# GDP, energy consumption and financial development in Italy

## Abstract

### Purpose

This study aims to explore the relationship among energy consumption, real income, financial development and oil prices in Italy over the period 1960-2014.

### Design/methodology/approach

Different econometric techniques – such as the General Methods of Moment (GMM) or the AutoRegressive Distributed Lags (ARDL) bounds test – are usually used in the empirical analysis. Moreover, both the Toda and Yamamoto causality tests and the Granger causality tests are applied to the data.

### Findings

The results of unit root and stationarity tests show that the variables are non-stationary at levels, but stationary in first-differences form, or I(1). The ARDL bounds *F*-test reveals an evidence of a long-run relationship among the four variables at 1% significance level. Moreover, an increase in real GDP and oil prices has a significant effect on energy consumption in the long run. The coefficients of estimated error correction term are also negative and statistically significant. In addition, the paper explores the causal relationship between the variables by using a VAR framework, with Toda and Yamamoto but also Granger causality tests, within both multivariate and bivariate systems. The findings indicate that energy consumption is affected by real GDP.

### Originality/value

The study also filled the literature gap of applying ARDL technique to examine this relevant issue for Italy.

## Keywords

#### Citation

Magazzino, C. (2018), "GDP, energy consumption and financial development in Italy", *International Journal of Energy Sector Management*, Vol. 12 No. 1, pp. 28-43. https://doi.org/10.1108/IJESM-01-2017-0004

### Publisher

:Emerald Publishing Limited

Copyright © 2018, Emerald Publishing Limited

## 1. Introduction

Italy has some indigenous production of oil and natural gas, but both oil and gas production will progressively decline in the coming years. In 2012, Italy’s total domestic oil production met only 7.7 per cent of its domestic demand. Italy relies heavily on imports and is the world’s largest net importer of electricity (Magazzino, 2014a). In 2011, 47.5 billion kWh was imported, and only 1.8 billion kWh exported, this net import level being typical of the past decade. As of 2011, the transport sector contributed most to Italy’s final energy consumption, with 30.4 per cent, with the residential (24.8 per cent) and industrial (22.8 per cent) sectors also contributing significantly. Thus, it should be clear why it is crucial for Italy to analyze the relationship among energy consumption, financial development and real income (Magazzino, 2012).

The nexus between economic growth and financial development, as well as energy consumption and economic growth, has been the subject of intense research in the past decades. Recent studies have documented that financial development can affect energy use. Financial development helps industrial growth, creates demand for new infrastructure and, thus, positively impacts energy use. Nevertheless, the empirical evidence remains controversial and ambiguous. Financial development may attract foreign direct investment, so as to boost economic growth and increase carbon emissions. Financial development is important because it can increase the economic efficiency of a country’s financial system and this can affect economic activity and the demand for energy. If financial development is found to affect the demand for energy, then this relationship can affect energy policy and carbon emissions strategies. Improvement in monetary transmission mechanism, as a result of financial liberalization, also encourages savings and investment and enhances economic growth. Literature shows that liberalization of financial markets leads to economic growth. Although several papers examine the relationship among energy consumption, economic growth and financial development, few studies concern Italy.

In the present paper, the relationship among real GDP, energy consumption and financial development in Italy has been investigated for the period 1960-2014, using time-series methodologies. The results might help to define and implement the appropriate energy development policies in Italy.

Different econometric techniques – such as the General Methods of Moment (GMM) or the AutoRegressive Distributed Lags (ARDL) bounds test – are usually used in the empirical analysis. Our study also filled the literature gap of applying the ARDL technique to examine this relevant issue for Italy. Moreover, we apply both the Toda and Yamamoto causality tests and the Granger causality tests to our data. In fact, previous studies devoted to the Italian case analyzed the relationship among economic activity, energy consumption and financial development through standard Granger causality approach.

Besides, the Introduction, the outline of this paper proceeds as follows: Section 2 provides a survey of the economic literature on the nexus among energy consumption, real income and financial development. Section 3 contains an overview of the applied empirical methodology and a brief discussion of the data used. Section 4 discusses our empirical results. Section 5 presents some concluding remarks, and finally, Section 6 gives suggestions for future research.

## 2. Literature review

The relationship between financial development and energy consumption has newly started to be discussed in energy economics literature (Çoban and Topcu, 2013). As shown in the survey of the literature presented below, both time-series and panel data studies have been published in the past decade.

As concerns time-series studies, Ali *et al.* (2015), using an ARDL bounds test framework, analyzed the dynamics of financial development, economic growth, energy prices and energy consumption in Nigeria for the period of 1972Q1-2011Q4. The error correction model (ECM) results show that in the short run, financial development has a significant negative impact on fossil fuel consumption. However, energy prices have a positive and significant influence on the consumption of fossil fuel. Mahalik and Mallick (2014) investigated the relationship among energy consumption, economic growth and financial development in India using annual data for the period 1971-2009. The ARDL approach suggests that energy consumption is positively and significantly affected by proportion of urban population in total population, while the same is negatively and significantly impacted by financial development, economic growth and proportion of industrial output in total output. Salman and Atya (2014) test the causality flow among financial development, economic growth and energy consumption in Algeria, Egypt and Tunisia. The ECM results are mixed. The study of Abalaba and Dada (2013) analyzed the relationship among energy consumption, real output, financial development, monetary policy rate and consumer prices, providing – via standard Granger causality tests – weak evidence in support of a long-run relationship between energy consumption and economic growth. Moreover, energy consumption positively influenced output growth in the short run. Islam *et al.* (2013) studied the relationship among financial development, energy consumption and GDP in Malaysia, covering the years 1971-2009. The results of the vector error correction model (VECM) Granger causality approach suggest that energy consumption is influenced by economic growth and financial development, both in the short and the long run, but the population – energy relation holds only in the long run. Ozturk and Acaravci (2013) examined the causal relationship among financial development, trade, economic growth, energy consumption and carbon emissions in Turkey for the period 1960-2007. The bounds test yields evidence of a long-run relationship among per capita carbon emissions, per capita energy consumption, per capita real income, the square of per capita real income, openness and financial development. Results for the existence and direction of Granger causality show that neither carbon emissions per capita nor energy consumption per capita causes real GDP per capita, but employment ratio causes real GDP per capita in the short run. Jalil and Feridun (2011) investigated the impact of financial development, economic growth and energy consumption on environmental pollution in China from 1953 to 2006, using the ARDL bounds testing procedure. The results of the Granger causality tests reveal a negative sign for the coefficient of financial development, suggesting that financial development in China has not taken place at the expense of environmental pollution. On the contrary, financial development has led to a decrease in environmental pollution. Farhani and Ozturk (2015) examined the causal relationship among CO_{2} emissions, real GDP, energy consumption, financial development, trade openness and urbanization in Tunisia over the period of 1971-2012. The results of the Granger causality tests reveal a positive sign for the coefficient of financial development, suggesting that the financial development in Tunisia has taken place at the expense of environmental pollution, with a positive monotonic relationship between real GDP and CO_{2} emissions.

With regard to panel data analyses, Chang (2015) analyzed the nonlinear effects of financial development and income on energy consumption in a sample of 53 countries for the period 1999-2008. It emerges a single-threshold effect on energy consumption when private credit, domestic credit, value of traded stocks and stock market turnover are used as financial development indicators. Zeren and Koc (2014) analyzed seven newly industrialized countries, over the period 1971-2010, reaching mixed results via Hacker–Hatemi causality tests and Hatemi-J asymmetric causality tests. Çoban and Topcu (2013) investigated the financial development – energy consumption nexus in the EU over the period 1990-2011 by using the system-GMM model. No significant relationship is found in the EU-27. The empirical results, however, provide strong evidence of the impact of the financial development on energy consumption in the old members. Shahbaz *et al.* (2013a) examined the linkages among economic growth, energy consumption, financial development, trade openness and CO_{2} emissions over the period 1975Q1-2011Q4 for Indonesia. The results of the VECM Granger causality analysis indicate that economic growth and energy consumption increase emissions, while financial development and trade openness compact it. Shahbaz *et al.* (2013b) investigated the relationship between energy use and economic growth in the case of China over the period 1971-2011. The ARDL bounds testing approach showed that energy use, financial development, capital, exports, imports and international trade have a positive impact on economic growth. The Granger causality analysis revealed a unidirectional causal relationship running from energy use to economic growth. Financial development and energy use Granger-cause each other. Al-mulali and Sab (2012a) explored the impact of energy consumption and CO_{2} emission on the economic and financial development in 19 countries in the period 1980-2008. The empirical findings show that energy consumption enables these countries to achieve high economic and financial development. However, the high development that these countries have achieved in the pate three decades increased the CO_{2} emission. A related study was conducted by Al-mulali and Sab (2012b), where the impact of energy consumption and CO_{2} emission on GDP growth and financial development in 30 sub-Saharan African countries was investigated, from 1980 to 2008. The results indicate that energy consumption had played an important role to increase both economic growth and the financial development in those economies, although with the consequence of high pollution. Shahbaz and Lean (2012) assessed the relationship among energy consumption, financial development, economic growth, industrialization and urbanization in Tunisia from 1971 to 2008. The ARDL bounds testing approach results confirm the existence of a long-run relationship among energy consumption, economic growth, financial development, industrialization and urbanization in Tunisia. The VECM Granger causality analysis shows the presence of a bidirectional causality between financial development and industrialization, which reveals that financial development and industrialization are complementary. Sadorsky’s (2011) study examined the impact of financial development on energy consumption in a sample of nine Central and Eastern European frontier economies. The empirical results show a positive relationship between financial development and energy consumption. Sadorsky (2010), using a GMM estimation technique, explored the impact of financial development on energy consumption in a sample of 22 emerging countries, in the years 1990-2006. The empirical results show a positive and statistically significant relationship between financial development and energy consumption.

With regard to studies on the Italian case, Magazzino (2012) explored the relationship between disaggregate energy production and real aggregate income in Italy by undertaking cointegration analyses and Granger causality tests using annual data from 1883 to 2009. The long-run causality analysis shows a bi-directional flow between each source of energy and GDP in the years 1946-2009, except for the nucleo-thermoelectric energy. Magazzino (2014c) examined the relationship between CO_{2} emissions, energy consumption and economic growth in Italy over the period 1970-2006, showing a lack of cointegration among these three variables. Lee and Chien (2010) studied the dynamic linkages among energy consumption, capital stock and real income in G-7 countries. A unidirectional relationship running from energy consumption to real income was observed, using Granger causality tests based on the Toda and Yamamoto (1995) procedure. Chontanawat *et al.* (2008) tested for Granger causality between energy and GDP using a data set of 30 OECD and 78 non-OECD countries. For Italy, they showed evidence of causality from energy to GDP. Zachariadis (2007) applied bivariate energy use–economic growth causality tests for G-7 countries. Bidirectional Granger causality emerges for most sectors. Lee (2006) explored the causality relationship between energy consumption and GDP in G-11 countries using the Toda and Yamamoto procedure. The results indicate that unidirectional causality running from GDP to energy consumption exists in Italy. Soytas and Sari (2006) analyzed the relationship between energy consumption and income in G-7 countries. For Italy, they found that Granger causality seems to run both ways. Soytas and Sari (2003) investigated the time-series properties of energy consumption and GDP in 10 emerging markets and G-7 countries. Granger causality running from GDP to energy consumption was discovered for Italy.

## 3. Methodology and data

The ARDL bounds testing approach of cointegration is developed by Pesaran and Shin (1999) and Pesaran *et al.* (2001). This approach has several advantages over the traditional cointegration approaches of Engle and Granger (1987), Johansen (1988) and Johansen and Juselius (1990). This takes care of small sample properties and simultaneity biasness in relationship among variables. The main constraint in the application of the conventional cointegration techniques is that they require all the variables included in the model to be non-stationary at levels but should be integrated of the same order. The present ARDL approach to cointegration method surmounts this problem, as it is applicable irrespective of order of integration of regressors, whether *I*(0) or *I*(1) or mixture of both. Apart from that, the ARDL model also has advantages in selecting sufficient numbers of lags to capture the data generating process in a general-to-specific modeling framework. These meritorious features justify the use of the ARDL model to obtain robust estimates.

Following the empirical literature (Farhani and Ozturk, 2015), the standard log-linear functional specification of a long-run relationship among energy consumption, economic growth, financial development and oil prices in Italy may be expressed as:

Basically, the ARDL bounds testing approach of cointegration involves two steps for estimating a long-run relationship. The first step is to investigate the existence of a long-run relationship among all variables in the equation. The ARDL model for equation (1) may follow as:

*ε*and

_{1t}*Δ*are the white noise term and the first difference operator, respectively. The bounds testing procedure is based on the joint

*F*-statistics or Wald statistics that tested the null of no cointegration,

*H*

_{0}: δ

_{r = 0}, against the alternative of

*H*

_{1}: δ

*≠ 0,*

_{r}*r*= 1, 2, …, 4. If the calculated

*F*-statistics lies above the upper level of the band, the null is rejected, indicating cointegration. If the calculated

*F*-statistics is below the upper critical value, we cannot reject the null hypothesis of no cointegration. Finally, if it lies between the bounds, a conclusive inference cannot be made without knowing the order of integration of the underlying regressors. The next step is to test for stability of the long-run coefficients as well as the dynamics of the short-run ones following Pesaran (1997), performing the general error-correction representation of the selected ARDL model of equation (2).

In this study, two causality tests are considered. First, the Granger non-causality test is carried out following the Toda and Yamamoto (1995) long-run causality test. Furthermore, a “standard” Granger causality analysis has been developed. A time series *X _{t}* is said to Granger-cause another time series

*Y*if the prediction error of current

_{t}*y*declines by using past values of

*X*in addition to past values of

*Y*(Granger, 1988).

In our analysis, the log transformations of the variables have been derived. The empirical analysis uses the time-series data of the fossil fuel energy consumption (per cent of total, *EC*), domestic credit to private sector (per cent of GDP, *FD*), real per capita GDP (2011 US dollars per capita, *RGDP*) and oil price (dollars per barrel, *OP*) for Italy in the period 1960-2014. The data are obtained from the World Development Indicator (WDI)[1]. The choice of the starting period was constrained by domestic credit to private sector data availability. Figure 1 shows the dynamic of our series. In the right-side panel, the first-differences series are graphed.

A visual inspection of the series in logarithmic form shows that there is an upward trend for real aggregate income and oil prices.

Some descriptive statistics are summarized in Table II as a preliminary analysis. The mean value of our four variables is positive. Moreover, financial development has positive value of skewness, indicating that the distribution is skewed to the right (Table I).

The energy consumption is statistically positively correlated with real GDP (*r* = 0.42), and negatively with financial development (*r* = −0.61), while the oil prices variable is statistically negatively correlated with real income (*r* = −0.66). In addition, these results are broadly confirmed by cross correlations analysis.

## 4. Empirical results

As can be grasped by the panel in the left-hand side of Figure 1 above, the four analyzed series do not seem to have stationary properties in the levels, contrarily to the relative first differences. Table II contains the results of common unit root and stationarity tests, to determine the order of integration of our variables. Here, we applied four different tests: in general, they indicate that energy consumption, real GDP, financial development and oil prices are all non-stationary at levels, but stationary at first differences.

In fact, for all series, we reject the hypothesis of non-stationarity at the 5 per cent level of significance, both with constant and constant plus trend deterministic specification. We therefore can conclude that all our series are integrated of order one, or *I*(1). The lag-order selection has been chosen according to the Akaike’s information criterion (AIC), the Schwarz’s Bayesian information criterion (SBIC) and the Hannan–Quinn information criterion (HQIC).

The bounds *F*‐test for cointegration yields evidence of a long-run relationship among energy consumption, real income, financial development and oil prices at 1 per cent significance level (Table III).

The long-run elasticity estimate of energy consumption with respect to economic growth is *β _{1}* > 0, at 1 per cent significance level, in line with previous findings in Shahbaz

*et al.*(2013b), but in contrast to Mahalik and Mallick (2014). This means that as per capita real income increases, per capita energy consumption increases as well. Moreover, it implies that energy demand plays a relevant role to enhance economic growth in Italy. Financial development variable has no significant effect on energy consumption in the long run. This result is similar to Ali

*et al.*(2015). Oil prices are negative and significant at 1 per cent, which means that a 1 per cent increase in oil prices could trigger consumption of energy to decrease by 0.32 per cent; therefore, an increase in oil prices leads to an increase in the cost of energy as well as a reduction in energy consumption in Italy in the long run (Table IV).

The coefficients of the estimated error correction term (ECT) are also negative and statistically significant at 1 per cent confidence level. These values indicate that any deviation from the long-run equilibrium between variables is corrected for each period to return to the long-run equilibrium level. In the long run, advancement of the financial sector has an insignificant adverse effect on energy consumption, while economic growth has a significant influence on energy consumption, and oil prices affected energy consumption in Italy (Table V).

The short-run dynamics show that financial development and oil prices do not have significant effects on energy consumptions, whereas economic activity has a significant positive effect on energy consumptions in Italy. This implies that an increase in the energy consumption should not lead to a decline in the financial development and oil prices in the short run.

The ECT is less than one, negative and significant as expected. Banerjee *et al.* (1998) reported that the ECM value confirms the integrity of a long-run relationship. This ratifies the above long-run nexus among the variables, which implies that energy consumption is corrected from the short-run toward reaching long-run equilibrium at 1.66 per cent every year.

Table VI presents the estimated ARDL model that has passed several diagnostic tests, which indicate no evidence of non-normality, serial correlation, heteroskedasticity and autoregressive conditional heteroskedasticity.

In addition, owing to the structural changes in the Italian economy – especially during the eighties (Magazzino, 2015, 2014b) – it is likely that macroeconomic series may be subject to one or multiple structural breaks. For this purpose, the stability of the short-run and long-run coefficients is checked through the cumulative sum (CUSUM) and cumulative sum of squares (CUSUMSQ) tests proposed by Brown *et al.* (1975). Unlike the Chow test, which requires break point(s) to be specified, the CUSUM and CUSUMSQ tests are quite general tests for structural change in that they do not require a prior determination of where the structural break takes place. Figures A1 and A2 in the Appendix present the plot of CUSUM and CUSUMSQ tests statistics that fall inside the critical bounds of 5 per cent significance. This implies that the estimated parameters are stable over the periods.

Granger causality test following the Toda and Yamamoto approach requires the estimation of an augmented VAR(*k + d*) model, where *k* is the optimal lag length and *d* is the order of integration of the series. For the multivariate specification, all tests suggest the inclusion of one lag in a VAR model and thus *k* = 1; hence, the final model to be estimated is VAR(2). To ensure that the VAR model is well specified and does not suffer from any normality or serial correlation problems, additional tests are carried out. Although the results are not reported to save space, diagnostic tests suggest the general absence of problems in the estimated VAR(2) model, with regard to normality and autocorrelation in the residuals, stability condition and lag-exclusion.

The results of Toda and Yamamoto Granger non-causality tests are presented in Table VII. If we want to reject the null hypothesis of “*X* does not Granger cause *Y*” at a 5 per cent level of significance, then the *P*-value should be less than 0.05. The left column in the table represents the dependent variable, while variables listed in the row are the independent variables (source of causation). To provide robust conclusions, both multivariate and bivariate tests are considered. For the multivariate model, empirical findings show that energy consumption is driven by financial development and oil prices. In addition, financial development is caused by energy consumption, too. Thus, energy consumption helped Italy to achieve high economic and financial development. However, countries with similar structure regarding issues of energy can achieve economic and financial development when they present high levels of energy consumption. The bivariate system exhibits a unidirectional causality flow running from real income to energy. This is in line with the findings of Lee (2006) and Soytas and Sari (2003). However, it is interesting to note that the bivariate results roughly confirm the previous of the multivariate system. In fact, again, energy consumption is caused by financial development and oil prices, while oil prices are sensible to financial development. In addition, energy consumption increased the financial development and GDP growth in Italy, with a high pollution consequence. Of course, apart from positive impacts, there are also negative externalities like environmental pollution, which could be the subject for a further research.

Furthermore, the results of the standard Granger causality tests are similar to those obtained via the Toda and Yamamoto approach. They indicate that energy consumption is affected by real GDP, both in the multivariate tests and in the bivariate ones. Again, in the bivariate system, real income is driven by financial development and oil prices.

In addition to the relationship between income and energy consumption, however, it is important for policymakers to take financial development into consideration when formulating energy policy. Financial development is particularly important to business investment because it allows businesses access to additional sources of funding and equity financing. The results from causality tests show that an increase in domestic credit to private sector affects energy consumption. Financial development can lower energy consumption by achieving efficiency in its use. In fact, financial development can provide efficient financial service in foreign banking markets, and improve the access of both foreign and domestic firms to financial goods and services. Developed financial market enhances participation by consumer and business, promotes economic activity and boosts energy use. Moreover, financial development enhances domestic production through investment activities and boost economic growth.

## 5. Concluding remarks and policy implications

This study has extended the research on the causal relationship among energy consumption, real income and financial development using annual data for Italy in the years 1960-2014. The results of unit root tests reveal that all variables are integrated of order one, *I(1)*, as each of them is non-stationary in its level form, and stationary in first differences. The ARDL bounds *F*‐test evidences the existence of a long-run relationship among energy consumption, real income, financial development and oil prices at 1 per cent significance level. The long-run coefficients estimation results show that an increase in real GDP and oil prices has a significant effect on energy consumption, although with an opposite sign. Moreover, the coefficients of estimated ECT are also negative and statistically significant. The results of the analysis reveal a negative sign for the coefficient of financial development, suggesting that financial development in Italy has not taken place at the expense of energy consumption. Finally, causality analyses in general reveal that energy consumption is driven by real income, financial development and oil prices.

Causality analyses indicate that a unidirectional causality running from real income to energy exists, in line with “conservation hypothesis”. This means that continuous economic growth simultaneously generates a continuous rise in energy consumption, and the policy of conserving energy consumption may be implemented with little or no adverse effect on economic growth, such as in a less energy-dependent economy. In this case, energy consumption is fundamentally driven by income, and as such, taking measures to conserve energy may be feasible without compromising economic growth. Beyond this, it is implied that a strategy for sustainable development with a lower level of CO_{2} emissions may, indeed, be appropriate in Italy (Magazzino, 2014c). Financial development Granger causes energy consumption, which reveals that adoption of energy conservation policy would not adversely affect economic growth. Again, the financial sector must fix its focus on the allocation of funds to those firms which adopt environment-friendly technologies and encourage the firms to use more energy-efficient technology for production purpose and hence to save environment from degradation.

The main recommendation suggested by this study, for Italy, is to increase energy productivity by increasing energy efficiency, implementation of energy-saving projects, energy conservation and energy infrastructure outsourcing to achieve its financial development and GDP growth. Therefore, it is important that the country increase its low investments on energy projects to achieve the full energy potential. This leads to reduction in emissions and improvement of the financial development factor. Moreover, as causality analysis pointed out, an energy policy that is focused solely on the relationship between energy demand and income would provide an inaccurate estimate of energy demand because it fails to consider the development of the stock market. Emerging market and developing countries whose stock markets continue to develop should thus anticipate growth in energy demand above and beyond that caused by increasing income alone.

## 6. Suggestions for future research

Further analysis might be conducted to analyze the effect of energy consumption, real income and financial development on carbon emissions in Italy, with the ARDL bounds test approach.

## Figures

Exploratory data analysis

Variable | Mean | Median | SD | Skewness | Kurtosis | Range | IQR | 10-Trim |
---|---|---|---|---|---|---|---|---|

EC | 4.5140 | 4.5237 | 0.0309 | −1.8793 | 6.5356 | 0.1531 | 0.0288 | 4.520 |

RGDP | 10.0188 | 10.1003 | 0.3701 | −0.6723 | 2.2310 | 1.2475 | 0.5951 | 10.060 |

FD | 4.1332 | 4.0801 | 0.1992 | 0.6364 | 2.4412 | 0.6983 | 0.2585 | 4.114 |

OP | 2.7479 | 2.9576 | 1.1033 | −0.2802 | 2.0159 | 3.5293 | 2.0602 | 2.788 |

IQR: Inter-quartile range

**Source:** Author’s calculations on WDI data

Results for unit roots and stationarity tests

Variable | Unit root and stationarity tests | ||||
---|---|---|---|---|---|

Deterministic component | ADF | ERS | PP | KPSS | |

EC | Constant | 0.2641 | −0.3642 | −0.1378 | 0.3477* |

RGDP | Constant | −5.7760*** | −1.1587 | −6.1585*** | 0.9246*** |

FD | Constant | −1.1517 | −1.0798 | −0.7128 | 0.3472* |

OP | Constant | −0.6667 | 0.3815 | −0.6606 | 0.8163*** |

EC | Constant, trend | 1.2620 | 0.3399 | −0.0891 | 0.2171*** |

RGDP | Constant, trend | −0.4072 | −0.1110 | 0.0730 | 0.2473*** |

FD | Constant, trend | −1.4320 | −1.5342 | −1.0398 | 0.2006** |

OP | Constant, trend | −1.7670 | −1.8064 | −1.8837 | 0.1199* |

ΔEC | Constant | −2.4261 | 0.4450 | −4.4619*** | 0.7296** |

ΔRGDP | Constant | −4.7346*** | 1.0069 | −4.7167*** | 0.9610*** |

ΔFD | Constant | −2.8399* | −2.7733*** | −3.4928** | 0.1880 |

ΔOP | Constant | −7.0112*** | −6.9330*** | −7.0111*** | 0.0816 |

ΔEC | Constant, trend | −3.6601** | −2.9090* | −6.2580*** | 0.1309* |

ΔRGDP | Constant, trend | −6.0832*** | −6.4744*** | −7.9507*** | 0.1063 |

ΔFD | Constant, trend | −3.5391** | −2.9500* | −3.4808* | 0.1047 |

ΔOP | Constant, trend | −6.9387 | −7.0560*** | −6.9385 | 0.0832 |

The tests are performed on the log-levels of the variables. ADF; ERS; PP; and KPSS refer, respectively, to the Augmented Dickey–Fuller test; the Elliot, Rothenberg and Stock point optimal test; the Phillips–Perron test; and the Kwiatkowski, Phillips, Schmidt and Shin test. When it is required, the lag length is chosen according to the HQIC;

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

ARDL bound test estimation results

Model for estimation | Lag length | F-statistics |
Significance level | Critical bound | |
---|---|---|---|---|---|

F-statistics |
|||||

I(0) |
I(1) |
||||

F_{EC}^{RGDP,FD,OP} |
2 | 8.1949*** | 1 | 3.65 | 4.66 |

5 | 2.79 | 3.67 | |||

10 | 2.37 | 3.20 |

Asymptotic critical value bounds are obtained from table *F*-statistic in Pesaran *et al.* (2001);

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

Long-run coefficients estimation

Regressors | Coefficient | Standard error |
---|---|---|

RGDP | 0.5074 | 0.1776*** |

FD | −0.6340 | 1.1532 |

OP | −0.3183 | 0.1123*** |

Constant | 2.6088 | 0.7637*** |

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

Estimated short-run coefficients from ECM

Regressors | Coefficient | Standard error |
---|---|---|

ΔEC_{t-1} |
0.2530 | 0.1220** |

ΔRGDP | 0.1638 | 0.0292*** |

ΔFD | −0.0108 | 0.0193 |

ΔOP | −0.0284 | 0.0337 |

ECM_{t-1} |
−0.0166 | 0.0026*** |

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

ARDL Diagnostic tests results

Test statistics | LM version |
F version |
---|---|---|

1: Serial correlation | χ^{2} = 0.6438 (0.7248) |
0.2663 (0.7676) |

2: Functional form | 11.1963 (0.0018)*** | |

3: Normality | 3.5866 (0.1664) | |

4: Heteroskedasticity | 0.9105 (0.4968) | 5.6398 (0.4647) |

5: ARCH | 0.0785 (0.7806) | 0.0818 (0.7749) |

1 = Lagrange-Multiplier test of residual serial correlation; 2 = Ramsey’s RESET test using squared of the fitted values; 3 = Jarque–Bera test for normality; 4 = Breusch–Pagan–Godfrey heteroskedasticity test; 5 = AutoRegressive Conditional Heteroskedasticity test;

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

Results of causality tests

Multivariate | Bivariate | ||||||||
---|---|---|---|---|---|---|---|---|---|

Independent variables | Independent variables | ||||||||

Dep. var. | EC | RGDP | FD | OP | Dep. var. | EC | RGDP | FD | OP |

Granger tests |
|||||||||

EC | – | 7.402*** (0.007) | 0.058 (0.810) | 10.302*** (0.001) | EC | – | 4.649** (0.031) | 7.548** (0.023) | 7.581*** (0.006) |

RGDP | 2.838* (0.092) | – | 0.091 (0.763) | 8.631*** (0.003) | RGDP | 4.035** (0.045) | – | 11.508*** (0.003) | 9.975*** (0.002) |

FD | 0.046 (0.830) | 0.895 (0.344) | – | 1.478 (0.224) | FD | 1.379 (0.502) | 1.561 (0.458) | – | 0.945 (0.331) |

OP | 0.246 (0.620) | 1.498 (0.221) | 0.107 (0.743) | – | OP | 0.040 (0.842) | 2.242 (0.326) | 0.006 (0.940) | – |

Toda and Yamamoto tests |
|||||||||

EC | – | 1.399 (0.497) | 6.363** (0.042) | 8.862** (0.012) | EC | – | 12.312*** (0.002) | 10.767** (0.013) | 28.467*** (0.000) |

RGDP | 1.007 (0.604) | – | 2.187 (0.335) | 4.095 (0.129) | RGDP | 3.500 (0.174) | – | 14.892*** (0.002) | 10.457*** (0.005) |

FD | 9.535*** (0.009) | 3.362 (0.186) | – | 8.167** (0.017) | FD | 3.333 (0.343) | 2.257 (0.521) | – | 2.077 (0.354) |

OP | 0.449 (0.799) | 4.926* (0.085) | 10.735*** (0.005) | – | OP | 0.706 (0.703) | 3.375 (0.185) | 5.290* (0.071) | – |

Wald tests (*p*-values in parentheses);

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

Correlation matrix

EC | RGDP | FD | OP | |
---|---|---|---|---|

EC | 1.000 | 0.4198** | −0.6135*** | −0.1882 |

RGDP | 0.4198** | 1.0000 | −0.3908** | −0.6601*** |

FD | −0.6135*** | −0.3908** | 1.0000 | 0.2522 |

OP | −0.1882 | −0.6601*** | 0.2522 | 1.0000 |

Sidak’s correction applied;

*p* < 0.01;

*p* < 0.05;

*p* < 0.10

## Note

See, for more details: http://data.worldbank.org/data-catalog/world-development-indicators

## References

Abalaba, B.P. and Dada, M.A. (2013), “Energy consumption and economic growth nexus: new empirical evidence from Nigeria”, International Journal of Energy Economics and Policy, Vol. 3 No. 4, pp. 412-423.

Ali, H.S., Yusop, Z.B. and Hook, L.S. (2015), “Financial development and energy consumption nexus in Nigeria: an application of autoregressive distributed lag bound testing approach”, International Journal of Energy Economics and Policy, Vol. 5 No. 3, pp. 816-821.

Al-mulali, U. and Sab, C.N.B.C. (2012a), “The impact of energy consumption and CO2 emission on the economic and financial development in 19 selected countries”, Renewable and Sustainable Energy Reviews, Vol. 16, pp. 4365-4369.

Al-mulali, U. and Sab, C.N.B.C. (2012b), “The impact of energy consumption and CO2 emission on the economic growth and financial development in the sub-Saharan African countries”, Energy, Vol. 39, pp. 180-186.

Banerjee, A., Dolado, J. and Mestre, R. (1998), “Error-correction mechanism tests for cointegration in a single-equation framework”, Journal of Time Series Analysis, Vol. 19 No. 3, pp. 267-283.

Brown, R.L., Durbin, J. and Evans, J.M. (1975), “Techniques for testing the consistency of regression relations over time”, Journal of Royal Statistical Society, Vol. 37, pp. 149-192.

Chang, S.C. (2015), “Effects of financial developments and income on energy consumption”, International Review of Economics and Finance, Vol. 35, pp. 28-44.

Chontanawat, J., Hunt, L.C. and Pierse, R. (2008), “Does energy consumption cause economic growth? Evidence from a systematic study”, Journal of Policy Modeling, Vol. 30 No. 2, pp. 209-220.

Çoban, S. and Topcu, M. (2013), “The nexus between financial development and energy consumption in the EU: a dynamic panel data analysis”, Energy Economics, Vol. 39, pp. 81-88.

Engle, R.F. and Granger, C.W.J. (1987), “Co-integration and error correction: representation, estimation, and testing”, Econometrica, Vol. 55 No. 2, pp. 251-276.

Farhani, S. and Ozturk, I. (2015), “Causal relationship between CO_{2} emissions, real GDP, energy consumption, financial development, trade openness, and urbanization in Tunisia”, Environmental Science and Pollution Research International, Vol. 22 No. 20, pp. 15663-15676.

Granger, C.W.J. (1988), “Some recent developments in a concept of causality”, Journal of Econometrics, Vol. 39 No. 1-2, pp. 199-211.

Islam, F., Shahbaz, M., Ahmed, A.U. and Alam, M.M. (2013), “Financial development and energy consumption nexus in Malaysia: a multivariate time series analysis”, Economic Modelling, Vol. 30, pp. 435-441.

Jalil, A. and Feridun, M. (2011), “The impact of growth, energy and financial development on the environment in china: a cointegration analysis”, Energy Economics, Vol. 33 No. 2, pp. 284-291.

Johansen, S. (1988), “Statistical analysis of cointegration vectors”, Journal of Economics Dynamics and Control, Vol. 12 Nos 2/3, pp. 231-254.

Johansen, S. and Juselius, K. (1990), “Maximum likelihood estimation and inference on cointegration - with applications to the demand for money”, Oxford Bulletin of Economics and Statistics, Vol. 52 No. 2, pp. 169-210.

Lee, C.C. (2006), “The causality relationship between energy consumption and GDP in G-11 countries revisited”, Energy Policy, Vol. 34 No. 9, pp. 1086-1093.

Lee, C.C. and Chien, M.S. (2010), “Dynamic modelling of energy consumption, capital stock, and real income in G-7 countries”, Energy Economics, Vol. 32 No. 3, pp. 564-581.

Magazzino, C. (2012), “On the relationship between disaggregated energy production and GDP in Italy”, Energy & Environment, Vol. 23 No. 8, pp. 1191-1207.

Magazzino, C. (2014a), “Electricity demand, GDP and employment: evidence from Italy”, Frontiers in Energy, Vol. 8 No. 1, pp. 31-40.

Magazzino, C. (2014b), “The relationship between CO2 emissions, energy consumption and economic growth in Italy”, International Journal of Sustainable Energy, Vol. 35 No. 9, pp. 844-857.

Magazzino, C. (2014c), “Electricity demand, GDP and employment: evidence from Italy”, Frontiers in Energy, Vol. 8 No. 1, pp. 31-40.

Magazzino, C. (2015), “Energy consumption and GDP in Italy”, Environment, Development and Sustainability, Vol. 17 No. 1, pp. 137-153.

Mahalik, M.K. and Mallick, M.H. (2014), “Energy consumption, economic growth and financial development: exploring the empirical linkages for India”, The Journal of Developing Areas, Vol. 48 No. 4, pp. 139-159.

Ozturk, I. and Acaravci, A. (2013), “The long-run and causal analysis of energy, growth, openness and financial development on carbon emissions in turkey”, Energy Economics, Vol. 36, pp. 262-267.

Pesaran, H.M. (1997), “The role of economic theory in modelling the long run”, The Economic Journal, Vol. 107 No. 440, pp. 178-191.

Pesaran, H.M. and Shin, Y. (1999), “Autoregressive distributed lag modelling approach to cointegration analysis”, in Storm S. (Ed.), Econometrics and Economic Theory in the 20th Century: The Ragnar Frisch Centennial Symposium, Cambridge University Press, Cambridge, MA.

Pesaran, H.M., Shin, Y. and Smith, R.J. (2001), “Bounds testing approaches to the analysis of level relationships”, Journal of Applied Econometrics, Vol. 16 No. 3, pp. 289-326.

Sadorsky, P. (2010), “The impact of financial development on energy consumption in emerging economies”, Energy Policy, Vol. 38 No. 5, pp. 2528-2535.

Sadorsky, P. (2011), “Financial development and energy consumption in Central and Eastern European frontier economies”, Energy Policy, Vol. 39 No. 2, pp. 999-1006.

Salman, D.M. and Atya, E.M. (2014), “What is the role of financial development and energy consumption on economic growth? New evidence from North African countries”, International Journal of Finance & Banking Studies, Vol. 3 No. 1, pp. 137-149.

Shahbaz, M. and Lean, H.H. (2012), “Does financial development increase energy consumption? The role of industrialization and urbanization in Tunisia”, Energy Policy, Vol. 40, pp. 473-479.

Shahbaz, M., Khan, S. and Tahir, M.I. (2013b), “The dynamic links between energy consumption, economic growth, financial development and trade in China: fresh evidence from multivariate framework analysis”, Energy Economics, Vol. 40, pp. 8-21.

Shahbaz, M., Hye, Q.M.A., Tiwari, A.K. and Leitão, N.C. (2013a), “Economic growth, energy consumption, financial development, international trade and CO_{2} emissions in Indonesia”, Renewable and Sustainable Energy Reviews, Vol. 25, pp. 109-121.

Soytas, U. and Sari, R. (2003), “Energy consumption and GDP: causality relationship in G-7 countries and emerging markets”, Energy Economics, Vol. 25 No. 1, pp. 33-37.

Soytas, U. and Sari, R. (2006), “Energy consumption and income in G-7 countries”, Journal of Policy Modeling, Vol. 28 No. 7, pp. 739-750.

Toda, H.Y. and Yamamoto, T. (1995), “Statistical inference in vector autoregressions with possibly integrated processes”, Journal of Econometrics, Vol. 66 No. 1-2, pp. 225-250.

Zachariadis, T. (2007), “Exploring the relationship between energy use and economic growth with bivariate models: new evidence from G-7 countries”, Energy Economics, Vol. 29 No. 6, pp. 1233-1253.

Zeren, F. and Koc, M. (2014), “The nexus between energy consumption and financial development with asymmetric causality test: new evidence from newly industrialized countries”, International Journal of Energy Economics and Policy, Vol. 4 No. 1, pp. 83-91.

## Acknowledgements

The author thanks the participants to the conference “Energy Quest 2016”, International Conference on Energy Production and Management in the 21st Century, 6-8 September 2016, Ancona, Italy, for their useful comments and advices. The author is also indebted to the anonymous referees and the editor for their valuable suggestions. However, the usual caveats apply.