從個案到基礎設施：基金通證化之市場設計

Abstract

臺灣共同基金市場資產管理規模（AUM）已超過 7.84 兆新臺幣，約當堆疊 1,700 棟臺北 101 大樓。若僅提升萬分之 1 的整體效率，即以總費用率（TER）降低 1 個基點計算，便可釋放近 8 億元的年度經濟效益。然而，支撐該市場運作的金融市場基礎設施（FMI）卻深陷結構性僵化，仍受前數位時代範式所拘束，長期陷於次優均衡的制度困境。 本研究採用「解釋性序列混合方法」（Quan → Qual → Design）。首先，量化模擬證實，若導入以分散式帳本技術（DLT）為基礎的新型 FMI，每年預估可釋放約 280 億新臺幣的潛在經濟效益（約當 TER 降低 35.93 個基點），顯示現行制度下存在重大「效率缺口」。其後，質性分析將此效率缺口的成因，歸納為「制度性雙重硬化症」：亦即由歷史路徑依賴所造成的「操作性硬化」，與來自既得利益與集體行動失靈的「策略性硬化」，在動態上相互強化；並進一步論證，在傳統中心化架構下，即便引入中立第三方，仍難以根本化解此一深層信任困境。 為填補現象描述與制度設計之間的鴻溝，本研究建構「算法信任雙柱模型」作為理論橋樑：將具法律問責的鏈下「治理性信任」與具數學確定性的鏈上「可計算信任」加以整合。基於此模型，提出「臺灣基金鏈（TWFC）」制度藍圖，設計以可信中立第三方為核心、由多元利害關係人共治的許可制聯盟鏈架構，不僅實現基金份額的通證化轉移與近乎即時的交易結算，亦內建權力制衡機制與透明規則，以達成「代碼即規則（Code-as-Rule）」的治理目標。本研究不僅為臺灣資產管理產業的數位轉型提供具體路徑，亦為全球面臨類似制度瓶頸的經濟體，展示一個具數位韌性的下一代 FMI 可複製原型。

Taiwan’s mutual fund market has assets under management (AUM) exceeding NT$7.84 trillion, roughly equivalent to 1,700 Taipei 101 towers. A one–basis-point (0.01 percentage point) reduction in the total expense ratio (TER) would generate approximately NT$0.8 billion in additional annual economic benefits. Yet the financial market infrastructure (FMI) underpinning this market remains structurally rigid, governed by pre-digital paradigms and locked in a persistently suboptimal institutional equilibrium. This study adopts an explanatory sequential mixed-methods design (Quan → Qual → Design). First, a quantitative simulation shows that introducing a distributed ledger technology (DLT)-based FMI could generate NT$28 billion in annual gains (equivalent to a reduction of 35.93 basis points in TER), revealing a substantial “efficiency gap” in the current system. Subsequent qualitative analysis traces this gap to an “institutional double sclerosis”: a reinforcing interplay between “operational sclerosis” driven by historical path dependence and “strategic sclerosis” rooted in vested interests and collective action failures. It further shows that, under traditional centralized architectures, even a nominally neutral third party cannot resolve this deep-seated trust dilemma. To bridge diagnosis and institutional design, the study develops an “algorithmic trust twin-pillars model”. This framework integrates off-chain “governance trust”, grounded in legal accountability, with on-chain “computable trust”, grounded in mathematical certainty. Building on this model, the dissertation proposes the “Taiwan Fund Chain” (TWFC): a permissioned alliance chain centered on a credibly neutral operator and co-governed by multiple stakeholders. TWFC is designed to enable tokenized transfer and near-real-time settlement of mutual fund shares while embedding checks-and-balances mechanisms and transparent rule sets to realize a “code-as-rule” governance regime. This research provides a pathway for the digital transformation of Taiwan’s asset management industry and a replicable prototype of a digitally resilient, next-generation FMI for jurisdictions facing similar institutional bottlenecks.