The various international conferences and agreements on the environment have stimulated economists and policymakers to assess the impact of corruption and renewable energies on climate change. Besides, the monopolistic competition models in international trade have demonstrated the importance of innovation as a factor that facilitates sustainable development, reducing greenhouse effects (e.g., Leitão 2021
; Leitão and Balogh 2020
In recent years, economists have assessed the relationship between corruption and economic growth, as well as corruption and carbon dioxide emissions. The empirical studies demonstrate that corruption affects economic growth, air quality, and climate change.
This causality between corruption, economic growth, and climate change is complex, as it involves political, economic, social, and environmental issues. In this context, this paper examines the linkage between renewable energy, corruption perception, and carbon dioxide emissions. The impact of trade openness and economic growth on air quality is also considered by panel data for Portugal, Spain, Italy, Ireland, and Greece, considering the period 1995–2015.
This article’s motivation is related to the fact that we have considered a group of countries in Europe with identical characteristics but with some particularities between them. In this context, it is possible to mention that the crisis that started in the USA in 2008–2009 had repercussions in Portugal, Spain, Italy, Ireland, and Greece. Moreover, it is essential to mention that Greece was the first country to resort to international aid in 2010, followed by Ireland in that same year. However, Ireland managed to cease relying on international support in 2013. In 2011, Portugal asked for international aid and, in 2014, exited the rescue program. Spain and Italy did not request a recovery plan; however, they implemented economic austerity measures.
After the most challenging years with measures of substantial restrictions in economic terms, it will be interesting to observe and understand how climate change reacts in the face of corruption and the relationship between innovation (via international trade and renewable energies) and carbon dioxide emissions.
This paper’s main objective is to assess the impact of corruption, trade openness, and renewable energies on climate change and air quality. iFrst, we present a literature review based on the link between renewable energies, corruption, international trade, economic growth, and carbon dioxide emissions. This first objective seeks to support the methodology to be applied, as well as the empirical study. In a second step, the econometric models’ applications evaluate the short and long-term impacts and the causality relationships between the variables used.
This article is formulated as follows. The next section presents the literature review. Section 3
contains the methodology, material, data and methods considered in this research in which we present the panel data, the sources from which the data were collected, the formulated hypotheses, and the respective econometric models. The empirical results and discussion are shown in Section 4
and the conclusions and policy implications are presented in Section 5
3. Materials and Methods
The effects of corruption, renewable energy, economic growth, and trade openness on carbon dioxide emissions are considered in this research. The methodology used in this study is panel data for the period 1995–2015, and we select the following European countries: Portugal, Spain, Italy, Ireland, and Greece. This group of countries have similar economic structures, namely in the distribution of per capita income, except for Ireland, which stands out from the other countries (Figure 1
Thus, concerning the emission of carbon dioxide, it is observed that of the economies in question, Italy is the one that has the highest emissions of carbon dioxide, followed by Spain, Greece, and then Portugal and Ireland (Figure 2
Regarding corruption, it is possible to infer that from 2010 onwards, there was an abrupt decline in all countries under analysis except for the Portuguese economy, which increased slightly (Figure 3
Concerning the use of renewable energies, it is observed that Portugal, of the economies under study, stands out in first place in terms of ranking. Ireland is one of the countries with less use of renewable energies; however, since 2005, there is an increasing trend towards the use of this type of energy (Figure 4
The dependent variable evaluates the emissions of carbon dioxide (CO2) expressed in Kilotons. The independent variables are income per capita, corruption perception index, renewable energy, and exports of goods and services as a percentage of GDP. It should be mentioned that the data for the renewable energy variable is only available until 2015 from the World Bank (2021) data. On the other hand, the corruption perception index collected from the International Transparency database has only been available since 1995.
We started with unit root tests to assess whether the studies’ variables show stationarity or are integrated into the methodology’s first differences.
Before applying the TSLS–Panel Two-Stage Least Squares estimator, it is necessary to apply the Hausman test to identify heterogeneity. The Hausman test compares random effects (RE) versus fixed effects (FE), indicating which estimators to use. After regression of the TSLS—Panel Two-Stage Least Squares model, it is necessary to apply the Breusch–Pagan postestimation test and the residual cross-section dependence test.
Therefore, the battery of panel unit root test such as Levin et al.
), ADF–Fisher Chi-square, Phillips–Perron, and Im et al.
) recommended by Maddala and Wu
), and Choi
) was applied to test and verify if the variables are stationary at the level or integrated into the first differences.
The next step pursued is to test the long-run panel data cointegration between the variables used in this study, considering Pedroni
) and the Kao
). As Leitão and Balsalobre-Lorente
) note, the Pedroni cointegration test is considered in two dimensions: with-dimension (panel cointegration) and between-dimension (mean panel). The panel v—statistics, panel rho—statistics, panel PP—statistics, and ADF—statistics represent the with-dimension. The classification between-dimension involves group rho—statistics, group PP—statistics, and group ADF—statistics.
Subsequently, the estimators of the panel fully modified least-squares (FMOLS) and panel dynamic least squares (DOLS) (e.g., Phillips and Hansen 1990
; Stock and Watson 1993
) are used to evaluate the long-run correlations between the variables used in this study.
Hypotheses and Model Specification
Considering the literature review, we formulate the following hypotheses:
Hypothesis 1 (H1).
Economic growth and economic activity are positively associated with carbon dioxide emissions.
The complexity of studying corruption and the different perspectives of this variable in the environment and climate change allows us to formulate two hypotheses as an alternative.
Hypothesis 2 (H2).
(a) Corruption causes adjustment costs in the environment; (b) the control of corruption reduces climate change.
This hypothesis was considered based on previous studies. As we analysed in the literature review, there are two different perspectives on the impact of corruption on carbon dioxide emissions: one of which argues that corruption’s control slows down climate change. However, as mentioned by the international transparency organization (https://www.transparency.org/en/what-is-corruption
, accessed on 18 April 2021), corruption has impacts or costs in social, political, economic, and environmental terms. In this context, corruption accelerates carbon dioxide emissions, and environmental damage (e.g., Dincer and Fredriksson 2018
; Ridzuan et al. 2019
; Akhbari and Nejati 2019
; Zhang and Chiu 2020
According to the international transparency organization, the corruption index (CPI) is a perception index ranging from 0 to 10. The highest corruption corresponds to level 0, and the value 10 indicates the lowest level of corruption.
Hypothesis 3 (H3).
The use of renewable energy promotes sustainable development and air quality.
In recent years, empirical studies have argued that the replacement of nonrenewable energy with cleaner energy promotes the reduction of environmental costs, with a proliferation of studies that assess the causality between renewable energy and air quality. The studies by Maneejuk et al.
), Leitão and Balsalobre-Lorente
), Koengkan and Fuinhas
), and Balsalobre-Lorente et al.
) observed that there is a negative relationship with statistical significance between renewable energies and carbon dioxide emissions.
Hypothesis 4 (H4).
International trade aims to decrease climate change.
Regarding the arguments of the monopolistic competition models applied to international trade and the link with the environment, the empirical studies of Wang et al.
), Ike et al.
), Shahzad et al.
), and Leitão
) demonstrated that international trade is negatively correlated with carbon dioxide emissions. The authors state that trade is associated with sustainable practices which aim to reduce CO2
emissions. As Leitão and Balogh
) and Leitão
) argue, international trade allows us to explain innovation and product differentiation. In this context, innovation factors make it possible to reduce climate change and improve air quality.
In this empirical study, we use as a dependent variable the logarithm of carbon dioxide emission (CO2) in kilotons to evaluate climate change, where i signifies the number of countries, and t signifies the time. Moreover, δt describes the common determinist trend, ηi are the specific effects, and εit represents the random disturbance.
The independent variables used in this study are the following:
LGDP—represents the logarithm of income per capita expressed in USD.
LCPI—signifies the logarithm of corruption perceptions index.
LREW—indicates the logarithm of a percentage of renewable energy use in total final energy consumption.
LTO—represents the logarithm of exports of goods and services in the percentage of GDP.
presents the data, the expected signs considering the literature review and the sources used in this research.
4. Results and Discussion
This section exhibits the econometric results to test the relationship between corruption, economic growth, renewable energy, international trade, and carbon dioxide emissions (climate change).
reveals the general statistics for all variables utilized in this empirical study. The variables of carbon dioxide emissions (LCO2
) and income per capita (LGDP) represent the maximum’s higher values. Besides, the variables of carbon dioxide emissions (LCO2
), income per capita (LGDP), and trade openness (LTO) present a positive skew. However, it is possible to observe that corruption index (LCPI) and renewable energy (LREW) exhibit a negative skew. The variable of carbon dioxide emissions (LCO2
) presents the low kurtosis value, but the variable assumes the high value of standard deviation (std. dev).
shows the unit root test results considering the methodology of Levin, Lin, and Chu, ADF–Fisher Chi-square, Phillips–Perron, and Im–Pesaran–Shin. Regarding the results, it is possible to infer that the variables of carbon dioxide emissions (LCO2
), economic growth (LGDP), corruption index (LCPI), renewable energy (LREW), and trade openness (LTO) are stationary at the first difference I (1).
illustrates the results of causality among the variables using the criterion of Dumitrescu–Hurlin pairwise causality.
We observe a unidirectional causality between economic growth (LGDP) and carbon dioxide emissions (LCO2
). The linkages between carbon dioxide emissions (LCO2
) and corruption (LCPI) and economic growth and corruption (LCPI) present a unidirectional causality. These results are according to the previous studies of Dincer and Fredriksson
), Ridzuan et al.
), and Lee et al.
). Besides, it is important to highlight that these results suggest that signs of the shadow economy occur and can be justified by the investigations of Bilan et al.
) and Nemec et al.
We also observe a unidirectional causality between renewable energy (LREW) and corruption (LCPI).
Furthermore, we see a bidirectional causality between renewable energy (LREW) and carbon dioxide emissions (LCO2). Our results also show a bidirectional causality between trade openness (LTO) and carbon dioxide emissions (LCO2).
demonstrates the results of the panel cointegration test and Kao residual cointegration. According to the results, we can conclude that the variables are cointegrated in the long term. In this context, the results show a cointegration relationship between the variables economic growth, corruption, renewable energy, international trade, and carbon dioxide emissions.
Before proceeding with the model specification, it is necessary to perform the multicollinearity test and Hausman test to identify the endogeneity. First, it is essential to mention that multicollinearity is frequent through the OLS estimator or multiple linear regression model. As a rule, the independent variables have tolerance problems when VIF (variance inflation factor) is higher than five. According to the results in Table 6
, we observe that all variables used in this research do not present a multicollinearity problem, VIF < 5.
When comparing random effects (RE) versus fixed effects (FE) using the Hausman test, Table 7
indicates that fixed effects are more appropriate.
presents the econometric results with panel two-stage least squares (TSLS). We observe that economic growth (LGDP), renewable energy (LREW), and trade openness (LTO) are statistically significant at a 1% level. Moreover, the coefficient of corruption is statistically significant at a 10% level. In the Durbin–Watson statistics, we can refer that the estimates do not present problems of serial correlation. It is still possible to observe the Breusch–Pagan test in the present table, which admits homoscedasticity as a null hypothesis. The postestimation test presented confirms the presence of heteroscedasticity.
The variables of income per capita (LGDP) and corruption index (LCPI) are according to H1 and H2. The empirical result confirms that economic growth can generate environmental problems, which the Kuznets environmental curve’s assumptions can explain. Besides, the result obtained shows that corruption causes ecological and economic costs. The previous studies of Zhang and Chiu
), Arminen and Menegaki
), and Lee et al.
) give support to our results. Nevertheless, the coefficients of renewable energy (LREW) and trade openness (LTO) are negatively correlated with CO2
emissions, showing that clean energies and international trade promote sustainable development, and subsequently the improvements of environmental factors (e.g., Ahmed and Shimada 2019
; Maneejuk et al. 2020
; Razzaq et al. 2021
; and Bouyghrissi et al. 2021
reports the estimates using the panel fully modified least squares (FMOLS) and panel dynamic ordinary least squares (DOLS).
According to our econometric results, it is possible to observe that the variables’ economic growth (LGDP) and trade openness (LTO) are statistically significant at the 1% level. These coefficients present a positive and negative impact on carbon dioxide emissions, respectively. Additionally, the renewable energy (LREW) variable is statistically significant at the 1% and 10% level with FMOLS (panel fully modified least squares) and DOLS (panel dynamic ordinary least squares).
The variable of corruption index (LCPI) presents a positive impact on CO2 emissions for both estimators with 10% and 5% levels, representing an increase in carbon dioxide emissions by 0.108% and 0.181%, respectively.
This result suggests that corruption can motivate climate change and environmental damage. The previous studies of Zhang and Chiu
), Lee et al.
), and Ridzuan et al.
) also found a positive impact of corruption on carbon dioxide emissions.
As the literature review reflects, trade openness (LTO) is negatively correlated with carbon dioxide emissions; our result found this association. The result can be explained by the trade intensity, as assessed by Leitão
), and this type of trade is associated with innovation and product differentiation. The empirical studies of Balsalobre-Lorente et al.
), Leitão and Balsalobre-Lorente
), Ike et al.
) also demonstrate that international trade present a negative effect on CO2
emissions. The authors concluded that international trade diminished climate change.
This study investigates the impacts of renewable energy, corruption perception, economic growth, and international trade on CO2 emissions considering panel data for 1995–2015. We selected five European economies: Portugal, Spain, Italy, Ireland, and Greece. We chose to focus on these countries because we intended to evaluate countries with relatively similar characteristics, as explained in this investigation’s introductiont. Besides, previous studies on this group of countries have not assessed the impact of corruption on climate issues.
The hypotheses were tested by panel unit root test (Levin et al. 2002
; Choi 2001
; Im et al. 2003
). The results showed that the variables used in this research are stationary into the differences, I (1). Furthermore, the Pedroni and Kao panel cointegration demonstrated that the variables economic growth, corruption, renewable energy, international trade, and carbon dioxide emissions are cointegrated in the long run.
Regarding the Dumitrescu and Hurlin
) pairwise causality test, we conclude that, in general, the variables used in this empirical study present a unidirectional causality between them. The results still suggest a link between economic growth and corruption, showing signs of a shadow economy. However, we observe a bidirectional causality between renewable energy and carbon dioxide emissions. The same is verified by the relationship between international trade and carbon dioxide emissions.
The econometric results using TSLS (Panel Two-Stage Least Squares), FMOLS (panel Fully Modified Least Squares), and DOLS (panel Dynamic Least Squares) are similar between them.
The studies of Dincer and Fredriksson
), Ridzuan et al.
), Akhbari and Nejati
), Zhang and Chiu
) confirmed that corruption can accentuate pollution emissions, reflecting a policy of less transparency in terms of industrial and government policy. Our results in accordance with this perspective. Thus, a positive effect of corruption on CO2
emissions contains several economic and fiscal policy implications for government policies.
Subsequently, our econometric results also show that renewable energy and trade openness negatively correlated with carbon dioxide emissions. This result is according to the EU agenda, the Kyoto Protocol to the United Nations Framework Convention on Climate Change
), and the Paris Agreement—UNFCCC
), i.e., the concept of sustainable development (e.g., Balsalobre-Lorente et al. 2021
; Leitão 2021
Therefore, based on the empirical results, we can present some suggestions and implications for policymakers. First, the study shows that corruption can stimulate carbon dioxide emissions and, consequently, impacts environmental damage. Thus, governments and their finance ministries should implement more regulatory, supervisory, and corruption control measures, as this unfair practice distorts competition laws and policy. Second, the governments of the economies under analysis should continue to develop acceptable practices regarding decarbonization, promoting them as they have done with the introduction of renewable energies and more sustainable trade that considers innovation and differentiation.
Regarding the extent of this investigation and clues for future work, we think it would be interesting to compare these results with the BRICS (Brazil, Russia, India, China, and South Africa) economies since they are emerging economies with great potential for international trade. In this context, it will be essential to note that India has invested a lot in cleaner energy production in recent years. On the other hand, China has reversed its international trade patterns due to the current pandemic crisis and its response (e.g., Leitão and Balsalobre-Lorente 2020