The 21 Levers Moving Bitcoin Price -Part I

Bitcoin is Rat poison. It is a Scumball activity.

It has no utility. It wastes energy. It is killing the planet!

We’ve all heard the incessant chatter, either dissing bitcoin or deifying it. Depending on who you ask, you are either stupid for investing in a bubble, or are helping usher in the second coming of money. So, what is the truth?

Does Bitcoin have any value? How can it be estimated?, and, What factors determine Bitcoin price?

Market Demand: Why do we need Bitcoin?

Let’s start with the context of Bitcoin’s creation. Contrary to popular opinion, Bitcoin wasn’t created because Satoshi Nakamoto hated Visa or Mastercard’s margins and wanted a “peer-to-peer” alternative to visa or Mastercard. Bitcoin was designed to be money that is not subject to the arbitrariness and manipulations of central banks, and resilient to attacks by governments, and was released at an unbelievably opportune time. It needed to be resistant to seizures and censorship, and to do that, it needed to be trust minimized and decentralized. Decentralization is a feature that helps us do this by eliminating trusted central parties that can arbitrarily change the rules of monetary issuance, and is a necessary means for separating the state‘s monopoly over money. Earlier attempts at private and anonymous online currencies such as David Chaum’s eCash/cyberbucks faced problems of centralization and were susceptible to censorship by banks and governments.

Broadly, Bitcoin satisfies the demand for two properties

1. Soundness (in Monetary Policy), and
2. Censorship Resistance (in Transactions)

Is “Sound Monetary Policy” needed?

Historically, the central authorities, like kings and central banks entrusted with issuing sovereign money have abused that privilege to varying degrees. They did this by transferring purchasing power from the population forced to use their currency to themselves and to themselves and those close to them (cantillon effect). In roman times, they debased their coins, but in modern times, they found a way to do this more subtly by constantly tinkering with money supply and interest rates. This has created and amplified credit bubbles and crashes, distorted the economy, and generally wreaked havoc on society. Bitcoin was Satoshi’s prescription for this pandemic.

The demand for “soundness” can be simply estimated by a thought experiment instead of bothering to collect accurate empirical evidence.

Would anyone like to have the purchasing power of their money decrease over time?

Do people care about Censorship Resistance?

There is no need for complicated measurements. Does any one want third parties to exercise arbitrary control over how much, how, and for what reason they can control their own money?

(For the data driven folks, here is a piece of data: we ran surveys and interviews with >3000 crypto users, many of whom were not technically or economically enlightened, >80% wanted coins with properties where their coins are unable to be seized by banks, governments, companies or other third parties, and >95% understood the importance of a money that does not lose value over time. The Caveat to this is that >63% of them also were speculating on cryptocurrencies that did not have the attributes of either censorship resistance, or a fixed supply schedule, but that discrepancy will not be addressed here)

Other problems that create demand for Bitcoin

1. Capital controls that prevent people’s own money from crossing sovereign borders
2. Wealth confiscations (ex: Civil Forfeiture laws in the US, recent demonetization policy in india etc)
3. Targeting the wealth of political dissidents
4. Wars and conflict that force people to abandon their wealth and flee

Value: What gives Bitcoin it’s value now?

Bitcoin is neither a reliable and stable store of value, nor a generally accepted medium of exchange nor a unit of account, just yet. Then what is Bitcoin’s current unique value proposition (UVP)?

1. For a majority of users, we believe that the primary value of Bitcoin is speculative.
2. Most are speculating that Bitcoin will appreciate in value, either because it becomes a store of value aka “digital gold”, Medium Of Exchange or Unit Of Account in the future.
3. Utility: For some, Bitcoin is already a medium of censorship resistant value transfer (MOE). (Gold and Hawala networks have been the traditional means to hedge against transfer risks. However, they are hugely inefficient, are prone to trust and security problems, and I believe that Bitcoin is cannibalizing them).

Granted, it is hard to precisely attribute how much of Bitcoin’s value comes from each of these value propositions, but it is possible to separate the factors that affect the 2 value props, thus determining Bitcoin price.

The Bitcoin Machine: Engines

(This analysis is heavily relies on Pagnotta, Emiliano and Buraschi, Andrea, An Equilibrium Valuation of Bitcoin and Decentralized Network Assets (March 21, 2018). Available at SSRN: https://ssrn.com/abstract=3142022)

Bitcoin network has three engines that keep the Bitcoin machine running, and they are essential to it’s equilibrium valuation:

1. Users buy and sell Bitcoin, implicitly valuing the Bitcoin network (demand for bitcoin). In simple terms, people use Bitcoin to buy, sell or transfer value, looking to maximize their expected utility.
2. Miners supply the hashpower that makes Bitcoin network secure and therefore, valuable. They do so by buying mining equipment like ASICs, paying for electricity and internet, AND, in return, get paid in Bitcoin. This property, described elegantly by Pagnotta and Buraschi as “Unity” is a key property of decentralized networks such as Bitcoin. This property essentially connects the Bitcoin network’s internal economy with that of the external real economy, and acts as a bridge that coordinates demand with supply.(If a miner provides resources to the network but does not get paid in the network token, there is no price coordination between the real economy and that closed decentralized network, except for the price the tokens have at exchanges. Several other decentralized networks, like Ripple and Stellar do not have this property because of a lack of mining, which makes them difficult to value based on this analysis.)
3. There is a liquid market into which the miners and users can sell/buy their bitcoin for fiat.

The Bitcoin Machine, the price levers:

#1. Transactional Demand: Bitcoin as a payment network:

(This analysis is largely borrowed from Arvind Narayanan, Joseph Bonneau, Edward Felten, Andrew Miller and Steven Goldfeder Bitcoin and cryptocurrency technologies,, Princeton University Press 2016)

Let’s say that Bitcoin is used by people wanting to transfer money without relying on the banking system, regulatory complications, paying fees etc. To value bitcoin based on that function, we need 4 variables.

A. Demand:

Let’s say that the total number of users who have this need are trying to transfer D dollars per day using Bitcoin. So, the demand for MOE is D dollars/day.

In other words, D dollars per day are waiting to be converted to Bitcoin so that they can be transferred, and the sender converts them to Bitcoin and transfers them to recipient. The recipient will then convert the Bitcoin back into USD or other fiat.

B. Price:

Let’s say that P is the current price of bitcoin in USD. The number of Bitcoins needed to mediate D dollars worth of transactions is D/P.

C. Lockup Time:

There is a duration of the transfer process includes the following elements:

• the sender needs to exchange USD to BTC,
• Sender sends BTC to the recipient,
• the recipient needs to wait for transaction confirmation,
• recipient sells the BTC back to USD.

All of the above take time, say T. Let’s assume that there is an average amount of time during which MOE Bitcoins will stay locked up in a transaction, before they are liberated and are available to service other transactions.

Let’s call this T the “lockup time”.

D. Supply:

Let’s say S is the total supply of Bitcoins available to mediate transactions (this would be the total supply of Bitcoins minus those that have been lost, locked up in cold storage by HODLers, etc).

Out of the total number of bitcoins available to perform the MOE function, only a fraction are available per unit time, because the rest are locked up in between transactions. How many Bitcoins are available to mediate transfers is determined by dividing the total supply by the average lockup time for transaction, therefore

Supply per unit time = S/T

Demand per second is the number of Bitcoins demanded, divided by the price of Bitcoin (because P is the price in USD and D is the total value of transactions that need to be mediated by Bitcoin), therefore,

Demand per unit time = D/P

The Equation:

Because markets clear and demand and supply continuously equilibrate, we can assume that for any unit of time, demand equals supply (If supply exceeds demand, price falls and vice versa).

Therefore,

D/P = S/T, or P=TD/S

This means that for a given available supply and transaction time, Price is proportional to the demand for Bitcoins as transaction mediation instruments. Now lets see what factors affect demand.

#2. HODL demand:

How can we separate the demand for tranfering value from the demand for HODLing?

A quick heuristic is to use the public blockchain and measure coin age from the last transaction in the UTXO set (input age) and calculate how the average and distribution of coin ages relate to changes in price. This assumes that people who buy and hold bitcoin for transferring value do so, for short periods of time, and that those that believe that bitcoin will replace fiat, are willing to buy and hold for much longer periods, largely insensitive to price changes. Unchained capital has an elegant heuristic that measures this HODL demand, and their methodology is called HODL waves.

With coin age analysis, we can infer the speculative vs investment demand. Two phenomena are noteworthy:

1. An increase in the fraction of newer coins is expected when prices are going up, as speculators move in and HODLers take their gains.
2. On the other hand, during periods of falling or stable prices, if the fraction of older coins increases, meaning HODLers are settling in for the long haul, it sets the stage for a price spiral when demand suddenly increases due to external factors or “triggers”, and the long term HODLers are resistant to sell.

The demand for Bitcoin as a store of value comes from those who believe that the fundamental value of Bitcoin comes from its ability to serve as a better store of value than fiat currencies over the long term (due to it’s fixed monetary policy), and the expectation that Bitcoin will eventually replace one or all fiat currencies. HODLers are a critical variable that determine Bitcoin’s value, as do full nodes, miners, and transaction throughput.

Hodler demand affects the price of bitcoin, as commonsense would indicate. But it does so in interesting and complex ways.

1. It reduces the supply of Bitcoin available to mediate transactions, increasing the price, (See section 2) and
2. As Bitcoin price rises, hash rate aggregate increases as more and more miners compete to construct blocks, which increases the network trust, making it more valuable in the eyes of HODLers, further increasing HODL demand and fueling a feedback loop.

#3. Network size:

In the case of Bitcoin, the size of network is determined by the number of full nodes and users. Because most networks are characterized by positive network effects, the greater the number of full nodes and users, the more the value of that network, all else being equal.

Valuing the size of networks based on is not a new, and there are many models and arguments as to which method is the best. However, it is generally agreed that networks grow exponentially more valuable as more and more participants join and use them.

Several proposed models include:

1. Reeds Law: The value of a network increases exponentially, proportional to the number of nodes N in that network (V ∝ 2ᶰ).
2. Metcalfe’s Law: The value of a network is proportional to the square of the number of nodes in that network (V ∝ N²).
3. Beckstrom’s Law: The value of a network equals the net value added to each user’s transactions conducted through that network, summed over all users (This may be the most accurate, albeit more difficult to model).

4. Odlyzko’s Law: V ∝ Nˡᵒᵍ ᴺ

I use Reed’s law here to estimate network value because it has been proven in a large scale studies. However, I believe it may still underestimate Bitcoin network value. In a social network, as the number of formable pairs increases, crowding and other negative network effects kick in, and the eventual number of network connections a user can form is limited by Dunbar’s number. In peer to peer financial networks, this limitation does not apply, negative externalities are only limited by the blockspace for on chain financial throughput. This is further mitigated by layer 2 solutions such as Lightning. Intuitively, as Bitcoin enables increasingly sophisticated financial contracts between many parties, and any number of subgroups are formable. Another well known criticism of Reed’s and Metcalfe’s laws also doesn’t apply to Bitcoin network analysis because all payment transactions can be safely assumed to be equally contributive to the network size and utility.

According to Reed’s law, the value of Bitcoin network grows proportional to (2^N)-N-1, where N is the number of participants. As N increases, the rate of change of V is proportional to e^(N log 2) -1. (Because, a network of N people has 2^N possible subsets, minus N individuals who are a set of one, and an empty set). This gives rise to a convex shape for Bitcoin network valuation.

Whatever heuristic is used to estimate the value of a network, whether it is Reed’s, Metcalfes or ), it is indisputable that network size determines its value more than linearly.

Tracking network size over time requires counting proxies such as the number of full nodes, and the number of wallets owned by individual entities. Full nodes are trackable, but wallets present a difficulty because HD wallets are not identifiable individually. A better proxy may be “individual” accounts at centralized exchanges, but that is often proprietary data not publicly obtainable.

Many thanks to Dan Held, Eric Voskuil, Donald McIntyre, Eric Martindale, Stephan Livera, Curtis Spencer and others who helped me with comments and feedback.

Continued: Network structure, Expectations etc in Part II……