摘要
已有文献普遍借助于空间计量模型来检验标尺竞争理论。本文指出,传统上基于外生空间权重矩阵的空间回归模型并不适合刻画当竞争标尺是内生决定的情形。为此,本文首次建立一类命名为空间秩次自回归(spatial rank-order auto-regression,SRAR)的模型来考察截面上的内生标尺竞争效应。P(1)阶SRAR模型表现了截面上任一个体与相对绩效优于或劣于自身的前P个近邻个体展开竞争的情景,因而刻画了一种超越传统地理意义的内生空间相关结构。在运用所提模型对国内81只开放式股票型基金2007~2009半年度复权净值增长率排名中可能存在的锦标赛特征进行估计后,我们发现:锦标赛效应在各截面上均显著存在;这一效应表现出年末强于年中,大盘向好时强于大盘向差时的特点。估计结果提示标尺竞争型激励机制分别存在于基金公司-基金经理,基民-基金这两层委托-代理关系中;投资组合等特征差异化不是大部分基金追求高收益的主要策略,追赶与模仿排名靠前且接近的同类基金可以解释其中大部分的原因。空间秩次自回归模型为研究各类有关标尺竞争与攀比效应的经济学问题提供了一个崭新的实证分析框架。
Existing literature often rely on spatial econometric models and techniques to test the "yardstick competition" theory.As argued in this paper,traditional spatial models based on exogenously specified spatial weights matrix may be not appropriate enough to capture the cross sectional competition especially when the yardstick is endogenously determined.To overcome this,we propose a new "spatial rank-order autoregressive(SRAR) model" where a unit competes against the neighboring units ranking ahead of or behind itself by their relative performances.We then apply the model to examining the tournament feature in funds performance ranking by using the data of 81 open-end stock funds in China from 2007-2009.The empirical results indicate that the tournament feature is significantly present indeed in funds ranking,and the feature is stronger at year end than in the interim ranking,for an upward trend market than for a downward trend market,suggesting that yardstick competition exists in the principal-agent relationships between fund corporations and fund managers,fund holders and funds.Our SRAR model provides an alternative approach to analyzing the economic problems about yardstick competition and copycatting behavior.
出处
《上海经济研究》
CSSCI
北大核心
2011年第7期3-14,共12页
Shanghai Journal of Economics
基金
上海市教育发展基金会晨光计划项目"广义空间计量模型前沿理论及其在微观金融市场个体策略性博弈中的实证研究"(编号10CG66)的资助
关键词
标尺竞争
空间秩次自回归模型
工具变量
基金业绩排名
Yardstick competition
Spatial rank-order autoregressive model
Instrumental variable
Fund performance ranking