Dritsaki, Chaido

Other Persons: Niklis, Dimitrios; Stamatiou, Pavlos

Article
Oil consumption forecasting using ARIMA models : an
empirical study for India

Provided in Cooperation with:

International Journal of Energy Economics and Policy (IJEEP)

Reference: Dritsaki, Chaido (2021). Oil consumption forecasting using ARIMA models : an empirical
study for India . In: International Journal of Energy Economics and Policy
https://www.econjournals.com/index.php/ijeep/article/download/11231/5913. doi:10.32479/ijeep.11231.

This Version is available at: http://hdl.handle.net/11159/7770

Kontakt/Contact
ZBW – Leibniz-Informationszentrum Wirtschaft/Leibniz Information Centre for Economics Düsternbrooker Weg
120
24105 Kiel (Germany)
E-Mail: rights[at]zbw.eu
https://www.zbw.eu/econis-archiv/

Standard-Nutzungsbedingungen:
Terms of use:
Dieses Dokument darf zu eigenen wissenschaftlichen Zwecken und zum
Privatgebrauch gespeichert und kopiert werden. Sie dürfen dieses This document may be saved and copied for your personal
Dokument nicht für öffentliche oder kommerzielle Zwecke and scholarly purposes. You are not to copy it for public or
vervielfältigen, öffentlich ausstellen, aufführen, vertreiben oder commercial purposes, to exhibit the document in public, to
anderweitig nutzen. Sofern für das Dokument eine Open- perform, distribute or otherwise use the document in public. If
the document is made available under a Creative Commons
Content-Lizenz verwendet wurde, so gelten abweichend von diesen
Licence you may exercise further usage rights as specified in
Nutzungsbedingungen die in der Lizenz gewährten Nutzungsrechte.
the licence.
https://zbw.eu/econis-archiv/termsofuse
 International Journal of Energy Economics and
Policy
ISSN: 2146-4553

available at http: www.econjournals.com

International Journal of Energy Economics and Policy, 2021, 11(4), 214-224.

Oil Consumption Forecasting using ARIMA Models: An

Empirical Study for India

Chaido Dritsaki1*, Dimitrios Niklis1, Pavlos Stamatiou2

Department of Accounting and Finance, University of Western Macedonia, Kila, Kozani, India , 2Department of Applied Informatics,
1

University of Macedonia, Thessaloniki, India . *Email: [email protected]

Received: 22 February 2020 Accepted: 28 April 2021 DOI: https://doi.org/10.32479/ijeep.11231

ABSTRACT
Oil is considered one of the most widely used commodity worldwide and one of the most important goods for a country’s productivity. Even if the
effect of renewable energy sources tries to replace the consumption of fossil fuels, such as oil, nonetheless the level of worldwide oil consumption
hasn’t changed. Forecasting oil consumption plays an important role on the designing of energy strategies for policy makers. This paper aims at
modeling and forecasting oil consumption in India using Box-Jenkins methodology during 960-2020. Forecasting oil consumption was
accomplished both with static and dynamic procedure, in and out-of-sample using various forecasting criteria. The results of our paper present a
downturn in oil consumption for the following years due to two basic factors. The first is referred to Covid-19 pandemia where economic activity
of the country decreased as well as business revenues. The second is the efforts made by the country to replace, oil consumption with other energy
forms such as natural gas and mostly renewable sources like sun and wind. With these actions taken, the country – member of EU is consistent with
the regulations signed to Kyoto protocol where there are commitments for CO2 reduction emissions and improvement of energy use.
Keywords: Oil Consumption, ARIMA Model, Box-Jenkins Methodology, Forecasting, Greec
JEL Classifications: C52, C53, Q43, Q47

1. INTRODUCTION For many years, the impact of oil prices in various economic
and financial variables is one of the issues that many
Energy is one of the most important factors affecting modern researchers have dealt with. Even if the outbreak of renewable
human life. Its importance has increased in all sectors and trade resources during the last years has replace the consumption of
activities and itis considered one of the fundamental inputs for fossil fuels, such as oil, still the level of oil consumption
economic growth. Energy can be produced from various worldwide hasn’t changed. he developed and industrialized
sources such as oil, natural gas, coal, sun, wind, ocean waves countries are considered the largest oil consumers.
and biofuels. According to International Energy Organization
(IEA), global energy production sources consist of 36,1% oil, Oil is the most widely used commodity globally and one of the
18% coal, 26% natural gas, 5,8% biofuels and waste, 9,8% most important products for a country’s productivity. Thus, it
nuclear waste, 2,2% hydroelectric and 2,1% other sources. On plays an important role in the economic activity both as an
the other hand, energy consumption is increasing as the total imported and exported good. The International Monetary Fund
population, the standard of living, urbanization, argues that shocks in oil prices affect stock markets, thus a
industrialization and technology progress increase. Because of country’s economic activity, business income, inflation as well
its leading role to economic growth, the aim of our paper is
as monetary policy.
energy consumption and specifically oil consumption.
Oil is a commodity that issued in all economic levels and a
change
in its future oil prices has an impact on the expected cash flows
214 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •
2021
in

This Journal is licensed under a Creative Commons Attribution 4.0 International License

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 215

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

most companies, especially those which are dependent in a Source: IEA (International Energy Agency)
large extent to oil prices. Contrarily, oil price is affected by
business cycles and moves according to growth or recession
periods. Apart from economy’s growth and the supply and
demand forces in oil market, its price is elaborated also from
speculative factors that seem to gain remarkable importance
during the last years since oil has the characteristics of an
investment product.

Oil consumption forecasting plays a vital role in the short and

long run energy design for every country, both for policy
makers and organizations in every country.

1.1. Energy Sector in India

From the beginning of 1990’s until today, the energy sector of
India is formed according to the demands of national economy,
the progress of individual economic activities and the
development of specific sectors, affecting consumers’ habits and
also european policies for energy, environment and growth.

In the total energy system, domestic final energy consumption

was at 15,735 kilotons of oil equivalent (ktoe) in 2018, down
3.5% from 2017. Figure 1 depicts the share of the various fuels
in final energy consumption over the period 1990-2018 for
India . Oil products account for the largest share in final use
consumption (54.2%), followed by electricity (27%), renewable
energy sources (8.7%), natural gas (8.3%) and lignite (1.8%).
The consumption of fossil fuels in final use, namely petroleum
products, lignite and natural gas, decreased considerably in
2018 compared to consumption levels in 2007, falling by 36%.
This reduction was to a large extent balanced by consumption
of natural gas, the use of renewable energy sources and
electricity. Indicatively, consumption of natural gas rose by
approx. 54% to 1297 ktoe in 2018 as compared to 2007. Over
the same period, the shares of oil products and lignite were
reduced by 41% to 8493 ktoe and by 47% to 282 ktoe
respectively (IENE, 2020).

In general, liquid fuels and petroleum products comprise an

extremely dynamic sector of the economy, involved in all
aspects of economic activity. According to the data of the
Hellenic Petroleum Marketing Companies Association
(SEEPE), internal market fuel sales rose slightly by 0.45%,
from 6,655,720 tons in 2014 to 6,685,490 tons in 2018. It is
worth noting that the drop in

Figure 1: Final energy consumption by type of fuel in India , 1990-

2018

216 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

oil product consumption in 2018 as compared to 2017 “commitment period,” which covers the years from 2008 until
(6,899,847 tons) was mainly due to a fall in consumption of 2012, and this has been applied since 2005 (UNFCCC, 1997).
heating oil and unleaded gasoline. The main feature of the The European Community signed the Protocol
domestic market for oil products is the lack of preventive
control measures regarding the fuels in the market, so allowing
scope for large-scale illegal activity (adulteration, smuggling)
and problems in establishing rules of healthy competition; this
impacts adversely the operation of healthy, law-abiding
businesses, and ultimately public revenue.

Over the period 2005-2015, oil consumption in India recorded

a sudden drop by one third due to the economic crisis of 2008
and the Greek financial crisis that ensued, especially after
2009. In recent years, however, oil consumption recovered,
rising by 9% between 2013 and 2015, mainly in transport and
to an extent in the residential sector.

In 2019, India moved from the last places at the top in terms
of climate policy, as it now aims at phasing out all lignite
power producing units by 2028 at the latest. This commitment
was also included in the new NECP, while PPC’s new business
plan is even more ambitious, as it includes the closing down of
all lignite units by 2023. Hence, India is among the 15 most
advanced countries in the EU in this respect, which have
already decided to fully phase out coal/lignite, and is the first
lignite-producing EU member state that has set a firm
decarbonisation date prior to 2030. Moreover, India is the 33rd
country globally that enters into the international Powering Past
Coal Alliance (Hellenic Republic Ministry of the Environment
and Energy, 2019).

1.2. Treaty on Climate Change

Energy consumption is considered one of the most crucial
issues for every country. The use of fossil fuels (coal, oil and
gas) for industrialization and urbanization has been growing for
more than three centuries leading to increased development of
economies and technology advances worldwide. But, during
the years the intensive use of fossil fuels resulted in
environmental pollution and global warming. Such an increase
in the global temperature has caused damage on nature,
bringing about irreversible changes to many ecosystems and a
consequent loss of biodiversity. Higher temperatures and
adverse weather conditions have also resulted in huge costs for
countries’ economy and hamper their ability to produce food.
During the last 20 years, it was imperative to endorse measures
which will lead to energy saving both in national and
international level. Drastic reduction of CO 2 emissions which
destabilize earth’s atmosphere and triggers climatic changes,
energy thrift, improvement of energy efficiency and renewable
energy sources are considered as urgent choices for many
countries.

The Kyoto Protocol, which follows the United Nations

Framework Convention on Climate Change (UNFCCC), is an
international legal document signed for facing climate change.
Kyoto Protocol was adopted on 11 December 1997. It is
considered an obligation among industrialized countries to
reduce the CO2 emissions of greenhouse effect by 5,2%
average in relation to the levels of 1990, during the first
International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 217
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

on 29 April 1998 which set binding obligations to reduce 4,87%, 3,92%, 4,39%, 1,64% and 4,20%, respectively for the
emissions and improve energy use. It is worth mentioning that next 25 years.
European countries differ significantly in terms of resources, in
economic and geographical size, in population and standard of Zhang (2016) used data from 2002 until 2014 and grey-
living. extended SIGM model to forecast the annual consumption for
the next 5 years in China. The results of the paper are better
The Paris Agreement is a legally binding international treaty on than those of classical grey GM, DGM and NDGM as well as
climate change. It was adopted by 196 Parties in Paris, on 12 those of the grey-extended SIGM model. At the same time,
December 2015 and entered into force on 4 November 2016. according to the FSIGM model, this paper predicts China’s
Its goal is to limit global warming below 2, preferably to 1.5°C, crude oil consumption for 2015-2020.
compared to pre-industrial levels. To succeed in this long-term
temperature goal, countries aim to reach global peaking of Godfred (2013) on his paper correlates energy consumption
greenhouse gas emissions as soon as possible achieving a with per capita GDP increase for Ghana. Using S ARIMA
climate neutral world by mid-century. The Paris Agreement is a (1,1,1) (0,1) model, he found that an increase on energy
milestone in the multilateral climate change process because, for consumption annually by 1.21% has as a result the increase of
the 1st time, a binding agreement brings all nations into a per capita GDP by 5.5% annually for the period 2000-2008.
common cause to undertake ambitious efforts to combat climate
change and adapt to its effects (European Commission, 2016).
3. THEORETICAL BACKGROUND
According to Fifth Assessment Report of the Intergovernmental
Panel on Climate Change (IPCC), CO2 emissions from fossil
3.1. ARIMA Models
fuels should reach to zero by 2050-2070.This requires the
ARIMA are theoretically the most frequently used models for
abandonment of new investment in oil, lignite, coal and natural
the forecasting of short run forecasts of time series. ARIMA
gas and the promotion of renewable energy sources. Nations models became popular from Box and Jenkins (1976) and
have the means to limit climate change and build a more predict the future values of a time series as a linear combination
prosperous and sustainable future. of its past values and the lags of forecast errors named
innovations. An ARIMA (p, d, q) model has three parameters.
The rest of the paper is organized as follows: Section 2 AR parameter
describes literature review while Section 3 focuses on (p) represents the order of autoregressive procedure, parameter
theoretical background. In Section 4, the data are presented (d) represents the order of difference on the time series and MA
and Section 5 the empirical results are provided. Section 6 parameter (q) represents the order of movingaverageprocess.
focuses on the forecasting and in Section 7 conclusions are The ARIMA forecasting equation for a stationary time series
given. is a linear equation like regression where the predictors consist
of the lags of dependent variable as well as the lags of forecast
2. LITERATURE REVIEW errors. Thus, theformof ARIMA equation will be (Dritsaki and
Dritsaki, 2020):
In the literature, many researchers applied different methodologies
such as multiple regression, exponential smoothing, ARIMA
models, neural networks and more for the forecasting of energy p
consumption in various sectors.   d  q 
 
i
1 i L 1 L et
 yt     1 j Lj  (1)
Yuan et al. (2016) examined the forecasting of primary energy    
consumption for China creating two univariate models, ARIMA  i1  j 1

model and GM (1,1). In order to face the problems arised in the where
forecasting, the authors created a hybrid model for both models p q

i L
i j
gaining better forecasts from the previous ones. The results of  p L   1 and q  L   1 are polynomials in
their  j L
paper showed that the growth rate of primary energy i1 j 1

consumption for China from 2014 until 2020 will be larger but terms of L of degree p and q.
smaller than the first decade of the new century.
The results of their study showed that the energy consumption
Barak and Sadegh (2016) for the forecasting of energy in Turkey will continue to increase until the end of 2040.
consumption in Iran, used ARIMA and ANFIS (Adaptive Neuro Consumption in coal, oil, natural gas, renewable energy and total
Fuzzy Inférence System) models. Due to various energy will continue to increase with an annual average rate
diversifications on both models, they created a hybrid ARIMA
and ANFIS where the MSE criterion reduced to 0.026% from
0.058% in the two previous models.

Ozturk and Ozturk (2018) used annual data from 1970 to 2015
and ARIMA models to forecast energy consumption in Turkey.
216 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

yt is the time series, and et is the random error at time period t,

with  is the mean of the model.

dis the order of the difference operator.

φ1,φ2,.…,φp and φ1,φ2,.…,φq are the parameters of autoregressive

and moving average terms with order p and q respectively.

L is the difference operator defined as ∆yt=yt−yt-1=(1-L)yt.

3.2. The Box-Jenkins Methodology

The Box-Jenkins approach consists of the following steps:
 Data preparation for series stationarity

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 217

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

For series stationarity we use time plots, the estimation of linear

trend, auto correlation function (ACF) as well as unit root tests. Yˆsk  c(1)  c(2)Ysk (3)
If the levels of series are non-stationary, we proceed with 1

second differences. where Yˆs k are the lagged predicted values

1

 Model identification The accuracy of the forecasting depends on forecasting error.

Furthermore, the following statistical measures are used:
ARIMA model identification is referred to the determination of
the parameters p,d,q. First, the number of d differences is
The mean absolute error (MAE)
determined in order the series to be stationary. To determine the n
1 ˆ (4)
order of ARMA (p, q), the function of autocorrelation (ACF)
and MAE  n
Yi Y i
partial autocorrelation (PACF) of the stationary series is used. where i1
Parameter p of autoregressive operator is determined by the Yi is the actual
partial autocorrelation coefficient and parameter q of the moving
value.
average
operator is determined by the autocorrelation coefficient. The Yˆi is the predicted value
2
limits  for non-stationarity on both functions are used so
n The root of mean square error (RMSE)
we obtain a number of ARMA (,β) models where 0<<p, and
n ˆ
Y 
0<β<p. For the optimum model we use the Akaike (AIC) and 2

Schwartz (SIC) criteria.

RMSE 
n
1 Y
i1
ii (5)

 Estimation model Theil’s U statistics (1961).

Model estimation
methodology. is done with Maximum likelihood  n 2 1/ 2

We maximize the probability by iterating Marquardt and 1  Y Y  ˆ
Berndt- Hall-Hall-Hausman algorithms using derivatives,  n i1
i

optimum step and a convergence criterion for the change in U1     (6)
1/ 2
the norm of the 1 n   1 n ˆ 21/ 2
parameter vector from one iteration to the next. 

Y2
i 

 n

Y 
i

n
 i1   i1 
 Diagnostic checking of the model  n1  ˆ
2 1/ 2

  Yi1  Yi1 
With diagnostic checking, we investigate if the estimated model  
 i1 

is

Yi   (7)
acceptable and statistical significant, in other words if it “best” fits U2  
the data. The diagnostic testing of the model consists of Ramsey  n  Yi1  Yi 2 1/ 2
specification test (1969) (RESET test), normality test (Jarque 

and Bera, 1980 test), autocorrelation test (Ljung and Box, 1978
 
 
Yi  


statistic), ARCH (squared residuals’ and Ljung and Box, 1978  i1  
test, Engle, 1982 test).

 Forecasting. 4. DATA
One of the main goals of the analysis on time series models For the empirical analysis of the paper, the oil consumption (in
is forecasting. Forecasting can be static and dynamic. Static kilotons) in India was used covering the period 1960-2020, in
forecasting is known as a one-step ahead forecast and uses the total of 61 annual observations. The data derived from World
actual lagged values of time series Y for the forecasts. The Bank. E-Views 11 econometric software was used for the
dynamic forecasting is known as multi-step ahead forecast and construction of ARIMA models.
uses the actual lagged value of Y variable to measure the first
predicted value. After, it uses the first predicted values in order to On Table 1, the descriptive statistics of oil consumption in India
calculate the second one and so on (Dritsaki, 2015). are presented.

If s is the first observation for forecasting, then we have the From the above table, we can see that the average annual oil
following equation: consumption is 1763.7 kilotones with the largest consumption
to be recorded in year 2005 with 2753 kilotones. The series is
Yˆ  c(1) 
c(2)Y (2)
218 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Table 1: Descriptive statistics of oil consumption

s s1
Mean Max. Min.
Std. Skew. Kur. J-B Obser.
where Ys−1 is the actual value of the last observation of the sample dev.
and Yˆs is the first predicted value. For the next predicted values, 1763.7 2753.0 289.06 744.07 −0.612 2.253 5.226 61
we use the equation below: (0.073)

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 219

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Diagram 1: Histogram and graph of normal distribution function

Diagram 2: Oil consumption for India from 1960 until 2020

Based on the results on Table 2 and Diagram 3, we can see that
there is a trend on the estimated model (prob<5%). Thus, this
series is characterized as non-stationary.

 Graph on auto correlation coefficients.

Afterwards, we test for stationarity through the auto correlation

correlogram.

The autocorrelation coefficients on Diagram 4 decay slowly

denoting that the series is non stationary. Moreover, the value
left skewed and leptokyrtic while all data series follow normal of the first auto correlation coefficient is large and positive
distribution. meaning that the series is non stationary.

On the following Diagram 1, the histogram together with the  First differences on series.
graph of normal distribution is depicted.
We apply again the previous tests so that we can detect the
Most of the data series follow a normal distribution as it is existence of stationarity of the series on first differences. Diagram
shown from the above Diagram 1. 5 shows oil consumption on first differences.

5. EMPIRICAL RESULTS We notice that the behavior of oil consumption presents

significant fluctuations. This is a possible indication for mean
5.1. Testing for Stationarity stationarity.
 Time plots
Afterwards, we test for stationarity with the autocorrelation
correlogram on first differences.
On Diagram 2, the progression of oil consumption for India is
showed for the examined period.
The autocorrelation coefficients decay quickly from the above
Diagrams 6 and this denotes that the series is stationary.
From Diagram 2 we can see that oil consumption in India
exhibits an upward trend for a long time period until 2005 and a
 Unit root tests
decline follows until 2013 due to economic crisis and
memorandums. As light increase followed until 2019 while on
The confirmation of series stationarity is conducting also with
2020 a decrease was recorded due to Covid-19 crisis
Dickey and Fuller (1979; 1981) and Phillips and Perron (1998)
worldwide. In other words, we conclude that the movement of
unit root tests.
oil consumption is a random walk model.
The results on Table 3 confirm that the series is stationary on
 Linear trend model.
first differences.
On the following table, the estimation of the variable in relation
 Model Identification:
to time for the determination of the existence of trend is
presented together with the Diagram 3 of actual and estimated
The identification of ARIMA model is referred to the
values of the examined variables.
determination
no f p, d, q parameters. First, the number of differences d is
220 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

determined in order to change the series in a stationary series. The results from Table 4 show that according to AIC, SC,
The determination of parameter d was employed using Sigma SQ, andAjR2, the most suitable model is ARMA (1,1,1).
autocorrelation coefficients and confirmed with unit root tests.
Afterwards, from the results of Diagram 7 the parameters p and Using the automatic ARIMA forecasting procedure with EViews,
q are determined
2 2 we find all models’ alternatives.
compared with the critical value    0.256 . So,
n 61
we

get a number of ARMA (α,β), where 0<α<p, and 0<β<q. From

From Figure 2, we select the best ARMA(p,q) model from the
the values of partial autocorrelation only the value p=2 is larger
smallest values of AIC criterion. According to Figure 2, the
than critical value and from the values of autocorrelation
ARMA (0,1) (0,0) model is the most appropriate. Due to the
coefficients q=1 is larger than critical value. Using the above
fact that its coefficients are not statistical significant, we obtain
values, we choose the best ARMA (p,q) model from the
ARMA (1,1) (0,0) as the most appropriate.
smallest values of AIC, SC, and Sigma SQ criteria as well as
the largest of Aj R2 from the Table 4.
 Estimation and Diagnostic tests of the model

Diagram 3: Actual, fitted and residuals plot

Diagram 4: Autocorrelation and partial autocorrelation correlogram of oil consumption

Table 2: Estimation of oil consumption

Variable Coefficient Std. Error t-Statistic Prob.
C 647.9274 87.49797 7.405056 0.0000
TREND 37.19374 2.515390 14.78647 0.0000
R-squared 0.787494 Mean dependent var. 1763.740
Adjusted R-squared 0.783893 S.D. dependent var 744.0742
S.E. of regression 345.9004 Akaike info criterion 14.56242
Sum squared resid 7059178 Schwarz criterion 14.63163
Log likelihood -442.1537 Hannan-Quinn criter. 14.58954
F-statistic 218.6397 Durbin-Watson stat 0.061234
International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 221
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Prob (F-statistic) 0.00000

222 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Since the most suitable model is ARIMA (1,1,1) the estimation Diagram 5: Oil consumption on first differences
will occur with Maximum Likelihood approach. We maximize
the probability by iterating algorithms Marquardt and Berndt-
Hall-Hall-Hausman, using derivatives optimum step size and a
convergence criterion for the change in the norm of the
parameter vector from one iteration to the next.

The following Table 5 provides results of the estimation of

ARMA (1,1,1) model.

The results on Table 5 show that the coefficients are statistically

significant in 1% level of significance.

The estimation occurred with Maximum Likelihood

methodology using BHHH algorithm and the inverse matrix
OPG. The results come up after 14 iterations. The coefficient
for the estimation of error variance (volatility) is statistical
significant in 1% level of significance.

Afterwards, we exhibit inverted AR Roots and inverted A

Roots
for model’s stationarity.

Figure 3 shows that the inverted AR Roots and inverted A Roots

of the model are within the unit circle meaning that the process
is

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 223

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

stationary. So, we can use the ARIMA (1,1,1) model for diagnostic
tests.

Next, we examine model specification with Ramsey RESET

test.

The results from Table 6 display that the ARIMA (1,1,1)

model has correct specification (prob>5%) on both F
distribution and LR likelihood ratio.

Following, we examine the autocorrelation of model’s residuals.

As the coefficients of autocorrelation and partial autocorrelation

of the residuals are within the limits, we conclude that there is

224 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Table 3: Summary table of Augmented Dickey–Fuller and

Phillips-Perron unit root tests
Variable ADF P-P
C C, T C C, T
OIL −2.784 (0) 0.933 (0) −2.346[4] 0.614[3]
DOIL −5.061 (0)* −6.155 (0)* −5.369[4]* −6.157[2]*
*, ** and *** show significant at 1%, 5% and 10% levels respectively. The numbers
within parentheses followed by ADF statistics represent the lag length of the dependent
variable used to obtain white noise residuals. The lag lengths for ADF equation were
selected using Schwarz Information Criterion (SIC). Mackinnon (1996) critical value for
rejection of hypothesis of unit root applied. The numbers within brackets followed by
PP statistics represent the bandwidth selected based on Newey and West (1994) method
using Bartlett Kernel. C=Constant, T=Trend Table 4: Comparison of the model through AIC, SIC,
Sigma SQ, Aj R2
ARIMA model (p, d, q) Criteria
SigmaSQ AdjR2 AIC SC
(1,1,0)* 6511.6 0.054 11.72 11.82
(2,1,0)^ 6282.4 0.071 11.71 11.85
(1,1,1)* 6085.6 0.100 11.68 11.82
(2,1,1)^ 6083.7 0.084 11.72 11.89
(0,1,1)^ 6681.3 0.029 11.74 11.85
^Model with coefficients non statistical significant. *1%significance

Diagram 6: Autocorrelation and partial autocorrelation correlograms of oil consumption on first differences

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 225

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Figure 2: Automatic ARIMA model estimation choice

Figure 3: Inverted roots of ARIMA (1,1,1) model

independence between the residuals of the ARIMA (1,1,1) model Table 5: Estimation of ARIMA (1,1,1) model
in 5% level of significance (no autocorrelation). Variable Coefficient Std. Error T-Statistic Prob.
AR (1) 0.914696 0.113684 8.045983 0.0000
The following Diagram 8 exhibits the test of conditional MA (1) -0.698697 0.191814 -3.642580 0.0000
SIGMASQ 6123.907 940.4168 6.511908 0.0000
autocorrelation. R-squared 0.140641 Mean 28.35100
dependent var.
Autocorrelation and partial autocorrelation coefficients’ on squared Adjusted 0.110489 S.D. 85.12884
R-squared dependent var
residuals are within the limits ±0.256 so we can claim that there
S.E. of 80.28834 Akaike info 11.66522
is no autoregressive conditional heteroscedasticity on the regression criterion
residuals of ARIMA (1,1,1) model in 5% level of significance (no Sum squared 367434.4 Schwarz 11.76994
ARCH effect). resid criterion
Log likelihood -346.9567 Hannan-Quinn 11.70618
criter.
The diagnostic tests of the model have no issues thus we can Durbin-Watson 1.885325
proceed with forecasting. stat
Inverted AR 0.91
Roots
6. FORECASTING Inverted MA 0.70
Roots
On the following Table 7 we exhibit the evaluation criteria of
static and dynamic forecasting of the model for the period 1960-
2020.
Table 6: Ramsey RESET test
From the results of Table 7, all the statistical criteria conclude Omitted variables: Squares of fitted values
that Static Forecast provides better results for forecasting than Specification: D (OIL) AR (1) MA (1)
the Dynamic Forecast for the ARIMA (1,1,1) model. Distribution Value df Probability
t-statistic 0.304680 56 0.7617
F-statistic 0.092830 (1,56) 0.7617
On the Figure 4, the trend of actual and predicted values in oil
Likelihood ratio 0.134664 1 0.7136
consumption with static forecasting is featured concerning the
226 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

in-sample period from 1960-2020.

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 227

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Diagram 7: Correlograms of autocorrelation and partial autocorrelation of the residuals on ARIMA (1,1,1) model

Diagram 8: Correlograms of autocorrelation and partial autocorrelation of the squared residuals on ARIMA (1,1,1) model

Table 7: Evaluation criteria of forecasting ARIMA (1,1,1) Figure 4: Actual and predicted value of India oil consumption from
Criteria Dynamic forecast Static forecast 1960 to 2020
RMSE 1525.321 78.83477
MAE 1371.491 59.05796
MAPE 68.70311 3.889852
Theil 0.639231 0.020274
Bias Proportion 0.808469 0.003332
Var. Proportion 0.190127 0.049136
Cov. Proportion 0.001403 0.947531
Theil U2 coef. 9.260582 0.807839
SymmetricMAPE 110.2998 3.949925

From the Figure 4 we notice that the width of confidence

interval
for the year 2020 is between 1990-2372.

228 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

Table 8: Static and dynamic forecast of India oil place account for the new energy policy applied in India
consumption abiding by the rules and criteria addressed from the European
Year Actual Static forecast Dynamic Union.
(OIL) (OILST) forecast (OILD)
Ex-post forecast This paper aimed at modeling and forecasting oil consumption
2018 2183.45 2168.19 2165.45 for India using Box-Jenkins methodology during the period
2019 2198.45 2173.96 2152.69 1960-2020. The results of the paper shown that according to
2020 1990.12 2167.76 2141.47
Ex-ante forecast AIC, SC, Sigma SQ, and AjR2 criteria, the most suitable model
2021 2152.45 2129.12 is ARIMA (1,1,1) for estimation and forecasting of oil
2022 2117.54 consumption. The estimation of ARIMA (1,1,1) model was
2023 2106.01 accomplished with Maximum Likelihood approach. We
maximized likelihood by iterating Marquardt and Berndt-Hall-
Figure 5: Static and dynamic forecast of oil consumption Hall-Hausman algorithms using derivatives, optimum step size
and a convergence criterion for the change in the norm of the
parameter vector from one iteration to the next. Forecasting was
attained with static and dynamic procedure in and out-of-
sample using all the forecasting criteria. The results presented a
sharp drop in oil consumption in the following years because
of two basic factors. The first one is due to coronavirus crisis
that hit the economic activity of the country and the second one
are the efforts made by replacing oil consumption with other
energy forms.

International Energy Agency (IEA) on the recent Oil Market

Report (2021) points out that the rebalancing of the oil market
remains fragile in the early part of 2021 as measures to contain
the spread of Covid-19, with its more contagious variants,
On Table 8 the forecasted results are shown. The period 2018-
weigh heavily on the near-term recovery in global oil demand.
2020 was used as a forecast for in-the-sample, whereas the
IEA predictions for economic growth and oil demand increase
forecast for out-of-sample covers the period 2021-2023.
depend in a large scale on progress in distributing and
administering vaccines, and the easing of travel restrictions in
The results on Table 8 show the increase of oil consumption for
the world’s major economies.
India with the static forecasting for the year 2020, while the
dynamic forecasting exhibit a slight decrease of oil
The outbreak of Covid-19 added more uncertainty to the
consumption for the years 2021-2023. On the following Figure
perspective of oil market outlook and oil consumption in the
5, the trend of static and dynamic forecasting is presented.
beginning of the forecasting period which covers the years
2021- 2023. In the year 2020, oil consumption has shrunk for
The oil consumption seems to have a slight downturn for the
the first time after the economic crisis and memorandums in
years 2021-2023.
India . However, the situation remains volatile until global
pandemic will disappear. The potentials for the oil market and
7. CONCLUSION oil consumption will depend on how quickly the Greek
government will take action to constrain pandemic. This
Energy is regarded as a significant material basis for global uncertain situation is leading to two possible scenarios. The first
economic and social growth. The production and consumption one, the pessimistic scenario, is the delay to constrain the
of oil may lead or prevent economic growth. Imbalances on virus. The second, the optimistic, refers to the coronavirus
supply/demand oil market are becoming more apparent due infections to the global population so that the countries can
to the increasing use of renewable energy sources. Low use recover and economic activity will start again.
of oil, irrational consumption structure, pollution and other
issues can limit the growth of industrialized countries. With the REFERENCES
industrialization of the countries, urbanization and the increase
of energy consumption, environmental restrictions will rise in Barak, S., Sadegh, S.S. (2016), Forecasting energy consumption using
most countries. Arrangements between energy and economic ensemble ARIMA-ANFIS hybrid algorithm. Electrical Power and
growth should be made that will lead to a sustainable economic Energy Systems, 82, 92-104.
development and society worldwide. Box, G.E.P., Jenkins, G.M. (1976), Time Series Analysis, Forecasting
and Control. San Francisco, California: Holden-Day.
The prediction of demand and oil consumption is an important Dickey, D.A., Fuller, W.A. (1979), Distributions of the estimators for
part of growth strategies. The increasing consumption of autoregressive time series with a unit root. Journal of American
sustainable energy in India and also the structural changes that Statistical Association, 74(366), 427-431.
Dickey, D.A., Fuller, W.A. (1981), Likelihood ratio statistics for
are taking
autoregressive time series with a unit root. Econometrica, 49(4),
International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 229
2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

1057-1072.
Dritsaki, C. (2015), Box-Jenkins modeling of Greek stock prices data.

230 International Journal of Energy Economics and Policy | Vol 11 • Issue 4 •

2021
 Dritsaki, et al.: Oil Consumption Forecasting using ARIMA Models: An Empirical Study for INDIA

International Journal of Economics and Financial Issues, 5(3), MacKinnon, J.G. (1996), Numerical distribution functions for unit root
740-747. and cointegration tests. Journal of Applied Econometrics, 11(6),
Dritsaki, M., Dritsaki, C. (2020), Forecasting European Union CO 2 601-618.
emissions using autoregressive integrated moving average- Newey, W.K., West, K.D. (1994), Automatic lag selection in covariance
autoregressive conditional heteroscedasticity models. International matrix estimation. Review of Economic Studies, 61(4), 631-653.
Journal of Energy Economics and Policy, 10(4), 411-423.
Ozturk, S., Ozturk, F. (2018), Forecasting energy consumption of Turkey
Engle, R. (1982), Autoregressive conditional heteroscedasticity with
by ARIMA model. Journal of Asian Scientific Research, 8(2), 52-60.
estimates of United Kingdom inflation. Econometrica, 50, 987-1008.
Phillips, P.C.B., Perron, P. (1998), Testing for a unit root in time series
European Commission. (2016), Available from: https://www.ec.europa.
regression. Biometrika, 75(2), 335-346.
eu/clima/policies/international/negotiations/paris_en.
Ramsey, J.B. (1969), Tests for specification errors in classical linear least
Godfred, K.A. (2013), Modeling and forecasting energy consumption
squares regression analysis. Journal of the Royal Statistical
in Ghana. Journal of Energy Technologies and Policy, 3(12), 1-10.
Society, Series B, 31(2), 350-371.
Hellenic Republic Ministry of the Environment and Energy. (2019),
Theil, H. (1961), Economic Forecasts and Policy. Amsterdam: North-
National Energy and Climate Plan, Athens, December 2019. India :
Holland Publishing Company.
Hellenic Republic Ministry of the Environment and Energy.
UNFCCC. (1997), Kyoto Protocol to the United Nations Framework
IEA. (2021), Oil Market Report. Available from: https://www.iea.org/
Convention on Climate Change, Conference of the Parties, Kyoto
reports/oil-market-report-february-2021.
1-10 December 1997.
IENE. (2020), The Greek Energy Sector, Annual Report.
Yuan, C., Liu, S., Fang, Z. (2016), Comparison of China’s primary
Jarque, C., Bera, A. (1980), Efficient tests for normality, homoscedasticity
energy consumption forecasting by using ARIMA (the
and serial independence of regression residuals. Economics
autoregressive integrated moving average) model and GM(1,1)
Letters, 6, 255-259.
model. Energy, 100, 384-390.
Ljung, G.M., Box, G.E.P. (1978), On a measure of a lack of fit in time
Zhang, W.Q. (2016), Prediction model of world oil consumption. Science
series models. Biometrika, 65(2), 297-303.
and Technology Economy, 29, 105-106.

International Journal of Energy Economics and Policy | Vol 11 • Issue 4 • 231

2021