landry 31 The Political Management of Mayors in Post-Deng China The Political Management of Mayors in Post-Deng China PIERRE F. LANDRY Abstract This article examines how the Chinese Communist Party (CCP) maintains political control over municipal elites in post-Deng China. First, it analyses the career paths of mayors for all prefecture-level municipalities between 1990 and 2000. It then shows how the dual processes of economic decentralization and the decentralization of the personnel management system has not led to a decline of the CCP's capacity to enforce key organizational norms. An ordered probit model of cadre promotion suggests that the economic performance of cities has little substantive impact on promotion or removal from office. However, strict enforcement of cadre retirement regulations has the effect of shortening the tenure of mayors, which facilitates the promotion of a greater share of secondary officials than was possible before the reforms of the personnel system, but weakens the link between good governance and political rewards. Introduction During the post-Maoist era, the centre of gravity of the Chinese politi- cal economy tilted decisively towards cities. The unprecedented pace of China's economic transformation favoured urban growth, which in turn increased the political relevance of municipalities and the officials who rule them. Cities now control a far greater share of the country's resources than at any point in the history of the People's Republic. In 2000, Chinese municipalities accounted for 51.8 percent of the coun- try's GDP, 50.1 percent of its industrial output and 76 percent of the value of services (Jiang and Cui 2001). This increased economic might was largely deliberate. As early as 1979, the Centre1 targeted some of its boldest policy initiatives at municipalities, symbolized by the early crea- tion of Special Economic Zones (Crane 1990; Kleinberg 1990). After 1984, economic decentralization was generalized to other areas, but Premier Zhao Ziyang stressed that coastal cities would enjoy economic privi- leges that would not be extended to the less developed and more rural Chinese hinterland (Yang 1990). The leadership's favourable bias to- wards cities survived the transition of 1989. Until 2002, Jiang Zemin Pierre F. Landry 32 The Copenhagen Journal of Asian Studies 17•2003 and Zhu Rongji—both former mayors and party secretaries in Shang- hai—presided over further reform initiatives that benefited cities (Naughton 1995; Wang and Hu 1999). In parallel with their rapid economic transformation, cities also en- joy greater formal institutional weight. Their number rose rapidly— from less than 200 in 1978 to over 660 today2—but more importantly their formal bureaucratic rank as well as those of the cadres who rule them has also increased. Many county seats have been elevated to the status of 'prefecture level municipalities' (diji shi), and 15 cities now have 'vice-provincial' rank (fushengji shi).3 Scholars who have examined China's decentralization often conclude that the economic power of the localities has eroded the political au- thority of the Centre.4 This article examines how this authority is exer- cised with respect to city mayors. My choice of focus is not simply be- cause cities are inherently important, but also because the terms of the debate on the political impact of China's economic decentralization rely excessively on provincial aggregate data. Critics of decentralization who focus on central–provincial fiscal relations (Hu and Wang 1996; Wang 1994, 1995, 1997) conclude that state capacity has weakened, while their detractors, who emphasize the role of institutional and political con- trols—particularly the power of appointment of central Party institu- tions (Brødsgaard 2002; Burns 1989; Harding 1981; Manion 1985), ar- gue that the political authority of the Centre is still very much intact in high-priority areas (Bo 2002; Huang 1995, 1996; Lampton 1992). As important as provinces are, they may not constitute the proper level of analysis to adjudicate the debate, because of their particular position in the Chinese political hierarchy. Both Huang (1996) and Bo (2002) find evidence that central control over provincial appointments facilitates policy enforcement, but that result is hardly surprising given that Beijing has always retained direct nomenklatura authority over pro- vincial leaders. The institutional capacity of the Chinese state should instead be tested against a tougher standard, namely Beijing's capacity to impose its policy preferences when principal–agent relations are not as straightforward as those linking central and provincial leaders. It seems more fruitful to focus on local political actors over which Beijing only exercises indirect control—such as mayors. The party control mechanisms differ vastly between provincial and municipal cadres. Although personnel management over city, county and township cadres has evolved over time, under the current system, top municipal leaders are appointed by the provinces, without direct 33 The Political Management of Mayors in Post-Deng China central control (Burns 1989, 1994; Landry 2000).5 Furthermore, focusing on municipal elites allows one to test hypotheses about regional differ- entiation that cannot be evaluated by looking at provinces alone. Spe- cifically, one can learn whether the Party's organizational practice vis- à-vis local officials is consistent across provinces, and whether its per- sonnel choices follow the logic of rewarding good governance in the localities. This article tests the hypothesis that municipal performance affects the political fate of mayors. The Political Control of Cadres The CCP has mobilized considerable resources to enhance its political control over local cadres. It sought to improve personnel management by gradually reshaping the institutions that collect information, moni- tor the performance of local governments, and sanction officials (Harding 1981; Huang 1995; Whiting 2001). These institutional reforms were designed to act as a counterpoise against the centrifugal forces of economic decentralization. The web of Party organization departments (zuzhi bumen), discipline inspection commissions (jiwei) and local CCP committees is expected to root out cadres who flaunt central policies, are guilty of 'localism', or who are shown to be corrupt.6 This strategy of reform has allowed economic decentralization to proceed, but seeks to reduce the costs of devolution. The political principals of the Chinese state recognize that effective governance is a necessary condition to maintain regime legitimacy among ordinary citizens (Tang and Parish 2000). They rely on the CCP's institutional dominance to enforce the norm among cadres that the goal of achieving a 'relatively wealthy society' (xiaokang shehui) is critical to the regime's strategy for long-term survival. They stress the need to recruit and promote officials who deliver good governance––defined, for the most part, in terms of economic growth (Wei 2002: 1733). Article 6-2 of the 'Regulations on the Work of Selecting and Appointing Lead- ing Party and Government Cadres' specifies that 'cadres should . . . be determined to carry out the reform and opening-up policy, be devoted to the cause of modernization, and work hard for the building of social- ism and the making of concrete achievements' (CCP Organization De- partment 2002). The Centre has devised various incentive mechanisms to reduce shirking and to improve performance among local cadres.7 The question of how best to select and reward local officials has al- ways vexed the CCP's 'organizational and personnel management sys- tem' (zuzhi renshi xitong), but the problem is especially acute when the Pierre F. Landry 34 The Copenhagen Journal of Asian Studies 17•2003 cadre management system is decentralized (Zhang 1994). In his classic analysis of the Maoist polity, Schurmann (1968) argued that reliance on ideology could effectively compensate for weakened state organizations during periods of decentralization. The decline of ideology in the re- form era has revived the role of formal organizations as instruments of political control. The Centre's challenge of controlling cadres is dual: local officials not only have broad leeway to steer the local economy in a direction that they choose, but leadership selection itself is also decen- tralized. Since Beijing controls directly only a handful of posts below the provincial level, it must trust that appointed provincial leaders have both the will and the capacity to implement personnel selection policy in a fashion that does not undermine the Party's political authority at the sub-provincial level. Operationalization and Data Collection This evaluation of the CCP's personnel management strategy is based on fairly systematic data about China's municipalities.8 Mayors (shizhang) constitute a natural pool to study how Party committees al- locate political power: they exercise broad responsibilities in economic management, but remain politically subordinate to provincial Party or- ganizations. They do not rank so high that they stand no chance of fur- ther promotion if they perform well. In recent years, the composition of the top echelon of party leaders has reflected the importance of experi- ence as mayors and municipal secretaries for promotion to higher po- litical office.9 Since mayors typically serve concurrently as (first) deputy secretary of municipal CCP committees (shiwei fushuji), promotion to the post of municipal secretary is a natural career move for analysis. Other types of promotions are also possible: mayors sometimes become provincial vice-governor, or are deployed in central ministries. Thus, the dependent variable of interest is promotion, within the same locality or elsewhere. The collection of local political and economic data in the PRC is a challenging task. Data collection about provinces has improved (Bo 2002; Huang 1996), but this is decidedly not the case below the provincial level. Although basic socio-economic performance indicators have gen- erally been available since the 1990s, systematic political data are much harder to collect. Until recently, even lists of local officials were typi- cally not available below the provincial level, let alone detailed bio- graphical information about the cadres in question. Thus, practical con- siderations have motivated the focus on mayors: They are the lowest 35 The Political Management of Mayors in Post-Deng China level of local cadres for whom reasonably systematic biographical in- formation is available. Since the late 1980s, the Chinese Urban Devel- opment Research Council (CUDRC) has been publishing biographical notices of mayors annually (Zhongguo chengshi fazhan yanjiuhui 1985- 2001; Li and Bachman 1989). This information has been combined in the present study with a cross-section time-series dataset of municipal performance, supplemented with data gathered from a variety of sources that allow one to track, with reasonable confidence, cadre careers after their terms as mayor.10 Municipal Governance in the Era of Marketization Prima facie, performance standards are clearly spelled out and uniform across cities. The Zhongguo chengshi fazhan yanjiuhui (2001) lists 33 socio- economic indicators of municipal modernization (see Appendix at end of article) and rates cities accordingly. The linkage between performance and promotion is even highlighted visually, with the mayor's name, picture and short biography prominently displayed above the table. Mayors have a broad mandate to 'govern well', but they must do so under market pressures that increasingly constrain the capacity of the lo- cal state to control local outcomes. In fact, many objectives of good gov- ernance tend to reduce bureaucratic control in favour of non-state actors. Several indicators measure the growing impact of market forces that are beyond the direct control of local officials. For example, the share of serv- ices in the GDP (Indicator 3) depends heavily on the size of the non-state sector in the local economy. Similarly, openness to international trade (In- dicator 5, calculated as Import + Exports/GDP) is affected in part by glo- bal market conditions which local officials can hardly control. These indi- cators probably measure the pace of the localities' economic moderniza- tion, but it is harder to conclude that they accurately measure the leader- ship's contribution to the modernization of the local economy. These indicators are powerfully biased towards GDP performance: the variables that seek to measure 'economic development' not only account for almost a third of the overall index, but other factors are also strongly correlated with output growth, either by construction (Indicator 9) or in- directly (Indicators 10, 11 and 12, inter alia). Simple co-linearity tests between the main components of the index confirm that the indicators are deceptively broad (see Table 1). Market forces also have the effect of exaggerating performance gaps across localities. The rise of regional disparities has been well documented across provinces (Wang and Hu 1999), but the differences are even more Pierre F. Landry 36 The Copenhagen Journal of Asian Studies 17•2003 pronounced among sub-provincial units. Table 6 (pp. 44-45) illustrates the large inter-municipal disparities in contemporary China, although the broad components of the overall performance exhibit some co-linearity.11 Large inter- and intra-regional disparities have disturbing implica- tions for the Party authority. At worst, persistently poor performance undermines the credibility of the regime's claimed successful transition to a 'socialist market economy with Chinese characteristics'. If the CCP is serious about penalizing poor governance and rewarding good gov- ernance, then career patterns among local cadres should reflect this heavily differentiated landscape. By this reasoning, officials posted in fast-growing cities ought to be promoted more frequently than their counterparts posted in regions in relative economic decline. In summary, China's rapid economic transformation has aggravated the Party's adverse selection problem. Provincial CCP committees may have the power to appoint and dismiss mayors, but this institutional strength remains theoretical in the absence of accurate ways to measure the leadership abilities of local officials. As marketization deepens, it is increasingly difficult to map the economic performance of the localities to the specific action of officials. :1ELBAT ehtfostnenopmoCyeKneewtebytiraenil-oCfotseT xednIecnamrofrePlapicinuMCDUC :elbairaVtnednepeD )401=N(srorredradnatstsuborhtiwnoissergeRxednIerutcurtsarfnI )201,1(F 68.32 F>borP 00.0 2R 22.0 ESMtooR 70.2 elbairaV .feoC EStsuboR xednItnemnorivnE 853.0 370.0 *** tnatsnoC 763.5 040.1 *** :elbairaVtnednepeD )401=N(srorredradnatstsuborhtiwnoissergeRxednIecnamrofrePcimonocE )201,1(F 57.73 F>borP 00.0 2R 62.0 ESMtooR 73.3 elbairaV .feoC EStsuboR xednIefiLfoytilauQ 296.0 311.0 *** tnatsnoC 749.6 115.1 *** :etoN srotacidniecnamrofrephcihwrof)401=N(seiticfotesbusehtnodesaberaatadehT .seiticesenihCllafotesetelpmocatneserpertonseodsihT.dehsilbupera :ecruoS .1002iuhuijnaynahzafihsgnehcouggnohZ 37 The Political Management of Mayors in Post-Deng China Who Are China's Mayors? Chinese mayors are typically well-educated men in their fifties and overwhelmingly Han.12 Although they share the broad characteristics of provincial officials (e.g. more than two-thirds have received some form of tertiary education), they are younger. Whereas Bo (2002) re- ports a mean age of 55.5 years, male mayors typically are just above 50 and their female counterparts just above 48. In contrast to provincial officials, ethnic minorities seem under-represented (under 5 percent)— a low number, but one that also reflects the concentration of munici- palities in coastal Han-dominated provinces. Formal personnel regulations (CCP Organization Department 1995, 2002) strongly suggest that the odds of promotion depend on a cadre's personal characteristics. The rule of retirement (60 for most cadres, 55 for women) constrains career prospects heavily. Age limits were ini- tially introduced to rejuvenate the post-Cultural Revolution leadership (Manion 1993; Shen 1994). This policy not only led to the replacement of 'old revolutionaries' by 'career bureaucrats' (Harding 1981; Lee 1991), but was also conducive to a considerable improvement in the overall level of education among cadres, since newly appointed officials were typically better trained than their predecessors (Landry 2000; Shen 1994) (see Tables 2 and 3). Enforcing the retirement age is no longer controversial: the mayors' average age has hovered around 50 since 1990, well within formal regu- latory limits. The data indicate strongly that they are always removed from office by the age of 60, unless they serve in deputy-provincial- 2ELBAT )0002-0991(noitacudEfoleveL:sroyaM: leveLnoitacudE elaM elameF sesaCllA )latotfo%sa( leveLetaudarG 792 9 )3.21( egelloC 941,1 22 )1.74( egelloClanoitacoV )nauhzad( 914 31 )3.71( *loohcShgiH 86 0 )7.2( **rehtO 405 2 )3.02( latoT 734,2 64 )%( )1.89( )9.1( :ecruoS .esabataD .atadgnissimgnidulcnI*;)nauhzgnohz(noitacudelanoitacovgnidulcnI** Pierre F. Landry 38 The Copenhagen Journal of Asian Studies 17•2003 level cities or centrally administered municipalities (CAMs), where the formal retirement age is higher. Even in this group, no one has remained in office beyond the age of 64.13 Mayors also tend to serve shorter terms: the 1990 average of 3.2 years that matched the regulatory standard of three- year appointments had declined to a mere 2.3 years by 2000. Since 1998, few mayors have served a full three-year term (see Table 4 and Figure 1). The norm of retirement is a powerful way to guarantee that relatively young cadres—who are arguably better trained and more attuned to the workings of a market economy—reach leadership positions in the localities. However, its enforcement can also come at the expense of :3ELBAT )0002-0991(sutatSytironiMdnaredneGyb,egA'sroyaM redneG egAnaeM .oN noitaiveD.dtS elaM 05.05 979,1 79.4 elameF 01.84 640,0 45.4 latoT 44.05 520,2 79.4 yticinhtE naH 34.05 029,1 89.4 ytironiM 85.05 501,0 47.4 latoT 85.05 520,2 79.4 :ecruoS .esabataD 4ELBAT )0002-0991(noitubirtsiDegA:sroyaM: raeY .oN naeM .veD.dtS .niM .xaM 0991 791 5.05 4.5 53 36 1991 891 4.05 9.4 63 36 2991 691 6.05 3.5 63 46 3991 671 3.05 8.4 73 26 4991 571 9.05 7.4 83 26 5991 381 9.05 1.5 43 36 6991 261 7.05 1.5 53 46 7991 461 8.05 1.5 93 16 8991 571 0.05 8.4 73 16 9991 491 2.05 7.4 83 26 0002 502 6.94 6.4 83 36 :ecruoS .esabataD 39 The Political Management of Mayors in Post-Deng China FIGURE 1: Mayors' Average Tenure (1990-2000) Note: Units are years of tenure at the time of observation. Source: Database. Y = -0.08t + 171.48 R = 0.70 2 2.2 2.4 2.6 2.8 3 3.2 3.4 1990 1992 1994 1996 1998 200 Year Average Length of tenure (in years ) Length of tenure Linear (Length of tenure) cadre accountability. O'Brien and Li's finding that cadre rotation among rural cadres weakens accountability (1999: 176) applies here as well. It is difficult to see how a mayor, no matter how effective he may be, can signal his contribution to local development if his expected tenure in office is barely above two years. Modelling the Relationship between Cadre Characteristics, Municipal Performance and Political Outcomes Dependent Variable: Promotion For the sake of tractability, a mayor's political fate is ordered along a single dimension and analysed as an ordered probability model: In a given year, he/she can 'exit'14 (coded 0), 'continue' as mayor or be trans- ferred to a position of identical rank (coded 1); or be 'promoted' to the rank of municipal party secretary (coded 2) in the same city or else- where.15 Recall that ordered probit models make no scaling assumption of any kind: one need not assume that the difference between an exit and continuation is more or less desirable than the difference between 'continue' and 'promoted'. One need only be satisfied that: Exit < Continue < Promoted 2 0 Pierre F. Landry 40 The Copenhagen Journal of Asian Studies 17•2003 While the definitions of 'promotion' and 'transfer' are not particularly controversial, the reader may question the treatment of an 'exit' as less desirable than continuing as mayor. Mayors disappear from the dataset for various reasons. In some cases, they die in office or retire from public life. Alternatively, cadres may be eased into non-executive positions at the local People's Congress or the People's Political Consultative Conference. This ranking assumes that such posts are less desirable than the post of mayor. However, a third possibility is decidedly more threatening to statistical inference: exit may also reflect promotions to unobserved posts in provincial departments or central ministries. Since there is no comprehensive database of Chinese leaders above the prefectural level, there is no way to confirm that this third case is rare enough to ignore. However, every effort was made to verify that officials coded as 'exit' do not reappear in executive positions elsewhere. A mayor's appointment marks the beginning of the last decade of his political career, barring exceptional circumstances. The dataset includes a handful of leaders who attained high office in the 1990s, including Jiang Zemin, Zhu Rongji, Li Ruihuan and Bo Xilai, but the presence of former mayors among China's top leaders should not mask the grimmer reality that promotions are relatively rare events. Independent Variables Mayors' Individual Characteristics The model accounts for the salient individual attributes of cadre pro- motion regulations. These variables include age (in years), level of edu- cation, gender, as well as whether the cadre is Han Chinese or not. The effects of these variables are discussed in detail in the next section. In addition, two contrasts capture the length of tenure up to the year of observation: three to five years, which corresponds roughly to a second term in office, and six years or longer.16 Municipal Bureaucratic Rank A second set of variables captures the bureaucratic rank of each munici- pality. 'Central Appointment' accounts for cities whose leadership is man- aged from Beijing. Furthermore, CAM party secretaries often sit as full or alternate members of the Central Committee, which enhances their vis- ibility and access to core decision-makers. Their peculiar position in the Chinese political hierarchy is likely to increase their odds of promotion. In addition, an interaction term for 'Vice-Provincial Cities' accounts for in- stances of extensive autonomy over economic policy combined with par- 41 The Political Management of Mayors in Post-Deng China 0 20 40 6 0 80 100 120 140 16 0 18 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 Year Frequency Exit Continue Promoted tial central political control. Such is the economic status of provinces, but only their mayor and party secretary are centrally appointed. Deputy- mayors and secretaries remain under provincial management, since they have the rank of prefecture-level cadres (diji ganbu). Municipal Performance At the current stage of the dataset development, it is necessary to restrict the set of municipal performance criteria to four indicators that are highly correlated with the municipal performance index published in 2001: • the city size (population); • the city's overall wealth (GDP per capita); • the extent of the transition to a market economy–measured by the share of services to the local GDP; and • openness to the world economy, measured by the magnitude of foreign direct investment (FDI) in the city. 17 FIGURE 2: Mayors: Distribution of Political Fate (1990-2000) Note: In order to avoid 'right censoring', the analysis ends in 2000, a year when the fate of almost all mayors is known. Source: Database. Pierre F. Landry 42 The Copenhagen Journal of Asian Studies 17•2003 :5ELBAT )0002-0991(ecnivorPybnoitaunitnoCdna,tixE,snoitomorP:sroyaM ecnivorP )%(yrogetaCnoitomorP )wor(N tixE eunitnoC detomorP gnijieB 81 54 63 11 nijnaiT 81 46 81 11 iebeH 12 26 81 021 ixnahS 33 45 31 16 uggnemieN 82 35 91 34 gninoaiL 52 16 41 251 niliJ 91 06 12 48 gnaijgnolieH 72 95 41 21 iahgnahS 9 37 81 11 usgnaiJ 61 17 31 731 gnaijehZ 82 26 11 401 iuhnA 51 66 91 421 naijuF 32 17 6 87 ixgnaiJ 42 16 51 66 gnodnahS 71 46 91 151 naneH 12 26 71 351 iebuH 32 76 11 39 nanuH 42 66 01 131 gnodgnauG 81 36 91 681 ixgnauG 22 86 01 78 naniaH 41 28 5 22 gniqgnohC 0 09 01 01 nauhciS 71 27 11 531 uohziuG 22 57 3 23 nannuY 42 36 21 14 tebiT 9 28 9 11 ixnaahS 22 56 31 77 usnaG 61 96 51 55 iahgniQ 81 37 9 11 aixgniN 42 67 0 52 gnaijniX 72 37 0 22 anihC 12 56 41 663,2 XnosraeP 2 .11.0:ytilibaborP81.37:)26( ecruoS .esabataD: 43 The Political Management of Mayors in Post-Deng China Provincial Contrasts A set of dummy variables tests the hypothesis of systematic disparities in the way provincial party committees exercise political authority. Most of the published research on disparities refers to economic disparities (Hu and Wang 1996; Tan 1997; Wang 1995, 1997; Wang and Hu 1999). Here, provincial contrasts relate instead to a political logic. Assuming that economic disparities are reasonably accounted for by municipal performance indicators, provincial contrasts capture systematic differ- ences across provincial party committees. Recall that apart from CAMs and vice-provincial cities, it is the provinces and not the Centre that exercise direct nomenklatura authority over municipal leaders. Princi- ples of organizational discipline suggest that personnel policy is imple- mented uniformly across provinces. If the hypothesis holds that cadre management practices are similar across provinces, these contrasts should be irrelevant in the multivariate model. Conversely, if system- atic variations exist across provinces, these contrasts should exhibit both statistical significance and powerful substantive effects. At face value, differences in the political fate of mayors exist across prov- inces. In Xinjiang and Ningxia, mayors (who usually belong to ethnic mi- norities) never become party secretary, while in Beijing, Jilin, Shandong, Anhui and Guangdong, promotions occur in about 20 percent of cases, which is well above the national average of 14 percent. It remains to be seen whether these differences are statistically significant and reflect true behavioural differences across provincial party committees. I shall return to this point in my discussion of the ordered probit model. Time Contrasts Finally, annual dummy variables18 account for two distinct processes. They control for time-specific shocks that are not specifically encapsu- lated in the model. Furthermore, annual dummy variables 'purge' the stochastic term of possible biases caused by omitted time-dependent variables and reduce autocorrelation among error terms, a major pitfall of cross-section time-series models.19 Results I present three closely related versions of the multivariate model of mayor promotion. Model 1 estimates the odds of the ordered outcome (exitp 2 000.0 000.0 000.0 selbairavtnednepednI epyT scitsiretcarahCs'royaM sraey5-3eruneT .d 191.0- 360.0 *** 381.0- 260.0 *** 191.0- 460.0 *** +sraey6eruneT .d 694.0- 741.0 *** 384.0- 341.0 *** 394.0- 841.0 *** 5691-7291,htribforaeY .c 920.0 800.0 *** 520.0 800.0 *** 720.0 800.0 *** elameF=1,redneG .d 210.0- 381.0 010.0 771.0 310.0- 181.0 ytironiM=1,ytironiM .d 620.0 891.0 140.0- 151.0 610.0- 251.0 etaudarG .d 060.0- 361.0 910.0- 551.0 610.0- 851.0 noitacudEegelloC .d 010.0 031.0 630.0 221.0 810.0 521.0 .cudEegelloC.coV .d 950.0- 451.0 460.0- 541.0 120.0- 841.0 scitsiretcarahCytiC tnemtnioppAlartneC .d 294.0 072.0 * 133.0 312.0 - - ytiClaicnivorPytupeD .d 664.0- 603.0 862.0- 872.0 630.0- 102.0 noitalupoP )elpoep000,01( .c 000.0 000.0 000.0 000.0 000.0 000.0 * atipacrepPDG )0001BMR( .c 110.0 500.0 ** 310.0 700.0 * 110.0 600.0 * secivresfoerahS )PDG%( .c 600.0- 300.0 * 700.0- 300.0 ** 700.0- 300.0 ** DF ( )0001$SU .c 000.0 000.0 000.0 000.0 000.0 000.0 tsartnoClaicnivorP s gniqgnohC .d - - - - 263.0- 573.0 iebeH .d 261.0 481.0 130.0 001.0 - - ixnahS .d 950.0- 542.0 - - 012.0- 691.0 uggnemieN .d 030.0- 161.0 - - 261.0- 391.0 gninoaiL .d 422.0 071.0 380.0 021.0 - - niliJ .d 073.0 832.0 - - 822.0 981.0 gnaijgnolieH .d 210.0- 181.0 - - 261.0- 431.0 usgnaiJ .d 641.0 191.0 710.0 501.0 - - gnaijehZ .d 211.0- 991.0 532.0- 721.0 * - - iuhnA .d 933.0 102.0 * - - 671.0 341.0 naijuF .d 130.0- 502.0 261.0- 041.0 - - ixgnaiJ .d 480.0 552.0 - - 550.0- 202.0 45 The Political Management of Mayors in Post-Deng China :6ELBAT ).tnoc(setamitsEytilibaborPderedrO 1ledoM 2ledoM 3ledoM secnivorPllA secnivorPlatsaoC secnivorPdnalnI 0871 478 149 doohilekiL-goL 9.3541- 1.476- 6.167- X>p 2 000.0 000.0 000.0 selbairavtnednepednI epyT gnodnahS .d 013.0 912.0 371.0 951.0 - - naneH .d 922.0 702.0 - - 950.0 541.0 iebuH .d 070.0- 591.0 - - 332.0- 031.0 * nanuH .d 641.0- 081.0 - - 903.0- 001.0 ** gnodgnauG .d 013.0 902.0 791.0 441.0 - - ixgnauG .d 650.0 691.0 - - 690.0- 841.0 naniaH .d 181.0 071.0 070.0 901.0 - - nauhciS .d 382.0 491.0 - - 501.0 821.0 uohziuG .d 580.0- 071.0 - - 722.0- 180.0 *** nannuY .d 830.0- 552.0 - - 891.0- 602.0 ixnaahS .d 743.0 662.0 - - 981.0 722.0 usnaG .d 945.0 572.0 ** - - 493.0 162.0 iahgniQ .d 552.0 702.0 - - 011.0 761.0 tsartnoCemiT s 9991 .d 855.0 721.0 *** 245.0 621.0 *** 445.0 521.0 *** 8991 .d 396.0 631.0 *** 866.0 331.0 *** 176.0 231.0 *** 7991 .d 137.0 741.0 *** 107.0 541.0 *** 017.0 341.0 *** 6991 .d 974.0 341.0 *** 154.0 041.0 *** 954.0 041.0 *** 5991 .d 159.0 731.0 *** 719.0 531.0 *** 529.0 431.0 *** 4991 .d 769.0 751.0 *** 129.0 251.0 *** 449.0 451.0 *** 3991 .d 097.0 861.0 *** 447.0 561.0 *** 667.0 761.0 *** 2991 .d 446.0 071.0 *** 195.0 661.0 *** 126.0 761.0 *** 1991 .d 090.1 671.0 *** 620.1 071.0 *** 860.1 271.0 *** 0991 .d 378.0 181.0 *** 097.0 771.0 *** 258.0 771.0 *** sretemaraPyrallicnA 1tuC 633.65 56.51 *** 484.74 558.32 *** 672.15 481.51 *** 2tuC 404.85 66.51 *** 435.94 378.32 *** 233.35 891.51 *** elbairaVtnednepeDehtfonoitubirtsiDdevresbO )tixE(ytilibaborP 91.0 81.0 02.0 )eunitnoC(ytilibaborP 76.0 76.0 66.0 )nitomorP(ytilibaborP 41.0 61.0 31.0 etoN foesoprupehtrofroatadgnissimoteuddettimoerastsartnoclaicnivorpemoS.sraey-royamerasnoitavresbO: llA.ymmuda'd',elbairavsuounitnocasetoned'c'.0002sistsartnocemitehtfoenilesabehT.noitacifitnediledomgnirusne .01.0