Like Father, Like Son: Social Network Externalities and Parent-Child Correlation in Behavior
构建代际交叠模型，说明当个人行为受邻居影响时，父母与子女行为的相关性可能源于社交网络重叠，而非直接亲子互动，并预测祖孙间也存在残余相关性。
We build an overlapping generations model where an individual sees higher returns to adopting a behavior as many neighbors adopt the behavior. We show that overlap in the state of a parent and child's neighborhood can lead to correlation in parent-child behavior independent of any parent-child interaction. Increasing the sensitivity of individual decisions to the state of their social community leads to increased parent-child correlation and less efficient (more costly) behavior on average in the society. We show this model is distinguished from a direct parental influence model, in that it predicts increased generational effects, implying residual correlation between children and grandparents after including parental information.