Earlier this year, Christian Catalini of MIT and Joshua Gans of University of Toronto, released a paper, Initial Coin Offerings and the Value of Crypto Tokens, that analyzes ICOs as a financing mechanism for a blockchain venture. They examine the attractiveness of this fundraising technique both from the point of view of the startup and from the point of view of potential investors.
The paper contains many novel findings. One critical idea from the paper that is applicable to blockchain startups looking to raise money is the tradeoff between creating a token that is attractive as an investment vehicle and one that is useful as a medium of exchange (i.e. used to buy and sell services on the platform). Many startups aim to design a token that fills both of these roles, but Catalini and Gans show that it may not be possible for a single token to do both.
Catalini and Gans study a traditional “utility token” setup. A new blockchain platform is looking to raise funds by holding an ICO for the token that will then be used to purchase goods or services on the platform once the platform has launched. Individuals (either future users of the platform or investors) can buy tokens during the ICO, or they can buy them once the platform has launched. Once they have purchased tokens, token holders can either save their tokens for later (hodl, in crypto lingo) or spend them on the goods or services provided by the platform (or its broader ecosystem). The goal of the token holders is to maximize the payoff they get from buying tokens by spending them at the time when they have the most value.
From this simple setup, Catalini and Gans derive a number of nuanced results. They show that the exchange rate — i.e. the real value of the token — is a function of three factors: the size of the platform’s user base, the size of the token supply, and the value of the product or service that the platform provides. The first two factors roughly correspond to demand and supply for the token. If a platform has more users, all other things equal, demand for the token in order to purchase the platform’s services increases and therefore the value of the token increases. As the platform issues more tokens, all other things equal, the supply of tokens increases and therefore the value of a single token decreases.
Faced with these exchange rate dynamics, token holders are going to choose to spend or save their tokens depending on when they believe their tokens will have the highest value. If they anticipate that the token will grow in value in the future — because the size of the user base will grow rapidly, for example, without a corresponding increase in the token supply — they will choose to save the token and spend it in the future rather than to spend it now. Conversely, if they anticipate that the value of the token will decrease in the future — for example, if the platform plans to release a large supply of tokens in the future but the user base is not growing — they will spend the tokens now rather than save them.
The token issuer now faces a challenging problem. If he or she wants the token to appreciate in value over time — i.e. he or she wants the token to be a good investment vehicle — it is clear what he or she should do in the Catalini and Gans setup: rapidly expand the user base while restricting the growth of the token supply. However, this means that token holders will not want to spend their tokens on the platform to buy the platform’s services.
They will be saving their tokens until they appreciate further in value. The platform will have a token that is a fine investment asset and have a dormant platform.
Conversely, if the token issuer wants an actively used platform — i.e. they want their token to be a good medium of exchange — he or she should maintain the token supply to encourage token holders to spend. However, token holders will only spend if they believe the token will not appreciate in value.
The token issuer will have a vibrant, well-used platform only if the token is not a good investment vehicle.
Catalini and Gans make a number of simplifying assumptions in their analysis, so we do not recommend that blockchain platforms use their mathematical model as a literal prescription for token policy. For example, they assume that the universe of things that can be done with a token is save or spend. This leaves out the very important option of selling them on the open market to users or to speculators. If token holders are limited only to individuals who will at some point use the platform, then this leaves out the possibility that the token’s exchange rate is — in part or in full — based on market speculation. Understanding better the relationship between speculation and spending is fertile ground for future research.
However, even with these simplifying assumptions, the core insight we outline above will continue to hold. Token holders seeking to maximize the value of their tokens will trade off between spending their tokens on the platform and saving them or selling them on a secondary market. If the token gives returns that make it more attractive to save or sell the token than to spend it, the underlying platform will not achieve adoption. Conversely, if spending the token is more attractive than saving or selling it — for example because it has a steady price — then it will not be appealing to potential investors.
A founding team must choose between setting token policy that will make the token a particularly attractive investment (either to save or to sell) and making one that will be an attractive medium of exchange (to spend).
It is plausible that new blockchain ventures could create a monetary policy that allows the price of their token to climb after an ICO and then level off once their platform has launched. This is certainly a path worth investigating further. However, there may also be less tricky ways of ensuring funding of a platform. Going the traditional venture funding route would be one way, and targeting the token sale specifically toward platform users rather than speculators — who will require a monetary return on investment — is another.