Professional Documents
Culture Documents
.
$
%&
% # !"
! ( # '
HAMAD A. AL-GHANNAM
Associate Prof.- Department of Economics
College of Administrative Sciences- King Saud University
Abstract
This paper aimes to study and investigate the time series of inflation in
Saudi Arabia covering the period from February 1980 to September
2004, in order to identify the change pattern of inflation rate and then
to build a model for forecasting in short run. The statistical time
series stationarity tests such as ADF test and the autocorrelation
function show that the inflation rate is stationary since the inflation
rate is calculated as a difference of the CPI in log form. The BoxJenkins methodology has been applied on selected models using some
statistical tests such as the stationarity tests of the residuals and the
Akaike and Schwarz criterions and the forecast errors to choose the
right model.
The results show that the appropriate model is an ARMA (13,33)
model which may imply some kind of seasonality.
-1-
) *
% &' ! $ (2004 1980
! "#
$
%
% 0 1
/ /+ ) *+ , ! - .
/
7+ *& 8+ 0
9+2: 5+3+ +6+ .20 &3
.+6 78+ $
(ADF)
=% +>
3 ?# @+'
/ A+ A3 ! + !' !
"%/ %
/
(Box-Jenkins)
(ACF)
+% , 53 ) +6 9+2: + B + $ "+ )+
!' %% ! D .1 C.
(Schwarz) (Akaike)
.% % 6 ! <& )
7++ 5+ )%
+ +
/ !,8, 8, " / 8, 1.+ 7
$
. "
/ 7,C B &
-2-
+ I
= ! +6 + / = .
+2=6 7- ' !
$% +2=6 7- B
/ E+,F 7 '
!,+
!,+% B J = .I +,6
G+ +. , 7- E A=8/
! 7
2+ +& +,F + ! = 0 76 !'
J% %, ! +,6 K+3 "+ ++ G+ "+
/ !' !
$% 8 +&: ,3 !M
+N = D2 +/ "
! (L
7+ !+. ++ 72=+ = 9 = !' + ' !
$ +/
,1% +% # &+: D2 !,
+ $ (&'
%,3 %6 !9
/
+,O
+:+ .)+: K+3
/ +N
% +2=
! & +2=6
/
+,O E .&+:
!+ ! 0 + 7++ 8 ! $ ++,F ! # +&8/ 7+
P0
, +2=6 +
+ B 7++ E 5+2 "+
( [1] % J% B# ! 9# +& + L J
G+
.[2]
%9 %3 !% B# 7+, J
/ L
+ 72'
2
" 7+.. $ B < + . $
2+ * +2=6 7 5= $ .0 +++ ' +
&' !
2 +C ! 70 + &' ! $ 5= !/ +3 70 + '
$ 0 1 ' $ &' ! + + 7+6
/'
% $% E% = ! D2 1 &' ! ?= )
0 ", 7+ 7/ ?
. ! .7++ E 7+1
.[7] ([6] ( [5] ( [4] ( [3] :+,
/ L 1
+ 9+2: 5+3+
?%3 7++%2= %
1 7+ ! / &
!% N% N*& 7 7+ 7
= .0 7++2=6
+ 9 +
%
/ 5+-
7+ E .
22 +2=6 7+ 7+'
-3-
"Philips Curve"
@
(Clementsm)
"
% ", 7+
IMF
% %8 .0 7++ <& ! /& 7.> [11]
Veiga)
.(Panal data)
%0 1% %&3
9+2: 5+3 B 7+ E 7 =
L .1
/ "0
#
9+2: 5+3 E ! +0 &3 '
$% .% .+6 ) & + 1 7 7+ E
% 7% (+%, %%%
.J' 9+2 5+' < ' E
%+ $%
ARIMA
(Bos, et al)
(ARIMA)
(Meyler, et al.)
)+% B% ?# 7' ." 7+6
5 + + 1
% %
. "# @# 1
/ "0
)+ ! +L
/
K+% &' !
ARIMA
)
VAR
ARIMA
(Valle)
ARIMA
) + 1 ?# 7' = . 0 1
%
.CPI !
% +' 1
7- +
2 7#
%0 1%
ARIMA
%,
ARIMA
(Hubrich)
(Bokil and
[16]
VAR
VAR
VAR
(Bermingham)
(Moser, et al)
ARIMA
%
Schimnelpfening)
.J3 #
?.+ +0 7++
20
-4-
7 N'
ARIMA
406
-5-
(%32.51)
. ,C
.. +/
.76+26 0
.D
(%1.26)
& + 20 J
1
/ /+ ) *+ &' ! $
$% .% % .+%6 ) + '
%+ + =
(Box-Jenkins)
& @
+%, *& D 0 E
+ .*&' ' 0 .) *+ .0
% G9+% %+, *%& B . & @ ) +0 TL +.:
%' %+3 .+%0
/ T 28 < *&
+ (+0
.G9+
-6-
+ ,- .' II
%
( t)
%
1
.0
+ + - = +
8 )+ '
$% .% % %6 )% )+% E ' ! .20 &3
& @ ) +N +' T Autoregressive Moving Average Model
(ARMA)
7%- 0 1
+N / - 7 )+ , ' ! ) c ([21]
A%
+:+ [23] ([22] +2=6 7+& <
B + , +2=6
%
!%#
!&
%
ARMA
) <& .[24] ([6] ([5] ([3] ([4]
:O+
t-i
t = + 1 t 1 + 2 t 2 + L + p t p
(1)
+% A
/ 1.+ 7
T+ ( 6
,
MA(q) AR(p)
(t
(2)
% ?
C.
ARMA
t = + 1 t 1 + 2 t 2 + L + p t p + t + 1 t 1 + 2 t 2 + L + q t q
-7-
+
t = + 1 t 1 + 2 t 2 + L + q t q
/ ?
. A
/ ($ .
(1)
(2)
+
) +&:
(3)
.5% %
/ $% . 6 78+ i i
$ . 6 ) !M
.0+ +
q
:+
/ 5
ARMA (p, q)
!
ARMA
)% 0 5
.
(d)
. ?#
ARMA
)
:O+ 5
ARIMA (p, d, q)
P
ARIMA
ARIMA(p, d, q)
(
%23
! + 8 !
(d)
= 0
! . +
& 8 ?# /
%
([26] ([25]
Unit Root
!M
+
7+ M
(Dickey - Fuller)
.[27]
5
.
> 0)
"&
. + >
!' +0 +N #2 !/
78+
-8-
(ACF)
9+2
2
Q = n k2 ~ (m)
k =1
. & =
7+% %
%2+
%
' +
n
Ljung-Box (LB)
9+2 ! +
: Q 9+2 ! 6
N [29] "-2
k2 2
~ (m)
LB = n(n + 2)
k =1 n k
m
?# *& 5
. AM
+ >
!' 7+6 E ' 8 ! D M
?%# %
(d)
(p)
p)
3 ". E ($ . .+6 ! +
Residuals
)
MA(q)
% %
PACF
(SBC)
'
rk
ACF
.p % +, = #2
(SBC) (Schwarz)
(AIC)
'
PACF
(ACF)
kq
0
+N 9+2 N#2 T+ 6
(AIC) (Akaike)
AR(p)
% = =' A T ) + 5+ ) +6
:(Estimation) ) 0 b5
(p + q + 1)
/
/ 2 ) 0 (q (
%& %& +
2+ . > 0 0. "+/ ("0 +
-9-
% "+%/
E
.%
> C.
/ T T $ .
+N % "%%0 %& ) + ! "+/ .+ !+0 +0 )+ "/ 0
)% "& @+0 = < [& +0 ! + ("0 #2 !/
.[33]
:(Diagnostic Cheking) 2 P# b)
! C
+ ) P
) +
<
E
% !' !% C J' "+ ($ . ' .+6 !
e + A'
.(White Noise) B3 C. . ?+. ) C.
% .+6 78+ .+6 78+ ?. !/ ) P
"+/
% %% %%0 !% + .
23
@% 0 ) = 9&
"0 + 8 / ! C
(The Sum of The Squared
%0 1%
)% % %, , 3 ) 0 .Forecast
Errors)
.[34]
00 = 0 ! 1 C. 5 $ 7+
:(Forecasting) 1 b
8% ?%. !/ $ 1
A% (E%0 5+ ) +
(% !%
C. %0 %0 =% <+ -
+ + 0
+%\ "% "# 1 % ((t+1) + C 3 %0
/ %2
$
3 0 8M
( t+2)
+, 0
/ 2 ! ((One period ahead forecasting)
MA
(t+1)
#
E
(
.
.#2 T+
)+
(Seasonality)
)8/
% 7%
8% %. 7%- (I # / 2+ ?+#: ,
"+/ .
!+ "+ 1
7- ! [ +& T1 ("%
(PACF) (ACF)
- 10 -
<% 7+%%+
4, 8, 12,
71.+ /
%2/ ' % % % 2/
+ 7- )8 .%
%
%
5%
. + .)
7# / $ .
< ?# ' 7++
/ +, ?# , +N +' ?# 3
)% 5 + E
("+ > ?#
+:+
2# 7++
:+ +
ARIMA (p, d, q) (P, D, Q)
AR
.
MA
D ! 7+, C. !+ !' B# "+/ A' +6+ "& +, .0
( * 7# / " 7+
0 ! 7
8 ! ,
.(Conditionally
Hetroskedastic)
1% "=
/ +N
,1 % !M
%
!+ 8% +%&
.)
> 8
N , !+ !+
.
!+ . ) < + [5]
(Engle)
7,
% %
/ .(ARMA) ) + !+
7-
/ T' 0 !+ :+
7+,
AR(p)
% %6 .% !+% 8 ) . e 2t = < +
.Autoregressive Conditional Heteroskedastic (ARCH)
)% %
/ . !+ !
ARCH
) [24]
(Bollerslev)
. +
)% A%
/ ?
. .+N $ . )
/ T
A% %
Generalized ARCH
%
/ % % .%
/
MA AR
5 q
ARMA
PACF ACF
- 11 -
%
GARCH
'
GARCH(p, q)
ARCH
%&
- 12 -
(t)
t = log(CPI t ) log(CPI t 1 )
(CPI)
.(IMF) % %0 ?2 [20] T 0 1 7++ +
% % %
(1980)
(296)
.(2004)
%
+ E+&6 L8 ! <
(1)
D
(%
%
!
/ N 1 . + ! < - 6
. !'
(PACF)
(PACF)
+%
(ACF)
(ACF)
33)
(ACF)
.&
G9+ 7
t
(ADF)
(Disturbance)
9+2 =
(Level)
+%
%
!' + #
71.+ /
J /
!'
(%1) (%5)
(1)
&
~ I ( 0)
#2 & !
+
- 13 -
0.06
0.04
0.02
0.00
-0.02
-0.04
-0.06
82
84
86
88
90
92
94
96
98
00
02
04
(ADF)
2.% (
4.
-11.639
-11.661
-9.571
-9.588
-7.604
-7.618
-6.764
-6.777
?%
%
(1-)
%& %
+ %
(PACF) (ACF)
%0
$ (+ E A T . + ! + ! N +N + & (
(PACF) (ACF)
7+&%
(ARMA)
(AR)
(PACF) (ACF)
- 14 -
& ! .+6
= #
ARMA (0, 0)
(ACF)
%
(33)
1.+% %/ $% ((13) 1.+ / 1 & L8 N'
!%/ %
+N % T' (0, "
)+ ! !
+ ((PACF)
E% %
/ *a +% .)% *+ / +/6
' " + #2
:+ )+ 0 0 5+& B 7+/6
ARMA (0,0)
ARMA(0,0) (2,0)
t = + t
t = + 13 t 13 + 33 t - 33 + t
ARMA(0,0) (0,2)
t = + 13 t 13 + 33 t - 33 + t
ARMA(0,0) (1,1)
t = + 13 t 13 + 33 t -33 + t
ARMA(0,0) (1,1)
t = + 33 t 33 + 13 t -13 + t
ARMA(0,0) (1,2)
t = + 13 t 13 + 11 t -11 + 33 t -33 + t
(2)
& P
%
/+% +3 .+0 B $+ !' 6 )+ ! N +N +
=%
Q(33)
(2)
%= [+%# 3 )% %0 ! D .5+ ) +
- 15 -
(PACF) (ACF)
(AIC)
(13)
% ! 3 ) *+,+ )+ 0 G9+ L8 + ." &
(SBC)
(Average Trend)
E+& . !/
.(1
!% =
PACF ACF
!3 $ (@+ )
MA(11)
)
<
<
<
<
>
>
2
)=
=
0.127
0.136
-0.27
0.158
0.10
-0.460
-0.443
-0.45
(-8.19)
(-8.79)
0.0074
0.0072
0.008
(2.211)
0.107
(2.08)
(-8.58)
0.00714
(-4.72)
(2.58)
(2.39)
(2.02)
6 <
0.00035
0.145
13
33
11
13
33
0.0074
0.0056
(2.46)
-0.26
(-4.58)
(1.07)
< [&
SSR
-10.562
-10.523
6.499
-10.459
-10.431
8.47
-10.555
-10.529
9.80
-10.507
-10.482
9.98
-10.457
-10.430
8.32
-10.381
-10.368
7.91
(0.59)
(0.49)
(0.37)
(0.35)
(0.50)
(0.72)
10.87
11.42
14.13
15.14
11.07
16.27
(0.93)
(0.93)
(0.82)
(0.77)
(0.95)
(0.80)
17.38
18.59
20.85
24.47
18.48
44.2
(0.97)
(0.96)
(0.92)
(0.79)
(0.96)
(0.09)
- 16 -
=
AIC/
SBC
Q(11)
Q(22)
Q(33)
SBC
AIC
(Forcast Error)
(0.00389 0.00376)
/ @+ < !&
@ % %& !
< ) +/ A
/ .< )
+%:+ [33] 1 + ! + , 3 )
/ + =3 ) # &
D%2 A
/ (#2 !/
6 +N / T+ 1.+
PACF ACF
!'
+ "+ !' + ) D
8,% "%# %+ 0
(Weighted Average)
(t)
9% *+%.3 . D& .+ + 9 C.
+:+ " /
()%
+
71.+ B L# .!%,8, 8, "%# +
+%L # !' ! A' + ( 7 > 7+ 71.+ 7 +
A' !
(Data-generating
E%# ! ) +9+0 !M $ ! >+ .[33] ! ) +N 0.
Processes)
!% <6 (" / +, 7#
L "+/ 7- +'
/
% 7+%+
7++ !' $
5 . 7++
I
%
% ?+#: 78+ 5
>' + (8) @+'
/
?+%#: % B
+:+ (&) 0
+N
%& + ! =' 0 !'
+ .8
N , !
,8, #
++ (0 !/
+N / '
"#
,C L # + (
+ T= $+ +N N !,8,
<%& %
!,8, 8, "#
D ,C $ .@+ )
" / +
- 17 -
% \ +N % !
% +3 +0 +=3 5+ <= !
.+N 0 )+
.)
+N
& !+ 0
!+% 8%6 <
7+ + + 5+ ) +
B% L8 (1 ) ! <
(t)
L 8 ! ..
.N%& "% 7 7+ !' 7+ 9' 7++, '
+/ - 7+
0
(2b & L) 0 )
= <
8,% %
*+,%+
(0.03)
(PACF) (ACF)
! ='
(PACF) (ACF)
% ! )
+ (Q 9+2 = B+# $ (0.099 ) !,8,
7
(LM)
0.06
% & B
= ++ &
TR
+ J 0+ "
= & E .. !+ 8
06 5 <& . 7++ +
6 7+ "
@ +3
&3 0
+:+ 7 +' 0 , ! (+ 2
.T +2=8 K+#6 "
D% "
0 1
+ ) ?. N'
% (FRLCPI = 0
RLCPI
23
= ?
(Wb)
1% %+ $ "
8 = 00 0 ! 5+0 D
.+\ " "#
- 18 -
IV
$ (
= E @9
% 0 1
/ /+ ) *+ , ! . - .
/ &3
0 1 ! + 7++ $ &' ! .20 &3
<+
(1980)
+, ! "#
$ ( 0 ?2 T
(ARMA)
+ , 5+ ) +6 9+2: + B + .(Box-Jenkins)
!+% 8 7+ 1 C.
(Schwarz) (Akaike)
.+ )
(ARCH)
- 19 -
7
PACF ACF "
& " % :(1-) (
- 20 -
% 2 PACF ACF " & " % :(2-) (
- 21 -
FRLCPI
= 0
RLCPI
23
= :(@#) "
.06
.04
.02
.00
-.02
-.04
-.06
1985
1990
1995
2000
RLCPI
.025
.020
.015
.010
.005
.000
-.005
-.010
1985
1990
1995
FRLCPI
- 22 -
2000
2(
(1) McNown, R. and Lee, D. (1976), "Economics in Our Time: Macro Issues", Science
Research Associates, INC.
(2) Krugman, P. (1995), "Dutch Tulips and Emerging Markets", Foreing Affairs, 74
(5): 28-44.
(3) Stockton, D. and Glassman, J. (1987), "An Evaluation of the Forecast Performance
of Alternative Models of Inflation", Review of Economics and Statistics, Vol. 69,
No. 1, February 108-117.
(4) Meyler, A, Kenny, G. and Quinn, T. (1998) "Forecasting Irish Inflation Using
ARIMA Models" Central Bank of Ireland, Techincal Paper.
(5) Engle, R. (1982), "Autoregressive Conditional Heteroskedasticity with Estimates
of Variance of U.K. Inflation". Econometrica 50, 987-1008.
(6) Engle, R. (1983), "Estimates of the Variance of U.S. Inflation Based on the ARCH
Model". Journal of Money, Credit and Banking, 15, 286-301.
(7) Valle, H. (2002), "Inflation Forecasts with ARIMA and Vetor Autoregressive
Models in Guatemala", Working Paper, Economic Research Department, Banco
de Guatemala.
(8) Stock, J. and Watson, M. (1999), "Forecasting Inflation", Journal of Monetary
Economics, Vol. 44 (2), pp. 293-335.
(9) Brave, S. and Fisher, J. (2004), "In Search of a Robust Inflation Forecost",
Federal Reserve Bank of Chicago Economic Perspectives, Vol. 28 (4), pp. 12-31.
(10) Clements, M. (2004), "Evaluating the Bank of England density Forecasts of
inflation", The Economic Journal, 144, pp 844-866.
(11) Aisen, A. and Veigo, F. (2005), "Dose Political Instability Lead to Higher
Inflation? A Panel data Analysis", IMF Research Department Working Paper No.
05/49.
- 23 -
(12) Bos, C., Franses, P. and Doms, M. (2001), "Inflation, Forecast Intervals and Long
Memory Regression Models", Tinbergen Institute, Discussion Paper TI-01-029/4
Amsterdam.
(13) Hubrich, K. (2003), Forecasting Euro Area Inflation; Does Contemporaneous
Aggregation Improve the Forecasting Performance?", ECB Working Paper 247.
(14) Moser, G. Rumler, F. and Scharler, J. (2005), "Forecasting Austrian Inflation"
Oesterreichische Nationalbank, Workshop No. 5.
(15) Bokil, M. and Schimmelpfennig, (2005), "Three Attempts at Inflation Forecasting
in Pakistan", IMF Working Paper No. WP/05/105.
(16) Bermingham, C. (2006), "How Useful is Core Inflation for Forecasting Headline
Inflation?", Central Bank and Financial Services Authority of Ireland, Research
Technical Paper, No. 11/RT/06.
(17) Krkoska, L. and Teksoz, U. (2006), "Forecasting Inflation for Transition
Countries: How Accurate are the EBRD forecasts? "European Bank for
Reconstrouction and Development, Working Paper No. 98.
(18) Shea, Thomas W. (1969), "The Riyal, A Miracle In Money", Saudi Aramco
World, Jan/Feb, edition.
(19) Keran, Michael W. and ALMalik, Ahmed A. (1979), "Monetary Sources of
Inflation in Saudi Arabia" Federal Reserve Bank of San Franciscom Economic
Review Suplement, Winter.
Bollerslev,
T.
(1986),
"Generalized
Autoregressive
- 24 -
Conditional
- 25 -