Underdetermination problem in methodology of economics

Abstract

This paper explores the Duhem-Quine (DQ) problem and its impact on economic methodology, focusing on how the reliance on auxiliary assumptions complicates the testing and validation of theories. The DQ problem shows that no hypothesis is tested in isolation, as it depends on additional assumptions and background knowledge, making it challenging to pinpoint where errors lie. This issue is particularly relevant in economics, where complex models and assumptions about human behavior play a significant role, and in finance, where the robustness of models is critical for decision-making under uncertainty.

The paper highlights two key gaps: (i) the limited discussion of the DQ problem in economic methodology, and (ii) the lack of alternative approaches to ensure rational methods in light of DQ. To address these issues, it proposes a multi-criterial framework for evaluating theories, emphasizing consistency, diverse data, localized testing, comparing models, and varying assumptions systematically.

Using examples like housing market models and the Ultimatum Game, the paper illustrates how addressing the DQ problem involves avoiding arbitrary changes to assumptions while adopting clear, rational strategies. By providing a stronger methodological foundation, this approach enhances the reliability of economic and financial theories, improving their influence on policy-making and practical applications.