Fair value measurements of financial assets and audit fees for the non-banking industry: The role of corporate governance

Abstract

本研究以台灣2016年至2021年的非銀行業上市櫃公司為研究樣本探討非銀行業的金融資產公允價值衡量與審計公費之關聯性，並進一步分析公司治理表現的調節效果。本研究發現總公允價值衡量的金融資產比重和審計公費成負相關：然而第三階公允價值衡量的金融資產比重則和審計公費成正相關。再者，本研究發現公司治理表現會影響審計公費，但不會影響總公允價值衡量的金融資產比重和審計公費之關聯性。除此之外，本研究的額外測試發現(1)第一階公允價值衡量的金融資產比重和審計公費成負相關，然而，第二階及第三階公允價值衡量的金融資產比重合計數則和審計公費成正相關；(2)移除沒有金融資產公允價值衡量的樣本公司，結果大致仍相同；(3)改變部分變數定義結果仍一致；(4)將樣本區分為新冠疫情前(2016-2019)及新冠疫情後(2020-2021)，兩期間的結果只有小部分不相同。最後本研究也採用固定效果模型來控制遺漏變數的內生性問題，實證結果和正文相同。本研究結果可以補充現有文獻的不足，並對公允價值會計提供啟示。

This research takes non-banking listed companies in Taiwan over the period 2016-2021 as samples to investigate the correlation between fair-valued measurements of financial assets and audit fees and to further explore the moderating effects of corporate governance performance. The paper documents that the proportion of total fair-valued financial assets is negatively associated with audit fees, but that the proportion of financial assets that are fair-valued using Level 3 inputs is positively associated with audit fees. In addition, my study finds that corporate governance performance influences audit fees, but it has no impact on the association between the proportion of total fair-valued financial assets and audit fees. Additional analyses find that (1) the proportion of financial assets that are fair-valued using Level 1 inputs is negatively associated with audit fees, however, the sum of the proportion of financial assets that are fair-valued using Level 2 and Level 3 inputs is positively associated with audit fees; (2) after deleting observations without fair-valued financial assets, the results mostly match my above tests; (3) after changing several variables in the equations, findings are consistent with prior results; and (4) separating my sample period into two sub- periods (pre-COVID-19 (2016 to 2019) and post-COVID-19 (2020 to 2021)), the empirical results have a few inconsistent findings between each other. Finally, I employ the fixed-effect model to mitigate the endogeneity that arises from omitted unobservable variables, and all findings are consistent with my main tests. The findings herein complement the literature and offer implications for adopting fair value accounting.