# Why Ethereum's Issuance Policy is [redacted]? | Devcon SEA

- Channel: [Devcon](https://streameth.org/devcon)
- Date: 2025-10-07
- Duration: 28:14
- Watch: https://streameth.org/watch/yt-cUgKXBq017g
- YouTube: https://www.youtube.com/watch?v=cUgKXBq017g

## Description

This talk explores the status quo of staking economics, its drawbacks as we see them and what the future of staking economics could look like.

Speaker(s): Caspar Schwarz-Schilling, Ansgar Dietrichs
Skill level: Intermediate
Track: Core Protocol
Keywords: ACD, Staking, Economics

Follow us: https://twitter.com/efdevcon, https://twitter.com/ethereum, https://warpcast.com/devcon
Learn more about devcon: https://www.devcon.org/
Learn more about ethereum: https://ethereum.org/ 

Visit the https://archive.devcon.org/ to gain access to the entire library of Devcon talks with the ease of filtering, playlists, personalized suggestions, decentralized access on Swarm, IPFS and more.

Devcon is the Ethereum conference for developers, researchers, thinkers, and makers. 
Devcon SEA was held in Bangkok, Thailand on Nov 12 - Nov 15, 2024.
Devcon is organized and presented by the Ethereum Foundation. To find out more, please visit https://ethereum.foundation/

## Transcript

[Music] thank you thanks for the info and thank you for coming so early um I know it's the first session um so yeah Ana and I going to talk about why ethereum's issuance policy is unsustainable in our opinion um and for some of you who might have been at Justin's talk um presenting his idea of the beam chain um one of the items that is part of that road map is smarter issuance as he refers to it and he put a little um spicy Emoji con next to it and it didn't take long to understand why um this topic tends to be spicy um it was the most upvoted question in the Q&A why are eth researchers obsessed with lowering issuance and we're here to kind of answer this question and argue for why we think we need to change issuance um so let's begin by just looking at the current staking Trend um it's been a relatively straight Affair up only um so here you see basically the amount of stake in the system and it's been yeah trending upwards um at a fairly um constant rate and um the question is obviously where does this trend go will it continue when will it stop what does it look like in the long run um and to kind of answer this question we're going to have to argue a bit about kind of staking supply and demand and to re to be able to reason about what an equilibrium looks like so in this setting the demand curve is what we call the issuance curve it's basically the protocol's willingness to pay um validators to secure the chain and that function is very well defined we know exactly at what point how much valid datas um earn as yield if they perform correctly of course and then the supply is basically the willingness of eth holders to do that work to do the validation um for um different levels of field and so um what we see here is basically um the demand curve which is the issuance curve um and to kind of quickly run through this intuition here why it's designed the way it's designed um right now is that for what you see here is that there's um a downwards Trend as more and more e is staked the protocol pays valid datas less and less yield um but at the same time one property here that you can see is for very low levels of staking the issuance yield is very high basically the the intuition here is the protocol really needs to make sure that um we have a minimum amount of security in the system and so it pays a lot for that and as more and more e um gets staked kind of the urgency of um um needing to kind of recruit more validators goes down um and it goes down but it kind of flattens out and so that leaves us in this awkward spot that we don't really know where we actually end up um in the long run um so what happened so far is basically um this upwards only Trend where more and more people stake and what that means in the supply and demand context is that the supply curve has been shifting downwards over time um and now the question is where does this end up in the long run and one reason of behind the fact that kind of why we might end up in a sort of High um percentage staked world is this what we call the flattening effect and the idea here is basically that um for solo stakers to stake the kind of costs are very heterogeneous and mostly it's the fact that some people are very good with computers some people not so much and so to recruit the last person on Earth or like the last e holder on Earth to sort of go through the pain of setting up a server and like running their own validating node is extremely costly you would need to pay a lot whereas with sort of delegated stake um that cost is very homogeneous meaning it's just a oneclick um solution for everyone and so the kind of obviously the Willing the preference is still different so it is um an upward sloping supply curve but it's basically uh significantly flattened and again in the supply and demand context what this basically means is that with a flatter supply curve you end up with more um stake in the system in equilibrium given a demand curve now um basically yeah the question is where do we end up in the long run and that's not um because we don't know the supply curve we can only observe it at any sort of point in time we know that today x amount of stake is in the system roughly 34 million so roughly 30% um um of all eth and so basically at the very least given the whole kind of trend with lsts and liquid staking it is very plausible that we over time um kind of continue trending to the right in terms of staking participation um and because the demand curve so the issuance curve is so kind of just relatively flat towards the end and doesn't kind of conclusively go to zero or anything like that we don't know where and you can see that a very small change in yield can have a dramatic impact on the sort of um equilibrium um staking level um and now I'm going to let Ana continue yeah and the the second section um we call that nominal versus real yield um so we we've been giving like talks like this over the past um half a year basically at at a few conferences and this is always the one section where in the end when we talked to people people are always just a little bit confused by it and didn't really understand what we were trying to say in that section so we did it tried to um restructure it a little bit experimentally um uh and and we'll still show like proper yield curves but I I wanted to start with like a very simple mental model first and specifically I really think that there's two I call them anomalies for proof of stake accounting that mean that I think most people's mental models for like how much yield you earn in proof stake are just fundamentally slightly off and to explain that a little bit more again simple model um so basically just let's look first at a normal as any real world example where you have a large group of people paying a small group of people right you everyone here takes some coins gives them over to the other group um very simple right and if you if you run the analysis like you could imagine I don't know that's a retirement fund or something a lot of people pay in few people get get get uh get a pay out or something H and you make the analysis like who basically gained who lost or like who paid who who received it's a very simple analysis right everyone here paid a little bit those two Gain gain quite a bit simple simple accounting here um if you go to First proof of work because that's where the first anomaly shows up um it's a bit of a different model right so we have these people they now like see they say these are all eth holders pre-merge right they all have their own balance you have these miners these miners need to incentivized for their work need to be incentivized for their work so what we do is we give them coins but this time those coins didn't come from anywhere right we minted them and so that's the first anomaly that's basically that we in in uh in um uh blockchain networks we often pay for expenses by just minting new coins now that's no magic right it's not magically coming out of nowhere it's still a payment from those people to those people right but it's an indirect payment basically because you're now being paid by being diluted right that that is accounting Wise It's accounting wise the same thing as if you just given coins over directly but it basically instead takes this indirect step and the way you can think about it is right like the balances here didn't change but because there's now more coins in existence in general everything shrinks a little bit right so that's the way basically same thing again it's just a different different way of doing accounting but it does make it a little bit harder to track what's going on now proof work that's still fine proof of stake adds one other kind of weird addition to that accounting model of course wait that's yeah that's obvious right like those people still paid a little bit and then those people gain so now proof of stake adds a second uh anomaly to this model and that is that there are no minus anymore right and that's great you know I think uh it's great that we were able to do that um but now who who actually is doing the work who's being paid here right um and it turns out that actually in improv stake of course the people that actually do the work are some of the stakers some of the the holders themselves right some of the people with eth in this case I don't know half of them of course right now we are more like at a third or something some of them actually are stakers are are doing the the work and just to to give a little bit more room let's separate them out a little bit um and so now same thing again right like they do stake they they do they fulfill their duties and they're getting paid right so let's pay them okay so then again same thing same question what happened here like who where where did this money come from right same answer of course dilution so let's shrink everything and then we can run the analysis again and we can run the analysis and we can say well everyone here paid for the for the security and then those those guys actually got uh got extra money and and and of course if you if you do the the accounting like everyone lost a little bit because they paid for the security those two gained more so on balance you still make a profit if you stake of course right that that makes sense but for example in this example we have four new coins we have four people so basically everyone got diluted by one coin on average right so basically it's I I chose this example just because actually if you look at this here it's very simple if you ask like how much did you actually earn you did not like you see two new new coins but the first coin here offsets the dilution the second coin is still earning so the actual answer would be your staking yields in this picture is one coin now there is one annoying fact I don't know how many of you all have ever paid taxes for staking but um if you're a solo Staker you just pay taxes for the two coins right because that is how much your balance went up now in many countries it's a bit better there are countries where like you know the top tax rate on income if that's taxed income is around 50% in that world you'd literally have lost the one coin that was your real income just now you know so 50% stake it's already getting quite uncomfortable with kind of this this um this delution effect and I I just wrote it down here so the two the two accounting Anon um that they make it a little bit hard to to to to talk about proof stake kind of the the payment via new coins and then the recipients are also paying um and so what you see in terms of yield is not what you get and actually I think you know how a couple years ago when we first had um icos people only looked at like floating um Supply right and at some point we all in the industry were like wait that's not the right way of doing accounting actually we need total like fully diluted um uh volume of the of the coins right and and that took a while but by now that's the standard metric and I would say we really need the same thing with staking rewards as well if someone tells you hey our network is pay 7% staking yields but we also have 70% staked you should actually think ah so I'm getting 2% staking yields not 7% right that that actually makes a big big big difference um now so far that's not an argument for like any anything about ethereum in particular it's just like some annoying aspect about how to you know do accounting for proof of stake and actually if someone tries to maybe convince the tax authorities that actually they should only pay taxes on one of those two coins I think that would in principle be a very sound argument but not not a tax Tax Advisor um but like now let's go back to actual yield curves you know and like proper economics like this is again this is the same curve from just now like this show tells you depending on how much total stake there's in the system we have roughly 120 million eth in total this tells you how much you earn or how much you think you earn right like this is the nominal that's what we call it the nominal yield how much the protocol how much your balance goes up every year now we can just take that and uh one one one point there is that your incentive to to come in and stake at any level right is is always the difference between earning that yield and earning zero right that's that's pretty trivial so so now let's let's adjust that for the real yield and that shifts the curve down and it should be relatively intuitive why it shifts the curve down like towards the left if barely anyone Stakes right there's barely any dilution so towards the left those two curves are basically the same if if we have 10% staked like the amount your balance goes up that's roughly what you actually earn in staking TOS the right if everyone Stakes that's not it's not like the system breaks like if everyone Stakes it still works but everyone's balance goes up by the same amount every year and that doesn't actually create any value right so you don't actually earn anything anymore at that point basically staking is just a duty for the entire network um and one one point I just wanted to make there because sometimes people ask about that that does not mean that the security of the network is in any way broken because your incentive to stake is still always the difference between staking and not staking and as you can see I kind of plotted the real eth yield here like so basically if you just told eth you know you're you you're getting diluted every year so like basically the the gap between those two curves still stays the same as before so the incentive to come in is still the same like as you you you you you you would have thought from the nominal graph but it just means that it comes from a different thing first it comes from the potential to earn profit and then towards the right side it comes just from avoiding to be diluted and so that's kind of the summary SL summary slide here so basically we have a spectrum of different paradigms of staking you can have staking set up as something that actually generates profits potentially of course you know for for stakers or you can set it up as just a duty for the entire protocol both of those work for the protocol so if we just reason from what's best for ethereum both of those work um but we have to be aware and I personally think talking to a lot of stakers that actually once you kind of like try to make that point people tend to really prefer the world where actually staking is a profitable activity um and one side argument there that we always make is like if we want to go to the other world instead and we want to just everyone participate in staking we should design ethereum explicitly with that in mind and then the last point of course it's that it's still all incentive compatible so yeah sorry this was like a little bit I tried again this now to give you the stickman um intuition hopefully that helped a little bit better understand what we mean when we talk about normal real yield but that's basically our first argument for like why if we do nothing we'll keep sliding more and more and change the character of staking from profit driven to just an accounting Mirage so you right um oh wait wait wait wait wait these are two because you know from uh Kasper is our M and Chief and so he always like comes up with these images and I always forget about mentioning them so that's kind of supposed to uh illustrate you know like this the staking mechanism once we get to like almost everyone's Stakes right it becomes this like purely cyclical thing where you think you know coins come down the waterfall but then it turns out that the same coins that just went in right so basically there like demonstrating that and then um you know if you have to pay taxes uh even worse so uh yeah and now now over to you thank you Ana for explaining my meme um third um okay final section um so what's the problem with all of this right um ana basically just explained the whole um yeah why we should pay attention to nominal versus real yield and how sort of um staking becomes sort of a profitable activity to kind of being quasy forced Onto You economically by basically if you don't want to be diluted um it becomes this Duty actually um and so I'm going to as you see there's a bunch of um reasons that we've gone through in various iterations of different talks um so if you're interested in like different points you can um Talk talk to us after the talk or um find it online but I'm going to focus on um basically the network effects of money and then we have a little bonus one and um I think the network effects of money are are really crucial um and so what do I mean when I talk about this um so basically um money has Network effects and that also applies to um eth it applies to lsts and um in some sense um um basically um one way to think about it is um what makes an what makes money attractive it makes money is attractive for example if you can trade it at good execution um sort of this kind of mental model is liquidity beets liquidity so as you are a liquid asset people are incentivized to come to your asset to have that good execution and then some sense e and lsts are in competition in the sense that they some somewhat represent similar things lsts are derivative obviously there's like slashing risk and sort of um counterparty risk if you do it in in a mediated way um but the whole point here is that basically lsts have Network effects and these Network effects kind of lead to this winner T's most dynamic where it's likely that in the in in the long run that one LST will kind of win as the money um on that um LST in that LST Market um and what that then implies is that if we end up in a world where a lot of the eth is staked say 80% um hypothetically um what does that mean so you have 20% raw e that is still in circulation and 80% is um mostly going to be dominated by lsts because unfortunately there's only so many um solo stakers um and um basically then what happens is that the dominant LST is basically going to be the de facto money of ethereum replacing raw eth and what that means is that we kind of we're adding a permanent trust layer to ethereum so as an as an e holder you come to etherum and you kind of faced with the question do I want to hold this very expensive maximally trustless acid that is eth or um am I going to kind of not be the only silly one in town that is getting diluted and in LST that is actually at that point even more liquid than eth because 80 like in a market where 80% of um um is dominated by lsts then um those lsts are going to be the ones um more most liquid on decentralized exchanges um it might even the the the nfts might be denominated in that dominant LST etc etc and so um it really becomes this permanent trust layer that we're kind of economically kind of obviously it's not enforced onto anyone but it's kind of economically quasi enforced and um yeah you kind of sort of bullied into holding this trusted asset um and we're losing this maximally trusted um maximally trustless raw eth as the main money of ethereum and um I mean as a side point also as for E the asset this is also a sort of huge concession in terms of monetary premium giving up on that kind of maximally trustless um uh Vision um on that point we're actually going to give a talk um next Saturday with a bit more focus on on that topic at at the bankler summit um and then since we're up with time let me quickly talk about this um sort of principal agent problem um so basic basically at the very beginning before um the be basically the initial people that staked were almost kind of crazy lunatics you didn't even know when uh when the merch would happen people still went ahead and um and staked they were like basically um yeah crazy cool and then um the more more kind of like sort of the less risk friendly um solo stakers came in and then um lsts kind of kicked in and sort of the what I'm trying to get out here is that there is a spectrum of sort of who is actually running the software uh who like the which of the stakers are actually sort of running the software themselves and as you kind of go on that Spectrum um there's more and more intermediation so lsts um are already Des people staking with um rocket pool or Lio for example on chain have are pretty close to kind of the risk tradeoff and are aware um and then coinbase and centralized exchanges uh it's it's gets it gets more and more abstracted away it becomes this oneclick thing you don't even know what's actually happening it becomes less and less a conscious decision and then one reason why I kind of we added this as a bonus was yesterday we there was the news that um one of the ETFs just bought a um staking service provider and effectively um the the capital that is being staked is less and less um sort of close to the actual staking they don't even know what happens in the in the in the back end anymore um and so I mean one obvious question here that the elephant in the room is so what can we do about all of this um is um yeah basically changing the issuance curve this is our tool of how we can kind of control the um staking ratio and with a high staking ratio as we just discussed there's like a bunch of uh negative externalities and so there's different kind of approaches there's a sort of progressive yield reduction which one of our colleagues um Anders is going to talk about I think in maybe two hours or something um and then there's also ideas of curves that actually go to zero um to sort of meaningfully um kind of constrain the amount of stake that would um plausibly enter the the the staking set um and then we have the second sort of question of um what kind of staking range do we actually find desirable um I think there's there's kind of arguments for why a low range is uh preferable than levels to similar today like 25% um or sort of more closer to 50% obviously there there is basically no um True Value it's a trade-off space and um it is is um it's it's a it's a big topic and we're going to talk a lot about about this in the coming sessions and future talks um but yeah I think we we wanted to really focus on why we think um we need to change it and then um hopefully we'll get to actually change it eventually thank you [Applause] thank you so much um we have some short minutes to answer um the most V question and that would be what should the ideal percentage of State e be for example if you are 20% what ideal number do we want to equalize it all right so that was a little bit on the last slide it's It's tricky because there's two qu two versions of the question the one is like if we were to design the system completely from scratch before anyone starts staking what number would we choose and there I think as we as we said like initially the system was designed with Solo stakers in mind with solar stakers only we would never go about 5 10% of of each stake maybe because just not so many people are sophisticated enough to run their own local setup um and that would be exactly be these like 5 10% of like enthusiasts that really also care about the network right those would be exactly the people you would want in the system the most um it's it's unclear whether you know from from the we now we're already at 30% Stak it's unclear whether we can actually get back down to that level in a way that doesn't actually hurt existing stakers too much or drives out solo stakers then the next best thing could be like maybe stay try to stay equalized at at today's level um except that for forever basically out of every three coins you earn one of them will be fake and just account for dilution um or potentially you know if it takes us for a long time to find a solution here or the the process with the community like of of All of Us coming to agreement takes a lot of time we might have to try to like basically just cut off as basically blate in the process as as we you know then find an agreement on which might be closer to 50% great let's jump on to the next one is an let me check what's the highest v um will restake in worse in this issue um we gave a talk about this I'm not sure if it's online is it online uh yes online there's also a website issuance what the where all resources on the topic are listed and also that talk and basically reaking will is only yet another source it's similar to me in a sense there's different types of reaking some of them don't really affect this at all the ones that do are what we call basically proposal specific reaking yield that's things like pre- confirmations things where basically your role in the protocol is what gives you access to that yield but those can be um addressed we didn't talk about it in this talk but there is this kind of upcoming proposal for a feature called me burn more the more precise term would actually be propos a specific yield burn and that would capture both the existing me and the re the relevant portion of the reaking burn a reaking yield so that basically yes it worsen the problem a little bit but it also we have already the tools for for addressing it okay then I think one last is an enshrined LST solution um yeah I mean I I'm not sure if I mentioned it but like basically the the the idea is here we just want to be intentional and also like you know the other question of like why didn't we do it at the start of the the beacon chain well because no one thought about the kind of the the tail end of staking everyone just thought about how do we make sure there's at least 5% of each Stak and so I think now the point is okay we have a mature understanding of proof of stake now we want to really design for the end game or at least you know like as close as we get to can get to that and for that basically we just have to make the intentional Choice hey do we want low most percent Stak and it is actually income you can earn do we want to embrace you know 100% stake but then let's redesign the system and enshrine some sort of LST mechanism which by the way would not mean that lier or rocket pool would go out of business you know like you still need all of that on top but basically the core mechanism we could enshrine which would make that part trustless so we would at least end up with an LST that is a trustless LST in the system um but then let's make that choice and in that world of course yes stakers no longer actually earn money directly right it becomes a duty for all eth holders if we if we like that if we prefer that world over the 10% or 20% State let's make that choice and properly designed for it instead of just being passively pushed into it because when the beacon chain launched we didn't anticipate we would go this way