Master Thesis
The Effects of Limited Attention on
Corporate Venture Capitalists’ Venture
Evaluation Process
Roy Visscher
Student number: 2484749
MSc. Business Administration - Small Business & Entrepreneurship Supervisor: Samuele Murtinu
Co-asessor: Michael Wyrwich Word Count: 15,798 University of Groningen
January 2019
Abstract: In 2017 the total value of corporate venture capital (CVC) deals increased to €38.2
billion globally. This implies an increase of 382% compared to 2012, hence CVC is clearly on the rise. CVCs provide ventures with complementary assets and are seen as one of the main drivers of the success of entrepreneurial firms. According to previous literature, CVCs use product development, market data, financial data and risk evaluation as decision criteria to evaluate ventures. It is generally assumed that CVCs come short in evaluating entrepreneurial ventures. However, literature about why CVCs come short and why CVCs use certain decision criteria is missing. In order to bridge this gap in the literature, this research is the first study that tests for CVCs’ limited attention by using a sample of 84,890 companies of which 3,802 have received CVC-backing. In order to test for limited attention, two models have been estimated. The outcomes of the analyses show a mismatch between the investment criteria CVCs use and the relation of these criteria to performance of ventures. The results suggest that CVCs’ attention is drawn to certain information, resulting in CVCs not investing in ventures with the characteristics that predict performance.
1. Index
2. INTRODUCTION ... 3
3. THEORETICAL FRAMEWORK ... 5
3.1.CORPORATE VENTURE CAPITAL ... 5
3.2.INVESTMENT SIGNALS ... 6
3.2.1.RESEARCH AND DEVELOPMENT ... 7
3.2.2.MARKET ... 7 3.2.3.FINANCIAL DATA ... 8 3.2.4.RISK EVALUATION ... 9 3.3.LIMITED ATTENTION ... 11 4. METHODOLOGY ... 13 4.1.DATA COLLECTION ... 13 4.1.1.LITERATURE ... 13 4.1.2.SAMPLE SELECTION ... 13 4.2.MEASUREMENTS ... 15 4.2.1.INDEPENDENT VARIABLES ... 15 4.2.2.DEPENDENT VARIABLES ... 18 4.2.3.CONTROL VARIABLES ... 19 4.3.ANALYSIS ... 19
4.4.VALIDITY AND RELIABILITY ... 20
5. RESULTS ... 21 5.1.DESCRIPTIVE STATISTICS ... 21 5.2.CORRELATION ANALYSIS ... 21 5.3.REGRESSION ANALYSES ... 24 5.3.1.MODEL 1 ... 24 5.3.2.MODEL 2 ... 25 5.4.ROBUSTNESS CHECK ... 28
6. DISCUSSION AND CONCLUSION ... 29
6.1DISCUSSION ... 29
6.2.THEORETICAL IMPLICATIONS ... 33
6.3.MANAGERIAL IMPLICATIONS ... 33
6.4.LIMITATIONS AND FUTURE RESEARCH ... 34
6.5.CONCLUSION ... 35
7. REFERENCES ... 36
8. APPENDICES ... 41
APPENDIX 1: ELABORATION ON SAMPLE SELECTION ... 41
APPENDIX 2: CODE BOOK ... 42
2. Introduction
Seen as one of the main drivers of the success of entrepreneurial firms (Bergemann & Hege, 1998; Croce, Martí, & Murtinu, 2013) venture capital has become an inseparable part of the funding options for entrepreneurs. Many entrepreneurs present their business plans to venture capitalists (VCs), convinced that with the desired capital and access to a network provided by the VC, their venture or business idea will be a commercial and financial success (Petty & Gruber, 2011). In most cases, entrepreneurial firms benefit from VCs in terms of growth and complementary resources such as management expertise and connections (Baum & Silverman, 2004; Bergemann & Hege, 1998; Blevins & Ragozzino, 2018). Therefore, one of the most sought-after milestones for entrepreneurs is obtaining venture capital (Blevins & Ragozzino, 2016).
To which extent do corporate venture capitalists invest in ventures whose characteristics predict future performance and how is this decision made?
This paper takes an inductive approach. Therefore, no hypotheses will be formulated. However, the theoretical framework will introduce three sub-questions that will help to answer the overarching research question. The sub-questions are as follows:
1. What criteria do corporate venture capitalists use to decide whether to invest in a venture?
2. Do these criteria predict venture performance?
3. How can the match or mismatch between investment criteria and performance predictors be explained?
In order to find an answer to the main research question and its sub questions, an extensive literature research has been performed, resulting in a framework including CVC decision criteria. Furthermore, an analysis of 84.890 companies have been made. 3.802 of these companies have received investment by a corporate VC. For this analysis two models have been built in order to directly test the limited attention effects on CVC investment decision making. The first model tests if the decision criteria that are widely regarded in the literature are actually used by CVCs. In addition to the first model, the second model tests the performance predictability of these same criteria on CVC-backed ventures. This study finds evidence for the effects of limited attention on the decision process of CVCs. Besides that, this study dives deeper in the decision criteria and gives an explanation for why CVCs choose for certain criteria. The results contribute to a deeper understanding of the CVC literature, specifically in the areas of performance, decision making and limited attention. Furthermore, this paper helps entrepreneurial ventures to understand the CVCs’ decision process and its limitations, which could help them to improve their application process for CVC-funding. Moreover, this paper can be a wake-up call for corporates, as it helps to better understand their own decision process.
3. Theoretical framework
In this section, prior literature in the field of CVC decision making will be discussed. Besides, the limited attention theory will be discussed as well. This section starts with defining corporate venture capital in paragraph 3.1. Following, a discussion on the previous work of scholars in the CVC decision making field follows in paragraph 3.2. And finally, a discussion on previous work in the limited attention field is shown in paragraph 3.3.
3.1. Corporate Venture Capital
Different types of venture capital exist, for instance: business angels; independent venture capital (IVC); banking venture capital (BVC) and corporate venture capital (CVC). Business angels are wealthy individuals who invest traditionally in the earlier stages of entrepreneurial ventures (Hsu, Haynie, Simmons, & McKelvie, 2014). Venture capitalists on the other hand, are funded by other means. These means include banks (BVC), non-financial corporations (CVC) and limited partnerships by financial institutions (IVC) (Chemmanur, Loutskina, & Tian, 2014). CVCs tend to have longer lifespan than IVCs and are less performance-driven, instead CVCs focus more on enhancing their parent’s firm performance (Chemmanur et al., 2014). IVCs tend to invest both in new ventures to create market value and in existing ventures to improve performance (Gifford, 1997). However, CVCs invest in more mature and potentially less risky ventures than in businesses that are in their seed-stage (Dushnitsky & Shapira, 2010). Investment banking (BVC) on the other hand, tends to solely invest in later stages of the venture (Chemmanur et al., 2014).
Although the differences between VCs, some similarities exist; CVCs itself perform at least as well as other VCs (Dushnitsky & Shapira, 2010). Moreover, Kim, Kim, & Lee, (2011) and Gompers & Lerner, (1998) observed no difference between the performance of ventures that were CVC-backed and ventures that were IVC-backed.
by creating a dedicated CVC-unit that acts as an intermediary between the corporate and the venture (Dushnitsky, 2009; Röhm, 2018). These CVC-units, or CVC-subsidiaries are stand-alone units but always behave in the interest of the mother company (Chemmanur et al., 2014). CVC investments are generally divided in two fundamental objectives: strategic investments and financial investments. Strategic investments are primarily aimed to create synergies between the corporate and the venture, enhancing sales, profits and knowledge of the corporate’s own business as a result. Financial investments are mainly made for attractive returns (Chesbrough, 2002). CVCs provide complementary assets and capabilities to the venture, helping the venture overcome financial constraints (Chesbrough, 2002; Röhm, 2018). CVCs evaluate many business plans and funding requests a year (Petty & Gruber, 2011), but not every request is suited for investment or does fit the goals of the corporate. Therefore, a strict selection is made based on certain investment criteria. Paragraph 3.3. elaborates on the decision criteria.
3.2. Investment Signals
Fund managers aim to build a portfolio of promising companies (Buchner, Mohamed, & Schwienbacher, 2017), therefore CVCs use a certain method to select these promising companies. This method can be divided into four phases: deal origination; deal screening; deal evaluation and deal structuring. In the first phase, deal origination, the CVCs main task is to discover investment opportunities. The second phase is where the CVC reduces the amount of opportunities to a manageable amount by declining the proposals that are right away not of interest to the CVC. In the third phase, deal evaluation, the CVC carefully evaluates the selection that was made in phase two. Lastly, in phase four, deal structuring, the CVC and the venture structure the terms and conditions of the deal (Block, De Vries, Schumann, & Sandner, 2014; Fried & Hisrich, 1994; Kollmann & Kuckertz, 2010). The decision process of the CVC belongs to the evaluation phase. Therefore, this research will mainly stick to the evaluation phase only.
with product or service development. Table 1 at the end of this section shows a summary of the presence of these criteria in the existing literature.
3.2.1. Research and Development
The first criterion is research and development (R&D). R&D is an important criterion for CVCs because of two distinct reasons, which can be linked to the two fundamental investment objectives, financial and strategic, developed by Chesbrough, (2002). Starting with the financial objective, ventures that are active in highly differentiated industries invest substantially in research and development. In high technology markets, such as the biotechnology industry, R&D is of great importance. Many projects are high risk and R&D-intensive (Champenois, Engel, & Heneric, 2006). Investing in R&D is necessary to stay ahead of competition. Deriving from this, R&D expense could therefore indicate competitive advantage, market leadership and involvement of the entrepreneur (Barker & Mueller, 2002). These are all potential indicators of performance of the venture, and therefore of interest for CVCs. CVCs tend to “pick winners” by investing in ventures which are actively pursuing R&D (Baum & Silverman, 2004). In other words, when a CVC invests in a R&D pursuing venture, chances are high that the it is a valuable investment. Miloud, Aspelund, & Cabrol, (2012) examined over a hundred ventures and would agree to the authors named above; they found that ventures that are active in these highly differentiated industries, receive higher valuation of Venture Capitalists. This indicates that CVCs use Product/Service Development as a decision criterion.
As regards the strategic objective, corporate venture capitalists often invest in ventures to gain new technological insights, thus enhancing their own R&D activities (Dushnitsky & Lenox, 2005; Sahaym, Steensma, & Barden, 2010). Ventures with high technological opportunity and weak appropriability are appealing to these CVCs because these ventures may yield higher returns on knowledge production compared to the CVC’s internal R&D (Basu, Phelps, & Kotha, 2011; Dushnitsky & Lenox, 2005).
In summary, R&D expense is of importance for CVCs because of two reasons. Not only because it can indicate that the venture performs well or will perform well in the future, but also because the knowledge of the venture could be of great interest for the corporate behind the CVC itself.
3.2.2. Market
market are amongst the most essential evaluation criteria for CVCs (Dushnitsky, 2009; Macmillan, Siegel, & Narasimha, 1985; Siegel et al., 1988).
As a rapidly growing market enables faster growth for ventures and it ensures that incumbents can maintain a strong financial performance (Porter, 1980), it also allows the entrepreneur to make some mistakes because they are not straight away fatal for the venture. Because of the maintaining of the strong financial performance and mistakes made by the entrepreneur being of less danger for the venture, the investment comes with less risk compared to investments in slower growing markets. Therefore, CVCs usually give ventures in these rapidly growing markets higher valuation. Besides rapidly growing markets, markets of considerable size are more preferred by CVCs (Macmillan et al., 1985; Miloud et al., 2012; Petty & Gruber, 2011). In summary, Corporate VC fund managers give the greatest emphasis to issues regarding the market, for example the potential of growth, the market need and the competition in the market (Mason & Stark, 2004). Some authors even state that the market potential is the first criteria that CVCs assess (McNally, 1997; Petty & Gruber, 2011).
3.2.3. Financial Data
On the contrary, several authors state that financial returns are not the most important, mainly for two reasons. First, many CVCs mainly focus on a strategic benefit instead of a financial return (Dushnitsky & Lenox, 2005; Macmillan et al., 1985). Second, although ventures with healthy financials and a stable amount of financial resources could be very appealing to a CVC, these ventures might not be in need of venture capital investment because they have sufficient capital available (Baum & Silverman, 2004).
Concluding, although some evidence exists that financial data is not of much interest for CVCs, the vast majority of scholars agrees that CVCs consider financial data before making an investment. Thus, financial data lies at the foundation of many CVC-investments, regardless of the main objective of the investment.
3.2.4. Risk Evaluation
Risk evaluation is closely related to financial data, because both criteria can on some surfaces overlap. To illustrate; a financial unhealthy company comes with more investment risk. Every CVC investment includes risk; for instance, the Dutch word of Venture Capital translates to “Daring Capital”. Risk reveals itself in many forms. For example; market risk, the risk of unforeseen competitive conditions (Fiet, 1995); product risk, the risk of unexpected negative conditions regarding the product, e.g. technical flaws or rise of resource costs (Sykes, 1986); human risk, the risk resulting from humans such as mistakes by employees or illness of employees (Sykes, 1986); agency risk, the risk that either the venture or the CVC will pursue its own interest rather than adhere to what was agreed between both parties (Fiet, 1995). Finally, economic risk, the uncertainty with regard to future outcomes such as survival of the company (Hsu et al., 2014). Risk often comes from high levels of uncertainty during the investment process, resulting from lack of quantifiable financial and market data and unanticipated competitors (Ruhnka & Young, 1991).
to selectively acquire information that indicates risk. However, it is impossible to take all possible kinds of risk into account. Therefore, CVCs specialise in certain risk types. This puts the CVCs in a better position to evaluate those types of risks (Fiet, 1995).
Concluding, because of the importance of risk for CVCs before and during the investment, many scholars agree that CVCs make an assessment of the risk involved before making an investment. No matter which investment-objective the corporate maintains, risk management will always be of importance (Chesbrough, 2002).
Altogether, decades of research in corporate venture capital and investment criteria, starting with the research done by Macmillan and colleagues in 1985 on investment criteria, brought a consensus. The decision criteria used by venture capitalists have been mapped (reported in table 1), and it seems that these theories are widely adopted. However, what almost all studies tend to neglect are the underlying motives for these criteria. Besides, little evidence exists on why corporates use these criteria in particular.
Table 1: Presence of decision criteria in literature
Research and
development Market data Financial data Risk evaluation
(Macmillan et al., 1985) x x x
(Sykes, 1986) x
(Sandberg & Hofer, 1987) x
(Siegel et al., 1988) x x x
(Ruhnka & Young, 1991) x
(Fiet, 1995) x
(McNally, 1997) x
(Zacharakis & Meyer, 1998) x
(Chesbrough, 2002) x x x
(Baum & Silverman, 2004) x x
(Mason & Stark, 2004) x x
(Dushnitsky & Lenox, 2005) x
(Champenois et al., 2006) x
(Dushnitsky, 2009) x
(Sahaym et al., 2010) x
(Basu et al., 2011) x
(Petty & Gruber, 2011) x x
(Miloud et al., 2012) x x
(Brush, Edelman, & Manolova,
2012) x
(Hsu et al., 2014) x x
(Wadhwa, Phelps, & Kotha, 2016) x
et al., 2014; Ruhnka & Young, 1991). A problem arises when the informed party has an incentive to misrepresent the given information. For instance, the entrepreneur could benefit from not stating certain flaws in his or her product (Block et al., 2014; Kollmann & Kuckertz, 2010). Besides unavailability of information, an abundance of information can also rise a problem for CVCs. For instance, it could leave the CVC unable to process all information. This results in CVCs making irrational investment decisions. Hence, the CVC has a limited amount of attention to divide among all available information (Boyaci & Akcay, 2017). The remainder of this section will explain the limited attention theory to a further extent and links this theory to CVCs.
3.3. Limited attention
In a utopian world, a corporate venture capital investor would be able to make its investment decision based on all available information. However, in the real venture capital world, CVCs face serious constraints in collecting all available information, human capital and financial resources (Gifford, 1997; Siegel et al., 1988). Due to limited attention, CVCs are not able to evaluate every venture into detail. For this reason, when a venture has been evaluated by a CVC, this same CVC may be restrained from evaluating another venture because it cannot evaluate every investment alternative (Wadhwa & Basu, 2013). This results in high opportunity costs when making the eventual investment (costs expressed in the yields of the best alternative investment) and uncertainty with each investment (Jääskeläinen et al., 2006; Wadhwa & Basu, 2013).
4. Methodology
In this section is described how the data was collected in paragraph 4.1, followed by how the was sample selected. Next in paragraph 4.2 is described what variables were measured in order to find an answer to the research question. Furthermore, in paragraph 4.3. is described how the variables were measured. In paragraph 4.4. the models that were built in order to find an answer to the research question are described. Finally, paragraph 4.5. covers issues regarding validity and reliability.
4.1. Data collection
4.1.1. Literature
Secondary data sources have been used for collecting the data for this research. First, an extensive study on venture capital literature has been done in order to find decision criteria that CVCs use in evaluating ventures. Only literature of highly acknowledged journals and books have been used. To verify the quality of the journal, a measurement of article influence by Eigenfactor has been used (Bergstrom & West, n.d.). Next, all decision criteria have been rated on presence in the literature over the time period ranging from 1985 up to 2018 (see Table 1). This resulted in a framework including the top five most mentioned decision criteria. One of these criteria has been dropped, because it is impossible to measure by the data available for this research, see section 6, paragraph 4 for a more in-depth explanation on the reason why this criteria has been dropped.
4.1.2. Sample Selection
This paragraph explains which steps have been undertaken to select the sample. Extra steps have been taken to ensure the quality of the data, these steps are reported in appendix 1. In order to obtain data about CVC investments, two databases have been used and the information of these databases have been carefully merged. To begin, the Thomson Reuters Eikon database has been used to retrieve information on CVC-investments. Eikon is one of the leading databases of financial market data (Refinitiv, n.d.).
Every single CVC investment included in the database, ranging from January first, 1968 up to December 31st, 2017 had been exported. This resulted in a dataset including 40,154 corporate
database resulted in a dataset including financial information on 110,206 companies, worldwide. In order to give every CVC-investment an own unique identifier, a new variable named ‘Company Code’ was created. Lastly, in both datasets all entries without a street address, company name or year were deleted. In order to be able to make the intended analyses, these two datasets from different sources need to be merged into one dataset. Below the process of merging is described in detail.
Creating dataset
Because both datasets origin from a different source, some extra steps had to be taken in order to merge these datasets. A matching protocol (Doherr, 2018) was used to ensure that the merge would be precise, careful and reliable. The first step to ensure that both datasets are safe to merge, was dropping duplicates. In the Compustat dataset all duplicates based on the unique company code labelled ‘gvkey’ were dropped using Stata SE 15 (Statacorp, 2017). Regarding the Eikon dataset, multiple investments by different CVCs in the same venture in the same year occur. This would result into problems when merging the two datasets. Therefore, it has been decided to delete duplicate entries resulting in a dataset of 19.177 entries, in which every entry represents a unique venture. This is still a substantial amount of data for a thorough analysis. The last step before merging both datasets was deleting all variables, keeping only year of investment, target company name, company address, and company code in order to not overload the computer software. To make the files suited for the matching protocol, the files have been converted to tab delaminated text files.
Using Search Engine Algorithm
The first step was to import the Eikon dataset as a base table and the Compustat dataset as the search table. After importing both files, the search fields had to be matched. All variables in the Eikon dataset were matched to their counterparts in the Compustat dataset. One of the powerful features of the matching protocol is that it can redesign the variable layout. The algorithm has been set up in a way that it equalled all variables by converting upper cases to lower cases; gathering single letters into one word; deleting umlauts and separating numbers from letters. This resulted that small differences in layout in both datasets could be ignored, resulting in more and better matches. The next step was to define weights to all variables, by using these weights the algorithm is instructed on which variables it should give the most importance during the merging process. After experimenting and several tries, the weighting allocated to each variable was as follows:
• Name 80%; • address 5%;
The company name is the most important matching variable because in both datasets, each company name only appears once and by using the redesign, all names are in the same layout. The address has a far lower weighting, because even though the layout between the two datasets was similar, the notation of the addresses was not. E.g.: Fifth Street versus 5th Street.
These notations have not been corrected with the matching protocol because it would make the results unreliable. The investment year has been taken into account to control for the company data not being from the same year as the investment date.
The limit for adding results to the results table has been set to 85% match, the results table included a match percentage column. After the merge had been executed by the algorithm, it has been decided to only keep data with a 95% match. The result table only consists out of the variables mentioned above, therefore the table was converted to a Stata file. By using a one-to-one merge based on ‘gvkey’ in Stata, all Compustat company info has been re-added to the merged dataset. Resulting in a sample including 84,890 companies, consisting out of 3,802 companies in which a CVC investment has been made and 81,088 companies which have not been backed by a CVC.
4.2. Measurements
All variables could be measured by many different indicators. The measurements used in this study are based both on existing literature and data available in the created dataset. Some measures were readily available, others have been calculated by using the available data. A codebook is presented in appendix 2.
4.2.1. Independent Variables
All decision criteria could be tested on several different proxies. It is likely that CVCs use multiple measures per decision criterion, therefore the models include multiple measures.
4.2.1.1. Measuring product development
4.2.1.2. Measuring market data
Market characteristics are hard to measure, because not many variables that measure the market characteristics are available within the used data. However, some variables can be used as indicators for the characteristics of the market. Total sales is generally regarded as a good measurement for the market (Gilbert, Mcdougall, & Audretsch, 2006). García-manjón & Romero-merino, (2012) found in their literature research that sales is an important indicator for growth and therefore of importance for measuring the market. Furthermore, market share has been used to measure for market potential of the venture. Note: all sales per market is not available, therefore sales of all available data of companies active in that main industry at the investment date is used. This data is also used to calculate market share, so market share is a relative measure that compares the venture to the other ventures available in the dataset. Stated all the above, total sales and market share are therefore used as a measurement for the investment criteria market data.
4.2.1.3. Measuring financial data
Companies use a series of financial data and measures in their financial statements; therefore, CVCs can easily access a broad pool of financial information and calculate ratios and indicators that they find of importance. This means that not simply one criterion is used. Most CVCs look for at least two things; a return on their investment and an exit strategy that reduces losses (Hsu et al., 2014; Macmillan et al., 1985). To conceptualize this, net income of the venture is the first measure for financial data. Also, a performance indicator has been included, to measure for venture health. This performance indicator is the return on assets (ROA) and is calculated as follows:
!"# = %&' )*+,-& .,'/0 /11&'1
When a venture has a lot of cash on hand, managers tend to act not in the best interest of the shareholders by overinvesting this cash. Long term debt plays a critical role in reducing abnormal investments because it has a disciplining role (D’Mello & Miranda, 2010). Therefore, the last two measures of the investment criterion financial data are long term debt and long- term debt – due in to one year.
4.2.1.4. Measuring risk evaluation
CVCs expect a venture to react well to risk, a worrying company health limits the ventures ability to react to risk (Harrison, Horngren, Thomas, & Suwardy, 2014; Macmillan et al., 1985) Three criteria that are measures for company health and therefore for investment risk are computed from the data.
Doubtful receivables can be best defined as all receivables that are most likely not able to be collected by the company. High amounts of doubtful receivables can be a sign of future problems, as the company is incapable of collecting payments for its products or services (Harrison et al., 2014). To conceptualize this, a new variable, the doubtful receivables ratio was created. This ratio indicates the amount of doubtful receivables compared to the total receivables. A higher value for this variable thus means that a bigger proportion of all receivables are doubtful. The doubtful receivables ratio is calculated as follows:
2,34'530 6&+&)7/40&1 6/'), =2,34'530 6&+&)7/40&1 .,'/0 6&+&)7/40&1
Furthermore, working capital is a measure of the company’s short-term financial health. Working capital has effects on a firm’s profitability and risk (Smith, 1980). The working capital ratio, or current ratio, indicates if a company has enough term assets to cover its short-term debts. When this ratio is lower than one, the company might have serious liquidity problems (Harrison et al., 2014). The current ratio is calculated as follows:
8! = 8366&*' #11&'1 8366&*' 9)/4)0)')&1
Besides the current ratio, also the change in working capital have been used as a measure for risk. A negative value for this variable indicates a decline in working capital, a big decline indicates that the company could run into liquidity problems (Smith, 1980).
(unqualified and unqualified with explanatory language) and a ‘0’ for a negative opinion (qualified, refusal of opinion and adverse opinion).
4.2.2. Dependent Variables
The first step is to test if the criteria from decision criteria framework formulated in paragraph 3 are actually used by CVCs. Consequently, model 1 has been estimated. This model estimates the impact of decision criteria on CVC-backing. Model 1 is constructed by using a dummy for CVC-backing as a dependent variable. This variable includes an ‘1’ if the venture is backed by a CVC and a ‘0’ if the venture is not backed by a CVC. Furthermore, model 1 includes measurements of the decision criteria as independent variables (see section 4.2.1.). The second step is to test if these same decision criteria tested in model 1 do predict performance of the CVC-backed venture, for this reason model 2 is estimated. Model 2 includes three different tests each on a different performance variable, these performance variables represent the dependent variables. Furthermore, the same independent variables as used in model 1 are included. However, model 2 tests only on the companies which are CVC-backed. The remainder of this section defines the three performance variables that are used.
Measuring performance
Performance is measured on three variables, first by measuring profitability by return on equity, secondly by measuring the ventures efficiency and third for initial public offering.
Shareholders are interested in return on equity and less interested in the return on assets, because assets include resources that are not owned by the shareholders. Since a CVC is shareholder of its portfolio company, it would be interested in ROE as a performance measure. ROE is calculated as follows:
!": = %&' ;*+,-&
.,'/0 #11&'1 − .,'/0 2&4'1
Efficiency is an important measure because it is the relation between the quality of the performance and the effort invested in (Kahneman, 1973). Higher efficiency results in better returns (Zacharakis & Meyer, 1998). The efficiency of a venture can increase as a result of capital investment. To measure efficiency, the asset turnover ratio (ATR) have been used. This ratio increases as the venture generates more capital per asset. ATR is calculated as follows:
Finally, initial public offering (IPO) was used as measure for performance. By creating a new variable that indicates if the year of IPO was later than the investment year, a performance measure based on the IPO year was created. This variable is named IPO timing and is constructed as follows; every CVC-backed company that went public after CVC-backing is represented by an ‘1’. All other companies that do not belong to this group are represented by a ‘0’.
4.2.3. Control Variables
Various aspects of the ventures and CVCs have been controlled for in order to eliminate both alternative explanations for the decision process of CVCs, and performance of CVC-backed ventures. The sample includes companies active in 74 main industries, some industries are more innovative than others, which could result to ventures performing better in these industries. Also, some industries are more preferred by CVCs than others, this could lead more investments in these industries. To control for industry differences, dummy variables for SIC main industry were created. Besides differences between industries, differences between countries need to be controlled for. Countries can for example differ in regulation, business climate and costs of currencies, all of these differences can be of influence for CVC preference and venture performance. Therefore, a dummy variable for countries in which the venture is based was created. An overview of these dummy variables is presented in appendix 2. Finally, a control variable for company size is created. As company size can also be of influence on the performance of the venture and the attractiveness for the CVC. Company size could be measured in different ways, for example in sales. However, a company could spend years to develop a product before going to market, while making no sales. Therefore, performance is measured in the number of employees (Gilbert et al., 2006).
4.3. Analysis
Two models have been developed in order to answer the research question. Model 1 measures which of the decision criteria found in the literature are actual used by CVCs. In other words, did ventures that scored well on the decision criteria receive CVC-backing? To find the relationships between CVC-backing and the decision criteria, a regression has been estimated in which the dependent variable is CVC-Backing. Because this variable is binary, a probit regression was used.
variable, a probit regression was used for the model in which IPO timing is the dependent variable. For both models a two-tailed test has been used. The two-tailed test is most appropriate because the relationship between the dependent and independent variables could be in both directions.
The interpretation of the results of both models led to insights regarding the decision criteria of CVCs. The results were used to indicate if a match or mismatch exists between the CVC decision criteria and performance of the venture, in section 5, the results of the tests on both models are reported. Moreover, section 6 shows an interpretation and argumentation for these results.
4.4. Validity and Reliability
5. Results
5.1. Descriptive Statistics
In Table 2 the characteristics of the data are presented. A more elaborate codebook is available in appendix 2. Surprising is that the two performance variables ROE and ROA have negative means. Indicating that many ventures do not perform well. On the other hand, the standard deviation of these variables is quite large, indicating a big spread in values.
Table 2: Descriptive statistics
Variable N Mean SD VIF
Dependent Variables
CVC-backing 84,890 .044 .204
Return on equity 19,234 .264 40.410
Asset turnover rate 57,673 1.078 4.271
IPO timing 84,890 .019 .137
Independent Variables
Research and Development
Product development expense 14,778 1332.058 51820.8 2.06
Market data
Total market sales 84,890 1.12*107 6.65*107 1.02
Market share 67,997 529.827 52553.99 1.09
Financial Data
Long-term debt due in one year 49,413 13461.26 749253.6 3.78
Long-term debt – total 66,448 128708.1 1.32*107 6.38
Return on assets 20,760 -.601 11.832 1.01
Net income 20,899 8.578 159.358 2.52
Risk Evaluation
Current Ratio 55,636 7.110 205.328 1.04
Doubtful receivables ratio 8,192 .101 1.277 1.03
Change in working capital 7,659 1787.66 100929.5 2.37
Auditor opinion 84,890 .181 .385 1.02
Control Variables
Company size – measured in
employees 20,563 3.004 15.552
2.78
Main industry dummies INCLUDED
Country dummies INCLUDED
5.2. Correlation Analysis
CVC-Backing weakly correlates significant with auditor Opinion (r = .108, p < .01). This seems to indicate that a more positive auditor opinion increases the change on receiving CVC-backing. Product development expense correlates significant with both long-term debt – total (r = .439, p < .01) and net income (r = .481, p < .01). This indicates that a higher long-term debt and a higher net income leads to higher product development expense. This could indicate that ventures use funds acquired by net income and loans for research and development. Furthermore, product development expense shows signs of a significant weak correlation with market share (r = .103, p < .01) and long-term debt due in one year (r = .141, p < .01). This seems to indicate that ventures with higher product development expense also have more long-term debts that are due in one year and have a bigger market share.
Remarkable is the evidence for a significant strong correlation between market share and long-term debt – total (r = .757, p < .01). This seems to indicate that ventures with a bigger market share do have more long-term debts. A reason could be that in order to acquire this market share, ventures used a lot of external capital such as loans. Next, long-term debt – total has a significant moderate correlation with change in working capital (r = .373, p < .01). It seems that higher long-term debts lead to a bigger positive in working capital, thus an increase in working capital. A reason for this could be that the financing used by ventures results in an increased cashflow.
Table 3: Pair wise correlation matrix 1 2 3 4 5 6 7 8 9 10 11 12 13 1. CVC-Backing 1 2. Product development expense -.007 1 3. Total sales per industry per
year
.002 .087*** 1
4. Market share -.002 .103*** -.002 1 5. Long-term debt due in one
year -.002 .141*** .049*** .061*** 1
6. Long-term debt – total -.002 .439*** .004 .757*** .081*** 1
7. Return on assets -.014** .007 -.002 .001 .001 .001 1
8. Net income -.014** .481*** .008 .005 .013* -.009 .008 1
9. Current ratio -.002 -.003 -.003 -.001 -.001 -.001 .004 -.003 1
10. Doubtful receivables ratio -.006 -.007 .013 -.016 -.001 -.001 -.047*** -.004 -.001 1
11. Change in working capital -.004 -.001 .096*** .025** .096*** .373*** .001 .030** -.002 -.005 1
12. Auditor opinion .108*** -.022*** -.008** -.005 -.012** -.005 -.030*** .004 -.004 .006 -.016 1 13. Company size – measured
in employees
5.3. Regression Analyses
5.3.1. Model 1
The results of the probit regression with CVC-backing as a dependent variable are reported in table 4. The likelihood chi-square has a p-value of p = .0001 < .01 with a McFadden pseudo R2 of .1283, which indicates that model 1 is statistically significant as a whole. 12.8% of the
variance of CVC-backing can be explained by the independent variables in this model. The results show a significant relationship for three out of the four decision criteria used by CVCs.
First, product development expense shows a strong significant positive relationship (p = .001 < .01) with CVC-backing. This seems to indicate that ventures that invest more in product development, are more likely to receive backing by CVCs. Next, long-term debt due in one year show evidence for a significant positive relationship (p = .064 < .1) with CVC-backing. This seems to indicate that ventures with a long-term debt that is due in to one year are more likely to receive CVC-backing. Remarkable is that a inverted relationship (p = .002 < .01). shows that ventures with a bigger total long-term debt, seem to have a slightly smaller change to receive CVC-backing. With this in mind, these results seem to indicate that CVCs value long-term debt that is shorter due.
Second, return on assets also shows strong evidence for a significant negative (p = .008 < .01) relationship with CVC-backing. This indicates that chances are higher for ventures with a high ROA that they will not receive CVC-backing. This is rather remarkable, since it indicates that it is not of importance for CVCs if a venture is efficient at using its assets. In fact, CVCs seem to rather not invest in those ventures.
Finally, two out of four measures for risk evaluation show a significant relationship with CVC-backing. First, change in working capital shows significant evidence for a positive relationship (p = .026 < .05) with CVC-backing. This seems to indicate that ventures that experience a positive change in working capital, are more likely to receive backing by CVCs. Second, auditor opinion shows a significant positive relationship (p = .095 < .1) with CVC-backing. This indicates that CVCs value a positive auditor opinion, thus are more likely to back ventures with a positive auditor opinion. The results show no evidence for a relationship between CVC-backing and total sales per industry per year and market share. Consequently, CVCs do not take market criteria into account in their decision process.
Table 4: Probit analysis model 1
CVC-backing Variable Model 1 Research and Development
Product development expense .104***
(.031)
Market
Total sales per industry per year -.000
(0.00)
Market share .119
(.567)
Financial Data
Long-term debt due in one year .127*
(.069)
Long-term debt – total -.042***
(.014) Return on Assets -.404*** (.153) Net income -.005 (.018) Risk Evaluation Current ratio -.002 (.008)
Doubtful receivables ratio -.827
(.802)
Change in Working Capital .027**
(.012)
Auditor opinion .255*
(.153)
Control Variables
Company size – measured in employees
-.274 (.168)
Main industry dummy INCLUDED
Country dummy INCLUDED
*** p < .01; ** p < .05; * p < .1
5.3.2. Model 2
observations out of 3802 totals. Hence, excluding number of employees would not bring up any issues on the regression analyses on performance. The results of model 2 are reported in table 5.
Return on Equity
Model 2.1 is not significant (p > .1) as a whole with a R2 of .0061 (adjusted R2 -.0643), this
indicates that no evidence is present that could indicate that the independent variables explain any of the variance of ROE. In other words, the variance of ROE because of other reasons than the decision criteria. Only one of the variables, namely, product development expense show significant evidence (p = .068 < .1) for a positive relationship with ROE. Although the model as a whole does not show a good fit, the results indicate that ventures with higher product development expense are more likely to show a higher return on equity. For every increase of a million in development expense, the average ROE increases with 13%.
Asset Turnover Rate
Model 2.2 is significant as a whole (p = .001 < .01) with a R2 of .411 (adjusted R2 .369), this
indicates that the model is a good fit and that 44% of the variance of ATR can be explained by the independent variables. It seems that the independent variables are fairly strong predictors of ATR. The results show evidence for two out of four decision criteria having a relationship with ATR. Beginning with net income; the results show evidence for a positive significant relationship (p = .000 < .01) between net income and ATR. Consequently, ventures with a higher net income are more likely to show a higher ATR. This relation makes sense hence, ATR is calculated by using sales of the venture.
Furthermore, significant evidence (p = .000 < .01) indicates a negative relationship between current ratio and ATR. This means that if the CR increases, the ATR decreases. This makes sense because in a situation where the current ratio increases due to increase of assets, the ATR would decrease if sales remain the same.
Moreover, significant evidence (p = .004 < .01) is present for a positive relationship between auditor opinion and ATR. This indicates that a positive auditor opinion increases the change of a positive ATR.
IPO timing
Model 2.3 is significant (p = .000 < .01) as a whole with a McFadden pseudo R2 of .1543.
15.4% of the variance of IPO timing can be explained by the independent variables. Two proxies show evidence for a relationship with IPO timing. First, the results show significant evidence (p = .001 < .01) for an inverted relationship between market share and IPO timing. This seems to indicate that ventures with a bigger market share are less likely to go public after a CVC-investment. Second, significant evidence (p = .000 < .01) exists for a positive relationship between auditor opinion and IPO timing. This indicates that ventures which received a positive auditor opinion (unqualified with or without explanatory language), are more likely to go public after CVC-backing. In short, only the market criteria and the risk evaluation criteria show evidence for a relationship with IPO timing. Remarkable is that the relationship between market share and IPO timing is inverted, this indicates that a smaller market share could be an indicator of better performance.
Table 5: Regression model 2
Variable 2.1 ROE 2.2 ATR 2.3 IPO Timing Research and Development Product development expense .132* (.072) .001 (.002) .005 (.004) Market
Total sales per industry per year -.000 (-.000) -.000 (.000) -.000 (.000) Market share 4.208 (18.331) -.062 (.581) -8.344*** (2.579) Financial Data
Long-term debt due in one year -.020 (.095) .000 (.003) -.005 (.0169)
Long-term debt – total -.006
(.013) -.0001483 (.000) -.001 (.001) Return on Assets -.001 (.063) .002 (.002) -.001 (.004) Net income .028 (.034) .005*** (.001) -.003 (.002) Risk Evaluation Current ratio .025 (.079) -.015*** (.003) (.006) -.002
Doubtful receivables ratio - - -
Change in Working Capital - - -
Auditor opinion -.596 (2.952) .189** (.094) 1.925*** (.233) Control Variables
Company size – measured in employees
Main industry dummy INCLUDED INCLUDED INCLUDED
Country dummy INCLUDED INCLUDED INCLUDED
5.4. Robustness Check
Before 2000, investing in new ventures was very popular. Per year, many CVC-investments were executed. As the economy turned in 2000, this came to a stop (Chesbrough, 2002). Only coming back to its old level when the economy was getting stronger in 2004 (Röhm, 2018). Likewise, (Gompers & Lerner, 1998) found that CVC-investments are strongly associated with the general condition of the economy. In order to see if periods of economic downfall are of influence for the results of model 1, a second probit regression was done including exactly the same variables. Because the sample consists over a period including two economic recessions, two time periods have been excluded. The first time period that is excluded is the time period of the first recession, ranging from 2000 up to 2003. The second recession, ranging from 2008 up to 2010, has been excluded as well. As reported in appendix 3, the results stay the same compared to the original analysis of model 1. It can be concluded that the economic downturns do not have a substantial impact on the results of the analysis.
Furthermore, second regressions have been performed for model 2, the results are also reported in appendix 3. The results of these regressions show minor differences with the results of the main regressions. For instance, the analysis excluding the economic recessions show a significant positive relationship between product development expense and ATR (p = .006 < .01) and between product development expense and IPO timing (p = .016 < .05). This is not a problem because product development expense already predicts performance measured in ROE in the original model. Furthermore, in the alternative analysis evidence is found for a relationship between net income and ROE (p = .034 < .05). This is not surprising because in periods of high economic prosperity it is likely that the net income of ventures is higher and therefore its part in performance measures could be more visible.
6. Discussion and Conclusion
6.1 Discussion
The aim of this study is to find evidence for limited attention of CVCs. For this purpose, two models have been estimated to find if a match or mismatch exists between the CVC decision criteria and the indicators for performance of ventures. The limited attention theory is used to explain these matches and mismatches. In this paragraph, an answer to the main question and sub questions is reported, starting with the first sub question:
1. What criteria do corporate venture capitalists use to decide whether to invest in a venture?
Four CVC decision criteria were carefully formulated as a result of a thorough literature research. These decision criteria include: research and development; market data; financial data and risk evaluation. Each of these criteria is conceptualized by using literature, resulting in at least one measurable variable per criterion. Surprisingly, not all decision criteria show strong evidence for CVC-backing. In other words, it cannot straightforwardly be assumed that the decision criteria that dominate the current literature are actually the decision criteria corporates prefer to use. To begin with, the results of the test on R&D are in line with what was expected. To explain, the used proxy product development expense is positively related to CVC-backing, for instance; CVCs use R&D as a decision criterion. This result is analogous to what has been found by many scholars (Baum & Silverman, 2004; Dushnitsky & Lenox, 2005; Miloud et al., 2012; Sahaym et al., 2010). Second, in contrast to work of many scholars in the past (Dushnitsky, 2009; Macmillan et al., 1985; McNally, 1997; Petty & Gruber, 2011; Siegel et al., 1988) CVCs do not seem to use market related information as a decision criteria. Both market share and total market sales that test the market criteria seem not to involve any relationship with CVC-backing. Consequently, CVCs do not seem to be influenced in their decision by a small or large market, neither do they seem to be influenced by the market share of the company.
that are due in more than a year, have a negative relationship with CVC-backing. This seems to imply that CVCs are somehow reluctant to ventures with longstanding debts. This suggests that ventures with financial obligations lasting over 12 months fit less in the CVC’s investment stategy. Furthermore, return to assets is negatively related to CVC-backing. This indicates that CVCs focus at entrepreneurial ventures with a smaller ROA thus, lower efficiency.
Finally, the analyses show contradicting results on the risk criterion. Neither current ratio nor doubtful receivables ratio seem to be used as risk-measures by CVCs. Although CVCs do not rely on the CR (or working capital ratio), they do use the absolute change in working capital as a measure for risk. Furthermore, a venture’s auditor opinion is also of influence on the decision process of CVCs. This is not unusual since CVCs use risk-avoidance strategies (Fiet, 1995; Ruhnka & Young, 1991), a negative auditor opinion indicates that the financial statements of the venture could be faulty. Investing in a venture with possibly faulty financial statements comes with more risk since the actual state of the venture is harder to estimate.
In summary, the results show that all decision criteria except the market criterion are used by CVCs in their evaluation phase. To enumerate, the following proxies; product development expense; long-term debt due in one year; total long-term debt; ROA; change in working capital; and auditor opinion are all of influence on the decision process of CVCs. The following section links the decision criteria and their proxies that CVCs use to venture performance. Consequently, the second sub question will be answered.
2. In what way do these criteria predict venture performance?
The results of the analysis of model 2 do not give a unanimous view on the impact of the CVC decision criteria on performance. Hence, three measures predict performance, two measures predict a decline in performance and four measures do not seem to have any relationship with performance. This section explains which measures for decision criteria indicate an increase or decrease of performance.
In line with what Barker & Mueller, (2002) and Baum & Silverman, (2004) found, product development expense seems to have a positive impact on performance. Specifically, ventures with bigger product development expense seem to generate a higher ROE, this is interesting because R&D seems to yield both strategic and financial benefits. Especially regarding that CVCs are interested in return on shareholder equity (after all CVCs are shareholders) and not in return on resources that are not owned by the shareholders.
large market share. The second market data criterion: total industry sales does not predict performance.
Third, entrepreneurial ventures with a higher net income seem to have a higher asset turnover ratio as well. It makes sense that ventures that have a higher income can be more efficient. The other proxies for financial data do not predict performance.
Finally, a higher current ratio implies that the efficiency of that venture will be lower. Moreover, the auditor opinion seems an important predictor of performance, as it involves a positive relationship with both ATR and IPO timing. A positive auditor opinion will probably not enhance performance directly, however one can expect more issues considering the financials that could lower performance when a ventures financial statements are faulty.
To summarise, the relationship between the decision criteria and performance seems to be ambiguous. Research and development and risk evaluation seem to be the only two criteria that fully predict performance, however these two criteria could not be measured on more than two proxies. As regards market data and financial data, less than half of their proxies seem to have a relationship with performance. Thus, the results give not enough evidence to assume that market data and financial data do predict performance.
3. How can the match or mismatch between investment criteria and performance predictors be explained?
In a situation in which CVCs are not limited by any sort of constraint, thus a fully rational decision process, CVCs would only use all the decision criteria that predict performance of the venture. In contrast, the results of this research prove that CVCs do not perform a fully rational investment decision process, the remainder of this paragraph gives an explanation on why CVCs do not make complete rational decisions.
CVCs seem to rely on financial data as an investment criterion, however none of the financial measures CVCs use show a match with performance of the venture. The results seem to indicate that attention is drawn to financial data, regardless of it predicts performance. Financial data in particular is concrete, easy understandable, easy comparable and it quickly shows a pattern. That CVCs’ attention is drawn to financial data makes sense, because according to Nisbett & Ross, (1980) attention is drawn to information that is concrete and can easily be summarized.
information. As a result, ventures that report high product development expenses will attract the attention of CVCs.
Furthermore, while previous literature shows evidence for market potential being the most important criteria for CVCs (McNally, 1997; Petty & Gruber, 2011), this study suggest the contrary. CVCs seem not to use the market criteria at all in their evaluation phase, even though market share seems to be an important predictor of performance. This occurrence can be explained by the limited attention of CVCs. CVCs pursue strategic objectives by acquiring knowledge by investing in entrepreneurial ventures active in markets unfamiliar to the CVC (Dushnitsky & Lenox, 2005; Wadhwa & Basu, 2013). However, attention is drawn to information that is familiar to the CVC (Kahneman, 1973), this leads to attention drawn away from information about these unfamiliar markets and drawn into more familiar information such as financial data and auditor opinions. Moreover, it is important to note that the attention a CVC is using evaluating a company on a certain criterion, cannot be used on other criteria or another company. Altogether, this results in the CVC being unable to process all information.
In summary, CVCs seem not to be able to fully evaluate the quality of the ventures they evaluate. The results of this paper suggest that the attention of CVCs is drawn to R&D criteria and financial data criteria, while the latter does not necessarily predict performance. Considering that attention can only spent once, the above stated results in a neglect of criteria that on their turn do predict performance. In case of a fully rational decision process, one would expect that CVCs only use decision criteria that predict performance. It can be concluded that due to CVCs limited attention, criteria that do not predict performance are used while other criteria that do predict performance are not used.
Table 6: Comparison of decision criteria on usage and performance.
DECISION CRITERION
PROXY EVIDENCE FOR
USAGE BY CVC
INDICATOR FOR PERFORMANCE
MATCH
Research and
development Product development expense Yes*** ROE* Match
Market data Industry sales No No Match
Market share No IPO timing*** Mismatch
Financial data Long debt 1 year Yes* No Mismatch
Long debt total Yes*** No Mismatch
ROA Yes*** No Mismatch
Net income No ATR*** Mismatch
Risk evaluation Current ratio No ATR*** Mismatch
Doubtful Receivables No Insufficient data -
Change working cap Yes** Insufficient data -
Auditor opinion Yes* ATR** & IPO
timing*** Match
Company size No No Match
With the findings of this research it can be tried to answer the research question: To which extent do corporate venture capitalists invest in ventures whose characteristics predict future performance and how is this decision made?
Prior research confirms that CVCs do valuate ventures higher than other VCs do because they pay a premium for strategic benefit (Dushnitsky & Lenox, 2006; Gompers, 1995; Gompers & Lerner, 1998; Röhm et al., 2018). However, this does not imply that CVCs do not aim at performance (Chesbrough, 2002; Dushnitsky & Lenox, 2006; Röhm, 2018). The results of this paper show that CVCs use a set of decision criteria to evaluate ventures that the CVC is considering investing in. However, these decision criteria do not all predict performance. While on the other hand, CVCs do not use certain criteria that do predict performance. This mismatch between use of these criteria and the relationship to performance can be explained by the vast amounts of information the CVC perceives. CVCs face constraints on processing this information resulting in the attention that is spent on evaluating a company or criterion cannot be used to evaluate another company or criterion. This results in attention drawn from criteria that do predict performance to criteria that are concrete, easily summarized and familiar, regardless of they predict performance.
In conclusion, CVCs are using decision criteria that might not always predict performance while neglecting criteria that might do predict performance. Hence, CVCs are not able to rationally evaluate entrepreneurial ventures because CVCs find constraints in processing all available information. The reason for these constraints is that CVCs’ attention is drawn to certain decision criteria, leaving no attention left for other criteria. Consequently, the limited attention of CVCs does influence the decision-making process.
6.2. Theoretical Implications
This paper provides several theoretical implications. First, the results of this paper seem to be in contrast with current believes on CVC decision criteria. For instance, contrary to what many scholars found, CVCs do not use market criteria for their evaluation of ventures. Furthermore, by executing a direct test on limited attention of CVCs, this study finds evidence that decision making of CVCs is influenced by limited attention. The results of this paper seem to indicate that CVCs’ attention is drawn to information that does not always predict performance. This paper provides a deeper understanding for the underlying reasons of the CVC decision process, therefore it is a valuable addition to the CVC decision making literature.
6.3. Managerial Implications
and to which criteria CVCs do not pay attention. This information can help entrepreneurial ventures structure their business proposals and adapt these in such a way that the information that CVCs value most, stands out from the rest of the information. By doing this, ventures attract the CVCs attention. Because CVCs cannot spend their attention twice, this could give the entrepreneurial venture an advantage over other ventures. Second, this research can be a wake-up call for CVCs because it provides insight into their own decision process. Consequently, these insights could help the CVCs to better understand their own decision process. This understanding could guide CVCs to make better decisions based on criteria that do predict performance. As a result, this could lower opportunity costs and enhance performance of the CVC.
6.4. Limitations and Future Research
This research has several limitations that could provide opportunities for future research. First, existing literature shows evidence of a fifth decision criterion; characteristics of the entrepreneur (Macmillan et al., 1985; Petty & Gruber, 2011; Sandberg & Hofer, 1987). Unfortunately, this criterion could not be included in this research because there was no possible way of measuring the characteristics of the entrepreneur with the data available. Future research could develop a framework that is able to quantitatively measure the characteristics of the entrepreneur. It would be interesting to find out if entrepreneurial characteristics do predict performance or that, as some suggest, the entrepreneur can easily be switched for a management team.
Second, the results of this research show that the general agreement on which decision criteria CVCs use has its limitations. For instance, the results show no evidence for the use of market criteria by CVCs. However, finding alternative decision criteria is beyond this research’s scope. Consequently, this raises the need for new research on decision criteria used by CVCs. Third, this research uses a substantial dataset, however per company the data available is representing only one period in time. It could be interesting to include data of multiple time-periods to compare investments into more detail.
6.5. Conclusion
7. References
Barker, V. L., & Mueller, G. C. (2002). CEO Characteristics and Firm R & D Spending. Management Science, 48(6), 782–801.
Basu, S., Phelps, C., & Kotha, S. (2011). Towards understanding who makes corporate venture capital investments and why. Journal of Business Venturing, 26(2), 153–171. Baum, J. A. C., & Silverman, B. S. (2004). Picking winners or building them? Alliance,
intellectual, and human capital as selection criteria in venture financing and
performance of biotechnology startups. Journal of Business Venturing, 19(3), 411–436. Bergemann, D., & Hege, U. (1998). Venture capital financing, moral hazard, and learning.
Journal of Banking & Finance, 22(6–8), 703–735. Bergstrom, C., & West, J. (n.d.). The Eigenfactor Project.
Blevins, D., & Ragozzino, R. (2018). An examination of the effects of venture capitalists on the alliance formation activity of entrepreneurial firms. Strategic Management Journal, 39(7), 2075–2091.
Block, J. H., De Vries, G., Schumann, J. H., & Sandner, P. (2014). Trademarks and venture capital valuation. Journal of Business Venturing, 29(4), 525–542.
Boyaci, T., & Akcay, Y. (2017). Pricing When Customers Have Limited Attention. Management Science, 1–20.
Brigl, M., Denhert, N., Groß-Selbeck, S., Roos, A., Schmieg, F., & Simon, S. (2018). How the Best Corporate Venturers Keep Getting Better.
Brush, C. G., Edelman, L. F., & Manolova, T. S. (2012). Ready for funding? Entrepreneurial ventures and the pursuit of angel financing. Venture Capital, 14(2–3), 111–129. Buchner, A., Mohamed, A., & Schwienbacher, A. (2017). Diversification, risk, and returns in
venture capital. Journal of Business Venturing, 32(5), 519–535.
Champenois, C., Engel, D., & Heneric, O. (2006). What kind of German biotechnology start-ups do venture capital companies and corporate investors prefer for equity
investments? Applied Economics, 38(5), 505–518.
Chemmanur, T. J., Loutskina, E., & Tian, X. (2014). Corporate venture capital, value creation, and innovation. Review of Financial Studies, 27(8), 2434–2473.
Chesbrough, H. W. (2002). Making Sense of Corporate Venture Capital. Harvard Business Review, 80(3), 90–99.
