Working Paper: NBER ID: w19784
Authors: Patrick M. Kline
Abstract: We derive the limiting distribution of the Oaxaca estimator of average treatment effects studied by Kline (2011). A consistent estimator of the asymptotic variance is proposed that makes use of standard regression routines. It is shown that ignoring uncertainty in group means will tend to lead to an overstatement of the asymptotic standard errors. Monte Carlo experiments examine the finite sample performance of competing approaches to inference.
Keywords: No keywords provided
JEL Codes: C01; J31; J7
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| treatment status (di) (I12) | scalar outcome (yi) (C29) |
| covariates (xi) (C20) | scalar outcome (yi) (C29) |
| treatment status (di) (I12) | covariates (xi) (C20) |
| mean untreated outcomes (I12) | Oaxaca estimator (C51) |
| odds of treatment (C22) | Oaxaca estimator (C51) |
| ignoring variability in group means (C20) | overestimate of asymptotic standard errors (C51) |