市场化进程中的"能力-庇护"双效应与冲突性偏差: 关于转型期中国共产党员收入优势的悖论. ; 关于转型期中国共产党员收入优势的悖论 ; Shi chang hua jin cheng zhong de "neng li-bi hu" shuang xiao ying yu chong tu xing pian cha: guan yu zhuan xing qi Zhongguo gong chan dang yuan shou ru you shi de bei lun. ; Guan yu zhuan xing qi Zhongguo gong chan dang yuan shou ru...
在数据分析的基础上,本文检验了两个理论假设--"政治资本假设与"冲突性偏差假设,以分析国家主导的市场化进程对共产党员收入优势的影响。主流观点认为,在中国,党员身份代表了一种政治资本,因而在经济转型期,共产党员收入优势的变化就相应地反映了市场化进程中政治资本收益(或党员身份政治资本含量)的变化,即关于党员身份的"政治资本假设。而"冲突性偏差假设则考虑了市场化进程中,政治资本和人力资本对个体的职业流动和收入回报的双向影响,以及这种双向影响在回归分析中造成的偏误。通过运用 Pearl(2000)有关冲突性偏差( collider bias,选择偏差的一种)的分析结论,笔者推断,在市场化进程中,控制"单位体制类型和"干部位置会导致党员身份和个人收入之间出现虚假的负偏差,因此,即使党员身份带来的政治资本收益不变,"党员身份对"个人收入的回归系数(控制"单位体制类型和"干部位置后)也会因为选择偏差的上升而呈现出下降甚至转变为负的趋势。如若"党员身份对"个人收入的回归系数在控制"单位体制类型和"干部位置后变为零,反而证明"党员身份对"个人收入仍存在积极的直接效应--这就是关于市场转型期中国共产党员收入优势的悖论。通过运用 2008年中国综合社会调查(CGSS2008),回归分析结果充分验证了"冲突性偏差的经验预测。这也提醒研究者:在对当代共产党员的收入优势进行理论解读时,主流观点忽视了冲突性偏差所造成的影响。在文章的最后,笔者讨论了此发现对将来研究的启发。 ; This study examines two hypotheses for the impact of deepening marketization on the income returns to party membership in socialist China: the political capital and collider bias hypotheses. The prevailing view holds that party membership brings political capital (positional power and connections) and changing income returns reflects the changing leverage of political capital in market transition. The collider bias hypothesis acknowledges the expanding role of unobserved ability in the determination of income in the presence of party clientelism. Drawing on Pearl's analysis of collider bias, we argue that recent marketization would lead to a paradoxical increase of negative bias on the income returns to party membership even if the political capital benefits of party membership persist as the prior literature maintains. Consequently, after conditioning on sector and cadre position, regression coefficient of party membership on income would be suppressed and become even negative. The paradox is that, if the income effect of party membership vanishes after controlling sector and/or cadre position, there must be a true positive party effect on income. Results based on the 2008 Chinese General Social Survey consistently support the predictions of collider bias hypothesis. The prevailing interpretation of previous findings is seriously undermined by ignorance of this pervasive selection bias. ...