Temporal Discounting in Technical Debt: How do Software Practitioners Discount the Future?

TL;DR

This study investigates how software developers value future outcomes in technical debt decisions, revealing widespread temporal discounting and individual differences, thus providing empirical insights into actual decision-making behaviors.

Contribution

First empirical study applying intertemporal choice methods to software engineering, revealing how practitioners discount future outcomes and highlighting variability in decision-making.

Findings

Widespread temporal discounting among developers

Significant individual differences in discounting behavior

Raises questions about causal factors influencing discounting

Abstract

Technical Debt management decisions always imply a trade-off among outcomes at different points in time. In such intertemporal choices, distant outcomes are often valued lower than close ones, a phenomenon known as temporal discounting. Technical Debt research largely develops prescriptive approaches for how software engineers should make such decisions. Few have studied how they actually make them. This leaves open central questions about how software practitioners make decisions. This paper investigates how software practitioners discount uncertain future outcomes and whether they exhibit temporal discounting. We adopt experimental methods from intertemporal choice, an active area of research. We administered an online questionnaire to 33 developers from two companies in which we presented choices between developing a feature and making a longer-term investment in architecture. The results show wide-spread temporal discounting with notable differences in individual behavior. The results are consistent with similar studies in consumer behavior and raise a number of questions about the causal factors that influence temporal discounting in software engineering. As the first empirical study on intertemporal choice in SE, the paper establishes an empirical basis for understanding how software developers approach intertemporal choice and provides a blueprint for future studies.