Difference‐in‐Difference Causal Forests With an Application to Payroll Tax Incidence in Norway
提出差分因果森林方法，用于估计动态处理效应下的异质性处理效应，并应用于挪威工资税对工资的归宿分析，发现企业及劳动力特征解释了税负转嫁的异质性。
ABSTRACT This paper introduces the difference‐in‐difference causal forest (DiDCF) method, which extends the causal‐forest technique for estimating heterogeneous treatment effects to settings with dynamic treatment effects. Regular causal forests require independence between treatment assignment and the outcome variable (after conditioning out observables). In contrast, DiDCFs provide consistent estimates with a parallel trend assumption. DiDCFs can be used to create event‐study plots. The method is applied to estimate payroll tax incidence on wages. We find that heterogeneity in incidence is explained by firm‐ and workforce‐level variables. Firms with a large and heterogeneous workforce are most effective in passing on the incidence of the payroll tax to workers.