THE USE OF ECONOMETRIC MODELS IN STUDIES OF ELECTRICITY GENERATION FROM BIOMASS: a bibliometric analysis

The present research investigated the utilization of econometric models in studies related to the generation of electricity from biomass, through a bibliometric analysis. The general objective of the study was to analyze the publications, from 1987 to 2018, that explored the potential of econometric models involving biomass for electricity generation. Additionally, it was intended to investigate the most cited articles and authors, seeking to verify the most relevant themes and the main econometric techniques used in the analyses; verify the countries most engaged in these researches, drawing a parallel with their current energy strategies; identify future trends of studies in this area. For this, the SCOPUS database was used, selecting articles in English, from the keywords “Econometrics, biomass and Electricity”. The data collected with the literature review were compiled on thematic maps, with the help of the Vosviewer software, which realized analysis of citation, co-citation, co-authorship and keywords. SciMat software was also used, which generated from key terms, longitudinal strategic maps that allowed identifying future trends on the theme discussed in this article. The results indicated that the main research fronts in this field are related to the use of econometrics to estimate the impacts of energy generation from biomass in variables such as economic growth, energy demand and greenhouse gas emissions.


Introduction
In the last decade, interest in understanding the relationships between economic and social variables associated with the production and use of electric energy through renewable resources, especially biomass, has grown significantly. (Arisoy & Ozturk 2014;Bhat 2018;Rehman, Deyuan, Chandio, & Hussain 2018). For energy related research, econometric models are versatile and can be used to analyze a multitude of problems, from issues related to energy efficiency and security (Alexander et al, 2006;Zhang et al. 2018), elasticities of energy products and inputs (Atalla, Bigerna, & Bollino 2018;Ko et al. 2010), climate change (Dong, Sun, & Dong 2018) as well as assisting in the development of long-term energy policies for countries. .
Biomass has been considered one of the most promising inputs to be used as a primary energy source to replace traditional fossil fuels. The current scenario, marked by population growth, food and energy demand, and the need to develop new paradigms related to the mitigation of environmental impacts caused by human activities, is a strong challenge to be faced by countries (Bulut & Muratoglu 2018;Martins et al. 2019). One of the solutions is the use of renewable energy sources to reduce greenhouse gas (GHG) emissions from fossil sources such as oil, coal and natural gas and to diversify the global energy matrix. An example is biomass and its waste, which has a 9.7% share in primary energy generation as demonstrated by data from the International Energy Agency (IEA 2018), volume larger than traditional and technically sources better developed, like nuclear (4.9%) and hydraulic (2.5%).
Despite being used since the mid-1950s, econometric models, before the last decade, when studying the electricity market, basically concentrated on the relationships between economic growth and energy consumption (Darmstadter 1971;Mainguy 1967;Mason 1955).
More recent studies, such as Hannesson (2009) have included other variables such as the price of oil, reaching more robust conclusions on the dependence on fossil fuels and broader geopolitical issues, such as energy security.
Only from 1983 began to emerge works with the use of more robust econometric models, such as (Bajracharya 1983) and Hosier & Dowd (1987), which analyzed the relationships of Martins Thus, the general objective of this research is to analyze the publications, from 1987 until 2018, that exploited the potentialities of the econometric models in the studies involving biomass for electricity generation. Additionally, it was intended to investigate the most cited articles and authors, seeking to verify the most relevant themes and the main econometric techniques used in the analyses; verify the countries most engaged in these researches, drawing a parallel with their current energy strategies; identify future trends of studies in this area.
For this purpose, a bibliometric research was conducted using the keywords "Econometrics, Bio-mass and Electricity", with the help of VOSviewer and SciMat software, as well as an analysis of time intervals, using the SCOPUS database, establishing, quantitatively and qualitatively, the main results of the research on the use of econometric models in studies of electricity generation from biomass, as well as the main trends in this field of study, given the development seen in this area in recent years. Martins

Bibliometric analysis
The bibliometric research refers a wide range of studies that emerged at the beginning of the 20th century, as an alternative to the need for further studies and evaluations of scientific production and communication and its direction. Among many other purposes, it can be used as a quantitative tool to evaluate the past contribution to science by research entities and also to predict their future research potential. In other words, bibliometric studies aim to demonstrate the direction of science in a given field of knowledge. (Gautam 2017

Choice of bibliometric indicators and methods
Quantitative indicators were used to measure the productivity of a researcher, periodical or country in terms of number of publications and citations, according to the methodology established by Cadavid Higuita, Awad, & Franco Cardona (2012). These indicators aim to mediate the frequency with which a work, author or periodic is cited in other researches, relating works, authors, institutions, countries and keywords according to the chosen method, mapping the units of analysis, in this case the keywords, according to this metric.
For the present research, the selected bibliometric methods were: Citation Analysis, which uses citation as a measure of influence, assuming that the most cited authors, papers and journals are more influential; Co-citation Analysis, seeks to answer which group of authors is systematically cited by a determined group of papers and which researches are jointly referenced, using the joint citation as a measure of similarity. It assumes that the more two works are cited together, the more their content is related, deducing the most influential authors of the studied area; Co-Author Analysis, answer which authors work together, as well as, which institutions or countries collaborated in a given field of research, i.e. identifies the measure of collaboration between the publications; Co-word Analyses, which identifies which words-key has been used more in each period of time, which of them are used together and as interest in research has changed over time. (Cobo, López-Herrera, Herrera-Viedma, & Herrera 2012;Zupic & Čater 2015).
The four techniques employed allowed the mapping of the established research field, i.e., the use of econometric models in studies of electric energy generation from biomass, especially through the Co-Word analysis technique that identified the current research front and possible future trends of studies related to the theme discussed in this review.

Softwares used
Two softwares were used in the process. graphical representations of the maps. This is particularly useful when viewing large maps, facilitating the interpretation and is mainly used in the creation of maps based on network data (Cobo, López-Herrera, Herrera-Viedma, & Herrera 2011;Jeong, Cho, Park, & Hong 2016). The second tool was the SciMAT: developed by the group "SECABA" of the University of Granada, which allows the construction of scientific maps, as well as a better visualization of evolution within a scientific area (Cobo et al. 2012).
VOSviewer is a computer program that has been developed to create, view and explore academic and scientific bibliometric maps. The program is available for free via the  In the operational environment of SciMAT, the following configuration was established: word as analysis unit, co-occurrence analysis as a tool to build networks, equivalence index as a measure of similarity to normalize networks, and simple centers algorithm such as the algorithm used in the detection of clusters and creation of strategic diagrams, which are subdivided into the four quadrants reported. (Cobo et al. 2012). The mapping, according to these criteria, allowed the recognition of patterns in the publications and also revealed trends in research in the area of interest.

Results
As   Regarding the number of publications by author, Lin, B. is that it has the largest number of articles with topic "Econometrics and Biomass and Electricity", with 17 works. However, Apergis, N. is the one with the highest number of quotes. Table 1 shows the 11 authors, who have at least 10 papers published on the subject of interest of a total of 2,121 authors. To earn the quality of the publications, the Hirsch Index, or h-index, was included in the analysis. The index h is determined after sorting the publications of a researcher in decreasing order, according to their frequency of citation. Thus, the researcher or research group is classified according to its h level, which is equal to or less than the number of quotes for its articles (Schreiber 2015). That is, it is a proposal to quantify the productivity and impact of the researchers, based on their most cited researches. As a complement, the average citation per author was also inserted. Making simultaneous analysis among the main authors, in terms of number of publications, with the relevance of the works, that is, the number of citations, noticed that the main authors work, basically, in isolation, except for the clusters led by Shahbaz, M., Bilgili, F.  Interestingly, this cluster could be working together with the group led by Assumadu-Sarkodie, since both research lines the impacts of carbon emissions on economic variables, including similar econometric models shall be used. However, as can be seen in Figure 4, the clusters are far apart and without links.
Another common and relatively well-cited author is Arabatzis G. The works carried out in the cluster led by him are interested in understanding the relationships between the use of biomass, specifically wood, and energy generation and economic growth using econometric  With a little more emphasis on the relevance of the authors' area and influence, a cocitation analysis was performed, which identifies which authors are most frequently cited simultaneously in a given study. Following the same parameters of the coauthor analysis, i.e., selection of a minimum of five articles, with 25 citations each, 395 authors were selected, divided into five clusters ( Figure 5).
The co-citation analysis´s demonstrates that despite a low joint production, but at the same time with several authors researching the same object, as seen in the analysis of coauthorship, the theme "Econometrics, Biomass and Electricity" shows similarity between the mentioned works, at least within the clusters. Figure    real GDP and renewable energy consumption, indicating that an increase of 1% in the latter would lead to an increase of 0.76% in GDP.
The second work of the duo, with 242 citations, titled "Renewable and non-renewable Consumption-Growth Nexus: Evidence from a panel error Correction model" (Apergis & Payne, 2012), re-leased by Energy Economics, the authors showed, also by means of panel analysis, using unitary root test, the relationship between economic growth and the total consumption of renewable energy and non-renewable energy and their impacts on the formation of fixed capital and labor. The results of the panel showed that there is bidirectional causality between renewable energy consumption and economic growth, as well as between non-renewable energy consumption and economic growth, both in the short and long term. Moreover, the results also show that renewable and non-renewable energy consumption can serve as a substitute for each other.
For the purposes of a systematized analysis by keywords, and using the SciMat software, in an attempt to capture the main trends for the field of study of interest, the initial analysis  Despite the scarcity of topics in the period, when the analysis is visualized through the links of the cluster, it is observed that "Global-Climate-changes" has strong relationships with themes that would become focal in later periods, illustrating that the area's development would be due to terms related to the energy structure of the sector, development of new technologies, econometric models, developing countries and computer optimization programs (Figure 7). Martins    concern of the academy in answering more precisely questions that worry society, such as environmental conservation for future generations and production of sufficient food to meet the population demand.

Discussion
The research showed that the countries with the highest number of publications on the subject of interest were the United States (163), followed by China (129) but the trend is to be replaced, gradually, by natural gas, which today has participation of 5% in the matrix, but it could be as much as 11%, according to the agency's predictions.
Despite The United Kingdom which produced 66 papers on the subject of interest and was the first country to use coal to generate electricity, as well as other European nations, has been implementing emission reduction policies, through the use of less polluting energy sources.
According to the DRAX Annual Report (2018), combined renewable energy capacity in the UK reached 42 GW, while fossil fuels reached 40.6 GW in the third quarter of 2018. According to the report, it was the first time that alternative sources had a greater share in electricity generation in Britain.
As regards the analysis of the periods, from 1987 to 2000, the seven studies presented focused on the theme related to global climate change, but were not limited to this scope.
Examples of this are the works of B. Ang (1992) and Meade (1984), which use growth curve methodology for energy studies, involving mathematical models, energy demand and studies of substitution of non-renewable fuel sources by renewables.
The main associated terms were energy structure, economy, econometrics, developing countries, BRICS, coal, technology, fuels and optimization models. Despite the scarcity of research associating the three terms of interest, "Econometrics", "Biomass" and "Electricity", it was possible to identify some emerging future trends from the connections of the focal themes, for example, energy structure and economy, which became focal themes in the periods 2001 to 2015 and 2016 to 2018, respectively. The most widely used models were linear regression analysis (Dowlatabadi, Hahn, Kopp, Palmer & DeWitt 1993;Gupta & Hall 1997), logistic regression (B. Ang 1992) and multinomial Logit model (Hosier & Dowd 1987).
The scarcity of studies on the subject of interest is possibly related to macro-economic factors. At the end of the eighties, the price of the barrel of oil was stabilized after the second oil crisis between 1982 and 1983, ranging between US$ 28.00 and US$ 30.00, the ton, in this period. In addition, the price of coal was also at levels acceptable to the foreign market, being in the amount of US$ 30.00 a ton in the period, removing the incentive for financial and academic investments in alternative energy. In the second period, 2001 to 2015, 423 publications were identified. The focal themes were: "Energy-Structure", "Forestry-sector", "Land-Management" and "Technology". The term with higher density and centrality was "Energy-Structure". This comprehensive cluster was associated with terms like "Economics", "Econometrics", "Renewable-Energy", "Fossil Fuels", An emerging theme caught the attention: "Water use". This cluster was associated with "Global Climate changes", "sugar Cane", "Resource-Consumption" and "health Risks".
Although it did not become, specifically, a focal theme in the later analyzed period (2016 a 2018), it amplified the discussion on the use of biomass for energy generation and its impacts on the conservation of water resources. The publications explored the relationship between electricity consumption and water for food production (Asgharipour, Mondani & Riahinia 2012;Gündoğmuş 2006;Thankappan, Mid-more & Jenkins 2006) and the impacts of climate change on water demand for electricity production (Arnell 2004;Ncube, Zikhali & Musango 2013).
From 2001 to 2010 the most widely used econometric models were estimates by OLS, used especially for decision-making in relation to the energy policy to be adopted (Cormio, Dicorato, Minoia & Trovato 2003;Sundqvist 2004). The estimates by OLS were also used to evaluate problems of energy demand, associated with the analysis of its substitution by alternative energy sources, in particular biomass from forest waste and municipal solid waste (Chambwera & Folmer 2007;Edwards & Langpap 2005;Jebaraj & Iniyan 2006). Another common technique in this sub period was the Logit models associated with optimization models.
This technique was especially used in the analysis of choice problems between producing food or energy biomass (Johansson & Azar 2007;Ouedraogo 2006 The relation of electricity production through biomass and its analogy with poverty and social inequality was the focus of studies by Mousavi-Avval, Rafiee, Jafari & Mohammadi (2011b) that through a two-stage error correction model, estimated the impact of the production of electricity from forest biomass on these social variables, reaching the conclusion that when the management for the energy exploitation of wood is implemented, its use tends to decrease social inequalities. The same method was used by Song, Aguilar, Shifley & Goerndt (2012), associated with analysis by cointegrated panels to study the North American residential demand for energy from wood. variables such as economic growth, agricultural production, in focus on food safety and scenario estimation to test whether electricity generation through different types of biomass, effectively has the ability to decrease greenhouse gas emissions, consolidating, so far, future trends in studies in this area.

Final Considerations
The bibliometric study revealed that there is an extensive and diverse field of study associated with keywords "Econometrics, Biomass and Electricity". In a period of only three decades, it was possible to identify 912 articles dealing, exclusively, directly and indirectly, with the use of biomass for energy and electricity generation, analyzed from the economic point of view, with the aid of econometric models.
The research also found that although most publications are concentrated in the area of energy and environment production, 31.5% and 26.2%, respectively, the field of study Economics, Econometrics and Finance, has been standing out (10.2%)especially with researches led by China, which according to what has been published, wants to understand the impacts of the use of biomass energy in its most diverse economic sectors, as well as knowing the emission potential of this type of energy compared to traditional sources.
As regards the use of econometric models in this type of study, the research demonstrated that each year, the interests of the researchers became more complex, requiring the use of increasingly structured models. In particular, in the first period (1987 to 2000) and in part of the second period (2001 to 2010), the prevalence of regression analysis models by the linear method, where the equations were mostly presented in a reduced form, and estimated by means of OLS.
From 2011, it was possible to observe more interesting models such as instrumental variables, generalized moment methods and two-stage minimum square methods. Interestingly, no articles were identified with estimation methodology for maximum likelihood, and only one by means of Bayesian estimation, which are common in strict research in the area of economics and finance.
Future research should be focused on the development of studies that associate terms such as biomass energy and its costs, waste from biomass and its relation to electricity demand and impacts of biomass use for electricity generation on economic and social variables such as Martins GDP, GDP per capita and social well-being. In addition, there is a need for further deepening on the cultivation of energy biomass and its impacts on food production, especially in countries with low supply of arable land, as well research to test the effective reduction of greenhouse gases from biomass-intensive energy processes, in detriment of processes based on traditional energy sources.