Autocallables - Early Redemption Structures

Level 3: Advanced | Module 3.3 | Time: 3 hours

🎯 Learning Objectives

By the end of this module, you will:

Master autocall mechanics and trigger levels
Design coupon structures (fixed, memory, step-up)
Understand worst-of autocallables
Price and risk manage autocalls
Identify when to use vs avoid

Prerequisites: Barrier Options & Exotics

What is an Autocallable?

A structured product that automatically redeems (“calls”) early if the underlying asset reaches a predetermined trigger level on scheduled observation dates.

The Core Concept

Traditional Note:
- Fixed term (e.g., 2 years)
- Hold until maturity
- Payoff determined at end

Autocallable:
- Maximum term (e.g., 2 years)
- Can redeem EARLY if conditions met
- Quarterly or monthly observations
- If triggered: Receive principal + coupon
- If not triggered: Continue to next observation

Think of it as: A series of knock-in opportunities to exit with profit.

Why Autocallables?

1. Enhanced Yield

Problem: Bond yields are low (3-4%)

Solution: Autocallable on Bitcoin
- Coupon: 10% per quarter (40% annualized!)
- If BTC rises: Called early, keep coupon + principal
- Risk: Downside if BTC falls significantly

2. Defined Upside, Limited Downside Exposure

Traditional Equity:
- Unlimited upside ✅
- Unlimited downside ❌

Autocallable:
- Capped upside (trigger level) ⚠️
- Barrier protection often included ✅
- Sweet spot: Moderate gains

3. Behavioral Appeal

Early redemption feels like “winning”

Investor psychology:
Month 3: "It autocalled! I got 10% in 3 months!"
Feels better than "I made 10% but it took a year"

This behavioral feature makes autocalls popular with retail

Basic Autocallable Structure

Standard Quarterly Autocall

Investment: $100,000
Term: 2 years (8 quarterly observations)
Underlying: Bitcoin (initial: $50,000)
Autocall trigger: 110% of initial ($55,000)
Coupon: 5% per quarter (if not called)
Barrier: 70% of initial ($35,000)

Observation Schedule:
Month 3, 6, 9, 12, 15, 18, 21, 24

How It Works: Scenario Analysis

Scenario A: Early Autocall (Best Case)

Observation 1 (Month 3):
Bitcoin: $56,000 (above $55k trigger) ✅

Result: AUTOCALLED!
Receive: $100,000 (principal) + $5,000 (1 coupon)
Total: $105,000

Return: 5% in 3 months (20% annualized)
Product terminates ✅

Scenario B: Later Autocall

Observation 1: BTC $52,000 → No autocall
Observation 2: BTC $54,000 → No autocall
Observation 3: BTC $57,000 → AUTOCALLED!

Receive: $100,000 + ($5,000 × 3 quarters) = $115,000

Return: 15% in 9 months (20% annualized)
Product terminates ✅

Scenario C: Never Calls, Final Payoff Above Barrier

All 8 observations: BTC stays $48k-$54k (never hits $55k)

Final observation (Month 24):
Bitcoin: $52,000 (above $35k barrier)

Receive: $100,000 + ($5,000 × 8) = $140,000

Return: 40% over 2 years (20% annualized)
Full term, full coupons ✅

Scenario D: Never Calls, Falls Below Barrier (Worst Case)

All 8 observations: BTC never hits $55k
Final: Bitcoin $30,000 (below $35k barrier)

Downside participation activates:
Loss = ($50,000 - $30,000) / $50,000 = 40%
Receive: $100,000 × (1 - 0.40) = $60,000

Plus coupons: $60,000 + ($5,000 × 8) = $100,000

Return: 0% (break-even despite 40% BTC loss)

Actually better than holding BTC! ⚠️
But still a disappointment

Coupon Structures

1. Fixed Coupon (Standard)

Pay same amount each period (if not called)

Example: 5% quarterly

Observation 1: $5,000
Observation 2: $5,000
Observation 3: $5,000
...

Simple, predictable

2. Memory Coupon (Investor-Friendly)

Unpaid coupons accumulate and pay when autocalled

Memory Feature:

Observation 1: BTC below trigger → No payment yet
Observation 2: BTC below trigger → Still no payment
Observation 3: BTC above trigger → AUTOCALL!

Receive: Principal + (3 × $5,000) = $115,000

Even though it took 3 quarters, you get ALL coupons ✅

This is generous to investors
More expensive for issuers to offer

Why Memory Matters:

Without Memory:
Obs 1-3: No autocall → No coupons paid
Obs 4: Autocalls → Get $5,000 (1 coupon)
Lost: $15,000 in coupons ❌

With Memory:
Obs 1-3: No autocall → Coupons accrue
Obs 4: Autocalls → Get $20,000 (4 coupons)
Lost: Nothing ✅

3. Step-Up Coupon

Coupon increases over time (compensate for time risk)

Observation 1: 3% ($3,000)
Observation 2: 4% ($4,000)
Observation 3: 5% ($5,000)
Observation 4: 6% ($6,000)
...

If autocalls late, you earn more
Incentivizes patience

4. Conditional Coupon

Coupon only paid if condition met (e.g., above barrier)

Quarterly Observation:

If BTC ≥ $45,000 (90% of initial): Pay 5% coupon
If BTC < $45,000: No coupon

Example:
Q1: BTC $52k → Pay $5,000 ✅
Q2: BTC $48k → Pay $5,000 ✅
Q3: BTC $43k → Pay $0 ❌
Q4: BTC $51k → Pay $5,000 ✅

Higher coupon rate (e.g., 7% vs 5%)
But conditional on performance

Worst-Of Autocallables

Autocall trigger based on WORST performing asset in a basket

Structure

Worst-of Autocallable on BTC + ETH + SOL

Initial Prices:
BTC: $50,000
ETH: $3,000
SOL: $100

Autocall: If ALL assets ≥ 110% of initial
Barrier: If ANY asset < 70% of initial

Coupon: 8% quarterly (high yield for multi-asset risk)

Scenario Analysis

Scenario A: All Perform Well (Autocall)

Observation 3:
BTC: $60,000 (+20%) ✅
ETH: $3,500 (+16.7%) ✅
SOL: $115 (+15%) ✅

All above 110% trigger → AUTOCALL!
Receive: $100,000 + ($8,000 × 3) = $124,000

Scenario B: One Laggard (No Autocall)

Observation 3:
BTC: $62,000 (+24%) ✅
ETH: $3,600 (+20%) ✅
SOL: $105 (+5%) ❌ (below 110% trigger)

One asset below trigger → NO autocall
Continue to next observation

Scenario C: One Crashes (Downside)

Final Observation:
BTC: $55,000 (+10%)
ETH: $3,300 (+10%)
SOL: $60 (-40%) ❌ (below 70% barrier)

Worst performer: SOL (-40%)
Your loss: 40% on worst performer

Receive: $100,000 × (1 - 0.40) + coupons
       = $60,000 + $64,000 = $124,000

Coupons offset principal loss ⚠️

Why Worst-Of?

Higher coupons for taking multi-asset tail risk

Single-asset autocall: 5% quarterly
Worst-of (3 assets): 8-10% quarterly

Why?
- You're exposed to WEAKEST link
- Higher probability of barrier breach
- Correlation risk (all can fall together)

Trade-off: Much higher yield for accepting worst-of risk

Pricing Autocallables

Monte Carlo required (path-dependent, multiple observations)

Python Implementation

def price_autocallable(S0, T, coupon_rate, autocall_level, barrier_level,
                       r, sigma, N_sims=10000, observations=8):
    """
    Price a quarterly autocallable note
 
    S0: Initial price
    T: Total term (years)
    coupon_rate: Per-period coupon (decimal)
    autocall_level: Trigger as % of S0 (e.g., 1.10)
    barrier_level: Barrier as % of S0 (e.g., 0.70)
    """
 
    obs_times = np.linspace(T/observations, T, observations)
    dt = T / observations
 
    payoffs = []
 
    for sim in range(N_sims):
        S = S0
        coupons_paid = 0
        autocalled = False
 
        for i, t in enumerate(obs_times):
            # Simulate to next observation
            Z = np.random.standard_normal()
            S = S * np.exp((r - 0.5*sigma**2)*dt + sigma*np.sqrt(dt)*Z)
 
            # Check autocall
            if S >= S0 * autocall_level:
                # Autocalled!
                principal = 100000
                coupons = 100000 * coupon_rate * (i + 1)  # All coupons
                payoff = principal + coupons
                discount_factor = np.exp(-r * t)
                payoffs.append(payoff * discount_factor)
                autocalled = True
                break
 
        # If never autocalled, final payoff
        if not autocalled:
            principal = 100000
 
            if S >= S0 * barrier_level:
                # Above barrier: full principal
                final_payoff = principal
            else:
                # Below barrier: downside participation
                loss = (S0 - S) / S0
                final_payoff = principal * (1 - loss)
 
            # Add all coupons
            coupons = 100000 * coupon_rate * observations
            payoff = final_payoff + coupons
            discount_factor = np.exp(-r * T)
            payoffs.append(payoff * discount_factor)
 
    avg_payoff = np.mean(payoffs)
 
    return avg_payoff
 
# Example
value = price_autocallable(
    S0=50000,
    T=2.0,          # 2 years
    coupon_rate=0.05,  # 5% per quarter
    autocall_level=1.10,
    barrier_level=0.70,
    r=0.04,
    sigma=0.80,
    N_sims=10000,
    observations=8
)
 
print(f"Autocallable Value: ${value:,.2f}")
print(f"Investment: $100,000")
print(f"Implied Return: {(value/100000 - 1)*100:.2f}%")
 
# Expected: $105,000-$110,000 (5-10% expected return)

Risk Analysis

Probability of Autocall by Observation

Simulate to understand early call likelihood

Simulation Results (10,000 paths):

Observation 1 (Month 3): 35% autocall
Observation 2 (Month 6): 48% autocall (cumulative)
Observation 3 (Month 9): 58% autocall
Observation 4 (Month 12): 65% autocall
...
Observation 8 (Month 24): 78% autocall

Never autocalls: 22%

Average autocall time: Month 8 (if it autocalls)

Insights:

High probability of early exit (78%)
Most autocall within first year
22% chance of full 2-year term

Downside Risk Assessment

What happens in crashes?

Stress Test Scenarios:

Scenario 1: Bitcoin -30% (to $35,000)
- At barrier exactly
- Receive: $100,000 + coupons
- Return: +40% (coupons offset)

Scenario 2: Bitcoin -50% (to $25,000)
- Below barrier
- Loss: 50% on principal
- Receive: $50,000 + $40,000 coupons = $90,000
- Return: -10% (despite 50% BTC crash!)

Scenario 3: Bitcoin -70% (to $15,000)
- Severe crash
- Loss: 70% on principal
- Receive: $30,000 + $40,000 coupons = $70,000
- Return: -30% (vs -70% BTC)

Coupons provide significant cushion ✅
But can't fully protect in severe crashes ❌

Greeks for Autocallables

Delta Profile

Delta changes dramatically by observation:

Early in term (Month 1):
- Delta ≈ +0.30 (moderate positive)
- Far from autocall, far from barrier

Near autocall (BTC at $54k, trigger $55k):
- Delta → +0.80 (very high!)
- Small move could trigger autocall

Near barrier (BTC at $36k, barrier $35k):
- Delta → -0.60 (negative!)
- Protection zone, inverse exposure

This is complex to hedge!

Vega Profile

Autocallables are SHORT volatility

Why?
- High vol → more likely to breach barrier (bad)
- High vol → less likely to autocall smoothly (less good)
- Issuer is short options → short vega

Implication:
- Autocallables suffer in vol spikes
- Perform well in calm markets
- Similar risk to selling puts

When to Use Autocallables

✅ Ideal Conditions

1. Moderate Bullish View

Market View: "BTC will rise 10-20% over next year"

Autocallable Fit:
- Autocall trigger at +10% → likely to hit
- Get enhanced yield
- Exit early with profit ✅

2. Low to Moderate Volatility

Volatility: 50-60% (below historical avg of 70-80%)

Why Good:
- Less barrier breach risk
- More predictable outcomes
- Higher probability of autocall

3. Seeking Enhanced Yield

Alternative: Bonds paying 4%

Autocallable: 20%+ annual coupon (if held to maturity)

Trade-off: Accept tail risk for 5x yield

❌ Avoid When

1. Very Bullish

Market View: "BTC going to $100k!"

Problem:
- Autocall caps upside at $55k
- Miss massive rally
- Opportunity cost huge

Better: Just buy Bitcoin

2. Very Bearish

Market View: "BTC crashing to $20k"

Problem:
- Barrier breach certain
- Principal loss
- Coupons don't offset enough

Better: Sell BTC or buy puts

3. High Volatility Environment

Current IV: 120% (crisis level)

Risk:
- Whipsaw through barrier
- Low autocall probability
- Tail risk elevated

Better: Wait for vol to normalize

4. Need Liquidity

Concern: "Might need cash in 6 months"

Problem:
- Autocallables are illiquid
- Can't exit early without penalty
- Locked in for term

Better: Keep cash or use liquid instruments

Autocallable Variations

1. Knock-In Autocallable

Add knock-in feature:

Downside only activates if barrier TOUCHED anytime

Structure:
- Autocall: $55k
- Knock-in barrier: $35k
- Principal protected UNLESS barrier touched

Benefit: Full protection if never touches $35k
Risk: If touches $35k once, downside exposure forever

2. Phoenix Autocallable

"Memory" on steroids + conditional coupons

Features:
- Coupons accrue even when not paid
- Pay all accrued coupons when autocalls
- Higher overall yield

Example:
Q1-Q3: BTC below autocall, above coupon barrier
      → Earn coupons but not paid yet
Q4: BTC hits autocall trigger
      → Receive principal + 4 quarters of coupons

Very investor-friendly

3. Digital Autocallable

Binary payoffs instead of smooth

Trigger: $55k
Digital payout if triggered: $20,000 (20% gain)

Outcome:
BTC at $54,999: $0 payoff
BTC at $55,001: $20,000 payoff

All-or-nothing mechanics
Higher coupons, more speculative

Real-World Example: 2021 Bitcoin Autocallable

Background

Date: January 2021
Bitcoin: $30,000
2-Year Autocallable Launched:

Structure:
Investment: $1,000,000
Autocall trigger: 120% ($36,000)
Coupon: 10% quarterly
Barrier: 50% ($15,000)
Observations: Quarterly for 2 years

What Actually Happened

Q1 2021 (April): BTC $60,000
→ AUTOCALLED at first observation! ✅

Payout: $1,000,000 + $100,000 = $1,100,000
Return: 10% in 3 months (40% annualized)

Investor Result: Very happy ✅

Issuer/Bank Result:
- Sold BTC exposure too cheap
- BTC went to $65k (should have sold higher)
- Lost on the trade ❌

This shows autocalls can work TOO well for investors!

Key Takeaways

1. Autocallables offer early exit at trigger levels

Quarterly or monthly observations
Automatic redemption if triggered
Principal + accrued coupons

2. High coupons compensate for downside risk

15-40% annualized yields common
But exposed to tail risk (barrier breach)
Coupons provide cushion

3. Memory and step-up features enhance value

Memory: Get unpaid coupons when autocalls
Step-up: Higher coupons over time
Conditional: Coupons depend on performance

4. Worst-of structures multiply risk

Higher yields (8-12% quarterly)
But exposed to weakest asset
Correlation risk in crashes

5. Best for moderate bull markets, low vol

Sweet spot: 10-20% upside
Calm markets (low volatility)
Avoid in extremes (very bullish/bearish)

6. Complex Greeks require active management

Delta changes dramatically near triggers/barriers
Short volatility exposure
Path-dependent behavior

What’s Next?

You’ve mastered autocallables! You now understand:

✅ Early redemption mechanics
✅ Coupon structures (fixed, memory, step-up)
✅ Worst-of autocallables
✅ Pricing via Monte Carlo
✅ When to use vs avoid

Ready to trade volatility itself?

Continue to: Volatility Trading →

Learn pure volatility strategies, dispersion trading, and advanced vol arbitrage.

Next Module: Volatility Trading →

Related Topics:

Barrier Options - Foundation for autocalls
Monte Carlo - Price autocallables
Volatility - Understand short vol exposure