Real Yield in a Debasement Era

The most popular onchain yield strategy in 2026 — parking fiat-pegged stablecoins like USDC, USDT, or Sky’s USDS in a savings vault for 3–4 percent APR — is structurally a wealth transfer, not a savings product.

The math is straightforward. US M2 is growing 4.9 percent year-on-year, having reaccelerated since the Fed ended quantitative tightening on December 1, 2025 and resumed balance-sheet expansion eleven days later. CPI is 3.3 percent. Federal debt held by the public sits near 100 percent of GDP, with CBO projecting 156 percent by 2055. A saver earning 3.75 percent in USDS while M2 expands 4.9 percent is losing roughly 1.1 percent of real purchasing power per year. Net of “yield.”

This isn’t an arcane observation. It’s the central, unspoken fact of a debasement era: every yield calculation has a denominator, and the denominator is not stable. The dollar itself is being debased. Measuring DeFi returns in dollars without correcting for what the dollar is doing is like measuring the speed of a car using a speedometer that’s slowly running fast.

The implications invert most of what we’ve been told about “safe” yield in DeFi. The supposedly riskier asset — ETH staked at the Reference Benchmark rate of 3 percent — is currently producing higher real yield than the supposedly safer one: USDC in a money market.

The denominator problem

Three frames are available for measuring how fast the dollar is being debased:

CPI is the official benchmark, currently 3.3 percent year-on-year headline, 2.6 percent core. CPI has well-documented limitations: hedonic adjustments, smoothed shelter components, basket revisions that bias the headline downward. Most inflation skeptics — across the political spectrum — argue CPI undermeasures real-world price increases by 1–3 points.

M2 money supply growth is the cleanest debasement signal. It doesn’t depend on basket-weighting decisions; it just measures how many dollars exist. Currently +4.9 percent year-on-year. By this frame, “neutral” yield needs to clear roughly 5 percent.

Real-world purchasing power is the third frame. The Big Mac Index is the most accessible cross-currency reference: a basket of beef, bread, cheese, and lettuce, priced consistently across countries. It’s most useful as a comparative purchasing-power tool — across cities and currencies — than as a single-country inflation series. Use it as a sanity check on whether nominal yield is keeping up with real-world prices.

For yield investors, M2 is the most defensible benchmark. It operates closest to what monetary economists in the Austrian tradition treat as the operational definition of debasement. Under 4.9 percent M2 growth, almost no fiat-stablecoin yield in 2026 clears the bar.

The inverted comparison

Here’s the worked example. Three identically-sized $100,000 allocations over twelve months at current 2026 conditions:

• USDC at Aave (~2% APR). Nominal: $102,000. M2-corrected at 4.9%: $97,235. Real yield: −2.7%.
• Sky USDS Savings (3.75% APR). Nominal: $103,750. M2-corrected: $98,904. Real yield: −1.1%. Worth flagging: a meaningful share of USDS yield (and of comparable fiat-stablecoin savings products) is sourced from tokenized US Treasuries and money-market instruments. That yield is the Fed funds rate — which is set by the same institution debasing the dollar. The saver is being paid by the Fed, in dollars the Fed is printing.
• ETH staked at 3% APR (Staking Rewards Reference Benchmark). ETH net supply grew 0.23% over the period — post-Dencun, ETH is mildly inflationary, not deflationary. Real yield in ETH-terms: +2.8%. In USD-corrected terms, assuming long-run ETH/USD price equilibration to monetary conditions, the real yield is approximately the same.

The “safe” yield strategy produces a real-yield decline. The “risky” strategy produces a positive real return. The framing — fiat stablecoins = safe yield, staking = risky yield — that drives most onchain capital allocation is structurally inverted in a debasement era.

This is not an argument that ETH is always the right asset. ETH carries volatility, regulatory, and idiosyncratic risk that stablecoins don’t. A treasury that needs to pay rent in USD next month should not hold ETH. But the binary risk framing is wrong on the real-yield axis. The choice of denominator matters more than most yield comparisons acknowledge.

The categorical breakdown of 2025 onchain yield (per Vadym’s analysis on X) makes the same point at scale. Roughly $8 billion of onchain yield was generated across DeFi in 2025:

• AMM trading fees (52.5% / ~$4.2B) — mostly real economic rent. But concentrated-liquidity LPs frequently lose to toxic order flow. Hard for passive investors to capture.
• Borrow interest (22% / ~$1.76B) — mostly nominal. Aave’s stablecoin borrow rates track Fed funds by construction. Half of borrowing is recursive looping; move the layer, you don’t move the underlying source.
• Staking + MEV (12.3% / ~$1B) — real relative to the denomination, but cross-frame comparisons require care. ETH at 3% APR with supply +0.23% ≈ +2.8% real yield in ETH-terms. SOL at ~6% APR with inflation ~4.7% ≈ +1.3% real in SOL-terms. The denomination matters more than the headline number.
• RWAs (9.4% / ~$700M) — mostly nominal, and worse than that on closer inspection. RWA yield is, almost by definition, the Fed funds rate plus a small credit spread. The same applies to the meaningful share of fiat-stablecoin savings yield that flows from tokenized Treasuries. When the Fed prints money to monetize fiscal commitments, the Fed funds rate is what flows back to RWA holders and stablecoin savers as “yield.” Calling that yield “real” is a category error.
• Perps funding (3.8% / ~$300M) — genuinely real but small and volatile. Returned 16% in 2021, 0.6% in 2022, 9% in 2023, 13% in 2024.
• Insurance underwriting (0.07%) — negligible today. Worth watching as DeFi matures.

Net: maybe a third to half of total onchain yield qualifies as real economic rent. The rest is debasement compensation in different wrappers — with the fiat-stablecoin layer being the most structurally compromised, because its yield is most directly the Fed funds rate.

The architectural question

Step back. Every analysis above takes the unit of account as given. We’re measuring yield in USD or ETH, complaining the USD is debasing, constructing portfolios to route around the limitation.

The deeper question: can the unit of account itself be designed to remove this calculation?

Hayek argued in Denationalisation of Money (1976) that competition among private monies, none protected by political privilege, would produce stable purchasing power as an emergent property. The classical gold standard between 1880 and 1914 produced roughly that outcome — US inflation averaged 0.1 percent annually — not because anyone managed it but because the institutional architecture was credibly neutral.

Whether modern monetary rails can reproduce that neutrality is an open research question. A new generation of so-called flatcoin designs aims at exactly this: stablecoins not pegged to the dollar but to spending power itself — explicitly distinct from the fiat-pegged stablecoins (USDC, USDT, USDS, PYUSD) that dominate today’s market and that, as we’ve seen, structurally cannot deliver real yield in a debasement era. Independent rating frameworks have begun evaluating monetary projects against criteria like neutrality, fiat-independence, debt-free issuance, and long-horizon spending-power stability — the properties Hayek identified, restated for crypto rails.

None of this is solved. The Austrian objection still holds: any politically administered index becomes the next instrument of debasement (Mises’ critique of Fisher’s compensated dollar, 1912). The harder design question is whether stability can emerge from the institutional architecture rather than be targeted by a central authority. A long history of currency debasement, from Roman coinage to modern M2 expansion, suggests the underlying problem isn’t new. Only the rails for solving it might be.

Four principles for a debasement-aware portfolio

1. Measure yield in multiple denominations. Re-denominate USD yields in ETH, BTC, gold, and Big Mac terms. The cross-denomination spread is itself a signal.
2. Discount fiat-stablecoin yields heavily during M2 expansion. A 2–4 percent USDC or USDS yield during 4.9 percent M2 growth is a −1 to −3 percent real return — especially when the underlying yield source is tokenized Treasuries (i.e., the Fed funds rate dressed up as DeFi). Treat it as a placeholder, not preservation.
3. Favor real-economic-activity yield over monetary-policy yield. AMM fees, MEV, and properly-priced perps funding are genuine economic rents. Borrow rates that track Fed funds, RWA yields that pass through tradfi, and fiat-stablecoin savings rates set administratively are mostly nominal.
4. Diversify across denominators, not just protocols. The biggest risk in onchain yield isn’t smart-contract failure. It’s that every position is denominated in a unit that’s slowly losing purchasing power. Spread the denomination, not just the protocol.

Closing

Onchain yield in 2025 was approximately $8 billion. That number is real. What’s less clear is how much of it represents purchasing power. The fiat-stablecoin layer almost none of it. The staking layer, more than the headline APR suggests. The economic-activity layer (AMM fees, MEV, perps funding), genuinely some.

Until the unit of account is stable, every yield calculation is a partial measurement. The job of a sophisticated investor is to remember which part is missing — and to be skeptical, in particular, of “yield” denominated in the same unit being debased to fund it.