企業社會責任與融資限制相關性探討-以中國上市公司為例

本文探討中國上市企業之社會責任與融資限制之關係。本文以《南方週刊》於2008年至2013年公佈之中國企業社會責任排行榜之公司為樣本（即CSR公司），並就與CSR公司相同產業、公司規模相近之條件篩選出配對樣本(Non-CSR公司)。本研究依據Baker, Stein and Wurgler(2003)改良的KZ融資限制指標、Whited and Wu(2006)建立的WW融資限制指標、Hadlock and Pierce(2010)建立的SA融資限制指標，作為企業融資限制的衡量，並將樣本分成全樣本，國有企業(SOEs)、非國有企業(Non-SOEs)，研究方法使用Heckman兩階段估計模型，進行迴歸。實證結果顯示，中國上市公司從事CSR對於融資限制指標並無顯著的關係，但在國有企業樣本下CSR對於融資限制有顯著的反向關係，而民營企業樣本下CSR對於融資限制有顯著的正向關係。全樣本的結果可能來自於國有企業與民營企業統計結果不同的抵銷，但經由國有企業樣本的統計結果，可以知道企業社會責任的履行，有助於提高企業資訊透明度降低代理問題，並減少融資限制。至於民營企業樣本的統計結果，可能在於中國CSR政策上，民營企業難以被分配到輔助資源，CSR成本加重企業營運的負擔，反而提高融資限制。

關鍵字

企業社會責任；融資限制

並列摘要

This paper discusses the relation between China's listed corporate social responsibility and financial constraint. Based on the sample of Chinese corporate social responsibility ranking (CSR) published on "Southern Weekly" from 2008 to 2013, the compared sample of the same industry and company size are selected (Non - CSR company). This research uses KZ, WW, SA financial constraint indicators as a measure of the company's financial constraint according to the research of Baker, Stein and Wurgler(2003) and Whited and Wu(2006), Hadlock and Pierce(2010). The sample is divided into state-owned enterprises (SOEs), non-state (Non-SOEs). Methods using the two-stage Heckman estimation model regression. The empirical result shows that there is no significant difference in financial constraint whether China's listed company is engaged in CSR or not, but in state-owned enterprises samples that there between CSR and financial constraint is significant inverse relationship. Results of samples may different statistical results from state-owned enterprises and private enterprises to offset. However, we can know the fulfillment of CSR that there is improving transparency of information to reduce agency problem, and reduce financing constraints. As for the statistics of private enterprises samples that China's CSR policy, private enterprises can be assigned to the secondary resource, CSR cost burden on business operations, but to improve the financial constraints.