Economics of Initial Coin Offerings

21 min readJun 2, 2017

This is an extract from this paper, providing an in-depth analysis on the structural and dynamic economic properties of Initial Coin Offerings.

This article is a thought experiment intended to shed light on several important aspects of Initial Coin Offerings (ICOs) and the economics of “Appcoins”. However, before proceeding we would first like to make the reader aware of a little known gambling phenomenon known as Pachinko.

Pachinko

Gambling for cash was illegal in Japan prior to legalization in late 2016. Until then some exceptions such as horse, boat, motorcycle and bicycle racing and lotteries existed, however this is now slowly changing [1]. Nevertheless, over the years determined entrepreneurs found simple yet effective methods to bypass Japanese gambling laws. One of the most popular methods of bypassing these laws was that utilized by parlors operating the pinball style arcade game known as Pachinko. Whilst almost unheard of outside of Japan, Pachinko caused the emergence of a thriving gambling market within Japan with approximately 11,000 Pachinko parlors nationwide generating gross revenues of $209billion in 2015, approximately 4–5% of Japan’s GDP [2] [3].

Pachinko works as follows [4]:

Players enter a Pachinko parlor and purchase/rent steel pachinko balls with Yen.
Players take the balls to a Pachinko machine (pictured above), which resembles a vertical pinball machine.
Like loading a traditional pinball machine with coins, players load their balls into the Pachinko machine and press and release a spring-loaded handle attached to a padded hammer, which then launches the ball into a metal track. The track guides the ball to the top of the machine where it loses momentum and falls into the playing field.
If one of the balls in the playing field hits certain targets during the fall through the Pachinko machine players have a chance to win more balls.
The more balls players have, the longer they can remain in the game, increasing their chances of winning more balls.
The balls, in and of themselves, have little intrinsic value as they belong to the gaming parlor. However, once players have finished they take their remaining balls (including any extra balls won in the game) to a prize booth in the parlor where they collect prizes equating to the value of their Pachinko ball winnings. If a player began with ¥10 of balls, and ended with ¥20 of balls he can collect a teddy bear worth ¥20. The key point here is that gambling for cash is not allowed in Japan, hence prizes must be anything but cash.
Now this is where it gets interesting. As the Pachinko parlor is forbidden by law from awarding cash prizes, it gives a “cash equivalent prize”. However, the Pachinko parlors have an associated, but normally separate, independent business located outside the parlor that “buys” the prizes for cash!
Effectively the steel balls and prizes are merely a promise for cash. This makes the parlors reputation of great importance as players must trust that the Pachinko balls will be redeemed for prizes that will in turn be redeemed for cash!

Pachinko parlors have built a thriving gaming (and closet gambling) industry operating subtly yet effectively in plain sight. Pachinko’s market size as stated earlier clearly shows there is either huge demand for this type of game, or that this game is a substitute for the simple and unadulterated concept of casino style gambling. The key question is which is it?

The Pachinko Effect

While the cultural phenomenon Pachinko may be as much fun as a Terminator pinball machine, we can also see that the gambling aspect of Pachinko is contrived and not an optimal substitute for the real thing. To some degree, Pachinko parlors have been one of the easiest ways for the Japanese market to fulfill its gambling demand by the only means it could easily do so, through a contrived mechanism designed to overcome legal restrictions on casino style gambling with cash. In so doing, Pachinko parlors have become a suboptimal substitute for an open gambling environment. Suboptimal because (i) Pachinko parlors can only offer one simple game that cannot compete with the wide variety and assortment of games offered by a casino and (ii) the process of gambling with Pachinko is a bit inconvenient if you just want to simply gamble!

The legalization of casino style gambling in Japan was the result of the Japanese government recognizing the opportunity to tax the sizable gambling market, increase tourism and stimulate economic growth [5]. With the introduction of these new laws, chances are the gambling market will expand as casinos start to provide a greater variety of gaming to meet a wider range of customer interests and requirements. Pachinko’s market share will likely decrease as the demand shifts toward a casino model of gambling. However there is a strong possibility that contrived business models, such as those around Pachinko parlors, will adjust to increase their appeal in the new gambling environment. Depending on the cost and availability of gaming licenses, existing Pachinko parlors may enable players to directly win cash prizes and or the parlors may shift to start offering other forms of gambling machines e.g. traditional slot machines and the more recent technologies of Fixed Odd Betting Terminals that have caused some stir in the UK [6].

It is key to note that under the new gambling laws, while establishments may be under tighter oversight and controls for the protection of vulnerable players, leading to higher operational costs, there is also greater potential for commercial opportunities and revenues, and more choice for gamblers.

Ok, so why are we talking about Pachinko in an article titled the Economics of ICOs? Well, because the phenomenon of Pachinko is an interesting one and draws similarities with the majority of contrived explanations and models in the crypto space designed to distance ICOs from securities. The construction of these “work-arounds” may be limiting the vision and creativity required to see the true scale of what ICOs may represent, and blinding many in the industry to possible risks if they take the wrong path.

Initial Coin Offerings

The introduction of Bitcoin [7] in 2009 gave us tools and infrastructure to transact primitive digital tokens of value (bitcoin in the case of the Bitcoin blockchain) over the open public Internet without trusted intermediaries. However, in order to create new tokens one either needed to deploy and scale a new blockchain network (likely forked from Bitcoin), or issue tokens on top of an existing blockchain network such as Bitcoin (through metadata encoded into raw transactions). The former was an uphill struggle due to challenges of scaling and achieving network effects for a new blockchain, and the latter was challenging due to the complexities of trying to encode sufficient information related to new tokens into raw Bitcoin transactions. Neither model was ideal.

However, with the introduction of Ethereum [8] in 2015 came the concept of decentralized smart contracts. The Ethereum blockchain not only provided the infrastructure for transacting primitive digital tokens (ether in this case) but also provided the capability for easily creating and autonomously managing other digital tokens of value over the open public Internet without trusted intermediaries.

Using this concept of smart contracts, which are effectively applications running atop a decentralized network, tokens can be created and allocated to users, and made to be easily tradable. This process of creating tokens and distributing them to users in return for a network’s primitive digital token is called an ICO process.

ICO Tokens

One of the most obvious and natural use cases for such ICO based digital token issuances is to represent some form of security e.g. equity, debt, participation in profit sharing, etc. As well as issuance, allocation and transferability being programmed into an immutable smart contract, one can also predefine a set of events such as cashflow rules that can be triggered either at set times or by specific external events. There are a number of reasons why a public blockchain infrastructure makes sense for the issuance and administration of financial securities, which are mostly related to regulations around how client money and assets are managed through their life cycle (which we will discuss in a future article).

However, since the issuance of securities is in and of itself highly regulated, several models have been devised by startups to enable the issuance of tokens whilst circumventing securities regulations. As well as the question around whether a token is a security or not there are also a number of other unanswered questions related to tax of capital gains and KYC/AML rules. These are some of the regulatory and statutory financial considerations that are currently an ongoing area of scrutiny and development [9] [10]. The focus of this article is not to delve into such consideration, but instead to try to understand the economic realities underpinning an ICO, the issued tokens and their uses.

Appcoins

One of the most prominent ICO models has been the creation and issuance of tokens called “Appcoins” [11]. This is where the issued digital token represents access to some product or service that either already exists or will exist in the future. Buying these tokens can naively be considered as purchasing a “software license” that gives the holder the “right” to access the final product or service. However, unlike normal licenses these issued tokens are easily transferable, either directly between users (over the counter) or through an established cryptocurrency exchange. This ease of transferability of the tokens on an exchange enables liquidity and thus drives price volatility based on the market’s perception of the issuing project.

The theory is that if appcoins provide access to a future product or service their value will increase as the product is launched and its usage increases. Therefore, early adopters (ICO participants) benefit from the upside through capital gains of the issued tokens. The economic dynamics and impacts on stakeholders of such appcoins are considered in this article.

Gnosis

As an example, one of the most prominent ICOs has been the Gnosis issuance, which raised the equivalent of $12million in Ether in less than 15 minutes of commencement, but issued only 5% of its total supply of tokens (GNO). This means the total market cap of the tokens post-issuance was ~$300million, and all before any commercially viable product was even built [12]! The Gnosis publications give a sense of their ICO structure, how the GNO tokens work and the possible value a buyer is getting [13]. The “Terms of the Token Sale” by Gnosis Limited [14] reveals the depth of legalese surrounding the issuance, the kind one would find in a securities private placement, but there is little or no discussion about the economic complexities and resulting risks of the ICO or the issued appcoins. However, even for someone with sufficient expertise in financial engineering, risk and technology it is not straightforward to quickly decipher the convoluted nature of how these tokens work, their possible economic dynamics and the potential value and risks they may represent.

The purpose of this article is not to provide a critique of the Gnosis issuance, but instead to provide a framework on analyzing the economic realities behind such appcoins in order to understand the value and risks they may represent. Ostensibly from the above Gnosis ICO this is not a simple process, so we will attempt to unbundle the complexity by using a simple and straightforward ICO thought experiment for creating a Laundry service.

Laundry Tokens

Imagine we are industry experts in providing laundry services, and we publicly announce a plan to open a new chain of laundry shops, where the development will be funded through an ICO process. The tokens would be issued at a price of 1 Laundry Token (LTX) for $1 and will enable holders to gain access to laundry services in all laundry shops that are a part of the new chain. However, we will only ever issue 1 million LTX, thereby creating artificial scarcity. In addition, this will be done through an independent third party to ensure we are unable to defraud the system i.e. the independent third party will ensure that we cannot issue more LTX and or recall LTX already issued to people, and ensure the LTX already issued will be honored for their intended purpose. Furthermore, imagine that the LTX price can go up and down based on supply and demand, and that they are easily transferable between users at the fair market rate. Under the efficient market hypothesis, the value at which the tokens are exchanged in a sufficiently liquid market will capture all the market information relating to the underlying project. The theory is that as we issue the LTX token and use the funds to build the laundry business, the value of the LTX will go up (on anticipation of a successful launch) and so LTX holders will enjoy the benefits of capital gains. On the other hand if the project begins to struggle and bad news enters the media, the value of the LTX will begin to decrease and LTX holders will suffer capital loss.

In order to ensure that the economics of the LTX works as an appcoin the issued tokens need to be convertible for their intended purpose of receiving laundry services i.e. holders should be able to take their LTX to any of the shops in the laundry chain and pay for required services. LTX can be divisible into subunits, but for simplicity in this example when a single LTX token is redeemed in a shop it is converted to laundry credits equivalent to the market rate in USD. Therefore, if 1LTX is now being traded for $50, then on redemption at the shop the holders can gain access to $50 of cleaning credits e.g. they can clean five jackets where the cost of cleaning is $10 per jacket. As the value of the tokens goes up, on the open market, holders can gain more laundry credits “for free” on redemption i.e. more jackets cleaned for the same units of tokens. Sound convoluted? Well this is simplified compared to how the Gnosis tokens work, where GNO=LTX and WIZ=laundry credits!

Token Fundamentals

However, the economics of the ICO and appcoin model needs to be viable to ensure sustainability i.e. a token skyrocketing in value may be damaging in subtle ways to the issuance business compared to the obvious issues of a token plummeting to zero. To build a basic framework for such analysis it is important to begin with reviewing some fundamentals related to tokens.

Monetary Policy

A token’s Monetary Policy means the model for supply release and the cap on total supply i.e. how many tokens are issued and how often, and what the total number of issued tokens will be. A capped and well-controlled supply release increases the chances of a small increase in demand driving token prices higher. Normally the monetary policy would be pre-defined as part of the issuance strategy, where a fixed number of tokens are created and issued. However, even though there is a total supply cap the issuer would only distribute a certain fraction of the available tokens to raise a fixed amount of capital for executing the business plan. The remaining tokens are then held in an “escrow” type service to finance operational costs or future connected projects. For example, in the previous case of the laundry LTX ICO the supply was capped at 1 million tokens, and then we may only distribute 500k LTX tokens and keep the remaining in an escrow account, which we can use to cover costs of running the business and or expanding laundry shops in the future. The escrow account would likely have some form of access/usage controls to provide comfort to investors that the tokens held will not be dumped (sold in one go), in turn causing a price crash; or they may be locked for a fixed period to allow sale in a controlled manner over a “sufficiently long” period of time. All of these aspects fall under the monetary policy of a token as they are related to directly managing the supply of tokens in circulation, and is a relatively well-understood concept in cryptocurrency.

Fiscal Policy

It is also important to understand and define the commercial benefits the ICO participants gain from holding tokens, beyond just the capital gains related to scarcity. This point is a key one, and one that is least talked about and or understood, but is just as important as a token’s monetary policy. In the case of the laundry ICO earlier, as we are ultimately developing a business we want to maximize the value being created so we will potentially offer laundry services in the issued LTX but also in USD fiat currency (and even in other cryptocurrencies BTC, ETH and ETC). Linking a commercial benefit (e.g. discounts) with token usage means customers would be more likely to access our services through LTX rather than any other form of payment, this is especially true if there is a large supply of tokens in circulation (resulting in less scarcity).

To drive continued customer interest in buying LTX, as issuers we can ensure that LTX holders always gain some benefits/discounts on the services offered e.g. rather than cleaning costs for a jacket being $10 the LTX holders may only pay $8. This discount may be adjustable so we can manage commercial benefits based on levels of external competition, changes in operational costs and other unknown factors. This becomes a way of managing the flow of the issued token without taking drastic actions related to monetary policy e.g. increasing/decreasing supply from circulation or even hoarding/dumping tokens. This control of flow of tokens and impacts on aggregate supply and demand is a form of “Fiscal Policy”. The fiscal policy actions highlighted in this paragraph are not directly connected to managing the supply of tokens in circulation but rather connected to managing the flow of tokens through indirect economic incentives.

One example of a potential use for such a fiscal policy mechanism is that an issuer can propose to increase the commercial benefit (e.g. cleaning discount in LTX), which will increase aggregate demand of the tokens (from D1 to D2 in diagram below). Such an action can then be combined with monetary policy decisions, for example as the aggregate demand of the tokens increases through the fiscal policy decisions, the issuing company could then also release further tokens, that may be held in escrow, increasing the total supply in circulation (from S1 to S2 in diagram below). This combined increase of supply in circulation and the demand due to increased commercial benefits may have a minimal impact on the current market price. This simple thought experiment can be visualized as in the following diagram:

This model, taking into account monetary and fiscal policies (which are fundamental in macroeconomics), and applying a framework, which allows them to interact, and effect aggregate supply and demand, shows how critical these policies are to understanding ICO structure, appcoins, and their impact on future issuing projects/companies and the capability for value creation and risk causation for token holders.

Supply and Demand

As seen above, from a monetary policy and fiscal policy perspective, the balance of Commercial Benefit and Supply Scarcity factors is critical in planning the issuance of a sustainable appcoin and the impacts on a business (see the diagram below). The economic analysis for such a balance can be modeled through assessing the supply and demand surfaces based on these explanatory factors. As a thought experiment it is viable to assume that highly scarce tokens with very high commercial benefit (e.g. offering steep product discounts) may result in hyper deflation of their value, leading to a hoarding mentality, as there will be a view of “falling prices” or being able to purchase more with the same tokens if they delay redeeming. While this may seem like a positive effect, as a higher purchasing capability of a token implies higher returns for a company when they sell a redeemed token back into the market, such a scenario also has the potential to detrimentally impact a business and its cashflow. One of the most obvious impacts is that whilst a hoarding mentality will result in people delaying the use of their tokens, accordingly customer perceptions of using fiat currency to pay for services could cause them to feel like they are getting a worse deal compared to buying with a token, hence driving them to also delay their purchase, or using substitutes and or competitive products/services. This business slow-down will, in turn (assuming an efficient market), lower the value of the token until an equilibrium point is reached. However, the equilibrium point may not be what makes this a highly profitable business. On the other hand, tokens with low scarcity and minimal commercial benefit may be of little interest to users either as “investments” or for access to services. Evidently there is a level of complexity here that is very challenging to assess without some historic and forecast sales revenue/cost data related to a company’s core product or service and its customer profiles.

The Optimal Model

From an issuer perspective, as the tokens are redeemed by holders for their relevant commercial purposes, for which they may receive company credit (e.g. coat cleaning in the laundry example) the company can resell the token in the market at the current market rate. Using this approach the company can recover operational costs and generate profits. This cycle of token purchase and redemption by customer and resale into the market by issuer can continue, meaning the appcoin tokens would become a form of money that is only locally accepted by the issuing company. From this context it becomes apparent why the above framework considering monetary and fiscal policies and their interactions makes sense, as “a long line of research emphasizes that separating monetary and fiscal policies overlooks policy interactions that are important for determining equilibrium” [15].

As an appcoin model, the above may be the most viable basic structure making business and economic sense. Conversely, an example of a nonsensical token structure would be a 1-to-1-access model wherein the issued tokens are just for single-use only, then their market value would intrinsically be capped at the value of substitute offerings. For example, if 1 token represents 1 unit of service, and competitors charge $10 for the same service (all other things being equal), then it is irrational for token purchasers to pay more than $10 for one token on the open market (assuming no commercial discount for token holders).

In addition, issuers must account for the time it takes to sell the token in the open market after being redeemed by a holder for its intended commercial purpose. In a sufficiently liquid market this may be “instantaneous”, thereby minimizing the risks associated with price volatility. From a commercial perspective these risks could be severe because hoarding tokens to control market supply may backfire if token value falls. In this scenario the issuing business may not be able to sell the tokens it holds to recover costs incurred delivering the services in return for redeemed tokens. While risks can be hedged if appropriate instruments/services are available, the optimal model would see smaller companies refrain from hoarding, instead preferring to sell in sufficient quantities to at least recoup costs. This is also the model used by blockchain miners.

Leading on from this is the challenge of “retiring” tokens, for example if an issuing organization at some point in the future decides to discontinue a particular product/service, or move away from a token model. In this scenario the token holders can effectively be seen to have some form of right or vote on the future of the product or service offered by the company! This challenge becomes greater than just a simple vote because as the value and or usage of these tokens increases, the holders will be seen to have more influence over company decisions. The optimal and simplest approach would be for the issuing company to buyback the tokens from holders at the market rate; a process which must be managed appropriately so that any announcements and buyback processes will not result in liabilities for the company. In this context, it must be noted that retiring any form of “money or security” from circulation (if these appcoin tokens are thought of as such) is a huge challenge. On a much larger scale this is evident in the issues that occurred in India with the heavy-handed banknote demonetization strategy in 2016 [14]. In the case of an appcoin the scale would be much smaller, but as a company grows and token values and or usage increases the challenges could be similar, especially when the tokens represent some form of access to a product or service that people have become attached to, or is critical to operations for which they still require some form of support from the issuing company (think Windows XP). A prudent way to manage this would be to have a buyback/demonetization strategy as part of an appcoin ICO, and even some functionality to manage this autonomously in a controlled manner.

Viable and non-Viable ICOs

Executing an ICO, issuing appcoins and managing the micro/macro-economic impacts on the business, key stakeholders and the market will mean companies in the future may not only need a CFO but likely also a Chief Economist, as running such a firm will be akin to running a small country! This will be more of a problem for a startup or a fledgling company, but maybe not so for an ICO executed by a larger firm with a more established product or service.

If a large, established company e.g. Spotify, Netflix, and even PornHub (which is probably the most likely considering the history of pornography and its influence in driving the uptake of new innovative business models [15]), executed an ICO and issued an appcoin, for managing operating costs or expansion of services, it may make commercial and economic sense to do so. In this scenario the economics can likely be modeled with an acceptable level of confidence due to good availability of sufficient historic and forecast operational and financial data. Therefore, an appropriate token structure can be setup and deployed for the ongoing benefits of participants and sustainability/growth of issuer’s business model.

In a startup context an appcoin ICO would likely only make sense commercially/economically for a series A (or greater round) when a decent product-market fit has been defined and scale up capital is required. From this context it can be seen how an appcoin ICO model may even be hugely disruptive for later stage venture capital firms, contrary to their current views [16].

In either of these cases, for established companies and or scale-up firms, the key question then is why use an ICO on a public blockchain if the appcoin tokens are not securities? If appcoins are not securities and they can be easily and widely marketed to retail investors, and the token life cycle can be managed without the challenges of complex client money and asset rules, why use a smart contract blockchain to execute an ICO? Why not just use cryptocurrency payments (or any form of payments) with the appcoin mechanism managed through a centralized system?

Some Legal Aspects

No discussion on ICOs would be complete without providing some legal points of views. We will touch on some of these here and go into more depth in an upcoming paper, especially regarding those less talked about aspects related to tort law and negligence.

In the laundry example, ultimately if we issued the LTX for our laundry project we may be raising risk capital. Why? Because the economic reality of this is that people are giving us money now to execute a project where they will gain some benefit in the future, and this benefit goes beyond a simple one off product or service. Unlike reward-based crowdfunding, such as a Kickstarter campaign, the upside for an ICO participant is not limited to receiving a single product or service. An ICO participant is essentially buying into the on-going commercial viability and success of the issuing company, and participants anticipate that the value of the token(s) they own will increase, in which case they can either earn a greater financial return by selling the tokens (capital gains) or gain access to more services at some discount (commercial benefits). Either way an ICO participant’s expected returns may be seen as linked directly to the pooled funds of the ICO and efforts of the issuer to use funds in executing the project and ensuring viability of the ongoing business. It is this expectation that drives participants to not only pay for a one-off product or service (as they may on Kickstarter) but to buy into an ongoing commercial proposition that is yet to be developed and commercialized, and all for the expectation of some form of “unlimited” (explicit/implicit) financial gain.

Due to such complex economic realities of ICOs and the resulting appcoins, it seems like these are a new form of asset class. If they do turn out to be a new type of primitive security, then a decentralized model for administration may be viable to ensure compliance with key client money and asset rules. However, this is a discussion for another time.

While the above complexities are related to statutory/policy/legal challenges, if an ICO is seen as a public offering and the resulting appcoin is considered a security, other less obvious challenges could result from common law issues related to contract or tort violations. These may become increasingly relevant as issuing companies expand and become profitable, at which point they would become bigger targets for consumer class action lawsuits. There may be legal risks related to issues seen as violations of the ICO terms of contract, or other forms of negligence directly linked to the losses suffered by token holders e.g. if a price plummets to zero based on an issuers perceived negligence. Such legal recourse may also become increasingly attractive as an issuing company changes its business model, becomes successful, and the tokens then become less relevant. This means that in the future VCs/investors/acquirers will likely need to perform appropriate due diligence related to such risks, thus impacting possible valuations for growth companies.

Final Thought

We must highlight that while the overview provided in this work is speculative and dares to touch on things through a “thought experiment” approach (a physicists and economists best friend) it is based on sound financial, commercial and economic reasoning. The fact is that ICOs and appcoins are a new phenomenon and while we may not have the benefits of the Pachinko Parlor industry to look outside of our little bubble and see what could be possible, we have our imagination, logic and intellectual honesty to determine what is happening in this industry and where it can go.

Some consider ICOs and tokens to be like reward based crowdfunding and software licences/APIs [17], which would potentially distance them from financial securities and or other forms of economically complex financial innovations. While this abstracts the legal and regulatory burdens and generate some short-term gains, chances are it may result in us missing a bigger opportunity or even delaying a ticking time bomb.

Acknowledgements

Thanks to Vic Arulchandran for his contributions to the refinement of this and upcoming work.
Thanks to Alexander Ainslie for his insights on the Japanese economy and pointing us in the direction of Pachinko Parlors and the analogy with ICOs.
Thanks to the rest of the Nivaura team for their valuable feedback and guidance on this and the ongoing research.