Working Paper: CEPR ID: DP13240
Authors: Myrto Kalouptsidi; Paul T. Scott; Eduardo Souza Rodrigues
Abstract: In structural dynamic discrete choice models, the presence of serially correlated unob-served states and state variables that are measured with error may lead to biased parameterestimates and misleading inference. In this paper, we show that instrumental variables canaddress these issues, as long as measurement problems involve state variables that evolveexogenously from the perspective of individual agents (i.e., market-level states). We definea class of linear instrumental variables estimators that rely on Euler equations expressed interms of conditional choice probabilities (ECCP estimators). These estimators do not requireobserving or modeling the agent’s entire information set, nor solving or simulating a dynamicprogram. As such, they are simple to implement and computationally light. We provideconstructive identification arguments to identify the model primitives, and establish the con-sistency and asymptotic normality of the estimator. A Monte Carlo study demonstrates thegood finite-sample performance of the ECCP estimator in the context of a dynamic demandmodel for durable goods.
Keywords: dynamic discrete choice; unobserved states; instrumental variables; identification; euler equations
JEL Codes: C13; C35; C36; C51; C61
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| ECCP approach (E61) | Superior performance in comparison to estimation techniques that overlook measurement issues (C51) |
| Presence of serially correlated unobserved states and measurement errors in state variables (C32) | Biased parameter estimates and misleading inferences in DDC models (C51) |
| Using IV methods (C36) | Address biases from serially correlated unobserved states and measurement errors (C32) |
| ECCP estimator (C51) | Simplifies estimation process by avoiding modeling the entire information set of agents (C51) |
| ECCP estimator (C51) | Provides consistent and asymptotically normal estimates of structural parameters (C51) |
| ECCP method (C59) | Mitigates biases from unobserved demand shocks correlated with observed prices (E39) |
| ECCP estimator (C51) | Enhances reliability of empirical findings in contexts such as dynamic demand for durable goods, land use choices, and technology adoption (D12) |