外生マスクを超えて：エネルギーシステムモデルにおけるフィードバック、政策閾値、感受性介入点の学習

Unofficial AI-generated summary based on the public title and abstract. Not an official translation.

🔬研究者:Provides a novel diagnostic approach to reveal policy thresholds and cross-technology coupling in energy system models.

🏢実務担当者:Offers insights for designing subsidy schemes and understanding deployment dynamics, useful for renewable project planning.

🏛政策担当者:Highlights the need to consider endogenous learning and sensitive intervention points when setting subsidy levels.

<title>Abstract</title> <p>Energy system models commonly represent technological cost reductions as exogenous functions of time, independent of the amount of capacity deployed. While computationally efficient, this removes the deployment–cost feedback that characterizes learning-by-doing, which prior work shows leads to systematic underestimation of renewable cost and deployment improvements. This paper argues that the misestimation framing is incomplete: the missing feedback also reshapes the relationship between policy intensity and deployment outcomes, introducing thresholds and cross-technology coupling that exogenous formulations cannot represent by construction. It proposes a complementary approach using deliberately stylized models with endogenous learning as a diagnostic instrument. Demonstrated on a stylized Great Britain power system in PyPSA, the diagnostic locates a sharp battery-subsidy threshold near £513/kW (within a threshold region of roughly £500–520/kW across the configurations tested), at which a sub-1% change in subsidy magnitude produces order-of-magnitude differences in deployment: a sensitive intervention point (SIP). This is one of three findings, each structurally invisible to the exogenous formulation of the same model, that together demonstrate the central point: exogenous costing smooths over the non-linear policy dynamics endogenous learning makes visible. The comparison also exposes policy–technology relationships, including the ineffectiveness of carbon pricing alone in launching emerging technologies and the need for a coordinated solar–battery subsidy corridor to reduce required expenditure. The methodology serves as a pre-deployment screen to flag whether a policy sits near a SIP, a retrospective screen for policies designed under exogenous assumptions, and a way to identify which questions exogenous integrated assessment models can and cannot reliably answer.</p>

• Research Square https://doi.org/10.21203/rs.3.rs-10792382/v1first seen 2026-08-26 04:20:14 · last seen 2026-09-08 04:21:28