Class MidPointCDOEngine

Namespace

NanoQL.Experimental

Assembly

NanoQL.dll

The mid-point engine for synthetic CDOs: each premium period's loss taken at its midpoint, the premium paid on the tranche's notional less the expected loss at the period's end.

[Experimental("NQL0001")]
public static class MidPointCDOEngine

Inheritance

object

MidPointCDOEngine

Inherited Members

object.GetType()

object.MemberwiseClone()

object.ToString()

object.Equals(object)

object.Equals(object, object)

object.ReferenceEquals(object, object)

object.GetHashCode()

Remarks

A static class: Calculate(SyntheticCDO, EvaluationContext, YieldTermStructure) takes the engine's inputs and returns the results. It reads the statistics through the CDO's basket, which must be of the engine's evaluation date. Premium accrual on default isn't paid, as in QuantLib's.

The legs and the upfront are levered by the CDO's leverage factor, and the fair premium and fair upfront are the unlevered rates: QuantLib's levers the premium leg alone, so a tranche at twice its notional reads half its fair premium, 0.02268 for 0.04535; and its additional "fairPremium" is −(protection + upfront)·rate/premium, −0.04535 where the instrument's is 0.04535 (Divergences/CdoLeverageTests.cs). The fair upfront is the side-invariant rate that zeroes the value, over the first accrual's discount: QuantLib's reads −0.2000 for a buyer where the seller's is 0.2000, and a forward start's value at it −0.22.

A basket with realized defaults is priced on the tranche's outstanding notional, its inception notional less the realized tranche loss and the amortized recovery: the premium leg pays on it less the live names' expected loss, the upfront is due on it, and a paid first coupon starts the losses from the realized tranche loss. A discount before the discount curve's reference date is 1, and a default probability before a default curve's 0, where QuantLib's reads refuse a seasoned CDO and curves dated T+2 ("negative time").

Methods

Calculate(SyntheticCDO, EvaluationContext, YieldTermStructure)

Prices a synthetic CDO.

public static SyntheticCDOResults Calculate(SyntheticCDO cdo, EvaluationContext context, YieldTermStructure discountCurve)

Parameters

cdo SyntheticCDO

The CDO.

context EvaluationContext

The evaluation context, of the basket's evaluation date.

discountCurve YieldTermStructure

The discount curve.

Returns

SyntheticCDOResults

The CDO's results.

Exceptions

ArgumentNullException

A reference argument is null.

NanoQLException

The basket's evaluation date isn't the context's, the basket has no loss model, or the model refuses a statistic.