Valuations for Cryptocurrencies I – The Equation of Exchange

Cryptocurrencies are a new asset class.  The last major new asset class, common stock in companies, existed for hundreds of years before Benjamin Graham fully elucidated metrics for their valuation. Likewise cryptos exist and have market prices established through very liquid markets, but one of the first questions on any newly interested party’s lips is “well, what should a Bitcoin be worth though?”  

It’s an important and incompletely-answered question.  In fact, as I write this, I can think of maybe only three people who I could call “buy-side analysts” in the crypto space and each one is more interested in the abstract notion of valuation more than churning out BUY/HOLD/SELL ratings for the rafts of new coins and tokens entering the space.  In this series of posts, I’ll try to acquaint readers with some of the terminology and ideas behind cryptoasset valuation.  While this entire introductory paragraph is a long-form warning that this is an incomplete theory, it will involve many fundamental ideas of the space and some light is better than none.

Possibly the most natural idea is to note that the name is not just branding: cryptocurrencies satisfy all of the economic criteria to be labeled as a currency.  As such, we can apply any valuation models people have developed for currencies or money supplies.  The most successful valuation model is the “Quantity Theory of Money” which traces roots at least back to Nicholas Copernicus who took a break from revolutionizing cosmology to write down a quantitative theory of how people value multiple currencies and the effects of circulating new, debased, coinage.  The contemporary form of the quantity theory of money is expressed in the nearly-Newtonian simplicity of Irving Fischer’s Equation of Exchange: MV = PQ.

          M is the money supply in active circulation
          V is the “velocity of money”
          P is the price level of the currency
          Q is the output level of economic activity transacted in said currency

This equation is not controversial at all, in fact, it’s an identity.  The left hand, or supply, side is simply a measures the total amount of currency changing hands during some time period.  The right hand or, or demand, side measures the total amount of goods being exchanged over the same time period.  These two quantities must be equal.  Any contentiousness arises from the estimation of the individual parameters.  

Here’s an example use in valuation of Bitcoin that should illustrate how it works, using not-unreasonable estimates of the values.  First, the left-hand side, or supply side of the equation.  Let’s assume that 20% of Bitcoin is in active circulation, the rest being in wallets with lost private keys or held as an investment.  This makes M = 3.3 million Bitcoins.  Since the blockchain is public, we can see that there’s a bit more than 250,000 BTC of daily on-chain transactions.  That makes about 100 million BTC worth of transactions per year.  The yearly velocity of money, therefore, can be estimated at V = 30.  That is, each Bitcoin in circulation is spent about 30 times a year.

On the demand side, we need to understand the economy that Bitcoin is being used as a transactional basis for.  In the case of national currencies, PQ is the GDP of the nation using that currency.  However, since cryptoassets are trans-national, the question becomes “who is using Bitcoin and for what?”  Instead of explicitly breaking this down in terms of the price level of goods or services (P) times their amount (Q), we can just – again – note that PQ is the total amount of services transacted during the same amount of time used on the MV side.

Let’s say, for example, that 25% of the half-trillion dollar/year remittance market is transacted in Bitcoin and that’s the only thing Bitcoin is used for. This means the total yearly transactional demand is for $125 billion of currency.  This quantity is the entire right-hand side of the equation: price per average remittance (P) times number of remittance payments (Q).   So, we have PQ, however it’s denominated in USD instead of Bitcoin.  This is what lets us establish an exchange rate (R) between USD and BTC.  

3,300,000 Bitcoin x 30 yearly tx = $125,000,000,000 yearly tx / R

Solving for the exchange rate, R, in this estimate, we get a price of $1300/BTC.  Again, the numbers are representative, I am not saying this is a fair value.  In this context, the equation can be used to get some general trends and estimates of the markets Bitcoin is supporting.  

Let’s say, for example, that the 3.3 million active circulation and 250k BTC/day numbers hold, but Bitcoin use expands to support transactions of $1 Trillion per year.  The price of Bitcoin would find a natural level of $10,000.

Now, let’s say for example that the that the $125 billion yearly transaction and 250k BTC/day numbers hold, but investors and speculators cash out and release another 5 million Bitcoin into active circulation (for a total of 8.3 million).  The price of Bitcoin would find a natural level of $500.

You can see some of the slipperiness in these examples.  The barrier to entry for a cryptocurrency user is low and dropping all the time.  You could send a remittance in Bitcoin this month and Ether next month if the difference in fees justifies it.  Further, the price volatility could also lead to sharp swings in the number of circulated coins.

Here’s an example use in valuation of Bitcoin that should illustrate how it works, using not-unreasonable estimates of the values.  First, the left-hand side, or supply side of the equation.  Let’s assume that 20% of Bitcoin is in active circulation, the rest being in wallets with lost private keys or held as an investment.  This makes M = 3.3 million Bitcoins.  Since the blockchain is public, we can see that there’s a bit more than 250,000 BTC of daily on-chain transactions.  That makes about 100 million BTC worth of transactions per year.  The yearly velocity of money, therefore, can be estimated at V = 30.  That is, each Bitcoin in circulation is spent about 30 times a year.

On the demand side, we need to understand the economy that Bitcoin is being used as a transactional basis for.  In the case of national currencies, PQ is the GDP of the nation using that currency.  However, since cryptoassets are trans-national, the question becomes “who is using Bitcoin and for what?”  Instead of explicitly breaking this down in terms of the price level of goods or services (P) times their amount (Q), we can just – again – note that PQ is the total amount of services transacted during the same amount of time used on the MV side.

Let’s say, for example, that 25% of the half-trillion dollar/year remittance market is transacted in Bitcoin and that’s the only thing Bitcoin is used for. This means the total yearly transactional demand is for $125 billion of currency.  This quantity is the entire right-hand side of the equation: price per average remittance (P) times number of remittance payments (Q).   So, we have PQ, however it’s denominated in USD instead of Bitcoin.  This is what lets us establish an exchange rate (R) between USD and BTC.  

3,300,000 Bitcoin x 30 yearly tx = $125,000,000,000 yearly tx / R

Solving for the exchange rate, R, in this estimate, we get a price of $1300/BTC.  Again, the numbers are representative, I am not saying this is a fair value.  In this context, the equation can be used to get some general trends and estimates of the markets Bitcoin is supporting.  

Let’s say, for example, that the 3.3 million active circulation and 250k BTC/day numbers hold, but Bitcoin use expands to support transactions of $1 Trillion per year.  The price of Bitcoin would find a natural level of $10,000.

Now, let’s say for example that the that the $125 billion yearly transaction and 250k BTC/day numbers hold, but investors and speculators cash out and release another 5 million Bitcoin into active circulation (for a total of 8.3 million).  The price of Bitcoin would find a natural level of $500.

You can see some of the slipperiness in these examples.  The barrier to entry for a cryptocurrency user is low and dropping all the time.  You could send a remittance in Bitcoin this month and Ether next month if the difference in fees justifies it.  Further, the price volatility could also lead to sharp swings in the number of circulated coins.

In a sense, the valuation example above is somewhat post-priori since it’s easy to estimate M and V from the blockchain itself, but (for probably obvious reasons) the PQ side is harder to determine precisely.  This analysis can, however, be used in a predictive way for what some are calling “cryptocommodities.”  Imagine there is a services market worth some amount of money.   This could be cloud computing, or file storage services, power from a decentralized electrical grid – anything that can be accounted for accurately on a blockchain. A cryptocommodity is a coin or token that gives you the right to use those services, secured by the blockchain, provided by a network of users paid in said tokens or coins.  We can now do a quantitative analysis of what those coins should be worth by simply comparing the market cap of the coins or tokens to the market share of the services network controlled by the blockchain.  

Here’s an example with numbers which, again, are just ballpark estimates, not careful analysis.  Imagine there’s a token which is created to control access to solar energy on a power grid that spans a few countries across the European Union.  The project’s pitch is that it will reduce payment friction by avoiding international bank settlements and that using a blockchain with a dedicated token will get sellers and buyers paired quickly and without a multiplicity of claims on the same output.  

The analysis would proceed like this; the entire solar energy market in the EU is about €3 billion.  If you think this company might get a 10% market share, you can value the PQ side of the exchange equation as €300 million.  Now, for the MV side, you need to get into the weeds a bit.  You’ll need some domain-specific knowledge to model the user behavior.  Specifically, an estimate of how frequently people transact across the grid to purchase power and what their average purchase might be.  Further, you’ll need to model investor behavior and know any token inflation or pre-mine policy in order to estimate M. It’s not a science, but it can let you know if things are outrageously over or under-valued.

Even though the regulatory status of initial coin offerings (ICOs) is still up in the air in many jurisdictions, the ICO fundraising genie is out of the bottle.  In the next few years, this will be the sort of analysis going on at investment research firms.

In the next article, I’ll discuss valuations derived from the decentralized networks themselves.