Utility Tokens Will Asymptotically Approach Zero

A large number of ‘utility’ tokens have been raised in recent ICOs, and yet their performance as an investment has all the characteristics of a bubble.

The classic shape of almost every utility token shows the following behavior:

Classic Bubble Graph (Ethereum actual data)

The graphs tend to show the classic slow rise, an initial bear trap, followed by exponential rise, a delusion crash, a temporary recovery followed by a long tail. The final death spiral can descend to zero, or bounce back over a significant period to a low level mean trend line indicating some fundamental basic value proposition.

In this article I will argue that in almost all cases the value of utility tokens will eventually approach the zero trajectory (albeit with a random walk along that path).

What is a Utility Token?

The Merriam-Webster definition is a useful start. It combines the two key concepts (and one that is missing but probably should have been there):

1. An organization issues utility tokens to raise funds for the development and deployment of a service or good that will be deployed at a later date;
2. The issued tokens can be used at a later date to purchase the service or good
3. The token issuance and redemption processes after the initial issue are designed to provide certain behavioral incentives as part of the process

Why a Token and Not Fiat Money?

The most obvious question (and the one least asked) is why not just charge for the service or good using fiat money? What is it about the idea of a token that creates additional value?

There are four possible reasons:

1. Buying A Future Service or Good — the most important thing about a utility token is that it conveys the right to purchase a service or good in the future, and is in essence a redeemable right. This may happen almost immediately, or at some distant point in the future. It begs the obvious question, why do people want to buy in advance, which we will discuss in more detail below;
2. Enabling Micro Purchases — buying services or goods with fiat money may be impossible if you wish to buy in units that are below the minimum denominations (such as $0.001). Using a token that is infinitely divisible with low transactions costs makes micro purchases possible;
3. Forced Barter — it is possible to force bartering between products and services using a token by refusing to redeem them for anything else but the token. This creates a closed community that is shielded from fiat conversion and where the exchange rates can be either set or can float with internal supply and demand. We will discuss the nuances of this rationale in some depth, as it is the second core reason to use utility tokens and yet is subject to many assumptions and issues.
4. Reduced Exchange Costs — it is possible that the token is one side of a service that involves some sort of exchange, and the token is required because existing methods of exchange are either too costly or too slow and the tokenized service offers a faster mechanism that requires a token exchange on one side of the transaction. We will discuss this briefly, especially given the recent volatility in XRP and Ripple.

Buying a Future Service or Good

A very common proposal is that the provider of the service is offering access (similar to a paid API key) but will only accept tokens as payment.

The obvious question is why not just buy the tokens just before you need them, and then redeem them for the service right there and then.

Which begs the second question, why not just charge for the service?

To justify advance purchase, the answer must either be:

1. You expect the future supply of the service to be constrained and therefore you are pre-purchasing access. This in turn suggests that the service or good access has a congestion or supply constraint, that implies that the provision is inelastic, and that the payment in advance is a form of congestion rent or constraint rent. This is a rather strange justification for a supposedly decentralized service where additional supply should flow smoothly and is therefore invalid or an indication of some fundamental poor design in the service; or
2. You expect the future price of the service to rise, and you are therefore insuring against this price rise by locking in access at a given current price. If we ignore the fact that almost every utility token is showing a dramatic price reduction over a 90 day time horizon and not a rise, we then need to question why would we expect the price of the service to rise in theory? One possible reason is that there are diseconomies of scale. So that as more people use the service the more per unit it will cost. Again in a decentralized environment this is rather hard to justify. Simply replicate. The second reason often given is to argue that there are increasing value returns to network size, so that as the network or service increases in size, its value per participant also increases, and this ‘value’ translates to prices. There are many variants of this argument circulating, and unfortunately all are spurious. One rationale is the Quantity Theory of Money argument, which is completely spurious, as discussed here and here. The second is that there is some sort of network affect akin to Metcalfe’s Law in play, and that as the number of users of the service increase, the value of the network will ‘obviously rise’ to each participant, and that this will translate to increased prices through an unspecified mechanism. The first point to make is a simple one. If there are 500,000 users and Metcalfe’s n-squared power law was really in play then we increased the number of users to 1 million then we doubled the per person value of the entire network. Really? Does that make sense to anyone? Okay what if it was nlogn as suggested recently (see here)? Then the per person value has risen by 30%?. In reality network effects per person should show rapidly decreasing utility past a threshold. The real economic analysis of network effect externalities is complex, as they are a public good and it is necessary to calculate both the value and the effect of pricing that value on the participation rate or demand rate. As soon as you attempt to price the externality the demand elasticity wipes out participation. This is discussed more carefully in this associated article.

So we are left with only one rational reason (as opposed to speculative larger fool price bubble reasons) why one would want to buy in advance — because congestion in the future may cause prices to rise. Now this congestion could come about because of two reasons:

1. Constraints on the supply of the product or service (so the tokens become a means of rationing); or
2. Artificial constraints on the token supply, so that the service is available but the consumer can’t get access to it because there are not enough tokens, so they bid up the price of the token (essentially as an artificial mirror constraint).

It is rather bizarre that reason (2) is often quoted as the rationale by token creators— why would anyone claim a system is efficient at provisioning a service if access to that service is artificially constrained by limiting the number of tokens, purely as a mechanism to drive up the price of tokens and force an under utilisation of the service?

Artificial constraints tend to have a way of being competed away, especially in a decentralized market and supply network. Observing that the token constraint is real but the supply constraint is not, surely the token will fork and more supply will be created — or an alternative token will be created at a lower price for the same service.

If we follow this argument, there is no rational reason to buy in advance. So in that case the rational price to pay for the service is the utility to you of the service at the time — its “price”. In a competitive market this will gradually fall to the marginal cost of supply — the long run equilibrium point (or pretty close to it). Have a look at each utility token network and ask what the marginal cost of supply of the service that it grants access to, and you will have found your long run settling point for the price of the token. In almost every case I can assure you it will be very low.

Forced Barter

The second large reason for issuing a token, is that it may be deemed necessary to have a system of forced barter, where it is desired to abstract from fiat money as the reference unit, and instead establish exchange rates for services that are denominated in the token and which participate in a closed market.

There may be very valid reasons for this. An extreme but relevant example is that a psychiatric hospital may reward patients for certain types of behavior by issuing tokens and then allow them to redeem the tokens in exchange for something of value (such as certain privileges). The hospital does not want any ‘leakage’ of the incentive arrangements, and so creates a closed system where the only way to get value from the tokens is to redeem them within the system. This enables somewhat arbitrary exchange rates to be established between things that may be very difficult or confusing to value using a fiat currency, and also allows for intertemporal exchanges between actions now and actions in the future. Importantly the creators of such systems want to deliberately exclude comparative value or exchange calculations from commodities or services outside of the barter group. There are non-trivial reasons for doing this (apart from avoiding confusion). One reason is that the token accumulation may have a virtual good aspect to it, where the value to the individual may substantially exceed what another individual perceives the value, and the exclusion of a common reference point is deliberate. The example is non-trivial and has a wide set of possibilities.

The key point to a barter exchange economy is that it must remain closed. As soon as the economy becomes open (in that you can buy, sell or price a token in fiat money or even some other external token) then it defeats the entire purpose of the closed barter — virtualisation of value and exchange. As soon as it becomes open, it raises the question “why not fiat?”, and relatively quickly people will assign the current fiat equivalent to the token, begging the question again, “why not just fiat?”.

In some cases the forced barter may be quite artificial, and even a way to try and avoid regulatory impositions. One example is where a network service will only accept tokens as deposit or a stake. This means that the person wishing to make a deposit has to either earn tokens or buy them. As soon as the network allows the person to buy the token, they then have established a fiat equivalence. Worse still, normal laws apply — such as that deposits are a financial liability on the deposit taker. This is also seen in many “game economies” where some games allowed tokens to be purchased or redeemed. As soon as they did this they opened the game to speculation (usually about deflation) and they also came within consumer laws that forced optional redemption, making the entire token supply a liability.

So the net of this discussion regarding barter as a use case, is that there is definitely a use case, but that the very use case that supports the concept also precludes a rational external price base — and more importantly a mechanism of purchasing. So you cannot have your cake and eat it. The tokens are not worthless, but they cannot be priced.

Enabling Micro Purchases and Reducing Transactional Costs

These are two valid reasons for establishing a utility token. There are cases where existing transactional systems are not sufficiently fungible at the right price to sustain diverse and potentially very important use cases. There are also cases where existing transactional systems are too expensive, or too complex, and improving these characteristics using tokens is a powerful use case.

However it is important to understand the value of a token in these cases. The value of a token used to facilitate an exchange is not related to the value of the product being exchanged. This most simple point seems lost on many people in the crypto community. As an example you can purchase an interbank settlement system and install it for maybe $100 million. That system may exchange $100 billion a night between banks. The size of the transfers has nothing to do with the value of the system or the tokens that are used to facilitate the transfer.

To further emphasize this point, imagine a system that transferred rights to gold bars in vaults. The token, say a “Goldy” is made of plastic worth a few cents, and has inscribed on it two keys and digital signature, with the signature done with the private key of the vault who issued it. The Goldy has an instruction on it transferring a certain number of gold bars between the strong boxes identified by the signature, one of which is a strong box owned by the first vault but located in the second vault. Owners of gold bars have strongboxes in various vaults. If Person A wishes to transfer gold bars between strongboxes in the same vault, they simply physically move them. If they want to transfer in another vault, they need to get a Goldy, and in the first vault they move the gold bars to the Vault strongbox, and get issued a Goldy for the second vault. They take the Goldy to the second vault, and it checks the signature for validity, and then moves the gold between the two strongboxes.

In this system, the Goldy may be the trigger that moves a few hundred thousand dollars, but the Goldy itself is worth a few cents.

Obviously the system can be made digital and simultaneos, with the changes being recorded on a digital register. Now what is the Goldy worth? Zero. It has no intrinsic worth whatsoever. If it is the only way of accessing the service, it has the value or price of the cost of the service — which is unrelated to the value of what is being exchanged. If the service is congested, then there may be congestion costs associated with the Goldy that act as a rationing agent (similar to the prior sub section).

Conclusion

In all the cases we have considered the price of a token in a utility network, in a competitive market with no congestion, will approach the marginal cost of supplying the service accessed by the utility token. No more and no less.

This is the long run position that will settle after any speculative frenzy associated with an ICO has died down.

And with modern computer systems the cost per transaction service is extremely low. Witness the decline in price for share trades (you can now do them for free), the decline in price for banking transactions (in many parts of the world you can do them for free), the decline in price for marketplaces for second hand good (mostly for free or at very low cost).

This has drastic ramifications for utility tokens. What it essentially means is that the service being offered by utility tokens can easily be competed by an alternative token over time — and so in the long run the most efficient network and token will win. Secondly it means that those networks that simply offer computer processing (such as Ethereum) are destined to almost zero value, because computing is simply so powerful and cost effective. Those systems that offer exchange tokens where the token provides access to the mechanism of exchange (such as Ripple) will be valued by the value of the boundary interchange (just as with Western Union) and the trust involved in these intermediaries, and it will not be related to value of transactions relayed. This implies that these exchange tokens will also tend towards zero, but will bounce to a low level long run equilibrium that places them in the competitive landscape with other suppliers. That is, there is no intrinsic high worth associated with the token itself — it is the boundary service that is being paid for.