Statistical Discrimination with Peer Effects: Can Integration Eliminate Negative Stereotypes?
将同伴效应引入劳动力市场统计歧视模型，发现社会网络隔离程度存在阈值，低于该阈值时负面刻板印象不可持续，技能水平可能大幅变化，并分析了平权政策的效果。
We introduce peer effects in the costs of human capital acquisition into a model of statistical discrimination in labour markets. This creates a link between the level of segregation in social networks and racial disparities in job assignment and wages. We show that this relationship is characterized by discontinuities: there is a threshold level of segregation below which negative stereotypes become unsustainable, and steady-state skill levels can change dramatically. This change can work in either direction: skill levels may either rise or fall in both groups. Which of these outcomes arises depends on the population share of the disadvantaged group and on the distribution of the costs of human capital investments. We also examine the effects of affirmative action policies in the presence of peer effects and provide conditions under which such policies eliminate negative stereotypes. Copyright 2008, Wiley-Blackwell.