Transaction Fees as Marketable Security?

[mickeyshrugged]

I've been researching and still quite can't get my head around the time series of transaction fees as an income stream. Does it necessarily terminate at some point or can a miner in theory be paid transaction fees into perpetuity? Can the rights to these fees be bought and sold like marketable securities?

If I understand it correctly, once a miner finds a block, he or she in some sense owns the block, or owns the destination address for transaction fees. What is that called? Is there a market for that property?

[DeathAndTaxes]

There is no time series.

A miner includes tx in a block, one of those transactions is a special one called the coinbase it has no input and only an output address(es).  The miner sets the value of the coinbase to the block subsidy plus the sum of all tx fees in the block.  When solved the miner is rewarded the sum.

It is a one time payment.  If a miner never solves another block there will be no future payments.

[Gabi]

Every transaction have its fee. Miners find a block, receive the new bitcoins, insert transactions and get the fees for these transactions. As long as bitcoin is used around the world and there are transactions, miners receive the fees.

[mickeyshrugged]

There is no time series.

A miner includes tx in a block, one of those transactions is a special one called the coinbase it has no input and only an output address(es).  The miner sets the value of the coinbase to the block subsidy plus the sum of all tx fees in the block.  When solved the miner is rewarded the sum.

It is a one time payment.  If a miner never solves another block there will be no future payments.

So when I send BTC to someone and add a fee to make it go faster, is that the moment in time the miner receives the fee? Or are you saying that the miner already received the fee when the block was solved?

Is there perhaps a diagram that shows, upon creation of a block, where the fee goes and at what time?

Because any time someone has something that receives payments, they have something that can be sold, even if it is not nominally worth any money. The price would be determined by the present value of whatever estimated future payments might be received, like an annuity or income from a business. But it sounds like what you're saying is that's not how it works. In order to receive any payments after solving a block, they must solve another block? Even so, does the miner not then have an asset that receives income, which can be bought or sold?

[DannyHamilton]

So when I send BTC to someone and add a fee to make it go faster, is that the moment in time the miner receives the fee? Or are you saying that the miner already received the fee when the block was solved?

When you send BTC to someone, the transaction is relayed throughout the network.  The transaction essentially has the fee embedded inside it and nobody has it yet.

Eventually a miner will create a block and include the transaction in the block that they create.  At that time, the transaction will then have "1 confirmation".  The miner that includes the transaction in the block that they create takes the fee for themselves at that time.  They do this by creating a special transaction in the block called a "coinbase" transaction.  This transaction has no inputs and has as the sum of its outputs a value that is no larger than the sum of the block subsidy and all the transaction fees from all the transactions that they included in the block that they created.

Is there perhaps a diagram that shows, upon creation of a block, where the fee goes and at what time?

There may be a diagram somewhere, I'm not sure.  I'm not sure that a diagram is necessary.  Upon creation of the block, the miner creates a coinbase transaction that pays the fees from the as yet unconfirmed transactions that they choose to include (and block subsidy) to an address that they own.

Because any time someone has something that receives payments, they have something that can be sold,

Right.  So once you receive the bitcoins (as a fee), you can sell them.

The price would be determined by the present value of whatever estimated future payments might be received, like an annuity or income from a business. But it sounds like what you're saying is that's not how it works.

Correct, that's not how it works.

In order to receive any payments after solving a block, they must solve another block?

Yes.  If they solve another block, they will receive the block subsidy from that block as well as the transaction fees of the as yet unconfirmed transactions that they choose to include in that block when they create it.

Even so, does the miner not then have an asset that receives income, which can be bought or sold?

Yes, they have hashing power.  There are miners that sell shares of their hashing power.

[mickeyshrugged]

Oh, I see. So the fees don't go to "the miners" per se, they go to future miners. Potentially all the existing miners could decide to give up on the whole enterprise, but the currency could still keep going because a new crowd of miners could come in and reap the fees of transactions waiting to be confirmed. That's really very fascinating and interesting. Thank you so much for helping me understand it.

[Gabi]

Yup, of course the fee will be collected by the miner that put your transaction in a block. Miners mine because it is profitable 

[gfoot]

It's instructive to actually have a look at a block in blockexplorer - here is the most recent one, as of writing:

http://blockexplorer.com/block/0000000000000224615e7dc4006fcc1baec60b7cc3fa102332575ef53d2b8be0

Note the list of transactions embedded in the block, and that the first transaction has no "from" address, this is the miner receiving the block generation subsidy plus the total value of transaction fees for all the other transactions in the table.  For the other transactions, for any with fees you should find that the sum of the outputs, plus the fee, is equal to the sum of the inputs.  That is where the money comes from, and if it didn't add up the transaction would have been rejected as malformed way before being included in a block.

[mickeyshrugged]

I see, thank you. You're right it is making a great deal of sense now.

Yeah, somehow I got the impression that the fee went to the miner who originally created the BTC you wanted to send, so you can imagine how confusing that would be, lol.

There's still much to be curious about, like the relationship between the quantity of transactions awaiting confirmation and the cost of solving a block.

[DannyHamilton]

- snip -
There's still much to be curious about, like the relationship between the quantity of transactions awaiting confirmation and the cost of solving a block.

And the maximum number of transactions that a single block can hold (actually maximum bytes), and the ratio of fees on waiting transactions per byte of transaction size.

Since there is a limit on how many bytes the block can hold, it is most profitable for the miner to choose the transactions for the next block that have the highest fee-per-byte.

It is possible that there are unconfirmed transactions that use other unconfirmed transactions as inputs.  In this case it could be more profitable to include a transaction that has a low fee-per-byte ratio IF it means that you can also include an additional transaction that has a very high fee-per-transaction ratio.  As far as I know there aren't any miners that are taking advantage of this opportunity to increase their profits, but as soon as they do, it will become possible for the receivers of bitcoins to add their own fee to a set of transactions if the senders fail to do so.

[gfoot]

I think the reason miners are not taking advantage of ways to increase transaction fees is that the transaction fees are miniscule compared to the block subsidies at the moment.  It will be more relevant in future, or would be now if more people voluntarily pay very high transaction fees.  But I don't think that's a good position to be in.

This is all as-far-as-I-know, please correct me if I'm wrong.  But -

The thing that bothers me in the longer term is that miners have absolutely no incentive to include a zero-fee transaction in the block, whether it's so-called "high priority" or not.  Even in the short term, because the subsidy outweighs the transaction fees by a significant margin, you might as well just never include any transactions in your blocks.  Not healthy for the network!  I noticed a block the other day which was the first one for about half an hour, and it only included six transactions.  That is not useful miner behaviour.

Is it not possible to offset the proof-of-work requirement by some metric involving number of transactions in the block and/or volume of transactions, so that miners have an incentive to do useful mining instead of just hoarding the subsidy?  Or is that too open to abuse?

[DannyHamilton]

I think the reason miners are not taking advantage of ways to increase transaction fees is that the transaction fees are miniscule compared to the block subsidies at the moment.

I think the average number of transactions that a 1 megabyte block can hold is somewhere around 4,200.  If all those transactions included the current recommended minimum fee of 0.0005 BTC, that would be 2.1 BTC.  That is an 8.4% premium on top of the subsidy.  At the current exchange rate (approximately $140), that's an extra $294.  Seems silly to leave $294 sitting out there unclaimed everytime you solve a block.  What's the benefit to a miner of not putting the extra $294 in their pocket?

The thing that bothers me in the longer term is that miners have absolutely no incentive to include a zero-fee transaction in the block, whether it's so-called "high priority" or not.

Well there is the collective benefit of increasing the popularity of bitcoin so it attracts more users, some of which whom may pay fees in the future.  (Like a loss leader in retail).

Even in the short term, because the subsidy outweighs the transaction fees by a significant margin, you might as well just never include any transactions in your blocks.

To what benefit?

Not healthy for the network!  I noticed a block the other day which was the first one for about half an hour, and it only included six transactions.  That is not useful miner behaviour.

Sure it is.  It prevents attacks that would modify a confirmed block.  Every new block contributes to that protection regardless of whether the new block has many new transactions in it.

[gfoot]

Maybe I was exaggerating, I guess it's not completely insignificant and it will become much more so in a few years' time, but it was clearly of no interest to the miner of that six-transaction block.  Maybe that was a particularly unusual case though.  You're right that even empty blocks do improve security by building on top of existing blocks, but it would be even better if it also dealt with some of the backlog that must have built up during that period.

For pooled miners, which receive transactions pre-hashed from a remote server, there seem to be distinct disadvantages to frequently breaking off work and asking the pool server for an update, especially with all the DDOS attacks against pools recently.

[mickeyshrugged]

- snip -
There's still much to be curious about, like the relationship between the quantity of transactions awaiting confirmation and the cost of solving a block.

And the maximum number of transactions that a single block can hold (actually maximum bytes), and the ratio of fees on waiting transactions per byte of transaction size.

I assume this is the main technical talking point regarding discussions about scalability?