Working Paper: NBER ID: w31632
Authors: Anirban Basu
Abstract: This chapter reviews the econometric approaches typically used to deal with the spike of zeros when modeling non-negative outcomes such as expenditures, income, or consumption.
Keywords: Econometrics; Nonnegative Outcomes; Zeros; Statistical Modeling
JEL Codes: C10; D0; I0
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| treatment variable d (C22) | nonnegative outcome y (C25) |
| treatment variable d (C22) | potential outcomes yj (j=0,1) (C25) |
| treatment variable d (C22) | expected observed outcome E[y] (C29) |
| zeros (Y60) | corner solutions in utility maximization (D11) |
| presence of excess zeros (C29) | overdispersion (C46) |
| understanding of data generation process (DGP) (C51) | accurate modeling (C59) |