Legacy Frax Dollar (FRAX) — Risk Report
High risk · 2.5/10 · Wind-down asset, existing holders should exit
This report covers Legacy Frax Dollar (FRAX,
0x853d955aCEf822Db058eb8505911ED77F175b99e), NOT frxUSD. They are separate assets with separate balance sheets and separate contracts. The 1:1 FRAX→frxUSD migration path was closed in April 2025 (FIP-430). frxUSD is Frax Finance’s current flagship; L-FRAX is deprecated.
Legacy FRAX is the original Frax stablecoin (launched Dec 2020), now in wind-down. Following the April 2025 “North Star” upgrade, Frax Finance separated the L-FRAX balance sheet from frxUSD and stopped offering the 1:1 swap that used to exit legacy holders into the successor. Today L-FRAX sits with ~$132M circulating supply, a 91% collateral ratio, a -$16M deficit, and only $24M of real (non-FRAX-denominated) backing — the rest of the “backing” is the protocol holding its own token.
| Peg | Yield | Exit | Status | Chains |
|---|---|---|---|---|
| $1 nominal (trades persistently ~$0.99) | None | Curve frxUSD/FRAX pool — $4.1M depth, frxUSD-denominated (not USDC) | Wind-down (legacy) | Ethereum + 6 chains |
Backing & solvency
L-FRAX is under-collateralized, and the realistic backing is shrinking.
| Metric | June 1, 2026 | August 23, 2026 |
|---|---|---|
| Headline collateral ratio | 90.99% | 91.05% |
| External-only CR (non-FRAX backing only) | 13.5% | 17.1% |
| Net deficit | −$16.3M | −$10.88M |
| Circular treasury (protocol holding its own token) | 77% of “backing” | 70.7% |
| Real external assets | $24.4M | $20.84M |
| Total liabilities | ~$180.8M | $121.57M |
The headline 91% CR moves slowly, but the External-only CR — the share of backing that is not L-FRAX, sFRAX, or Frax-ecosystem tokens — fell from 15.8% in March to 13.5% in June, as real external assets went from $28.6M → $24.4M while supply contracted less.
⚠️ Updated 2026-08-23, and the ratio has since reversed — for a reason that is not good news. External-only CR is now 17.1%, up from 13.5%. But external assets did not grow. They fell again, $24.4M → $20.84M, another 14.6%. The ratio improved because liabilities collapsed faster: ~$180.8M → $121.57M, down 32.8% in twelve weeks.
Read the numerator across all three readings and the picture is unambiguous:
| March | June | August | |
|---|---|---|---|
| Real external assets | $28.6M | $24.4M | $20.84M |
| period change | — | −14.7% | −14.6% |
| External-only CR | 15.8% | 13.5% | 17.1% |
External backing has fallen by essentially the same proportion in each of two consecutive periods — about 15% a quarter, 27% since March — while the ratio measuring it fell, then rose. The reversal is entirely a denominator effect.
The deficit is not widening on the axis that matters, and the reason needs stating precisely. In dollar terms the external asset base is still shrinking, steadily and at a consistent rate, which is what that sentence was pointing at. But the metric it cited has moved the other way, and a reader checking the External-only CR today would find it improving and conclude the report was wrong. Both facts are true and they point in opposite directions; the numerator is the one that describes the asset.
⚠️ What that adds up to is a book retiring, not one recovering or deteriorating. Liabilities down a third, deficit down a third, circular share falling from 77% to 70.7% — every ratio improving, on a backing base that is still about 71% the protocol’s own tokens. This report already describes L-FRAX as in wind-down; what is new is that the wind-down is now visible in the ratios themselves, and the improving ratios are a symptom of it rather than evidence against it. A holder should not read the improvement as repair.
⚠️ Provenance, stated because it bears on how much weight these figures carry: they come from the issuer’s API plus on-chain reads, and the independent third-party cross-check is 21.7 days old and was not attempted on this run — it runs monthly. These numbers are not independently corroborated, unlike most figures in this coverage.
The on-chain CR oracle is frozen. The on-chain collateral_ratio() value reads 94.5% and has not changed since June 19, 2023 — predating the “100% CR” vote and the North Star upgrade entirely. Any protocol reading this oracle for risk parameters is consuming a value more than three years old with no relationship to current state. The actual CR was 90.99% at the June reading — 3.5 percentage points below the stale figure — and 91.05% at 2026-08-23, a gap of 3.4 points. ✅ That re-read strengthens this finding rather than ageing it: across 83 days in which liabilities fell by a third and every other ratio on this page moved, the oracle did not move at all and the gap stayed within a tenth of a point. It is not drifting toward correctness; it is simply frozen, and the passage of time keeps making that more demonstrable.
There is no reserves audit. The balance sheet is on-chain and publicly visible via the live dashboard and Frax Facts, but no third-party CPA firm attests to the AMO positions, and S&P rated FRAX 5/5 (Weak) in December 2023 — the worst score among the eight stablecoins it assessed at the time.
Exit liquidity
The primary exit venue is Curve, the depth is thin, and the quote is in frxUSD, not USDC.
| Venue | Depth | Notes |
|---|---|---|
| Curve frxUSD/FRAX (Ethereum) | $4.11M total ($3.87M frxUSD, $248K FRAX) | The main price-setting pool. Most FRAX has been drained to the buy side. A $1M+ sell moves the pool meaningfully. |
| Bridged FRAX on other chains | Varies, mostly thin | Cross-chain inventory exists on Fraxtal, Arbitrum, Optimism, Polygon, BSC, Avalanche via FraxFerry. Bridging back to Ethereum to exit is the practical path. |
| CEX | None to speak of | Not a listed asset on major centralized venues. |
Exit takes two hops for USDC: FRAX → frxUSD (Curve) → USDC (further DEX route). Each hop carries slippage, and the $4.1M Curve pool is the binding constraint — exiting a sized FRAX position pushes price through that pool before reaching frxUSD, then through frxUSD’s own liquidity to reach USDC.
Practical guidance: exit in small tranches. The pool is already below the $5M depth threshold where round-trip slippage becomes material for retail-scale positions.
Peg dynamics
L-FRAX has traded at a persistent discount to par for months and has been widening in late May. Over the last 30 days (May 2 – Jun 1, 2026, 710 hourly readings), the average price was $0.9928 (-0.72% deviation), with 92% of readings below $0.995 and 5.2% below $0.990. The week of May 25 alone had 17% of readings below $0.990 — the worst concentration since the Feb-Mar baseline window — and the latest reading on Jun 1 was $0.9908 (-0.92%). The discount is widening as the balance sheet’s external-CR slide (15.16% → 13.5% in two months) gets repriced; the peg is the market’s running estimate of real backing.
The peg is held by passive inertia, not active defense. The AMO Minter contract — the system designed to mint and deploy new FRAX into peg-stabilizing positions — has been dormant since February 2023 (3+ years). Today’s price is the equilibrium between existing legacy LP positions (about $55M in a Convex FRAX/USDe pool, $37M in a Compound borrow position, and a handful of smaller LPs) absorbing marginal trading flow. Nobody is actively rebalancing or defending the peg.
Governance has been clear about the priority. Every AMO-related proposal since mid-2025 has been frxUSD/sfrxUSD-specific (FIP-434, 437, 439, 443, 444). The only L-FRAX-touching governance is a May 2026 TempCheck for early withdrawal of inactive Convex pools — i.e. proposals about unwinding L-FRAX positions, not defending them.
Admin & multisig
The FRAX token contracts and AMO comptroller are controlled by the same Frax 3-of-5 multisig with no timelock (0xB1748C79…3f27) that runs frxUSD. The multisig is active (last transaction May 28, 2026) but its activity does not prove L-FRAX peg defense — it manages both ecosystems and the policy direction is firmly toward frxUSD.
There are no new audits of the legacy AMO stack. The original CertiK audit (Oct 2020, 39 issues — 3 critical, 8 major — all addressed) predates most current deployments.
Who it’s for · Who should avoid
For (genuinely no one in good conscience):
- Holders with existing positions that need to be exited methodically — see Exit liquidity above.
- Specialized fixed-income operators evaluating the related Frax Bonds (FXBs), which are zero-coupon claims on L-FRAX at specific future dates and trade at discounts. The wrapper isolates from new-issuance risk but inherits L-FRAX credit risk at maturity. FXBs require their own per-maturity assessment.
Avoid if (the general case):
- Looking for a stablecoin allocation. The 13.5% external-only CR and the absence of any active peg defense mean L-FRAX should not be treated as a $1-pegged asset in any sized position.
- Holding L-FRAX as collateral inside a third-party protocol that reads the on-chain CR oracle. The oracle is frozen at 94.5% from June 2023 — your protocol may be using a value that overstates actual collateralization by ~3.5 percentage points.
- Considering FRAX as a frxUSD substitute. They are different assets; the 1:1 migration is closed.
What to watch
- External-only CR. Currently 13.5% and falling (about $1.5M/month of external assets disappearing). If it drops another $5M without supply contraction, the realistic backing position becomes meaningfully worse.
- Curve frxUSD/FRAX pool depth. Live on the dashboard. If pool depth drops below $3M, exit liquidity is in a different regime.
- Comptroller multisig activity. Track the comptroller address for sudden large withdrawals from the Convex FRAX/USDe position (~$55M) or the Curve frxUSD/FRAX LP — those are the legacy peg-support positions, and there’s no governance lock preventing their reallocation to frxUSD.
- Governance proposals touching L-FRAX. The May 2026 TempCheck for early Convex pool withdrawal is the only L-FRAX-touching governance currently in motion. Any FIP proposing further balance-sheet separation or accelerated wind-down would be material.
- Compound AMO leverage. $20M USDT borrowed against $37M sFRAX collateral. The position has been steady for 2 months; a sudden change (forced unwind, liquidation pressure) would worsen the deficit.
A note on Frax Bonds (FXBs)
L-FRAX has a related instrument worth disambiguating: FXBs (Frax Bonds) are zero-coupon ERC-20 tokens that redeem for exactly 1 L-FRAX at specific future dates (2026-12-31, 2027-12-31, 2029-12-31, 2055-12-31). The L-FRAX collateral is locked inside each FXB contract at mint, so redemption is contract-guaranteed for already-issued FXBs — but the asset you receive at maturity is L-FRAX, with all of the risk described above. FXBs trade at discounts that price in the L-FRAX trajectory. They are a separate decision from holding L-FRAX directly; the per-maturity dynamics matter.
This report describes Legacy FRAX as of June 1, 2026, with balance-sheet figures re-read 2026-08-23. Live balance sheet, CR, AMO positions, and Curve pool depth are on the live dashboard. Frax Finance has not engaged on this report; figures are sourced from on-chain reads, the Frax Facts API, and the Frax governance forum. Corrections welcome at info@tidresearch.com.
Revision history: 2026-08-23 — balance sheet re-read after 83 days; no score change. Liabilities ~$180.8M → $121.57M (−32.8%), real external assets $24.4M → $20.84M (−14.6%), net deficit −$16.3M → −$10.88M, circular treasury 77% → 70.7%, and External-only CR 13.5% → 17.1%. ⚠️ The ratio improved and the thing it measures did not. External backing fell again, by almost exactly the rate it fell the previous period — −14.7% March→June, −14.6% June→August, −27% since March — while liabilities collapsed twice as fast. The reversal is a denominator effect, and this is the fourth figure in this coverage to move favourably for a reason that does not mean what it looks like, after Figure, wYLDS and reUSD. When a ratio moves in your favour, check which side of it moved. ”
“The previous conclusion — “the deficit is widening on the axis that matters” — is restated rather than deleted: the dollar external base is still shrinking steadily, which is what that sentence was pointing at, but the metric it cited has reversed, and a reader checking the External-only CR today would find it improving and conclude the report was wrong. The numerator is the figure that describes the asset. ”
“The reading this adds is that the book is retiring rather than recovering or deteriorating — every ratio improving on a backing base still about 71% the protocol’s own tokens. This report already described L-FRAX as in wind-down; what is new is that the wind-down is now visible in the ratios, and the improving ratios are a symptom of it. ”
”✅ And the frozen on-chain oracle finding strengthens: collateral_ratio() still reads 94.5% against a computed 91.05%, a gap of 3.4 points — unchanged within a tenth of a point across 83 days in which liabilities fell by a third. Not drifting toward correctness; frozen, and more demonstrably so with time. ”
“⚠️ Provenance caveat, stated on the page: these figures come from the issuer’s API plus on-chain reads, with the independent third-party cross-check 21.7 days old and not attempted on this run — it runs monthly. They are not independently corroborated, unlike most figures in this coverage. ”
“No axis moves: Backing 2.0 is near the floor of the scale, and a wind-down is not obviously better for a holder than a widening deficit — the exit remains the same and the backing base remains overwhelmingly circular. last_verified is not bumped: the balance sheet was re-read but the peg series, AMO positions and Curve depth in this body still date from the 2026-06-01 pass. This report remains staging-only — it carries no production flag and is not published — though it is the analytical reference behind a public dashboard.