Since Xi Jinping (習近平) launched his anti-corruption campaign, the investigation and punishment of local officials has been viewed as a successful way to restore “legitimacy” to the Chinese Communist Party and local state governments. However, after several years of this campaign, many policymakers and scholars have pointed out the counter-intuitive ways in which this reduction in local official discretion has negatively impacted both the economy and governance. For example, a former central bank official, David Li, argues that the anti-corruption campaign has exacerbated economic slowdown by closing off informal mechanisms for using interpersonal ties to allocate resources and forge contracts. Additionally, Xuelian Chen and Christian Göbel contend that the campaign has altered local incentive structures encouraging innovation thereby reducing local innovation, often described as the source of the adaptability of the CCP to changing economic, social and political conditions.

Policy experimentation is an important source of bottom-up, endogenous change in democracies as a way to craft individualized (versus one-size fits all) policies better addressing local conditions and a way to “test” policies before being adopted by other local governments or nationally. Regarding the United States, “laboratories of democracy”[1] scholars contend that elections serve as the mechanism to incentivize policy experimentation under conditions of political risk. Creating new policies often challenges vested interests and carries the possibility of failure, so is inherently a risky action. The promise of electoral gains offsets this risk in some situations.

However, in an authoritarian context without strong electoral incentives, policy experimentation (and possible failure) might be seen as too risky, maybe even a potential challenge to the leader at the top of the hierarchal political system. Despite the perceived threat of sources of bottom-up policy change, policy experimentation in this type of system might be even more vital to outcomes like durability, social stability, and good governance than in a democracy, because these systems lack other sources of feedback mechanisms from lower levels of government and broader society. Top leaders in these systems might recognize this since, despite the high levels of risk and uncertainty, examples of policy experimentation exist in many states such as Russia, Vietnam, Chile under Pinochet and even totalitarian states like North Korea initiate policy experiments with economic reforms. Scholars often characterize these policy experiments as designed and implemented by central leaders; however, many of these experiments are actually piloted by local leaders first and then co-opted by central leaders, or are designed and implemented in collaboration with local officials.

Many scholars have noted the adaptability of the Chinese Communist Party (CCP) in responding to governance challenges as the key factor behind their authoritarian resilience and others have pointed out the importance of this adaptability in economic reforms. However, despite the rapid economic growth and regime durability experienced by China being attributed to policy experimentation, most studies of large state bureaucracies predict that local officials would act as rational bureaucrats and not engage in risky and uncertain policy experimentation. Innovation is uncertain because subnational officials usually have terms less than three years, which is often not long enough to show results; innovation is risky because officials do not know if the results will be positive or negative, or challenge elite power or other vested interests. Policy change that is seen as ineffective or causing negative outcomes might cause the official to lose his/her position. Despite the uncertainty and risk, however, we observe a great deal of policy innovation at the subnational level in China, including pilot programs to reform the household registration system, village- and township-level elections, and social welfare programs. Thus, the puzzle motivating this research is what causes local officials in China to do something new – create a new policy, launch a pilot, adopt an experimental policy – especially when innovation is uncertain and risky?

The extant literature on policy innovation in China relies on two vertical causal mechanisms to create incentives for policy experimentation. First, the mechanism from “above” is the Target Management System (考核/考评制度) which is used to evaluate leading cadres annually for promotion or punishment. Central policymakers encourage local officials through the promise of promotion to innovate in response to governance challenges in order to then integrate the local experiences back into national policy. Second, the mechanism from “below” is social pressure for policy solutions to local problems. For example, Joseph Fewsmith contends that local political needs for governability influence innovation, and Xuelian Chen argues that concerns about party-state reputation and legitimacy drive policy experimentation. Scholars debate whether local officials will innovate to either compete for promotion or maintain their position by minimizing social unrest. This debate corresponds well with the American politics literature that relies on the electoral mechanism to incentivize innovation for officials seeking promotion or avoiding punishment.

These vertical causal mechanisms are unsatisfying both from a theoretical and empirical perspective. Theoretically, this account of policy experimentation assumes that all subnational officials respond in similar ways to these incentives; however, given the risky and uncertain nature of policy innovation, we might predict that these incentives do not structure behavior perfectly. Policy decisions made under conditions of risk and uncertainty evidence more variability in outcomes depending on loss aversion. Empirically, evidence from many studies of policy innovation and diffusion demonstrate that many factors influence decisions of local officials to engage in policy experimentation. The evidence from studies of policy experimentation in China mirror the general literature, so does the presence of multiple causal pathways mean that we cannot understand or model local officials’ motivations for policy innovation? How do we theorize this evidence of variation?

In response to this variation, Roberts and King add individual-level attributes into this structural understanding of policy innovation by arguing that there is an “innovative personality” that encourages certain officials to innovate even if the structural factors are not conducive. Studies have shown that these individuals are charismatic, intelligent, and creative; however, are also “conventionally neurotic”, and crave achievement to the point that they often lack “concern for the feelings of others” and will do anything necessary “including bending the rules and outright illegality (in at least one case).” This utilization of individual-level attributes corresponds with the use of prospect theory in the fields of economics, policy studies and international relations scholars to understand how loss aversion and other psychological factors distinguish among different risk thresholds and subsequent behaviors. Mintrom and Norman developed a “policy entrepreneur” framework to theorize about these individuals who are willing to accept high degrees of risk: “Policy entrepreneurs distinguish themselves through their desire to significantly change current ways of doing things in their area of interest.”

As Ahlers, Heberer, and Schubert note in the case of China, scholars need to study policy innovation in a more interactive way that captures individual-level attributes:

“Local state maneuvering in contemporary China is shaped as much by centrally designed policies and institutional control mechanisms (tax competition, cadre and performance evaluation and promotion) as by strategic agency on the part of local cadres which must deliver results to the upper levels in order to ensure their future careers, or is promotion is not likely, at least in order to maintain their current rank and social status. This tension entails, as we argue, the increasingly successful reconciliation of local state agents’ collective interests amidst ever-increasing complexity and challenges. Policy-making in China is determined neither by top-down guidance nor by bottom-up collusion and “muddling through”, but rather by a delicate mixture of central state “signaling”, institutional constraints, and strategic agency on the part of local cadres across all administrative tiers. Therefore, current local governance arrangements and effective policy implementation are the product of central state stimuli under conditions of re-emergent claims at steering and standardization, and local responses, which are increasingly influenced by and take into account relevant non-governmental stakeholders with specific relevance for policy implementation.”

This quotation captures the interactive and strategic nature of policy innovation, which we contend necessitates an evolutionary approach that focuses on individual-level attributes interacting with environmental stimuli.

We, therefore, propose to use an evolutionary approach to analyze how the interaction of individual preferences and existing institutions in China create incentives for local officials to act as policy entrepreneurs. An evolutionary approach requires three factors interacting with each other: mutation, selection mechanisms, and replication. In the context of policy innovation in China, we use the individual-level attributes such a base preferences for innovation as the “mutation”, in that local officials naturally vary in attributes such as risk tolerance and personal efficacy that impact a preference for innovation. Selection mechanisms refer to anything that narrow the scope of variation within the system. In this case, scholars focus on the institutional incentives that reward or punish officials. Our approach is similar to prospect theory in that the same set of selection mechanisms may be perceived differently; however, we argue that individuals’ risk tolerance combined with feelings of personal efficacy interact with the selection mechanisms. This means that officials with a higher risk tolerance and higher personal efficacy perceive the benefits of successful innovation to be higher and those punishing failure to be lower; while officials with lower risk tolerance and lower personal efficacy perceive the same set of incentives in the opposite direction. This type of framework provides better empirical information on the variation in local agent preferences, analyzes the interaction between these preferences and the incentive structure facing policy entrepreneurs, and explains why local officials experiment with new policies.

Based on this in literature, as well as other studies the psychology, we posit the following framework for understanding the interaction of the individual level characteristics — namely one’s own belief in their personal efficacy and their fundamental risk tolerance. These ideal types are detailed in figure 1. Officials with a high risk tolerance and high belief in personal efficacy, are assumed to be of innovative personality. Those who believe in there and efficacy but are unwilling to except individual risk, tend to consult with others, and are perceived of as democratic in their basic predisposition. Low efficacy but high risk tolerance might characterize the corrupt cadre; whereas, lower risk tolerance and Low personal efficacy most likely to be associated with a deference to authority, and a desire to follow rules for promotion and personal gain.


As an initial test, this project first intends to test the plausibility of individual level variation amongst government officials with respect to fundamental personality characteristics and preferences. Initial surveys of 90 local officials find that—even under conditions of strong authoritarianism—individuals fundamentally very in terms of their personalities and base level preferences. However, these personality characteristics are insufficient for understanding pathways to innovation. In order to do this, we must understand how individual characteristics interact with political institutions. In other words, how do we understand the interaction of preferences and institutions? We argue it is important to understand the determinants of policy innovation in a single-party, authoritarian bureaucracy, because, as seen in cases ranging from village-level elections to environmental regulations, this experimentation has the potential to reshape policies and institutions across the country. Thus, we intend to survey local officials to measure preferences and understand how they view their incentive system. This is area is vital to understanding regime resilience and adaptation of the CCP to social and economic changes, and whether it will continue to do so under the current anti-corruption campaign and reduction of local discretion.

[1] Justice Brandeis popularized the term “laboratories of democracy” in the 1932 case of New State Ice Co. v. Liebmann to describe how a “state may, if its citizens choose, serve as a laboratory; and try novel social and economic experiments without risk to the rest of the country.”

This piece is based on a larger project with Orion Lewis and Reza Hasmath financed through SSHRC funding.

This article was originally published in CPI Analysis.

Editor: Edward White