Throughout China's reform era, local governments have pioneered policies aimed at removing the shackles of socialist economic planning, introducing market-oriented initiatives, pursuing gradual privatization, and absorbing foreign investment. These moves have helped offset the central state's greater inertia against change from the command economy.
 
As China entered a consolidated phase of steadfast growth in recent years, these local governments embarked on new initiatives that transcend the economic realm and encompass a variety of governance issues. Indeed, the restructuring of the Chinese state and society--two key forces in public governance beyond markets and firms--has become an inevitable result of China's rapid economic transition and growth.
 
A recent and widely noted local governance reform method in China has been the practice of the "democratic consultation assembly" (minzhu kentan hui, or DCA) in the city of Wenling in Zhejiang Province. Since February, regular DCAs have been held in all towns, townships and villages in Wenling, with four major variations:
 
1) Face-to-face assemblies, involving the local party and state, and individual citizens or their representatives participate in collective discussions, deliberations, consultations and decision making regarding public affairs in a given community;
 
2) A significantly revamped budget process in Wenling's township-level People's Congress, primarily in the town of Xinhe, in which the local legislature exercises a solid influence on the executive branch's annual expenditures--a rare process in China;
 
3) An emerging mechanism, currently tried in the town of Zeguo, whereby sampled citizens interact with local officials, in ways determined by direct and deliberative democracy, to arrive at joint decisions on pre-selected categories of overall government spending;
 
4) Annual collective bargaining over wages, coordinated by local officials, between manufacturers' associations and the local labor union in the garment and footwear industries, as practiced in towns such as Xinhe and Wenqiao.
 
Although tangible discrepancies exist among these variations in terms of scope, formality, participation and substantive effects, they share a number of core, common features that make DCAs a distinct approach to local governance reform in China. Such features include expanded grassroots participation that produces substantive impact, enhanced transparency, less information asymmetry, and the strengthened ability to monitor the executive branch. Other features are more equality among participants, greater tolerance, greater willingness to compromise, and components of deliberative democracy. These common denominators stem from an evolutionary path in which the last three variations have, in fact, been extensions or adaptations of the first by applying fundamentals to various issues and within myriad institutions. These continuities result in a coherence among the four variations that makes a concept like the "Wenling model" analytically, if not practically, meaningful.
 
It is not surprising that the overwhelming majority of literature on Wenling's DCAs has concentrated, as the name of the practice suggests, on their significance for grassroots democracy and political reform in China. Nonetheless, it can be argued that while the promotion of democracy paves the way for governance reform in Wenling, the overall impact of reform can be more fully assessed through the additional perspective of public sector accountability--a crucial dimension recognized in extant literature on governance. This article attempts to answer two interrelated research questions. First, how do DCAs heighten public accountability in the current institutional setting of China's sub-province localities? And what can be learned from the Chinese case, in relation to achieving public accountability elsewhere?
 
To address the first question, this article will explore the first two variations of DCAs outlined above, and will focus on the interplay between managerial and democratic accountability orientations to address the second question.

Accountability in the Public Sector
 
Considering the specific nature of DCAs and the research questions to be examined, a highly relevant definition of public accountability offered by researcher named Bovens is that it is "a social relationship in which an actor feels an obligation to explain and to justify his or her conduct to some significant other." Such an obligation arises by virtue of the fundamental nature of public accountability as a chain of principal-agent contractual relations among the citizenry (and all stakeholder groups therein), legislative and judicial organs, and the executive branch. At each link in the chain, the principal is entitled to acquire sufficient information about whether the agent fulfils all delegated duties required by the contract. In view of the conceptual lens above, the subsequent analyses of Wenling's DCAs will elaborate selectively on the following four of myriad characteristics integral to the concept of public accountability:
 
1) How the contract comes to be delimited between citizens and local cadres in the first place;
2) How the relationship between various local party-state organs undergoes adaptation;
3) How local cadres become more responsive to popular demands and more capable of explicating and justifying their performance;
4) How local cadres can be subject to certain consequences, if sanctions are imposed due to failure in carrying out contractual obligations.
 
From the standpoint of public accountability and governance--notions originating from relatively established liberal democracies--Wenling's DCAs manifest uniqueness in that the mechanism has emerged and matured in a political system that lags behind most others in terms of democratization, but ranks high in state-led deployment of resources, in combination with market forces, to attain socioeconomic development effectively. This distinctive feature makes DCAs prominently instrumental in probing interactions between the pursuits of democratic participation and managerial efficiency; two divergent accountability orientations, which is a leading issue in the literature on public management and governance.
 
Managerial accountability is widely seen as a defining feature of New Public Management (NPM), and is conceived as underlining the more plentiful, prompt and efficient provision of public goods and services, normally with a special emphasis on result-based measurement by means of performance evaluation of public managers who are held accountable. Strongly influenced by the managerial focus in the private sector, managerial accountability is sometimes also associated with market-based convictions besides efficiency, such as privatization, competition and entrepreneurship.

In contrast, democratic accountability is often advocated by those critical of NPM, who call for citizens to be brought closer to the government and its policymaking process, usually through participative institutions and a variety of forums that nurture direct and deliberative democracy. Among other things, proponents of the two orientations disagree over the extent to which, and in what ways, long-standing democratic institutions in liberal democracies ought to be adapted or reshuffled, as post-industrial societies necessitate greater managerial efficiency in the public sector. Given their aforementioned uniqueness, Wenling's DCAs will provide a new avenue for comprehending and assessing the present views involved in the debate.
 
Principal-Agent Contract
 
In liberal democracies, public officials are held accountable, often in accord with the policy platforms previously committed to by elected political parties or candidates during the campaign stage. These pledges, together with popular demands articulated through  a wide variety of opportunities to influence bureaucratic and legislative actors provided by a structurally pluralistic society, serve as key yardsticks to which public managers' performances are expected to conform. Therefore, they are also the defining foundation for the principal-agent contractual relationship epitomizing public accountability.
 
In the absence of sufficiently competitive elections in China's villages and towns and townships, public accountability appears to be a questionable concept, since the contract lacks a sufficient constituent mandate. To fill this vacuum, DCAs are a prime conduit through which citizen demands are channeled to public officials, whereby both principals and agents share a common knowledge of what the latter are obliged to deliver. In this article, the first variation of DCAs at the village level will be examined as an example of this instrumentality.
 
The specific DCA studied was held on June 24, 2005, in the village of Shangdun, in the town of Wenqiao. This particular DCA was organized to discuss the dismantling and relocation of about 280 residential houses, accommodating about 70 percent of Shangdun's population, to make the land available for a major roadway through the village. The responsibility for implementing this project, handed down from the county as a major public construction project, was assigned to the Wenqiao party committee as a so-called "political mission." This mission for the Shangdun Village Party Committee (VPC) required it to ensure that the massive dismantling of residences and relocation of residents, unprecedented in the village's history, was completed smoothly, and to minimize the possible social turbulence often ensued in similar cases elsewhere.
 
The chief concerns of the villagers whose houses were affected by the project related to the living conditions and the impacts on their regular jobs (not limited to farming) resulting from the relocation and the new housing. The villagers' opinions, expressed during pre-DCA correspondence with the Shangdun Village Committee (VC), converged on the insistence that, unless all new conditions would make them better off in general, the relocation was to be turned down.
 
More than 90 percent of Shangdun's villagers attended the DCA in June 2005, with three popularly determined representatives of the villagers articulating the stance and terms demanded by those households affected by the relocation. In response to pressures both from inside and outside, the VPC secretary embraced a two- pronged tactic. While stressing the vitality of the mission, delegated from above, to facilitate the roadway's construction, he also committed himself to requesting fiscal assistance from Wenqiao in covering the costs of the relocated households. Despite this tactic, which seemed to involve both carrots and sticks, suspicions and debates still arose over whether the demands from above were justifiable, dragging the DCA into a prolonged tug-of-war.
 
Two rounds of voting were taken to resolve the disagreements among villagers. After more than five hours of bargaining, compromises, voting and mutual persuasion, the DCA finally concluded by passing the relocation, accompanied by a set of concrete criteria for compensating villagers for myriad costs, either by assigning them new residences or by covering their financial losses. The DCA's conclusions were mostly put into effect during the relocation, except that three households filed complaints, in subsequent DCAs, about compensatory payments not being made on time.
Legislative-Executive Relations
 
In parliamentary or presidential systems, federal and non-federal legislatures and political parties are institutional devices that act on behalf of constituents, in terms of public accountability, by monitoring and judging performances in public. In China, however, such devices lack binding power, as exemplified by the more symbolic than substantive role played by China's legislature at various levels, namely the People's Congress system. Compared to its counterparts in liberal democracies, the system remains largely submissive to, if not entirely a "rubber stamp" for, the party-state. Deviating from this national pattern, the second variation of DCAs, which commenced in July 2005 and is found in the Xinhe People's Congress (XPC) in Wenling, is a mechanism that extends the participatory decision-making of DCAs to strengthen XPC's regular function of budget review, as well as its approval and audit supervision. In adopting this system, XPC vastly altered Xinhe's legislative-executive relations, beefing up its instrumentality as the agent of citizens to better ensure the local party-state's political accountability.
 
Given what the DCA had stood for in terms of governance reform, following the DCA formality was a key factor in justifying XPC's innovation. However, it was in fact at a regular session of the XPC that the innovative budget process unfolded. Nevertheless, since then the extent to which XPC has made substantive changes to the budget plan drafted by the town government (TG) of Xinhe has not been matched by any comparable case at the township level in China.
 
Chief among the innovations was a vast increase in information and details in the TG's draft budget proposal presented to the XPC, primarily the specifications in terms of revenues and expenditures of all concrete items and projects. A lack of sufficient information in executive branch budget proposals at all administrative levels has been a main barrier to exercising adequate and meaningful oversight by the entire apparatus of the People's Congress in China since 1949.
 
Furthermore, during the DCA, XPC's representatives exercised their powers of interpellation, in relation to obtaining more detailed budget information, and obtained responses from TG officials on specific items and relevant criteria. Immediately after the DCA, a committee, composed of selected XPC members, collated all views and information obtained and proposed revisions to the TG's draft budget.

After reaching an agreement with TG over these suggestions, the committee brought the conclusions back to XPC's plenary session for a final vote. Finally, XPC established a standing committee on fiscal and economic affairs to monitor, when the XPC was not in session, implementation of the finalized budget plan resulting from DCA.
 
In March 2006, the example of XPC in continuing a DCA on public finance demonstrated exactly how some core aspects of the budget process were handled. XPC's first-stage DCAs for a preliminary review of the TG's draft budget, including the divisions on industrial, agricultural, and social developments, was convened two days before the official XPC session. DCA participants included members of the newly established standing committee and representatives who had chosen one of three divisions specified above.
 
Expenditures for the draft budget, totaling 92.97 million yuan, included 15 major categories that, in turn, were subdivided into 110 specific projects. For example, the expenditure on infrastructure for urban development, in the amount of 21.91 million yuan, the largest chunk of all, contained 15 projects, such as payments for land expropriation (7 million yuan), development of the Jiangbin Industrial Zone (5 million yuan), renovation of village infrastructure (2 million yuan), erection of a new processing center for environmental and public health affairs (580,000 yuan), renovation of aging urban areas (500,000 yuan), renovation of historic and cultural assets zone (400,000 yuan), building of the main town area's roadway loop and drainage system (500,000 yuan), and land-use planning adjustment costs (300,000 yuan). TG's administrative outlays of 15.87 million yuan--second only to urban infrastructure in terms of spending--included 17 items, such as the salaries and personnel fees for active TG employees (5.3 million yuan) and for retired employees (2.45 million yuan), the costs of conferences and meetings (1.2 million yuan), administrative costs and office supplies (800,000 yuan), utility and phone bills (800,000 yuan), transport and business trip costs (700,000 yuan), a reserve fund for employee housing (500,000 yuan), a Xinhe Labor Union subsidy (250,000 yuan), and new vehicles (650,000 yuan). After reviewing the itemized draft budget, each division produced its own preliminary report.
 
The preliminary reports then became the benchmark for discussions during the second stage of DCA, which was literally a plenary session of the XPC, and for a comparison with the TG's briefing on the draft budget. In the plenary DCA, which lasted a whole afternoon, there were questions or suggestions for revision from 13 XPC representatives relating to 18 specific projects or items in the draft budget. Notably, their suggestions for revisions conformed to a rule of thumb, set by XPC in advance, that any proposed increase in the budget for a specific item was to be accompanied by a proposed reduction in some other item, so that the total expenditure of 92.97 million yuan would remain unchanged.
 
The next stage was a joint meeting of the XPC's presidium, its standing committee and TG, in which negotiations took place to generate a draft proposal for budget adjustment. The principal adjustments included converting the environmental and public health center into a transfer station for garbage (in the same amount), expanding the budget for renovation of aging urban areas (from 500,000 to 1 million yuan), and cutbacks in the protection and refurbishment of historic buildings and streets (from 1 million to 500,000 yuan). This draft for budget adjustments was then presented on the floor at the XPC, where representatives were free, with the endorsement of at least five representatives, to initiate amendments to any specific part of the draft.
 
This was a new procedure and central to the overarching goals of boosting XPC's substantive budget influence and promoting more participation from among the rank and file. As in the previous plenary session, the amendments were required to keep the overall expenditures unaltered by proposing simultaneously both increases in some items and decreases in others.
 
Eight amendments were submitted to the plenary DCA at this stage. However, six were not actually qualified for voting because of the absence of any concrete sums proposed for the budget change. Both the remaining amendments to be voted upon in the plenary DCA were passed, registering a 100 percent success rate. TG made the corresponding adjustments and presented the finalized budget plan based on these two amendments. The first adjustment was a cutback in the items of surgery fees and inspection fees under the category of birth control (from 1 million to 500,000 yuan), combined with a growth in the straightening and dredging of village settings under the category of urban infrastructures (from 2 million to 2.5 million yuan). The second adjustment was an increase of 500,000 yuan for garbage cleanup, balanced by a decrease of the same amount in a fund in the urban infrastructure category. After the overall DCA process ended, the standing committee established in 2005 continued its oversight of TG by regularly reviewing TG's quarterly reports of the budget's implementation, and by taking part in forming the following year's budget plan.
Responding to Popular Demands
 
The public sector's responsiveness to societal demands is more accurately gauged by its actual problem-solving activity and policy outcomes than by the mere creation of certain programs or projects. Likewise, the dissemination of information relating to these actions and outcomes also facilitates the evaluation by the principals, as to the degree to which the performance of the agents is justifiable. The specific scheme for revealing an executive branch's concrete actions and ensuing outcomes found in Wenling's DCAs at the town and township level, and being part of the first variation, shows how the need for responsiveness and justification are dealt with. Most towns and townships in Wenling have formed substantive and procedural rules on DCAs consistent with the Regulations on Democratic Consultation Assembly in the Songmen Town. Heralding DCAs, the Songmen regulations were chosen in March 2004 by the Wenling party committee as a model for institutions and practices, and were to be introduced to all other towns and townships. Under these regulations, four rounds of DCAs are to be convened each year, preferably one round in each quarter.
 
The Songmen regulations entail the simultaneous presentation of issues and preliminary proposals for their solution at town DCAs. This procedure is rarely found in the rules and practices of village DCAs. In addition, the concrete policy proposals to be discussed in DCAs are required by the Songmen regulations to be based upon sufficient operability and feasibility. Taken together, these rules highlight a distinctive feature of town DCAs as opposed to the village DCAs. Most town DCAs are not as inclined as village DCAs to solicit, from the general public, reminders about the issues and fundamental goals or values that are not incorporated into the government agenda. This is partly because many of these reminders often stand a good chance of being raised by the town and township People's Congress (TTPC), which enjoys greater institutional legitimacy than the villagers' assembly (VA), the de facto legislature in villages. Town DCAs usually focus on the issues subsumed under, or derived from, some pre-existing policies, programs, or issues already under consideration or in progress. In contrast with village DCAs, they are more about the choice or adjustment of concrete policy instruments and the resulting allocation of public resources.
 
By virtue of the ensuing abundance of information disclosed in town DCAs on technical details and feasibility, local cadres responsible for handling an issue or policy deliberated ex ante are often subject to public evaluation ex post. Such evaluation, usually conducted in a structured question-and-answer fashion between citizens and cadres, is in most cases the last part of the session on pre-scheduled issues and bills in DCAs. Responsible cadres are required to file written and oral reports on the implementations authorized during previous DCAs. All participants are free to raise further questions and to request cadres' replies after a predetermined team, composed of both the TTPC representatives and citizens, has done so. If cadres are unable to offer a satisfactory answer to any question, the issue would be revisited at the next regular DCA.
 
Accountability through Consequences
 
The ultimate, and perhaps the most critical, issue involved in holding agents accountable is the actual consequences they may face if they fail to realize contractual obligations to the principals. Such consequences, often resulting from sanction mechanisms inherent in the accountability institutions themselves, such as legislative impeachment or judicial prosecution, involve various costs which the agents would want to avert. In Wenling, a loss of employment for the cadres--one of the most severe possible consequences in an electoral democracy--has not yet happened as a direct result of the DCAs. Nevertheless, DCAs can impose credible costs on the cadres by augmenting intended sanctions of other existing institutions that were inadequately applied under China's particular political and administrative environments, as shown by two types of cost. First, while no administrative or legal sanction is formally stipulated for an apparent discrepancy between a local cadre's disclosed performance and evaluation by the masses in DCAs, the pressure perceived by local cadres has brought considerable costs. In general, the stronger the challenges and condemnations a cadre confronts in the publicized gathering of the DCAs, the less competent, reliable and persuasive he or she appears in the eyes of the masses. A tarnished image and reputation often weakens the legitimacy of a cadre's leadership and, thus, the future development of his or her career. According to six of eight village leaders and all five of the town and township leaders interviewed between 2005 and '07, this threat is more disturbing than anything else, including legal or administrative sanctions.
 
Notably, the perceived cost from this threat varies among the two types of DCAs previously discussed. Pressures from XPC representatives were hardly considered as creating a negative effect, in either of the values above, for the head of the TG. In the first variation of DCAs, the local heads of the VC, VPC, TG and town or township party committee (TTPC), being accountable to village and township DCAs often discover a boosted legitimacy, thanks to greater public participation and mutual persuasion. But they also perceive a mounting defiance among disgruntled groups of citizens on certain zero-sum issues, such as land expropriation and ensuing compensation. The crucial implication of this finding is that, thanks to DCAs, powerful VPC and TTPC cadres who are not popularly elected and thus not accountable to citizen voting decisions as in liberal democracies now have a heavy personal stake to the extent to which they can meet the expectations of the masses. This remedies the absence of penalty through ballots as an ultimate safety valve for party officials' accountability. The DCAs substantially increase publicity about their performance which, in turn, provides critical input in the vertical personnel control system that dictates their career prospects.
 
Second, holding the first variation of DCAs in compliance with what is stipulated in all relevant official regulations has become an established category in the formal annual cadre evaluation of VPC and TTPC secretaries.
 
Cadres in the party's propaganda department at county, town and township levels not only check the frequency and formality of DCAs held, but also personally attend most DCAs. Therefore, failure to conform to the regulations would jeopardize official evaluation records of party secretaries. This institutionalized rule is thus crucial to preventing opportunistic behavior by averting the first type of cost by dismantling the DCAs. To date, there has been only one case of this behavior. The party secretary of Songmen indicated, in an interview with the author in 2006, that the "disruption of local development" by the masses through DCAs could have been forestalled, or at least have become less popularized, if the DCAs had not provided an outlet for their grievances. As a result, he simply suspended all town-level DCAs in 2006, regardless of the possible adverse impacts resulting from the suspension on the outcome of his annual cadre evaluation. However, this case showed how influential the first type of cost can be in holding the cadres to account.

Managerial Vs. Democratic Accountability
 
As initially noted, China in the reform era has stood in contradistinction to other countries by prioritizing managerial over democratic accountability. Specifically, its quest for managerial efficiency relies on a carefully crafted system of cadre management and evaluation, imposed from the top leadership all the way through the state hierarchy, to ensure the fulfillment of national policy goals across diverse levels of the party-state apparatus. During China's reform era, this system grew increasingly institutionalized as the foundation for making personnel decisions, and was substantially shored up by the introduction of a variety of merit-based performance measures under the rules of former President Jiang Zemin and President Hu Jintao. Such measures are designed to reflect how successfully local cadres reach policy targets set by the officials at the next level up. Hence, the principals who really hold all state agents accountable, given China's pursuit of managerial efficiency, are not the citizens, but the cadres' supervisors.
 
Therefore, the first implication of DCAs for the two contending views on accountability is that they (and their many functional equivalents elsewhere in China) amount to a belated, yet inevitable, evolution toward a situation that has long been consolidated in well-established democracies and has precipitated NPM stress on managerial accountability. China has long embraced NPM efficiency-oriented objectives, often in the context of deliberate promotion of inter-jurisdictional competition at all administrative levels, akin to market dynamics. Nonetheless, this approach in China to ensuring accountability results from an undersupply of democracy, as opposed to the need to tackle an oversupply of democracy in other countries influenced by NPM.
 
Nevertheless, this transition from managerial to democratic accountability in China should not be comprehended merely as a reversal of, or an exception to, an evolution toward NPM in general.
 
For example, when juxtaposed with some developing countries in Latin America, China's case fits squarely with the reality that the public sectors of different countries have already employed varying approaches to assigning the relative weights between managerial efficiency and democratic governance while, in general, pursuing NPM as their common goal. Chile and Uruguay have organized their public sector performance evaluations around the budgeting cycle and within the realm of budget allocations, thereby focusing on managerial reinvigoration. Conversely, performance evaluations in Colombia and Costa Rica have prioritized strategic planning and decision making at the policy or organizational level, in ways guided by incumbent political leaders' electoral commitments, according to which the citizenry hold public servants accountable. Therefore, as democratic initiatives in China sprout in an incipient stage amid an entrenched dominance of managerial accountability, China's case stands, in the broader terms of NPM, somewhere between the two contrasting groups of Latin American cases.
 
Second, from the perspective of public accountability as a chain of principal-agent ties, managerial accountability--relations between distinct levels in public organizations--is in fact an intermediary node of this chain, and thus cannot be divorced altogether from the public's mandate that governs the contract between the political leaders and their constituents, which is the starting point in this chain. Indeed, the enhancement of democratic accountability in Wenling, thanks to DCAs, goes hand-in-hand with a marked headway in the instrumentality of managerial accountability for the job performance of local officials. In the author's field interviews, 22 of 27 officials reported positive responses when questioned about whether their jobs had been facilitated by their supervisors' adaptations of criteria and/or procedures for job oversight in line with outputs from DCAs.
 
Likewise, when asked about whether it was, in their experience, more effective to monitor subordinate performances when public decisions and guidelines emanating from DCAs were incorporated, 19 of 21 officials agreed it was. Whereas this finding in China alludes to the possible compatibility between two accountability orientations elsewhere, one has to consider China's peculiar background when interpreting the finding's external validity. For example, the relative paucity of managerial autonomy among public servants, under the restraints imposed by democratic institutions in the West, which is a key source of debate between the two competing views, is not at all outstanding in China. The democratic headway in Wenling may appear remarkable in comparison to most of China's other localities, but it remains far short of reversing the domination of the vertical control of officials in local governance.
 
Conclusion
 
Wengling's DCAs provide lessons far beyond China's democratization. In the enhancement of public accountability, DCAs play a pivotal role in defining the fundamental principal agent contract between the citizens and the local public sector, empowering the legislature to request greater political accountability upon the executive branch, increasing local cadre responsiveness and compelling justification of their performances, while imposing substantial costs on officials assessed unfavorably by the masses. Having remained, since inception, a mechanism alien to the formal political institutions, DCAs have nonetheless significantly altered relations at all nodes of the principal-agent chain and in the functioning of formal institutions, bringing about progress in the pursuit of public accountability.
 
Also, this Chinese case demonstrates more comparability than incongruence with the general trends and the ensuing conceptual perspectives of accountability in public sectors elsewhere. On the other hand, it entails a more nuanced comprehension of the relationship between various aspects of accountability, including the democratic and managerial dimensions that preoccupied academic discussions. To what extent the mushrooming initiatives of China's local governance, such as DCAs, will proliferate upward within the administrative hierarchy remains uncertain, as do all the sweeping questions about China, such as its democratic future and long-term sustainability of its hyper-growth.

However, what is certain is that no matter what the pace or mode by which the proliferation may move forward, the practical issues and the analytic concerns of accountability and governance common to liberal democracies are likely to have an increasing applicability to China. These will also confront a growing need for conceptual refinement, necessitated by their application to new empirical realities, as this analysis of a Chinese city's case has revealed.
 
S. Philip Hsu is a political scientist at National Taiwan University in Taipei.