Working Paper: NBER ID: w27198
Authors: Aprajit Mahajan; Christian Michel; Alessandro Tarozzi
Abstract: Time-inconsistency may play a central role in explaining inter-temporal behavior, particularly among poor households. However, little is known about the distribution of time-inconsistent agents, and time-preference parameters are typically not identified in standard dynamic choice models. We formulate a dynamic discrete choice model in an unobservedly heterogeneous population of possibly time-inconsistent agents. We provide conditions under which all population type probabilities and preferences for both time-consistent and sophisticated agents are point-identified and sharp set-identification results for naïve and partially sophisticated agents. Estimating the model using data from a health intervention providing insecticide treated nets (ITNs) in rural Orissa, India, we find that a little over two-thirds of our sample comprises time-inconsistent agents and that both sophisticated and naïve agents are considerably present-biased. Counterfactuals show that the under-investment in ITNs attributable to present-bias leads to substantial costs that are about five times the price of an ITN.
Keywords: Time Inconsistency; Identification of Types; Partial Identification; Mixture Models; Expectations; Malaria; Bednets; Dynamic Discrete Choice
JEL Codes: D91; I12; I3
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| time-inconsistent preferences (D15) | demand for insecticide-treated nets (ITNs) (O15) |
| present bias (D15) | investment in ITNs (H54) |
| time inconsistency (D15) | lower purchases of ITNs (H53) |
| time inconsistency (D15) | lower retreatment rates for ITNs (R20) |
| lower purchases of ITNs (H53) | higher expected costs associated with malaria (I18) |