政府推動「公共建設計畫跨域加值方案」機制，將溢價回收（Value Capture）機制引入於財務補助評估架構。無論租稅增額融資（Tax Increment Financing, TIF）或土地開發收益皆屬預期收益；倘若，TIF或土地開發收益不如預期，將影響建設計畫財務自償能力，衍生財務風險及財政資源分配負擔問題；然而，過去文獻對此課題研究十分有限。基此，本文以地方政府角度，構建多目標數學規劃財務模式，將前述效益項目內部化於自償率（Self-Liquidation Ratio, SLR），可得地方政府應承擔風險與政府補助建設經費，此可補充過去文獻的研究限制。本文以捷運建設計畫「三峽鶯歌線（三鶯線）」為例，採用蒙地卡羅方法，模擬TIF與土地開發收益之機率分配，利用「整體準則法」求解多目標數學規劃財務模式。實證顯示，計畫自償率最佳妥協解為33.25%，土地開發收益為16.23億，TIF為35.09億；同時，地方政府承擔風險期望值為0.14億，可獲得政府補助經費214.7187億元。若地方政府追求政府補助最大化下，計畫自償率隨之遞增，風險期望值亦隨之遞增；若地方政府偏好承受較低風險，計畫自償率隨之降低，風險期望值隨之遞減。實證分析顯示，本文模式具有可操作性，可應用於其他建設計畫，實施溢價回收機制之財務自償與風險分析之參考。

The government aims to incorporate value capture mechanisms into Taiwan's financial assistance evaluation framework by introducing the Project for Cross-Field Value-Adding in Public Works Financial Planning ("the Project"). Given that the tax increment financing (TIF) and land development (LD) returns are expected returns, the self-solvency of infrastructure projects is affected when the TIF and LD return fails to meet expectations, bringing financial risk and distribution of financial resources into question. However, research into this matter is lacking. We developed a financial model for local governments using a multi-objective mathematical programming method to measure the TIF and LD returns derived from the Project, internalize the returns into a self-liquidation ratio (SLR), calculate the financial risk and burden of the local government, and overcome the limitations of prior literature. Using the "Sanying MRT Line" Infrastructure Project as an example, we adopted the Monte Carlo simulation method to simulate the probability distribution of TIF and LD returns. Additionally, we adopted the "global criteria method" to solve the multi-objective mathematical programming model. Empirical results showed that the best compromise solution of the Project's SLR was 33.25%. In this instance, LD return was NT$1.623 billion, and TIF return was NT$3.509 billion. At the same time, the expected financial risk was NT$14 million, and the expected central government subsidy was NT$21.47187 billion. If the local government sought to maximize central government subsidy, the Project's SLR and expected risk would increase incrementally. Conversely, if the local government sought to minimize the central government subsidy, the Project's SLR and expected risk would decrease incrementally. Empirical analysis results validated the application and operability of the proposed model and highlighted the reference value of the model in measuring the financial self-compensation and risk of implementing the value capture mechanisms in other infrastructure projects.