analyzing-exotic-options

# Analyzing Exotic Options

## When To Use

- Pricing or valuing barrier options (knock-in, knock-out, single/double barrier)
- Analyzing Asian (average-price or average-strike) options for hedging or structured product design
- Evaluating lookback options (floating or fixed strike) for optimal entry/exit analysis
- Pricing digital (binary/cash-or-nothing) options for event-driven payoffs
- Structuring exotic payoffs within structured notes, CLNs, or autocallable products
- Comparing exotic option economics against vanilla alternatives for a given hedge or exposure

## Inputs To Gather

- **Underlying specification**: asset class, ticker/identifier, spot price, dividend/carry assumptions
- **Option type and subtype**: barrier (up-and-in, down-and-out, etc.), Asian (arithmetic vs. geometric, discrete vs. continuous averaging), lookback (floating vs. fixed strike), digital (cash-or-nothing vs. asset-or-nothing)
- **Strike, barrier levels, and observation conventions**: barrier monitoring frequency (continuous vs. discrete), averaging dates/windows, lookback observation period
- **Expiry and exercise style**: European, American, Bermudan; early exercise provisions if any
- **Market data**: volatility surface (skew/term structure), risk-free rate curve, correlation matrix (for multi-asset exotics), repo/borrow rates
- **Model preferences**: Black-Scholes analytic (where closed-form exists), Monte Carlo (path-dependent), PDE/finite-difference, tree-based methods
- **Notional, currency, and settlement terms**
- **Purpose**: trading desk pricing, risk management hedge analysis, structured product design, or model validation

## Workflow

1. **Classify the exotic structure**
   - Identify the payoff type: path-dependent (Asian, lookback, barrier) vs. event-triggered (digital)
   - Map the exact payoff formula — express mathematically before coding or modeling
   - Determine whether closed-form solutions exist (e.g., Merton for continuous barriers, Rubinstein for lookbacks) or if simulation is required

2. **Select and configure the pricing model**
   - For continuous-barrier Europeans with lognormal dynamics: use analytic formulas (Reiner-Rubinstein for single barriers) [VERIFY applicability to discrete monitoring]
   - For discrete barriers: apply Broadie-Glasserman-Kou continuity correction or use Monte Carlo with fine time-stepping
   - For arithmetic Asian options: no exact closed-form — use Monte Carlo with variance reduction (control variate using geometric Asian) or moment-matching approximations (Turnbull-Wakeman, Levy)
   - For lookback options: use analytic formulas for continuous observation (Goldman-Sosin-Gatto for floating strike) [VERIFY whether observation is truly continuous or discrete]
   - For digitals: Black-Scholes closed-form with spread approximation for risk management; adjust for skew using call/put spread replication
   - Specify simulation parameters if Monte Carlo: number of paths (minimum 100k for convergence), time steps (align to observation dates), random number generation (Sobol recommended for path-dependents), seed for reproducibility

3. **Calibrate inputs and run sensitivities**
   - Fit the volatility surface — exotic prices are highly sensitive to skew and term structure; use market-implied vols at relevant strikes/tenors, not ATM flat vol
   - For barrier options: compute vanna and volga adjustments if using a local/stochastic vol model [VERIFY model choice aligns with desk conventions]
   - Run Greeks: delta, gamma, vega, theta; for barriers also compute barrier delta (sensitivity to barrier level) and gap risk near knock-out
   - For Asians: assess the impact of fixing schedule changes and partial averaging (options mid-averaging period)
   - Stress test: shift vol surface, spot, rates, and correlation; identify largest P&L drivers

4. **Analyze path-dependency and risk characteristics**
   - For barriers: quantify pin risk near the barrier at expiry, assess knock-in vs. knock-out economics, evaluate rebate value
   - For Asians: compare arithmetic vs. geometric pricing gap, assess vol reduction from averaging, evaluate Greeks evolution as fixings accumulate
   - For lookbacks: analyze the premium over vanilla (lookback premium), assess whether cost is justified by the hedge benefit
   - For digitals: quantify discontinuity risk at the strike, evaluate overhedge cost of call/put spread replication at various widths

5. **Benchmark and validate**
   - Cross-check analytic prices against Monte Carlo (where both are available) — differences should be within 2-3 standard errors
   - Compare against Bloomberg DLIB, FinCAD, or internal library pricing [VERIFY available systems]
   - Validate Greeks numerically (bump-and-reprice) against analytic Greeks
   - If structured product: decompose into component exotics and vanilla pieces, verify sum matches full-structure price

6. **Document findings**
   - State the pricing model used, key assumptions (vol surface, dividend treatment, barrier monitoring convention), and limitations
   - Present fair value, Greeks, and scenario analysis in tabular form
   - Highlight model risk: where does the chosen model likely under/overstate value (e.g., local vol vs. stochastic vol for barriers)
   - Flag any approximations and their estimated impact

## Output

- **Exotic option valuation summary**: fair value (mid), bid/ask adjustment methodology, and confidence interval (for MC-based prices)
- **Greeks table**: delta, gamma, vega, theta, rho; plus exotic-specific sensitivities (barrier delta, fixing sensitivity for Asians)
- **Scenario matrix**: P&L under spot/vol/rate shifts and correlation stress
- **Model comparison** (if applicable): side-by-side pricing from analytic vs. MC vs. alternative models with variance explained
- **Risk commentary**: key risks (pin risk, gap risk, model risk, discrete monitoring bias), hedging recommendations, and replication cost estimates
- **Payoff diagram**: mathematical payoff specification and, where useful, illustrative payoff profiles at expiry

## Quality Checks

- Payoff formula matches the term sheet or trade confirmation exactly — verify barrier direction, observation convention, averaging methodology, and settlement type
- Monte Carlo standard error is below 0.5% of option value; increase paths or apply variance reduction if not
- Greeks are internally consistent: delta integrates to option value across spot range, put-call parity holds for European-style exotics where applicable
- Barrier option price converges to vanilla price as barrier moves far from spot (knock-in) or to zero as barrier approaches spot (knock-out)
- Asian option price is bounded between the geometric average option price (lower) and the vanilla option price (upper) for calls
- Digital option price equals the negative of the derivative of the vanilla call price with respect to strike — verify numerically
- All jurisdiction-specific regulatory constraints on exotic sales or structured product classification are noted [VERIFY: MiFID II target market, Dodd-Frank SEF eligibility, ISDA documentation requirements]