1Short-rate modelling framework(3)  QuantLib  Short-rate modelling framework(3)
2
3
4

NAME

6       Short-rate modelling framework -
7

Detailed Description

9       This framework (corresponding to the ql/ShortRateModels directory)
10       implements some single-factor and two-factor short rate models. The
11       models implemented in this library are widely used by practitionners.
12       For the moment, the ShortRateModels::Model class defines the short-rate
13       dynamics with stochastic equations of the type dx_i = (t,x_i) dt +
14       ma(t,x_i) dW_t ] where $ r = f(t,x) $. If the model is affine (i.e.
15       derived from the QuantLib::AffineModel class), analytical formulas for
16       discount bonds and discount bond options are given (useful for
17       calibration).
18

Single-factor models

20       The Hull & White model.RS 4 dr_t = (heta(t) - lpha(t) r_t)dt + ma(t)
21       dW_t ] When $ lpha $ and $ ma $ are constants, this model has
22       analytical formulas for discount bonds and discount bond options.
23

The Black-Karasinski model.RS 4 dr_t} = (heta(t) - lpha r_t})dt + ma dW_t ] No

25analytical tractability here.
26

The extended Cox-Ingersoll-Ross model.RS 4 dr_t = (heta(t) - k r_t)dt + ma

28t{r_t} dW_t ] There are analytical formulas for discount bonds (and soon for
29discount bond options).
30

Calibration

32       The class CalibrationHelper is a base class that facilitates the
33       instanciation of market instruments used for calibration. It has a
34       method marketValue() that gives the market price using a Black formula,
35       and a modelValue() method that gives the price according to a model
36
37       Derived classed are QuantLib::CapHelper and QuantLib::SwaptionHelper.
38
39       For the calibration itself, you must choose an optimization method that
40       will find constant parameters such that the value: V = t{_{i=1}^{n}
41       ac{(T_i - M_i)^2}{M_i}}, ] where $ T_i $ is the price given by the
42       model and $ M_i $ is the market price, is minimized. A few optimization
43       methods are available in the ql/Optimization directory.
44

Two-factor models

Pricers

47       Analytical pricers.RS 4
48
49
50If the model is affine, i.e. discount bond options formulas exist, caps are
51easily priced since they are a portfolio of discount bond options. Such a
52pricer is implemented in QuantLib::AnalyticalCapFloor. In the case of single-
53factor affine models, swaptions can be priced using the Jamshidian
54decomposition, implemented in QuantLib::JamshidianSwaption.
55

Using Finite Differences.RS 4

57
58
59(Doesn't work for the moment) For the moment, this is only available for
60single-factor affine models. If $ x = x(t, r) $ is the state variable and
61follows this stochastic process: dx_t = (t,x)dt + ma(t,x)dW_t ] any european-
62style instrument will follow the following PDE:
63
64ac{artial
65P}{artial
66t} +  ac{artial
67P}{artial
68x} + ac{1}{2} ma^2 ac{artial^2
69P}{artial
70x^2} = r(t,x)P ].PP The adequate operator to feed a Finite Difference Model
71instance is defined in the QuantLib::OneFactorOperator class.
72

Using Trees.RS 4

74
75
76Each model derived from the single-factor model class has the ability to
77return a trinomial tree. For yield-curve consistent models, the fitting
78parameter can be determined either analytically (when possible) or
79numerically. When a tree is built, it is then pretty straightforward to
80implement a pricer for any path-independant derivative. Just implement a class
81derived from NumericalDerivative (see QuantLib::NumericalSwaption for example)
82and roll it back until the present time... Just look at QuantLib::TreeCapFloor
83and QuantLib::TreeSwaption for working pricers.
84
85   Classes
86       class AffineModel
87           Affine model class.
88       class TermStructureConsistentModel
89           Term-structure consistent model class.
90       class ShortRateModel
91           Abstract short-rate model class.
92       class OneFactorModel
93           Single-factor short-rate model abstract class.
94       class OneFactorAffineModel
95           Single-factor affine base class.
96       class BlackKarasinski
97           Standard Black-Karasinski model class.
98       class CoxIngersollRoss
99           Cox-Ingersoll-Ross model class.
100       class ExtendedCoxIngersollRoss
101           Extended Cox-Ingersoll-Ross model class.
102       class HullWhite
103           Single-factor Hull-White (extended Vasicek) model class.
104       class Vasicek
105           Vasicek model class
106       class TwoFactorModel
107           Abstract base-class for two-factor models.
108       class G2
109           Two-additive-factor gaussian model class.
110
111
112
113Version 0.8.1                     29 Oct 2007Short-rate modelling framework(3)
Impressum