Large Is Beautiful: Political Performance Assessment under Economic Decentralization
In December 1978 the CCP held the Third Plenum of the Eleventh Central Party Committee. Although no concrete reform measures were laid out in the meeting, it was decided that the focus of government policy should shift from class struggle to economic development. That decision is widely deemed to be the starting point of economic reforms in the post-Mao era. It also had important implications for the personnel policy and practice of the CCP, which had long focused on noneconomic considerations.1
During the three decades that followed, the CCP’s Organization Department (CCPOD), which is the command center of personnel controls in the party-state, issued a series of directives and guidelines that redefined and tweaked the political performance assessment system of the party-state in line with the dominant policy agendas of the CCP leadership.  Four important changes are especially noteworthy.
First, concrete criteria of performance were adopted, and quantified indicators were used to facilitate the assessment of party and government leaders at various levels of public administration. In particular, economic performance occupied the central place in the new metrics of assessment. It was not until 2006 that more diverse criteria were adopted by the CCPOD to guide political performance assessment in a “scientific” manner. Second, political performance assessment was not only consequential to the career movements of leading officials but coupled with material incentives in such forms as performance-based bonus and salary. Third, while central strategic control remained and ad hoc intervention from higher levels of authority was not infrequent, performance assessment and personnel decisions on local (sub-provincial) officials were delegated to provincial, prefectural, and county authorities over their immediately subordinate levels. Accordingly, centrally introduced guidelines and indicators were adapted to address local policy priorities that varied over time and across jurisdictions and administrative levels. Fourth, subordinate officials’ evaluative feedback on local leaders was incorporated as part of the performance assessment process. This practice, known as minzhu pingyi (democratic appraisal), was initiated in 1983 and institutionalized in 1988 with the use of standardized feedback forms and procedures.
The relationship between economic growth and local officials’ behavior under the new and evolving performance assessment system and the concurrent decentralization of economic (especially fiscal) authority has been analyzed from different angles. Oi (1992, 1999) discerns a positive link and explains it from a local developmental state perspective. Montinola, Qian, and Weingast (1995) (see also Maskin, Qian, and Xu 2000) argue that the combination of fiscal decentralization and political rewards for economic performance amounts to a form of federalism that induces promarket and progrowth policies of local governments. Zhou Li-An (2007) uses a tournament model to examine interlocale political performance competition as a key causal mechanism and expounds on its implications for local economic growth (see also Li and Zhou 2005).
Evidence in support of the positive link view mainly comes from data analyses on provincial leaders’ promotion (e.g., Bo 2002; Chen, Li, and Zhou 2005; Li and Zhou 2005). Some other empirical studies of local leaders (e.g., Guo 2007; Landry 2008), however, reveal less definitive effects of economic output growth on elite mobility. Still other studies find no significant effects at all. An analysis of a revised set of the data used by Li and Zhou (2005), for example, yields results with sharply reduced significance in the effect of GDP growth on provincial leaders’ promotion (Tao et al. 2010). Another study incorporating more variables (Shih, Adolph, and Liu 2012) finds no close link between economic output growth and political ranking among local leaders selected into the CCP central leadership. Instead, factional ties and factors like education and revenue are found to have mattered more in the allocation of political rewards.
Despite these different findings, none of the data analyses explicitly addresses the question whether economic growth was a necessary condition for upward political mobility. Indeed the focal concern they share instead is whether the political reward for local economic performance corresponds incrementally to variations measured on an interval scale. While a useful question, it may not lead to findings that will shed clear light on promotion decisions using much less differentiated (e.g., binary or ordinal) metrics. Such a situation is not uncommon, especially given the multifactor considerations involved in political performance evaluation and the temporal and spatial variations in local policy environment and priority under centrally defined guidelines that emphasize the importance of economic performance (Feng Junqi 2010; Wang Hansheng and Wang Yige 2009).
During 1995-2009 I interviewed a total of 137 officials at different levels of sub-provincial government. Although there was no consensus as to whether those who excelled in local economic performance were more likely to be promoted, 135 of the interviewees agreed that poor management of the local economy in comparison with regional peers and with predecessors was a major negative factor in the assessment of political performance. In addition, 129 of them considered it imperative not to let local economic performance fall below the average level of regional peers and the attainment of predecessors. As one of them opined (informant, 29/1996), “You have to do well enough economically to move on, however important other factors may be [for promotion decisions].” More systematic evidence consistent with this opinion comes from a panel data analysis of the career movements of the party secretaries and mayors in 296 prefectural cities during 1990-2004 (Chen and Lin 2016). The results show that, other things being equal, local leaders failing to achieve economic performance at par with that of regional peers and predecessors had significantly lower probability of being promoted than those with better performance, though the difference among those with average and above-average performance was insignificant.
These findings suggest that, although local officials might not all strive to be stellar performers in economic growth, they had to be mindful ofthe implications of the “pass-fail” test, implied by the results reported above, for their careers. The means that they used to address this concern and related self-interest agendas during the 1980s and 1990s is the focus of the analysis in this chapter. What I seek to explore is how local officials responded to the new political incentives they faced under evolving structural conditions (especially fiscal decentralization) and what consequences their responses had for the decline of public ownership. Instead of separating their strategies into those seeking political (career) benefits and those seeking economic (e.g., revenue related) benefits (Li and Zhou 2005), I see them as closely related and oftentimes mutually reinforcing pursuits.
The point of departure for my exploration is Susan Whiting’s (2000) pioneering study of the evolution of rural property rights institutions from the start of economic reform to the mid-1990s. Based on comparative case studies, Whiting argues that the new political performance assessment system and fiscal revenue sharing with higher-level government in the 1980s and early 1990s drove rural officials to promote the growth of industrial output. Yet the extent to which rural officials pursued this strategy through local public enterprises (i.e., TVEs) depended on the strength of collectively owned rural industry before the reform. Locales with a strong organizational legacy tended to seek further expansion of TVEs. They also adopted tactics to hide fiscal revenue in public enterprise accounts that local officials could manipulate. Locales with a moderately strong organizational legacy shared the tendency to develop public enterprises, whereas at the same time they also granted limited tolerance to private businesses but made them absorb the transaction costs of revenue extraction. Locales with weak organizational legacy tended to take a more accommodating approach to private business. To tackle the transaction costs of revenue extraction they devised simplified metrics of tax estimation. The overall policy bias toward promoting output growth of TVEs led to dysfunctional consequences, however. The expansion of TVEs was facilitated with soft credit, which spawned inefficiency and overproduction and exacerbated the difficulties of centralized control over state bank lending. Revenue hiding contributed to the deteriorating fiscal capacity of the central government. These problems compelled the central government to restructure public finance and banking in the mid-1990s. Coupled with growing competition, the restructuring hardened the budget constraints of local governments and forced them onto the path of privatization.
This is an insightful analysis that has broad implications. It identifies political performance assessment and fiscal relations as key variables to explain local officials’ behavior; it also relates the interplay of these variables to the initial conditions of the local political economy to account for spatial variations and illustrates important links in a complex process of endogenously induced institutional change. All these are useful analytic elements that can also be extended to the study of local political actors beyond the context of townships and villages. At the same time Whiting’s work also points to several issues that need to be further explored for understanding and clarifying the causal mechanisms of privatization in the post-Mao era.
First, striving for industrial output growth, which Whiting emphasizes, represents a strategy different from promoting industrial sales growth, which is the focus of this chapter. Noting this seemingly minute difference is not a matter of splitting hairs. The former was practically difficult to sustain because of cash flow constraints in a marketizing environment (see next section). More importantly, it could have fallen short of serving the best interests of local officials. The main reason for this is that output by itself was not a revenue-generating tool. Although increasing industrial output would directly correspond to a key performance evaluation indicator (i.e., gross industrial and agricultural output before the mid-1990s and GDP thereafter) under the new political incentive system, any unsold output would not help address the revenue imperative under the reforming fiscal structure, where some 70% of the budget revenue came from indirect taxes realized in or through sales during 1980-1993.  In contrast, promoting industrial sales would generate more immediate relief to the cash flow problem (although not without limit). And it could boost output growth as a corollary and at the same time address the revenue imperative, including the agenda of expanding discretionary resource pools (e.g., parking of exempted taxes in enterprise accounts). Still another benefit of sales growth is that could help generate and sustain nonfarm employment, which was another important policy imperative for local governments, as long as the volume of output sold could keep growing.
Second, Whiting’s analysis notices that the “dysfunctional” consequences from the use of local public enterprises to boost political performance represented a moral hazard problem, yet why the problem was pervasive among local officials needs to be explained.11 Part of the answer, I argue, lies in the broad interests that the strategy of promoting sales growth served to address. Other factors include transient tenures and lack of clear cost accounting in public administration, both of which further spawned opportunism.
While local leaders might be personal rivals and had different factional ties (Edin 2003; Shih 2008), promoting sales growth tended to be the dominant strategy not only because of its superiority in terms of narrow economic calculus but because of its low vulnerability to local politics. It had higher payoffs, for example, than subversive strategies that spoilers might pursue. The reason is that the common career benefits it furnished often assumed the key features of public goods—nonexcludability and nonrivalry (Samuelson 1954)—for those in the top echelon, especially given that achieving growth through local public enterprises was widely deemed as the most legitimate means to advance the new economy (hence also raising the political cost of opposition). Moreover, local leaders were not the only beneficiaries of the expansion of discretionary revenue pools. Subordinate officials in charge of economic affairs had access too. Many of them were directly involved in the management and administration of local public enterprises and thus held a key to the implementation of the sales growth strategy. Under the new political performance assessment system, their evaluative feedback on local leaders was also a factor that local leaders could not afford to ignore. Promoting sales growth of local public enterprises provided an avenue for co-opting these political actors, thereby reinforcing the dominant strategy of local leaders.
Furthermore, the political performance assessment system itself was afflicted with inherent problems in the personnel practice of the party-state. An important feature of the CCP’s nomenclatural system in the reform era is the transient tenure of local leaders. Factors leading to transient tenures were many, including concerns about nepotism and factionalism, strategic positioning (e.g., in the name of huoqu jiceng gongzuo jingyan, or “gaining grassroots level work experience”) of officials slated for subsequent promotion, alignment of loyalty with higher-level authorities and policies, rotation for diversification of work experience and environment, reduction of redundancy, responses to contingencies (e.g., personnel reshuffling in antigraft campaigns), repartitioning (merger or division) of administrative boundaries (of townships and urban districts), among others (Edin 2003; Feng Junqi 2010;
Wang Hansheng and Wang Yige 2009). A major consequence of transient tenures of local leaders is the prevalence of a high time discount rate in their decision-making. With short time horizons in their political expectations, incumbent local leaders tended to seek maximization of short-term political and personal benefits, even at the cost of creating or exacerbating longer- term problems, such as deterioration of the financial health of local public enterprises.
Adding to the growth of moral hazard among local officials was the lack of clear cost accounting for administrative decisions under the political performance assessment system. The system was geared toward assessing the tangible “achievements” by local officials on a yearly basis, rather than administrative efficiency in terms of cost-benefit comparison. There was a built-in mechanism calledyipiao foujue (disqualification for promotion on account of one major dereliction), which was introduced in the early 1980s as a way to strengthen the implementation of family planning and subsequently extended to include other issues, such as major accidents, social unrest, food contamination, and so on. But it was event-specific and not well suited for ascertaining responsibilities for the cumulative negative effects of decisions by past local leaders, which was made even more difficult by transient tenures. In fact it was not uncommon for incumbent local leaders to gloss over growing problems (e.g., through borrowing new loans to service loans already incurred by local public enterprises to expand sales despite diminishing returns), to blame any exposed problems on their predecessors, and even to claim credit for finding “solutions” to problems that were part of their own doing.
Third, the consequences of moral hazard for privatization need to be further explored. Whiting offers two useful pointers: growing financial liabilities of local public enterprises due to overexpansion, and post-1993 hardening of financial and fiscal budget constraint as a result of the central government’s efforts to address problems such as lax banking controls and central revenue shortfalls, which had been exacerbated by local officials’ opportunistic behavior in the early years of reform. The shocks from the 1994 banking and fiscal restructuring were indeed important for understanding the debacle of public enterprises after the mid-1990s, as I will discuss later. Yet there was more to the story. The decline of public enterprises was an uneven process. The sales growth strategy ran out of force in numerous locales even before 1994, which therefore cannot be explained by the shock effect. Whiting traces the variations in the development of local public enterprises to the significance of commune and brigade enterprises in the local economy before the reform. The focus of such an account of path dependence, however, is on the initial policy preferences of local officials and their subsequent strategies of revenue extraction rather than the differences in the ability to carry forward the course of TVE-led growth. Nonetheless, it can be broadened to help illuminate the latter.
The adverse effect of moral hazard on local public enterprises could be enhanced by three major factors: poor ability of local officials to provide strategic guidance to local public enterprises, managerial corruption, and collusion between local officials and public enterprise managers for purely private purposes. As the stakeholder in local public enterprises, where and how the local government made its investment mattered greatly to the near-term sustainability of these enterprises. For the sales growth strategy to work, cash flow was critical. A minimum condition for this was to make products that could sell, which was by no means a matter of autopiloting in an increasingly competitive environment. Also important was the ability of the local government to monitor the behavior of enterprise managers and prevent them from engaging in self-seeking at the expense of their enterprises. This task would become more challenging when local public enterprises experienced fast expansion in number, size, and functional complexity. The sales growth strategy, therefore, generated forces that threatened the governance of local public enterprises. A further challenge came from the widespread practice of hiding taxes in the accounts of local public enterprises, which entailed collusion with managers. Although initially intended mainly as a way to stash away funds for addressing at least some public spending needs, such collusive relationship could degenerate into purely private pursuits harmful to the enterprises concerned.
Where and when local public enterprises suffered from poor strategic investment and/or serious deficiency in supervision, the sales growth strategy would run into difficulties unless there was unlimited supply of credit. A vicious cycle could kick in when supervising officials ceased to see local public enterprises as the goose able to lay golden eggs (i.e., facilitating the sales growth strategy) and thus slipped into full private collusion mode with managers to make a quick killing. For many local public enterprises, the 1994 restructuring of banking and public finance was probably what tolled their death knell. But for others the driving forces for movement to the final stretch might have come into play earlier, largely because of poor organization and management by supervising officials. In this connection, understanding the role of the organizational legacy from the prereform era in internal governance helps us discern the mechanisms at work (other than an extended proclivity in administrative behavior or policy preference), as locales with more experienced administrators and stronger organizational infrastructure would more likely have the ability to sustain the sales-expansion strategy for longer time on the road to self-destruction.
Last but not least, what explains the variations in local governments’ initial tendency to pursue the sales growth strategy or in their ability to sustain it does not offer an adequate explanation for the variations in local governments’ tendency to support private business. Whiting argues that the strength of commune and brigade enterprises (CBEs) in the prereform era holds a common explanation for the varying degrees of tolerance toward private ownership in different locales. Plausible as it may seem, this argument needs to be qualified. Supporting private business was not the fallback option by default for locales with weak organizational legacies from the Mao era. The main reason is that supporting private business beyond centrally set limits involved a political risk. The ability to manage this political risk did not bear any directional correlation with the strength of organizational legacy. Other factors, especially those that helped justify the need for supporting private business, should be considered. I will leave this issue to the next chapter, though. The focus here instead is how the sales growth strategy played out.
-  For an analysis of personnel policies under Mao, see Lee 1991.
-  During 1977-1983 a series of personnel reshuffling took place throughout the party-state apparatusto replace officials purged for ties with Maoist factions and to fill the positions held by retiring officials.For accounts of the changes after the start of reforms, see Burns 1989; Edin 2003; Manion 1993; andWhiting 2004. Related discussions are also offered in Huang 1996; Cheung, Chung, and Lin 1998; andLandry 2008. Tao et al. 2009 provide a useful description of more recent developments.
-  A further note on the four changes highlighted here is posted at the book site.
-  The Chinese government did not compile GDP statistics until 1993. Industrial and agricultural output instead was the main output measure used in political performance evaluation before the mid-1990s.
-  For examples at the grassroots level, see Oi 1999 and Whiting 2000, 2004.
-  For examples, see Edin 2003; Landry 2008; and Whiting 2004.
-  A profile of the interviewees, including rank, age, gender, education, work history, etc., is posted atthe book site.
-  Contrary to the more widely shared view mentioned earlier, Cai and Treisman (2006) argue that thenew political incentive system and fiscal decentralization did not have a causal effect on China’s economic growth. Rather, factional rivalry among central leaders holds the key to explaining it. Since thefocal issue of my study is privatization rather than economic growth, I will not address that debate here.
-  As shown in the next section, there appears to be little evidence that TVEs had massive overproduction of output, which is often associated with output-maximizing behavior. Even SOEs were movingaway from emphasizing meeting output targets to turning out marketable products during the secondhalf of the 1980s, when the central planning system was in decline (Naughton 1995). What eventuallybrought down TVEs was not cumulative glut of industrial output—which is what Whiting’s analysis alludes to, but sales expansion without a close link to profitability. Focusing on industrial outputgrowth therefore may obscure this important causal linkage.
-  Whiting does pay attention to the role of the post-1980 fiscal system. But her discussion focuseson the impetus from the self-financing mandate for township governments rather than the compositestructure of revenue sources.
-  Whiting brings up the issue of moral hazard in her book (2000) and a subsequent paper (2004) butoffers no further analysis.
-  As to the discretionary pools of revenue resulting from the expansion of sales by local public enterprises, Whiting notices the use for private purposes but seems to de-emphasize this by pointing to thefact that part of it was used to finance regular public administration functions. What is interesting,though, is that the blurred boundary between public and private uses also created a cushion against thepolitical risk involved and helped neutralize spoilers (Lin 2001).
-  Guo (2007), for example, reports that during 1994-2002 only 42% of the county chief executivescompleted their five-year terms and only one-fifth of them remained in the same county to serve as thelocal CCP secretary. Landry (2008) finds that the average tenure of prefectural mayors decreased from3.2 years in 1990 to 2.5 years in 2001. A further examination of the data spanning 1990-2004 revealsthat 60% and 70% of the CCP secretaries and mayors of prefectural cities had a tenure of no more thanthree years respectively (percentages are calculated from data set).
-  For some interesting examples of the tactics used by local officials to finesse political credit and shedresponsibilities, see the in-depth case study of a county government by Feng Junqi (2010).