Economic growth and electricity consumption: MS‐VAR and MS‐Granger causality analysis

• DOI: http://doi.org/10.1111/opec.12011
• Date: 01 December 2013
• Author: Melike E. Bildirici
• Published date: 01 December 2013

Economic growth and electricity

consumption: MS-VAR and MS-Granger

causality analysis

Melike E. Bildirici

Professor, Department of Economics, Yildiz TechnicalUniversity, Barbaros Bulvari, 34349, Bes¸iktas¸/

Istanbul, Turkey.Email: melikebildirici@gmail.com

Abstract

This study estimates the causality relationship between electricity consumption and economic

growth by Markov Switching VAR (MS-VAR) method for some emerging countries: Argentina,

China, India, Brazil, Mexico, Turkey and South Africa. Knowledge of the direction of causality

between electricity consumption and economic growth is of primary importance if appropriate

energy policies and energy conservation measures are to be devised. The results from MS-VAR

models show that in first, second and third regime, electricity consumption is the Granger cause of

the economic growth and economic growth is the Granger cause of the electricity consumption. In

sum, we found some evidence of bidirectional Granger causality between the electricity consump-

tion and the economic growth.

1. Introduction

A significant amount of the literature focuses on the relation between energy consump-

tion and economic growth which could be collected with respect to their approaches

towards energy and its impact on economic growth. The first group of studies investi-

gate the energy consumption by evaluating it as a measure of economic growth. The

second group of papers also utilise the energy consumption, but their approach focuses

on the evaluation of energy consumption by taking it as one of the important factors of

production besides capital, labor and factors such as raw materials and technology. By

evaluating models derived from a specified production function with variables related to

energy, land, labor and capital for the United States, show that the increases in the

energy prices stimulated negative trends on gross national product of the country.

Further, a third group of studies focus on evaluation of income elasticity and price elas-

ticity of electricity demand as Houthakker (1951), Fisher and Kaysen (1962), Baxter

and Rees (1968), Anderson (1973), Houthakker and Taylor (1970), Wilson (1971),

Bakırtas¸ et al. (2000).

447

© 2013 The Author. OPEC Energy Review © 2013 Organization of the Petroleum Exporting Countries. Published by

John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA 02148, USA.

Many papers take the electricity consumption as measure of economic growth by

deriving bidirectional or/and unidirectional causality analysis. Among many, Kraft and

Kraft (1978), Berndt (1978), Akarca and Long (1980), Proops (1984), Yu and Hwang

(1984), Yuand Choi (1985) and Erol and Yu (1987) are the pioneering studies. If an over-

look is to be provided it is observed that the relationship between economic growth and

electricity consumption is examined in many studies both for different countries. In

accordance with the progress in the econometric theory, Sims causality, Johansen Cointe-

gration, Engle–Granger cointegration are evaluated intensively. Furthermore, new tech-

niques include genetic algorithms and neural networks; structural time series,

autoregressive distributed lag (ARDL) methodology and panel cointegration tests.

One of the shortcomings of these studies is the avoidance of the non-linear structure of

the time series under consideration, especially gross domestic product (GDP) series that

has been extensively evaluated as a measure of economic performance under business

cycles. As a result, the ongoing perspective, though followed the new technical improve-

ments in terms of linear econometrics by assuming that the parameters were constant over

the sample period, the introduction of non-linear econometric analysis will revealsignifi-

cant points with respect to the relationship between economic growth and energy or elec-

tricity consumption.

In the study, the relationship between electricity consumption and economic growth

in perspectives of business cycle is suggested to be evaluated by Markov Switching

(MS) models: MS-VAR and MS Granger causality analysis. Hamilton (1989) proposed

a simple non-linear framework for modelling economic time series with a permanent

component and a cyclical component as an alternative to a stationary linear autoregres-

sive model. In his framework, recessions are due to permanent negative shocks. Another

type of business cycle asymmetry is due to Kim and Nelson (1998, 1999a,b) and this

type of asymmetry is known as the ‘plucking model’ of business cycles. In their

approach, Kim and Nelson show that the recessions are consequences of temporary

deviations from the long-run level of GDP as occasional ‘plucks’ whereas expansions

reject permanent shocks. MS-AR models are also commonly used for the analysis of oil

shocks with respect to energy analysis. Clements and Krolzig (2003), Holmes and Wang

(2003), Cologni and Manera (2008) and Bildirici et al. (2011) used MS-AR and

MS-VAR models to test the impact of oil shocks. Falahi (2011) used MS-VAR model

for analysis of the relationship between electricity consumption and economic growth in

the United States.

In the study, MS-VAR models are estimated and evaluated to analyse the relationship

between electricity consumption and economic growth.This study can be defined as com-

plementary to the previous empirical papers. However, it differs from the existing litera-

ture for some aspects. Firstly, as being distinguished from the previous works, it employs

MS-VAR method. Secondly, the study introduces MS Granger causality analysis.

Melike E. Bildirici448

OPEC Energy Review December 2013 © 2013 The Author.

OPEC Energy Review © 2013 Organization of the Petroleum Exporting Countries