Autoregressive conditional heteroskedasticity

Editor-In-Chief: C. Michael Gibson, M.S., M.D. [1]

Overview

"ARCH" redirects here. For the children's rights organization, see Action on Rights for Children.

In econometrics, an autoregressive conditional heteroskedasticity (ARCH, Engle (1982)) model considers the variance of the current error term to be a function of the variances of the previous time period's error terms. ARCH relates the error variance to the square of a previous period's error. It is employed commonly in modeling financial time series that exhibit time-varying volatility clustering, i.e. periods of swings followed by periods of relative calm.

Definition

Specifically, let <math> ~\epsilon_t~ </math> denote the returns (or return residuals, net of a mean process) and assume that <math> ~\epsilon_t=\sigma_t z_t ~</math>, where <math> z_t\sim iid~ N(0,1) </math> and where the series <math> \sigma_t^2 </math> are modeled by

<math> \sigma_t^2=\alpha_0+\alpha_1 \epsilon_{t-1}^2+\cdots+\alpha_q \epsilon_{t-q}^2 = \alpha_0 + \sum_{i=1}^q \alpha_{i} \epsilon_{t-i}^2 </math>

and where <math> ~\alpha_0>0~ </math> and <math> \alpha_i\ge 0,~i>0</math>.

How to estimate an ARCH(q) model

An ARCH(q) model can be estimated using ordinary least squares. A methodology to test for the lag length of ARCH errors using the Lagrange multiplier test was proposed by Engle (1982). These steps show us how to do it:

1. Estimate the best fitting AR(q) model <math> y_t = a_0 + a_1 y_{t-1} + \cdots + a_q y_{t-q} + \epsilon_t = a_0 + \sum_{i=1}^q a_i y_{t-i} + \epsilon_t </math>.
2. Obtain the squares of the error <math> \hat \epsilon^2 </math> and regress them on a constant and q lagged values:

   <math> \hat \epsilon_t^2 = \hat \alpha_0 + \sum_{i=1}^{q} \hat \alpha_i \hat \epsilon_{t-i}^2</math>

   where q is the length of ARCH lags.

3. The null hypothesis is that, in the absence of ARCH components, we have <math> \alpha_i = 0 </math> for all <math> i = 1, \cdots, q </math>. The alternative hypothesis is that, in the presence of ARCH components, at least one of the estimated <math> \alpha_i </math> coefficients must be significant. In a sample of T residuals under the nullhypothesis of no ARCH errors, the test statistic TR2 follows <math> \chi^2 </math> distribution with q degrees of freedom. If TR² is greater than the Chi-square table value, we reject the null hypothesis and find that there is no ARCH effect in the ARMA model. If TR² is smaller than the Chi-square table value, we accept the null hypothesis.

GARCH

If an autoregressive moving average model (ARMA model) is assumed for the error variance, the model is a generalized autoregressive conditional heteroskedasticity (GARCH, Bollerslev(1986)) model.

In that case, the GARCH(p, q) model (where p is the order of the GARCH terms <math> ~\sigma^2 </math> and q is the order of the ARCH terms <math> ~\epsilon^2 </math>) is given by

<math> \sigma_t^2=\alpha_0 + \alpha_1 \epsilon_{t-1}^2 + \cdots + \alpha_q \epsilon_{t-q}^2 + \beta_1 \sigma_{t-1}^2 + \cdots + \beta_p\sigma_{t-p}^2 = \alpha_0 + \sum_{i=1}^q \alpha_i \epsilon_{t-i}^2 + \sum_{i=1}^p \beta_i \sigma_{t-i}^2 </math>

Generally, when testing for heteroskedasticity in econometric models, the best test is the White test. However, when dealing with time series data, there means to test for ARCH errors (as described above) and GARCH errors (below).

Prior to GARCH there was EWMA which has now been superseded by GARCH. Some people utilise both.

How to estimate a GARCH(p, q) model

This methodology show how to find the lag length p of a GARCH(p, q) process:

1. Estimate the best fitting AR(q) model

   <math> y_t = a_0 + a_1 y_{t-1} + \cdots + a_q y_{t-q} + \epsilon_t = a_0 + \sum_{i=1}^q a_i y_{t-i} + \epsilon_t </math>.

2. Compute and plot the autocorrelations of <math> \epsilon^2 </math> by

   <math> \rho = {{\sum^T_{t=i+1} (\hat \epsilon^2_t - \hat \sigma^2_t) (\hat \epsilon^2_{t-1} - \hat \sigma^2_{t-1})} \over {\sum^2_{t=1} (\hat \epsilon^2_t - \hat \sigma^2_t)^2}} </math>

3. The asymptotic, that is for large samples, standard deviation of <math> \rho (i) </math> is T^{-½}. Individual values that are larger than this indicate GARCH errors. To estimate the total number of lags, use the Ljung-Box test until the value of the these are less than, say, 10% significant. The Ljung/Box Q-statistic follows <math> \chi^2 </math> distribution with n degrees of freedom if the squared residuals <math> \epsilon^2_t </math> are uncorrelated. It is recommended to consider up to T/4 values of n. The null hypothesis states that there are ARCH or GARCH errors. Rejecting the null thus means that there are no such errors in the conditional variance.

IGARCH

Integrated Generalized Autoregressive Conditional Heteroskedasticity IGARCH is a restricted version of the GARCH model, where the sum of the persistent parameters sum up to one, and therefore there is a unit root in the GARCH process. The condition for this is

<math> \sum^p_{i=1} ~\beta_{i-t} ~\sigma^2_t = 1 </math>.

EGARCH

The exponential general autoregressive conditional heteroskedastic (EGARCH) model by Nelson (1991) is another form of the GARCH model. Formally:

<math>\log\sigma_{t}^{2}=\omega_{t}+\sum_{k=1}^{\infty}\beta_{k}g(Z_{t-k})</math>

where <math>g(Z_{t})=\theta Z_{t}+\lambda(|Z_{t}|-E(Z_{t}))</math>, <math>\sigma_{t}^{2}</math> is the conditional variance, <math>\omega</math>, <math>\beta</math>, <math>\theta</math> and <math>\lambda</math> are coefficients, and <math>Z_{t}</math> is a standard normal variable.

GARCH-M

The GARCH-in-mean (GARCH-M) model adds a heteroskedasticity term into the mean equation. It has the specification:

<math> y_t = ~\beta x_t + ~\sigma_t^{1/2} + ~\epsilon_t </math>

The residual <math> ~\epsilon_t </math> is defined as

<math> ~\epsilon_t = ~\sigma_t^{1/2} z_t </math>

QGARCH

The Quadratic GARCH (QGARCH) model by Sentana (1995) is used to model asymmetric effects of positive and negative shocks.

In the example of a GARCH(1,1) model, the residual process <math> ~\sigma_t </math> is

<math> ~\epsilon_t = ~\sigma_t z_t </math>

where <math> z_t </math> is i.i.d. and

<math> ~\sigma_t^2 = K + ~\alpha ~\epsilon_{t-1}^2 + ~\beta ~\sigma_{t-1}^2 + ~\phi ~\sigma_{t-1} </math>

GJR-GARCH

Similar to QGARCH, the The Glosten-Jagannathan-Runkle GARCH (GJR-GARCH) model by Glosten, Jagannathan and Runkle (1993) also models asymmetry in the GARCH process. The suggestion is to model <math> ~\epsilon_t = ~\sigma_t z_t </math> where <math> z_t </math> is i.i.d., and

<math> ~\sigma_t^2 = K + ~\delta ~\sigma_{t-1}^2 + ~\alpha ~\epsilon_{t-1}^2 + ~\phi ~\epsilon_{t-1}^2 I_{t-1} </math>

where <math> I_{t-1} = 1 </math> if <math> ~\epsilon_{t-1} \ge 0 </math>, and <math> I_{t-1} = 0 </math> if <math> ~\epsilon_{t-1} < 0 </math>.

TGARCH model

Finally, the Threshold GARCH (TGARCH) model by Zakoian (1994) is similar to GJR GARCH, and the specification is one on conditional standard deviation instead of conditional variance:

<math> ~\sigma_t = K + ~\delta ~\sigma_{t-1} + ~\alpha_1^{+} ~\epsilon_{t-1}^{+} + ~\alpha_1^{-} ~\epsilon_{t-1}^{-} </math>

where <math> ~\epsilon_{t-1}^{+} = ~\epsilon_{t-1} </math> if <math> ~\epsilon_{t-1} > 0 </math>, and <math> ~\epsilon_{t-1}^{+} = 0 </math> if <math> ~\epsilon_{t-1} \le 0 </math>. Likewise, <math> ~\epsilon_{t-1}^{-} = ~\epsilon_{t-1} </math> if <math> ~\epsilon_{t-1} \le 0 </math>, and <math> ~\epsilon_{t-1}^{-} = 0 </math> if <math> ~\epsilon_{t-1} > 0 </math>.

References

• Tim Bollerslev. "Generalized Autoregressive Conditional Heteroskedasticity", Journal of Econometrics, 31:307-327, 1986.

• Enders, W., Applied Econometrics Time Series, John-Wiley & Sons, 139-149, 1995

• Robert F. Engle. "Autoregressive Conditional Heteroscedasticity with Estimates of Variance of United Kingdom Inflation", Econometrica 50:987-1008, 1982. (the paper which sparked the general interest in ARCH models)

• Robert F. Engle. "GARCH 101: The Use of ARCH/GARCH Models in Applied Econometrics", Journal of Economic Perspectives 15(4):157-168, 2001. (a short, readable introduction)

• Gujarati, D. N., Basic Econometrics, 856-862, 2003

• Nelson, D. B. (1991). Conditional heteroskedasticity in asset returns: A new approach, Econometrica 59: 347-370.

External links

• ARCH and GARCH models for forecasting volatility, quantnotes.com
• SAS examples of estimating GARCH models, support.sas.com