corp financial

Crowdfunding and Venture Capital: Substitutes or Complements?

Author(s): Mario D'Ambrosio and Gianfranco Gianfrate

Source: The Journal of Private Equity , WINTER 2016, Vol. 20, No. 1 (WINTER 2016), pp. 7-20

Published by: Euromoney Institutional Investor PLC

Stable URL: https://www.jstor.org/stable/44396812

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

is collaborating with JSTOR to digitize, preserve and extend access to The Journal of Private Equity

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Crowdfunding and Venture

Capital: Substitutes or

Complements? Mario D'Ambrosio and Gianfranco Gianfrate

Mario D'Ambrosio

is a former analyst at H.I.G. Capital in Milan, Italy. [email protected]

Gianfranco

Gianfrate

is a Giorgio Ruffolo Fellow in Sustainability Science at Harvard

University in Cambridge, MA.

gianfranco_gianfrate @ hks

.harvard.edu

Despite to scale of crowdfunding encourage action the growing by the it as and U.S. importance a the source Congress large- of

of crowdfunding and the large- scale action by the U.S. Congress to encourage it as a source of

capital for new ventures, the phenomenon remains relatively unexplored. Specifi- cally, the nature of the interactions between crowdfunding and other traditional forms of entrepreneurial financing - venture capital, especially - has received scant attention. Is crowdfunding an alternative or comple- mentary source of funding with respect to traditional venture capital and angel invest- ments? Will venture capital be swept away by crowdfunding?

The question of whether crowdfunding is or will become a meaningful alternative to angel finance and venture capital still needs to be explored: There is uncertainty about whether crowdfunding efforts rein- force or contradict existing theories on how ventures raise capital and achieve success (Mollick [2013]). Moreover, if it is accepted that the success of traditionally-funded new ventures is often highly constrained by geog-

raphy (Chen et al. [2009]), then the role of geography in crowdfunding represents a noteworthy subject. Some researchers and academics have noted that crowdfunding has the potential to mitigate many of the dis- tance effects found in traditional fundraising efforts. Some evidence has been found that

crowdfunding relaxes geographic constraints

among funders (Agrawal, Catalini, and Goldfarb [2011]), but the effect of geography on project founders is less clear. Other authors

(Mollick [2013]) document the geographical concentration of crowdfunding but neglect a fundamental aspect: whether crowdfunding is concentrated in the same main centers as

venture capital activity or in different areas. The research question we attempt to

answer is whether crowdfunding is a comple- ment to or rather a substitute for traditional

venture capital investments. For instance, research findings may suggest that new ventures receiving capital through crowd- funding typically withstand the market without asking for further rounds of venture capital financing, classifying this new source of funding as a substitute of venture capital; or that crowdfunding typically asks for and enhances further rounds of investments from

venture capitalists, classifying it as a comple- ment to the venture capital model. Therefore, this article suggests that crowdfunding is a substitute for venture capital for entrepreneurs

at very beginning of their entrepreneurial activity and that it needs and triggers further

rounds of venture capital financing after- ward, classifying itself as a complement to venture capital in after seed stages.

The second question addresses the geo- graphical determinánts of entrepreneurial financing: We argue that if it holds true that crowdfunding represents a substitute

Winter 2016 The Journal of Private Equity 7

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

for venture capital in the seed stage and a complement afterward, we do not expect significant differences in their geographical patterns. In other words, we expect crowdfunding to have the same geographical drivers as venture capital investments and to be concentrated in the same main areas. Therefore, while crowdfunding relaxes geographic constraints among funders, making it easier to provide capital from any location through the Internet, it does not mitigate the effects of geography on project founders.

Since crowdfunding is based on online platforms that enable interactions with friends and relatives, some

scholars (e.g., Schwienbacher and Larralde [2010]) have described crowdfunding as an alternative way to finance projects for small, entrepreneurial ventures - neglecting the potential implications of traditional fund- raising models. However, in recent years, crowdfunding has proved to be a means to raise funds not only for small projects but also for high-growth start-ups typi- cally financed by venture capital and angel investors. Therefore, a strong need has emerged to understand the degree to which crowdfunding will ultimately sub- stitute for other forms of more formal venture funding

(Mollick [2013]).

DATA AND METHODS

In our analysis, we use data extracted from Kickstarter - the leading crowdfunding platform globally - to proxy for crowdfunding investment amounts and geographical dispersion. We specifically gathered data related only to those projects supporting business ideas and eventually linked to the creation of new ventures. Specifically, we include only projects related to the following categories: games, food, fashion, product design, and technology. Moreover, given the need of this study to make a comparison with venture

capital financing, only successful projects were taken into consideration.1 Following this first screening, we further limited the projects to those looking for $5,000 or more in capital, in order to further restrict our data to

initiatives looking for more substantial funding and thus aimed at supporting long-lasting enterprises.2 For each project, data were gathered providing information about entrepreneurs and their plans; location of the founder (and thus of the potential new venture) in terms of city and state; amount pledged by investors, percentage of funding, and implicitly, initial fundraising goal;3 and end date of the funding window. Exhibit 1 provides a summary of this data by category.

To tackle the second research question of this article

and examine geographical clustering of crowdfunding, investments have been grouped at the Combined Sta- tistical Area (CSA) and Metropolitan and Micropolitan Statistical Area (MSA) level for each year. The U.S. Census Bureau provides a list of Core Based Statistical Aras (CSBAs), MSAs, and CSAs, as of February 2013. Therefore, the location of each project in the dataset was assigned at the CSA level and, in cases where a city is not located in a CSA,4 it was assigned to the appro- priate MSA.5 This allows us to study the geographical clustering of crowdfunding at the CSA and MSA levels. In 2013, 105 CSAs and 59 MSAs not located in a CSA had significant crowdfunding activity.

For venture capital financing activity, data were drawn from Thomson One Investment Banking. In this study, venture capital activity is divided into four main investment stages: (1) seed, that is, the first stage of venture capital financing, aimed at raising relatively modest amounts of capital to finance the early devel- opment of a new product or service; (2) early stage, for companies that are able to begin operations but are not yet at the stage of commercial manufacturing and sales; (3) expansion, in which financing is provided for

Exhibit 1

Summary Statistics

(1) (2) (3) (4) (5) (6)

AU Fashion Food Games Product Design Technology

Projects (n°) 5.954 976 242 1.958 1.602 1176 Goal ($ min) 165.8 15.1 4.0 60.9 38.7 47.1 } Pledged ($ min) 512.9 33.8 6.6 186.3 136.4 149.8

Funded

8 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

major expansion such as the purchase of physical plants, product improvement, and marketing; (4) later stage, which refers to capital provided after commercial manu- facturing and sales to companies that boast stable growth and before any initial public offering. In 2013, venture capital activities were distributed across 81 CSAs and 38 MSAs not located in a CSA.

From the point of view of the empirical analysis, we measure crowdfunding and venture capital invest- ments as follows:

• CF. The natural log of the total capital raised through crowdfunding to finance all the successful projects on Kickstarter in a given month;

• VC_SEED. The natural log of the total amount of funding provided as seed capital from venture capitalists in a given month; and

• VC_AFTERSEED. The natural log of the total amount of funding provided by venture capitalists in a given month, excluding the capital raised as seed financing. In other words, it makes reference to only the three investment stages after seed: early

stage, expansion, and later stage.

Therefore, monthly observations of crowdfunding and venture capital investments (with a distinction, for the latter, between seed and after seed stages) are used to explore the relationship between the two fundraising models.

Empirically, we built a vector autoregressive model, in which each variable is regressed on one or more lags of every other variable. As discussed in the next sec- tion, a VAR model is particularly useful in describing the relationship between crowdfunding and venture capital in after seed stages, because it helps us understand how an increase in crowdfunding investing today can eventually enhance after seed venture capital activities at a future time - providing some evidence for the idea that crowdfunding is a potential substitute for the ven- ture capital model in seed financing, and a complement to venture capital in the following stages. The port- manteau test and the lag length criteria (streamlining the Box-Jenkins approach6) were used to decide the optimal length of our VAR model, even though at the end we limited our analysis to one lag because, for our purposes, we only want to get some insights about the impact of crowdfunding on venture capital and not to build the best predicting model. Finally, to summarize

the relationships built into our VAR model, we used Granger causality tests (GCT) and impulse response functions (IRF).

RESULTS

Which Relationship between Crowdfunding and Venture Capital?

A first insight into the implications of the crowd- funding model on traditional venture capital activity is provided by a graphical analysis. Exhibits 2, 3, and 4 depict monthly crowdfunding and venture capital

Exhibit 2

Monthly Crowdfunding Investments in the U.S.

4"l

-6 1 1 1 1 1 1 1 1 1 ■ 1 1 1 1 ■ ■ i ■ ■ i ■ ■ 1 1 ■ i ■ 1 1 ■■ 1 1 1 1 1 1 1 ■■ ļ ■ ■ 1 1 » 1 1 1 1 1 1 1 ■ 1 1 ■■ i » ■ 1 1 » i » i

limivi limivi nmivi nniivi nmivi nmiv

2009 2010 2011 2012 2013 2014

Exhibit 3 Monthly Venture Capital Investments in Seed Financing in the U.S.

ÖT

1 -

0-

-1 1 1 » ■ I ■ ■ I ■ 1 1 ■ ■ 1 1 ■ I ■ 1 1 1 1 1 1 ■ I ■ » I ■ ■ I » ■ 1 1 ■ I » 1 1 II 1 1 1 1 . ■ I ■ 1 1 ■ ■ 1 1 » I » » I ■ » I ■ »

ii m iv i n iii iv i ii in iv i ii ni iv i ninivi nmiv

2009 2010 2011 2012 2013 2014

Winter 2016 The Journal of Private Equity 9

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 4

Monthly Venture Capital Investments in After Seed Stages in the U.S.

g 5

6.5-

6.0-

5.5 i ■ 1 1 ■ » i ' » i ■ ■ i ■ ■ i ■ 1 1 ■ ■ i " i ■ ■ i ' * i ■ 1 1 ■ ■ i " i ■ ■ i " i ■ ' i ■ ■ i * ■ i ■■ i ■ ■ i " i "

ii in rv i ii iii iv i ii iii iv i ii iii iv i ii iii iv i ii m iv

2009 2010 2011 2012 2013 2014

investments in the United States, for the period June 2009 to December 2014.

The graphs suggest that, from the inception of the

crowdfunding model, amounts of seed capital provided by venture capitalists have been decreasing, while the capital provided in further rounds of financing has been slightly increasing. Of course, this may be owing in part to the recent financial crisis, which has rendered it more difficult for entrepreneurs seeking seed capital to obtain financing. But, as suggested by other authors (e.g., Agrawal, Catalini, and Goldfarb [2011]), we argue that crowdfunding plays an important role and that the shortage of capital engendered by the current global crisis is just one of the factors behind the success of the crowdfunding model. Therefore, to examine the role of crowdfunding as a potential substitute for rather a complement to venture capital investments, we start our analysis by regressing VC_SEED and VC_AFTER- SEED on CF. The results of this ordinary least squares (OLS) regression are shown in Exhibits 5 and 6.

We find the existence of a negative relationship between VC_SEED and CF, and a positive relation- ship between VC_AFTERSEED and CF. The results suggest that relevant changes in crowdfunding invest- ments somehow impact venture capital, negatively affecting seed financing and positively enhancing after seed investments.

In Exhibits 7 and 8, we report the cross-correlo- grams of our variables. Even in this case, the estimated coefficients are statistically significant, providing further

Exhibit 5

OLS Regression: Seed Stage

Dependent Variable: VCSEED Method: Least Squares Sample: 2009M06 2014M12 Included Observations: 67

Variable Coefficient Std. Error t-Statistic Prob.

C 3.9092*** 0.08841 44.2142 0.0000

CF -0.1390*** 0.03166 -4.3922 0.0000

/^-squared 0.228873 Mean dependent var. 3.866050 Adjusted R-squared 0.217009 S.D. dependent var. 0.812808 S.E. of regression 0.719227 Akaike info criterion 2.208117 Sum squared resid 33.62370 Schwarz criterion 2.273929 Log likelihood -71.97193 Hannan-Quinn criter. 2.234159 F-statistic 19.29221 Durbin-Watson stat. 1.361622

Prob (F-statistic) 0.000042

Note: ***Indicates statistical significance at the 1% level.

Exhibit 6

OLS Regression: After Seed Stages

Dependent Variable: VC_AFTERSEED Method: Least Squares Sample: 2009M06 2014M12 Included Observations: 67

Variable Coefficient Std. Error t-Statistic Prob.

C 7.3746*** 0.03799 194.099 0.0000

CF 0.0706*** 0.01360 5.19530 0.0000

^-squared 0.293411 Mean dependent var. 7.39650 Adjusted R-squared 0.282540 S.D. dependent var. 0.36488 S.E. of regression 0.309066 Akaike info criterion 0.51887 Sum squared resid 6.208900 Schwarz criterion 0.58468 Log likelihood -15.38214 Hannan-Quinn criter. 0.54491 F-statistic 26.99118 Durbin-Watson stat. 1.29062

Prob (F-statistic) 0.000002

Note: ***Indicates statistical significance at the 1% level.

confirmation of the evidence that emerged from the regression analysis. Cross-correlations in Exhibit 7 show that venture capital investments in seed stages are nega- tively correlated with crowdfunding activity. But the most interesting result lies in Exhibit 8, which suggests that an increase in crowdfunding investments today predicts a rise in venture capital financing in after seed rounds. This finding supports the idea that crowdfunding

potentially substitutes venture capital as a means to

10 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 7

Cross-Correlograms: Seed Stage

Sample: 2009M06 2014M12 Included Observations: 67

Correlations are asymptotically consistent approximations

Notes: The figure reports cross- correlograms for monthly crowdfunding and venture capital investments in the United States, for the period from June 2009 to September 2014. They capture the lead- lag relationships between crowdfunding and venture capital proceeds up to 24 months back /ahead. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter. VC_SEED is the natural logarithm of the monthly level of financing provided by venture capital firms as seed capital. Bars outside the interval indicate statistical significance at the 5% level.

raise seed capital and therefore enhances venture capital investments in further rounds of financing.

Exhibits 9 and 10 use a bivariate vector autore-

gressive model to represent the dynamic process that simultaneously describes crowdfunding and venture capital activity. Again, it is worth noticing that, for the purposes of this study, we are only interested in how crowdfunding predicts venture capital, and not in building the best VAR model that leaves zero autocorre- lation in the residuals. For this reason, we limit our anal-

ysis to one lag, but results would not change if we added

more lags in the model to capture some further autocor- relation left in the residuals. The second column of both

models reports an important finding of this study. The coefficient on CF (-1), which is statistically significant, indicates that a rise in crowdfunding investments pre- dicts lower venture capital investments in seed financing and, at the same time, higher venture capital investments in after seed stages. The Granger causality test in Exhibits

11 and 12 confirms that crowdfunding is a significant predictor of venture capital activity. This is further evi- dence that supports Hypothesis 1 of this work.

Winter 2016 The Journal of Private Equity 11

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 8

Cross-Correlograms: After Seed Stages

Sample: 2009M06 2014M12 Included Observations: 67

Correlations are asymptotically consistent approximations

Notes: The figure reports cross- correlograms for monthly crowdfunding and venture capital investments in the United States, for the period from June 2009 to September 2014. They capture the lead- lag relationships between crowdfunding and venture capital proceeds up to 24 months back/ahead. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter. VC_AFTERSEED is the natural logarithm of monthly amounts of capital provided by venture capital firms in stages other than seed financing. Bars outside the interval indicate statistical significance at the 5% level.

Finally, a key point about the first research ques- tion of this study is how an isolated shock to crowd- funding activity affects our forecasts of venture capital investments. In Exhibits 13 and 14, we report the impulse response functions associated with our vari- ables. By looking at the responses of VC_SEED and VC_AFTERSEED to CF, we find that an unanticipated increase in crowdfunding investments leads to negative forecast revisions for seed amounts provided by venture capitalists and raises the expectations about the capital financed by venture capital firms in the early stage, expansion, and later stages.

The Geographical Patterns

Many authors have documented the geographical clustering of venture capital investments (e.g., Chen et al.

[2009]; Kantor [2013]; Mollick [2013]). In particular, by analyzing data from 2010 to the first half of 2011, Kantor

[2013] noticed that only five regions accounted for roughly 76% of total venture capital investments, with 39% of total venture capital funding invested in Silicon Valley. Similarly, based on a dataset from 1975 to 2005, Chen et al. [2009] pointed out that approximately 50% of venture capital investments are made in companies

12 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

0 0.5417 0.5417 1 0.4748 0.5011

2 0.4862 0.4349 3 0.5051 0.4170

4 0.4259 0.3860

5 0.4010 0.3205

6 0.3620 0.3542

7 0.2942 0.3573

8 0.2679 0.2996

9 0.2367 0.3024

10 0.2020 0.2922

11 0.1673 0.2726

12 0.1658 0.3120

13 0.1280 0.3049

14 0.0931 0.2543

15 0.0885 0.2366

16 0.0481 0.2191

17 0.0028 0.1382

18 0.0063 0.1099

19 -0.0128 0.1525

20 -0.0304 0.0993

21 -0.0525 0.1469

22 -0.0453 0.1628

23 -0.0556 0.1638

24 -0.0700 0.1421

25 -0.0625 0.1470

26 -0.0759 0.1380

27 -0.0851 0.1336

28 -0.0833 0.1278

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 9

Vector Autoregressive Model: Seed Stage

Vector Autoregression Estimates Sample: 2009M06 2014M12 Included Observations: 67

Standard errors in ( ) & t-statistics in [ ]

CF VCJSEED

CF (-1) 0.9348*** -0.0734*** (0.03560) (0.03553) [26.2610] [-2.06459]

VCJSEED (-1) -0.0648 0.3203*** (0.12156) (0.12133) [-0.53279] [2.63978]

C 0.3958 2.6264***

(0.48213) (0.48121) [0.82095] [5.45787]

^-squared 0.935969 0.2556120 Adj. ^-squared 0.933936 0.2319810 Sum sq. resids 30.98901 30.869740 S.E. equation 0.701348 0.6999970 F-statistic 460.4471 10.816640

Log likelihood -68.7012 -68.57396 Akaike AIC 2.172764 2.1689080

Schwarz SC 2.272294 2.2684380

Mean dependent 0.395803 3.8440850 S.D. dependent 2.728670 0.7987490

Determinant resid 0.232705

covariance (dof adj.) Determinant resid 0.2 1 203 1

covariance

Log likelihood -1 36. 1 1 6 Akaike information criterion 4.306547

Schwarz criterion 4.505607

Notes: The figure presents the output of a one-lag vector autoregressive model estimated using Eviews. The sample consists of 64 monthly obser- vations for venture capital and crowdfunding investments in the United States, between June 2009 and December 20Î4. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter.

VC_SEED is the natural logarithm of the monthly level of financing provided by venture capital firms as seed capital. C represents the intercept of our model.

***Indicates statistical significance at the Î % level.

located in three main CS As: San Jose- San Francisco (CA), Boston (MA), and New York (NY).

Exhibits 15 and 16 present a distribution of the geography of venture capital and crowdfunding invest- ments from our dataset. More specifically, we show the top ten CS As by amount invested, grouping the remaining CSAs in the category Other. Approximately

Exhibit 10

Vector Autoregressive Model: After Seed Stages

Vector Autoregression Estimates Sample: 2009M06 2014M12 Included Observations: 67

Standard errors in ( ) & t-statistics in [ ]

CF (-1) 0.9729*** 0.0431*** (0.03663) (0.01587) [26.5615] [2.71837]

VC_AFTERSEED (-1) -0.4018 0.3247*** (0.27853) (0.12067) [-1.44277] 2.69143

C 3.1077 4.9901***

(2.05662) (0.89098) [1.51111] [5.60069]

R-squared 0.937737 0.338573 Adj. /¿-squared 0.935761 0.317575 Sum sq. resids 30.13302 5.655446 S.E. equation 0.691594 0.299615 F-statistic 474.4219 16.12429

Log likelihood -67.7768 -12.5678 Akaike AIC 2.144753 0.471750

Schwarz SC 2.244283 0.571280

Mean dependent 0.395803 7.403812 S.D. dependent 2.728670 0.362690

Determinant resid 0.04 1 1 79

covariance (dof adj.) Determinant resid 0.037521

covariance

Log likelihood -78.9654 Akaike information criterion 2.574708

Schwarz criterion 2.773768

Notes: The figure provides the output of a one-lag vector autoregressive model estimated using Eviews. The sample consists of 64 monthly obser- vations for venture capital and crowdfunding investments in the United States, between June 2009 and December 20Î4. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter.

VC_AFTERSEED is the natural logarithm of monthly amounts of capital provided by venture capital firms in stages other than seed financing. C represents the intercept of our model.

***Indicates statistical significance at the 1 % level.

70% of venture capital investments between 2009 and 2014 in the United States are concentrated in the top five regions, with roughly 40% invested in San Jose-San Francisco-Oakland (CA) CSA. Moving beyond the first five CSAs, 80% of the capital providejd by venture capi- talists has been invested in the top ten CSAs, and San Francisco, Boston, and New York represent the three

Winter 2016 The Journal of Private Equity 13

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 11

Granger Causality Test: Seed Stage

VAR Granger Causality/Block Exogeneity Wald Tests Sample: 2009M06 2014M12 Included Observations: 67

Excluded Chi-sq df Prob.

Dependent Variable: CF VCJSEED 0.2839 1 0.5942 All 0.2839 1 0.5942

Dependent Variable: VC_SEED CF 4.2625 1 0.0390

All

Notes: The figures present the results of a Granger causality test. We test the null hypothesis that all lags of one variable in the VAR model are not statistically different from zero. The sample consists of 64 monthly obser- vations for venture capital and crowdfunding investments in the United States , between June 2009 and September 2014. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter.

VC_SEED is the natural logarithm of the monthly level of financing provided by venture capital firms as seed capital.

Exhibit 12

Granger Causality Test: After Seed Stage

VAR Granger Causality/Block Exogeneity Wald Tests Sample: 2009M06 2014M12 Included Observations: 67

Excluded Chi-sq df Prob.

Dependent Variable: CF VC_AFTERSEED 2.0815 1 0.1491 All 2.0815 1 0.1491

Dependent Variable: V C_AFTERSEED CF 7.3895 1 0.0066

All

Notes: The figures present the results of a Granger causality test. We test the null hypothesis that all lags of one variable in the VAR model are not statistically different from zero. The sample consists of 64 monthly observations for venture capital and crowdfunding investments in the United States, between June 2009 and September 2014. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter. VC_AFTERSEED is the natural logarithm of monthly amounts of capital provided by venture capital firms in stages other than seed financing.

main centers. Only 20% of the investments are located in other areas, that is, 98 CSAs and 72 MSAs are not located in a CSA. These results clearly describe the geographical clustering of venture capital. Turning to crowdfunding, our findings based on

Kickstarter data are not that different. The first key evi-

dence arising from Exhibit 15 is that, among the top 10 CSAs by crowdfunding investments, 9 are exactly the same as in the case of venture capital. This finding would

be already sufficient to state that crowdfunding activity tends to be concentrated in the same main areas as ven-

ture capital. Of the $462 million invested in the United States between 2009 and 2014 through Kickstarter, approximately 55% is clustered in five main areas, with the city of San Francisco with the highest ranking in this case as well. The top 10 CSAs account for more than 70% of total investments, while the remaining 30% invested outside of the main centers is distributed across

103 CSAs and 88 MSAs not located in a CSA. San Fran-

cisco, Los Angeles, and New York represent the three main cities. With respect to venture capital, these are the main differences:

• Crowdfunding does seem to be slightly less con- centrated than venture capital funding, although the difference is not great in magnitude (70% of investments clustered in the top ten CSAs, against 80% for venture capital). This result is consistent with a previous analysis performed by Mollick [2013], based on locational Gini coefficients.

• While New York and San Francisco are among the top three centers for both fundraising models, Los Angeles and Boston represent the second city in the

top three ranking for crowdfunding and venture capital, respectively.

However, these differences are not large enough to conclude that crowdfunding and venture capital follow different geographical patterns and cluster in different areas. Conversely, our findings show that crowdfunding activity does not mitigate the geographical constraints to traditional fundraising methods and tend to cluster in the same main geographical areas as venture capital, although to a lesser extent.

DISCUSSION AND CONCLUSIONS

Prior to the development of crowdfunding, ven- ture capital firms acted as a key player in the institutional environment of technology entrepreneurship, as even alternative fundraising sources - such as angel investors - often served as first step toward the Subsequent pursu- ance of venture capitalists. However, in recent years, the

14 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 13

Impulse Response Functions: Seed Stage

Notes: The charts depict the impulse response functions of our variables. They capture the response of one variable to an isolated shock to the other variable over a 24-month time period. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter. VC_SEED is the natural logarithm of the monthly level of financing provided by venture capital firms as seed capital. The dashed lines give a 95% confidence interval.

rapid growth of crowdfunding as a means to raise capital for new ventures has triggered important changes in the

institutional context of entrepreneurship. Crowdfunding differs from venture capital in

many aspects. First of all, it allows millions of indi- viduals of uncertain background to provide funding for entrepreneurial projects and business ideas, while ven- ture capital firms are typically run by a few people with a strong expertise in a given business. Second, it oper- ates through the Internet, while venture capital strongly relies on a network of offices located in the main centers

in order to identify and support start-ups. Finally, unlike venture capitalists, it is not necessarily rewarded with equity participation or monitoring rights in the financed company, but it does allow several forms of reward.

Nevertheless, since its inception in 2009, crowd- funding has proved to be a reliable tool for entrepreneurs to seek financing, and thus, it has become a meaningful alternative to business angels and venture capital firms as a source of finance for new ventures and high- growth start-ups. There have been many cases of suc- cessful startups financed, totally or partially, through Kickstarter and other online platforms.

The reasons behind the success of this new fund-

raising model lie in the fact that crowdfunding boasts a set of features that make it more attractive than venture

capital for seeking seed capital. First, it is relatively easy for entrepreneurs to raise capital and for funders to pro- vide money with crowdfunding: pounders just need to provide a description of their project and can rely on millions of potential investors (whereas, in the case of

Winter 2016 The Journal of Private Equity 15

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 14

Impulse Response Functions: After Seed Stage

Notes: The charts depict the impulse response functions of our variables. They capture the response of one variable to an isolated shock to the other variable over a 24-month time period. CF is the natural logarithm of the monthly level of crowdfunding proceeds from Kickstarter. VC_AFTERSEED is the natural logarithm of monthly amounts of capital provided by venture capital firms in stages other than seed financing. The dashed lines give a 95% confidence interval.

venture capital, some more formal and structured reports

are required, such as a business plan, and founders can rely only on a few venture capital firms located in the area), and funders can easily pledge whatever amount they want through their credit cards and from any region of the world (although, on many platforms such as Kickstarter, U.S. citizenship is required to finance U.S. projects). Therefore, as argued by many scholars (e.g., Agrawal, Catalini, and Goldfarb [2011]), crowd- funding relaxes geographic constraints among funders and increases the average distance between entrepre- neurs and investors.

Second, by allowing entrepreneurs to reach net- works of millions of individuals more easily, crowd- funding can be used to test or demonstrate demand for a given product, because it asks for financing directly of

those people that are potential consumers. Some authors (e.g., Lambert and Schwienbacher [2010]; Mollick [2013]) have described the crowdfunding model as a means to test new products or even run marketing campaigns and expand awareness of the initiative through social media.

Third, crowdfunding can be used not only to get public attention but also to receive feedback on a given product/service from the crowd, in order to develop the project based on the suggestions arising directly from potential customers (Belleflamme, Lambert, and Schwienbacher [2014]).

Fourth, crowdfunding protects start-ups against the implications of venture capital selection. Indeed, when venture capital firms do not invest in a new venture, it can

have far deeper consequences for start-ups than simply

16 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

Exhibit 15

Geography of Venture Capital Investments

Number of Amount Invested Share of Total

CSA Deals ($mln) Amount Invested

San Jose-San Francisco, CA 5.248 47.136 0.43

Boston, MA 2.041 11.501 0.10

New York, NY 2.011 11.363 0.10

Los Angeles, CA 1.068 6.355 0.06 Washington, DC 783 3.808 0.03 Seattle, WA 533 2.671 0.02

Chicago, IL 368 2.257 0.02 Dallas, TX 238 2.032 0.02 Denver, CO 440 1.839 0.02

Atlanta, GA 282 1.492 0.01 Other 5.492 20.691 0.19

Total 18.504 111.145 1

Notes: The table describes the geographical clustering of venture capital investments and ranks the top ten CSAs in terms of total amount invested. It also reports information about the number of venture capital deals. The sample consists of 18,504 venture capital deals in the United States, between June 2009 and September 2014, and accounts for approximately $111 billion of capital invested. Geographic locations are assigned at the Combined Statistical Area (CSA) level. In cases where a city is not located in a CSA, we assign the deal to the appropriate Metropolitan Statistical Area (MSA).

Of note: (1) the venture capital deal is intended to be the match of the venture asking for financing with the bunch of venture capital investors providing capital in a given round of financing; (2) the amount invested is defined as the total amount of money invested by venture capital firms in the CSA, regard- less of the stage of financing; (3) share of total amount invested equals the percentage of total investments located in the CSA; (4) Other includes 98 CSAs and 12 MS As not located in a CSA.

Exhibit 16

Geography of Crowdfunding Projects

Number of Amount Pledged Share of Total CSA Projects ($ min) Amount Pledged

San Jose-San Francisco, CA 852 94 0.20

Los Angeles, CA 796 66 0.14 New York, NY 1,011 45 0.10 Seattle, WA 280 30 0.07 Portland, OR 277 23 0.05 Boston, MA 397 22 0.05 Dallas, TX 177 15 0.03

Atlanta, GA 157 15 0.03

Chicago, IL 376 15 0.03 Washington, DC 261 15 0.03 Other 518 123 0.27

Total

Notes: The table describes the geographical clustering of crowdfunding investments and ranks the top ten CSAs in terms of total amount pledged. It also reports information about the number of projects. The sample consists of 5,102 crowdfunding projects (with fundraising goal above $5,000) in the United States, between June 2009 and September 2014, and accounts for approximately $462 million of capital invested. Geographic locations are assigned at the Combined Statistical Area (CSA) level. In cases where a city is not located in a CSA, we assign the deal to the appropriate Metropolitan Statistical Area (MSA). » Of note: (1) the amount pledged is defined as the level of crowdfunding proceeds raised in the CSA; (2) shąte of total amount pledged equals the percentage of total crowdfunding investments located in the CSA ; (3) Other includes 103 CSAs and 88 MSAs not located in a CSA.

Winter 2016 The Journal of Private Equity 17

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

depriving them of capital, because of the reputational effects of venture capital endorsement (Mollick [2013]). As suggested by Ferrary and Granovetter [2009], "by selecting start-ups, venture capitalists implicitly prevent the other agents in the institutional network of innova- tion from collaborating with new ventures that do not get VC funding," and it sometimes happens that poten- tial valuable innovations do not reach customers simply because they did not get venture capital financing to get connected to the network.

Finally, unlike venture capital, entrepreneurs are typically more likely to maintain full control over the project with crowdfunding (Gerber, Hui, and Kuo [2011]). Legitimately, it is not surprising that concerns have been raised by critics about the fact that the quality

of entrepreneurial projects or the viability of new ven- tures may not be as clear or influential to funders in crowdfunding settings, compared with venture capital firms that have strong expertise and business background

(Bogost [2012]). However, Mollick [2013] found that signals of project quality typically used by venture capitalist to select portfolio companies are also used by crowdfunders, so that good projects get financed while lower-quality projects receive few to no backers.

Overall, we argue that, in the entrepreneurial finance environment, crowdfunding represents a mean- ingful substitute for traditional venture capital activity as a way to raise seed capital for entrepreneurs at the very beginning of their entrepreneurial initiative. This argument leads to the consistent belief that, over the next years, crowdfunding might have a strong impact on traditional fundraising models by further decreasing the amount of money invested by venture capitalists as seed capital. Therefore, as expected, the results of our analysis are totally consistent with the idea that crowdfunding is a worthwhile alternative to venture capital in seed financing. Indeed, our findings (broadly described in the previous section) clearly show the exis- tence of a negative relationship between crowdfunding and venture capital investments in seed stages, strongly suggesting that a further expansion of the crowdfunding

model may negatively affect the provision of seed capital by more traditional methods.

From this perspective, the results of our analysis demonstrate that crowdfunding has the ability to scale down the venture capital phenomenon, serving as a first step toward further rounds of financing provided by venture capital firms. The fact that crowdfunding might

further evolve as the main actor in the provision of seed capital is a very interesting (and, in a way, even sur- prising) finding because it sheds some light on another important topic: the role of institutional changes in the entrepreneurial environment.

Some authors (e.g., Mollick [2013]) have already discussed the potential implications of the radical changes brought by the crowdfunding phenomenon in the institutional context of entrepreneurship. Of course, we must be very careful in interpreting the results of our

study, since the shortage of capital engendered by the financial crisis in recent years may have played a key role in leading the growth of crowdfunding.

While crowdfunding appears as a challenger to venture capital in the provision of seed capital, it is less so for after seed stages. Conversely, and in line with previous

evidence (e.g., Mollick [2013]), successful crowdfunding initiatives seem to provide a signal to venture capital firms of potential good long-term investments, even- tually leading to additional rounds of future financing. This is because, while crowdfunding has some charac- teristics that make it attractive to entrepreneurs seeking

seed capital, it is not as appropriate as venture capital firms in supporting portfolio companies in after seed stages, such as in further expansion initiatives.

More specifically, unlike crowdfunding, venture capital boasts a few distinctive properties. As a matter of fact, venture capitalists do not limit their activity to the provision of funding. They have strong expertise and deep business knowledge, and play an active role granting portfolio firms monitoring, governance and even reputational effects from venture capital endorse- ment (Stuart, Hoang, and Hybels [1999]; Hsu [2004]). Venture capital firms have experience in running a company and thus participate in the active manage- ment of the venture, providing managerial support in sales, accounting, marketing, strategy, distribution, and any other field (Schwienbacher and Larralde [2010]). In other words, in the case of venture capital, portfolio companies not only receive financing but also help pro- fessionalizing and connecting to vital resources (Baum and Silverman [2004]).

Moreover, the amount of funding provided by venture capital firms can be much higher with respect to crowdfunding. This is because, when seeking capital through the Internet, entrepreneurs rely only on small contributions from millions of individuals, and it is

very difficult to raise huge amounts of money. Just to

18 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

provide some numbers, out of the 5,102 Kickstarter proj- ects taken in consideration in our analysis, individual projects that received more than $1 million represents less than 10% (and this percentage is going to decrease significantly if we add all those initiatives excluded to make our dataset as consistent as possible with venture capital investments).

Conversely, venture capital can grant portfolio firms a greater amount of financing and also access to additional resources, making introductions that help lead to more business or funding down the road (Kantor [2013]). From this perspective, it is reasonable to state that venture capital is more appropriate than crowd- funding in after seed stages, where the amount of capital

to be provided (for instance, to support the company in major expansion through the purchase of new physical plants) is typically larger than that required at the very beginning of an entrepreneurial initiative.

For the reasons explained here, in our study, we argue that crowdfunding mainly serves as a first step for

the provision of seed capital to start-ups, signaling new ventures as potential good long-term investments and enhancing venture capital investments in further rounds of financing. In other words, we discuss crowdfunding as a meaningful alternative to traditional fundraising models for the provision of seed financing, while when it

comes to after seed stages, it acts as a complement (and not

a substitute anymore) to venture capital activity, because in these stages, ventures need additional resources, services, and competencies that only venture capitalists can bring.

Again, as expected, the results of our analysis strongly support our arguments. Indeed, the empirical findings described earlier provide evidence about the existence of a positive relationship between crowd- funding and venture capital investments in after seed stages, shedding light on how in the past crowdfunding has led to additional capital provided by venture capitalists in further rounds of financing. This sug- gests that a further expansion of crowdfunding might positively affect venture capital investments in after seed

financing. Crowdfunding has thus become an important player in the institutional system of technology entre - preneurship and represents a critical factor in affecting the dynamics of traditional fundraising efforts.

Many authors have documented that venture capital investments are highly concentrated in just few areas, where both pools of successful entrepreneurs

and venture capital firms are located (e.g., Chen et al. [2009]; Shane and Cable [2002]). The main reasons behind this clustering lie in the benefits of spillover and in the requirement of face-to-face interaction. Venture capitalists provide not only capital but also management and monitoring services, at a cost that is considerably lower when firms are within a driving distance from the

venture capital office; for these reasons, entrepreneurs may prefer direct access to portfolio firms.

In fact, the main consequence of the strong con- centration of venture capital activity in just few areas is that many start-ups located in less prevalent venture capital areas go unfunded. For this reason, many wonder whether crowdfunding mitigates the geographical con- straints found in traditional fundraising models, giving access to financing to all those start-ups not selected by venture capital firms because not located in venture capital geographical clusters.

We find that, in entrepreneurial financing, crowd- funding investments are not only clustered in some geographical areas (though they are slightly less con- centrated than venture capital ones),7 but also clustered in the same main cities as venture capital activity. There- fore, the empirical findings of our analysis suggest that crowdfunding has the same geographical determinants as venture capital. Interestingly, crowdfunding gets more concentrated in those areas with higher concentration of venture capital firms, demonstrating that, from the founders' perspective, it does not overcome the geographical limit represented by the proximity of venture capital offices to portfolio companies. Thus, we argue that this novel fundraising method does not relax geographical constraints to traditional venture capital activity among entrepreneurs.

ENDNOTES

1^Writh successful projects, we reference those projects that raised at least their initial target. Indeed, Kickstarter fol- lows an all or nothing model, so money pledged is collected only if the fundraising goal is reached.

2The suggestion to restrict the sample to crowdfunding initiatives with a fundraising goal above $5,000 was first given by Mollick [2013], in order to build a database that is as comparable as possible with venture capital investments.

3For instance, if the pledged amount is $100,000 and the percentage of funding 200%, the initial goal is $50,000. Obviously, since we take into consideration only successful

Winter 2016 The Journal of Private Equity 19

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms

projects, the percentage of funding of each initiative in our dataset will be at least 100%.

For instance, San Diego, CA. Our use of CSA as the main unit of location is driven

by the narrower definition of MSAs. For instance, New York City and Greenwich, CT, belong to the same CSA, but they are located in different MSAs.

6In time series analysis, the Box-Jenkins approach is a generalized method that applies to autoregressive moving average models to find the best fit of a time series model to past value of a time series.

7This result is consistent with a previous research car- ried out by Mollick [2013], based on locational Gini coef- ficients. Therefore our analysis adds further evidence to the existing literature.

REFERENCES

Agrawal, A., C. Catalini, and A. Goldfarb. "The Geography of Crowdfunding." Working paper, University of Toronto, 2011.

Baum, J., and B. Silverman. "Picking Winners or Building Them? Alliance, Intellectual, and Human Capital as Selection Criteria in Venture Financing and Performance of Biotech- nology Startups." Journal of Business Venturing , Vol. 19, No. 3 (2004), pp. 411-436.

Belleflamme, P., T. Lambert, and A. Schwienbacher.

"Crowdfunding: Tapping the Right Crowd "Journal of Busi- ness Venturing , Vol. 29, No. 5 (2014), pp. 585-609.

Bogost, I. "Kickstarter: Crowdfunding Platform or Reality Show?" Fast Company, July 18, 2012.

Chen, H., P. Gompers, A. Kovner, and J. Lerner. "Buy Local? The Geography of Venture Capital." Journal of Urban Economics, Vol. 67, No. 1 (2009), pp. 90-102.

Ferrary, M., and M. Granovetter. "The Role of Venture Cap- ital Firms in Silicon Valley's Complex Innovation Network." Economy and Society, Vol. 38, No. 2 (2009), pp. 326-359.

Gerber, E. M., J. S. Hui, and P.-Y. Kuo. "Why People Are Motivated to Post and Fund Projects on Crowdfunding Plat- forms." Working paper, Northwestern University, Creative Action Lab, 2011.

Hsu, D. "What Do Entrepreneurs Pay for Venture Capital Affiliation?" The Journal of Finance, Vol. 59, No. 4 (2004), pp. 1805-1844.

Kantor, R. "Why Venture Capital Will Not Be Crowded Out by Crowdfunding." Working paper, Chicago-Kent College of Law-Illinois Institute of Technology, 2013.

Lambert, T., and A. Schwienbacher. "An Empirical Analysis of Crowdfunding." March 2010: http://wenku.baidu.com/ view/2de46229b4daa58da0114ac8.

Mollick, E. "Swept Away by the Crowd? Crowdfunding, Venture Capital and the Selection of Entrepreneurs." SSRN Working Paper #2239204, March 2013: https://ssrn.com/ abstrae t= 2239204.

Schwienbacher, A., andB. Larralde. "Crowdfunding of Small Entrepreneurial Ventures." SSRN Working Paper #1699183, September 2010: https ://ssrn.com/abstract= 1699 183.

Shane, S., and D. Cable. "Network Ties, Reputation, and the Financing of New Ventures." Management Science, Vol. 48, No. 3 (2002), pp. 364-381.

Stuart, T., H. Hoang, and R. Hybels. "Interorganizational Endorsements and the Performance of Entrepreneurial Ven- tures." Administrative Science Quarterly, Vol. 44, No. 2 (1999), pp. 315-349.

To order reprints of this article, please contact Dewey Palmieri at [email protected] or 212-224-3675.

20 Crowdfunding and Venture Capital: Substitutes or Complements? Winter 2016

This content downloaded from �������������194.34.78.248 on Sat, 19 Jun 2021 12:29:22 UTC�������������

All use subject to https://about.jstor.org/terms