[Signal Request] Lower the Flash Mint Fee to 0%

The Flash Mint Module was activated on July 3rd and still is not seeing any usage. As a personal user of flash loans this comes down to who has the cheapest solution to use as well as cost / benefit of having to change the code to a new source. Currently dYdX has the cheapest flash loans at 0%. Previously I thought there may be an argument for paying our fees because we can guarantee liquidity is always available, but since the original poll dYdX has improved its liquidity from 138k to ~10M+. This is large enough now to cover most use cases.

I suggest we consider removing the fee and do so permanently. As there seems to be no integrations so far we can benefit from removing the logic for the fee altogether if we never plan to use it again and re-deploy. Flash loans are highly gas-sensitive products and so we want to squeeze out as much as possible.

Pros

  • Better arbitrage for DAI
  • Better UX for front-ends, can lever up in one go instead of looping, re-financing, etc

Cons

  • Potential loss of revenue
Should we remove the Flash Mint Module fee?
  • Yes, remove the fee permanently (requires redeploy, new contract address, saves gas)
  • Yes, set the fee to 0% (can be raised again later)
  • No, leave the fee as-is
  • Abstain

0 voters

Poll will run until Wednesday, August 25th. If successful an on-chain vote will be held to confirm the change followed by an executive with the change.

11 Likes

I must admit I do not know where to find or how to use the flash mint module.

2 Likes

I’m in favor of the cheap gas and free borrow contract. Flash loans are commodified at this point and since DYDX is free and liquid, I think the train has left the station for flash loan fees in DeFi.

If I were building an arb bot right now I’d choose DyDx, which means I’d be reading their docs and learning their ecosystem. It’s technically not free to borrow from there, it’s 2 wei, so we’d be undercutting them by what seems like a tiny amount, but which would save a developer a couple of lines of additional math (and gas) to process the fee amount. The net result is that developers would be reading our docs and learning our ecosystem.

Not to mention that providing liquid Dai on demand can help arb price discrepancies across markets to smooth out market price and generally increase Dai usage in the ecosystem.

If we’re wrong and we do need to charge fees later, we can always de-auth a new mint module and re-auth the old one, so permanent doesn’t necessarily mean permanent.

16 Likes

It’s in the docs: Flash Mint Module - Maker Docs

It’s more of a developer service though, so non-technical folks shouldn’t really be directly using it anyways. It will just make their life easier on the front ends.

5 Likes

A few questions:

How much are these flash mint products are being used?
To do what?
How much profit is being skimmed from where doing this?

Without understanding the above I think it is pretty hard to talk about fees when I don’t even know what the market looks like for this product. I mean why make a product that will earn us no fees that someone else has? It is like competing with someone else on a product that earns no revenue. If this is a mindshare, PR thing perhaps there are better ways to get developers looking at Maker protocol.

Entirely unclear to me what the real goals are with this.

Voting No for the above reasons. Why spend more time on something when the real reward is unclear at best.?

Those two aren’t super compatible. Abstain is a lack of decision, no is a decision. You can vote for both, but we end up discounting the abstain vote in cases where people combine an abstain vote with a non-abstain vote.

This is the kind of insight I like to read because it wasn’t intuitive to me

2 Likes

good point. I was thinking no, and abstain from the point of view of - goals need to be clarified. Edited to voting no as I don’t see any tangible goal case being made, much less a measurable or data on market use of these facilities by dydx or any other provider.

1 Like

The tangible goal is improved DAI arbitrage and improving user experience on front-ends. Flash loans can be used to take out leverage in one go as opposed to have to lock, mint, trade, lock, mint, trade, over and over. Having 500M in liquidity means even big whales can do this and not have to worry about whether dYdX has liquidity or not. Removing the fee gets rid of user friction. The choice is they are presented with a big lever up button and either only have to pay the gas or some seemingly random fee attached.

1 Like

Where is the data even showing DAI ‘arbitrage’ that will then lead us to measuring some ‘improvement’. My point is that I don’t see any data making a business case here, much less one that should be done at 0 cost.

Show me some data regarding flash loan use that gives us a handle on why it is important to spend time on it and I will change my view. I can’t count the times someone has told me, “we need this”, and when it was all done we found that no-one used it, because there was no use case. Just because something sounds great doesn’t mean you have to do it. Show me some or any data regarding what flash loan arbitrage is doing, how much they are making, and where they are sourcing what and I am happy to change my view.

I mean if the argument is - its already done - and just because we can - and people are using it - and it doesn’t reduce system security. I mean if this is a way to get business going so later we can and would raise fees to be competitive in a space just starting off that is cheap to do, maybe I can buy that. But we are still going to want to have some data here so we can make some kind of informed decisions on this later.

The dydx pool is pretty heavily used. You can view the internal transactions against their solo margin pool at https://etherscan.io/tokentxns?a=0x1e0447b19bb6ecfdae1e4ae1694b0c3659614e4e
It’s a zero-sum game though, they’re getting the traffic because they’re providing the cheapest access.

I view this as a developer tool where we have the advantage of being able to originate Dai at no cost. From a business perspective, you don’t want to use developer tools as an income source, you want to give them away to encourage downstream integration of your core product.

13 Likes

Thanks for the pointer @brianmcmichael I think this is my key point. clarity on goals here. If this is a cheap, easy and no implications for security (we have this done) way to basically pull developers that is one thing easy to measure. If it is to enable more arbitrage that is another (that to me requires some data to back up).

I just want to have some clarity regarding the goal, and then the overall cost. I also like things to have this - and later if markets warrant we can easily increase fees. Which brings me to a question. If we do this change can we have a on chain fee variable so we can later (if usage warrants this) manage the fee like we do other system parameters. I think it is the latter point that would make me a bit more excited to do this as it would then have the potential to generate system revenue…

I really would like to see some analysis of what the flash loan arbitragers are doing and how much they are making… But if this is a get foot in game with devs, this alone can get me to change to a yes as long as there is a provision for easy fee control via governance in this for later.

1 Like

The modifications would be trivial in this case. One might argue that by trimming out some of the moving parts in favor of constants we’d be making this slightly safer.

1 Like

Sounds reasonable. My thinking pretty much above. Based on discussion just going to go from no to abstain due to lack of clarity on goals and market analysis/needs on this.

1 Like

If users migrate to our Flash Mint module instead of staying with dydx, will this free up any DAI locked into the liquidity pool of dydx?

If yes, then it seems like a nice side benefit would be to increase DAI supply in the open market?

1 Like

It might, but not necessarily since the DAI used for flash loans is typically locked for other reasons (i.e. used as collateral similar to AAVE’s source of flash loan funds).

I support reducing the fee to 0% in order to remain competitive in the flash loan market. Flash loans are unlikely to generate meaningful revenue, and having a fee ostracizes users since other offers are almost strictly better. At this point in time, the goal should be onboarding users and expanding the use of DAI, which is best accomplished by reducing friction in the ecosystem.

8 Likes

I totally agree with you, let it be of massive use, let it be a better service and then progressively place commissions necessary for its maintenance and operation.

1 Like

The 0 fee can be something attractive in the short term, we have to promote this, make more people aware of this functionality and work on the integration with other platforms where we can get benefits.

1 Like

But lowering the fee to 0 is just making it the same as dYdX. Given that dYdX supports multiple tokens rather than just DAI, and the liquidity there is getting deep, it sounds that the best strategy for a developer is to integrate with dYdX still.

2 Likes

The issue with the liquidity on dYdX is that it tends to disappear when people most need it. This is true of all the money markets. When there is an extreme market move, there is a lot less liquidity on these platforms, and often that is when people most want to make use of flash loans to quickly adjust their positions.

4 Likes