GMX Fee reduction proposal

[breallyant]

Summary

This proposal introduces a discussion about implementing fee discounts for GMX traders. The objective is to address concerns raised by the community regarding trading costs, with the goal of making GMX more competitive and attractive to both existing and new users. By lowering fees, the proposal seeks to enhance the overall user experience and increase trading volume, which would benefit the protocol in the long run.

Rationale

One of the most frequent complaints from GMX traders revolves around costs, with feedback being split 50/50 between concerns about fees and price impact. By introducing fee discounts, GMX can improve trader satisfaction.

Reducing fees could also help GMX stay competitive in the increasingly crowded perp DEX space, where traders are looking for platforms with lower costs and higher efficiency. With rising competition from centralized and decentralized exchanges, reducing fees could increase trader retention and acquisition. Moreover, addressing this pain point could foster long-term loyalty, leading to higher trading volumes and protocol growth.

Specification

There are multiple approaches GMX could take to reduce fees and increase trading volume. The options being considered include:

1. Volume Discounts – A tiered system where fees decrease based on traders’ monthly or biweekly trading volume, offering incentives for higher volume traders.

2. Flat Rate Reduction for Open/Close Fees – A uniform fee reduction for all traders, ensuring everyone benefits from lower costs regardless of their trade size.

3. Volume-Based Discounts for GMX Holders – Offering lower fees to traders based on the amount of GMX tokens they hold, encouraging more engagement with the GMX ecosystem.

4. GMX Holding Rebates – Implementing trading rebates paid in GMX tokens, where a portion of weekly trading fees is returned to GMX holders, up to a percentage of their trading costs.

Open/Close fees competition comparison

In comparison to its close competitors taker fees: Hyperliquid, Vertex, DyDx, Gains. GMX fees look as follows:

x-axis: Trader size based on 14-day trading volume
y-axis: Taker fee, avg for gmx

GMX’s fees are meaningfully higher than some of these competitors, depending on the trader’s size

Volume Discounts

Volume discount is supposed to increase the overall trading volume at GMX by:

1. Attracting traders from other platforms or Re-engaging churned traders – Traders are more likely to favor GMX due to reduced fees and the opportunity to secure or maintain better rates.
2. Incentivizing traders to increase their trading activity in order to reach higher discount tiers or to maintain their current discount level, creating a reason to trade more frequently and in larger amounts.

Pros

• This scheme is familiar in the market, it helps build “stickiness”, encouraging users to stay on the platform. Theoretically, increased trading volumes could potentially cover the cost of offering additional discounts.
• GMX’s strengths for large traders, who are aware of their costs and consider pricing as a factor in their decision-making. Many large traders operate across multiple platforms, meaning there’s potential to attract their volume through discounts.

Cons

• Users might perceive it as unfair, though our existing referral program already works the same way.
• It positions us directly against CEXs and Hyperliquid and others, shifting the perception of GMX from an ecosystem to primarily a trading platform.
• The benefits of volume discounts overlap with the referral system, which will require us to implement “caps” on the total incentives a trader could receive when participating in both programs.

Flat Rate Discount

Decrease flat rate fees across the board for all traders. It is simple, easy to promote, and ensures that the platform treats all traders, regardless of size, equally, which contributes to a sense of fairness.

Volume Discounts for GMX Holding

Using GMX holdings as an anchor to provide volume discounts.

Could also be combined with other incentive schemes. For example, with volume discounts, we could implement an “or” structure, where traders must meet one of the thresholds to get the discount, for example:

14-day Volume		GMX Holdings	Fee
0 - 3M	or	some value 1	0.06%
3M - 25M	or	some value 2	0.052%
25M-100M	or	some value 3	0.04%
100M+	or	some value 3	0.03%

GMX Holding Rebates

Implement trading rebates paid in GMX tokens for holding GMX tokens. A portion of the weekly trading fees can be paid out to traders holding GMX up to a % of their trading fees.

• Effect is the same as a discount, but the cost comes out of GMX distributions instead of LP share
• Probably requires allocating a greater % of revenue to GMX vs LPs which, unless we can show other changes drive more value, could be viewed negatively.
• This is now possible also because of BB&D, which creates available GMX tokens

Calculation Examples

Volume Discount approximation

Since the main hypothesis is that volume discounts could attract traders from competitors or bring back churned users, it’s important to understand how trading volume is distributed among competitors based on trader size. The volume distribution for the first 14 days of September 2024 is as follows:

x-axis: Trader size based on 14-day trading volume
y-axis: Cumulative trading volume from traders of various sizes, with the number displayed at the top of each bar indicating the count of traders in that size category

According to this data, “wealthiest” segments are those with trading levels above $3M.
Let’s take a look at an example of how the proposed volume tiers would impact the fees collected. For example, for the following volume tiers:

14-day Volume	Fee
0 - 3M	0.06%
3M - 25M	0.052%
25M-100M	0.04%
100M+	0.03%

In comparison to the competition, the fees would look as follows (represented by the blue line):

Assuming that the open close fee historically accounts for 30-60% of total collected fees, total fees will be lowered by approximately 9%

Fee Structure	Total Fees	Open/Close fee collection change	Total fees Change
Current: 6bps	833283	0.00%
Volume tiers	684473	- 17.86%	- 8,7%

Flat Rate reduction approximation

If we review flat drop of the open/close fees to 5bps, 4bps and 3bps respectively

Based on the current volumes generated at GMX, the 14-week open/close fees would approximately change as follows, with data calculated for the period from 01.09 to 14.09:

Fee Structure	Total Fees	Open/Close fee collection change	Total fees Change
Flat – 6bps	833283	0%	0%
Flat – 5bps	694403	- 16.67%	- 8.3%
Flat – 4bps	555522	- 33.33%	- 16.5%
Flat – 3bps	416642	- 50.00%	- 25%

Conclusion

This proposal seeks community feedback on the outlined fee reduction strategies. Once input is gathered, a more formal proposal will be drafted and implemented, taking into account the thoughts and preferences of the GMX community.

[Jonezee]

Thanks for your elaborate research into this, Breallyant. Let’s consider the various options and their implications, then communally decide on the best strategy.

[Seraph]

Thank you breallyant
I look forward to lowering of fees (volume tier) to attract the large whales over to trade.

not your wallet, not your money, FTX taught us some expensive lessons.
just hope that the lower fees make it easier for whales to switch over.

[Jonezee]

Personal thoughts that come to mind:

• Volume Discounts: a proven system to reward loyalty. Pretty effective. Adds a bit of complexity and requires a bit of management.

• Flat Rate Reduction for Open/Close Fees: transparent and simple, but can have a large impact on protocol revenue that the increased competitiveness may not balance out.

• Volume-Based Discounts for GMX Holders: stimulates loyalty and benefits the community. But with GMX tokens being readily available on lending markets and Perp DEXs nowadays, users could just borrow the required GMX or short equivalent GMX to offset any holding requirement.

• GMX Holding Rebates: Not really a fan of it, as trading fee rebates are after-the-fact rewards rather than upfront lower costs. This makes it a psychologically less effective strategy imho.

[Alik1975]

I do not think we need to focus on whales. Whales come and go. Fees need to be adjusted on the number, not the amount of trades. This will eventually attrack volume in number thus increase the fees for the platform, in my opinion.

[Tano]

Personally, I am more of a fan of the volume tiers than straight up reduction of fees.

I think our fees will struggle a bit too hard if we generalize the reduction all across the board in the first few weeks, so I’d rather go with volume tiers unless other delegates come with their opinions.

No fan of the GMX holder discounts or rebates.

[gmsolq]

I will support Volume Tiers or GMX Holding, but more tiers can be added to more precisely correspond to users of different levels.

If the 14-day trading volume is used as the threshold, the specific parameter for GMX Holding can be set at a ~1% ratio (referencing Binance):

[Vaultka]

Thanks for proposing new suggestions to improve GMX’s competitiveness, Breallyant.

The Vaultka team supports offering discounts based on volume tiers instead of fiat tiers. This volume-based approach can attract more whales to use GMX by incentivizing them to trade more, which would bring larger volumes to GMX and increase the protocol’s revenue. Moreover, the volume tier-based system is more flexible regarding discount adjustments compared to a flat rate.

[SolvProtocol]

Thanks for the suggestions, Breallyant!

Solv Protocol supports offering discounts based on volume tiers, similar to how centralized exchanges structure their fees. This approach can enhance competitiveness by attracting more traders and increasing overall trading volume, while also fostering loyalty and rewarding users for sticking with the same platform. We believe that such incentives can significantly enhance the trading experience on GMX.

It’s crucial that GMX communicates a more comprehensive go-to-market strategy, rather than solely marketing itself as a Tier 1 perpetual DEX. Users should have clear ways to display or showcase their status (e.g. VIP 1, VIP 2 as an example) on GMX, adding a layer of engagement.

We encourage the community to monitor the impact of these changes over time for necessary adjustments. However, we’re not too fond of the other suggestions, such as flat rate discounts and GMX holding rebates.

[coinflipcanda]

Reflecting on GMX’s Evolution: Preserving Fairness in Our Fee Model

As we reflect on GMX’s journey, from the early days of Gambit to the advanced platform we’ve built today, it’s important to revisit our core mission: creating deep, efficient markets on-chain. Our goal was never to replicate centralized exchanges (CEXs), but to innovate and build new, decentralized markets.

However, with the possible introduction of VIP tiers, I’m concerned we might be moving in the wrong direction. GMX has always aimed to create an equitable model for all traders and liquidity providers (LPs), but this shift risks undermining that principle.

V2 was designed from the ground up to better manage risk, allowing for greater capital efficiency and more aggressive pricing. With improvements like enhanced data streams, the GLV for automated liquidity allocation, and the Chaos Risk oracles for real-time market management, GMX is now more efficient than ever. The on-chain market has also become increasingly competitive, which means being prepared to share those efficiency gains with market pariticipants.

Yes, rewarding big traders with discounts is a well-established practice for retaining users, particularly on order book DEXs where large traders and MMs (sometimes the same) are incentivized to help fill a book which allows the exchange to profit off small traders. But if we follow this path, I worry that GMX could start to resemble a less attractive version of a CEX, moving away from the decentralized infrastructure and liquidity we’ve been building, especially when our approach of abstracting away the order book removes some of the compulsions that CEXs have.

Fees on GMX are designed to fairly compensate LPs for the risks they take by providing passive liquidity across our markets. These fees also ensure traders benefit from transparent, non-manipulable execution. By offering the biggest traders discounts, we risk enabling toxic flows, where LPs are paid less for trades that may increase risk. Over time, this could lead to spreads or other mechanisms that socialize costs, which defeats the purpose of creating a risk-based model where each dollar of volume contributes equitably.

Instead, why not focus on lowering GMX’s fees across the board to match or exceed the competitiveness of other leading venues, like Binance? WE can be positioned as competitive with CEXs, on fees and price execution, while also keeping GMX positioned alongside top DeFi protocols—Uniswap, AAVE, Curve, Compound—we can create efficient markets that cater to both large and small traders alike. If you are a whale, you don’t get a discount on Uniswap or AAVE what you get is markets that have grown to support the largest users just like GMX has as well. I’m not saying this path is easier, but we should be aiming to create markets that move away from privileged access and privileged pricing of tradefi (and cexs) and create something better.

This doesn’t mean we shouldn’t reward GMX stakers or loyal traders. Mechanisms at the DAO level, such as directing a portion of protocol fees or funding rebates, could achieve this without creating VIP tiers within the core protocol itself, a change which practically speaking would be hard to reverse. Once you build into the protocol mechanisms to favour some users over others, there will be a constant battle for people to gain benefits and negotiate and push for more.

Some may argue that referral fees already create a tiered system, and while they’ve been successful, we should be cautious about adopting every practice simply because CEXs do it, and maybe in V3 or before we should rethink how and what we want our refereral program to do.

With the capital efficiency of V2 markets, our current pools are already well-positioned to sustain much higher volumes. Let’s stay focused on making GMX a decentralized leader perps with our liquidity and efficiency giving us an edge, not a tiered system that risks compromising the vision.

[FallingSands]

Hell yeah brother.

GMX should be a resource that any entity can instantly tap into, with the same priority and guarantees enjoyed by all. This is the permissionless, equitable ethos of crypto. This is the entire value proposition of an open, decentralized protocol.

End-game success is for the protocol to provide a ubiquitous, composable marketplace that connects traders with LPs, big and small, anywhere, any time. A foundational, decentralized and credibly neutral resource.

Universal fees make sense long term. If we provide volume tiers, someone will create a wrapper that pools traders together to reach the highest tiers. This is not an issue CEX need to worry about.

If we look at the most successful protocol, the internet, it was the ethos of net neutrality that created a flourishing resource whos unfettered access is today regarded as a critical human right. A lack on neutrality would have lead to a failed experiment.

Ok, ok, enough pontificating.

Failing to thrive in a competitive market is a serious risk, potentially a lethal one. Unfortunately, its not always the best ethos that wins in the open market. You can argue the open internet was uncompromising only because it lacked the aggressive competition that we face from both incumbents and contemporaries. Its true we need a real life solution.

As an LP and a GMX staker, I appreciate the need to tread the line between offering the lowest fees possible and still providing competitive yield to LPs.

We already offer a great product. A strategy that may increase competitiveness is to introduce a curve that marginally decreases fees as the trading volume rises across a market. This creates a positive feedback loop, where increased trading volume decreases trading fees, incentivizing further volume. If volumes grow enough, our fees could potentially become the most competitive in the market, without unduly hurting our LPs when volume is low, as a low fixed fee would. This would also provide a gradual path to compensating the loss of yield with increased volumes, without creating a short term shock that Tano and others discussed.

The curve would need a floor so that during times of low volume, fees don’t become completely uncompetitive, as well as a smoothing function to keep fee volatility reasonably stable.

Unfortunately, I’m not a gigabrain statistician like breallyant, but I do feel like this strategy could be a middle ground that offers a fair and equitable marketplace, protects LPs, while also providing increased competitiveness.

[Jonezee]

A curve to determine fees could be an interesting approach, though it doesn’t reward loyalty on the individual level or offer a strong marketing angle. Food for thought; I’m gonna think about it some more.

[coinflipcanda]

@FallingSands

If volumes grow enough, our fees could potentially become the most competitive in the market, without unduly hurting our LPs when volume is low, as a low fixed fee would. This would also provide a gradual path to compensating the loss of yield with increased volumes, without creating a short term shock that Tano and others discussed.

The curve would need a floor so that during times of low volume, fees don’t become completely uncompetitive, as well as a smoothing function to keep fee volatility reasonably stable.

while there is probably some complexity from multiple pools that trade the same assets, i think there is a great idea in there. Essentially everyone can trade at a certain assured maximum fee but as volumes continue to go up, the benefit of the gained efficiency is shared back between traders and LPs, making the process virtuous.

If say in a pool the rolling 7 day verage daily trading volume in a pool is above various thresholds of volume/ TVL

0.0x - 0.5x of TVL, 6bps
0.5x - 1.0x of TVL, 5bps
1.0x - 2.0x of TVL, 4bps
2.0x+ of TVL, 3bps

Do you get an individual loyalty bonus? no but what you get is a situation where fees continue to come down as the overall infastructure is better used. Comparable for exampel to how as ZK chains usage goes up, the cost to post the proof to L1 remains similar so the higher volume results in everyone getting a better deal.

In theory a model like this could actually work by charging a 6bps fee universally, paying LPs on a running basis 3bps, and then do weekly rebate claims for traders and LPs using those proceeds in GMX tokens. Basically would become a further extension of BB&D.

To be clear i’m not endorsing the idea, but definitely love the idea of efficiency pricing with the benefits going to all.