在全球環境、社會與治理（Environmental， Social， and Governance， ESG）永續發展浪潮下，薪酬公平已成為衡量企業社會責任與公司治理 的重要指標。台灣薪酬揭露制度歷經三階段演進：2010 年強制設置薪 資報酬委員會、2019 年擴及非主管職全時員工薪資揭露、2024 年《證 券交易法》修正進一步要求上市櫃公司於章程中訂明基層員工盈餘分配 比率，使ESG 中的「社會（S）」與「治理（G）」從自願揭露走向法 律強制義務。 然而，強制揭露政策是否真的縮小了薪酬差距？差距的變化又如何 影響員工留任行為與企業績效？現有理論對此存在截然對立的預期： 「公平—規範派」主張揭露有助縮小差距，差距過大將引發員工不公平 感、推升流動率並損害績效；「誘因—效率派」則主張揭露反而擴大差 距，且差距本身透過錦標賽機制強化留才效果並提升績效。本研究以同 一資料集依序裁決兩派預期。 本研究以2014 至2024 年台灣1，711 家上市（TSE）及上櫃（OTC） 公司為樣本，共計16，210 筆公司—年觀測值，採公司固定效果面板迴 歸（H1）及雙向固定效果面板迴歸（H2、H3）進行實證分析。 實證結果三組方向一致，均支持「誘因—效率」典範：其一，2019 年薪酬揭露政策實施後，高管與基層員工薪酬差距不減反增（β = 0.020， p < 0.1），且此一正向關聯在資訊透明度較低的上櫃公司中更為明顯（β -ii- = 0.056，p < 0.01）；其二，薪酬差距較大的企業員工流動率顯著較低 （β = −2.363，p < 0.01），與錦標賽理論預期一致，跨市場與跨產業 方向均一致；其三，薪酬差距與資產報酬率（ROA）、股東權益報酬率 （ROE）及毛利率均呈顯著正向關聯（ROA β = 1.498，p < 0.01），以 製造業效果最強，非製造業（排除金融業）ROA 不顯著。此一發現揭 示強制揭露政策的非預期後果：以ESG 社會構面（S）為設計目標的政 策，實際運作邏輯更接近治理構面（G）的誘因機制，政策意圖與企業 實際薪酬結構之間存在可觀察的落差。須說明的是，由於 2019 年政策 為全國上市櫃公司同步實施、缺乏可作為對照基準的控制組，本研究三 項假設之實證結果均定位為關聯性分析，而非因果關係。 關鍵詞： 薪酬揭露政策、薪酬差距、員工流動率、企業績效、錦 標賽理論、相對剝奪感、固定效果面板迴歸、ESG;As ESG has become a global benchmark for corporate sustainability， pay equity has grown into a key indicator of corporate social responsibility and governance. Taiwan′s pay disclosure rules have evolved through three stages: mandatory compensation committees in 2010， disclosure of non-managerial employee pay in 2019， and a 2024 amendment requiring listed firms to earmark a portion of annual earnings for rank-and-file employees. These changes have progressively shifted the "S" and "G" dimensions of ESG from voluntary commitment to legal obligation. Against this background， this study asks three questions: Did mandatory disclosure narrow or widen the pay gap between executives and rank-and-file employees? Is a wider pay gap associated with higher or lower employee turnover? And is the pay gap related to firm performance， and does the relationship hold across industries? For each question， we pit the equity-fairness view against the incentiveefficiency view and let the data decide. The sample covers 1，711 listed firms in Taiwan (TSE and OTC) from 2014 to 2024， yielding 16，210 firm-year observations. We use firm fixed-effects panel regressions for H1 and two-way fixed-effects panel regressions for H2 and H3. -iv- The results consistently favor the incentive-efficiency view. First， the pay gap widened rather than narrowed after the 2019 disclosure rules took effect (β = 0.020， p < 0.1)， with a stronger effect among OTC firms with lower pre-existing transparency (β = 0.056， p < 0.01). Second， firms with wider pay gaps had significantly lower employee turnover (β = −2.363， p < 0.01)， consistent with tournament theory， with the direction holding across market segments and industries. Third， a wider pay gap was positively associated with ROA， ROE， and gross margin (ROA β = 1.498， p < 0.01)， with the strongest effect in manufacturing and an insignificant effect on ROA in services. These findings reveal an unintended consequence of mandatory disclosure: a policy designed around the social dimension of ESG (S) appears to operate through the logic of the governance dimension (G)， with a measurable gap between policy intent and corporate pay-setting behavior. It should be noted that， because the 2019 reform applied simultaneously to all listed firms and offers no clean control group， all three sets of findings are interpreted as associations rather than causal effects. Keywords: pay disclosure; pay gap; employee turnover; firm performance; tournament theory; relative deprivation; fixed-effects panel regression; ESG
