8.1 Stablecoins: Mechanisms and Stability
The $180 Billion Stability Problem
On May 9, 2022, at 11:00 PM UTC, Terra's UST stablecoin traded at $0.998—just 0.2% from its $1.00 peg. Thirty-six hours later, UST traded at $0.30. Within a week: $0.10. By month's end: $0.01.
The cascade:
May 7: UST = $1.00, LUNA = $80
May 9: Large UST sells begin, peg wobbles to $0.98
May 10: Panic spreads, UST → $0.60, LUNA → $30
May 11: Death spiral accelerates, UST → $0.30, LUNA → $5
May 12: System collapse, UST → $0.10, LUNA → $0.50
May 13: Both tokens effectively worthless
Total value destroyed: $60 billion in market cap evaporated—one of the largest financial collapses in history. Hundreds of thousands of retail investors lost life savings. The DeFi ecosystem entered a year-long bear market.
Yet paradoxically, stablecoins have become the most successful crypto product:
Market size (2024):
Total stablecoin supply: $180B+
Daily trading volume: $100B+
Number of unique holders: 50M+
By token:
USDT (Tether): $110B (61%)
USDC (Circle): $35B (19%)
DAI (MakerDAO): $5B (3%)
Others: $30B (17%)
Why stablecoins matter:
1. Bridge to crypto: Entry point for new users 2. Store of value: Escape volatility without leaving crypto 3. Trading pairs: 80% of CEX volume is vs stables 4. DeFi backbone: Required for lending, derivatives, payments 5. Global payments: Fast, cheap cross-border transfers
But maintaining stability is harder than it looks.
This lesson explores the fundamental tension in stablecoin design: the stability trilemma.
Stability
▲
/ \
/ \
/ \
/ \
/ \
Decentralization ◄─────► Capital Efficiency
Pick two:
- Stable + Decentralized = Capital inefficient (DAI ~150% collateral)
- Stable + Capital efficient = Centralized (USDC: need Circle's reserves)
- Decentralized + Capital efficient = Unstable (Terra UST: death spiral)
We'll explore:
- Why maintaining a peg is an economic, not just technical, problem
- The mathematics of collateralization and leverage
- Game theory of stablecoin attacks and bank runs
- Why algorithmic stablecoins keep failing (and will likely continue to)
- The stability-decentralization-efficiency trilemma
- Whether truly stable, decentralized, capital-efficient stablecoins are possible
Understanding stablecoins is essential because they're simultaneously DeFi's greatest success and greatest vulnerability. Let's explore why.
What Are Stablecoins?
The Volatility Problem
Cryptocurrency volatility makes them poor money:
Bitcoin daily volatility: ~4%
Ethereum daily volatility: ~5%
Traditional currency (USD): ~0.3%
Consequence: Can't use for:
- Pricing goods (price changes hourly)
- Storing value (loses 50% in weeks)
- Receiving salary (value uncertain)
- Business contracts (unhedgeable risk)
Example of the problem:
Scenario: Alice runs online store
Monday: Customer orders $100 product, pays 0.05 ETH
(ETH = $2,000)
Wednesday: Alice's supplier bill is due: $90
But ETH now = $1,800
0.05 ETH = $90 ✓ Just breaks even
Friday: Alice wants to buy inventory
ETH now = $1,600
0.05 ETH = $80 ✗ Lost $10 (10% of revenue!)
Result: Can't run business on volatile currency
Solution needed: Currency with crypto benefits but stable value.
Definition: What Makes a Stablecoin?
A stablecoin is a cryptocurrency designed to maintain a stable value relative to a reference asset (typically $1 USD).
Key properties:
1. Stability: Maintains peg within tolerance (±1-2% typical)
2. Liquidity: Can be bought/sold at peg price
3. Convertibility: Mechanism to maintain peg
4. Redeemability: Can be exchanged for reference asset (or equivalent)
Stability mechanisms:
All stablecoins need mechanisms to restore peg when it deviates:
Price > $1.00 (premium):
→ Increase supply (sell/mint stablecoins)
→ Decreases price back to $1.00
Price < $1.00 (discount):
→ Decrease supply (buy/burn stablecoins)
→ Increases price back to $1.00
The devil is in the details: How do you increase/decrease supply reliably?
Classification of Stablecoins
Three main categories:
1. Fiat-Collateralized (Custodial)
Backed by real USD in bank account
Examples: USDT, USDC, BUSD
Collateral ratio: ~100%
2. Crypto-Collateralized (Over-collateralized)
Backed by crypto assets (ETH, BTC)
Examples: DAI, sUSD, LUSD
Collateral ratio: 150-300%
3. Algorithmic (Under/uncollateralized)
Stabilized by algorithmic supply adjustment
Examples: UST (failed), FRAX (hybrid), UXD
Collateral ratio: 0-100%
The fundamental trade-off:
Fiat-collateralized:
✓ Most stable
✓ Most capital efficient (1:1 backing)
✗ Centralized (requires trusted custodian)
✗ Regulatory risk
✗ Censorship possible
Crypto-collateralized:
✓ Decentralized (trustless)
✓ Transparent (on-chain collateral)
✗ Capital inefficient (need 150-300% collateral)
✗ Less stable (collateral price volatility)
Algorithmic:
✓ Capital efficient (minimal/no collateral)
✓ Potentially decentralized
✗ Unstable (death spiral risk)
✗ Game theory vulnerabilities
✗ Track record: 100% failure rate for pure algorithmic
Historical Context
Evolution of stablecoins:
2014: Tether (USDT) launches
First fiat-collateralized stablecoin
Supply: $1M → $110B (2024)
2017: MakerDAO launches (DAI)
First crypto-collateralized stablecoin
Supply: $0 → $5B (2024)
2018: USDC launches (Circle/Coinbase)
"Compliant" fiat-collateralized
Supply: $0 → $35B (2024)
2020: Basis Cash, Iron Finance (algorithmic)
Both failed (death spirals)
2021: Terra's UST launches
Largest algorithmic attempt
Peak: $18B supply
2022: UST collapses
$60B market cap destroyed
Algorithmic stablecoins discredited
2023-2024: Consolidation
USDT, USDC, DAI dominate
Remaining algorithmic coins <$2B total
Survival rates:
Fiat-collateralized: 95%+ survive
(Failures mostly regulatory issues, not depeg)
Crypto-collateralized: 80% survive
(Failures from oracle issues, exploits)
Algorithmic: 5% survive
(Near-universal failure, death spirals)
Fiat-Collateralized Stablecoins
Mechanism: Trust-Based Backing
Design:
User → Deposit $1 USD → Company → Mint 1 USDC
Wire transfer (Circle) On-chain
User → Redeem 1 USDC → Company → Withdraw $1 USD
Burn token Custody Wire transfer
Key components:
1. Custody:
Circle (USDC): Holds USD in bank accounts + T-bills
Tether (USDT): Holds USD + T-bills + other reserves
Binance (BUSD): Held by Paxos (now deprecated)
Custodian responsibilities:
- Hold reserves
- Process deposits/withdrawals
- Monthly attestations
- Regulatory compliance
2. Minting/burning:
Minting: Only issuer can mint (Circle for USDC)
Burning: Anyone can burn to redeem
Supply = Outstanding deposits
Example:
Deposits: $1B → Mint 1B USDC
Redemptions: $100M → Burn 100M USDC
Supply: 900M USDC
3. Peg maintenance:
Market price > $1.01:
→ Arbitrageurs deposit USD, mint USDC
→ Sell USDC for $1.01
→ Profit $0.01 per coin
→ Supply increases, price falls to $1.00
Market price < $0.99:
→ Arbitrageurs buy USDC for $0.99
→ Redeem for $1.00 USD
→ Profit $0.01 per coin
→ Supply decreases, price rises to $1.00
Arbitrage ensures peg (if issuer is solvent).
USDC: Case Study
Circle's USDC mechanism:
Reserve composition (2024):
Total reserves: $35B
Breakdown:
- Cash in banks: $10B (29%)
- U.S. Treasury bills: $25B (71%)
- Other: $0B (0%)
Monthly attestations by Grant Thornton (auditor)
Publicly viewable at circle.com/usdc
Minting/redemption process:
Minting (institutional):
1. Register with Circle (KYC)
2. Wire $1M+ USD
3. Receive USDC within 1 business day
4. Minimum: $100k
Redemption (institutional):
1. Send USDC to Circle
2. Receive USD within 1 business day
3. Minimum: $100k
Retail users:
- Can't mint/redeem directly
- Trade on exchanges (price ~$1.00 due to arbitrage)
Stability analysis:
USDC price history (2020-2024):
- 99.5% of time: $0.995 - $1.005 (±0.5%)
- 99.9% of time: $0.99 - $1.01 (±1%)
- Max deviation: $0.88 (March 2023, SVB crisis)
SVB crisis (March 2023):
March 10: Silicon Valley Bank fails
Circle has $3.3B stuck in SVB
March 11: USDC depegs to $0.88 (panic)
March 12: Fed backstops SVB deposits
USDC recovers to $0.97
March 13: Full recovery to $1.00
Lesson: Even "safe" fiat stablecoins have centralization risk
Advantages:
+ Very stable (±0.5% typical)
+ Capital efficient (1:1 backing)
+ Fast minting/redemption
+ Regulatory compliant (licenses in multiple jurisdictions)
+ High liquidity (major exchanges)
Disadvantages:
- Centralized (Circle controls everything)
- KYC required for direct minting/redemption
- Censorship possible (Circle can blacklist addresses)
- Bank risk (SVB crisis showed vulnerability)
- Regulatory risk (could be shut down)
- Not trustless (must trust Circle's reserves)
Tether (USDT): Controversy and Dominance
Despite controversies, USDT is the largest stablecoin ($110B).
Reserve composition (claimed, 2024):
Total reserves: $110B
Breakdown:
- Cash & cash equivalents: ~$5B (5%)
- U.S. Treasury bills: ~$90B (82%)
- Other investments: ~$8B (7%)
- Corporate bonds: ~$3B (3%)
- Secured loans: ~$2B (2%)
- Precious metals: ~$2B (2%)
Controversies:
2019: Attorney General investigation
Claims Tether used reserves for other purposes
Settlement: $18.5M fine
2021: Misleading reserve statements
Claimed "100% backed" but included commercial paper
Settlement: $41M fine (CFTC)
2023: Still no full audit
Only attestations (weaker than audits)
Community skepticism remains
Yet: USDT maintains peg better than most
Trading at $0.998-$1.002 consistently
Why USDT dominates despite issues:
1. First mover advantage (launched 2014)
2. Highest liquidity (most trading pairs)
3. Available everywhere (even unregulated exchanges)
4. No KYC for secondary market trading
5. Network effects (everyone uses it)
Result: "Too big to fail" status
If USDT collapsed, entire crypto market affected
The Tether paradox:
Most controversial stablecoin
+
Most used stablecoin
=
Systemic risk to crypto
Regulatory Considerations
Legal status varies by jurisdiction:
United States:
Not explicitly regulated yet (as of 2024)
Proposed legislation:
- Stablecoin Transparency Act
- Digital Asset Market Structure Act
Would require:
- 100% reserves
- Regular audits
- Federal oversight
- Deposit insurance (possibly)
European Union:
MiCA (Markets in Crypto Assets) 2024:
- Requires authorization
- Reserve requirements
- Redemption rights
- Daily volume limits
- Quarterly audits
Some U.S. stablecoins may not meet requirements
The regulatory trend:
2020-2022: Light regulation, rapid growth
2023-2024: Increasing scrutiny, compliance requirements
2025+: Likely heavy regulation, consolidation
Outcome: Favors compliant issuers (Circle, Paxos)
Challenges offshore issuers (Tether?)
Crypto-Collateralized Stablecoins
The Decentralization Solution
Problem with fiat-collateralized:
Requires trusting a centralized entity
What if you want:
- No KYC/censorship
- Transparent reserves
- Permissionless participation
- No regulatory risk
Solution: Collateralize with crypto assets
User → Deposit 1.5 ETH → Smart Contract → Mint 1,000 DAI
(Worth $3,000) MakerDAO (Worth $1,000)
Over-collateralized: 300% collateral ratio
Why? ETH is volatile, need safety buffer
Key innovation: Trustless liquidation
If ETH price drops:
Collateral value falls below threshold
→ Liquidation bot repays debt
→ Seizes collateral
→ System stays solvent
No trust required, all algorithmic
MakerDAO and DAI: Deep Dive
Maker protocol architecture:
Components:
1. Vaults (formerly CDPs):
Users create "Vaults":
- Deposit collateral (ETH, WBTC, USDC, etc.)
- Mint DAI (stablecoin)
- Pay stability fee (interest)
- Can add collateral or repay DAI anytime
Example:
Deposit: 10 ETH at $2,000 = $20,000
Mint: 10,000 DAI
Collateralization: 200%
Minimum: 150% for ETH
Liquidation if drops below 150%
2. Collateral types:
Multiple collateral types supported:
ETH-A: 150% ratio, 0.5% stability fee
ETH-B: 130% ratio, 2% stability fee (riskier)
WBTC-A: 150% ratio, 0.5% fee
USDC-A: 103% ratio, 0% fee (stablecoin collateral)
Each has independent risk parameters
3. Stability mechanisms:
Stability fee:
Annual interest rate paid by borrowers
Accrues continuously
Paid when repaying DAI
Purpose: Control DAI supply
High fee → Expensive to mint → Supply decreases
Low fee → Cheap to mint → Supply increases
Adjustable by Maker governance
Typical range: 0-10%
DAI Savings Rate (DSR):
Interest paid to DAI holders who lock in DSR contract
Funded by stability fees
Purpose: Control DAI demand
High DSR → Attractive to hold → Demand increases → Price increases
Low DSR → Less attractive → Demand decreases → Price decreases
Typical range: 0-5%
4. Liquidations:
When vault collateral ratio < 150%:
1. Collateral auctioned off
2. DAI debt repaid
3. Liquidation penalty charged (13% typical)
4. Remaining collateral returned to owner
Dutch auction mechanism:
- Starting price high
- Decreases over time
- First liquidator to bid wins
Example:
Vault has: $15,000 collateral, $10,500 DAI debt
Ratio: 142.8% < 150% threshold
Auction:
- Debt to cover: $10,500
- Penalty: 13% = $1,365
- Total to raise: $11,865
- Collateral available: $15,000
Liquidator pays $11,865 DAI
Receives $11,865 worth of collateral
Owner keeps: $15,000 - $11,865 = $3,135 collateral
Mathematical Model: Collateralization Ratios
Stability condition:
For DAI to maintain $1 peg, the system must be solvent:
Total Collateral Value ≥ Total DAI Debt
Let:
C = total collateral value (in USD)
D = total DAI supply
R = collateralization ratio
Requirement: C ≥ D, or R = C/D ≥ 1
Actually need buffer: R > 1 + safety_margin
Typical: R = 1.5 to 3.0 (150-300%)
Why overcollateralization?
ETH volatility: ~80% annualized
Daily moves: ±5% common
Flash crashes: -20% possible
Need buffer so that:
Even if ETH drops 30% in a day
System stays solvent before liquidations execute
Example:
Initial: $2M collateral, $1M DAI, R = 200%
ETH drops 30%: $1.4M collateral, $1M DAI, R = 140%
Still solvent, but underwater vaults liquidate
After liquidations: System back to healthy R
Relationship: Collateral ratio and maximum DAI supply
Given collateral C and minimum ratio R_min:
Max DAI supply = C / R_min
Example:
Total ETH collateral in Maker: 4M ETH
ETH price: $2,000
Total collateral value: $8B
Minimum ratio: 150%
Max DAI: $8B / 1.5 = $5.33B
Actual DAI supply: ~$5B (2024)
Utilization: 94%
Capital efficiency:
Capital efficiency = DAI issued / Collateral value
Fiat-collateralized (USDC): 100%
$1B reserves → $1B USDC
Crypto-collateralized (DAI): 33-66%
$1B ETH → $330M - $660M DAI (at 150-300% ratios)
This is the cost of decentralization
DAI Stability Analysis
Historical peg performance:
2019-2021: Very stable
- 99% of time: $0.98 - $1.02
- Max deviation: $1.04 (brief spikes)
2022-2023: More volatile
- 80% of time: $0.97 - $1.03
- Max deviation: $1.08 (USDC depeg contagion)
Why less stable recently?
1. More USDC collateral (centralization)
2. Lower ETH collateral (less decentralized)
3. Market stress (UST collapse, USDC depeg)
Peg recovery mechanisms:
DAI > $1.00 (premium):
Actions:
1. Lower stability fee → Cheaper to mint → Supply increases
2. Lower DSR → Less attractive to hold → Demand decreases
3. Both push price down to $1.00
Example:
DAI = $1.02
Governance lowers stability fee: 3% → 1%
More users mint DAI (cheaper now)
DAI supply increases by 10%
Price falls: $1.02 → $1.00
DAI < $1.00 (discount):
Actions:
1. Raise stability fee → More expensive to mint → Supply decreases
2. Raise DSR → More attractive to hold → Demand increases
3. Both push price up to $1.00
Example:
DAI = $0.98
Governance raises DSR: 1% → 5%
Users lock DAI in DSR to earn 5%
DAI demand increases
Price rises: $0.98 → $1.00
The USDC dilemma:
Problem: DAI increasingly backed by USDC
2020:
- ETH collateral: 80%
- USDC collateral: 5%
- Decentralized: ✓
2024:
- ETH collateral: 30%
- USDC collateral: 50%
- Decentralized: Questionable
Reason: USDC more capital efficient (101% ratio vs 150% for ETH)
Market prefers capital efficiency over decentralization
Result: DAI stability comes at cost of centralization
"Decentralized" stablecoin backed by centralized stablecoin
Other Crypto-Collateralized Stablecoins
Liquity (LUSD):
Immutable protocol (no governance)
ETH-only collateral
110% minimum collateralization (more efficient than DAI!)
Redemption mechanism maintains hard peg
Mechanism:
Can always redeem 1 LUSD for $1 of ETH from riskiest vaults
This creates hard floor price at $1.00
Supply: $200M (2024)
Stability: Excellent ($0.998-$1.002)
sUSD (Synthetix):
Collateralized by SNX token
400-500% collateralization (very conservative!)
Part of Synthetix synthetic asset platform
Supply: $50M (2024)
Stability: Good when SNX stable, poor when SNX volatile
Comparison:
Property DAI LUSD sUSD
─────────────────────────────────────────────────────
Collateral Multi ETH-only SNX-only
Min ratio 150% 110% 400%+
Governance Yes No Yes
Supply $5B $200M $50M
Stability Good Excellent Fair
Decentralization Decreasing High Medium
Algorithmic Stablecoins: The Failed Dream
The Vision
Goal: Stablecoin without collateral (or minimal collateral)
Fiat-collateralized: Need $1B to back $1B stablecoin
Crypto-collateralized: Need $1.5-3B to back $1B stablecoin
Algorithmic: Need $0 to back $1B stablecoin (?)
If possible → Infinite capital efficiency!
Core idea: Algorithmic supply adjustment
Price > $1.00:
→ Mint new stablecoins
→ Supply increases
→ Price falls to $1.00
Price < $1.00:
→ Burn stablecoins
→ Supply decreases
→ Price rises to $1.00
The problem: Who would burn stablecoins at a loss?
Terra/UST: The Biggest Failure
Mechanism: Dual-token system
Two tokens:
UST: Stablecoin (target $1.00)
LUNA: Volatile governance token
Mint/burn relationship:
$1 UST can always be minted by burning $1 of LUNA
$1 LUNA can always be minted by burning 1 UST
Example:
UST = $0.98 (below peg)
Action: Burn 1 UST → Mint $1 LUNA
Profit: Buy UST at $0.98, burn for $1 LUNA, profit $0.02
Result: UST supply decreases, price rises to $1.00
Why it worked (temporarily):
2021: Crypto bull market
LUNA price: $1 → $80
UST supply: $0 → $18B
20% yield on UST (via Anchor protocol)
Positive feedback loop:
High LUNA price → Confidence in UST
High UST demand → Need to mint more UST
Minting UST → Burning LUNA (supply decreases)
Lower LUNA supply → Price increases
Higher LUNA price → More confidence
(Repeat)
The death spiral:
May 7, 2022:
Large UST holder sells $285M UST on Curve
UST depegs to $0.98
May 8:
Panic spreads, more sell UST
UST = $0.95
Arbitrageurs burn UST for LUNA
LUNA supply increases (dilution)
LUNA price falls: $80 → $60
May 9:
Death spiral accelerates:
UST = $0.70 → More burning → More LUNA minted
LUNA = $40 → Lower LUNA market cap
Can't absorb UST supply → UST falls more
May 10:
UST = $0.30, LUNA = $5
Spiral is irreversible
Market cap of LUNA < Value of UST that needs backing
May 11-13:
Both tokens collapse to near zero
UST = $0.01, LUNA = $0.0001
$60B in value destroyed
Mathematical analysis of death spiral:
Let:
P_UST = UST price
P_LUNA = LUNA price
S_UST = UST supply
M_LUNA = LUNA market cap
Stability condition:
P_UST = $1 maintained if:
Demand to burn UST = Supply of UST being sold
When P_UST < $1:
Arbitrage burns UST → Mints LUNA
LUNA supply increases by: S_UST × (1 - P_UST) / P_LUNA
LUNA dilution: δ_LUNA = S_UST × (1 - P_UST) / M_LUNA
Example (pre-collapse):
S_UST = $18B
P_UST = $0.95 (5% depeg)
M_LUNA = $30B
LUNA dilution = $18B × 0.05 / $30B = 3%
Manageable, LUNA absorbs
Example (during collapse):
S_UST = $18B
P_UST = $0.70 (30% depeg)
M_LUNA = $10B (LUNA already falling)
LUNA dilution = $18B × 0.30 / $10B = 54%
Unsustainable! LUNA can't absorb this
→ LUNA price crashes further
→ Makes problem worse
→ Death spiral
The feedback loop:
UST depeg
↓
LUNA minting (to restore peg)
↓
LUNA price falls (dilution)
↓
Market cap falls
↓
Reduced ability to absorb UST sells
↓
UST depegs more
(Loop back to top)
Why death spiral was inevitable:
Fundamental flaw:
System relies on LUNA having value
LUNA value comes from belief in system
If confidence breaks → LUNA value evaporates
→ System can't maintain peg
→ Confidence breaks further
→ Self-fulfilling collapse
Game Theory: Self-Fulfilling Attacks
Based on currency crisis models (Obstfeld 1996):
Setup:
N users hold UST
Each user i can:
- Hold (maintains peg)
- Sell (breaks peg if enough others sell)
Payoffs:
If <50% sell: Peg holds, everyone wins
If >50% sell: Peg breaks, sellers escape, holders lose
Classic coordination game
Multiple equilibria:
Equilibrium 1 (Good):
Everyone believes peg will hold
→ No one sells
→ Peg does hold
→ Beliefs confirmed
Stable equilibrium
Equilibrium 2 (Bad):
Everyone believes peg will break
→ Everyone sells
→ Peg does break
→ Beliefs confirmed
Also stable equilibrium!
The coordination problem:
Each user reasons:
"If I think others will sell, I should sell too (escape first)
If I think others will hold, I should hold (benefit from stability)"
This creates two self-fulfilling prophecies:
1. Confidence → Stability → More confidence
2. Fear → Collapse → More fear
Small shock can flip from equilibrium 1 to 2
Application to Terra:
May 7: Small depeg ($0.98)
In "good" equilibrium, this would be absorbed
May 8: Fear spreads (large holder sold, maybe more selling coming?)
Users coordinate on "bad" equilibrium
Rush to exit
May 9-13: Self-fulfilling collapse
Everyone sells because everyone else is selling
Post-mortem: No single cause
Coordination failure triggered by fear
Why Algorithmic Stablecoins Keep Failing
The fundamental problem: Reflexivity
Stablecoin value depends on confidence
Confidence depends on stablecoin value
Circular dependency creates instability
In traditional stablecoins:
Value backed by external assets (USD, ETH)
Breaking confidence doesn't destroy backing
In algorithmic stablecoins:
Value backed by future confidence
Breaking confidence destroys backing
Death spiral ensues
List of failed algorithmic stablecoins:
2018: Basis (closed before launch, regulatory)
2020: Empty Set Dollar (ESD) - failed
2021: Iron Finance (IRON) - death spiral in 48 hours
2021: Fei Protocol (FEI) - abandoned peg mechanism
2022: Terra (UST) - catastrophic collapse
2022: DEI - depegged
2023: Many others (most unknown)
Success rate: ~0%
Common failure modes:
1. Death spiral (UST, IRON):
Depeg → Minting → Dilution → Further depeg
2. Low adoption (ESD, Basis Cash):
Need critical mass for stability
Without users, no stabilization mechanism works
3. Bank run (all):
Coordination on "bad" equilibrium
Self-fulfilling panic
4. Oracle manipulation:
Attacker manipulates price oracle
Exploits mint/burn mechanism
5. Governance attack:
Attacker gains control
Changes parameters to exploit system
Why the dream persists despite failures:
Potential benefits are enormous:
- Infinite capital efficiency
- True decentralization
- No external dependencies
- Scalable to any size
But benefits remain theoretical
All real-world attempts have failed
Likely a mathematical impossibility
Hybrid Approaches and Innovations
FRAX: Partially Collateralized
Mechanism: Fractional collateralization
FRAX = α × Collateral + (1-α) × Algorithmic
Where α = collateral ratio (dynamic)
Example at α = 90%:
To mint $100 FRAX:
- Provide $90 USDC (collateral)
- Burn $10 FXS (algorithmic token)
Total: $100 value backs $100 FRAX
Dynamic collateral ratio:
If FRAX > $1.00:
System decreases α (less collateral needed)
Makes minting easier, increases supply
If FRAX < $1.00:
System increases α (more collateral needed)
Makes minting harder, decreases supply
α adjusts by 0.25% per hour
Range: 50-100% typically
Why hybrid might work better:
Partially collateralized:
- Some safety from collateral
- Some efficiency from algorithmic
- Middle ground between extremes
When α = 100%: Fully collateralized (like USDC)
When α = 0%: Fully algorithmic (like UST - dangerous)
FRAX stays in safe middle ground
Currently: α ≈ 90-95%
Track record:
Launch: December 2020
Major test: May 2022 (UST collapse)
Performance: Held peg at $0.97-$1.00
Supply: $650M (2024)
Best performing algorithmic/hybrid so far
But still <10% crypto-collateralized (DAI = $5B)
RAI: Unpegged "Stablecoin"
Radical idea: Don't peg to $1, float with dampening
RAI doesn't target $1.00
Instead: Target "redemption price" that adjusts
Redemption price moves slowly based on market price
If RAI trading at premium:
→ Redemption price increases
→ Borrowing becomes more expensive
→ Supply increases
→ Price falls
If RAI trading at discount:
→ Redemption price decreases
→ Borrowing becomes cheaper
→ Supply decreases
→ Price rises
PID controller:
Borrowed from control theory
Proportional-Integral-Derivative controller
Adjusts redemption rate based on:
P: Current deviation from target
I: Accumulated past deviations
D: Rate of change of deviation
Result: Smooth adjustments, reduced volatility
But doesn't target $1.00 precisely
Performance:
Launch: Feb 2021
Redemption price: Started at $3.14, now ~$3.00
Market price: Tracks redemption price within ±2%
Supply: $10M (2024) - very small
Use case: Unclear (not good for payments if value drifts)
Interesting experiment, limited adoption
Ethena (USDe): Delta-Neutral Stablecoin
2024 innovation: Fully collateralized via derivatives
Mechanism:
1. Hold $1 of ETH
2. Short $1 of ETH perpetual futures
3. Net position: Delta-neutral
4. Earn funding rate (typically positive)
5. Issue $1 USDe backed by this position
Result:
- Fully collateralized by ETH
- Market-neutral (no ETH price exposure)
- Generates yield from funding rates
Why it works:
ETH exposure: +$1 (spot)
ETH exposure: -$1 (futures)
Net: $0 (delta neutral)
If ETH ↑ to $1.10:
Spot: +$0.10 gain
Futures: -$0.10 loss
Net: $0 (collateral stays $1)
USDe stays backed 1:1
No need for overcollateralization!
Risks:
1. Funding rate risk:
If funding rate goes negative (rare)
Position loses money
Need buffer collateral
2. Counterparty risk:
Futures contracts with exchanges
Exchange failure = loss of hedge
3. Liquidation risk:
If futures position liquidated
Delta neutrality breaks
4. Operational risk:
Must actively manage positions
Rebalance as needed
Track record:
Launch: July 2023
Supply: $3.5B (April 2024) - rapid growth!
Stability: Excellent ($0.998-$1.002)
Yield: 8-15% APY from funding rates
Most successful new stablecoin launch since UST
But much safer design (fully collateralized)
The Stablecoin Trilemma
Three Desirable Properties
Following Klages-Mundt et al. "While Stability Lasts":
1. Decentralization
- No central authority
- Censorship-resistant
- Trustless operation
- Open participation
2. Capital Efficiency
- Minimal collateral required
- Low opportunity cost
- Scalable supply
- 1:1 or better backing
3. Stability
- Maintains peg reliably
- Low volatility
- Quick peg recovery
- Resistant to attacks
The Impossibility Result
Theorem (informal): You can have at most two of three.
Decentralized + Capital Efficient = Unstable
Examples: Terra UST, Basis Cash
Result: Death spirals, bank runs
Decentralized + Stable = Capital Inefficient
Examples: DAI, LUSD
Result: Need 150-300% collateral
Capital Efficient + Stable = Centralized
Examples: USDC, USDT
Result: Need trusted custodian
Why the trilemma exists:
Stability requires backing:
- Either real assets (centralized)
- Or excess crypto collateral (inefficient)
- Or algorithmic promises (unstable)
Decentralization requires trustless operation:
- Can't have trusted custodian
- Must use crypto collateral or algorithms
- Crypto is volatile → need overcollateral
- Algorithms are reflexive → unstable
Capital efficiency requires minimal collateral:
- Can't afford large buffers
- Must be centralized (to maintain confidence)
- Or algorithmic (which fails)
Visualizing the Trade-off Space
Stability (Peg Maintenance)
▲
│
│
USDC ● USDT
│ (Centralized,
│ stable,
│ efficient)
│
│
│ DAI
│ ● (More centralized now)
│
│ LUSD
│ ● (Decentralized,
│ stable,
──────┼───────────────────────────► Capital Efficiency
│ inefficient)
│
│
│ UST (failed)
│ ● (Tried to be all three,
│ achieved none)
│
Decentralization
Current state of major stablecoins:
Stablecoin Decentralization Capital Efficiency Stability
──────────────────────────────────────────────────────────────
USDC ★☆☆☆☆ ★★★★★ ★★★★★
USDT ★☆☆☆☆ ★★★★★ ★★★★☆
DAI ★★☆☆☆ ★★☆☆☆ ★★★★☆
LUSD ★★★★☆ ★★★☆☆ ★★★★★
FRAX ★★☆☆☆ ★★★★☆ ★★★★☆
RAI ★★★★☆ ★★☆☆☆ ★★★☆☆
UST ★★★☆☆ ★★★★★ ☆☆☆☆☆ (failed)
Can We Escape the Trilemma?
Possible paths forward:
1. Accept trade-offs:
Different stablecoins for different use cases:
- USDC for stability (payments, savings)
- DAI for decentralization (censorship resistance)
- Neither is perfect, both are useful
2. Minimize trade-offs:
Gradual improvements:
- DAI working to re-decentralize (less USDC backing)
- Better oracle technology (reduces needed collateral)
- More efficient capital deployment
3. Novel mechanisms:
Ethena (USDe): Delta-neutral derivatives
- Achieves full collateralization without overcollateral
- But introduces new risks (funding rates, counterparties)
Future: Zero-knowledge proofs of reserves?
Better algorithmic models?
Combination of multiple approaches?
4. Regulation as partial solution:
If all fiat-collateralized stablecoins:
- Must hold 100% reserves
- Must allow redemptions
- Must be audited regularly
Then: Centralization becomes less risky
Trust replaced by verification
But: Still not permissionless/censorship-resistant
The Optimal Point?
No consensus, depends on priorities:
For payments/commerce:
Stability > Capital Efficiency > Decentralization
Winner: USDC
For DeFi collateral:
Stability ≈ Decentralization > Capital Efficiency
Winner: DAI or LUSD
For speculation/trading:
Stability > Capital Efficiency > Decentralization
Winner: USDT (liquidity matters)
For ideology/censorship resistance:
Decentralization > Stability > Capital Efficiency
Winner: RAI or future innovations
Systemic Risks and Future Outlook
Concentration Risk
Current market share:
Top 3 stablecoins: $150B / $180B total = 83%
If any of top 3 fail:
- Would affect entire crypto ecosystem
- Billions in losses
- Potential contagion to other stablecoins
- Regulatory crackdown likely
Tether systemic risk:
$110B USDT outstanding
Used on nearly every exchange
Most trading pairs denominated in USDT
If Tether collapsed:
1. Immediate: $110B market cap gone
2. Secondary: Can't trade on many exchanges
3. Tertiary: Panic across all crypto
4. Result: Potential crypto winter
"Too big to fail" but also
"Too central to allow to fail"
Bank Run Dynamics
From "While Stability Lasts" paper:
Model setup:
N users hold stablecoin
Stablecoin partially backed (or algorithmic)
Each user can:
- Hold and earn yield: u = r (interest rate)
- Redeem early: u = v (liquidation value)
If too many redeem:
v < 1 (insufficient backing)
Later redeemers get less
Strategic game: When to redeem?
Equilibrium analysis:
Multiple equilibria possible:
Equilibrium 1 (No run):
- Each user believes others will hold
- No one redeems early
- Stablecoin maintains peg
- Everyone earns r
Equilibrium 2 (Bank run):
- Each user believes others will redeem
- Everyone redeems immediately
- Stablecoin depegs
- Later redeemers get v < 1
Transition: Small shock can flip equilibria
Practical implications:
All partially-backed stablecoins vulnerable:
- Algorithmic stablecoins (high risk)
- Fractional stablecoins like FRAX (medium risk)
- Even crypto-collateralized if illiquid collateral
Only immune: 100% liquid, audited reserves (USDC, theoretically)
But even USDC suffered mini-run (SVB crisis)
→ No stablecoin is truly immune
Regulatory Future
Likely regulatory trajectory:
2024-2025: U.S. framework emerges
Proposed requirements:
- 100% reserves (cash + T-bills)
- Monthly audits (not attestations)
- Federal oversight (OCC or Fed)
- Deposit insurance (potentially)
- Redemption rights guaranteed
Impact on different types:
Fiat-collateralized (USDC):
✓ Can comply easily
✓ May benefit from clarity
✗ Higher costs (insurance, audits)
Crypto-collateralized (DAI):
? Unclear status
May be exempt (decentralized)
Or may need to register as something else
Algorithmic:
✗ Likely banned or heavily restricted
"Unbacked" = securities? commodities?
Regulatory hostility after UST
Global fragmentation:
Different jurisdictions, different rules:
US: Strict requirements
EU: MiCA regulations
Asia: Varies by country
Offshore: Permissive
Result: Stablecoin fragmentation
US-regulated vs offshore vs truly decentralized
The Future of Stablecoins
Predictions for 2025-2030:
1. Consolidation:
Current: 50+ stablecoins
Future: 10-15 major ones
Winners likely:
- USDC (compliance winner)
- USDT (too big to fail)
- DAI (decentralization winner)
- 2-3 new innovations
Losers: Small algorithmic coins, compliance failures
2. Institutional adoption:
Central Bank Digital Currencies (CBDCs):
- Many countries launching
- May compete with private stablecoins
- Or may coexist
Stablecoins become regulated infrastructure
Similar to money market funds
Boring but essential
3. DeFi integration deepens:
More sophisticated use cases:
- Cross-chain stablecoins (native on multiple chains)
- Yield-bearing stablecoins (default interest)
- Programmable stablecoins (smart contract enabled)
- Privacy stablecoins (ZK-SNARK transfers)
4. Better mechanisms:
Innovations in collateral:
- Real-world assets (RWA) as collateral
- Delta-neutral strategies (like Ethena)
- Hybrid approaches (FRAX evolution)
- Better oracle technology
5. Market maturation:
$180B (2024) → $500B+ (2030)
- More institutional use
- More countries adopting
- More payment use cases
- Integration with TradFi
Open Questions
Can we escape the trilemma?
Theory says no
Practice may find approximations
Likely: Multiple stablecoins, each optimizing differently
Will algorithmic stablecoins ever work?
Track record: 100% failure rate
Theory: Reflexivity problem seems fundamental
Verdict: Probably impossible without some collateral
What's the role for decentralized stablecoins?
If regulation makes fiat-collateralized reliable,
Do we still need crypto-collateralized?
Yes, for:
- Censorship resistance
- Permissionless access
- Trustless operation
- DeFi composability
But market share may remain small (<20%)
Will stablecoins remain stable?
Fiat-collateralized: Yes (if properly regulated)
Crypto-collateralized: Mostly (some volatility)
Algorithmic: No (death spirals inevitable)
Overall: Yes, but with caveats
Not all stablecoins are created equal
Conclusion: The Stability Paradox
Stablecoins are crypto's greatest success and greatest vulnerability.
Success:
- $180B market cap (larger than many fiat currencies)
- 50M+ users globally
- Essential infrastructure for DeFi
- Proven product-market fit
Vulnerability:
- Concentration in 2-3 tokens
- Systemic importance (too big to fail)
- Impossible trilemma (can't have it all)
- Regulatory uncertainty
Key lessons:
1. Stability requires backing:
No free lunch: Either hold real assets, crypto assets, or face instability
Algorithmic stablecoins are mathematical impossibility
2. Decentralization has costs:
Crypto-collateralized: 150-300% overcollateralization
Capital inefficiency is the price of trustlessness
3. Centralization has risks:
Fiat-collateralized: Trust required
But verification can mitigate (audits, regulation)
4. Multiple solutions needed:
No single stablecoin solves all problems
Different stablecoins for different use cases
USDC for stability, DAI for decentralization
5. Game theory matters:
Self-fulfilling prophecies are real
Coordination failures can break any system
Confidence is fragile, especially for algorithmic designs
The future:
- Regulatory clarity (2024-2025)
- Consolidation around winners
- Continued innovation in collateral mechanisms
- Integration with TradFi
- CBDCs as competition/complement
For users:
- Understand what backs your stablecoin
- Diversify across types (don't hold only USDT or only DAI)
- Watch for signs of stress (depeg, bank runs)
- Remember: Not all stablecoins are equally stable
For builders:
- Accept the trilemma (can't optimize all three)
- Focus on specific use case
- Design for failure modes (bank runs, depegs)
- Be honest about trade-offs
Stablecoins are here to stay, but perfect stability remains elusive. The search continues.
Prerequisites: Lesson 8.1 (Stablecoins), basic understanding of monetary policy
Key formulas and concepts recap:
Collateralization ratio: R = Collateral Value / Debt Value
Health factor: HF = (Collateral × LTV) / Debt
Capital efficiency: Issued / Collateral
Liquidation penalty: Typically 5-15%
Death spiral condition: LUNA dilution > Market cap absorption capacity
Bank run equilibrium: Multiple stable equilibria in coordination game
Trilemma: Decentralization + Capital Efficiency + Stability → Choose 2
Practice problems:
-
Calculate the maximum DAI you can mint with 10 ETH at $2,000/ETH if liquidation threshold is 150%.
-
Analyze a death spiral: If UST has $10B supply, trades at $0.80, and LUNA market cap is $8B, calculate the required LUNA dilution and explain why it triggers a death spiral.
-
Design a stablecoin mechanism that maximizes two vertices of the trilemma. Specify which two you're optimizing and explain the trade-offs.
-
Model a bank run: With 1000 users, partial backing at 80%, what's the breakeven point where running becomes dominant strategy?
-
Compare capital efficiency: Calculate how much capital is needed to issue $1B worth of stablecoin for USDC (100% backing), DAI (150% collateral), and a hypothetical algorithmic coin (0% backing). What are the risks of each?
The search for the perfect stablecoin continues—understanding why it's so hard is the first step toward better solutions.