social insurance), China's forestry sector enjoys several advantages. In the period under consideration here, severe floods on the Yangzi and Nujiang rivers in 1998 shocked the leadership of the Chinese Communist Party into a strong conservation-minded consensus, and since then, conservation and afforestation initiatives have enjoyed a degree of political support and degree of investment unusual in either the developing world or, for that matter, in other policy areas in China.

Within China, significant parts of the forestry portfolio, notably tree planting campaigns, have a well-established repertoire of administrative practices to fall back on, unlike new social insurance systems that have to be built from scratch. Forestry's institutional and spatial scales are also unusually broad: in geographic coverage, in the local diversity of tenure arrangements, in the tendency to be concentrated in areas of officially designated poverty, and in the long-standing tensions within forestry's core institutional goals of conservation, tree planting, and wood product production. But equally, the ways in which forestry is atypically broad, locally diverse, and subject to cross-cutting policy agendas mirror many of the features of the wider Chinese state, making it an excellent illustrative case study for these ongoing processes.

There are substantial literatures within forestry and natural resources management that look to either market incentives (Hyde, Belcher, and Xu Reference Hyde, Belcher and Xu2003) or community involvement (Peluso, Turner, and Fortmann n.d.) as possible ways to resolve the wide range of issues within forestry portfolios, in both China and other countries. Although China's forestry sector is not subject to the influence of foreign multinational organizations, nor is China a net forest product exporter, markets and communities both matter, albeit in somewhat different ways than elsewhere in the developing world: the relevant markets are regional and national rather than international, and nongovernmental organizations in China play a relatively smaller role in community participation and governance. I suggest that irrespective of whether a program is geared toward markets and production, community participation, or some combination of the two, local administration is a key factor in program effectiveness, or the lack thereof. In this respect, the dilemma of China's forestry sector is repeated throughout the developing world and in other policy areas in China: how to fundamentally transform and dramatically expand the capacity of state institutions at local levels when the existing fiscal and administrative bases for so doing are inadequate. I argue that under conditions of chronic resource shortages and imperfectly functioning systems of bureaucratic control and oversight, the “softer” elements of politics and culture, including appeals to norms, the mobilization of moral commitments through particular rhetorics, and the deployment of symbols are an important complement to the “harder” mechanics of bureaucratic and economic incentives for the transformation of local state capacity. Structural incentives clearly matter, but so do normative values. Reform in formal organization both reflects and fosters changes in administrative capacity; informal organization and leadership have the potential to either support or subvert the policy imperatives of formal organization. It is through the interaction of these different elements that China's administrative capacity will (or will not) be decisively transformed.

Structural, Normative, and Institutional Legacies of the Twentieth-Century Chinese State

China's efforts to modernize and expand institutional capacity in the forestry sector have been informed by a long-lived repertoire of administrative strategies employed by the aspiring state makers of the twentieth century, from the regents for the last two emperors, through warlords, the Republican-era Guomindang, and the People's Republic from 1949 onward. In formal dance-like lockstep, central and provincial levels of the state alternate between stressing the need for stronger formal systems of control and audit and appealing to the moral norms of local officials; chronically underresourced and ill-supervised local officials alternate between concealment, bargaining, feigning compliance, and partial implementation. The central state's oscillation between attempts to “rationalize” formal systems of reporting, auditing and control, and informal appeals to the moral and normative commitments of local cadres, and the local state cadres' corresponding efforts to formally indicate compliance while pleading their own special circumstances, are by no means empty rhetoric. In practice, these are very old integrative devices that do more than paper over glaring administrative gaps in a still nonrationalized and underfunded system; they provide enough surface cohesion, shared language, and shared values to enable administrative measures (including reform programs) to be worked through on the ground. With its ethnic near homogeneity, vast scale, coercive apparatus, centuries-old tradition of rulership and monocratic authority, and ongoing lack of anything approaching open political pluralism (save in Taiwan), the Chinese state may look strong, unitary, and tough in comparison to states in other developing countries. In practice, however, the Chinese state is “pervasive but weak” (Bendix Reference Bendix1959, 112, after Weber), with limited real powers in systemic taxation, auditing, and bureaucratic oversight. Lacking a sufficient basis in tax revenue, it has been impossible to develop an impartial and salaried cadre of central and provincial administrators capable of directly monitoring what goes on at more local levels of government and providing the financial incentives for local governments to implement the types of programs that higher levels would like to see. For their part, local governments have been caught in a triple bind: prevented from levying new taxes, expected to be largely self-financing, and subjected to increasing barrages of circulars and targets for raising standards in a wide range of service provision, from local education to road building to environmental protection (Wong Reference Wong1997, Reference Wong and Donald1998). In the absence of both systemic monitoring and financial support from above, local governments have resorted to filling in gaps in a variety of semiformal and informal ways. The most important of these has been through a proliferation of off-budget revenues that may (or may not) be perceived as legitimate, but are in any case beyond the purview of formal supervisory control from above. The government's current commitment to rationalizing fees and surcharges through tax-for-fee reform is attempting to deal with this problem at present, but the prognosis for uniform implementation of this program is poor, as success depends heavily on the unproven ability of the central government to provide comprehensive funding (Yep Reference Yep2004).

In addition, a particular set of historical legacies from Maoist state socialism have exacerbated these more general dilemmas of institution building: the majority of local cadres reside permanently in the localities they administer; lines of authority and accountability continue to remain ambiguously split between functional hierarchies (tiao) that send down directives and receive reports, while territorial units (kuai) with often weak commitment to environmental initiatives fund functional bureaus and have final authority over them; and property rights are statutorily unclear and vigorously contested by different state bureaus (Ho Reference Ho2001). Finally, throughout the 1990s, the ongoing conflation of production and government functions in the northeastern and southwestern state-owned forests continued to result in heavy extra burdens on state forestry bureaus, which were responsible for economic performance as well as the governmental functions of social welfare, public security, infrastructural provision, and public health (Harkness Reference Harkness1998; Linye Bu Zhengce Fagui he Tizhi Gaige Si 1995, 140–44).

The current challenge for China's central government revolves around holding together such a large, underresourced unitary state while meeting rising demand for increasing provision of state services; the corresponding challenge for local state administrators is to fulfill (or at least to be seen to fulfill) enough of the center's core programs and dictates enough of the time to avoid sanctions and, for the fortunate and talented, to win positive recognition and upward mobility despite chronic underfunding. Central strategies for governing under these conditions have been remarkably consistent since late imperial times: (1) publicly and bureaucratically touted formal and standardized rules and regulations that are often deliberately left vague to allow for local circumstances and flexibility in implementation, and (2) the manipulation of moral incentives through exhortation as positive, normative tools to rally local officials into promoting central state dictates while signaling which set of sent-down rules and dictates are to be taken particularly seriously. For their part, local administrators rely on an equally long-standing repertoire for fulfilling (or seeming to fulfill) central and provincial directives: concealing local information (ranging from who actually uses the land to how much deforestation has taken place), bargaining with higher levels of the state over targets, and resorting to semiformal and informal supplements in funds, labor, and local elite involvement to get the system to work at all.

Structural Constraints and Chronic “Leakiness” in the Forestry Regime prior to 1998

As of the Fifth National Survey on China's Forest Resources (1994–98), China had roughly 159 million hectares of forest land and forest cover of 16.55 percent—a notable improvement on the 133.7 million hectares of forest land and forest cover of roughly 14 percent from the previous survey of 1989–93.Footnote ¹ But this aggregate improvement was from a baseline of severe deforestation and one of the lowest per capita averages of forest cover ratios in the world (Harkness Reference Harkness1998, 188; Shi et al. Reference Shi, Li, Lin and Zheng1997; State Environmental Protection Administration of China 2000). China's post-1949 forestry regime has been divided into two types: state-owned forests (guoyou lin) of large, state-run forestry bureaus and smaller, more fragmented collective forests (jiti lin) that are, in theory, managed by village collectives. The former dominate in the mostly coniferous forests of the northeast (Heilongjiang, Liaoning, Jilin, as well as Inner Mongolia); the latter are more common in the mixed broadleaf deciduous forest of central and southern China, and the ecologically sensitive subtropical and Himalayan mountain forests of the southwest (Sichuan and Yunnan) are roughly equally divided between the two types of regimes.

The state forestry administration's portfolio has historically been one of goals and programs that are in tension with each other: extraction of forest resources (logging) to fill national requirements and keep forestry production units in business, afforestation to replenish those resources and contain erosion, and the conservation of existing forest to preserve biodiversity. In aggregate, both state-owned forests and collective forest regimes remained subject to chronic “leakiness,” despite a growing rhetoric of environmental sustainability and an accompanying range of administrative initiatives to dampen logging, strengthen conservation, and promote afforestation throughout the 1980s and 1990s. In particular, the administrative systems to regulate the consumption of extant forest resources, enforce logging quotas, implement a permit system for forest land use (lindi xuke zheng), and dampen illegal forest land occupation were weakly and inconsistently enforced (Li Changjian Reference Li Changjian1998).

In fin-de-siècle China, party, state, and enterprise functions continued to be fused in state forestry corporation bureaus. Throughout the reform period of the 1980s and 1990s, it was up to these production-oriented forestry bureaus to enforce logging quotas handed down from above through an administrative system of integrated official permits for logging, transport, and sale called the sanzong liang, or “The Three Integrated Measurements.” Although some informants insisted that “the system was severe” (zhidu yanzhong), a good deal of evidence suggests that while the formal regulations of the sanzong liang may have been severe, the effective implementation of those regulations was anything but. Putatively strict logging quotas were regularly surpassed in Heilongjiang and elsewhere (interviews, Special Inspector J, Harbin; Special Inspector S, Dongjingcheng). Written records at least indirectly confirm the oral accounts, as official logging statistics for 1996 conspicuously omit all of the northeast save Jilin from its final tables (CFY 1997, 67), in turn suggesting that official evasion of quotas was chronic and widespread elsewhere in the northeast. Although the figures are not broken down, a terse official summary for 1997 notes that “a minority of provinces surpassed their logging quotas by over 5,000,000 cubic meters of wood, the forestry corporations of the northeast and Inner Mongolia regularly went over quota, that permits were only issued for around 40% of the trees cut, and that average provinces only checked up on half of the logging operations” (CFY 1998, 106). Creative accounting whereby logging would be counted against the following year's quotas was commonplace, as was after-the-fact negotiation of more relaxed quotas, and collusion throughout the system in underreporting timber harvests (Official W, Daxing Anling; Official J, Heilongjiang Forestry Corporation; Zhang Reference Zhang2000). Almost all officials pointed to the difficulty of reducing logging in areas where forestry corporation bureaus were both primary employers and responsible for high social welfare bills, with many reiterating that “conservation can't work if people aren't eating,” and one official going as far as to bluntly state that “local enterprise interests and national interests simply aren't the same” (Official J, Heilongjiang Forestry Corporation).

This system of complete state control was monitored by highly trained and knowledgeable “special forestry inspectors” (tezhuan jianduyuan), of which there were two types: a small number specially dispatched to provincial capitals from the central Forestry Bureau in Beijing, and many more sent from either the provincial capital or the county-level forestry bureau to assess operations on the ground. Despite their comprehensive knowledge of local conditions, the inspectors were hampered in three ways: their writ only extended to temporarily stopping unauthorized logging and recommending action to their superiors, their salaries were paid and offices provided by the very production units they were monitoring, and their workloads far outstripped their capacity (Special Inspector B, Xilinji; Special Inspector M, Hailin). In the words of one inspector, “There was no way for the system to be enforced … you just can't properly supervise people you work for!” (Special Inspector W, Daxing Anling). Possessing neither real fiscal independence nor the direct authority to assess fines for over-quota logging, inspectors, whether dispatched from the central, provincial, or county-level forestry bureau, tended to be at least partially “captured” by the districts they were ostensibly regulating. Even the Harbin-based centrally appointed special inspector admitted that “there is no one reason for over quota cutting. But when enterprises are bankrupt, people can't eat, and the forestry bureau has no choice but to let people cut the trees; after investigating and finding these difficulties, special inspectors also have no way to enforce these policies, as people's livelihoods and social insurance depend on income from logging” (Special Inspector J, National Forestry Bureau). Field inspectors also commonly concurred with the local bureau view that subsistence had to come before conservation, and that local people needed to be able to make a living first.

The administrative situation in the collective forest regions of southern and western China was quite different, but led to similarly dismal outcomes for conservation of the extant forest. In contrast to the heavily officially controlled state forests of the northeast, where officials were infrequently challenged and indeed seldom needed to issue forest land use permits to individuals, because “no one up here thinks ‘this land is mine’” (Special Inspector S, Dongjingcheng), decollectivization in central and southern China combined with legally vague property rights and unclear property boundaries to prompt exactly the opposite: a great deal of conflict and resistance to local cadre attempts to curb unregulated logging and land use. Despite the announcement of a regulatory regime of forest land use permits (lindi xukezheng), the implementation of this program was weak to nonexistent. For example, in Guangxi, the land use permit regime as an effective means for containing logging and regulating the use of forest land seems to have been dead on arrival: in many areas, local forestry authorities were simply unable or unwilling to contain “unauthorized” appropriation of forest land (Guangxi University Forestry Institute Roundtable; Officials P and H, Guangxi Forestry Bureau).

In 1998, a central review of forest land use revealed how poorly administrative oversight functioned in the sector: in aggregate, local forestry departments reviewed and approved 953 of 1,527 projects, only gave permits for roughly two-thirds of the forest actually logged, collected less than one-fifth of the reforestation fees due after logging had taken place, and were unable to stop more than 200 illegal projects even after they were banned in 1998. The heavily state forestry–dominated provinces in the northeast had the fewest administrative problems with compliance—probably because at that time, there were virtually no forest land use initiatives outside of the state-managed regime. There was, however, significant regional variability within the collective forest regimes of central and south China. Some areas did quite well; Fujian had a compliance rate over 99 percent, was paid almost all of the reforestation fees due, and had no outstanding illegal projects in operation. But virtually all of the provinces with low rates of permit issuance and reforestation fee recovery had significant or exclusively collective sectors. According to official statistics, the important province of Sichuan only reviewed and approved 11.7 percent of its projects, issued logging permits for 5.7 percent of the trees actually felled, recovered 34.2 percent of the afforestation fees due, and was unable to stop twenty-seven unauthorized projects. The much smaller province of Jiangxi only assessed 17.7 percent of the projects it ought to have, issued logging permits for a paltry 1.3 percent of the wooded areas it ought to have regulated, recovered only 7.5 percent of afforestation fees due, and had thirty-nine banned projects still in operation when the review was conducted. Other provincial-level units such as Chongqing had mixed records; Chongqing scored highly on approval of projects (at nearly 85 percent), but issued less than 1 percent of the logging permits that it ought to have, and only recovered 27 percent of the afforestation fees it was owed (CFY 1999–2000, 112–13).

Ongoing mismanagement (or nonmanagement) of forest land use permits and unapproved conversion of forest land use both exacerbated disputes over land use. In the collective forest regions, where individuals strongly held the view that “this land is mine—no one should tell me what to do with it,” chronic contestation and disputes that often flared into open violence were the norm (Guangxi University Forestry Institute Roundtable). Cash-strapped local governments who resorted to extra levies on forest products to keep afloat could only add to the local resentment; one survey in Guangxi revealed that counties levied anywhere from seven to nineteen different kinds of fees that typically accounted for between 30 percent to 50 percent of the product value (Li Mingfeng Reference Li Mingfeng1999), and elsewhere in south and south central China, surveys have suggested that although this varied from place to place, “approved taxes and fees absorb 42–50% of the sale value of wood and bamboo cane. Unofficial charges absorb up to 10% more” (Liu and Landell-Mills Reference Liu, Landell-Mills, Hyde, Belcher and Xu2003, 51).

Thus, while the dynamics of the weak regulation could not have been more different, neither the state forestry bureaus in the northeast nor the collective forestry administration of the south and southwest did a particularly good job in either containing logging or protecting old-growth forest. The state-owned and state-managed forestry regime presumed that chronically underresourced and self-reliant production units would be able to police themselves, while the collective forestry regime presumed that underresourced, self-reliant, and thinly stretched local officials would be able to police a society with increasingly vigorous claims to land use. The “leakiness” of these two assumptions was suddenly laid bare with the floods of 1998, which laid the ground for the sudden concentration of political will behind both conservation and afforestation from then on.

The crisis of the 1998 summer floods on the Yangzi and Nujiang rivers sparked the overnight emergence of a strongly conservation-minded consensus within the top political leadership, which renounced the incrementalism of previous years in favor of strongly assertive conservation measures. In late July 1998, the State Council signaled a new hard line with a circular on the importance of conservation of natural resources that banned “indiscriminate felling of trees” (luankan luanfa) and “chaotic squatting on forestry land” (luanzhan lindi). This was followed up by a high-profile announcement by Premier Zhu Rongji in August, promoting a widely publicized “32 character policy” of “planting trees to close off mountain passes; returning agricultural land to forest” (fengshan yulin, tuigeng huanlin); these sentiments were repeated when Zhu toured the flood disaster area in Heilongjiang at the end of the month (CFY 1998, 471). These general statements were supplemented by a barrage of administrative fiats from above: countrywide logging ban on slopes steeper than 25 degrees and sharp national reductions in logging quotas. Sichuan opted for a total ban on logging in parts of fifty-seven counties in the ecologically sensitive Himalayan highlands (CFY 1998, 40–42, 471). Initiatives to “return agricultural land to forest” (tuigeng huanlin), more colloquially known as the “Grain for Green” program, were launched in 1999, along with sharp crackdowns on “indiscriminate squatting on forest land.” The autumn and early winter of 1998 also saw a flood of nationally televised environmental awareness campaigns that stressed the dangers of erosion and desertification, and eventually a large commitment of resources to accelerate afforestation projects in areas particularly vulnerable to soil erosion.

In Heilongjiang and Sichuan, forestry officials unanimously applauded the central government's commitment to afforestation and conservation, and welcomed the likelihood of increased central investment in their districts. But most also had private reservations about the way in which the central government was setting about conservation policies; after all, in the early to mid-1990s, the central state had already failed quite miserably in implementing a meaningful system of forest land use permits. Given conditions of chronic underresourcing, lack of clear functional administrative oversight and auditing, and the sheer diversity of the conditions in the countryside, how would these even more ambitious policies become translated into action?

Strategies for Enhancing Capacity: Expanding and Strengthening Formal Bureaucratic Systems

Higher (central and provincial) levels of the Chinese state have historically had two options at their disposal in their efforts to enhance administrative capacity and responsiveness at lower functional levels in the bureaucracy: (1) the assertion of rationalization and control over lower levels of the state (county and subcounty) through strengthening the formal and technical instruments of reporting, auditing, and supervision, and (2) the mobilization of the normative commitments of its thinly stretched ground agents through a combination of education, training, exhortation, and the (limited) provision of incentives.

The first strategy that the central state has deployed to enhance accountability and capacity has been classic in its Weberian methods of institution building from above. The central forestry bureau in Beijing has systematically attempted to assert more control from above by “strengthening administrative systems” (jiaqiang guanli)—a rhetoric that has been replicated further down the line in provincial capitals toward county and township layers of government.Footnote ² What this means in practice is an acceleration of the rhetoric of central control through a barrage of administrative directives, circulars, and orders continually fired off in the direction of subordinate offices to disseminate knowledge about policy and good practice, in combination with efforts from above to formalize and regularize (zhengguihua) systems of control. In addition to spreading the word about what higher levels want done, the attempt to formalize and regularize has been promoted in a number of ways.

Particularly in the state forestry regime of northeast China, the first and most important measure of “administrative strengthening” has been movement at the provincial level to further centralize and upgrade the status of special forestry inspectors. To the relief and unanimous applause of special forestry inspectors in Heilongjiang and Daxing Anling, their salaries were taken over in full by the relevant provincial-level office in Harbin or Jiagedaqi in late 1998, with the result that, for the first time, they could reasonably resist pressure from the units they were monitoring by shrugging “I'm just doing my job” and getting on with things. One inspector reported that “things have been going much better since our complete independence was established” and another suggested that “without this kind of [administrative] strengthening, anything else is just empty talk” (Special Inspector B, Xilinji; Special Inspector W, Daxing Anling).

The second part of “administrative strengthening” has been to create greater ease of monitoring through simplification, standardization, and the selective use of technical instruments such as geographic information systems. The most notable example of this has been a central bureau effort to push through a complete reclassification of China's forest land. The new forestry classification scheme drastically simplifies a complex scheme of five different types of forestry land to only two: ecological public forests (shengtai gongyi lin) in which logging is completely forbidden, and commercial and economic forests (shangpin jingji lin) that permit some cutting of commercial tree crops. The guidelines for the new classification scheme suggest a ratio of somewhere between 60:40 and 70:30 of ecological to commercial and economic forests (CFY 1999–2000, 59–60; Zhang and Wang Reference Honming and Ling1998). In the short term, this kind of draconian simplification has forced local officials to take stock of the forest land under their regulatory management and pass a large amount of information about forest land use up to higher levels in the bureaucracy. In the longer term, the effect of radical administrative simplification has been to undercut extant wiggle room for local officials to squirm out of environmental targets and make for greater ease of monitoring from above.

And finally, the process of “formalization” has been both directly and indirectly supported by concrete financial incentives through grants from higher levels of government, often in partnership with international organizations such as the World Bank. These programs have provided financial carrots (access to centrally disbursed investment funding for afforestation) in return for strict monitoring of compliance from above. While the specific terms for each grant vary, local authorities typically apply for afforestation grant and loan projects and commit to providing some portion of the finance and all of the land and labor; the province, the central government, or an organization such as the World Bank commits to disbursing a percentage for the total project, with subsequent funding conditional on seedling survival rates of between 75 percent and 85 percent. Of particular importance is the way in which the provision of funds from above makes it possible for higher levels of the government to contain self-reporting (and potential concealment) from lower levels and demand independent and systemic monitoring from above, normally from a combination of the provincial-level forestry bureau and a supplementary independent provincial level auditing organization. In locations as different as the state-owned forest bureaus of Heilongjiang and the mixed state and collective region of Sichuan, programs have explicitly established positive incentives by linking seedling survival rates to the disbursement of the following year's funds and, critically, have backed up those positive incentives with more systemic and robust methods of monitoring and auditing. In the words of senior engineers involved with these new programs, “With more grant aid coming in from above for tree planting, the supervision of management is becoming increasingly strict” (Senior Engineer Z, Sichuan Planning and Investigation Institute; Official L, Chengdu Municipal Forestry Bureau; Special Inspector S, Dongjingcheng). Thus, the wider process of centralization and formalization concurrently attempts to make its directives clear, simplify administration, and pair positive incentives (more funds) for lower administrative levels with negative ones (an expanded monitoring presence from higher levels) to ensure compliance.

These steps to formalize and extend administrative systems of control from above are significant and can be expected to continue. But the fiscal and administrative resources for taking this process to its logical conclusion through full rationalization of local government finance and administration do not exist in the present, nor are they likely to come into being in the foreseeable future. Given the still large difference between what higher levels of the state wish to see implemented at the local level and what it can afford to directly subsidize or administratively monitor on a routine and ongoing basis, the central and provincial layers of the state still have to find relatively inexpensive, indirect ways to get their programs realized. They do so much as the Maoist, Republican, and late imperial regimes of the Chinese state did in their time: by attempting to mobilize the positive normative commitments of their agents on the ground. Higher levels of the state disseminate positive and negative models through specialist publications, they “bring cadres to the center (or provincial center)” by convening special training courses and annual forestry conferences to socialize them into the procedures and the norms of new programs and targets decided from above, and they continue to disseminate ideal types of the “good cadre” through old fashioned political study sessions.

While these methods seem to smack of a naive, almost utopian faith in the power of individuals to overcome objective circumstances, forestry officials are seldom utopian and almost never naive. Rather, they integrate subjective human factors for effective leadership with objective factors such as local per capita income, economic diversification, availability of investment, and laws and directives from above. Virtually every policy piece written on the subject of forestry administrative reform includes references to administrative weaknesses; the prevalence of individual cadres who lack a basic understanding of the law, training, and commitment to conservation; and the need to inculcate both knowledge and consciousness in order to “overcome the interests in front of one's eyes in favor of long term interests, and the parochial interests of the department in favor of the interests of the whole” (Li Changjian Reference Li Changjian1998; Jing Reference Jing Jialiang1998). Most explicitly call for the “strengthening of leadership,” and pair the normative responsibility of cadres to “conscientiously study, unify thought, raise levels of knowledge, and thoroughly understand the spirit [of the given directive]” with the extension of formal organizational systems of responsibility, control, and auditing. In this view, effective implementation of new conservation policies is directly linked to the activism, spirit, commitment, and sense of responsibility of individual leading cadres in getting these new systems to work (CFY 1999–2000, 40, 42).

Rhetorical and empty as these formulae may sound, attempting to mobilize the normative commitments of individual cadres into the values and policies established above is at least as adaptive a strategy as pretending that the forestry sector is tightly controlled from above, with central and provincial levels of the state meting out harsh punishments to underresourced local officials who fall short of the mark. In practice, for as long as local cadres are entrusted with implementing state directives without correspondingly comprehensive financial commitments and auditing systems from above, they will retain substantial real discretion and must be “brought in” to higher-level state goals, programs, and norms through a complex process of formal control, normative manipulation, and subterranean deal cutting. Local officials can and do legitimately plead special local circumstances in target setting once concrete plans are disseminated from on high, and in times of policy uncertainty, they do their level best to negotiate the best possible deal for their localities. But this does not by definition mean that local officials are able to refuse to implement directives from higher levels of the government.

The change in the classification of forest land offers a good example of these dynamics. As is the case with most implementation of policy, the new forest classification system relies completely on county- and township-level forestry officials to make the actual decisions on which parcel of forest land falls where in the new classification categories. While in some ways, reclassification of forest land is a technical and not particularly politically sensitive administrative reform, the strict 70:30 ratio of ecological to commercial forest of this particular reclassification scheme has a set of implications for local officials: in areas with a large amount of commercial and economic forest, much productive land has to be taken out of cultivation entirely in order to meet these bureaucratic targets. Perhaps in advance recognition of likely resistance, almost immediately after setting the basic 70:30 target ratio in 1999, the provincial forestry bureau in Sichuan brought county-level officials and special surveyors to a province-wide training conference at Dujiangyan to educate the actual implementers of the policy about these new standards (Director S, Santai County Forestry Bureau).

The widespread publicity and conference-going certainly educated local implementing officials as to the new rules, but did not prevent loud squeals of pain and attempts to negotiate better deals from below. For example, in Le'an Township, Santai County, Sichuan, local cadres objected strenuously to the uniform implementation of the new rule that demanded a 70:30 ratio of ecological to commercial forest, in which the former was to remain completely untouched while the latter could be harvested. Le'an forestry officials clearly participated in the discourse and categories established by higher levels of the state in accepting the reasoning for the new policy and the importance of combating soil erosion through the reforestation of degraded hillsides. They even thought that “returning hilly areas to forest land presented no problems.” But at the same time, they were practically apoplectic with outrage over quotas that required them to extend this practice to forest stands on the township plain, where virtually all designated “forest” was “economic forest” interspersed with regular agricultural plots. In their representations to higher levels in the bureaucracy, Le'an officials employed a range of rhetorical tactics that were part of a broadly understood discourse of economic justice, objective special circumstances, and, where it suited, their own limited administrative capacity. First raised was the specter of widespread county poverty and the need to ensure social and economic fairness (thus arguing that they needed to retain more forest land for crop cultivation in the commercial forest stands on the plain). Second were the “objective” local conditions of unusually high population pressure and the way in which fragmented forest stands were interspersed with agricultural land (thus making the policy both socially unfair and difficult to administer). Finally, they warned that such an inflexible system could well establish incentives for cadres to designate the most prized, high-yielding old stands as “commercial forest” (Le'an Township Roundtable).

The Santai County Forestry Bureau to which Le'an Township reported clearly symphathized with Le'an's plight to some degree, characterizing the new classification system as “a real headache … while our natural resources section is currently managing this transition, we started off the process with only 35% [of our total] as ecological forest. But while we're permitted to raise the ratio of ecological to commercial forest, we're not allowed to lower the ratio” (Director S, Santai County Forestry Bureau). In turn, officials in the Mianyang Municipal Forestry Bureau to which Santai County reported were also at least implicitly sympathetic: when asked about the new forest classification scheme, they fairly snorted with derision over the targets that they had been handed from above as “inappropriate and unscientific” (bu heshi bu kexue) (Vice Bureau Chief Y and Senior Engineer X, Mianyang). What exactly eventually transpired in Le'an Township and Santai County remains unclear. At the time of these interviews, officials in Mianyang suggested that the reclassification process even in Santai was already “88% complete,” the Santai County forestry bureau chief thought that Santai would likely “come in a bit under target,” and officials in Le'an Township were still trying, without anything like the success that they sought, to argue that the scheme should not be applied to their agricultural areas at all. This narrative of policy making can be read as either as a classic case of bureaucratic bargaining, or a blunting of policy implementation by foot-dragging local officials. But equally, it can be understood as the working through of some degree of local flexibility beneath a larger rubric of standardization and central directives. However doubtful subprovincial administrative units may have been about the appropriateness of the new classification scheme for their areas, it is clear that the new classification was implemented at local levels, and often in a way that was not to the liking of local officials. Insofar as everyone shares the same surface commitments, accepts the same organizational hierarchies, is socialized into the same formal goals, and speaks the same normative language, informal negotiation below the surface can be worked through in a way that builds in some flexibility while maintaining the broad thrust of policy.

Variation by Program, Sector, and Locality

By the end of the twentieth century, China's forestry regime exhibited a great deal of variation in implementation in both programmatic and spatial terms. Some of forestry's subgoals have been implemented with much more consistent success than others, and these successes are often linked to some combination of positive bureaucratic incentives, ease of monitoring and defining success, and prior historical experience with campaigns. Thus, in aggregate, forestry administration has been much more effective with resource extraction than resource replenishment and much better at getting new seedlings into the ground than preserving old forests. Since the time of the Second National Survey on China's Forest Resources in 1980, forest cover has steadily grown from 12 percent (1980), to 14 percent (1993), 16.55 percent (1998), and, most recently, 18.21 percent (2003), with corresponding rises in standing forest volume. Interestingly, both state-owned and collective forestry regimes have contributed to this expansion, although collective forests have done somewhat better (Jiang Reference Jiang2007; Rozelle, Benziger, and Huang Reference Rozelle, Benziger, Huang, Hyde, Belcher and Xu2003, 118–23). However impressive this increase in afforestation may be, the forestry sector continues to do very poorly in either preserving old-growth forest or ensuring biodiversity, as virtually all of the new trees are monocultural stands or commercial tree crops (Rozelle, Benziger, and Huang Reference Rozelle, Benziger, Huang, Hyde, Belcher and Xu2003, 131).

There are good reasons for this. Until the late 1990s, there were very weak bureaucratic incentives to limit the cutting of trees. From the late 1990s on, increasingly strong financial incentives, when combined with the relative ease of monitoring from above and the historical legacy of tree planting campaigns, produced substantial results in getting viable trees into the ground. But these factors do not fully answer the question of why some areas seem to do so much better at implementing conservation policies that are devoid of the boost provided by strong bureaucratic incentives, ease of monitoring, extra resources, or particular historical legacies that lend themselves to effective implementation. And all of the available evidence suggests that some places in China do quite well with conservation and dampening logging and illegal land use, some do quite miserably, and others come out somewhere in the middle.

At present, neither Chinese nor foreign scholars have systematically investigated the empirical bases for this variation. Even highly knowledgeable forestry officials and experts in China do not possess the comprehensive knowledge to definitively answer the age-old question of why some places perform well and others poorly. One set of provincial-level forestry officials frankly admitted, “We have so many counties to keep track of that we can't thoroughly know what is going on in all of them; really conservation is mostly in the hands of local officials” (Official Y and Division Head L, Sichuan Forestry Bureau). But when asked to speculate about general tendencies associated with good or poor local forestry administration in the collective forest regions, most provincial- and municipal-level officials referred to a surprisingly consistent set of factors: (1) the degree of diversification and development of the local economy—officials frequently stated that “the more developed and diversified, the greater the reduction of pressure on natural resources” (and the greater the resource base and amount of fees that local authorities could collect without provoking strong resistance); (2) good transport links; (3) the relative degree of formalization (zhengguihua) and systematization (guifanhua) of local administration; and (4) the “quality” of local cadres (ganbu suzhi haohuai), reflected in their knowledge of central state forestry objectives and their degree of commitment in implementing those directives (Official Y and Division Head L, Sichuan Forestry Bureau; Official L, Chengdu Municipal Forestry Bureau; Senior Engineer X and Vice Bureau Chief Y, Mianyang Forestry Bureau; Officials P and Q, Nanning Municipal Forestry Bureau; Officials P and H, Guangxi Autonomous Region Forestry Bureau).

It is possible that this standard recitation of factors reveals nothing more or less than the degree to which local forestry officials have absorbed the dominant rhetoric of their superiors, and in any case, “systematization” and local cadre “quality” would be nearly impossible to measure in an objective manner even if data were available.Footnote ³ But two kinds of evidence support some of these claims. Regression analysis suggests that the two most important variables in determining changes in forest cover are population (the greater the population density, the worse for forest cover) and industrialization (the greater the proportion of light industry to the local economy, the better for forest cover) (Rozelle, Benziger, and Huang Reference Rozelle, Benziger, Huang, Hyde, Belcher and Xu2003, 127). What can be indirectly gleaned from statistical yearbooks and county gazetteers further suggests that the best guesses of forestry officials are at least plausible, if not yet confirmed. In Guangxi, where forestry officials and academics were the least reluctant to name well-performing counties and shame poorly performing ones, the frequently cited “good” areas of Guilin, Wuzhou, Yulin, and Rong County were uniformly in the economically better off east of Guangxi, with good transport to the metropoles of Guangzhou and Hong Kong, and the “bad” localities of Baise and Hechi were in the historically impoverished interior, with very poor communications and transport. Guilin is on a major north–south railway line and is in its own right a favorite destination for foreign and domestic tourists; Wuzhou a historic river port and entrepôt for Chinese medicine products with rapid links to the Pearl River Delta and Hong Kong; Yulin is served by a second-tier railway with links to the main north–south line, as well as having three major highways and the new cross province superhighway run through it (Zhongguo Guoqing Congshu 1993, 228–29). Equally, at the time of these interviews, the poorly performing inland areas of Baise and Hechi were without railroad transport or good road links, although the then unfinished cross-province high-speed highway was slated to reach Baise in the near future.

External indicators of local wealth and economic development only partially confirm the best guesses of the Guangxi forestry experts. In 2001, rural incomes in Guangxi varied from a high of 3,052 yuan in Wuzhou rural districts to a low of 970 yuan in Qinyun County. There was a significant gap in average rural incomes between the “good” and “bad” localities singled out by Guangxi provincial officials, with Wuzhou and Yulin averaging at 2,445 yuan and Baise and Hechi at 1,303 yuan. But neither were Baise and Hechi at rock bottom; slightly less than 20 percent (18 of 102) of the counties in Guangxi had even lower average rural incomes (Guangxi Nianjian 2002, 508–14). There are two possible reasons for this. Either the very poorest counties are so remote and out of the way that the forestry experts and cadres in the provincial capital simply do not have enough knowledge and contact with these areas to classify them one way or the other, or the commercial and transport links of these areas are so truncated that there is little economic incentive to cut the forest, other than for strictly local consumption.

Much more research with paired case study comparisons is needed to ascertain process and to determine how causality flows between these different factors. Is good transport an indicator of a diversified and increasingly wealthy local economy, or does it promote an increasingly diversified and wealthy economy? Does the existence of relative formalization and systematization create the appearance of local cadre effectiveness, or do good local cadres actually foster norms and expectations conducive to relative formalization and systematization? Is a diversified and developed economic resource base a prerequisite for effective administration and for good cadres to reach their full potential, or do good cadres see opportunities to promote a more diversified local economy? (e.g., through the promotion of tourism). Or do wealthier counties simply possess the educational and social resources to produce consistently “better” cadres? At present, there are no definitive answers to these questions.

The oft-repeated trope of “local cadre quality” is suggestive in two ways. First, the sheer number of interviewees and articles that integrate the human factor into administrative practice hint at something that goes beyond either cultural or structural factors, but integrates both. The ongoing structural gap between what China's administration claims to be able to do in principle and what it can effect in practice reinforces a particular tradition of a Mencian political philosophy that privileges the position of the good and moral man in state administration. But these deeply held norms and values about governance (“good administration by good men”) in turn make it possible for those gaps to be bridged by effective local leadership. Together, they result in a system in which the commitments of individual state administrators on the ground disproportionately continue to matter. Given the importance of individuals in getting the system to work and the way in which those individuals are embedded in local society, it is little wonder that higher levels of the Chinese state are so concerned with the question of “cadre quality,” and spend so much energy on attempting to ensure cadre quality control through its two-pronged strategy of formalizing administrative control and mobilizing local agent commitments and capabilities.

The term “cadre quality” (ganbu suzhi), while ambiguous, captures how the state attempts to fill in gaps in structural administrative capacity. While the term suzhi has come into very wide currency to reify hierarchies, to push individuals to be responsible for raising their own “quality” through self-improvement and education, and to justify authoritarian and statist categories (Kipnis Reference Kipnis2006; Murphy Reference Murphy2004), suzhi discourse is also relevant to state administration itself, particularly in within the local state. When local cadres discuss suzhi, they mean two subtly different things. In the first instance, “cadre quality” denotes the possession of a formal and quantifiable qualification (such as a technical college or university degree), which in turn presumes a high degree of technical knowledge and awareness of environmental policy. But suzhi also carries an active connotation of “effectiveness in policy implementation”—how the cadre in question functions in practice (does he or she get policies from above implemented?). Individual effectiveness in policy implementation, in turn, is seen to revolve around a variety of factors, including knowledge of the relevant laws and regulations, responsiveness to central dictates, personal commitment to implementing conservation initiatives, and, perhaps most important of all, how well the cadre in question “gets on with local people” (Officials P and H, Guangxi Forestry Bureau; Officials P and Q, Nanning Municipal Forestry Bureau; Senior Engineer Z, Sichuan Forest Investigation and Planning Institute). In short, “good” cadre suzhi is defined by a combination of book knowledge, people skills, and the capacity to link formal state policies from above to on-the-ground action.

In Mianyang, Sichuan, officials in a bureau with an unusually good reputation for forestry and conservation implementation echoed this position: “If the people curse and hate cadres, there is no way that anything can be done … as most of conservation is down to education of the population.” In Mianyang, cadres suggested that the bureau's “good working relationship with the masses” went back to a series of decisions taken in the early 1970s to concentrate on ensuring food subsistence, sufficient forest products for local use, and adequate water supplies. This, in turn, set in motion a long-standing pattern of openness (“if people see what is going on, this keeps everyone honest and aboveboard”), consultation, and discussion with local farmers in advance of drawing up plans for a range of different projects. This combination of attention to subsistence, openness, and consultation established a baseline of trust on all sides, so that Mianyang officials were able to begin conservation initiatives as early as the 1980s, well before most other areas. In 1999, Mianyang forestry officials were still heavily involved in promoting rural livelihood and income generation projects, from testing new forms of bamboo cultivation and marketing to diversification into medicinal woodland products. “So that this way we know when we can count on the masses for their share of the investment … and this means that we're also in a better position than most of Sichuan to implement [more stringent] conservation projects now” (Senior Engineer X and Vice Bureau Chief Y, Mianyang Municipal Forestry Bureau).

Given the still limited reach of the higher levels of the state, and its inability to fully (or even mostly) subsidize local governments even in as high priority an area as forestry and conservation, formalization, systemization, rulemaking, and monitoring can go only so far. The balance of success or failure still lies with local administrators, or cadres, and how “good or bad” the suzhi of the cadre in question happens to be. The trouble is that “goodness” or “badness” in terms of quality is conceived of in multiple, and often subtly conflicting, ways. As far as the higher reaches of the state are concerned, what makes for a cadre of good quality is a combination of formal qualification and responsiveness to orders, directives, and policies from above. Among his or her peer group, the cadre of “good quality” is one who listens, consults, and “gets on well with local people.” How local people truly view cadre “quality” will have to await future survey and case study work, but the case of Mianyang suggests that cadres who are open, consultative, and make clear efforts to guarantee subsistence and raise local incomes are likely to build up significant reservoirs of goodwill that can then later be drawn on in a variety of ways.

This suggests that the “good [local] cadre” has a very delicate balancing act to carry off. He or she needs to steer an effective course between the Scylla of state conservation policies and the Charybdis of ensuring local livelihoods—a portfolio that requires the wisdom of Socrates, the diplomatic skills of Talleyrand, the political alertness of a party activist, the technical competence of an advanced degree holder, and the creative accounting skills of a CPA. At a minimum, the effective local cadre needs to be able to persuade and negotiate well with higher levels of the state and with local society. This, in turn, involves exercising an authority that is as much moral as it is implicitly coercive. Despite efforts to further formalize and rationalize formal administration in contemporary China, structures alone cannot account for success or failure in a state in which so much of what happens on the ground takes place in a shadowy area that is neither formal state nor autonomous society, but somewhere in between. This mirrors findings in the wider community forestry literature suggesting that conservation projects that enlist the support of legitimating preexisting semiformal bodies and norms will be much more effective than those that do not (Chi Reference Chi1999, 60–69).