The Effect Of Venture Capitalists In The Long Run: The
Performance Of Initial Public Offerings For Technology
Stocks In The United States
Amsterdam Business School
Name Tsz-Tian Lu
Student number 10832904
Program Economics & Business
Specialization Finance & Organization
Number of ECTS 12
Supervisor Ilko Naaborg
Target completion 31 / 01 / 2018
Abstract
This paper examines the impact venture capitalists (VCs) have on the long run
performance of technology IPOs in the US, by comparing the three-year buy-and-hold returns of the IPOs stocks. The analyses are based on a sample of 103 VCs backed IPOs and 40 non-VCs backed IPOs in the US from 2004 to 2007. The difference between the long run performance of non-VCs backed and VCs backed IPOs is statistically
insignificant, which suggests that the effects of VCs on the IPOs firms stock performance are insignificant in the long-run. In effect, due to the time-period of the dataset, the negative effect of market timing might have an impact on the role of VCs.
Statement of Originality
This document is written by Student Tsz-Tian Lu who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Table of Contents
1. Introduction 3
2. Literature review 5
3. Methodology and Data
3.1. Methodology 9
3.2. Data and descriptive statistics 12
4. Empirical result
4.1. Empirical Results 18
4.2 Robustness check 21
5. Conclusion and discussion 24
1. Introduction
Venture capitalists (henceforth VCs) are one of the external funding options for
entrepreneurs to finance and grow their business. In return for the risky capital the VCs put into the startups, the expected rate of returns on the investments is also high.
Historically, 37% of the initial public offerings (IPOs) of firms headquartered in the US are backed by VCs. In particular, VCs play a significant role in the life of many
technology startups in the US, as from 1980 to 2016, 58% of the technology IPOs are backed by VCs (Ritter, 2017). The questions regarding whether VCs-backed IPOs demonstrate less IPO underpricing as well as superior long-run performances have attracted a great amount of attention from both practitioners and academics. Many researchers have already identified the prevalent phenomenon of IPOs long-run underperformance regardless of the time-windows chosen, the regions or the methodologies employed (Coakley, Hadass, and Wood, 2007). When the factors concerning market timing and investor sentiments are included, the patterns are further identified. Many researchers find that IPOs tend to cluster in the hot market period, which is a phenomenon that firm tend to go public during the period when the investor
sentiment is relatively high (Ritter, 1984). It is also proved that the subsequent long-run performances of these hot market IPOs are inferior in comparison to the IPOs launched during other periods (Ljungqvist et al., 2006). Yung, Çolak, and Wang (2008)
demonstrate that the market cycle present in the IPO market and the mediocre long-run performances for the hot market IPOs stocks are an IPO phenomenon, not a market-wide one.
However, when the IPOs are categorized by whether VCs are involved in the creation of the public firms, previous literature exhibits inconsistent findings concerning whether the participation of the VCs in the process of IPOs and the business life of a firm is advantageous in the long-run, as compared to the non-VCs backed ones. That is, it is unclear whether the VCs backed IPOs issuers outperform the ones without VCs support, and the results differ especially between different regions. Jain and Kini (1995) use the US data from 1976 to 1988 to demonstrate that in terms of operating performance, the presence of VCs is value-added to the IPOs firm’s long-run performance. On the other
hand, Coakley, Hadass, and Wood (2007) use the data from the UK stocks market from 1985 to 2003 to show that the difference in the long-term performance between VCs backed and non-VCs backed IPOs firms are not significant. Besides, Coakley, Hadass, and Wood (2007) also find that for the IPOs that go public during the bubble years, VCs backed IPOs performed significantly worse in the long-run compared to non-VCs backed ones.
Therefore, this research specifically focuses on the long-term performance of IPOs stocks to examine the role of VCs. Moreover, instead of inspecting the effects of VCs in general, this research focuses on the technology industry. The technology IPOs have shown several unique patterns in previous research. The research by Coakley, Hadass, and Wood (2007) further shows that within the group of IPOs which take place in the hot market period, the long-run operating performances of the technology IPOs stocks are notably worse compared to other industries. That is, technology stocks are more sensitive to the market cycle, and low-quality firms are more likely to go public at inappropriate timing which results in more severe long-run underperformance compared to the IPOs firms in the other industries that go public under the same period. It is thus worth
investigating the role of VCs in the technology industry, to detect whether the presence of VCs facilitates the quality of decisions and the long-term operating performance of the IPOs firms. This research employs the IPOs data from the year of 2004 to 2007 in the US technology industry and attempts to analyze the dynamic between the entrepreneurial business and the VCs that invest in them. A set of cross-sectional data and the slightly- adjusted OLS regression model initially proposed by Brav and Gompers (1997) are employed in this event study. The research question this research address is, how do venture capitalists effect IPOs long-term performance in the technology industry in the US in the years from 2004 to 2007, and did VC-backed issuers outperform non-VC backed ones?
The main contribution of this research is that it explores the long-term effects VCs have on the IPOs firms in a VCs-intensive, highly competitive and rapidly growing industry. The results are generated using a model similar to the one in Brav and Gompers (1997), but an entirely different sample set which provides more insights into the roles of VCs. Alternative perspectives on the results are further provided to thoroughly inspect
the possible explanation of the findings, which are different from the findings of Brav and Gompers (1997).
This paper is organized as follows: In the next section, the functions of VCs are explained, and two broad categories of theories concerning the long-run effects of VCs on the IPOs firms are discussed. The hypothesis and its implication are addressed at the end of Section 2 as well. Section 3 presents the methodology, the regression model and the descriptive of the data employed in this research. The regressions outcomes, the interpretations of the results as well as some limitations of the research are elaborated in Section 4. Section 4 also includes robustness checks in which two additional regressions are run, one with fixed effects and the other with the data with different time-windows. The last section briefly summarizes this research, and also discusses possible implications of the findings.
2. Literature review
There are various models and theories attempting to describe the role of VCs in the life of their startups as well as in the IPOs market. Essentially, VCs identify startups which are in their early stages of business with the potential for growth, then invest in the startups financed mostly through closed-end fund and limited partnership (Sahlman, 1990). Thus, VCs serve as one of the methods for entrepreneurs to finance their business and possibly facilitate the growth rate of the startups through both the VCs grant and the access to the expertise VCs possess. VCs also tend to be specialized in investing in a specific industry and/or in specific stages of a startup (Gorman and Sahlman, 1989; Barry et al., 1990). Since the startups VCs invest in provide high investment returns but also high risk, return on investments is decisive for the VCs success (Gompers, 1996). It follows that VCs should function in a way that enables them to maximize the return on investment and minimize their exposure to risks. From the investment procedures of VCs, the
compensation scheme that VCs design to reward the individual venture owners, to the contracts signed between the two parties, the way how VCs function disclose a lot about what roles and impact VCs have on the startup (Sahlman, 1990). Based on these, the two main theories presented below thus use the dynamics between the VCs and the
entrepreneurs to explain the influence of VCs.
2.1 Certification/monitoring model on the role of VCs
Theories regarding the effect of VCs on the performance of the startups are categorized into two broad categories. The first category being theories concerning the monitoring/certification roles VCs play both before and after the creation of the public firms, while the other concentrates on theories about the influence of asymmetric
information and adverse selection on the type of startups that VCs attract. Within the first category of theories, the monitoring model is the most well-adapted one in the existing literature. The monitoring model indicates that venture capitalists are active investors and large block-holders who seek to add value to the startups they invest in by, for example, the network the VCs has in the specific industry or the expertise they possess that could facilitate the operation and performance of the startup. VCs profit from the growth firms they hold in their portfolio through exit options such as initial public offerings (IPOs) or merger and acquisition, thus, theoretically speaking, the VCs will put forth the majority of its resources in monitoring and mentor the startups under its funding, and the VCs should design the incentive systems for the entrepreneurs that are incentive-compatible with the VCs’ ultimate aim -- maximizing the value of the startups when VC cash out on the investment. The term monitoring is sometimes used interchangeably with the
certification model, in the sense that both models suggest that the presence of VCs certify the value of the startups in the IPOs market where information asymmetry is particularly severe (Megginson and Weiss, 1991). The certification/monitoring model, therefore, indicates that the presence of VCs in a startup's business lifecycle is value-adding to the startup.
In regard to the evidences of the certification/monitoring role of VCs in the startups, research results show that VCs devote most of their time on supporting the startups, and the quality of VCs services is dependent on the reputation of the VCs. The assumption that VCs spend most of their time monitoring the startups is verified in a study by Gorman and Sahlman (1990) based on the survey sent out to the VCs in the United States. The aim of their research is to provide insights into how VCs allocate their time
and what types of activities they focus on the most. It turns out that VCs spend the
majority of their time monitoring and supporting the operations of the startups they invest in, which is in line with the monitoring model. Another common way that VCs monitor their investments is that VCs are also involved extensively in the management team (i.e. the board of directors) of the startups, supervising the decisions entrepreneurs make and take action when necessary (Barry et al., 1990; Gorman and Sahlman, 1990). Because of the services and incentives the VCs provide, it is proved that the participation of VCs could also signal that the startups are of better quality comparing to the non-VCs backed ones (Wang, Wang, and Lu, 2003). As to how credible the signal is is dependent on the reputation, or the track record of the VCs, according to Megginson and Weiss (1991) and Gompers (1996).
2.2 Adverse selection models on the role of VCs
The other category of theories, which hold a contrary point of view in comparison to the certification/monitoring model, are models which base their reasoning on the negative effects of VCs due to adverse selection, also known as the lemon problem. The presence of asymmetric information between the investor and investee (i.e. the VCs and the startups) is often two sides of the same coin. In the certification/monitoring model, it is inferred that the participation of VCs can alleviate the impacts of asymmetric information on external investors' concern over the quality and future prospect of the startups. On the other hand, since adverse selection is essentially an ex-ante asymmetric information problem, in this context it is argued that only the less promising ventures will seek and accept VCs funding. The funding that VCs provide comes with rules that would sometimes be too risky for the entrepreneurs to agree on. One example is that VCs
conventionally incorporate exceptionally high discount rates when evaluating the value of the individual entrepreneur in question, and VCs do so in order to ensure that only the entrepreneurs who are highly confident in the true value and the success of their ventures will agree on the contract. Nevertheless, it is often difficult to judge if the discount rates VCs implement are too high for the best startups to accept the funding (Sahlman, 1990). Consequently, according to Sahlman (1990), only the startups that are less profitable and
have no other finance option will agree on the contracts VCs propose. If within the non-VCs backed firms the ones with better quality reject non-VCs funding and the ones with an inferior prospect are rejected by VCs; the startups backed by VCs are therefore not the most profitable ones comparing to non-VCs backed startups (Wang, Wang, and Lu, 2003).
The other model which also starts from adverse selection is the grandstanding model first documented by Gompers (1996). The word "grandstanding", in this context, refers to the acts that VCs do to impress the investors and the capital market. Since roughly 80% of the VCs raise their capital through a limited partnership with predetermined lifetime, the track record and reputation are crucial determinants of the VCs future success (Gompers and Lerner, 1996). Gompers argues that VCs, especially young VCs without any history of successfully bringing a firm in their portfolio public, have an immense incentive to grandstand. The result of grandstanding is that the startups younger VCs bring public are more likely to be premature. As a consequence of early IPOs, these startups backed by young VCs tend to go public at relatively inferior market timing and more likely to suffer from long-run underperformance in terms of returns on assets (ROA) and net profit change (Wang, Wang, and Lu, 2003). The grandstanding model thus suggests that not all startups which are backed by VCs will benefit from the VCs service. If the IPOs firm is supported by young VCs, the grandstanding model predicts that the long-run performance of VCs-backed IPOs may be inferior to non-VCs backed ones.
2.3 Summary of the literature and hypothesis
The two broad categories of theories on the role of VCs both start from the presence of asymmetric information, but arrive at contradicting predictions of the role of VCs. Despite the conflicting findings in the existing literature, logically it is expected that even with the presence of potential negative effects, the services VCs provide to the
entrepreneurs to build and manage the firm should outweigh the negative aspects in the long run. This lead to the hypothesis of this research: The presence of VCs has a positive effect on the IPOs stock’s three-year long-term performance. Rejecting the null
hypothesis means that the positive impacts VCs bring to the IPOs firms facilitate the firm's performance in the long-run, and this is also aligned with the
monitoring/certification model. On the other hand, if the null hypothesis is not rejected, the possible explanation including the adverse selection model, the grandstanding model and/or the market timing of VCs IPOs during the hot market period.
3. Methodology and Data
3.1 Methodology
To evaluate whether the participation of venture capital has a positive impact on the IPOs stock’s three-year long-term performance, an econometric model will be employed here. The model is a modification of the one utilized in the article by Brav and Gompers (1997), in which they intend to determine the factors that potentially have an impact on a stock’s five-year wealth relatives1 for all the IPOs between 1972 and 1992. Since this cross-sectional model contains the ingredients allowing for testing the significance of the presence of venture capital, it is used in this research with some slight adjustments which I will explain in detail in the section below. The dependent variable is the three-year buy-and-hold abnormal returns of the US technology stocks of the firms that went public in the US during the year of 2004 to 2007. The dependent variables including a dummy variable VC indicating whether an IPO issuer is backed by venture capital(s), and two control variables which are the logarithm of size and the logarithm of the book-to-market ratio of the issuing firm. The regressions model is formulated as
!ℎ#$$%&$'#(%)*+,( = ./+ .123 + .4ln 789$ + .:ln%()<<=%><%?'#=$>%#'>8<) + A The buy-and-hold abnormal returns (BHARs) is defined as the three-year buy-and-hold returns less the market return, namely, the expected return B(,CD), throughout the same period. The formula for BHARs is
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
1!Five-year wealth relatives is defined as the buy and hold return for a stock over the buy and hold return on
% )*+,CE = % [1 + ,CD E DH1 ] − [1 + B ,CD ] E DH1
in which 8 represents a specific investment in the month >, cumulating across K periods (Barbera and Lyon, 1997). In this research K is being set at 36 (i.e. three years), and the S&P 500 index as well as the NYSE/AMEX equal weighted index (including dividends) are used as proxies to measure the benchmark returns. As for the control variables; size is measured by multiplying the IPO firms number of shares outstanding by the closing price on the last day of the issuing month, and the book-to-market ratio is defined as the first book value available within one year of the offer date over the market value within the same period of time. The two control variables, size and book to market ratio, are proven to have significant impact on the performance of a security respectively (French and Fama, 1993). Thus, the coefficients .4 and .:are expected to be statistically significant, and consequently the inclusions are necessary in order to separate the effect of venture capitalists from the effect of size and the book to market ratio of the IPOs firms. The coefficients on the dummy variable VC, .1, is the variable of interest in this research. If the coefficient is positive and significantly different from 0, the null hypothesis is rejected and it indicates that the participation of venture capitalists has a positive impact on the IPOs stocks’ long-term performance.
This regression model is different from the original model in two respects. First of all, regarding the dependent variable of the model, Brav and Gompers (1997) use natural logarithm of wealth relative2 instead of the buy-and-hold abnormal returns. Secondly, the dataset employed here is completely different in terms of the time window chosen, the industry focused on, as well as the source of the data. Regarding the first aspect, Brav and Gompers (1997) use wealth relative and incorporate lagged dividend-price ratio in the regressions as one of the independent variables to control for the general impact of the market price level. In this research BHARs is used instead, since it is a more commonly !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
2!Wealth relative is defined as the buy-and-hold return on an investment divided by the buy-and-hold return
used method in similar event studies. Also, the element assumably containing
information for general price level is removed since the use of BHARs should already capture the market price level. For evaluating long-term stock returns in an event study, another prevalent way of testing abnormal returns is applying cumulative average
abnormal returns (CAARs)3. However, BHARs is favored here, since the aim is to obtain the abnormal returns on an IPOs stock over three years, not the mean monthly abnormal return of an IPOs stock over three years. In short, due to the difference between CAARs and BHARs result from the effect of (monthly) compounding, CAARs is a biased estimator of BHARs, and the bias magnifies as the time window increases (Barber and Lyon, 1997).
It seems ambiguous as to what extent the correlation between the presence of venture capitalists and the long-term IPOs stocks returns can be interpreted as a causal
relationship rather than merely a correlation. This could potentially give rise to an endogeneity problem caused by simultaneous causality between the explanatory variable VC and the dependent variable, abnormal returns. On one hand, since VCs tend to continue to hold a significant amount of equity in the IPOs firms after the firm goes public (Gorman and Sahlman, 1989; Megginson and Weiss, 1991), it could contribute to the IPOs firm’s long-run stock performance through actively monitoring the overall operations of the firm, providing necessary funds vital for growths and influencing the firm's future direction and managerial decisions by taking sits on the board of directors. On the other hand, it could also be that the venture capitalists are skilled at picking the potential winners in the stock market, thus the startups with higher expected returns after IPOs attract more venture capital funding. If this is the case, the coefficient on the variable VC can only depict the correlation between the returns and the existence of venture capital rather than saying that the venture capitalists’ engagement lead to the success of the IPOs firms, that is, saying that there is a causal relationship between the dummy variable VC and the returns on the IPOs stocks. Within the two effects of VCs, the first one is discussed extensively in many of the similar academic papers,
furthermore, it is shown that VCs hold a firm’s equity for 5 years on average (Sahlman, !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
3!CAARs is formulated as the mean of cumulative abnormal returns (CARs), in which the CARs is the
1990). Therefore, it is plausible to assume that if the coefficient on VC and the stock returns is statistically significant, a causal relationship between VC and the returns on the IPOs stocks is a more feasible explanation. If the venture capitalists’ main concerns are more about cashing out their investments when the return is maximized, then
fundamentally their role can be best describe merely as activist investors. Nevertheless, this explanation contradicts with the finding that verified the monitoring and mentoring roles of the venture capitalists (Barry et al., 1990), thus, the indication is that the venture capitalists are value added to the startups they invested in rather than being merely activist investors.
3.2 Data and descriptive statistics
This research uses a cross-sectional dataset including firm characteristics and stock returns. Ordinary least squares (OLS) regression is used to obtain the test statistics necessary to examine the hypothesis. The list of firms that went public in the period between January of 2004 to December of 2007 is obtained from the database of Thomson One. At this stage of retrieving data, I apply the criteria which are aligned with most of the recent IPOs research to get the data that is relevant in this context. The criteria are as follows: within all the data already identified as technology stocks listing in the US stock market in terms of SIC code4, the data is excluded if the offering has an offer price below
$5 and if it is unit offers, American Depositary Receipts or limited partnership (i.e. closed-end fund). The list contains essential information regarding the name and identifiers of the issuing firms, IPOs offer date and whether the firm is supported by venture capital when the IPOs took place.
The source of the returns on IPOs stock is DataStream, in which the monthly return is retrieved first and then the three-year buy-and-hold return is calculated with the help of Excel. Observations without ISIN number or without a record of returns in DataStream !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
4!These SIC codes are: 3571, 3572, 3575, 3577, 3578 (computer hardware), 3661, 3663, 3669
(communications equipment), 3671, 3672, 3674, 3675, 3677, 3678, 3679 (electronics), 3812 (navigation equipment), 3823, 3825, 3826, 3827, 3829 (measuring and controlling devices), 3841, 3845 (medical instruments), 4812, 4813 (telephone equipment), 4899 (communications services) and 7371, 7372, 7373, 7374, 7375, 7378, and 7379 (software). The definition is taken from the research of Loughran and Ritter
are removed from the database. In order to obtain BHARs, the returns on two benchmark portfolios are downloaded from CRSP database and then the buy-and-hold returns are calculated. BHARs is readily available after matching each stock to the market returns in the corresponding period. It is worth noting that one of the market indexes used here as a proxy for expected investment returns is a value-weighted index (S&P 500 index) while the other, NYSE/AMEX index, is equal-weighted. Incorporating both of them into the calculation process allows this research to provide a conclusion that is more robust. Nevertheless, since most of the IPOs firms are small in terms of market capitalization (Brav and Gompers, 1997), theoretically the results calculated using value-weighted index might lead to flawed interpretations and/or conclusions. Thus, the BHARs determined using the S&P 500 index should be handled cautiously in the next section. Two control variables are collected from the Center for Research in Security Prices (CRSP), in the CRSP/Compustat Merged database. If the data does not exist in the quarterly updated database, the annual ones are used instead. The size of the IPOs firms is measured as the number of shares outstanding at the end of the month in which the equity offerings take place, multiplying by the price per share on the same date. The book-to-market ratio is defined as the book value of the IPOs firms, available within one year of the IPOs, over its market value under the same fiscal year. In particular, the book value is defined as the sum of two balance sheet items available in Compustat: the book common equity plus the balance sheet deferred taxes and investment tax credits. These definitions are aligned with the ones in the research of Brav and Gompers (1997) as well as most research concerning stock returns, for instance, the paper of Fama and
French(1992) which looks into the combining effects of firm size and book to market ratio on average stock returns. The observation is excluded from the dataset if the item(s) required to calculate the control variable is missing. Observations with a negative book to market ratio will also be removed. As a result, 21 out of 168 observations for firms that went public between the year of 2004 and 2007 are eliminated due to the lack of data on the variable size. On top of that, 4 out of 168 observations are deleted as a result of a negative book to market ratio or missing balance sheet data on book value/market value. Table I reports the numbers of IPOs by year, from 2004 to 2007. During this time period 143 technology firms in total conducted an IPO, in which 103 out of 143 offerings
were supported by venture capitals. As discussed in section 2, the majority (72%) of the IPOs in the technology industry are VCs backed. The numbers of IPOs increase by year and is at its peak in the year of 2007, the year before the financial crisis in 2008 hits the stock market, one-third of the samples went public in that year. The fact that 34% of the IPOs take place in 2007 signifies that the results of the regression obtained using this dataset might be affected by the financial crisis, since the three-year buy-and-hold abnormal returns of these IPOs stocks are calculated during the periods which comprised largely of the crisis period.
Table I
Numbers of VCs backed and Non-VCs backed IPOs by year, 2004-2007
Year Venture Capital Backed Capital Backed Non-Venture Total
2004 22 (21.36%) 8 (20%) 30 (20.98%) 2005 17 (16.5%) 14 (35%) 31 (21.68%) 2006 24 (23.3%) 10 (25%) 34 (23.78%) 2007 40 (38.83%) 8 (20%) 48 (33.57%) Total 103 (72.03%) 40 (27.97%) 143 In the parentheses: the percentage of that particular year over 4 years in total. Source: Thomson ONE database
Descriptive statistics containing the information essential for inspecting the extent to which the impacts of the financial crisis has on the variables are reported in Table II. Panel A shows the BHARs, size and book to market ratio of non-VCs backed firms, and Panel B shows the ones of VCs backed firms. Three-year buy-and-hold abnormal returns determined by the S&P 500 index and the NYSE/AMEX equal-weighted index are
presented respectively in the first two rows of Panel A and Panel B. In both cases, the returns are higher when the value-weighted index (S&P 500 index) is used as the benchmark portfolio. The size of the IPO firms, in terms of mean, is larger mostly for VCs backed IPOs, although the variation is enormous as illustrated by the exceptionally large standard deviation and the gap between the mean and the median of the size, stated in the third row of Panel A and B. The returns almost doubled under the BHARs based on the value-weighted index, but the explicit interpretation of this result is unclear since the nature of event studies like this research gives equal weight to all the observations in the sample when calculating the returns on investments. Nevertheless, the abnormal returns measured by the two methods both demonstrate two attributes that are different between VCs backed IPOs and non-venture capital backed ones. First of all, the mean of three-year abnormal returns is higher for non-VC backed IPOs, as reported in the first column in both panels. However, it is not appropriate to infer that the stocks of non-VC backed firms outperformed VC backed ones since the median of the abnormal returns is significantly higher for VC backed IPOs stocks. Secondly, the variation of the returns is larger for non-VC backed IPOs stocks, as shown in the column "Standard deviation". One noteworthy point is that, under the abnormal returns calculated using NYSE/AMEX equal-weighted index, the difference in the median of the abnormal returns for IPOs stocks is 14.4%. The exceptionally large discrepancy between the mean and the median of the abnormal returns could potentially be explained by the presence of some extreme values in the dataset. Most of the stocks with negative three-year buy-and-hold returns turn out to be the firms that went public in the year of 2007 and late 2006; which makes sense since their long-term returns are calculated largely across the crisis period.
The fact that most of the firms in the dataset went public in 2007 (33.57%, as stated in the column “Total” in table I) indicates that the distribution of the BHARs is not a normal distribution in which the majority of the observations symmetrically centered around its mean. Instead, the distribution of BHARs is positively skewed (i.e. right skewed), with relatively more returns cluster toward to the left. Without further
inspections regarding the specific circumstance of the firms with relatively more extreme long-term buy-and-hold returns, the extent to which these observations can be
from the dataset when the OLS regression is performed in the next section, nevertheless, extra treatment (presents in Section 4.2) which separates the effects of the year of IPOs from the effect of venture capitalists on the three-year abnormal returns of the IPOs stocks should be employed in order to obtain a less noisy result. The specific measure applied will be discussed in detail in the next section when the regression results are present.
Table II
Descriptive statistics on three-year buy-and-hold abnormal return (S&P 500 index and NYSE/AMEX equal weighted), size and book-to-market ratio for VC backed
and non-VC backed IPOs firms, 2004-2007
Variables Mean Median Std. dev. Min Max Sample size
Panel A: Non-venture capital backed IPOs firms BHARs (S&P 500 index) 17.35% -2.02% 0.9180 -110.72% 357.54% 40 BHARs (NYSE/AMEX equal-weighted) 9.17% -15.73% 0.9143 -111.63% 354.69% 40 Size 985.368 0 598.925 6 1614.76 25.8209 10040.9 40 B-to-M ratio 0.2114 0.1335 0.2374 0.0028 0.9557 40
Panel B: Venture capital backed IPOs firms BHARs
(S&P 500 index)
BHARs (NYSE/AMEX equal-weighted) 7.37% -1.29% 0.7757 -150.94% 252.67% 103 Size 75722.5 7 388.048 5 411482. 7 44.47602 4010854 103 B-to-M ratio 0.1968 0.1617 0.1720 0.0035 1.3522 103
Sources: Center for Research in Security Prices (CRSP) and CRSP/Compustat Merged database.
To further analyze the divergence between the two samples summarized in Table II, Table III presents a test for statistical difference in mean in stocks returns and IPOs firm characteristics grouped by whether the firm is backed by VCs. The t-statistics suggest that the three-year buy-and-hold returns of the IPOs stocks, the size and the book-to- market ratio of the IPOs firms are not statistically different under 10% significance level. Although the gap of the size between the two samples is seemingly large in absolute value, the substantial variance in sizes of the firm (as shown in table II, the standard deviation of the variable size is 1615 for non-VC backed firms and 411483 for VC backed ones) lead to a test statistic that is not significant enough.
Table III
Descriptive statistics and t-test for difference between VC backed and non-VC backed IPOs, 2004-2007
Variables Venture Capital Backed Capital Backed Non- Venture Difference in Means t-stat
IPOs returns 6.27% [-0.11%] 15.8% [-0.04%] 0.409 Size 75722.57 [388.05] 985.37 [598.93] -1.146 B-to-M ratio 0.20 [0.16] 0.21 [0.13] 0.408
Observations 103 40
Size is in 1000s; IPOs returns: 3-year buy-and-hold returns. In [ ]: the median
4. Empirical results and analysis 4.1 Empirical results
Table IV presents the correlations between each independent variables. Notably, the correlation between the logarithm of firm size and the logarithm of book-to-market ratio is positively correlated with 5% significance level. Therefore, in the regression results present below, these two independent variables should not be included in the same regression, or it might result in multicollinearity problem. Table V presents the OLS regression results and t-statistics calculated using the heteroscedasticity-consistent method for testing the hypothesis that IPOs firms with venture capital funding perform better than the ones without VC funding. Table V demonstrates the results of the
regression of three-year BHARs on VC, ln(Size) and ln(Book-to-market ratio), using the dataset consisting of IPOs stocks in the US technology industry from the beginning of 2004 to the end of 2007. The dependent variable three-year buy-and-hold abnormal returns with NYSE/AMEX equal-weighted index as the benchmark for returns of the market portfolio. The coefficient on VC is not statistically significant with 10% significance level as reported in all the regressions in Table V, indicating that the
supports of venture capitals do not have an effect on the IPOs stocks three-year abnormal returns, after controlling for firm size or book-to-market. The null hypothesis that the participation of the VCs does not enhance the performance of the long-term stock of the firms they take public is thus not rejected.
Table IV
Correlations between independent variables: VC, Size and Book-to-Market ratio, 2004-2007
Panel A: Without log transformation
VC Size B-to-M ratio
VC 1.00
B-to-M ratio 0.1064 0.0245 1.00 Panel B: With log transformation
VC ln(Size) ln(B-to-M ratio)
VC 1.00
ln(Size) 0.0642 1.00
ln(B-to-M ratio) 0.0194 0.1846** 1.00
* Significant at the 0.10 level ** Significant at the 0.05 level *** Significant at the 0.01 level
The coefficients associated with the control variables, the natural logarithm of size and book-to-market ratio, are positive and significant under both regressions shown in the second to the third column in Table V. For instance, the logarithm of firm size has a positive impact on the IPOs stocks’ long-term abnormal stocks returns, and is significant under 5% significance level. The general interpretation is that the larger the size, and the smaller the book-to-market ratio of the firm (due to the nature of the natural logarithm, as the value of the book-to-market ratio is mostly positive decimals between the value of 0 and 1), the higher the three-year abnormal returns are predicted to be. These results regarding the control variables are aligned with what has been hypothesized in section 3. The results of regressions exhibited in Table V are mostly coherent with the findings of the paper of Brav and Gompers (1997), in which the original version of this model5 is established. Nevertheless, contradicting to their outcomes in which the coefficient associated with VCs participation is still slightly significant6, in this research, the firms with the support of VCs in the US technology industry did not demonstrate long-run performance that is better than the others. One potential explanation is that the time- windows of the dataset of this research cover the years that are sometimes considered the “bubble years” before the financial crisis in 2008. This might negatively affect the long- term returns of VC-backed firms, since it is plausible that the drawbacks of VC
participation may offset or even outweigh the benefits when they can exploit extra returns from the IPOs. It is shown that technology IPOs stocks launch during the bubble years exhibit mediocre long-term stock returns comparing to the IPOs in other sectors. Furthermore, VCs backed IPOs performed severely worse than non-VCs backed ones specifically during the bubble years (Coakley, Hadass, and Wood, 2007).
various markets especially after the subprime crisis in 2007 as the capital flowed from the housing market to the other markets; and the US stock market is one of them (Phillips and Yu, 2011). In particular, IPOs markets are relatively more sensitive to the change in investor sentiments, and thus during the bubble years, more firms take advantage of the hot market period (i.e., the bubble years in the US stock market) to launch their stock when the valuation of firms’ value tend to be inflated (Ritter, 1984). As shown in Table I in section 3, around one-third of the firms in the sample went public in the year of 2007, which is consistent with the ‘hot issue phenomenon’ characterized by Ritter (1984). The research of Coakley, Hadass, and Wood (2007) reveals that a higher proportion of low- quality IPOs are present during the bubble years since they are more likely to time the market compared to high-quality startups. Therefore, summing up previous findings and established theories, the effect of VCs is insignificantly different from 0 due to the nature of technology IPOs throughout the IPO market cycle and the effect of asymmetry
information during hot market overshadows the value-added features of VCs.
Table V
Cross-sectional regressions of Buy-and-Hold Abnormal Returns, 2004-2007
Table V demonstrates the regressions results (robust standard error applied) of the model% !ℎ#$$%&$'#(%)*+,( = ./+ .123 + .4ln 78 $ + .:ln%()<<=%><%?'#=$>%#'>8<) + A , benchmark is NYSE/AMEX equal-weighted returns.
Independent Variables
Dependent Variable: Three-Year Buy-And-Hold Abnormal Returns Venture Capital backed dummy variable -0.1798 (-0.11) -0.0694 (-0.44) 0.0083 (0.05) Logarithm of firm size 0.0409** (3.10) Logarithm of book-to-market ratio 0.1373** (1.90) Constant 0.0917 (0.64) -0.7439** (-2.52) -0.2144 (-1.28)
Adjusted-R2 0.0001 0.1011 0.0387
Observations 143 143 143
* Significant at the 0.10 level ** Significant at the 0.05 level *** Significant at the 0.01 level
4.2 Robustness Check
The statistical results of long-term stock performances are sensitive to the measurement methods, econometric methodologies applied and the sample period chosen (Ritter and Welch, 2002), therefore additional OLS regressions are performed in this section to inspect whether the conclusion and interpretations derived in the main empirical results section are robust. Firstly, to prevent possible omitted variable(s) from contaminating the regression results obtained, an extra feature of the regression is introduced to test whether the conclusion remains unchanged. Then, a different dataset containing the US
technology IPOs from January 2010 to December 2013 will be utilized to test whether the conclusion drawn is applicable to another time period as well.
Table VI reports the regression results using the same dataset as section 4.1 but in addition, it incorporates year-fixed effects. As the IPOs long-term returns are highly correlated in terms of the issuance year of IPOs, implementing the fixed effects model eliminates the potential bias caused by the high correlation in returns. The results indicate that the effects of VCs on the long-term abnormal stock returns remains indifferent from zero.
Table VI
Cross-sectional regressions on Buy-and-Hold Abnormal Returns, 2004-2007 with cohort year fixed effect
Independent Variables
Dependent Variable: Three-Year Buy-And-Hold Abnormal Returns
Venture Capital backed dummy variable -0.558 (-0.36) -0.0927 (-0.62) -0.0219 (-0.14)
Logarithm of firm size 0.1179**
(3.24) Logarithm of book-to-market ratio 0.1383** (2.37) Constant 0.1190 (0.92) -0.6294* (-2.40) -0.1948 (-1.06) Adjusted-R2 0.0001 0.1008 0.0384 Observations 143 143 143
* Significant at the 0.10 level ** Significant at the 0.05 level *** Significant at the 0.01 level
Table VII tabulates the results of the regressions with the same regression model as proposed, but a different set of data is employed. The dataset consists of 114 observations (within which 75% of the IPOs are backed by VCs) for firms in the US technology industry which went public between the year of 2010 and 2013. The sources and the definition of variables of the dataset are the same as described in Section III of this research. The purpose of running this additional regression is to assess the external validity of the inferences made in section 4.1. As shown in the first column of Table VI, the coefficient associated with VC returns a heteroscedasticityconsistent tstatistics of -0.36, which infers that the coefficient is insignificant from 0. Thus, the conclusion that the effect of VCs is insignificant under this model is robust, confirmed by the fixed effect model and this regression with a dataset from different time-window.
Table VII
Cross-sectional regressions of Buy-and-Hold Abnormal Returns, 2010-2013
Table VII demonstrates the regressions results (robust standard error applied) of the model!ℎ#$$%&$'#(%)*+,( =%
./+ .123 + .4ln 789$ + .:ln%()<<=%><%?'#=$>%#'>8<) + A, benchmark is NYSE/AMEX equal-weighted returns.
Independent
Variables Dependent Variable: Three-Year Buy-And-Hold Abnormal Returns Venture Capital backed dummy variable -0.0848 (-0.36) -0.0774 (-0.32) -0.1075 (-0.44) Logarithm of firm size -0.0473 (-0.53) Logarithm of book-to-market ratio -0.0471 (-0.62) Constant 0.1898 (0.89) 0.4859 (0.77) 0.0935(0.41) Adjusted-R2 0.001 0.0039 0.0045 Observations 114 114 114
* Significant at the 0.10 level ** Significant at the 0.05 level *** Significant at the 0.01 level
The methodologies applied in this event study, nevertheless, presume several strong assumptions concerning the characteristics of the samples employed. These underlying assumptions can potentially cause the estimators to be biased and/or misleading. For instance, the use of BHARs, as other methodologies in calculating long-term stock returns, suffers from several biases due to the way it is calculated and the way it defines
the "normal" returns on investment. Lyon and Barber (1997), although promote the use of BHARs over CARs for event studies investigating the long-term returns on stocks, they specifically point out that BHARs are still vulnerable to rebalancing bias and skewness bias. The former refers to the overstating of the returns on benchmark portfolio (i.e. the equal-weighted market index) resulted from the periodic rebalancing of the composition of the index, and the latter refers to the positive skewness of the distribution of BHARs which is magnified as the BHARs takes the compounding effect into account. The presence of skewness bias may lead to an inflated standard deviation of the BHARs, thus potentially produces flawed test statistics when testing the hypothesis.
Another source that could induce imprecise outcomes is the application of OLS regression model and the assumption of normality in this research. Although the
regressions are run with robust standard errors so that the assumption of homoscedasticity still holds, assumption such as no autocorrelation is likely violated due to the nature of time series regression. Furthermore, the use of t-test from the OLS regression assumes a normal distribution of the underlying data, which is not aligned with the dataset of this research, producing less reliable results. Therefore, although feasible procedures have already been incorporated in this research, the above-mentioned constraints of the methodologies employed and the nature of the dataset used in this research may hinder the precision of the regression results and the actual magnitude of the effects of
explanatory variables on the abnormal returns.
5. Conclusion
This paper investigates whether the presence of VCs has positive effects on the long-run performance of the IPOs stocks in the US technology industry, using the data on the firms that went public between 2004 and 2007. According to the OLS regression results, it is shown that the influence of VCs is not significant under 10% significance level. To verify the results, additional regressions incorporate fixed-year effects and a dataset under a different time period (between 2010 and 2013) are run. The effects of VCs are indifferent from zero under all three regression results. That is, there is no difference in the long run performance between VCs backed and non-VCs backed stocks. This is aligned with the
findings of Coakley, Hadass, and Wood (2007), but contrary to the results report by Jain and Kini (1995). The findings of this paper do not support the certification/monitoring model nor the adverse selection model, and one of the possible interpretations is that the negative effects of VCs cancel out the positive ones, resulting in the long-term
performances that are no different from non-VCs backed firms. Another possibility is that the time period of the dataset covers the bubble years. Thus, the negative effect of market timing might have an impact on the role of VCs. The roles of VCs are complicated, so the statistical tests and the regression model employ in this research have limited ability to explain the effects of VCs. To further explore the influential factors of the role of VCs in the long run, one should consider incorporating the characteristics of the VCs such as reputations, firm ages or market shares. Both firm-specific factors of the VCs and the startups, as well as the indicators of the corresponding market, need to be taken into account when attempting to evaluate the role of VCs in the longer term.
REFERENCES
Barber, B. M., & Lyon, J. D. (1997). Detecting long-run abnormal stock returns: The empirical power and specification of test statistics. Journal of financial
economics, 43(3), 341-372.
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 going-public process. Journal of Financial economics, 27(2), 447-471.
Brav, A., & Gompers, P. A. (1997). Myth or reality? The long-run underperformance of initial public offerings: Evidence from venture and nonventure capital-backed companies. The Journal of Finance, 52(5), 1791-1821.
Coakley, J., Hadass, L., & Wood, A. (2007). Post-IPO operating performance, venture capital and the bubble years. Journal of Business Finance & Accounting,
34(9-10), 1423-1446.
Fama, E. F., & French, K. R. (1992). The cross-section of expected stock returns. The
Journal of Finance, 47(2), 427-465.
Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of financial economics, 33(1), 3-56. ISO 690
Gompers, P. A. (1996). Grandstanding in the venture capital industry. Journal of
Financial economics, 42(1), 133-156.
Gompers, P., & Lerner, J. (1996). The use of covenants: An empirical analysis of venture partnership agreements. The Journal of Law and Economics, 39(2), 463-498. Gorman, M., & Sahlman, W. A. (1989). What do venture capitalists do?. Journal of
business venturing, 4(4), 231-248.
Megginson, W. L., & Weiss, K. A. (1991). Venture capitalist certification in initial public offerings. The Journal of Finance, 46(3), 879-903.
Jain, B. A., & Kini, O. (1995). Venture capitalist participation and the post&issue
operating performance of IPO firms. Managerial and decision economics, 16(6), 593-606.
Loughran, T., & Ritter, J. (2004). Why has IPO underpricing changed over time?.
Financial management, 5-37.
Ljungqvist, A., Nanda, V., & Singh, R. (2006). Hot markets, investor sentiment, and IPO pricing. The Journal of Business, 79(4), 1667-1702.
Ritter, J. R. (1984). The" hot issue" market of 1980. Journal of Business, 215-240. Ritter, J.R. (2017). Initial Public Offerings: Technology Stock IPOs. Retrieved from
https://site.warrington.ufl.edu/ritter/files/2017/04/IPOs2016Tech.pdf.
Ritter, J. R., & Welch, I. (2002). A review of IPO activity, pricing, and allocations. The
Journal of Finance, 57(4), 1795-1828.
Phillips, P. C., & Yu, J. (2011). Dating the timeline of financial bubbles during the subprime crisis. Quantitative Economics, 2(3), 455-491.
Sahlman, W. A. (1990). The structure and governance of venture-capital organizations.
Journal of financial economics, 27(2), 473-521.
Wang, C. K., Wang, K., & Lu, Q. (2003). Effects of venture capitalists’ participation in listed companies. Journal of Banking & Finance, 27(10), 2015-2034.
Yung, C., Çolak, G., & Wang, W. (2008). Cycles in the IPO market. Journal of Financial
