I keep getting asked some version of the same question about ETH sleeves, and it usually arrives dressed up as a math problem. If staking pays roughly 3 to 4 percent, and I hold for ten years, and it compounds, isn't leaving it unstaked just lighting money on fire. The math half of that is easy and mostly correct. The part people skip is that staking yield is not free money sitting on the table. It comes with a tax treatment, a set of risks, and a few tradeoffs that quietly eat into the number that looked so clean in the spreadsheet. So before you stake anything, it is worth being honest about both sides.
What the compounding actually buys you
Start with the friendly version, because it is real. Take an ETH position and assume it earns something in the low single digits per year, paid in more ETH, and you restake the rewards so they compound. Over a decade, a yield in the 3 to 4 percent range roughly grows your ETH count by something like a third to a half, before you touch the price of ETH itself. That is not nothing. If your whole thesis is that you are going to hold this ETH for ten years regardless of what happens, then accumulating more units of the thing you already decided to hold is close to the definition of a good deal.
The key phrase there is more units of ETH. Staking yield is denominated in the asset, not in dollars. So the compounding works on your ETH balance, and then the dollar outcome rides entirely on the price. If ETH goes nowhere for a decade, a 35 percent larger stack of something flat is still a modest win. If ETH falls by half, you have more of a thing that lost value, and the yield softened the fall a little. The yield is a real edge, but it is a second-order edge. It never rescues a bad call on the underlying, and I have watched people talk themselves into a position mostly because of the yield, which is the tail wagging the dog.
The costs nobody puts in the spreadsheet
Here is where the clean number starts leaking. Three things, roughly in order of how often they surprise people.
First, tax. In a lot of jurisdictions, staking rewards are treated as ordinary income at the moment you receive them, valued at the price on receipt. That matters more than it sounds. You owe tax on rewards you may not have sold, in a year the token might later crash, and you owe it at your income rate rather than a long-term capital gains rate. Then, when you eventually sell that reward ETH, you pay capital gains on any further appreciation, with a cost basis set at the receipt-date price. So a single reward can get taxed twice in effect, once as income when it lands and again as a gain when it leaves. In a taxable account, this drags the real after-tax yield well below the headline number, and it creates a recordkeeping chore, because you are logging a taxable event every time rewards hit. The specific rules differ by country and change over time, so treat this as the shape of the problem rather than tax advice, and check your own regime.
Second, liquid staking and depeg risk. Plenty of people do not stake natively because they do not want to lock ETH, so they use a liquid staking token that represents their staked position and trades freely. Convenient, and usually fine. The failure mode is that these tokens can trade below the value of the ETH they represent, sometimes meaningfully, during stress. If you need to exit during exactly the kind of panic where you most want out, you may have to sell at a discount, or wait in an exit queue, which can take days when everyone hits it at once. The peg is a soft promise, not a hard one.
Third, lockup and custody. Native staking means your ETH is committed and not instantly liquid, and unstaking runs through an exit queue that lengthens when demand spikes. If you stake through an exchange or a custodial service to avoid the hassle, you have handed over custody, which reintroduces the counterparty risk a lot of people came to self-custody to escape. Every convenience here trades away either liquidity or control, and you should know which one you are paying with.
When not staking is the right call
This is the part that gets lost, because the default framing assumes staking is always correct and any hesitation is laziness. There are real cases where leaving ETH unstaked is the better portfolio decision.
- You might sell within your tax year's short-term window, or you want the flexibility to. Locking up or holding a liquid staking token adds friction and, potentially, a discount at exit, right when you want to move fast.
- The position is small enough that the yield is a rounding error but the tax paperwork is a real annual cost. Below some threshold, the admin is not worth the extra ETH.
- You are already close to a rebalancing band and expect to trim. Staking a slice you plan to sell soon just means unstaking it, and possibly eating a queue or a depeg to do so.
- You do not want to take on smart-contract or validator risk on top of price risk. Native staking has slashing risk if the operator misbehaves, and liquid staking adds contract risk. If your whole reason for holding ETH is a clean, boring long-term bet, piling on operational risk for a few points of yield is a fair thing to decline.
None of those are exotic. They are just the ordinary situations where the yield does not clear the costs it drags in.
A stake-or-not framework by account type
The cleanest way I have found to cut through it is to decide by the account the ETH lives in, because the account mostly settles the tax question, which is usually the deciding factor.
In a tax-advantaged account, if your rules allow crypto and staking at all, staking is close to a default yes for ETH you intend to hold long-term. The income-on-receipt problem largely disappears inside that wrapper, so the after-tax yield stays close to the headline, and compounding does its job. The main things left to check are custody and lockup, so favor a setup where you are not exiting during a panic.
In a regular taxable account, treat it as a real decision, not a default. Stake only the core, long-hold portion of the sleeve that you are confident you will not touch for years, so the yield has time to outrun the tax and paperwork drag. Keep a trading or flexible slice unstaked and liquid. If you use a liquid staking token here, size it so you are never forced to sell the whole thing during a depeg, and keep clean records of every reward's receipt-date value, because you will need them.
If the ETH sits on an exchange or with a custodian, the real question is whether you are comfortable with that counterparty holding a committed position, not whether the yield is attractive. If you are not, the right fix is usually to move to self-custody first and decide about staking second, rather than staking because the button is right there.
A rule of thumb that has held up for me. Stake the part of the position that behaves like a ten-year holding, in an account where the tax treatment does not gut the yield, using a method where you can afford to sit through a bad exit queue. Everything that fails one of those three tests, leave unstaked and stop feeling guilty about it. The yield is a nice tailwind on a decision you already made for other reasons, and it is worth the least when it is the reason.