Are Crowdfunding Platforms Active and Effective Intermediaries?

Transcription

1 Are Crowdfunding Platforms Active and Effective Intermediaries? Douglas Cumming Professor and Ontario Research Chair York University - Schulich School of Business 4700 Keele Street Toronto, Ontario M3J 1P3 Canada dcumming@schulich.yorku.ca Yelin Zhang York University - Schulich School of Business 4700 Keele Street Toronto, Ontario M3J 1P3 Canada ZhangY12@schulich.yorku.ca June 25, 2016 * We are indebted to the crowdfunding platforms for providing their detailed data. Also, we are indebted to Craig Asano, Sofia Johan, and the seminar participants at the 2016 National Crowdfunding Association of Canada Annual Summit and York University for helpful comments. This paper is scheduled for presentation at the EMLyon Conference on Entrepreneurial Finance, July 2016.

2 2 Are Crowdfunding Platforms Active and Effective Intermediaries? Abstract Crowdfunding platform due diligence comprises background checks, site visits, credit checks, cross checks, account monitoring, third party proof, and register checks. We expect that crowdfunding platform due diligence reflects the busyness of the platform employees, the fee structure or incentives of the platform, the type or complexity of campaigns, and campaign project uncertainty. By screening lower quality or fraudulent projects and mitigating information asymmetries between entrepreneurs and their funders, we further expect that crowdfunding platform due diligence positively correlates with entrepreneurs meeting their capital goals. Based on data from 51 platforms for each of the years , we find strong evidence consistent with these predictions. Further, we show that strategic fundraising advisory services provided by platforms are associated with a higher likelihood of successful crowdfunding campaigns. Keywords: Crowdfunding, platforms, entrepreneurship, entrepreneurial finance JEL Codes: G23, G24, L26

3 3 I am excited to be joining the OurCrowd team as it continues its growth into Canada s burgeoning tech hub, said David Shore, Director of Investor Relations, OurCrowd Canada. The Canadian market presents a significant opportunity for OurCrowd since startup investing is on the rise in Canada and most investors until now have not had access to high quality deals such as those available on OurCrowd. I look forward to bringing additional trailblazers from the Canadian tech ecosystem onto the OurCrowd platform, and to opening up the global technology market to accredited Canadian investors. - Crowdfund Insider, July 14, Introduction It is widely recognized in oft-repeated media releases such as the one above that crowdfunding has drastically changed the entrepreneurship and entrepreneurial finance ecosystem. Unlike the more well studied traditional forms of finance involving banks, angels, or venture capitalists, crowdfunding is the democratization of capital, with the frequency and success of capital campaigns that are equitably distributed across gender and project types, spurring on the creative process and enabling innovation and entrepreneurship at new levels of vigor not previously possible through traditional forms of entrepreneurial finance, such as banks, angel investors, and venture capitalists. Crowdfunding has experienced at least a doubling in growth every year in recent years around the world, and in 2015 was a $17.25 billion market in North America, an $85.74 million in South America, a $6.84 billion market in Europe, a $

4 4 million market in Africa, a $10.54 billion market in Asia, and a $68.60 million market in Ocenia (Massolution, 2015). With the growth in crowdfunding markets around the world, there are a number of new questions for which answers are not easily transferrable from other types of entrepreneurial finance. One question hitherto unexplored is very basic: are all types of platforms the same and merely provide an ease of connection from individual investors to those that need capital, or are they in fact different in the sense that, for example, one venture capital fund may be different from another which in turn implies massive differences for the success of the venture and the returns to the investor. In view of the massive information asymmetries between investors with capital and entrepreneurs that need capital, which is attributable to the scant or non-existent disclosure requirements when an entrepreneurial firm is not publicly listed on a stock exchange, it becomes quite important to understand what it is that crowdfunding platforms actually do, and whether or not it matters what they do in respect of entrepreneurial outcomes and investor returns. Similarly, as many regions around the world wrestle with legislation pertaining to crowdfunding, it is important to know what role platforms can, should, and/or might play in the governance of crowdfunding marketplaces. In this paper, we address these questions with a unique dataset gathered from 51 Canadian crowdfunding platforms over the years The data have the usual limitation in the sense that they are hand-collected survey data, but nevertheless comprise a majority of the marketplace representing 71% of the total number of crowdfunding platforms in Canada in 2015 (and 100% of the active platforms with a nontrivial presence in the marketplace), and were collected by a reputable third party (The National Crowdfunding Association of Canada, or

5 5 NCFA 2 ). The data enable a first-ever look at what it is that crowdfunding platforms do. In particular, we examine crowdfunding due diligence, or screening of possible projects that get listed on their platforms. We also examine other value-added services provided by platforms that go beyond due diligence. We assess the factors that influence the application of due diligence and other services, as well as whether or not due diligence and other services provided by platforms are associated with the success of projects. The data examined indicate that crowdfunding platform due diligence comprises background checks, site visits, credit checks, cross checks, account monitoring, third party proof, and register checks. The data indicate that crowdfunding platform due diligence is less pronounced for platforms with busy employees that list too many projects, that do not have incentive fee structures tied to the success of projects, and less complex campaigns with less campaign project uncertainty. The data further indicate that there is a strong positive relationship between crowdfunding platform due diligence and the likelihood that entrepreneurs meeting their capital goals, even after controlling for the endogeneity of due diligence. The intuition is that platforms screen lower quality or fraudulent projects and mitigate information asymmetries between entrepreneurs and their funders, thereby increasing the fraction of successful campaigns. Similarly, the data indicate that other types of advice provided by platforms such as strategic fundraising advisory services is associated with a higher likelihood of successful crowdfunding campaigns. Our paper is related to a growing number of papers on rewards based crowdfunding (Agrawal et al., 2015; Bayus, 2013; Belleflamme et al., 2013, 2014; Colombo et al., 2016; Cumming et al., 2014; Kuppuswamy and Bayus, 2014; Lin and Viswanathan, 2014; Mollick, 2

6 6 2014; Qiu, 2014) equity crowdfunding (Ahlers et al., 2015), and crowdfunding regulation (Cumming and Johan, 2013; Griffin, 2012; Hornuf and Schwienbacher, 2014, 2016). A common feature of these papers, however, is that the differences across platforms are not empirically studied, as the data typically come from just one platform. Our paper is distinct insofar as we exploit differences across platforms and have unique proprietary data for the first time ever on what it is that platforms do. Our paper is likewise related to other forms of entrepreneurial finance, such as work on investor effort. Perhaps most directly, our paper is related to work on the importance of due diligence (Yung, 2009) and investor value added (Kanniainen and Keuschnigg, 2003, 2004; Keuschnigg, 2004; Andrieu and Groh, 2012) and reputation (Nahata, 2008) in venture capital and private equity. There is evidence of massive heterogeneity across private equity funds in the extent of due diligence carried out prior to investment, and a positive correlation and even a causal connection between the extent of due diligence and subsequent performance of the investee firm (Cumming and Zambelli, 2016). This paper is organized as follows. Section 2 summarizes the hypotheses. Section 3 presents the data. Section 4 presents the main regression evidence. Section 5 discusses limitations and extensions of the data and tests carried out in this paper. The last section offers concluding remarks as well as lessons for crowdfunding practice and policy.

7 7 2. Institutional Details and Hypotheses on the Crowdfunding Market and Platforms Crowdfunding involves sourcing capital from a large number of (typically) retail investors through an Internet webpage known as a platform. Entrepreneurs post projects for which they need capital. Anyone in the crowd (the pool of possible investors) can see the project on the platform, and decide whether or not to invest. Some platforms facilitate donations and rewards based crowdfunding without offering equity shares, while others offer equity shares in the entrepreneurial firms. Rewards based crowdfunding is the most common worldwide (Massolution, 2015), and involves small rewards, such as early product access, to the crowd in exchange for capital contributions. In Canada, over the sample of our period, there are some platforms that offer equity for sale to accredited retail investors (only). The issues with equity crowdfunding are more complex (Ahlers et al., 2015), in the sample period there was no legislation that afforded equity crowdfunding to non-accredited retail investors. Our first hypothesis pertains to the simple notion of time constraints of agents, and draws from the venture capital literature on portfolio size per manager. Value-added active investors face a tradeoff of adding more investees to their portfolio insofar as having a benefit of network externalities across portfolio firms, but with a cost of having less time for proper due diligence and time to add value to each investee company in respect of providing financial, administrative, human resource, and network and other forms of advice (Kanniainen and Keuschnigg, 2003, 2004; Keuschnigg, 2004; Cumming. 2006; Bernilie et al., 2007; Fulghieri and Sevilir, 2009). The same notion applies to crowdfunding platforms that can spend time on carrying out due diligence on entrepreneurs that want to list a campaign on a platform. Platforms do not have unlimited resources, and it takes time to do background checks, site visits, credit checks, cross checks, account monitoring, third party proof, and register checks. Likewise, strategic

8 8 fundraising advisory services for those firms that are listed also takes time. The greater the number of projects for a given number of platform employees, the less likely it is that there will be sufficient time to carry out all of these activities. Hypothesis 1: Crowdfunding platform due diligence reflects the busyness of the platform (due diligence declines with the number of campaigns on a platform, but increases with the number of platform employees). Different crowdfunding platforms have different fee structures. Different types of fee structures may align the interests of the platform and the crowd that invests in projects listed on the platform. If platforms receive a fee regardless of whether or not the project campaign is successful, then the platform has a greater incentive to list the project regardless of its quality. Conversely, if the platform only receives a fee for funds raised from successful projects, then the platform has more incentive to carry out extensive due diligence and only list projects that are likely to come to fruition. Hypothesis 2: Crowdfunding platform due diligence reflects the fee structure of the platform: platforms that receive fees regardless of the success of the project are less likely to carry out extensive due diligence. Not all entrepreneurial projects are created equal. A project campaign that involves selling equity in an entrepreneurial firm, as opposed to rewards based on donations crowdfunding, is much more complex in respect of contractual terms and necessary monitoring of the platform (Ahlers et al., 2015), and hence requires more due diligence. Similarly, entrepreneurial firms in high-tech industries with more intangible assets on average involves greater information asymmetries between the entrepreneur and the crowd of potential investors,

9 9 and as such likewise requires greater platform due diligence to avoid listing projects that involve misrepresentation. Hypothesis 3: Crowdfunding platform due diligence reflects project campaign uncertainty, with equity crowdfunding requiring more due diligence, and projects in industries with more intangible assets requiring more due diligence. Finally, we expect crowdfunding platform due diligence to be positively associated with improved performance of fundraising campaigns. Effective due diligence removes the left tail of the quality distribution, and such low quality entrepreneurial projects will not appear on the platform. Without the left tail, the average quality is higher. Furthermore, the platform due diligence processes encourage entrepreneurs to present a more transparent campaign to clear the due diligence hurdle, which in turn mitigates information asymmetries between the entrepreneur and the crowd. Entrepreneurs faced with extensive due diligence checks are more likely to take costly steps to signal their quality, such as by preparing quality project descriptions, which in turn lowers information asymmetries faced by investors and signals quality to the crowd (Spence, 2002). Hypothesis 4: Crowdfunding platform due diligence mitigates information asymmetries between entrepreneurs and their funders, and screens lower quality or fraudulent projects, and hence positively influences the success rate of entrepreneurs in terms of meeting their capital goals.

10 10 3. Data, Summary Statistics, and Univariate Comparison Tests The data were provided by National Crowdfunding Association of Canada (NCFA Canada). The data contain information for 51 crowdfunding platforms, representing 71% of total Canadian crowdfunding market. 3 The data cover the time period for each of the four years ; information about 2016 was estimated by each platform, and information for 2013, 2014, and 2015 were based on their historical data. The data are recorded by platform; that is, each row in the data is a different platform (and some of the variables also vary by platform and year, such as average success rates on the platform by year, as detailed below). Table 1 summarizes the different categories of platform information reported in the data. The data indicate the general status of the platform, such as the registration date and status, crowdfunding type, number of employees, website address, and related details. The data indicate the type of due diligence checks carried out by the platform, such as background checks (i.e. verification of government-issued ID), personal meeting or site visit, financial or credit checks, cross-checks from social media connections (such as Facebook or LinkedIn), monitoring of account activities, and requests of third party certificates or proof (such as accredited investor status). Only 49% (25 of 51) platforms acknowledged that they regularly carried out any form of due diligence. The average number of the different types of due diligence checks amongst the seven types (background checks, site visits, credit checks, cross checks, account monitoring, third party proof, and register checks) was 1.2, with a median of 0 and a maximum of 7. The data indicate the services available to subscribers, such as pre-evaluation before listing on the platform, strategic fundraising guidance, business or financial planning, facilitation 3 As of December 2015, there were 72 crowdfunding platforms in Canada. There were fewer platforms in earlier years, implying we have coverage of more than 72% of the market for earlier years.

11 11 in crowdfunding contract design, and marketing or promotional services. A total of 35.3% of platforms provided regular updates to users (investors and entrepreneurs), while 29.4% of platforms offered pre-evaluation to start-ups before listing, 27.5% offered fundraising guidance to entrepreneurs, 17,7% offered marketing and promotional services, 15.7% offered business and financial planning services, and 7.8% offered contractual help to start-ups. The data comprise information on each platform s operating conditions: number of projects launched by year, average successful fund-raising rate, average fund-raising duration, total amount of money raised on platform by year, and the industry composition of listed projects. The median platform has projects that take between 7-9 weeks for funding, and the median platform has entrepreneurs that achieve 21-30% of their funding goal. The median platform spends between $2,500 to $10,000 on compliance annually, has 10 employees, investors per year, has entrepreneurial projects listed per year, and total capital raised between $50,000 - $100,000 per year. The more common industries include non-profit (median 25%), business and professional services (median 15%), education and research (median 10%), art (median 10%), life science (median 10%), cleantech and energy (median 5%), and hardware and software (median 5%), followed by manufacturing, media, real estate, and social enterprise each with medians per platform at 0% but averages across platforms between 2-5% per year (details are in Table 1). Fee structure/revenue models of the platform are in the data, including information on whether or not the platform charges a one-time platform listing fee, periodical subscription at different levels/tiers, fixed percentage of total amount raised (whether funding is successful or not), fixed percentage of total amount raised (only if funding is successful), and management fees and carry percentages. There are 11.8% of platforms with a one-time listing fee, 19.6% of

12 12 funds with a periodical subscription at different levels, 13.7% of funds with a fixed percentage of the total amount raised regardless of success, 15.7% of funds with a fixed percentage raised only if funded was successful, and 11.8% of funds with a mixed management fee and a carry percentage fee structure. Other variables and detailed summary statistics (means, medians, standard deviations, minimum, and maximums) are shown in Table 1. Table 1 About Here The comparison tests in Table 2 provide a first impression of some distinct patterns in the dataset. Although comparison tests do not show the precise relationship among variables controlling for other things being equal because variables are analyzed separately and in isolation, the comparison tests nevertheless present a general picture on relationships in the data and the relationships between some key variables of interest. The joint effect of the same variables is discussed later in the regression analysis in the next section. The data in Table 2 are based on the 2013 information provided by the platforms. We applied the same tests on data in later years; the results are consistent with what we report for 2013, although the magnitude of the differences and statistical significance slightly varies. Table 2 About Here Table 2 Panel A shows that there is a higher probability of due-diligence application

13 13 among platforms with a smaller number of projects (when the number of campaign projects is greater than the median across platforms only 17.9% of platforms carry out due diligence, and when the number of campaign projects is below the median then 87.0% of platforms carry out due diligence), consistent with Hypothesis 1. Similarly, there is a higher probability of duediligence among platforms with a great number of employees (when the number of employees is greater than the median across platforms there are 86.4% of platforms carrying out due diligence, and when the number of employees is below the median then 20.7% of platforms carry out due diligence), again consistent with Hypothesis 1. There is a higher probability of due-diligence among platforms that spend more on compliance (when compliance expenditures are greater than the median across platforms there are 95.5% of platforms carrying out due diligence, and when the compliance expenditures are below the median then 13.8% of platforms carry out due diligence). Due diligence is more likely among platforms that have fee models with periodical subscription and different levels (73.5% with these fees, versus 30.7% without this fee structure), consistent with Hypothesis 2. There is a higher probability of due-diligence among equity crowdfunding platforms (80.1% of equity crowdfunding platforms carry out due diligence, while 41.5% of non-equity crowdfunding platforms carry out due diligence), consistent with Hypothesis 3. There is a higher probability of due-diligence among platforms with art, lifesciences, and non-profit and charities, which is in part consistent with Hypothesis 3 at least for the life-sciences industries. Each of these differences are statistically significant at at least the 5% level, with the exception of the differences for art industries which is significant only at the 10% level. Table 2 Panel B presents the possible advantages of due-diligence application, consistent with Hypothesis 4. There is a higher level of fully funded projects and larger amount of money

14 14 raised through platforms that carry our due diligence, and these differences are significant at the 1% level. Due-diligence does not have a significant effect on fundraising duration. Table 3 reals the correlations among variables of interest. Table 3 Panel A shows that duediligence application is positively correlated with resource on compliance, employee numbers, equity crowdfunding, user subscription at different levels, and a higher industry composition of art, life science and charities. Due diligence application is negatively correlated with the number of campaign projects on a platform. Table 3 Panel B shows the correlation among due-diligence application, platform performance, and different types of platform services. Consistent with comparison test results in Table 2 Panel B, due-diligence application is positively associated with higher percentage of fully funded projects and larger amount of money raised through a platform and does not exhibit strong correlation with fund-raising duration. The number of types of duediligence is positively correlated with percentage of fully funded project, total amount of money raised through the platform, and several platform services. The number of projects per employee ratio is negatively correlated with total amount of money raised through the platform, periodical updates, and strategic guidance. Further detailed correlations among platform services are also presented in Table 3. Table 3 About Here

15 15 4. Multivariate Analyses In this section, we use detailed analysis to reveal the factors influencing platform duediligence application and demonstrate how due-diligence application benefits crowdfunding platforms. We used two methods in the analysis: logit regressions to examine the factors that influence platform due diligence in Tables 4 and 5, and ordered logit regressions to examine whether due-diligence, among other things, influences platform performance in Table 6, accounting for the non-random application of due diligence in the first step. For reasons of conciseness, we do not show regressions of all types of due diligence one by one. Instead, we first report regressions for the overall due-diligence (all kinds of due-diligence combined), and then report regressions for the three most common due-diligence subcategories: background check, site visit and cross check. Tables 4 and 5 About Here Table 4 shows the factors influencing due-diligence application in general. Due-diligence is applied when at least one of the following actions is taken: background check, site visit, credit check, cross-check from social media connections, monitor account activities, and request third party certificates or proof. We applied the same logit regressions on data in different years; the results are consistent over time. The data indicate that number of projects in each year is negatively correlated with due-diligence application, and this effect is statistically significant at the 10% level in Models 1 and 3 for 2013 and 2015, respectively, and at the 5% level in Models 2 and 4 for 2014 and 2016, respectively. The economic significance is such that on average, an increase by one categorical unit (the number of projects is an ordinal variable; see Table 1) in

16 16 the number of projects is associated with a 13.85% (Model 4) to 23.79% (Model 1) reduction in probability of due-diligence, consistent with Hypothesis 1. Similarly, the number of employees is positively and significantly correlated with due-diligence: a one standard deviation increase in the number of employees results in a 2.52% (Model 1) to 6.75% (Model 4) increase in the probability of due-diligence, again consistent with Hypothesis 1. These results remain even when controlling for resources spent on compliance, which is positively and significantly correlated with due-diligence: an increase by one unit in resources spent on compliance (resources is an ordinal variable; see Table 1) leads to an 8.55% (Model 2) to 10.35% (Model 4) increase the probability of due-diligence. Platforms with periodical subscription at different levels are more likely to carry out due diligence by 4.28% (Model 2) to 5.85% (Model 4), consistent with Hypothesis 2. Also, platforms that offer equity crowdfunding are more likely to carry out due-diligence, consistent with Hypothesis 3. On average, a platform with equity crowdfunding is 14.68% to 16.98% more likely to carry out due-diligence than a platform without equity crowdfunding. These effects are robust across each of the years, and robust to controls for industry (the industry effect, although not explicitly reported in Table 4 for conciseness, are consistent with that which was reported in Table 2 above). Table 5 further analyzes three main types of due-diligence application: background checks (Panel A), site visits (Panel B), and cross-checks from social media connections (Panel C). The data in Table 5 Panel A indicate the following. A one unit increase in the number of campaign projects reduces the probability of background checks by 13.36% (Year 2013) to 16.95% (Year 2014). A one unit increase in the amount of resources spent on compliance annually increases the probability of background checks by 3.49% (Year 2016) to 5.30% (Year

17 ). A one standard deviation increase in the number of employees increases the probability of background checks by 2.55% (Year 2013) to 4.06% (Year 2015). Equity crowdfunding increases the probability of background checks by 8.24% (Year 2013) to 12.24% (Year 2014). Advanced fee structures increase the probability of background checks by 1.94% (Year 2013) to 4.56% (Year 2016). Table 5 Panel B indicates that a one unit increase in the number of projects reduces probability of site visits by 7.88% (Year 2015) to 10.46% (Year 2014). A one standard deviation increase in the number of platform employees increases the probability of site visit by 1.82% (Year 2013) to 3.13% (Year 2014). Equity crowdfunding is associated with an increase in the probability of site visit by 2.12% (Year 2013) to 3.23% (Year 2015). The fee structure, however, does not have statistically significant impact on site visits. Table 5 Panel C indicates that a one unit increase in resources spent on compliance increases the probability of cross-checks by 9.58% (Year 2013) to 16.49% (Year 2015). A one standard deviation increase in the number of employees increases the probability of cross-checks by 3.39% (Year 2013) % (Year 2016). Equity crowdfunding is associated with an increase in the probability of cross-checks by 2.85% (Year 2016) to 4.53% (Year 2013). Advanced fee structures increase the probability of cross-checks by 3.80% (Year 2016) to 6.06% (Year 2015). The number of campaign projects and number of investors do not have a noticeable impact on cross-checks from social media connections. Also, platform age has no influence on duediligence application, likely because most platforms are very young in our sample. Table 6 presents the results for the impact of due-diligence application on platform performance (controlling for the determinants of due diligence from Table 4), measured by percentage of fully funded projects, total amount of money raised annually, and average fund-

18 18 raising duration. Since it is costly for platforms to apply due-diligence, the according benefit should outweigh the cost to justify for the expenses. We use the fitted values of due diligence from Table 4, with the instruments that include the fee structure and resources spent on compliance. Fee structures and resources spent on compliance will be directly connected with due diligence, but only indirectly related (if at all) to project outcomes since fee structures and compliance expenditures are unknown to the crowd investors. We note that the results without fitted values for the potentially endogenous due diligence variables are not materially different when we use the non-fitted raw variables, and likewise similar with different controls in the first stage regressions. Table 6 About Here Table 6 Panel A shows that due-diligence application is associated with higher percentage of fully funded projects, controlling for all types of services offered by the platform. Specifically, the application of due-diligence increases the percentage of fully funded projects by 40.96% (Year 2014) to 60.45% (Year 2016). This effect is statistically significant at 5% level in each Model in Table 6. Also, notice that project/employee ratio has a negative impact on the percentage of fully funded projects: the higher the ratio, the less resource will be devoted to each project, the more competition there is across projects, and the lower the success of projects on average. Table 6 Panel B shows that due-diligence application is associated with larger amount of money raised through a platform. Specifically, the application of due-diligence increases the total amount of money raised by 27.62% (Year 2014) to 39.06% (Year 2015), controlling for number

19 19 of projects listed on the platform and services provided by the platform. Table 6 Panel C does not present robust negative relationship between due-diligence application and fund-raising duration (the effect is only marginally significant at the 10% level in Model 4 and insignificant in the other models); nevertheless, the negative coefficient for duediligence application shows that as due-diligence applied, fund-raising becomes quicker, which is consistent with what we would expect. Notice that promotion and marketing service does have a significant impact on the efficiency of fund-raising: on average, fund-raising becomes 20.88% (Year 2013) to 29.51% (Year 2014) quicker when promotion and marketing service is offered. 5. Conclusion The past decade leading up to 2016 has witnessed a massive growth in the popularity of crowdfunding as a viable form of entrepreneurial finance. In Canada, thousands of new projects are launched on different fund raising websites every year. Connecting donors and investors with beneficiaries, borrowers and entrepreneurs, crowdfunding platforms help idle money find its value. But exactly what do crowdfunding platforms do? Do they simply provide a cheap online spot for business soliciting? Or, do they apply due-diligence on listed projects and help reduce information asymmetry between projects initiators and subscribers? What advantage can platforms obtain through carrying out due-diligence? In this paper, we assess the factors that influence the application of due diligence, as well as whether or not due diligence by platforms is associated with success of projects. The scope of crowdfunding due diligence comprises background checks, site visits, credit checks, cross checks, monitoring accounts, third party proof, and register checks. For the first time ever, we

20 20 examine empirical data on topic, made possible from the innovation data collection efforts of the National Crowdfunding Association of Canada. The summary statistics and comparison tests showed a transparent picture in the data, as do the regression results controlling for other things being equal. The application of duediligence is associated with more affluent platform resources, either in compliance expenditure or in employee number; and more sophisticated management structure, indicated by different levels of subscription service. Due-diligence is also more likely to be applied when a platform contains projects pertaining to higher intangible assets or information asymmetry, including the life science industries, as well as in the arts and charities. Due-diligence is less likely to be applied when platform employees expected working load is heavier, as shown by larger number of campaign projects launched on a platform. The data further indicate that the application of due-diligence in general has very strong positive influence on the fund-raising successful rate and amount in the platform, controlling for all service offered by a platform. Among all service offered by platforms, only strategic fund raising guidance is significantly positively related with fund-raising successful rate and total amount raised through platform. The strong positive association between due diligence and fundraising success shows an important value for crowdfunding platforms in limiting the number of lower quality projects on a platform through active due diligence. The evidence herein suggests that there is a strong case for policymakers to impose standards on platforms to act with greater stringency in carrying out due diligence. Also, the evidence suggests that further research on crowdfunding could play careful attention to differences across platforms, as there appears to be massive heterogeneity in respect of what it is that platforms actually do.

21 21 References Agrawal, A., Catalini, C., and Goldfarb, A. (2015). Crowdfunding: Geography, Social Networks, and the Timing of Investment Decisions, Journal of Economics and Management Strategy, 24 (2), Ahlers, G.K.C., Cumming, D.J., Günther, C., and Schweizer, D. (2015). Signaling in Equity Crowdfunding, Entrepreneurship Theory and Practice, 29, Andrieu, A., Groh, A. (2012). Entrepreneurs' Financing Choice between Independent and Bank- Affiliated Venture Capital Firms, Journal of Corporate Finance 18, Bayus B. (2013). Crowdsourcing New Product Ideas over Time: An Analysis of the Dell IdeaStorm Community. Management Science 59 (1), Belleflamme, P., Lambert, T., and Schwienbacher, A. (2013). Individual Crowdfunding Practices, Venture Capital: An International Journal of Entrepreneurial Finance 15 (4), Belleflamme, P., Lambert, T., and Schwienbacher, A. (2014). Crowdfunding: Tapping the Right Crowd, Journal of Business Venturing 29(5), Bernile, G., Cumming, D.J., and Lyandres, E The Size of Venture Capital and Private Equity Fund Portfolios Journal of Corporate Finance 13, Colombo, M.G., Franzoni, C., Rossi-Lamastra, C. (2015). Internal Social Capital and the Attraction of Early Contributions in Crowdfunding Projects, Entrepreneurship Theory and Practice, 39, Cumming, D.J. (2006). The Determinants of Venture Capital Portfolio Size: Empirical Evidence. Journal of Business 79, Cumming, D.J., Leboeuf, G., and Schwienbacher, A. (2014). Crowdfunding Models: Keep-it-All vs. All-or-Nothing, Working Paper, York University and University of Lille. Cumming, D.J., and Johan, S.A. (2013). Demand Driven Securities Regulation: Evidence from Crowdfunding, Venture Capital: An International Journal of Entrepreneurial Finance 15, Cumming, D.J., and Zambelli, S. (2014). Due Diligence and Investee Performance, European Financial Management, forthcoming. Fulghieri, P., and Sevilir, M. (2009). Size and Focus of a Venture Capitalist s Portfolio, Review of Financial Studies, 22,

22 22 Griffin, Z.J. (2012). Crowdfunding: Fleecing the American Masses, Working paper, Available at: Auteurs : Hakenes, H. and Schlegel, F. (2014). Exploiting the Financial Wisdom of the Crowd, Working paper, Available at: Hornuf, L. and Schwienbacher, A. (2014). The Emergence of Crowdinvesting in Europe, Working paper, Available at: Hornuf, L. and Schwienbacher, A. (2016). Which Securities Regulation Promotes Crowdinvesting? Small Business Economics, forthcoming. Kanniainen, V., Keuschnigg, C. (2003). The optimal portfolio of start-up firms in venture capital finance, Journal of Corporate Finance, 9, Kanniainen, V., Keuschnigg, C. (2004). Start-up investment with scarce venture capital support, Journal of Banking and Finance 28, Keuschnigg, C., (2004). Taxation of a venture capitalist with a portfolio of firms. Oxford Economic Papers, 56: Kuppuswamy, V., and Bayus, B. (2014). Crowdfunding Creative Ideas: The Dynamics of Project Backers in Kickstarter. UNC Kenan-Flagler Research Paper No Lin, M. and Viswanathan, S. (2014). Home Bias in Online Investments: An Empirical Study of an Online Crowdfunding Market, Working Paper. Massolution, Crowdfunding Industry Report, Massolution/Crowdsourcing.org 2015CF, URL: Mollick, E.R. (2014). The Dynamics of Crowdfunding: Determinants of Success and Failure, Journal of Business Venturing 29, Nahata, R., (2008). Venture Capital Reputation and Investment Performance, Journal of Financial Economics, 90, Qiu, C. (2013). Issues in Crowdfunding: Theoretical and Empirical Investigation on Kickstarter, Working Paper. Spence, M. (2002). Signaling in Retrospect and the Informational Structure of Markets, American Economic Review 92, Yung, C. (2009). Entrepreneurial Financing and Costly Due Diligence, Financial Review, 44,

23 23 Table 1. Definitions and Summary Statistics This table provides definitions of the main variables, as well as summary statistics. Variable Definition Obs Mean Median Std. Dev Min Max Any Due Diligence is applied? (Yes=1) Dummy Variable: Is Any of the Following Due Diligence Regularly Applied: Background Check, Site Visit, Credit Check, Cross-check from Social Media Connections, Monitor Account Activities, Request Third Party Certificates or Proof? (Yes=1, No=0) Dummy Variable: Does a Platform Provide Periodical (weekly, bi-weekly, monthly) Platform Provides Periodical Updated Platform Updates and Activities to Users (Investors and Startups)? Information to Users (Yes=1, No=0) Platform offers pre-evaluation before Dummy Variable: Does a Platform Offer Pre-evaluation to Startups before Their Listing? listing Startups (Yes=1, No=0) Platform offers strategic fundraising Dummy Variable: Does a Platform Offer Strategic Fundraising Guidance to Startups? guidance Platform helps with business and financial planning Platform offers contractual help to Startups Platform offers marketing or promotion service Total Number of types of Due Diligence applied Average Fund-raising Duration Percentage Composition of Fully Funded Project Resource Spent on Compliance Annually (Yes=1, No=0) Dummy Variable: Does Platform Helps Startups with Business and Financial Planning? (Yes=1, No=0) Dummy Variable: Does a Platform Offer Contractual Help to Startups? (Yes=1, No=0) Dummy Variable: Does a Platform Offer Marketing or Promotion Service to Startups? (Yes=1, No=0) Total Number of Types of Due Diligence Applied by a Platform. Types of Due Diligence refer to: Background Check, Site Visit, Credit Check, Cross-check from Social Media Connections, Monitor Account Activities, Request Third Party Certificates or Proof Ordinal Variable: Average Fund-raising Duration Level 1: 1-3 Weeks; Level 2: 4-6 Weeks; Level 3: 7-9 Weeks; Level 4: Weeks; Level 5: More than 12 Weeks Ordinal Variable: the percentage of fully funded project with respect to all projects launched on the platform. Level 1:0-10%; Level 2:11-20%; Level 3: 21-30%; Level 4: 31-40%; Level 5:41-50%; Level 6: More than 50% Ordinal Variable: Total Resource Spent on Compliance Annually? Level 1: less than $2500; Level 2: $ ; Level 3: $ ; Level 4: $ ; Level 5: Larger than $ Number of Employees Number of Employees Working for a Crowdfunding Platform Equity Crowdfunding on the Platform? (Yes=1) Dummy Variable: Does a Platform Handles Equity Crowdfunding? (Yes=1, No=0) Platform Age Number of Months between Platform Setup Month and February Fee Structure: One-time Platform Dummy Variable: Is the Main Service Charge a One-time Listing Fee? Listing Fee (Yes=1, No=0) Fee Structure: Periodical Subscription at Dummy Variable: Is the Main Service Charge based on Periodical Subscription at Different Levels/Tiers Different Levels/Tiers? (Yes=1, No=0) Fee Structure: Fixed percentage of total Dummy Variable: Is the Main Service Charge based on Fixed Percentage of Total Amount amount raised, whether funding is Raised, whether Funding is Successful or Not? (Yes=1, No=0) successful or not Fee Structure: Fixed percentage of total amount raised, only if funding is successful Fee Structure: Management fee and carry percentage Dummy Variable: Is the Main Service Charge based on Fixed Percentage of Total Amount Raised, only if Funding is Successful? (Yes=1, No=0) Dummy Variable: Is the Main Service Charge based on Management Fee and Carry Percentage? (Yes=1, No=0)

24 24 Table 1. (Continued) Variable Definition Obs Mean Median Std. Dev Min Max Ordinal Variable :Total Number of Investors in Each Year from 2013 to 2016 (estimated) Number of Investors in Respective Year Level 1: less than 100; Level 2: ; Level 3: ; Level 4: ; Level 5: ; Level 6: ; Level 7: ; Level 8: ; Level 9: Larger than (Number of Investors in 2013) (Number of Investors in 2014) (Number of Investors in 2015) (Number of Investors in 2016: estimated) Ordinal Variable :Total Number of Projects/Financings/Loans Launched in Each Year Number of Projects in Respective Year from 2013 to 2016 (estimated) Level 1: less than 20; Level 2: 21-50; Level 3: ; Level 4: ; Level 5: ; Level 6: Larger than 500 (Number of Projects in 2013) (Number of Projects in 2014) (Number of Projects in 2015) (Number of Projects in 2016: estimated) Ordinal Variable: Total Number of Projects Launched in Each Year from 2013 to 2016(estimated) divided by Number of Employees for Each Platform: Level 1: less than 10 Project/Employee Ratio in Respective Year projects per employee; Level 2: projects per employee; Level 3:21-50 projects per employee; Level 4: projects per employee; Level 5: Larger than 100 projects per employee (Project/Employee Ratio in 2013) (Project/Employee Ratio in 2014) (Project/Employee Ratio in 2015) (Project/Employee Ratio in 2016: estimated) Ordinal Variable: Total Amount of Money Raised in Each Year from 2013 to Total Amount of Money Raised 2016(estimated) Level 1: Less than 2.5K; Level 2: 2.5K-10K; Level 3: 10K-50K;Level 4: 50K-100K; Level 5: 100K-500K; Level 6: 500K-1 M; Level 7: 1M-5M; Level 8: More than 5M (Total Amount in 2013) (Total Amount in 2014) (Total Amount in 2015) (Total Amount in 2016: estimated) Industry Composition in respective year: Art Percentage of Projects Launched on the Platform Categorized as Art (Art in 2013) % 10% % 100% (Art in 2014) % 10% % 100% (Art in 2015) % 10% % 100% (Art in 2016:estimated) % 10% % 100% Industry Composition in respective year: Business and Professional Service Percentage of Projects Launched on the Platform Categorized as Business and Professional Service (Business and Professional Service in 2013) % 15% % 30% (Business and Professional Service in 2014) % 15% % 30% (Business and Professional Service in 2015) % 15% % 70% (Business and Professional Service in 2016: estimated) % 15% % 70%

25 25 Table 1.(Continued) Variable Definition Obs Mean Median Std. Dev Min Max Industry Composition in respective year: Education and Research Percentage of Projects Launched on the Platform Categorized as Education and Research (Education and Research in 2013) % 10% % 10% (Education and Research in 2014) % 10% % 10% (Education and Research in 2015) % 10% % 30% (Education and Research in 2016: estimated) % 10% % 30% Industry Composition in respective year: Clean Tech and Energy Percentage of Projects Launched on the Platform Categorized as Clean Tech and Energy (Clean Tech and Energy in 2013) % 5% % 10% (Clean Tech and Energy in 2014) % 5% % 10% (Clean Tech and Energy in 2015) % 5% % 10% (Clean Tech and Energy in 2016: estimated) % 5% % 15% Industry Composition in respective year: Life Science Percentage of Projects Launched on the Platform Categorized as Life Science (Life Science in 2013) % 10% % 20% (Life Science in 2014) % 10% % 25% (Life Science in 2015) % 10% % 25% (Life Science in 2016: estimated) % 10% % 30% Industry Composition in respective year: Manufacturing Percentage of Projects Launched on the Platform Categorized as Manufacturing (Manufacturing in 2013) % 0% % 5% (Manufacturing in 2014) % 0% % 10% (Manufacturing in 2015) % 0% % 20% (Manufacturing in 2016: estimated) % 0% % 20% Industry Composition in respective year: media Percentage of Projects Launched on the Platform Categorized as Media (Media in 2013) % 0% % 10% (Media in 2014) % 0% % 30% (Media in 2015) % 0% % 30% (Media in 2016: estimated) % 0% % 30%

26 26 Table 1.(Continued) Variable Definition Obs Mean Median Std. Dev Min Max Industry Composition in respective year: Non- Profit and Charities Percentage of Projects Launched on the Platform Categorized as Non-Profit and Charities (Profit and Charities in 2013) % 25% % 100% (Profit and Charities in 2014) % 25% % 100% (Profit and Charities in 2015) % 25% % 100% (Profit and Charities in 2016: estimated) % 25% % 100% Industry Composition in respective year: Real Percentage of Projects Launched on the Platform Categorized as Real Estate Estate (Real Estate in 2013) % 0% % 15% (Real Estate in 2014) % 0% % 25% (Real Estate in 2015) % 0% % 100% (Real Estate in 2016:estimated) % 0% % 100% Industry Composition in respective year: Social Percentage of Projects Launched on the Platform Categorized as Social Enterprise Enterprise (Social Enterprise in 2013) % 0% % 50% (Social Enterprise in 2014) % 0% % 50% (Social Enterprise in 2015) % 0% % 50% (Social Enterprise in 2016: estimated) % 0% % 50% Industry Composition in respective year: Percentage of Projects Launched on the Platform Categorized as Hardware and Software Hardware and Software (Hardware and Software in 2013) % 5% % 10% (Hardware and Software in 2014) % 5% % 20% (Hardware and Software in 2015) % 5% % 100% (Hardware and Software in 2016: estimated) % 5% % 100%

27 27 Table 2. Comparison Tests on Variable Impact This table shows the impact of different platform characters on due diligence application (Panel A) and the impact of due diligence application on platform performance (Panel B) using comparison tests. The table is based on platform activities in year We observe similar results when analyzing platform activities in year 2014, 2015 and 2016(estimated). *, **, *** Significant at the 10%, 5%, and 1% levels, respectively. Panel A: Factors Affecting Due Diligence Application (Year 2013) Project Number (Below Median) Project Number (Above Median) Z Value 4 Compliance Expenditure (Below Median) Compliance Expenditure (Above Median) Probability of Due Diligence Application *** *** Employee Number (Below Median) Employee Number (Above Median) Z Value Equity Crowdfunding (Not Available) Equity Crowdfunding (Available) Probability of Due Diligence Application *** ** Periodical Subscription at Different Levels/Tiers (No) Periodical Subscription at Different Levels/Tiers (Yes) Z Value Industry Composition: Art (Below Median) Industry Composition: Art (Above Median) Probability of Due Diligence Application *** * Industry Composition: Life Science (Below Median) Industry Composition: Life Science (Above Median) Z Value Industry Composition: Non-Profit and Charities (Below Median) Industry Composition: Non-Profit and Charities (Above Median) Probability of Due Diligence Application ** *** Z Value Z Value Z Value Z Value Panel B: Impact of Due Diligence Application on Platform Performance (Year 2013) Average Percentage of Fully Funded Project (in Levels/Ordinal Scale) Average Amount of Money Raised (in Levels/Ordinal Scale) Average Fund-raising Duration (in Levels/Ordinal Scale) Due Diligence (Not Applied) Due Diligence (Applied) T Value *** *** Two-sample test of proportions 5 Two-sample t test with equal variances