The Economic Limits of Bitcoin

[Yba Muse]

http://faculty.chicagobooth.edu/eric.budish/research/Economic-Limits-Bitcoin-Blockchain.pdf

I will caveat this by stating that I am not a economist and have not spent time working through the formulas the author proposes to understand Bitcoin's equilibrium states.

TLDR on the article is that bitcoin does not scale because fees required to secure large value transaction, consistent with Bitcoin as a store of value, destroy it's use case for small value payments. The author claims that Bitcoin's economic limit is probably around $1.5-2.0B of value - far far less than the value of digital gold. The author's arguments rely on the potential of a 51% attack to double spend transactions, and the honest income required to offset this risk - the larger the value of the payments, the larger the income required.

At a conceptual level, it makes sense. However, there are developments in Bitcoin, such as layer 2 and sidechains that may solve these issues.

My ask is for a technical/academic person to review this and provide your thoughts so I can more fully understand this paper.

[pebwindkraft]

As a non statistician or economist I have my own critique to this paper, and I get the feeling it is trying to ride the waves of bitcoin, but the system is not completely understood, or it is riding the waves of big blockers. It starts at the very beginning with a phrase like this:

Bitcoin is an electronic payment system that relies on a combination of cryptography and a large, anonymous, decentralized collection of participants, called miners, to verify transactions, without the need of any trusted third party.

The role of non mining (full) nodes is not covered here... without the relaying nodes an important part for the security of the bitcoin eco system is missing in the equations (remember last year, when miners wanted to have 2mb blocks, and UASF prevented this?)

Further down the text we are informed about a misnomer:

while this problem is called the “double spending” problem, the “double” part is a misnomer — the attacker can re-spend his Bitcoins arbitrarily many times.

WoW! Hats off! Fantastic, for an economic paper on bitcoin we learn s.th.about the usage of the word « double ». What do these guys think they are?

Next quote:

In words: the equilibrium per-block payment to miners for running the blockchain must be large relative to the one-off benefits of attacking it.

Who would have guessed it? So the creation of the word « per block payment equilibrium » is no misnomer, and economical incentive to drive an attack must be way higher than the block rewards? Is this the essence of the paper?

Equation (3) places potentially serious economic constraints on the applicability of the Nakamoto (2008) blockchain innovation. By analogy, imagine if users of the Visa network had to pay fees to Visa, every ten minutes, that were large relative to the value of a successful one-off attack on the Visa network.

Doesn’t this smell like big blocker arguments? This continues on page 8, where a 30% tax is derived. Funny.

Further on page 11 we find several attack scenarios, with varying assumptions. At the end follows a conclusion which states that decentralized trust is ingenious, but expensive. Followed by a derivation, that  a high level of trust requires high running costs of the blockchain relative to the amount for attacking the system. As if the scenarios haven’t been discussed up and down in the relevant forums.

The whole paper relies on some mathematical models, which are based on someone’s imagination, and achieve a certain level of complexity. Ok. But the assumptions haven’t been checked against any period of the ten years history of blockchain, e.g. why was blockchain not attacked? How is blockchained secured by technological progress through history? What is the incentive model for the attacker at specific periods in time? ... This could proof relevance and applicability of the text... as there is no such thing, it looks like an ivory tower model, that fails to link to reality. And the conclusions are well known facts - blockchain is expensive, and attacking it requires enormous amounts. And the economic limits, as found in the headline, are not calculated. Btw: is it on purpose that Lightning is not mentioned at all? Because it doesn’t fit in the 30% tax model?

I am missing relevance in this text.

[Yba Muse]

Yeah - to be honest, I'm not sure what to make of this because I was not able to follow the math.

From what I gathered, the author is making a claim that, based on his mathematical model, the fees required to secure transaction on the bitcoin blockchain (or any POW blockchain for that matter) are so large that they place an limit on the economic value that can be secured by them. For example, he  makes a claim that a $1M transaction would required over $10K in fees to secure, and that because $1M transactions are being secured on the chain at this fee, this is the same fee that would be required of smaller transactions. And because smaller transactions do not work, the chain cannot be used as a store of value. How he justifies that jump isn't clear to me.

It's similar to arguments we've heard from Roger and others that if fees price out small transactions the whole chain isn't valuable. IMO, the vision of lightning is meant to change this and the author does not explore in any detail how layer two technologies impact his conclusions.

I'm left wondering what to make of the paper because I can't tell if i'm just not getting it because I'm not doing the math or because he's looking at Bitcoin wrong, or both.

[pebwindkraft]

I was trying to follow the math as well, but didn’t get it. Maybe due to my limited understanding of making a very easy bitcoin calculation so difficult in formulas 
I tried to reverse the idea, and think about attack vectors and the validation of the system (nodes validations and miner validations, where a single wrong tx makes also end up in an invalid block, destroying potential revenues). There is no view on this “from the bottom”, instead it postulates a model, and derives “limits of blockchain”. Without links to reality. So I believe it is a speculative paper to prove at a “scientific” level, that bitcoin cannot scale. And this smells like fiat banks or big blockers behind the paper...

[AdolfinWolf]

I was trying to follow the math as well, but didn’t get it. Maybe due to my limited understanding of making a very easy bitcoin calculation so difficult in formulas 
I tried to reverse the idea, and think about attack vectors and the validation of the system (nodes validations and miner validations, where a single wrong tx makes also end up in an invalid block, destroying potential revenues). There is no view on this “from the bottom”, instead it postulates a model, and derives “limits of blockchain”. Without links to reality. So I believe it is a speculative paper to prove at a “scientific” level, that bitcoin cannot scale. And this smells like fiat banks or big blockers behind the paper...

I *think* the author is trying to say that bitcoin cannot grow due to the fact that if it will, the incentive to do a 51% attack would become too big.

To counter this his argument is that the transaction value of each block must be atleast X amount of $, ( to make it unprofitable to do such an attack, (Right?)), which means that relative small transactions will become impossible, ( due to the high fees and limited blocks) and thus bitcoin would never "work" on a larger scale/be adopted due to the fact that there are much better/cheaper options out there.

I've skimmed through the text, so i might be completely wrong here though.

I'm not too sure if his math checks out either. We've seen bitcoin work reasonably well at a marketcap of >200 billion, so i'm not quite sure how accurate all his statements are.

[hatshepsut93]

I don't think that this paper uncovered something new, all these game-theoretic possibilities were discussed by Bitcoin enthusiasts since early days.
I think the author misses the fact that Bitcoin is more than just its hashing power, its main part is the community. Bitcoin community have already successfully defended itself against major attacks by miners and big companies - SegWit2x, Bcash and earlier forks. Maybe in the future banks and governments will try to launch 51% attack and mine empty blocks for a few days to scare users, but they can't keep it up forever, because it would cost millions every day, but users on the other hand will just have to wait, and with LN they won't be even bothered, since they will rarely need on-chain transactions.

[spartacusrex]

This guys clearly not a moron..

BUT..

In Section 2.1, the double spend section, where he calculates the fees you need to pay to deter someone attacking the network..

For example, suppose Bitcoin is used as a “store of value” akin to gold, .. Suppose it is possible to convert $1,000,000 of Bitcoin into other forms of wealth in the largest transactions in the system.

Ok - you can already do that today.

Then, even with an escrow period of e = 100 blocks, and an attack period of just one block’s worth of transactions, then, focusing now on A = 1.05, you would need ptransaction > 1/9.2*$1M ≈ $108700 for the system not to induce a majority attack.

He says you need to pay ~100K to not induce an attack...

And yet I see no attack.

He concludes :

This means that for the blockchain system to be robust against this double-spending attack would require that the pertransaction
payment to miners for running the blockchain exceeds 1/3.35 ≈ 30% of the highest-value transactions that are possible through the system.

Hmm..

I'm missing something or he did.

[nicosey]

He doesn't seem to be aware about what happened last year with the segwit2x debate, meaning there is not mention of the value of the network nodes.  This paper is all about the mining.

[dafunkizdiz]

Does the paper talk about offchain transactions? If not, then not worth reading.

[Spendulus]

http://faculty.chicagobooth.edu/eric.budish/research/Economic-Limits-Bitcoin-Blockchain.pdf

I will caveat this by stating that I am not a economist and have not spent time working through the formulas the author proposes to understand Bitcoin's equilibrium states. ....
My ask is for a technical/academic person to review this and provide your thoughts so I can more fully understand this paper.

I've read the thread but not the paper, and I have one immediate and relevant comment.

The peer review process of that paper is quite obviously THIS THREAD.

It is not the various "experts" to whom the paper was sent for peer review.

Various errors and questions have been raised in this thread. I'll try to have a comment in a couple of days that's more than just a quickie.

[buwaytress]

I *think* the author is trying to say that bitcoin cannot grow due to the fact that if it will, the incentive to do a 51% attack would become too big.

To counter this his argument is that the transaction value of each block must be atleast X amount of $, ( to make it unprofitable to do such an attack, (Right?)), which means that relative small transactions will become impossible, ( due to the high fees and limited blocks) and thus bitcoin would never "work" on a larger scale/be adopted due to the fact that there are much better/cheaper options out there.

But if you increase the value to $, then the gains from the attack would similarly increase dollar value. So as long as the Bitcoin value is profitable, the attack will always be profitable at any price, I would assume.

http://faculty.chicagobooth.edu/eric.budish/research/Economic-Limits-Bitcoin-Blockchain.pdf

I will caveat this by stating that I am not a economist and have not spent time working through the formulas the author proposes to understand Bitcoin's equilibrium states. ....
My ask is for a technical/academic person to review this and provide your thoughts so I can more fully understand this paper.

I've read the thread but not the paper, and I have one immediate and relevant comment.

The peer review process of that paper is quite obviously THIS THREAD.

It is not the various "experts" to whom the paper was sent for peer review.

Various errors and questions have been raised in this thread. I'll try to have a comment in a couple of days that's more than just a quickie.

Now that you point this out, I wonder when all these experts will realise this and actually come here on this forum seeking peer review. If you think about it too, a lot of the so-called peer reviewing done in academic research is self-serving. To be cynical, having this forum's experts probably disqualifies peer review since they're probably a lot more competent than the academicians typically doing these reviews (peer reviewers generally should be those of similar competence ).

[GSTremor]

I *think* the author is trying to say that bitcoin cannot grow due to the fact that if it will, the incentive to do a 51% attack would become too big.

To counter this his argument is that the transaction value of each block must be atleast X amount of $, ( to make it unprofitable to do such an attack, (Right?)), which means that relative small transactions will become impossible, ( due to the high fees and limited blocks) and thus bitcoin would never "work" on a larger scale/be adopted due to the fact that there are much better/cheaper options out there.

But if you increase the value to $, then the gains from the attack would similarly increase dollar value. So as long as the Bitcoin value is profitable, the attack will always be profitable at any price, I would assume.

http://faculty.chicagobooth.edu/eric.budish/research/Economic-Limits-Bitcoin-Blockchain.pdf

I will caveat this by stating that I am not a economist and have not spent time working through the formulas the author proposes to understand Bitcoin's equilibrium states. ....
My ask is for a technical/academic person to review this and provide your thoughts so I can more fully understand this paper.

I've read the thread but not the paper, and I have one immediate and relevant comment.

The peer review process of that paper is quite obviously THIS THREAD.

It is not the various "experts" to whom the paper was sent for peer review.

Various errors and questions have been raised in this thread. I'll try to have a comment in a couple of days that's more than just a quickie.

Now that you point this out, I wonder when all these experts will realise this and actually come here on this forum seeking peer review. If you think about it too, a lot of the so-called peer reviewing done in academic research is self-serving. To be cynical, having this forum's experts probably disqualifies peer review since they're probably a lot more competent than the academicians typically doing these reviews (peer reviewers generally should be those of similar competence ).

There is no limit to bitcoin as it can cost quadrifoliate equivalent so a cent so I don't see its aisles.

[jayanath123]

This is the my idea the economic limits of Bitcoin decided future activity of economic,   

[Quickseller]

I would argue, as the paper noted, that if miners were to perform any kind of large scale double spending attack, that Bitcoin would be largely worthless if the POW was not changed.

I would also point out that there is not (and never has been) a large scale rental market for miners, so an anon malicious user cannot simply rent a large portion of the mining network, they would need to actually buy (or create themselves) the underlying miners to perform this kind of attack.

The paper seemed to incorrectly assume that old, inefficient miners are available in unlimited quantities, which is not the case, especially at the below-cost-to-manufacture prices they will sell for. When a miner is no longer profitable, they will often be sold, and the manufacturer will not producing them (if they have not already), causing the prices of these miners to decline. When the new model of miners are manufactured in the same quantity, they will be able to "out mine" the prior class of miners.

If mining capacity is being rented out to a pool (via PPS), or a miner rental site, that is using said capacity in a way that is harmful to Bitcoin, the owner of those miners will move their capacity elsewhere in order to increase the chances of recovery of their "sunk costs" of their miners.

[Kakmakr]

I would argue, as the paper noted, that if miners were to perform any kind of large scale double spending attack, that Bitcoin would be largely worthless if the POW was not changed.

I would also point out that there is not (and never has been) a large scale rental market for miners, so an anon malicious user cannot simply rent a large portion of the mining network, they would need to actually buy (or create themselves) the underlying miners to perform this kind of attack.

The paper seemed to incorrectly assume that old, inefficient miners are available in unlimited quantities, which is not the case, especially at the below-cost-to-manufacture prices they will sell for. When a miner is no longer profitable, they will often be sold, and the manufacturer will not producing them (if they have not already), causing the prices of these miners to decline. When the new model of miners are manufactured in the same quantity, they will be able to "out mine" the prior class of miners.

If mining capacity is being rented out to a pool (via PPS), or a miner rental site, that is using said capacity in a way that is harmful to Bitcoin, the owner of those miners will move their capacity elsewhere in order to increase the chances of recovery of their "sunk costs" of their miners.

Exactly. We saw what happened to GPU cards when ASIC chips was introduced. For a while those cards were flooding the market and nobody was interested in buying them. Then some Alt coins came out with a much lower difficulty and these cards sold again, but they were never used for Bitcoin mining, because the difficulty increase rendered them obsolete.

The same thing happens with outdated ASIC miners. So I do not know where all this hashing power is going to come from to launch this attack.  <You have to hope that other role-players will not counter this attack, like they did with GHash>

[fredo123]

In relation with "Economic limits of bitcoin" While technological improvements in data collection could, in theory, improve central bank performance, Their benefits would be even greater under a system of free banking backed by some market - determined money. Individual banks are best equipped to access and quickly meet changes in the demand for inside money among their customers. If they could harness the power of big data to streamline this process, it's very likely the economy would benefit and the more relevant information banks have on loan applicants and potential investment opportunities. The better job they are likely to do in allocating their customers scarce savings. Any improvements in the efficiency of the process of financial intermidiation would serve to push out the economy's sustainable production possibility frontier and accelerate economic growth.