政治連結與銀行貸款利率

本研究根據2004～2012年台灣上市非金融產業公司的資料檢驗是否公司政治連結（political connection）程度愈高，愈能夠獲得較低的銀行貸款利率，以致於使公司有較低廉的銀行資金借貸成本。實務上企業具有特定政治傾向是相當普遍的現象，例如公司聘請政治人物進入董事會或擔任高階管理者、在選舉時公開表態支持特定候選人，亦或給予政黨政治獻金等，可以推測企業之所以尋求政治連結，或許是因政治連結給公司帶來助益，使公司有較佳的財務績效或較低的營運風險。既有研究已經指出公司政治連結對於公司將產生助益但也會產生成本，然而對於公司政治連結是否可以在金融市場上取得較低資金成本的討論相對較少，特別是台灣地區受政府政策主導的官股銀行在銀行資金借貸市場上佔有重要地位，台灣的公司是否能夠在政治上與政黨或政治人物有所連結進而取得較為低廉的銀行貸款資金，便成為一個有趣的議題並成為本研究的主要研究動機。另外本研究在量化公司的政治連結程度上亦有所創新，透過在台灣地區發行的報紙中，使用文字搜尋（公司名稱、董事長姓名、政黨名稱、執政黨之總統姓名與在野黨之黨主席姓名）的方式並統計特定公司與特定政黨交疊之政治新聞筆數以建構公司政治連結的衡量指標。本研究的基本實證結果發現，較高程度的整體政治連結對應著較低的銀行貸款利率，而考慮採用兩階段估計以矯正公司樣本需求政治連結的選擇偏誤問題、進一步使用分量迴歸估計以分析不同利率水準下政治連結的效應、以政治連結變數落遲項進行估計以降低同期相關謬誤等穩健性測試，實證結果皆仍然支持公司政治連結有助於降低銀行貸款利率的觀點。

關鍵字

政治連結；銀行貸款利率

並列摘要

Based on data of listed-companies on the Taiwan Stock Exchange (TWSE) covering the period of 2004~2012, this paper examines whether higher firm's political connection enjoys cheaper financing cost on bank loans in terms of lower bank loan rates. Practically, enterprise has a specific political orientation is a common phenomenon, for example, company may hire politician or retired government officials to be board of director or senior management, publicly expressed support for specific candidates during election, or make political contribution to specific politician or party. Existing research has indicated that the political connection brings benefits ranging from better financial performance or lower operating and financial risk, thus may enjoy lower cost of capital. On the contrary, part of existing studies has shown that political connection worsens the problem of asymmetric information and fosters agency cost. When political-connected company no longer receives government subsidies, the performance will be deteriorated more than unpolitical-connected one. Thus, firm's political connection incurs cost in terms of higher risk and higher cost of borrowing funds. To construct measures of political connection, the number of political news is screened and computed from the Parliamentary Library of the Legislative Yuan (News and Knowledge Management System) by the intersection with the name of listed companies, board chair, party, president and party chair. Cost of bank loan is measured by average interest rate, highest interest rate and lowest interest rate. Empirical result generally shows that higher degree of political connection is corresponded to lower interest rates on bank loans. Robustness check under Heckman's two stage estimation to correct for sample selection bias, quantile regression estimation and lagged explanatory variables to address contemporaneous correlation drive similar result. Totally, principal outcome supports the view that banks favor political-connected firms.