影響新創公司長期表現之因素 - 以教育網絡為例 - 政大學術集成

全文

(1)國立政治大學社會科學學院經濟學系碩士論文 National ChengChi University Department of Economics, College of Social Sciences Master Thesis. 影響新創公司長期表現之因素-以教育網絡為例. 政治大 What立Factors Affect New Ventures’. ‧ 國. 學. Long-term Performance?. ‧. - A Case of Educational Network. n. er. io. al. sit. y. Nat. 項柏維. i n Po-Wei Hsiang C hengchi U. v. 指導教授：王信實、李文傑 Advisor：Shinn-Shyr Wang & Wen-Chieh Lee 中華民國 104 年 4 月 April, 2015.

(2) 摘要本研究提出了影響新創公司長期表現的架構，並以教育網絡連結作為衡量社群網絡連結的依據，探討其解決資訊不對稱問題的效果。本論文透過建構台灣的新創公司、創業投資基金和國家發展基金三個群體的經理人資料，計算出各教育網絡間的連結關係；依此可衡量各教育網絡關係對新創公司長期表現的影響。. 政治大. 實證結果顯示，新創公司和創業投資基金間的雙向教育網絡連結對新創公. 立. 司的長期表現有正向影響；然而當教育網絡連結和政府部門(本研究定義為國發. ‧ 國. 學. 基金)有關係時，其效果則延伸出許多議題，其一為政府部門與被投資公司產生. ‧. 教育連結時，其對被投資公司的長期績效表現效果為負向，顯示出產生政治連. Nat. io. sit. y. 結時政府部門的投資行為會受到影響。另一方面，當創業投資基金和政府部門. er. 產生教育連結時，其對被投資公司的長期績效表現為正，解釋了政府單位在選. n. al. i n C 擇創業投資基金時是以長期績效表現作為選擇依據。 hengchi U 關鍵字：教育連結、新創、創業投資、國發基金. 2. v.

(3) Abstract In this research, educational links are hypothesized as the key factors for solving the information asymmetry problem. By using the uniquely constructed dataset with key information of invested companies, venture capital firms and National Development Fund (NDF) in Taiwan, we quantify the effect of network linkages on invested companies’ long-term performance. Our empirical results show that a linkage between invested companies and VCs positively affects invested companies’ long-term performance. However, the political linkage formed with NDF has non-trivial two-fold. 政治大 long-term performance adversely, 立 the educational link between venture capital firms effects. While an educational link formed between invested companies and NDF impact. ‧ 國. 學. and NDF in turn has positive effects on the long-term performance. We thus confirm the existence of inappropriate investment decisions frequently taken because of the. ‧. political link between invested companies and National Development Fund.. sit. y. Nat. al. n. Fund. er. io. Key words: Educational Link, New Venture, Venture Capital, National Development. Ch. engchi. 3. i n U. v.

(4) INDEX CHAPTER 1 INTRODUCTION ......................................................................... 8 1.1 THE ORIGIN OF A NEW BUSINESS ........................................................... 8 1.2 THE SOCIAL NETWORK LINKAGES TO SOLVE INFORMATION ASYMMETRY PROBLEM .................................................................................................................................. 10 1.3 THE DISTINCTIVE POWER TO SOLVE INFORMATION ASYMMETRY PROBLEM .......... 12 1.4 THE LEVEL OF LINKAGES AND DISTINCTIVE POWER ............................................. 14 1.5 THE POLITICAL RELATED ISSUE IN TAIWAN .......................................................... 16 CHAPTER 2 LITERATURE REVIEW .............................................................17. 政治大. 2.1 INFORMATION ASYMMETRIC AND AGENCY PROBLEM ........................................... 17 2.2 DIFFERENT TYPES OF LINKAGES .......................................................................... 17. 立. ‧. ‧ 國. 學. 2.2.1 Educational link .......................................................................................... 17 2.2.2 Background link .......................................................................................... 18 2.2.3 Political link ................................................................................................ 18 2.3 DISTINCTIVE POWER............................................................................................ 18 2.3.1 Human capital ............................................................................................. 18 2.3.2 Reputation ................................................................................................... 19. y. Nat. io. sit. 2.3.3 Political power ............................................................................................ 19. n. al. er. CHAPTER 3 DATA SOURCES AND DESCRIPTIVE STATISTICS.................20. i n U. v. 3.1 SOCIAL NETWORK DATABASE .............................................................................. 21 3.2 DEPENDENT VARIABLES ...................................................................................... 21 3.3 INDEPENDENT VARIABLES ................................................................................... 22 3.4 OTHER CONTROL VARIABLES .............................................................................. 24 3.4.1 Invested company ....................................................................................... 24 3.4.2 Venture Capital ............................................................................................ 25. Ch. engchi. 3.4.3 Investment records ...................................................................................... 25 3.5 DESCRIPTIVE STATISTICS ..................................................................................... 25 CHAPTER 4 HYPOTHESIS AND REGRESSION MODEL SETTING ............27 4.1 HYPOTHESIS ........................................................................................................ 28 4.1.1 Invested companies & Venture capitals ...................................................... 28 4.1.2 Invested companies & National Development Fund .................................. 29 4.1.3 Venture capitals & National Development Fund ........................................ 30 4.2 REGRESSION MODEL SETTING .............................................................................. 31 4.

(5) CHAPTER 5 METHODOLOGY AND RESULTS .............................................32 5.1 METHODOLOGY .................................................................................................. 32 5.2 RESULTS .............................................................................................................. 32 5.2.1 Results – invested companies & VCs ......................................................... 33 5.2.2 Results – invested companies & NDF ........................................................ 33 5.2.3 Results – VCs & NDF................................................................................. 33 5.2.4 Results – different independent variables ................................................... 34 CHAPTER 5 CONCLUSION .............................................................................43 REFERENCES ...................................................................................................44 APPENDIX.........................................................................................................47. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 5. i n U. v.

(6) LIST OF TABLES TABLE 1: DEFINITION OF NEW VENTURE STAGES ........................................................... 9 TABLE 2: THE SOCIAL NETWORK DATABASE .................................................................. 21 TABLE 3: DEPENDENT VARIABLES ................................................................................. 22 TABLE 4: INDEPENDENT VARIABLES .............................................................................. 23 TABLE 5: CONTROL VARIABLES – INVESTED COMPANIES .............................................. 24 TABLE 6: CONTROL VARIABLES - VENTURE CAPITALS .................................................. 25 TABLE 7: CONTROL VARIABLES – INVESTMENT RECORDS ............................................. 25 TABLE 8: DESCRIPTIVE STATISTICS ............................................................................... 26 TABLE 10: TABLES OF THE ENGAGED PARTIES OF LINKAGES ......................................... 27 TABLE 11 RESULT 1 - INVESTED COMPANIES & VENTURE CAPITALS ............................. 35 TABLE 12 RESULT 2 - INVESTED COMPANIES & NATIONAL DEVELOPMENT FUND ........ 36 TABLE 13 RESULT 3 – VENTURE CAPITALS & NATIONAL DEVELOPMENT FUND ........... 37 TABLE 14 RESULT 4 – TYPE 1 LINKAGE ........................................................................ 38. 立. 政治大. ‧. ‧ 國. 學. TABLE 15 RESULT 5 – TYPE 2 LINKAGE ........................................................................ 39 TABLE 16 RESULT 6 – TYPE 3 LINKAGE ........................................................................ 40 TABLE 17 RESULT 7 – TYPE 4 LINKAGE ........................................................................ 41 TABLE 18 RESULT 8 – TYPE 1 ~ TYPE 4 LINKAGE ......................................................... 42. n. er. io. sit. y. Nat. al. Ch. engchi. 6. i n U. v.

(7) LIST OF FIGURES FIGURE 1: TREND OF NEW VENTURES IN TAIWAN ........................................................... 8 FIGURE 2: STAGES OF NEW VENTURES ........................................................................... 9 FIGURE 3: SOLUTIONS TO SOLVING INFORMATION ASYMMETRIC PROBLEM ................... 14 FIGURE 4: THE EFFECTIVENESS AREA OF LINKAGES....................................................... 15. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 7. i n U. v.

(8) Chapter 1 Introduction 1.1 The origin of a new business New startups play an important role in the process of economic growth. According to White Paper On Small And Medium Enterprises In Taiwan 2014, there were more than a million Small and Medium Enterprises (SMEs) in Taiwan in 20131, employing 8.588 million or 78.30 percent of the total number of employees in Taiwan at that time. Although these numbers seem quite large and might be an important factor of economic. 政治大 Only around 48.69% of new ventures last over 10 years. This high failure rate shows 立. growth, more than one half of them disappear before they mature and contribute to GDP.. the harsh environment investors in Taiwan confront. Investors like VCs or other funds. ‧ 國. 學. face unforeseen future because of uncertainties of earnings and sustainable cash flows.. ‧. Figure 1: Trend of New Ventures in Taiwan. n. er. io. sit. y. Nat. al. 1. Ch. engchi. i n U. v. According to the Ministry of Economic Affairs, the actual number of SMEs is 1,331,182 in Taiwan in 2013. 8.

(9) Startup companies grow amazingly fast and play the role of boosting the economy, no doubt. But what kind of difficulties do they face that induce a high failure rate. Most of them actually face the problem of funding. They are unable to raise funds at the early stages 2 of a new venture. Young startups with no creditworthiness have little possibility of getting mortgages from banks, and also have little chance to raise funds from specific financial institutions or general public. They hold abundant ideas and advanced technologies in vain, without proper investors. Thus, funding support from specialized sources like venture capital firms or a public sector institution like the. 政治大. sovereign wealth fund is important for their long-run development.. 立. Figure 2: Stages of New Ventures. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. i n U. v. Table 1: Definition of New Venture Stages. 2. Stage. Definition. Stage 1. Seed Stage. Stage 2. Startup Stage. Stage 3. Expansion Stage. Stage 4. Mezzanine Stage. Stage 5. Turnaround Stage. Early stage means seed stage, startup stage and expansion stage here 9.

(10) While startups do not always succeed in delivering the promised returns which are based on uncertain future growth, decisions to choose promising startups are full of uncertainty and information asymmetry. Investors like VCs or SWF have to put efforts in gathering information and evaluating each investment. Thus, what is needed is the ability to decide which target to invest in and to distinguish the probability of different returns materializing. To solve the information asymmetry problem and to avoid the agency problem, a series of platforms try to balance and exchange information from the two sides of the sectors. We call this kind of platform the social network.. 政治大 Social network links provide an effective way to deal with information asymmetry. 立. There are two main branches having their own explanations. First is the social network. ‧ 國. 學. linkages between the investing and invested companies. Literature has covered. ‧. educational, background and political linkages, all of which are based on the. sit. y. Nat. relationships among the members of parties engaged. The second issue is the distinctive. io. er. power that the managers or board of directors hold. Different kinds of power are held, like specialty managers, a unique human capital, or a company’s outstanding reputation.. n. al. i n Creturn These result in different levels of to investments. hengchi U. v. 1.2 The Social network linkages to solve information asymmetry problem The social network linkages reveal the relationships between the managers and board of directors within investing and invested companies. Some of them have the same educational background, some have the same intersection of past working experience in private sector and others may have some time overlapped in public sector. No matter which kind of linkages exist, we classify them in the category of social network links. 10.

(11) The first kind of linkages between investing and invested companies is the educational links among managers and board of directors of both sides of companies. The idea first came up in 2007 with a benchmark paper, Cohen (2007).3 Cohen (2007) proposed the idea that educational link serves as a bridge to investing and invested companies, and defined four types of linkages between managers within investing and invested companies. He proved that the same educational background leads to higher investments in invested companies and leads to a higher return. And of course, a stronger connection or linkage may bring out a relatively significant result. Thus, the. 政治大 It does help the investing company to find out investment-worthy targets. Here we 立. educational link becomes a good explanation to eliminate the information asymmetry.. venture market and expect to obtain a convincing outcome.. 學. ‧ 國. decide to replicate a similar method to measure the educational link in Taiwan’s new. ‧. y. Nat. The second kind of the linkages follow the concept of background link. As. er. io. sit. Saxenian (2002) has proved, managers’ past experience may bring some effect to the current position. For example, Saxenian (2002) proved that a common experience in. al. n. v i n Silicon Valley is positive to startC a new venture in Hsinchu h e n g c h i U and Shanghai. Also, if there’s. some connection between managers or board directors in terms of past experience, the information asymmetry problem may be solved. Taiwanese market is actually a good target to measure the background link. Unlike the giant market, it contains fewer companies and managers. Besides, most of the technology companies are located in the same cities, such as Hsinchu. All the people, companies, news and issues seem related to each other in this condition.. 3. Cohen 2007, 9. Cohen L., Frazzini A., Malloy C. (2008). The small world of investing: Board connections and mutual fund returns. Journal of Political Economy 116 (5), 951-979. 11.

(12) The third kind of linkages is political links, which is definitely the most important and powerful link. Once an enterprise has some linkages with the authorities, it might probably has a great chance of providing abnormal returns because it may obtain some profitable news which others don’t know. And most of the time, those who are privy to firsthand information before others tend to execute specific actions in advance and react to the industry policy ahead of their competitors. Since every industrial policy and any government related projects involve giant sums of money, its implications are worth inquiry and research. The main purpose here is to judge the unreasonable profit earned. 政治大. because of the social network system, and further uncover the veil of the complicated social network relationships.. 立. ‧ 國. 學. 1.3 The distinctive power to solve information asymmetry problem. ‧. Linkages among managers of investing and invested companies can effectively. sit. y. Nat. eliminate a portion of information asymmetry problem. But what if there are no links. io. er. between them? To choose invested companies which reveal some positive signals would be the closest approach. These positive signals not only contain firms’ features. al. n. v i n Ch but also the human capital’s characteristics. We classify them into three categories. First, engchi U the specialty of human capital. Second, the reputation of the invested companies, the underwriters and the VC. Third, the political patronage enjoyed by specific invested companies.. Under the human capital studies, educational background and past experience are the main ideas to discuss. Whether managers of the invested companies have MBA or a professional degree may cause different leading styles. Also, if the managers have some special professional expertise, it may signal some positive messages to potential investors. For example, managers with past experience of VCs or startup may signal a 12.

(13) profession in new venture. Managers with special financial backgrounds may take a specialized view of financial operations or management. Above all, the characteristics of managers not only reveal signals to investors but are also directly related to the standing of the company and its corresponding industries.. Besides human characteristics, reputation may also be a good measure of a firm’s characteristics. The invested companies’ own reputation may release some signals to its investors. A good reputation may generate a high goodwill and make it easier to raise. 政治大 past literature, some variables like VC market share may lead to a positive effect to 立 funds. VCs’ characteristics also have some impact on the companies’ performance. In. invested companies’ long run performance. Other variables such as VC age, VC capital. ‧ 國. 學. or IPO frequency may not have discernible effects on performance but are still widely. ‧. discussed. Apart from the VC, reputation of underwriters also affects performance. A. sit. y. Nat. good underwriter perfectly audits a firm’s financial condition and then decide to. io. al. n. who lack information.. er. underwrite a capital issue. Reputations of different parties provide a signal to parties. Ch. engchi. i n U. v. In addition to these two categories, political power might be the most powerful and valuable. Managers or members of board of directors may sometimes hold a concurrent post in public sector. In that sense, companies who hold some political power may gain the possibility to set some “rules” which are profitable for themselves. In this case, there is no doubt that generating a corresponding return is a piece of cake for those who hold political power on their own.. 13.

(14) 1.4 The level of Linkages and distinctive power To deal with the information asymmetry problem, we define two ways of finding solutions. First is the own distinctive power that a particular invested company has. It may reflect in characteristics the company has, including human capital, reputation of all parties concerned and the political power. But if there’s no power, then the second one, social network linkage is used to deal with the problem. The linkages include educational links, background links and political links. The closer the linkages appear, the more significant is the effect. The figure below shows the classification of these linkages and power.. 立. 政治大. Figure 3: Solutions to solving information asymmetric problem. ‧. ‧ 國. 學 er. io. sit. y. Nat. al. n. v i n From past experience, weC suppose the investingUparties choose the investment hengchi. targets following rationality which means the decisions are mainly based on invested companies’ performance. But if the performance cannot be seen or considered, the. managers have to make investment decisions based on educational link or even overdrive the effect of the performance. It means that managers choose investment targets based upon linkages with others rather than performance. Furthermore, the effect of political link may be the strongest. It drives not only the effect of educational links but also the performance-oriented behavior. In order to clarify the level of these linkages, we draw a level figure to show them. As the figures below, there are three situations showing the effectiveness of the linkages. 14.

(15) Figure 4: The effectiveness area of linkages. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 15. i n U. v.

(16) 1.5 The political related issue in Taiwan The political issue is always an unknown factor in literature, because many of the details and references are not available to the general public. But inside the black box of the government sector, we often hear of an amazing performance in National Development Funds. For example, the asset management of National Development Fund has had a 3700% growth in 25 years. How do they pick the investment targets and how to efficiently manage their huge positions? Is the profitability reasonable or just fit in the minimum requirement of legal compliance? It is an important issue to dig out. 政治大 general public paid. We here consider the National Development Fund as the political 立. how National Development Fund is managed, which is originally from the tax that the. sector because it holds the right to affect and maintain the operation of the market.. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 16. i n U. v.

(17) Chapter 2 Literature review 2.1 Information asymmetric and agency problem Akerlof (1970) first came up with the idea of information asymmetry and explained the situations that investors and invested companies encounter. Investors lack information and hence may face a possibility of moral hazard and adverse selection. Only the new ventures know the conditions and opportunities they face. In order to balance the availability of information between each party, both may try to obtain more information of the other with their respective resources, be it the educational. 政治大. background, past experience or any kind of linkages. Therefore, a series of publications. 立. in extant literature have tried to propose these ideas for solving the information. ‧ 國. 學. asymmetry problem.. ‧. 2.2 Different types of linkages. y. Nat. io. sit. 2.2.1 Educational link. n. al. er. Among all kinds of linkages, educational background may be the easiest to. i n U. v. measure because most of the managers’ educational qualifications are open to public. Ch. engchi. through annual reports. To practically measure the effect, Cohen (2007) defined 4 types of educational links which he thought explain the social network system. They are educational background in Type1-same school, Type2-same department in same school, Type3-time overlapping in same school and Type4-same department with a time overlapping in same school. Besides the investing and the invested parties, Hwang (2014) also tested the invested companies and the underwriters. He shows that not only the educational links between invested and investing parties matter, but all relevant links play a role in investment decisions.. 17.

(18) 2.2.2 Background link In addition to educational linkages, a background linkage which stands for managers’ past work experience may be another approach. Saxenian (2002) pointed out that a past work experience in Silicon Valley may lead to a better performance in founding a high-tech company.. 2.2.3 Political link The benchmark idea of political link first came up as the Suharto government got. 政治大 political links. It surprisingly proves that a higher closeness with Suharto might result 立. corrupted in the late 1990s. Fisher (2001) used Suharto Dependency Index4 to measure. in higher dependence of performance on the news which the authorities spread out.. ‧ 國. 學. Another stream of literature has put an effort on measuring the donations to the. ‧. authorities. McDonald IV (2013) used the data of donations from stock analysts and. sit. y. Nat. proved that higher donations lead to higher chance of making accurate estimation.. io. er. Published literature has also mentioned that companies usually sponsor some election candidates for large campaigns or outdoor advertising. Feijen (2008) and Michael (2010). al. n. v i n examined whether an enterpriseC sponsoring a specific U h e n g c h i candidate or party would lead to an abnormal return.. 2.3 Distinctive power 2.3.1 Human capital Zarutskie (2008) showed that having a manager or a director with VC background positively affects the performance of a new venture.. 4. Suharto Dependency Index is an index created by Castle Group in 1995 which measures the political links by classifying the relationship with Suharto into 5 categories by closeness. 18.

(19) 2.3.2 Reputation Except for the human capital an invested company holds, its own reputation or its funding by venture capital firms having their own reputations may also be a strong signal to reveal positive information. Gompers (1996) considered VC’s age as its reputation and discovered that a young VC may be in a hurry for an IPO in order to accelerate the accumulation of IPO cases. This may lead to misevaluation of the new venture. Nahata (2008) considered Cumulative IPO Market Share as the VC reputation and showed that it positively affects the possibility of an initial public offering. There. 政治大 invested company’s own reputation; Gompers and Lerner (1999) mentioned VC Capital, 立 are still many variables chosen in past literature in attempts to value VC reputation and. ‧ 國. 學. Gompers (1996) and Lee and Wahal (2004) mentioned IPO frequency, and Nahata (2008) mentioned accumulated IPO market share. Though the effects are not all. ‧. significantly positively related, they are all measures to deal with the information. sit. y. Nat. asymmetry problem. A new venture may choose a better VC with reputation for. io. al. er. rationality and finally reduce the agency cost from information asymmetry and enhance. n. the long term performance.. Ch. engchi. i n U. v. 2.3.3 Political power Pramuan (2010) showed that a company with real political power (maybe without a guarantee in law) has an incentive to support election of and hold the real legal political power. Further, the M/B value rises significantly after the election victory.. 19.

(20) Chapter 3 Data Sources and Descriptive Statistics In this study, we examine the social network between invested firms, venture capital firms and National Development Fund by educational links and test whether educational links play an important role in enhancing the long term performance of the invested firms. Based on the research of Lin (2014) and Ho (2014), we use the database they constructed for measuring the educational links of 390 start-up firms. We added more data and extended the database to the field related to political authority. Now the database includes 263 firms (including 40 venture capital, 222 invested companies and. 政治大. National Development Fund), 390 investing cases and a total of 11,327 individuals. We. 立. take political link, the linkages with NDF, as our key variables which we believe drive. ‧ 國. 學. the effects of educational links between invested companies and VCs.. ‧. The database constructed by Lin (2014) and Ho (2014) contained all VC. y. Nat. io. sit. investment cases based on the Thomson Financial SDC. They also include the. n. al. er. educational backgrounds of board directors and main managers, extracted from annual. i n U. v. reports from the Market Observation Posting System. Here, besides the original. Ch. engchi. educational background of those companies, our contribution to the database mainly focuses on two points. First, the background of VC managers. Second, the list of names of members of three committees of National Development Fund. We contacted the executive unit of National Development Fund in Oct. 2014 and obtained the members list of NDF’s three main decisive committees. We then put all the members of invested companies, VC and NDF together and calculated the educational link. Here we consider NDF as the political data.. 20.

(21) 3.1 Social network database Our database conducted as panel data. The format is as below. Table 2: The social network database Year. Investing. Invested. Dependent. Independent. Company. Company. Variables. Variables. Invest ID. 1. 112. Y. X. ⋮. 1. 1. 112. Y. X. 2011. 1. 1. 112. Y. X. 1952. 2. 2. 188. Y. X. ⋮. 2. 2. Y. X. 2011. 2. 2. 188. Y. X. 1952. 11. 3. 200. Y. 學. X. ⋮. 11. 3. 200. Y. X. 2011. 11. 3. 200. Y. ‧ 國. 1. 立. 政 188 治大. X. Nat. io. sit. y. ‧. 1952. n. al. er. 3.2 Dependent variables. Ch. i n U. v. To examine the long term performance of the investee companies, we use some. engchi. variables such as ROE/ROA, cash flow and cumulative abnormal return. The data of ROE, ROA and cash flow are directly downloaded from TEJ and Market Observation Post System. Cumulative abnormal return is our own calculation based on the FamaFrench three-factor model, as below. 𝑅 = 𝑅𝑓 + 𝛽1 (𝐾𝑚 − 𝑅𝑓 ) + 𝛽2 𝑆𝑀𝐵 + 𝛽3 𝐻𝑀𝐿 + 𝛼 Where R is the expected return, 𝑅𝑓 is the risk free rate calculated by “One month fixed deposit interest rate in First Bank”, 𝛽1 (𝐾𝑚 − 𝑅𝑓 ) is the premium of the market risk, 𝛽2 𝑆𝑀𝐵 is the premium for firm size, 𝛽3 𝐻𝑀𝐿 is the book value premium and 𝛼 stands for the abnormal return. Here we first run the linear regression and get 21.

(22) 𝛽1 , 𝛽2 and 𝛽3. Then we estimated the individual abnormal return. For robustness check, we use both monthly data and daily data for regression. Table 3: Dependent Variables. Variables. Sources. Meaning. ROE. TEJ, M.O.P.S.. Return on Equity. ROA. TEJ, M.O.P.S.. Return on Asset. Cash Flow. TEJ, M.O.P.S.. Cash Flow in certain year. CAR (monthly). TEJ, M.O.P.S.. Cumulated abnormal return, calculated by the. FAMA Model. model of FAMA three factors5.. CAR(daily). 治 abnormal return, calculated by the 政Cumulated 大. TEJ, M.O.P.S.. 立. FAMA Model. model of FAMA three factors6.. ‧ 國. 學 ‧. 3.3 Independent variables. Following Cohen (2007), we categorize educational links into 4 types. First, Type. y. Nat. io. sit. 1 means there exists a same school linkage. Second, Type 2 means that there exists a. n. al. er. same school and a same department linkage. Third, Type 3 means that there exists a. Ch. i n U. v. same school and a time overlap in educational background. Finally, Type 4 means there. engchi. exists a same school, same department and a time overlap in educational background. Under the concept of 4 types of educational links, we have three parties to measure the educational link. First, the invested companies and venture capital. Second, the invested companies and National Development Funds. Third, the venture capital and the National Development Funds. With 4 types of educational links and 3 combinations of firms, we can evaluate a total of 12 variables.. 5. Formula: monthly return – risk free rate (One month fixed deposit interest rate in First Bank) = Market premium + firm size premium + book value premium + abnormal return 6 Formula: monthly return – risk free rate (One month fixed deposit interest rate in First Bank) = Market premium + firm size premium + book value premium + abnormal return 22.

(23) Table 4: Independent Variables. Variables. Sources. Meanings. NVVCtype1. Social network. Type 1 linkages within invested companies. Database. and VCs.. Social network. Type 2 linkages within invested companies. Database. and VCs.. Social network. Type 3 linkages within invested companies. Database. and VCs.. Social network. Type 4 linkages within invested companies. Database. and NDF.. Social network. Type 2 linkages within invested companies. Database. and NDF.. Type 3 linkages within invested companies. Social network. io. n. al. and NDF.. Database VCNDFtype1. Ch. Social network. Social network. ‧. Database. Database NVNDFtype4. 學. NVNDFtype3. Type 1 linkages within invested companies. Nat. NVNDFtype2. 立. Social network. ‧ 國. NVNDFtype1. 治政 and VCs. 大. y. NVVCtype4. sit. NVVCtype3. er. NVVCtype2. i n U. v. Type 4 linkages within invested companies. e nandgNDF. chi. Type 1 linkages within VCs and NDF.. Database VCNDFtype2. Social network. Type 2 linkages within VCs and NDF.. Database VCNDFtype3. Social network. Type 3 linkages within VCs and NDF.. Database VCNDFtype4. Social network. Type 4 linkages within VCs and NDF.. Database 23.

(24) 3.4 Other Control variables Other control variables are separated into 3 categories, invested companies, venture capitals and investment records.. 3.4.1 Invested company Table 5: Control Variables – Invested companies. Variables. Meanings At least one member of invested companies graduated from NTU,. f_twtop NTHU, NCTU or NCKU in a specific year.. executives program, graduate school of business administration, National Chengchi University in a specific year. The percentage of invested companies’ members to obtain a degree in Commerce.. ‧. f_MBA. 學. f_NCCUBaep. ‧ 國. f_ustop. 政治大 League or University of California in a specific year. 立 At least one member of invested companies graduated from the. At least one member of invested companies graduated from Ivy. The percentage of invested companies’ members to obtain a degree in Science.. f_Taipei. The official registration location of invested company is Taipei.. f_Hsinchu. The official registration location of invested company is Hsinchu.. f_Age. Firm age in specific year.. f_1. The invested company belongs to manufacturing industry.. f_2. The invested company belongs to high technology or biotechnology industry.. f_3. The invested company belongs to financial or services industry.. n. al. er. io. sit. y. Nat. f_Science. Ch. engchi. 24. i n U. v.

(25) 3.4.2 Venture Capital Table 6: Control Variables - Venture Capitals. Variables. Meanings. vc_twtop. At least one member of VCs graduated from NTU, NTHU, NCTU or NCKU in a specific year. At least one member of VCs graduated from Ivy League 8 or University of California in a specific year. At least one member of VCs graduated from the executives program, graduate school of business administration, National Chengchi University in a specific year.. vc_ustop vc_NCCUBaep. vc_MBA. The percentage of VC members to obtain a degree in Commerce.. vc_Science. The percentage of VC members to obtain a degree in Science.. vc_Hsinchu. 學. ‧ 國. vc_Taipei. 治政 The official registration location of大 VC is Taipei. 立 The official registration location of VC is Hsinchu.. 3.4.3 Investment records. y. al. n. AR. sit. io. The record of as if a same investment case was invested by at least two VCs. The accumulated amount of investments of invested company in a specific year.. er. Syn. Meanings. Nat. Variables. ‧. Table 7: Control Variables – Investment records. Ch. engchi. i n U. v. 3.5 Descriptive Statistics The descriptive statistics are as follows. Our dependent variables include ROE, ROA, CAR (monthly), CAR (daily) and cash flow. Our dependent variables include 12 types of educational linkages with invested companies, venture capital and National Development Fund. Our control variables include the ratio of top school, ratio of NCCU BAEP, ratio of major in MBA, ratio of major in science, VC market share and syndication. 8. Ivy League includes Brown University, Columbia University, Cornell University, Dartmouth College, Harvard University, University of Pennsylvania, Princeton University and Yale University. 25.

(26) Table 8: Descriptive Statistics. Variable. Obs. Mean. Std. Dev.. Min. Max. (240.5500) (78.6300) (2.6845) (0.9956) (1,376,200). 113.6200 64.0600 4.0709 8.0174 330,000,000. ROE ROA Z AA CashFlow. 3,067 4.1065 31.3264 2,880 3.8889 13.8555 2,389 0.0641 0.4211 2,323 0.0167 0.2050 2,910 4,151,976 15,900,000. NVVCtype1 NVVCtype2. 8,170 8,170. 0.4316 0.4229. 0.4953 0.4940. 0 0. 1 1. NVVCtype3 NVVCtype4. 8,170 8,170. 0.2962 0.2834. 0.4566 0.4507. 0 0. 1 1. NVNDFtype1 NVNDFtype2 NVNDFtype3 NVNDFtype4. 8,170 8,170 8,170 8,170. 0.4886 0.4761 0.3357 0.3246. 0.4683. 0 0 0 0. 1 1 1 1. 0.5804 0.5799 0.5624 0.5605. 0.4935 0.4936 0.4961 0.4964. 0.4756 0.0585 0.1967 0.1796 0.2784. 0.4994 0.2346 0.3975 0.2279 0.3147. n. al. Ch. 8,175 8,175 8,175 8,175 8,175. IPO_mkt SZn. 8,175 129.8620 8,175 0.0281. 1 1 1 1. 0.2894 0.5000 0.2361 0.1987. 0 0 0 0 0. 1 1 1 1 1. 119.1707 0.1654. 0 0. 489 1. 0.5072 0.0922 0.5072 0.2619 0.2079. 26. y. 1 1 1 1 1. e n g0.5000 chi. vc_twtop vcNCCUBeap_1 vc_Ustop vcMBA vcScience. 0 0 0 0 0. sit. io. 8,175 8,175 8,175 8,175 8,175. 0 0 0 0. er. Nat. f_twtop fNCCUBeap_1 f_Ustop fMBA fScience. 8,170 8,170 8,170 8,170. ‧. VCNDFtype1 VCNDFtype2 VCNDFtype3 VCNDFtype4. 學. ‧ 國. 立. 0.4999 治政 0.4995 大 0.4723. i n U. v.

(27) Chapter 4 Hypothesis and regression model setting This paper mainly focuses on the effects of educational links between managers and promoters of new ventures, venture capital firms and National Development Fund and tests as if the linkages do enhance the long-term performance of the new venture. We define the educational linkages into 4 types and 12 kinds of linkages. Table 9: Types of Educational Link. Type. Meanings. Type 1. A linkage of educational background in same school.. Type 2. A linkage of educational background in same department, same school.. A linkage of educational background in same time, same department and same school.. 學. Type 4. ‧ 國. Type 3. 政治大 A linkage of educational background in same time, same school. 立. Linkages. ‧. Table 10: Tables of the Engaged Parties of linkages. Engaged Parties. Nat. NVVCType2. Invested companies & Venture Capital. NVVCType3. Invested companies & Venture Capital. NVVCType4. Invested companies & Venture Capital. NVNDFType1. Invested companies & National Development Fund. NVNDFType2. Invested companies & National Development Fund. NVNDFType3. Invested companies & National Development Fund. NVNDFType4. Invested companies & National Development Fund. VCNDFType1. Venture Capital & National Development Fund. VCNDFType2. Venture Capital & National Development Fund. VCNDFType3. Venture Capital & National Development Fund. VCNDFType4. Venture Capital & National Development Fund. Ch. engchi. 27. sit er. n. al. y. Invested companies & Venture Capital. io. NVVCType1. i n U. v.

(28) 4.1 Hypothesis 4.1.1 Invested companies & Venture capitals 1.. 𝐻0 ：A Type I link (same school) between managers of invested companies and VCs has no effect on invested companies’ long term performance. 𝐻𝑎 ：A Type I link (same school) between managers of invested companies and VCs has an effect on invested companies’ long term performance.. 2.. 𝐻0 ：A Type II link (same school, same department) between managers of invested companies and VCs has no effect on invested companies’ long term performance.. 政治大 companies and VCs has an effect on invested companies’ long term performance. 立. 𝐻𝑎 ：A Type II link (same school, same department) between managers of invested. 𝐻0 ：A Type III link (same school, same time) between managers of invested. 學. ‧ 國. 3.. companies and VCs has no effect on invested companies’ long term performance.. ‧. 𝐻𝑎 ：A Type III link (same school, same time) between managers of invested. io. y. sit. 𝐻0 ：A Type IV link (same school, same department, and same time) between. er. 4.. Nat. companies and VCs has an effect on invested companies’ long term performance.. managers of invested companies and VCs has no effect on invested companies’. n. al. long term performance.. Ch. engchi. i n U. v. 𝐻𝑎 ：A Type IV link (same school, same department, and same time) between managers of invested companies and VCs has an effect on invested companies’ long term performance.. 28.

(29) 4.1.2 Invested companies & National Development Fund 5.. 𝐻0 ：A Type I link (same school) between managers of invested companies and NDF has no effect on invested companies’ long term performance. 𝐻𝑎 ：A Type I link (same school) between managers of invested companies and NDF has an effect on invested companies’ long term performance.. 6.. 𝐻0 ：A Type II link (same school, same department) between managers of invested companies and NDF has no effect on invested companies’ long term performance. 𝐻𝑎 ：A Type II link (same school, same department) between managers of invested. 政治大 𝐻 ：A Type III link (same school, same time) between managers of invested 立 companies and NDF has an effect on invested companies’ long term performance.. 7.. 0. companies and NDF has no effect on invested companies’ long term performance.. ‧ 國. 學. 𝐻𝑎 ：A Type III link (same school, same time) between managers of invested. 𝐻0 ：A Type IV link (same school, same department, and same time) between. y. Nat. io. sit. managers of invested companies and NDF has no effect on invested companies’. er. 8.. ‧. companies and NDF has an effect on invested companies’ long term performance.. long term performance.. al. n. v i n 𝐻𝑎 ：A Type IV link (same department, and same time) between Cschool, h e nsame h gc i U. managers of invested companies and NDF has an effect on invested companies’ long term performance.. 29.

(30) 4.1.3 Venture capitals & National Development Fund 9.. 𝐻0 ：A Type I link (same school) between managers of VCs and NDF has no effect on invested companies’ long term performance. 𝐻𝑎 ：A Type I link (same school) between managers of VCs and NDF has an effect on invested companies’ long term performance.. 10. 𝐻0 ：A Type II link (same school, same department) between managers of VCs and NDF has no effect on invested companies’ long term performance. 𝐻𝑎 ：A Type II link (same school, same department) between managers of VCs. 政治大 𝐻 ：A Type III link (same school, same time) between managers of VCs and NDF 立 and NDF has an effect on invested companies’ long term performance.. 11.. 0. has no effect on invested companies’ long term performance.. ‧ 國. 學. 𝐻𝑎 ：A Type III link (same school, same time) between managers of VCs and NDF. ‧. has an effect on invested companies’ long term performance.. y. Nat. 12. 𝐻0 ：A Type IV link (same school, same department, and same time) between. er. io. performance.. sit. managers of VCs and NDF has no effect on invested companies’ long term. al. n. v i n 𝐻𝑎 ：A Type IV link (same department, and same time) between Cschool, h e nsame h gc i U. managers of VCs and NDF has an effect on invested companies’ long term performance.. 30.

(31) 4.2 Regression model setting 𝐽. 𝐾. 𝑅𝑂𝐸𝑖,𝑡 = 𝑇𝑦𝑝𝑒𝑙,𝑖,𝑡 + ∑ 𝑏𝑗 𝑁𝑉𝑗,𝑖,𝑡 + ∑ 𝑏𝑘 𝑉𝐶𝑘,𝑖,𝑡 + Macro𝑡 + Other𝑡 + 𝜀𝑖,𝑡 𝑗=1. 𝑘=1. Where 𝑇𝑦𝑝𝑒𝑙,𝑖,𝑡 stands for the 4 types of linkages among managers and directors of the 3 parties. ∑𝐽𝑗=1 𝑏𝑗 𝐼𝐶𝑗,𝑖,𝑡 stands for all variables of invested companies.. 政治大. ∑𝐾 𝑘=1 𝑏𝑘 𝑉𝐶𝑘,𝑖,𝑡 stands for all variables of venture capital firms.. 立. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 31. i n U. v.

(32) Chapter 5 Methodology and results 5.1 Methodology Since the data used are panel data, so we use Hausman test to verify whether to use Random-effect model or Fixed-effect model to run our regression. The Hypothesis of Hausman test can be listed as below, ̂ 𝑭𝑬 → 𝜷𝟎 , 𝜷 ̂ 𝑹𝑬 → 𝜷𝟎 ) 𝑯𝟎 = 𝒓𝒂𝒏𝒅𝒐𝒎 𝒆𝒇𝒇𝒆𝒄𝒕 (𝜷 ̂ 𝑭𝑬 → 𝛽𝟎 , 𝜷 ̂ 𝑹𝑬 ↛ 𝜷𝟎 ) 𝑯𝟏 = 𝒇𝒊𝒙𝒆𝒅 𝒆𝒇𝒇𝒆𝒄𝒕 (𝜷 And its test statistics is. 政治大. 𝑊 = [𝛽̂𝐹𝐸 − 𝛽̂𝑅𝐸 ]′ 𝜓 −1 [𝛽̂𝐹𝐸 − 𝛽̂𝑅𝐸 ]~𝜒 2 𝑘−1. 立. 學. ‧ 國. Where 𝜑 = 𝑉𝑎𝑟(𝛽̂𝐹𝐸 ) − 𝑉𝑎𝑟(𝛽̂𝑅𝐸 ) ̂ 𝐅𝐄 , 𝛃 ̂ 𝐑𝐄 two estimators follow consistency, and 𝛃 ̂ 𝐑𝐄 still follows efficency H0 = 𝛃. ‧. ̂ 𝐅𝐄 follows consistency H1 = only 𝛃. If the regression rejects 𝐻0 , then use the Fixed-effect model; if the regression does. y. Nat. io. sit. not reject 𝐻0 , then use Random-effect model to estimate.. n. al. er. So, by the result of Hausman test and rejection of the null hypothesis, we found our data fit the Fixed-effect model.. Ch. engchi. i n U. v. 5.2 Results To measure the effect of educational links, this paper used panel data with Fixedeffects obtained from the Hausman test in the previous hypothesis test. We put types of linkages as independent variables X, ROE for long term performance is the dependent variable Y in the regression and other references for control variables such as ratio of MBA, ratio of top school. The results of ROA and CAR are quite similar, so we only show the table of ROE in this paper.. 32.

(33) 5.2.1 Results – invested companies & VCs Table 10 shows the measured effects of educational links. Theoretically we expect that closer links among invested companies and venture capital firms may lead to a positive effect on the ROE. Here the effect of Type 1 and Type 2 linkages are negative in real data but not significant. It shows that the linkages of same school or same department are not close enough to bring a positive effect on the invested companies, or may even obtain a negative effect. But concerning Type 3 and Type 4 which means there is some time overlap in the educational link, the effect becomes positive and. 政治大 managers of invested companies and venture capital firms have positive effects on the 立. significant. The results follow our expectation that a closer relationship between. ROE.. ‧. ‧ 國. 學. 5.2.2 Results – invested companies & NDF. y. Nat. Unlike invested companies & VCs, the linkages between invested companied and. er. io. sit. NDF reveal a negative effect on ROE when there’s a light linkage. In Type 1 and Type 2 linkages, the effects are significantly negative at 1% and 5% significance. But also. al. n. v i n C hlinkages like TypeU 3 or Type 4 turn out to be unlike the previous result, closer engchi insignificant. These results may contribute to the investing decisions of NDF. National. Development Fund stands for the government and has the duty to support the growing industry in Taiwan. Thus, NDF may not choose the investment target purely based on ROE, but to follow the industrial policy or their own social network system.. 5.2.3 Results – VCs & NDF As the results in 5.2.2 suggest, the linkages within invested companies and NDF reveal a negative effect on ROE. Here, the linkages of venture capital and National Development Fund show a positive effect. We obtain a significant positive result in 33.

(34) Type 1 and Type 2 linkages, and an insignificant positive result in Type 3 and Type 4 linkages. These results match the essence of NDF’s existence; it needs to assist not only the development of national industry but also provide some extra earnings to payback the government. In order to ensure some return, NDF chooses professional venture capital firms to bring a reasonable return. Previous results of invested companies have explained the assisting of the industry, and the result of VC and NDF here then explain the positive effect on ROE.. 政治大 After reviewing the individual effects of the linkages, we then put three kinds of 立. 5.2.4 Results – different independent variables. Type 1 independent variables in the regression model simultaneously and further. ‧ 國. 學. evaluate the effect. As Table 14 shows, the effect is quite similar with putting them in. ‧. the regression individually. Then we put Type 2, Type 3 and Type 4 of the linkages. We. n. al. er. io. sit. y. Nat. found that as the connection becomes closer, the effects become more significant.. Ch. engchi. 34. i n U. v.

(35) Table 11 Result 1 - Invested companies & Venture capitals Reg1 -4.403 [2.874]. NVVCtype1 NVVCtype2. Reg2. 3.055* [1.676]. NVVCtpye4 f_twtop fNCCUBeap_1 f_ustop. 立. ‧ 國. IPO_mkt SZn Constant Observations Number of invest_id R-squared. sit. n. vcScience. al. er. io. vcMBA. y. Nat. vcNCCUBeap_1. 政治大. ‧. vc_ustop. 3.671** [1.642] -7.820*** -7.938*** -9.565*** -9.702*** [2.974] [2.959] [2.907] [2.905] -2.355 -2.334 -2.736 -2.818 [2.615] [2.614] [2.626] [2.624] 0.62 0.553 -0.007 0.028 [1.585] [1.579] [1.577] [1.573] 6.232 6.06 4.622 4.696 [3.939] [3.915] [3.807] [3.806] 0.384 0.348 -0.608 -0.625 [4.293] [4.291] [4.267] [4.265] -9.877*** -9.825*** -10.872*** -11.034*** [3.059] [3.062] [3.067] [3.067] -0.196 -0.188 -0.943 -1.064 [2.680] [2.681] [2.683] [2.683] 9.933* 9.303 8.395 8.522 [5.754] [5.657] [5.549] [5.544] 4.574 4.743 -0.165 -0.379 [5.780] [5.824] [5.636] [5.611] 0.010* 0.010* 0.009* 0.009* [0.005] [0.005] [0.005] [0.005] 9.936*** 9.965*** 9.948*** 9.920*** [2.919] [2.919] [2.918] [2.917] 10.718*** 10.756*** 11.422*** 11.454*** [1.202] [1.197] [1.144] [1.144] 3067 3067 3067 3067 183 183 183 183 0.04 0.04 0.04 0.04. 學. fScience. Reg4. -4.051 [2.686]. NVVCtype3. fMBA. Reg3. Ch. engchi. 35. i n U. v.

(36) Table 12 Result 2 - Invested companies & National Development Fund Type 1 -9.874*** [2.679]. NVNDFtype1 NVNDFtype2. Type 2. -0.57 [1.441]. NVNDFtype4 f_twtop fNCCUBeap_1 f_ustop. 立. ‧ 國. IPO_mkt SZn Constant Observations Number of invest_id R-squared. sit. n. vcScience. al. er. io. vcMBA. y. Nat. vcNCCUBeap_1. 政治大. ‧. vc_ustop. -0.54 [1.422] -4.105 -6.045* -8.740*** -8.764*** [3.163] [3.152] [2.924] [2.917] -1.77 -2.014 -2.226 -2.254 [2.612] [2.615] [2.617] [2.615] 0.605 0.495 0.384 0.387 [1.569] [1.571] [1.589] [1.592] 9.345** 7.368* 4.773 4.737 [4.005] [3.986] [3.816] [3.812] 3.707 1.751 -0.292 -0.278 [4.398] [4.364] [4.271] [4.273] -9.652*** -9.964*** -10.270*** -10.267*** [3.047] [3.050] [3.050] [3.050] 0.051 -0.145 -0.405 -0.437 [2.672] [2.676] [2.686] [2.680] 8.237 8.019 7.503 7.491 [5.521] [5.530] [5.530] [5.530] 1.919 1.727 2.006 2.035 [5.507] [5.516] [5.523] [5.526] 0.011** 0.011** 0.010* 0.010* [0.005] [0.005] [0.005] [0.005] 8.254*** 8.967*** 9.758*** 9.767*** [2.946] [2.945] [2.936] [2.935] 10.895*** 11.040*** 11.283*** 11.284*** [1.145] [1.147] [1.143] [1.143] 3067 3067 3067 3067 183 183 183 183 0.04 0.04 0.03 0.03. 學. fScience. Type 4. -5.710** [2.521]. NVNDFtype3. fMBA. Type 3. Ch. engchi. 36. i n U. v.

(37) Table 13 Result 3 – Venture capitals & National Development Fund Type 1 Type 2 Type 3 Type 4 VCNDFtype1 30.776*** [7.310] VCNDFtype2 24.205*** [6.664] VCNDFtype3 1.19 [3.252] VCNDFtype4 2.014 [3.258] f_twtop -9.139*** -9.215*** -8.940*** -8.942*** [2.879] [2.882] [2.887] [2.887] fNCCUBeap_1 -2.629 -2.474 -2.293 -2.316 [2.608] [2.610] [2.616] [2.616] f_ustop 0.199 0.175 0.267 0.246 [1.565] [1.567] [1.571] [1.571] fMBA 5.221 4.788 4.641 4.609 [3.800] [3.801] [3.811] [3.811] fScience -0.075 -0.384 -0.323 -0.291 [4.255] [4.258] [4.269] [4.269] vc_ustop -16.341*** -15.285*** -10.699*** -11.088*** [3.367] [3.343] [3.280] [3.332] vcNCCUBeap_1 0.161 -0.05 -0.511 -0.52 [2.670] [2.671] [2.674] [2.674] vcMBA -25.599*** -17.647** 6.643 6.255 [9.600] [8.850] [5.992] [5.877] vcScience -24.526*** -18.688** 1.108 0.479 [8.354] [7.910] [5.965] [6.001] IPO_mkt 0.012** 0.011** 0.010* 0.010* [0.005] [0.005] [0.005] [0.005] SZn 9.290*** 9.396*** 9.929*** 9.961*** [2.914] [2.916] [2.922] [2.922] Constant 10.692*** 10.944*** 11.279*** 11.265*** [1.148] [1.144] [1.143] [1.143] Observations 3067 3067 3067 3067 Number of invest_id 183 183 183 183 R-squared 0.04 0.04 0.03 0.03. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 37. i n U. v.

(38) NVVCtype1 NVNDF type1 VCNDFtype1 f_twtop fNCCUBeap_1 f_ustop fMBA. 立. SZn Constant Observations Number of invest_id R-squared. sit. er. al. n. IPO_mkt. io. vcScience. y. Nat. vcMBA. Reg4 -3.658 [2.923] -9.782*** [2.720] 32.895*** [7.328] -3.463 [3.189] -2.233 [2.607] 0.786 [1.577] 11.161*** [4.076] 4.595 [4.395] -15.844*** [3.361] 1.015 [2.675] -25.101*** [9.618] -24.172*** [8.423] 0.013*** [0.005] 7.682*** [2.943] 9.775*** [1.210] 3067 183 0.05. ‧. vcNCCUBeap_1. 政治大. 學. ‧ 國. fScience vc_ustop. Table 14 Result 4 – Type 1 Linkage Reg1 Reg2 Reg3 -4.403 [2.874] -9.874*** [2.679] 30.776*** [7.310] -7.820*** -4.105 -9.139*** [2.974] [3.163] [2.879] -2.355 -1.77 -2.629 [2.615] [2.612] [2.608] 0.62 0.605 0.199 [1.585] [1.569] [1.565] 6.232 9.345** 5.221 [3.939] [4.005] [3.800] 0.384 3.707 -0.075 [4.293] [4.398] [4.255] -9.877*** -9.652*** -16.341*** [3.059] [3.047] [3.367] -0.196 0.051 0.161 [2.680] [2.672] [2.670] 9.933* 8.237 -25.599*** [5.754] [5.521] [9.600] 4.574 1.919 -24.526*** [5.780] [5.507] [8.354] 0.010* 0.011** 0.012** [0.005] [0.005] [0.005] 9.936*** 8.254*** 9.290*** [2.919] [2.946] [2.914] 10.718*** 10.895*** 10.692*** [1.202] [1.145] [1.148] 3067 3067 3067 183 183 183 0.04 0.04 0.04. Ch. engchi. 38. i n U. v.

(39) Table 15 Result 5 – Type 2 Linkage Reg1 -4.051 [2.686]. NVVCtype2. Reg2. NVNDFtype2. Reg3. -5.710** [2.521]. VCNDFtype2. 立. vc_ustop. io. IPO_mkt SZn Constant Observations Number of invest_id R-squared. al. n. vcScience. Ch. engchi. 39. ‧. vcMBA. Nat. vcNCCUBeap_1. 學. fScience. ‧ 國. fMBA. 政治大. y. f_ustop. -6.045* [3.152] -2.014 [2.615] 0.495 [1.571] 7.368* [3.986] 1.751 [4.364] -9.964*** [3.050] -0.145 [2.676] 8.019 [5.530] 1.727 [5.516] 0.011** [0.005] 8.967*** [2.945] 11.040*** [1.147] 3067 183 0.04. sit. fNCCUBeap_1. -7.938*** [2.959] -2.334 [2.614] 0.553 [1.579] 6.06 [3.915] 0.348 [4.291] -9.825*** [3.062] -0.188 [2.681] 9.303 [5.657] 4.743 [5.824] 0.010* [0.005] 9.965*** [2.919] 10.756*** [1.197] 3067 183 0.04. er. f_twtop. 24.205*** [6.664] -9.215*** [2.882] -2.474 [2.610] 0.175 [1.567] 4.788 [3.801] -0.384 [4.258] -15.285*** [3.343] -0.05 [2.671] -17.647** [8.850] -18.688** [7.910] 0.011** [0.005] 9.396*** [2.916] 10.944*** [1.144] 3067 183 0.04. i n U. v. Reg4 -3.403 [2.719] -6.067** [2.561] 26.046*** [6.688] -5.349* [3.176] -2.277 [2.610] 0.612 [1.576] 8.802** [4.043] 2.454 [4.366] -14.992*** [3.347] 0.634 [2.680] -17.468** [8.904] -18.120** [8.114] 0.013** [0.005] 8.458*** [2.945] 10.194*** [1.200] 3067 183 0.04.

(40) Table 16 Result 6 – Type 3 Linkage Reg1 3.055* [1.676]. NVVCtype3. Reg2. NVNDFtype3. Reg3. -0.57 [1.441]. VCNDFtype3. fMBA. vcNCCUBeap_1. Nat. vcMBA. io. IPO_mkt SZn Constant Observations Number of invest_id R-squared. al. n. vcScience. Ch. engchi. 40. ‧. ‧ 國. vc_ustop. 學. fScience. 立. 政治大. y. f_ustop. -8.740*** [2.924] -2.226 [2.617] 0.384 [1.589] 4.773 [3.816] -0.292 [4.271] -10.270*** [3.050] -0.405 [2.686] 7.503 [5.530] 2.006 [5.523] 0.010* [0.005] 9.758*** [2.936] 11.283*** [1.143] 3067 183 0.03. sit. fNCCUBeap_1. -9.565*** [2.907] -2.736 [2.626] -0.007 [1.577] 4.622 [3.807] -0.608 [4.267] -10.872*** [3.067] -0.943 [2.683] 8.395 [5.549] -0.165 [5.636] 0.009* [0.005] 9.948*** [2.918] 11.422*** [1.144] 3067 183 0.04. er. f_twtop. 1.19 [3.252] -8.940*** [2.887] -2.293 [2.616] 0.267 [1.571] 4.641 [3.811] -0.323 [4.269] -10.699*** [3.280] -0.511 [2.674] 6.643 [5.992] 1.108 [5.965] 0.010* [0.005] 9.929*** [2.922] 11.279*** [1.143] 3067 183 0.03. i n U. v. Reg4 3.525** [1.755] -1.512 [1.515] 1.052 [3.280] -9.190*** [2.932] -2.721 [2.627] 0.189 [1.592] 4.808 [3.815] -0.428 [4.272] -11.390*** [3.298] -0.75 [2.690] 7.832 [6.029] -1.032 [6.054] 0.010* [0.005] 9.674*** [2.937] 11.395*** [1.145] 3067 183 0.04.

(41) Table 17 Result 7 – Type 4 Linkage Reg1 3.671** [1.642]. NVVCtype4 NVNDFtype4 VCNDFtype4 f_twtop fNCCUBeap_1 f_ustop fMBA. vcNCCUBeap_1. SZn Constant Observations Number of invest_id R-squared. sit. n. IPO_mkt. al. er. io. vcScience. y. Nat. vcMBA. 政治大. ‧. ‧ 國. vc_ustop. Reg3. 學. fScience. Reg4 4.078** [1.705] -0.54 -1.567 [1.422] [1.481] 2.014 1.803 [3.258] [3.283] -9.702*** -8.764*** -8.942*** -9.340*** [2.905] [2.917] [2.887] [2.926] -2.818 -2.254 -2.316 -2.885 [2.624] [2.615] [2.616] [2.626] 0.028 0.387 0.246 0.255 [1.573] [1.592] [1.571] [1.592] 4.696 4.737 4.609 4.782 [3.806] [3.812] [3.811] [3.811] -0.625 -0.278 -0.291 -0.349 [4.265] [4.273] [4.269] [4.273] -11.034*** -10.267*** -11.088*** -11.889*** [3.067] [3.050] [3.332] [3.349] -1.064 -0.437 -0.52 -0.942 [2.683] [2.680] [2.674] [2.686] 8.522 7.491 6.255 7.546 [5.544] [5.530] [5.877] [5.907] -0.379 2.035 0.479 -1.65 [5.611] [5.526] [6.001] [6.062] 0.009* 0.010* 0.010* 0.010* [0.005] [0.005] [0.005] [0.005] 9.920*** 9.767*** 9.961*** 9.666*** [2.917] [2.935] [2.922] [2.935] 11.454*** 11.284*** 11.265*** 11.413*** [1.144] [1.143] [1.143] [1.145] 3067 3067 3067 3067 183 183 183 183 0.04 0.03 0.03 0.04. 立. Reg2. Ch. engchi. 41. i n U. v.

(42) Table 18 Result 8 – Type 1 ~ Type 4 Linkage NVVC NVNDF VCNDF f_twtop fNCCUBeap_1 f_ustop fMBA. io. IPO_mkt SZn Constant Observations Number of invest_id R-squared. al. n. vcScience. sit. Nat. vcMBA. er. vcNCCUBeap_1. y. ‧. ‧ 國. vc_ustop. 立. 政治大. 學. fScience. type1 type2 type3 type4 -3.658 -3.403 3.525** 4.078** [2.923] [2.719] [1.755] [1.705] -9.782*** -6.067** -1.512 -1.567 [2.720] [2.561] [1.515] [1.481] 32.895*** 26.046*** 1.052 1.803 [7.328] [6.688] [3.280] [3.283] -3.463 -5.349* -9.190*** -9.340*** [3.189] [3.176] [2.932] [2.926] -2.233 -2.277 -2.721 -2.885 [2.607] [2.610] [2.627] [2.626] 0.786 0.612 0.189 0.255 [1.577] [1.576] [1.592] [1.592] 11.161*** 8.802** 4.808 4.782 [4.076] [4.043] [3.815] [3.811] 4.595 2.454 -0.428 -0.349 [4.395] [4.366] [4.272] [4.273] -15.844*** -14.992*** -11.390*** -11.889*** [3.361] [3.347] [3.298] [3.349] 1.015 0.634 -0.75 -0.942 [2.675] [2.680] [2.690] [2.686] -25.101*** -17.468** 7.832 7.546 [9.618] [8.904] [6.029] [5.907] -24.172*** -18.120** -1.032 -1.65 [8.423] [8.114] [6.054] [6.062] 0.013*** 0.013** 0.010* 0.010* [0.005] [0.005] [0.005] [0.005] 7.682*** 8.458*** 9.674*** 9.666*** [2.943] [2.945] [2.937] [2.935] 9.775*** 10.194*** 11.395*** 11.413*** [1.210] [1.200] [1.145] [1.145] 3067 3067 3067 3067 183 183 183 183 0.05 0.04 0.04 0.04. Ch. engchi. 42. i n U. v.

(43) Chapter 5 Conclusion In this paper, we propose a social network database based on Ho (2014) and Lin (2014). It now contains the data of invested companies, venture capital firms and the National Development Fund. With these three parties, we measure the individual social network effect on invested companies’ long-term performance. This paper is based on the educational links method of Cohen (2007) and defines four types of linkages. We found that a closer linkage between invested companies and venture capital affects the long-term performance positively because venture capital firms may find a target. 政治大. worthy of investment with the social network system. However, when there is a political. 立. link between invested companies and National Development Fund, there exists a. ‧ 國. 學. negative effect. This appears to be an interesting phenomenon because the result implies that the National Development Funds chooses investment targets through its own social. ‧. network system but its approach is not profitability-oriented, which leads us to think. y. Nat. sit. that NDF is not playing the role of a government sector institution. Another result is the. n. al. er. io. positive effect of educational links between VC and NDF on invested companies’ long-. i n U. v. term performance. The reason behind this effect may be that the NDF choosing good. Ch. engchi. VCs to invest in order to match the standard of refunding to government.. As the results above show, whether educational linkages have a positive or negative effect on invested companies’ long-term performance, they do provide a platform for all parties to exchange their information. Therefore, the social network links successfully solve part of the information asymmetry problem, and do impose an effect on invested companies’ long-term performance.. 43.

(44) References Akerlof, G. A. (1970). The market for" lemons": Quality uncertainty and the market mechanism. The quarterly journal of economics, 488-500. Bai, C. E., Hsieh, C. T., & Song, Z. M. Crony Capitalism with Chinese Characteristics. Barber, B. M., & Lyon, J. D. (1996). Detecting abnormal operating performance: The empirical power and specification of test statistics. Journal of financial Economics, 41(3), 359-399. Barry, C. B., Muscarella, C. J., Peavy, J. W., & Vetsuypens, M. R. (1990). The role of venture capital in the creation of public companies: Evidence from the goingpublic process. Journal of Financial economics, 27(2), 447-471. Boot, Arnoud W A and Thakor, Anjan V (1993). Security Design. Journal of Finance, 48(4): 1349-78. Brav, A., & Gompers, P. A. (1997). Myth or reality? The long‐run underperformance of initial public offerings: evidence from venture and nonventure capital ‐backed. 立. 政治大. ‧ 國. 學. companies. The Journal of Finance, 52(5), 1791-1821.. ‧. Bunkanwanicha, P., & Wiwattanakantang, Y. (2009). Big business owners in politics. Review of Financial Studies, 22(6), 2133-2168. Carleton, W.T., and I.A. Cooper. (1976). Estimation and Uses of the Term Structure of Interest Rates. Journal of Finance, 31: 1067-1083.. Nat. y. sit. n. al. er. io. Claessens, S., Feijen, E., & Laeven, L. (2008). Political connections and preferential access to finance: The role of campaign contributions. Journal of Financial Economics, 88(3), 554-580. Cohen L., Frazzini A., Malloy C. (2008). The small world of investing: Board connections and mutual fund returns. Journal of Political Economy 116(5)951-979. Cooper, M. J., Gulen, H., & Ovtchinnikov, A. V. (2010). Corporate political contributions and stock returns. The Journal of Finance, 65(2), 687-724. Eckbo, B. E. (Ed.). (2008). Handbook of Empirical Corporate Finance SET (Vol. 1). Elsevier. Faccio, M. (2006). Politically-connected firms. American Economic Review, Fama, E. F., & French, K. R. (1992). The cross‐section of expected stock returns. the. Ch. engchi. i n U. v. Journal of Finance, 47(2), 427-465. Fisman, R. (2001). Estimating the value of political connections. American Economic Review, 1095-1102. Goldman, E., Rocholl, J., & So, J. (2009). Do politically connected boards affect firm value? Review of Financial Studies, 22(6), 2331-2360. Gompers, P. A. (1995). Optimal investment, monitoring, and the staging of venture capital. The journal of finance, 50(5), 1461-1489. 44.

(45) Gompers, P. A. (1996). Grandstanding in the venture capital industry. Journal of Financial economics, 42(1), 133-156. Gompers, P. A., & Lerner, J. (1999). What drives venture capital fundraising? (No. w6906). National bureau of economic research. Grossman, S. J., & Hart, O. D. (1986). The costs and benefits of ownership: A theory of vertical and lateral integration. The Journal of Political Economy, 691-719. Ho (2014), Social Network and R&D Productivity of the Venture Invested Entrepreneurial Team：the Case of Educational Network Jensen M.C. and Meckling W.H. (1976). Theory of the Firm: Managerial Behavior, Agency Costs and Ownership Structure. Journal of Financial Economics 3(4)： 305-360. Kaplan, S. N., Sensoy, B. A., & Strömberg, P. (2009). Should investors bet on the jockey or the horse? Evidence from the evolution of firms from early business plans to public companies. The Journal of Finance, 64(1), 75-115. Krishnan, C. N. V., Ivanov, V. I., Masulis, R. W., & Singh, A. K. (2011). Venture capital reputation, post-IPO performance, and corporate governance. Journal of Financial and Quantitative Analysis, 46(05), 1295-1333. Lee, P. M., & Wahal, S. (2004). Grandstanding, certification and the underpricing of venture capital backed IPOs. Journal of Financial Economics, 73(2), 375-407. Leuz, C., & Oberholzer-Gee, F. (2006). Political relationships, global financing, and. 立. 政治大. ‧. ‧ 國. 學. Evidence from. Indonesia. Journal of Financial. y. Nat. corporate transparency:. sit. n. al. er. io. Economics, 81(2), 411-439. Lin, T. H., & Smith, R. L. (1998). Insider reputation and selling decisions: the unwinding of venture capital investments during equity IPOs. Journal of Corporate Finance, 4(3), 241-263. Lin (2014) Social Network and the Venture Capital Investment Target：The Case of. Ch. engchi. i n U. v. Educational Network Mahmood, I. P., Zhu, H. and Zajac, E. J., 2011, “Where can capabilities come from? Network ties and capability acquisition in business groups,” Strategic Management Journal, 32, 820–48.Megginson and Weiss. McDonald, M. B. (2014). Politically Connected Analysts. McDonald, R. L. (2013). Contingent capital with a dual price trigger. Journal of Financial Stability, 9(2), 230-241. Megginson, W. L., & Weiss, K. A. (1991). Venture capitalist certification in initial public offerings. The Journal of Finance, 46(3), 879-903. Nahata, R. (2008). Venture capital reputation and investment performance. Journal of Financial Economics, 90(2), 127-151. Petersen, M. A. (2009). Estimating standard errors in finance panel data sets: 45.

(46) Comparing approaches. Review of financial studies, 22(1), 435-480. Ritter, J. R. (1991). The long‐run performance of initial public offerings. The journal of finance, 46(1), 3-27. Saxenian A. (1994). Regional Advantage：Culture and Competition in Silicon Valley and Route 128. Harvard University Press, Cambridge, MA. Saxenian, A. (2002). Brain Circulation and Chinese Chipmakers: The Silicon ValleyHsinchu-Shanghai Triangle. Berkeley: University of California at Berkeley. Singh, A. K., Dark, F. H., & Carter, R. B. (1998). Underwriter Reputation, Initial Returns, and the Long-Run Performance of IPO Stocks. The Journal of Finance, 53(1). Y. Chan, D. Siegel, A.V. Thakor (1990). Learning, corporate control and performance requirements in venture capital contracts. International Economic Review, 31, pp. 365–381. Zarutskie, R. (2010). The role of top management team human capital in venture capital markets: Evidence from first-time funds. Journal of Business Venturing, 25(1), 155-172.. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 46. i n U. v.

(47) Appendix. 立. 政治大. ‧. ‧ 國. 學. n. er. io. sit. y. Nat. al. Ch. engchi. 47. i n U. v.

(48)