ournal of Economic Studies

83 

Full text

(1)

European Journal of Economic Studies, 2016, Vol.(18), Is. 4

E

UROPEAN

of Economic

Journal

Studies

Has been issued since 2012.

ISSN 2304-9669. E-ISSN 2305-6282

2016. Vol.(18). Is. 4. Issued 4 times a year

Impact Factor MIAR 2016 – 5,602

EDITORIAL BOARD

Dr. Vidishcheva Evgeniya – Sochi State University, Sochi, Russian Federation (Editor-in-Chief)

Dr. Simonyan Garnik – Scientific Research Centre of the Russian Academy of Sciences, Sochi, Russian Federation

Dr. Levchenko Tatyana – Sochi State University, Sochi, Russian Federation

Dr. Tarakanov Vasilii – Volgograd State University, Volgograd, Russian Federation

Dr. Balatsky Evgeny – Central Economics and Mathematics Institute (RAS), Moscow, Russian Federation

Dr. Dinh Tran Ngoc Huy – Banking University HCMC Viet Nam – GSIM, International University of Japan, Japan

Dr. Gerasimenko Viktor – Odessa State Economic University, Odessa, Ukraine

Dr. Gvarliani Tatjana - – Sochi State University, Sochi, Russian Federation

Dr. Gunare Marina – Baltic International Academy, Riga, Latvia

Dr. Kryshtanovskaya Olga – Institute of Sociology of the Russian Academy of Sciences, Moscow, Russian Federation

Dr. Minakir Pavel – Economic Research Institute of the Far Eastern Branch Russian Academy of Sciences, Khabarovsk, Russian Federation

Dr. Papava Vladimir – Ivane Javakhishvili Tbilisi State University, Tbilisi, Georgia

Dr. Prokopenko Olga – Sumy State University, Sumy, Ukraine

Dr. Vishnevsky Valentine – Institute of Industrial Economics of the National Academy of Sciences of Ukraine, Donetsk, Ukraine

The journal is registered by Federal Service for Supervision of Mass Media, Communications and Protection of Cultural Heritage (Russia). Registration Certificate

ПИ № ФС77-50465 4 July 2012.

Journal is indexed by: CrossRef (UK), EBSCOhost Electronic Journals Service

(USA), Electronic scientific library (Russia), Global Impact Factor (Australia), Index Copernicus (Poland), Open Academic Journals Index (Russia), ResearchBib

(Japan), ULRICH’s WEB (USA).

All manuscripts are peer reviewed by experts in the respective field. Authors of the manuscripts bear responsibility for their content, credibility and reliability.

Editorial board doesn’t expect the manuscripts’ authors to always agree with its opinion.

Postal Address: 26/2 Konstitutcii, Office 6 354000 Sochi, Russia

Website: http://ejournal2.com/ E-mail: ejoes@inbox.ru

Founder and Editor: Academic Publishing House Researcher

Passed for printing 15.12.16. Format 21  29,7/4.

Headset Georgia.

Ych. Izd. l. 4,5. Ysl. pech. l. 4,2.

Order № 110.

© European Journal of Economic Studies, 2016

А

E

ur

ope

an

J

ou

rna

l of Econ

om

ic Studies

4

2016

201

0

(2)

European Journal of Economic Studies, 2016, Vol.(18), Is. 4

E

UROPEAN

of Economic

Journal

Studies

Издается с 2012 г. ISSN 2304-9669. E-ISSN 2305-6282 2016. № 4 (18). Выходит 4 раза в год

.

Impact Factor MIAR 2016 – 5,602

РЕДАКЦИОННЫЙ СОВЕТ

Видищева Евгения – Сочинский государственный университет, Сочи, Российская Федерация (Гл. редактор)

Левченко Татьяна – Сочинский государственный университет, Сочи, Российская Федерация

Симонян Гарник – Сочинский научно-исследовательский центр Российской академии наук, Сочи, Российская Федерация

Тараканов Василий – Волгоградский государственный университет, Волгоград, Российская Федерация

Балацкий Евгений – Центральный экономико-математический институт РАН, Москва, Российская Федерация

Вишневский Валентин – Институт экономики промышленности Национальной академии наук Украины, Донецк, Украина

Гварлиани Татьяна – Сочинский государственный университет, Сочи, Российская Федерация

Герасименко Виктор – Одесский государственный экономический университет, Одесса, Украина

Гунаре Марина– Балтийская международная академия, Рига, Латвия

Динь Чан Нгок Хай – Банковский университет Хошимин Вьетнам - GSIM, Международный университет Японии, Япония

Минакир Павел – Институт экономических исследований ДВО РАН, Хабаровск, Российская Федерация

Крыштановская Ольга – Институт социологии РАН, Москва, Российская Федерация

Папава Владимир – Тбилисский государственный университет имени Иване Джавахишвили, Тбилиси, Грузия

Прокопенко Ольга – Сумский государственный университет, Сумы, Украина

Журнал зарегистрирован Федеральной службой по надзору в сфере массовых коммуникаций, связи и охраны культурного наследия (Российская Федерация). Свидетельство о регистрации средства массовой информации ПИ № ФС77-50465 от 4 июля 2012 г.

Журнал индексируется в: CrossRef (Великобритания), EBSCOhost Electronic Journals Service (США), Global Impact Factor (Австралия),Index Copernicus (Польша), Научная электронная библиотека (Россия), Open Academic Journals Index (Россия),

ResearchBib (Япония), ULRICH’s WEB (США).

Статьи, поступившие в редакцию, рецензируются. За достоверность сведений, изложенных в статьях, ответственность несут авторы публикаций.

Мнение редакции может не совпадать с мнением авторов материалов.

Адрес редакции: 354000, Россия, г. Сочи, ул. Конституции, д. 26/2, оф. 6

Сайт журнала: http://ejournal2.com/ E-mail: ejoes@inbox.ru

Учредитель и издатель: ООО «Научный издательский дом "Исследователь"» - Academic Publishing House Researcher

Подписано в печать 15.12.16. Формат 21  29,7/4.

Гарнитура Georgia.

Уч.-изд. л. 4,5. Усл. печ. л. 4,2.

Заказ № 110.

© European Journal of Economic Studies, 2016

А

E

uropean

J

ourn

al of Econ

omic Stu

di

es

4

2016

(3)

C O N T E N T S

Articles and Statements

Government Expenditure, Defense Expenditure and Economic Growth: a Causality Analysis for BRICS

Salman Ali Shah, Chen He, Mao Yu, Wang Xiaoqin ... 447 The Impact of Energy Consumption, Trade Openness and Financial Development on

Economic Growth: Empirical Evidence from Turkey (1980-2014)

Murat Cetin ... 459 Fostering the Sustainable Development of the Economy of the Russian Federation

via the Creation of Small Innovation Enterprises at Institutions of Higher Learning

Mikhail N. Dudin, Natalia P. Ivashchenko ... 470 Features of Touristic Territory Branding on the Example of Sochi City

(Russian Federation) and Jurmala City (Latvia)

Marina Gunare, EvgeniyaV. Vidishcheva ... 476 Characteristics of Basel Principles and Standards in Banking

Branimir Kalaš, Nada Milenković, Jelena Andrašić, Miloš Pjanić ... 486 The Effect of Credit Risk Management on Banks’ Profitability in Kosovo

Aliu Muhamet, Sahiti Arbana ... 492 Impact of OPEC Policies over the Global Economy: Case of USA

(4)

Copyright © 2016 by Academic Publishing House

Researcher

Published in the Russian Federation

European Journal of Economic Studies

Has been issued since 2012.

ISSN: 2304-9669 E-ISSN: 2305-6282

Vol. 18, Is. 4, pp. 447-458, 2016

DOI: 10.13187/es.2016.18.447

www.ejournal2.com

Articles and Statements

UDC 33

Government Expenditure, Defense Expenditure and Economic Growth: a Causality Analysis for BRICS

Salman Ali Shah a , *, Chen He a , Mao Yu a , Wang Xiaoqin a

a Huazhong University of Science and Technology, Wuhan, Hubei, P.R China

Abstract

This paper empirically examines the effects of civilian and military portions of government expenditure on economic growth of five key emerging economies Brazil, Russia, India, China and South Africa (BRICS). We ran separate Cointegration and Granger causality tests for each country using data taken from WDI and SIPRI while taking account of the limitations of time series data. We got interestingly different effects of military expenditure on economic growth across countries especially for the three nuclear powers Russia, India and China. India and Brazil showed negative, Russia and China showed positive while South Arica showed no effect on economic growth in terms of government civilian expenditure.

Keywords: government expenditure, military expenditure, economic growth, BRICS, cointegration, granger causality, unit root, emerging economies, one way causality, feedback relationship.

1. Introduction

‘‘The single and most massive obstacle to development is the worldwide expenditure onnational defense activity.’’*

The traditional gun-butter tradeoff claims that military spending is a non-productive expenditure. The logic behind this argument is the fact that military expenditure consumes a lot of resources thus leaving little for other economic activities, for instance, investment in public infrastructure, private consumption and investment, social security programs, etc., and thus slows down economic growth (Shieh, Lai, & Chang, 2002). Moreover, substantial military imports can also cause problems in balance of payments. On the other hand, the following quotation puts questions for researchers that need empirical answers;

“There is no way of telling from economic theory whether a greater military effort will slow down or accelerate output growth.”*

* Corresponding author

E-mail addresses: salmanalishah@hust.edu.cn (Salman Ali Shah)

(5)

Nonetheless, studies like (Benoit, 1973; Benoit, 1978; Yildirim†, Sezgin, Öcal, 2005) and

(Yildirim, Öcal, 2014) have proved empirically wrong the conventional belief that military

expenditure negatively affects economic growth. . On the other hand, a plethora of studies do empirically support this argument (Faini, Annez, & Taylor, 1984b), (Lim, 1983b), (Abu-Bader &

Abu-Qarn, 2003), (Galvin, 2003), (Klein*, 2004), and (H.-C. Chang, Huang, & Yang, 2011). Studies

that found out mixed results in cross country analysis include among others, (Chowdhury, 1991),

(Kusi, 1994), (Kollias, Manolas, & Paleologou, 2004), (Chang et al., 2013) and (Pan, Chang,

Wolde-Rufael, 2014).

There are several channels from both supply and demand point of view that show positive effect of military expenditure on economic growth. Regarding the supply-side effect, the defense sector provides a variety of public infrastructure (e.g., dams, communication networks, roads, airports, highways, and other transportation networks), and enhances human capital through education, nutrition, medical care, and training. Moreover, military research and development experience created by arms imports positively affects private production. From the demand-side point of view, defense spending reduces unemployment and increases aggregate demand, thus promoting economic growth. Furthermore, defense spending may favor economic growth since it provides both internal and external security, and therefore enhances private investment and attracts foreign investment. This form is known as military spending growth hypothesis. Growth hypothesis is a one-way Granger causality running from military spending to economic growth. The second form is that military spending is detrimental to economic growth (‘guns or butter’). This hypothesis is built upon the belief that if taxes or borrowings are used to finance military expenditure, it will crowd-out private investment. Otherwise, it takes the resources away from more productive government expenditures, for instance education and health services (Deger &

Smith, 1983); (Lim, 1983a) (Dunne & Vougas, 1999). The second form is called the military

spending growth detriment hypothesis. Growth detriment hypothesis is also a one-way Granger causality running from military spending to economic growth. The relationship between economic growth and military expenditure is bidirectional; that is to say, economic growth is caused by military spending and high military spendings are associated with economic growth. Furthermore, military sending is not exogenous when we consider changes in economic growth (Cappelen, Gleditsch, & Bjerkholt, 1984), (Kusi, 1994), (Kollias et al., 2004). The third form is a feedback hypothesis, which is a two-way Granger causality between military expenditure and economic growth. Finally, there is a fourth form of the relationship between military expenditure and economic growth which states that there is no relationship between military expenditure and economic growth (Biswas & Ram, 1986), (Grobar & Porter, 1989). The fourth form is called neutrality hypothesis, no causal relationship between military expenditure and economic growth. If the relationship between military spending and economic growth is either growth (detriment) hypothesis or feedback hypothesis, then reduction in (increase) military spending may lead to negative economic growth. For this reason, policy-makers need to analyze the relationship between military spending and economic growth to make an appropriate military strategy.

Military spending is qualitatively different from other government spending in many ways. Firstly, military procurements follow more strict acquisition processes and quality requirements than non-military spending (Hartley, 2004). Secondly, military spending is generally sanctioned by the government, independently from other types of spending. Thirdly, there is comparatively little flexibility in shifting military spending to other uses, unlike other spending. Fourthly, in almost every country, there is centralized allocation of military spending, while non-military spending may be allocated by central, state or local governments. While centralization might present different oversight, decentralization can involve more middlemen (Teobaldelli, 2011). Thus, it is highly likely that military and non-military spending have different effects on the economy. There has been an ongoing debate on the relationship between government spending and economic growth. The celebrated “Wagner’s law” postulates that government spending is income elastic and that the ratio of government spending to income tends to grow with economic development. Furthermore, government provides public goods and services (for non-military purposes) such as education,

* The authors point towards the study of Benoit (1973) that claims a positive effect of military expenditure on

(6)

infrastructure, and laws, are often considered as important variables in economic growth. The effects of economic growth on government expenditure have been examined by a plethora of empirical studies using different testing procedures and different measures of government spending (Peacock & Scott, 2000). Since the 1990s, it has become a common practice to test Wagner’s law using times-series techniques such as unit-root and co integration tests (Narayan,

Nielsen, & Smyth, 2008). Using the Swedish data, (Henrekson, 1993) finds no evidence for

Wagner’s law; he also finds that earlier results from time-series studies may be spurious because they did not test the stationarity properties of the data. On the other hand, (Akitoby, Clements,

Gupta, & Inchauste, 2006) empirically supports Wagner’s law by using the co-integration method

to a sample of 51 developing countries. Moreover, a number of studies have examined the effect of government spending on economic growth assuming that an inverted-U relationship exists between the scale of government and economic growth e.g. (Ram, 1986); (Dar & AmirKhalkhali, 2002). (Hansson & Henrekson, 1994) utilize disaggregated data and find that government transfers, consumption and total outlays have negative effects, while educational expenditure has a positive effect, and government investment has no effect on private productivity growth. In a framework of endogenous growth, (Barro, 1990) presented two kinds of predictions; unproductive government expenditure will have a negative effect on economic growth while the role of productive government expenditure on economic growth is unclear; it depends on how the government reacts and how much is the ratio of government spending to GDP. Later on, other studies also find support for negative effect of government spending on economic growth e.g.

(Barro, 1991). The current body of literature generally suggests that developed countries may

confirm Wagner’s law but it is less likely to find support for it in developing countries (Akitoby et al., 2006).

On the other hand, another strand of literature suggests that government spending could have a positive effect on economic growth if it involves public investment in infrastructure, but could have a negative effect if it involves only government consumption. Yet, previous studies have not reached a consensus on the relationship between government spending and economic growth, owing to their differences in the specification of econometric models, the measurement of government expenditures, and the selection of samples (e.g., (Agell, Lindh, & Ohlsson, 1997). As argued by (Abu-Bader & Abu-Qarn, 2003), typical regressions for explaining government spending or economic growth generally focus on the relationship between government spending and economic growth, rather than providing insight into the direction of causality. One popular approach to investigating the causal relationships between the two variables has been using the tests (Granger, 1969). Over the past decades many studies have applied the Granger causality tests to test the causal relationship between government spending and economic growth. (Halicioĝlu, 2003) applies the Granger causality tests to the Turkish data over 1960–2000 and finds neither co-integrated nor causal relationships between per capita GDP and government spending shares. In contrast, several studies find evidence on the Granger causality running from national income to government expenditure, and thus provide support for Wagner’s law e.g.,(Abu-Bader & Abu-Qarn, 2003). In particular, (Dritsakis, 2004) provides evidence on such a causal relationship for Greece and Turkey. By applying the unit-root, co-integration, and the Granger causality tests to panel data,

(Narayan et al., 2008) find that Wagner’s law is supported by the panel of sub-national data on

China’s central and western provinces, but is rejected by the full panel consisting of all Chinese provinces. Using the U.S. data since 1792, (Guerrero & Parker, 2007) find evidence supporting Wagner’s law but not supporting the hypothesis that the size of the public sector Granger causes economic growth.

A wave of literature concerning the BRIC countries has erupted since the term’s creation in 2001 by (O'neill & Goldman, 2001) e.g. (Armijo, 2007); (Cheng, Gutierrez, Mahajan, Shachmurove,

& Shahrokhi, 2007); (Cooper, 2006); (Glosny, 2010); (Macfarlane, 2006). In (Wilson,

Purushothaman, & Goldman, 2003) predicted that in less that forty years, or by 2050, the BRICs’

(7)

economic growth and, more importantly, the prospect of becoming a global power. The acronym thus provoked an ever growing body of literature, many concerning the accuracy of including one BRIC or another in the group and the feasibility for a certain country to realize its ‘BRIC potential’ e.g. (Cooper, 2006); (Desai, 2007); (Macfarlane, 2006); (Sotero & Armijo, 2007). Later on in 2010 South Africa was included in the group of major national economies and thus the acronym is now known as “BRICS”.

The objective of this paper is to test the four hypotheses of government (military or non military) spending in case of five major emerging economies Brazil, Russia, India, China and South Africa. These four hypotheses are,

Growth hypothesis: a one-way Granger causality running from government (military or non military) spending to economic growth. (Positive)

Growth detriment hypothesis: also a one-way Granger causality running from government spending to economic growth. (Negative)

Feedback hypothesis: a two-way Granger causal relationship between government spending and economic growth.

Neutrality hypothesis : No causality between them

We believe our findings will add up to the existing body of literature in two ways. One, our Granger causality analysis will test the causality while our cointegration analysis will determine the direction as well as the nature of the relationship whether it’s positive or negative. Two, our findings will help the policy makers of these rapidly growing economies identify what could be slowing down their growth.

2. Data and Methodology

Annual data ranging from 1988 to 2013 is used in our study for all the countries. All the variables are measured in million dollars and are expressed in logarithms. Data for Gross domestic product and Government consumption is taken from World Development Indicator (WDI) while Military Expenditure’s data is taken from Stockholm International Peace Research Institute (SIPRI). The list and symbols of variables used in our study are as follows.

LGDP: Log of Gross domestic Product used an indicator for economic growth. LGE: Log of Government expenditure

LME: Log of Military Expenditure 2.1 Econometric Methodology:

Our econometric methodology consists of the following steps. 2.1.1 Augmented Dickey Fuller Test:

Since our data set includes time series data, thus we have to test the properties of the time series. In order to find out whether the data is stationary or not, we use Augmented Dickey Fuller test. This test was proposed by Dickey-Fuller (1979) and is widely used in the literature. Economic time series is typically non stationary and non stationary data can give us misleading results. Therefore, such time series should be made stationary or in other words such data should be differenced d times. The time series which is made stationary after differencing is called integrated of order d. When the test value comes out to be greater than the critical value, we interpret that the time series is stationary and vice versa.

2.1.2 Optimal Lag selection:

After testing for stationarity; if the variables are integrated of the same order, the next step is to choose optimal lag length. Different criterions have been used for lag selection in the literature but the most widely used method is to select the lag length suggested by majority of the criterion.

2.1.3 Johansen Co Integration Test:

In order to find the cointegrating relationship among the variables, we use Johansen (1988) test.

Johansen’s procedure starts with VAR of order p and is given by

(8)

t p i i t i i t

t y y e

y      

    1 1

Where I Ai p i   

1 And

j p

i j

i

A

 

1

If the coefficient matrix Π has reduced rank r<n, then there exist nxr matrices α and β each with rank r such that Π = αβ′ and β′y t is stationary. Johansen proposes two different likelihood ratio tests: the trace test and maximum eigenvalue test,

2.1.4 Multivariate VECM Granger Causality Test:

If cointegration exists between the variables then there is causality running between these variables in at least one direction (Granger, 1988). In order to test the causal relationships among the variables we use Granger causality test proposed by Engle and Grnager (1987).

The null hypothesis of Granger causality can be formulated as: H0: Y does not Granger cause X

As per the definition of Granger causality, Y does not cause X if,

0 ...

3 2

1

ti

j

And

X does not cause Y if,

0 ...

3 2

1

ti

j

Granger causality can be interpreted as Y is Granger caused by X if current value of Y can be forecasted with the help of past values of X.

3. Empirical Results

3.1. Augmented Dickey Fuller Test:

In the first step of our analysis, we run Augmented Dickey Fuller test so as to test stationary of our variables. In order to go further with our analysis, our variables should be integrated of the same order. Thus, we present the results of unit root test in Table 1. As evident from the table, all of the variables are non-stationary at first level and are shown stationary after differencing it once. In other words, our variables become stationary at first difference, therefore, we can apply further tests in our analysis.

Table 1. Augmented Dickey Fuller Test

Country Variables Trend Intercept Lag

Length T value value/critical Order Integration of

Brazil lgdp Yes Yes 8 -2.27

(-4.498) Level

∆lgdp Yes Yes 1 -6.34

(-4.39)*** First difference

ge No Yes 5 -0.96

(-4.37) Level

∆lge No Yes 5 -5.05

(-4.39)*** First difference

me Yes Yes 2 -1.60

(-3.61) Level

(9)

Russia lgdp yes yes 5 0.89

(4.37) Level

∆lgdp yes yes 5 4.89

(4.41)*** First difference

ge Yes yes 5 0.37

(4.37) Level

∆lge yes Yes 5 4.80

(4.39)*** First difference

me yes yes 5 2.76

(5.37) Level

∆lme yes yes 5 5.38

(4.39)*** First difference

India lgdp Yes Yes 5 -2.26

(-4.37) Level

∆lgdp yes Yes 5 -4.63

(-4.39)*** First difference

ge Yes Yes 2 -2.60

(-3.61) Level

∆lge No Yes 6 -3.44

(-2.99)** First difference

me No Yes 3 -2.33

(-3.60) Level

∆lme No Yes 3 -3.99

(-3.61)** First difference

China lgdp Yes Yes 5 2.64

(5.39) Level

∆lgdp Yes Yes 5 5.35

(4.41)*** First difference

ge Yes Yes 6 3.14

(4.39) Level

∆lge Yes Yes 6 4.64

(4.41)*** First difference

me No Yes 2 2.16

(3.73) Level

∆lme No Yes 2 4.07

(3.75)*** First difference

S.Africa lgdp Yes Yes 5 3.00

(4.37) Level

∆lgdp Yes Yes 5 4.84

(4.39)** First difference

ge Yes Yes 5 1.33

(4.37) Level

∆lge Yes Yes 5 3.41

(3.26)* First difference

me Yes Yes 5 2.29

(4.37) Level

∆lme Yes Yes 5 3.57

(3.24)* First difference

3.2. Optimal Lag Selection:

(10)

Table 2. Optimal Lag Selection

Country Lag LogL LR FPE AIC SC HQ

Brazil 0 -340.5371 NA 5.44e+08 28.62809 28.77534 28.66716

1 -325.8713 24.44294* 3.42e+08 28.15594 28.74497* 28.31221* 2 -315.9379 14.07237 3.31e+08* 28.07815* 29.10895 28.35163

Russia Lag LogL LR FPE AIC SC HQ

0 -325.7297 NA 1.58e+08 27.39414 27.54140 27.43321

1 -271.6568 90.12148 3737226. 23.63807 24.22710 23.79434 2 -255.5505 22.81729* 2157727.* 23.04588* 24.07667* 23.31935*

India Lag LogL LR FPE AIC SC HQ

0 81.23805 NA 2.96e-07 -6.519838 -6.372581 -6.480770

1 180.1608 164.8713* 1.66e-10* -14.01340 -13.42438* 13.85713* 2 189.5540 13.30705 1.68e-10 -14.04617* -13.01537 -13.77270

China Lag LogL LR FPE AIC SC HQ

0 68.53548 NA 6.73e-07 -5.698737 -5.550630 -5.661489

1 96.50269 46.20670* 1.31e-07 -7.348060 -6.755628* -7.199065 2 108.1699 16.23268 1.09e-07* 7.579994* -6.543239 7.319253*

S. Africa Lag LogL LR FPE AIC SC HQ

0 -109.0955 NA 2.288299 9.341295 9.488552 9.380363

1 -49.76022 98.89220* 0.034821* 5.146685* 5.735712* 5.302954* 2 -43.47541 8.903489 0.045577 5.372951 6.403748 5.646421 3.3. Johansen Cointegration Test:

We present our findings of Trace statistics and Eigen Value statistics in Table 3. Furthermore, cointegration equations for all the 5 countries obtained from Vector Error Correction Model are shown in the same table. Null hypotheses of “no cointegration” among the three variables (Economic growth, Government expenditure and Military expenditure) are rejected in case of our sample countries. Thus it is inferred, there is one cointegrating vector in case of each of the trivariate system of our variables.

Table 3. Johansen Cointegration Test

Country Hypothesized Trace Critical Max-Eigen Critical

No. of CE(s) Statistic Value at

0.05 Statistic Value at 0.05

Brazil H0: r = 0 32.51* 29.84 23.56* 21.13

H0: r ≤ 1 8.94 15.49 8.88 14.26

H0: r ≤ 2 0.05 3.84 0.05 3.84

Cointegrating equation

Lgdp = 9.66 - 2.33 lge*** - 0.03 lme*** (1)

(5.74) (-5.14)

Country Hypothesized Trace Critical Max-Eigen Critical

No. of CE(s) Statistic Value at

0.05 Statistic Value at 0.05

Russia H0: r = 0 46.42* 29.79 37.94* 21.13

H0: r ≤ 1 8.48 15.49 8.38 14.26

H0: r ≤ 2 0.09 3.84 0.09 3.84

Cointegrating equation

(11)

Country Hypothesized Trace Critical Max-Eigen Critical No. of CE(s) Statistic Value at

0.05 Statistic Value at 0.05

India H0: r = 0

55.03* 29.79 43.69* 21.13

H0: r ≤ 1 11.34 15.49 11.32 14.26

H0: r ≤ 2 0.01 3.84 0.01 3.84

Cointegrating equation

Lgdp = -7.16 - 1.23 lge* + 1.07 lme** (3) (-2.48) (-2.33) (3.09)

Country Hypothesized Trace Critical Max-Eigen Critical

No. of CE(s) Statistic Value at

0.05 Statistic Value at 0.05

China H0: r = 0 39.54* 24.27 32.13* 18.51

H0: r ≤ 1 11.80 12.32 10.90 12.20

H0: r ≤ 2 0.24 4.12 0.24 4.12

Cointegrating equation

Lgdp = 5.57 + 0.57 lge* + 1.50 lme (4) (-3.00) (-2.28) (-1.90)

Country Hypothesized Trace Critical Max-Eigen Critical

No. of CE(s) Statistic Value at

0.05 Statistic Value at 0.05

S. Africa H0: r = 0 32.62* 29.79 24.83* 21.13

H0: r ≤ 1 7.79 15.49 7.66 14.26

H0: r ≤ 2 0.13 3.84 0.13 3.84

Cointegrating equation

Lgdp = -16.49 - 0.43 lge - 3.16 lme*** (5) (-2.23) (1.11) (5.42)

Our cointegration equation for Brazil shows statistically significant and negative relationship of military expenditure and government expenditure with the economic growth. Further, a 0.03 percent change in military expenditure will reduce the economic growth by one percent while the same decrease in the economic growth of Brazil is caused by a 2.33 percent change in the government civilian expenditure. The equation for Russia shows a positive effect of government civilian expenditure on economic growth while the defense expenditure causes the economic growth to reduce. In quantitative terms, 1.07 percent increase in defense expenditure causes the economic growth to reduce by one percent. On the other hand, economic growth is enhanced by one percent with a 4.80 increase in the government civilian expenditure. It is evident from Table 3 that the economic growth of India reduces by one percent with the increase in government expenditure by 1.23 percent while it is increased by one percent when military expenditure is increased by 1.07 percent. The cointegration results show the relationship between economic growth and military expenditure of China is statistically insignificant while a one percent increase in economic growth is observed when government civilian expenditure is increased by 0.57 percent. Finally, our cointegration equation for South Africa shows no statistically significant relationship of government expenditure with economic growth while it shows the economic growth is reduced by one percent when the military expenditure is increased by 3.16 percent.

3.4. VECM Granger causality test:

(12)

helps us to determine the weak exogeneity among variables. This test suggests us the causal relationship of one variable with the other variable. The results of VECM Granger causality test are reported in Table 4. The significant chi-square statistic shows the dependent variable is Granger caused by the independent variable. Table 4 shows bidirectional causality between economic growth and government expenditure in case of Brazil. Unidirectional causality running from government expenditure to economic growth has been found in case of Russia, China and India while no statistically significant relationship can be detected for South Africa. In our trivariate analysis, we found unidirectional causality running from growth to military expenditure for Brazil, unidirectional causality running from military expenditure to growth in case of Russia and India while no relationship was found between military expenditure and growth for China. We found bidirectional causality between growth and military expenditure in case of South Africa.

Table 4. Multivariate Granger Causality Test

Country Independent

Variables

Brazil Independent

Dependent

lgdp lge lme

lgdp --- 5.22* 3.04

lge 14.80*** --- 0.79

lme 5.77* 2.02 ---

Russia Independent

Dependent

lgdp lge lme

lgdp --- 12.55* 14.83**

lge 9.28 --- 13.47**

lme 8.79 16.29*** ---

India Independent

Dependent

lgdp lge lme

lgdp --- 7.31** 12.02***

lge 0.14 --- 1.89

lme 1.44 11.94*** ---

China Independent

Dependent

lgdp lge lme

lgdp --- 5.45** 0.77

lge 0.16 --- 0.38

lme 0.08 0.30 ---

S. Africa Independent

Dependent

lgdp lge lme

lgdp --- 0.88 11.95***

lge 1.28 --- 0.026

(13)

4. Results

We will sum up our findings from the statistical analysis for all the five countries in this section. Our trivariate analysis for Brazil reveals there is a negative long run causality running from government civilian expenditure to economic growth which means our growth detriment hypothesis holds true for Brazil. Furthermore, two ways causality between government expenditure and economic growth was also found in case of Brazil, thus accepting our feedback hypothesis. To abridge our findings for Russia, one way positive causality from government expenditure to economic growth while negative causality from military expenditure to growth is detected. Therefore, growth hypothesis is accepted for government civilian expenditure and growth detrimental hypothesis is accepted for government military expenditure. Our findings for India affirm growth detrimental hypothesis for government civilian expenditure, i.e. unidirectional negative causality running from government spending to economic growth. These findings further affirm growth hypothesis for government military expenditure, i.e. unidirectional positive causality running from government spending to economic growth. We found government civilian expenditure to positively affect economic growth in case of China, thus proving growth hypothesis true. No statistically significant relationship was found between military spending and economic growth for Chinese data. Summarizing our findings for South Africa, bidirectional causality between military spending and economic growth is detected which confirms feedback hypothesis.

5. Conclusion

Our aim in this study was to find out whether there is any causal relationship between economic growth and both civilian and military portions of government expenditure in five emerging economies recently known as BRICS, i.e. Brazil, Russia, India, China and South Africa. Since it is generally believed that military expenditure can slow down economic growth, we examined the effects of military expenditure on economic growth of these five major emerging economies of the world. Our results for the 3 nuclear powers in our analysis, i.e. Russia, India and China were interestingly different from each other. Russian data showed negative effect of military spending on economic growth, Indian data showed positive effect while Chinese data suggested insignificant impact of military spending on economic growth for our sample period. The implications for these findings are straightforward; our sample period starts from 1988 and ends on 2015 which was a particularly rough period for Russia. The Afghan war and the separation of 6 central Asian states from USSR forced Russia to spend serious money on military which shook its economy. Chinese economy has been boosting for the last few decades and our findings might imply that Chinese economy is too strong for its military expenditure to affect it. The implication for positive impact of military spending on Indian economy might be the investment on public infrastructure, hospitals, education and etc. by military organizations. Our findings for Brazil and South Africa indicate that military spending slow down economic growth of both the countries.

Government civilian expenditure of India and Brazil showed negative effect on economic growth, therefore we suggest the policy makers of these countries to reduce their government spending and/or reallocate it to productive projects. In case of Brazil, shifting resources from military to civilian spending may not enhance economic growth since government civilian expenditure itself is reducing economic growth. Thus, the government should look for civilian productive activities to foster economic growth. Russian and Chinese data gave positive response to economic growth for our sample period. Therefore, we conclude that only the military portion of government spending has been a burden on Russian economy while Chinese economy was being neutral to military spending. Our analysis for South Africa suggested statistically insignificant relationship of government civilian expenditure with economic growth, hence we conclude by suggesting reduction in its military spending which is causing its economy to slow down.

References

Abu-Bader, Abu-Qarn, 2003 – Abu-Bader, S., & Abu-Qarn, A.S. (2003). Government

expenditures, military spending and economic growth: causality evidence from Egypt, Israel, and Syria. Journal of Policy Modeling, 25(6), 567-583.

Agell et al., 1997 – Agell, J., Lindh, T., & Ohlsson, H. (1997). Growth and the public sector:

(14)

Akitoby et al., 2006 – Akitoby, B., Clements, B., Gupta, S., & Inchauste, G. (2006). Public spending, voracity, and Wagner's law in developing countries. European Journal of Political Economy, 22(4), 908-924.

Armijo, 2007 – Armijo, L.E. (2007). The BRICs countries (Brazil, Russia, India, and China)

as analytical category: mirage or insight? Asian Perspective, 7-42.

Barro, 1990 – Barro, R.J. (1990). Government Spending in a Simple Model of Endogeneous

Growth. Journal of Political Economy, 98(5, Part 2), S103-S125. doi: 10.1086/261726

Barro, 1991 – Barro, R.J. (1991). Economic Growth in a Cross Section of Countries. The

Quarterly Journal of Economics, 106(2), 407-443. doi:10.2307/2937943

Benoit, 1973 – Benoit, E. (1973). Defense and economic growth in developing countries.

Benoit, 1978 – Benoit, E. (1978). Growth and defense in developing countries. Economic

development and cultural change, 26(2), 271-280.

Biswas, Ram, 1986 – Biswas, B., & Ram, R. (1986). Military expenditures and economic

growth in less developed countries: An augmented model and further evidence. Economic development and cultural change, 34(2), 361-372.

Cappelen et al., 1984 – Cappelen, Å., Gleditsch, N. P., & Bjerkholt, O. (1984). Military

spending and economic growth in the OECD countries. Journal of Peace Research, 21(4), 361-373.

Chang, Yang, 2011 – Chang, H.-C., Huang, B.-N., & Yang, C.W. (2011). Military expenditure

and economic growth across different groups: A dynamic panel Granger-causality approach.

Economic Modelling, 28(6), 2416-2423.

Chang, Chu, 2013 – Chang, T., Lee, C.-C., & Chu, H.-P. (2013). Revisiting the Defense–

Growth nexus in European countries. Defence and Peace Economics(ahead-of-print), 1-16.

Cheng et al., 2007 – Cheng, H. F., Gutierrez, M., Mahajan, A., Shachmurove, Y., &

Shahrokhi, M. (2007). A future global economy to be built by BRICs. Global Finance Journal, 18(2), 143-156.

Chowdhury, 1991 – Chowdhury, A. R. (1991). A causal analysis of defense spending and

economic growth. Journal of Conflict Resolution, 35(1), 80-97.

Cooper, 2006 – Cooper, J. (2006). Of BRICs and brains: Comparing Russia with China,

India, and other populous emerging economies. Eurasian Geography and Economics, 47(3), 255-284.

Dar, AmirKhalkhali, 2002 – Dar, A. A., & AmirKhalkhali, S. (2002). Government size, factor

accumulation, and economic growth: evidence from OECD countries. Journal of Policy Modeling,

24(7), 679-692.

Deger, Smith, 1983 – Deger, S., & Smith, R. (1983). Military expenditure and growth in less

developed countries. Journal of conflict resolution, 27(2), 335-353.

Desai, 2007 – Desai, R. (2007). Dreaming in technicolour? India as a BRIC economy.

International Journal, 62(4), 781-804.

Dritsakis, 2004 – Dritsakis, N. (2004). Defense spending and economic growth: an empirical

investigation for Greece and Turkey. Journal of Policy Modeling, 26(2), 249-264.

Dunne, Vougas, 1999 – Dunne, P., & Vougas, D. (1999). Military Spending and Economic

Growth in South Africa a Causal Analysis. Journal of conflict resolution, 43(4), 521-537.

Faini et al., 1984 – Faini, R., Annez, P., & Taylor, L. (1984). Defense spending, economic

structure, and growth: Evidence among countries and over time. Economic development and cultural change, 32(3), 487-498.

Galvin, 2003 – Galvin, H. (2003). The impact of defence spending on the economic growth

of developing countries: a cross-section study. Defence and Peace Economics, 14(1), 51-59.

Glosny, 2010 – Glosny, M.A. (2010). China and the BRICs: A Real (but Limited) Partnership

in a Unipolar World. Polity, 42(1), 100-129.

Granger, 1969 – Granger, C.W. (1969). Investigating causal relations by econometric models

and cross-spectral methods. Econometrica: Journal of the Econometric Society, 424-438.

Grobar, Porter, 1989 – Grobar, L. M., & Porter, R. C. (1989). Benoit revisited: defense

spending and economic growth in LDCs. Journal of conflict resolution, 33(2), 318-345.

Guerrero, Parker, 2007 – Guerrero, F., & Parker, E. (2007). The Effect of Federal

Government Size on Long-Term Economic Growth in the United States, 1792-2004.

(15)

Hansson, Henrekson, 1994 – Hansson, P., & Henrekson, M. (1994). A new framework for testing the effect of government spending on growth and productivity. Public Choice, 81(3-4), 381-401.

Hartley, 2004 – Hartley, K. (2004). The economics of military outsourcing. Defence Studies,

4(2), 199-206.

Henrekson, 1993 – Henrekson, M. (1993). Wagner's law-a spurious relationship? Public

Finance, 46(3).

Klein, 2004 – Klein*, T. (2004). Military expenditure and economic growth: Peru 1970–

1996. Defence and Peace Economics, 15(3), 275-288.

Kollias et al., 2004 – Kollias, C., Manolas, G., & Paleologou, S.-M. (2004). Defence

expenditure and economic growth in the European Union: a causality analysis. Journal of Policy Modeling, 26(5), 553-569.

Kusi, 1994 – Kusi, N.K. (1994). Economic growth and defense spending in developing

countries: a causal analysis. Journal of conflict resolution, 152-159.

Lim, 1983 – Lim, D. (1983a). Another look at growth and defense in less developed countries.

Economic development and cultural change, 31(2), 377-384.

Macfarlane, 2006 – Macfarlane, S. (2006). The ‘R’in BRICs: is Russia an emerging power?

International Affairs, 82(1), 41-57.

Narayan et al., 2008 – Narayan, P. K., Nielsen, I., & Smyth, R. (2008). Panel data,

cointegration, causality and Wagner's law: Empirical evidence from Chinese provinces. China Economic Review, 19(2), 297-307.

O'neill, Goldman, 2001 – O'neill, J., & Goldman, S. (2001). Building better global economic

BRICs.

Pan et al., 2014 – Pan, C.-I., Chang, T., & Wolde-Rufael, Y. (2014). Military spending and economic growth in the Middle East countries: bootstrap panel causality test. Defence and Peace Economics(ahead-of-print), 1-14.

Peacock, Scott, 2000 – Peacock, A., & Scott, A. (2000). The curious attraction of Wagner's

law. Public Choice, 102(1-2), 1-17.

Ram, 1986 – Ram, R. (1986). Government size and economic growth: A new framework and

some evidence from cross-section and time-series data. The American Economic Review, 76(1), 191-203.

Shieh et al., 2002 – Shieh, J.-y., Lai, C.-c., & Chang, W.-y. (2002). The impact of military burden on long-run growth and welfare. Journal of Development Economics, 68(2), 443-454.

Sotero, Armijo, 2007 – Sotero, P., & Armijo, L.E. (2007). Brazil: To be or not to be a BRIC?

Asian Perspective, 43-70.

Teobaldelli, 2011 – Teobaldelli, D. (2011). Federalism and the shadow economy. Public

Choice, 146(3-4), 269-289.

Wilson et al., 2003 – Wilson, D., Purushothaman, R., & Goldman, S. (2003). Dreaming with

BRICs: The path to 2050 (Vol. 99): Goldman, Sachs & Company.

Yildirim, Öcal, 2014 – Yildirim, J., & Öcal, N. (2014). Military expenditures, economic

growth and spatial spillovers. Defence and Peace Economics(ahead-of-print), 1-18.

Yildirim† et al., 2005 – Yildirim†, J., Sezgin, S., & Öcal, N. (2005). Military expenditure and

(16)

Copyright © 2016 by Academic Publishing House

Researcher

Published in the Russian Federation

European Journal of Economic Studies

Has been issued since 2012.

ISSN: 2304-9669 E-ISSN: 2305-6282

Vol. 18, Is. 4, pp. 459-469, 2016

DOI: 10.13187/es.2016.18.459

www.ejournal2.com

UDC 33

The Impact of Energy Consumption, Trade Openness and Financial Development on Economic Growth: Empirical Evidence from Turkey (1980–2014)

Murat Cetin a , *

a Faculty of Economics and Administrative Sciences, Namik Kemal University, Tekirdag, Turkey

Abstract

The developments in Turkish economy indicate that energy, trade openness and financial development are critical determinants of economic growth. This study aims to investigate the impact of energy consumption, trade openness and financial development on economic growth in case of Turkey over the sample period 1980-2014. The results of unit root tests reveal that the variables are integrated at I(1). The results of ARDL bounds test and Johansen-Juselius technique reveal that there exists a long-run relationship among energy consumption per capita, trade openness, domestic credit provided by banking sector and real GDP per capita. Energy consumption and financial development have a positive impact on economic growth while there do not a statistically significant relationship between trade openness and economic growth in the long run. The VECM Granger causality results show that there exist a uni-directional causal linkage running from energy consumption, trade openness and financial development to economic growth in the long run. The empirical findings can provide several policy implications for Turkish economy over the period.

Keywords: energy consumption, trade openness, financial development, economic growth, cointegration, causality, Turkey.

1. Introduction

The determinants of economic growth has long been argued by theoretical and empirical literature. It is well known that economic growth has been affected by energy, trade and financial development (Goldsmith, 1969; Yu, Choi, 1985; Barro, Sala-i-Martin, 1997). Energy-growth literature reveals the existence of four hypotheses on the link between energy consumption and economic growth. These theories explain the causal linkages between the variables. According to the growth hypothesis energy consumption is very important for economic growth implying that there exists a uni-directional causality running from energy consumption to economic growth

(Altinay, Karagol, 2004).The conservation hypothesis suggests that there exists a uni-directional

causality running from economic growth to energy consumption (Payne, 2010). The feedback hypothesis implies a bi-directional causality between energy consumption and economic growth

(Soytas, Sari, 2003).The neutrality hypothesis assumes that there exist no causal linkages between

energy consumption and economic growth (Zhang, Xu, 2012).

(17)

Trade openness is an important determinant of economic growth. (Bhagwati, 1978; Romer

1986; Grossman and Helpman, 1990; Rivera-Batiz, Romer, 1991; Taylor, 1993) are the main

theoretical studies investigating the link between international trade and economic growth. Generally, these literature reveal the presence of a consensus that trade openness causes economic growth. Financial development is linked with economic growth. Several theoretical studies such as

(McKinnon, 1973; King, Levine, 1993; Warman, Thirlwall, 1994) discuss the link between financial

development and economic growth. According to these literature, financial development will cause economic growth through productive and efficient use of financial resources.

Kraft and Kraft (1978) is the first study investigating the relationship between energy consumption and economic growth. The study reveals a uni-directional causality running from economic growth to energy consumption. Erol and Yu (1988) show that there exists a bi-directional causality between energy consumption and economic growth for Japan. The study finds a uni-directional causality running from energy consumption to economic growth for Canada. The study also finds a uni-directional causality running from economic growth to energy consumption for Germany. Masih and Masih (1996) find a uni-directional causality running from energy consumption to economic growth for India and Indonesia. In the study, a uni-directional causal linkage running from economic growth to energy consumption is found. In addition, there exists a bi-directional causality between the variables in Pakistan. Asafu-Adjaye (2000) finds no causality for Indonesia and India. Soytas and Sari (2003) indicate that there exists a uni-directional causality running from economic growth to energy consumption in Italy and Korea. The study aslo indicates that there exists a uni-directional causality running from energy consumption to economic growth in France, Germany, Japan and Turkey. Farhani and Rejeb (2012) reveal that there exists a uni-directional causality running from economic growth to energy consumption in low and high income countries. This study also reveals that there exists a bi-directional causality between the variables in upper-middle income countries.

Barro (1991), Edwards (1998) and Frankel and Romer (1999) examine the link between trade openness and economic growth through the cross-country regression analysis. Empirical results show that trade openness is positively correlated with economic growth. Musila and Yiheyis (2015) investigate the effect of trade openness on economic growth in case of Kenya. Regression analysis reveals that trade openness is positively linked with economic growth. But, the impact is found to be statistically insignificant.

Applying the Johansen-Juselius cointegration method and Granger causality test, Jenkins and Katircioglu (2010) explore the long run relationship among international trade, financial development and economic growth for Cyprus. Empirical results imply that there exists a long run relationship between international trade, financial development and economic growth. Empirical results also imply that there exists a uni-directional causality running from economic growth to financial development and international trade. Gokmenoglu et al. (2015) deal with the links between international trade, financial development and economic growth in Pakistan. The Granger causality analysis indicates that there exists a uni-directional causality running from financial development to economic growth.

In recent years, several studies such as Shahbaz et al. (2013), Muhammad et al. (2015) and Kumar et al. (2015) examine the relationship between energy consumption, trade, financial development and economic growth. However, these studies provide inconclusive findings and do not investigate Turkish economy. For example, Shahbaz et al. (2013) investigate the relationship between energy consumption, trade, financial development and economic growth in China by using the ARDL bounds testing approach to cointegration and VECM Granger causality method. The empirical results show that energy consumption, trade openness and financial development positively affect economic growth. The Granger causality analysis indicates that a uni-directional causality running from energy consumption to economic growth exists. The Granger causality analysis also indicates that there exists a bi-directional causality between international trade and economic growth and, financial development and economic growth.

(18)

In addition, a bi-directional causality between energy consumption and economic growth, and a uni-directional causality running from trade and financial development to economic growth are found in the long run.

Kumar et al. (2015) examine the effect of energy consumption, trade openness and financial development on economic growth in case of South Africa. Using the ARDL bounds and the Bayer and Hanck cointegration tests, the study shows that trade openness and energy consumption positively affect economic growth in the long run. The study also shows that financial development negatively affects economic growth in the long run. The Toda-Yamamoto causality analysis indicates that there exists a bi-directional causality between trade openness and economic growth. The Toda-Yamamoto causality analysis also indicates that there exists no causal linkage between energy consumption and economic growth. In addition, financial development does not cause economic growth.

Energy consumption, financial development and trade openess are crucial factors for economic growth of Turkish economy. Therefore, the objective of present study is to explain the impact of energy consumption, financial development and trade openess on economic growth in case of Turkey over the period of 1980-2014. The stationarity properties of the variables are investigated through different unit root tests. The study implements the ARDL bounds testing approach to cointegration to examine the long run relationship among the variables. In addition, the study applies the VECM Granger causality approach to explore the causal linkages between the variables. The findings are expected to present several implications for energy, financial and trade policies to sustain economic growth in Turkey.

The rest of the study is organized as follows. Section 2 deals with econometric specification and data description. Section 3 describes the methodology used in the study. Secton 4 reports the empirical findings. Conclusion and policy implications are offered in Section 5.

2. Econometric Specification and Data Description

In this study, the standard log-linear model is used to investigate the impact of energy consumption, trade openness and financial development on economic growth as it can present more efficient results. Following Shahbaz et al. (2013) and Kyophilavong et al. (2015) the long run relationship between the variables is specified as follows:

where, gdpt is per capita real GDP (constant 2010 US$), energy is per capita energy

consumption (kg of oil equivalent), finance is financial development (domestic credit to private sector, % of GDP) and trade is the openness ratio (foreign trade, % of GDP). µt is the regression

error term. The annual data covers the sample period 1980-2014. The Turkish economy has witnessed many radical changes and structural reforms since the 1980s (Terterov and Rosenblatt, 2006). Therefore, this sample period is selected to analyze the links among the variables. The data is obtained from the World Development Indicators (WDI) online database. All the series are converted to their logarithmic form.

The parameters, βi, i=1, 2, 3, indicate the long-run elasticities of per capita real GDP with

respect to per capita energy use, domestic credit to private sector (% of GDP) and trade openness, respectively. Under the energy, finance and trade-led eonomic growth hypotheses, the signs of β1,

β2 andβ3 are expected to be positive (Shaw, 1973; Levine, 1997; Payne, 2010; Yenokyan et al.,

2014).

(19)

Table 1. Descriptive Statistics and Correlation Matrix (Time Series Data: 1980-2014, Observations=35)

Statistics/Variables lgdp lenergy lfinance ltrade

Mean 8.909 6.984 3.119 3.692

Median 8.910 7.026 2.897 3.765

Std. dev. 0.257 0.239 0.468 0.295

Min. 8.473 6.559 2.609 2.838

Max. 9.327 7.353 4.312 4.094

Skewness 0.008 -0.145 1.324 -0.874

Kurtosis 1.931 1.994 3.548 3.446

Observations 35 35 35 35

lgdp 1.000

lenergy 0.994 1.000

lfinance 0.753 0.730 1.000

ltrade 0.879 0.889 0.605 1.000

8.4 8.6 8.8 9.0 9.2 9.4

1980 1985 1990 1995 2000 2005 2010

lgdp

6.4 6.6 6.8 7.0 7.2 7.4

1980 1985 1990 1995 2000 2005 2010

lenergy

2.4 2.8 3.2 3.6 4.0 4.4

1980 1985 1990 1995 2000 2005 2010

lfinance

2.8 3.2 3.6 4.0 4.4

1980 1985 1990 1995 2000 2005 2010

ltrade

Fig. 1. Trends of the Series in Turkey 3. Methodology

The present study aims at examining the relationship between energy consumption, financial development, trade openness and economic growth over the period 1980-2014. The unit root properties of the variables are determined by different unit root tests. The ARDL bounds testing approach to cointegration is applied to investigate the presence of long run relationship among the variables. In addition, the VECM Granger causality framework is applied to determine the causal links between the variables.

3.1 Cointegration Analysis

(20)

for small sample. In addition, a dynamic unrestricted error correction model (UECM) includes the short run and the long run dynamics (Pesaran and Shin, 1999; Pesaran et al., 2001). The equation of UECM model is expressed as follows:

where, α0, Δ and εt are the constant, the first difference operator and the random error term,

respectively. The appropriate lag order is selected by the Akaike Information Criterion (AIC). The ARDL bounds test uses F-statistic to determine the existence of cointegration between the variables. This test compares the computed F-statistic with the upper critical bound (UCB) and lower critical bound (LCB). These critical bounds are presented by Pesaran et al. (2001) and

Narayan (2005). Here, the null and the alternative hypotheses are and

, respectively. There exists a cointegration between the variables when the computed F-statistic exceeds the UCB. There exists no cointegration between the variables

when the computed F-statistic below the LCB. The finding is uncertain when the computed

F-statistic falls between the UCB and LCB.

Several diagnostic tests can be used to examine the robustness of the ARDL model. These are serial correlation, functional form, normality of error term and heteroskedasticity tests. Additionally, the cumulative sum of recursive residuals (CUSUM) and the cumulative sum of squares of recursive residuals (CUSUMsq) tests developed by Brown et al. (1975) can be applied to investigate the stability of the ARDL parameters.

3.2 Granger Causality Analysis

The cointegration methods do not provide any information about the direction of causality. This study uses the VECM Granger causality test as this method examines the long run and the short run causality between the variables. The VECM specification is expressed as follows:

)

3

(

lg

)

1

(

lg

)

1

(

4 3 2 1 1 1 1 1 1 44 43 42 41 34 33 32 31 24 23 22 21 14 13 12 11 1 4 3 2 1

    

t t t t t t t t t i i i i i i i i i i i i i i i i p i t t t t

ECT

ltrade

lfinance

lenergy

dp

x

b

b

b

b

b

b

b

b

b

b

b

b

b

b

b

b

L

a

a

a

a

ltrade

lfinance

lenergy

dp

L

where (1- L) is the lag operator and ECTt−1 is the lagged error correction term. This term is

obtained from the long run specification. ε1t, ε2t, ε3t andε4t are error terms assumed to be N (0,σ).

A significant F-statistic on the first differences of the variables implies the presence of a short run causality between the variables. In addition, a significant t-statistic on the coefficient of ECTt-1

implies the existence of a long run cusality between the variables. 4. Empirical Findings

The study applies several unit root tests such as DF-GLS, PP and Ng-Perron methods to explore the unit root properties of the series. Ng-Perron tests provide more reliable results compared to classical unit root tests. Additionally, it can be more suitable for small sample size (Alimi, 2014).

(21)

integrated at I(1). The results also imply that the ARDL bounds testing approach to cointegration can be applied to test the presence of cointegration between the variables.

Table 2. The Unit Root Tests Results

Regressor DF-GLS PP Ng-Perron

(t) (Adj. t) MZa MZt MSB MPT

lgdp 0.742 -0.479 1.473 1.434 0.973 72.223

lenergy 0.474 -0.886 1.180 1.136 0.963 67.146

lfinance 1.038 0.570 2.636 1.332 0.505 28.477

ltrade -0.720 -2.925 -0.074 -0.041 0.560 21.852

Δlgdp -6.572*** -7.857*** -16.128*** -2.836*** 0.175** 1.530***

Δlenergy -5.700*** -6.628*** -16.336*** -2.857*** 0.174** 1.499***

Δlfinance -3.925*** -4.395*** -14.176*** -2.659*** 0.187** 1.740***

Δltrade -3.983*** -5.801*** -14.089*** -2.644*** 0.187** 1.775***

Notes: The model with constant and trend is used for unit root analysis. The optimal lag length is selected automatically using SBC for ADF test and the bandwidth is selected using the Newey-West method for PP test. *** and ** denote the significant at 1 % and 5 % level of significance, respectively.

Table 3 reports the results of bounds F-test for cointegration. As noted in Table 4, we use

critical bounds obtained by Pesaran et al. (2001) and Narayan (2005). According to Pesaran et al. (2001) critical values, the results show that calculated F-statistic is greater than UCB at 1 per cent. According to Narayan (2005) critical values, the results show that calculated F-statistic is greater than UCB at 5 per cent. All the findings indicate that the series are cointegrated implying that there exists a long run relationship between per capita energy consumption, domestic credit to private sector, trade openness and per capita real GDP for Turkish economy over the period of 1980–2014. The results for diagnostic tests of ARDL model are also reported in the lower part of Table 3. The findings show that the ARDL model passes all the tests successfully.

Table 3. Cointegration Test Results Bounds testing approach to cointegration

Model ARDL lag order Calculated F-statistics

F(lgdp/lenergy, lfinance,

ltrade) [2,1,0,0] 7.411

Peseran et al. (2001) critical value bounds of the F-statistic: unrestricted intercept and unrestricted trend

Significance level Lower bounds, I(0) Upper bounds, I(1)

1% 5.17 6.36

5% 4.01 5.07

10% 3.47 4.45

Narayan (2005) critical value bounds of the F-statistic: unrestricted intercept and unrestricted trend (T = 35)

Significance level Lower bounds, I(0) Upper bounds, I(1)

1% 6.38 7.73

5% 4.56 5.79

10% 3.80 4.88

Diagnostic tests

R2 0.988

Adjusted-R2 0.930

F-statistic 17.259***

Breusch-Godfrey LM test 3.320(0.142)

Figure

Updating...

Download now (83 pages)