Friday, February 3, 2023

Hey bitcoin owners, how are you paying for bitcoin's energy costs?

Bitcoin miner in La Doré, Quebec via L'Étoile du Lac

In what form do bitcoin holders bear bitcoin's huge energy costs?

For the Bitcoin network to be secure, it requires miners to do a lot of work, and those miners will only do that work if they are compensated. The creation of new bitcoins every 10 minutes is the main method of paying them. The question in this post is how these mining costs get passed on to people who hold bitcoin. Bitcoin is bloody expensive, after all. Owning it can't be free.

Let's explore the problem by looking for an analogy in traditional finance. In the place of bitcoin, let's introduce CashCo. CashCo owns $100 in cash. It has 100 shares outstanding. For simplicity's sake, let's assume that the shares trade at fundamental value, so each share is worth $1 ($100 / 100 shares) and the company's market capitalization is $100.

Let's introduce Jack, who holds one share of CashCo, worth $1.

Next, let's make CashCo resemble Bitcoin by introducing a mechanism that functions like bitcoin rewards. On January 1, CashCo announces that it will henceforth issue a single new share at the end of every day to an independent entity to validate CashCo's database. CashCo will do this each day for the next 30 days.

Like Bitcoin rewards, the additional CashCo shares are created out of nowhere and are paid to an external validator. Furthermore, the policy is time-limited, in the same way that bitcoin rewards will no longer be paid after some terminal point in time.

This is how it looks. At the end of Day 1, CashCo will issue one new share to the independent entity. At that point CashCo will have 101 shares outstanding. The fundamental value of each share will be $0.99 ($100 cash in the bank / 101 shares). But the fundamental value of the company as a whole would stay the same, since CashCo would still have $100 cash in the bank.

As for Jack, he still holds one share at the end of Day 1, but the fundamental value of his share will have fallen to $0.99.

At the end of Day 2, CashCo will issue another share to the validator, who now owns two shares. CashCo will now have 102 shares outstanding. The shares will have a fundamental value of 98 cents ($100 / 102 shares). But the total fundamental value of the company will still be $100, since it remains backed by $100 in cash.

The fundamental value of Jack's single share will have fallen to $0.98.

Fast forward 30 days, and there will be 130 shares outstanding. Each share will have a fundamental value of 76.9 cents, but the fundamental value of the firm remains at $100.

Jack is a forward-thinking individual. When the new policy is announced on Day 0, he quickly runs through the above calculations and sees that the fundamental value of his single share will be marked down from $1 to 76.9 cents over the next 30 days. Aghast, he immediately tries to sell it. But the policy being common knowledge, everyone else will make this calculation this too. No one will pay Jack more than 76.9 cents for his share, knowing that in 30 days its fundamental value will be 76.9 cents.

And so on Day 0, the moment the announcement is made the value of CashCo shares falls to around 76.9 cents. That is, the new information about future costs of paying the validator gets brought forward in time and is quickly baked into the current price of CashCo shares.

Bitcoin operates along the same principles as CashCo.  

The schedule of new bitcoins to be paid to miners is already known. Because bitcoin buyers are like Jack and forward-thinking, this cost is effectively brought forward in time such that it is already built into bitcoin's price. That is, the original bitcoin owners (and all owners after them) prepaid for security the moment that the Bitcoin network was brought into existence.

So let's conclude by getting back to my original question. In what form do bitcoin holders bear bitcoin's huge energy costs?

Holders bear bitcoin's costs the same way that Jack bears CashCo's validation costs: in the form of foregone price appreciation.

Jack's share is worth 76.9 cents, but if CashCo suddenly found a way to avoid paying an external validator, his shares would immediately vault from 76.9 cents to $1. So the form in which Jack absorbs validation costs is through a lower-than-potential price for CashCo.

The same goes for bitcoin. We can think of the price of bitcoin as being much lower than it would otherwise be in a world where those costs didn't exist.

So bitcoin owners, the form in which you absorb bitcoin's huge energy prices is via a permanent discount on the value of your bitcoin stash. Instead of bitcoin being worth, say, $35,000 or $45,000, it's only worth $23,000 you're effectively missing out on a big one-time jump in the price.


Here's an exercise for you. Does this same logic apply to proof-of-stake coins like ether or tezos? Is the schedule of future Ethereum staking rewards already baked into today's price of ether, just like bitcoin mining rewards are baked into bitcoin's price? Yes? No? Provide your work. 


  1. Isn't the supply of Bitcoin theoretically finite ie capped at 21 million? And as the number of Bitcoin issued increases, it takes more and more mining activity to generate new Bitcoin? Doesn't disprove your point, just raises mute issues about sustainability of the model

    1. Hi Frances,

      It's capped at 21 million, but that level won't be reached till around 2140 AD. Right now there are around 19.3 million in circulation. As for the amount of mining activity to generate new bitcoin, that's generally related to the price. The higher the bitcoin price, the more computers will come online to compete for the reward.

  2. > Holders bear bitcoin's costs [...] in the form of foregone price appreciation.

    You could argue bitcoin price would jump up if we decided to ditch miner subsidy tomorrow, but the diminishing amount of bitcoin pending distribution that is discounted in the price is only a temporary way for holders to pay for trust minimization.

    In the long run, subsidy will disappear and yet trust minimization will not be provided for free to bitcoin holders. A comparatively greater volatility (lower risk-adjusted return) can be expected to be the cost that bitcoin holders will need to pay for the convenience of trust minimization.

    Miners will keep spending electricity to secure new transactions and get bitcoin fees in exchange, but we should not confuse the transaction ordering *service* that miners provide with the convenience yield that bitcoin (the *asset*) provides and the associated cost offsetting it at the margin.

    1. "A comparatively greater volatility (lower risk-adjusted return) can be expected to be the cost that bitcoin holders will need to pay..."

      In this post I was trying to get into how mining costs get passed on to end users, specifically holders. I agree that volatility is a cost/drawback of holding bitcoin, but I don't believe that volatility is the manifestation of a resource outlay incurred by miners and subsequently passed on to holders. That is, users can pay miners with fees, or they can pay them with rewards (which gets manifested as a lower-than-otherwise price, as I argue in this post), but users can't pay miners with volatility.

    2. I understand you are looking at the issue from a mining costs perspective, but I felt some unrelated things were being mixed together. Bitcoin’s energy costs derive from two different sources of miner revenue—coins distribution and the transaction ordering service.

      The transaction ordering service is subsidized as a side effect of coin distribution, but the set of users benefiting from that subsidy is a different one —despite being overlapping—. The “maintenance work” done to keep the ledger working is a service that holders don’t need most of the time. They only need it when they become senders or recipients of new transactions.

      The creation of a trust-minimized asset implies a temporary cost, which is the mechanism used to ensure a fair distribution of its supply. This is part of the nature of bitcoin, which could be changed to end any further distribution if this were less costly than the potential appreciation resulting from the change. But it seems that is not the case, so there is no incentive for holders to push for it.

      It may not be straightforward to see the link between the greater volatility of bitcoin and the dilution suffered by holders, but these two costs must add up and offset the convenience yield that bitcoin provides as a trust-minimized digital asset. Even if holders saw no risk in the fork required to freeze supply, at the margin, they would still need to pay the same cost for the convenience yield they get. The only change being that there would be more volatility instead of dilution, thus they have not much incentive to change it. I’m afraid lower energy consumption does not weigh much in market decisions.

    3. "...if this were less costly than the potential appreciation resulting from the change. But it seems that is not the case, so there is no incentive for holders to push for it."

      I've explained the foregone price appreciation by comparing our world to an alternative world in which it's possible to get just as much security at less cost, and so the price of bitcoin is higher. I'm not sure that this hypothetical option actually exists, so holders can't push for it, but that doesn't mean that they aren't feeling the costs in the form of a lower bitcoin price.

      I don't really get your volatility point. Bitcoin users can pay miners with transaction fees or rewards (which get discounted into the price of bitcoin), but I don't see how they pay with volatility.

    4. You assume holders benefit from the extra PoW ("security") funded by mining subsidy and that is what prevents the change. But it is only a minority of them, who also happen to be new transaction recipients (Bitcoin network users), that benefit from the extra PoW. Most bitcoin holders don’t care about how much PoW secures new transactions. Their coins are deep enough and secure against reorgs regardless of PoW.

      The way I understand it, holders pay a cost for holding bitcoin equal to the convenience yield they receive. If the dilution cost disappears and the convenience yield remains the same, holders will be willing to pay in other ways for that scarce resource. Accepting a lower risk-adjusted return (i.e. greater volatility) is the obvious outcome (no, miners don't get any reward derived from this—miners don’t serve bitcoin holders, they serve Bitcoin network users). I think this is why stopping any further coin distribution has no significant potential to make bitcoin price jump up. Ethereum migration to PoS can serve as an empirical proof.

    5. "You assume holders benefit from the extra PoW ("security") funded by mining subsidy and that is what prevents the change. But it is only a minority of them, who also happen to be new transaction recipients (Bitcoin network users), that benefit from the extra PoW."

      I don't think it makes a difference either way to my argument whether holders benefit or not. The fact is that they are paying for the costs of mining.

      If a stream of new units are being created out of nothing and paid to third-parties, then by definition the total future value of that stream will be discounted into the asset's present price, just like what happened with the shares of CashCo in my example. That discount represents a cost, whether you are Jack holding CashCo shares or a bitcoiner holding bitcoins.

      You'd have to refute my CashCo example to get me to think otherwise. But my CashCo example is just Finance 101. Surely you wouldn't try to argue me out of it.

    6. Yes, holders are paying for most of the costs of mining. I'm not trying to dispute that, which is obvious.

      The point I'm trying to make is about the trust minimization service that bitcoin holders consume having the same cost regardless of whether it funds mining expenditures or not.

      If holders decided not to subsidize mining anymore (they could do that), they would suffer additional volatility, until total cost equals the unaltered convenience yield for the marginal bitcoin holder. Thus, cancelling mining subsidy has little potential to affect bitcoin price.

  3. In POW Crypto, unlike Fiat, there is no Centralized Sovereign Treasury Chest that is protected by guns and a private army. This lowers the costs for the customer and transfers it to willing miners

    So is it true if you'd like to earn, you mine and sell whenever the Hash-Rate gets low?

    Stocks have
    1) Earning per share
    2) Price to Earnings ratio or
    3) Market cap as a Multiple of Annual Revenue

    If we treat Bitcoin as a Stock, do we have comparable metrics?
    Although it may not be fair to compare Bitcoin to a stock issued by a company,
    the company in this case could still be the hardcoded "code" or fixed or predefined deflation rate and fixed or predefined total number of shares.. 21 mil BTC.

    1. Bitcoin and other crypto coins are basically ledgers that grow over time as new ledger entries are created out of nothing to reward folks who do maintenance work. The reason I use a simple firm, CashCo, as an analogy is because a stocks are ledgers too, and like a crypto ledger they grow as new stock is conjured up into existence to pay contractors and such. It may not be the perfect analogy, but it's probably the best one we've got.

  4. One problem with this argument is bitcoin's future energy costs are not known, since they're tied to bitcoin's market price. Therefore it's hard to argue that these costs are priced in.

    1. I'm not sure I agree.

      Say that CashCo's assets were not made up of $100 cash but something much more random, like $100 worth of bitcoin. We have thus created something very similar to the premise of your argument; an asset who's future validation costs (paid in shares) are not known because they are tied to bitcoin's market price.

      Even so, this doesn't prevent the market from properly pricing in the future costs of validation. CashCo shares still fall to 76.9 cents the day of the announcement.

      Likewise with bitcoin, the market should be able to price in all future rewards.

  5. Although you make good points worth pondering, and your analogy is stretched to its limit, it matters very little in the grand scheme of things. It's just the next iteration of money -- a better version then fiat.