Turn on Taker Fees at 0.1%

This proposal is to set the default Taker Fees at 0.1%.

Previously Prop 530 passed a signaling proposal to set Taker fees at 0.15% but there were some concerns about the impact this could have on multihop transactions and potential impacts on volume. Since that time, Osmosis has moved to Supercharged Concentrated Liquidity Pools which reduced spread factor fees from 0.3% to 0.2% on many pools, mitigating some of the concerns.

In a twitter poll 43% of respondents thought fees should be set at 0.1% or higher while 39% thought the fees should be set at 0.05%.

Further governance discussion took place on the Osmosis forum with mixed opinions between 0.05% - 0.15%:

It appears reasonable to take the median response in these polls and set the default Taker Fee at 0.1% at this time to get meaningful data back to determine if there should be any adjustments.

Other Factors

In Prop 629, a subdao has been created to delegate the setting of NON-DEFAULT Taker Fees to a subdao. This subdao will be able to reduce Taker Fees of individual pairings and make exceptions to the rules set forth by governance.

By passing this proposal and establishing 0.1% Default Taker Fees, this subdao will have more data to determine appropriate course of action for individual pairings. 0.1% will be at a meaningful level but also reduces the incremental jump to 0.15% if required.

Target on chain Date: Monday, September 25th.

3 Likes

Supportive of this! Looking at the text of Proposal 629, it seems that the subdao can’t even really start its work until a default fee is implemented by governance.

I don’t think there’s universal support for a 0.1% taker fee, but this seems like a good baseline compromise number for the subdao to start with, adjusting the pools where needed to account for pools where a higher fee might drive away volume (Stables, LSTs, etc).

As @whitemarlin pointed out, I feel like the most important thing right now is to start generating some impact data of a taker fee on pool volume. There have been a lot of discussions stating that this will impact volume, but the extent of this isn’t really knowable until a fee is actually implemented and you have a few weeks of historical data to measure volume deltas.

If there’s a substantial impact, Osmosis governance (or potentially even the subdao itself) could return the fee to 0 and come up with a plan from there.

1 Like

As I mentioned elsewhere, I think 0.1% is a good place to start (it wasn’t my preferred fee) considering the concerns of increasing the fee at any point could be worse than just setting the default to a midpoint (0.1%).

IF there was strong opposition to this, we could do a dual proposal with 0.1 and 0.05% although I think this will pass.

To summarize, in favor and appreciate you bringing this forward.

As I mentioned in the previous post, I think this is mistimed. For two-pool hops, the taker fee should only be applied once AND the multihop discount should not be removed, OR we should wait till crypto/USDT pools have better liquidity because right now they have very poor liquidity. Otherwise, effective fees for osmosis will jump from 0.2% now to 0.6%. 0.6% fees are 12 times as high as dexes like Uniswap and twice as high as those on astroport, which seems wholly unreasonable.

To the point towards volume, I believe that you can get a decent idea by looking at the volume in pool 1 (ATOM normal) and pool 1135 (ATOM supercharged). The normal pool has much higher liquidity and the supercharged one is more efficient. However, the ATOM normal has 2x higher fees (0.4%) due to lack of incentives/multihops which is lower than what is suggested to be implemented. Thus it’s routed through less on the frontend and has 6x less volume. AND ATOM has been very volatile. If it’s not volatile, the price will constantly move in the range that the price is unfavorable due to fees (this range will be 1.2% with the proposed fees unless trader are illogical).

2 Likes

We’re basically discussing how much to price gouge people for swapping on Osmosis, instead of focusing on being competitive.

1 Like

All platforms charge fees. Spreadfactors have been reduced from 0.3% to 0.2% and the proposed Taker Fee is 0.1%. Most transactions (excluding arb trades) are for below $1,000. A $1,000 transactions would have a $2 spreadfactor fee (0.2%) + $1 Taker Fee (0.1%) in this model. A $3 fee on a $1,000 transaction seems very competive against other CEX exchanges as well as DEX’s which have higher gas fees as well. This will help create a sustainable model for the OSMO token and has received a lot of support from the community.

1 Like

Please for the love of Osmosis, stop disincentivizing people from using this exchange.

We’re pushing updates (removing multihop, now taker fees) that look like they’re adding value to stakers on the surface, but they’re making it more expensive to transact on the exchange. This not only reduces the amount of volume stakers are generating revenue from, but it reduces the price/value of osmosis as an ecosystem AND a coin by removing utility.

Essentially we’re moving towards stakers taking a bigger part of a consequentially shrinking pie, rather than trying to make a bigger pie. This platform needs to rethink its vision. Make it attractive to use, rather than prohibitive.

There’s cheaper DEXes, why would any DEX user continue to use osmosis after these greedy updates?

(repost of my msg on the topic last night, as discussion moved here soon after).

But adding to that, is there any logical/economic reason to make the DEX more prohibitive than it already is when more economic options ALREADY exist?

1 Like

Most people do not trade from osmo to a crypto, they go from USDC or USDT to a crypto (same as an exchange). So the current spread factor is 0.2%. Since we will lose the multi hop bonus, the same swap from USDC to a crypto will go from 0.2 to 0.2+0.1+0.2+0.1 = 0.6%. For the biggest CEX, binance, the taker fee to do a trade is 0.075%, which means osmosis will have fees 8!!! times higher than binance. Is that not unreasonable? This is before we even consider that the price on binance is always going to be better or the same as osmosis (otherwise a bot will make the trade, presumably)… Which successful dex has 0.6% fees? Now if the multi-hop discount is not removed, then 0.3% fees are toe-ing the line of reasonable, and we can have a discussion…as it stands, the removal of the multi-hop discount makes even a 0.0000000001% taker fee crazy.

2 Likes

Has there been any thought to the obvious conclusion that lower fees would lead to increased volume to compensate for the lower % charged? It’s consistent too, arbitrageurs would react to an earlier margin, and the exchange would consequently become more useable for actual swappers as well.

Uniswap Interface You can see that volume scales with lower % pool fees.

We should be considering market GIVER instead of market TAKER given the currently already prohibitive pool fees (or just scrap this death spiral inducing idea and consider reducing pool fees to make osmosis attractive versus DEXes people actually use).

Just for the sake of transparency; the conversation in the subDAO (which is indeed meant for analyzing and maintaining the non-zero taker fee) was still at a stage whether it should be turned on now or when we have thought about how to monitor.

There was nothing like a consensus on that field (rather, everyone was still testing the waters, until 1 made a decision to just go ahead). So, be aware that when this goes live there is NO method to monitor what happens or any framework how the subDAO will steer the taker fee in such a way that it won’t drive away traders.

For more sake of transparency; before the subDAO started I already left if because of this atmosphere. The taker fee passed governance, that is very true. So @whitemarlin is very right that it should also be executed. It did not include how to implement it though and just switching it on without giving a thought on how to monitor is a very bad idea imo. One idea mentioned in the DAO was to do it in small incremental steps of 0.05% every month, which gave the rare opportunity to monitor the change in volume and TVL and the value accrual. However, this never made the cut while multiple people were in favor. In the current format my vote will still be a NO to switching on any form of taker fees until we have thought about how to monitor, how to steer and how to adjust quickly if needed.

1 Like

There are two things going on here.

  1. The default Taker Fee is up to Osmosis governance and should be subject to Osmosis governance. Osmosis governance previously signaled support for 0.15% with consensus.

  2. The subdao would have ability to monitor and adjust if necessary, however per prop 629 changes can still be vetoed by Osmosis governance if necessary.

By taking the median rate, this is providing the SubDao team more flexibility to adjust up or down, and is ultimately Governance’s decision.

I also would not say there is No method to monitor what happens. This proposal will not be on chain until September 25th and will take time to pass as well. This gives the SubDao 9 days to formulate models and monitoring. I have also shared some manual spreadsheets that I had that tracked roughly 40% of the TVL. This can also be completed by anyone in the community as well.

While I agree that the end state would be nice to have elaborate models, I think there is a happy median here in the meantime. It is not that difficult to baseline trading volume and observe the data in the short term and develop further elaborate models at a future date.

My personal opinion is that appropriate measures have been taken and compromises have been met such as 1) waiting until Supercharged Liquidity Pools were established reducing spread factor from 0.3% fees to 0.2% fees, and 2) lowering the amount from what governance signaled approval for from 0.15% to 0.1%. As Robo pointed out, there is not Universal Support for 0.05%, 0.1% or 0.15% but there is room to meet in the middle if there is willingness to compromise. I think it should be up to governance to decide whether or not they would like to turn on Taker Fees or if they want to wait for elaborate models and wait for Future fee reduction programs (in addition to the fee reductions that already took place for this).

By voting yes for this proposal, the community is signaling its support to turn on Taker Fees and provide data to the subdao. The SubDao should develop its monitoring activities in this time (over 1 week).

1 Like

I think this should depend on Permissionless creation going live around the same time.
Since Permissionless will be proposed before this, any popular single hop routes can be made and cater for most swaps with very low liquidity, preventing this issue.

1 Like

Hugely in favor of this prop to set the default taker fee at 0.1%. Some quick observations:

-This needs to happen to support OSMO price. A default 0.1% taker fee would generate roughly $2.2M in revenue per year, making Osmosis hands-down the highest revenue chain in Cosmos.

-If OSMO price goes up, DEX liquidity goes up and user experience improves.

-Regarding competition - there is none. Given the efficiency gains from concentrated liquidity, Osmosis is by far the best DEX in the world for trading:

  • ATOM
  • OSMO
  • AKT
  • STRD
  • MARS
  • EVMOS
  • SCRT
  • BLD
  • And soon: DYDX, TIA, and whatever the Namada token is

-Finally, note that several major Osmosis pools have a super low spread factor already, such as USDC/USDT and OSMO/USDT. So when adding the 0.1% taker fee to these pools, the total fee is still lower than the original 0.2%.

1 Like

“-This needs to happen to support OSMO price. A default 0.1% taker fee would generate roughly $2.2M in revenue per year, making Osmosis hands-down the highest revenue chain in Cosmos.”

Yeah, assuming volume is unchanged. Why not just make the fee 1%, we’ll make 22M. Or 10? that’s a whopping 220M. That’s unheard of! Also, nobody wants a stablecoin swap pool that has 0.11% fees. I can do that for free on binance. The OSMO/USDT pool at 0.05% is a step in the right direction but its liquidity is poor compared to the other pools on the platform. Actually, what I think is more likely here is that people will slowly migrated to Crypto/Stable pools (but it might take months). During then, all the OSMO currently in LPs will be dumped on the open market (though some people might keep it to try to get a piece of the shrinking taker fee * volume pie). + the osmo from emissions will mostly be dumped on the market too. I fail to see how this is beneficial for osmo price, if I’m honest. Maybe I’m just salty after holding osmo from $2 to now. Even if the taker fee is turned on at 0.0001% today and we lose the multi-hop, the volume will halve overnight.

If we had many GOOD LIQUIDITY (500k+) USDC/CRYPTO 0.05% fee pools, I would not be against a 0.1% taker fee. But I want to hammer home my main issue: current 2 pool swaps will go from 0.2% to 0.6% fees. It is the wrong time to turn this on! If we wait until there are some established stable/crypto 0.05% spread pools and add the fees, I think 0.1% is reasonable. But let’s not jump the gun and make the dex unusable for weeks or months just because.

1 Like

People still won’t migrate their liquidity instantly. I mean, USDT/ATOM currently exists but there’s no migration buttons and it doesn’t get routed through on the frontend so most people don’t see it.

@whitemarlin I don’t think you addressed any of the concerns in my previous post.

1 Like

Historical pool volumes are queriable I believe, so no method to evaluate impact need be in place prior to passing this proposal. When that method is put into place (hopefully soon after the subdao proposal passes), the subdao will now have more historical data off of which to base its analysis.

I think the assumption here is that all volume will immediately flee the day we turn these fees on and that’s simply not going to be the case. There is a risk of losing volume over longer periods of time, but those periods are certainly long enough to let the subdao get its monitoring rails in place and do a comparative analysis.

1 Like

I think this is misconstruing my argument – obviously not all volume is going to leave immediately. Let me present you a scenario:

If the ratio of the tokens in a pool prior to any spread factor or taker fees is 1 token >osmo > 1 USDT, with the current structure, we could either sell 1 token for 0.998 USDT or buy 1 token for 1.002 USDT (0.2% fees). When fees are increased to the proposed 0.6%, that all of sudden becomes sell for 0.994 and buy for 1.006. So even though the ratio of the tokens has not changed, there is a larger range where the price doesn’t make sense to use the dex at all (let’s assume the pool values are pegged to a CEX price of $1, else bots or people would have taken advantage). Thus, it doesn’t make sense to trade in the pool (let’s assume traders and bots alike are logical actors) for a range of 1.2% of the price movement (0.994 to 1.006) versus a smaller range of 0.4% prior. The market the past few months most days has been pretty dead and sometimes there’s not even a 1.2% fluctuation on bad days. So yes, volume will go down - obviously it’s not going to zero. Most of the volume is driven by bots on larger pools taking advantage of this fluctuation, and manual trades on smaller pools. If you go to the main USDC pool Osmosis Info , you can see how much of the main volume on osmosis is driven by lower fees. Offering traders a 3x larger range of unfavorable price ranges does little to make the dex attractive to use.

2 Likes

Hey sorry I was responding to @LeonoorsCryptoman 's post, not yours.

My response to you is far simpler. Imo the taker fee (vs the maker fee I feel this probably should have been) is an acknowledgement and acceptance of the fact that a volume reduction will happen. Governance accepted this when it voted in the taker fee in the first place. We (as a chain) are fine with that. The time to have the argument that we shouldnt implement fees has passed.

From here, what matters is monitoring what the extent of that volume reduction is, and ensuring that it isnt too large relative to the benefit to the protocol that the fee brings.

I tend to agree with you that 0.1% extra is likely too large for stable pairs long-term. This is why we have a subdao that will evaluate and make reductions on a pool-by-pool basis. I agree less that the fee is too high for other pairs though, however I can accept there’s a chance i’m wrong about that. We just won’t know with any verifiable certainty until we actually turn the fees on.

As a general estimation, what is an acceptable reduction in volume? Not just to you, but all other proponents of the fee as well. It seems difficult to come to a consensus on a fee unless we know what each others number is for that.

A few things:

  1. The title of the proposal does not match its content.

No where in the text of the proposal does it say to turn on the taker fee. The text, in my opinion, only seems to be asking governance to reset the default fee level to 0.1%. Only in the title does it actually call for the taker fee to be turned on at 0.1%. As such I am confused about what exactly is being voted on here. It would be very helpful if the language could get cleaned up a bit.

  1. If the proposal is asking for the taker fee to be turned on, when is it asking for it to be turned on? You mention the subDAO would have 9 days to formulate models and monitoring. Does that mean the fee will be go live on Sunday, 10/8 (9 days from 9/29, when voting on the prop would be closed is 10/8) At what time? Shouldn’t it just go live on Monday, 10/9?

I would suggest that language be included in the proposal on what day the taker fee will go live. No one should have to be trying to guess what day this will go live.

I would also suggest that langue be added to this proposal to direct the subDAO to work with the necessary parties to ensure users are provided with at least seven (7) days advance notice of when the fee will go into effect. My expectation would be that this at a minimum would include an official blog post on the Osmosis blog that provides users with the date it will go into effect along with an explanation of how it works, screen shot examples, where the collected fees are going, and background info and contact information for the subDAO. A notification banner on the front end with a link to the blog post that replaces the current notification banner that reads “Osmosis Frontier has merged with app.osmosis.zone. You can now manage all your tokens in one place!” is also something I would like to see.

I also have the expectation that the subDAO will reach out to "content creators”, their fellow validators, and ecosystem chains/projects with tokens that are primarily found only on Osmosis (e.g. Quasar, Stride, Regen, OmniFlix, Stargaze, Gitopia, Nolus, etc.) to help get the word out on what day the taker fee will going into effect. With there being 7 members to the subDAO, this should be easily accomplished in 9 short day before the fee goes live and hopefully can lean on subDAO member @L1am-Stakecito great communication skills.

Overall, communication and transparency are of the utmost importance here and I hope become the watchwords of the subDAO. Speaking of which…

  1. I am very curious to see how the fee will be disclosed to users.

@JohnnyWyles perhaps you can shed some light on this and share some of the mock ups of what has been developed/designed so that we can all see how the fee will be disclosed to users when conducting a swap on the front end. Also, do you know if aligned alternative front ends (e.g. TFM, Calc.Fi, etc) are ready or will be ready to provide such disclosures about the cost of the fee on their front ends to their users?

The fee rate and the actual cost should be fully disclosed and made abundantly clear to the user when conducting a swap. This is more important to me than what the actual fee rate is. I strongly believe that the fee should not be “hidden” in a drop down option that users have to click to discover that it is there. The current interface which requires the user to click a button to be able to see what the swap fee is has long annoyed me as an unethical practice that needs to be changed. It may not violate the letter of any consumer protection laws, but it surely does violate the spirit of such laws.

I have commented about this a couple times in the past, and the implementation of the taker fee provides opportunity for Osmosis to remedy this very easily correctable flaw.

  1. In @JohnnyWyles X poll, once you exclude the respondents that chose “see results”, 47% that believe that the taker fee should be set at 0.05%, meaning that 53% felt that it should be 0.1% or higher.

If we all pretend that respondents were a statistically representative sample, because of the small sample size, you would have a likely margin of error of at least +/ -5%, which puts it at a statistical 50/50 split between those who believe that the fee rate should be 0.05% and 0.1% or higher.

If you add the respondents to @JohnnyWyles poll on this forum that he had posted in a different discussion thread, and assuming those respondents didn’t partake in his X poll (I know I only participated in his poll on here and not on X), you get the same 47/53 split. (The 9 person sample size of the poll conducted on here would easily have a margin of error of +/-10%, which would again make it a statistically 50/50 split. Even if you averaged the two polls you would get a 46/54 split, which is again is a statistically 50/50 split because of the +/-5% margin of error.

Just wanted to point that out. It does however also beg the question why the compromise position should be 0.1%, when the plurality of that participated in the polls were in favor of initially implementing a 0.05%, and statistically speaking, the community is evenly divided between a 0.05% fee and 0.1% or higher initial fee, particularly if turning it on across the board for all pools is for the subDAO to collect data to make adjustments.

Implementing a 0.1% fee across the board wastes a completely perfect opportunity for the subDAO to conduct a quasi-experimental difference in difference design study to estimate the CAUSAL effect the fee rates has on various outcomes of interest (e.g. volume, liquidity, users, prices, etc.) by turning the fee for different pools on at different times and at different rates. Evidence is always better than data. :smiley: after all. It also shouldn’t be that difficult or time consuming for the subDAO to do an even more rigorous regression discontinuity design study or even an instrumental variable design study. I think liquidity incentives could be used as an instrument and am actually curious whether it would have a strong enough F statistic to be valid.

The subDAO also doesn’t even necessarily need to have the fee turned on to be able to have some data to start looking at gain some insight into how it may effect certain outcomes of interest. Just recently the multi-hop swap discount was turned off for at least a few days, the removal of liquidity incentives from low volume pools, the time when USDC.grav and USDT.grav were receiving liquidity incentives and enjoyed the benefits of the multi-hop discount, and the difference 0.15% swap fee for the USDT.grav pool compared to all other stablecoin/OSMO pools are all natural experiments that the subDAO could look at to gain an understanding of the effect the fee could have.

There are also very some relevant studies that I strongly believe allows us to guesstimate what the effect turning on the fee could have. In fact, the findings from a study published in May from researchers from Haas at Berkley, Sloan at MIT, and Wharton at Penn can be extrapolated to suggest that we should in the worst case scenario be prepared to see an 18% decline in the number of retail users on Osmosis and a 30% drop in retailer user volume. (See: Fee the People: Retail Investor Behavior and Trading Commission Fees)

  1. Monday, September 25th is Yom Kippur.

I would be grateful if you could delay posting any proposal on chain till Tuesday, September 26th.

2 Likes