Bounds for average treatment effect on couples with an application to the retirement-consumption puzzle
提出了当分析单位为夫妻或个体在夫妻中时，平均处理效应（ATE）的界，并应用于退休-消费之谜，发现首次和联合退休导致支出小幅下降最多14个百分点，二次退休对消费的影响在-5到13个百分点之间，生物性别似乎不驱动消费支出变化。
We present bounds for the Average Treatment Effect (ATE) of an exposure, Z , on an outcome, Y , when the units of interest are couples or individuals in a couple. We illustrate the usage of the bound in an application to the retirement-consumption puzzle. • New bounds for the Average Treatment Effect (ATE) when the units of interest are couples or individuals in a couple. • We illustrate the use of the bounds in an application to the retirement-consumption puzzle. • First and joint retirements lead to small drops in expenditure of up to 14 percentage points (pp), • Second retirements will cause a small variation in consumption of [-5pp,13pp]. • Biological gender does not seem to drive variations in consumption expenditure.