New initiatives and private investors are showing a growing interest in innovative forms of fundraising. ICO stands for Initial Coin Offering and represents an innovation in entrepreneurial finance. However, no study has ever developed a taxonomy of academic and non-academic discourse related to this type of innovative financial instrument. Moreover, no other academic article has dealt with relating this phenomenon with the sports sector. This thesis aims to fill this gap by developing a taxonomy to investigate and classify documents discussing the phenomenon of initial coin supply and at the same time study how linguistic content of white paper can reveal the main topics of each ICO. We explored, furthermore, a possible correlation between the positive emotion and amount of money raised. This study is developed using a mixed methodology. The first phase of the research protocol concerns the definition and description of the dataset. In the second phase we adopted the taxonomy process developed by Nickerson. The aim of this work is to first develop a taxonomy with a set of dimensions each consisting of a set of characteristics describing objects in a specific study, and then compare these results with an empirical analysis as a result of Meaning Extraction Method. We identified “Research topics” as the set of dimensions to be analyzed. It comprises eight dimensions: field of investigation, focus, actors, token type, extra topic, research problem, ICO stage, blockchain. We assigned a single value to each dimension during the taxonomy process.

The LIWC software (Meaning Extraction Method) is used again in the final section of the work to compare the results of the text analysis of the white papers with the results of the taxonomy.

Chapter 1: The ICOs Phenomenon: Initial Coin Offering

1.1 General framework

Technological developments and the growing interest in cryptocurrencies have supported the emergence of new forms of investment and financing, especially ICOs.

Startups and private investors are showing a growing interest in innovative forms of fundraising. In addition, economies are becoming more digital than ever and local businesses are going international urging the need for a borderless and efficient flow of capital (Rrustemi & Tuchschmid, 2020).

ICO is short for Initial Coin Offering and represents an innovation in the field of entrepreneurial finance (Fish, 2019; Block, Colombo, Cumming, & Vismara, 2018).

The easiest way to define an ICO is that it can be thought of as a funding activity that allows online projects and startups to raise the necessary funds with the backing of venture capitalists

The definition bears many similarities to the term crowdfunding campaign: in crowdfunding, an entrepreneur raises outside funds from a large audience (the crowd), with each individual contributing a very small amount, rather than recruiting a small group of experienced investors (Belleflamme, Lambert and Schwienbacher, 2012).

Initial coin offerings, like a crowdfund, usually take place in the early stages of a technology project. The difference is that a crowdfund is often a donation, while an ICO consists of investors who want to see a return. Some academics define ICOs as crowd sales, distinguishing them from crowdfunding.

At the same time, an initial money offering, as the name suggests, is the process by which a company sells its stock to the public, similar to ordinary IPOs or regular stock IPOs.

The public token sale, colloquially known as an initial coin offering, is a powerful new tool to create decentralized communities, kickstart network effects, incentivize participants, provide investors with faster liquidity, and accumulate capital for creators ((Batiz-Benet, Clayburgh and Santori, 2017).

Compared to traditional IPOs, which are very standardized and require significant legal effort, ICOs are very different.

In IPOs, investors pay for a company’s stock with fiat money or cash in exchange for some level of control over the company.

ICOs are different, there is no need for an investment bank to manage fundraising activities, shares and voting rights are barred for investors and there is no government involvement. In addition, projects or startups rarely have a business history or assets placed in the market.

ICOs enable startups to raise large amounts of funding with minimal effort while avoiding compliance with strict rules and intermediation costs (Kaal and Dell’Erba, 2017).

In particular, an ICO can be described as a mechanism through which new companies raise capital by selling tokens to a crowd of investors (Fisch et al., 2019).

A token is generally a cryptocurrency, a digital medium of exchange based on distributed ledger technology (DLT). Tokens become future functional units of the enterprise project in the form of ownership rights, royalties, or other utility functions.

Distributed ledger technology is a database distributed among different nodes or IT devices, individually involved in network replication and storing a copy of the ledger. There is no central command authority, no arbitrator, and each node that proceeds with registration and rescue operates independently.

The currently most widespread type of DLT is blockchain technology (BCT). Blockchain startups have adopted Initial Coin Offerings (ICOs) as a tool to raise seed capital. The crypto tokens offered in these sales are said to fulfill very different roles on different platforms (Conley, 2019).

The idea behind BCT is that it allows actors in a system (called nodes) to conduct digital transactions over a P2P network that stores those transactions distributed across the network (Back, Corallo and Dashjr, 2014). This innovative technology offered ICOs the opportunity to raise large amounts of money.

When evaluating ICOs by size, we can consider both the amount of money raised in the ICO, and the return on investment. Sometimes ICOs with a relative return on investment do not represent the projects that earn the most and vice versa. The Ethereum ICO in 2014, one of the early pioneers, raised $18 million in 42 days. Ethereum has been instrumental in the development of ICOs thanks to its innovations in decentralized applications. More recently, ICOs have generated significantly larger amounts in terms of total funds raised. The biggest ICO in this regard is Filecoin, a decentralized cloud storage project. During a month-long ICO that ended in September 2017, Filecoin managed to raise around $257 million (Frankenfield, Investopedia.com).

In 2018, 2,284 initial coin offerings were concluded and investors could choose, on average, from 482 token sales. The total amount raised in 2018 was almost $11.4 billion (ICObench database, which includes over 5,100 ICOs as of August 2015).

1.2 ICOs and other forms of funding

Funding comes from token sales by cutting-edge tech companies and token purchases by investors around the world (Chiu and Greene, 2019). Therefore, investors can buy tokens directly from the new company without the involvement of a third party. This feature can serve as a replacement for third parties such as banks and financial institutions, since the security of the system is guaranteed by the blockchain code, and the process is much less expensive.

According to Howell, Niessner, & Yermack (Howell, Niessner and Yermack, 2019), There are three main categories of tokens: (i) currency tokens: used as a medium of exchange and stored as cryptocurrency; (ii) Security Token: Used as a traditional security but recorded and traded on a blockchain. The underlying asset of this type of token can range from company shares (typical stocks) to commodities, real estate or even currencies and (iii) utility tokens: the most common type of token that gives the buyer consumer rights to access a product or service. According to Kranz et al.(J Kranz, Nagel and Yoo, 2019), there is a fourth type of token, namely donation tokens, which do not grant the investor any rights and are used to raise funds for entrepreneurial and non-profit projects. According to Brochado (Campino, Brochado and Rosa, 2021), there are also hybrid tokens that combine more than one of the properties mentioned above. In addition, new types of tokens are expected to appear in the future (Fisch, 2019). The characteristics of the project are collected in the white paper, which is not regulated but follows certain characteristics and can be compared to a regulated prospectus. The white paper is also a measure of the credibility of the project as it contains technical information, business information and information about the team (Chiu and Greene, 2019).

An ICO project might have a minimum or maximum capital threshold that should be raised(J. Kranz, Nagel and Yoo, 2019): (i) no-cap: project without funding limits; (ii) Soft-Cap: minimum capital limit reached to proceed with the project; (iii) Hard-Cap: maximum amount of capital accepted; (iv) Collecting and Returning: A fixed cap will be set and if exceeded, the tokens will be distributed taking into account the ratio between the fixed cap and the total funds received; (v) dynamic upper limit: Several upper limits are defined and kept secret, and (vi) a combination of several named characteristics.

The ICO market is still largely unregulated as new initiatives are not required to follow many internal rules. This lack of regulation is a distinctive feature of ICO projects.

They have been compared to crowdfunding, venture capital (VC) and initial public offerings (IPO) which are traditional ways to fund the project, but there are still significant differences. The concept of crowdfunding can be defined as an open call for the collection of resources (funds, money, material goods, time) from the general population via an internet platform. In exchange for their contributions, the crowd may receive a range of tangible or intangible assets, depending on the type of crowdfunding.

In exchange for their contributions, the crowd may receive a range of tangible or intangible assets, depending on the type of crowdfunding. It usually takes place on crowdfunding platforms, i.e. Internet-based platforms that connect fundraisers with donors. ICOs share similarities with crowdfunding, mainly when the former uses utility tokens and the latter follows some sort of premium crowdfunding, as both concepts allow the investor to use the end product after launch, although unlike crowdfunding, most ICOs do not confer equity rights (OECD, 2019).

In addition, both types of funding make it possible to support a project at a very early stage in its lifecycle. In fact, ICOs share several characteristics with crowdfunding even if they use a blockchain system. However, ICOs differ from crowdfunding as the former are mainly decentralized by replacing intermediaries with a blockchain and the latter use a centralized platform that also performs some due diligence on the project. Businesses using crowdfunding may need to price products upfront, unlike ICO (OECD, 2019). The main difference between the two concepts is the existence of liquidity and a secondary market for ICOs (Campino, Brochado and Rosa, 2021).

Venture capital (VC) firms fill an important gap in corporate finance by fostering innovation and entrepreneurship by funding new companies with growth prospects but few intangible assets, few or no guarantees and high uncertainty about their future (Gompers and Lerner, 2001). Traditional lenders will not find themselves in such a risky situation of financing these projects with insufficient collateral. VC companies also rely on the network of investors who trust the VC’s work in selecting good business projects with high return potential. These investors will regularly entrust their money to a VC company investing them. VC firms also oversee business ventures and should eventually sell their interest, usually convertible securities or stock, at a premium through an IPO or other entity (Nippa and Reuer, 2019). ICOs and VCs have been compared, although they remain distinct instruments. They complement each other through the involvement of VC firms in the early stages of ICOs (OECD, 2019). This complementarity is all the more important as the scrutiny performed by VC firms, their experience, industry knowledge and networking are paramount in early stages of projects (Fried, Bruton and Hisrich, 1998). Again, the major benefit of ICOs is liquidity and the secondary market, as tokens are liquid and can be easily traded, which differs from VC investments, which take several years to become liquid. ICOs have also been favored over projects with a higher risk involved, as VC firms require guarantees to fund such risky projects, while leveraging the founders’ network and helping to build a potential future consumer base and brand awareness (OECD, 2019).

Companies complete IPOs later in life because they need capital expansion at a stage when other sources of capital may not suffice. The capital raised through an IPO could be used to pay off previous debts or to invest in new projects. Greater transparency after an IPO could be important for a company to compete in the market when its activities and accounts are much more scrutinized (Khurshed, 2019). The ICO and IPO concepts were compared, but compared to previous concepts they are the most differentiated form of financing. First, ICOs have a much lower cost of selling tokens than equity in an IPO, and there is also a lower initial threshold for investing. These processes take place in different places, in fact an ICO is done virtually in a crypto exchange while an IPO has to take place on the exchange. As such, the IPO market is highly regulated, and although ICOs have attracted the attention of regulators (Howell, Niessner and Yermack, 2019), they are still largely unregulated as well as the Whitepaper, which can be compared to the prospectus of an IPO (Zhang et al., 2020). Although largely unregulated, ICO regulation has increased and in some markets it is a regulated activity like the Swiss market (Spinedi et al., 2019). Stocks and tokens differ mainly in that the former grant their holders ownership rights along with possible voting rights in the company’s decisions, while tokens do not. Furthermore, ICOs fund an idea or project at a very early stage, while IPOs take place when a company is growing and at a more mature stage of its life cycle, and therefore the funding received is mainly based on a solid company balance sheet with good performance. Finally, there is an opportunity to trade a fraction of a token in secondary markets, but the same does not happen with stocks (OECD, 2019).

Chapter 2: The Blockchain

2.1 Basic components

Before presenting the role of the blockchain and how it actually works, it is necessary to define its basic components:

Nodes: A node consists of a physical network device (although there are some specific cases where virtual nodes are required) that acts as a redistribution point or communication endpoint, where a message can be created, received, or transmitted. With a network of nodes, digital currencies are traded without intermediaries, regardless of physical distance;
Transaction: A blockchain transaction is the operation generated by the values being exchanged. Each transaction must be approved – to be validated – and then entered the chain;
Block: is a set of merged files (transactions) related to the network and recorded permanently and then stored as a page of a ledger or log book. Each completed block gives way to the next block in the blockchain and cannot be modified or removed;
Ledger: It is a database containing all verified transactions – in chronological order – shared consensually and synchronized between multiple locations, institutions or accessible by multiple people. Participants in each node of the network can access and have a copy of the shared recordings;
Hashing/Hash: A hash is a function (non-reversible, i.e. not decipherable) that converts an input of letters and numbers (of variable length) into an encrypted output (of fixed length). It is created using an algorithm and provides a unique and secure identifier for each block.

2.2 The Blockchain’s mechanism

In today’s society, trust was created through intermediaries. Third parties have been and are used because people and businesses trust them to store and protect sensitive assets and data while sending the right amount, to the right person, when needed. In this context, blockchain has replaced the need for intermediaries and redirected that reliance to decentralized systems.

Central banks are good examples of such companies and are therefore mainly influenced by this technology as they have changed their business model from an hourly rate to a rate per item due to how the blockchain itself works. Although there is no universally accepted definition, blockchain is a subset of what is known as distributed ledger technology (DLT), which is a way of recording and sharing data between ledgers, d managed and controlled by a distributed network of Computer servers called nodes. In addition, the blockchain mechanism uses an encryption method known as cryptography, which uses specific sets of algorithms to create and then verify an ever-growing data structure. In this structure, data can only be added and not removed, making it a chain of transaction blocks that function like a DT, i.e. distributed ledger.

Fig.1: The Blockchain mechanism (https://www.researchgate.net/figure/Blockchain-mechanism_fig7_335163861)

The blockchain is then expanded by each additional block and represents a complete ledger of transaction history.

As mentioned above, a key element is the trust factor. In this structure, trust is monitored by cryptography. With encryption, each block is securely wrapped in a protective layer and can be validated by the network. In addition, each of them contains a timestamp; the hash value of the previous block (parent); and a nonce, which is a random number, to verify the hash. This procedure ensures the integrity of the entire blockchain and is intended to effectively prevent fraud. Overall, the key identifiable concepts that ensure the functioning of the system are blocks and hashing, mining, and proof of work and consensus.

Finally, from a practical point of view, blockchain is a multifaceted technology as it covers a wide range of systems ranging from fully open without authorization to authorized consortium blockchains that combine elements of both. Most cryptocurrencies currently in circulation are based on an open, no-authorization blockchain, where a person can join or leave the network at will without requiring prior approval from a central authority that just requires a computer with the appropriate software needed. There is no central ownership for both the network and the software. In stark contrast to public blockchains, private chains (or permissioned blockchains) require transaction validators (i.e. nodes) to be pre-selected by a network administrator in order to join the network.

In this authorized environment, the administrator sets rules for the ledger that the chain can see and write, and also defines and verifies the identity of the network participants. Furthermore, this system can be divided into two subcategories, public (or open) and enterprise (or closed) permissioned blockchains. In the first case, anyone can access and view the ledger, but cannot generate transactions or update the status of the ledger unless authorized. However, it should be noted that transactions can be validated on an approved open blockchain and then executed without the intervention of a trusted third party.

Finally, the consortium blockchain sits halfway between public and private chains, combining elements of both, with significant consensus differences. In fact, it works with closed and cryptographically protected databases, which means that the ledger is accessible to the nodes participating in the network and from there different rules apply about who can update the status of the ledger.

Blockchain technology is particularly associated with digital or virtual currency systems, financial services and payments, but could be applied in different sectors and has numerous potential applications e.g. Identity services, such as identity tracking for passport management, birth and marriage certificates, etc. or government services that support the delivery of public needs for citizens and stakeholders, especially in healthcare management (secure storage and access to patient data is a crucial part of the medical industry). Other examples relate to education, as records of students and teachers in education (such records are kept and shared with selected stakeholders), or to public elections, as electoral processes could be managed through distributed registers used to conduct voting processes and identity fraud or theft and many others. In addition, it can apply to the pledging of collateral, the registration of bonds, shares or other assets.

Chapter 3: Cryptocurrencies

3.1 General overview

Financial services and information technology have enabled the development and establishment of both the blockchain technology system and cryptocurrencies. The concept of cryptocurrency first appeared in a 1998 article written by software engineer Wei Dai, perhaps a pseudonym entitled B-Money, An Anonymous Distributed Electronic Payment System. Current cryptocurrency systems are essentially based on their principles, which they have defined as a scheme for a group of untraceable digital pseudonyms to pay each other with money and enforce contracts between them unaided. In the same period, other developers tried to create concepts similar to those of Wei Dai, including Nick Szabo: his project had as a protagonist Bit-Gold, a crypto-value whose functioning was practically identical to B-Money. In this era, cryptocurrencies are essentially applications of blockchain technologies. These applications represent a digital asset whose main purpose is to be a medium of exchange that uses cryptography (i.e. the technique of protecting information by encrypting / converting it into an unreadable format) to protect all transactions through sophisticated public and private digital key systems.

Cryptocurrencies are defined as virtual currencies and a form of unregulated digital money, typically issued and controlled by its developers and accepted and used by members of a specific virtual community: the network. In the absence of trust between parties and intermediaries (therefore they are not managed by a specific institution), virtual currencies use distributed ledgers to enable peer-to-peer remote exchange of electronic values and can be divided into two basic types: convertible and non-convertible virtual currencies. Convertible bonds have real currency equivalent and can be exchanged back and forth for real currency. Non-convertible securities are specific to a particular domain/virtual world, are subject to the rules of inter nos (i.e. the rules governing their use) and therefore cannot be exchanged for fiat currency.

3.2 Main characteristics

In addition, in order to better understand cryptocurrencies and how they work, it is necessary to define some other main characteristics. Cryptocurrency can have all or some of the following characteristics: decentralization, anonymity, reliability and security, problem limitation, accessibility, convenience and speed.

Decentralization: As mentioned, cryptocurrencies emerge from the need to move away (or separate) from the traditional system of control of central authorities and create a decentralized and fully autonomous one. From a practical point of view, this currency is an encrypted code calculated by a certain algorithm that determines its output. The latter controls a network of linked servers, rather than being managed by a single entity like the central bank. This process of creating money is called mining;
Anonymity: It is difficult to be aware of and to know the identity of the natural person or person conducting a cryptocurrency transaction. In fact, transactions do not require identification by another person’s name, address, or information. This means that person can purchase and use numerous services without being detected. However, it is important to note that this does not mean that there is no trace of it, on the contrary, the blockchain works precisely because transactions are recorded regularly, everyone can see them and check the exchanges so far. The fact that virtual currencies and transactions are considered pseudo-anonymous and not fully anonymous means that the sender and recipient can only be represented by a long code made up of letters and numbers (the public key). From this it is understandable that through these public keys it is not possible to easily identify the natural person performing the operation, but at the same time have access to the balance of, for example, bitcoins of a specific public key;
Reliability and Security: The cryptocurrency system works because it does not require trust between parties; Anyone can access and view ongoing transactions at any time, even for a specific user. These currencies are transparent, all operations are stored in the ledger system and remain immutable and irreversible. The structure is secure because there is a common consensus, it is the network server system that takes care of validating the integrity of transactions. Even if it is not possible to trust all participants, it is difficult not to trust the network as a whole, because in order to change the system and gain full control, a single node would have to have extremely high computing capacity;
Limitation: Some cryptocurrencies have coin creation caps, which means that their issuance will stop when a maximum number set by the creators of the same coin is reached. For example, the most famous and widely used currency, Bitcoin (BTC), will cease to be created when it reaches 21 million units, presumably in 2140; this also applies to Litecoin: its coin creation will stop when it reaches 84 million units;
Accessibility: You only need an internet connection to use cryptocurrency. Through this connection, it is possible to create an electronic wallet, also called a web wallet, containing cryptocurrency available for transactions;
Convenience: cryptocurrencies bypass the intermediaries that the traditional system usually uses; They do not require payment of high fees or other charges for executing transactions. Small fees are usually paid to miners for these trades (these fees are variable and there is generally a market price to execute the trade in a timely manner: the higher the commission, the sooner the operation is executed) – which allows transactions to be validated as a priority;
Speed: Unlike banks, which sometimes take a few days to transfer funds, cryptocurrencies take around 10 minutes to verify and validate a transaction;

In summary, as in the case of blockchain technology, there is no universally accepted definition of the term cryptocurrency. However, most policymakers (such as the European Central Bank, the Financial Action Task Force, the International Monetary Fund, the World Bank, etc.) have attempted to define them as a subset or form of virtual or digital currencies, or as a summary of the preceding discussion as a digital representation of Value intended to represent a peer-to-peer (P2P) alternative to government-issued legal tender, used as a generic medium of exchange (independent of any central bank) secured by a mechanism known as cryptography and can be converted into a legal means of payment are converted and vice versa.

The term cryptocurrency is often incorrectly used in a very broad sense; Therefore, some clarifications are needed. Cryptocurrencies can be divided into two other main categories, namely altcoin, which was born as a hard fork of the Bitcoin code that has undergone modifications like Litecoin and Dogecoin, and altcoins based on other blockchains like Ethereum. All software needs constant updates to fix problems or improve performance. In the cryptocurrency world, these updates are called forks. Since cryptocurrencies are decentralized networks, all participants in the network must follow the same rules in order to work together properly. This set of rules (like block size or miner rewards) is called a protocol. There are two types of forks in the cryptocurrency world, soft forks and hard forks. A soft fork is a change in a cryptocurrency protocol that is backward compatible: Legacy nodes can continue to process transactions and add or move new blocks to the chain as long as they don’t violate the new protocol’s rules.

On the other hand, a hard fork is a change in a cryptocurrency protocol that is incompatible with previous versions. Nodes that have not been updated to the new version cannot process transactions or add/send new blocks to the chain. These forks can be used to modify or improve an existing protocol or to create a new and independent blockchain. As a result, additional categories and / or subcategories emerged from these classic currencies, both for technological evolution and for the evolution of needs. Among these, tokens come into play and can be defined as a subset of cryptocurrencies that do not have their own specific blockchain but indirectly use other chains. Their peculiarity is that they offer a different and superior functionality than a generic medium of exchange. An example of this are utility tokens, typically introduced by ICOs, asset or security tokens, and payment tokens. Security tokens are also known as crypto-securities because they use blockchain technology to register, issue, and transfer stocks or other commercial securities using cryptography. What tokens and cryptocurrencies have in common is that they both implement blockchain technology.

But how many cryptocurrencies are there? Currently, according to CoinMarketCap, the number of existing currencies is extremely high, exceeding 5000 species, and it is a growing number. Coinlore shows a notable difference, listing a total of 4390 cryptocurrencies as many of them can be linked to projects that have been abandoned, failed or scammed. In fact, not all cryptographic projects can have a long lifespan, or in any case were not created with serious intentions. The first cryptocurrency, certainly the best known and most used, is Bitcoin (BTC). They are joined by other cryptocurrencies that are based on the same blockchain technology but are designed differently than Bitcoin because they are created for different purposes. The other most traded so far are Ethereum (ETH), Tether (USDT), Ripple (XRP) and Bitcoin Cash (BCH).

Chapter 4: ICO’s components

As mentioned above, Initial Coin Offerings are a cutting-edge way to raise money for startups by allowing them to exchange digital tokens for funding from investors. Companies that want to start a token sale to fund the development of a project produce and distribute their own cryptocurrency tokens using a platform that uses blockchain technology. It is clear that both the advent of cryptocurrencies and technology have made ICOs possible. Therefore, it is crucial to break down token sales into their core components and explain each component in detail, including blockchain technology, platform, smart contracts, tokens, and white papers, to fully understand how they work.

4.1 Blockchain technology

Initial Coin Offerings are a phenomenon related to Global Distributed Ledger Technology (DLT), specifically its application Blockchain. In fact, token issuance (through ICOs) and subsequent transactions are recorded in a distributed, encrypted and protected digital ledger. Let’s look at these two technologies in detail.

A distributed digital ledger represents a database that does not depend on a central computer, but is shared and synchronized by a network of independent computers (called nodes) that are connected to each (Belleflamme, Lambert and Schwienbacher, 2012). In other words, DLT is a register that can be read and modified by the many actors involved in the network. As this network of nodes grows, the registry becomes more reliable and less vulnerable to cyberattacks and fraud.

Blockchains, on the other hand, are a subdivision of the more general distributed ledger technology, with the particularity that they are ledgers organized as a blockchain. A blockchain is technically defined as a distributed ledger that stores chronologically ordered data in an ever-growing list of blocks. These blocks of information are linked together so that each block is linked to the next.

Individual blocks contain transaction and business information and are verified by a consensus mechanism distributed to some or all nodes in the network. In other words, the entire network uses a consensus-based algorithm to come to an agreement on which blocks can be added to an existing, pre-verified blockchain. For this reason it is called blockchain.

Consensus protocols are the rules of the game for agents operating within this distributed system. It is precisely thanks to these rules that the system enjoys the trust of all system participants, even if there are many heterogeneous block verifiers (or miners). Realizing decentralized consensus is actually the greatest innovation that blockchain technology can bring. Consensus, in this case, is not only about transaction agreements, but also about dispute resolution protocols, the history of events and the information in general that is recorded on the blockchain.

The advantages of this technology are many.

Decentralization: the registry is decentralized and not controlled by a single central authority;
Transparency: the history of transactions made on the ledger can be viewed by anyone who wishes;
Automation: Smart contracts automate tasks that are quick and impossible to change;
The ledger is distributed across multiple nodes, making it more resilient to cyberattacks;
Reduction of costs and transaction times compared to traditional systems, for example for some types of transactions in banks.

The first blockchain in history was theorized in 2008 by one or more people under the pseudonym Satoshi Nakamoto and the technology was released in October of the same year under the name Bitcoin. To facilitate bitcoin transactions, Nakamoto developed a peer-to-peer network (a network of computers connected to each other via the Internet that allows files to be exchanged directly without the presence of a central server) that allows users to exchange currencies while allowing the entire network to record all transactions on a publicly accessible but secure ledger (blockchain), which is collectively verified by a central authority.

This digital ledger is called decentralized precisely because there is no central authority verifying the authenticity of transactions or managing data, and it is stored on all private computers on the network, distributed around the world. The individual computers that represent the nodes not only keep copies of the public ledger, but depending on the type of blockchain, are also tasked with verifying and updating the information contained in the blocks of the digital chain.

There are three types of blockchains: so-called public or permissionless blockchains, permissioned blockchains, and private blockchains. The difference between the three lies in the distribution of the consensus mechanism among the different nodes of the network, i.e. who can participate in the consensus building process.

With permissionless blockchains, this mechanism is distributed to all nodes in the network. This means that, as a rule, all agents can participate in the creation and verification of information blocks. However, permissively, agreements are only distributed among the nodes that are pre-authorized to participate in this process, while in private blockchains the management of the ledger occurs at the level of the individual owner of a given blockchain.

While most cryptocurrencies (e.g. Bitcoin) are based on public blockchains, many business applications rely on authorized blockchains as they are controlled by the respective companies that own them. Traditionally, even central entities such as governments and banks have relied on consensus, but this is achieved through costly and time-consuming methods, subject to change due to centralized control by a single authority.

Blockchain offers a more efficient and decentralized alternative to making agreements about shared information. Now let’s analyze in more detail the main elements of the blockchain and the actions required for verification and subsequent data collection.

Nodes: are the individual participants in the network and are physically represented by a dedicated computer for each participant. The sum of all nodes is the network itself. Each node has a copy of the public ledger, and it is the network’s job to verify transactions so that they can be recorded in the blockchain;
Transaction: consists of a data packet containing information about the subject of the exchange (date and time, public address of the recipient, characteristics of the transaction, cryptographic signature). This data must be verified, approved and finally collected by the network;
Block: represented by a grouping of transactions (can be viewed as a series of transactions). Each transaction must be linked to other transactions to form a block that must be verified and approved by a miner before being added to the blockchain;
The Ledger: is the public ledger where all transactions are stored in an orderly (in blocks) and sequential manner. The ledger is represented by the union of blocks chained together by a cryptographic function (the hash);
Hashing: Consists of an irreversible operation that converts a variable-length text and/or numeric string into a single, unique, fixed-length string. This converts the transaction data contained in a block into a single code of the same length. This ensures that each block is uniquely associated with a specific code and that converting a block of 100 transactions into a single code requires the same effort as converting a block into a single transaction. The text that generates the hash value (transaction data) is also not recognizable. Finally, each block contains a hash value that not only records all information about the block itself, but also contains information about the previous block, and each block is combined with the previous block.

This data-logging mechanism effectively prevents fraudulent transactions, since any change to a block inevitably changes the corresponding hash value. Therefore, to reliably modify a block, all blocks containing that hash must be modified. This task requires enormous effort and energy. Blockchain, or more generally the advanced system of recording and managing information made possible by distributed ledger technologies, is used by startups, banks, insurance companies, governmental and non-governmental organizations, and many other banking, payment, and transactional facilities smart contract services, property rights management and many other possible uses of such a system.

In initial coin offerings, companies enter their data and rely on the blockchain to create and issue their tokens.

4.2 Platform

The central moment of the ICO is the creation and issuance of tokens. For this purpose, companies use platforms based on blockchain technology. In the case of Initial Coin Offerings, platform refers to the digital infrastructure on which companies distribute and issue tokens and the distributed ledger that stores transaction data. Issuing companies can choose between two methods of creating cryptocurrencies: create tokens from scratch and develop their own blockchain platform, or leverage existing services and platforms. Both options have advantages and disadvantages. One option is not necessarily better than the other, but it is in each company’s interest to weigh and compare the pros and cons, also in relation to the type of project they wish to develop.

In particular, creating a coupon from scratch has the benefit of giving greater freedom in terms of functionality. This means companies can customize it and tailor it to their needs and goals. On the other hand, however, this requires a high level of commitment. In addition to time and money, they must have the technical skills to develop not only the token but also the blockchain on which the token is based. On the other hand, when using existing platforms, much of the programming is done by existing services, which allows companies to reduce their overhead, but the degree of flexibility and customization of the tokens created in this way is less. Typically, companies looking to develop projects in the technology sector (with funds raised through an ICO) choose the first option. This is because they likely already have the skills needed to develop their own blockchain platform and the cost of development is lower than the cost of acquisition.

The most common platform used by companies to conduct ICOs is Ethereum. It allows the development of any decentralized application and has the advantage of simplifying and speeding up the creation of any blockchain-based application. In addition, it is also the most popular platform for executing smart contracts.

4.3 Smart contracts

When an ICO is launched, investors buy the issued tokens in exchange for cryptocurrencies or other fiat currencies, as determined by the issuing company. The number of these transactions to buy (sell) tokens makes it impractical for companies to manage them individually. In fact, in addition to the enormous amount of time it takes to process each exchange transaction, errors and fraud can also occur. Therefore, companies use smart contracts to achieve efficient operations at this stage.

Smart contracts are one of the most innovative applications of blockchain technology and are at the heart of initial coin offerings. Specifically, smart contracts are codes that define the rules for transactions between two or more parties that are recorded on the blockchain and continuously converted into code. Once the terms predefined by the parties have been implemented and automatically validated by the smart contract, the terms of the contract will be executed automatically. In other words, a smart contract consists of a digital medium that reads not only the terms agreed upon by the parties, but also the operating conditions under which those terms should occur and is executed when the data specifying those terms Data match the predefined terms.

In the case of a contract offer, before launching the ICO, the company establishes in the smart contract code a set of rules for interaction between itself and investors, which may relate to the type of currency accepted (fiat or other cryptocurrencies) and the price of the token , the duration of the offer, the total number of tokens to be issued and any other conditions that must be met by the parties involved in order for the transaction to be carried out correctly. Smart contracts make it easy to review and conduct contract negotiations without the need for intermediaries as everything is managed through code. Also, they are immutable. The contract is concluded when the default clauses and operating conditions match and there are no possibilities for manipulation. In other words, the contract once entered into cannot be canceled or changed, making fraud less likely.

4.4 Tokens

Another important element of an initial coin offering is the token. These are digital assets that companies sell to investors in exchange for cryptocurrencies or other fiat currencies (during ICOs). The tokens are based on blockchain technology and can be transferred between multiple parties without the involvement of a central government and traded without restrictions on virtual cryptocurrency markets. In this sense, tokens can be equated with cryptocurrencies, but tokens are not only a means of payment and value transfer, but they also confer various rights on their holders.

There are therefore different types of tokens depending on the rights and functions granted to their holder.

Buying tokens does not create ownership of the vouchers themselves, but rather royalties on specific projects. Therefore, it is necessary to classify tokens issued in ICOs into three groups:

1) Payment Tokens: currency generated vouchers that can be used as tokens and exchanged for other tokens or fiat currency depending on their value. They can be exchanged for tokens or other types of tokens or other currencies, and their value is determined by the market. Examples of this type are bitcoin and ether, which are real currencies that represent a means of payment; they are general purpose currencies used to purchase goods and services;

2) Investment tokens: this is another type of token that represents a set of rights (e.g. the right to vote or receive payments). Those who decide to buy this type of token do so in order to invest in the token, believing that the value of the token might increase in the future. Investment tokens may be subject to regulation in many countries. In some they are treated as financial instruments, but in most cases they are considered financial instruments;

3) Utility tokens: tokens that allow the purchase of goods or the use of certain services of the issuer and exclude speculation, money or participation activities. The difference is that utility tokens can only be used to purchase specific goods or services from the issuer, while payment tokens are used as a general means of payment for any goods or services. The buyer of a utility token intends to use it to purchase or access goods and services within the system in which the token operates. In this case, the token is an access key to enter the network and use its tools.

Therefore, given the variety of tokens that can be created, there are different types of ICOs, which in practice can be divided into two groups:

(a) Coin ICO (Currency ICO): in this case, the Token Offering consists of creating a new cryptocurrency that is exchanged for an existing cryptocurrency such as Bitcoin as a payment token that will be circulated in a new mode.

b) Project ICO: this type of initial coin offering is so called when the tokens are for investment or use. In fact, by purchasing tokens, you acquire certain rights or the ability to use the services offered by the company. The issuing company’s own tokens are distributed on an existing blockchain platform.

4.5 Whitepaper

At the outset of a token offering, startups typically create and publish a document called a white paper that provides details about the project and information about the cryptocurrency assets they intend to issue. This document will be distributed through the company’s communication channels (website, telegram groups, twitter and other ICO review sites) before the launch of the ICO.

Through their public announcement, the companies try to attract as many investors as possible, and those interested in the project can theoretically understand what they want to invest in or simply want to participate in. The information is published by many companies in their white papers, including:

Description of the project, business plan, amount of funds raised, development status and progress of the project (roadmap), etc.;
Technical specifications, including the mechanism and type of blockchain used to create and issue the tokens, and the rationale for choosing the ICO over other means of fundraising;
The details of the ICO, including the terms of purchasing the tokens, the start and end dates of the ICO, the total amount of crypto assets to be issued, the number of tokens to be allocated to the team, and the issue price;
Composition of the project team, distribution of roles, previous successes of the founders and their background.

As initial coin offerings are not regulated, there is currently no minimum level of information that companies are required to provide to potential investors (Amsden and Schweizer, 2019). As a result, the information contained in the whitepaper varies from ICO to ICO, and companies may choose not to publish a whitepaper at all. This is the main risk of initial coin offerings (Adhami, Giudici and Martinazzi, 2018). Investors have no protection available, especially when the tokens they hold cannot be defined as securities and there is no regulator overseeing the issuer’s activities.

However, for investors, this document can be a very important tool to better inform them about the objective of the investment or, more simply, about the participation. In fact, logically, the more information that is published and the more complete it is, the greater the confidence of potential investors in the investment opportunity and the greater the company’s money will be raised during the ICO.

However, as several studies have shown, few companies provide a detailed and complete introduction to the project they intend to develop or to the company itself through white papers. In addition, several econometric analyzes have shown that the probability of success of an ICO, measured by the amount of capital raised, does not increase with white papers. This is because white papers are not considered very relevant for potential investors, if only because the amount and quality of the information they contain varies, and above all because these documents do not have a certification or administrative function, so their mere existence is not considered to be of great relevance for potential investors, although they can play a positive role for marketing and communication purposes.

4.6 ICO’s stages

After a detailed analysis of the main components, it is helpful to examine the stages of the Initial Coin Offering development process. However, ICOs are a very heterogeneous and unregulated reality, so the order in which the different phases take place is not predetermined and the process of developing the token offering does not necessarily unfold through them. When a startup has a project that needs to develop funds, they should first consider whether launching an Initial Coin Offering is the most convenient and efficient way to raise the necessary funds. At a very early stage of this evaluation, the company should consider reasonable and sensible ways to incorporate the later issued tokens into the project and make them an integral part of the final product. This move is crucial to drive demand for tokens and stay ahead of the competition.

The second important step in preparing an ICO is choosing a blockchain platform and creating the infrastructure. You can use an existing platform or create your own blockchain from scratch. This decision varies from ICO to ICO as there is no real preferred platform. In practice, the choice is made by a dedicated startup team that weighs the pros and cons of the two options and compares them to the project they want to launch.

Building your own infrastructure is therefore more flexible in terms of functionality, but also more expensive than alternatives, as it requires specialized technical know-how and significant capital. On the other hand, the use of existing blockchain platforms guarantees cost savings, but at the expense of flexibility as they usually offer predefined and standardized features.

On the other hand, using existing blockchain platforms guarantees cost savings, but at the expense of flexibility, since they usually offer predefined and standardized functions. Although there are many potential existing platforms suitable for ICOs, most token offerings are structured using the Ethereum platform. Once a blockchain platform is chosen, smart contracts are programmed onto it to actually raise funds. This programming consists of defining the conditions of the token sale and the characteristics of the token itself. Once the technical infrastructure of the ICO is in place, the next step is to create a white paper that describes the key elements of the project and offers icons to the public. The creation of this document follows its publication and dissemination through the communication channels available to the company.

At this stage, the company’s professional website is usually created. This website usually contains information about the team, project roadmap, success stories and previous projects etc. After the launch of the ICO, this website is usually used to post updates about the project. The creation and publication of white papers is not regulated in an ICO at the regulatory level, however many companies choose to provide white papers to inform potential investors about the project. After publishing the whitepaper, the company can decide to launch a pre-ICO. A pre-ICO is like an ICO with a smaller amount of capital where investors who buy the tokens get a discount. This is similar to the financial incentives offered during the securities offering phase of a public offering. It encourages the purchase of tokens and is unfamiliar to investors due to its nature (in terms of Initial Coin Offerings).

The pre-ICO (or pre-sale) has several functions: covering the start-up costs of the main ICO, including marketing campaigns and software development; it can increase the credibility of the issuer with potential investors, especially when institutional investors participate in the pre-ICO and successfully raise a large amount of capital. ; It allows you to get a timely assessment of the demand for the token and its fair price. The pre-ICO can also be viewed as a mechanism to gather information from potential investors to increase the efficiency of the ICO. After coming up with an idea, building a dedicated team, writing and publishing a white paper, creating discount vouchers and ensuring compliance with applicable laws, startups need to carry out promotional activities.

This phase is crucial because the more potential investors you can reach, the greater the chances of an ICO being successful. Companies use social media to communicate with investors and announce coupon deals, mainly via Telegram and Twitter, Reddit or a highly professional company website. Once the previous steps are completed, the ICO proceeds. There are no specific rules as to when or how long an O IC can begin: some O ICs last less than a day, while others can last a year or more. The company then decides on the duration of the offering and whether it will set a minimum or maximum amount of funds to be raised. This information is part of the conditions defined in the smart contract.

In many ICOs, token buyers send payments to the issuing company’s blockchain address. Payments are usually made in other cryptocurrencies, but sometimes virtual currencies can also be accepted. The entire public offering process is managed automatically and autonomously by smart contracts. The issuing company has no control at this point, once the smart contract is activated all token trading terms are irrevocably set by the company itself. If the capital sent by the investor is accepted, the smart contract sends the corresponding token (always according to the predefined conditions in the contract) to the investor’s blockchain address. On the other hand, the unaccepted token is automatically transferred to its holder. If the company has not reached the minimum amount (if any) by the time the public offering is completed, the ICO will be considered bankrupt and the company itself will have to return all the raised capital to investors through smart contracts.

Once the public offering phase is complete, companies can decide to list their digital assets on virtual cryptocurrency markets. The listing allows the token to be traded even after the ICO is complete, making it the largest source of liquidity. This liquidity attracts new investors and allows the token to be used as a standalone currency. However, for most tokens, the lifecycle ends upon termination, meaning the project essentially disappears as there is no platform to trade tokens. After these phases, if the offer is successful, the company can further develop the project with the support of investors.

Chapter 5 : ICOs: a taxonomy of academic literature

5.1 ICO’s theoretical framework

The literature on ICOs is still quite limited. Most of the work relates to the success factors of ICOs and links success to the amount of money raised. Adhami, Giudici, and Martinazzi (Adhami, Giudici and Martinazzi, 2018) examined the specific characteristics of an ICO that determine success. They argue that an ICO’s likelihood of success is greater when the source of the code is available, when a token pre-sale is organized, and when tokens allow contributors to access a specific service. Fisch (Fisch, 2019) analysed the factors that determine the amount collected. The results examined by Fisch showed that technical white papers and high-quality source code increase the amount raised, while patents are not associated with higher funding amounts. Fisch, Masiak, Vismara, and Block (Fisch et al., 2019) identified and ranked motivations for investing in ICOs using factor analysis. They show that investors are driven by ideological, technological and financial models. Furthermore, Fisch and Momtaz (Fisch and Momtaz, 2020) examined the role of institutional investors in ICOs. They argue that the superior screening and coaching skills of institutional investors allow them to overcome the information asymmetry of the ICO context. They believe institutional investor support is associated with higher post-ICO performance.

Roosenboom, Van der Kolk and De Jong (Roosenboom, van der Kolk and de Jong, 2020) provide evidence that ICOs are more successful in raising funds when they disclose more information to investors, have a higher quality rating of cryptocurrency experts, a organize Presale, have shorter planned token sale periods and have a larger project team.

Some studies have focused on aspects of the company such as: B. fundraising, others have examined the characteristics of the investor or the investee company or both. Hsieh and Oppermann (Hsieh and Oppermann, 2020) examine how ICO characteristics, cryptocurrency markets, jurisdictions, the ICO industry, and traditional financial markets affect initial ICO returns. Additionally, they found that a short bidding period, no pre-sales, a well-written whitepaper, and the creation of an independent blockchain positively impact initial ICO returns.

Momtaz (Momtaz, 2020) focused on asymmetric investor-investee information. He believes that loyal CEOs should offer fewer financial incentives to attract investors and still be able to generate more revenue and be less likely to fail. Another mainstream study analyses ICOs in relation to IPO (initial public offering) and crowdfunding phenomena.

An, Duan, Hou, and Xu (An et al., 2019) examined the impact of founder traits on company success in ICOs and drew parallels between ICOs, crowdfunding, and venture capital, with a large body of literature examining the relationship between trait traits Founder and company performance. They found that disclosure of the founders’ personal information is associated with a larger amount of funds raised in ICOs. Huang, Meoli, and Vismara (Huang, Meoli and Vismara, 2020) have shown that the availability of investment-based crowdfunding platforms is positively associated with the growth in the number of ICOs, while the debt and private equity markets offer no similar effects.

Block, Groh, Hornuf, Vanacker and Vismara (Block et al., 2020) made a comparison between crowdfunding and ICOs. Their study showed that while the two market segments may initially appear similar, there are significant differences between them. Their discussion focused on stakeholders, microstructures, regulatory frameworks and market development. Collomb, De Filippi, and Sok (Collomb, De Filippi and Sok, 2019) compare initial public offerings (IPOs) and equity crowdfunding to ICOs and examine the associated risks and limitations of these different fundraising practices. They found that many ICOs share many similarities with traditional IPOs and equity crowdfunding; So they should be adjusted in a similar way.

Hashemi Joo, Nishikawa and Dandapani (Hashemi Joo, Nishikawa and Dandapani, 2019) recognize the advantages of the ICO as a means of raising funds and present a comparison between the ICO and the IPO to realize the future possibilities of this innovative financing method. The structure of the ICO is much more elastic and represents a faster and cheaper way to raise capital than IPOs.

Other academics also consider it essential to go into blockchain when analyzing the phenomenon of ICOs. Kher, Terjesen, and Liu (Kher, Terjesen and Liu, 2020) systematically reviewed 152 articles on blockchain and its applications, and synthesized five themes: computer science, economics, entrepreneurship, law, and governance.

According to Boreiko, Ferrarini, and Giudici (Boreiko, Ferrarini and Giudici, 2019), ICOs are a new way for blockchain startups to fund project development by issuing coins or tokens in exchange for fiat money or bitcoin or other cryptocurrencies. They compared European and American regulation and highlighted the major differences between Europe and the United States that make Europe less friendly to blockchain startups.

According to Yan Chen (Chen, 2018), blockchain tokens can democratize entrepreneurship by offering entrepreneurs new ways to raise funds and engaging stakeholders, and offering innovators a new way to build decentralized applications.

Lo and Medda (Lo and Medda, 2020), the blockchain token-related company, and developed the analysis through a stepwise test of four hypotheses using the panel’s ordinary squares with robust clusters -Standard Errors. They show that token features are statistically significant in relation to token prices.

Mangano (Mangano, 2018) illustrated the pros and cons of using blockchain technology in finance. The issuance of blockchain securities creates a divide between the world where securities are issued, offered and sold and the world where the law applies. Albrecht, Lutz, and Neumann (Albrecht, Lutz and Neumann, 2020) examine whether blockchain initiatives can reduce investor/investee information asymmetries by leveraging reporting mechanisms on Twitter and see how the resulting impacts differ from those in traditional market environments.

5.2 Methodology

In conducting the review, I followed the approach of Za (Za et al., 2018), which includes four main phases: 1) collection of materials, 2) collection of analyses, 3) development of taxonomy (selection of structural dimensions and categories based on an established theory), 4) preliminary evaluation and interpretation.

Although I have followed this process, I have collected the first two steps so that the study has three main phases as shown in fig.2.

This study was developed using a mixed methodology. In the first phase, he selected bibliographic sources using the Scopus database because academics and professionals consider this tool a complete database of citations and abstracts edited by experts. The first phase of the research protocol concerns the definition and description of the dataset. I selected search terms using Scopus and then performed a hands-on and methodological screening that provided a description of the dataset, possibly a refinement selection.

In the second phase, I used the Nickerson et al. developed taxonomy process (Nickerson, Varshney and Muntermann, 2013). The goal is to develop a taxonomy with a set of dimensions, each consisting of a set of characteristics that describe the objects in a given study. It consists in the iteration of empirical-conceptual and conceptual-empirical approaches in the analysis of the articles collected in the data set. The process of iteration must be carried out until the values and attributes of the theoretical dimensions appear clear (Cipriano and Za, 2021). The purpose is to identify a set of dimensions and their values to better classify the items in the data set. A “useful” taxonomy is as defined by Nickerson et al. (Nickerson, Varshney, and Muntermann, 2013): Dimensions and values must be concise, robust, complete, and extensible. Finally, in the search, each paper has a value for each dimension, so we didn’t assign two or more values to a dimension.

In the third section, we report the results of a preliminary interpretation of studies on the ICO phenomenon. As part of this analysis, we look at the theoretical background of ICOs to understand how research type and topics are related to this innovative financial tool. From the preliminary analysis, we report the number of papers involved for each value, distinguishing four dimensions on the research type (research approach, research design, data collection, philosophical perspective) and eight dimensions on the different topics (field of study, focus, actors, token type, additional topics). , research question, ICO phase, blockchain).

5.3 Material collection and sample description

In a literature review, academics select different types of criteria to collect documents. In order to cover all studies on the ICO phenomenon, we conduct our search using the Scopus database, as this platform is widely used by academics and researchers in the field of social sciences (Cipriano and Za, 2021). The first phase of our work involved data collection to identify a congruent source of scientific literature. We used the Scopus database to collect relevant studies on the ICO phenomenon. To conduct the search, we specified a query on Scopus to find documents containing the string Initial Coin Offering* in the article title, abstract, or keywords with no time constraint. We’ve used wildcards to include plural words, grammar and spelling variants as well.

The first query returned 226 publications published between 2017 and 2021. There is growing interest in this innovative financial tool. Articles made up 63% of the papers, followed by conference papers, which made up 20% of the total. Commonly used keywords are initial coin offering (130), blockchain (108), and cryptocurrency (67). Major publishing outlets included Small Business Economics, Economist United Kingdom, Journal of Alternative Investments, European Business Organization Law Review, Journal of Corporate Finance. We also included only English and excluded the publisher Economist United Kingdom as it is a newspaper. The first refined dataset comprised 133 articles. We analyzed all abstracts to further refine our dataset to only include work consistent with the ICO phenomenon. The final refinement resulted in 99 relevant articles for our purposes.

5.4 Taxonomy description

In this study, we developed a taxonomy to better understand and describe the emerging phenomenon of ICOs. A taxonomy is the result of a design research approach; it consists of dimensions containing features that are mutually exclusive and overall exhaustive (Nickerson, Varshney and Muntermann, 2013). Our goal is to synthesize the literature on the ICO phenomenon and examine what type of research academics are employing and what topics are being studied. We have identified research topics as a set of dimensions. It comprises eight dimensions: field of investigation, focus, actors, token type, additional topic, research question, ICO phase, blockchain. In the taxonomy process, we assigned a single value to each dimension.

Research topics

Field of investigation. We used ICO bench categories to classify a specific study area for each work. ICO Bench is a review platform that identifies 26 categories (Investment, Cryptocurrency, Banking, Software, Legal, Communication, etc.).

Focus. We mean the main aspects discussed in our data set: technological, organizational, individual, social, geopolitical and legal aspects.

Actors. Each paper content can be analyzed from the point of view of two main actors: capital seekers (Investee), capital providers (Investor) or both.

Token type. We considered the general classification of token type: utility, security and payment. Utility tokens give holders access to a product or service, but do not grant holders rights equivalent to those granted by specific investments. A security token offers rights and obligations similar to securities or investments such as stocks or debt securities. Payment tokens are used as an alternative means of payment and exchange.

Extra topic. This dimension includes papers that discuss not only the ICO phenomenon, but also other financial tools such as crowdfunding, IPO, and venture capital.

Research issue. This dimension relates to the type of aspects that each work seeks to examine, each work covers gaps in the literature in a specific area and its purpose is to analyze: technical aspects, token market, information asymmetry (investor-investee) and ideological aspects ( Entrepreneurial behavior, risk tolerance of investors).

ICO phase. Basically, every ICO can be divided into three phases: Pre-ICO, ICO-Launch and Post-ICO. Each work can be assigned to a specific ICO phase. Pre-ICO refers to ICO planning and marketing services. The ICO phase refers to the actual launch and development process. Once the ICO launch and development process is complete, it will be opened for ICO token sale and exchange (post-ICO).

Blockchain. This dummy dimension relates to the presence or absence of blockchain as an issue.

Table 1. Research topic dimensions

RESEARCH TOPICS
*DIMENSION*	*VALUES*	*DISTINGUISHING ATTRIBUTES*
Field of Investigation (Icobench categories)	Investmentquantiative	According to finance, the practice of investment refers to the buying of a financial product or any valued item with anticipation that positive returns will be received in the future.
	Cryptocurrency	Cryptocurrencies are digital financial assets, for which ownership and transfers of ownership are guaranteed by a cryptographic decentralized technology (Boreiko, Ferrarini and Giudici, 2019).
	Banking	“Banking means accepting, for the purpose of lending or investment, of deposits of money from the public, repayable on demand or otherwise, and withdrawable by cheque, draft, order or otherwise.” (Banking regulation act, 1949).
	Software	Software comprises the entire set of programs, procedures, and routines associated with the operation of a computer system. (www.britannica.com)
	Media	The means of communication that reach large numbers of people, such as television, newspapers, and radio. (Collins’s dictionary)
	Legal	It comprises papers that explores legal issues.
	Platform	Platform stands for Blockchain platform. It represents a platform that exists to support a particular flavour of Blockchain such as Ethereum, R3, Ripple.
	Internet	It comprises papers that explores only internet field as a media tool.
	Smart contract	It comprises papers focused on Smart contracts. Smart contracts are computerized transaction protocols that execute terms of a contract. (Nick Szabo, 1998). Smart contracts permit trusted transactions and agreements to be carried out among parties without the need for a central authority or a legal system.
	Artificial intelligence	The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. (Oxford dictionary, 2017)
	Communication	It comprises papers that explores almost one type of communication: verbal and non-verbal communication and written communication.
	Energy	It comprises papers that explore energy sector. The energy sector is a category of stocks that relate to producing or supplying energy.
Focus	Technological	Papers that explore technical aspects related to the ICO: innovative tools, bonus, soft and hard cap, platform.
	Organizational	Papers that point out organizational structure of companies: functional, divisional, matrix, flat.
	Individual	Documents that consider more aspects related to the psychology of the individuals involved.
	Social	Documents that explore social projects. Socially responsible ICOs aim to improve public wellbeing in education, environment, health and poverty.
	Geopolitical	Documents that explore geopolitical issue related to ICOs: how governments can affect the development of new ICOs.
	Legal	Documents that explore legal issues: how ICOs face different legal issues in different countries.
Actors	Investor	The person who invests money in order to make a profit.
	Investee	The business entity in which an investment is made.
	Both	Both investor and investee.
Token type	Utility	Utility Tokens grant holders access to a current or prospective product or service but do not grant holders rights that are the same as those granted by specified investments.
	Security	A Security Token provides rights and obligations similar to securities or investment like share or debt instruments.
	Payment	Payment Tokens are used as an alternative means of payment and exchange.
Extra topic	Crowdfunding	Crowdfunding is most commonly defined as “the efforts by entrepreneurial individuals and groups—cultural, social, and for-profit—to fund their ventures by drawing on relatively small contributions from a relatively large number of individuals using the internet, without standard financial intermediaries” (Mollic, 2014).
	IPO	We refer to IPO as the first offer of shares of a private company to public.
	Venture Capital	Venture capital investment consists in the purchase of shares of young, privately held companies by outsiders for the primary purpose of capital gain (The Oxford handbook of entrepreneurship, 2006 chapter 14)
	Not present	Absence of extra topics in documents.
Research Issue	Information Asymmetry	Documents that explore investor-investee relationship.
	Ideological aspects	Documents that explore ideological aspects (entrepreneur behaviour, investor’s risk tolerance).
	Technical aspects	Documents that cover literature gap in technical aspects: platforms, smart contracts, token price.
	Token market	Documents that give an overview of token market.
Ico phase	Pre ICO	It refers to a sale of a limited number of Tokens or Coins before the actual ICO (Initial Coin Offering) takes place.
	ICO	It refers to the launch of ICO.
	Post ICO	It refers to post ICO performance.
Blockchain	YES	Documents that deepen Blockchain topic.
Blockchain	NO	Blockchain topic is not explored.

5.5 Preliminary evaluation and discussion

From a preliminary analysis of our dataset, we identified several outcomes. First, the papers examine issues mainly related to the token market (42), technical aspects (32), and information asymmetry between capital seekers and funders (19). Record documents mainly focus on technological (30), social (19) and legal (18) aspects. Several characteristics and goals discussed in our dataset documents can be examined by combining two or more dimensions.

We have proposed a preliminary analysis of the ICO literature using a subset of 3 dimensions in two different combinations:

In the first analysis we considered the following dimensions: focus, research issue, actors (Fig.4).
In the second analysis we considered the following dimensions: focus, research issue, ICO phase (Fig.5).

This analysis can provide insight into the discussion of the ICO phenomenon in the posts that focus on the relationships between funders and seekers (stakeholders) and what phase of the ICO process is.

Figure n.4 shows the resulting diagram combining three specific dimensions: Focus (technological, social, organizational, legal, individual, and geopolitical aspects), Research issue (ideological aspects, information asymmetry, technical aspects, and token market) and Actors (capital seekers and capital providers). Looking at the diagram, it appears that most documents consider both investor/investee relationships (61), most of these focus on technological, social and legal aspects, exploring the token market and technical aspects.

Figure n.5 shows the resulting diagram combining three specific dimensions (the first two are the same of the first diagram): Focus (technological, social, organizational, legal, individual, and geopolitical aspects), Research issue (ideological aspects, information asymmetry, technical aspects and token market) and ICO phase (pre Ico, ICO, post ICO). Looking at the diagram, it appears that most documents explore ICOs without considering pre or post ICOs launch (74), most of these focus on technological and legal aspects, exploring technical aspects and the token market. Few documents (6) explore post ICO services.

5.6 Considerations and conclusions regarding the proposed taxonomy

The proposed taxonomy could be useful for both practitioners and academics. These results offer several theoretical contributions to the ICO phenomenon. Initial previous studies have not developed a taxonomy of academic and non-academic discourse related to this innovative financial tool. Second, this study contributes to innovation research by identifying eight dimensions in the research topic, which combined appropriately can reveal any gaps in the literature or identify the most researched topics. Third, the study helps point out the emerging research problem related to information asymmetry and investor/investee relationships. The results of this study also have implications for investors and venture capital companies to raise awareness of this funding vehicle, as well as some policy implications for countries where this phenomenon is unregulated.

Although the present work fills some of the gaps in the literature review on ICOs, it also presents several limitations that offer opportunities for future research. First, the dimensions of research topics relate to a personal interpretation of the present authors, so the dimensions can differ significantly from those of other researchers. Second, the dataset is limited to 99 documents and growing interest in the ICO phenomenon may determine a larger dataset. Third, a second set of dimensions related to research type (research approach, research design, data collection) could be examined. Therefore, a further literature search process is recommended to enrich the dimensions of the research topics as well as the number of contributions of the dataset. Future work could also explore several new dimensions related to research approach and research design adopted by researchers. Finally, it might be useful to use the present taxonomy to examine other innovative financial instruments such as Initial Exchange Offering and Security Token Offering (Za et al., 2018).

Chapter 6: The Role of White Papers’ linguistic content

6.1 Teoretical framework

Drawing on signaling theory (Spence, Heleich and Stapp, 1973), which aims to reduce information asymmetry in the investor-investee relationship, the purpose of my research is to understand how white papers can help reduce investor skepticism. Spence showed that, under certain conditions, well-informed agents can improve their market performance by sharing their private information with ill-informed agents (G. Akerlof, M. Spence, and J. Stiglitz 2001).

Signaling is the idea of one party (the agent) passing information about itself to another party (the principal). In Michael Spence’s job market signaling model, (potential) employees send a signal to the employer about their performance level by improving their educational skills. Informational value comes from the fact that the employer believes there is a positive association with greater ability. In a similar context, signaling theory has been used to explain what types of non-technical information drives investors to invest in startups. This stream of literature mainly focused on the signaling of startups in terms of technical aspects, board and top management characteristics, gender, the presence of venture capitalists or angel investors, and the involvement of founders.

This research enriches the existing literature on signaling theory by introducing white papers on linguistic content analysis, more specifically how positive emotions emanating from white papers can affect the amount of money raised.

A white paper is an official document that describes the problem that the project is trying to solve. White papers are relevant to provide information for potential ICO investors, which is a crucial tool to build investor confidence and consequently could be an important factor in determining the amount of funds raised.

White papers contain a range of information on IT protocols, adopted public blockchain, token delivery, pricing and distribution mechanism and details on the project to be developed, finally a business plan including team description (Adhami, Giudici and Martinazzi, 2018). So it can be said that this tool provides detailed information about the plan of a project, technical details, budget, goals and the distribution of the tokens.

More precisely, the content of a white paper should include several points: introduction, disclaimer, table of contents, description of the market and the problem, description of the product and how it will solve this problem, tokens (how many, why, how, when), how those collected Funds are used, the team and the roadmap (cointelegraph.com). While previous scholars have found that technical white papers attract higher amounts of funding (Fisch et al., 2019), little is known about how non-technical content can help reduce information asymmetry.

This study examines whether using signals related to positive emotion words mentioned in the document can predict the amount of money raised by a company. As a result, two key questions arise:

How can emotional words mentioned in white papers affect ICO success?
More specifically, how positive emotions used in the document can improve fundraising?

6.2 Methodology

This study attempts to introduce a quantitative linguistic analysis to explore how the use of positive emotional words can boost investor confidence and consequently the total amount of tokens sold and funds raised.

To examine these questions, I used an empirical quantitative method, compiling a sample manually using mostly data from the Icobench website, which is considered the number one ICO evaluation platform.

Typically, to promote the projects, the entrepreneurial team will register on various ICO tracking websites, which will then evaluate the plans based on the information submitted. Lee (Lee and Shin, 2018) considered Icobench, based on its data and one-in-five ratings, to be the major online hub for evaluating ICOs, and found that the probability of successful fundraising increased by 19.8% for every 1 % after control increases for other project functions. I added more information from several ICO listing websites like www.icodrops.com, www.coinmarketcap.com, www.tokenmarket.net, www.foundico.com, www.icomarks.com, www.icorating. com, www.trackico.io, www.findico.io to collect the top 100 ICOs (best rating) conducted between 2015 and 2019 in the top three countries in the world by number of ICOs: USA, Singapore and Great Britain.

Rating is the result of the combination of:

– Icobench assessment algorithm that uses more than twenty different criteria on which each ICO can earn more than thirty points

– The rating that independent experts give to the ICO following its rating methodology suggestions

When an ICO is first listed on the site, an automated scoring algorithm called Benchy calculates its score based on several objective criteria that take into account: team, ICO information, product presentation and marketing, and social media. Initially, the Benchy Score represents 100% of an ICO rating on ICObench, but as the ICObench experts start giving their opinions, the importance of the Benchy Score decreases significantly (https://icobench.com/ratings).

I have collected the top 100 ICOs completed in the US, Singapore and the UK and collected their respective white papers. I used generalized linear regression analysis to find the relationships between the amount of funds collected (dependent variable) and the positive emotion score (independent variable) from Linguistic Inquiry and Word Count software.

LIWC is a transparent text analysis program that counts words in psychologically meaningful categories. Empirical results using LIWC demonstrate its ability to discern meaning in a variety of experimental settings, including depictions of attentional focus, emotionality, social relationships, thinking styles, and individual differences (Tausczik and Pennebaker, 2010).

Research suggests that LIWC accurately identifies emotions in language use. For example, positive emotion words (e.g. good, nice, successful) are used to write about a positive event and more negative emotion words (e.g. hurt, ugly, bad) are used to write about a negative event (Kahn et al., 2007). LIWC ratings of positive and negative emotion words correspond to human ratings of the written excerpts (Alpers et al., 2005).

The impact of realizing that language can be quantified to reveal suggestions about a person’s underpinning psychology is hard to overdo. The more a person habituated words from certain vindicated language orders (like words about wrathfulness or family), the further these generalities sounded to be a central dimension of that person’s psychology.

But by the late 1990s, psychologists began to discover that language was further is than its content. Up to this point, nearly every textbook analysis fashion in psychology has concentrated on what we call happy words (also called open class words because they can take numerous different forms, similar as run, run, and run). Content words convey some type of meaning (e.g. who, what, where, etc.) and make up the maturity of the words in a person’s vocabulary (Boyd, 2017).

Variables

Dependent variable: amount raised (ln)

The amount of funding raised in the ICO is the dependent variable (in USD). This kind of dependent variable is generally used in entrepreneurial finance research (Bepari, Rahman and Mollik, 2014).

The data have been extracted from Icobench when available. I used a natural log transformation to consider the skewness of the variable (Fisch et al., 2019).

The study examines the potential positive emotion signals that constitute the independent research variable.

The words used in the top 100 white papers for the US, Singapore and UK were analyzed using LIWC (Linguistic Inquiry Word Count) software. LIWC reads a given text and counts the percentage of words that reflect different emotions, thought styles, social concerns, and even parts of speech. Because LIWC was developed by researchers with interests in social, clinical, health, and cognitive psychology, the language categories were created to capture people’s social and psychological state (http://liwc.wpengine.com/how-it-works /).

The way LIWC categorizes Posemo words has been described below, for example by testing the NAGA white paper, an American ICO rated 4.1 by Icobench.

Figure – Excerpt from NAGA white paper

LIWC highlights Posemo words in the document, as a result the software gives the percentage of words that reflect this type of emotion (in this case: 2.71%).

Control Variables: Characteristics of the venture

To rule out confounding effects and examine additional determinants of the amounts raised in ICOs, the analysis includes control variables. Two variables relate to the company’s characteristics, while one relates to the rate assigned by Icobench.

Team members (ln). I looked at the natural log of the number of components in the ICO team. Icobench provides first and last names, pictures and Linkedin URL of the entire team.

Category. Icobench identifies 29 ICO categories listed below:

1-Art

2-Artificial Intelligence

3-Banking

4-Big Data

5-Business services

6-Casino & Gambling

7-Charity

8-Communication

9-Cryptocurrency

10-Education

11-Electronics

12-Energy

13-Entertainment

14-Health

15-Infrastructure

16-Internet

17-Investment

18-Legal

19-Manufacturing

20-Media

21-Other

22-Platform

23-Real estate

24-Retail

25-Smart Contract

26-Software

27-Sports

28-Tourism

29-Virtual Reality

Control Variables: Characteristics determined by external evaluatators

Ico rate (ln). I used the natural log of the ICO rate assigned by Icobench. All ICOs are scored under the same condition by the same scoring algorithm that takes into account four different areas: team, ICO information, product presentation, marketing and social media.

6.3 Results

The table below shows the matrix correlation. We can see a negative correlation between the amount of funds raised and the use of positive emotions in white papers.

CORRELATION	lnAMOU~D	POSEMO	lnTEAM~S	lnRATE
lnAMOUNTRA~D	1.0000
POSEMO	-0.0931	1.0000
lnTEAMMEMB~S	-0.0144	-0.1209	1.0000
lnRATE	0.0236	-0.0341	-0.0842	1.0000

In this research, I adopted a generalized linear model to find possible relationships between variables.

Only control variables were included in the first table.

Model 1				No. of obs = 103 Log pseudolikelihood = -180.1868396
lnAMOUNTRA~D	Coef.	Robust Std. Err.	z	P>\|z\|	[95% Conf. Interval]
lnTEAMMEMB~S	-.021974	.3961703	-0.06	0.956	-.7984535	.7545055
lnRATE	1.219464	3.062748	0.40	0.691	-4.783411	7.22234
Categories (1-29)	yes	yes	yes	yes	yes	yes

All variables were included in the second table.

Model 2				No. of obs = 103 Log pseudolikelihood = -178.8665086
lnAMOUNTRA~D	Coef.	Robust Std. Err.	z	P>\|z\|	[95% Conf. Interval]
POSEMO	-.2673694	.1611828	-1.66	0.097	-.5832818	.0485431
lnTEAMMEMB~S	-.1072171	.4218471	-0.25	0.799	-.9340221	.719588
lnRATE	1.342171	2.955611	0.45	0.650	-4.450719	7.135062
Categories (1-29)	yes	yes	yes	yes	yes	yes

Surprisingly, we can see a negative correlation between positive emotions and the amount of money raised. How should we interpret a negative correlation? Why don’t positive emotions mentioned in white papers seem to imply fundraising?

We must remember that Venture’s goal is to earn the trust of investors. White papers are a useful tool for building trust, especially when venture capital companies use positive language.

Australia’s Markets Authority has issued a document stating that ICOs have the potential to make an important contribution to companies’ ability to raise capital and investment opportunities for investors, while clarifying that an ICO must be conducted in a manner that promotes investor confidence and trust and complies with relevant legislation (Adhami, Giudici and Martinazzi, 2018).

The goal of new ventures is to reduce information asymmetry. They believe that writing a good white paper is a way to achieve your goal, especially if you include positive emotional words. The investor is probably much more culturally prepared in terms of financial risk and is not fooled by sheer dialectics. Potential investors are aware of the grey areas surrounding the financial regulatory system.

As a result, the positive words in a white paper are often not enough, as there is often a lack of information that inspires confidence. Therefore, positive emotional words do not play a sufficiently large role in reducing the information asymmetry between investors and the ICO team.

6.4 Implications, Limitations and future research

This study will contribute to the corporate finance literature by introducing a quantitative linguistic analysis related to the white paper of ICOs. More specifically, the study aims to capture the psychological states of investors, reflecting their different types of emotions. The research will also have an impact on practitioners as they inform potential companies on how to write a white paper that can attract larger funds.

Several issues related to data accessibility and quality limit the generalizability of the results. Most of the data was collected manually and some ICOs were excluded due to lack of white paper or amount of money raised.

I wasn’t able to collect data for some variables that could affect the amount raised at ICOs, such as media activity.

In addition, the number of observations is limited to three countries and one hundred ICOs. The results will be statistically more significant as more ICOs are added.

Future research could explore other variables related to psychological states in text analysis, such as B. Sadness, anger and fear.

In addition, academics could interview investee companies to collect data. Researchers could contact the CEO of ICO via Linkedin, Twitter, Facebook, Reddit, Telegram to collect information not only on technical but also on psychological aspects.

Examining the relationship between ICO and post-ICO performance could be another interesting point. While most companies intend to develop a product, these companies often do not have a working product at the time of the ICO (Fisch, 2019).

Chapter 7: Meaning Extraction Method: an approach for extracting a topic from a Whitepaper document

7.1 LIWC and Meaning Extraction Method

The main strength of the LIWC wordbook is sustained by the fact that it has been precisely developed using established standard psychometric approaches similar as confirmation of external cerebral data, as well as ways that insure high internal trustability from a statistical point of view.

Utmost textbook analysis styles can be distributed as top-down or bottom-up. Top-down styles like LIWC start with a predefined set of words that researchers search for that they know are related to important cerebral processes. Popular top-down styles range from wordbook- grounded approaches to the most sophisticated supervised literacy algorithms in the natural language processing world (www.liwc.com).

You might want to know what the individuals in a data collection are discussing, though. What are the recurring topics or problems raised in a sample? This is where bottom-up techniques can be useful, notably a group of techniques known as topic modelling. Entwistle (Entwistle et al., 2021) sought to identify the categories of relationship issues that are frequently discussed online. They discovered that persons who discussed their relationship problems online frequently discussed topics including money, physical closeness, and leisure activities using a topic modeling methodology known as the meaning extraction method.

The Meaning Extract Method (MEM) is a way of automatically inferring which words are used together, basically performing in a wordbook of word- to- order mappings from a collection of textbooks. Simply put, this is achieved by chancing words that naturally combine into motifs using some introductory statistical ways. In practice, the MEM can be viewed as a series of way taken by a researcher. This process begins with a collection of textbooks and leads to psychologically significant word clusters (Boyd, 2017):

Identify words in a corpus of text that occur relatively frequently;
Make a table showing which texts use which common words;
Statistically identify words that frequently occur together throughout the corpus

The Meaning Extraction Method (MEM), created by Chung and Pennebaker (Kim and Chung, 2018), is one of many topic modeling approaches used in text analysis. The MEM approach, on the other hand, is distinct in that it is based on psychometric theory and statistical techniques commonly used in the social sciences. The MEM has evolved over time and now consists of a few steps to go from text to extracting topics.

In this study, we follow the same process used by Boyd to identify topics and themes in a word text:

Identify words in a corpus of text that occur relatively frequently. The MEM operates in this manner to ascertain how words are commonly employed in each sample, similar to other topic modelling techniques. In order to understand trends across different people and texts, rather than just one person’s peculiar usage, we must first identify the words that appear in enough different texts.

Additionally, when conducting the MEM, we are more interested in the ideas that each word expresses than in specific word variations. Lemmatization, which reduces words to their most fundamental form, can be used to account for this (e.g. “write” , “written” and “writing” are all converted to “write”). The majority of English words can be automatically lemmatized by the LIWC software.

After lemmatizing words, we create a list of all terms and calculate the proportion of texts that use each word. Following the collection of all this data, we choose which terms to keep and which to remove depending on the proportion of texts that each word appears in, often with a minimum of about 5% (Boyd, 2017). See table 2.

Make a table showing which texts use which common words. Once the most frequent terms have been chosen, we need to compile that data into a dataset so that we can use some simple statistical methods to identify common themes. The simplest approach to achieve this is to grade each text using a binary system, which determines if a text uses each word or not. See table 3.
Statistically identify words that frequently occur together throughout the corpus. We used Principal Component Analysis (PCA) as a method for finding correlation groups, essentially finding groups of words that tend to be used together. Statistical packages (such as Stata or R) have ways to run a PCA rather easily. For example, a PCA would likely find that the words “token”, “bitcoin” and “cryptocurrencies” form a meaningful theme about “token”. See table. So what the PCA has done for us is the identification some word clusters that can be thought of as new language categories to be measured for psychological purposes or topic modelling. See table 4.

Table 2 Example of a frequency list generated by MEM

Table 3 – An example of Step 2 in the Meaning Extraction Method: a binary table that reflects which texts used common words in a corpus

Table 4 – Example results from a Principal Components Analysis

7.2 Methodology

We consider a dataset of 434 ICO White Papers from UK, USA and Singapore (the three main countries) between 2017 and 2019.

In the first step, we extract the frequency list generated with the Meaning Extraction Method using the LIWC software. LIWC allows you to select options when the analysis is started. More precisely, we have selected the following options:

N-Gram Setting: N-grams refer to phrases of “N” length. For example, “happy” is a 1-gram, “very happy” is a 2-gram, and so on. In LIWC, we can extract up to 3-grams for our meaning extraction. We have chosen 1-grams (recommended by software developers);
Skip texts with word count < …: since the meaning extraction method is based on analysis of common occurrences of words, extremely short texts tend to add a lot of noise to our results. We consider 8 words as the threshold;
Omit words: topic models generally work by finding clusters of words that share some sort of common meaning. However, a large class of words known as “function words” have no meaning in and of themselves and should thus be excluded from our MEM analysis. Stop words are words that are filtered out during text analyses; a list of stop words is referred to as a “stop list.” LIWC includes a few prebuilt stop lists that you can use to save time and avoid having to create your own. We have loaded a pre-built English list.
Pre-processing: most natural language processing methods require some “preprocessing” – that is, some “cleaning” and adjustments to the text before beginning more formal analysis to perform. Changing the forms in which words appear across a collection of texts is a very useful part of text pre-processing. Sometimes this entails standardizing a word’s spelling (for example, “color” versus “colour”), correcting spelling errors (“boyfreind” instead of “boyfriend”), and so on. The conversion of words to their simplest forms is a useful part of this pre-processing for the MEM. Lemmatization can be accomplished in a variety of ways, ranging from highly sophisticated and elegant machine learning methods to straightforward “find and replace” procedures. LIWC takes a more straightforward approach, increasing transparency and allowing users to customize their list of words to be adjusted during pre-processing. In our research, we used a pre-built conversion list.
Output type: there are several scoring systems we can use when creating a dataset that shows which texts contain which words. In the first two steps of our research, we used three output models:
1. Raw counts. This output type shows how many times each word appears in each text. This output format is quite versatile, and it is also the most commonly used format for other types of topic models, such as Latent Dirichlet Allocation.
2. Relative frequencies. This output format reflects the percentage of words in each text that each specific word accounts for. When we have a small sample size and medium-length texts, relative frequencies are usually the best method to use (e.g., 500 to 1,000 words). This output format is also useful for analyzing very long texts or samples with texts that do not have a consistent word count.
3. Binary output. This output format simply indicates whether a given text contains or does not contain each word by using 1s and 0s. This is the most commonly used form of output for the MEM and works best with short texts, such as tweets or social media posts.

First step: Identify words in a corpus of text that occur relatively frequently.

In this first step we have derived the most recurring words in our data set, considering the options indicated above. Then we refined the search by eliminating words that represent verbs, conjunctions, nouns or words that represent synonyms of our search topic (Initial Coin Offering).

Table 5. Words frequency list.

We have selected the first forty most recurring words, among these it will be our task to do a refinement, in the following steps, to select the most relevant topics of our dataset.

Second step: Binary table

For each document of our data set (therefore for each white paper) it is indicated with 1 if a certain topic is present, with 0 if that topic is absent. For example, WP number twelve sees the absence of some of the main topics such as: market, system, technology.

Table 6. Binary table that reflects which texts used common words in a corpus

This second step can give us further guidance on how to refine the data.

Third step: Principal Component Analysis (PCA)

Principal Component Analysis (PCA) and Linguistic Inquiry and Word Count (LIWC) are two different methods used in the analysis of text data. We combined these two techniques in our study to find topics in the dataset.

PCA is a dimensionality reduction technique that is used to extract the underlying structure of high-dimensional data and reduce the number of variables while retaining as much information as possible. In the context of text data analysis, PCA can be used to reduce the dimensionality of the term-document matrix, which represents the frequency of words in a text document. The goal of PCA in this context is to identify a set of principal components that capture the most important information in the term-document matrix.

LIWC, on the other hand, is a text analysis tool that categorizes words into different categories based on their meaning and usage. It provides a quantitative analysis of a text document by counting the frequency of words in predefined categories, such as emotions, personal concerns, and linguistic styles. LIWC can be used to gain insights into the content and style of a text document and to compare the language used in different texts.

7.3 Criteria adopted and results

The three heuristic criteria used to determine the number of components are as follows (Williams, 2010):

Select only those components that account for 80% of total variability
Apply “Kaiser’s Rule”: select only those components with eigenvalues greater than or equal to one, or, equivalently, those with variance greater than the mean.
The eigenvalue graph or “Screen Plot” can be used to select the number of components (enough to reproduce the starting data with a good approximation). The eigenvalue graph or “Screen Plot” can be used to select the number of components (enough to reproduce the starting data with a good approximation). Within the graph, you select the number of components that correspond to the break’s “elbow” point. The Screen Plot is created by placing the order numbers of the eigenvalues on the abscissa axis (k) and the corresponding eigenvalues on the ordinate axis (λ). Segments connect coordinate points. The smallest k such that the trend of j is strongly decreasing to the left of k and strongly increasing to the right of k will be used to determine the number of principal components to be used.

First criterion

Based on the first criterion, a number of principal components equal to the number of eigenvalues that can cover 80% of the total variability is chosen. A table is proposed below showing the percentages of each component, as well as their cumulates.

Table7. Eigenvalue percentages

It is clear that the eigenvalues are ordered descendingly and correspond to the estimates of the sample variances of the Y.

Table n. show that the first 23 values cover 80% of the total variance. So, with the first criterion, the dimensions decrease from 40 to 23, and each component has different weights for each variable. The linear combination coefficients coincide with the eigenvector components corresponding to the 23 chosen eigenvalues.

Second criterion

According to Kaiser’s rule, the first 14 eigenvalues should be taken because their value is greater than 1.

Table 8 Eigenvalues

In this case, the dimensions to be considered decrease from 40 to 14.

Third criterion

The Screen Plot shows that the number of eigenvalues must be five, as there is an abrupt change in slope at that value.

Fig.6 Screen plot after pca

It is useful to reduce the number of dimensions as much as possible for the purposes of our analysis, so it is suggested to find five dimensions (topics) in accordance with the third criterion.

The first information on the interpretation of the new variables is given in Table n. , which shows the eigenvectors. The coefficients of the linear combinations correspond to the eigenvector components corresponding to the eigenvalues.

Table 9. Principal components (eigenvdectors)

The table analysis yields five dimensions that correspond to the five main topics identified through the text analysis of white papers. To highlight the sub-dimensions and groupings made, we refined the data and inserted it into a new table.

Table 10: Example results from a Principal Components Analysis

Technically, all words will have a ‘loading’ onto each of the themes extracted from a corpus of texts. Usually, a manually chosen cutoff is used (typically around 0.20/0.25) to determine which words belong to which themes. In this table, shaded values indicate the theme onto which each word best loads using a conservative cutoff value. Words may also commonly load onto multiple themes.

As a result, our model has identified five major topics.:

Investment. Component 1 resulting from the combination of the following variables: market, cryptocurrency, exchange, fund, currency, financial, investment.
Technology. Component 2 resulting from the combination of the following variables: service, technology, system, payment.
Smart contract. Component 3 resulting from the combination of the following variables: transaction, exchange, ethereum, contract, smart.
Blockchain. Component 4 resulting from the combination of the following variables: blockchain, business, risk.
Token. Component 5 resulting from the combination of the following variables: token, platform, sale.

CONCLUSION

The goal of this Thesis was to first examine the phenomenon of Initial Coin Offerings and then the role of the White Paper in terms of linguistic content. We took, after providing a taxonomy of the academic literature, the following steps:

Academic literature taxonomy that provided us with the main topics addressed in the papers;
Analysis of positive emotions in white papers to investigate potential relationships between the amount of money raised and the use of positive words in the document;
LIWC and STATA were used to conduct a topic analysis of white papers using two different techniques: Meaning Extraction Method and Principal Component Analysis.

Following this research, we can see that the result obtained using the LIWC and Stata is very close to the result obtained manually using the paper taxonomy. However, empirical research on the topics has allowed us to narrow the number of dimensions and thus the topics.

BIBLIOGRAPHY

Adhami, S., Giudici, G. and Martinazzi, S. (2018) ‘Why do businesses go crypto? An empirical analysis of initial coin offerings’, Journal of Economics and Business. Elsevier, 100(March), pp. 64–75. doi: 10.1016/j.jeconbus.2018.04.001.

Albrecht, S., Lutz, B. and Neumann, D. (2020) ‘The behavior of blockchain ventures on Twitter as a determinant for funding success’, Electronic Markets. Springer, 30(2), pp. 241–257. doi: 10.1007/s12525-019-00371-w.

Alpers, G. W. et al. (2005) ‘Binocular rivalry between emotional and neutral stimuli: A validation using fear conditioning and EEG’, International Journal of Psychophysiology, 57(1), pp. 25–32. doi: 10.1016/j.ijpsycho.2005.01.008.

Amsden, R. and Schweizer, D. (2019) ‘Are Blockchain Crowdsales the New “Gold Rush”? Success Determinants of Initial Coin Offerings’, SSRN Electronic Journal. doi: 10.2139/ssrn.3163856.

An, J. et al. (2019) ‘Initial coin offerings and entrepreneurial finance: The role of founders’ characteristics’, Journal of Alternative Investments, 21(4), pp. 26–40. doi: 10.3905/jai.2019.1.068.

Back, A., Corallo, M. and Dashjr, L. (2014) ‘Enabling blockchain innovations with pegged sidechains’, URL: http://www., pp. 1–25. Available at: http://newspaper23.com/ripped/2014/11/http-_____-___-_www___-blockstream___-com__-_sidechains.pdf.

Batiz-Benet, J., Clayburgh, J. and Santori, M. (2017) ‘The SAFT Project: Toward a Compliant Token Sale Framework’. Available at: www.saftproject.com.

Belleflamme, P., Lambert, T. and Schwienbacher, A. (2012) ‘Crowdfunding: Tapping the Right Crowd’, SSRN Electronic Journal, (April 2018). doi: 10.2139/ssrn.1578175.

Bepari, M. K., Rahman, S. F. and Mollik, A. T. (2014) ‘Firms’ compliance with the disclosure requirements of IFRS for goodwill impairment testing: Effect of the global financial crisis and other firm characteristics’, Journal of Accounting and Organizational Change, 10(1), pp. 116–149. doi: 10.1108/JAOC-02-2011-0008.

Block, J. H. et al. (2020) ‘The entrepreneurial finance markets of the future: a comparison of crowdfunding and initial coin offerings’, Small Business Economics. doi: 10.1007/s11187-020-00330-2.

Boreiko, D., Ferrarini, G. and Giudici, P. (2019) ‘Blockchain Startups and Prospectus Regulation’, European Business Organization Law Review, 20(4), pp. 665–694. doi: 10.1007/s40804-019-00168-6.

Boyd, R. L. (2017) ‘Psychological Text Analysis in the Digital Humanities’, pp. 161–189. doi: 10.1007/978-3-319-54499-1.

Campino, J., Brochado, A. and Rosa, Á. (2021) ‘Success Factors of Initial Coin Offering (ICO) projects’, Economics Bulletin, 41(2), pp. 252–262.

Chen, Y. (2018) ‘Blockchain tokens and the potential democratization of entrepreneurship and innovation’, Business Horizons, 61(4). doi: 10.1016/j.bushor.2018.03.006.

Chiu, I. H. Y. and Greene, E. F. (2019) ‘The Marriage of Technology, Markets and Sustainable (and) Social Finance: Insights from ICO Markets for a New Regulatory Framework’, European Business Organization Law Review, 20(1), pp. 139–169. doi: 10.1007/s40804-019-00138-y.

Cipriano, M. and Za, S. (2021) ‘Exploring the Discourse on Digital Transformation in the Domain of Non-profit Organisations’, pp. 198–231. doi: 10.1007/978-3-030-87842-9_15.

Collomb, A., De Filippi, P. and Sok, K. (2019) ‘Blockchain Technology and Financial Regulation: A Risk-Based Approach to the Regulation of ICOs’, European Journal of Risk Regulation, 10(2), pp. 263–314. doi: 10.1017/err.2019.41.

Conley, J. P. (2019) ‘Blockchain and the Economics of Crypto-tokens and Initial Coin Offerings’, IT Professional, 21(3), pp. 63–66.

Entwistle, C. et al. (2021) ‘Dirty laundry : The nature and substance of seeking relationship help from strangers online’, 38(12), pp. 3472–3496. doi: 10.1177/02654075211046635.

Fisch, C. (2019) ‘Initial coin offerings (ICOs) to finance new ventures’, Journal of Business Venturing, 34(1), pp. 1–22. doi: 10.1016/j.jbusvent.2018.09.007.

Fisch, C. et al. (2019) ‘Motives and profiles of ICO investors’, Journal of Business Research. doi: 10.1016/j.jbusres.2019.07.036.

Fisch, C. and Momtaz, P. P. (2020) ‘Institutional investors and post-ICO performance: an empirical analysis of investor returns in initial coin offerings (ICOs)’, Journal of Corporate Finance, 64. doi: 10.1016/j.jcorpfin.2020.101679.

Fried, V. H., Bruton, G. D. and Hisrich, R. D. (1998) ‘Strategy and the board of directors in venture capital-backed firms’, Journal of Business Venturing, 13(6), pp. 493–503. doi: 10.1016/S0883-9026(97)00062-1.

Gompers, P. and Lerner, J. (2001) ‘The venture capital revolution’, Journal of Economic Perspectives, 15(2), pp. 145–168. doi: 10.1257/jep.15.2.145.

Hashemi Joo, M., Nishikawa, Y. and Dandapani, K. (2019) ‘ICOs, the next generation of IPOs’, Managerial Finance, 46(6), pp. 761–783. doi: 10.1108/MF-10-2018-0472.

Howell, S. T., Niessner, M. and Yermack, D. (2019) ‘Initial Coin Offerings: Financing Growth with Cryptocurrency Token Sales’, The Review of Financial Studies. doi: 10.1093/rfs/hhz131.

Hsieh, H.-C. and Oppermann, J. (2020) ‘Initial coin offerings and their initial returns’, Asia Pacific Management Review. National Cheng Kung University. doi: 10.1016/j.apmrv.2020.05.003.

Huang, W., Meoli, M. and Vismara, S. (2020) ‘The geography of initial coin offerings’, Small Business Economics, 55(1), pp. 77–102. doi: 10.1007/s11187-019-00135-y.

Kaal, W. A. and Dell’Erba, M. (2017) ‘Initial Coin Offerings: Emerging Practices, Risk Factors, and Red Flags’, SSRN Electronic Journal, pp. 0–21. doi: 10.2139/ssrn.3067615.

Kahn, J. H. et al. (2007) ‘Measuring emotional expression with the Linguistic Inquiry and Word Count’, American Journal of Psychology, 120(2), pp. 263–286. doi: 10.2307/20445398.

Kher, R., Terjesen, S. and Liu, C. (2020) ‘Blockchain, Bitcoin, and ICOs: a review and research agenda’, Small Business Economics. doi: 10.1007/s11187-019-00286-y.

Khurshed, A. (2019) INITIAL PUBLIC OFFERINGS-: The mechanics and performance of IPOs. Edited by H. H. Limited.

Kim, M. S. and Chung, J. Y. (2018) ‘Sustainable growth and token economy design: The case of Steemit’, Sustainability (Switzerland), 11(1). doi: 10.3390/su11010167.

Kranz, J, Nagel, E. and Yoo, Y. (2019) ‘Blockchain Token Sale: Economic and Technological Foundations’, Business and Information Systems Engineering. Springer Gabler, 61(6), pp. 745–753. doi: 10.1007/s12599-019-00598-z.

Kranz, J., Nagel, E. and Yoo, Y. (2019) ‘Blockchain Token Sale: Economic and Technological Foundations’, Business and Information Systems Engineering, 61(6), pp. 745–753. doi: 10.1007/s12599-019-00598-z.

Lee, I. and Shin, Y. J. (2018) ‘Fintech: Ecosystem, business models, investment decisions, and challenges’, Business Horizons. Elsevier Ltd, 61(1), pp. 35–46. doi: 10.1016/j.bushor.2017.09.003.

Lo, Y. C. and Medda, F. (2020) ‘Assets on the blockchain: An empirical study of Tokenomics’, Information Economics and Policy. doi: 10.1016/j.infoecopol.2020.100881.

Mangano, R. (2018) ‘Blockchain Securities, Insolvency Law and the Sandbox Approach’, European Business Organization Law Review, 19(4), pp. 715–735. doi: 10.1007/s40804-018-0123-5.

Momtaz, P. P. (2020) ‘Entrepreneurial Finance and Moral Hazard: Evidence from Token Offerings’, Journal of Business Venturing. doi: 10.1016/j.jbusvent.2020.106001.

Nickerson, R. C., Varshney, U. and Muntermann, J. (2013) ‘A method for taxonomy development and its application in information systems’, European Journal of Information Systems. Nature Publishing Group, 22(3), pp. 336–359. doi: 10.1057/ejis.2012.26.

Nippa, M. and Reuer, J. J. (2019) ‘On the future of international joint venture research’, Journal of International Business Studies. Palgrave Macmillan UK, 50(4), pp. 555–597. doi: 10.1057/s41267-019-00212-0.

OECD (2019) ‘Initial Coin Offerings (ICOs) for SME Financing’, Organisation for Economic Cooperation and Development, p. 32. Available at: http://www.oecd.org/finance/ICOs-for-SME-Financing.pdf.

Roosenboom, P., van der Kolk, T. and de Jong, A. (2020) ‘What determines success in initial coin offerings?’, Venture Capital. doi: 10.1080/13691066.2020.1741127.

Spence, J. T., Heleich, R. and Stapp, J. (1973) ‘A short version of the Attitudes toward Women Scale (AWS)’, Bulletin of the Psychonomic Society, 2(4), pp. 219–220. doi: 10.3758/BF03329252.

Spinedi, G. et al. (2019) ‘Guideline for enterprise to a Value Plan through Blockchain and ICO’, in Proceedings – 2019 IEEE International Conference on Engineering, Technology and Innovation, ICE/ITMC 2019. doi: 10.1109/ICE.2019.8792601.

Tausczik, Y. R. and Pennebaker, J. W. (2010) ‘The psychological meaning of words: LIWC and computerized text analysis methods’, Journal of Language and Social Psychology, 29(1), pp. 24–54. doi: 10.1177/0261927X09351676.

Williams, L. J. (2010) ‘Principal component analysis ´’, pp. 433–459. doi: 10.1002/wics.101.

Za, S. et al. (2018) ‘Exploring foundations for using simulations in IS research’, Communications of the Association for Information Systems, 42(1), pp. 268–300. doi: 10.17705/1CAIS.04210.

Zhang, S. et al. (2020) ‘Does policy uncertainty of the blockchain dampen ICO markets?’, Accounting and Finance. doi: 10.1111/acfi.12639.

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

It’s so easy to contact us

How to place an order

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order:

How to place an order: