A Bibliometric Study on the Nexus of Economic Growth and Renewable Energy in Brazil

,


Introduction
The evolution of population well-being and projections of reduced poverty are intimately linked to economic growth; as a result, their determinants are critical to public administrators and private economic agents (Barro & Sala-i-Martin, 1995;Gregorio & Lee, 1999). The study of the relationship between economic growth and energy consumption became vital because it is a fundamental component of the productive sectors and the consumption pattern of families. electricity matrix between 2005 and 2017 was relatively stable, given that the increase in the installed capacity of renewables was also accompanied by the increase significant increase in the installed capacity of non-renewable thermal sources. As of 2017, one consideration of renewables passed the mark of 82%, particularly by the increase in wind and solar sources, driven by the greater competitiveness of these sources in the electricity auctions.
It should be noted that policies encouraging the adoption of renewable technologies have increased the contribution of other energy sources in the nation, such as those solar and wind power. In 2015, for example, the Development Program for Distributed Electricity Generation (ProGD) was created. The program's objective was to stimulate energy generation from solar panels inside the consumer units, which can be shared with the energy distribution system.
In 2021, the Brazilian Government launched yet another program to encourage solar energy. -Pró-sol‖ was the name given by the federal government to the program that forms part of its policy to encourage the use of solar energy. The initiative, which goes beyond renewing current incentives for installing solar panels, gained momentum after the World Economic Forum in Davos in 2020. One of the main incentives of the program is the reduction of taxes.
Also, in 2021, the Environment Commission (CMA) approved, in a final decision, a project establishing the Incentive Program for the Development of Wind Energy and Solar Photovoltaics (Pides). The text provides that Pide funding will derive from appropriations from the Federal Budget. The Union will grant an economic subsidy to the National Bank for Economic and Social Development (BNDES) to equalize interest rates for financing the program. Financing contracts from the Federal Government to BNDES with a view to Pides will have a financial cost equivalent to the Long-Term Rate (TLP). In January, the Brazilian Senate approved the project (PL 3.386/2021) that creates the Incentive Program for the Development of Wind Energy and Solar Photovoltaics (Pides). The proposal provides Union with up to R$ 500 million annually for clean energy development projects.
Brazil is the world's fifth-largest and fifth-most populous country. As the tenth-largest economy in the world, it is among the global giants of mining, agriculture, and manufacturing (European Parliament, 2022). The energy industry plays a significant role in Brazil's economy. The country is among the top 10 largest oil producers in the world. In addition, it has significant renewable generation capacity energy (84%, while the world average is 38%), mainly from hydroelectric energy, but also from solar energy and wind energy. The International Energy Agency informs that the total demand for primary energy has not doubled since 1990, driven by strong growth in electricity consumption and demand for transportation fuels due to robust economic growth and an expanding middle class. However, the US Energy Information Administration notes that there are issues that challenge power generation and transmission, not least the mix on the reliability of the national system electricity generation system (Brazil depends up to 66% on hydroelectric power, which was achieved in 2021 by severe drought); the great distances between the centers of generation and demand; and ongoing droughts and deforestation.
Brazil's Gross Domestic Product (GDP) grew by 2.9% in 2022, in comparison with 2021, as can be seen in graph 1. In 2021, it showed recovery, compared to 2020, when the COVID-19 crisis spread worldwide. As can be seen, in 2015 and 2016, the country experienced a downturn for two consecutive years in its economy. This sequence of two consecutive years of decline was only verified in Brazil in 1930 and 1931, when the retreats were 2.1% and 3.3%, respectively. This study aims to conduct a bibliometric analysis of the literature on economic growth and energy consumption under the umbrella of renewable energy sources for Brazil to guide researchers and public policymakers effectively. This study is divided into five sections in addition to this introduction. The critical elements related to the energy-growth-sustainability nexus will be presented in section 2. The clusterization and bibliometric methods are presented in Section 3. In section 4, the results are analyzed. Following that, the final considerations are conducted in Section 5.

Energy-Growth-Sustainability Nexus (ECS)
As previously mentioned, job and wage creation and second-plane consumption are one of the connections between energy and economic growth, according to International Energy Agency (IEA, 2021a). Between 1990 and 2018, the final electricity consumption related to the residential and industrial sectors increased by 138.5% and 107.2%, respectively. This movement will lead to the comparison of data of global electricity consumption: According to the Renewable Energy and Jobs report from International Renewable Energy Agency, nearly 1.27 million new jobs were created in the renewable energy sector in Brazil in 2021 (IRENA, 2022), an increase of about 25% from the previous year. Approximately 67% of this total relates to the biofuel industry, making it the biggest employer in Brazil. Expansions in the wind energy sector led to record increases in the installed productive capacity of this source, reaching the accumulated mark of 21.2 GW. The agency estimates that around 63 800 workers are employed in the construction and operations & maintenance (O&M) of such projects and are located mainly in the country's northeast region, where the sector demands industries that manufacture equipment. The photovoltaic energy sector also increased installed generation capacity by 5.5 GW in 2021, accumulating 14 GW in the country's energy matrix. Most of the installed capacity is located in the south and southeast. According to IRENA, one-third of the total photovoltaic energy installed in the country comes from distributed systems (up to 5MW).
Between 1996 and 2021, the average growth rate of the Brazilian economy was 0.55%, and since 1990, the country's overall demand for primary energy has doubled. According to a report from the International Energy Agency, this increase was caused mainly by the transportation sector's demand for electricity and fuel (IEA, 2022). Access to electric energy became almost ubiquitous thanks to private investments and primarily to government initiatives like the -Luz para Todos‖ program (Ministé rio de Minas e Energia, 2019, 2021) and Social Electricity Tariff (ANEEL, 2020).
Hydroelectric generation is responsible for 60% of the energy generated in the country (EPE, 2021). It imposes uncertainties about the ability of energy demand to be met in the face of the resurgence of climate change. Despite this, hydroelectric generation is projected to increase by 36% by 2024 (Hunt et al., 2018). Studying the effects of investments in renewable energy on economic growth has become a crucial topic for Brazil, given the sector's enormous potential for growth.
Currently, 66 projects are in the licensing process in the country. Together they add up to 169 GWthe states of Rio Grande do Sul and Ceará lead the ranking in estimated power. In addition, Brazil is a potential player in the green hydrogen market, which, because it can be transported, has proven to be an alternative for European countries with scarce energy resources (EPBR, 2021). In addition, with the worsening of the effects and projections related to climate change and the global coordination around the reduction of CO 2 emissions exclusive to the burning of fossil fuels, the replacement of hydrocarbons by renewable energy sources such as wind (onshore and offshore wind ), water (hydroelectric, tidal energy) and biofuels (blue hydrogen, green hydrogen, ethanol, biodiesel) can not only promote sustainable economic growth but also represent an opportunity to lead the international carbon allowance system and increase Brazil's competition with more present countries (Fareed & Pata, 2022).

Bibliometrics and Clustering Method
The bibliometric method is an interdisciplinary scientific approach to quantifying academic output from individuals and institutions concerning a specific topic. VOSviewer, CiteSpace, Histcity, and Bibexcel are the programs used for statistical estimations and visual analysis tools (Hu et al., 2022).
In the first step, we elaborated the keywords to extract the articles relevant to our study. In the second stage, we quantitatively analyzed the collected sample and the proportion of types of publications included. In the third step, we used the VOSviewer software to analyze co-citation, co-authorship, regions/countries, and ijef.ccsenet.org International Journal of Economics and Finance Vol. 15, No. 4; co-occurrence of keywords. Then, we will perform the analysis of benchmarks and cluster analysis. Cluster analysis, or clustering, is a Multivariate Statistics procedure that aims to partition elements into two or more clusters considering their similarity according to pre-established criteria (Santos et al., 2020). The dissimilarity between objects is measured by a distance matrix whose components resemble the distance between two points.
Clustering methods can be described by a matrix containing a measure of dissimilarity or proximity between each pair of objects. Each p ij entry in the matrix is a numerical value demonstrating how close objects i and j are. The presented dissimilarity coefficients are functions d: d: Γ X Γ ⇒ R, where Γ represents the set of things of interest. These functions allow the transformation of the data matrix, being d(i,j) the calculated distance between the elements i and j. The dissimilarity functions need to follow some criteria, namely: After meeting the properties listed above, if the metric also has the property ( , ) = | | ( , ), it is called the norm. The hierarchical method was used to construct the clusters, consisting of identifying groups and the probable number g of groups by a series of successive mergers or consecutive divisions. VOSviewer calculates the score per author using the count and fractional method. The first equally scores the authors of a document, and the fractional criterion divides the score by the number of collaborating authors. And then, the force-association algorithm is used to normalize the raw data and build a distance and graph-based literature visualization map (Hu et al., 2022).

Results
This article suggests using a bibliometric analysis of the ECS network to track the literature scene. The database for this study's sample was compiled using the Web of Science indexing basis. Using it, we are compiling studies related to influential periodicals that provide significant results for this field of research [15]. The publications were obtained through the keywords [(-energy consumption‖) AND (-economic growth‖) AND (-renewable energy‖) AND (-Brazil‖)] so that they should appear in the title, abstract, or keywords of the articles.

Figure 2. Publication types -keywords -energy consumption‖ AND -economic growth‖ AND -renewable energy‖ AND -Brazil‖
We will follow the approach taken by Hu et al. (2022) using CiteSpace and VOSviewer. VOSviewer and CiteSpace are bibliometric analysis software based on information visualization written in Java. From its results, it is possible to trace the development of the literature as well as trends and research frontiers (Chen et al., 2009).
As a result, 177 documents were collected, and their types are distributed as shown in Figures 1 and 2. 79% of the sample is made up of peer-reviewed articles, which is helpful for the implications of our analysis of the results. The documents cover the period from 1995 until 2022.   As seen in Figure 1, it is not surprising that China is at the top of the list of studies related to renewable energy, given that it is also at the top of the list of investments in RD&D for renewable energy in 2021 (Bhutada, 2022). This has the effect of encouraging countries in the Asia-Pacific region to pursue similar energy transition goals, diversifying their energy portfolios, and providing energy security to nations with high levels of hydrocarbon reserve depletion. Countries like Uzbekistan, Mozambique, Bangladesh, and Nigeria have contemporary literature in this field of study and strongly resemble Turkey and China, albeit with less quantitative evidence.
The ten most essential affiliations in the literature on Brazilian economic growth and renewable energy consumption are listed in Table 2, ranked by the number of articles published. This list is led by the International University of Cyprus (11 articles), the University of Gelisim in Istanbul (6 articles), and the University of Sakarya (5 articles). It is important to note that even though fewer articles have been published by organizations like King Abdullah University of Science and Technology, Huaqiao University, and National Chiao Tung University, these organizations have more citations overall than the organizations that topped the ranking. We used VOSviewer to visualize cooperation networks in the ESG Nexus literature (Figure 7). The node size represents the number of documents per institution. The width of connections indicates the degree of cooperation between organizations. The wider the link, the greater the existing collaboration between institutions. The set of nodes of the same color represents a cluster, organizations with a substantial degree of cooperation. We can see from Figure 4 that there are six institutional groupings, and those in green and red have the most significant number of connections. The red cluster has Sakarya University with greater centrality in relationships and has extensive collaboration with several institutions, with no one standing out due to recurrence. On the other hand, the University of Coimbra coopers firmly with Evora University, Beira Interior University, and Fluminense Federal University. Other institutions are grouped, as we can see in Figure 4, forming a broad network of cooperation and collaboration in research. These studies relate to the environmental Kuznets curve, CO 2 emissions, and empirical analyzes of energy consumption and economic growth in E7 countries. Figure 5 shows research collaboration networks, including authors with at least one article. As can be seen, collaboration in this field presents a large set of clusters. However, few groups are connected. It is worth paying attention to the restriction of the research field we carried out, which may reduce the possibility of overlapping ijef.ccsenet.org International Journal of Economics and Finance Vol. 15, No. 4; themes. In addition, some research areas suffer from the phenomenon known as the small world effect, in which collaborations are restricted to people in the immediate circle of the authors, making it challenging to share methods, resources, and knowledge (Moody, 2004).

Figure 5. Cooperation networks between different authors
Note. Prepared by the authors based on data collected on the Web of Science.
An author's productivity is an essential indicator of their influence on literature. Table 3 lists the ten most productive authors for literature. The list is led by author Tomiwa Adebayo, followed by Andrew Alola and Murad Bein. These authors are linked to the International University of Cyprus, located on the Island of Cyprus. Cyprus is currently developing essential projects in the energy sector alongside Europe. It has attracted researchers and entrepreneurs from the industry to the region, a process that became more intense after discovering hydrocarbon sources on the coast of the Island. The journal co-citation analysis allows observing the major journals in the research field. When at least one article from two journals occurs concurrently in a cited paper, both journals will have a citing relationship in the joint. Figure 6 shows the visualization of journal co-citation networks. The node size represents the number of common citations: the more significant this number, the greater the importance of the journal for the topic. Table  4   The analysis of authors' co-citations aims to identify authors with a high citation rate, configuring it as essential to identify multidisciplinary training and related research areas. The view of co-citation relationships is shown in Figure 7. Nodes indicate authors, and connections represent co-citation relationships. As in previous analyses, the size of the nodes means the citation number for the author that labels the nodei.e., the more prominent, the more influential the author. ijef.ccsenet.org International Journal of Economics and Finance Vol. 15, No. 4; The ten primary cited documents are listed in Table 5, along with information on the total number of citations, year of publication, and respective DOI code. The articles by Pesaran, Sebri & Ben-Salha, and Apergis & Payne are, respectively, the three most cited articles by authors in the literature on economic growth and renewable energies in Brazil (Apergis & Payne, 2012;Pesaran, 2007;Sebri & Ben-Salha, 2014). Keywords are words extracted from the text (title, abstract, or body of the document) that seek to synthesize the main content of the research in terms of hypotheses, methods, and/or evidence. Thus, analyzing the co-occurrence of specific keywords among the documents of a research field can indicate a possible theoretical and empirical approach predominant in the literature, as well as likely trends and frontiers to be developed. In Figure 8, we present the co-occurrence of keywords grouped by clusters. We see five main groupings, with very defined themes interconnected by nodes that refer to economic methods. CiteSpace's burst detection can assess the most referenced keywords and their importance.  rational. If S > 0.7, clustering is convincing. Both the modulus value of Q and S should be as close to 1 as possible. For our application, Q = 0.8223 and S = 0.9435. The results point to 10 knowledge clusters based on data on the co-occurrence of keywords. In Table 11, the details related to each group are organized.
We see that only clusters 0 and 3 have S < 0.7, which demonstrates some precision in clustering data related to the literature on the nexus of renewable energy and economic growth in Brazil.  Figure 9 and Table 7 show that -biomass energy consumption‖ is the first cluster, with 55 articles and a silhouette of 0.692. The second and third clusters are -carbon neutrality‖ and -top-10 polluted countries‖, including 48 and 32 articles, with silhouette values of 0.738 and 0.833.

Final Considerations
The relationship between economic growth and the energy consumption is an important topic in the literature on energy economy and economic development. The need to accelerate the decarbonization of economies is discussed, as is the importance of analyzing the empirical relationship between emissions of greenhouse gases and economic growth, considering technological advancements, regulatory changes, and international capital flows. This study aimed to perform a bibliometric analysis of Brazil's economic growth and renewable energy sources. Complete document data from the Web of Science database were used. Data description and visualization were obtained using the programs CiteSpace and VosViewer.
An extensive and comprehensive co-authorship and co-citation analysis are performed, and hot research and research frontiers are discussed. The crucial role of topics related to the financial sector, the application of new econometric methods of causal identification on unconsolidated empirical evidence, and the importance of technological innovations for the frontier of research on this topic are highlighted.