Solving Dynamic Discrete Choice Models: Integrated or Expected Value Function?

TL;DR

This paper compares the expected value function and the integrated (ex ante) value function in solving Dynamic Discrete Choice Models, demonstrating that the integrated approach is more practical and efficient in real-world applications.

Contribution

The paper provides a comparative analysis showing the integrated value function outperforms the expected value function in solving DDCMs, supported by empirical benchmarks.

Findings

Integrated value function is more practical than expected value function.

Benchmarks show the integrated approach is more efficient.

Expected value function is less effective in models with large state spaces.

Abstract

Dynamic Discrete Choice Models (DDCMs) are important in the structural estimation literature. Since the structural errors are practically always continuous and unbounded in nature, researchers often use the expected value function. The idea to solve for the expected value function made solution more practical and estimation feasible. However, as we show in this paper, the expected value function is impractical compared to an alternative: the integrated (ex ante) value function. We provide brief descriptions of the inefficacy of the former, and benchmarks on actual problems with varying cardinality of the state space and number of decisions. Though the two approaches solve the same problem in theory, the benchmarks support the claim that the integrated value function is preferred in practice.