既有研究多半只用捐款金額衡量企業慈善，本研究換一個問法：企業慈善的成本，究竟是公司出錢，還是員工出時間？多數研究以單一揭露總額作為分析單位，本研究改為逐筆閱讀並編碼每一項揭露的公益活動內容。以2020-2024年臺灣50指數成分股企業的ESG 與永續報告書為樣本，人工編碼建立238筆公司年度觀察值，每項活動依主題分為教育與人才培育、社區福祉、環境倡議、社會共融與平權、緊急救援五類，並依資源型態分為財務捐贈、員工志工與組織支持三種。這項資源組成結構直接回答了「誰在付錢」：財務捐贈占社會參與投入的61.0%，員工志工與組織支持合計占39.0%；有揭露志工活動的企業，年均志工時數達11，276.7小時，顯示員工時間並非慈善支出的邊緣項目，而是實質可觀的一部分。本研究進一步直接檢驗「企業是否慷員工之慨」：簡單相關顯示捐款依賴度越高的企業，志工人數顯著越少，表面上像是以志工替代捐款；但在控制產業別與公司規模後，這項關聯不再達統計顯著。資料顯示，金融控股業捐款比例偏高，消費通路物流業捐款比例則偏低，但兩產業的志工投入並未呈現對應的明顯落差，顯示捐款依賴度與志工強度是跟著產業一起變動，而非同一家公司內部彼此替代。在控制產業與公司規模後，本研究未發現明確證據支持員工志工替代財務捐贈。作為次要的補充分析，控制產業別後，財務捐贈依賴程度較高的企業，亦揭露較多與本業相關的公益活動，但本研究資料無法確認個別捐款金額是否確實投入這些本業相關活動。本研究以逐筆編碼取代單一總額衡量，建立台灣上市企業公益活動的活動層級資料庫，具體且直接地回答了企業慈善究竟由誰買單。;This study examines how Taiwanese listed firms allocate resources for corporate philanthropy， focusing on financial donations， employee volunteering， and organizational support. Rather than relying on a single aggregate donation total， this study asks who pays for corporate philanthropy， the firm’s cash or its employees’ time， by coding each disclosed philanthropic activity individually. Using ESG and sustainability reports from Taiwan 50 Index firms (2020–2024)， 238 firm-year observations were built from the activity level up， with each activity classified into one of five categories (Education and Talent Development， Community Wellbeing， Environmental Initiatives， Social Inclusion and Equity， and Emergency Relief) and coded by resource type. Financial donations account for 61.0% of disclosed social participation input， while employee volunteering and organizational support together account for the remaining 39.0%; firms disclosing volunteer activity report an average of 11，276.7 volunteer hours per year， indicating employee time is a substantial component of what firms report as philanthropy. A simple correlation initially suggests firms substitute employee time for cash (donation reliance is negatively correlated with volunteer headcount)， but this pattern does not survive once industry composition and firm size are held constant: donation reliance and volunteer intensity move together across industries rather than trading off within firms， and this study finds no evidence that volunteering substitutes for financial giving. As a secondary， exploratory finding， donation reliance is also positively associated with business-related philanthropic activity once industry composition is accounted for. By coding activities individually rather than relying on a single reported total， this study builds an activity-level dataset of Taiwanese corporate philanthropy and offers a direct answer to who pays for it.
