Measuring the Forecasting Power of ARIMA Modeling for

Report
Jamaladeen Abubakar
Department of mathematics and statistics
Hussaaini Adamu Federal polytechnic, kazaure
08034067081, 07053555571
[email protected]
Nafiu Bashir Abdussalam
Department of Economics
Bayero University, Kano
07037880962
[email protected]
Introduction
 Literature Review
 Methodology
i)
Sources of Data
ii)
Econometric tools
iii) Result Presentation and Analysis

The Role of electricity in the global economy
 The role of electricity in Nigeria
 Electricity Demand
i)
Residential Sector
ii)
Commercial sector
iii) Street light and
iv)
Industrial Sector

The rule of thumb for electricity consumption
is 1 gigawatt (1000 megawatt) for every one
million population (RMFR, 2010)


Demand and supply gap of eletricity: the
global exprience
Countries
Population
of the
contries (M)
Demand
(GW)
Supply (GW)
Excess DD
(GW)
Germany
80
80
120
40
UK
60
60
80
20
S. Africa
50
50
40
(10)
Egypt
80.5
80.5
24
(56.5)
Algeria
40
40
11
(29)
Brazil
200
200
100
(100)
Nigeria
150
150
3
(157)

Demand and supply gap of eletricity: the
Nigerian exprience
Year
Demand
(GW)
Suppy
Excess DD
2005
19.85
4
(15.85)
2006
15.59
7
(8.59)
2007
19.06
13
(6.06)
2008
22.11
8
(13.11)
2009
22.11
5
(17.11)
2010
21.92
3
(18.92)
2011
21.92
3.5
(18.42)
2012
20.13
4
(16.13)
The Government efforts to address the demand- Supply
Gap in Nigeria
i)
Fulfilling the imperative of the electric power sector act
reform
ii)
Improving service delevary the transition by
Fuel to power
Generation
Transimition
Distribution
Human Capital Development and
Energy consumption effeciency
iii) Removing obstacles to private sector investment through
a)
Establishment of appropiate price regime
b)
The estaablishment of a bulk purcher



Nature of electricity market in Nigeria
Does incentives existance in the Nigeria
Electricity sector?
Users
Types of
Product
Price paid
Current triff
Difference
Poor people
Burning
candles and
kerosene
N 80/kwh
N 8.5/kwh
N 71.5
Manufactures Diesel or
LPFO
generation
N 60/kwh
N 8.5/kwh
N 51.5
Others
N 50-70/kw
N 8.5/kwh
N 61.5
Diesel or
petrol

Empirical evidence from developed countries,
for instance, Halicioglu (2007) for Turkey;
Zachariadis and Pashourtidou (2007) for
Cyprus; Narayan and Smyth (2006) for
Australia; Galindo (2007) for Mexico;
Holtedahl and Joutz (2004) for Taiwan;
Filippini and Pachauri (2004) for India; Hunt
et al. (2003) for the United Kingdom; Sa’ad
(2009) for South Korea; Donatos and Mergos
(1991) for Greece
Evidence from developing countries
De Vita et al. (2006) for Namibia; Ziramba (2008)
for South Africa; Babatunde and Shuaibu (2009),
Ayodele A.() and Adams et-al (2011) for Nigeria
 However,
an
aggregated
analysis
that
incorporates other uses of electricity such the
industrial and commercial sectors to obtain
robust
estimates
of
electricity
demand
parameters for policy decision-making is
considered pertinent with this paper filling the
vacuum.

Sources of Data
 Econometric Model Specification
 The traditional enonomics/economiteric
anlysis and its modeling procedure
 Modern econometrics/time series analysis
Box-Jenkis Methodology
The basic ARIMA (1 1 1) Model

Y t u  0Yt 1  1et 2et 1...........(3.1)
2012
2010
2008
2006
1000
2004
2002
2000
1998
1996
1994
1992
1990
1988
1986
1984
1982
1980
1978
1976
1974
1972
1970
2500
2000
1500
year
Electricity
consumption
500
0
Mean
Median
Max
Electricit 951.5130 946.6000 1921.00
y
Min.
Std. Dev Skewnes Kurtosis Jarques
bera
145.3000 524.4172 0.3319
2.1560
2.0656
Prob.
Obs
0.3559
43
variables
Electricity
Demand
Level
First Difference
KPSS Statistic Decision
KPSS Statistic Decision
0.7975
0.026581
I(1)
I(0)

Diagnostic Checks
MODELS
ARIMA (111)
ARIMA (211)
ARIMA (212)
ARIMA (311)
ARIMA (411)
ARIMA (412)
12.9306
12.7116
12.5364
12.93493
13.01389
12.77921
12.8520
0.9285
12.8788
0.9360
12.7454
0.9489
13.14604
0.9214
13.26983
0.9163
13.07780
0.9371
AIC
SIC
Adjusted r2

Evaluation of the forecasting power of the
models
Statistic
RMSE
MASE
MAPE
Teil U- Stat
Bias
proportion
Variance
proportion
Covariance
proportion
Models
ARIMA (1 1 1)
ARIMA (2 1 1)
ARIMA(2 1 2)
ARIMA (3 1 1)
ARIMA(4 1 1)
ARIMA(4 1 2)
179.5966
152.0166
98.5064
156.5794
198.0987
89.4728
134.5973
104.5003
114.0913
119.9807
154.5071
120.5003
16.38568
27.0032
68.3078
59.0879
23.7896
25.38938
0.079443
0.10987
0.009143
0.07653
0.067540
0.083632
0.250202
0.5432
0.066577
0.98765
0.253725
0.742937
0.024101
0.02387
0.010875
0.08765
0.012311
0.046834
0.725697
0.9670
0.453765
0.89765
0.784302
0.928336

Result For Serial Correlation Test
F-Statistic
0.612803
Obs*R-square 1.425557
Probability
0.547178
Probability
0.490280



Government should consider alternative
models for forecasting both demand and
supply of electricity in Nigeria
In order to reach the minimum standard for
electricity supply, huge amount of money
need to be Investment on electricity sector of
Nigeria
The sources of electricity in Nigeria should be
diversified as it exist the hydro-electric as the
main sources of electricity in Nigeria

similar documents