let's think about what creates market for transacion fees

[niko]

I am still trying to form an opinion about the block size limit.  One of the problems is that most of related threads attempt to discuss this issue in general, and are therefore extremely broad; a flood of arguments on various facests makes it impossible to stay focused on any one of the important facets.

One particular detail that I would like discussed is the question of the economics of transaction fees. This is what got me thinking:

Without a sharp constraint on the maximum blocksize there is currently _no_ rational reason to believe that Bitcoin would be secure at all once the subsidy goes down.

Bitcoin is valuable because of scarcity. One of the important scarcities is the limited supply of coins, another is the limited supply of block-space: Limited blockspace creates a market for transaction fees, the fees fund the mining needed to make the chain robust against hostile reorganization.

It seems to me that this is not quite the case. limited block space does not create a market for transaction fees.  If it did, there would have been no transaction fees paid nor requested in all the years up to now, because we almost never got close to the limit. In other words, the blockspace has practically not been limited (due to relatively low transaction volume), and yet there are fees.

If the above is true, what drives the market for transacion fees?  It seems to me that it is the need to validate transactions and have them included in the blockchain: ultimately, it is miner's cost of hashing that creates the market for the fees. Not the (not so) limited block size.

What are your thoughts?

[markm]

Sounds reasonable.

The block size limit makes it possible to come to market in the first place; it goes along with things like knowing the marketplace is not a trap laid by the guild of assassins, who have snipers on rooftops all around waiting for anyone to be fool enough to come to the market to be shot.

-MarkM-

[misterbigg]

One particular detail that I would like discussed is the question of the economics of transaction fees. This is what got me thinking...What are your thoughts?

You need to think some more and come back when you understand Bitcoin a little better.

The only reason that transaction fees are paid for now is because clients will not relay messages unless they have a minimum fee (to prevent spam). This minimum fee can be zero for unspent coins that have sufficiently aged. Aside from this minimum fee for relaying transactions, there is currently NOTHING that creates a market for transaction fees, because:

1) There is no scarcity of transaction space in the block chain (we are not reaching the 1MB limit at all)

2) The block subsidy of 25BTC is orders of magnitude larger than the sum of transaction fees in a block.

In short, there is very little incentive for anyone to pay fees now. Which is why the total amount of fees per block is so low.

Anyone can figure this out on their own by just reading the available literature. This goes for most of the people who are posting in these max block size threads.

To answer your question, limited block space DOES NOT create a market for transaction fees TODAY; But there are still some fees because of the rules clients impose on relayed transactions. However, when the average transaction volume regularly exceeds the space available in a block (i.e. we always reach the maximum block size), this will create a market for transaction fees.

[niko]

To answer your question, limited block space DOES NOT create a market for transaction fees TODAY; But there are still some fees because of the rules clients impose on relayed transactions. However, when the average transaction volume regularly exceeds the space available in a block (i.e. we always reach the maximum block size), this will create a market for transaction fees.

Miners can choose and prioritize processing based on fees paid. They don't today, for the same reason you brought up - block subsidy is still high - but they will in the future. Doesn't this create a market for fees, without limiting the network processing power to the ridiculous ~7 tps?

[misterbigg]

...when the average transaction volume regularly exceeds the space available in a block (i.e. we always reach the maximum block size), this will create a market for transaction fees....

Miners can choose and prioritize processing based on fees paid....Doesn't this create a market for fees, without limiting the network processing power to the ridiculous ~7 tps?

No. Unlimited transaction rate means unlimited block size. If the block size was unlimited, miners would simply include all transactions they know about, as long as the fee was non zero. In the short run, this would maximize their revenue. In the long run, users would quickly discover that their transactions always make it into the block chain quickly as long as they is any fee (even down to the minimum tx fee for relaying). This would drive the fees down to the minimum. That's not a "market for transaction fees."

This analysis ignores the drop in network hash rate that comes from pushing bigger blocks around through the overlay network (a separate but important problem).

tl;dr; Without a maximum block size, fees will trend towards zero.

[niko]

I only now realized that hashing is done on the block header only, so it doesn't involve raw transactions. Indeed, there can be no fee market driven by the "scarcity of hashing power" as I claimed above. Unless the algorithm is changed to involve hashing all transactions, the only option is to have a predetermined, and sufficiently small, block size limit (or a predetermined formula for a dynamic limit).

[nevafuse]

tl;dr; Without a maximum block size, fees will trend towards zero.

I completely disagree.  Mining isn't free.  Therefore once the block reward disappears & assuming no block size limit, miners will rely on transaction fees to pay for equipment, electricity, etc.  There may be some miners that process free transactions, but they will most likely have small hashrates & find blocks rarely.  The miners with higher hashrates will require a fee to pay their expenses & increasing your chances of getting into a block sooner.  We may also get to a time where trying to include every pending transaction in a block increases your orphan risk because it doesn't propagate fast enough.  Your orphan risk is the only block size constraint necessary.  And this will fluctuate naturally with bandwidth improvements.

[markm]

It is not the only constraint necessary, even if it is the only one necessary for motivating / profiting miners.

We still should have an absolute limit, and (possibly unfortunately for miners) it does seem prudent that the absolute limit be at least slightly larger than whatever limit suffices in slow seasons to support miners. Christmas might only come once a year, but it does come. So the max limit should be high enough to accomodate Christmas even if miners find it convenient or expedient to create blocks smaller than the max the rest of the year.

-MarkM-

[Technomage]

There is obviously going to be a market even if there is no scarcity with the block size. It's not going to be much of a market, the fees would go really low and the hash power really low as well, but there would be a "lowest point" so to speak. This is not what we want however, the hash power would be very much inadequate for a monetary system as valuable as Bitcoin.

To create a functioning market with some scarcity there has to be some changes to how it works now. Well, there would be a very efficient market for sure if we just do nothing, but then we simply limit the capabilities of Bitcoin for no good reason.

Some sort of a compromise between total scarcity and no scarcity needs to be found with the block size, it was never meant as something with total scarcity. The total amount of bitcoins was meant to be like that though. Anyone thinking otherwise are completely deluding themselves, changing the block size max in some way was always the idea since day 1.

[hazek]

I am still trying to form an opinion about the block size limit.  One of the problems is that most of related threads attempt to discuss this issue in general, and are therefore extremely broad; a flood of arguments on various facests makes it impossible to stay focused on any one of the important facets.

One particular detail that I would like discussed is the question of the economics of transaction fees. This is what got me thinking:

Without a sharp constraint on the maximum blocksize there is currently _no_ rational reason to believe that Bitcoin would be secure at all once the subsidy goes down.

Bitcoin is valuable because of scarcity. One of the important scarcities is the limited supply of coins, another is the limited supply of block-space: Limited blockspace creates a market for transaction fees, the fees fund the mining needed to make the chain robust against hostile reorganization.

It seems to me that this is not quite the case. limited block space does not create a market for transaction fees.  If it did, there would have been no transaction fees paid nor requested in all the years up to now, because we almost never got close to the limit. In other words, the blockspace has practically not been limited (due to relatively low transaction volume), and yet there are fees.

But we aren't paying any fees.. well not really. All we are paying is a fraction of a cent which is a spam protection measure more than anything else, but essentially there are no fees.

[markm]

Yes. Personally I happen not to care whether there are any fees to be had in making a block or not, only that processing blocks remains viable even if the only fees one could ever hope to make doing it come from all the many interesting applications that could spring up in the merged-mining space.

For example I do not want namecoin to be unable to be merged-mined alongside bitcoin due to bitcoin being such a massively bloated high speed neverending stream of worthless spam/dust that even given the huge hashing power it brings to the problem of securing blockchains it is just insanely impractical to process the damn thing, let alone in the context where it is not even your main profit-centre, just a chain-securing resource you would like to use to secure your cancer-curing genes study or whatever application you happen to be using a blockchain for and are looking to secure that blockchain.

One could merged-mine against the highest difficulty alternative high difficulty hashing provider, but if one's app turned out profitable, that would in effect be to support a potential attacker of bitcoin instead of supporting bitcoin.

It seems better to hope to be able to have the bulk of the world's SHA256 hashing power merged mining alongside bitcoin than to plan on diverting enough of it away from bitcoin to make all other chains secure.

-MarkM-

[hazek]

tl;dr; Without a maximum block size, fees will trend towards zero.

And not only fees, but so would Bitcoin security in the form of hashing power.

Just think about it, why would a miner shoot himself in the foot and reject transactions below a certain fee? While he is left with no fees, other miners take what's available and maintain at least some sort of a profit (I'm talking once the subsidy is gone..). It's the same argument for why a market strictly regulated by consumption i.e. a free market would never support a racist business. While the racist business discriminates against the customers, another business opens up down the road who doesn't and simply outcompetes it.

The only way miners could set a minimum transaction fee with an unlimited block size is if it was just one miner or if they form a cartel and somehow are able to prevent any other miners to enter the scene and compete with them.

[justusranvier]

There is obviously going to be a market even if there is no scarcity with the block size. It's not going to be much of a market, the fees would go really low and the hash power really low as well, but there would be a "lowest point" so to speak.

Let me make sure I correctly understand what you're saying.

If miners are not limited by protocol to a fixed block size, space in the blocks will no longer be scarce. Miners have an infinite amount of bandwidth, disk space, and processing power so they will be able to include an infinite number of transactions in a block with absolutely no marginal cost, because if any of these things are not true then space would be scarce.

Because miners have the infinite capability to generate blocks of unlimited size, they will automatically do so. This will make space in the blockchain is free nobody will include fees in their transactions. The only possible option miners will have to respond to this by quitting mining entirely, strangely even though we already established they have infinite free bandwidth and processing power to mine with.

Bitcoin will fail because mining, like every other business in the world, becomes cripplingly unprofitable if the industry is allowed to increase capacity to match higher demand. The amount of transactions that can be processed must remain fixed in order for people to be willing to invest capital equipment into mining and thus keep the network secure. It's just like how we limit the worldwide car market to 100 new cars per year in order no matter how many people want to buy. It's the best way to keep the industry decentralized.

[hazek]

There is obviously going to be a market even if there is no scarcity with the block size. It's not going to be much of a market, the fees would go really low and the hash power really low as well, but there would be a "lowest point" so to speak. This is not what we want however, the hash power would be very much inadequate for a monetary system as valuable as Bitcoin.

To create a functioning market with some scarcity there has to be some changes to how it works now. Well, there would be a very efficient market for sure if we just do nothing, but then we simply limit the capabilities of Bitcoin for no good reason.

Some sort of a compromise between total scarcity and no scarcity needs to be found with the block size, it was never meant as something with total scarcity. The total amount of bitcoins was meant to be like that though. Anyone thinking otherwise are completely deluding themselves, changing the block size max in some way was always the idea since day 1.

Glad to see some sense out of you for a change, I thought I wasn't ever again going to read anything you write that I'd agree with nearly 100%.

[Technomage]

There is obviously going to be a market even if there is no scarcity with the block size. It's not going to be much of a market, the fees would go really low and the hash power really low as well, but there would be a "lowest point" so to speak. This is not what we want however, the hash power would be very much inadequate for a monetary system as valuable as Bitcoin.

To create a functioning market with some scarcity there has to be some changes to how it works now. Well, there would be a very efficient market for sure if we just do nothing, but then we simply limit the capabilities of Bitcoin for no good reason.

Some sort of a compromise between total scarcity and no scarcity needs to be found with the block size, it was never meant as something with total scarcity. The total amount of bitcoins was meant to be like that though. Anyone thinking otherwise are completely deluding themselves, changing the block size max in some way was always the idea since day 1.

Glad to see some sense out of you for a change, I thought I wasn't ever again going to read anything you write that I'd agree with nearly 100%.

My opinion on this particular issue has been along those lines ever since this debate started. Sometimes the discussions can get sidetracked and people don't really know the whole argument though.

[jgarzik]

tl;dr; Without a maximum block size, fees will trend towards zero.

And furthermore, create solutions that incentivize block chain spam, and discourage creation of alternative, more efficient solutions outside the block chain.

[markm]

It's just like how we limit the worldwide car market to 100 new cars per year in order no matter how many people want to buy. It's the best way to keep the industry decentralized.

I think it is probably more like how we limit the speed limit to a hundred thousand miles per hour to limit the number of cars that can make the journey between point A and point B per unit of time, thus limiting the number of cars built because the slower the speed limit the number of people actually interested in going from point A to point B by car declines.

Or something.

-MarkM-