Risk Quantification for Lending

Breaking down the risk components:

LENDING RISK DECOMPOSITION:

CATEGORY 1: SMART CONTRACT RISK

Definition:
├── Bugs in protocol code
├── Vulnerabilities exploited
├── Unintended behavior
├── Upgrade failures
└── Technical failure modes

Factors Affecting Probability:
├── Code complexity (more code = more bugs)
├── Audit coverage and quality
├── Time in production without incident
├── Bug bounty payouts and participation
├── Formal verification status
└── Development team quality

Probability Estimation Framework:

Tier 1 (Established, 2+ years, multiple audits):
├── Base rate: ~1% annual probability of significant loss
├── Range: 0.5% - 2%
├── Confidence: Moderate (track record exists)

Tier 2 (Maturing, 6-24 months, audited):
├── Base rate: ~3% annual probability
├── Range: 1% - 5%
├── Confidence: Low-moderate

Tier 3 (Emerging, <6 months, basic audit):
├── Base rate: ~8% annual probability
├── Range: 3% - 15%
├── Confidence: Low

Tier 4 (Experimental, new, minimal audit):
├── Base rate: ~20% annual probability
├── Range: 10% - 40%
├── Confidence: Very low

CATEGORY 2: ECONOMIC/DESIGN RISK

Definition:
├── Flawed economic model
├── Unsustainable tokenomics
├── Oracle manipulation
├── Governance attacks
├── Economic spirals (bank runs)
└── Model failure modes

Factors Affecting Probability:
├── Economic model complexity
├── Oracle design quality
├── Liquidation mechanism robustness
├── Token dependency
├── Historical stress test performance
└── Professional economic audits

Probability Estimation:
├── Simple models: +1-2% to base risk
├── Complex models: +3-5% to base risk
├── Token-dependent yields: +5-10% to base risk
├── Novel mechanisms: +5-10% to base risk
└── Additive to smart contract risk

CATEGORY 3: LIQUIDATION RISK (Borrowers)

Definition:
├── Collateral value drops
├── Position becomes undercollateralized
├── Liquidation triggered
├── Loss of collateral and penalty
└── Borrower-specific risk

Factors:
├── LTV chosen
├── Collateral volatility
├── Liquidation threshold
├── Liquidation penalty
├── Time horizon
└── Monitoring capability

Estimation Method:
├── Use historical volatility data
├── Calculate probability of X% drop
├── Map to liquidation threshold
└── See detailed calculation below

CATEGORY 4: PLATFORM/CHAIN RISK

Definition:
├── Underlying chain failure
├── Bridge exploits (if applicable)
├── Chain-wide vulnerabilities
├── Network congestion during crisis
└── Infrastructure-level issues

For XRPL:
├── Chain itself is mature (10+ years)
├── Chain risk: ~0.5% annual probability
├── Hooks infrastructure newer: +1-2%
├── No major bridge risk (native assets)
└── Relatively lower platform risk
```

How individual risks combine:

RISK COMBINATION:

INDEPENDENCE ASSUMPTION:

If risks are independent:
P(no loss) = P(no smart contract loss) × P(no economic loss) × ...
P(any loss) = 1 - P(no loss)

Example:
├── Smart contract risk: 3%
├── Economic risk: 2%
├── Platform risk: 1%
├── P(no loss) = 0.97 × 0.98 × 0.99 = 0.941
├── P(any loss) = 1 - 0.941 = 5.9%
└── Combined risk slightly higher than sum (risk correlation)

CORRELATION CONSIDERATION:

Risks are often correlated:
├── Market crash → Higher liquidation risk + stress on protocols
├── Smart contract exploit → Often triggers economic spiral
├── Chain congestion → Liquidation failures + cascade risk
└── Correlation increases tail risk

Practical Adjustment:
├── Add 20-50% to combined probability for correlation
├── Example: 5.9% × 1.3 = ~7.7% adjusted risk
└── This is rough but directionally correct

SIMPLIFIED APPROACH:

For practical use:
├── Estimate single "total loss probability"
├── Adjust for protocol tier
├── Add premiums for special risks
└── Don't over-engineer

Total Loss Probability Estimates:

Tier 1: 2-4% annual
Tier 2: 5-10% annual
Tier 3: 10-20% annual
Tier 4: 25-50% annual

These include all major risk categories.
```

For borrowers specifically:

LIQUIDATION PROBABILITY ESTIMATION:

HISTORICAL VOLATILITY METHOD:

Step 1: Determine Liquidation Buffer
├── Example: 60% LTV, 80% liquidation threshold
├── Buffer = (80% - 60%) / 60% = 33%
├── Price must drop 33% for liquidation
└── (Actually: Liquidation price = Debt / (Collateral × Threshold))

Step 2: Estimate Volatility
├── Use historical price data
├── Calculate annualized volatility
├── XRP example: ~100% annualized volatility (high)
├── ETH example: ~80% annualized volatility
└── Stablecoins: ~2% (minimal)

Step 3: Calculate Probability
├── Assume log-normal distribution
├── Calculate probability of 33%+ drop
├── Time horizon matters (longer = higher probability)
└── Use statistical tables or calculators

PRACTICAL ESTIMATION TABLE:

For XRP Collateral (100% annual volatility):

Time Horizon: 1 Month
├── 20% drop probability: ~25%
├── 30% drop probability: ~15%
├── 40% drop probability: ~8%
├── 50% drop probability: ~4%

Time Horizon: 3 Months
├── 20% drop probability: ~35%
├── 30% drop probability: ~22%
├── 40% drop probability: ~13%
├── 50% drop probability: ~7%

Time Horizon: 12 Months
├── 20% drop probability: ~45%
├── 30% drop probability: ~32%
├── 40% drop probability: ~22%
├── 50% drop probability: ~14%

APPLICATION:

If liquidation occurs at 30% price drop:
├── 1-month position: ~15% liquidation probability
├── 3-month position: ~22% liquidation probability
├── 12-month position: ~32% liquidation probability
└── These are SIGNIFICANT probabilities

Implications:
├── Aggressive LTV is genuinely risky
├── Time horizon matters enormously
├── XRP volatility = High liquidation risk
├── Conservative LTV (40-50%) much safer
└── Position size should reflect this

---

The core calculation:

EXPECTED LOSS FORMULA:

Basic Formula:
Expected Loss = Probability × Loss Severity × Exposure

Components:
├── Probability: Likelihood of loss event (annual %)
├── Loss Severity: Percentage of position lost if event occurs
├── Exposure: Amount at risk
└── Result: Expected annual $ loss

EXAMPLE CALCULATIONS:

Example 1: Supply-Side (Lending RLUSD)

Position: $10,000 RLUSD in Tier 2 protocol
├── Probability of total loss: 7% annual
├── Loss severity if total loss: 100%
├── Expected loss = 7% × 100% × $10,000 = $700/year
└── Expected loss = 7% of position

Example 2: Borrow-Side (Leveraged Position)

Position: $15,000 XRP collateral, $10,000 RLUSD borrowed (67% LTV)
├── Protocol risk: 7% annual (same Tier 2)
├── Liquidation risk: 20% annual (based on volatility)
├── Combined risk: ~25% (with correlation adjustment)
├── Loss if protocol fails: 100% of collateral = $15,000
├── Loss if liquidated: ~20% of collateral = $3,000 (penalty + slippage)
├── Expected protocol loss = 7% × $15,000 = $1,050
├── Expected liquidation loss = 20% × $3,000 = $600
├── Total expected loss = ~$1,650/year
└── Expected loss = 11% of collateral

SEVERITY ESTIMATION:

Protocol Failure Severities:
├── Total hack: 100% loss
├── Partial exploit: 30-70% loss
├── Economic failure: 20-80% loss
├── Temporary lock: Opportunity cost, minimal principal loss
└── Average across scenarios

For simplicity, often assume:
├── Catastrophic failure: 100% loss
├── Moderate failure: 50% loss
├── Weighted average: 70-80% loss severity
└── Conservative: Assume 100% for total loss probability
```

What yield compensates for risk:

REQUIRED YIELD FRAMEWORK:

BREAK-EVEN YIELD:

At minimum, yield must cover expected loss:
Break-even yield = Expected loss percentage

Example:
├── Expected loss: 7% annual
├── Break-even yield: 7% annual
├── At 7% yield: Expected return = 0
├── Below 7%: Expected NEGATIVE return
└── Above 7%: Expected positive return

RISK PREMIUM:

Beyond break-even, require risk premium:
├── Compensates for uncertainty in estimates
├── Compensates for illiquidity
├── Compensates for effort/monitoring
├── Compensates for psychological cost
└── Typical: 2-5% above expected loss

Required Yield = Expected Loss + Risk Premium

Example:
├── Expected loss: 7%
├── Risk premium: 3%
├── Required yield: 10%
└── Below 10%: Not attractive for this risk

COMPARISON FRAMEWORK:

Calculate risk-adjusted return:
Risk-Adjusted Return = Offered Yield - Expected Loss

Example Comparison:
├── Tier 1 protocol: 5% yield, 2% expected loss → 3% risk-adjusted
├── Tier 2 protocol: 10% yield, 7% expected loss → 3% risk-adjusted
├── Tier 3 protocol: 20% yield, 15% expected loss → 5% risk-adjusted
└── Tier 3 has highest risk-adjusted return (if estimates correct)

BUT: Uncertainty matters:
├── Tier 1 estimate confidence: High
├── Tier 3 estimate confidence: Low
├── Tier 3's 15% expected loss could be 10% or 30%
└── Consider estimate uncertainty in decisions

PRACTICAL YIELD REQUIREMENTS:

By Protocol Tier (including premium):

Tier 1:
├── Expected loss: 2-4%
├── Risk premium: 2%
├── Required yield: 4-6%
└── If below: Consider opportunity cost

Tier 2:
├── Expected loss: 5-10%
├── Risk premium: 3%
├── Required yield: 8-13%
└── If below: Probably not worth it

Tier 3:
├── Expected loss: 10-20%
├── Risk premium: 5%
├── Required yield: 15-25%
└── If below: Definitely not worth it

Tier 4:
├── Expected loss: 25-50%
├── Risk premium: 10%
├── Required yield: 35-60%
└── Most offers in this tier ARE this high (red flag)
```

Aggregating across positions:

PORTFOLIO RISK CALCULATION:

INDIVIDUAL POSITION CONTRIBUTIONS:

For each position:
├── Expected loss $ = Position size × Expected loss %
├── Sum across positions = Portfolio expected loss
└── Helps identify risk concentration

Example Portfolio:

Position 1: $20,000 in Tier 1 protocol
├── Expected loss: 3%
├── Expected loss $: $600

Position 2: $8,000 in Tier 2 protocol
├── Expected loss: 7%
├── Expected loss $: $560

Position 3: $2,000 in Tier 3 protocol
├── Expected loss: 15%
├── Expected loss $: $300

Portfolio Total: $30,000
├── Total expected loss $: $1,460
├── Portfolio expected loss %: 4.9%
└── Risk-weighted average

RISK CONCENTRATION ANALYSIS:

From example above:
├── Tier 1 contributes: $600 / $1,460 = 41% of expected loss
├── Tier 2 contributes: $560 / $1,460 = 38% of expected loss
├── Tier 3 contributes: $300 / $1,460 = 21% of expected loss

Despite being only 6.7% of portfolio value ($2K/$30K):
├── Tier 3 contributes 21% of expected loss
├── Small position, outsized risk contribution
└── This is appropriate—accepting concentrated risk for upside

CORRELATION ADJUSTMENT:

If protocols share risks:
├── Common platform (Ethereum): Correlated
├── Common oracle (Chainlink): Correlated
├── Common collateral type: Correlated
└── Correlation increases tail risk

Practical adjustment:
├── Add 10-30% to portfolio expected loss
├── For heavy Ethereum concentration: Higher adjustment
├── For diversified platforms: Lower adjustment
└── Accounts for "everything fails at once" scenario

---

Apples-to-apples comparison:

OPPORTUNITY COMPARISON FRAMEWORK:

STEP 1: GATHER DATA

For each opportunity, determine:
├── Offered yield (APY)
├── Protocol tier
├── Estimated expected loss
├── Position size (from allocation framework)
└── Time horizon

STEP 2: CALCULATE METRICS

For each opportunity:
├── Expected return = Yield - Expected loss
├── Return per unit risk = Expected return / Expected loss
├── $ expected return = Position size × Expected return %
└── Compare across opportunities

COMPARISON EXAMPLE:

Opportunity A: Tier 1 Protocol
├── Yield: 5%
├── Expected loss: 2.5%
├── Expected return: 2.5%
├── Return/risk ratio: 1.0
├── Position: $15,000
├── $ expected return: $375

Opportunity B: Tier 2 Protocol
├── Yield: 12%
├── Expected loss: 7%
├── Expected return: 5%
├── Return/risk ratio: 0.71
├── Position: $5,000 (limited by tier)
├── $ expected return: $250

Opportunity C: Tier 3 Protocol
├── Yield: 25%
├── Expected loss: 15%
├── Expected return: 10%
├── Return/risk ratio: 0.67
├── Position: $1,500 (limited by tier)
├── $ expected return: $150

ANALYSIS:

By expected return %:
├── C > B > A (Tier 3 has highest expected return %)

By return/risk ratio:
├── A > B > C (Tier 1 has best risk-adjusted ratio)

By $ expected return:
├── A > B > C (Tier 1 contributes most to portfolio)

INTERPRETATION:

├── Higher tier = Better risk efficiency
├── Lower tier = Higher expected return % but position-limited
├── Optimal: Mix across tiers per allocation framework
├── Don't chase % returns—consider position limits
└── Risk-adjusted, Tier 1 often wins
```

Applying traditional risk metrics:

SHARPE RATIO FOR LENDING:

TRADITIONAL SHARPE RATIO:

Sharpe = (Return - Risk-free rate) / Standard deviation

Measures: Excess return per unit of volatility

LENDING APPLICATION:

Adaptation for DeFi:
├── Return = Expected yield - Expected loss
├── Risk-free rate = Traditional savings/T-bills (~4-5%)
├── "Volatility" = Uncertainty in outcome
└── Modified Sharpe = (Return - Risk-free) / Risk estimate

Example Calculation:

Tier 2 Protocol:
├── Expected yield: 12%
├── Expected loss: 7%
├── Net expected return: 5%
├── Risk-free rate: 5%
├── Excess return: 5% - 5% = 0%
├── Risk (expected loss): 7%
├── Modified Sharpe: 0% / 7% = 0
└── Interpretation: No better than risk-free after adjustment

Tier 1 Protocol:
├── Expected yield: 5%
├── Expected loss: 2.5%
├── Net expected return: 2.5%
├── Risk-free rate: 5%
├── Excess return: 2.5% - 5% = -2.5%
├── Risk: 2.5%
├── Modified Sharpe: -2.5% / 2.5% = -1.0
└── Interpretation: Worse than risk-free on risk-adjusted basis

IMPLICATIONS:

In high risk-free rate environments:
├── DeFi lending may not beat traditional options
├── Must compensate for smart contract risk
├── "Safe" DeFi may underperform T-bills
├── Need either higher yields or higher risk tolerance
└── Be honest about opportunity cost

When DeFi Lending Makes Sense:
├── Risk-free rates are low (not current environment)
├── Protocol yields are high relative to risk
├── You have specific reasons beyond pure return
├── Tax efficiency, ecosystem participation, etc.
└── Or you're accepting suboptimal risk-adjusted returns
```

Portfolio-level risk metric:

VALUE AT RISK (VaR) CONCEPT:

DEFINITION:

VaR answers: "What's the maximum loss I might experience
with X% confidence over Y time period?"

Example:
├── 95% VaR of $1,000 over 1 year
├── Means: 95% confident loss won't exceed $1,000
├── 5% chance of losing more than $1,000
└── Provides risk boundary estimate

LENDING VAR CALCULATION:

Simplified Approach:

Portfolio: $30,000 in lending positions
├── Expected loss: 5% = $1,500
├── Standard deviation of loss estimate: ~3%
├── 95% confidence = 1.65 standard deviations
├── 95% VaR = Expected + 1.65 × StdDev
├── 95% VaR = 5% + 1.65 × 3% = 10%
├── 95% VaR $ = $30,000 × 10% = $3,000
└── 95% confident annual loss won't exceed $3,000

PRACTICAL APPLICATION:

Use VaR for:
├── Setting maximum acceptable loss
├── Determining if portfolio fits risk budget
├── Comparing portfolios with different compositions
└── Communication about risk

Decision Rule:
├── Calculate VaR for proposed portfolio
├── Compare to risk budget (from Lesson 15)
├── If VaR > Risk budget: Reduce exposure
├── If VaR < Risk budget: Current allocation OK
└── Adjust positions to meet risk target

EXAMPLE APPLICATION:

Risk Budget: $5,000 (what I can afford to lose)
Current Portfolio VaR: $3,000 (95% confidence)
├── VaR < Budget → Room for more risk if desired
├── Could add positions up to $5,000 VaR
└── Or maintain buffer for unexpected correlation

Risk Budget: $5,000
Current Portfolio VaR: $7,000
├── VaR > Budget → Overexposed!
├── Must reduce positions
├── Or accept exceeding stated risk tolerance
└── Don't ignore this mismatch

---

Creating a personal calculation tool:

PERSONAL RISK MODEL TEMPLATE:

INPUTS:

Protocol Assessment:
├── Name: _______________
├── Tier: 1 / 2 / 3 / 4
├── Base loss probability: _%
├── Adjustments: +% for _______
├── Total expected loss: ____%

Position Parameters:
├── Position size: $______
├── Offered yield: ____%
├── Time horizon: ______

CALCULATIONS:

Expected Values:
├── Expected loss $ = Position × Loss% = $______
├── Expected yield $ = Position × Yield% = $______
├── Net expected return $ = Yield$ - Loss$ = $______
├── Net expected return % = Net$/Position = ____%

Risk-Adjusted Metrics:
├── Risk-adjusted return = Yield% - Loss% = ____%
├── Return per unit risk = Return% / Loss% = ____
├── Sharpe-like ratio = (Return% - 5%) / Loss% = ____

Decision Support:
├── Required yield (Loss% + 3% premium) = ____%
├── Offered yield vs. required: Above / Below
├── Recommendation: Proceed / Pass

PORTFOLIO AGGREGATION:

Sum across all positions:
├── Total position value: $______
├── Total expected loss $: $______
├── Total expected yield $: $______
├── Portfolio expected loss %: %
├── Portfolio 95% VaR: $__
├── Compare to risk budget: $______
└── Status: Within budget / Over budget
```

Practical implementation:

SPREADSHEET STRUCTURE:

TAB 1: PROTOCOL DATABASE

| Protocol | Tier | Base Loss % | Adjustments | Total Loss % | Notes |
|----------|------|-------------|-------------|--------------|-------|
| Aave     | 1    | 2.0%        | +0%         | 2.0%         | Established |
| Protocol X| 2   | 5.0%        | +2% complexity| 7.0%       | New oracle |
| XRPL Y   | 3    | 12.0%       | +3% emerging | 15.0%       | Early stage |

TAB 2: POSITIONS

| Protocol | Amount | Yield % | Loss % | Exp Return % | Exp Return $ |
|----------|--------|---------|--------|--------------|--------------|
| Aave     | $15,000| 4.5%    | 2.0%   | 2.5%         | $375         |
| Protocol X| $5,000| 11.0%   | 7.0%   | 4.0%         | $200         |
| XRPL Y   | $2,000 | 22.0%   | 15.0%  | 7.0%         | $140         |
| TOTAL    | $22,000| -       | 4.4%   | 3.3%         | $715         |

TAB 3: RISK SUMMARY

├── Total deployed: $22,000
├── Expected annual yield: $1,935 (8.8%)
├── Expected annual loss: $1,220 (5.5%)
├── Net expected return: $715 (3.3%)
├── Portfolio expected loss: 5.5%
├── Estimated 95% VaR: ~$2,400
├── Risk budget: $5,000
├── Status: Within budget ✓

TAB 4: DECISION TRACKER

| Date | Protocol | Decision | Rationale | Outcome (later) |
|------|----------|----------|-----------|-----------------|
| 1/15 | XRPL Y   | Enter $2K| Exp return 7%, fits tier 3 budget | TBD |
| 1/20 | Protocol Z| Pass    | Yield 8% < required 13% | N/A |

Update monthly, review outcomes quarterly.

Improving accuracy over time:

ESTIMATE CALIBRATION:

TRACKING ACTUALS VS. ESTIMATES:

For Each Protocol/Position:
├── Record initial expected loss estimate
├── Track actual outcomes
├── Compare over time
├── Adjust estimation methodology
└── Build personal track record

Outcome Categories:
├── No loss: Expected, confirm estimate
├── Partial loss: Note amount, update model
├── Total loss: Record, major update to estimates
└── All outcomes inform future estimates

BAYESIAN UPDATING:

Prior Belief:
├── Tier 2 protocol: 7% expected loss

New Evidence:
├── Tier 2 protocol (different one) hacked
├── Industry-wide: Suggests maybe tier 2 = 10%?

Updated Belief:
├── Adjust Tier 2 estimate toward 8-9%
├── Weight: Prior evidence vs. new evidence
├── Don't overreact to single events
├── But don't ignore new information
└── Gradual adjustment

CALIBRATION QUESTIONS:

Quarterly Review:
├── Were my estimates too high or too low?
├── Did I miss any risk factors?
├── Was my tier classification accurate?
├── How did my recommendations perform?
├── What should I adjust?

Annual Review:
├── Calculate: Total expected losses vs. actual losses
├── If actual > expected: Estimates too optimistic
├── If actual < expected: Estimates too conservative
├── Adjust base rates accordingly
└── Improve methodology

HUMILITY:

Remember:
├── Small sample sizes mean high variance
├── No losses ≠ estimates too conservative
├── Single loss ≠ estimates too optimistic
├── Need years of data for calibration
├── Stay humble about prediction accuracy
└── Range estimates better than point estimates

---

✅ Quantification enables better decisions - Even rough numbers beat pure intuition for comparing opportunities and sizing positions.

✅ Expected value framework is sound - Probability × Impact is mathematically valid for risk assessment.

✅ Diversification reduces portfolio risk - Spreading across uncorrelated risks improves outcomes.

⚠️ Actual probabilities - Our estimates are educated guesses. Actual probabilities may be significantly different.

⚠️ Correlation structure - How risks correlate is hard to estimate. "Everything fails at once" scenarios are hard to model.

⚠️ Tail risks - Extreme events (black swans) may be more common than our models suggest.

🔴 False precision - Treating 7.3% as meaningfully different from 7.5%. Use ranges.

🔴 Model overconfidence - "The model says X" doesn't mean X is correct. Models are tools, not truth.

🔴 Ignoring qualitative factors - Some risks can't be easily quantified. Don't ignore them because they don't fit the spreadsheet.

Risk quantification is useful but imprecise. The value is in the process—forcing yourself to think through probabilities, expected losses, and comparisons—not in the specific numbers. Treat your estimates as rough approximations, use ranges rather than point estimates, and maintain humility about your predictive accuracy. A simple model used consistently beats a sophisticated model used poorly.

---

Assignment: Build a personal risk quantification spreadsheet for your lending activities.

Requirements:

Part 1: Protocol Database (20%)

• Protocol name
• Tier classification
• Base loss probability estimate
• Risk factor adjustments
• Total expected loss probability
• Confidence level in estimate

Include at least 5 protocols (real or hypothetical).

Part 2: Position Calculator (25%)

• Position size

• Protocol expected loss %

• Offered yield %

• Expected loss $

• Expected yield $

• Net expected return

• Required yield (with premium)

• Decision recommendation

Part 3: Portfolio Aggregator (25%)

• Sums individual positions
• Calculates portfolio expected loss %
• Estimates 95% VaR
• Compares to risk budget
• Identifies risk concentration

Part 4: Decision Log (15%)

• Date of decision
• Protocol and position
• Key estimates used
• Decision made
• Outcome (to be filled later)

Part 5: Documentation (15%)

• How you estimate base loss probabilities

• What adjustments you apply and why

• Your risk premium requirement

• How you'll calibrate over time

• Model functionality (30%)

• Methodology soundness (25%)

• Practical usability (25%)

• Documentation clarity (20%)

Time investment: 3-4 hours
Value: This spreadsheet becomes your ongoing risk management tool for all lending activities.

---

For Next Lesson:
Lesson 17 covers Monitoring and Position Management—the ongoing work required to maintain healthy lending positions.

---

End of Lesson 16

Total words: ~6,500
Estimated completion time: 60 minutes reading + 3-4 hours for deliverable exercise