ARCH and GARCH models

Fulvio Corsi

SNS Pisa

5 Dic 2011

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 1 / 21
Asset prices

S&P 500 index from 1982 to 2009

1600

1400

1200

1000

800

600

400

200

0
1980 1985 1990 1995 2000 2005 2010

asset prices are typically integrated of order one processes I(1)

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 2 / 21
Asset returns

the standard solution is to take the first difference of prices. Two different type of returns:

1 simple net return

Pt − Pt−∆
Rt =
Pt−∆

2 log return or continuously compounded returns

rt = log Pt − log Pt−∆ = pt − pt−∆

over short horizon ∆, rt is typically small (|rt | << 10%) so that Rt ≈ rt being
1 2
1 + Rt = exp(rt ) = 1 + rt + r + ...
2 t

The main advantage of the log return is that a k-period return rt (k) is simply:
rt (k) = pt − pt−k∆ = rt−(k−1)∆ + ... + rt−∆ + rt
hence multi-period log returns are simply the sum of single-period log returns.

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 3 / 21
Asset returns dynamics
USD/CHF 1989−2001
6

std deviations
0

−2

−4

−6

−8
0 500 1000 1500 2000 2500 3000

Gaussian noise
6

2
std deviation

−2

−4

−6

−8
0 500 1000 1500 2000 2500 3000

S&P 500 1990−2003

6

2
std deviations

−2

−4

−6

−8
0 500 1000 1500 2000 2500 3000

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 4 / 21
Time changing volatility

Dynamics in the volatility of asset returns has paramount consequences in important finance
applications:

asset allocation

risk management

derivative pricing

What makes volatility change over time? Still unclear.

“event-driven volatility”: different information arrival rate, consistent with EMH

“error-driven volatility”: due to over- and underreaction of the market to incoming information

“price-driven volatility”: endogenously generated by trading activities of heterogeneous

agents ⇒ strong positive correlation between volatility and market presence

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 5 / 21
Different volatility notions

Different types of volatility approaches:

Parametric: volatility measure is model-dependent

- Discrete-time: ARCH/GARCH models

- Continuous-time: Stochastic Volatility models

Non-Parametric: volatility measure is model-independent (or model-free)

- Realized Volatility (exploiting the information in High Frequency data)

Different notions of volatility:

ex-ante conditional volatility

spot/instantaneous volatility

ex-post integrated volatility

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 6 / 21
Basic Structure and Properties of ARMA model

standard time series models have:

Yt = E[Yt |Ωt−1 ] + t
E[Yt |Ωt−1 ] = f (Ωt−1 ; θ)
h i
Var [Yt |Ωt−1 ] = E 2t |Ωt−1 = σ2

hence,

Conditional mean: varies with Ωt−1

Conditional variance: constant (unfortunately)

k-step-ahead mean forecasts: generally depends on Ωt−1

k-step-ahead variance of the forecasts: depends only on k, not on Ωt−1 (again
unfortunately)

Unconditional mean: constant

Unconditional variance: constant

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 7 / 21
AutoRegressive Conditional Heteroskedasticity
(ARCH) model

Engle (1982, Econometrica) intruduced the ARCH models:

Yt = E[Yt |Ωt−1 ] + t
E[Yt |Ωt−1 ] = f (Ωt−1 ; θ)
h i
Var [Yt |Ωt−1 ] = E 2t |Ωt−1 = σ (Ωt−1 ; θ) ≡ σt2

hence,

Conditional mean: varies with Ωt−1

Conditional variance: varies with Ωt−1

k-step-ahead mean forecasts: generally depends on Ωt−1

k-step-ahead variance of the forecasts: generally depends on Ωt−1

Unconditional mean: constant

Unconditional variance: constant

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 8 / 21
ARCH(q)
 
How to parameterize E 2t |Ωt−1 = σ (Ωt−1 ; θ) ≡ σt2 ?

ARCH(q) postulated that the conditional variance is a linear function of the past q squared
innovations
Xq
σt2 = ω + αi 2t−i = ω + α(L)2t−1
i=1

Defining vt = 2t − σt2 , the ARCH(q) model can be written as

2t = ω + α(L)2t−1 + vt
Since Et−1 (vt ) = 0, the model corresponds directly to an AR(q) model for the squared
innovations, 2t .

The process is covariance

P stationary if and only if the sum of the positive AR parameters is
less than 1 i.e. qi=1 αi < 1. Then, the unconditional variance of t is

Var(t ) = σ2 = ω/(1 − α1 − α2 − ... − αq ).

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 9 / 21
ARCH and fat tails
Note that the unconditional distribution of t has Fat Tail.

In fact, the unconditional kurtosis of t is

E(4t )
E(2t )2
where the numerator is

h i h i
E 4t = E E(4t |Ωt−1 )
h i
= 3E σt4

= 3[Var(σt2 ) + E(σt2 )2 ]
= 3[Var(σt2 ) +E(2t )2 ]
| {z }
>0

> 3E(2t )2 .

Hence,
E(4t )
Kurtosis(t ) = >3
E(2t )2

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 10 / 21
ARCH and fat tails: intuition

rt ∼ N(µ, σt ) is a mixture of Normals with different σt ⇒ fatter tails

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 11 / 21
AR(1)-ARCH(1)

Example: the AR(1)-ARCH(1) model

Yt = φYt−1 + t
σt2 = ω + α2t−1
t ∼ N(0, σt2 )

- Conditional mean: E(Yt |Ωt−1 ) = φYt−1

- Conditional variance: E([Yt − E(Yt |Ωt−1 )]2 |Ωt−1 ) = ω + α2t−1

- Unconditional mean: E(Yt ) = 0

1 ω
- Unconditional variance: E(Yt − E(Yt ))2 = (1−φ2 ) (1−α)

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 12 / 21
Other characteristics of ARCH(1)
for the ARCH(1) model

Yt = µ + t
σt2 = ω + α2t−1
t ∼ N(0, σt2 )
we also have:

- Excess kurtosis:
E(4t ) 1 − α2
Kurtosis(t ) = 2
= 3
E(t )2 1 − 3α2
kurtosis is equal to 3 iff α = 0

- Stationarity condition for finite variance of 2 (or for finite kurtosis of )

1
α < √ ≈ 0.577
3

- Autocorrelation of σt : Corr(σt , σt−h ) = αh ⇒ difficult to replicate empirical persistence of σt .

⇒ ARCH(q) models but ...

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 13 / 21
GARCH(p,q)

ARCH(q) problem: empirical volatility very persistent ⇒ Large q i.e. too many α’s

Bollerslev (1986, J. of Econometrics) proposed the Generalized ARCH model.

The GARCH(p,q) is defined as

q
X p
X
σt2 = ω + αi 2t−i + 2
βj σt−j 2
= ω + α(L)2t−1 + β(L)σt−1
i=1 j=1

As before, defining vt = 2t − σt2 , the GARCH(p,q) can be also rewritten as

2t = ω + [α(L) + β(L)] 2t−1 − β(L)vt−1 + vt

which defines an ARMA[max(p, q),p] model for 2t .

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 14 / 21
GARCH(1,1)
By far the most commonly used is the GARCH(1,1):
t = σt zt with zt ∼ i.i.d.N(0, 1),
σt2 = ω + α 2t−1 + βj σt−1
2
with ω > 0, α > 0, β > 0

Stationarity conditions:
being

σt2 = ω + (αz2t + β)σt−1

2

and hence,

E[σt2 ] = ω + (α + β)E[σt−1
2
]

then the process is covariance stationary iff

α+β < 1

Unconditional variance: ω
Var() =
1 − (α + β)

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 15 / 21
GARCH(1,1)

By recursive substitution, the GARCH(1,1) may be written as the following ARCH(∞):

∞
X
σt2 = ω(1 − β) + α β i−1 2t−i
i=1
which reduces to an exponentially weighted moving average filter for ω = 0 and α + β = 1
(sometimes referred to as Integrated GARCH or IGARCH(1,1)).

Moreover, GARCH(1,1) implies an ARMA(1,1) representation in the 2t

2t = ω + (α + β)2t−1 − βvt−1 + vt

From the ARMA(1,1) representation we can guess that

ρh ≡ Corr(σt , σt−h ) ≈ (α + β)h .
The precise calculations give:
α(1 − β 2 − αβ)
ρ1 = , ρh = (α + β)ρh−1 for h>1
1 − β 2 − 2αβ

Forecasting. Denoting the unconditional variance σ2 ≡ ω(1 − α − β)−1 we have:

2
σ̂t+h|t = σ2 + (α + β)h−1 (σt+1
2
− σ2 )
showing that the forecasts of the conditional variance revert to the long-run unconditional
variance at an exponential rate dictated by α + β
Fulvio Corsi ARCH and ()
GARCH models SNS Pisa 5 Dic 2011 16 / 21
Asymmetric GARCH
News Impact Curve
8
symmetric GARCH
asymmetric GARCH
In standard GARCH model:

conditional variance
6

σt2 =ω+ 2
αrt−1 + 2
βσt−1
4

σt2 responds symmetrically to past returns.

2

The so called “news impact curve” is a parabola

0
−5 0 5
standardized lagged shocks

Empirically negative rt−1 impact more than positive ones → asymmetric news impact curve

GJR or T-GARCH

1 if rt < 0
σt2 = 2
ω + αrt−1 2
+ γrt−1 2
Dt−1 + βσt−1 with Dt =
0 otherwise
- Positive returns (good news): α
- Negative returns (bad news): α + γ
- Empirically γ > 0 → “Leverage effect”

Exponential GARCH (EGARCH)

rt−1 rt−1
ln(σt2 ) = ω + α +γ 2
+ βln(σt−1 )
σt−1 σt−1

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 17 / 21
GARCH in mean

In the GARCH-M (Garch-in-Mean) model Engle, Lilien and Robins (1987) introduce the
(positive) dependence of returns on conditional variance, the so called “risk-return tradeoff”.

The specification of the model is:

rt = µ + γσt2 + σt zt
σt2 = 2
ω + αrt−1 2
+ βσt−1

Given the inherent noise of financial returns rt , the estimates of γ are often very difficult,
typically long time series are required to find significant results.

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 18 / 21
Engle test for heteroscedasticity or ARCH test

The ARCH test of Engle assesses the null hypothesis that a series of residuals t exhibits no
conditional heteroscedasticity (ARCH effects),

The test is performed by running the following regression

2t = c + a1 2t−1 + a2 2t−2 + ... + aL 2t−L

then computes the Lagrange multiplier statistic T × R2 , where T is the sample size and R2 is the
coefficient of determination of the regression.

Under the null, we have that

T × R2 → χ2L

i.e. the asymptotic distribution of the test statistic is chi-square with L degrees of freedom.

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 19 / 21
Estimation
A GARCH process with gaussian innovation:
rt |Ωt−1 ∼ N(µt (θ), σt2 (θ))

has conditional densities:

 
1 1 (rt − µt (θ))
f (rt |Ωt−1 ; θ) = √ σt−1 (θ) exp −
2π 2 σt2 (θ)

using the “prediction–error” decomposition of the likelihood

L(rT , rT−1 , ..., r1 ; θ) = f (rT |ΩT−1 ; θ) × f (rT−1 |ΩT−2 ; θ) × ... × f (r1 |Ω0 ; θ)

the log-likelihood becomes:

T T
T X 1 X (rt − µt (θ))
log L(rT , rT−1 , ..., r1 ; θ) = − log(2π) − log σt (θ) −
2 t=1
2 t=1 σt2 (θ)

Non–linear function in θ ⇒ Numerical optimization techniques.

When innovations not Normal → PMLE standard errors (“sandwich form”)

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 20 / 21
Stochastic Volatility models (idea)

In ARCH-GARCH models, the variance at time t, σt2 is completely determined by the

information at time t − 1, i.e. σt2 it is conditionally deterministic or Ft−1 measurable

Another possibility is to have σt2 being also a (positive) stochastic process i.e. variance is
also affected by an idiosyncratic noise term ⇒ Stochastic Volatility models

Example: the Heston model

p
dP(t) = µP(t)dt + h(t)P(t)dW P (t)
p
h(t) = k(θ − h(t)) + ν h(t)dW h (t) CIR process

where dW P (t) and dW h (t) are two (possibly correlated) Brownian processes.

Fulvio Corsi ARCH and ()

GARCH models SNS Pisa 5 Dic 2011 21 / 21