Modelling and Local Agency in XNCJS Implementation

XNCJS Piloting and Modelling in Practice

XNCJS was found to be of considerable importance in all three localities. The organizational setting of XNCJS differed in each county. Although they had all established a XNCJS leadership small group (xinnongcun jianshe lingdao xiaozu 新农村建设领导小组), the operational core of XNCJS implementation below this ad hoc body varied considerably, from an omnipotent XNCJS bureau under the department of rural work (nonggongbu 农工部), a Party organ in Dingnan, to a much more loosely organized mode of departmental cooperation in Qingyuan.Footnote ¹⁰

Terminology, Concepts and Strategies of Piloting and Modelling XNCJS in Three Counties

Following the usual trajectory of policymaking in the PRC, XNCJS programmes become more and more specific the further down in the governmental hierarchy that they are issued. This also holds true for the instructions and requirements for piloting and modelling. However, we did not find a uniform mode conceptualizing these practices in our three cases, but rather two different grand schemes.

In the case of Mizhi, promoting model villages is still the dominant strategy for XNCJS implementation. Higher up, Shaanxi province and Yulin 榆林 city strongly support the promotion of pilots and models in their respective XNCJS strategies, and interviewees at both levels confirmed that they constitute the focal points of rural development. This also applies to Mizhi county, where a leading cadre of the county's department of rural work described the establishing of model villages as “the essence of XNCJS.”Footnote ¹¹ On our first visit to Mizhi back in 2008, it was impossible for us to avoid being driven to Huihuang villageFootnote ¹² which enjoys province-wide fame as a showcase for successful rural development. A proud township Party secretary, whose sphere of administrative jurisdiction included Huihuang, recounted the history of the village while showing us the neatly laid out date plantations and reforested terraces stretching around the place, the refurbished houses, solid village roads and, last but certainly not least, the cultural centre of Huihuang with its own small library and museum. Huihuang was described to us as the embodiment of Mizhi's XNCJS spirit, a model that would lead and inspire the other villages in their efforts to implement the “20 character objective” passed down from the centre. The county Party secretary later “encouraged” us to leave aside the poor townships and villages in the county and focus our research solely on Huihuang village, because it showed the way that Mizhi county was going to go.Footnote ¹³ The promotion of model villages has been much more prominent in Mizhi than in the other two counties we visited.

Mizhi started with 17 county-level XNCJS model villages in 2006. Their total number increased to 30 in 2009. These villages received 50,000 yuan per year. A number of city-level “demonstration” or “model” villages partially overlapped with those that already enjoyed model status at county level. Of the 168 shifancun 示范村 (demonstration villages) that Yulin city supported at the time of our visits, 16 were located in Mizhi county.Footnote ¹⁴ Each of these villages received 500,000 yuan per year. The higher the administrative level that the “model village” belongs to, the higher the subsidies. This means that in Mizhi, city-level demonstration villages receive about 450,000 yuan more than normal villages.Footnote ¹⁵ Under a programme called “one red flag (five stars) and ten flowers” (yi mian hongqi shiduohua 一面红旗十朵花, which was later changed to yi mian hongqi wuxing shiduohua 一面红旗五星十朵花), Mizhi set up a special modelling scheme according to which one model village would serve as a blueprint for more model villages. However, these “premium” model villages were the result of substantial comparative advantages pertaining to geographic conditions that had enabled them, over a long period of time, to develop at a faster pace than the other villages in the area (such as Huihuang). Most of them, therefore, existed long before XNCJS came into being and are now, complemented by some new measures introduced under the umbrella of XNCJS, propagated by local officials as prime examples of “new socialist villages.”

In interviews as well as in official documents, the term shidiancun 试点村 is somewhat ambiguously merged with the term shifan 示范 when used to refer to model villages, but shidian can also refer to the implementation of “new” XNCJS projects within a village and thus carries a more limited meaning than shifancun. At the time of our last visit in 2009, specific XNCJS shidian projects comprised 106 villages out of a total number of 396 villages in Mizhi. As a first step, only 30 shidiancun were selected for these projects. Depending on the results, the measures were then extended to 76 other villages later on. Local officials explained that the usual life-span of a shidian is three to five years. Shidian in this sense pertains to a pilot that experiments by carrying out specific development projects within a village, although this usage was often blurred by equating the term shidian with shifan, as mentioned above. This makes it difficult to distinguish between the two cases and determine the precise numerical relation between models and pilots in Mizhi. However, the crucial fact is that, in Mizhi, XNCJS is primarily seen as an effort to create an increasing number of template or model villages, and every new measure that is launched offers privileges to those villages that are already on the road towards achieving this status.

Qingyuan and Dingnan follow a different trajectory. Although they also employ modelling strategies, these are clearly not the focal points of overall XNCJS implementation. Officially, Zhejiang province had begun to abolish the practice of establishing model villages in 2008. This means that the yangbancun 样板村 or mofancun 模范村 categories have vanished from official usage. However, the term shifancun still appears here and there in city and county-level documents, mostly with regard to villages that figure as models with respect to specific policies or projects. In each case, we only identified the existence of “keypoint villages” (zhongdiancun 重点村)Footnote ¹⁶ which receive special funding and incentive structures because they are classified as disproportionately poor;Footnote ¹⁷ however, these measures run under specific poverty alleviation schemes and must therefore be distinguished from XNCJS-related projects. In fact, the main XNCJS modelling in both counties is based on what we would call a project-based strategy according to which specific projects (xiangmu 项目) are spread over a couple of villages to spur homogeneous spatial development.

On the basis of this general approach, Dingnan county features a project-site (dian 点) strategy that differs significantly from the strategy of establishing model villages as in the Mizhi case. In Dingnan, a project is linked to as many sites as possible, conditioned by the amount of distributable money and villagers' support. One site can apply for as many projects as are laid out in the village development plan and are affordable through the matching funds allocated among the relevant segment of the population. A dian usually comprises a natural village (ziran cun自然村), or one or more village small groups (cunmin xiaozu 村民小组), that is, at least 30 households. According to Jiangxi province guidelines, each selected dian receives the fixed amount of 160,000 yuan for XNCJS projects.Footnote ¹⁸ Thus, one village can have about 20 dian. Initially, the village must set up a so-called village XNCJS administration council (cunmin xinnongcun jianshe lishihui 村民新农村建设理事会) (hereafter VAC) that is responsible for a specific project.

The VAC consists of three to five of the most influential individuals in a village, excluding the heads of the formal village leadership, i.e. the Party secretary (cunzhibu shuji 村支部书记) and the village director (cunzhuren 村主任). It applies for XNCJS project funds and is later responsible for implementing the XNCJS-related measure in the selected dian. VAC members communicate closely with the county and township officials in the relevant government departments who are responsible for ensuring that the project is realized within a one-year period. For example, a county-wide XNCJS-related initiative to install flush toilets and pipe systems for potable water targets all villages and must be successfully accomplished in a number of villages each year according to the county's development plan. A further example is the construction of new houses for peasants in so-called “new communities,” another XNCJS project in Dingnan which enjoys high priority. Construction is subsidized for any village households that are ready to tear down their old dwellings and move out, although this process is supposed to be incremental since the limited government funding means that the building of new communities can only be carried out consecutively in the county. In addition, each government department also tutors one dian for full-fledged XNCJS implementation. By 2010, Dingnan county had 481 dian. In 2010 alone, there were 77 newly identified dian at province level and 22 at county level.

In Qingyuan county, XNCJS implementation generally applies the same strategy as in Dingnan, although the selection and distribution of XNCJS projects is not primarily connected to specific sites. Rather, Qingyuan follows a “project management model” (xiangmu guanli de moshi 项目管理的模式) logic. A preference for the construction of basic infrastructure guides the decision on each project. Villages are categorized according to their needs and financial capacities. Different village categories receive different amounts of financial support to realize projects under the heading of XNCJS according to pre-set standards. Moreover, special projects set up to refer more directly to the ershi zi mubiao 二十字目标 (objectives summarized in 20 characters) are spread over a number of villages in the county, promoting, for example, “village remediation” (zhengzhi cunzhuang 整治村庄). Another programme, called “ten demonstration villages and one hundred renovated villages” (shi cun shifan, bai cun zhengzhi 十村示范, 百村整治), has been launched in response to a province-wide initiative for village hygiene and greening. It aims to renovate 200 villages within a four-year period between 2008 and 2012. These villages will then become sites for selected projects coordinated by the involved county-level departments and the township governments.Footnote ¹⁹ County cadres from various departments stressed that eligible project sites are selected on the basis of “sustainability,” meaning that villages should be able to run these projects themselves after the period of official sponsoring and supervision has ended so that governmental support does not have to be constant and can move on to other sites. In contrast to Dingnan, Qingyuan's modelling strategy does not come with the establishment of new participatory and coordinating bodies like the VACs, but is generally conferred on existing institutions, most notably the village committees.

Mizhi's modelling of XNCJS is most typical of those cases that have aroused much scepticism among domestic and foreign observers in the past: overly beautiful villages and seemingly content peasants in an environment that is much less beautiful and developed. Our interview partners in Mizhi stressed repeatedly that the major goal of the county government is to “stimulate driving forces for rural development by promoting trial sites (shidian).”Footnote ²⁰ However, they found it difficult to explain exactly how selected villages would successfully create “best practices” and how these would be emulated by other villages later on, given the differences in natural conditions, historical trajectories and economic development between the privileged few and all the other villages in the county. It was impossible to identify any linkage effect between the model villages, most notably Huihuang, and their surroundings. Even the Party secretary could say little more than that Mizhi's model villages represented a new spirit that other villages would (have to) assume in order to progress.Footnote ²¹ This, however, seemed to be wishful thinking at best. To implement XNCJS by strategically setting up model villages served the objective, first and foremost, of stimulating competition between townships and villages for the scarce funding.Footnote ²² County leaders noted that well performing shidian could develop into shifan villages, thus becoming eligible for more subsidies. The criteria applied to decide on the distribution of these funds, however, were more oriented towards the given stability of villages and good working relations between county, township and village leaders than towards a village's objective success in creating or emulating “best practices.”

In Dingnan and Qingyuan, as discussed above, county and township cadres do not focus on “traditional” modelling, that is, on selecting single villages as XNCJS showcases, but on the effective implementation of one project or different measures in many villages. Qingyuan has also embarked on what can be called complementary site implementation, according to which different villages located in adjacent vicinities are strategically linked by one specific project – for instance, the planting of particular crops. After one village is seen to receive preferential treatment for developing and marketing a certain crop, other villages are supposed to join in (and are subsidized) once market success becomes visible, resulting in agricultural intensification and economic gains for all villagers in the wider territory. In Dingnan, as in the whole province of Jiangxi, the dian strategy is considered a “preferential system” (puhui 普惠) in the sense that those sites which have already successfully implemented XNCJS projects can apply for more projects and funding than others.Footnote ²³

It is controversial in the development literature if development is best pursued by concentrating resources versus spreading them more thinly over wider areas.Footnote ²⁴ Closer to the later variant, Dingnan and Qingyuan are clearly aiming at systematically linking up their different XNCJS projects, particularly in the realm of agricultural modernization and village reconstruction. Moreover, latecomers may become eligible for government XNCJS funding as well. Whereas in Mizhi formerly successful villages have simply been affixed with a new label and now figure as XNCJS shidian or shifan villages, Dingnan and Qingyuan have at least tested new initiatives or creatively modified existing projects, bringing about a number of innovative “best practices” (most notably in the realm of land management, rural cooperatives, township administration and local urbanization schemes). These initiatives were steered by the county government in close cooperation with the townships and villages. The experimentation was not initiated at immediate grassroots level. Furthermore, we did not see any evidence that the county governments had lobbied for their “implementation models” or “innovations” at higher levels, as described by Heilmann.Footnote ²⁵

Interestingly, our respondents in all three counties disagreed on the more abstract question of whether “model tourism” was helpful to rural development. Almost all of them had taken part at least once in a trip organized by either Party schools or departments to a model village somewhere either nearby or in a different province, sometimes even as far away as South Korea and Taiwan. While many cadres praised these “inspirations,” the extent to which they affected local policy implementation back home remained unclear. Many cadres openly doubted the value of model emulation – at least in technical terms. Others voiced the opinion that it was important for maintaining an “innovative spirit” and “competition” among cadres; some simply stated that nowadays model emulation is kexue 科学 (scientific) and must therefore be taken seriously.

Altogether, we found that county and township cadres in the counties we studied entertained very different ideas and concepts concerning experimentation and modelling. They were certainly ordered to apply specific strategies by upper-level guidelines and thus did not have sole discretion on what form these strategies would take. However, at the same time county governments were still allowed substantial leeway when following these guidelines and were able to enrich them through local experimentation and modelling strategies. They also had specific resources and the means to implement them.

Implementation Features of XNCJS Piloting and Modelling

XNCJS policymaking at and below county level is very much agency-driven and quite autonomous from upper-level interference, as was confirmed by one township Party secretary in Mizhi county: “At the stage of formulation, the state's (project) policy [XNCJS] is not biased. But when implementing it, we have to consider some other factors.”Footnote ²⁶ A clear picture of these “other factors” can be gained by looking at the way that XNCJS modelling is executed at grassroots level. At least in the counties that we investigated, it was not bottom-up experimentation in villages that later inspired and informed township or county-wide implementation, but counties that determined sites where specific XNCJS measures – and sometimes a whole array of them – were to be implemented.

How has local XNCJS implementation been sequenced in our field sites? According to the given procedure, villages apply for XNCJS project funds annually. The township screens these applications and then passes them to the county government which has the final say in project selection and funding allocation, usually made on the basis of some 20 criteria (including population size, developmental capacity, etc.). In Mizhi, our respondents were divided on the issue of the selection criteria. For instance, some leading cadres emphasized the fact that a village's economic development or capacity was not of major importance,Footnote ²⁷ but others claimed that just this was the case.Footnote ²⁸ Some interviewees declared that city government officials and Mizhi county's department of rural work cooperatively selected shidian, while other county departments did not participate in this selection, suggesting a rather hierarchical process which is arguably less “scientific.”Footnote ²⁹

Selection criteria in Qingyuan and Dingnan were described to us as based on strictly objective standards. County officials emphasized that the selection of projects and project sites was horizontally coordinated between the relevant government departments at all stages of the process, in full accordance, we were told, with the requirements of scientific (kexue) procedures and homogenous development over the whole county. This was particularly emphasized in Qingyuan, where the nomination of model villages had officially been abolished in the whole province of Zhejiang in 2008. In Jiangxi, where shifancun still have some – at least symbolic – importance, the selection of project sites was tied to giving preference to “favourable conditions” which would strengthen the site's future shifan potential.

In addition to the sustainability of a project, the capability of village leaders and the enthusiasm of village residents in supporting XNCJS, our respondents named other, more specific criteria which confirm strategic modelling by county governments. Funding decisions, for instance, depend substantially on the degree of “peaceful” cooperation or “solidarity” (tuanjie 团结) within a village, between the village and Party committees, and on the quality of communication between the cadre bureaucracies at village, township and county level. This translates into a clear advantage for villages and townships that have established a good working relationship with the county departments and the county lingdao banzi 领导班子 (leadership team).Footnote ³⁰

Generally speaking, all XNCJS funding at county level is organized differently in each case, as is the overall coordination of XNCJS. In Mizhi, XNCJS funds were passed down from city level as a lump sum and allocated to the relevant departments under the supervision of the county's XNCJS leadership small group, while the department of rural work was responsible for all further coordination. In Dingnan, parts of these funds were shifted directly to the project sites which had their own “accounts” (zhanghu 账户), with the remainder being allocated to the county departments which included them in their annual operating budgets. Qingyuan did not receive specifically marked XNCJS transfers, but annually increasing earmarked funds for each government department (to be used for their respective projects) were given an XNCJS label in accordance with the county's XNCJS blueprint.

In relative terms, Mizhi's defined model villages apparently received the largest portion of XNCJS funds passed down to the county. In order to achieve quick results and induce other townships and villages to strive for shifan-status themselves, payments were made in instalments, depending on a project's implementation progress.Footnote ³¹ A leading cadre in Mizhi's department of finance stated that annual investment for one shidiancun was about 450,000 yuan per year, compared to approximately 100,000 yuan for “normal” villages.Footnote ³² In Dingnan, shifan support usually “only” involved minor awards and was thus said to be more symbolic in nature. Project dian received 160,000 yuan (see above), but “selection supremacy” was ensured by the fact that the county could decide to give several dian to a certain village and none to others. Zhejiang province, as has been mentioned, no longer engages in traditional modelling. Despite this fact, however, a kind of sub-provincial model category seems to have been maintained in order to create incentives for effective policy implementation. Cadres of the Lishui 丽水city bureau of rural work inferred that they would “award” (yi jiang dai bu 以奖代补) an extra amount of 50,000 yuan to villages on the basis of good evaluation results in XNCJS implementation.Footnote ³³ In Qingyuan, leading county cadres also made it clear that although the official use of “models” had been officially terminated, policy implementation on the ground still required “some kind of shifan work.”Footnote ³⁴ Thus, in all three counties, there was some funding of models that bypassed regular budget channels and was sometimes directly disbursed to the model unit itself or to the responsible government one level up. External observers, however, do not have access to information on the exact figures involved.

County governments certainly dominate the process of XNCJS implementation, but they still depend on the township authorities in many ways. In Mizhi, county officials ascribed townships a “missing link” or nexus function (chengshang qixia 承上启下) between the county and village levels. According to a leading cadre of the county's department of rural work, because the villagers and village cadres' suzhi 素质 (personal/educational quality) is relatively low, township officials must strongly support project work in the villages:

Although they are in a subordinate position to the counties, township governments have a say when it comes to project adjustment (xiangmu tiaozheng jianyi 项目调整建议) or proposals for eligible new shifancun.Footnote ³⁶ In Qingyuan and Dingnan, however, differentiation was often even more clear-cut, downgrading the townships to a kind of “cooperating unit” instructed to assist villages through the application process, but without any say in the later prioritization of these applications at county level. Apart from this formal division of responsibilities, county and township officials cooperate closely with village leaders to ensure smooth project implementation and are on high alert to avoid any problems that might lead to social instability (shehui bu wending 社会不稳定), since maintaining social stability is crucial for the positive evaluation of any government department or individual cadre.

The often mentioned positive reaction (jijixing 积极性) of villagers to XNCJS projects remained rather abstract to us. As far as direct participation was concerned, we were not able to confirm any meaningful contribution on the part of the villagers to project application and design in Mizhi. In Qingyuan, we were told that more than half of the members of the village assembly had to vote for a project before it could be inserted in the village's annual development plan. In Dingnan, new VACs had been set up to institutionalize villager participation in the XNCJS process. One of their major tasks is to ensure an 85 per cent threshold support in the village assembly for the village's XNCJS blueprint. Moreover, VACs assist village officials and the township government to meet project targets and overcome any potential obstacles. From this perspective, VACs act as the long arm of county, township and village officials, helping local governments to make XNCJS happen on the ground. From a different perspective, a VAC can also be seen as an elected body that takes care of villagers' specific demands concerning XNCJS project implementation and serves as an important feedback mechanism for county and township governments to minimize possible resistance at grassroots level. VACs in Dingnan (and the whole of Jiangxi province) can therefore be seen as both a steering instrument for county and township cadres and as an institution for strengthening the bottom-up participation of villagers in defining and implementing XNCJS projects. Since village residents often have to help finance XNCJS projects by either bringing in complementary funds (peitao zijin 配套资金) or by providing voluntary labour, VACs can be a crucial means of local cadres' strategic implementation as they ensure some degree of functional participation and, arguably, legitimacy in the implementation process.Footnote ³⁷

Our interviewees were quite sensitive to the underlying question of “distributive fairness” concerning the different modelling strategies of XNCJS implementation. In Mizhi, where project distribution was very selective and clearly favoured shifan and shidian villages, local officials underlined the fact that the county government would not neglect the other villages and would pursue a strategy of homogeneous spatial development: shidiancun would “receive some extra XNCJS subsidies, but other villages can still receive funding, for example, for road building projects.”Footnote ³⁸ However, it was admitted that the model villages also have an advantage when it comes to applying for “regular” project money, since their performance sheets are usually far better than those of the other villages.Footnote ³⁹ For Mizhi's officials at county and township level, identifying early developers and latecomers was legitimized, at least rhetorically, by the emulation effect that outstanding villages were supposed to have on the other villages, helping their economic and social development.Footnote ⁴⁰

In Qingyuan and Dingnan, where projects were more broadly spread but still prioritized spatially, all cadres justified this procedure by referring to strong funding constraints and assured us that all villages in the county would be covered within a reasonable period of time. In Qingyuan, one leading official in the department of rural work stated:

Selective implementation, especially via modelling, is often said to result from China's excessive target, project and cadre evaluation systems which set the wrong incentives and work against effective policy implementation.Footnote ⁴² Particularly in Mizhi, it was quite obvious that model villages served as showcases for successful policy implementation which would boost the cadres' evaluation record. The same was true for a number of “beacon-like” project sites in Dingnan and Qingyuan. Given the set targets for XNCJS pilots and model sites that can be found in official documents,Footnote ⁴³ we asked our respondents at all levels about their perception and understanding of the evaluation process. Although most of the respondents complained that the evaluations constituted a heavy burden for them, some noted that local policy implementation would not work without this kind of pressure and control. Upper-level monitoring of project budgets was reported to have become increasingly strict over the years and the bulk of XNCJS funding came down to the counties as highly regulated, earmarked funds or transfer payments. Diversion of funds had been made an absolute “one vote veto” (yipiao foujue 一票否决) criterion in Jiangxi's XNCJS evaluation, from province level down to dian level.Footnote ⁴⁴

Some respondents mentioned that the county evaluation of the townships would be much stricter than the evaluation of the county by higher levels.Footnote ⁴⁵ At the same time, some cadres hinted that county and township governments “cooperated” informally to work through an evaluation.Footnote ⁴⁶ For instance, some township officials in Dingnan indicated that they are able to negotiate the eventual assessment of target fulfilment with the county government if they encounter certain unforeseeable hardships (for example, natural disasters or a financial crisis, as in 2008/2009) during the evaluation period. Others said that suboptimal results in one project area could be outweighed by an excellent performance in others. Friction between counties and townships could thus be avoided, whereas the pressure exerted by both tiers on village leaders remained high to ensure smooth project implementation even under disadvantageous conditions and to prevent villages developing any wrong ideas about a possible negotiation space when dealing with higher levels. In the words of one township cadre in Dingnan, “The method of evaluating villages [by the township] differs from the way the county evaluates us. If the result of the evaluation is very bad, we will certainly punish the village cadres.”Footnote ⁴⁷

Confirming the criticism found in much of the cited literature on policy implementation in the local state, quality, at first sight, seems to have been compromised in the evaluation of pilots and models. For example, in Mizhi, one leading county cadre said quite bluntly that although the city level was imposing a huge workload on county officials as a result of its continuous demand for evaluation, the pressure was still tolerable, as “their [higher-level governments'] main concern is a fair number of XNCJS model villages.”Footnote ⁴⁸ A look into official documents does not reveal much about possible sanctions in cases where modelling was unsuccessful. This confirms the generally negative opinion of many China scholars of the evaluation system. During our interviews, cadres reported that sanctions ranged from individual career setbacks and salary cuts to the reduction of bureau funds and a simple, but hurtful, “loss of face.” We would thus argue that project evaluation, which extends to models and pilots as well, is as much a top-down instrument that places severe constraints on the autonomy of local cadres as it is a useful means for county governments to a) ensure compliance throughout the bureaucratic hierarchy down to village level, and b) ensure horizontal and vertical cooperation between county and township cadres.

Assessing the Results of “Strategic Modelling”

Since we focused on aspects of political agency in the context of “strategic modelling,” we did not undertake a thorough investigation of XNCJS outcomes in our three field sites. Nevertheless, some observations concerning the results of piloting and modelling should be reported at this point. As has been noted above, model villages in Mizhi county enjoyed considerable advantages when it came to the distribution of scarce government funds and were the main destinations of the regular inspection teams sent down by the county, city and even provincial government. XNCJS measures thus strengthened the privileged economic position and showcase functions which these villages already enjoyed, but did not display any visible linkage effect on the development of adjacent (non-model) villages. Mizhi county officials, certainly aware that their strategy left them open to criticism, were quick to assure us that the county government takes care of Mizhi's overall development and does not just focus on model villages. We found sufficient reasons to believe that this was true, at least with respect to the construction of roads, irrigation systems and water reservoirs all over the county. Looking at specific measures for the purpose of spurring agricultural intensification and modernization, Mizhi is still a county of a few better-off islands in a sea of rural poverty, and the way that modelling is undertaken there does at least create some doubt as to whether the county has left the conventional trajectory of image-building.

XNCJS modelling in Dingnan and Qingyuan follows a strategy of spatial development via selective project site implementation. This strategy, we would argue, is more complex and demanding in terms of coordinating and assessing XNCJS than that of promoting model villages which should then take the lead. More villages play a part in the selection and implementation process, as model enclaves do not figure particularly prominently in the local development strategy and are intentionally avoided. Consequently, XNCJS projects can be found all over the counties and most villages have seen at least one or two projects implemented during recent years. Against this background, we hypothesize that Dingnan and Qingyuan have been more effective in implementing XNCJS over a wider territory than Mizhi has so far. We would also argue that this is not only related to very different points of departure in terms of economic development and geographical advantage, but also to the greater developmental spirit displayed by officials in those southern counties in comparison with their counterparts in north-western Mizhi. However, further research is necessary to substantiate these assumptions.Footnote ⁴⁹