Skip to main content
Pharos

Programmatic peg defense, with different collateral models

Protocol-level supply rules and arbitrage defend the peg; current designs can hold real collateral, unlike historical pure mint/burn systems.

This mechanism family uses programmatic issuance, redemption, or market operations to defend a target value. Collateral is design-specific: FPI's documented design uses legacy-FRAX-linked balances and AMO positions, and HBD converts against HIVE under a debt-ratio haircut. An algorithmic mechanism does not itself mean an unbacked token. The current collateralized variants are described below.

Pharos taxonomy no longer treats algorithmic as a tracked backing bucket; that historical backing label remains for off-catalog PSI assets. The walkthrough below covers the historical pure mint/burn variant, exemplified by UST/LUNA, not the reserve mechanics of FPI. Pure uncollateralized algorithmic stablecoins are not a live design pattern at scale in 2026.

Users burn a governance token to algorithmically mint STBL; an autonomous mint/burn module defends the peg through arbitrage incentives; the system has no 1:1 reserve backing, so confidence in the governance token is critical.01Burn governance tokenalgorithmic mint02Mint/burn AMOdefends peg via arbitrage03STBL mintedno 1:1 backing

Diagram: historical UST-style pure mint/burn, not the collateralized variants below.

How it works

The flow, step by step

  1. Historical pure mint/burn: burn governance token

    In the UST/LUNA-style variant, a user burns governance tokens worth a given dollar value, and the protocol mints that same value of stablecoin. There is no segregated 1:1 collateral: the loop relies on the governance-token market float, itself partly determined by the stablecoin's success. The reverse trip exists too: redeem the stablecoin for the same value of newly minted governance token. Pharos now tracks this as mint-authority and mechanism risk rather than a FreezeWatch freeze tier.

  2. Historical pure mint/burn: arbitrage loop

    Above $1, arbitrageurs mint at $1 of governance-token value and sell higher; below $1, they buy the stablecoin and redeem for $1 of newly minted governance token. This historical pure mint/burn loop has no segregated collateral buffer, and each defense changes the governance-token float. It is not the same as collateral-backed AMOs, which can deploy existing reserves.

  3. Historical pure mint/burn: reflexive stress

    In this uncollateralized variant, peg stability depends on governance-token liquidity, market confidence, and arbitrageur willingness. In a panic the governance token sells off, newly minted governance tokens lose value before arbitrageurs can sell them, and the peg breaks reflexively. That UST-style dilution loop does not describe ZSD, which never mints ZEPH to defend its peg.

Tracked examples

Tracked coins using this design

Decommissioned

Designs that broke and stayed broken

  • NuBits2018-03

    One of the first stablecoins ever created (2014), NuBits held its peg for two years before holders dumped it to chase Bitcoin gains. A pioneering cautionary tale about algorithmic pegs backed by volatile assets.

  • Empty Set Dollar2021-01

    Pioneered the "seigniorage shares" model in DeFi. When ESD traded below $1, users could buy coupons (burning ESD) for future redemption at a profit. The mechanism worked during expansion but collapsed when confidence evaporated: coupons expired worthless, and ESD fell to $0.01.

  • Dynamic Set Dollar2021-01

    A fork of ESD with faster epoch cycles (2 hours vs 8 hours), designed to stabilize more quickly. Instead, the shorter cycles amplified volatility. DSD spiked to $3 during expansion then collapsed to $0.24 in the same month.

  • Basis Cash2021-01

    An anonymous fork of the Basis design, BAC lost its peg within weeks of launch. Later revealed to be co-founded by Do Kwon under a pseudonym. He learned nothing before building the even more catastrophic TerraUSD.

  • IRON2021-06

    Dubbed crypto's "first large-scale bank run." IRON was partially collateralized (75% USDC, 25% TITAN). When whales dumped TITAN at its peak, a flawed redemption mechanism sent TITAN from $65 to zero in hours, dragging IRON down with it.

  • Neutrino USD2022-04

    Waves-chain algorithmic dollar backed by WAVES via a Terra-style mint/burn loop. Lost its peg after a reflexive crash in WAVES price and never recovered; the loop ran in reverse and never closed back.

  • Beanstalk v12022-04

    Credit-based algorithmic stablecoin governed by on-chain seigniorage. Killed by an $182M flash-loan governance attack that passed a malicious proposal in one block and drained the protocol's reserves.

  • TerraUSD2022-05

    The canonical algorithmic-stablecoin failure. UST went from $1 to ~$0.10 across May 9-13, 2022; LUNA went from above $80 to fractions of a cent in the same window; ~$40B of combined market value evaporated. The mechanism worked as designed under normal conditions and as critics had predicted under coordinated stress.

  • TerraKRW2022-05

    Terra's KRW-pegged sibling, swept away in the same death spiral as UST. Inherited every assumption of the LUNA-burn loop and broke at the same moment.

  • DEI2022-05

    Deus Finance's fractional algorithmic dollar, partly backed by USDC and partly by the protocol's DEUS governance token. Collapsed alongside UST in May 2022 when the algorithmic portion of backing lost confidence.

  • SpiceUSD2022-09

    SpiceTrade's algorithmic dollar. Lost its peg and faded out of trading without a single catastrophic event: a slow attrition under thin liquidity rather than a UST-style breakdown.

  • USN2022-10

    Near Protocol's algorithmic dollar. A double-mint bug left the protocol roughly $40M short of its claimed backing; the Near Foundation eventually wound USN down rather than recapitalize.

  • Bean (v2)2024-05

    Beanstalk's relaunched credit-based stablecoin. The second planting failed for the same structural reason as the first: protocol-issued credit cannot defend a $1 peg without an exogenous reserve.

  • Pinto2025-10

    A direct Beanstalk fork on Base, launched November 2024. Lost its peg in 2025 after the same supply/credit feedback loop drifted negative and refused to close. Same algorithm, same end.

Full obituaries, peak market caps, and post-mortems in the stablecoin cemetery.

Where the design fails

Known failure modes

Reflexive collapse / death spiral
Terra/UST went from $1 to roughly $0.10 over four days between May 9 and May 13, 2022; LUNA went from above $80 to fractions of a cent across the same window. Roughly $40 billion of combined market value evaporated. The mechanism worked exactly as designed under normal conditions, and exactly as critics had predicted under coordinated stress. UST is retained as a Pharos shadow asset and replays through the Stability Index changelog so the May 2022 event remains visible in PSI history.
Pure mint/burn arbitrage breakdown
In the historical pure mint/burn variant, the peg only holds when arbitrageurs believe newly minted governance tokens can be sold for at least the redemption value. Once that confidence breaks, the under-peg trade does not close. Collateralized variants instead depend on their actual reserves, issuance thresholds, or PSM capacity; FPI's FPIS dilution backstop is distinct from ZSD's reserve-ratio enforcement without new ZEPH issuance.
Recursive use as collateral
UST was deeply embedded across DeFi: Anchor's ~$14B deposit base, the Curve 4pool, lending markets, and synthetic asset protocols all carried direct UST exposure. When UST broke, every protocol holding it was simultaneously impaired. Algorithmic designs that allow themselves to be used as collateral multiply the contagion surface, a property the design family shares with no other archetype on Pharos.
Privileged mint authority retained
Many coins carrying this archetype retain uncapped or admin-controlled mint authority without an explicit holder-balance freeze surface. That is no longer a FreezeWatch tier; the Mint Authority module carries the supply-control review while FreezeWatch stays focused on freezes, blocks, wipes, pauses, and upstream value blocks.

Variations

Sub-flavors within the archetype

Pure mint/burn (uncollateralized)
The canonical UST model: burn governance token, mint stablecoin, with no segregated collateral. The cleanest version of the design, and the one with the most decisive failure mode. Not a live design at scale today; UST itself survives only as a Pharos shadow asset for PSI replay.
Fractional and conversion-based
FRAX v1 paired partial collateral with an algorithmic remainder before migrating to a fully collateralized model. HBD mints against a single protocol-native base coin under conversion rules with a debt-ratio haircut. These designs limit the death-spiral surface by capping issuance rather than relying purely on arbitrage.
CDP-plus-PSM hybrids
USDD 2.0 is structurally a CDP with a Peg Stability Module; Pharos classifies it under the CDP archetype rather than algorithmic. FPI's documented design targets a 100% collateral ratio and includes FPIS sales as a backstop; the target does not establish its observed reserve composition.

What to watch on Pharos

Signals that matter most

  1. 01

    Active depeg cap on the Safety Score. Algorithmic designs that trade meaningfully below peg are hard-capped at D (≥10% deviation) or F (≥25% deviation).

  2. 02

    DEWS supply-velocity signal on the Depeg Tracker. Algorithmic designs frequently trigger this sub-signal first; contraction in circulating supply is the on-chain symptom of the burn-and-redeem loop running against the peg.

  3. 03

    Mint Authority review. Most coins under this archetype retain privileged mint authority; sudden mint bursts and admin-key activity are supply-control risk rather than holder-freeze events.

  4. 04

    PegScore and the 7-day deviation chart on each coin page. The composite peg score caps lower for chronic deviators, and the chart shows whether a depeg actually closed or just went stale at a discount.

  5. 05

    PSI Top Contributors table. Chronic depegs decay through the PSI factor floor (0.25 after roughly 120 days), so a long-running algorithmic depeg is downweighted in the system-wide stability index rather than amplified.

  6. 06

    Live Reserve view where available. FPI's composition should be read from the dated reserve evidence, separately from its collateral-ratio target. Quarterly snapshots are too coarse for this design family.

Tracked universe

4 tracked stablecoins in this archetype

Case studies

When this mechanism met a stress test

Continue reading