Initial Coin Offerings, Information Disclosure, and Fraud Cases

Transcription

1 Initial Coin Offerings, Information Disclosure, and Fraud Cases Armin Schwienbacher (joint work with L. Hornuf from the Univ. Bremen) AMF/Paris June 7, 2018

2 INTRODUCTION Many firms have used the opportunities of an ICO to raise capital. Digitalization: ECF a few years ago, now ICOs First fraud cases have arisen. They will affect the reputation of the whole market. Different regulatory measures taken internationally Find the right balance between costs and benefits of regulation Many similarities with equity crowdfunding Our study is a first step is to understand the different forms of fraud and chances to detect fraud early on. This is crucial for the proper design of effective regulation, and ensure a sound development of the ICO markets.

3 AN EXAMPLE OF ICO SCAM: THE BENEBIT Source:

4 AN EXAMPLE OF ICO SCAM: THE BENEBIT Source:

5 METHODOLOGY Our method is more conservative to ensure reliability in the cases we classify as fraud, at the expense of possibly missing some fraudulent issuers out. We adapt the method developed by Cumming et al. (2017) for crowdfunding: Extensive search on Factiva and Google for the following four words: 1) ICO, 2) Initial Coin Offering, 3) Crowdsale, as well as 4) ICO together with SEC and combined them with the word fraud and different synonyms for fraud such as cheat, untruth, deception, scam, dishonesty, trickery, double-dealing, unlawful. To capture conceivable fraud cases as well, we also searched for the four words in combination with a set of extra keywords. All the identified cases are finally examined by the authors and one research assistant to decide which ICOs qualify as "fraud". These lead to suspected fraud cases, where 90% are currently investigated by a national regulator.

6 CALCULATED INDICATORS Availability of a hard cap Writing style: Average sentence length, number of words, Flesch Reading Ease Index, Flesch-Kincaid Grade Level, percentage passive sentences Availability of specific information: Allocation of funds (pre-sale, ICO, reserve, team, bounty, advisors, marketing, rewards, early contributors, others) Founders (numbers), privacy info, legal info, country, typology Content and structure of white paper: Figures, charts, tables, pictures, pictures on humans, TOC, timeline, reference to global player

7 ICO FRAUD TYPES (N=40) Regular Fraud / Ponzi Scheme Securities Fraud Exit Fraud Pretended Hack Phishing Attack Pump and Dump Scheme

8 TYPES OF ICO FRAUDS 1) Regular fraud/ponzi scheme: substantial misstatements about the nature of the business and the founder team, or even a Ponzi scheme 2) Securities fraud: subject to investigations by an national securities regulator, because they did not register the ICO under the respective securities law but solicited the general. In this case, the funds have not necessarily been misused or expropriated. 3) Exit fraud: strong indication that project owners misappropriated funds, over even disappeared immediately after the ICO 4) Pretended hack: issuer argues there was a hack and the money got stolen. 5) Phishing attack: fraudster presumably not the founder of the ICO pretended to be member of the founder team and advised investors to send their money to a fraudulent Bitcoin wallet. 6) Pump and dump scheme: release of an important negative news after the ICO, leading to major drop of value and thus losses. Whether planned or not is often not clear in most of the cases.

9 MAIN SAMPLE STATISTICS Our final sample contains 503 ICOs. Average amount raised of USD 11 million and a median of USD 3.6 million. Differences in informational variables not significant between ICOs with higher/smaller amounts. Differences may appear in the quality of the information. Pre-sales as a signal: An issuer that sold 10% of the token during the presale has an 18% greater probability to end up raising an amount above the median sample value. Founder information, for example the number of founders involved in the project, is included in only two-thirds of the whitepapers. Information on how much was ultimately raised could be retrieved in 55% of the cases only, despite extensive and systematic search efforts.

10 MAIN FINDINGS ON FRAUD (1/3) Occurrence of suspected fraud cases: 2-3% of the cases (depending on reference sample used) this percentage may become higher as most ICOs are too recent to classify them as suspected fraud But up to 16% when weighted by the amount raised ICOs that are suspected to be scams are by far larger in terms of funds raised. Sample differences: USD 74 million vs. USD 10 million The positive relationship confirmed in a multivariate regression setting Fraud cases are less likely to have sold coins/tokens in a presale or have early contributors prior to formal issuance.

11 MAIN FINDINGS ON FRAUD (2/3) No meaningful difference in degree of information disclosure in their whitepaper except for detailed information on the allocation of tokens that is significantly less often provided in the fraudulent cases. There may, however, be differences in the quality of information not was not captured in our study This requires similar to venture capital investments good understanding about the ICO process and due diligence of risky projects by investors.

12 MAIN FINDINGS ON FRAUD (3/3) Besides the cases considered in our sample, we have identified several out-of-sample fraud cases that show different characteristics. The reason that they are out of our original sample is that they raised capital via an ICO only very recently. Others are not necessarily globally advertised because not in written in English. The largest fraudulent ICO we could identify runs under the name Modern Tech Loss for backers of USD 660 million through exit scam by Vietnamese company Modern Tech has encouraged people to invest in two new crypto-currencies Ifan and Pincoin promising a profit rate of 48 percent a month. Unlike many ICO, Modern Tech held local conferences.

13 MAIN CONCLUSIONS While text content helps detect fraud in crowdfunding, it seems not to be the case for ICOs. The quality of information however may remain crucial, something investors may find difficult to see. This casts doubt on proposed solutions by some websites. Investors may not be able to collect more information nor do the needed checks for information reliability. Need for regulators to step in.

14 TAKE-AWAYS Similarities with equity crowdfunding in terms of development of practices The current discussion on how to deal with ICOs resembles in many ways the one we had on equity crowdfunding. Adopt similar regulatory approaches? Emergence of platforms? Or existing platforms entering the asset class when regulation becomes clearer?

15 PROF. ARMIN SCHWIENBACHER SKEMA BUSINESS SCHOOL DEPARTMENT OF FINANCE AND ACCOUNTING (LILLE CAMPUS) AVENUE WILLY BRANDT EURALILLE (FRANCE) ARMIN.SCHWIENBACHER@SKEMA.EDU