How to think about Tokenomics

As a Math and Operations research guy (PhD Stanford) , I want to start off a formal discussion on how to model “tokenomics” (the economics of token prices).

I want to start out with the Kyle Samani “velocity paper” which I think is a good start but not the final answer on this. But start reading that before diving into the thought process below

Here are the variables

TS — Total circulating token supply, eg: 97,000,000 for ETH
P — Current price per token, eg: 834 for ETH
HT — Hold time of a a Token in Ecosystem (fraction of year). e.g: 0.1 if the average ecosystem participant holds the coins for a bit over one month (one tenth of a year)
TV = transaction volume per year measure in $. Let’s look at a marketplace doing 100 Million dollars in transaction volume per year, and assume all of this is in the token Then TV = 100,000,000. In the case of ETH, TV = 1 Trillion USD.
GTV = Growth in TV per year. You want TV to be growing as much as possible. A growth rate of 100% or more implies that there are new entrants who need your tokens. A zero growth rate means whatever equilibrium price you are at, it’s an equilibrium — it won’t go up
TT = transaction time. How long does a transaction last in terms of time.
TMCAP = token market cap = TS * P
R = TV / TMCAP = ratio of transaction volume to transaction market cap. As we will show below, its about 0.1 for ETH
R1 = R / TT. The R valuation metric is adjusted for the transaction time.
R2 = R / HT. A different valuation based on hold time

Law 1: You want HT to be as big as possible.

As Kyle Samani points out, there is no value in investing in a token that is only held for a short amount of time. The example would be buying a ticket with tokens. If I need the token to buy the ticket, but the transaction with the seller is immediate and he can cash out immediately with tokens, there is zero long term demand for tokens, no matter how big the transaction volume (TV) is.

Law 2: HT > TT.

They are not equal. Many reasons can exist for HT to be significantly bigger than TT.

In the case of pure Bitcoin transactions, each transaction takes a very small amount of time (under a day). But buying BTC with fiat is a nightmare, and case take a week or longer. So the average person needing BTC will tend to “buy in” and “fund their accounts” with BTC. The same is true for ETH, and indeed for any good alt-coin.

So we see here clearly a reason for hoarding coins: difficulty of buying and selling them. Altcoins can further dis-incentivize selling by adding fees for small coin redeemtions (the stick) or rewards for accumulation (the carrot). Both work.

Law 3: Tokenomics are variable, they are not fixed at the white paper stage.

A lot of people (including Kyle Samani) are quick to knock companies where these tokenomics are not 100% nailed in the white paper. But as the Law 2 discussion shows, many of these economics tweaks can be done post ICO launch. The fact that you did not plan ahead for them does not mean they are not doable.

Law 4: Once the token is fully launched, R should be below significantly below 10.

Currently, ETH is generation about 1 Million transactions per day, at 5k / transaction. Thats a volume of 5 Billion transactions per day or a bit more a 1 Trillion dollars per year run rate. At its peak, ETH was processing 1 Billion in transactions per hour .

ETH has a current market cap of 81 Billion, so it is trading at a ration R = TV / TMCap of order of magnitude 0.1. This is way below 10, but again the transactions are extremely short.

In general if your ecosystem is doing 10 million in annual turnover and there are 100 million dollars in tokens, there will be very little upward pressure on tokens. On the other hand, if like ETH, the annual turnover is 10x the market cap, the pressure is on.

In general R is a very crude measure of token value — but does not take into account TT or TH.

Law 5: You “probably” need to have a reason to BUY tokens as opposed to simply earn them.

In the previous analysis, it was assumed that there in fact was turnover — in otherwords tokens were not just given away, they were bought. But not all ICOs are modeled on this. Many new participants are arguing that everything should be given away, including tokens, with zero revenue model.

This is a tricky one to analyze. Until recently if you looked at STEEM, the chart looked like a disaster

The 1.6 Billion market cap of STEEM was whittled down to a low of 17 million on March 16, 2017.

But then, this happened.

After making new highs in December, the cap of STEEM is still 1 BILLION. Hence the word “probably” in the title.

Law 6: R1 and R2 is a better measure than R of pure value.

If I am doing 100 MM in volume per year, and the average transaction time its 0.1 years, then 10MM USD in volume is tied up at any one time. If the market cap is 100 Million, then R = 1, but R1 = 0.1, which means that any of the 90% of tokens not in transactions can be sold in theory to finance new entrants to the ecosystem.

However, this ignores hold times. If the average hold time is actually 6 months, then 50% of all tokens are being held in the ecosystem. In such an equilibrium, any new growth translates into a surge in prices.

Law 7. Growth matters. A LOT

In any Equilibrium, growth in transactions will cause growth in prices all other things considered. There are so many other considerations in play that this is hard to quantify, but as long as a fair amount of tokens are held in the system either inside of transactions or held by participants, prices will tend to move up.

This is really the explanation for STEEMs recent revival. It’s growing like crazy. STEEM is still not Reddit, but it’s growth is amazing.

Law 8. You really want an economy where sellers become buyers

One of the real goals of an ecosystem is for all parties to be both buyers and sellers. This is true for Ethereum and Bitcoin. As an ETH user, I frequently both send and receive ETH. In a marketplace, you want your sellers to find uses for your coin as opposed to just cashing out for fiat (or another coin). This plays into Law 1. You want to maximize HT.


It’s really hard to model token prices. But there are clearly things to look for

  • high growth (or high possibility of growth)
  • incentives for people in the ecosystem to hold (either long transaction times, which force holding, or general incentives to hold /penalties to lqioudate)
  • a reasonable transaction volume (or possibility of seeing one), relative to market cap.