Topic: "Surprisingly, Tail Emission Is Not Inflationary" -- A post by Peter Todd - page 8. (Read 2760 times)

I don't see how this isn't true for Bitcoin's model as well. This miner "enrichment" seems to be a function of security in $ value.
It doesn't matter whether the miners get their reward from the block subsidy, transaction fees or a combination of both. If you end up with a secure PoW system, then you must be enriching miners.
As far as I can tell, the only difference is that Bitcoin's long term plan is to do this only through transaction fees. The assumption is that we'll have a constant backlog of high fee paying txs that will substitute the vanishing block subsidy.
Even if this turns out to be correct, the miner reward will stay the same (assuming the same security). I'd argue the fee-only model might be worse on that front because transaction fees directly transfer value from the existing
users to the miners while block subsidy prints new coins. Unless I'm missing something, the latter seems to be less of an enrichment.

Grin, at 3.5 years since launch, is only 3.5% into its soft total supply [1].
Compare that to Bitcoin, which at 3.5 years after launch was already 44% into its total supply,
Or to Monero, which was already over 80%.
Grin's emission is over an order of magnitude slower than any other coin.
So yeah, it will do miserably on marketcap for many more years to come.

Bitcoin reached the same 3.5% of its total supply in less than 3.5 months,
distributing only to a handful of people, and trading for fractions of cents.

I don't measure Grin's success by its marketcap, but by its elegance, simplicity [2], scalability, and long term survival.
For the short term, it may be only of academic interest in a crypto world that's rather dominated by speculation.


You have this weird idea that a coin's distribution is split into  a later phase of distribution to the "mining industry" which is enriching itself, preceded by an earlier phase of distribution to "everyone else", who are now being unfairly diluted.
Never mind that in Bitcoin "everyone else" were receiving much larger block rewards at much lower effort.

The "winner" that can't be displaced by dilution in their lifetime is Satoshi, not some tail mining operation that in a competitive market makes a few % profit at best and would need many lifetimes to match Satoshi's stack.

Miners are simply the mechanism that helps to fairly distribute coins to more people, and in Grin's case to more generations of people.

[1] https://john-tromp.medium.com/a-case-for-using-soft-total-supply-1169a188d153

[2] https://www.reddit.com/r/CryptoTechnology/comments/kyhgcv/are_there_any_public_cryptocurrencyblockchain/

It works, but then the limit is how long you are alive, there will be enough people bringing coins in circulation and if they dont, it wont impact the economy, as they’re like lost coins for the time theyre not being used. Its unrealistic to think that a human will never spend/ invest any considerable amount of money in their lifetime, when they have the opportunity to do so. And if theres a few people who saved till the end of their lifes, then idk if theres a justification for devaluing their wealth artificially, just because of this.

Also i get that inflation is an economic tool to incentivise spending, but we shouldnt act like it doesnt work without it. No inflation works as an incentive for wiser spending, and i prefer this alternative as its something the world could really need.

Could possibly happen, but then a coin with fixed supply might attract more people again, just like now.

Getting coins from other place than transaction fees means that you have to take them from the hodlers. And that can be done by using reversed coinage: you can check, in which block number some coin was created, and take a fee based on the number of confirmations. So, if you want to take one satoshi per 0.01 BTC from everyone, then if someone has 1 BTC, it means taking 100 satoshis per block. So, that means all coins will be taken after 1000000 blocks. That means, on average after 10 million minutes (around 19 years), each hodler will lose everything. Then, that hodler will have a choice: sign those coins, and send them as fees, or not sign those coins, and effectively get them excluded from the circulation.

So, that simple example means that getting a fixed tail supply fee as a single satoshi per 0.01 BTC in every block is too much. You can change amounts, make it percentage-based, again, it is all about proportions.

Edit: That's not true. If you are a miner of block number one, you have 50 BTC. And there are only 50 BTC in existence, so you have 100% of the coin supply. Then, over time, you have less and less percent of the supply, until reaching the total fixed supply, then you have a constant fraction of the supply. Later, coins are lost or burned, so if you still have your keys, then you have bigger and bigger fraction of all coins. So, hodling Bitcoin is a long time strategy that always works by design.

But if there would be tail emission, then it would be the same as with fiat currencies: we would start with 21 million coins. Then people would find out, that "we need more coins, because ". And then, more coins will be produced. Then, people will come and say: "this is still not enough". So, there will be endless "bailouts", not for banks, but for miners this time, and we will return to the point when we started, so burning coins will be proposed as a solution, because "hey, we have more coins, and the situation is not better, let's burn some of them, to make our coins worth more". And regular coin burning is what you can see on many altcoins.

How would tail emissions solve distribution? As whales with that much economic power will be able to acquire more mining equipment/ can afford the best suited locations and then getting more subsidies additionally. This could actually have the opposite effect, more concentration of wealth in the hands of early whales, than without tail emissions.

As Bitcoin grows in adoption, hodling will bring less and less gains, forcing these old hodlers to participate in actual economic activity/ spending if they want to benefit from their wealth, which will jump start redistribution even without tail emissions.

The problem might be the cost of mining in the future, making it too expensive for locations without close to 0 energy costs. If mining will be too centralised, a tail emission might not impact security as much, as it doesnt matter if the same small group of miners can afford to run more machines. And it might not actually be able to allow for better distribution of mining in the first place.

I wouldn't be surprised if a dozen well placed high altitude EMP optimized nuclear bombs couldn't destroy access to substantially more than half of all coins without killing anyone (well, okay, killing some through long term fall out, but not directly).

It also doesn't have to happen all at once to result in totally distorted rates over time, though obviously all at once makes it worse. 4 25% loss events is a 70% loss.

"Equal to what people lose by accident" is only an equilibrium reached after an unspecified, arbitrary, and potentially very long time.

You've posted an overly technical argument that after an unspecified timeframe constant subsidy will eventually match coin loss.  Yet in this discussion you are acting as though your conclusion guarantees that the subsidy is a tiny % of your wealth,  it doesn't.  This is a slight of hand.   I get that you intuitively and informally believe that would be the case, and I would agree that it's possible for it to be the case.  But it is also unambiguously possible for it to be very much not the case.

There have been many altcoin forks to change inflation rates, including several of ethereum's, as the highest profile and highest value example.  (there have been many others that have been forked to change the subsidy schedule, there was even an attempt to create a 50 BTC forever fork of bitcoin back around the first halving).

There are many possible ways to bring coins into existence.

The subject being discussed here is redirecting wealth from the existing users of the system in the long term in order to subsidize providing security to the system.

Tail subsidy as discussed by the author of this article doesn't solve that except to the extent that it forever enshrines an additional industry in that privileged position.

The potential for abuse can be irrelevantly small.

For example, allowing your coins to be spent by another party if they go 120 years-- which will cover the expected lifetime of the owner *and* their heir-- without moving appears to me to have an inconsequential risk for abuse (and the risk could be mitigated by making it somewhat longer).  And it would be economically rational for a very long lived entity to pay people small amounts to encumber their coins accordingly, and economically rational to accept those offers.

I think I've given a completely realistic way that it could become extremely low.  I think it's inevitable that if bitcoin continues to be widely used that such schemes will be adopted at some scale, though I admit I'm less confident that they'd be universal.

At the time what I'd considered was writing far-future timelocked spends as soon as the coins were received and sending them off via tor to a host that collected them.  Though today that kind of thing is better accomplished by having an alternative spending path with a relative timelock on it.

How's that been working out for you so far?    Speculation brings its own varrious annoyances, but it's an important part of adoption too.  I believe that a high inflation rate Bitcoin would never have taken off at all-- and, if anything, just been replaced by a low inflation coin.

It is, of course, impossible to prove such conjectures about alternative histories-- but as far as I can tell every high inflation rate altcoin  (or things with similar economic policy) have been total adoption failures so far.  I can't attribute this to Bitcoin's first mover advantage because they've been unsuccessful compared to other bitcoin alternatives not just bitcoin, including when they had more to offer.

I take it you would argue that this trend will reverse in the long term.  I doubt it: I think network effect is more important than jealousy (not that this doesn't conflict with the above: network effect doesn't happen until something is already successful).  It's just a fact of life that the people that came before you had myriad opportunities you missed.  Imagine how wealthy you could have if you bought apple shares in 1985 and held it till today, San Francisco real-estate in 1982, or traded your silver for gold in the year 1200?

[And you could apply this argument to real estate-- $/acre can differ by 1000 fold based on just the network effect of people already in a location.  Those who were there first are enriched by this.  People could abandon the high price place and go elsewhere-- they do to some extent, but seldom enough to change which places are expensive and which places are cheap.]

Economic activity diffuses wealth over time-- particularly if some systematic effect isn't sticking it back in the hands of the few as our central banks do today or as tail subsidy might in some cryptocurrency schemes. It might not diffuse as quickly as we might like.  The OP's argument was fine with reasoning about the state of the system arbitrarily far in the future: if we adopt that approach we can argue it doesn't even matter how the wealth was initially distributed, since at some arbitrarily far point in the future it won't matter.  But that doesn't hold true in the presence of tail subsidy:  Tail subsidy will always continue to enrich the mining industry (and those close to it in the economy) at the expense of everyone else, creating a "winner" that can't be displaced by diffusion.

What you've described here is different to what you have described previously. Taking a fee from every address, proportional to the amount of bitcoin being held on each address, and attributing it to future blocks is very different to taking a proportion of the transaction fees and attributing those to a future block. The former is sustainable, as you are always taking the same proportion of the same amount of bitcoin (21 million), creating a constant tail emission. The later is unsustainable, as you eventually reach a point where x% of the fees from the current block that you are locking for the future is smaller than the x% of fees previously locked which can now be claimed. With this arrangement you would eventually reach the point where the whole network is secured by fees alone (albeit with a small proportion of those feels delayed to a later block), which is the very thing you start out by saying is unsustainable.

You either have to create new coins and breach the 21 million limit, or you have to siphon coins from somewhere other than the block fees.

You are free to send coins to addresses like bc1qqph8gusf2x7ch4xjs8vnp7hy449r929wnv5jggmy678gam85l6rqgajus9. It has "1000000 OP_CHECKLOCKTIMEVERIFY OP_DROP OP_TRUE" script, so it can be claimed after a block 1000000 by any miner.

So make it a soft-fork proposal, where N out of M coins are locked in every coinbase transaction, to cover future rewards. It will have higher chances of being deployed than any other hard-fork, and will also increase future mining rewards. You want to get a 0.01 BTC tail emission? No problem, just send 0.01 BTC to block 1000000, then 0.01 BTC to block 1000001, and so on.

What about RSK? Of course it is not an altcoin, but it works only on transaction fees.

Edit: It is also possible to raise funds for each block by encouraging people to join. Let's test it on testnet3:
It simply means: "I agree to donate my 0.00071000 tBTC to the block 2360000, if other people will also put their coins in, and we want to collect 0.01 tBTC as a tail supply for that block". And it can be done for each block number separately, no fork is needed. But if some soft-fork will be deployed, and if some current fees will be locked for the future, the effect will be the same. Of course, to fully implement tail supply as a soft-fork, getting single satoshis from hodlers is also needed, but that can be achieved by increasing their fees as a punishment, then they will have a choice: hodl and not pay anything, or move coins, and pay a tax called "tail supply".

Let's suppose a truly devastating war somehow kills half of the global population. For reference, WW2 only managed to kill about 3-4%. We'll also suppose that every single person who was killed lost their private keys, and absolutely no-one left backups to surviving relatives.

If 10% of the world's wealth/year was going to mining after that catastrophic event, 5% of the world's wealth/year must have been getting destroyed in boating accidents.

Sorry, but that's a ridiculous example. And not because of the specific amount: the proportions mean that to get a measly doubling of reward - a reward so low it matches what people lose by accident - half the entire coin supply has to somehow be lost in a massive catastrophe.


In no world is 2x unconscionable where 1x was just fine. Especially when the 1x happened to be what people lose by accident on a regular basis.


Again, there is no way that we can be "disruptively overpaying" with a tail emission approximately equal to what people lose by accident.

...and yes, obviously the continued subsidy isn't guaranteed to be high enough to support security. But as long as Bitcoin is in the billions to trillions market cap range, the level of threat a 0.1% to 1% subsidy is protecting against is the kind of very expensive outside attack that's difficult to predict anyway. All you can do is spend an affordable level of wealth to protect yourself, and hope it'll be enough.

If it's not, there's a good chance that Bitcoin is entirely nonviable anyway, as it's just worth enough to be worth protecting. If Bitcoiner's need to spend 10%+ per year of their entire wealth to protect the network, it may not be worth it at all.


That's absurd. Overpaying isn't a concern when you're paying a tiny % of your wealth. What Satoshi did was take a perfectly good system, and build into it a massive long-term unknown in how we'll pay for security.

It's notable that people regularly pay hundreds of times more than strictly necessary in fees. Just look at any block explorer: https://mempool.space/block/000000000000000000030da833111fd3c4ade500b3b96963fadb4523475fe529

That block has a median fee of 15sat/vB, yet some transactions are paying as much as 2007sat/vB. Why? When you're moving millions of dollars fixing your fee estimator to save $10 isn't high on the priority list.

I sure don't care about spending 0.1% or 1% per year of my wealth to keep Bitcoin secure. But I do care that everyone else also chips in so we don't have a tragedy of the commons.


If you were right, that'd still be happening in alt-coins. So where are the examples?

Being first is much stronger attractor than zero... Especially when "first" in this case could also have just as easily marketed itself as "zero"


Yes, and by seeding the alternatives in peoples' minds now, it's much easier to have that discussion later. It's also more likely that alternatives will exist that don't make the mistakes Satoshi did.

I think people don't understand that clearly. Maybe there is a need to create a calculator, where people will type some amount in satoshis, and they will see, how fractions change over time. For example:
And then, some messages like that would make it more clear:
Then, they should click on "show tail emission", and see this kind of messages:

That's another problem, but to get the full picture, you can add punishments in a reversed way than coinage: the older coins you have, the more coins you have to pay to move them. And to preserve old signatures, it should be possible to require paying as a fee, no matter who will pay that, so for example if some old address has to pay 0.01 BTC, then it should be checked globally on the coinbase level, to make it possible to delegate payment to someone else, without invalidating old signatures from old transactions.

Edit: Let's see how many decimal digits are needed to express that precisely enough:
Yes, 32 decimal places seems to be a good choice for the start. But maybe reversed uint256 as a fraction would be better, or maybe some difficulty-like system, it is still an open question, how to display it easily, maybe storing amounts in satoshis and calculating it on the fly is the simplest solution, but still, there is a need to always display all fractions precisely enough to show, how it really works, because users should see, how their coins are affected by inflation, instead of seeing the same amounts and falsely thinking they don't lose anything from tail emission.

You don't get single a specific amount of satoshis from all accounts, though. As you said, from the transaction fee, 0.5 sat/vb goes for the tail emission and the other 0.5 sat/vb for the reward. This means that only those who make transactions are those who fund the future reward. And since you can't charge satoshis from all the addresses without the respective signatures, increasing the supply is the fairest option.

It is all about proportions. You can have two systems:
1) with fixed supply, where everyone will lose some satoshis in explicit way, and they will be taken by miners, because of tail emission
2) with infinite supply, where everyone will lose some satoshis in obscured way, because miners will be always rewarded by new coins, because of tail emission

Maybe you can't see that, so I will try to make some more extreme example, you can adjust numbers to reach some real-life scenarios:

Imagine there are 21 million coins, distributed to many different users, and the block reward is zero. Then, imagine that 21 million coins are produced, because of tail emission. Then, you can have two systems:
1) with fixed supply, where everyone will lose half coins in explicit way, and they will be taken by miners, because of tail emission: 10.5 million coins will remain in users' hands, 10.5 million coins will be taken by miners, 21 million coin limit is untouched
2) with infinite supply, users will have exactly the same amounts, but they will be worth 50% less than before, because miners will be always rewarded by new coins, because of tail emission

And then, imagine more real-life numbers. What result you will reach? You can do some simulations, then you will see that taking single satoshis from all accounts, and building new block rewards from that, is an equivalent to tail emission. It is the same, if you know all addresses and all amounts, then you can exactly calculate, what percentage of the total supply each user has, and then you can calculate, how many coins each user should own, to reach an equivalent situation, in a fixed supply coin.

It is simple. The total supply does not matter, all that matters is that is a fixed number. You can create a coin, where all amounts will be numbers from [0;1] range, then you can add fractions, and then you will explicitly see, that if we have 21 million coins, and if you burn 10.5 million coins, it is the same situation as doubling all balances on all accounts. And by contrast, if we have 21 million coins, and if you produce additional 21 million coins, it is exactly the same, as halving all balances on all accounts.

Then you should understand, that adding for example 0.01 BTC to the system as a tail emission, is effectively the same situation, as taking single satoshis from all accounts, and giving that to the miners. The difference is simple: it is done in a more obscure way.

If the tail supply fee is redacted from the transaction fee, then there's no actual tail emission, as described by o_e_l_e_o. The supply remains 21 million.

I think we should burn more Monero, for two reasons:
1) to show them that finite supply is better
2) to protest against such proposals
I don't know if Monero has some OP_RETURN equivalent (to avoid spam, because sending to some trap address will force all nodes to process that), but if they don't have anything like that, then it may be possible to burn them in a coinbase transaction, so it could be publicly noticed, how many coins were burned.

I don't know if on-chain transaction fees will drop below 1000 sats/kvB. What if they won't drop, forcing people to use second layers, like Lightning Network and sidechains, to not pay 1000$ for coffee?

Another thing is that even if tail subsidy will win, then coins still could be always burned. Adding new coins to the system is hard, but possible (see: zero satoshis). Removing coins from circulation is easy and straightforward, you can burn them by using OP_RETURN, or burn them by taking less amount in the coinbase transaction. So it is perfectly valid to join any inflationary network, and start burning coins. Because that's what will be eventually needed: tail supply will cause problems, that will eventually require a solution, for example based on burning coins.

And of course, claiming more coins explicitly in the coinbase transaction will obviously not work, because it will be unnecessary hard-fork, where there are soft-fork or no-fork solutions available.

That doesn't really solve the problem, though. Let's say you want the tail emission to be a fixed amount, for example 0.01 BTC per block. After the 13th halving, the block subsidy drops to 0.00610351 BTC, meaning you can no longer commit that amount to future blocks. If you say instead that you want the tail emission to be a variable amount, and to commit 10% of all transaction fees from each block to block n+6,000,000 (for example), then there will still be plenty of blocks which are empty or near empty and therefore can only lock up trivial amounts or none at all for future blocks.

Even if you decide to commit 10% of the total block reward (subsidy + fees) to future blocks, really you are just delaying the inevitable. If fees alone are not sufficient to sustain the network, then simply taking a cut of those insufficient fees and allocating them to the future will not be sufficient either.

No, it is preferable to rise the transaction flow instead, by compressing things, and by joining transactions. It doesn't matter if you have one transaction paying 0.01 BTC fee, or if that transaction was created by joining many small transactions into one. If you have a single user paying 0.01 BTC, or if there are 1000 users, paying 1000 satoshis each, it doesn't matter, if the size of that transaction is identical.

Exactly, tail emission is preferable, when blocks are almost empty, so there are almost no fees, and if the block subsidy is zero, then there is a problem, because there is no incentive to keep mining.

Also true. If you assume 1 sat/vB as the minimal fee, and if you assume 4 MvB as the maximum block size, you will get 0.04 BTC per block. And by looking at coinbase transactions, you can quickly notice that the current fees are higher than that. I guess even locking 0.01 BTC per block for the future rewards will be rejected, but if someone wants to do that, then that person can start mining, and then see, how it is to voluntarily reject some reward, and lock it for the future. And if someone is not a miner, then that person can try to do that as a user, because anyone can send coins to addresses like bc1qqph8gusf2x7ch4xjs8vnp7hy449r929wnv5jggmy678gam85l6rqgajus9, then miners could claim that later by creating a transaction that will send all of that as a fee. Also, using timelocked transaction with some high fee, and publishing that, will have the same effect. It is a matter of fee policy, adding anything to the minimal 1 sat/vB, and creating any rule to lock N satoshis out of all fees to M block would solve that, without any hard-fork.

Yes, it is in practice an attack. Imagine a situation, where you will have 1 BTC, and you will lose single satoshis, one-by-one, because of some soft-fork. That would be exactly the same situation, if you will have tail supply, the difference is that if you have a fixed supply, you will explicitly see that you lose something, but with the infinite supply, the same thing is done in a more obscure way.

Any miner can require a minimal fee of X. Currently, we have 1 sat/vB as a reasonable default. But nothing stops miners from saying, that for example now we will have 0.5 sat/vB minimal fee for them, and 0.5 sat/vB locked for the future. As you can see, no fork is needed.

Exactly. But you don't know, which miner will pick that later, because any miner will have a chance to do that.

Exactly. And the best thing is that you can do that as a user, and you can do it now, so no forks are needed to start it.

Three questions:

• If mining incentive is the problem, isn't it preferable to rise the block size instead?
• Isn't tail emission preferable when there's little transaction activity? Isn't the exact opposite happening on bitcoin?
• Wouldn't a tail emission be a direct attack on one of the principles, which is non-arbitrary inflation schedule?

How can you force a user pay a "tail supply fee" in a no hard-fork way? Don't you need their signature?

Isn't this going to just delay the miners' reward? As far as I understand, you're taking the current reward (transaction fees) and send it to a future block.

Exactly. So, if someone really wants a tail supply, then funding addresses like bc1qqph8gusf2x7ch4xjs8vnp7hy449r929wnv5jggmy678gam85l6rqgajus9 is an option, as vjudeu said. And if someone really wants to make it a consensus rule, then it can be simply done by getting all transaction fees, and then making a soft-fork that will lock some part of the coinbase transaction into some future block number (for example, the current block number plus 210000). That's all what is needed to make a tail supply.

Sure. But in both cases the outcome is the same, in that people holding bitcoin will have their holdings effectively taxed, either by losing 1 satoshi per 0.01 BTC as per your example, or by having a constant low-level block reward. Not that I think we should have either, but if you were going down this path then I think that the constant block reward would be the simpler (and more emotionally acceptable to most people) of the two to implement.

Enforcing "all transactions" rather than "all addresses" creates a tax on transactions, though, which disincentivizes bitcoin's use as a currency.