Two vaults can advertise the same headline number and mean completely different things by it. One pays 8% from interest that borrowers actually owe. The other pays 8% mostly in a token the protocol mints out of thin air to attract deposits. Both numbers say "8% APY." Only one of them is a durable return, and the difference between the two is the single most important distinction in evaluating any DeFi yield strategy.

This is not a new problem, but it remains the most common analytical gap institutional allocators make when entering DeFi. The phrase real yield was popularized after the late-2022 collapse of several DeFi protocols whose triple-digit advertised returns were funded almost entirely by token emissions; as the reward token's price fell, the actual dollar value of the yield collapsed with it, often to zero. The mechanism that caused those failures has not gone away. It has simply gotten better disguised inside more sophisticated vault structures.

This guide explains what separates real yield from subsidized yield, the mechanics of each, how to identify which one a vault is actually paying, and where the line gets genuinely complicated. For the broader evaluation discipline this fits into, our beyond-APY evaluation guide and our real returns after fees companion cover adjacent ground in depth.

What Is the Difference Between Real Yield and Subsidized Yield?

Real yield is paid from cash flow a protocol earns through actual economic activity: trading fees, lending interest, liquidation penalties, or other revenue that users pay because they want the service. If a perpetuals exchange earns a million dollars in trading fees this week and distributes a portion of it to stakers in a productive asset like ETH or stablecoins, that distribution is real yield. The protocol did not create new supply to pay it; it routed money that came in from real demand.

Subsidized, or inflationary, yield works differently. The protocol mints new units of its own governance token and distributes them to depositors, regardless of whether the protocol generated any revenue at all. Early liquidity mining programs, where lenders and borrowers received governance tokens simply for participating, are the canonical example. The yield is real in the sense that the tokens land in a wallet, but it is paid by dilution: existing token holders absorb the cost as the new supply enters circulation, and the dollar value of the reward depends entirely on the token maintaining its price under that selling pressure.

The practical test is simple: strip the incentive token component out of the headline APY and look at what remains. If the residual is competitive with a reasonable benchmark, like the prevailing Treasury bill rate for a stablecoin strategy, the yield is durable. If the residual collapses to near zero once incentives are excluded, the advertised number was mostly a temporary subsidy rather than a return.

How Do You Identify Which Type of Yield a Vault Is Paying?

Three checks separate genuine yield from dressed-up emissions, and all three are answerable from public data.

The first is the revenue-to-distribution ratio. Pull the protocol's actual fee revenue from a tracker like DeFiLlama Fees or Token Terminal and compare it to what is being paid out to depositors. If a protocol generates a million dollars in fees and distributes 700,000 dollars of that to stakers, the math reconciles and the yield is grounded in something real. If a protocol generates a hundred thousand dollars in fees but is paying out millions in rewards, the gap is being filled by emissions, and that gap is the size of the subsidy.

A concrete comparison makes the distinction tangible. Picture two stablecoin pools both advertising 8% APY. Pool A pays that 8% entirely from borrower interest on a major lending market, with a visible multi-year history of the rate moving with utilization. Pool B pays roughly 2% from underlying lending activity and the remaining 6% in a newly issued governance token distributed regardless of how much the pool actually earns. Both display the same number on a yield aggregator. Only one of them would still be paying 8% a year from now if the token's price fell by half, and that is the entire difference an allocator needs to price into the decision.

The second is the payment asset. Real yield is typically paid in a major asset, stablecoins, ETH, or a fee-bearing token, things with value independent of the protocol's own success. Subsidized yield is typically paid in the protocol's own governance token. GMX, for instance, pays stakers in ETH on Arbitrum and AVAX on Avalanche, never in inflationary GMX tokens, which is a structural choice that ties the protocol's payout obligation directly to its own trading volume rather than to its ability to keep its token price elevated. A vault or protocol paying primarily in its own native token is a flag worth investigating, not a disqualifier, but it changes what question you need to ask next.

The third is emissions schedule and unlock pressure. A token reward program has a finite emissions schedule, and as that schedule winds down or as large allocations unlock and hit the market, the effective yield can fall sharply even if the headline rate has not changed. Checking the emissions calendar and the unlock schedule tells you whether today's rate is likely to persist or is a temporary bootstrapping phase the protocol is actively trying to wind down.

Why Do Protocols Use Subsidized Yield at All?

Subsidized yield is not inherently dishonest. It is a legitimate, widely used bootstrapping mechanism. A new lending market or liquidity pool has no organic activity yet, no borrowers paying interest and no traders generating fees, and emissions are the standard way to attract initial liquidity that lets the protocol reach the scale where real activity can take over. Used during an early growth phase, this can be a reasonable tool, and many durable protocols started exactly this way.

The problem arises when the reward token remains the primary reason users deposit long after the protocol should have transitioned to organic revenue, or when the protocol never builds genuine activity and the entire yield structure depends indefinitely on continued emissions. In that scenario, liquidity is mercenary: it arrived for the subsidy and will leave the moment the subsidy declines or the token price falls, often taking the protocol's apparent stability with it. This is precisely the failure mode that defined the DeFi 2.0 collapse cycle, where triple-digit advertised yields evaporated as soon as token emissions could no longer outpace sell pressure.

A useful market-level signal is the shift institutional capital has been driving since 2025: traditional finance firms entering DeFi have been acquiring protocol tokens and capital directly in protocols with demonstrated revenue models, signaling confidence in sustainable business economics over speculative emissions rather than in token-incentive programs. Protocols that have successfully transitioned from emissions-funded growth to fee-funded sustainability are increasingly the ones capturing serious institutional allocation, while protocols still dependent on emissions face the risk of capital flight as that distinction becomes more closely scrutinized.

Where Does the Distinction Get Genuinely Complicated?

The real-yield-versus-subsidized-yield framework is useful but not always clean, and a few categories deserve specific attention.

Engineered or layered yield complicates the picture further. A vault that uses recursive borrowing, rehypothecation, restaking, or a delta-neutral basis trade is constructing yield from multiple onchain legs rather than a single fee stream, and that construction can be entirely legitimate while still being opaque. The full risk framework for yield farming now explicitly separates emissions risk from impermanent loss, oracle risk, and depeg risk specifically because blending them into one APY number obscures which risks an allocator is actually taking on. A useful starting discipline is this: if you cannot explain who is paying the yield, why they are paying it, and what risk you are taking on to receive it, treat the advertised APY as unproven until you can. A synthetic-dollar basis trade paying yield from positive perpetual funding rates is real yield in the sense that real counterparties are paying real funding costs, but it is not riskless, and that funding rate can turn negative.

Protocol-level fees that route to token holders rather than depositors are another nuance. Curve's admin fees, for example, flow from real trading activity to long-term lockers, while gauge rewards layered on top of that base are emissions and should be evaluated separately even though both appear in the same dashboard number. A vault that blends a lending protocol's organic supply APY with a separate incentive program needs both components disclosed, not a single blended figure that obscures which part is durable.

Liquid staking and restaking yield sit closer to the real-yield end of the spectrum because the underlying reward, network validation, is a genuine economic function with real demand, but the layered incentive programs that liquid restaking protocols often add on top follow the same emissions logic as any other token reward and should be assessed the same way.

What Should an Allocator Check Before Committing Capital?

The discipline reduces to a short, repeatable process. Start with the source: is the yield paid from fees, interest, or genuine protocol revenue, or from newly minted tokens? Separate the base rate from any incentive layer rather than accepting a single blended APY. Check the payment asset, since yield paid in a major asset or stablecoin carries different sustainability than yield paid in the protocol's own token. Review the protocol's actual revenue and usage data against what it is distributing. Study the emissions schedule and any pending token unlocks that could affect the reward token's price. And size the position based on the durable portion of the yield, not the headline number.

This discipline matters most for vault strategies specifically, because a vault can layer several yield sources, lending interest, a liquidity-mining incentive, a restaking reward, into one blended APY that looks like a single coherent number but is actually three different risk and durability profiles stacked together. The 10-question framework for picking a vault treats yield-source decomposition as a first-order diligence question precisely because the blended number on the landing page tells you almost nothing about what you are actually being paid for. The same discipline applies directly to the DeFi lending markets that underpin many vault strategies, where the utilization-driven base rate and any curator-layered incentive need to be evaluated as separate line items, and it connects to the broader question of what a vault's reported performance actually nets out to after every cost and subsidy is stripped away.

Conclusion

The question that matters is never simply how high the APY is. It is where the yield comes from and whether that source can continue. Real yield, funded by fees, interest, and genuine economic activity, is durable in proportion to the activity that generates it. Subsidized yield, funded by token emissions, is durable only as long as the protocol keeps minting and the market keeps absorbing the new supply without the token price collapsing.

Neither category is automatically safe and neither is automatically a red flag; emissions have a legitimate bootstrapping role, and real yield still carries smart-contract, liquidity, and market risk of its own. What matters is knowing which one you are looking at before you commit capital, and decomposing any blended APY into its constituent sources rather than trusting the headline. Platforms like Lucidly Finance build this decomposition into vault curation directly, because an allocator should never have to guess what is actually paying the yield they are being offered.