Working Paper: NBER ID: w26375
Authors: Amit Gandhi; Jean-François Houde
Abstract: We study the estimation of substitution patterns within the discrete choice framework developed by Berry (1994) and Berry, Levinsohn, and Pakes (1995). Our objective is to demonstrate the consequences of using weak instruments in this non-linear GMM context, and propose a new class of instruments that are designed to avoid weak IV and can be used to estimate a large family of models with aggregate data. We argue that strong instruments should reflect the (exogenous) degree of differentiation of each product in a market (Differentiation IVs), and provide a series of examples to illustrate the performance of simple instrument functions.
Keywords: No keywords provided
JEL Codes: C35; C36; L13
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| Weak instruments in nonlinear GMM contexts (C51) | Significant biases in estimating substitution patterns (C46) |
| Proposed differentiation IVs (C36) | Improved precision of estimates (C51) |
| Use of differentiation IVs (C36) | Improved estimation outcomes (C51) |
| Strong instruments (C36) | Ability to reject the IIA hypothesis (L15) |