Introduction

Grassland is a dominant landscape in China, accounting for 40 % of the nation’s area. Although there is debate over the exact amounts of grassland degradation, scientific research on diverse pastoralist communities (Xinjiang, Qinghai, Gansu, Tibet, Ningxia, Sichuan and Inner Mongolia) is unanimous regarding the presence of massive grassland degradation, albeit of varied degree, depending on region (Ho 2000; Harris 2010; Wang et al. 2010; Li et al. 2007; Taylor 2006). Official estimates regarding the proportion of Chinese grasslands with some degree of degradation are 90 %, with degradation reported to be increasing at a rate of 200 km2/year (State Council 2002). Accelerating grassland degradation is widely perceived as a threat to the sustainability of Chinese grassland use and to poverty alleviation initiatives (Banks 2001; Banks and Richard 2003; Wang et al. 2010), as grassland desertification and degradation is occurring mainly in arid, poverty-stricken parts of northwest China. In these regions pastoralism is widespread and local people are highly dependent on natural resources for their livelihood. Several technical measures, such as aerial sowing and grass planting, were employed by the central government in the 1990s to address grassland degradation but their impact was limited.

One issue that has been widely debated is the unclear role of land tenure systems in fostering degradation. During the Mao period (1950s–1970s), agricultural operations were organized in these regions under a production-team system, with each team consisting of 20–30 neighboring households (Lin 1992). After the demise of the communes system, commune based tenure persisted until the de-collectivization reform (also known as Household Responsibility System reform) in the early 1980s. With the Rangeland Law of 1985 (later slightly amended in 2002), Chinese pastoralist land-use tenure policy aimed to generate a shift toward individual household tenure, which was presumed necessary for improving incentives for sustainable rangeland management (Banks 2001). Pastoralists signed long-term leases for grassland use with administrative communities, while ownership remained “collective”. By the end of 2003, about 70 % of usable rangeland in the country was leased through long-term contracts to households or groups of households (Jian et al. 2011).

It is widely presumed by Chinese policymakers that these individual household husbandry reforms, where livestock was transferred to private ownership, combined with use of commonly held grasslands, has triggered widespread overgrazing, due to a “tragedy of the commons” (Yamaguchi 2011; Ho 2000; Banks 2001; Yan et al. 2005; Taylor 2006). Grassland privatization has thus been regarded as an essential strategy for solving the degradation problem (Wang et al. 2010). It is expected, in keeping with conventional economic theory, that it is possible to eliminate pastureland degradation by assigning its costs directly to those who are using the grassland, thereby encouraging more responsible behavior. Sustainable use of grasslands would then be rewarded and overgrazing would be punished through market mechanisms because environmental costs now appear in the herders’ own production cost calculations. However, we propose that it is incorrect to presume a simple “tragedy of commons” in the case of Chinese pastoralism.

The widely cited concept of “commons” in Hardin’s (1968) influential theory is actually “open access” instead of “common property” (Bromley and Cernea 1989; Ostrom 1990; Taylor 2006; Ho 2000; Williams 2002), which leads to confusion. However, instead of an absence of institutions, which would be an ‘open access’ situation, current pastoral land use management in China includes clear access restrictions: use-rights have long been assigned to administrative villages and use itself is usually coordinated through a combination of informal institutions at the local level. In addition, the assumption that natural resources cannot be used sustainably by a collective has lost support in recent years, with ample evidence indicating that local communities are capable of jointly using natural resources on a sustainable basis (Agrawal and Lemos 2007; Agrawal and Gibson 1999; Banks 2003; Ho 2000; Jones and Craswell 2004; Li et al. 2007; Li and Yu 2011; Ostrom 1990, 1999; Ostrom et al. 1999; Runge 1986; Tyler 2006; Wade 1987). Local decision makers are expected to have lower transaction costs because of lower overhead and operating costs, as compared to central decision makers (Agrawal and Lemos 2007); they can contribute to crafting better-adapted rules for governance of their particular local setting, with the unique time- and place-specific information they possess (Krister and Ostrom 2008; Agrawal and Lemos 2007; Waudby et al. 2012; Gadgil et al. 1993); and local residents, who share a collective interest in sustainable management, can solve internal free riding problems among themselves without having to assign private property rights (Ostrom 1990).

We propose that exclusive focus on the role of property rights in achieving sustainable management of China’s rangelands ignores the wider economic and political institutional contexts within which herders and their related resources (grassland and livestock) are embedded. Existing studies show clearly that privatization does not mitigate pastureland degradation (Banks 2001; Williams 1996, 2002; Taylor 2006; Li et al. 2007). Following the implementation of privatization oriented grassland reform, degradation has continued throughout China (Jian et al. 2011). Experiences from pastoral areas of other countries (Sylla 1995; Fernandez-Gimenez 2002) suggest that practices aimed at promoting rigidly defended property boundaries in grasslands are inclined to fail. We propose that the degradation problem in China’s grasslands does not necessarily result from a lack of secure property rights but is rather a consequence of policy that fails to take into account the particular practices and needs of the affected local communities.

In this paper we explore changes in herders’ grazing behaviors in Yanchi County, in the Ningxia Hui Autonomous Region of China, taking into account the broader context of the institutional arrangements within which the privatization of pastoralist land-use tenure has been implemented. We ask: is it the case that de-collectivization reform has ruined the practice of cooperative herding, leading to a tragedy of the commons? If not, then what is the main reason for the increasingly individualized use of pastureland that we observed in the study region? This paper does not attempt to evaluate the impact of the private property regime on the ecological environment but instead focuses on the impact of the wider institutional context (i.e., grazing policies and their implementation) on both individual and collective herding behaviors. Our overall aim is to contribute toward the work of developing a better understanding of the drivers of land degradation in the regions of China where pastoralism is practiced.

Research Background

Research Site

Yanchi County is located in the east of Ningxia Hui Autonomous Region (Fig. 1), in the transition zone between the Loess Plateau and Ordos Desert, where semi-arid grassland eventually becomes the arid steppe of northwestern China. It has a total area of 10,695,000 muFootnote 1 (713,000 ha), of which 8,355,000 mu (557,000 ha) is grassland. During the last three decades grassland degradation has become increasingly severe in this region, with 79 % of the land suffering serious desertification as of 2002. Yanchi County is one of China’s most impoverished counties with comparatively little arable land, most of which is alkaline soil, making farming difficult. Livestock grazing is the dominant form of local land use and the main agricultural activity, representing 31 % of the per capita net farmer incomeFootnote 2 in 2009. Due to the importance of pastureland for the regional economy and the increasing desertification, Yanchi has repeatedly been chosen by the government of Ningxia Hui Autonomous Region as a pilot County for experimenting with supports for the sustainable use of pastureland.

Fig. 1
figure 1

Map of Yanchi County in Ningxia Hui Autonomous Region, China

Full size image

Grassland Related Policies

Over the past 20 years the regulation of grassland use in Yanchi County has passed through three distinct phases: the privatization of pastoral land use rights, which began there in the early 1990s; the Grazing Ban, which was adopted in 2002; and the Open Grazing pilot project, which allowed restricted use of pastoral lands in selected parts of Yanchi County between 2006 and 2007. In this section we provide a brief overview of each phase, to give the reader a general picture of the dynamics of institutional change in the study area, before then proceeding to present our research methods and empirical results.

  1. 1)

    Land-use Privatization

    Responding to local environmental problems, grassland decentralization reform was implemented in the 1990s in Yanchi County. The aim was to lease grassland use rights to individual households, while maintaining community ownership. The degree of grassland allocation varied (Ho 2000; Banks and Richard 2003; Nelson 2006): some grassland areas were contracted to individual households, as encouraged by the policy; in some areas with quite limited grassland, land was assigned to groups of several households; while in still others the status quo situation was maintained (i.e., land use rights continued to be owned and managed by the community). As of 2003 use rights for 4,755,000mu (317,000 ha) of grassland, about 86.6 % of Yanchi County’s total grasslands, had been leased to households or groups of households, with 840,000 mu (56,000 ha) allocated to individual households, and 3,915,000 mu (261,000 ha) contracted to groups of households.

  2. 2)

    Grazing Ban

    Based on the presumption that only complete elimination of livestock could restore land productivity, a Grazing Ban policy was adopted in Ningxai Hui Autonomous Region in November 2002, with Yanchi County serving as the pilot case for its implementation. The policy covered the entire County, impacting all 98 administrative villages,Footnote 3 comprised of 675 natural villages associated with the County’s eight towns. All activities related to grassland use were banned, with a special emphasis on the prohibition of animal husbandry. Rearing sheep in sheds, using fodder, was strongly recommended. The policy was subsequently extended to other parts of the Ningxia Hui Autonomous Region. After the first few years of the Grazing Ban policy there was some recovery of vegetation (Table 1). Locals who we interviewed observed an improvement in the ecological environment during this period, which they understood to be a result of the implementation of the Grazing Ban. On a scale from 0 to 10 (0 = worst; 10 = best), our interviewees’ average assessment of the pastureland situation in 2010 was 7.4, well above the average assessment of 4.3 given for 2000. They also observed a return of desertification in recent years. It is likely that persistent drought is a key factor contributing to renewed desertification and the herders we interviewed do not dispute this. However, they also believe that grassland quality has been deteriorating in recent years because grass that is not being eaten by livestock dries out and covers the ground, making it more difficult for new grass to come up in the next spring, especially after this dry grass has accumulated over several years. They are also concerned that there is loss of biodiversity in the rangeland, which has been observed by Fang et al. (2012), and is an observation consistent with the results of other studies concerning the impacts of animal husbandry on grassland quality (Walker et al. 1981; de Knegt et al. 2008): namely that the presence (or absence) of herbivores has significant impacts not only on the volume but also on the composition of biomass in rangelands.

    Table 1 Impacts of grazing ban on grassland indicators in Yanchi County (Maximum values appear in bold)
    Full size table

    In recent years conflicts between the Yanchi County government and local people have intensified in Grazing Ban areas, as the policy has forced these people to abandon their most important traditional means for securing livelihood. The policy has also accelerated rural out-migration, which is found throughout China and is considered to be an important factor contributing to the nation’s present rural development problems. In the case of Yanchi County this is an important factor influencing the strategies and practices of the actors in the study region.

  3. 3)

    Open Grazing Pilot Project

    Taking into account this complex set of problems, in November 2006 ten natural villages associated with one administrative village in Yanchi County were selected by the County government as pilot areas for an Open Grazing program, where the Grazing Ban was temporarily lifted. The pilot area initially covered roughly 15,000mu of grassland and was expanded in the following year to cover 15 administrative villages associated with one town, comprising 96 natural villages and covering a total of 823,600mu of grassland. Farmers in these pilot villages were permitted to use grasslands from the period May to November. A grazing quota was allocated to each natural village, based on the quality of their grassland, which was then distributed equally, by the village committee, among all registered persons of the village. Here it is important to note that villagers who had migrated to cities were included in this population count and were able to transfer their rights to other villagers. Grazing licenses, which specified information such as the grazing quota of the household, family size and household name were to be carried by the herder during livestock grazing, with failure to present the appropriate license resulting in a fine of 10 RMB (about 1.60 dollar) per sheep. This pilot project was suspended in 2008.

Research Methods

The data reported here were collected in July and August, 2010 in various communities in Yanchi County. The responses of villagers in the study region, first to land-use privatization, then to the Grazing Ban, and, where applicable, to the pilot Open Grazing project, constitute the empirical focus of our research. A combination of qualitative and quantitative data was collected. Taking into account the diversity of attributes influencing grazing activities, 12 representative natural villages were selected for study. Interviews were conducted with numerous local residents from each village and with one local Non-governmental Organization (NGO). Selection criteria for the villages included total grassland area, style of fencing, type(s) of grazing policy in the village, village location and characteristics of the villages’ arable land (see Table 2). Style of fencing was used as an indicator for de facto property rights, with three basic fencing styles being observed: for the village; for groups or for individual households. Four of the villages we studied were under the pilot Open Grazing policy in 2007; eight were non-pilot villages, which had been subject to the Grazing Ban since 2002. Location and land characteristics of villages were recorded based on whether or not the territory was beside a main road or an administrative quarter of the County and on whether the terrain was hilly or plains. Both factors indicate the remoteness of a village, as the hilly areas of this region are usually remote and hard to reach from the administrative quarters.

Table 2 Land-use attributes of case villages
Full size table

Participatory rural appraisal was employed in each village in order to develop a general impression of the setting, including geographical information, population, history and local modes of organization. Village-level data was collected through key-informant interviews and household-level data was collected through semi-structured, questionnaire based interviews with local herders. All interviews were conducted in Mandarin without translation and on the basis that reported data would be kept anonymous. At the household level, random selection was used to choose interviewees, where possible, with the target sample size being ten respondent households per village. However, due to out-migration, in some villages it was difficult to find ten households raising sheep. Data collection targets were therefore adjusted, according to the available population of sheep herders in each village. In the end, 111 questionnaires were collected, with 42 drawn from the four pilot villages and 69 drawn from the eight non-pilot villages. Additional information about village populations and husbandry intensities and the typical types of land management in the study area, is presented in Tables 3 and 4, respectively.

Table 3 Demographic and husbandry data for case villages
Full size table
Table 4 Land usage per household by cultivation type (111 observations)
Full size table

Results

In order to evaluate the collected data, we used a combination of institutional and discourse analysis, with reference to interview data and supporting literature concerning the legal and environmental situation in the region. Our two main observations are: 1. herder response to the Grazing Ban is more individualistic than was their response to land-use rights privatization. This individualization, typified by the practice of illegal night grazing, appears to have been accompanied by negative social and environmental impacts, including the breakdown of what we may, for lack of a better name, call the ‘traditional common property resource management systems’ of these villages. Here we are referring to informal and formal collaborative grassland use practices that have been present since the privatization of livestock in the early 1980s. Since it was not possible for us to determine if, or if so to what degree, these local institutions pre-dated the Mao-period, we take the 1980s situation as a pragmatic baseline for what today constitutes traditional practice among these communities. We presume that this reflects a mixture of pre-Mao-period and Mao-period influences. 2. That herders’ responses to the pilot project seem to reflect an increased sense of ownership and curatorship of the grasslands and an increased ability to employ and to adjust their traditional common property resource management systems.

Frequent Occurrence of Illegal (Night) Grazing

In the first few years after the grazing ban, when monitoring was intensive and rigorous, illegal grazing rarely existed. The official sanction criterion was 10–20 RMB per sheep; with variations depending on herding time, size of flock and the age of the herders involved. However, the cost of rearing sheep in sheds was much higher than for grazing sheep, with average costs per sheep increased by around 100 RMB (roughly 16 dollars) mainly due to the need to purchase fodder from central stores (Qi and Hu 2006). Because of high monitoring costs and frequent conflicts between herders and the government monitoring committee, enforcement of the Grazing Ban has tended to be much looser in recent years, resulting in more frequent illegal grazing. The “rush hour” for such illegal grazing is the middle of the night: from 11 p.m. to 4 a.m. Herders are familiar with local geography, while the government monitoring committee usually is not, so it is relatively easy for herders to avoid monitoring at night. More than 54 % of all interviewees admitted that they had been grazing almost every day; 22 % grazed sheep frequently; and only 15 % gazed seldom or never. Among all the interviewees, half were fined less than two times, 30 % between two to five times, and 3 % more than 10 times (see Table 5).

Table 5 Interviewee responses: average number of sanctioned for illegal grazing in 2010
Full size table

The implementation of the Grazing Ban varied among villages, in terms of both monitoring and sanctions. In remote areas enforcement was relatively lax. As an old Chinese saying puts it: “The mountains are high and the emperor is far away” (Shan Gao Huang Di Yuan). In contrast, implementation of the policy in villages adjacent to administrative headquarters or near to highways was comparatively thorough and these areas were heavily targeted for policy fulfillment, as monitoring is much easier and less costly for the city and provincial governments. After the implementation of the Grazing Ban policy, most of the villagers living near a highway or close to a county administrative quarter moved to cities or changed over to rearing sheep in sheds.

Disappearance of Cooperative Herding

Along with the adoption of night grazing practices, we also observed a shift away from cooperative herding practices in response to the Grazing Ban. In the study areas cooperative herding existed widely in the 1980s and 1990s. Households realized economies of scale with respect to herd supervision through group herding arrangements or by cooperatively contracting commercial herders. Interviewees indicated to us that there had been no significant change in the household herding strategies of their communities following the end of commune system. Cooperative grazing in Yanchi County usually involves a small group of several households, bound mostly by close kinship ties, pooling their sheep into a single herd and taking care of sheep herding in rotation, one household after the other. When the number of sheep per household varies dramatically within a herding group, herding tasks are assigned to each household based on the number of sheep they have. By herding cooperatively individual workload is decreased substantially, freeing up some of the labor force for additional work. However, the Grazing Ban technically made such cooperation illegal and practices were seen to change after its implementation, with substantial decreases in both the collective contracting of commercial herders and in group grazing. This can be seen by looking at the household grazing patterns in the pilot villages during different periods (see Table 6). Statistics based on our research results from the 42 interviews conducted in the four pilot villages show that before the Grazing Ban households realized economies of scale with respect to herd supervision through reaching group herding arrangements (31 %) or contracting cooperatively with commercial herders (17 %). Following a sharp decrease in group herding arrangements (from 48 to 7 %) during the Grazing Ban, this then increased again, to 21 % during the pilot Open Grazing period, falling again to 5 %, after the Open Grazing policy was suspended.

Table 6 Pilot Village Households: herding types during different periods
Full size table

Contrary to what was expected, the privatization reform did not lead to a boom in the construction of fences. Instead it would seem that the tradition of common use of grasslands persisted as an informal institution, with several of the herders we interviewed continuing to adhere to the belief that private use of grazing land was “hard to accept”. However, the Grazing Ban appears to have finally undermined this tradition of cooperation by encouraging individual herding and intensifying individualized use of grassland. In recent years households have increasingly built individual pastureland fences, which are now formalizing the privatization that was expected to come with the grassland reforms of the early 1990s. This increased fence construction sheds light on how the Grazing Ban and its implementation have influenced grassland use.

In Yanchi County village fences were built at the beginning of the grassland reform period in almost all villages, with financial support from the County government, whereas construction of group fences was voluntary and required the approval of all group members. In the first few years after implementation of the Grazing Ban, when monitoring was regular and strict, the economic incentives for further fence construction were low. Household pastureland fences have, however, been increasingly built in recent years, due to the popularizing of night grazing. For safety reasons, the pastureland close to residential areas is usually favored by herders during night grazing. It is also easier to escape governmental monitoring when using this pastureland, because sheep can be herded back quickly, if and when the monitoring committee is observed. However, this land is also readily accessible to others in the community and it is difficult to exclude them from using one’s land when it is dark. Fences have therefore been built to prevent potential free-riders from using ones individual lands. In some villages sheep are brought to fenced individual grasslands in the evening, where they can be left during the night without supervision. Farmers then return home and herd their sheep back in the early morning. Interviewees explained that, compared with night grazing on village grasslands - where herders supervise their sheep throughout the whole night - this grazing method greatly reduces working hours and workload, allowing them to take other jobs during the day.

Decreased Self-Organizing Activities

In Yanchi County local herders have shown ability to address emerging and changing land use problems with locally generated solutions. In some villages, before implementation of the Grazing Ban, informal rotational grazing systems were formed through group discussion in village committee meetings, or in daily chatting, and implemented as oral agreements. In contrast to some other pastoralist regions in China, (e.g., Inner Mongolia and Qinhai-Tibet), rotational grazing is not common in Yanchi County and there is no current practice of seasonal migration. However, rotational grazing emerged in some areas when local land use came under pressure from accelerating land degradation, as herders gained an increased awareness of the importance of their ecological environment. For example, in 2000 in village 9 (Table 2), where pastureland fences were constructed for groups of households, herders formed a rotational grazing system by contracting the use of two grasslands located far away. According to self-designed rules, they decided that the grasslands near the village would be available for use exclusively in winter and those far away from the village would be open for grazing in summer. These rules were, according to a discussion held with one of these groups, widely accepted within the groups and can be understood to have been successfully established. Although there was no written agreement or specific punishment mechanism, only a few violations occurred, and these were during a time of extremely bad weather. Here formal sanction and official monitoring had been substituted by reliance upon mutual trust, built on long-term relationships and guaranteed by the prestige of the village leader. Locals respected these rules because, as they put it, “Violation of community rules might result in moral criticism and hatred from community members. The risk of being disapproved of by community members is the last thing people would like to face.” However, this self-formed rotational grazing system ended after the implementation of the Grazing Ban in 2002.

We also found innovative strategies in pilot area villages, organized by local resource users, such as the arrangement of community-based grassland leases that pooled together allocated use rights. This illustrates the ability of locals to adjust institutional settings, based on local social and ecological circumstance, in order to manage their common resource use.

Increasing out-migration has left a great amount of grassland unused. Grassland leasing bridges a gap between the amount of grassland actors wish to use and the amount of grassland for which they actually own use rights. Grazing quotas, established for the pilot scheme villages, provided a well-defined arena for grassland leasing by getting around the problem of unclear household pastureland boundaries. In some pilot villages grassland leasing was regulated by community-designed and implemented rules. For example, village 11 (Table 2) had, during the pilot period, 76 households with approximately 300 people. Among these, only 24 households were permanently living in the village (i.e., had not migrated to cities to work) and their livelihood was highly dependent on sheep grazing. In 2006, when the pilot Open Grazing policy was in place, based on scientific tests of grassland quality conducted by the local government, 20mu of grassland could support one sheep in this village. The village has 33,000mu of grassland, giving an assigned grazing quota of 6 sheep per registered person. Considering that approximately two thirds of villagers were out of the village nearly all year round, locals decided to adopt a grazing quota lease arrangement. Grassland quota leasing among community members was encouraged, with the commonly accepted price being 10 RMB per sheep per year. In 2007, about one third of village 11 households were involved in grassland leasing (i.e., either selling or buying). This kind of lease existed only in pilot areas, where grazing was legalized and a grazing quota was assigned by the County government via a literal grazing license. After suspension of the pilot project, grassland leasing dropped substantially. In 2010, no grassland leasing took place in any of the pilot villages studied.

Generally, the legal grazing quota per-person is not enough to support household livelihood. Based on empirical measurements of grasslands in Yanchi County, it is possible to herd one sheep for every 20mu of pastureland in 90 % of villages. In the remaining 10 % of villages quotas varied from one sheep for every 10mu up to one every 30mu. Since the average household grassland area in Yanchi County is 170mu, average household grazing quotas are quite limited, coming in at around 8 sheep per household. However, during the survey, 78 % of interviewees in village 11 expressed support for the quota system.

We can see from this case that both the Grazing Ban and its suspension have influenced the way locals are dealing with environmental degradation, producing not only conflict (e.g., between resource users and government), but also coordination (e.g., among resource users and in balancing the requirements of eco-– and human systems).

Discussion

Small scale herders in pastoralist societies characteristically rely on cooperative labor. Group grazing arrangements, whether they employ the collective contracting of commercial herders or coordinated grazing as a group, imply joint, co-operative use of grassland and therefore also the presence of institutions. This widely observed institutional regime of local community property offers actors an arena for cooperation (Hagedorn et al. 2002; Hagedorn 2008). Where resource utilization patterns place considerable demands on individual labor, local community property can be managed with greater efficiency. This can be understood to reflect a strong demand for the positive effects of labor investments in pastoralist production (Næss 2010, 2012). In the villages we studied, the efficiencies provided through cooperation were especially important, due to the limited grassland area allocated to each household and the small numbers of livestock (approximately 50 sheep per household). Where flocks are small, it is not difficult for one herder to take charge of the sheep of several households, whereas initial herding costs are roughly the same. Self-organized cooperation among pastoralists, such as the leasing we observed in pilot villages, allows for more flexible policy implementation and locally appropriate adjustments. It also enhances the capability of locals to jointly cope with risks (e.g., of increasing climate change related impacts) through mutual assistance and can help reduce production costs, by allowing for the use of pooled labor and contracted commercial herding that matches labor effort to herd size. Perhaps most importantly in rural China, such cooperation increases the grassland utilization rate per person, helping to compensate for the problem of a declining rural labor force. It also contributes toward a fair and efficient redistribution of grassland use rights, as herders still in the villages can access the grassland resources of out-migrants, which might otherwise be either under or over utilized in the absence of a management regime. In addition, cooperation facilitates learning and information sharing among resource users.

Some observers have argued that privatization of China’s grasslands use can be expected to prevent effective cooperation among herders (Banks 2001; Jian et al. 2011; Ho 2000). We believe it is too early to draw that conclusion. In Yanchi County the impacts of privatization reform in pastoral areas are hard to determine because of inconsistencies between the property regime in use (de facto) and in law (de jure). The special characteristics of grassland use, e.g., labor time demands, as well as the heterogeneity of pastoralist communities, with respect to available grassland area, location and modes of cultivation, mean there are a wide diversity of de facto property regimes. For example, in a previous study of pastureland use in China, some villages were found to have established partial contract systems based on rough areas assigned to individual households (Longworth and Williamson 1993). In the case of Yanchi County there are at least three land tenure systems working simultaneously: a quasi-private property regime, with individually owned land use rights linked to specific pieces of pastureland; a common property regime where groups of households collectively own use rights and sometimes contract the use of additional plots; and a common property regime where use rights are assigned collectively to all members of the community. The general property regime in the surveyed area can therefore not be defined as simply private or common. This diversity of de facto property rights regimes developed by the local communities in Yanchi County, including the practice of trading herding quotas, demonstrates the limited impact of the official privatized property rights regime.

In considering these results, we find that the institutional context of the Grazing Ban policy and its implementation, rather than that of privatization reform, seems to have been the main driver leading to increased individualization of grassland use and to increased short-term considerations of profit maximization among herders. We found that it has not been privatization reform but rather the Grazing Ban policy that has contributed most significantly to the increased construction of household pastureland fences, which is inconsistent with traditional forms of land use cooperation in Yanchi County. However, increased night grazing in response to the Grazing Ban, has rendered traditional land sharing rules-in-use invalid. Risks of sanction during illegal grazing seems to have reduced the willingness of herders to engage in cooperative herding practices. As McCay and Jentoft (1998) have pointed out, while resource decline may be a consequence of the imperfect allocation of property rights, it can also be caused by a collapse of social bonds, which they believe leads to community failure, e.g. conflict among competing groups or opportunism by privileged elites. The implementation of the Grazing Ban in Yanchi County appears to have, on the one hand, undermined the initiative to encourage responsible husbandry through privatization, and on the other, to have resulted in herders making a break with cooperative tradition. Rather than attributing use driven land degradation in Yanchi County to a ‘tragedy of the commons’, we find it more plausible that economic constraints and institutional pressures originating from the Grazing Ban have destroyed a mutual trust that had lasted for generations among these resource users, undermining their ability to collaborate proactively in the sustainable management of the grasslands upon which their livelihoods depend.

Moreover, the Grazing Ban has exacerbated out-migration, with the majority of residents in Yanchi County now made up of the elderly, women and children. Sheep herding is usually the last remaining economic option for these ‘left-behind’ people, especially for those who are not receiving subsidies from the government or remittances from children. Local pastoralist in the villages we studied acknowledge that something has to be done to solve the ongoing severe desertification in the region, otherwise, as one interviewee put it: “the thing we leave to the younger generation is desert instead of the pastureland that we inherited from the older generation”. They have also argued that, although grazing quotas are not sufficient for maintaining a household, increasing out-migration makes it possible to maintain a household with grazing, by using voluntary grassland leasing systems organized by the herders, such as in village 11.

The popularity of night grazing highlights the need for more attention to be given to how both herders and the lands they use are responding to both private property and policy factors. The proposition that properly implemented land-use privatization would eliminate a ‘tragedy of the commons’ in Yanchi County is not supported by the results of our research. Instead we find that a combination of pressures, including the Grazing Ban policy, depopulation of the area, and the demographics of the ‘left-behind’ community, have led to pronounced changes in herding behavior, which have had impacts on both herders and grazing lands ecosystems. The possibility of cooperative herding is substantially decreased under the Grazing Ban. However, grazing continues, as ‘left-behind’ people see few other options. The Grazing Ban also makes contracting a commercial herder complicated and considerably more expensive than in the past. First, since the herding action itself has become illegal, but also because night grazing has a profoundly destructive impact on a herder’s life, so that demand for contract herders is increasing at the same time that interest in offering this service is decreasing: unsocial working hours impair both health and community life; differences in temperature between day and night are large and it is quite cold during the night, even in summer; constant attention looking for official monitoring teams makes herders hyper sensitive to any movements and fearfulness and anxiety negatively influence both physical and mental health. Under these conditions, the remaining local population is increasingly less willingness and less able to contract herding services, whether individually or collectively, and the tendency toward individualized grazing is further reinforced.

Pastoralist systems have been found to be capable of developing flexible institutions and practices that make persistent success in rangelands possible over the long term (Scoones 1994; Li and Yu 2011; Butt 2011; Fernandez-Gimenez 2002; Li et al. 2007). In our field research we saw that flexible responses to changing conditions (e.g., self-organizing grazing, night grazing) arose in Yanchi County both before and during the Grazing Ban period, and during the pilot period of Open Grazing. Oral agreements were applied in most cases and most self-organized activities were coordinated between households within villages, with kinship being a key factor. However, these self-organizing activities proved quite vulnerable to pressures from formal institutional arrangements and largely disappeared following reintroduction of the Grazing Ban.

Local people are currently excluded from the grazing land management decision-making process in China and this lack of communication between policymakers and local people can be seen to make top-down policies difficult to implement, as illustrated by the practice of illegal night grazing. This can also lead to confrontations and conflict between the government and local people, and in the case of the Grazing Ban in Yanchi County, to grassland users focusing more on short-term considerations and less on long-term sustainable use of resources. Bromley (2008) argues that a coherent institutional regime is one that serves to secure expectations, so that forward-looking economic behavior is facilitated. If the institutional conditions that parameterize village life are of low quality, meaning that the resulting signal is perverse, it is difficult for individuals to form plausible expectations concerning investments and future resource use. Where this is the case, forward-looking actions with respect to the management of individual assets will be undermined (ibid) and privatization will not lead to better long term management of the resource in question.

Privatization does not occur in a vacuum. Far from being a panacea, land-use privatization has been proven insufficient for ensuring the sustainable use of pasturelands and at times can even prove counter-productive (Dong et al. 2007; Fernandez-Gimenez 2002, Hobbs et al. 2008, Lesorogol 2008; Sylla 1995; Ojima and Chuluun 2008).

No property regime fits for all situations: “What actions are physically possible, what outcomes can be produced… are affected by the world being acted upon in a situation” (Ostrom 2005). Therefore, it is essential to examine the physical properties of land-use related economic transactions (buying or selling land and land-use rights and directly using resources) on a case-by-case basis. The physical world (and the related physical properties of transactions) is as important for institutional analysis as the social world (and the related physical characteristics of actors’ behaviors) (Hagedorn 2008). In the case of Yanchi County, it is clear that unsustainable grazing patterns have complex and interlinked social, economic and ecological causes and so we find it generally implausible that there is a single solution to the County’s problem of anthropogenic land degradation. Based on our study of the responses of herders to initial privatization reforms, the Grazing Ban policy and the pilot Open Gazing project, we find the proposition that overgrazing is related exclusively to privatization is too narrow in its scope.

Conclusions and Recommendations

Our analysis strongly suggests that the effort to implement privatization of pastureland is only one among several factors changing herding behavior in Yanchi County. De-collectivization reform of pasturelands has been mediated by long standing informal herding tradition, the physical attributes of the targetted pastoral areas and demographic change in these regions. Futhermore, the Grazing Ban policy appears to have intensified individualized grazing behaviors.

Under the tremendous pressure of accelerated and large-scale land degradation, the focus of the government turned from one extreme (i.e., concentrating purely on economic growth) to the other (i.e., focusing exclusively on ecological protection through a Grazing Ban), failing in the processes, to identify ways to balance the demands and pressures of economy, society and environmental systems.

The experiences of the pastoralists we studied in Yanchi County illustrate that they are both aware of and capable of developing effective strategies for managing the impacts of their grazing activities on the grasslands. We find that the increasingly individualized, short-term oriented grazing behavior of these pastoralists is related to the institutional setting within which they are operating. In the face of a prohibition on grazing, these people nonetheless have to attain the basic materials required for securing their livelihood, which is leading to two negative impacts: policy implementation is failing and sustianable use of the grassland is not being achieved. If the key element that determines the success or failure of an institution is the extent to which it fosters coordinated expectations in relation to a particular physical and social environment (Runge 1992; Bromley 2008), then the exiting pasturelands policy in Yanchi County, with the exception of the Open Grazing pilot program, has hardly been a success.

We propose that the situation in Yanchi County could be improved if policy makers acknowledge that institutional diversity may be as important as biological diversity for ensuring the long-term survival of these communities (Ostrom et al. 1999). A more appropriate long-term institutional arrangement would, we suggest, give local pastoralists the opportunity to co-ordinate grassland use from the bottom-up, based on locally-specified strategies and criteria. At the very least, it seems critically important that communication between “the top” and “the bottom” is improved, so that the interests of local pastoralists are taken into account when policy is being formulated. These people, who depend directly upon natural resource for their livelihoods, will use the lands as best they can, whatever the policy, and in full knowledge that their use patterns are harming both them and their environment, not because they do not know better but because they have no other realistic alternatives.

Based on the above discussion, we conclude that the special characteristics of pastoral lands, with their fuzzy boundaries, highly diverse geographical and social situations, sparse human populations, and climate variability imply that there is limited functionality to be expected from a rigid and uni-dimensional land-use management approach. Instead, the situation in Yanchi County calls for more flexible patterns of resource use management and institutional arrangements that support flexible cooperation. This would allow herders to experiment with various types of innovation, as we saw happening during the pilot Open Grazing period. Replacing single-objective oriented regulations, such as the Grazing Ban or the land use privatization reform, with county and national institutions that facilitate, or at the very least enable locally developed land use institutions to emerge (e.g., co-operative herding, self-organized management, quota leasing), may help not only to achieve sustainable use of Yanchi’s grasslands but also to improve the quality of life of those who have been ‘left-behind’ to preserve them.