Vitalik’s Concerns Miss a Crucial Point
Vitalik Buterin has recently published a very interesting piece regarding medium-of-exchange token valuations. In that piece he points out several serious problems with what many people consider to be a very good idea: raising money for the creation of technologies by binding those technologies to a specific token that they then mint and sell.
Full disclosure: I’m currently in the process of doing exactly that. So I have a very clear motivation to defend my side and to show that Vitalik was too harsh. That his warning does not apply to all such projects.
I know that Vitalik isn’t always easy to understand, and that being so well grounded in theory he sometimes neglects to mention the real world aspects (perhaps taking them for granted) that can help you and me make the distinction between a viable economic model and a Ponzi scheme. So I’ll try to add that part. And I’ll start with my conclusion:
While the problems he highlights exist to some degree in all medium-of-exchange tokens, from “good old ether” (as Vitalik calls it) to bitcoin and even to fiat moneys (still the most common form of medium-of-exchange tokens), they pose a serious threat only to a very specific (and suddenly very common) type of token — the type whose value is chiefly speculative.
The bad news is that all token sales, by definition, start that way. The good news is that they don’t have to stay that way forever.
First Problem: Stream of Buyers and Sellers
Vitalik uses a simplified yet insightful economic model to illustrate his point. If you haven’t read it, go read. It’s well worth your time. I’ll just quote the part that relates to our problem:
Now, let’s look at the story with a “medium of exchange” token. N people value a product that will exist in a decentralized network at $x; the product will be sold at a price of $w < x. They each buy $w of tokens in the sale. The developer builds the network. Some sellers come in, and offer the product inside the network for $w. The buyers use their tokens to purchase this product, spending $w of tokens and getting $x of value. The sellers spend $v < w of resources and effort producing this product, and they now have $w worth of tokens.
The first thing that bothered me was how the description keeps going back to “product” (as opposed to “service”). In a sales pitch, this would be a big warning sign. A product, generally speaking, receives its value once at the stage we call “production”. That is when effort and other resources are poured together and mixed to create a valuable product. If the token sale you are considering is only dealing with exchanging products and tokens back and forth, beware. No new value is being added to the system.
But the most important thing here is “$w < x”. This seems wrong. Logic dictates that people be willing to buy the token when $w = x. The resources they are willing to spend to obtain a product or service are exactly and by definition what they think its actual value to be. This value is often referred to as the token’s “utility value”.
Let’s imagine a network of masseuses who provide massages in exchange for a specific token. The massage-coin. The token’s utility value is derived from the masseuses’ willingness to accept the token in exchange for giving massages — a service that requires an effort on the masseuse’s side. The concept of effort is crucial here, as effort generates value that’s constantly added to the system. So the value of the masseuse-coin is derived from the sweat of the masseuse.
But Vitalik isn’t really talking about that kind of value because he is not talking about the functioning masseuse network. He is talking about the fundraising stage, long before masseuses and sweat are a part of the equation. And this is a very important distinction. So let’s describe the fundraising phase.
In order to build the masseuse network, its developers mint a bunch of tokens and sell them to the public. The buyers in this phase, being investors, have to buy the token for less than what they believe its worth is going to be, so that they can sell it at a profit once the network is up and running. This is called “speculation”, and is the reason tokens are sold at a discount. This is when “$w < x” makes sense.
Notice that here, the cycle is not complete, and in fact it never will be; there needs to be an ongoing stream of buyers and sellers for the token to continue having its value. The stream does not strictly speaking have to be endless; if in every round there is a chance of at least v / w that there will be a next round, then the model still works, as even though someone will eventually be cheated…
This sounds almost trivial. Of course there has to be a stream of buyers and sellers. And such a stream is to be expected when providing a valuable service. And no one needs to be cheated. But that’s only true at the utility phase, when the token receives its value from what it can actually buy. During the speculative phase, on the other hand, when the token receives its value from what people speculate it to be worth in the future, such a stream is often a dangerous thing that ends with a bubble bursting and with people losing money.
And here comes the bad news: when using a token to raise funds it will always be sold for a speculative value and at a discount, i.e. for less than what the buyer speculates the value to become. And as long as the token remains at that speculative phase (often referred to as “fundraising” and “bootstrapping”) it suffers from all of the problems Vitalik describes. The traps he enumerates can only truly be avoided by increasing the utility component of the token’s value and decreasing the speculative component. By getting people to value the token for its actual utility.
The speculative component can never be completely avoided. When I was growing up, hiding dollars under the floor was a socially acknowledged channel of investment. But as long as it overshadows everything else, as is almost always the case when raising funds, there is volatility and risk.
Second Problem: Market Cap vs. Velocity
Next, Vitalik sets up some very important equations. Again, read them if you haven’t, and I’ll take the liberty of jumping straight to the bottom line:
MC = TH
The left term is quite simply the market cap. The right term is the economic value transacted per day, multiplied by the amount of time that a user holds a coin before using it to transact.
So the all-important market cap drops as the velocity increases. This sounds counter-intuitive. We are used to thinking about both market cap and velocity as positive indicators that should therefore correlate. So let’s illustrate it with a silly example:
Let’s imagine a single masseuse providing service for massage-coins. And let’s imagine a perfect, frictionless exchange for said token. When people want to use the service they give the exchange fiat money that’s immediately converted to tokens which are immediately sent to the masseuse who immediately converts them back to fiat. Now riddle me this: how many massage-coins are required to run such a system?
The answer is of course 1. A single coin is enough. Even if some folks want to buy an extra special massage that costs 10 massage-coins, you can simply run your single token ten times through the frictionless exchange. And that’s how we get phenomenal cosmic velocity with an ity bity market cap.
There are, of course, mechanisms for slowing down the velocity, making sure more tokens remain “in the pipes” at any given time, to increase the number of tokens the system requires. But that’s not always a good idea. As a rule of thumb, investors care about market cap, so it is important during the speculative phase, while users care about velocity, which matters at the utility phase. This can’t be stressed enough: there are huge differences between these phases. And any project that doesn’t give a lot of thought and planning to the transition between the two is downright dangerous.
To make this transition it is not enough for the token to have a utility value, it needs an overwhelming utility value. The massage-coin I described certainly has utility value, but neither masseuses nor their clients have a preference for massage-coins over dollars, bitcoin, or any other acceptable token. For this to happen, the massage-coin has to give something that other tokens do not. Either better massages, or more massages, or massages that can’t be directly bought in any other way.
Third Problem: Equilibria and the Hidden Fee
Vitalik briefly explores a completely different, but still important problem with medium-of-exchange tokens. The hidden fee. By holding the token the holder is exposed to fluctuations in its value against whatever token the holder would hold otherwise. I do not think this requires much clarification. Yes, binding a market to a specific token in the absence of a perfect exchange (and in real life nothing is perfect) forces the participants in that market to hold that specific token. This is a well known problem for all medium-of-exchange tokens, including fiat moneys (remember the dollars under the floor!). You can always regret holding a certain token instead of a better performing one (which would probably be bitcoin or ether).
But Vitalik takes this one step further, pointing to a more specific problem:
One immediate conclusion from this particular insight is that appcoins are very much a multi-equilibrium game.
Equilibrium theory is a bit out of the scope of this article, but the bottom line is that if there is more than one equilibrium then there are unstable equilibria, which may be exploited to the detriment of the market. Add this to the effect of the number of users on velocity, and therefore on market cap, and therefore on price, and you could find yourself with a feedback loop that makes market manipulations easier and more advantageous.
This becomes less of a problem the more widely accepted the token is - because small markets are easier to “game” than big markets. It is also more of a problem when the price isn’t stable. So once more, a huge problem during the speculative phase, when the budding market is still small and the price is volatile and disconnected from the balancing effect of the network’s utility value.
What’s that balancing effect? Imagine the value of the massage-coin drops. Masseuses find that the massage-coins they are getting are no longer worth their effort. So they quit the network. Now, if we were only discussing simple massages, masseuses could simply sell them for another token. But what if the network has an overwhelming utility value? What if people desire those special massages that can only be bought in massage-coins?
In that case, masseuses quitting the network will reduce supply. And when supply drops, prices soar. And when prices soar, more masseuses join the network, and prices drop again. And before you know it — prices stabilize around the utility value. That’s the balancing effect.
Moreover, during the utility phase, a token can often withstand a shift in price, as tokens often do when the market is in flux and as is often seen with fiat moneys in foreign exchanges. It is possible that massages become a huge fad. Suddenly everyone wants a massage and there aren’t enough masseuses to satisfy the demand. Prices soar. Then people get sick of massages and the price plummets. It doesn’t matter — as long as masseuses are willing to put in effort in exchange for the token, it has a “sensible” value.
What this all serves to show is that relying purely on the medium-of-exchange argument to support a token value, while attractive because of its seeming ability to print money out of thin air, is ultimately quite brittle. Protocol tokens using this model may well be sustained for some time due to irrationality and temporary equilibria where the implicit cost of holding the token is zero, but it is a kind of model which always has an unavoidable risk of collapsing at any time.
This is an apt description of the speculative phase. Yes, there is an unavoidable risk. And there should be! Speculation entails risk. That’s why it pays off. Sometimes! If you invest in something because you believe it has a chance of paying off, you also implicitly believe it has a chance of not paying off.
In his conclusion, Vitalik proposes using sinks. These cause deflation which can counter the hidden fee effect and can help with the vicious cycle described in the first part, with the stream of sellers and buyers, when the deflation can counter the negative gradient operating on the token’s price.
This is a fantastic idea, but it isn’t enough. The negative gradient has to subside with the transition from the speculative phase to the utility phase, (when $w = x replaces $w < x). If this doesn’t happen, the project has no financial redeeming qualities and no right to exist.
So first of all, make sure your project has an overwhelming utility value. I know this sounds obvious, but judging from the gold-rush state of the market — the message is yet to be internalized.
And secondly, make sure there is a clearly planned path to making the transition between the speculative phase, which is fun but dangerous, and the actual “real-world” utility phase, when the value actually makes sense. You can use sinks, you can use stakes, but don’t just assume its gonna happen. Be a responsible and successful entrepreneur.