This post appeared originally on The Defiant under the title “‘Rich Getting Richer in PoS Chains:’ By Chainflow’s Chris Remus.” You can read the original on The Defiant,here.
There’s a lot of excitement in the staking economy these days. Existing staking networks are starting to attract developers and users. Staking networks continue launching at a fast pace. Staking token prices continue soaring in this current bull run.
Yet this enthusiasm washes over an important weakness. Stake in these networks isn’t decentralized. In fact, stake is highly centralized.
This post’s title is inspired by the site arewedecentralizedyet.com. I first came across it in 2018 when Chris Burniske tweeted and Fred Wilson wrote about it.
The site intends to indicate how centralized proof of work (PoW) network mining is. Fred’s answer back then was that PoW wasn’t really decentralized yet. A look at the site now indicates not much has changed since then.
This analysis has stuck in my mind since entering the staking economy as a Cosmos validator operator in October 2017. I took this step after discovering proof of stake (PoS) in mid-2016. At the time, I felt optimistic about PoS as a more inclusive alternative to PoW.
At the same time, I also held reservations about staking’s rich get richer problem. I published those thoughts after calling attention to the problem during early Cosmos testnets.
The issue manifested very soon after the mainnet launch, when stake consolidated among a very small number of validators. Large whale validators also emerged, after showing little to no presence in the testnets leading to launch.
Since then, it’s become clear that stake centralization isn’t unique to Cosmos. Choose a popular staking network and you’ll likely see high stake concentration among a very small number of validators.
Sure, you may see networks with 100’s of validators. This is a step in the right direction. However, look closer and you’ll see that stake is still highly concentrated at the top.
You’ll even see 1,000s of Ethereum validators. If you look closer there, you’ll see that the number of unique operators is much smaller. In fact, I’ve heard anecdotally that staking-as-a-service providers are offering implementations that can scale to 100,000s of ETH validators for their largest clients.
Staking’s supporters often cite the need for validators to have “skin in the game.” The argument goes that the more skin a validator has in the game, the less likely the validator is to manipulate the network.
This argument carries some weight. However, at a certain point, it becomes a paradox. More skin means more stake. So skin accumulation works against decentralization.
Furthermore, relying too heavily on the skin in the game argument provides cover for larger operators who use it to accumulate huge and disproportionate stakes. They do this while simultaneously claiming to be supporters of decentralization.
Check the carefully “crafted” messaging of Coinbase’s acquisition of Bison Trails in the Coinbase and Bison Trails announcement as one indirect example of what I’m talking about.
I don’t see how this acquisition “…strengthens the entire crypto ecosystem.” as the Coinbase announcement claims. If you’ve read this far, you probably won’t be surprised that I see this acquisition more as a step toward a decentralized doomsday scenario.
It’s hard to support decentralization when your business model and investors expect you to capture every token of delegation possible. This zero-sum-game mentality, so prevalent in the legacy economy, has unfortunately carried over and proliferates in these early days of the staking economy.
Well, if you’re coming from the legacy financial services industry, with big pockets and lots of money to spend, it’s not a big deal at all. The current state of stake centralization is actually exactly the exploitable condition that can be leveraged by such participants to make their next million, millions or billions. I met lots of operators in this category during New York City Blockchain Week 2019.
These validators can accumulate large stakes through their privileged access to capital, run a validator, pay someone to do it for them or delegate to one of the big validator operators. After that, they can sit back and watch their bag grow, without participating in the community in any meaningful way.
Yet if you truly care about decentralization, stake centralization is a problem. To begin, controlling 33% of a network’s stake means you are able to also control whether a network runs or not.
Controlling over 50% of the stake means you can pretty much control the network’s governance process. It’s also easier for a small number of operators to be manipulated by any number of bad actors or even governments.
These are scary scenarios for those of us that are working hard to build a more open, equitable and inclusive decentralized economy.
Why is stake centralized? Why is stake continuing to centralize? Is it really that hard to decentralize stake?
In theory, as I mentioned above, some, if not most, staking protocols encourage the rich get richer scenario, using the skin in the game argument as cover. Staking protocol designers often say there is no other way.
This isn’t quite good enough though. We’re too early in the staking game to start using the “this is the way it’s always been done” excuse, to throw up our hands as if there’s no other solution. It’s not an easy problem to address, yet with so many brilliant people doing really disruptive things in crypto, it’s possible.
In practice, often starting in the testnet stage, community members are understandably hopeful. If potential stake centralization issues are brought up by a brave community member (more on this later), the community —including the core protocol development team— often responds with an answer along the lines of, “well that might happen, yet people are good and they wouldn’t do that.”
They hope that validators will act altruistically. (My gratitude to Collin Myers for introducing me to the altruistic perspective during this episode of the Staking Defense Podcast.) Yet when real money starts flowing in mainnet, people indeed do “that,” e.g. try and grab as much stake as possible.
Altruism gets lip service and nothing else. Validators talk the decentralized talk, yet don’t walk the walk.
After the protocol has left the gate and mainnet launches, validators are best positioned to address the issue. Yet most validators are silent.
The bigger validators among whose stake is centralized have a good thing going. Why would they want to disrupt it, especially if they’ve raised large funding rounds and have investors to satisfy? Smaller validators often feel too powerless or maybe even scared to say anything too loudly.
Many smaller validators rely on foundation grants for survival. As centralization is a touchy subject, smaller validators may fear losing those delegations if they say something negative related to centralization.
Staking community members, particularly delegators, also contribute to the problem. Many, if not most delegators, want to passively earn income staking. This is understandable.
However, when the passive approach is taken too far, it often results in delegators not doing their homework when it comes to choosing validators. In a way, delegators take the easy route when choosing their validator.
They look at a block explorer, which stacks the odds in large validators’ favor by default sorting validators by stake, with the largest stakes at the top of the list. Delegators see who has the lowest fee among the top 3 to 5 validators by total stake and pile their delegation on to the already disproportionately large holdings.
Delegators don’t often consider spreading their stake among a diverse group of operators. They don’t do the research to understand the values of the validator they delegate to and if those values align with theirs. They prioritize short-term returns over long-term network sustainability, health and security.
Recognizing that staking is in its infancy, you may wonder if stake will ever decentralize. I don’t believe it’s too late. I do believe it’s possible for stake to decentralize over time.
Yet changes are needed for that to happen. When I think about these changes, I wonder what catalyst is required to start elevating this issue to levels that encourage staking economy participants to take real action.
Human nature dictates that as long as prices keep rising, income remains passive and there are no unpleasant shocks to the system, the staking economy will keep humming along, while stake continues to centralize, weakening the economies’ foundations as a result.
In fact, the Kava network, went down for a day or so last year. I’m only aware of a single validator that tried valiantly, yet futilely, to call attention to the outage. The network published a short post-mortem after the outage and nothing else. Although now that I tried to find the link to the post-mortem, it seems to have been removed.
My sense is that for those who are unaware of or choose to not engage in this centralization dialogue to become engaged, they’ll have to experience a disruption. I don’t know what that disruption would look like, yet it would have to be big enough to jar them out of their current “everything is great,” rose-colored glasses perspective.
At the same time, experience tells me that there are people out there who truly care about fully manifesting the power of a decentralized future. I’ve met and spoken to them. They’re starting to contact me a little more often now than they did six months ago. This gives me hope.
Yet, these voices can’t sit passively in the background. They have to become louder. Because, as we’ve seen with PoW, the longer the benefits of centralization compound to that small number of participants that have privileged access to capital which gives them a head start, the harder it becomes to decentralize.
And if all we accomplish with this decentralized movement is to rebuild the current economy and the inequitable and exploitative power structure it enables, which is controlled by a shrinking number of increasingly wealthy participants, well, then what’s the point?
You can read the original version of this post, and more, on The Defiant.