End value formulations for mitigating finite-horizon distortions in Ramsey-based integrated assessment models
Year: 2026
Type: preprint
Abstract: Approximating infinite-horizon problems with finite-horizon optimization is a persistent challenge in Ramsey-type integrated assessment models (IAMs), often leading to end-of-horizon distortions. As the end of the optimization horizon approaches, the social planner has no incentive to invest and instead consumes all output. Extending the optimization horizon, a common practice in numerically solved IAMs, can mitigate these terminal effects but increases computational costs and is unsuitable for deliberately myopic models. An alternative is to add an end value, representing the value of remaining stocks, to the objective function. However, existing formulations have been limited to simple Ramsey growth models. This article proposes two approaches for formulating end values, applicable to both standard Ramsey growth models and hybrid energy-economy models. The first approach employs an explicit integrated tail of future utility flows, while the second uses an iterative algorithm based on shadow prices. Using a nonlinear programming framework, we evaluate both formulations across varying horizon lengths. The results demonstrate that while both approaches converge to the infinite-horizon solution over sufficiently long horizons, shorter horizons reveal distinct behaviors. These differences do not represent errors, but rather reflect alternative assumptions about how the social planner treats the post-horizon future. Consequently, the end-value choice should be treated as a deliberate modeling decision tailored to the specific case study.
Source: Research Square
Institution: Potsdam Institute for Climate Impact Research
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W7236298821
