In Defense of Utility Tokens

On the kermacline, predictable consumption and other things

“X can be swapped for cash or spent in shops most easily in Côte d’Ivoire, Egypt, Ghana and Uganda (…) X is commonly used as money in Nigeria, too. (…) But even in places like Kenya, X [is] being used as currency (…) X’s value does not rely directly on a government’s stability or ability to hold down inflation by, say, showing restraint printing money. Opening a (…) money account typically requires waiting for days after showing your ID. In contrast, X can often be purchased and sent immediately and anonymously.” (The Economist, 19-Jan-13, “X” substituted for the real term by the author of this article)

X is clearly Bitcoin, right? And good on the Economist to cover its utility as a means of exchange as early as January 2013!

Not quite. X are actually prepaid mobile minutes. They seem to provide some of the same advantages as Bitcoin, as mentioned in the quote above. Not explicitly noted above but surely also part of their attraction as a currency are characteristics such as the divisibility / small denomination, ease of exchange (between phones), and ease of storage. Unlike Bitcoin, mobile minutes clearly do not live on a decentralized system, but it appears Africans are happy to trust that the central bank, pardon, mobile operator, practice sound ‘monetary policy.’

This may be a somewhat exotic example of the use of a utility item as general ‘money’, but clearly not the only example — everybody knows that cigarettes are used as money in prisons (I always wondered what happens to that system if and when a corrupt guard comes back from holidays and loaded up at the duty free shop), albeit they recently seem to have been replace by mackerel.

What I find interesting here is that the populations in these examples opted to use items with a specific use case (i.e. with tangible utility) as money rather than coming up with something that has no practical use. Say, e.g. the prisoners could theoretically just take a book from the prison library, tear it up, and denominate each page as a currency unit — and, pronto, they would have a limited-supply currency with no significant risk of inflation, easy divisibility, easy storage, etc. Presumably the Africans could have come up with other such alternatives, too. (Clearly, you do not have the decentralized ledger system in these examples either — and that is a fair and important point, but not the focus of this article).

One reason I find these facts interesting, from the angle of cryptocurrencies, is, of course, the fact that Bitcoin does not seem to have tangible utility (except as a potential store-of-value, which we shall discuss imminently), at least currently not even as a feasible means of exchange — given the high transaction fees at the time of this writing. Bitcoin advocates will argue that Bitcoin’s use is now as a pure-play store-of-value and point to the fact that gold is a very popular store of value and has few tangible uses — and, indeed, probably few people that choose gold as a store-of-value do so thinking that they may also end up using the gold to make jewelry, some electronic circuit, a dental filling, or to gold-plate their Lamborghini (the latter primarily if they are also crypto whales). One question I keep asking myself these days, though, is whether, if there was something that could do both, fulfil the roles of money incl. being a store of value and have a practical use case (utility), whether people on average would not prefer that to something that can only play one role? Two quick observations on this:

(1) I ask myself this question with a specific view on cryptoassets — I do not suggest, nor believe, that something will emerge in the non-crypto world to dethrone gold, partly due to its track record of millennia.

(2) In the world of traditional investments, there is actually a very prominent asset class that is both a store-of-value and has a very clear use case: real estate (worth over $200 trillion as of 2015, leaving gold far behind).

Meanwhile, in the crypto world, some advocates of the store-of-value use case do not only believe that a tangible use case is unnecessary for a coin to work as a store-of-value but go even further: they say that coins with utility (“utility tokens”) may actually be worth very little or even worthless. The rest of this article will investigate their argument. It provides some brainstorms, but does not claim to be anywhere near a complete analysis. Please also note that I will use some examples that are not decentralized utility tokens — the discussion which use cases even need a decentralized protocol is another big topic and worthy of its own article. Lastly, this is not meant to be a BTC-bashing piece (I am a BTC hodler) — rather, I have grown so tired of utility coin-bashing that I felt the urge to make some counterpoints.

Utility tokens & their valuation

There is a broad, informal classification that, I think is fair to say, many people within the crypto community make with regard to cryptocurrencies, namely into store-of-value coins (substantially Bitcoin/BTC) and utility tokens (some people further sub-divide into “protocol tokens”, but I include those within utility tokens in this article). There are also coins with financial security-like aspects, in that they give certain economic rights, but these are not within the scope of this article. Utility coins, as the name implies, have a use case (other than store-of-value), e.g. pay for computations on the Ethereum network (Ether), file storage (Filecoin), or even webcam porn (Spankchain). This classification of cryptocurrencies is also reflected in common approaches to their valuation. Bitcoin, being nowadays a store-of-value pure play, is frequently valued making an assumption about the global store-of-value market share it may be able to gain (especially in relation to gold). Utility tokens, on the other hand, are often valued using the monetary equation of exchange:

M*V = P*Q

where: M = the nominal amount of money supply

V = the velocity of money

P = the price level

Q = the economic output

Since we are interested in valuing the currency, we solve for M:

M = P*Q / V

Actually, as, being investors, we are really interested in valuing 1 unit of the currency, we also have to divide M by the number of currency units in issuance (i.e. the currency supply). This is a point we shall return to at the end of this article.

Note that P, Q and V are all with regard to the ‘economy’ in which the currency is used. Traditionally, this would be substantially within the borders of the country issuing the currency. One thing crypto, with its utility tokens, has done is to create currencies that are linked to use cases (such as file storage) rather than geographies. Note also that P*Q is effectively the nominal GDP of that “token’s economy.”

The critics that argue that utility tokens have little or no value typically do not disagree with the approach of using the equation of exchange but rather argue that P*Q (GDP) may be low and even declining, and that V is high. Let’s take these in turn.

Low & declining PQ?

The argument that PQ is low and declining was made, inter alia, by John Pfeffer with regard to Ethereum (and its cryptocurrency Ether), arguing that Ether is used to pay the ‘gas’ to perform computations for applications (distributed applications — ‘dapps’) that run on the Ethereum network. As computational cost has been, and is widely expected to continue, declining, the argument states that PQ will decline, too.

I have two main counterarguments to this:

First, speaking of utility coins in general (and not just Ether), we need to consider the growth of the underlying use case of the coin. If e.g. the use case is data storage (e.g. Filecoin), one should note that Dropbox is said to have annual revenue growth of approx. 30%. That is, the underlying ‘data storage economy’ (PQ) is growing 30% p.a., despite what has historically been significant deflation in the cost of data storage equipment. Some will now argue that this represents revenue, and that, at least for some utility coins, the coin is technically only required to effectively pay for the infrastructure cost of the service (as John Pfeffer argues) — which brings me to my second counterpoint.

I think this definition of PQ / coin usage — expect coins to be used to pay infrastructure cost only — is far too narrow, even for something like Ether. I have compared utility tokens, and specifically Ether, to the fiat currency of a country before, with one key difference being that the domain of the former is defined by a use case while the domain of the latter is defined, more or less, by geographical boundaries. If I run with this fiat-currency-of-a-country metaphor, then saying that e.g. Ether will only ever be used to pay for gas for computations is more or less equivalent to saying that one expects e.g. the Swiss Franc to only ever be used to pay for Swiss infrastructure (roads, public services, etc.), while ignoring all of the ‘apps’ that have evolved to run on top of the infrastructure — in the Swiss Franc metaphor, world-beating companies like Nestle, Novartis, Roche, Swatch, etc. The Swiss infrastructure is, of course, paid for by taxes and sometimes that is how I think of Ethereum’s gas — a tax payment. Note that, thinking in this way, you may also regard that fiat currency in your pocket as a utility token for things like public roads, police protection, possibly public health (depending on where you live), but we digress. Importantly, those Swiss companies do not only use the Swiss Franc to pay taxes, but also e.g. to charge for their products, to pay their suppliers and workers, and so on. So, in the crypto world, e.g. for Ether, the key question will be how much Ether can expand beyond its pay-for-computations core use and gain usage for other purposes as well (like charging for products, paying workers, etc.), before it may or may not hit a ‘boundary’ beyond which fiat money is used. I have always wanted to coin (pun intended!) a new word from ancient Greek, so I will call this boundary between crypto and fiat use ‘kermacline’ from the ancient Greek words for coin and slope. I, for one, am pretty optimistic that e.g. Ether can go far beyond its pay-for-computations use — notice that e.g. for participating in most ICOs, you must have Ether on hand. There is also a trend now of paying compensation in cryptocurrency (see e.g. here and here), which, of course, makes sense for companies who also expect their revenue base to be mostly or entirely in cryptocurrency. You can now pay some utility bills in Australia in Bitcoin or even taxes in some places in Switzerland. There is now also at least one crypto hedge fund that is denominated in a cryptocurrency (XRP) rather than a fiat currency. Are these examples sporadic and almost gimmicky for now? Maybe. Do I seem to see more and more of them over time? I think so — the kermacline seems to expand outwards into the fiat economy, and with it the potential for PQ to be higher than some people think. I actually happen to think that it is our responsibility, as the crypto community, to help push out the kermacline further and further. Let us now move on to velocity.

High velocity?

The second key argument that critics of utility tokens make is that these tokens may have very high velocity, as people may prefer to hold something more liquid, like fiat or digital cash, and just swap into and out of utility tokens whenever needed, leading to high velocity (V), and hence leading to a smaller monetary base (M) required to support the token economy (PQ). By the way, for reference, the velocity of key fiat currencies is currently between 1–2 (you can find data e.g. here, albeit note this table gives you the inverse of V, i.e. M/PQ). Some people are also already calculating the velocity of cryptocurrencies, e.g. for BTC, see here.

You hopefully will have noted the assumption made above: people prefer to hold cash. Sounds intuitively reasonable, but let us take a step back: why is that?

Cash as an option on everything

Personally, I like cash because it gives me flexibility, and I need that flexibility because I am relatively uncertain about many of my future consumption needs. Cash allows me the luxury of not having to decide right now whether I buy a new TV, a piece of jewelry for my wife, a flight somewhere, a Lambo, or whatever. Cash is basically a financial option on (almost) everything, with an infinite exercise period (it even goes beyond your death, as you can leave the cash as inheritance). You can exercise it whenever you want (what we call an “American option” in the financial market). Pause for a minute and marvel at this: cash solves the problem of uncertainty of future consumption. That is hardly a trivial problem to solve — it is probably one of the key reasons that socialism/communism does not work, as the centrally-planned 5-year-plans do not work.

Let us run with the cash-is-an-option model. Financial theory tells us you have to pay a premium to have an option. Now you might say “wait, what are you talking about — I earn interest on cash” (well at least in normal, non-ZIRP, times). True, you might and indeed should earn interest, but that actually fits with financial theory: if you lend out your cash for a period of time, you cannot use that cash for your own consumption needs during the term of the loan, i.e. you give up part of your exercise period, and you should be compensated for that, via interest (that holds true even if we assume zero credit risk — which is one of the other reasons for interest). However, do not delude yourself for a minute that you are not usually paying a price for holding on to your cash and flexibility. If you doubt me on that, try e.g. going to your local gym and see how much you pay, per month, for a 1-month membership and how much you pay (also per month) if you commit for a year (by the way, this example holds true even here in Switzerland, where interest rates are zero).

Predictable consumption

Does that then mean that, if consumption was predictable, people would be willing to part with cash and its flexibility in exchange for a “voucher” for the specific future consumption (wait — that is a utility token!) or the actual good itself (if it can be stocked)?

I think so, for a number of reasons, only one of which is the potential saving of the option premium for cash, as noted above.

A second reason may be ‘hedging’, as people may have uncertainty about whether the good or service they want to consume may be available in the future and at what price: if they know with high certainty that they will want to consume the good/service in the future, they might as well ‘lock it in’ now. This is what we see in some unstable countries where people hoard food — consumption is pretty predictable when it comes to food!

Third, one can actively try to conduct this hedging of future consumption at favorable rates. E.g., I travel a lot and expect to continue doing so. Whenever there is a chance to somehow buy airmiles at what I consider a favorable rate, I do so. What are these airmiles if not a utility token (albeit a centralized one where I take the risk that the central bank, pardon, airline, may misbehave in its monetary policy — and they occasionally do). People sometimes do this sort of thing if the consumption item is significant enough and there is price volatility — e.g. my mum seems to happily give up cash in order to fill up her 5000-liter heating oil tank in her house every time she judges the price of heating oil to be low, given her 50 years or so of experience of heating oil prices (I just realized my mum is effectively an energy trader — scary.)

Fourth, sometimes people may actually prefer to part with their cash and lock in a certain type of consumption for psychological reasons. This may include avoiding the temptation to spend cash on something else, feeling the joy of ownership or anticipation of ownership, or both. There have been extreme examples of this, where people effectively paid too much / gave their cash away too cheaply for these psychological effects: e.g. in 2007–2013, there was a business in Brazil called “Baú da Felicidade”, which basically required people to make monthly cash payments for a certain period in return for the right to exchange the accumulated paid-in value (which did not earn market interest) for a limited set of goods in a company-owned store.

All this is to say — yes, people seem to happily give up their cash and hodl utility vouchers (or the actual good) if future consumption is predictable enough. As I am writing this, I just remembered that, besides airmiles, I have a prepaid Starbucks card, a stash of “Disney Dollars” (which you can only use in Disney Parks) somewhere and also another stash of car wash tokens. That is right: I am a Disney Dollar hodler (I blame that one on my wife).

[Really quick side note here: is all of the above perhaps only descriptive of people’s behavior, while, prescriptively, people should not behave this way? One specific argument that is sometimes made here is that people should hold the minimum ‘inventory’ possible, just as companies are often expected to do. I actually do not believe this line of thinking is correct even for companies, but the comparison is misplaced to begin with — people are not companies, they have no owners that provided them with capital and who expect the return on that capital to exceed a certain hurdle rate, and to whom they could return that capital otherwise. However, these are all topics beyond the scope of the current article.]

Do the examples above again all seem sporadic and gimmicky? Yes, they probably are. There are probably at least three issues here that prevent the examples, for now, from being more common:

First, these existing ‘utility tokens’ (like e.g. airmiles and car wash tokens) are smallish, unconnected, not directly fungible amongst each other, illiquid, intransparent, etc.

Second, a lot of consumption still seems unpredictable.

Third, even where consumption is more or less predictable, it seems a hassle to actively trade a utility token to optimize it — and this is a fair point, as you have to invest time, which, after all, is the most valuable and constrained resource of them all (the makers of the 2011 movie “In Time”, which depicts a world where time is used as currency, were onto something).

A brave new world of utility tokens and predictable consumption

The first problem will be solved by traded utility tokens: you already can (and will increasingly be able to) trade one utility coin against another. So, if your predictions on your future consumption change and you think that you need less data storage and more Cryptokitties in the future, you may sell some Siacoin and buy some Ether. If all goes well, the markets for these coins will have deep liquidity and transparent pricing.

How about the problem of predicting consumption? Well, is consumption really that unpredictable? Look here for the typical budget of an average American making $75k per year. There are a lot of things in here that would seem to be more or less predictable, like housing, transportation, food, healthcare and insurance. Even if we add up all the rest (apparel and services, entertainment, cash contributions, and all other expenditures), we get to less than 20%. And even within those latter categories, there may be many predictable expenditures, like e.g. haircuts.

This brings us to the third problem mentioned above, namely that it may just be too much of a hassle to predict something like your haircutting needs and strategically trade/invest in “Haircutcoin.” In a way, it seems like there may actually be something like an ideal size of token utility case: if a token is small, it becomes a hassle to predict use and hence, even if the token exists, virtually nobody will try to hoard it, but rather simply swap in and out of it when necessary (unless the token use case, and therefore the average value tied up in it, is so small, that the utility of having some on hand outweighs any efficiency losses of having too much — as is e.g. the case with my car wash tokens). If a token use is too broad, on the other hand, it may fail in its role as a specific consumption hedge, besides other potential problems. So, the ideal utility coin may be one with a use case that is fairly predictable and that is big enough (in monetary terms) for you to care about — perhaps e.g. something like personal transport / UberCoin that you could use to pay for your daily commute to and from work.

Do you think it would still be a hassle to try to predict your personal transport needs and actively trade UberCoin in order to optimize and hedge your personal transport expenditure? Well, maybe you do not have to do any of these things yourself in the future. While cryptocurrency is a technological megatrend right now, it is not the only one, and it does not live in isolation. One of the other megatrends is the rise of artificial intelligence, which includes machine learning models for prediction. Prediction — as in e.g. the prediction of your consumption. Does this sound like science fiction? Consider that as early as 2014, Amazon filed a patent regarding ‘anticipatory shipping’, where basically they stock up warehouses near you with goods that they predict you will buy, based on your past consumption behavior. Perhaps, in the future, an AI assistant/app will e.g. rapidly learn your parking behavior (initially based on average population data that it was trained on, but then refining its model on an ongoing basis, based on your specific behavior) and then use its predictive model to foresee your future parking needs and ‘hedge’ them buy buying ‘ParkingCoin’, perhaps in conjunction with another AI system that specializes in trading such utility coins (all within parameters you can set — e.g. to prevent that a good part of your net worth is spent on buying the next 50 years’ of parking, even if the market price for ParkingCoin is really good). If we actually manage to really automate this type of process, then even utility coins for small uses cases may work, as the hassle is transferred from human to machine. In any event, some entrepreneurs should start thinking about these kinds of apps — they may make a lot of money. Internet of things (IoT), by the way, is a huge related topic here. (Caveat: I do have a graduate degree in AI, so my AI ‘hammer’ tends to see AI use case ‘nails’ everywhere!)

Now we have engaged in some proper futurology and, while very exciting, let us get back to the original flow of thought of this article. The last few paragraphs were basically all here to say that the velocity of utility coins may not have to be high, as people may actually choose to hoard them, for various reasons. I have given some of the potential reasons. Note that there are potential others, e.g. the very utility token itself may have built-in features that encourage hodling it; or, the token may even be hodl’d as a store of value, if people regard it as such.

A counterattack of utility coins on pure store-of-value coins?

Now that we have hopefully explored a line of argument that shows that utility coins do have value, can we counterattack the pure store-of-value coins and perhaps say that we do not need them, as utility coins may function as a store-of-value by themselves? That could easily be an article by itself and I will not try to cover this topic here, but rather just leave a few thoughts, in no particular order.

Stores-of-value should have low volatility — here, utility coins may actually have something like a built-in automatic stabilizer: if they become too expensive, the use case-driven demand declines and soon the coin price will also decline, and vice versa. Of course, the actual use case-driven demand may itself be volatile, in which case that will feed through to price volatility (in the same way that some country’s currencies are more volatile as others — e.g. currencies of countries with diversified economies vs. countries that are e.g. substantially exporters of certain commodities).

Furthermore, one has to consider the supply policy of the utility coin, of course — anything too inflationary is not great. The supply aspect may include features such as e.g. ‘burning’ utility coins every time they are used.

Lastly, and as we already noted, if a utility coin actually does become a store of value, note that its velocity will further decline.

Hopefully this article has given you some inspiration of how to think about utility coins and that any rumors of their imminent death may be greatly exaggerated. I have clearly not covered every angle of the utility coin debate. For example, another important topic is the argument that token protocols can just be forked if the token becomes too pricey — this is worthy of a whole article in itself, which I may or may not write in the future. The same goes for the discussion of which utility token use cases even need a decentralized protocol, as already noted above. However, I will excuse myself now, as I have some airmiles to trade.

In Defense of Utility Tokens was originally published in Hacker Noon on Medium, where people are continuing the conversation by highlighting and responding to this story.