Crowdfunding, a proxy for firm quality?

(1)

Crowdfunding, a proxy for firm quality?

by

Wiebe Bonekamp

S3273377

University of Groningen

Faculty of Economics and Business

(2)

ABSTRACT

Obtaining financial capital is often a challenge for small businesses. Besides these difficulties, the financial crisis has caused declining bank lending to small businesses. As a consequence, alternative sources of finance emerged. This study focusses on crowdfunding, where firms are financed by a large group of individual investors, known as “the crowd”. The tremendous growth of crowdfunding volume in the recent past reflects the increasing importance of crowdfunding in entrepreneurial finance. However, in economic transactions, information asymmetries between entrepreneurs and investors exist. Drawing on signaling theory, from the firm perspective, the asymmetry of information can be reduced by sending quality signals to potential investors.

This study aims to examine if crowdfunding serves as a quality signaling mechanism to future debt providers. I use leverage after crowdfunding as a proxy for firm quality. In addition to this main research question, I explore three contingencies of the main relationship where the power of the signal might be relatively stronger: 1.) equity and debt crowdfunding, 2.) firm age, and 3.) firm size. To test the formulated hypotheses, I collected data on 367 debt and equity crowdfunded Dutch small businesses with a total fundraising volume of EUR 80.8 million.

My results support the view that crowdfunded firms show a significant higher leverage than non-crowdfunded firms. This indicates that crowdfunding serves as quality signal. However, robustness needs to be tested further by future researchers to increase validity. I also find a significant higher leverage for medium-sized firms compared to small firms. Large firms compared to medium-sized firms on the other hand show no significant difference in leverage. Thus, as firms become reasonably large, the power of the quality signal becomes stronger, but not indefinitely. As for equity crowdfunding as contingency, no significant results were found between equity and debt crowdfunding. The same holds for firm age, where I could not find direct evidence for my hypothesis. I discuss the implications for theory and practice.

(3)

1. INTRODUCTION

Financial capital is a required resource for entrepreneurs to launch and subsequently grow their businesses (Cassar, 2004). Most SMEs rely on internal sources of finance (Berger & Udell, 1998). Next to internal finance, bank financing has traditionally been the most widely used source of debt financing for small and medium-sized enterprises (SMEs), even in early stages of the firm (Kraemer-Eis et al., 2018; Robb, 2002). However, after the start of the global financial and economic crisis, the availability of financial capital for SMEs got considerably worse (Drakos, 2012; Ferrando & Griesshaber, 2011). Consequently, alternative sources of entrepreneurial finance have started to emerge. The most common ones of these are microfinance, peer-to-peer lending and crowdfunding (Bruton et al., 2015; Khavul, 2010; Schwienbacher et al., 2013). Likewise, the volume of venture capital in Europe shows strong growth over the past years (Kraemer-Eis et al., 2018).

This study focusses on crowdfunding as promising finance alternative for SMEs. Crowdfunding is a rapidly expanding alternative way to finance firms by a large group of individual investors, who are connected through platforms (Mollick, 2014). In Europe alone, the total business-related transaction volume on crowdfunding platforms increased strongly since 2012, from less than 100 million euros to almost 1.2 billion euros in 2018 (Kraemer-Eis et al., 2018). This impressive growth has also aroused interests among researchers in the recent past. However, despite the fact that the amount of crowdfunding literature is steadily enlarging, a lot more research is needed in order to contribute the research area, especially among SMEs.

Next to declining bank lending, SMEs have traditionally face difficulty in attracting financial capital. Therefore, it is important for SMEs to discover other sources of finance. However, in entrepreneurial finance, significant information asymmetries exist. Often, entrepreneurs possess more knowledge about the firms than the investors (Ahlers et al., 2015; Akerlof, 1970). Information asymmetries arise “when different people know different things” (Stiglitz, 2002, p. 469). Due to this asymmetric information, investors are usually not fully aware of the true quality of a firm and the behavioural intentions of the entrepreneurs (Shane & Cable, 2002). Also in crowdfunding transactions, information asymmetries exist between the entrepreneurs and the potential backers.

(6)

Thus, it is important for SMEs to show their firm value to investors in order to receive funding. In addition to previous mentioned firm-specific quality signals, the source of finance chosen by entrepreneurs plays a prominent role in the investment decision-making process by investors.

Due to the screening, contracting, and monitoring process of venture capitalists, information asymmetries can be reduced (Berger & Udell, 1998). Following signaling theory, due to the extensive selection process carried out by venture capitalists, the firm quality as perceived by other potential future investors increases (Connelly et al., 2011). Likewise, the presence of an angel investment in a firm increases the chance to obtain a follow-on venture capital investment (Hellmann et al., 2013; Kerr et al., 2014). Subsequently, questions arise if crowdfunding might also serve as a quality signaling mechanism.

This study aims to explore crowdfunding as a new firm-specific quality signal to future debt providers. In existing crowdfunding literature, most studies have focused on the factors that affect fundraising performance on crowdfunding platforms (e.g. Ahlers et al., 2015; Block et al., 2017; Mollick, 2013; Piva & Rossi-Lamastra, 2017; Vismara, 2016), but a detailed understanding of the influence of a crowdfunding campaign to the investment decision-making process of investors in the first place is lacking. An attempt to investigate the motivation of firms to launch a crowdfunding campaign was made by Walthoff-Borm et al. (2018). These authors argue that firms are more or less forced to crowdfunding as a last resort. Their study shows that more unprofitable firms and firms that show excessive debt levels are more likely to raise funding with use of crowdfunding. However, these authors focus on firm characteristics before funding. To the best of my knowledge, the research about the presence of crowdfunding as possible quality signal to future investors is limited to one notable study of Ryu & Kim (2016). In their study, the authors examined 822 projects from Kickstarter, where they investigate the relationship between crowdfunding campaign success and a start-ups follow-on funding performance. The authors found that venture capitalists may perceive a successful crowdfunding campaign as quality signal to reduce risks in their investment decision making process, especially under circumstances of large information asymmetries.

In this study, I aim to extend the paper of Ryu & Kim (2016), where I build on signaling theory as initially developed by Spence (1973). My findings can be considered as a step towards filling the literature gap on the presence of crowdfunding as quality signal and the influence of a crowdfunding campaign to further financing of the company. I expect that leverage serves as a proxy of a signaling effect exerted by crowdfunding. Given the current research gap, this study will address the following main research question:

Does the presence of crowdfunding serve as a proxy for firm quality to future debt providers?

(7)

the effect to leverage. Thus: Does equity crowdfunding send a stronger power signal to future investors than debt crowdfunding? Secondly, financing start-ups is riskier than non-start-ups due to high failure rates. Therefore, I expect that once a start-up got crowdfunding, the quality signal providers will be stronger, which results in a higher leverage. To address this issue, I investigate the relationship between firm age and leverage. So: Is the strength of the power signal of crowdfunding stronger at start-ups than non-start-ups to future investors? Lastly, SMEs often do not possess a lot of resources in comparison to larger firms. Little resources can be considered as a weak quality signal. So: Do larger firms send a stronger power signal to future investors than smaller firms?

Next to the main research gap, this study contributes in two more ways to crowdfunding literature. Firstly, existing studies are mainly focused on one type of crowdfunding, while my data set covers both business-related types of crowdfunding, namely equity and debt funding. Secondly, to prove my hypothesis empirically, I examined 367 Dutch crowdfunding projects. This data is provided by Crowdfundmarkt.nl, who collects the data of all major crowdfunding platforms in the Netherlands. Therefore, this data set offers an accurate reflection of the Dutch crowdfunding market, in contrast to existing studies, who are mainly context specific.

To continue on the managerial contribution of this study, the results are in the first place relevant for SME owners, because it gives insights to the influence of a crowdfunding campaign to the decision-making process of potential future investors, which is currently unknown. Attracting sufficient financial capital is essential to the future development of an SME. Secondly, the results of this study serve as a guideline for investors who are planning to grant loans to SMEs. Because of the increasing share of crowdfunding in entrepreneurial finance, it is of value to investors how to qualify a crowdfunding campaign in their investment decision-making process to reduce information asymmetries.

(8)

2. LITERATURE REVIEW

2.1 Sources of entrepreneurial finance Entrepreneurial finance

Finance is an important and inevitable function in any business. “Businesses typically need finance to pay for assets, to fund the operation of the business and to grow the business” (Storey & Green, 2010, p. 312). However, access to finance is often a challenge for SMEs. According to corporate finance literature, most common reasons for these problems of business finance include: high information asymmetries, agency risks, basic financial statements, short track records as well as lack of securities (Beck & Demirgüç-Kunt, 2008; Berger & Udell, 1998; Jõeveer, 2012). In particular start-ups, where business growth and chances of survival are impeded by insufficiency of resources (Block et al., 2018).

Moreover, SMEs heavily depend on bank loans (Kraemer-Eis et al., 2015). As a result, SMEs suffered a lot from stagnant lending by banks as a result of the financial crisis (Artola & Genre, 2011). Consequently, SMEs became interested in alternative sources of finance. There are multiple sources of alternative finance available for SMEs in start-up, growth and mature stages. The most common ones of these are microfinance, peer-to-peer lending and crowdfunding (Bruton et al., 2015; Khavul, 2010; Schwienbacher et al., 2013). Also, business angels and venture capitalists are increasingly providing funding to SMEs (Storey & Green, 2010). Business angels are wealthy people who use their personals savings to fund early-stage businesses and, besides money, provide the entrepreneurs also with knowledge and expertise (Hsu et al., 2013; Ibrahim, 2008), whereas venture capital is often a fund where fund managers provide funding to new firms whose focus is on short-term returns (Hsu et al., 2013). This study focusses on crowdfunding as a relatively new source of entrepreneurial finance, which will be introduced in the next section.

2.2 Crowdfunding

Introduction of crowdfunding

(9)

One of the main drivers of crowdfunding is the emergence of the internet, which can be considered as the backbone of the global crowdfunding market (Brabham, 2008; Kleemann et al., 2008).

Since the 2008 global financial crisis, crowdfunding has accelerated as an alternative finance source. As a result of this economic headwind, banks reduced their credit engagement (Drakos, 2012; Ferrando & Griesshaber, 2011), where SMEs suffered most under the stricter lending policy of banks. Consequently, crowdfunding has grown into a small but significant alternative source of funding for SMEs (Kraemer-Eis et al., 2018).

The high growth rates of crowdfunding over the past few years has also captured increasing interest by scholars. A widely used definition of crowdfunding in literature is composed by Belleflamme et al. (2014):

“Crowdfunding involves an open call, mostly through the Internet, for the provision of financial resources either in the form of donation or in exchange for the future product or some form of reward to support initiatives for specific purposes” (Belleflamme et al., 2014, p.4).

Types of crowdfunding

While the umbrella term ‘crowdfunding’ is used often to describe the process of fundraising from a large number of individuals, crowdfunding is typically categorized into four basic models, which vary in the goals of both the founders and supporters (Mollick, 2014). However, the idea of obtaining capital from a crowd of individuals in exchange for intangible or tangible returns commonly exists within these models (Frydrych et al., 2014). The four crowdfunding models that have been observed in literature are elaborated below:

(1) Donation-based crowdfunding: donation-based crowdfunding is known as the least complex of all four observed models. As the term ‘donation’ suggests, the funders do usually not receive any gain from their investment, but are motivated by charitable giving and creating social image (Block, et al., 2018, Paschen, 2016). Hence, donation-based funding is relatively well-aligned with social-entrepreneurship (Lehner, 2013).

(2) Reward-based crowdfunding: in reward-based crowdfunding, investors typically receive a non-financial benefit in exchange for their investment (Ahlers et al., 2015). Most commonly, investors receive a product or service as compensation. A common example is the so-called pre-selling, where investors receive a free sample of the final product (Mollick, 2014).

(10)

(4) Lending-based crowdfunding: is also known as debt financing. In lending-based crowdfunding, funding can be considered to be a loan. Investors receive fixed periodic income and the full loan is repaid at the end of the term (Ahlers et al., 2015, Paschen, 2016). Therefore, lending-based funding is more or less similar to debt provided by banks (Mollick, 2013). Lending-based crowdfunding is also referred to as debt crowdfunding.

Taken these types into account, donation and reward crowdfunding are not really applicable in a business context. Therefore, I ignore donation and reward-based crowdfunding and in this study I focus on lending and equity crowdfunding.

The crowdfunding process

Crowdfunding is either possible in one of two ways: either as a direct transaction, where a firm obtains funding directly, for instance by sending emails or using social media to approach investors (Lambert & Schwienbacher, 2010). The other way in which firms can raise money with crowdfunding is by an indirect transaction, where an intermediary, commonly a specialized platform, arranges the match between the capital-seeking firms and the possible investors. The large number of crowdfunding platforms (more than 50 in the Netherlands) indicates this last method as the most preferred. The key players in an indirect crowdfunding transaction include the capital seekers (firms), the capital providers (the individuals who invest and form the crowd) and the intermediaries (the crowdfunding platforms) (Ahlers et al., 2015, Beugré & Das, 2013).

The crowdfunding platform function as intermediaries who connect the entrepreneur with the investor (Burkett, 2011). According to (Ahlers et al., 2015, p.1): “The open call and the investments take place on an online platform provides the means for the transactions (the legal groundwork, pre-selection, the ability to process financial transactions, etc.”. Moreover, financial intermediaries like crowdfunding platforms are parties who fulfil process of contracting, monitoring and screening (Berger & Udell, 1998).

2.3 Signaling in entrepreneurial finance Information asymmetries

(11)

information and a high degree of uncertainty causes difficulties to distinguish high and low-quality firms. Information asymmetries play a key role in the process of crowdfunding, where the sellers (the entrepreneurs) often possess more knowledge about the quality of the crowdfunding projects than the buyers (the individual investors) (Ahlers et al., 2015; Akerlof, 1970).

Signaling theory

As explained in previous section, the presence of information asymmetries has major effects on the investor-entrepreneur relationship. An important way to overcome information asymmetries is signaling (Connelly et al., 2011). Signaling theory holds a leading position in research in the context of entrepreneur-investor relations (e.g. Arthurs et al., 2009; Cassar et al., 2015; Connelly et al., 2011). Information asymmetries form the fundamental assumption of signaling theory. Signaling theory originates from the research of Spence (1973), where the concept of signals is presented as visible attributes to reduce information asymmetries by transmit information from one actor to another. It is important to note that signaling theory intentionally focuses on the communication of positive information by insiders aiming to distinguish themselves from other actors with unobservable quality (Connelly et al., 2011).

In their study, Connelly et al. (2011) present an overview of signaling theory’s primary elements: the signaler, signal, receiver, and feedback. Firstly, the signaler refers to insiders (e.g. managers of a firm), who possess information about the quality of a subject (Spence, 1973) and convey signals to the receivers. Next, the signal is a message about information communicated to the receivers (Connelly et al., 2011). Thirdly, the receiver receives the signal as communicated by the signaler. Lastly, feedback is provided by the receiver to the signaler (Bergh et al., 2014; Connelly et al., 2011).

Applying signaling theory to the crowdfunding context, we can easily draw parallels. Firstly, the entrepreneurs act as signalers: they own critical information regarding the firms that are about to be crowdfunded as well and they send signals to the investors. Secondly, the project information on the crowdfunding platforms are considered to be the signals. Finally, the individual investors (the crowd), form the receivers. They receive the signals as transmitted by the entrepreneurs.

Quality signals

(12)

As mentioned in this definition: signals differ in terms of observability, and can therefore be distinguished as “strong” or “weak” (Gulati et al., 2003). According to signaling literature, a wide range of signals exist. As already mentioned in the introduction, examples of quality signals that weight heavily in the investment decision-making process of venture capitalists are the presence of alliances, patents and human capital (Baum & Silverman, 2004; Gompers et al., 2010; Hoenig & Henkel, 2015; Hsu & Ziedonis, 2013).

In the context of crowdfunding, several factors are indicated that affect fundraising performance on crowdfunding platforms. Included videos in the project description on platforms, frequent updates and disclosure of financials are known as signals that influence project success positively (Lukkarinen et al., 2016; Mollick, 2014). On the other hand, spelling mistakes made in the project description reduce the change of project success (Mollick, 2014).

2.4 Crowdfunding as quality signal

Putting all the arguments together, it is essential for entrepreneurs to signal firm quality to potential investors in order to receive funding. Based on previous story, the question arises whether a crowdfunding campaign has the ability to reduce information asymmetries between entrepreneurs and potential future debt providers. Drawing on signaling theory, I’m interested if the presence of crowdfunding serves as quality signal to future debt providers following after the crowdfunding campaign. I use financial leverage after funding as a proxy of the willingness of future investors. I expect that leverage serves as a proxy of a signaling effect exerted by crowdfunding. Thus:

H1 Crowdfunded SMEs show a higher leverage than non-crowdfunded firms

2.5 Contingencies of the main relationship

In addition to the baseline hypothesis, I highlight some contingencies of the main relationship where the power of the quality signal may be relatively stronger. These contingencies include: equity and debt crowdfunding, firm age and firm size.

2.5.1 Equity and debt

Equity and debt crowdfunding are both related to business finance. As explained earlier, debt crowdfunding can be considered as a loan where investors receive fixed periodic income (Ahlers et al., 2015, Paschen, 2016) and is more or less similar to bank debt (Mollick, 2013). In equity crowdfunding, investors receive an ownership in the company (Ahlers et al., 2015; Mollick, 2014).

(13)

2018). In debt crowdfunding, information asymmetry is caused by the chance whether the entrepreneur is able to supply the product. Meanwhile, in equity crowdfunding, the asymmetry is strengthened by generating equity value, instead of only the ability to supply the product (Agrawal et al., 2014).

Secondly, in equity crowdfunding, entrepreneurs sell a stake in their company. This means that besides ownership, also entrepreneurial risk is transferred to investors (Berger and Udell, 1998).

Based on previous arguments, I expect that once a firm received equity crowdfunding, the signal of overcoming the problem of information asymmetries and the risk transfer yields as a stronger signal to potential debt providers than debt crowdfunding.

H2 Equity CF firms show a higher leverage than debt CF firms

2.5.2 Firm age

Early stage firms, like young start-up companies, are less attractive by debt providers due to high information asymmetry, lack of collateral and negative cash flow (Cosh et al., 2009). Most often, crowdfunding is associated with start-ups. Crowdfunding plays an important role in the early stages in a start-up’s origin and growth (Paschen, 2016). However, start-ups face difficulty in obtaining external finance (Berger & Udell, 1998). As a result, one of the main difficulties for start-ups is their lack of capital, especially risky and innovative ones (Carpenter & Petersen, 2002). Therefore, it is important for start-ups to signal quality to investors.

Research shows that a start-up backed by angel funding is an important quality signal (Stuart et al., 1999). Likewise, start-ups that raised a large amount of money by crowdfunding are more likely to attract venture capitalists then start-ups obtaining funding from angel investors (Ryu & Kim, 2016). To build on this, I’m interested if future debt providers view young firms backed by crowdfunding as a stronger quality single than older crowdfunded firms.

H3 Younger firms show a higher leverage than older CF firms

2.5.3 Firm size

The financial sources of small businesses are really different compared to the funding of large businesses (Storey & Green, 2010). Large firms have access to public markets, which SMEs generally not have (Berger & Udell, 1998). Instead, small firms are traditionally dependent on internal finance and bank loan (Berger & Udell, 1998; Kraemer-Eis et al., 2018). As already mentioned, early-stage firms are less attractive to investors (Cosh et al., 2009). Given these arguments, I assume that future investors prefer to finance larger firms over smaller firms. Therefore:

(14)

(15)

3. RESEARCH DESIGN

3.1 Data description

The data required to test the hypotheses is gathered from different sources. The initial data set is provided by Crowdfundmarkt.nl and consists of information about Dutch crowdfunding campaigns characteristics in the years 2015 and 2016. Crowdfundmarkt.nl (founded in 2015) offers investors an overview of all Dutch crowdfunding projects that have been approved by The Netherlands Authority for the Financial Markets. Next, the crowdfunded firms are merged with additional information on firm financials from Bureau van Dijk’s Orbis database. Orbis is a company database comprising detailed information about 280 million companies worldwide, including 5.4 million Dutch companies.

Disclosure requirements in the Netherlands require privately held firms to publish annual financial statements. Small businesses must send an abbreviated balance sheet with additional limited explanatory notes to the Dutch Chamber of Commerce each year. The Dutch government defines small businesses as firms that have any two of the following: a turnover of 12 million EUR or less, 6 million EUR or less on its balance sheet, or 50 employees or fewer (Kamer van Koophandel, 2019). Despite the limited information available, the abbreviated financial statements offer information about debt and equity, which is sufficient to determine the capital structure.

3.1.1 Non-crowdfunded firm control group

To test if crowdfunded firms show a higher leverage than non-crowdfunded firms, a comparable non-crowdfunded control group must be created. As a first step to create a control group of similar charactertics, I selected firms that fall within the same age and size range as the crowdfunded firms. Using the ORBIS database, I’ve downloaded the financials of Dutch private firms that meet the following criteria: founded in and after 1947 and a total asset value between EUR 0 and 10 million in the year 2017. After removing missing values, 507,871 firms meet these criteria.

To visualize correlation, a correlation matrix between variables of both the crowdfunded sample and a random sample of equal size to the crowdfunded sample has been made (Appendix I). When looking at the correlation matrix, it is observable that some of the sectors are strongly correlated to whether a firm is crowdfunded or not. Looking at for instance the hotels and restaurants sector, where crowdfunded firms are more likely to be than for instance the banks sector. This unequal distribution of sectors among the crowdfunded firms and non-crowdfunded firms can cause multicollinearity.

(16)

omitted to avoid the so called ‘dummy trap’. The best I can do in cases of non-perfect multicollinearity is to reduce the effects.

In this study, I reduce multicollinearity between firm sector and whether or not a firm is crowdfunded by generating a non-random control group in terms of sectors, where each sector has the same relative share of observations in both the crowdfunded sample and the non-crowdfunded control group. To reduce the effects of multicollinearity even further, I also make sure that the control group is equally distributed in terms of firm size as measured by 5-year average assets.

To create this non-random control group, a matching algorithm has been developed using the statistical computing environment R. The algorithm works as follows: I use the quantile classification method to compute the asset quantiles of the crowdfunded sample (25th, 50th, and 75th percentiles). By using the quantile distances of the asset classes, the desired amount of non-crowdfunding firms can be determined for each sector regarding firm size.

Next to this main control group, four more control groups have been created aimed to check whether the results from the main hypothesis are robust to changes between control groups. In addition to the equal-sized non-random control group, a second non-random control group that is ten times the size of the crowdfunded sample has been created. Furthermore, two randomly selected control groups are created (1:1 and 1:10 to CF sample). As a final robustness check, all non-CF firms will be selected as control group in the analysis. The following table summarizes the five different control groups.

Table 1: Control groups

Control group Sample size Random or non-random1 Description

1 367 Non-random Main control group

2 3,670 Non-random 1:10 CF firm sample

3 367 Random 1:1 CF firm sample

4 3,670 Random 1:10 CF firm sample

5 507,871 Random All non-CF firms

In Appendix II appears a graphical overview of the distribution in terms of sector and firm size of the control groups who are equally-sized to the crowdfunding sample. The first thing that strikes the eye when comparing the sector distribution of the crowdfunded firms with the sector distribution of the random control group is that the distribution is really different. The sector textiles, wearing apparel and leather that does appear in the crowdfunding sample is not even present at all in the random control group. Contrary to the random control group, the non-random control group shows a similar sector distribution to the crowdfunding sample. To continue, all three graphs of asset distribution show a right

(17)

skewed histogram. However, when analysing the graphs in more detail, it is clearly visible that the distribution of the random control group is truly different from the crowdfunded firms. The histograms of the non-random control group and crowdfunded firms show roughly the same distribution. This indicates that the selection process by the algorithm worked well.

3.1.2 Dependent variable

For my research purpose, leverage is my key interest and will be used as dependent variable. According to corporate finance literature, leverage refers to the amount of debt a firm uses to finance its assets, where debt is the sum of both short- and long term liabilities (Clayman et al., 2012). Short term debt refers to any type of firm loan that is expected to be repaid within one year. Long term debt includes any amount of outstanding debt lasting over one year. In the analysis, I adopt long term debt (long-term debt-to-total-assets ratio) as measurement for leverage. There are two reasons for this choice. Firstly, the average duration loan in the data set as provided by Crowdfundmarkt is 53 months or 4.4 years. This average loan term is much larger than the loan term of one year as suggested by Clayman et al. (2012). Secondly, for multiple reasons, SMEs face difficulties to attract long term debt (Berger & Udell, 1998). Following the thoughts of Connelly et al. (2011), once SMEs managed to attract long term debt, receiving long term debt might serve as a stronger signal to future debt providers. Thus, for the purposes of this study, it is useful to use long term debt to measure leverage.

3.1.3 Independent variables Equity vs debt crowdfunding

To test my second hypothesis, the binary variable equity or debt crowdfunding will be used to distinguish between the two types. In addition, the continuous variable equity share will be used in a second regression analysis.

Firm age

To be able to perform a t-test to test hypothesis 3, firm age is categorized in different groups. The first group contains of firms founded between 2010 and 2017, and I call this group “start-ups”. Next, firms founded between 2000 and 2009 form the second group, who are called “mature firms”. The remaining firms who are founded before 2000 form the last group, which I label as “old firms”.

Firm size

(18)

3.1.4 Control variables

There are multiple control variables used in the analysis. The first one is leverage before crowdfunding. It is important to control for past leverage, because it is expected as an important explanatory variable for future leverage. The second control variable is industry. Despite the fact that I ensured that each sector has the same relative share of firms in both the crowdfunded sample and the main control group, I control for this variable because firm characteristics may still differ. The firms in the data set are categorized into the following industries (Orbis format): 1.) Banks, 2.) Chemicals, rubber, plastics and non-metallic products, 3.) Construction, 4.) Education and health, 5.) Food, beverages and tobacco, 6.) Gas, Water and Electricity, 7.) Hotels and restaurants, 8.) Machinery, equipment, furniture and recycling, 9.) Metals and metal products, 10.) Other services, 11.) Post and telecommunications, 12.) Primary sector, 13.) Publishing and printing, 14.) Textiles, wearing apparel and leather, 15.) Transport, 16.) Wholesale and retail trade, 17.) Wood, cork and paper. The last control variables are firm size (as measured by 5-year average assets) and funding year.

3.1.5 Data cleaning

(19)

Figure 1: Leverage after crowdfunding before and after data trimming

The original data shows a huge right trial (Figure 1). After the trimming procedure, the histogram shows a much more concentrated distribution of the data. Although, the data is still asymmetric due to a large amount of zero values in the data set. The zero values are caused by firms who are not leveraged at all in the year after funding.

3.2 Analysis

The statistical analysis in this study will be performed using R. The main variables that are used in the R script are summarized in appendix III. The R script is available on request.

Hypothesis 1 and 2

Firstly, I’m going to build a multiple regression model to research whether crowdfunded firms show a higher leverage than non-crowdfunded firms (hypothesis 1). The regression equation I use is the following:

𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔/0 = 𝛽3+ 𝛽5𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐵𝑒𝑓𝑜𝑟𝑒𝐶𝑟𝑜𝑤𝑑𝑓𝑢𝑛𝑑𝑖𝑛𝑔/0+ 𝛽:𝑖𝑠𝐶𝑟𝑜𝑤𝑑𝑓𝑢𝑛𝑑𝑒𝑑/0+

(20)

where the dependent variable 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 is the leverage after crowdfunding of firm i which was funded in year t. For crowdfunded firms, 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 is defined as the leverage in the year following the funding year. For example, if a firm has received crowdfunded in 2016, the value 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔/0 is equal to the leverage in 2017. To compare the leverage

after funding to the non-crowdfunded firms, the variable 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔/0 must be defined

using some sort of proxy. Because all crowdfunded firms are funded in 2016 or 2017, the chosen value of 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 for a non-crowdfunded firm is the average of the leverage in 2016 and 2017. To control for past leverage, the variable 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐵𝑒𝑓𝑜𝑟𝑒𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 is defined. This variable measures the leverage in the year of the funding. Similar to the 𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 variable, for non-crowdfunded firms, this variable is set to the average of the leverage in the years 2015 and 2016.

Next, to determine if equity-crowdfunded firms show a higher leverage than debt-crowdfunded firms, two different regression models will be built. First, I extend the regression equation of model 1, where I use a binary variable to distinguish between equity and debt crowdfunding. Here, 𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑌𝑒𝑎𝑟_/0 is added as control variable.

𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑓𝑡𝑒𝑟𝐹𝑢𝑛𝑑𝑖𝑛𝑔_/0 = 𝛽₃+ 𝛽₅𝑙𝑒𝑣𝑒𝑟𝑎𝑔𝑒𝐵𝑒𝑓𝑜𝑟𝑒𝐶𝑟𝑜𝑤𝑑𝑓𝑢𝑛𝑑𝑖𝑛𝑔_/0+

𝛽_:𝑖𝑠𝐸𝑞𝑢𝑖𝑡𝑦𝐶𝑟𝑜𝑤𝑑𝑓𝑢𝑛𝑑𝑒𝑑_/0+ 𝛽_<𝑎𝑣𝑒𝑟𝑎𝑔𝑒𝐴𝑠𝑠𝑒𝑡𝑠_/0+ 𝛽₌𝐹𝑢𝑛𝑑𝑖𝑛𝑔𝑌𝑒𝑎𝑟_/0+ 𝛽_F𝑆𝑒𝑐𝑡𝑜𝑟_/0+ 𝜀_/0

Additionally, instead of a binary variable to distinguish between debt and equity crowdfunding, a linear function of equity share will be used to improve hypothesis validity. Therefore, in above equation, the variable 𝑖𝑠𝐸𝑞𝑢𝑖𝑡𝑦𝐶𝑟𝑜𝑤𝑑𝑓𝑢𝑛𝑑𝑒𝑑_/0 will be replaced by the continuous variable 𝑆ℎ𝑎𝑟𝑒_/0.

Hypothesis 3 and 4

To research if there is a significant difference between firm age and leverage, t-test will be used. To be able to compare means, the data will be subdivided into three different groups. Similar to the approach of hypothesis 3, to test if there is a positive relationship between firm size and leverage (hypothesis 4), t-tests are used.

Additional analysis

(21)

4. RESULTS

4.1 Descriptive statistics

The resulting sample, so after data cleaning, consist of 367 crowdfunded firms. The descriptive statistics of both the crowdfunded sample and non-random control group appear in appendix IV (after data trimming). Next, a bird’s-eye view of some key crowdfunding statistics from the data set (before data trimming):

• The average funding amount raised per crowdfunding project was EUR 220,160 • The average age of the crowdfunded firms is 10.4 years

• On average, crowdfunded firms have an average leverage of 0.40 (median 0.27) in the year of crowdfunding, which increases to 0.39 (median 0.18) in the year after crowdfunding.

• Most of the crowdfunded firms are operating (after other services) in the wholesale and retail trade sector (21%)

4.2 Crowdfunding as a proxy for future debt

To test hypothesis 1, I run a regression of leverageAfterCrowdfunding on leverageBeforeCrowdfunding, isCrowdfunded (dummy variable), averageAssets (in EUR thousands) and a set of dummy variables indicating the sector of the firms. To avoid the earlier explained multicollinearity problem, I set the sector ‘banks’ as reference category and omit it as an explanatory variable. This implies that the intercept and sector regression coefficients correspond to the sector ‘Banks’. The output of the regression analysis is summarized in next figure.

(22)

To interpret the regression output in figure 2, I see that a hypothetical non-crowdfunded firm in the banking sector with an average number of assets of zero and an average leverage equal to zero, has a leverage of 0.05305. Then the following holds: each EUR 1 million of additional leverage total assets increases the leverage by 0.02708, ceteris paribus2. The effect of increasing average total assets is not that strong, but it is slightly statistically significant at the 90% level (p-value is 0.07). Moreover, I observe that leverage before funding indeed plays an important role in predicting the leverage after funding: the regression coefficient of 0.4855 is highly statistically significant (p-value is < 0.01), which shows the importance to control for this variable.

The coefficients of the sectors indicate that nearly every sector does not deviate significantly from the banking sector. To summarize, after controlling for past leverage, firm sector and average assets, I find that crowdfunded firms have significantly higher leverage than comparable non-crowdfunded firms (p-value is < 0.01). To be precise, ceteris paribus, non-crowdfunded firms have a leverage that is 0.03823 higher than non-crowdfunded firms. This result verifies hypothesis 1. The R-squared of the model (0.40) is not that high. The adjusted R-squared, which penalizes for having many variables, is not much lower at 0.42 in the model. Therefore, my model is in line with Occam’s Razor principle, which means that simpler models should be preferred over sophisticated and complex ones (Rasmussen & Ghahramani, 2001).

To test the reliability of the regression, three robustness checks are conducted. As a first robustness check, I run the regression analysis again with four other control groups (see Table 1 in Section 3.1.1). As a next robustness check, I do not trim the variables at all. Lastly, to verify robustness, I focus on the firms that have no leverage before funding at all. As already discovered in section 3.1.5, there are quite a few firms in the data set that are not leveraged at all before funding (also indicated by the large difference between mean and median in the summary statistics). The scatterplot in Appendix V shows an overview of the leverage after crowdfunding. To research the effect of this zero leverage firms to our study, I run a separate regression with these zero leverage firms. The output of the robustness tests appears in Appendix VI.

The results of the robustness check show that the results of my previous model are not robust at all. Firstly, when I test the crowdfunded sample against the four other control groups, the crowdfunded sample does no show a significant higher leverage any more. Except for the regression of the non-trimmed control group 2, I see a significance results for the variable leverage before crowdfunding. This shows that leverage before crowdfunding is the most important predictor value of the model. Moreover, it is visible that data trimming has a major effect to the output of the regression model. When I do not trim the data, the p-values in all models increase. Finally, only the results of the last regression with the

2_{Ceteris paribus means ‘other things equal’, so in my model: holding all variables in the model constant and}

(23)

zero leverage firms are in line with my previous results and show a significant higher leverage for crowdfunded firms than non-crowdfunded firms (after trimming).

4.3 Contingencies of the main relationship

4.3.1 Equity and debt crowdfunding

Two different regression models have been built to test if equity crowdfunded firms show a higher leverage than debt crowdfunded firms. In the first model, I run a regression of leverageAfterCrowdfunding on leverageBeforeCrowdfunding, isEquityCrowdfunded (dummy variable), amountCrowdfunded, averageAssets, Funding.year and a set of dummy variables indicating the sector of the firms. The output of this regression analysis is summarized below.

Figure 4: Output regression analysis R hypothesis 2 (binary variable for equity and debt crowdfunded firms)

(24)

Figure 5: Output regression analysis R hypothesis 2 (equity share)

The regression output shows that equity share is not significant to leverage (p-value is 0.51). Following the results of both testing methods, I cannot conclude that a significant difference in leverage exists between the two types of crowdfunding. Hence, I cannot adopt hypothesis 2.

4.3.2 Firm age

To test if younger crowdfunded firms show a higher leverage than older firms, t-tests are used. As null hypothesis I assume that no statistical differences exist in leverage between the means of the different groups, which is exactly the opposite of hypothesis 3. The alternative hypothesis is contrary to the null hypothesis, in my case in line with hypothesis 3. So, if the p-value of the t-tests exceeds the significance level of 0.05, I can reject the null hypothesis and adopt the alternative hypothesis, which is similar to hypothesis 3. To remain consistent with hypothesis 1 and 2, I use the variable leverage after crowdfunding as measurement for leverage. The output of the t-tests is included in Appendix VIII.

(25)

4.3.3 Firm size

Similar to hypothesis 3, t-tests are used to test if there is a positive relationship between firm size and leverage. Actually, this hypothesis is already confirmed by looking at the first regression model where assets are positive and statistically significant, but I zoom in a bit further to get a more comprehensive picture of the phenomenon. As null hypothesis I assume that no statistical differences exist in average leverage between the different asset categories. Again, the null hypothesis is the opposite of hypothesis 4. So, in case the t-test is statistically significant, I can reject the null hypothesis and adopt the alternative hypothesis (in line with hypothesis 4). The output of the t-tests is included in Appendix IX.

I first run a t-test of the mean leverage on medium-sized firms over small firms. The p-value is smaller than 0.01, therefore I can reject the null hypothesis. This indicates strong evidence in favour of hypothesis 4. The difference in means is 0.14 (0.24 – 0.10). Next, a t-test is conducted on large firms over small firms. The p-value is again smaller than 0.01, which allows me again to reject the null hypothesis. The difference in means is 0.16 (0.26 – 0.10). Lastly, a final t-test between the mean leverage of large firms on medium-sized firms shows a p-value of 0.34. This high p-value indicates that no difference exists in average leverage between large firms and medium-sized firms.

In conclusion, the conducted t-tests show strong evidence that firm size is positively correlated to leverage, which was already visible from the regression of hypothesis 1. However, the effect is only statistically different between small firms and medium-sized firms. No more significant difference exists between medium-sized and large firms. Thus, as firms become reasonably large, the leverage increases, but not indefinitely.

4.4 Platform return and leverage

So far, the statistical testing in this study offered interesting results and gave insights in the role of crowdfunding as firm quality indicator. However, still some interesting data remained unused in the analysis. To explore this unused data, I provide a statistical analysis as addition to this study.

(26)

Table 2: Net return crowdfunding platforms

Platform Projects Net return (%) Rounded net return Fundingcircle 8 6.22 6

Geldvoorelkaar 89 4.94 5 Kapitaalopmaat 52 4.31 4 Collincrowdfund 109 4.09 4

Platform return is calculated by subtracting the interest paid from platform fees and any write-offs for loans from companies that are unable to repay the borrowed money to the investors. Based on this, I’m interested whether firms that are crowdfunded by a platform with a higher return show a higher leverage compared to firms that are crowdfunded by platforms with a lower return. Yet, a remark must be made in the way platform return is calculated. The return is based on historical data and crowdfunding projects whose loan term has not yet been exceeded. It is not known whether a crowdfunded firm will not be able to meet the financial obligations of the loan (for any reason). This indicates that the current returns have little predictive power since it is subject to change. Despite this limitation, I want to explore if a statistically significant relationship exists between platform return and leverage.

To research whether a higher platform return leads to a higher leverage, I have conducted one-by-one t-tests and an overall F-test. My methodological approach of testing is as follows: I compare the means of the leverage after crowdfunding per platform pair. When the return of platform 1 is greater than platform 2, I use a one-tailed test to test if the average leverage of platform 1 differs significantly from platform 2. In case the return of platforms is equal (see Collin Crowdfund and Kapitaalopmaat), I use a two-tailed t-test. Finally, an ANOVA test has been performed. The output of the statistical testing is summarized in Appendix X. I do not trim any data in this analysis.

Fundingcircle

Fundingcircle has the highest return (6%) of all available platforms. Therefore, I test the alternative hypothesis if the average leverage of Fundingcircle is on average higher than Kapitaalopmaat, Collincrowdfund and Geldvoorelkaar. The null hypothesis claims that no differences exist in leverage between the platforms. According to the output of the t-tests, the high p-values of 0.95, 0.96 and 0.85 indicate that we cannot reject the null hypothesis.

Geldvoorelkaar

(27)

Collincrowdfund and Kapitaalopmaat

Lastly, the return for both the platforms Kapitaalopmaat and Collincrowdfund is equal (4%). Hence, I cannot make a prediction of the direction of testing. To deal with this, I use a two-sided t-test instead of a one-sided t-test. This t-test shows a p-value of 0.87. This again indicates that there is no evidence that the means of these platforms differ from each other.

ANOVA

Finally, a one-way analysis of variance (ANOVA) has been performed. An ANOVA is the generalization of a two sample t-test and allows me to compare the mean of multiple groups. I use an ANOVA test because I am interested if there are any differences between the four platforms simultaneously as a final test. As null hypothesis I assume that the average leverage of the different platforms is equal. Contrary, the alternative hypothesis claims that at least one platform leverage is not equal to others. Given the results, as already expected, no significant differences exist between the different platforms.

Results

(28)

5.

DISCUSSION

The amount of crowdfunding literature is steadily enlarging, however, a lot more research is needed to contribute to the crowdfunding research area. This study has provided insights into quality signals. I followed signaling theory, as originated by Spence (1973), as a theoretical approach to overcome information asymmetries. Due to lack of financial capital and declining bank lending, SMEs already have shifted the focus to alternative sources of finance. In order to receive funding, sending quality signals to investors is fundamental for SMEs. This study has provided new insights into the quality signaling mechanism between entrepreneurs and investors. In this study, the main research question was: “Does the presence of crowdfunding serve as a proxy for firm quality to future debt providers?”. I can answer this question with a “yes”. After performing the analysis, I found evidence in favour of the main research question.

Hypothesis 1 suggested that crowdfunded firms show a higher leverage after funding than non-crowdfunded firms, as a proxy for firm quality as perceived by future debt providers. My results show that crowdfunded firms indeed show a significant higher leverage than non-crowdfunded firms. This result is in line with the study of Ryu & Kim (2016), where the authors found that venture capitalists may perceive a crowdfunding campaign as quality signal. Due to my findings, I build on this paper in the sense that the presence of crowdfunding can be seen as a signal of firm quality. Also, crowdfunding can be added to the main quality signals as used by venture capitalist in their investment decision-making process, besides the presence of alliances, patents and human capital (Baum & Silverman, 2004; Gompers et al., 2010; Hoenig & Henkel, 2015; Hsu & Ziedonis, 2013).

However, the robustness checks show conflicting results to the main analysis. When I change the control groups, the results of the statistical testing show not enough evidence any more to support the main hypothesis. Also, when I run an additional regression without trimmed data, the results are not in line any more to hypothesis 1 due to insignificant results. Therefore, it is important that future researchers dive into this topic to gain a deeper understanding of the main relationship.

When looking at the regression model, some remarks need to be made. Firstly, an explained portion of the variance of 40% is quite low, but I observed that the model is at least better than a model with only one intercept term. This indicates that perhaps I miss some important variables in the model. Secondly, by looking at the R2, it shows the importance of controlling for past leverage. Removing this variable reduces the R2 from 40% to lower than 10%. This indicates that future researchers must focus on potentially improving the R2 by including more refined estimates of past leverage, for example long term debt, or other observed control variables.

(29)

and the non-crowdfunded sample. I also made sure that the control group is equally distributed in terms of firm size as measured by 5-year average assets. However, to reduce the possible effects of multicollinearity even further, future researchers can consider to expand the selection algorithm with other firm-specific factors.

To continue to the contingencies of the main relationship, I found different results. Firstly, based on my results, equity crowdfunding cannot be considered as a stronger quality signal towards future investors than debt crowdfunding. My expectation that overcoming a larger degree of information asymmetries might serve as a stronger quality signal, as stated by Agrawal et al. (2014) and Vismara (2018), cannot be proven in this study. In other words, hypothesis 2 cannot be proved.

Next, I was not able to find evidence for hypothesis 3, were was stated that younger firms show a higher leverage than older crowdfunded firms. All three performed t-tests where I tested the average leverage of start-ups, mature and old firms against each other show high p-values, so really no evidence can be presented in favour of these hypotheses. To summarize, no relationship exists between firm age and leverage, so also no stronger quality signal to future investors.

(30)

6. CONCLUSIONS

In this study, I aimed to examine if the presence of crowdfunding serves as quality signal to future debt providers. In addition to this idea, I investigated three contingencies of the main idea where the power of the quality signal might be relatively stronger: debt and equity crowdfunding, firm age and firm size. I used leverage after crowdfunding as a proxy for firm quality. Drawing on signaling theory, it was found that crowdfunded firms indeed show a higher leverage than non-crowdfunded firms. However, robustness needs to be tested by future researchers to increase validity of the analysis.

Moreover, the outcome of the tests of the contingencies differ. It was found that larger crowdfunded firms are able to send a stronger signal to future investors than smaller firms. Furthermore, I could provide proof that small crowdfunded firms show a higher leverage than medium-sized firms. However, the results provide no evidence that medium-sized firms show a higher leverage than large firms. Therefore, I can say: as firms become reasonably large, the power of the quality signal increases, but not indefinitely. Lastly, I could not prove that younger crowdfunded firms show a higher leverage after crowdfunding than larger firms.

6.1 Limitations

This study is subject to certain limitations which I will summarize in this section. Firstly, the data availability on firm-level was limited. The ORBIS database provided only abbreviated financial statements, which did not offer a very detailed overview of the firm financials. Moreover, the initial data set consists of firms that have received funding quite recently, which left me no choice but to select only year after crowdfunding in my analysis. Taken together previous arguments, future researchers may consider more extensive data sets to use a more comprehensive financial overview of crowdfunded firms for the analysis.

Furthermore, the data set consist of quite a few firms that do not show leverage before funding. To better investigate the effect of these firms to my analysis, I conducted a separate zero leverage analysis. However, future research may conduct more research to these firms to get a more comprehensive picture of the phenomenon. Moreover, the number of equity crowdfunded firms in our data set is quite different from debt crowdfunded firms. In future research, a more equally distributed amount of both types is desirable.

(31)

6.2 Future research

Based on the findings of this study, multiple promising research areas for further research are identified. Firstly, prior work in the area of entrepreneurial finance has mainly concentrated on finance sources separately. However, crowdfunding can also be used by SMEs to raise funding in addition of traditional sources like bank debt or other alternative forms of finance in order to increase financial leverage. For future researchers, it could be interesting to dive deeper into this topic and enlarge the emerging stream of crowdfunding research within the entrepreneurial finance context.

A second future research direction is a deeper understanding of ‘the wisdom of the crowd’. As opposed to business angels and venture capitalists, a crowdfunded firm is usually backed by a high number of individual investors who all small amount of money. Therefore, in addition to money, crowdfunding offers the opportunity to exchange knowledge and expertise between the entrepreneurs and the investors (Ferrary & Granovetter, 2009; Belleflamme et al., 2014). Following the thoughts of the knowledge-based view, which is a recent extension of the resource-based view, the accumulated knowledge of the individual crowdfunding investors can potentially be considered as a very special strategic resource and is likely to increase firm performance (DeCarolis & Deeds, 1999). Further work needs to be done to establish whether the SME owners can benefit from the ‘wisdom of the crowd’ and are able to use this knowledge to improve their business operations.

6.3 Practical implications

(32)

REFERENCES

Agrawal, A., Catalini, C. & Goldfarb, A. (2014). Some simple economics of crowdfunding. Innovation Policy and the Economy, 14(1), 63-97.

Ahlers, G. K., Cumming, D., Gunther, C., & Schweizer, D. (2015). Signaling in Equity Crowdfunding. Entrepreneurship Theory and Practice, 39(4), 955-980.

Akerlof, G. (1970). The Market for “Lemons”: Quality Uncertainty and the Market Mechanism. Quarterly Journal of Economics, 89, 488-500.

Anderson, C. (2006). The long tail: Why the future of business is selling less of more. New York: Hyperion Books.

Arthurs, J.D., Busenitz, L.W., Hoskisson, R.E. & Johnson, R.A. (2009). Signaling and initial public offerings: The use and impact of the lockup period. Journal of Business Venturing, 24(4), 360-372.

Artola, C. & Genre, V. (2011). Euro Area SMEs Under Financial Constraints: Belief or Reality? CESifo Working Paper Series No. 3650. Available at: http://www.econstor.eu/handle/10419/53136 (accessed 15 May 2019).

Backes-Gellner, U., & Werner, A. (2007). Entrepreneurial Signaling via Education: A Success Factor in Innovative Start-Ups. Small Business Economics, 29, 173-190.

Baum, J. A. & 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.

Beck, T. & Demirgüç-Kunt, A. (2008). Finance, Firm Size, and Growth. Journal of Money, Credit and Banking, 40(7), 1379–1405.

Belleflamme, P., Lambert, T., Sahm, M. & Schwienbacher, A. (2014). Crowdfunding: Tapping the Right Crowd, Journal of Business Venturing, 29(5), 585-609.

(33)

Bergh, D. D., Connelly, B. L., Ketchen, D. J. & Shannon, L. M. (2014): Signaling theory and equilibrium in strategic management research: An assessment and a research agenda. Journal of Management Studies, 51(8), 1334-1360.

Beugré, C. D., & Das, N. (2013). Limited capital and new venture creation in emerging economies: A model of crowd-capitalism. S.A.M. Advanced Management Journal, 78(3), 21-27.

Block, J., Hornuf, L., & Moritz, A. (2017). Which updates during an equity crowdfunding campaign increase crowd participation? Small Business Economics, 1-25.

Block, J., Colombo, M. G., Cumming, D. J. & Vismara, S. (2018). New players in entrepreneurial finance and why they are there. Small Business Economics, 1-12.

Brabham, D.C. (2008). Crowdsourcing as a Model for Problem Solving: An Introduction and Cases, Convergence: The International Journal of Research into New Media Technologies, 14(1), 75–90.

Bruton, G., Khavul, S., Siegel, D. & Wright, M. (2015). New Financial Alternatives in Seeding Entrepreneurship: Microfinance, Crowdfunding, and Peer-to-Peer Innovations, Entrepreneurship Theory and Practice, 39(1), 9-26.

Burkett, E. (2011). A Crowdfunding Exemption? Online Investment Crowdfunding and U.S. Securities Regulation. The Tennessee Journal of Business Law, 13, 63-106.

Carpenter, R. E. & Petersen. B. C. (2002). Capital Market Imperfections, High-Tech Investment, and New Equity Financing. Economic Journal, 112, 54–72.

Cassar, G. (2004). The financing of business start-ups. Journal of Business Venturing, 19(2), 261–283.

Cassar, G., Ittner, C.D. & Cavalluzzo, K.S. (2015). Alternative information sources of information asymmetry reduction: Evidence from small business debt. Journal of Accounting and Economics, 59(2-3), 242-263.

(34)

Clayman, M. R., Fridson, M. S., & Troughton, G. H. (2012). Corporate Finance: A Practical Approach. New Jersey: CFA Institute Investment Series.

Connelly, B. L., Certo, S. T., Ireland, R. D. & Reutzel, C. R. (2011). Signaling theory: A review and assessment. Journal of management, 37(1), 39-67.

Cosh, A., Cumming, D. & Hughes, A. (2009). Outside Entrepreneurial Capital. Economic Journal, 119, 1494-1533.

DeCarolis, D. M. & Deeds, D. L. (1999). The Impact of Stocks and Flows of Organizational Knowledge on Firm Performance: An Empirical Investigation of the Biotechnology Industry. Strategic Management Journal, 20(10), 953-968.

Drakos, K. (2012). Bank Loan Terms and Conditions for Eurozone SMEs. Small Business Economics, 41(3), 717–732.

Elitzur, R., & Gavious, A. (2003). Contracting, signaling, and moral hazard: A model of entrepreneurs, “angels,” and venture capitalists. Journal of Business Venturing, 18, 709-725.

Ferrando, A. & Griesshaber, N. (2011). Financing Obstacles among Euro Area Firms: Who Suffers the Most? ECB Working Paper Series No 1293. Available at:

http://www.ecb.europa.eu/pub/pdf/scpwps/ecbwp1293.pdf (accessed 7 May 2019).

Ferrary, M. & Granovetter, M. (2009). The Role of Venture Capital Firms in Silicon Valley's Complex Innovation Network. Economy and Society, 18, 326-359.

Frydrych, D., Bock, A. J., Kinder, T. & Koeck, B. (2014). Exploring entrepreneurial legitimacy in reward-based crowdfunding. Venture Capital, 16(3), 247-269.

Gompers, P., Kovner, A., Lerner, J. & Scharfstein, D. (2010). Performance persistence in entrepreneurship. Journal of Financial Economics, 96, 18-32.

Gulati, R., & Higgins, M. C. (2003). Which ties matter when? The contingent effects of

(35)

Hellmann, T., Schure, P. & Vo, D. (2013). Angels and Venture Capitalists: Complements or Substitutes? NBER Working Paper. Available at:

https://pdfs.semanticscholar.org/45a4/1415253a5e57f346414904c7f61c0be964e3.pdf

Hoenig, D. and Henkel, J. (2015). Quality signals? The role of patents, alliances, and team experience in venture capital financing. Research Policy, 44(5), 1049-1064.

Howe, J. (2006, June 1). The Rise of Crowdsourcing, Wired Magazine. Retrieved from:

http://www.wired.com/wired/archive/14.06/crowds.html?pg=1&topic=crowds&topic_set=_The_ Rise_of_Crowdsourcing.pdf

Hsu, D.K., Haynie, J.M., Simmons, S. A. & McKelvie A. (2014). What matters, matters differently: a conjoint analysis of the decision policies of angel and venture capital investors. Venture Capital, 16(1), 1-25.

Hsu, D.H. & Ziedonis, R.H. (2013). Resources as dual sources of advantage: Implications for valuing Entrepreneurial-firm patents. Strategic Management Journal, 34(7), 761-81.

Ibrahim, D. M. (2008). The (Not So) Puzzling Behavior of Angel Investors. Vanderbilt Law Review, 61(5), 1405–1452.

Jõeveer, K. (2012). What Do We Know about the Capital Structure of Small Firms? Small Business Economics, 41(2), 479–501.

Kamer van Koophandel. (2019, May 2). What do the financial statements comprise? Retrieved from: https://www.kvk.nl/english/filing/what-do-the-financial-statements-comprise/

Kerr, W. R., Lerner, J. & Schoar, A. 2014. The consequences of entrepreneurial finance: Evidence from angel financings. Review of Financial Studies, 27(1), 20-55.

Khavul, S. (2010). Microfinance: Creating opportunities for the poor? Academy of Management Perspectives, 24(3), 57–71.

(36)

Kleemann, F., Voß, G. & Rieder, K. (2008). Un(der)Paid Innovators: The Commercial Utilization of Consumer Work through Crowdsourcing, Science, Technology & Innovation Studies, 4(1), 5–26.

Kraemer-Eis, H., Botsari, A., Gvetadze, S., Lang, F. & Torfs, W. (2018), European Small Business Finance Outlook, EIF Research & Market Analysis. Retrieved from:

http://www.eif.org/news_centre/publications/eif-wp-50.pdf

Lambert, T. & Schwienbacher, A. (2010). An Empirical Analysis of Crowdfunding. Working paper. Available at: http://ssrn.com/abstract=1578175

Lehner, O. M. (2013). Crowdfunding social ventures: a model and research agenda. Venture capital: An International Journal of Entrepreneurial Finance, 15(4), 289-311.

Lukkarinen, A., Teich, J. E., Wallenius, H., & Wallenius, J. (2016). Success drivers of online equity crowdfunding campaigns. Decision Support Systems, 87, 26-38.

Mollick, E. (2013). Swept Away by the Crowd? Crowdfunding, Venture Capital, and the Selection of Entrepreneurs. SSRN Working Paper No. 2239204. Available at:

http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2239204

Mollick, E. (2014). The dynamics of crowdfunding: An exploratory study. Journal of Business Venturing, 29(1), 1-16.

Mollick, E. & Nanda, R. (2016). Wisdom or Madness? Comparing Crowds With Expert Evaluation in Funding the Arts. Management Science, 62(6), 1533-1553.

Myers, S. (1977). Determinants of Corporate Borrowing, Journal of Financial Economics, 5(2), 147– 175.

Myers, S. & Majluf, N.S. (1984). Corporate Financing and Investment Decisions when Firms Have Information that Investors Do Not Have, Journal of Financial Economics, 13(2), 187–221.

Nassr, I.K. & Wehinger, G. (2014). Non-Bank Debt Financing for SMEs: The Role of Securitisation, Private Placements and Bonds, OECD Journal: Financial Market Trends, 2014(1), 139–159.

(37)

Piva, E. & Rossi-Lamastra, C. (2017). Human capital signals and entrepreneurs’ success in equity crowdfunding. Small Business Economics, 1-20.

Rasmussen, C. E., & Ghahramani, Z. (2001). Occam's razor. In Advances in Neural Information Processing Systems, 294-300.

Robb, A. (2002). Small Business Financing: Differences between Young and Old Firms, Journal of Entrepreneurial Finance, 7(2), 45–64.

Ryu, S. & Kim, K. (2016). Crowdfunding Success as a Quality Signal to Venture Capital. AMCIS.

Schwienbacher, A., Belleflamme, P. & Lambert, T. (2013). Crowdfunding: Tapping the right crowd. Journal of Business Venturing, 29(4), 585–609.

Shane, S. & Cable, D. (2002). Network ties, reputation, and the financing of new ventures. Management Science, 48(3), 364–381

Spence, M. (1973). Job Market Signaling. Quarterly Journal of Economics, 87, 355-374.

Stiglitz, J. E. (2000). Capital market liberalization, economic growth, and instability. World development, 28(6), 1075-1086.

Stiglitz, J. (2002). Information and the Change in the Paradigm in Economics. The American Economic Review, 92(3), 460-501.

Storey, J.D. & Greene F. (2010), Small Business and Entrepreneurship. Harlow: Pearson Education Limited

Stuart, T. E., Hoang, H. & Hybels, R. C. (1999). Interorganizational endorsements and the performance of entrepreneurial ventures. Administrative science quarterly, 44(2), 315-349.

Surowiecki, J. (2004), The Wisdom of Crowds. Anchor Books, New York.

(38)

Vismara, S. (2018). Signaling to overcome inefficiencies in crowdfunding markets. The Economics of Crowdfunding, 29-56.

Yoo, W., Mayberry, R., Bae, S., Singh, K., Qinghua, P. & Lillard, J. W. (2014). A Study of Effects of MultiCollinearity in the Multivariable Analysis, International Journal of Applied Science and

Technology, 4(5), 9-19.

(39)

Crowdfunding, a proxy for firm quality?