Derivatives and Hedging
A receive-fixed overlay is sized to the 10-year key-rate gap only. Twist the wings, add volatility, or add basis between the hedge index and the liability crediting rate, and residual risk appears. Teaching key-rate durations and a constant vega, not a vendor DV01 tape.
01
Why a parallel hedge is not enough
Duration matching closes a parallel move. Key-rate duration asks which part of the curve moved. Embedded guarantees add vega: the option to lapse, annuitize, or put a floor on income cheapens or dearens when implied volatility moves. Basis is the leftover when the swap index is not the crediting rate.
Analogy: you insured the middle of the roof. A storm that hits only the eaves still leaks. The eaves are the 2-year and 30-year key rates. Wind damage is volatility. A policy written on a different house is basis.
Closing the 10-year gap makes a 10-year-only shock look tidy. A 30-year twist, a vol jump, or a 10 basis point index mismatch all reopen residual surplus.
02
Key-rate shock and residual
Teaching lab
Key-rate shock and residual
The receive-fixed hedge is sized to the 10-year key-rate gap only. Twist the wings, add volatility, or add basis (hedge index versus liability crediting rate) and the residual appears. Teaching KRDs, not a vendor risk system.
- Unhedged surplus
- $1,000,000
- Residual
- $0
- Hedge effectiveness
- 100%
Closing the middle of the curve is not closing the whole curve. Wings, volatility on the guarantee, and a different index on the swap all leak. That leak is residual risk.
03
Behind the scenes
- Read the three key-rate sliders, the vol slider, and the basis slider.
- Revalue assets and liabilities with the teaching KRDs at 2y, 10y, and 30y.
- Apply the 10-year-only hedge and the constant vega coefficient.
- Subtract a basis cost when the hedge index is not the crediting rate.
- Residual is the leftover surplus change. Effectiveness compares it with the unhedged change.
04
Greeks, dynamic hedging, and slippage
Delta (or DV01) is the first-order rate or underlying move. Gamma is how that delta changes — the reason a static overlay slips after a large move. Vega is the volatility Greek on the embedded guarantee or the swaption overlay. Rho is residual discount-rate sensitivity after the designed hedge. This lab’s KRD stack is a bucketed delta. Its vol slider is a bucketed vega. Gamma is not re-solved on every tick; that is the point of calling the overlay static.
Dynamic hedging rebalances as delta changes. Hedge slippage is the cash that leaks between rebalances, plus transaction costs, plus basis. Interest-rate and equity stress on the same book live on the Sensitivity & Stress Testing lab. Swaption mechanics stay on swaptions and interest-rate hedging. Position Greeks for listed options stay on Greeks.
05
A real-world ALM use
A hedge desk claims 95% effectiveness on a +100 basis point 10-year move. Leave the 10-year slider at +100 and raise the 30-year slider. Effectiveness falls because the overlay was not sized to the wing. That is the conversation an ALCO has before it accepts a single-number hedge ratio.
06
Limitations
Three key rates, constant KRDs, no partial DV01 tape, no smile, no counterparty, no initial or variation margin, and no hedge accounting overlay. Effectiveness can print above 100% or go negative when residual and unhedged have opposite signs — a teaching ratio, not an accounting test. Do not treat the residual as a live P&L.
Related pages: Advanced ALM hub · Advanced quantitative finance · Shock and net impact.
