Temporal Discounting and Intergenerational Equity: A Pre-registered Adversarial Collaboration on Delay-of-Gratification Trade-offs

Abstract

Individual differences in temporal discounting (preference for immediate rewards) are proposed to underlie resistance to climate action and resistance to intergenerational resource allocation. This pre-registered adversarial collaboration (N = 2,106) tested competing theories: hyperbolic discounting versus domain-specific risk attitudes. Participants completed incentivised delay choices, environmental risk surveys, and allocation tasks. Bayesian model comparison strongly favoured a dual-process account (BF₁₀ = 34.2), with risk tolerance—not discounting—predicting climate action support. Computational models revealed distinct neural substrates in intertemporal choice.

Introduction

The tragedy of the commons unfolds partly through systematic undervaluation of future consequences. Temporal discounting—the tendency to weight immediate payoffs more heavily than delayed ones—has long been implicated in phenomena ranging from addiction to climate inaction. Yet decades of research have produced conflicting evidence about whether discounting reflects a unitary construct or comprises domain-specific decision processes. This disagreement has practical implications: interventions targeting "temporal impatience" may fail if the underlying mechanism differs across domains.

We conducted a large-scale pre-registered adversarial collaboration to adjudicate competing computational models of intertemporal choice and to test whether temporal discounting—or risk tolerance—better predicts real-world intergenerational resource allocation decisions. This work was motivated by growing evidence that climate risk perception, rather than impatience per se, drives policy opposition.

Method

Participants

We recruited 2,106 participants (M_age = 34.8, SD = 12.4; 52% female) from four geographic regions via Prolific with stratified sampling for socioeconomic diversity: Canada (n = 537), United States (n = 531), United Kingdom (n = 519), and Australia (n = 519). Participants completed informed consent procedures approved by the Magic Institute Behavioural Research Ethics Board (REB #2025-098). A pre-registered power analysis specified N = 1,850 for 80% power to detect interaction effects (α = .05).

Procedure

Participants completed four tasks across two sessions (two weeks apart). Session 1 included 27 incentivised intertemporal choice trials (real stakes; payment via Amazon gift card) with delays ranging from 1 week to 5 years, combined with a domain-generality assessment (choices across financial, health, and environmental domains). Session 2 involved a climate risk perception survey (EPA risk ladder), an allocation task (distribute $100 across immediate personal gain vs. ocean acidification mitigation fund), and a computational modelling battery. All data were analysed using Bayesian hierarchical regression and model comparison via Bayes factors. Pre-registered exclusion criteria removed 31 participants (1.5%) for inconsistent responding.

Results

Temporal discounting exhibited substantial individual variation (discount rate k: M = 0.052, SD = 0.089) and was significantly correlated within domains (r_financial_health = 0.58, 95% HDI [0.52, 0.63]) but only weakly across domains (r_financial_climate = 0.31, 95% HDI [0.24, 0.38]). Bayesian model comparison strongly favoured a dual-process model (BF₁₀ = 34.2 vs. hyperbolic discounting; BF₁₀ = 18.7 vs. exponential discounting), wherein financial and health discounting loaded on a single factor, while environmental discounting loaded separately. Risk tolerance (measured via domain-specific risk appraisals) strongly predicted intergenerational allocation decisions (β = 0.67, 95% HDI [0.59, 0.75]), whereas discount rate did not (β = 0.08, 95% HDI [−0.02, 0.18]). A Bayesian mediation model revealed that risk tolerance, not temporal impatience, accounted for 76% of the effect linking climate salience to allocation choice.

Discussion

Our adversarial collaboration reveals that temporal discounting and risk tolerance are dissociable constructs, with profound implications for policy. The weak domain generality of discounting suggests that interventions addressing "impatience" will fail to shift environmental decision-making; instead, reframing climate risk as visceral and immediate—rather than distant—should be prioritised. The robust dual-process structure aligns with recent neuroimaging work showing dissociable ventromedial prefrontal cortex (vmPFC) and insular activations for financial versus climate-relevant decisions.

Future research should investigate whether risk-framing interventions produce sustained behaviour change, and whether neural signatures of risk representation can identify individuals most responsive to climate messaging. Our Bayesian hierarchical framework, made publicly available via OSF, provides a template for resolving theoretical disputes in decision science.

References

Camerer, C. F., & Loewenstein, G. (2004). Behavioural decision theory. In D. Koehler & N. Harvey (Eds.), Handbook of judgment and decision making (pp. 3–51). Blackwell.
Frederick, S., Loewenstein, G., & O'Donoghue, T. (2002). Time discounting and time preference: A critical review. Journal of Economic Literature, 40(2), 351–401.
Osei, D., Whitfield, R., & Yusuf, H. (2026). Risk tolerance versus temporal discounting in intergenerational allocation. Magic Institute Working Paper, No. 2026-04.
Slovic, P., & Peters, E. (2006). Risk perception and affect. Current Directions in Psychological Science, 15(6), 322–325.
Sunstein, C. R. (2023). Deciding who decides: How collective institutions can overcome the inadequacies of individual judgment. Journal of Economic Literature, 61(2), 411–432.
Thaler, R. H., & Johnson, E. J. (1990). Gambling with the house money and trying to break even: The effects of prior outcomes on risky choice. Management Science, 36(6), 643–660.
Tremblay, M. R., & Osei, D. (2024). Bayesian hierarchical models for multi-domain intertemporal choice. Computational Behavioural Science, 7(3), 245–278.