碳交易與地緣政治因素對能源市場之影響:基於 MCP 模型與 Sobol 敏感度分析之歐洲天然氣市場研究

Abstract

本篇論文旨在探討各類成本對天然氣市場中公司決策行為的影響，並透過建立數理模型以模擬不同市場結構下的資源配置結果。本文選取歐洲天然氣市場作為研究對象，鑑於其歐洲的碳排放交易制度運作成熟且涵蓋高碳排產業，並面臨地緣政治與能 源供應結構變動等挑戰，具有分析價值。 研究方法採用 Nash-Cournot 博弈模型建構多企業產量競爭機制，運用 KKT 條件轉化為混合互補問題形式，並使用 GAMS 進行模型求解。在模型中，公司的最終目標 為最大化自身利潤，並納入生產成本、運輸成本、碳權許可證價格及減碳技術成本作為 考量，進一步模擬在不同碳權政策與成本條件下，最適的產銷量與減碳策略的變動情形。 實證部分則是分別模擬單一企業獨佔與多企業庫諾競爭兩種市場結構，設定多項 情境進行模擬比較與政策變動分析，包括生產成本變動、減碳排放量的調整、免費碳許可證的額度變化、俄烏戰爭對天然氣供應鏈影響等等，並進一步去使用全局敏感度分析，評估各個成本參數對總成本的影響。 本篇論文提供決策者在規劃碳排政策與公司營運策略時之量化分析依據，亦可作為未來擴充模型納入儲氣、淡旺季等等的因素之基礎參考。

This study investigates the impact of various cost components on corporate decision-making behavior in the natural gas market and utilizes a mathematical modeling approach to simulate resource allocation outcomes under different market structures. The European natural gas market is selected as the case study, given the maturity of the EU ETS, its coverage of high- emission industries, and the market's exposure to geopolitical tensions and shifts in energy supply structures, making it a valuable subject for analysis. The research employs a Nash-Cournot game-theoretic framework to model quantity competition among multiple firms. The optimization problem is reformulated into a Mixed Complementarity Problem using KKT conditions and solved through the GAMS. In the model, companies aim to maximize individual profit, accounting for production costs, transportation costs, carbon permit prices, and decarbonization technology costs. The study simulates optimal production, sales, and emission reduction strategies under varying carbon policies and cost structures. The empirical analysis includes simulations of both monopolistic and Cournot-competitive market structures, with a series of policy and cost-related scenarios, such as changes in production costs, emission reduction volumes, the allocation of free carbon permits, and the impact of the Russia-Ukraine war on natural gas supply chains. A global sensitivity analysis is conducted to evaluate the relative influence of each cost parameter on total cost outcomes. This study provides a quantitative foundation for policymakers and corporate decision-makers in designing carbon pricing strategies and operational planning. It also serves as a baseline for future model extensions, including storage dynamics, seasonal fluctuations, and intertemporal decision-making.