Inmathematical finance, theCox–Ingersoll–Ross (CIR) model describes the evolution ofinterest rates. It is a type of "one factor model" (short-rate model) as it describes interest rate movements as driven by only one source ofmarket risk. The model can be used in the valuation ofinterest rate derivatives. It was introduced in 1985^[1] byJohn C. Cox,Jonathan E. Ingersoll andStephen A. Ross as an extension of theVasicek model, itself anOrnstein–Uhlenbeck process.

The CIR model describes the instantaneous interest rate${\displaystyle r_t}$ with aFeller square-root process, whosestochastic differential equation is

${\displaystyle d{r}_t=a(b-r_t)dt+\sigma \sqrt{r_t}d{W}_t,}$

where${\displaystyle W_t}$ is aWiener process (modelling the random market risk factor) and${\displaystyle a}$,${\displaystyle b}$, and${\displaystyle \sigma }$ are theparameters. The parameter${\displaystyle a}$ corresponds to the speed of adjustment to the mean${\displaystyle b}$, and${\displaystyle \sigma }$ to volatility. The drift factor,${\displaystyle a(b-r_t)}$, is exactly the same as in the Vasicek model. It ensuresmean reversion of the interest rate towards the long run value${\displaystyle b}$, with speed of adjustment governed by the strictly positive parameter${\displaystyle a}$.

Thestandard deviation factor,${\displaystyle \sigma \sqrt{r_t}}$, avoids the possibility of negative interest rates for all positive values of${\displaystyle a}$ and${\displaystyle b}$.An interest rate of zero is also precluded if the condition

${\displaystyle 2ab\ge {\sigma }^2}$

is met. More generally, when the rate (${\displaystyle r_t}$) is close to zero, the standard deviation (${\displaystyle \sigma \sqrt{r_t}}$) also becomes very small, which dampens the effect of the random shock on the rate. Consequently, when the rate gets close to zero, its evolution becomes dominated by the drift factor, which pushes the rate upwards (towardsequilibrium).

In the case${\displaystyle 4ab={\sigma }^2}$,^[2] the Feller square-root process can be obtained from the square of anOrnstein–Uhlenbeck process. It isergodic and possesses a stationary distribution. It is used in theHeston model to model stochastic volatility.

• Future distribution

The distribution of future values of a CIR process can be computed in closed form:

${\displaystyle r_{t+T}=\frac{Y}{2c},}$

where${\displaystyle c=\frac{2a}{(1-e^{-aT}){\sigma }^2}}$, andY is anon-central chi-squared distribution with${\displaystyle \frac{4ab}{{\sigma }^2}}$ degrees of freedom and non-centrality parameter${\displaystyle 2c{r}_t{e}^{-aT}}$. Formally the probability density function is:

${\displaystyle f(r_{t+T};r_t,a,b,\sigma )=c{e}^{-u-v}{\left(\frac{v}{u}\right)}^{q/2}{I}_q(2\sqrt{uv}),}$

where${\displaystyle q=\frac{2ab}{{\sigma }^2}-1}$,${\displaystyle u=c{r}_t{e}^{-aT}}$,${\displaystyle v=c{r}_{t+T}}$, and${\displaystyle I_q(2\sqrt{uv})}$ is a modified Bessel function of the first kind of order${\displaystyle q}$.

• Asymptotic distribution

Due to mean reversion, as time becomes large, the distribution of${\displaystyle r_{\mathrm{\infty }}}$ will approach agamma distribution with the probability density of:

${\displaystyle f(r_{\mathrm{\infty }};a,b,\sigma )=\frac{{\beta }^{\alpha }}{\mathrm{\Gamma }(\alpha )}{r}_{\mathrm{\infty }}^{\alpha -1}{e}^{-\beta r_{\mathrm{\infty }}},}$

where${\displaystyle \beta =2a/{\sigma }^2}$ and${\displaystyle \alpha =2ab/{\sigma }^2}$.

• Mean reversion,
• Level dependent volatility (${\displaystyle \sigma \sqrt{r_t}}$),
• For given positive${\displaystyle r_0}$ the process will never touch zero, if${\displaystyle 2ab\ge {\sigma }^2}$; otherwise it can occasionally touch the zero point,
• ${\displaystyle \mathrm{E}[r_t\mid r_0]=r_0{e}^{-at}+b(1-e^{-at})}$, so long term mean is${\displaystyle b}$,
• ${\displaystyle \mathrm{Var}[r_t\mid r_0]=r_0\frac{{\sigma }^2}{a}(e^{-at}-e^{-2at})+\frac{b{\sigma }^2}{2a}(1-e^{-at}{)}^2.}$

• Ordinary least squares

The continuous SDE can be discretized as follows

${\displaystyle r_{t+\mathrm{\Delta }t}-r_t=a(b-r_t)\mathrm{\Delta }t+\sigma \sqrt{r_t\mathrm{\Delta }t}{\epsilon }_t,}$

which is equivalent to

${\displaystyle \frac{r_{t+\mathrm{\Delta }t}-r_t}{{\sqrt{r}}_t}=\frac{ab\mathrm{\Delta }t}{{\sqrt{r}}_t}-a{\sqrt{r}}_t\mathrm{\Delta }t+\sigma \sqrt{\mathrm{\Delta }t}{\epsilon }_t,}$

provided${\displaystyle {\epsilon }_t}$ is n.i.i.d. (0,1). This equation can be used for a linear regression.

• Martingale estimation
• Maximum likelihood

Stochastic simulation of the CIR process can be achieved using two variants:

• Discretization
• Exact

Under theno-arbitrage assumption, a bond may be priced using this interest rate process. The bond price is exponential affine in the interest rate:

${\displaystyle P(t,T)=A(t,T)e^{-B(t,T)r_t}}$

where

${\displaystyle A(t,T)={\left(\frac{2h{e}^{(a+h)(T-t)/2}}{2h+(a+h)(e^{h(T-t)}-1)}\right)}^{2ab/{\sigma }^2}}$

${\displaystyle B(t,T)=\frac{2(e^{h(T-t)}-1)}{2h+(a+h)(e^{h(T-t)}-1)}}$

${\displaystyle h=\sqrt{a^2+2{\sigma }^2}}$

The CIR model uses a special case of abasic affine jump diffusion, which still permits aclosed-form expression for bond prices. Time varying functions replacing coefficients can be introduced in the model in order to make it consistent with a pre-assigned term structure of interest rates and possibly volatilities. The most general approach is in Maghsoodi (1996).^[3] A more tractable approach is in Brigo and Mercurio (2001b)^[4] where an external time-dependent shift is added to the model for consistency with an input term structure of rates.

A significant extension of the CIR model to the case of stochastic mean andstochastic volatility is given byLin Chen (1996) and is known asChen model. A more recent extension for handling cluster volatility, negative interest rates and different distributions is the so-called "CIR #" by Orlando, Mininni and Bufalo (2018,^[5] 2019,^[6][7] 2020,^[8] 2021,^[9] 2023^[10]) and a simpler extension focussing on negative interest rates was proposed by Di Francesco and Kamm (2021,^[11] 2022^[12]), which are referred to as the CIR- and CIR-- models.

• Hull–White model
• Vasicek model
• Chen model

• Hull, John C. (2003).Options, Futures and Other Derivatives. Upper Saddle River, NJ:Prentice Hall.ISBN 0-13-009056-5.
• Cox, J.C.;Ingersoll, J.E.;Ross, S.A. (1985). "A Theory of the Term Structure of Interest Rates".Econometrica.53 (2):385–407.doi:10.2307/1911242.JSTOR 1911242.
• Maghsoodi, Y. (1996). "Solution of the extended CIR Term Structure and Bond Option Valuation".Mathematical Finance.6 (6):89–109.doi:10.1111/j.1467-9965.1996.tb00113.x.
• Damiano Brigo; Fabio Mercurio (2001).Interest Rate Models — Theory and Practice with Smile, Inflation and Credit (2nd ed. 2006 ed.). Springer Verlag.ISBN 978-3-540-22149-4.
• Brigo, Damiano; Fabio Mercurio (2001b)."A deterministic-shift extension of analytically tractable and time-homogeneous short rate models".Finance & Stochastics.5 (3):369–388.doi:10.1007/PL00013541.S2CID 35316609.
• Open Source library implementing the CIR process in python
• Orlando, Giuseppe; Mininni, Rosa Maria; Bufalo, Michele (2020). "Forecasting interest rates through Vasicek and CIR models: A partitioning approach".Journal of Forecasting.39 (4):569–579.arXiv:1901.02246.doi:10.1002/for.2642.ISSN 1099-131X.S2CID 126507446.

• Interest rates
• Fixed income analysis
• Stochastic models
• Short-rate models
• Financial models

• Articles with short description
• Short description matches Wikidata