民營化與經理人薪酬對聯貸條件之影響

全文

(1)國立政治大學財務管理研究所博士論文. 立. 政治大. ‧ 國. 學 ‧. 民營化與經理人薪酬對聯貸條件之影響. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. 指導教授：張元晨博士研究生：吳周燕. 撰. 中華民國 104 年 6 月.

(2) 謝辭. 這篇論文的完成，首先要感謝我的指導教授張元晨老師，在研究的過程中提供我許多寶貴的意見，也從張元晨老師身上學到了許多人生與學習的態度，看到了對研究的積極與熱愛，並且誠摯感謝口試委員林修葳老師、張士傑老師、陳思寬老師、以及盧秋玲老師在口試時針對我的博士論文提供了許多寶貴的建議，讓我的博士論文內容更完整。回首六年來的博士班生活，剛進來那段充實又忙碌的日子，感謝同學楚彬和美菁的幫忙與相伴，以及感謝玉美學姊、胤哲學長、偉劭學長、曉梅學姊、淑華. 政治大聽我訴苦，在過程中給我建議，我會記得大家一起在研究室討論課業的日子、吵立學姊、依婷學姊、淑惠學姊等學長姐，在修課、準備學科考與做研究的那段日子. 吵鬧鬧地一起慶生、元宵節一起吃湯圓、逛夜市、小旅行，也感謝博士班的學弟. ‧ 國. 學. 妹們、之寧、采彤助教的幫忙，研究的路上雖然孤獨，也曾經徬徨過，但是有大家在身邊陪伴，給我鼓勵與打氣，讓我的博士班生活更豐富。. ‧. y. Nat. 我還要感謝政大財管博士班及學校的教授們，在博士班的課程中汲取了豐富. sit. 的知識。另外，我還要感謝對我伸出援手並提供工讀機會的所有老師們，除了減. n. al. er. io. 輕了經濟負擔，也從中學習了準備授課的過程，以及期刊投稿的經驗。. Ch. i n U. v. 最後，我要感謝永遠在我身旁支持我的家人，在這段期間總是牽掛著我的博. engchi. 士學位，擔心著是否有吃飽，是否有好好睡覺，陪我走過這段博士班生涯，感謝妹妹在我念博士班的期間獨自承擔起家中的經濟重擔，感謝母親、外婆和妹妹對我在精神上的支持與包容，才能讓我得以心無旁騖地完成學業，你們的關心與支持是推動我更努力付出的動力，如今我終於從博士班畢業了。在此，謹將這篇論文獻給所有幫助過我的家人、朋友及老師們一同分享，感謝你們！. 吳周燕. 於. 2015 年 7 月 6 日. i.

(3) 中文摘要【第一部份論文中文摘要】本文使用 1993 年至 2007 年間 67 家實施部分及完全民營化公司為樣本，探討國營企業民營化後對聯貸條件之影響。實證結果顯示，民營化後若政府沒有控制權，則聯貸利差擴大，若完全民營化則會被銀行要求提供擔保品，此外，借款公司的信用風險比較會透過聯貸利差及相關費用反映出來，比較不會透過聯貸金額或到期日等條件反映。這些實證結果與本文假說一致，顯示政府保證效果在民營化過程中對於銀行借款是一項重要因素。. 政治大. 關鍵字：民營化、政府股權、聯合貸款、聯貸費用. 立【第二部份論文中文摘要】. ‧ 國. 學. 本部分論文以美國 1992 年至 2010 年有銀行聯貸之 1,560 家借款公司及 71 家銀行為樣本，探討銀行 CEO 薪酬對於銀行風險及聯貸條件的影響。實證結果. ‧. 顯示非銀行業借款公司 CEO 的 Vega 與股票報酬變動度、系統性風險及非系統性. y. Nat. 風險間為顯著負向關係，但是其 Delta 則與股票報酬變動度為顯著正向關係，隱. sit. 含 CEO 的 Vega 不會顯著提高股票報酬的風險，但 Delta 則會顯著提高股票報酬. n. al. er. io. 的風險。在探討 CEO 薪酬對聯貸條件影響時，結果顯示 Vega 及 Delta 與聯貸利. i n U. v. 率加碼間為顯著負向關係，但與財務限制個數間並無顯著關係存在。CEO 相對. Ch. engchi. 負債權益比與聯貸利率加碼為顯著負向關係，隱含當 CEO 相對負債權益比較高時，的確會傾向降低公司風險，因此聯貸利率加碼較低。銀行業借款的實證結果顯示，銀行 CEO 的 Vega 及 Delta 均會顯著提高股票報酬的風險，且與 Z-Score 有負向關係但並未具統計上顯著性。此外，銀行 CEO 的 Vega 均與聯貸利率加碼及財務限制條款之個數間為正向關係，Delta 則與聯貸利率加碼有正向關係，與財務限制條款之個數間為負向關係，但以上結果不具有統計上之顯著性。實證結果隱含，由於銀行業彼此間有資訊優勢，CEO 的風險性薪酬對銀行營運之影響反而無法如同一般產業之公司明顯反映於聯貸條件上。. 關鍵字：經理人薪酬、聯貸、聯貸利率加碼、財務限制條款 ii.

(4) 英文摘要【第一部份論文英文摘要】 Using a sample of 67 partially and fully privatized firms, this paper investigates the effect of privatization on loan conditions from 1993 to 2007. The empirical results show that loan spreads widen when governments have no control right after privatization. In addition, loans are more likely to be secured when firms are fully privatized. The empirical results also show that credit risk of borrowing companies is more likely reflected through price-term conditions. These results are consistent with the hypothesis that implicit government guarantee is an important factor for bank. 政治大. loans during the privatization process.. 立. Keyword: privatization, government ownership, syndicated loan, fee. ‧ 國. 學【第二部份論文英文摘要】. ‧. Using sample of syndicated loans by 1,560 borrowing companies and 71 banks. y. Nat. from 1992 to 2010, this paper investigates how CEO risk-taking incentives affect the. sit. syndicated loan conditions. For the results based on industrial firms’ syndicated loans,. n. al. er. io. we find that CEO’s Vega is negatively related with stock return volatility, systemic. i n U. v. risk, and idiosyncratic risk. Besides, there is a significantly positive relation between. Ch. engchi. CEO’s Delta and stock return volatility. The results imply that CEO’s Vega lowers stock return volatility. However, CEO’s Delta raises the risk of stock return volatility, which is consistent with the findings in previous literature. We also find that both CEO’s Vega and Delta have negative relation with loan spreads but insignificantly related to financial covenants. Moreover, CEO’s debt to equity ratio to firm’s leverage has negative impacts on loan spreads. This implies that higher CEO’s debt to equity ratio to firm’s leverage promotes CEO to reduce firm risk resulting in lower spreads. For the results based on banks’ syndicated loans, we provide evidence that both bank CEOs’ Vega and Delta have significantly positive relations on the risk of stock returns. In additions, they have insignificantly negative relation with banks’ Z-Score. We also find that higher CEOs’ Vega leads to higher loan spreads and more covenants. iii.

(5) Higher CEOs’ Delta are associated with higher spread but fewer covenants. The results imply that banks in the financial industry are familiar with each other, thus CEO risk-taking incentives may have insignificantly impacts on loan conditions relative to industrial companies.. Keyword: executive compensations, syndicated loan, spread, covenants. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. iv. i n U. v.

(6) 目錄緒論 ............................................................................................................................... 1. 【第一部份論文】 The Impact of Privatization on Syndicated Loan Conditions ................................. 3 1.. Introduction ............................................................................................................. 3. 2.. Literature Reviews and Hypotheses ....................................................................... 5 2.1 The Impact of Privatization on the Cost of Debt .............................................. 5 2.2 Literature on Syndicated loan ........................................................................... 6 2.3 Hypotheses........................................................................................................ 7. 政治大. 3. Data and Methodology............................................................................................... 8. 立. 3.1 Data and Sample Selection ............................................................................... 8. ‧ 國. 學. 3.2 Methodology..................................................................................................... 9 4. Empirical Results ..................................................................................................... 11. ‧. 4.1 Summary Statistics ......................................................................................... 11 4.2 Changes in Loan Conditions following Privatizations ................................... 13. y. Nat. sit. 4.3 Effects of Retained Government Ownership and Syndicated Loans ............. 16. al. er. io. 5. Conclusions .............................................................................................................. 18. v. n. References .................................................................................................................... 19. Ch. engchi. v. i n U.

(7) 【第二部份論文】經理人薪酬對聯貸條件的影響 ................................................................................. 44 1. 研究背景、動機與目的 ........................................................................................ 44 2. 文獻回顧與研究假說 ............................................................................................ 49 2.1 美國薪酬揭露義務之相關規範 ................................................................... 51 2.2 高階經理人薪酬與公司經營績效之相關文獻 ........................................... 52 2.3 高階經理人薪酬與公司風險之相關文獻 ................................................... 53 2.4 高階經理人薪酬與聯貸借款條件之相關文獻 ........................................... 54 2.5 研究假說 ....................................................................................................... 54 2.5.1 CEO 風險性薪酬對股票報酬風險之影響 ......................................... 55. 治政大 3. 研究樣本與研究方法 ............................................................................................ 57 立 3.1 非銀行業 ....................................................................................................... 57 2.5.2 CEO 風險性薪酬對聯貸條件之影響 ................................................. 56. ‧ 國. 學. 3.1.1 研究樣本與資料來源 ......................................................................... 57 3.1.2 研究方法 ............................................................................................. 58. ‧. 3.2 銀行業 ........................................................................................................... 60. y. Nat. 3.2.1 研究樣本與資料來源 ......................................................................... 60. sit. 3.2.2 研究方法 ............................................................................................. 60. n. al. er. io. 4. 實證結果 ................................................................................................................ 62. i n U. v. 4.1 非銀行業 ....................................................................................................... 62. Ch. engchi. 4.1.1 敘述性統計 ......................................................................................... 62 4.1.2 CEO 風險性薪酬對股票報酬波動風險之影響 ................................. 63 4.1.3 CEO 風險性薪酬對聯貸條件之影響 ................................................. 64 4.2 銀行業 ........................................................................................................... 66 4.2.1 敘述性統計 ......................................................................................... 66 4.2.2 銀行 CEO 風險性薪酬對銀行風險之影響 ....................................... 67 4.2.3 銀行 CEO 風險性薪酬對聯貸條件之影響 ....................................... 69 4.3 穩健性測試 .................................................................................................... 70 4.3.1 CEO 薪酬與聯貸之資料頻率不同 ..................................................... 70 4.3.2 Vega 與 Delta 之內生性問題 .............................................................. 70 5. 結論與未來研究建議 ............................................................................................ 71 vi.

(8) 附錄 1. 變數定義 ...................................................................................................... 74. 附錄 2. 經理人薪酬之 Vega 與 Delta 計算方式 ..................................................... 77. 附錄 3. 相關係數表－非銀行業 .............................................................................. 79. 附錄 4. 相關係數表－銀行業 .................................................................................. 85. 附錄 5. 非銀行業 CEO 薪酬之相對負債權益比對聯貸條件之影響 .................... 89. 附錄 6. 非銀行業 CEO 之風險承受性薪酬對聯貸利率加碼之影響（年資料） .... ....................................................................................................................... 91. 附錄 7. 非銀行業 CEO 之風險承受性薪酬對聯貸之財務限制條款個數之影響（年資料）........................................................................................................... 93. 附錄 8. 銀行業 CEO 之風險承受性薪酬對聯貸條件之影響（年資料） ............ 95. 附錄 9. 非銀行業 CEO 薪酬與股票報酬風險之關係—兩階段迴歸模型 ............ 97. 學. ‧ 國. 附錄 10. 治政大銀行業 CEO 之風險承受性薪酬與股票報酬波動度間之關係—兩階段迴立歸模型....................................................................................................... 101. 參考文獻 ................................................................................................................... 104. ‧. n. er. io. sit. y. Nat. al. Ch. engchi. vii. i n U. v.

(9) 表目錄【第一部份論文】 Table 1 The definitions of variable ........................................................................... 22 Table 2 Sample descriptions ..................................................................................... 23 Table 3 Distribution of privatizations ....................................................................... 26 Table 4 Summary statistics for government ownership and syndicated loan conditions ..................................................................................................... 28 Table 5 Price terms and non-price terms of syndicated loan conditions, 1980-2010 ... ...................................................................................................................... 29. 政治大 development of countries ............................................................................. 31 立. Table 6 Differences in means tests for price terms and non-price terms by. Table 7 Differences in means tests for price terms and non-price terms by. ‧ 國. 學. government ownership ................................................................................. 33 Table 8 Effects of retained government ownership and price terms of syndicated. ‧. loans conditions, Whole sample ................................................................... 36. sit. y. Nat. Table 9 Effects of retained government ownership and non-price terms of. io. n. al. er. syndicated loans conditions, Whole sample ................................................. 41. Ch. engchi. viii. i n U. v.

(10) 表目錄【第二部份論文】表1. 敘述性統計－非銀行業................................................................................. 108. 表 2 CEO 薪酬與股票報酬風險............................................................................ 111 表 3 CEO 薪酬與股票報酬風險在不同時期間之變化........................................ 114 表 4 CEO 薪酬與股票報酬風險之關係................................................................ 117 表 5 CEO 薪酬與公司規模之關係........................................................................ 119 表 6 CEO 薪酬對聯貸利率加碼之影響................................................................ 121 表 7 CEO 薪酬對聯貸之財務限制個數的影響.................................................... 123 表8 表9. 政治大銀行業與非銀行業之樣本特性比較............................................................. 128 立敘述性統計－銀行業..................................................................................... 125. 表 10 1992-2010 年 CEO 薪酬與銀行風險之概況 .............................................. 133. ‧ 國. 學. 銀行業 CEO 之風險承受性薪酬在不同時期之關係 ................................. 136. 表 12. 銀行業 CEO 之風險承受性薪酬與股票報酬波動度間之關係................. 139. 表 13. 不同聯貸時期下之銀行業 CEO 風險承受性薪酬與聯貸條件、銀行風險的. ‧. 表 11. sit. y. Nat. 關係 ................................................................................................................ 141. 表 15. 銀行業 CEO 之風險承受性薪酬對聯貸條件之影響................................. 144. n. al. er. 銀行業 CEO 之風險承受性薪酬與銀行風險在不同聯貸期間之差異..... 142. io. 表 14. Ch. engchi. ix. i n U. v.

(11) 圖目錄【第二部份論文】圖 1 1992-2010 年 CEO 總薪酬 ............................................................................ 146 圖 2 1992-2010 年 CEO 的 Vega ........................................................................... 147 圖 3 1992-2010 年 CEO 的 Delta ........................................................................... 148. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. x. i n U. v.

(12) 緒論本論文架構主要包含兩個部分，第一部分係探討全球民營化對民營化公司之聯貸條件的影響，第二部分則包含兩篇論文，係探討經理人薪酬對聯貸條件的影響，當一般產業（非銀行業）與銀行業的公司，其經理人同樣在面臨激勵性薪酬時，一般產業與銀行在公司的風險及聯貸條件的影響上是否有所不同。雖然第一部分與第二部份論文皆與聯貸有關，但內容差異性較大，因此目錄上分開呈列，且第一部分論文已發表於國內期刊《中山管理評論》。第一部份與第二部份論文皆是探討公司向銀行團聯貸時，借款條件的差異。. 治政於不少民營化公司來自歐洲及開發中國家，因此此部份論文之聯貸資料來自大立 Dealogic 資料庫；第二部份論文的借款公司則是來自美國 S&P1500 指數的成分第一部分論文的借款公司為過去在全球進行部分民營化或完全民營化的公司，由. ‧ 國. 學. 股公司，Execucomp 資料庫提供了這些公司詳細的 CEO 及高階經理人薪酬資料，此部份論文之樣本包含了銀行及非金融業的借款公司聯貸資料，並分別進行討論，. ‧. 由於第二部份論文係以美國上市公司為主體，Dealscan 資料庫涵蓋較為完整的美. y. Nat. 國借款公司聯貸資料，並且提供其他更為詳盡的聯貸資訊，因此第二部分的聯貸. n. al. er. io. 進行說明。. sit. 資料選擇使用 Dealscan 資料庫，文獻中針對 Dealogic 及 Dealscan 資料庫之差異. Ch. i n U. v. 第一部分論文中，約有將近 30%左右的民營化樣本來自於銀行及保險公司等. engchi. 金融業，而且在資料整理過程中，我們發現金融業者在民營化過程中或民營化後的併購活動，無論擔任主併或被併公司，似乎相對於其他產業的民營化公司來得多。在研究第二部分論文前，我們分析 1982 年至 2010 年美國的借款公司於 Dealscan 資料庫中的所有聯貸資料，若按 Fama-French 產業分類分成 12 個產業，來自金融公司借款共 16,596 筆，占整體 12 個產業，共 142,215 筆聯貸中的 11.67%，其聯貸的數量均高於其他個別產業，顯示金融業雖然由於產業特性，通常被排除在某些研究論文的樣本之外，但事實上仍是聯貸市場中重要的參與者。在第二部分論文中，將同時研究銀行業與非銀行業的經理人薪酬對聯貸條件之影響，並獨立探討銀行業經理人薪酬對聯貸條件之影響，其理由除了上述原因外，可再歸納出以下四點。第一，過去探討 CEO 薪酬的相關文獻多將金融業剔除，而研究銀行業的文獻則著重於銀行業 CEO 薪酬對銀行風險的影響，少數探討 CEO 薪酬與 1.

(13) 公司借款條件間關係的文獻（Dezso and Ross, 2012; Anantharaman, Fang, and Gong, 2013）則僅著重於非金融產業的公司。第二，銀行業向來是各國管制最為嚴格的行業，銀行必須在符合監管條件下營運以獲得利潤。第三，銀行係透過高度槓桿操作以獲得利潤，經理人薪酬對銀行營運、風險及借款條件的影響程度，可能與一般產業公司有所差異。第四，銀行業高階經理人薪酬結構與一般產業公司有很大差異。未來在後續的研究中，嘗試結合金融業民營化、經理人薪酬、併購活動或銀行聯貸之相關議題進行研究。第一部分係研究全球的國營企業民營化後，其聯貸條件在民營化前後之改變。此部分論文使用 1993 年至 2007 年間 67 家實施部分及完全民營化公司為樣本，實證結果顯示，民營化後若政府沒有控制權，則聯貸利差擴大，若完全民營化則. 政治大. 會被銀行要求提供擔保品，此外，借款公司的信用風險比較會透過聯貸利差及相. 立. 關費用反映出來，而非透過聯貸金額或到期日等條件反映，其實證結果隱含政府. ‧ 國. 學. 保證效果在民營化過程中對於銀行借款是一項重要因素。. 第二部分論文以 1992 年至 2010 年有銀行聯貸之 1,560 家借款公司及 71 家. ‧. 銀行為樣本，探討非銀行業及銀行業 CEO 薪酬對於公司風險及聯貸條件的影響。. sit. y. Nat. 銀行業經理人的 Vega 及 Delta 均會顯著提高銀行的股票報酬風險，與過去文獻一致；非銀行業經理人的 Vega 則會顯著降低股票報酬風險，但 Delta 則會顯著. io. n. al. er. 提高股票報酬風險，與過去文獻有所差異。進一步探討經理人激勵性薪酬對聯貸. i n U. v. 條件的影響時，我們發現當銀行業經理人的 Vega 越高時，則會使得聯貸利率加. Ch. engchi. 碼的利差及財務限制條款個數增加，但 Delta 與聯貸利率加碼間存在正向關係，與財務限制條款個數間有負向關係，然而上述結果在統計上並不具有顯著性，此結果可能是因為銀行間對於彼此都相當了解，因此經理人薪酬對聯貸條件的影響並不明顯。非銀行業的實證結果則與預期不一致，經理人的 Vega 及 Delta 越高時，聯貸利率加碼反而會顯著較低，但與財務限制個數間並無顯著關係存在。 CEO 相對負債權益比與聯貸利率加碼為顯著負向關係，此實證結果與文獻一致，隱含當 CEO 相對負債權益比較高時，的確會傾向降低公司風險，因此聯貸銀行會將利率加碼調低。. 2.

(14) 【第一部份論文】. The Impact of Privatization on Syndicated Loan Conditions 1. Introduction Privatization has been a well established method of ownership transfer by the government for state-owned enterprises (SOEs) to the private sector after the very first successful case by the British government back in 1979. As of the end of year 20121, privatization has taken place in 23 developed countries and 130 developing countries, mostly for its significant characteristic of improving profitability and. 治政 government guarantee as a result of privatization may 大simultaneously increase the 立 default risk of privatized firms and hence results in an increase in the cost of debt. For operation efficiency as discussed in previous literature. However, the decreasing. ‧ 國. 學. example, on April 1 2013, Jean-Cyril Spinetta, the CEO of Air France-KLM, argued that airlines in European Unions would face the threat of competition if the. ‧. restrictions on Gulf airlines flying into European airspace were removed. Gulf airlines,. y. Nat. Etihad, Dubai’s Emirates, and Qatar Airways rapidly expand their long haul tour and. sit. compete with European airlines in recent years. Their rival European airlines. n. al. er. io. criticized that government-owned airlines have lower financing costs and enjoy subsidies paid by the government.. Ch. engchi. i n U. v. The existing literature on financing costs for privatized firms mainly focus on the changes in bond spreads following privatizations (Faccio et al., 2006; Borisova & Megginson, 2011; Borisova et al., 2012). For example, Borisova & Megginson (2011) find that government ownership has a non-linear relationship with bond spread. Fully privatized firms have lower spreads while partially privatized firms have higher credit spreads. They argue that credit spreads are higher due to decreasing government guarantee in the privatization process. Previous literature suggests that government guarantee. is. associated. with. government. ownership,. especially. for. the. government-owned companies with poor operating performance and high leverage. 1. Privatization data are collected from Privatization Barometer and the World Bank Privatization Databases.. 3.

(15) Fan, Wong, and Zhang (2007) find that newly partially privatized firms wih political connected CEOs have underperform of the post-IPO stock returns. Borisova et al. (2012) find that higher government ownership results in higher cost of debt, which on the other hand is lower in the period of economic recession or firm distress when the effects of implicit government guarantee and default risk are stronger. Bank loan is one of the major financing sources for firms and this enables banks in accessing more information about firms than other lenders (Dass & Massa, 2011), and information asymmetry is also reflected by syndicated loan conditions (Sufi, 2007). Rajan & Winton (1995) document that bank directly monitor borrowing companies through loan conditions to improve the information transparency of. 政治大. borrowing companies. It also avoids borrowing companies investing in high risk. 立. investment plans, especially for those companies with low information transparency.. ‧ 國. 學. Moreover, Ahn & Choi (2009) find that the degree of improvement in corporate governance has a positive relationship with bank monitoring effect, especially for. ‧. firms with heavy reliance on bank loans. Roberts & Sufi (2009) find that 75% of syndicated loans are renegotiated prior to the stated maturity and rarely end up with. y. Nat. sit. distress or default. They show that most renegotiation contains maturity, loans, and. er. io. spreads based on the information about credit quality, investment opportunities, and. al. v i n important because fee paymentsCare higher than interest h e n g c h i U payments for some loans and n. collaterals. Finally, Berg et al. (2012) suggest that the fees in syndicated loans are. their empirical results show that relationship lending often charges lower fees.. This paper aims to investigate whether the syndicated loan conditions change significantly after privatizations. Previous literature on the impact of privatizations focuses on the changes of bond spread and cost of equity, and bank loan should be investigated since it is also a major financing source for firms. Thus this paper could serve to fill the gap in literature on the relationship between government ownership and financing costs for privatized firms. In order to investigate the effect of loan conditions by credit risk, we adopted the methodology of Berg et al. (2012) by re-categorizing loan conditions into different price terms (e.g., AISD, AISU, spreads, commitment fee, and facility fee) and non-price terms (e.g., loan amount, maturity, 4.

(16) secured, structure of syndicated loan). In this study, we examine 67 partially and fully privatized firms, which include 138 observations of privatization programs from 1993 to 2007 and 678 observations of syndicated loans from 1980 to 2010. There are 374 observations and 304 observations of loans before and after the initial privatization respectively. By excluding the requirement of collaterals, the empirical results show that credit risk of borrowers is reflected more on price terms of loan conditions but less on non-price conditions. We also find that an opposite interaction with negative and positive effect for companies with and without control right of government ownership after privatizations resulting in the insignificant results in price terms conditions. Lower spread and lower fees are associated with strong implicit. 治政大increase spreads and fees to privatizations. In contrast, we also find that banks would 立 reflect an increase in credit risk of borrowing companies due to weaker implicit government guarantee when borrowing companies retain government ownership after. ‧ 國. 學. government guarantee. Thus, the results in this paper support the implicit government guarantee effect after privatization in the syndicated loan market.. ‧. The rest of the paper is organized as follows. Section 2 presents the literature. y. Nat. sit. reviews. Section 3 describes the data and methodology. Section 4 provides empirical. n. al. er. io. results and Section 5 concludes.. Ch. engchi. i n U. v. 2. Literature Reviews and Hypotheses. 2.1 The Impact of Privatization on the Cost of Debt There is a large body of literature focuses on the impact of privatization on stock prices and operating performance for privatized firms, but the results on the issue of whether privatization affect the funding costs for privatized firms are sparse. Using data from 38 countries between 1987 and 2006, Ben-Nasr et al. (2012) find that government ownership decreases approximately 10% on average after privatization, and governments tend not to release their control rights in the year of privatization. They also find that the cost of equity increases dramatically as government ownership 5.

(17) decreases in the year of privatization. Faccio et al. (2006) find that companies obtain more government support if they have good political relationship. Brown & Dinc (2005) find that state-owned banks have lower default risk before privatization, but their profitability declines after privatization because of higher moral hazard due to implicit government guarantee (Borisova & Megginson, 2011; Borisova et al., 2012). Borisova & Megginson (2011) investigate the impact of privatization on credit spread of bonds by using partially and fully privatized firms in Europe from 2001 to 2009. They find that the relationship between government ownership and bond spreads is non-linear. On average, the credit spreads increase 0.75% when government ownership decreases 1%, but the credit spreads of fully privatized firms. 政治大. are lower than those of partially privatized firms. Their empirical results also suggest. 立. that lower credit spreads are associated with the speed of privatizations. Borisova et al.. ‧ 國. 學. (2012) find that higher government ownership leads to lower credit spread during financial crisis period, while higher government ownership leads to higher credit. ‧. spreads in normal period. Their empirical results suggest that the implicit government guarantee leads to lower cost of debt only in the period of economic recession.. y. Nat. sit. Borisova et al. (2012) show that the shares held by central and domestic governments. al. er. io. are associated with lower credit spread, while shares held by sovereign wealth funds. v. n. and foreign government are associated with higher credit spreads.. Ch. 2.2 Literature on Syndicated loan. engchi. i n U. We examine the impact of privatization on syndicated loan conditions since bank loan is one of major financing sources for companies. Prior research documents that banks have private information about borrowing companies that bondholders do not have (Dass & Massa, 2011). Thus, the information asymmetry between banks and borrowing companies is minor than that between bondholders and borrowing companies. Literature also shows that the firms with higher information transparency have lower ex ante information risk so they often have better loan conditions (Diamond & Verrecchia, 1991; Baiman & Verrecchia, 1996). Roberts & Sufi (2009) find that lenders have the right to increase loan maturity and terminate the unused 6.

(18) revolving credit facilities in times of technical default. Berg et al. (2012) document that 80% of U.S. syndicated loans contain more than one type of fees, and the amount of fee payments is larger than interest payments in some loans. Previous literature shows that higher information asymmetry and moral hazard result in less favorable loan conditions (Diamond, 1984). Sufi (2007) suggests that information asymmetry also affects the lender structure of syndicated loans (e.g., structure of lead and participant banks, and lead bank shares). Focarellia et al. (2008) find that lead banks increase their shares to reduce the agency problem for borrowers with higher information asymmetry. Both loan conditions and structure of syndicate reflect operations, credit risk, and degree of information trenchancy of borrowing. 政治大. companies. This paper investigates the impact of privatization on syndicated loan. 立. conditions. Loan contract terms are used to compare the changes in loan conditions. ‧. ‧ 國. 2.3 Hypotheses. 學. before and after privatizations.. y. Nat. Borisova & Megginson (2011) show that there is a non-linear relation between. sit. government ownership and the cost of bond yields. Borisova et al. (2012) also find. er. io. that lower cost of debt is associated with a reduction in default risk through increasing. al. n. v i n C stronger in the period implicit government guarantee is h e n g c h i U of economic recession, but. implicit government guarantee in the period of economic recession. This implies that. lower government ownership promotes operating performance in the normal period.. Privatizations not only enhance operation efficiency but also decrease the government guarantee. Thus, what the actual impacts of privatizations and how privatizations affect operation efficiency and the government guarantee are empirical question. The main hypothesis of this paper is that the price and non-price terms of loan conditions deteriorate due to less implicit government guarantee when governments release their control right for privatized firms after privatization. Lower government ownership means weaker implicit government guarantee. The credit risk for privatized firms will increase if the government release control right 7.

(19) after privatization; therefore, loan conditions are expected to be higher and privatized firms is expected to provide collaterals.. 3. Data and Methodology 3.1 Data and Sample Selection We compile our sample of syndicated loans from 1980 to 2010 using privatization and Dealogic databases. Our sample consists of firms with initial privatization over 1993-2007 and with syndicated loans both before and after the initial privatization.. 立. 政治大. Privatized data is obtained from Privatization Barometer database and the World. ‧ 國. 學. Bank Privatization database. Privatization Barometer database provides the information on European state-owned firms from 1977 to 2011. The World Bank. ‧. Privatization database provides the information on privatized state-owned enterprises. sit. y. Nat. in developing countries from 1980 to 2010. We also obtain the information of undisclosed privatized programs from both annual reports of sample firms and SDC. io. n. al. er. Merger and Acquisition Database. Syndicated loan data are obtained from Dealogic. i n U. v. database which contains large portions of loans for borrowing companies in Europe. Ch. engchi. and in developing countries. Percentage of shares held by government is an important variable in this paper. We hand-collected this data from several data sources, which include Privatization Barometer database, SDC Merger and Acquisition Database, Datastream, Worldscope Bankscope and OSIRIS, and various firms’ annual reports and websites. Accounting data are mainly obtained from OSIRIS and COMPUSTAT Global Vantage databases. Countries are classified by the World Bank definition according to the level of development. Our initial sample contains 67 partially and fully privatized firms, which include 138 observations of privatization programs and 720 observations of syndicated loans. We winsorize the top and the bottom 1% observations according to loan amount or 8.

(20) loan spreads. We also exclude 3 observations with incomplete government ownership data after privatizations. The final sample consists of 67 partially and fully privatized firms, which include 138 observations of privatization programs and 678 observations of syndicated loans. We obtain both loan conditions before and after privatizations. Borisova & Megginson (2011) compile a sample of 1,651 observations of credit spreads of bonds issued by 60 privatized firms over 2001 to 2009. Moreover, their sample firms must have at least one privatization in the period and are represented in Privatization Barometer database. Our sample contains syndicated loans both before and after the privatization. The number of sample firms in this paper is comparable to the number. 政治大. of sample firms in Borisova & Megginson (2011).. 立. 3.2 Methodology. ‧ 國. 學. Following Berg et al. (2012), we define all-in-spread-drawn (AISD) as the sum. ‧. of spread and facility fee, measured as lending costs of drawn. All-in-spread-undrawn. y. Nat. (AISU) is defined as the sum of facility fee and commitment fee, measured as lending. n. al. Ch. er. io. credit facilities, unlike the usual payments for term loan.. sit. costs of undrawn. Moreover, facility fee and commitment fee are paid for revolving. i n U. v. In order to investigate the impact of privatizations on loan conditions, we follow. engchi. Borisova & Megginson (2011) using a multivariate regression framework: 𝐿𝑜𝑎𝑛𝑖,𝑡 = 𝛼 + 𝛽1 𝐺𝑜𝑣𝑡50𝑖,𝑡 + 𝛽2 𝐿𝑜𝑎𝑛𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑖,𝑡 + 𝛾𝑋𝑖,𝑡−1 + 𝛿𝑌𝑒𝑎𝑟𝑡 + 𝜀 (1) 2. 𝐿𝑜𝑎𝑛𝑖,𝑡 = 𝛼 + 𝛽1 𝐺𝑜𝑣𝑡ℎ𝑜𝑙𝑑𝑖,𝑡 + 𝛽2 (𝐺𝑜𝑣𝑡ℎ𝑜𝑙𝑑𝑖,𝑡 ) + 𝛽3 𝐿𝑜𝑎𝑛𝐶𝑜𝑛𝑡𝑟𝑜𝑙𝑖,𝑡 + 𝛾𝑋𝑖,𝑡−1 + 𝛿𝑌𝑒𝑎𝑟𝑡 + 𝜀. (2). where Loan denotes the loan conditions, including price terms and non-price terms. Price terms contains AISD, AISU, spread, and commitment fee. Non-price terms include loan amount and maturity2. Govt50 is a dummy variable, taking a value of one 2. We also use other non-price terms, Number of lead banks and Number of participants, as dependent variables; however the results of main exogenous variables, Govt50, Govthold, and Govthold2, are. 9.

(21) if the percentage of government ownership is higher than 50% 3 after the initial privatization and zero otherwise. Govthold is defined as shares held by central government, local government, and agencies of government (e.g., Ministry of Finance and Ministry of Economic Affairs) divided by total capital. (Govthold)2 is the square of government ownership. Prior literature defines government ownership as common shares held by government divided by outstanding shares. There are three reasons which we use this definition. First, some data sources only provide basic ownership information. For instance, a lot of annual reports only disclosed the value of shares held by shareholders divided by total shares, but they did not disclose the value of common shares held by shareholders divided by common shares. Second, privatized. 治政大government typically holds shares of government-owned enterprises. For example, 立 both common shares and preferred shares with special rights. Some governments hold. firms are from various countries and some governments use different methods to hold. ‧ 國. 學. only one preferred shares (e.g., golden share) with special rights after full privatization. Common shares, preferred share, or both are sold by government during. ‧. privatized process. For instance, shares or assets of privatized firms could be sold. y. Nat. privately to specific buyers. LoanControl denotes the set of control variables of. sit. syndicated loans, including loan types, number of syndicate lenders, and number of. n. al. er. io. participants. X is other control variables, including privatized variables, firm character. i n U. v. factors, industry factors, and country level variables. Other privatized variables cover. Ch. engchi. immediate privatization and times of privatizations. Firm character factors include Size, ROE, Leverage ratio and interest coverage. Industry factor is Bank, which takes. 3. insignificant. In other words, government ownership does not have significant relation with Number of lead banks and Number of participants. We define the government have control right if more than 50% of shares held by government according to few literature’s definitions. For instance, defined by Megginson et al. (1994), D’Souza & Megginson (1999), and D’Souza et al. (2007), the government releases control of operations if less than 50% of shares held by government. Bortolotti & Faccio (2009) document that government have control right if controlling shareholders which percentage of shares held is more than 10% are central government, local government, or other government agencies. Otherwise, the government do not have control right. Both Borisova & Megginson (2011) and Borisova et al. (2012) do not explicitly define the control right of government, they only observe whether the government hold golden share. However, considering the definitions provided by Bortolotti & Faccio (2009) and other literature on ownership (La Porta et al., 1999; La Porta et al., 2002), shareholder has control right if ownership is more than 10% or 20% depending on the share diversification. We find that there are insignificant difference in classifications of 10%, 15%, 20%, and 25%; therefore, the definition which the government have control right if more than 50% of shares held by government is used in this paper.. 10.

(22) a value of one if the sector of privatized firms is bank sector, and zero otherwise. Country level variables is the development of countries defined as dummy variable that takes the value of one if privatized firms are in developing countries, and zero otherwise. Year denotes year fixed effect. The detailed definitions are summarized in Table 1. [Insert Table 1 here]. 4. Empirical Results 4.1 Summary Statistics. 政治大. Table 2 presents the annual distributions of privatization programs and. 立. syndicated loans. Panels A and B show the distributions of privatization programs and. ‧ 國. 學. syndicated loans. The sample contains 67 privatized firms, which include 15 banks and 52 non-banks, as shown in Table 2. There are 138 observations of privatization. ‧. programs and 678 observations of syndicated loans contributed by 67 privatized firms. Panel A shows that many privatization programs are initiated in 1990s, especially in. Nat. sit. y. 1997 and 1999. Moreover, the largest number of deals is in France (17 privatized. al. er. io. deals). Panel B shows that the largest number of syndicated loans are initiated in. n. 1990s (274 observations of syndicated loans) and the second largest subsample is. Ch. i n U. v. those loans came after 2000 (243 observations of syndicated loans). Besides, Indian. engchi. privatized firms have the largest number of syndicated loans (116 observations of syndicated loans). The second largest and the third largest number of syndicated loans are initiated in Russian Federation and China, each contributes 62 observations and 61 observations of syndicated loans. [Insert Table 2 here] In Table 3, we report the percentage of shares held by government in a privatization transaction. Panel A shows the results sorted by the year of privatizations. Privatized firms are separated into developed and developing countries, as shown in Panel B. In Panel A, privatizations seem to concentrate in 1990s 11.

(23) (52.17%). Overall, the mean (median) of percentage of shares held by government for sell is 25.89% (20.18%). The mean of percentage of shares held by government for sell is 34.54% in 1993, while the mean of percentage of shares held by government for sell is only 12.89% in 2002. Panel B shows that there are 74 observations and 64 observations of privatization deals from developed and developing countries, respectively. As for developed countries, France, Finland and Spain are the major three with the largest numbers of privatization deals. While for developing countries, the title comes to Russian Federation. The mean of percentage of shares held by government for sell in developed countries is 26.68%, and it is 1.7% more than the number in developing countries (24.98%).. 立. 政治大. [Insert Table 3 here]. ‧ 國. 學. Table 4 documents the summary statistics of government ownership and. ‧. syndicate loan variables. Panels A and B present the summary statistics of price terms. y. Nat. and non-price terms respectively. Panel C shows the summary statistics of the. sit. percentage of shares held by government and Panel D shows the summary statistics of. er. io. other variables. Table 4 reports that term loans is the major type of syndicated loans in. al. n. v i n C mean (median) of facility fee. For example, in Panel A, the h e n g c h i U fee and commitment fee. our sample and borrowing companies often do not pay facility fee and commitment. are only 6.83 basis points (0 basis point) and 1.21 basis points (0 basis point), respectively. Moreover, the mean of AISD and spread are 142.39 basis points and 141.19 basis points, and both median of them are 70 basis points. AISD is mainly contributed by spreads in our sample, so there is only a small difference between AISD and spread. Table 4 also shows that most of syndicate loans are participated by more than one lender. Panel B shows that the mean and median of sole lender loans are only 0.4 and 0 respectively. On average, the funding of syndicated loans of privatized firms is consisted of around 15.97 lenders, which contain 3.84 lead banks and 12.13 participant banks. 12.

(24) Panel C demonstrates that the average weight of government ownership for the whole sample is 60.57%. There are 365 observations (approximately 64.60%) in which government holds more than 50% of shares. Besides, the mean and median percentage of this group are 85.03% and 95.95% respectively. More than 30% of the 365 observations are coming from post-privatizations period. On the contrary, there are 200 observations where government ownership is less than 50%, and more than 90% of them are coming from post-privatizations period. [Insert Table 4 here] Table 5 demonstrates the annual mean value of both price terms and non-price. 政治大 column 8th-14th are results for non-price terms. In Table 5, AISU fluctuates but it 立 terms of syndicated loan conditions. Column 3th-7th are results for price terms, and. seems to decrease slightly over the years. For instance, AISU is 21.74 basis points in. ‧ 國. 學. 1981 and less than 5 basis points after 2005, excluding the period of financial crisis in 2008. Maturity also seems to be shorter over the years since 1994, where the average. ‧. of maturity is less than 5 years. In the same periods, loan amount increase gradually. sit. y. Nat. and the average amount is more than 220 million dollars after 1998. The average loan. io. er. amount in 2005, 2007, and 2010 are more than 1.3 billion. Number of lead banks, number of participants, and number of syndicate lenders also increase in the sample. n. al. Ch. i n U. v. period. This trend of loan structure may be associated with the larger loan amount in. engchi. the recent years because banks have to reduce their liquidity risks. Both AISD and spread tend to fluctuate as compared with other loan conditions in the sample period. [Insert Table 5 here] 4.2 Changes in Loan Conditions following Privatizations In this section, the changes on price terms and non-price terms of loan conditions following privatizations are discussed. Table 6 displays the difference in mean tests for price terms and non-price terms. Panel A is the result of whole sample. Panel B and Panel C are the results of developed and developing countries, respectively.. 13.

(25) In Panel A, for the price terms, only facility fee has a significant increase after privatizations, by 1.95 basis points at 5% significance. Maturity decreases by 2.24 years while both loan amount and secured dummy rise significantly by 539.70 million and by 0.08 respectively. Number of participants has an insignificant rise after privatization; however, both the number of lead banks and the number of syndicate lenders significantly increase by 4.37 and 5.44 respectively. There are similar results in Panel B and C. It is observed that the results in Panel A are mainly contributed by the loans in developing countries. For example, facility fee increases significantly by 2.83 basis points after privatizations in developing countries, while it only decreases insignificantly by 0.1 basis points in developed countries.. 立. 政治大. [Insert Table 6 here]. ‧ 國. 學. In Table 7, the sample following privatizations is separated into government with. ‧. control right, which the government ownership is more than 50%, and government. y. Nat. without control right, which the government ownership is less than 50%. Panel A is. n. al. Ch. er. io. from developed and developing countries, respectively.. sit. the mean difference tests for the whole sample. Panel B and Panel C are the samples. i n U. v. For price terms, it is clearly observed that both spreads and fees significantly. engchi. decline if government still has control right after privatizations. For instance, AISD, AISU, spread, and commitment fee decrease by 62.72 basis points, 4.40 basis points, 63.75 basis points, and 5.43 basis points respectively, after privatizations. In contrast, there are significant increases on spreads and fees if government loses control right after privatizations. For example, AISD, spread, and facility fee increase by 44.19 basis points, 41.61 basis points, and 2.58 basis points respectively, after privatizations; AISU and commitment fee insignificantly increase by 4.21 basis points and 1.23 basis points. The results imply that government ownership is associated with the implicit government guarantee, which dominates the changes in loan spreads and fees after privatizations. Both loan spreads and fees significantly decrease for those firms with more than 50% government ownership. In contrast, loan spreads and fees have 14.

(26) significant increases for those firms with less than 50% government ownership. The results in Table 7 also imply that privatization has opposite influences on both samples where government either with or without control right, therefore leading to insignificant changes shown in Table 6. The changes in non-price terms are consistent with the results in Table 6, where changes in loan amount, number of lead banks, and number of syndicate lenders are significantly positive, whether or not there are changes in government stakes during privatizations. The results shown in Panel B and Panel C of Table 7 are consistent with what its Panel A suggests. Besides, the results in Panel A of Table 7 are mostly contributed by the sample of developing countries. The changes of loan conditions in developed. 政治大. countries are smaller than the changes in developing countries.. 立. In summary, Table 7 reports that there is a positive relationship between. ‧ 國. 學. government ownership and the implicit government guarantee during privatized process. Then, the implicit government guarantee has positive influences on credit. ‧. Nat. y. risk, loan spreads, and loan fees for borrowing companies after privatizations.. er. io. al. sit. [Insert Table 7 here]. n. These difference-in-mean tests provide a preliminary support that banks tend to. Ch. i n U. v. adjust the price terms of syndicated loan conditions, which reflect the credit risk of. engchi. privatized firms, rather than the non-price terms. Banks also reach the same targets through setting the requirement of collaterals. However, the results of loan amount, maturity, number of syndicate lenders, and number of lead banks all have insignificant. differences. between. different. privatized. firms.. Furthermore,. privatization has opposite influences on price terms conditions for both subsamples where government either with or without control right, therefore leading to the insignificant changes for the whole sample. For instance, stronger implicit government guarantee make loan spreads and fees lower when government still hold over 50% of stakes of privatized firms after privatization. In contrast, sharp declines in government ownership also leads to the declines in the implicit government guarantee when the percentage of shares held by government is less than 50% after 15.

(27) privatization. Thus, banks tend to raise both loan spreads and loan fees, and require privatized firms providing collaterals. The results are partially consistent with Borisova & Megginson (2011)4 that there is a positive relation between government ownership and the implicit government guarantee, thus resulting in the increases of financing costs. Our results also imply that lenders of syndicated loans are attracted by credit risk of privatized firms while bondholders are attracted by operating performance after fully privatization. 4.3 Effects of Retained Government Ownership and Syndicated Loans In this section, we discuss the effects of privatization on retained government. 政治大 conditions versus government ownership for the whole sample. The price-term 立 ownership and syndicated loans. Table 8 reports the regression results of price term. dependent variables include AISD, AISU, spread, and commitment fee. In model 7,. ‧ 國. 學. there is a significantly negative relation between AISU and Govt50. AISU for firms with over 50% government ownership is lower than AISU for firms with below 50%. ‧. government ownership at 6.6 basis points. Moreover, there is a significantly negative. sit. y. Nat. relation between Govt50 and commitment fee (in Model 19), and this relation is still. io. er. strong even through adding firm character control factors (in Model 20). In most of the models, there are significantly negative relations among Govthold and AISD,. n. al. i n U. v. AISU, spread, and commitment fee. Furthermore, Govthold2, as an important variable. Ch. engchi. introduced by Borisova & Megginson (2011), is only significantly positive with AISD in both Model 5 and Model 6, and significantly positive with spreads in both Model 17 and Model 18. Partially privatized, as another important variable in Borisova & Megginson (2011), has significantly negative relations to AISD and spread, but has insignificantly positive relation to commitment fee after considering firm character control factors. This is inconsistent with Borisova & Megginson (2011). Finally, npriv. 4. In Borisova & Megginson (2011), Partially privatized and Fully privatized are both important variables. However, they do not have significantly impacts on loan conditions in this paper (not shown in this paper).. 16.

(28) has insignificantly positive relations with loan spreads and fees, hence meaning that speed of privatizations does not promote banks to lower spreads and fees. As for loan control variables, there is a positive relationship between Maturity and price terms, meaning that the longer the maturity, the higher the spreads and fees. This is inconsistent with Berg et al. (2012). Secured has both insignificantly positive relations with AISD and spread; however, Secured has a significantly negative relation to commitment in Model 19-24. This is partially consistent with Berg et al. (2012). In Model 3 and Model 15, AISD and spread of highly leveraged loans are higher than those of investment grade loans, at 280.2 basis points and 279.1 basis points respectively. In addition, AISD and spread of leveraged loans are higher than. 政治大. those of investment grade loans, at 73.02 basis points and 73.39 basis points. 立. respectively. It is in line with the expectation that both lenders of highly leveraged. ‧ 國. 學. loans and leveraged loans can bear higher default risks of companies than those lenders of investment grade loans. For other control variables, Developing countries is. ‧. significantly positive in Model 1-6 and Model 13-18. It means that AISD and spread in developing countries are higher than those in developed countries. Finally, AISU. y. Nat. al. [Insert Table 8 here]. er. io. sit. and commitment fee for banks are significantly lower than those for other industries.. n. v i n C h of non-price term Table 9 presents the regressions e n g c h i U conditions on government. ownership for the whole sample. The dependent variables include loan amount and. maturity. Govt50 has significantly negative relation to loan amount in Model 1, therefore meaning that loan amount is smaller when government has control right. Model 3, 5, and 6 represent that Govthold has significantly negative relation to loan amount, thus meaning that higher government ownership is associated with smaller loan amount. Both impacts of highly leveraged loans and leveraged loans on non-price terms are smaller than those impacts on price terms. The default risks for highly leveraged loans and leveraged loans are insignificantly higher than others, so loan amount are expected to be smaller, and maturity are expected to be shorter. There is a 17.

(29) significantly positive relation between Number of syndicate lenders and Loan amount5. It means that larger number of lead banks and larger number of participants would lead to larger loan amount. Finally, borrowing companies in developing countries have smaller loan amount and shorter maturity than those companies in developed countries. Following Borisova & Megginson (2011), we only consider post-privatization and the results are similar to Table 8 (not shown in this paper). Regression results also show that the impact of privatization on price-terms is stronger, and are consistent with Borisova & Megginson (2011), which support the implicit government guarantee hypothesis.. 立. 政治大. ‧ 國. 學. 5. Conclusions. ‧. This paper investigates the impact of privatization on loan conditions using 678 observations of syndicated loans from 1980 to 2010. The empirical results show that. Nat. sit. y. credit risk of borrowing companies is more likely to be reflected through price terms. al. er. io. conditions, as compared with non-price terms conditions except for the requirement of. n. collateral. We also find that privatization has opposite influences with negative and. Ch. i n U. v. positive effect for companies with and without control right of government ownership. engchi. after privatizations, resulting in insignificant results for the full sample. Besides, when governments have control right after privatizations, the price terms significantly decline and companies do not pledge assets as collaterals for loans. In contrast, there is a significant change on the price terms and companies need to pledge assets as collaterals for syndicated loans, in case governments release control right after privatizations. In addition, these results are mainly driven by loans from developing countries. Our results are partially consistent with Borisova & Megginson (2011), which asserted that there is positive relation between government ownership and the implicit government guarantee, thus resulting in increases in financing costs. In other 5. If Number of lead banks and Number of participants are the dependent variables (not shown in this paper), they are insignificantly related with Govt50, Govthold and Govthold2.. 18.

(30) words, loan spread and fees are lower if borrowing companies have over 50% of government ownership after privatizations. These results imply that implicit government guarantee is an important factor for bank loans during the privatization process.. References Ahn, S. and Choi, W., 2009, “The Role of Bank Monitoring in Corporate Governance: Evidence from Borrowers’ Earnings Management Behavior,” Journal of. 治政 Baiman, S. and Verrecchia, R. E., 1996, “The Relation 大 Among Capital Markets, 立 Financial Disclosure, Production Efficiency, and Insider Trading,” Journal of Banking and Finance, Vol. 33, No. 2, 425-434.. ‧ 國. 學. Accounting Research, Vol. 34, No. 1, 1-22.. Ben-Nasr, H., Boubakri, N., and Cosset, J., 2012, “The Political Determinants of the. ‧. Cost of Equity: Evidence from Newly Privatized Firms,” Journal of Accounting. y. Nat. Research, Vol. 50, No. 3, 605-646.. sit. Berg, T., Saunders, A., and Steffen, S., 2012, “The Total Costs of Corporate. n. al. er. io. Borrowing in the Loan Market: Don't Ignore the Fees.” Working paper, Bonn. i n U. v. University, New York University, and ESMT European School of Management and Technology.. Ch. engchi. Borisova, G. and Megginson, W. L., 2011, “Does Government Ownership Affect the Cost of Debt? Evidence from Privatization,” Review of Financial Studies, Vol. 24, No. 8, 2693-2737. Borisova, G. F., V. Holland, K. V., and Megginson, W. L., 2012, “Government Ownership and the Cost of Debt: Evidence from Government Investments in Publicly Traded Firms.” Working paper, Iowa State University, Bocconi University, and University of Oklahoma. Bortolotti, B. and Faccio, M., 2009, “Government Control of Privatized Firms,” Review of Financial Studies, Vol. 22, No. 8, 2907-2939.. 19.

(31) Brown, C. Q. and Dinc, I. S., 2005, “The Politics of Bank Failures: Evidence from Emerging Markets,” Quarterly Journal of Economics, Vol. 120, No. 4, 1413-1444. D’Souza, J. and Megginson, W. L., 1999, “The Financial and Operating Performance of Privatized Firms during the 1990s,” Journal of Finance, Vol. 54, No. 4, 1397-1438. D’Souza, J., Megginson, W. L., and Nash, R., 2007, “The Effects of Changes in Corporate Governance and Restructurings on Operating Performance: Evidence from Privatizations,” Global Financial Journal, Vol. 18, No. 2, 157-184. Dass, N. and Massa, M., 2011, “The Impact of a Strong Bank-Firm Relationship on. 治政大 Liquidity, and the Cost of Diamond, D. W. and Verrecchia, R. E., 1991, “Disclosure, 立 Capital,” Journal of Finance, Vol. 46, No. 4, 1325-1359. the Borrowing Firm,” Review of Financial Studies, Vol. 24, No. 4, 1204-1260.. ‧ 國. 學. Diamond, D. W., 1984, “Financial Intermediation and Delegated Monitoring,” Review of Economic Studies, Vol. 51, No. 3, 393-414.. ‧. Faccio, M., Masulis, R. W., and McConnell, J. J., 2006, “Political Connections and. y. Nat. Corporate Bailouts,” Journal of Finance, Vol. 61, No. 6, 2597-2635.. sit. Fan, J., Wong, T. J., and Zhang, T., 2007, “Politically Connected CEOs, Corporate. n. al. er. io. Governance, and Post-IPO Performance of China’s Newly Partially Privatized. i n U. v. Firms,” Journal of Financial Economics, Vol. 84, No. 2, 330-357.. Ch. engchi. Focarellia, D., Pozzolob, A. F., and Casolaroc, L., 2008, “The Pricing Effect of Certification on Syndicated Loans,” Journal of Monetary Economics, Vol. 55, No. 2, 335-349. La Porta, R., Lopez-de-Silanes, F., and Shleifer, A., 1999, “Corporate Ownership around the World,” Journal of Finance, Vol. 54, No. 2, 471-517. La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R., 2002, “Investor Protection and Corporate Valuation,” Journal of Finance, Vol. 57, No. 3, 1147-1170. Megginson, W. L., Nash, R., and Van Randenborgh, M., 1994, “The Financial and Operating Performance of Newly Privatized Firms: An International Empirical Analysis,” Journal of Finance, Vol. 49, No. 2, 403-452. 20.

(32) Rajan, R. and Winton, A., 1995, “Covenants and Collateral as Incentives to Monitor,” Journal of Finance, Vol. 50, No. 4, 1113-1146. Roberts, M. R. and Sufi, A., 2009, “Renegotiation of Financial Contracts: Evidence from Private Credit Agreements,” Journal of Financial Economics, Vol. 93, No. 2, 159-184. Sufi, A., 2007, “Information Asymmetry and Financing Arrangements: Evidence from Syndicated Loans,” Journal of Finance, Vol. 62, No. 2, 629-668.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 21. i n U. v.

(33) Table 1. The definitions of variable. Variable. Definition Takes a value of one if percentage of capital of the company owned by the government is higher than 50% after the initial privatization and zero otherwise Govthold Percentage of capital of the company owned by the central government, local government, government departments, and government agencies Private_yr A dummy variable that the value is one if the credit year is the year of privatization year, and zero otherwise npriv Times of privatizations Immediate privatization A dummy variable that the value is one if the company was fully privatized in one transaction Partially privatized A dummy variable that the value is one if the company has government ownership after the initial privatization, and zero otherwise Before the initial privatization Takes a value of one if the year is before the initial privatization of the company, and zero otherwise After the last privatization Takes a value of one if the year is after the last privatization of the company, and zero otherwise AISD All-in-spread-drawn (AISD) defined by Berg et al.(2012) is sum of spread and facility fee AISU All-in-spread-undrawn (AISU) defined by Berg et al. (2012) is sum of facility fee and commitment fee Spread All-in-pricing of the syndicated loan shown in basis point Commitment fee It is a fee paid on the unused amount of loan commitments Facility fee It is a fee paid on the entire amount of loan commitments Loan amount Total deal amount of syndicated loan shown in millions of dollars Ln(Loan amount) The natural log of total deal amount of syndicated loan Maturity Years of maturity of the syndicated loan Secured Takes a value of one if the company is requested by banks to provide a collateral for the syndicated loan, and zero otherwise Sole Lender A dummy variable that the value is one if number of lenders for the syndicated loan is one, and zero otherwise # of syndicate lenders Number of lenders for the syndicated loan # of lead banks Number of arrangers (lead banks) of the syndicated loan # of participants Number of participants (participant banks) of the syndicated loan Loan Type A dummy variable that the value is one if the tranche type is -Highly Leveraged highly leveraged, and zero otherwise Loan Type A dummy variable that the value is one if the tranche type is - Leveraged leveraged, and zero otherwise Developing countries A dummy variable that the value is one if privatized firms are in developing countries, and otherwise. Bank Takes a value of one if the sector of privatized firms is bank sector, and zero otherwise Size The natural log of total assets ROE Pre-tax profits / Shareholder equity Leverage (Total assets – Shareholder equity) / Shareholder equity Interest coverage Earnings after tax / Interest expenses Privatization variables are obtained from Privatization Barometer, The World Bank Privatization Database, firms’ annual reports, and SDC Merger and Acquisition Database. Syndicated loan variable are from Dealogic database. Government ownership data are collected from several data sources, including Datastream, Worldscope, Bankscope, OSIRIS, and companies’ annual reports. Financial data are mainly from OSIRIS and some are from COMPUSTAT Global Vantage database. The level of economic development of countries is classified by the definition of World Bank. Govt50. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 22. i n U. v.

(34) Table 2. Sample descriptions. Panel A: Distribution of privatizations Country. Number of Companies Total. Number of Observations. Banks Nonbanks. Total Banks Nonbanks 2007 2006 2005 2004 2003 2002 2001 2000 1999 1998 1997 1996 1995 1994 1993. Belgium. 1. 0. 1. 2. 0. Brazil. 8. 6. 3. 2. 1. Colombia. 1. 0. Croatia. 3. 2. Czech Republic. 1. 0. Finland. 5. 0. France. 7. 0. Ghana. 1. Greece. 0. 0. 0. 0. 0. 0. 0. 0. 2. 0. 0. 0. 0. 0. 0. 2. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 1. 0. 政治大 8. 0. 0. 0. 0. 0. 1. 0. 2. 0. 2. 1. 1. 0. 1. 0. 6. 立. 5. 1. 0. 2. 3. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 1. 1. 0. 1. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 1. 8. 4. 4. 1. 1. 1. 0. 1. 2. 0. 2. 0. 0. 0. 0. 0. 0. 0. 1. 1. 0. 1. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 5. 12. 0. 12. 2. 0. 2. 2. 0. 2. 0. 0. 0. 0. 0. 1. 2. 1. 0. 7. 17. 0. 17. 1. 1. 2. 2. 0. 1. 0. 1. 3. 3. 1. 0. 2. 0. 0. 0. 1. 3. 0. 3. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 1. 1. 0. 2. 0. 2. 4. 0. 4. 0. 0. 0. 0. 1. 1. 1. 0. 1. 0. 0. 0. 0. 0. 0. Hungary. 1. 0. 1. 8. 0. 8. 0. 3. 0. 3. 0. 0. 0. 0. 0. 1. 1. 0. 0. 0. 0. India. 6. 3. 3. 3. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 2. 0. 2. 2. 0. Indonesia. 2. 1. 1. 2. 0. 0. 1. 2. 0. 0. 0. 2. 0. 0. 0. 0. 0. 0. Italy. 3. 0. 3. 9. a l4. 4. 0. 1. 1. 1. 1. 1. 2. 1. 1. 0. 0. Kazakhstan. 1. 1. 0. 1. 1. 0. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. Pakistan. 1. 0. 1. 1. 0. 1. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. Philippines. 2. 0. 2. 3. 0. 3. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 2. 0. Poland. 1. 1. 0. 1. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. Portugal. 1. 0. 1. 4. 0. 4. 0. 0. 0. 0. 0. 0. 1. 0. 1. 1. 1. 0. 0. 0. 0. Romania. 1. 1. 0. 2. 2. 0. 0. 1. 0. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. ‧ 國. 0. 0. Nat 7 6. 0. Ch. 9 0 0 0 e0n g0 c h0 i U 0. 23. 學. 6. China. 0. y. 2. sit. 2. er. 0. ‧. 1. n. 1. io. Argentina. 1v i n0.

(35) Russian Federation. 6. 1. 5. 13. 1. 12. 1. 1. 0. 0. 0. 0. 0. 0. 1. 1. 2. 3. 1. 2. 1. Slovak Republic. 1. 0. 1. 3. 0. 3. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 2. 0. 0. Slovenia. 1. 1. 0. 1. 1. 0. 1. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. Spain. 5. 0. 5. 11. 0. 11. 1. 0. 1. 0. 0. 0. 3. 1. 3. 0. 0. 1. 1. 0. 0. Thailand. 2. 1. 1. 2. 1. 1. 0. 0. 0. 2. 0. 2. 2. 0. 67. 15. 52. 138. 25. United Kingdom Total. 0. 0. 0. 0. 0. 1. 0. 0. 0. 0. 1. 0. 2. 0. 0. 0. 0. 0. 0. 0. 0. 1. 0. 0. 0. 1. 0. 0. 113. 9. 9. 11. 9. 5. 8. 7. 8. 16. 9. 13. 8. 14. 11. 1. 政治大. 立. 學. Number of Companies. Brazil. 6. China. 3. Colombia. 1. Croatia. 3. Czech Republic. Nonbanks. 7. 0. 1. 4. 0. 6. 94. 2. 1. 61. 0. 1. 7. 1. 12. 2. al. iv 26 n Ch U 8e n g c h i 59 1. After 2000. 1990s. 1980s. 7. 0. 1. 6. 0. 0. 4. 1. 3. 0. 0. 94. 28. 30. 36. 53. 8. 11. 28. 22. 0. 7. 1. 1. 5. 6. 6. 7. 5. 0. 0. 8. 4. 4. 0. 0. 26. 9. 12. 5. 0. 59. 26. 21. 12. ‧. 1. Banks. 0. n. Belgium. Total. 1. io. 1. Nat. Argentina. Nonbanks. y. Banks. Number of Observations. sit. Total. er. Country. ‧ 國. Panel B: Distribution of syndicated loans. 1. 0. Finland. 5. 0. France. 8. 0. Ghana. 1. 0. 1. 8. 0. 8. 3. 5. 0. Greece. 2. 0. 2. 30. 0. 30. 2. 11. 17. Hungary. 1. 0. 1. 15. 0. 15. 7. 8. 0. India. 6. 3. 3. 116. 35. 81. 32. 65. 19. Indonesia. 2. 1. 1. 12. 4. 8. 6. 6. 0. 8. 5. 24.

(36) Italy. 3. 0. 3. 32. 0. 32. 7. 2. 23. Kazakhstan. 1. 1. 0. 10. 10. 0. 8. 2. 0. Pakistan. 1. 0. 1. 2. 0. 2. 1. 1. 0. Philippines. 2. 0. 2. 18. 0. 18. 8. 7. 3. Poland. 0. 6. 6. 0. 0. 2. 4. 7. 1. 3. 3. 1. 1. 0. Romania. 1. 1. Russian Federation. 6. 1. Slovak Republic. 1. 0. Slovenia. 1. Spain. 4. Thailand. 2. United Kingdom. 2. 政治大 1. 7. 0. 0. 6. 6. 0. 4. 2. 0. 5. 62. 5. 57. 52. 10. 0. 1. 8. 0. 8. 2. 6. 0. 1. 0. 8. 8. 0. 7. 1. 0. 0. 4. 28. 0. 28. 12. 10. 6. 1. 1. 25. 18. 7. 0. 19. 6. 0. 2. 7. 0. 7. 3. 4. 0. 立. 學. ‧. ‧ 國. 1. Portugal. n. al. er. io. sit. y. Nat. Total 15 52 158 520 243 274 161 67 678 This table documents the distributions of privatizations and syndicated loan in our sample by year. Panel A and Panel B represent the distribution of privatizations and distribution of syndicated loan respectively. Syndicated loan data are collected from Dealogic database for privatized firms listed mainly in the Privatization Barometer and the World Bank database. Syndicated loan data cover the period 1980 – 2010 and privatization programs cover the period 1994 – 2007. Only one privatization event in 1993 collected from SDC M&A database is Gazprom which is energy sector and is located in Russian Federation.. Ch. engchi. 25. i n U. v.

(37) Table 3 Distribution of privatizations Panel A: by year Year. Number of Observations. Percentage of government holding for sale Mean Median. 1993. 1. 15. 15. 1994. 11. 28.3273. 29. 1995. 14. 32.3214. 23.8. 1996. 8. 18.34. 13.655. 1997. 13. 34.54. 34.31. 1998. 9. 21.63. 17.83. 1999. 16. 26.5563. 24. 2000. 8. 33.305. 27. 2001. 7. 23.4. 25. 8. 12.8875. 11.6. 5. 28.938. 25. 9. 30.7322. 17.86. 2005. 11. 24.0336. 10.5. 2006. 9. 24.2211. 16.7. 2007. 9. 15.7222. 6.95. Total. 138. 25.8905. 2002 2003. 學 y. n. al. Czech Republic Finland. Ch. er. io. Belgium. Percentage of government holding for sale Mean Median. Number of Observations. Country. 20.175. sit. Nat. Panel B: by development of country. ‧. ‧ 國. 2004. 立. 政治大. 2. e n g c121h i. i n U. v. 43.5. 43.5. 7. 7. 24.8292. 15.5. France. 17. 25.2494. 17.5. Greece. 4. 17.2475. 15.395. Hungary. 8. 28.6325. 10.865. Italy. 9. 22.9378. 17.6. Portugal. 4. 22.59. 25.3. Slovak Republic. 3. 28.6667. 27. Slovenia. 1. 48.1002. 48.1002. 11. 24.4509. 25. 2. 77.55. 77.55. 74. 26.6784. 18.305. Spain United Kingdom Developed countries. 26.

(38) Argentina. 2. 25.5. 25.5. Brazil. 8. 38.4975. 38.615. China. 6. 16.0367. 14.17. Colombia. 1. 24. 24. Croatia. 8. 29.255. 25. Ghana. 3. 11.7333. 10.2. India. 7. 15.5543. 10.5. Indonesia. 6. 18.5583. 14.18. Kazakhstan. 1. 33.3. 33.3. Pakistan. 1. 73. 73. Philippines. 3. 22.1333. 20. Poland. 1. 30. 30. Romania. 2. 34.94. 34.94. 13. 24.7877. 24.84. 2. 17.6. 17.6. 64. 24.9795. 23.35. Russian Federation Thailand. 立. Developing countries. 政治大. ‧. ‧ 國. 學. This table shows the mean and median percentage of government holding for sale in a privatization deal. Panel A represents it by year and Panel B represents it depending on level of development of countries for private companies. Syndicated loan data are collected from Dealogic database for privatized firms listed mainly in the Privatization Barometer and the World Bank database. Syndicated loan data cover the period 1980 – 2010 and privatization programs cover the period 1994 – 2007. Only one privatization event in 1993 collected from SDC M&A database is Gazprom which is energy sector and is located in Russian Federation. The level of economic development of countries is classified by the definition of World Bank.. n. er. io. sit. y. Nat. al. Ch. engchi. 27. i n U. v.