Stochastic Modelling

Exercises on Time Series

Dr. Iqbal Owadally

March 3, 2003

Elementary Problems
Q1. Rewrite the following time series models using the backward shift notation. Classify
each of them as an ARIMA(p, d, q) process (that is, determine p, d and q in each case).
State whether each is (i) stationary, (ii) invertible.

(i) Xt = 0.5Xt1 + Zt
(ii) Xt 0.5Xt1 = Zt 1.3Zt1 + 0.4Zt2
(iii) Xt 1.5Xt1 + 0.6Xt2 = Zt
(iv) (Xt 0.2) 1.2(Xt1 0.2) + 0.2(Xt2 0.2) = Zt 0.5Zt1

Q2. The following autoregressive processes are stationary. Calculate 1 , 2 and 3 .

(i) Xt + 0.5Xt1 0.1Xt2 = Zt ;

(ii) Xt = 0.6Xt2 + Zt ;
(iii) (1 1.1B + 0.18B 2 )Xt = Zt ;
(iv) Xt = Xt1 2 Xt2 3 Xt3 + Zt .

Q3. Calculate 1 and 2 for the following MA processes:

(i) Yt = Zt Zt1
(ii) Yt = (1 + 2.4B + 0.8B 2 )Zt

Q4. (i) Describe the key difference between the correlograms of a stationary AR process
and an MA process of the same order.
(ii) Derive the autocorrelation function for the stationary ARMA(1, 1) process:
(Xt ) (Xt1 ) = Zt Zt1 .
(iii) Comment on the correlogram of the ARMA(1, 1) process above.

Elementary problems should be attempted first. Past exam questions are included for exam practice and
could be attempted later. They are adapted from papers set by the Exam Board of the Institute and Faculty
of Actuaries. Papers set by the Exam Board, Faculty of Actuarial Science and Statistics, Cass Business School,
City University, are separately available.

Contact details: Cass Building Room 5071, extension 8478, iqbal@city.ac.uk.

1
Q5. The AR(1) process Xt 0.4Xt1 = Zt is started arbitrarily at t = 0 with initial condition
X0 = x0 R. The sequence {Zt } is a set of independent and identically distributed
random variables with zero mean and variance Z2 .

(i) Derive the mean of the process for t 0.

(ii) Derive the autocovariance function of the process for t 0.
(iii) Explain why the process is non-stationary for 0 t < .
(iv) Explain why the process is stationary in the limit as t .
(v) Explain how the initial condition x0 may be chosen so that the process is stationary
for t 0.

Q6. (i) Show that {Wt } is (weakly) stationary if Wt = Xt + Yt and {Xt } and {Yt } are
independent and both are (weakly) stationary.
(ii) The logarithmic (or geometric) rate of inflation Xt is calculated from the consumer
prices index CPIt as follows: CPIt = CPIt1 exp(Xt ). It is frequently modelled as
follows: (Xt ) = (Xt1 ) + Zt .
Show that ln CPIt is then an ARIMA(1, 1, 0) process.

Past Exam Questions

Q1. (i) (a) Calculate the autocovariance function {k : k 0} and autocorrelation func-
tion {k : k 0} of a first-order Moving Average process

Xt = + et + 1 et1 ,

where {et : t 0} is a sequence of uncorrelated, zero-mean random variables

with common variance e2 .
(b) State the conditions on the values of the parameters such that the process is
invertible.
(ii) A sequence of observations x1 , x2 , . . . , xn has sample variance 0 = 14.5, sample
lag-1 autocovariance 1 = 5.0. Show that there is more than one first-order moving
average process which can be fitted to these data, but verify that only one of the
fitted processes is invertible.
(Exam Board of the Institute and Faculty of Actuaries)

Q2. A stationary second-order autoregressive process X, which may be assumed to be in

equilibrium at time 0, is defined by

Xt = + 1 (Xt1 ) + 2 (Xt2 ) + et ,

where {et : t 1} is a sequence of independent, zero-mean Normal random variables,

each with variance e2 .

(i) (a) Obtain an equation for 1 in terms of 0 and 2 by substituting for Xt in the
equation 1 = Cov [Xt , Xt1 ].
(b) Derive similar equations for 2 and 0 .
(c) State the autocorrelation function k of X for k = 0, 1, 2.

2
 (ii) Suppose that the equations derived in (i) for 1 and 2 are used as the basis of
an estimation procedure: estimates
1 and 2 are defined to be the solutions of
those equations when is replaced by a suitably-defined sample autocorrelation
function r.
Solve these equations.
(Exam Board of the Institute and Faculty of Actuaries)

Q3. A stationary stochastic process {Yt : t = 0, 1, . . . } satisfies the relationship

Yt = + 0.8(Yt1 ) 0.4(Yt2 ) + et ,

where {et : t = 0, 1, . . . } is a sequence of independent, zero-mean Normal random

variables with common variance 2 .

(i) Calculate the autocorrelation function k , and the partial autocorrelation function
k of Y for k = 1 and 2.
(ii) State, without performing additional calculations, what you would expect to find
if you were to calculate k and k for larger values of k.
(Exam Board of the Institute and Faculty of Actuaries)

Q4. (i) State the purpose of each of the stages of identification, estimation and diagnosis
of an observed time series, according to the Box-Jenkins approach to time series
modelling.
(ii) Describe briefly three criteria you would use to determine the most appropriate
value for the parameter d, the order of integration.
(iii) In the context of fitting an ARIMA time series model, what is meant by the term
parsimony ?
(iv) Fitting an AR(1) model to the annual force of price inflation in the UK over the
period 19191997, the sample autocorrelations of the 78 model residuals for lags 1
to 10, denoted by r1 , r2 , . . . , r10 are found to be:
-0.07, 0.08, 0.09, 0.20, 0.07, 0.18, -0.01, 0.05, 0.13, 0.01.
Carry out a portmanteau test and state your conclusion.
(v) Also, the set of model residuals contains 48 turning points. Carry out another
goodness of fit test on the model and comment on your results.
(Exam Board, Faculty of Actuarial Science and Statistics, Cass Business School, City University)

Q5. The movements of a consumer price index are to be subjected to time series analysis
with the aim of forecasting future behaviour. The index is calculated monthly.

(i) Explain whether you would expect to fit a model which included (a) a trend term,
(b) a seasonal effect.

The values {xt : 1 t n} are the residuals which remain once any trend or seasonal
variations have been removed. An ARIMA(1, 1, 1) model is to be fitted to the {xt }.

(ii) (a) Assuming the ARIMA(1, 1, 1) model is correct, write down an equation for
Xn+1 in terms of the white noise process {et : 1 t n + 1} and the
observations {xt : 1 t n}.

3
 (b) State the parameters of the model.
(iii) The Box-Jenkins procedure defines the k-step-ahead forecast for X to be

x
n (k) = E [ Xn+k | xn , xn1 , . . . , x1 ] .

(a) Derive the 1-step-ahead and 2-step-ahead forecasts for X for the ARIMA(1,1,1)
model, assuming that the values of the parameters and the value of e0 are
known exactly.
(b) Evaluate the prediction variance Var(Xn+1 x n (1)), again assuming that the
values of the parameters are known.
(iv) The most elementary form of the technique known as exponential smoothing pro-
duces at time n a 1-step-ahead forecast xn defined by

xn = xn + (xn1 xn ).

for some (0, 1) which may be chosen by the user.

Show that, for particular values of the autoregressive and moving average param-
eters, the Box-Jenkins forecasts above coincide with the forecasts produced by
exponential smoothing.
(v) (a) State whether the ARIMA(1, 1, 1) model is I(0), I(1) or neither.
(b) Discuss whether there is a difference between an I(0) model and an I(1) model
in terms of the conditional distribution of Xn+k given {xt : 1 t n} for
large values of k.
(vi) It is suggested that a salaries index might be cointegrated with the consumer price
index.
(a) Explain what is meant by the suggestion.
(b) Comment on whether it is a reasonable suggestion.
(Exam Board of the Institute and Faculty of Actuaries)