经济绩效和省级领导人的政治升迁

薄智跃222次双周记录

主持人（盛）：今天我们非常荣幸的请到薄智越教授。薄教授在芝加哥大学获得了政治学博士学位，在许多国内外著名

刊物上发表了很多论文。他现任教于圣约翰·菲舍大学。他今天学术演讲的题目很有意思，在国内研究得还比较少。天则所今天请到薄教授来作这个演讲，表明我国现代研究领域已经推进到这个层次，即在对政治学及政府研究方面能不断的突破各种禁区，不断的往前推进。下面欢迎薄教授作演讲。时间为一到一个半钟头，然后是评议和讨论。

薄教授的发言我 1966年上小学，1976年高中毕业，是“浪费的一代”（笑）。在部队干了四年，80年从部队出来后，

81年上北大国政系。我在北大呆的时间比较长，85年本科毕业后又上了研究生。研究生毕业后留校教书教了两年。90年去美国留学，最初是在美国大学（American University），呆了一年后转到芝加哥大学政治系，95

年获博士学位。其后是在各个大学教书，比如罗斯福大学，芝加哥大学及美国大学。从 98年开始在圣约翰·菲舍大学国际关系系任系主任，我们设立了本科生和硕士研究生项目。我今天的演讲取材于我这本书（？），很可惜还没有中文版。人大的一些老师和博士很感兴趣，想把它翻译过来，这要和出版社商量， 另外这个题目比较敏感，翻译了能不能出也是未知数。这个题目的中文意思是“1949年以来中国省级领导人的政绩与升迁”，主要从四个方面进行大概的阐述：研究对象、研究动因、资料和研究方法、研究结果。

第一， 研究对象主要是省一级（省、自治区和直辖市）的领导人，包括四类：（1）省委书记（即文革时“第一书记”）；（2）省委副书记（即文革时“第二书记”以下）；（3）省长（包括自治区主席、直辖市市长和文革时的革命委员会主任）；（4）副省长（包括自治区副主席、直辖市副市长和文革时的革命委员会副主任）。 这本书本来只涵括了 1949年至 1998年的内容，但是最近我应国内外机构的要求又进行了进一步的研究，收录了一些最新的资料。要说明一点，我的研究对象中不包括省人大、省政协和省军区的领导，虽然省人大领导问题已经引起了一些学术关注，我在近期也就省人大领导和党政领导进行了一些对比研究，但是我的主要研究对象还是上述四类省级领导。

研究时间为 1949年到现在，分为三个时期：（1）1949－1966年，文革前；（2）1967－1977年，文革期间；（3）1978年至今，改革开放以来。

研究的主要内容为“升迁”问题（political mobility），具体包括：（1）提升（promotion），包括调升；（2）降职（demotion），包括被开除；（3）调动（transfer），主要指平调；（4）退休（retirement），这是改革开放以来即 80年代以后出现的新的现象；（5）停留原职（staying），即不升不降，没有动。

第二，研究动因，即为什么要研究省级领导。因为：（1）研究省级领导非常重要。首先，“省”在中国政治中作用举足轻重。尤其从 1949年以来特别是 1954年取消 6大行政区域后，“省”就成为中央下属的、直接与中央发生互动的层次。我国的省人口庞大，有好几个省如河南、山东、四川比欧洲的七大国家的人口都多。去年 11月美国的一个内部讨论会邀请了世界各地一些中国问题专家讨论十四大召开中国政局人事变动，绝大多数学者关注的是中央一级将要出现的变动，只有我是把关注点集中在省级领导上的。这表明这方面的研究还很欠缺。而中国是一个以省为主要建制的国家（a country of province）。（一个学者）如果不研究中国的省，那么可以说他对中国并不了解。其次，省级领导在中国政治中的作用非常重要。一方面在中共中央委员会中省级领导干部作为中央委员或候补委员的比例是相当大的（是主要成分之一）。加州大学圣地亚哥分校的一位教授于 1993年所著的《中国经济改革的政治逻辑》一书中提到，中共中央委员会是一个极其重要的核心机构，而省级领导在这个核心机构中起的作用非常大；另一方面在省级领导结构中，省级领导当然是起决定性作用。（2）国外对中国省级领导的研究也不足。最早比较系统的研究中国省级领导的是 Frederick Teiwes，他是哥伦比亚大学政治学系毕业的博士生，对中国省级领导非常感兴趣，曾写了两本专著研究这一方面，缺点是资料比较陈旧（多是文革时期的资料），不够准确和完整。在某种意义上他是开创者，但是有很大的局限性，主要是资料的不足和研究方法的某些不科学性（如局限于一般的推断）。他的两本专著是：“Provincial Party

Personnel in Mainland China, 1956-1966”(1967)；“Provincial Leadership in China: the Cultural Revolution and the

Aftermath”(1974)。后者写的是文革期间省级领导的变动，但他所说的文革仅到 1969年。另一个是中国学者，叫Huang Yasheng，现在是哈佛商学院的副教授。他的专著“Inflation and Investment Controls in China”(1996)，研究的是中央和地方尤其是与省一级领导之间的互动，涉及到了相关的问题，但是对于省级领导的政治升迁陈述的非常有限，篇幅很少。第三个比较重要的资料是 Peter Cheung、Jae Ho Chung和 Lin Zhimin所著的“Provincial Strategies of Economic Reform in Post-Mao China: Leadership, Politics and Implementation”(1998)。这本书采用个案分析的方法，主要研究改革开放以来 8个省的各界领导的改革举措及成效。因为仅仅选取了 8个省研究，所以并不具有很强代表性，是有局限的。

总体上说，（国外对中国省级领导的研究存在的）不足包括：（1）资料陈旧。如 Teiwes。（2）资料不够系统。如Huang 和 Cheung。（3）资料不完整。这既指其研究时间上的不完整，又指其仅限于某一时间某一部分领导，如仅研究省级第一把手。（4）对政治升迁（political mobility）缺乏研究。首先对省级领导的研究即所谓“精英研究”（elite studies）仅限于选用与构成的研究，是静态的，并不注重变动与发展。其次，缺乏对各省具体条件及领导政绩对其升迁的影响的研究。

我研究的特点是：（1）全面而系统。从 1949年开始逐年研究，涵盖每个省的四类省级领导人。（2）以“政治升迁”为研究重点。运用计量分析的方法，把“政治升迁”作为因变量（dependent variable）。（3）把升迁与政绩联系起来。（4）研究方法的科学性。采用量化分析方法。（5）1996年我发表了同一主题的文章，介绍我的研究成果，在相关学术领域引起一定的反响。 第三，资料和方法。资料包括：（1）省级领导。我运用了十个变量，如年龄、党龄、籍贯、性别、民族、原籍工作或在外地工作、与中央的关系等。我的资料库中收录了 1949年至今各省担任过省级领导职务的 2600多人的资料。（2）各个省市自治区的基本情况（包括人口、经济增长和向国家上缴的税收或国家给予的补贴）。这些资料在目前是最全面的。

我的理论模式，也就是研究角度和思维方式从一般而言是“精英研究”（elite studies）。（我认为）目前政治理论界存在的一些理论模式，不太适用。现在做一个简单介绍，然后指出其不足。（1）派系模式

（factionalism）。这在国外非常流行。最早的代表人物是 Jorgen Domes，他是加州伯克利的政治学教授，他把中国政治的派系分为两类：一类是“结构性”派系（structural factions），即指这些人在生活上有共同经历，比如是同学、同乡或同行；另一类是“功能性”派系（functional factions），即指在存在部门利益的各个部门基础上形成的派系，如军队与政府。哥伦比亚大学的Andrew Nathan教授对其理论框架提出了质疑，认为派系与正式组织之间的权力关系不同，所谓派系都是建立在个人关系（personal tie）上，而建立在部门之间的利益或者领导－被领导关系之上的是“正式权力关系”（formal power relationship），不属于派系。我在研究过程中倾向于第二种定义。用此种派系模式研究中国政治尤其是中国省级领导的政治升迁存有许多不足，比较牵强附会，比如分析建立在猜测和小道消息而不是系统资料的基础上，对许多现象无法解释（例如，许多同乡在政治上互相争斗而非同乡则相互提携）。所以我认为派系只是升迁的一个必要条件（necessary condition），而不是充分条件（sufficient condition）。派系模式不足以解释政治升迁。（2）地方主义模式（localism）。这也是一种较流行的研究地方领导的模式。Teiwes提出了一个基本公式，来说明是否存在地方主义，以及用比例说明地方主义的强弱。这个公式是：在原籍任职的领导与外地派来的领导之比，其比例大则地方主义倾向强，反之亦然。这里暗含的一个判断就是本地领导即在原籍任职的领导等同于地方主义。这个公式似乎尚无人提出挑战，影响比较大。又有学者提到了明清时期的“回避制度”（law of avoidance），认为中共在选派地方领导人时往往无视传统的回避制度。我认为这样的研究模式不够严谨，因为把地方主义（通常是贬义）与出生地主义等同起来了。我 96年对此问题作了一个专题研究，我认为地方主义与是否在原籍担任职务没有直接关系。另外，地方主义无法解释升迁的问题，这是其很大的缺陷。 我因此提出了不同于以上的“政绩模式”（performance model）。升迁受到各种因素的影响，其中非常重要的因素是政绩。这是我的一个假设（hypothesis）。具体内容包括：（1）若一个省经济发展好，则省领导得到提升的可能性大；（2）若一个省上缴国家税收多，则省领导得到提升的可能性大。

我的研究方法（methods）是芝加哥大学的“试验主义”，量化分析。以政治升迁为因变量，自变量则包括很多，政绩外还有党龄、年龄、性别、民族、学历、原籍甚至不同历史时期、有无政治运动、省的人口、经济发展状况、向国家上缴税收等。我的统计模式是“多项式选择模式（multinomial logic）”。这种模式取决于因变量，因变量可以是多项选择（如某一领导人于某一时间面临数种可能：升迁、降职、调往外地或者退休）。自变量分为两类：分类变量（categorical variable）和数量变量（numerical variable），两种变量都可以。统计结果是 RRR（relative risk ratios），即相对选择比例（率）。举个简单的例子，选择A或者选择 B，选择A的概率是 5.25％，选择 B的概率是 2.15％。这两个概率的比率 2.44即是 RRR。如果省长被提升的概率是5.25％，而省委书记被提升的概率是 2.15％，则 2.44意味着省长被提升的机率是省委书记的 2.44倍。

第四，研究结果。我选择三方面介绍。（1）学历和升迁。A、基本特点。加州大学伯克利分校的教授Hong Yung Lee出版了专著“From

Revolutionary Cadres to Party Technocrats in Socialist China”(1991)。这本书主要研究中国领导，特别是改革开放背景下，如何转型。他的资料非常丰富，提到了中国领导层的学历和教育程度。他的基本判断是中国的领导阶层是从没有文化的农民转变成有文化的知识分子。这种说法不够准确。我们研究的结果表明，从 49年至今的中国省级领导的文化程度和学历呈现“V”形，两头高中间低。文革前省级领导的文化程度较高，50％以上的省级领导受过大专以上的高等教育，文革期间急遽下降。原因一是中共领导者中也有很多是知识分子或小知识分子，受过一定的高等教育；二是解放时期的统一战线政策的作用，许多省级领导是原来国民党政府的官员或者进步民主人士。而文革时这些人都被打倒了，取而代之的是两类人：军队干部和造反派，相对文化程度较低。改革开放后，文化程度大幅提高，尤其是 95年“选拔任用党政领导干部暂行条例”（最近颁布了“正式条例”）的颁布实施，规定省级领导必须具备大学本科学历。我统计的数据显示：1983年主要省级领导中有大学本科学历的占总人数的 1/3，1993年是 2/3，2002年是 82％，甚至有 4个书记 5个省长是研究生毕业，其中

河南省省长李克强还是北大的经济学博士。有的省级领导干部有外国学历，如浙江省委书记张德江，是金日成综合大学经济系毕业的。没有大学本科学历的省级主要领导也占一定比例，如 19％的书记、16％的省长不具备这一条件。B、学历与升迁。学者Hong Yung Lee持“专家治国论”（Technocracy），从这一理论推出一个假设：学历越高被提升的可能越大。真是这样吗？我通过研究发现，学历与升迁并不直接相关，甚至大学学历是进一步提升的障碍，有大学学历的提升机率是没有大学学历提升机率的 60％。（说明一下，这是指进了省级领导这个层次之后进一步升迁的问题，与前述省级领导有大学学历的比例大幅升高的整体性论述是不矛盾的）。原因在于，中国的省很大，情况复杂，专家精通某一方面但往往缺乏基层工作经验和把握全局的能力，这是一个机会成本的问题。因此，我的结论是大学学历不是提升的主要条件，省级领导干部还称不上“专家治国”。

（2）地方主义与升迁。A、在原籍任职的基本特点。中国古代的回避制度自曾国藩建湘军始被废除，地方主义从此抬头。后来军阀混战是典型的地方主义。解放后，中央采用“以中央的地方势力来反对不忠于中央的地方势力”的方法，委派大批干部回原籍任职，如叶飞到福建，黄克诚到湖南，李先念到湖北。从 49年至今，约 1/3的领导在原籍任职。还有一个有趣的现象就是，一般而言省委书记是外地人，省长是本地人。这样搭配比较和谐，省委书记与中央联系密切，省长深谙当地情况。图形分布也是“V”形，文革前和改革开放后干部在原籍任职的比例较高，文革时较低。现在（2002年）比例又大幅下降，省长和省委书记中的 18％是本地人，其中省长中的 29％和书记中的 3％是本地人。B、地方主义与升迁。我们的假设是：在原籍任职者更有可能被降职。研究结果表明：原籍与外来在升迁上没有本质区别，但原籍比外来更少被调动。所以，我认为地方主义理论不成立，应当注意的是任期长短。意即，原籍并不是关键，而在于任期长短产生的“认同”心理（identity）。如果在一个地方任职很久，则与该地认同心理日强，关系日深，这比原籍更容易产生所谓的“地方主义”。

（3）政绩与升迁。我运用了经济增长率（包括省增长率与在任期间平均增长率）与向中央上缴税收（包括省缴税与在任期间平均上缴税）两个变量。我的第一模式是“对比”，即用原职不动（staying）和提升、降职比较，发现政绩越好更可能留任。我的第二模式研究显示：经济增长率与提升无关，但上缴税越多越能提升。因此我的结论就是：中央关注的是上缴税收而不是单纯的经济增长率。上缴税收越多越能提升，少交税则更可能被降职。

评议主持人（盛）：感谢薄教授的讲演。这是一个很新鲜的话题，因为在我们的习惯印象中，政治升迁是一

件很偶然的事，并没有一定的规则，而且要在政治领域中突破禁区寻找规律也需要很大的勇气。今天我们请来了两位评论人。一位是中国人民大学社会学系的周孝正教授，另一位是北京大学的赵晓博士。 赵晓：很高兴在天则经济研究所听到这样的报告。这虽是一个政治学的报告，但是也可看作经济学的报告。因为他的研究方法已非常接近经济学的研究方法，如建立模型，进行计量分析，然后得出结论。可以说当今国际范围内，政治学、经济学和社会学的边缘正在变得模糊。

我先说一下对这个题目的看法。目前中国转轨时期存在两大难题。我们的精力都集中在第一个难题上。第二个难题是，根据标准产权理论，只有产权清晰，经济才能有效率（如道格拉斯·洛斯在许多著作中都已阐述的）。但是中国经济产权并不清晰，却一直保持着高增长。为什么呢？另外，根据西方标准的公共选择理论，中国政治的公共选择模式是非标准的，是没有效率的，而在没有效率的公共选择模式下经济能高速增长是为什么？主要是两方面的原因。一是从一些经济学的文献对后一个问题所作的解释看——如张维迎分析了中国地方竞争如何促进中小企业的变革——地方为了提高经济绩效，获取更多的税收资源，积极推进了中小企业的变革。

二是中央强调以经济建设为中心，把经济绩效作为考察地方干部的前提。基于以上两点，在大的政治框架没有改变的前提下经济保持了增长。另外，我在中国“抓大放小”的案例中也发现：即使大的政治框架没有改变然而效率也能改变。因为以经济建设为中心，所以原来的政治官僚由技术型转为财经型，懂得经济建设的官僚更容易得到提升。但是这方面的研究无论是在经济学还是政治学领域都太少了。今天博教授提出的政绩模式，力图说明经济增长与省级干部升迁的一些内在关系，对于我国处于转轨期，公共选择模式又不标准的情况下如何促进经济增长能作出一些解释，从学术上看似乎可以算作这一支流。

对于研究方法，我没有不同意见。但是对于选取指标的真实性，我存在一些疑问。毕竟现在造假太猖狂了。另外对于你刚才提到的“任期长造成地方主义”的观点我认为还有讨论的余地。据我所知，国民经济统计核算对地方经济增长关注不大，主要是用税收的数据，因为统计部门认为其真实性高。还有就是从 49年至今时间跨度很大，应当分时期研究。特别是改革开放以后也能分出很多小的阶段，不同阶段对学历、专业的要求都不同。

薄：我书的第六章建立在第五章“基本模式”的基础上，阐述了三个阶段的情况，其中改革开放以后更是年年不同，有较细的分类。

周孝正：研究很有意义，虽然数据真实性有待考问，但是对于趋势的分析是对的。我认为应该人文型领导和技术型（侧重自然科学的）领导并重，提倡政治文明。另外可以与发达国家的领导升迁做一个横向比较。举个例子，比如中国历史上共有 300多个皇帝，分了 36个姓，但日本 2000多年就是一个姓，所以被神化，所以能不负责任了，而中国则改朝换代频繁（笑）。我们在编《国家资源报告》时曾讨论应该用何种指标和指标体系来考察我国的干部。国际上有人认为应对一定任期后该领导所在地区的生物物种多样性进行考察，作为核心指标。江总书记也说过“当今世界是丰富多彩的”，所以这种说法是有一定道理的。（笑）这是我的一点感想。

蔡：我认为，研究中国领导的升迁问题更重要的是研究模式的变化。从整个中国国情发展的过程所引动的人事制度的模式、干部升迁制度的模式的变化或者某一阶段国家领导层的侧重点转移的角度来摸索干部升迁的规律，可能更准确一些。对于干部个人资料的收集与统计当然也是必要的，但是通过对历史的回顾与思考，我认为如果能将今后的研究切入到干部升迁制度的变迁会触及问题的实质。薄教授远在美国，观察是远距离的，让人更多的感觉到一种理论上的灰色，理论与实际脱节比较厉害，因为生活在急遽变化，权力失去制约的市场经济滋生出大量的权钱交易，通过种种不正当手段获得官职和提升的现象已经扭曲了我们的干部提升制度。应当拉近观察的距离，深入本土，完全采取数理统计的研究方法太遥感了。中央一级赤裸裸的“跑官”比较少，因为受到的制约较多，但是省级这种现象比较泛滥。您的研究是纯学术的，但真正想要在世界政治文明的发展史上留下这一段历史的记录，必须转换模式的切入点，要更贴近中国的国情。比如，徐匡迪为什么突然从政治舞台上消退，他一直是很多人心目中下届总理的最佳人选，他的离去反映了我国干部选拔体制中存在的问题。

黎鸣：我怀疑你用纯理性的办法来研究中国政治会有什么有价值的成果。中国的政治是“人治”，它与理性主义偏差得很远。你把 49年至今的数据放在一起研究，可能出现许多相关因素相互抵消的情形。在中国很多升迁情况与政绩和上税多少没有必然的关系。即使按照相关计算得出相关性，但也是值得怀疑的。所以用纯理性的方法来研究中国政治，并不有助于中国政治的变迁，反而把其中一些重要的因素掩盖了。用纯理性的方法来研究政治我是很怀疑的。对经济学的研究也是如此。现在的经济学家就进入了这样一个巨大的误区。经济不是纯理性主义的。比如，对心理学的研究在凯恩斯理论中占据重要地位。又如，马克思和哈耶克都不仅仅是运用数字来建立理论的。我认为应当将研究的对象与理论密切结合，用纯理性的方法来研究与人、社会、政治、经济相关的现象则是进入了理论的误区，不会在学术上走得很远。

孟翔：相关性并不等同于因果性。相关性背后的原因还是要找到中国官僚升迁的机制，在这方面博教授没有下手。而且还很可能存在全然没有因果关系的情况，比如强省的领导因为上缴税收多而升迁机会大，但是他是如何到强省去担任领导的呢？很可能他是中央的亲信。因此这里升迁的原因是“任人惟亲”，恰恰不是政绩。所以仅从表象研究是搞不清中国的升迁机制的。

赵农：对，目前中国的干部升迁模式是只要得到他顶头上司的赏识就能提升。另外正如孟先生所说，有时候是为了提升一个干部才派他到强省去的。这是中国的国情，研究中必须触及。另外，我认为省委常委也是应当注意的一类省级领导类型，因为从一个副省长变为省委常委也是升迁。提到省级领导的学历，我认为存在原来学历很低但是担任领导职务后弄到大学学历的情形。

主持人（盛）：各位讨论的很激烈。我想说的是，薄教授的研究有一个既定的命题，各位提出的涉及面非常之广，已经远远超出了他这本书所研究的范围。刚才赵晓提到政治学的这种研究方法与经济学相近，我说它运用数据统计、回归、相关系数、显著性等，仅是经济学研究方法中的一种。刚才黎先生批评的应该是这样一种方法，但是经济学研究方法还有很多别的种类。我知道政治学学科中有政治经济学的门类，我建议薄教授在研究中应当加入制度分析的方法，比如分析正式及非正式的晋升制度，这可能使你通过数据分析得出的结论更加丰满。

另外，我对你的某些结论存有怀疑。比如你认为大学学历是进一步提升的障碍，你给出的“机会成本”的解释，不能令人信服。又如上缴的税收和经济增长某些时候可能存在负的相关性，浙江省义乌小商品市场年销售额 100亿（现在是 200亿），收的税是 6千万，税率仅 0.6％。现在很多地方往往隐瞒增长率就是要少缴税，税收高的地方恰恰就是经济增长差的地方。除以上两个因素外，“政治忠诚”也是一个重要的变量，它可通过税收来表达。税收不能太高，否则影响地方经济发展和安定团结，中央也不喜欢。中央和地方存在一个均衡问题，中央并不欣赏完全没有地方主义，不顾地方利益的领导。

还有一个问题值得注意，就是领导自身是否想晋升。为什么有些政绩特别好的领导留在本地拒绝晋升，例如凡是经济发达地区的省级领导都不很愿意调往中央。事实上，央行的许多领导都是从落后地区调来的。在中国有许多现实问题，可能在美国远距离观察是不太容易涉及到的。

蔡：我补充一个例子。一位广东省的市级干部对我说，在他们那里要想跟某位官员过不去就提升他，这样他就失去了人际关系等等既得利益，而且要提升就要考察，一考察他那些暗无天日的事情就容易暴露出来。（笑）

薄教授对评议的回应感谢大家提出宝贵的意见，这对我是一次很珍贵的学习机会。我在国外呆的时间比较长，对国内的许多

情况确实了解得不够，需要进一步学习和体会。关于制度的研究，比如国家干部管理制度，我在这本书里虽然没有涉及，但我也开展了一些工作，如 99

年为一本英文杂志《中国法律和中国政府》编译了一组文章，关于党政领导人选择和任用的基本条例等的介绍。我的这本书其实是提出了一个问题，作了一些基础性研究，由此也提供了一个供大家讨论的平台。

我认为研究省级领导的政绩要注重两个方面的情况：一是研究各省的情况；二是研究各位省领导的相应举措。这样可能比较全面和彻底。谢谢各位。

北京天则经济研究所版权所有。如您有何建议或问题，请与我们联系北京市海淀区紫竹院公园报恩楼二层，100044 　Tel:86-10-88413052　　Email:　unirule@mx.cei.gov.cn

Title: Economic performance and political mobility: Chinese provincial leaders.

Subject(s): CHINA -- Politics & government; POLITICAL scienceSource: Journal of Contemporary China, Jul96, Vol. 5 Issue 12, p135, 20p,

7 charts, 1 diagramAuthor(s): Bo, ZhiyueAbstract: Focuses on the political mobility of Chinese provincial leaders.

Personal characteristics; Provincial characteristics.AN: 9608225137

ISSN: 1067-0564Full Text Word Count: 8675

Database: Academic Search PremierPersistent Link to this

Article: http://search.epnet.com/direct.asp?an=9608225137&db=aph

* * *

Chinese provincial leaders, unlike their counterparts in a democratic system, are not elected but selected. Hence the criteria by which the center uses to select and retain provincial leaders would be critical for the political mobility of provincial leaders. As China has been a developmental state since 1949, it is not unreasonable to expect that the center would retain or promote or demote provincial leaders according to the economic performance of their provinces. To test the hypothesis, I conducted multinomial regressions with a data set of Chinese provincial leaders between 1949 and 1994 in thirty provincial units. As the results show, the political mobility of provincial leaders is determined not only by the political movements of the PRC but also by the economic performance of the provincial leaders. The worse the economic performance record the more likely the provincial leader will be demoted. Moreover, the revenue contributions of the province during the provincial leader's tenure are also a determinant of the political mobility of the provincial leader.

Introduction

Chinese provincial leaders are a group of important players in Chinese politics. They form the largest bloc in the central committee of the Chinese Communist Party (CCP), and they constitute a major link through which the center implements its social, economic, and political policies. Yet these powerful leaders do not operate without constraints. They are the central government appointees, and their political mobility is dependent on their superiors in the center. Studies of Chinese provincial leaders are abundant,[1] yet few have

ECONOMIC PERFORMANCE AND POLITICAL MOBILITY: CHINESE PROVINCIAL LEADERS

systematically analyzed the political mobility of Chinese provincial leaders. Those who studied Chinese provincial leaders have paid almost exclusive attention to the recruitment and composition of political leaders in China. But for an individual, admittance to the club of political leaders is not the end of a political career. What happens afterwards may reveal much about the dynamics of the political system. Hence this study starts where most studies end. Moreover, existing studies of provincial and municipal leaders have paid little systematic attention to the impact of provincial conditions and performances on mobility. This study seeks to fill this vacuum.

Data and method

Because until recently the PRC government did not publish the rotations of officials systematically, one had to scan the PRC media to 'identify' which person holds what post. This method of 'positive identification'[2] has been used by Wolfgang Bartke[3] and David S. G. Goodman to great effect.[4] But, as Donald W. Klein pointed out, the problem with this method is that the 'earliest-identification' date in the various directories may be months or even years off the mark',[5] i.e. the appointment date. Often a provincial leader did not appear in the media until long after his or her appointment. For instance, Buhe, a son of Ulanhu, was appointed the Governor of Neimenggu (Inner Mongolia) in December 1982, but he was not identified by the media as such until four months later.[6] Chi Biqin was a deputy secretary of Beijing between August 1966 and April 1967, but this totally escaped the media.[7]

In the last few years, however, provinces began to publish systematic data on their leaders, making it feasible to revisit an old topic. Two book series on provincial leaders and provinces have been published. One is Dangdai Zhongguo [Contemporary China], which includes lists of provincial leaders from 1949 to 1987-1989. This series is supplemented by another series entitled Shengqing [Provincial Data].[8]

These data are superior because they provide appointment dates instead of identification dates. With these data we can more confidently reconstruct the political career of provincial leaders. We then can see to what extent the career of a provincial leader is affected by local conditions. There are thirty provincial units in China excluding Taiwan (twenty-two provinces, five autonomous regions, and three centrally administered municipalities). Out of the data on the thirty provincial units, we have produced a set of data of Chinese provincial leaders with 19,627 cases.[9]

The major tool in this study is multinomial logit analysis. This is because the dependent variable--political mobility--in the study is a discrete variable with several options, and the independent variables include both discrete and continuous variables. The results are relative risk ratios (RRR) which are ratios of probabilities of different features. A simple reading of the coefficients is that, if RRR is less than 1, it can be understood as a coefficient of a negative sign; or, if RRR is greater than 1, it can be seen as a coefficient of a positive sign.

The political mobility of provincial leaders

The PRC provincial leaders[10] in this study include: provincial party secretaries; provincial deputy secretaries; governors; and vice governenors.[11] The political mobility of Chinese provincial leaders refers to five kinds of movement (or non-movement). The first is promotion. A provincial leader is considered promoted if he or she moves from a lower rank to a higher rank within the party or government apparatus, or from a government position to

a party position of the same rank (Figure 1). A deputy secretary is promoted if he or she is appointed governor or secretary. A governor is promoted if he or she becomes a secretary. The second kind of political mobility is demotion. This, in fact, includes both demotions and purges or dismissals. The third is lateral transfers with no changes, in rank. The fourth is retirement. This is particularly true of provincial leaders since the 1980s.

Table 1 shows that the Chinese provincial leadership has been very stable.[12] Out of 19,627 cases, 16,093 (or 82%) are cases of 'continue', i.e. staying in present positions.[13] This image of stability, however, has been shattered three times between 1949 and 1994. The first serious blow came in 1967 when 97% of the provincial leaders were purged at the beginning of the 'cultural revolution'. The second major change occurred in 1983 when a majority of the provincial leaders were retired to make a way for younger and more energetic leaders under the reforms. The third shift came in 1993 when the center restructured the provincial leadership as part of the political transition preparing for the post-Deng era.

It is surprising that the provincial leadership during the 'cultural revolution' period was the most stable, with more than 90% of provincial leaders staying where they were. This contradicts the conventional wisdom which has portrayed the 'cultural revolution' period as one of turbulence when one's political career was very unpredictable. Once one survived the first blow of the 'cultural revolution' or entered into the provincial leadership at the beginning of the 'cultural revolution', one would very likely survive in the following years. The provincial leadership of the economic reform era, on the other hand, was the most mobile.

Over 1949-1994, only a small number of provincial leaders experienced upward mobility. Out of 19,627 cases, only 519 (or 3%) could be identified as cases of promotion. It has been difficult for provincial leaders to climb the ladder of success further because they are already quite high in the hierarchy. The distribution of opportunities over the years, however, has not been even. In some years there were almost no promotions, while in others the proportion of provincial leaders who were promoted surpassed 10%. In the 1950s, 2-5% of the total provincial leaders got promoted each year. In the following years, however, the chance for further promotion became smaller, less than 1% in most years. Since the 1980s, however, the chances for upward mobility have increased dramatically. Every few years, there has been a year of promotions for more than 5% of the provincial leaders. This has probably resulted from the Party's effort to institutionalize the cadre system. As older leaders gave up their positions, the opportunities for younger leaders dramatically increased.

Lateral transfers among provincial leaders are even more rare than promotions. There are only 541 (or 2.76%) cases of lateral transfers out of 19,627 cases. There are several big waves of lateral transfers, nonetheless. The first wave lasted from 1952 to 1954 during which 61 provincial leaders were transferred. Several factors may explain this first wave. First of all, provinces were in a formative stage in those years and some provincial leaders who had worked in established provinces were moved to newer provinces. Second, the center may have been trying to reduce the power of certain regional leaders by transferring their followers out. Thirdly, the center also transferred some provincial leaders to the central government. From 7 to 13% of the total provincial leaders was transferred each year during this period.

The second largest wave of lateral transfers occurred over 1977-1979. In absolute numbers these years witnessed even more lateral transfers. More than 30 provincial leaders were transferred each year, totalling more than 100. These numbers represent about 5% of the

provincial leadership each year. This wave was directly related to the reshuffle of provincial governments in favor of the rehabilitated leaders in the center.

The third wave took place immediately after 1989 when the center decided that local governments had too much independence and transferred provincial leaders among provinces. In 1990, for instance, three governors were exchanged between provinces. Li Changchun of Liaoning was transferred to Henan, Yue Qifeng of Hebei to Liaoning, and Cheng Weigao of Henan to Hebei. Transfers to the center also became frequent during this period.

For Chinese provincial leaders, demotions and purges are more likely to occur than either promotions or lateral transfers. Out of 19,627 cases in our sample, there are 1,458 cases of demotions or purges. This represents about 7% of the total. Among these 1,458 cases, 114 cases are governors, 103 secretaries, 774 vice governors, and 467 deputy secretaries.

The most massive demotions or purges occurred in 1967, when 96% of the provincial leaders were purged. This is followed by 1977, when 32% of provincial leaders lost their posts. In 1958, many provincial leaders were also demoted or purged (12%). The reasons for these three waves are well-known: 1967 was the beginning of the 'cultural revolution' at the provincial level; 1977 witnessed the return of those who had been purged at the start of the 'cultural revolution'; and 1958 was the year of the 'great leap forward'. The victims of these years, thus, were 'capitalist roaders', the followers of the Gang of Four, and rightist provincial leaders, respectively.

For a long time in the PRC's history, there was no retirement system for high-ranking officials.[14] Veteran revolutionaries did not leave their positions until they died or were purged or promoted. This mentality has been described as one of 'only going up but not down, (neng shang bu neng xia). For this reason, upward mobility for young and ambitious officials of lower ranks was blocked. Moreover, many provincial leaders who had been purged during the 'cultural revolution' recovered their positions in the late 1970s. They saw the 'cultural revolution' as nothing but an interruption of their political careers. Yet these rehabilitated provincial leaders were much older by the late 1970s. In 1965, the average age of provincial leaders was 55. In 1980, the average age of provincial leaders had risen to 62, and many of them were well over 70. As China shifted its focus from politics' to economics in 1978, more energetic and better-educated provincial leaders were needed. To make way for a younger generation of provincial leaders, the Chinese central leadership decided to institutionalize a system of retirement.

As the veteran leaders were reluctant to leave the power structure entirely, some transitional measures were adopted. First, like central leaders who left the formal structure to join the Central Advisory Commission, provincial leaders 'retired' to similar commissions at the provincial level. This was politically consequential. It is well-known that the Central Advisory Commission members intervened in the formal power structure and were instrumental in purging general secretary Hu Yaobang in 1987. Local advisory commissions may have played a similar but less intrusive role at the provincial level. Second, many veteran provincial leaders 'retired' from their positions in the party or government to work in provincial People's Political Consultative Conferences or provincial People's Congresses. Finally, retiring provincial leaders left the power structure completely but retained all benefits and perks.[15]

Retirement age from office is rank-based. Generally speaking, male cadres are supposed to retire at the age of 60, and female cadres at 55. But if a cadre is a vice governor or deputy secretary, he or she may retire at 60. Provincial secretaries and governors retire at

65.[16] We classify all these cases (as well as those who died in office) as 'retirement'. Before 1979 there were few who actually retired. From 1979 to 1994, a system of retirement was initiated and institutionalized. 1983 was a watershed. Before 1983, few veteran provincial leaders were retired. In 1980 and 1981, only 2-5% of provincial leaders were retired. In 1983, however, more than 60% of provincial leaders were retired. Since 1983, retirement of provincial leaders has been more or less institutionalized and three waves have occurred in 1985 (20%), 1988 (15%), and 1993 (20%). As retirement proceeded, the average age of provincial leaders has dramatically declined.

To summarize, Chinese provincial leaders are the most likely to stay in the same positions in the same provinces. When they do exit from their positions, the most likely causes are demotions or purges. Next comes retirement, then promotion. Transfers are rare.

Personal characteristics and political mobility

The differences in patterns of provincial leaders' mobility may be explained by three groups of variables. The first are social and political movements. Provincial leaders fared worse in certain years of political movements than in other 'normal' years.

The second group of variables is personal characteristics. These include the following (Table 3): (1) educational background (variable edu2); (2) party membership (variable nonCCP); (3) number of years of being a CCP member (variable paryears); (4) age (variables age49 through age90);[17] (5) gender (variable female); (6) nationality (variable nation2); (7) native or outsider (variable homeprov); (8) central or local origins (variable center); and (9) specific offices (gov through depsec).

Education

In traditional China, education was the ladder to success. After 1949 however good educational backgrounds were no longer assets for upward mobility. During the Anti-Rightist movement and the 'cultural revolution', good educational backgrounds even became liabilities.

Since 1978, however, education has become an important criterion for selecting cadres and thus has become a major asset for upward mobility. Scholars of China studies have found that the composition of political elites has experienced a transformation 'from revolutionaries to bureaucratic technocrats' (Hong Yung Lee, 1991). Here we want to examine whether education is a determinant of the political mobility of provincial leaders.

Age

Age may have an impact upon mobility. In American politics, for instance, there is a minimum age requirement for presidential candidates. In PRC, for most of the period we examine, there is neither minimum nor maximum age requirements for entrance to provincial leadership. The question here is not the age requirement for provincial leaders but the impact of age on political mobility. Were older leaders more likely to get promoted than younger ones? Does the impact of age vary from one historical period to another?

Party membership

Provincial leaders are not all communist party members. Some started as non-communists

and then joined the party later on. Wu Han, former vice mayor of Beijing, for instance, started out as a non-communist in 1949 but was admitted into the party in 1957.[18] Do communist provincial leaders have more chance for upward mobility than non-communists? Do non-communists suffer more than communists?

In our sample, 970 cases (7%) out of 13,217 with known party status are non-communist provincial leaders. The majority (744) assumed their positions during the first period while only 30 could be found during the second period. During the third period, there were 196 non-communist provincial leaders. We expect that they did worse in terms of political mobility than their communist colleagues.

Party-years

Party-years is a measure of political seniority, which is a more important seniority measure than age. Being a long time member of the party, however, could have diverse implications for political mobility. One may be able to establish extensive personal connections and thus is more likely to get promoted. One may also have more chances to encounter conflicts with central leaders as well, increasing the likelihood of being demoted. Being long in the party may also mean that a person may be old enough for retirement during the third period.

Home province advantage?

It is well-known that presidential candidates in the US have a vote advantage in their home states over 'outsiders' because the home state favors its native son over his out-of-town rival.[19] In traditional China, however, officials were forbidden to serve in their home towns and even an adjacent town within 500 li (155 miles) in any direction.[20] In the PRC, the party not only ignored this rule but even sent officials to their home provinces. The question is, do those leaders who serve in their home provinces have any advantage in political mobility?

Gender

We know that Chinese women are extremely underrepresented in the provincial leadership, accounting for only 3% of the provincial leadership. Do these female elites fare better than their male colleagues in political mobility? Or are they a disadvantaged group? In a male-dominated bureaucracy, we expect that female leaders would do less well than their male colleagues.

Minority leaders

Minorities are over represented in the provincial leadership. Minorities are 6% of the total population, but minority provincial leaders account for 9% of the provincial leadership (1,790out of 19,264). Do these minority provincial leaders do better than their Han counterparts? What are their prospects for further promotions? Are they equally likely to be demoted as their Han colleagues? Or are they more vulnerable to purges because of their minority origins?

Since minority provincial leaders may have to overcome language and cultural barriers to work in the center and elsewhere, we would expect that minority provincial leaders are less mobile than Han provincial leaders.

Central or local origins[21]

One group of provincial leaders stand out in their relationship with the center. They usually start in the center as ministers or vice ministers (or in a similar rank) and then are sent down to provinces to gain some local experience before being promoted back to the center.

Among these people, we find: Zhu Rongji, former mayor of Shanghai and now first vice premier; Hu Jingtao, former party secretary of Tibet and now a member of the politburo standing committee; and many others. We label this small group of provincial leaders as people of central origins. In contrast, the majority of provincial leaders thrive and wither where they work and are labelled for their local origins. We expect that these two groups of provincial leaders have different patterns of political mobility. Provincial leaders of central origins are more likely to be promoted because they have stronger connections with the center. But those of central origins may also be more likely to be demoted since they may not perform well due to their ignorance of local conditions.

Provincial characteristics and political mobility

The third group of variables pertain to provincial characteristics and performances. They include: (1) provincial population (variable pop); (2) provincial income (variable income); (3) provincial economic performances (variables growth and may); and (4) provincial revenue contributions (variables revenue and revave).

Population

Provinces of different population sizes differ in central representation. The bigger the population, the larger the representation.[22] In this sense, population size is an indicator of provincial political weight in China. The question is what is the impact of population size on the political mobility of provincial leaders? Do leaders of more populous provinces, for instance, have more chances for upward mobility than those of smaller provinces? Or are they more likely to be demoted than their colleagues from smaller areas?

Because we are interested in cross-sectional variations holding time constant, we use the percentage of one province in the total population of a particular year to indicate the relative standings of different provinces across time. Because populations are political resources, we would expect that leaders of bigger provinces have certain political advantages over those in smaller provinces.

Provincial income

Different provinces of China are at different stages of development. Some provinces, such as Guangdong, Jiangsu and Zhejiang, may be quite advanced, while others such as Qinghai and Ningxia are still developing. Both total provincial income and provincial income per capita count in the national politics of China. In fact, total provincial income is probably even more important than provincial income per capita in determining directions of the political mobility of provincial leaders because the more aggregate provincial income a province has, the more leverage the province may have over the total national income. The national government may thus treat leaders of rich provinces differently from those of poor provinces. Provincial income per capita, on the other hand, is less important because it does not have direct impact on the welfare of the central government. However, we would expect that provincial leaders from less developed provinces to have more incentive to

change their places than those in more prosperous provinces.

Economic performance: economic growth

Economic growth (variable growth) measures provincial economic performance. Average growth rate (variable may) measures the provincial economic performance during the tenure of the provincial leader.[23] The former is yearly based and varies from province to province, while the latter is more individually oriented and varies from one person to another in the same province. Since personnel changes may occur any time of a year, we use the growth rates of current years if mobility occurred in the late months of a year (beyond June) and those of the previous year if changes took place in the early months of a year (before July).[24]

Although the economic growth of a province may have some impact on the political mobility of provincial leaders, the economic record of a provincial leader may be more relevant. In other words, we would expect that a provincial leader who has a better record of economic performance may be more likely to get promoted than a provincial leader whose province is doing well in a particular year. This is because the economic record is an indicator of ability while the good performance for one year may be the result of sheer luck.

Economic performance: revenue contributions

Provincial financial contributions (variable revenue) measure how much a province contributes to the central coffer in a year (the contribution is negative when the province gets a subsidy). The average revenue contribution (variable revave) measures how much a province contributes to the central coffer during the tenure of a provincial leader. As in the case of growth, we will look at the impact of these measures according to which month the change occurred. Again, we would expect provincial financial contributions to be relevant, but the records of financial contributions of provincial leaders may be more important in determining the orientation of their political mobility.

Findings

Education

Surprisingly, even during the third period, college education still does not make any difference with respect to the political mobility of Chinese provincial leaders. As Table 4 shows, most coefficients for education are not statistically significant. Provincial leaders were not transferred, demoted or retired by educational level. An important exception is that provincial leaders with college education were less likely to be promoted than those with no college education by a factor of two. This contradicts with the common sense that, during the third period, college-educated people were more likely to get promoted than those with less education. There may be two reasons for this phenomenon. On the one hand, college education is a credential for the entrance to the leadership, but not for further upward mobility. On the other, investing a lot of time in education may have cost those provincial leaders opportunities to make connections.

This finding is important. It shows that recent studies of Chinese elites overestimated the role of educational backgrounds in the political mobility of political elites. A good education, as mentioned above, may be a ticket to the officialdom during the third period, but not a credential for further political mobility. Indeed, provincial leaders with college education are

less likely to be promoted than those with no college education.

Basic determinants of political mobility

Since education is not a significant determinant of the political mobility of provincial leaders, a baseline model can be established without this variable. The results of this baseline model are reported in Table 5. The political mobility of Chinese provincial leaders, as Table 5 shows, is determined by social and personal variables. In terms of upward mobility, several factors are very important. First, a provincial leader is about three times as likely to get promoted during the third period as before. Second, a communist provincial leader is about ten times as likely to get promoted as a non-communist. Third, governors are more likely to get promoted than other types of provincial leaders. Fourth, provincial leaders of central origins are more likely to get promoted than those of local origins. One is also better off if one is a Han Chinese, male, and older.

In terms of lateral transfers, several variables make differences. Senior provincial leaders (such as governors and secretaries) are generally more likely to be transferred than junior provincial leaders (such as vice governors and deputy secretaries), probably because the center does this to ensure its control over the provinces. Minority provincial leaders and those who serve in their provinces, however, are both less likely to be transferred. Minority provincial leaders are less likely to be transferred due to cultural barriers, while natives are less likely to be transferred because of their rich local experiences.

Several factors account for downward mobility. During the third period, provincial leaders were much less likely to be demoted or purged than before. Junior provincial leaders, not surprisingly, were less likely to be demoted than senior provincial leaders because they are less visible. Provincial leaders of central origins are twice as likely to be demoted as those of local origins. Central experience does not help in making a good provincial leader. Those from the center may have conflicts with local leaders, and they may not perform well because of their ignorance of local conditions. Younger provincial leaders are also less likely to be demoted than older provincial leaders.

Not surprisingly, older provincial leaders are more likely to be retired during the third period than younger ones. What is interesting about retirement is that minority provincial leaders are less likely to be retired than Han provincial leaders and that non-CCP provincial leaders are less likely to be retired than CCP provincial leaders. Minority provincial leaders are less likely to be retired because they are scarce resources, while non-CCP provincial leaders are less likely to be retired because they are mostly new entrants. Moreover, it seems that junior provincial leaders are more likely to be retired than senior provincial leaders. This finding, however, is biased because leaders of different ranks have different age requirements for retirement and the likelihood that junior provincial leaders were retired was greatly overestimated. Controlling for different ages for different provincial leaders, we find no statistically significant differences between governors and vice governors and deputy secretaries.

Performance and mobility

Although the above determinants are important, they can not fully explain the dynamics of the political mobility of Chinese provincial leaders. A full account has to include provincial conditions and provincial performance. The results of the full model are reported in Table 6. Clearly, provincial conditions and performances do have impact on the political mobility of provincial leaders. First, as expected, provincial leaders of bigger or richer provinces are

more likely to get promoted than those of smaller or poorer provinces.[25] It seems that both political and economic resources count in upward mobility for provincial leaders in China.

Second, individual-oriented performance indicators account for the political mobility of Chinese provincial leaders better than provincial performance indicators. Those provincial leaders whose provinces do well economically are not statistically significantly different from those whose provinces do not do well. A minor exception is that those whose provinces did well in a particular year were more likely to be retired than those whose provinces did not do so well. This may be because in a year of abundance the center can afford to devote attention to retirement issues. In contrast, two individual-oriented performance variables (performance record--may and revenue record--revave) both generated significant results. Provincial leaders whose provinces grew well or contributed a lot to the central coffer during their tenure are less mobile than those whose provinces grew less well or contributed less to the center. It seems that the center likes to retain in their original positions those who had good performance or revenue records.

Finally, three municipalities under central control stand out in terms of political mobility.[26] Leaders of these municipalities are significantly more likely to get promoted than those of other provincial units. The leaders of Beijing, for instance, are more than twice as likely to get promoted as those of other provincial units. This is also the case for the leaders of Tianjin. When population share is controlled, the leaders of Shanghai are more than seven times as likely to get promoted as those of other provincial units (Model 2). The effect of Shanghai is outstanding. Many former or current leaders of Shanghai are among the most important players in national politics. Jiang Zemin, Zhu Rongji, Wu Bangguo, and Huang Ju are all politburo members. Jiang Zemin is the general secretary of the party and president of the state. Zhu Rongji is a member of the politburo standing committee and first vice premier in charge of economic affairs. Wu Bangguo was recently elected vice premier and put in charge of industrial construction in April 1995. Huang Ju is now party secretary of Shanghai and a member of the politburo. These is much talk in the mass media about a 'Shanghai clique'. It is evident that Jiang Zemin has promoted a lot of his followers from Shanghai since the late 1980s, and this makes Shanghai an outlier.

What is interesting about Shanghai, as well as other municipalities, however, is that its leaders are also more likely to be demoted than those of other provincial units. Shanghai leaders, for instance, are more than four times as likely to be demoted as those of other provincial units. Leaders from Beijing and Tianjin are more than twice as likely to be demoted as those of other provincial units. This indicates that being a leader of municipalities under central control also runs a bigger risk of being demoted. The leaders of Shanghai are also more than four times as likely to be retired than those of other provincial units.

These findings show that municipalities under central control, Shanghai in particular, are dramatically different from other provincial units in terms of their leaders' political mobility. The leaders of these cities are generally more mobile than those of other provincial units. They are more likely to be promoted and also more likely to be demoted. In the case of Shanghai, the leaders are also more likely to be retired.

Discussion

Performance and mobility

It is not surprising that those with better economic performance or revenue records are less likely to be demoted or retired because of their better performance. It is difficult to understand why these people are also less likely to be promoted.[27] To find out what is going on, we reran the regression (Model 2 of Table 6) with different comparison groups. As a result, we found that provincial leaders with better performance records are different from those with better revenue records (Table 7). With promotion as the comparison group (Model 2.1 of Table 7), we find that economic performance records make no difference in terms of demotions, while revenue records do. In other words, those with better revenue records are more likely to be promoted than to be demoted.

These findings seem to show that whether a provincial leader is to be promoted or demoted is not dependent on his or her economic performance record. Provincial leaders with better growth records are as likely to be demoted as to be promoted. But given their economic performance records, the provincial leaders with better revenue records were more likely to be promoted than to be demoted. In other words, the center cares more about revenue contribution than about economic performance.

As for retirement, provincial leaders with better economic performance records are also different from those with better revenue contributions. The former are more likely to get promoted or to be demoted than to be retired, while the latter are more likely to be retired than to be demoted. In other words, a better revenue contribution record may secure the provincial leader a nice arrangement after all, while a better economic performance record may leave the provincial leader's career uncertain in the end.

Conclusion

What have we learned so far? To what extent does this study contribute to our understanding of Chinese politics? In the following, we summarize the major findings of the study and explore their implications.

The chinese political system is a centralized system

This is nothing new about the Chinese political system, but it sheds light on central-local relations. Since provincial leaders in China are controlled by the center, their political mobility depends on what criteria the center adopts to manage provincial personnel.

This finding implies that although Chinese provincial leaders are diverse in terms of personal experiences and characteristics, they do not represent independent interests. Since their political careers depend on the center instead of their provincial general public, they are very unlikely to challenge central policies due to 'local interests'. In this sense, pluralist models which are products of analyses of democratic systems are largely nonapplicable to Chinese politics.[28]

Chinese provincial leaders had different fates during different historical periods

In the history of the PRC, there are three distinctive periods: 1949 to 1967; 1968 to 1977; and 1978 to the present. Provincial leaders had difficult times during the first and second periods because they were targets in numerous political campaigns. During the third period, however, provincial leaders were able to leave their posts without losing face. They retired to take positions in provincial People's Congresses or provincial Political Consultative

Conferences. Or they left the power structure with great retirement benefits. At the same time, younger leaders have enjoyed greater chances for upward mobility. During this period, no evidence suggests that provincial leaders were massively demoted or purged in the wave of the purges of certain central leaders. Hu Yaobang's departure as general secretary in 1987 had no effect on the downward mobility of provincial leaders. The same is true of Zhao Ziyang's purge as general secretary in 1989.

These findings show that the appointment system at the provincial level has been largely institutionalized during the third period. Since provincial leaders retire at the retirement age, it becomes relatively predictable who will leave provincial leadership in which year. This should direct our attention more to the institutional aspect of the Chinese political system than to the personal aspect of the system.

Chinese politics is basically the politics of big provinces

In China although provincial leaders of different provinces have the same ranks,[29] leaders of big provinces are usually considered more powerful than those of smaller provinces. Big provinces such as Sichuan, Guangdong, and Shandong are major producers of the Central Committee (CC) members or even politburo members.

This finding directs our attention to provincial variations and indicates the importance of big provinces in Chinese politics.

Performance counts in the era of economic reforms

In the era of economic reforms, this study shows, performance counts in the political mobility of provincial leaders. Provincial leaders who had better economic growth records or revenue contribution records during their tenure are less likely to be demoted. They are also less likely to be retired. In other words, those who did not perform well during their tenure are more likely to be demoted or retired.

Revenue contributions, it should be noted, are more important than economic performances in terms of promotions for provincial leaders. Those who contributed a lot to the central treasury are more likely to be promoted than to be demoted, while economic growth records make no difference. The center cares more about revenue contributions than about economic growth records.

All these findings direct our attention to structural factors. In studies of communist systems, it is not uncommon for scholars to overestimate the role of personal factors in the political mobility of elites. For politics in these political systems is usually perceived as politics behind the scenes. Speculations or even fabrications of personal connections between elites are thus inevitable. Although this study does not provide full support for the performance model, it does show that performance counts in the political mobility of provincial leaders.

The leaders of Shanghai are a distinctive group

The leaders of Shanghai are much more likely to get promoted than those of other provincial units. Three reasons may account for this difference. One is that Shanghai is an economic powerhouse in China. As richer provinces produced more promotions than poorer ones, Shanghai was no exception. Second, Shanghai has been the most important

contributor to central revenue since 1959. Third, since 1989, Jiang Zemin has promoted a great number of his followers from Shanghai. Being a leader in Shanghai, however, one also has greater chance to be demoted or retired. This indicates that Shanghai is a more competitive place, and its retirement system is also more institutionalized.

These findings show that any understanding of Chinese politics may be inadequate without some understanding of the role of Shanghai in the system. In the history of the People's Republic of China, 'Shanghai clique' experienced three rises. Before the 'cultural revolution' one finds Chen Yi and Ke Qingshi in the politburo. During the 'cultural revolution', Zhang Chunqiao, Yao Wenyuan, and Wang Hongwen (three of the Gang of Four) were among the standing members of the politburo. Now Jiang Zemin, Zhu Rongji, and Wu Bangguo are general secretary, the first vice premier, and vice premier, respectively.

In sum, this study shows that in a centralized system, the political mobility of provincial leaders is also related to the economic performance of provincial leaders in an era of economic reforms. This is not because provincial leaders had to win votes from their constituents in provinces by 'delivering the goods'. Instead, provincial leaders had to perform to win recognition from central leaders.

[1.] These include Frederick C. Teiwes (1966,1967,1971 and 1974), Victor C. Falkenheim (1972), Lynn T. White, III (1972), Donald H. McMillen (1979), David S. G. Goodman (1984), and Ezra F. Vogal (1989). For a recent literature review, see Jae Ho Chung, 'Studies of central-provincial relations in the People's Republic of China: a mid-term appraisal', China Quarterly 142, (June 1995), pp. 487-508.

[2.] David S. G. Goodman, China's Provincial Leaders, 1949 1985. Volume 1: Directory (Atlantic Highlands, NJ: Humanities Press International, 1986), p. vi.

[3.] Wolfgang Barkte, Who's Who in the People's Republic of China (Armonk, NY: M.E. Sharpe, 1981), p. vii.

[4.] Goodman, Volume I: Directory, op. cit., p. iii.

[5.] Donald W. Klein, 'Sources for elite studies and biographical materials on China', in Robert A. Scalapino, ea., Elite in the People's Republic of China (Seattle: University of Washington Press, 1972), p. 621.

[6.] Goodman, Volume 1: Directory, op. cit., p. 76. See also Directory of Chinese Officials: provincial organizations 1984, p. 122.

[7.] Ibid., p. 207.

[8.] These two series are further supplemented by two compilations of provincial data. One is Zhongguo Shengshi Zizhiqu Ziliao Shouce [Handbook of Materials of Provinces, Municipalities, and Autonomous Regions of the People's Republic of China] (Beijing: Zhongguo Tongji Chubanshe, 1990) and the other is Quanguo Gesheng Zizhiqu Zhixiashi Lishi Tongji Ziliao Huibian, 1949-1989 [Historical Data of Provinces, Autonomous Regions and Municipalities of the People's Republic of China, 1949-1989](Beijing: Shehui Kexue Wenxian Chubanshe, 1990). The former compiles information about provincial leaders (1949-1989) and the latter gives information about provincial economic performance (1949-1989). As for the data from 1989 to 1994,1 have mainly utilized Taiwan sources (such as Yearbook on Chinese Communism, Mainland China Studies, and Studies on Chinese

Communism), with occasional reference to Mainland sources.

[9.] These are individual-years. A governor of a five-year tenure, for instance, is counted as five cases.

[10.] The phrase 'provincial leaders' is used here instead of 'provincial elites' because 'provincial elites' in the Chinese political system may include many other political actors such as regional and provincial military leaders, leaders of provincial people's congresses, and leaders of provincial political consultative conferences.

[11.] During certain historical periods such as that of the 'cultural revolution', these leaders may have different titles. Party secretaries were called provincial first secretaries; deputy secretaries include party secretaries and deputy secretaries except for first secretaries; governors include mayors of the municipalities under central control and chairmen of the autonomous regions as well as chairmen of the 'revolutionary committees' during the 'cultural revolution'; and vice governors include vice mayors of the municipalities under central control and vice chairmen of the autonomous regions as well as vice chairmen of the 'revolutionary committees' of provinces during the 'cultural revolution'.

[12.] For a detailed description of political mobility, see Zhiyue Bo, 'Chinese provincial leaders: economic performance and political mobility', PhD dissertation, Department of Political Science, University of Chicago, (1995), ch. 4.

[13.] It should be indicated that some of those provincial leaders did experience some changes in their same post. For instance, a deputy secretary may be promoted from a junior position to a senior position without changing the title of deputy secretaryship. Or a vice governor may be demoted from first vice governor to second vice governor. Cases of these kinds are not included in this study.

[14.] For a detailed study of retirement of cadres in China, see Melanie Manion, Retirement of Revolutionaries in China: public policies. social norms, private interests (Princeton, NJ: Princeton University Press, 1993).

[15.] For detailed presentations on the benefits of retirement for high-ranking of officials in China, see Hong Yung Lee, From Revolutionary Cadres to Party Technocrats in Socialist China (Berkeley, CA: University of California Press, 1991), pp. 234-245; and Melanie Manion, op. cit.

[16.] Ibid., p. 65.

[17.] People of different ages are divided into four groups: age49 includes age 18 to age 49; age59 includes age 50 to age 59; age64 includes age 60 to age 64; and age90 includes age 65 to age 90.

18. See Su Shuangbi and Wang Hongzhi, Wu Han Zhuan (Beijing: Beijing Chubanshe, 1984), p. 102.

[19.] Michael S. Lewis-Beck and Tom W. Rice, 'Localism in Presidential elections: the home state advantage', American Journal of Political Science 127(3), (August 1983), pp. 548-556.

[20.] See Wei Hsiu-mei, Qingdai zhi Huibi Zhidu [The Avoidance System of the Ch'ing

Dynasty] (Taibei: Institute of Modern History Academia Sinica, 1992), ch. 1, pp. 5-56.

[21.] I thank Dali Yang for pushing me to think hard along this line.

[22.] For a definition of central representation and its provincial variations, see Zhiyue Bo, 'Chinese provincial leaders', ch. 2.

[23.] This is in fact a comulative average. It is the same as the growth rate of the year in the first year of the tenure of the provincial leader. It is a mean of the first year and second year in the second year. It is a mean of the first three years in the third year, and so on.

[24.] I thank Bill Parish for his suggestions in this regard.

[25.] Since provincial populations are highly related to provincial total income (0.71), their impacts are examined m separate models.

[26.] Cheng Li and David Bachman (1989) included the mayors of these municipalities in their study of Chinese mayors. This is, I believe, a serious mistake because the mayors of the three municipalities under central control enjoy the rank of provinces (shengji) while mayors of other cities enjoy the rank of municipalities (shiji).

[27.] I thank Adam Przeworski for his comments in this regard.

[28.] Scholars of China studies have tried to approach Chinese politics from the pluralist perspective since the 1970s. Although their efforts have generated a more realistic picture of Chinese politics than a totalitarianist perspective dictates, they often exaggerated the independence of provincial leaders. Prominent examples include Paris H. Chang's Power and Policy in China (University Park, PA: The Pennsylvania State University Press, 1978); and Susan L. Shirk's The Political Logic of Economic Reform in China (Berkeley, CA: University of California Press, 1993).

[29.] Governors and secretaries enjoy the rank of ministers and vice governors and deputy secretaries have the rank of vice ministers.

Table 1. Political mobility of provincial leaders (1949-1994)

MobilityYear Continue Promoted Lateral Demoted Retired Total

1949 111 2 6 2 0 1211950 145 4 7 8 2 1661951 154 3 9 1 1 1681952 144 11 26 15 1 1971953 175 9 15 6 3 2081954 201 11 20 274 2 2581955 264 9 9 18 2 3021956 375 6 10 9 0 4001957 402 2 3 12 2 4211958 363 15 13 53 5 4491959 439 12 1 19 2 4731960 472 0 7 14 1 4941961 485 7 11 7 2 512

1962 502 3 11 22 2 5401963 480 4 15 35 2 5361964 487 5 11 25 3 5311965 496 18 16 17 6 5531966 453 10 4 85 8 5601967 13 3 3 464 3 4861968 276 1 3 22 0 3021969 295 2 3 4 0 3041970 355 3 2 23 2 3851971 451 8 7 26 1 4931972 475 7 9 17 4 5121973 503 5 22 28 2 5601974 517 8 7 15 2 5491975 492 19 21 27 5 5641976 512 5 1 36 2 5561977 308 37 33 181 6 5651978 503 26 34 58 14 6351979 407 91 37 39 80 6541980 536 33 10 12 31 6221981 567 14 17 7 14 6191982 504 20 10 8 101 6431983 162 28 11 22 343 5661984 302 2 2 0 23 3291985 220 27 8 11 68 3341986 290 12 3 9 14 3281987 276 9 16 11 14 3261988 246 11 12 16 52 3371989 281 17 9 10 9 3261990 282 8 18 8 11 3371991 300 12 10 4 12 3381992 304 11 13 8 15 3511993 219 31 25 18 71 3641994 349 9 1 2 2 363

Total 16093 590 541 1458 945 19627

Table 2. Political mobility of provincial leaders (1949-1994)

MobilityProvince Continue Promoted Lateral Demoted Retired Total

Beijing 549 25 27 54 32 687Tianjin 445 26 24 34 37 566Hebei 700 22 21 52 41 836Shanxi 604 27 24 49 33 737Neimeng 609 13 18 42 30 712Liaoning 691 23 15 68 35 832Jilin 642 21 18 54 32 767Heilongi 559 15 17 74 34 699Shanghai 601 30 10 64 37 742Jiangsu 504 20 13 42 32 611Zhejian 478 25 10 41 28 582Anhui 553 12 12 59 44 680Fujian 571 15 18 48 32 684Jiangxi 512 16 14 41 32 615Shandong 576 26 26 50 37 715

Henan 496 25 23 47 23 614Hubei 605 22 27 48 36 738Hunan 553 26 20 50 23 672Guangdon 614 33 33 48 39 767Guangxi 522 17 14 43 35 631Scihuan 615 30 10 54 41 750Guizhou 458 10 17 61 27 573Yunnan 509 15 20 41 39 624Xizang 592 15 18 47 15 687Shaanxi 539 17 18 58 39 671Gansu 543 22 21 62 35 683Qinghai 454 11 21 45 25 556Ningxia 390 8 17 38 24 477Xinjiang 547 23 12 40 23 645Hainan 62 0 3 4 5 74Total 16093 590 541 1458 945 19627

Table 3. Summary statistics

Variable Obs Mean SD

Dependent variablesContinue 19627 0.8199419 0.3842456Promoted 19627 0.0300606 0.1707585Lateral 19627 0.0275641 0.1637243Demoted 19627 0.0742854 0.2622415Retired 19627 0.048148 0.2140843Independent variablesgov 19627 0.0656748 0.2477192sec 19627 0.0631273 0.243198vicgov 19627 0.5447088 0.4980098depsec 19627 0.326489 0.4689404paryears 12247 33.80011 9.295372edu2 5882 0.6414485 0.4796159nation2 19627 0.0930351 0.290489homeprov 14520 0.3447658 0.4753082center 19627 0.0159884 0.1254335female 19627 0.0333214 0.1794792nonccp 13217 0.0733903 0.260786age49 14536 0.2436709 0.4293112age59 14536 0.4271464 0.4946809age64 14536 0.1867089 0.389691age90 14536 0.1424739 0.3495475period1 19627 0.3757579 0.4843303period2 19627 0.2440516 0.4295344period3 19627 0.3801906 0.4854459growth 18195 7.576831 12.09922income 18823 3.6175 2.21076revenue 18479 6.362395 21.88478pop 18216 3.572421 2.464938mav 19627 6.420887 6.844274revave 19627 5.517292 19.19251Beijing 19627 0.0350028 0.1837915Tianjin 19627 0.0288378 0.1673548Shangha 19627 0.0378051 0.1907294

Variable Min. Max.

Dependent variablesContinue 0 1Promoted 0 1Lateral 0 1Demoted 0 1Retired 0 1Independent variablesgov 0 1sec 0 1vicgov 0 1depsec 0 58paryears 0 1edu2 0 1nation2 0 1homeprov 0 1center 0 1female 0 1nonccp 0 1age49 0 1age59 0 1age64 0 1age90 0 1period1 0 1period2 0 1period3 0 1growth 58.29205 76.69566income 0.1246062 10.76967revenue - 32.99 155.55pop 0.1681375 11.31764mav - 45.133326 50.81633revave - 32.16 155.42Beijing 0 1Tianjin 0 1Shangha 0 1

Table 4. Educational backgrounds and political mobility

Variable Promoted Lateral Demoted Retired (Continue)

sec 0.338[C] 1.592 0.534 1.301vicgov 1.222 0.570 0.303[C] 8.259[C]depsec 0.716 0.974 0.281[C] 3.210[B]gov -- -- -- --nonccp 0.064[B] 1.017 0.510 0.546[A]ccp -- -- -- --college 0.544[C] 0.716 0.861 0.978no college -- -- -- --Minority 0.820 0.765 0.891 0.646Han -- -- -- --natives 1.228 0.489[B] 0.636 1.026outsiders -- -- -- --central 5.786[C] 1.330 1.276 1.11e-14local -- -- -- --female 0.502 1.031 2.454[B] 0.850

male -- -- -- --age49 0.829 1.000 0.773 0.158[A]age64 1.530[A] 0.703 0.807 11.393[C]age90 2.845[C] 0.518 0.580 28.412[C]age59 -- -- -- --

Number of obsChi[sup2](48) = 425.44Pseudo R[sup2] = 0.1147

[A] p <0.10;[B] p <0.05;[C] p <0.01.

Notes: 1. Continue is the comparison group. 2. Period3(1978-1993) only.

Table 5. Determination of political mobility

Variable Promoted Lateral Demoted Retired (Continue)

sec 0.268[C] 1.159 0.857 0.776vicgov 0.840 0.456[C] 0.508[C] 3.511[C]depsec 0.576[C] 0.777 0.587[C] 2.206[C]gov -- -- -- --nonccp 0.103[C] 0.624 0.712 0.453[C]ccp -- -- -- --Minority 0.592[C] 0.467[C] 0.887 0.637[B]Han -- -- -- --natives 1.037 0.647[C] 0.838 0.965outsiders -- -- -- --female 0.450[A] 0.904 1.615 1.400male -- -- -- --central 3.217[C] 1.846[B] 1.904[B] 0.405[A]local -- -- -- --age49 1.209 1.188 0.749[B] 0.361[B]age64 1.446[C] 0.789 0.805 10.244[C]age90 1.544[C] 0.737 0.873 34.385[C]age59 -- -- -- --period1 0.376[C] 1.737 6.194[C] 0.734period2 0.345[C] 1.734 3.808[B] 0.484[B]period3 -- -- -- --year78 0.823 4.235[C] 7.572[C] 0.852year79 3.627[C] 5.312[C] 5.164[B] 3.994[C]year81 0.393[B] 1.708 1.859 0.455[A]year82 0.797 1.966 2.103 4.253[C]year83 2.311[C] 3.204[A] 12.084[C] 38.284[C]year84 0.101[C] 0.227 4.15e-15 3.918[C]year85 1.949[B] 2.198 3.554[A] 26.866[C]year86 0.654 0.660 2.225 4.800[C]year87 0.518 3.229[B] 3.094[A] 4.601[C]year88 0.673 2.662[A] 5.210[B] 23.612[C]year89 0.903 1.984 3.602[A] 4.045[C]year90 0.463[A] 3.856[C] 3.160[A] 4.866[C]year91 0.507[A] 2.201 1.584 4.030[C]year92 0.558 2.706[A] 2.603 4.424[C]

year93 2.804[C] 8.170[C] 12.328[C] 37.545[C]year80 -- -- -- --Number of obs = 12281Chi[sup2](112) = 2619.61Pseudo R[sup2] 0.1659

[A] p <0.01;[B] p <0.05;[C] p <0.01.

Note: 1967 and 1977 are excluded.

Table 6. Provincial performance and mobility

Variable Promoted Lateral Demoted Retired (Continue)

Model 1growth 0.999 1.004 1.006 1.013[A]income 1.105[C] 0.963 1.007 1.031revave 0.980[C] 0.992 0.967[C] 0.984[C]Beijing 1.900[C] 1.126 1.764[B] 1.293Tianjin 2.062[C] 1.419 2.312[C] 1.429Shangha 3.844[C] 0.894 3.034[C] 4.214[C]Others -- -- -- --Number of obs: 11113Chi[sup2](136): 2548.48Pseudo R[sup2]; 0.1830Model 2mav 0.931[C] 0.936[C] 0.945[C] 0.908[C]pop 1.093[C] 0.967 0.991 1.048[A]revave 0.982[C] 0.991 0.968[C] 0.986[C]Beijing 2.895[C] 1.583 2.862[C] 1.484Tranjin 2.165[C] 1.295 2.107[B] 1.337Shangha 7.069[C] 0.969 4.240[C] 4.092[C]Others -- -- -- --Number of obs: 11103Chi[sup2](136): 2783.76Pseudo R[sup2]: 0.2004

[A] p <0.10;[B] p <0.05[C]; p <0.01.

Note: The items in Table 5 are all included but not reported.

Table 7. Provincial performance and mobility

Variable Promoted Lateral Demoted Retired Continue

Model 2.1 (promoted as the comparison group)mav -- 1.007 1.016 0.976[A] 1.075[C]pop -- 0.885[C] 0.907[C] 0.959 0.915[C]revave -- 1.009 0.985[B] 1.004 1.018[C]Number of obs: 11103Chi[sup2](136): 2783.76

Pseudo R[sup2]: 0.2004Model 2.2 (demoted as the comparison group)mav 0.985 0.991 -- 0.961[C] 1.058[C]pop 1.103[C] 0.977 -- 1.058 1.010revave 1.015[B] 1.025[C] -- 1.019[B] 1.034[C]Number of obs: 11103Chi[sup2](136): 2783.76Pseudo R[sup2]: 0.2004

[A] p <0.10;[B] p <0.05;[C] p <0.01.

Note: The items in Table 5 are all included but not reported.

DIAGRAM: Figure 1. Power structure of provincial government.

~~~~~~~~

By ZHIYUE BO[*]

[*] Zhiyue Bo is a visiting assistant professor of political science at Roosevelt University. He is presently completing a book called [Economic Performance and Political Mobility: Chinese Provincial Leaders. This paper is derived from his dissertation completed at the University of Chicago. His thanks go to committee members Adam Przeworski, Bill Parish, Dali Yang, and Guy Alitto for thier advice and comments. Thanks also go to his colleague at Roosevelt University, Jeff Helgeson, for helping make the paper readable.

Copyright of Journal of Contemporary China is the property of Carfax Publishing Company and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.Source: Journal of Contemporary China, Jul96, Vol. 5 Issue 12, p135, 20pItem: 9608225137

2002 “Governing China in the early 21st Century: Provincial Perspective,” Journal of Chinese Political Science (forthcoming).

  2002 “Cong Shengyiji kan Ershiyishiji de Zhongguo Zhengzhi” (Chinese), Chinese

Social Sciences Review (forthcoming).   2002 “The Provinces: Training Ground for National Leaders or a Power in Their Own

Right?” a chapter in David M. Finkelstein and Maryanne Kivlehan, eds., Chinese Leadership in the Twenty-first Century: The Rise of the Fourth Generation Transition (M.E. Sharpe, forthcoming).  

2002 “Institutionalization of Elite Management in Post-Deng China,” a chapter in Barry

Naughton and Dali Yang, eds., Holding China Together (accepted by the editors for publication as a contributing chapter, and the publisher of the book is yet to be identified).

  2000 “Economic Development and Corruption: Beijing Beyond ‘Beijing’” Journal of

Contemporary China 9, no. 25: 467-487.   2000 “Managing Political Elites in Post-Deng China” Asian Profile, Vol. 28, No. 5, pp.

349-370. Under the invitation by International Political Science Association based in France, an abstract of this article is published in International Political Science Abstracts, the main reference journal in the discipline of political science in the world.

  1999 “The Stratification of Social Entitlements in China” Asian Profile 27, no. 3: 189-

204.   1998 “Provincial Power and Provincial Economic Resources in China” Issues &

Studies 34, no. 4: 1-18.