Herding in Equity Crowdfunding

Transcription

1 Herding in Equity Crowdfunding Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Research in Behavioral Finance Conference, Amsterdam 2018 Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 1 / 23

2 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

3 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

4 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Duration of the campaign typically 60 days. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

5 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Duration of the campaign typically 60 days. Goal: funds to be raised during the campaign. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

6 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Duration of the campaign typically 60 days. Goal: funds to be raised during the campaign. Backers pledge (any positive amount). Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

7 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Duration of the campaign typically 60 days. Goal: funds to be raised during the campaign. Backers pledge (any positive amount). All-or-nothing: If by the end of the campaign the goal is not reached, all backers get their money back. If by end of the campaign, the goal is reached, the campaign succeeds, funds are invested and backers get shares of the new firm. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

8 Equity crowdfunding: key elements Entrepreneurs issue shares through the web to finance their early stage projects: Fundraising campaign runs on an internet platform. Description of the investment project. Duration of the campaign typically 60 days. Goal: funds to be raised during the campaign. Backers pledge (any positive amount). All-or-nothing: If by the end of the campaign the goal is not reached, all backers get their money back. If by end of the campaign, the goal is reached, the campaign succeeds, funds are invested and backers get shares of the new firm. If the campaign succeeds, entrepreneur can extend the campaign and sell a larger fraction of the firm. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 2 / 23

9 Equity crowdfunding on Seedrs Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 3 / 23

10 Pros and cons Facts: Equity crowdfunding is increasingly popular (in 2015 about 35% of all seed-stage investment deals in UK used crowdfunding) Project information is very limited Not possible to do deep due diligence Campaign open to unsophisticated investors Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 4 / 23

11 Pros and cons Facts: Equity crowdfunding is increasingly popular (in 2015 about 35% of all seed-stage investment deals in UK used crowdfunding) Project information is very limited Not possible to do deep due diligence Campaign open to unsophisticated investors Pros: Exploit the wisdom of the crowd for detecting profitable projects without recurring to experts Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 4 / 23

12 Pros and cons Facts: Equity crowdfunding is increasingly popular (in 2015 about 35% of all seed-stage investment deals in UK used crowdfunding) Project information is very limited Not possible to do deep due diligence Campaign open to unsophisticated investors Pros: Exploit the wisdom of the crowd for detecting profitable projects without recurring to experts Cons: The sequential nature of a crowdfunding may induce Pledge herding: investors pledge, or pledge more, because previous investors pledged Abstention herding: investors not pledging because previous investors did not pledge Information cascades: Investors actions provide no information about the project s profitability (the whole crowd is not wiser than the first few investors) Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 4 / 23

13 Research questions Are crowdfunding campaigns gathering the wisdom of the crowd about new business ideas, or are they inducing investors, and particularly unsophisticated investors, to herd? Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 5 / 23

14 Research questions Are crowdfunding campaigns gathering the wisdom of the crowd about new business ideas, or are they inducing investors, and particularly unsophisticated investors, to herd? Theory Is (rational) herding theoretically possible in equity crowdfunding? If yes, what type of herding? Which type of herding generates information cascades? Empirics Do real life investors engage in rational herding? What type of investor is more likely to herd? Do information cascades occur? Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 5 / 23

15 Paper main findings Both pledge and abstention rational herding are theoretically possible, but only abstention herding can generate an information cascade. Actual backers do engage in rational herding and in general their behavior is in line with our theory predictions. Using IV we can exclude alternative models: naïve herding, independent investments, exogenously correlated investments. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 6 / 23

16 Formal of plengding in equity crowdfunding Project of unknown quality: good ( NPV > 0) or bad (NPV < 0) Log-utility backers with wealth W : Arrive following a Poisson process with intensity 1 Upon arrival a backer decides whether and how much to pledge. Private information: Private conditionally i.i.d. signals θ {g, b, u} Informed backers: P(θ = g good project) = P(θ = b bad project) = λq (0, 1) Uninformed backers P(θ = u good project) = P(θ = u bad project) = 1 λ Public information: h t: history of past strictly positive pledges until time t. Public belief: π t := P(good project h t) Informed backer s belief π θ t := P(good project h t, θ) All or Nothing: Pledge are actually invested iff by the end of the campaign total pledge exceed the pre-announced goal G. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 7 / 23

17 Some definitions Definition A investor engages in pledge herding if the amount he pledges is increasing in the amount of past pledges A investor engages in abstention herding if he he does not pledges after seeing no pledges for long enough The campaign is in an information cascade at time t if the pledging history after t provides no information about the project quality: t t, P(π t = π t ) = 1 Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 8 / 23

18 Equlibrium pledging behavior (qualitative descrpition) A backer arriving at time t with signal θ invests only if πt θ is large enough and the amount he/she invests is increasing in πt θ and W. The public belief reaction to a pledge is increasing in the pledge size. The public belief decreases in the time between two pledges. Information cascade: A pledge size always has some information content: no pledging information cascade. There is π > 0 such that as soon as π t < π, no backer pledges an abstention information cascade occurs. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 9 / 23

19 Empirical Implications 1-2 A pledge size and the probability of observing a new pledge should be increasing in the size of the most recent pledge; decrease in the time elapsed since the most recent pledge. 3 Uninformed backers react more to prior pledges than informed backers. 4 The probability of success of a campaign is highly dependent on early campaign dynamics: increasing with the number of backers and the amounts invested during the first days of a campaign. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 10 / 23

20 Data Universe of campaigns and investments from SEEDRS, one of the 2 leading UK equity crowdfunding platforms Detailed data on backers and their individual pledges made to each campaign by the second Period: October March 2016 Pledges: 69,699 Campaigns: 710 Investors: 22,615 Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 11 / 23

21 Descriptive statistics Campaigns Pre-money valuation (median) 5,031,415 Equity offered (median) 7.66 Campaign goal (median) 387,334 % Raised # investors # Pledges Type of Investor (Share) Authorized 0.79 High net worth 0.14 Sophisticated 0.07 % Recurrent investors 0.73 Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 12 / 23

22 Estimation Workhorse: Distributed Investment Lags log I n,c = β 0 + K β k log I n k,c + αw n,c + γz n,c + η c + ɛ nc k=1 I n,c : amount pledged by the n th investor after the start of a campaign c and k is the lag to the pledge by the n k th investor positive values of β k will indicate positively correlated pledges over time W n,c : dummy variables indicating the type of investor (e.g. sophisticated or unsophisticated ) Z n,c : campaign and time-varying variables capturing the information available to investor n at time of decision. (e.g. cumulative amount of funding, number of days since the campaign started, number of investors) η c : Campaign fixed effects allows us not to care about the number of entrepreneurs, the pre-money valuation, whether there was a video or not, etc. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 13 / 23

23 Common shocks problem and Instrumental variable Correlation across pledges could result from the arrival exogenous information observed by the backers but not by the econometrician. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 14 / 23

24 Common shocks problem and Instrumental variable Correlation across pledges could result from the arrival exogenous information observed by the backers but not by the econometrician. Instrumental variables approach: information on the characteristics of investors that is not revealed to potential follow-on investors but known to the researcher used to instrument lagged pledges Three instruments of the endog. lagged investment I n k,c : Amount of money returned to the backer if the last campaign he/she supported failed. Number of pledges made by the investor across all previous campaigns before he/she made that n k th investment Maximum amount pledged across all previous campaigns on SEEDRS by the investor making the lagged pledge before he/she made that n k th investment, zero if no prior investment Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 14 / 23

25 Prediction 1: Reactions to the most recent pledge A pledge size should be increasing in the size of the most recent pledge; decrease in the time elapsed since the most recent pledge. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 15 / 23

26 Prediction 1: Reactions to the most recent pledge A pledge size should be increasing in the size of the most recent pledge; decrease in the time elapsed since the most recent pledge. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 16 / 23

27 Prediction 2: Probability of a pledge The probability of observing a new pledge should be increasing in the size of the most recent pledge; decrease in the time elapsed since the most recent pledge. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 17 / 23

28 Prediction 3: Uninformed backers react more to prior pledges Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 18 / 23

29 Prediction 4: What happens early is very important The probability of success of a campaign is highly dependent on early campaign dynamics: increasing with the number of backers and the amounts invested during the first days of a campaign. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 19 / 23

30 Conclusion Equity crowdfunding popularity is rapidly increasing both from the entrepreneurs and investors perspectives. Rational herding in equity crowdfunding is both theoretically possible and empirically observed in our equity crowdfunding data. The herding exists but concerns about lemmings type pledging information cascades appear unwarranted. Abstention cascade can occur theoretical and do occur in reality. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 20 / 23

31 Realted Literature Rational herding theories: Banerjee (1992); Welch, 1992; Bikhchandani et al., (1992); Smith and Sorensen, (2000); Horner and Herrera, (2013), Avery and Zemsky, (1998); Park and Sabourian, (2011),... Empirical analysis of herding in crowdfunding: Zhang and Liu (2012) and Bursztyn et al. (2014),... Crowdfunding theories: Belleflamme et al. (2014); Ellman and Hurkens (2016); Chemla and Tinn (2016); Strausz (2017); Cong and Xiao (2017),... Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 21 / 23

32 Equlibrium pledging behavior Proposition Once the goal has been reached 1 A type θ backers arriving at time t pledges { } απ θ ˆσ(θ, π t ) = max t 1 α 1 W, 0 (1) 2 The public belief evolves according to the following rule: 1 The jump in the public public belief resulting from a pledge is increasing in the pledge size. 2 If between t and t > t no pledge is observed then π t π t. 3 Information cascade: There is π > 0 such that as soon as π t < π, no backer pledges and for all t > t, π t = π t. That is, an abstention information cascade occurs. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 22 / 23

33 Prediction 1: Reactions to the most recent pledge A pledge size should be increasing in the size of the most recent pledge; decrease in the time elapsed since the most recent pledge. Thoams Åstebro, Manuel Fernàndez, Stefano Lovo, Nir Vulkan Herding in Equity Crowdfunding 23 / 23