est.cir: Estimates the parameters of the CIR model.

In SMFI5: R functions and data from Chapter 5 of 'Statistical Methods for Financial Engineering'

Description

Estimates the parameters of the CIR model

dr = alpha(beta-r)dt + sigma sqrt(r) dW

with market price of risk q(r) = q1/sqrt(r) +q2 sqrt(r). The time scale is in years and the units are percentages.

Usage

est.cir(data, method = "Hessian", days = 360, significanceLevel = 0.95)

Arguments

data	c(R,tau) (n x 2), with R: annual bonds yields in percentage, and tau: maturities in years.
method	'Hessian' (default), 'num'.
days	Number of days per year (default: 360).
significanceLevel	95%(default).

Value

theta	Parameters (alpha, beta, sigma, q1,q2) of the model.
error	Estimation errors for the given confidence level.
rimp	Implied spot rate.

Note

Translated from Matlab by David-Shaun Guay (HEC Montreal grant).

Author(s)

Bruno Remillard

References

Chapter 5 of 'Statistical Methods for Financial Engineering, B. Remillard, CRC Press, (2013).

Examples

data(data.cir)
out = est.cir(data.cir,method='num')

Example output

Loading required package: ggplot2
Loading required package: reshape
Loading required package: corpcor
Are you a satisfied with the graph? 


The estimated coefficients correspond to the annualized spot rate (##)
 Fisher information computed with the numerical gradient (Appendix B.5.1)


 alpha = 0.5092 /+ 1.0909 

  beta = 2.4562 /+ 1.3785 

 sigma = 0.3486 /+ 0.0203 

    q1 = 0.3244 /+ 9.3209 

    q2 = -0.2471 /+ 3.3111 

   phi = 0.9986,  phiest = 0.9986 /+ 0.0030 

SMFI5 documentation built on May 2, 2019, 10:25 a.m.