Econometric Analysis of Brazilian Scientific Production and Comparison with BRICS

Abstract

This article proposes a study that placed Brazilian scientific production at the core of the debate, and Brazil as a member of BRICS framework emphasises its importance. It starts from the argument that scientific production is a key factor considering that it is through knowledge that societies evolve. This issue becomes even more relevant when pointing to the capitalist production model that transforms knowledge into wealth. That was the starting point for leading an econometric study to formulate a model to explain Brazilian scientific production from 1994 to 2014. The least square method was used to estimate the parameters of a production function to identify the variables that most influenced, statistically, the behaviour of Brazilian scientific production. The main variables that strongly explained the number of published articles, proxy of the scientific production, are number of post-graduate programmes, number of masters and doctors, Brazilian population and expenditure on research and development (R&D). The present study showed that postgraduate programmes are the ones that most influenced the behaviour of Brazilian published articles. Therefore, it is necessary to equip Brazilian postgraduate system to create the basis of a structure, which, in the near future, will produce knowledge in large scale and with high competitiveness.

Introduction

Scientific production is an important indicator of research productivity and of Scientific and Technology systems as it depicts innovation potential, an intangible good, into something tangible that serves as basis of wealth creation in a society. Individual institutions and researchers are classified based on scientific productivity, which is typically based on the number of published papers, journal impact factors and number of citations (Almeida, Pais & Formosinho, 2009, pp. 134–142). Scientific and technological research is vital to secure economic prosperity and social development particularly given that competition of knowledge instituitons in capitalism has become an important criterion of the system that gives comparative advantage of nations. Transforming knowledge into wealth is a goal of capitalism based on the premise that capital accumulation forms the essence of a system based on the formation and distribution of wealth.

Knowledge, skills and creativity are valuable assets that, if exploited well, allows success in the modern economic world. They are the key to designing high-value goods and services and advanced business practices (Jessop, 2000, pp. 63–78). Thus, scientific production is an indicator that is closely linked to the wealth-generating capacity from knowledge base of a nation as the economic system guides this knowledge through to the formation of wealth and social welfare.

In addition to the dissemination of knowledge itself, scientific production determines what good science should be, it demarcates centres of excellence in time and space, it reveals trends that guide the progress of science and technology and, it acts as an ‘invisible hand’ in the allocation of financial, physical and human resources focused on technical and scientific training, in turn optimising scarce resources and applying them in optimal ways for maximum benefit. The publication of scientific articles is widely used in academia as a measure of the productivity of researchers and institutions. Conventional bibliometric methods generally evaluate research trends by investigating the publication outputs of different countries (Tian, Wen & Hong, 2008, pp. 65–74).

For the scope and purposes of this study, however, only the number of published articles is used to assess scientific production, as this is a more appropriated measure. Brazil plays an important role in global scientific production, occupying the 23rd position in the global ranking and placing first among Latin American countries in 2014 (Nature Index, 2015). With the adoption of a new model for evaluating Brazilian postgraduate programmes developed by the Coordination for Improvement of the Higher Education/CAPES in 1998, scientific production has become a crucial element for measuring the quality of teachers and programmes with an emphasis on scientific production (Balbachevsky, 2009). This emphasis shows that, overall, Brazil is on the progressive path, as the National System of Science, Technology and Innovation (SNCTI) has achieved good results in relation to policies guiding the advancement of Brazilian science. Nevertheless, when analysing the country in a segmented manner, it is evident that Brazil suffers from regional issues among institutions, areas of knowledge and other aspects that form the structure of Brazilian science.

The application of a National Innovation System (NIS) reinforces the need to create an institutional arrangement, based on the strengthening of interactions that should occur not only between less developed and more developed economies, but above all, in local interactions between firms, universities and centres of research, financial and legal systems and government among others, to transform the social potential into economic growth. NIS also provides an important framework for comparison with other regions and countries. The concept of NIS linked to neo-Schumpeterian thinking and disseminated by authors like List (1841), Rosemberg (1976), Nelson (1977), Freeman (1987) and Lundvall (1992) explains the learning process towards innovation influenced by culture and history and by institutional, political and economic dimensions that form a country’s NIS (Cassiolato & Soares, 2013; Motta e Albuquerque, 2004). However, structural features that explain this situation are not easily identifiable. Although development agencies in Brazil have dedicated 30 per cent of all funds to conduct research in northern and northeastern regions since 2003. It is known that the southern and southeastern regions control the bulk of scientific activities regardless of whether we consider this a regional issue. This reflects a strong concentration of investment in science, technology and innovation in Brazil (Arora, David & Gambardella, 1988, pp. 49–50). These disparities can also be seen in knowledge field structures. The hard sciences receive more funding than the soft sciences and publish more. This reflects a worldwide phenomenon due to aspects of the development of science itself. The article is structured as follows. In the next section, we will briefly review Brazilian science and technology policy strategies. Following this, the section will present the econometric research on Brazilian science production. Then the next section will draw some comparison with BRICS. Some concluding remarks are presented in the last section.

Brazilian Science and Technology Policy Strategies

The twenty-first century began with a concern with the strategic issues to put Brazil in the forefront of the world’s leading nations with Brazilian policy on Science, Technology and Innovation (ST&I). Taking this as a milestone, the first National Conference of ST&I held in Brasilia in 2001 set an agenda that contributes to transform Brazilian society into a knowledge society (CGEE/MCT/ABC, 2002). Brazil’s acknowledged ability to generate highly qualified human resources in areas such as aeronautics and space, telecommunications, energy, oil and petrochemicals, agriculture, health, etc., faces the challenge of transforming this accumulated knowledge into innovation. It is necessary to identify strategic areas that could be developed to strengthen the national scientific and technological structure and drive it to optimise the result of public and private investments. In recent years some advances have been observed, such as the implementation of the National Strategy for Science and Technology from 2012 to 2015 (MCTIC, 2012), which put into practice the main conclusions and guidelines drawn from the debate of the First National Conference of ST&I and subsequent conferences in the following years.

Recognition of Science, Technology and Innovation as the structuring axis of Brazil’s development has indicated the need to articulate governmental policies to raise Brazil’s development pattern where the basis is public investment (federal and state universities, state-owned enterprises and state institutions for research support/FAPs). This strategy identified challenges, structuring axis, priority and complementary programmes committing approximately US$ 23.7 billion¹ (MCTIC, 2012). As this policy document observes:

Brazil has the challenge of reducing the scientific and technological gap that still separates Brazil from the more developed nations; The expansion of the consolidation of the Brazilian leadership in the knowledge economy; The broadening of the bases for environmental sustainability and the development of a low carbon economy; The consolidation of Brazil’s new international insertion pattern; Overcoming poverty and reducing social and regional inequalities. (MCTIC, 2012, pp. 33–38)

In accordance with this policy strategy, several programmes have been created, including new funding patterns for S&T development, scientific and technological infrastructure and qualification of human resources with priority in pharmaceuticals and industrial health complexes, oil and gas, industrial defence complexes, aerospace, nuclear, frontiers for innovation (biotechnology, nanotechnology and new materials), green economy (energy, biodiversity, climate change, oceans and coastal areas) and ST&I for social development. Some complementary programmes were also created, such as sustainable agricultural production, water resources, Amazon and the semi-arid, Pantanal and Cerrado programmes. The new National S&T Strategy for the period 2016–2019, in partnership with the scientific community and the productive sector, is currently in force, defining eleven priority areas for the strengthening of the National System of Science, Technology and Innovation to increase competitiveness (MCTIC, 2016). The understanding that knowledge is a key point for a society to evolve is a permanent concern of governments engaged in social welfare and reveals the intention to maintain governmental policies aimed at the organisation and consolidation of the National System of ST&I committed to the long-term vision. As already highlighted, one of the main bottlenecks that Brazil faces in terms of its ST&I system is the transformation of knowledge into innovations. The Innovation Research (Pintec, 2014) carried out by the Brazilian Institute of Geography and Statistics/IBGE showed the status of innovation driven by companies in Brazil. While there is an industry interest in acquiring ready-made technologies through the acquisition of capital goods, there is also a concern to generate the country’s own technologies through R&D expenditures.

As Graph 1 indicates, the behaviour observed during the period 2000–2014 is of concern because the rates related to the incidence of internal R&D have stabilised. Although the rate in 2000 was 10.3 per cent there was a loss of competitiveness of the industry given the low investment in R&D. When analysing the role of public sector in innovation activities, the participation of universities and research institutes is predominant. With regard to public incentive to industry innovations, public sector has been fundamental, including investment in R&D (programmes and projects) and policy formulation to encourage innovation (laws and guidelines). Public policies allied to private strategies have been used to increase the competitiveness of Brazilian economy, where the investment in private R&D becomes a long-term strategy.

Graph 1

Source: Brazilian Institute of Geography and Statistics/IBGE, Research of Innovation 2014.

Graph 2 above, shows the share in percentage between internal and external expenditures made by companies in selected years.

Other important indicators of Brazilian National S&T System are considered to build a broad vision of this sector in the country.

As seen in Graph 2, although internal R&D expenditures (generated by companies) were higher than external expenditures (acquired from other companies), the behaviour in the period was stable in terms of percentages in relation to the country’s gross domestic product (GDP). The conditions indicate the need of a greater participation of companies in this issue, mainly because there was a reduction of internal R&D expenditures from 0.50 to 0.45 per cent when comparing 2008 and 2014.

Graph 2

Source: Technical Note IPEA (2016).

When we explore the issue of human capital and training in academic institutions there is an increase in the number of masters and PhDs of 589.88 per cent between 1994 and 2014. The figure stood at 9708 in 1994 which was increased to 66974 in 2014. (Table 1). This fact reinforces the investment, mainly public, in the training of human resources to improve the development of research in Brazil. The consolidated structure of Brazilian postgraduate programmes also has contributed to this growth. The country also experienced a period of sustained growth in the number of enrolments in higher education institutions and increased the number of scholarships granted to undergraduate students beginning scientific research, stimulating the involvement with its activities. The importance of this result points to the creation of an increasing potential for post-graduation, including grants for research and scholarships. It is important to point the need to create a contingent of researchers in areas more sensitive to innovation, a gap that is being filled by federal government’s ‘Science without Borders Program’, which has sent almost 100,000 students abroad for training to major universities worldwide.

Table 1

Variables of the econometric model (1994–2014)

Period	Number of Published Articles	Number of Post-Graduate Programs	Number of Master (MsC) and Doctor (PhD)	Brazilian Population	Expenditure on Research and Development (R&D) (t - 2)
1994	4,561	1,243	9,708	156,430,949	10,276,486,030
1995	5,298	1,315	11,479	158,874,963	13,620,099,715
1996	6,321	1,209	13,484	161,323,169	20,724,821,051
1997	7,595	1,274	15,542	163,779,827	23,790,794,227
1998	8,911	1,259	16,266	166,252,088	24,763,264,881
1999	9,928	1,363	19,812	168,753,552	26,950,058,646
2000	10,851	1,439	23,139	171,279,882	28,060,025,491
2001	11,487	1,550	26,053	173,808,010	29,556,262,540
2002	12,841	1,683	31,338	176,303,919	36,879,906,582
2003	13,822	1,818	35,743	178,741,412	38,068,775,061
2004	15,962	1,930	34,751	181,105,601	37,388,915,135
2005	16,955	2,057	39,623	183,383,216	37,471,654,582
2006	18,939	2,265	41,627	185,564,212	38,008,770,475
2007	23,400	2,408	42,805	187,641,714	40,839,299,478
2008	28,394	2,567	46,725	189,612,814	42,048,259,362
2009	30,307	2,718	50,156	191,480,630	48,625,813,114
2010	31,992	2,840	50,904	193,252,604	53,430,013,519
2011	34,690	3,128	55,554	194,932,685	52,833,073,032
2012	36,928	3,342	61,050	196,526,293	58,862,619,778
2013	38,402	3,528	66,034	198,043,320	60,112,760,312
2014	39,503	3,678	66,974	199,492,433	61,709,480,766

Sources: 1) Published articles: The Web of Science/WoS.

2) Post-Graduate Programmes.: Coordination for Improvement of the Higher Education/CAPES.

3) Number of masters (MsC) and doctors (PhD): Coordination for Improvement for Higher Education/CAPES.

4) Brazilian populations: Brazilian Institute of Geography and Statistics/IBGE.

5) Expenditure on R&D: Ministry of Science, Tecnology, Innovation and Communications/MCTIC.

The growth in the number of higher education institutions in that period accompanied the increase in the number of enrollments, reinforcing the strategy of creating an infrastructure for the training of human resources with potential for scientific and technological research. There is an incidence of a greater number of private higher education institutions compared to public institutions, which are concentrated in some federal and state universities and public research centres, where research is much more prolific.

Regarding the patents there has been an increase over time in the number of Brazilian patents granted in the United States Patent and Trademark Office (USPTO), from 98 in 2005 to 362 in 2014. The number is encouraging considering a behaviour that may indicate an increase in Brazil’s economy innovative capacity. On the other hand, there is a concern about the low absolute number of total patents (2,022 in 10 years). As a better indicator for Brazilian innovation, the analysis should include the number of patents granted internally. There was a registration of 18,486 patents in 2005, reaching 30,435 in 2014, according to data from the National Institute of Industrial Property (INPI). In general, the country has increased the number of patents, but there is a need to intensify innovation activities to increase its global participation.

This article examines the determinants of scientific production in Brazil in the period from 1994 to 2014 to identify variables that have strongly influenced its generation. It is not intended to present an analysis of Brazilian scientific production indicators but to determine the main factors that explain such production indicated by the number of articles published in indexed journals, to help policy makers formulate policies for the development of Brazilian science. As a complementary analysis, this article presents a comparison between Brazil and BRICS framework to evaluate Brazilian ST&I pattern when comparing to countries with similar level of development in an acknowledged scheme in the global scenario.

Econometric Analysis of Brazilian Science Production

Methods and Sources

The methodology that was adopted formulated a theoretical model that determines Brazilian scientific production and then, through simulations that have selected variables with high levels of statistical significance, defined an econometric model that measures the influence of determinant variables of scientific production. Many studies using the econometric approach have been conducted in order to understand the relations of the variables that affect R&D and scientific production (see Castellani, Montresor, Schubert & Vezzani 2015; Pilinkiené, 2015).

While there are no direct measures of new knowledge, previous studies have used a variety of proxies. The most commonly used proxies are publications and citations (Geuna & Crespi, 2008, pp. 565–579). For the present study, we used variable in the econometric model such as the number of published articles as a proxy of scientific production. We performed statistical tests (Student’s t-test, F-Snedecor and Durbin–Watson) as well as statistical analysis (heteroscedasticity, homoscedasticity and correlation matrix) to evaluate consistency levels to validate the model. We employ the least squares method to obtain estimated parameters of the studied population from observed data. The regressions were designed to evaluate the relationship between dependent and independents variables. Table 1 shows those variables that were selected to form the econometric model after simulations were conducted. The analysis covered the period between 1994 and 2014.

In addition to the explained variable (published articles), which is representative of scientific production, two of the four explanatory variables were included in the scientific and academic fields (number of postgraduate programmes and number of masters and doctors), one from the demographic field (Brazilian population) and the other from the economics field (expenditure on R&D).

Econometric Results and Discussion

The econometric results show that the model that was formulated to represent Brazilian scientific production is statistically significant at 95 per cent. Our analysis of the coefficient of multiple determination (R²) shows that 99.18 per cent of the variations observed in published scientific papers are explained by variations that were observed in the independent variables. Table 2 shows the econometric results after running the regressions.

This result highlights the good fit of the model. In short, the number of post-graduate programmes, the number of masters and doctors, Brazilian population and the expenditures on R&D are linearly correlated with the evolution of published articles in Brazil for the studied period. Figure 1 shows the fit of the observed data to the regression model. The straight line denotes the curve generated by the regression production function, and dots denote the dataset (observed). The model was found to be statistically significant.

Table 2

Econometric result output

ANOVA
	gl	SQ	MQ	F	F significance
Regression	4	2814411772	703602943.1	483.5212936	1.83483E– 16
Residual	16	23282629.41	1455164.338
Total	20	2837694402
	coeficients	standart error	t stat.	p-value	95% confidence interval
	coeficients	standart error	t stat.	p-value	lower	upper
Intercept	–93215.6066	24498.81081	–3.804903323	0.001556547	–145150.7538	–41280.45952
PG_Progr.	18.0468744	2.120211896	8.511825863	2.45792E–07	13.55222703	22.54152187
MsC and PhD	–0.75638971	0.203641226	–3.714325045	0.001884488	–1.188089723	–0.324689689
Population	0.00050789	0.000156494	3.24539122	0.005069941	0.000176132	0.000839639
Expend. R&D	2.7464E–07	1.01755E–07	2.699038723	0.015803085	5.89295E–08	4.90349E–07

Source: Generated by statgraphics software.

Figure 1

Source: Generated by statgraphics software.

Strong adjustment was confirmed through an F Snedecor test, which yielded a calculated value of 483.52, a result that is higher than the standard value of 3.01. This result indicates that overall, the independent variables explain the dependent variable. To examine isolated effects of each independent variable on the dependent variable (partial individual regressions), Student’s t-test was conducted. According to the regression results, all of the variables are statistically significant with 95 per cent confidence, as all calculated values are higher than 2.120. To support the results of the t-test, we used the p-value. Statistically, ‘p-values’ of less than 0.05 denote a strong correlation (Gujarati, 2000). The p-values in this study confirm the results of the t statistical analysis. We also performed an elasticity analysis to determine the sensitivity of quantitative variables of the linear econometric model. The results reveal inelastic behaviours in three of the four coefficients, denoting that changes in the number of published articles are proportionately less significant than changes observed in the number of postgraduate programmes, in the number of masters and doctors and in R&D spending. The explained variable proved to be less sensitive. This inelasticity, however, did not compromise the explanatory power of the independent variables as they exhibit high explanatory power regarding the behaviour of the number of published articles representing Brazilian scientific production. The only variable that showed elastic behaviour was the Brazilian population, which generated a coefficient of elasticity of higher than 1. This coefficient shows that the number of published articles is highly sensitive to this variable.

Empirical Analysis and Discussion on Results

As shown through our analysis of the regression results with 95 per cent confidence, the model explains behaviours observed in relation to the number of published articles in Brazil between 1994 and 2014. As the series on article publication represents Brazilian scientific production, the econometric model can identify which variables explain production patterns. Through our validation of the econometric model, we are able to draw conclusions that help us understand environments surrounding the processes of knowledge generation in Brazil through the publication of scientific articles and to propose ways to orient the Brazilian science system towards increased production.

Number of Postgraduate Programmes

The number of postgraduate programmes was influenced heavily by the number of published articles for the period, showing that the organisation of knowledge production processes for related areas involves a highly productive scheme. This format has allowed for the social organisation of researchers and scientists employed mainly at public universities, concentrating the brightest minds of respective areas of knowledge in one place and enhancing knowledge production mechanisms. Postgraduate programmes and their respective areas of research specialty are environments in which researchers with their skills and interests engage in joint, integrated and convergent activities, in turn collectively producing knowledge (Bernardino & Alentejo, 2014). These are true centres of knowledge production, and due to their high levels of productivity, they are also centres of excellence that publish scientific articles.

Our finding that these programmes have the greatest influence on scientific production can guide policy makers’ decisions. Based on this argument, investments made to improve academic infrastructure, develop researcher abilities, strengthen research funding mechanisms, and increase research incentives (grants/productivity grants) among others constitute strategic actions that should be geared towards the strengthening of Brazilian postgraduate programmes to accelerate national scientific production. With the creation and expansion of postgraduate programmes in Brazil since the 1970s, university research activities have involved strong participation from research groups and postgraduates (Parker, 2011, pp. 26–61), increasing the need for more investment in postgraduate programmes in Brazil. Despite the low sensitivity of this variable due to the coefficient of elasticity (|0.11|), the high explanatory effect of postgraduate programmes on scientific production warrants attention to this sector, highlighting the need for university activities that support agents that are directly involved in scientific research. In addition, macro-policymakers (Ministry of S&T,I; Secretariats of State; national, state and local funding agencies) must help other system agents strengthen Brazilian postgraduate programmes in science. Such investment should follow two paths. First, investments should be made by universities, whereby the strengthening of less structured programmes that require more institutional support should guide institutional policies. Second, government agencies (municipal, state and federal) should formulate policies that strengthen universities and the postgraduate system as a whole.

Unequal distribution strategies of financial resources must be employed in public universities with a less pronounced tradition of research excellence. Federal and state incentives must be applied depending on levels of administrative dependence. In regard to the private sector, institutions themselves must promote incentives for research with funds raised through economic activity, promote investment in academic and laboratory infrastructure and train faculty, in turn stimulating the development for research efforts and capacities for innovation. Public administrations can create an environment that is conducive to new investments and that requires the participation of other governmental structures, in the field of education and other fields such as economic planning, infrastructure and financing organisations, to create incentives for private investment in scientific research.

Number of Masters (MsC) and Doctors (PhD)

From our analysis of the number of masters and doctors, we conclude that their influence on the publication of scientific articles relates to their contributions to the postgraduate system. While postgraduate training is not sufficient for the insertion of researchers in scientific research positions, it helps postgraduates become researchers. The actual contingent allocated to scientific research from the number of masters and doctors is lower as new doctors and masters can follow other paths. Part of the new contingent of PhDs, while also engaging in academia, is engaged strictly in teaching activities; some leave the system by performing activities in the nonacademic government sector; others enter the formal non-academic market, and finally, others are driven to work directly for companies (public or private). Therefore, while this group of individuals has the potential to conduct ample research, only some of these individuals are engaged in this activity.

Among masters who are still undergoing research training, some continue their training to complete a PhD while others leave the system or teach, as was observed for graduated doctors. Numerous new masters leave the academic research system, as the private sector of higher education calls for high-quality teaching based on market logics, ascribing the new graduate a key role in the private education production process. Already by 2002, the private sector constituted the main market for Brazilian academic masters, employing 70 per cent of teachers with this title (Soares, 2002). In addition, some masters become PhDs. This may have generated a model specification bias, as the same individuals would have been included in the group of PhDs for subsequent periods. This potential source of duplicity may have overestimated the total number of potential researchers and may explain inconsistencies observed in the variable parameter as shown by our analysis of regression signals.

We therefore found an even smaller contingent of masters who effectively conduct research after graduating. Their academic status as emerging researchers accounts for this. It is known that the role of masters and doctors in the research environment involves contributing to potential research training. Even when understanding that the real contingent of new graduates engaged in scientific research is smaller than the number of potential researchers generated from the Brazilian graduate system, this is an important variable of scientific production to consider. As the regression statistically confirms the high explanatory power of this variable, increasing capacities for research justifies the permanence of the number of masters and PhDs in our model. The following question arises when evaluating the Brazilian postgraduate system and its capacity to receive new researchers. Does the Brazilian postgraduate system support the absorption of the entire contingent of masters and PhDs in postgraduate programmes to facilitate national scientific production? The absorption of new researchers will involve engagement in research groups, and so it is natural for postgraduate programmes to create competitive minimum conditions for acceptance; this requires improving the physical, laboratory and financial conditions of academic programmes. Thus, this raises the need for more investment in training and/or strengthening of postgraduate programmes by financial resource holders and for new shares to be created by policy makers of academic and scientific policies. This necessitates improving our scientific system and making it more productive.

Brazilian Population

Our analysis of the econometric results for Brazilian population shows that the number of people influences growth in the number of scientific papers through a directly proportional relationship. In principle, as the country’s population grows, the number of articles published in Brazil also increases according to the econometric model. It is expected that increasing the population rate should increase the national scientific production when focusing on the influence of the population. However, demographic policies have discouraged population growth globally. This has been a consequential decision given the effects of population growth on various systems with which we are involved. Many scholars have listed population growth as a threat to ecological, social, political and economic well-being (Babones, 2002; Foley et al., 2005; Hodren & Ehrlich, 1974; Jones, 1998; Meadows, Meadows, Randers & Behrens, 1972; Solow, 1956; Vorosmarty et al., 2000), warranting birth control policies and highlighting future uncertainties. This raises concerns regarding the role of Brazilian educational systems and policies considering the observed influence of population growth on scientific production. The following question is thus formulated. What are the issues related to population growth in Brazil, and what, contrary to incentives for population growth in the country, could justify the importance of Brazilian population for scientific production?

The answer lies in the fact that regardless of the size of the population or even the pace of its growth, its contribution to scientific production concerns the existence of people. The inherent logic of this statement centres on the fact that scientific production exists as a result of knowledge generation, which in turn is possible because rationality is involved in this process. Rationality is a human trait and is of course rooted in demographic processes. This underscores the importance of populations for scientific production. In analysing other major knowledge-producing nations that have larger populations and even those with low population indices, we found that in all of them, scientific production is directly dependent on the population’s existence and not on the population size or rates of change. Table 3 shows data for publication of scientific articles in 2014 (Reddit, 2014) and population estimates for 2015 (United Nations [UN], 2015) for selected countries.

Table 3

Comparison Between Population and Published Articles (Selected Countries)

Country	Population (thousands)	Population ranking (2015)	Published articles	Publication ranking (2015)
USAJapanThe NetherlandsBrazilSwitzerland	321,774126,57316,925207,8488,299	41167 697	310,20668,30829,29627,80821,372	1 5131518

Source: Created by the authors.

From the ranking of published articles for 2015, the selected countries beyond Brazil held the first, fifth, 13th and 18th places. In both cases, for the USA and Japan, which are more populous countries, and for Switzerland and the Netherlands, which are less populous, the total number of articles published in each country was impressive. Brazil with an estimated population of 207,848,000 inhabitants in 2015 also had a good performance with 27,808 published articles, occupying the 15th position. These examples demonstrate that it is not the size of a population or the rate of its change but a country’s capacity to generate knowledge that determines a country’s global ranking among producers of knowledge. Scientific production is thus influenced by population that researchers are part of and is therefore linked to educational policies that encompass all levels of individual training. This is dependent on long-term public policies that require investment in education systems and that if well managed through well-defined projects can generate positive results in the long run.

Unique are those countries with low population indices such as Switzerland and the Netherlands, which have continually invested heavily in education and which have created strong educational systems and knowledge generation structures. These data lead us to believe that what determines large-scale qualified scientific production is the educational aspect and not the population. Thus, as already expressed by those in charge of this process, there is a need to improve the Brazilian educational system at all levels.

Although Brazil’s postgraduate system is quite evolved and structured, it is highly concentrated among a small group of large public universities, strengthening postgraduate and scientific elitism in the country. In addition, lower levels of educational training must be addressed through state policies that strengthen the Brazilian educational system at the earliest levels of individual training (elementary and high school). Policy makers should thus promote a deep restructuring of the country’s educational system, and the government should identify gaps and propose solutions to support real changes in Brazilian education.

Expenditures on Research and Development (R&D)

Given that the production of knowledge requires investment to provide support in physical and financial terms, we include expenditures on R&D as a percentage of the GDP in the econometric model. This variable is considered as an important input of knowledge generation. More than simply representing direct investments in knowledge, this variable is also critical for economic growth through the generation of knowledge that once applied, increases wealth and prosperity through the generation of new technologies. Studies show that R&D spending is highly strategic (Akcaldi & Sismanoglu, 2015; Gonzales-Brambila, Reyes-Gonzales, Veloso & Perez-Angón, 2016; Rosenbloon, Ginther, Juhl & Heppert, 2015; Yan, 2015). We further found that increased R&D spending leads to higher growth rates in the number of citations, which in turn can be explained by an increase in publications. After matching disciplinary citation dates with R&D expenditures, it was found that disciplinary citation growth rates align citations with growth rates in global R&D expenditures, thus providing evidence revealing the impact of R&D expenditures on knowledge production (Yan, 2015, pp. 1–29).

There is a positive relationship between research resources and knowledge production while assessing the publication of scientific articles (Rosenbloon et al., 2015). Inter-regional R&D networks are critical for the production of new knowledge in Europe (Wanzenbock & Piribauer, 2016). It was necessary to use a lag time to demonstrate that effects of investment during period t will be felt in subsequent periods. Therefore, the model is based on a 2-year period so that effects on current knowledge production through the publication of scientific articles are a reflection of investments made two years earlier. Thus, the period considered in the case of R&D spending runs from 1992 to 2012 to the maturity of investments and the generation of published articles. A time lag of one to several years must be considered in studies on knowledge production and other scholars have studied 6- and 7-year periods when examining the publication of scientific articles as a result of R&D spending (Wanzenbock & Piribauer, 2016). However, there is little evidence of any positive effects occurring over two years in the case of publications (Geuna & Crespi, 2008, pp. 565–579). We found that, in this study, a lag time of 2 years best fits the observed data and generates satisfactory econometric results.

Our econometric results confirm the international literature’s finding that expenditures on R&D are an important source of scientific production. Intuitively speaking, one might assume that R&D investments increase the number of scientific articles published from research projects. On average, it takes two years for the initial results to be transformed into articles or into other forms of scientific dissemination. In Brazil, R&D spending as a percentage of GDP has increased, confirming a direct relationship between these variables and an increasing number of published articles. When taking a time lag of two years into account, the average growth rate in the number of articles published between 1994 and 2014 was 10.83 per cent, while the average growth in expenditures on R&D for 1992–2012 was 8.91 per cent, revealing a positive effect. Therefore, to support scientific production, higher levels of R&D spending are needed. After calculating the elasticity between these variables, we found that for every 1 per cent increase in R&D spending, there is 0.31 per cent increase in the number of articles published. Despite this low degree of sensitivity, in conjunction with the other variables considered in the model, expenditures on R&D are critical for further national scientific production. This statement points to the need to formulate investment policies in science, technology and innovation that may encourage scientific production and economic growth in the future.

Brazil’s S&T and Comparison with BRICS

In this section we have analysed the participation of Brazil in BRICS-STI perspective. BRICS (Brazil, Russia, India, China and South Africa) is the group of five countries leading the emerging economies, with the exception of Russia, that counts with approximately three billion people, which is around 42 per cent of world’s total population of about 7.074 billion, and hold about US$ 23.386 billion of combined GDP (PPP) of the world, with China on top (US$ 12.269 billion) followed by India (US$ 4.793 billion), Russia (US$ 3.373 billion), Brazil (US$ 2.365 billion) and South Africa (US$ 586 million) (Hasan & Luthra, 2014). The importance of BRICS in the global scenario has highlighted the issues surrounding economic growth based on the progress of their R&D systems and have become crucial to their respective societies. Based on the argument that innovations are at the basis of economic growth and hence of the development of nations, the First BRICS Science, Technology and Innovation Ministerial Meeting was held in Cape Town, South Africa in 2014 ‘to discuss and coordinate positions of mutual interest and identify future directions of institutionalizing cooperation in science, technology and innovation within the framework of BRICS’ (BRICS-STI, 2014, p. 1). The transformation of knowledge into wealth and social development is one of the great challenges that BRICS countries have faced, since the particularities that surround their societies range from differences between culture and society to economy and politics. Therefore, the path to be followed must take into account the relations between the state and the respective National Innovation Systems. Thus ‘an evolutionary approach has been adopted in order to capture the nature of the state in the respective countries and thus understand the historical and ideological basis for its role in the evolution of the NSI in the five countries’ (Cassiolato & Soares, 2012, p. iv). For these authors there is a dichotomy, starting in the 1980s, between market and state-based development goals in developing countries. The following Tables 4–7 show data about important scientific and technological indicators of BRICS.

Overall, BRICS countries have been publishing more along the period of 2008–2014. There is a ratio of 1.93 per 1 in comparison between 2014 and 2008, that is, for each article published in 2008 almost two were published in 2014. It strengthens the significant development observed in the national systems of science of BRICS members, although China has strongly influenced this trend. A comparative analysis between a percentage share of BRICS countries and the Brazilian contribution in the publication of scientific articles shows that Brazil had a good competitive output. The relative participation of Brazil in BRICS total was inferior when compared to China and India but superior when compared to Russia and South Africa. Brazil indicated a participation of 12.41 per cent with 240,216 articles, Russia indicated 9.65 per cent with 186,803, India 16.46 per cent with 318,462 articles while China presented the best performance with 58.29 per cent with 1,127,966 articles. South Africa presented a percentage below two digits with only 3.18 per cent of BRICS total with 61,512 published articles. Despite the good performance of Brazil, it is necessary to enhance Brazil’s scientific and technological systems by encouraging researchers and institutions via sectorial policies that include the improvement of the infrastructure of the research centres and universities and stimulating the private sector to invest more in innovation. Besides that, it is necessary to increase the percentage of R&D/GDP to finance the knowledge production sector as observed in China.

Table 4

Number and percentage share of BRICS countries in publishing research papers 2008–2014

Period	BRASIL			RUSSIA			INDIA			CHINA			SOUTH AFRICA			BRICS TOTAL
Period	Published articles	% of 240,216	% of Total Brics	Published articles	% of 186,803	% of Total Brics	Published articles	% of 318,462	% of Total Brics	Published articles	% of 1,127,966	% of Total Brics	Published articles	% of 61,512	% of Total Brics	Published articles	% of 1,934,959
2008	28,394	11.820	14.385	25,671	13.742	13.005	36,520	11.468	18.502	100,324	8.894	50.826	6,480	10.535	3.283	197,389	10,201
2009	30,307	12.617	13.917	26,121	13.983	11.995	38,233	12.006	17.557	116,026	10.286	53.279	7,083	11.515	3.253	217,770	11,255
2010	31,992	13.318	13.554	25,377	13.585	10.751	41,492	13.029	17.579	129,516	11.482	54.871	7,661	12.454	3.246	236,038	12,199
2011	34,690	14.441	12.935	26,862	14.380	10.016	45,361	14.244	16.914	152,351	13.507	56.808	8,921	14.503	3.326	268,185	13,860
2012	36,928	15.373	12.448	26,080	13.961	8.791	47,590	14.944	16.042	176,483	15.646	59.490	9,580	15.574	3.229	296,661	15,332
2013	38,402	15.986	11.357	27,641	14.797	8.175	52,322	16.430	15.474	209,434	18.567	61.939	10,331	16.795	3.055	338,130	17,475
2014	39,503	16.445	10.374	29,051	15.552	7.629	56,944	17.881	14.954	243,832	21.617	64.034	11,456	18.624	3.009	380,786	19,679
TOTAL	240,216	100.000	12.415	186,803	100.000	9.654	318,462	100.000	16.458	1,127,966	100.000	58.294	61,512	100.000	3.179	1,934,959	100,000

Source: The Web of Science (updated on October 13, 2017).

Table 5

Research papers published by BRICS countries per selected areas in 2008 and 2014

Subject area	Brazil		Russia		India		China		South Africa		BRICS Total
Subject area	2008	2014	2008	2014	2008	2014	2008	2014	2008	2014	2008	2014
Agricultural Science	820	2,074	205	191	2,113	2,015	1,787	4,512	222	327	5,147	9,119
Astronomy/Astrophysics	437	802	1,253	1,449	660	1,095	1,093	2,036	138	486	3,581	5,868
Biochemistry Molecular Biology	1,230	1,473	999	930	1,567	2,835	5,236	11,852	183	226	9,215	17,316
Chemical Science	470	815	4,995	5,439	7,132	12,517	19,849	46,140	386	757	32,832	65,668
Computer Science	2,243	3,154	297	382	709	1,612	3,613	10,374	92	127	6,954	15,649
Geoscience	238	446	1,033	1,087	657	987	1,544	3,970	171	301	3,643	6,791
Mathematical science	859	1,233	1,876	2,047	1,075	1,606	6,984	14,388	276	499	11,070	19,773
Medical science	6,276	8,476	746	1,010	5,346	7,321	11,101	34,596	1,399	2,236	24,868	53,639
Physical science	2,328	2,909	7,863	8,127	5,211	7,751	18,020	28,883	331	696	33,753	48,366
Psychology	155	352	33	84	45	131	426	1,140	212	332	871	2,039
Social science	2	13	34	70	46	125	260	726	108	305	450	1,239
Total	15,058	21,747	19,334	20,816	24,561	37,995	69,913	158,617	3,518	6,292	132,384	245,467

Source: The Web of Science (updated on October 13, 2017).

Table 6

Gross Domestic Expenditure on R&D in BRICS (2008–2014)

(Millions of current PPP)
	2008/2009	2009/2010	2010/2011	2014
Brazil	22.2	23.4	25.3	35.4
Russia	30.1	33.5	32.8	42.6
India	Na	Na	Na	66.5
China	120.7	154.0	178.2	409.0
South Africa	4.7	4.4	4.0	4.8

Sources: RSA (2013).

For 2014: Brazil (World Bank); Russia (OECD); India (2016 Global R&D Funding); China (National Bureau of Statistics); South Africa (OECD).

Table 7

Gross Domestic Expenditure on Research and Development in BRICS as a Percentage of GDP (2008–2014)

	2008	2009	2010	2011	2012	2013	2014
Brazil	1.129	1.12	1.159	1.141	1.128	1.198	1.168
Russia	1.044	1.252	1.130	1.022	1.046	1.056	1.088
India	0.867	0.845	0.822	0.831	Na	Na	Na
China	1.443	1.252	1.710	1.022	1.046	1.991	2.021
South Africa	0.888	0.835	0.737	0.735	0.735	0.723	Na

Source: World Bank.

When comparing the intra-block countries and selected areas, the following situation is observed between 2008 and 2014. Brazil increased 44.42 per cent, Russia increased 7.66 per cent, India increased 54.69 per cent, China increased 126.87 per cent and South Africa increased 78.85 per cent. Overall all countries achieved an exceptional result, particularly China. Brazil’s result was mainly due to its robust postgraduate system and an increase in public funding for research, placing the country at the forefront of the major nations that publish scientific articles in spite of being a developing country. The challenge for Brazil is to transform knowledge into new technologies which, despite the increase in R&D investments (Table 6), still demands mechanisms that promote innovation, such as technological and industrial policies and the insertion of private enterprises in the generation process of innovations.

When analysing R&D investments in the period 2008–2014, it was observed that China not only had the largest amount, but also had the highest growth rate with 238.85 per cent from 2008 to 2014, while Brazil presented 59.45 per cent. Russia performed 51.5 per cent and South Africa indicated only 2.13 per cent. India spent US$ 66.5 million in 2014, ranking above Brazil, Russia and South Africa. Data as to India’s investments for the previous period were not available. These data demonstrate a Chinese aggressive policy of ST&I, especially in the issue of innovation. Brazil needs not only to increase its level of R&D spending, but also to create policies aimed to encourage the participation of private sector in the innovation generation system, such as the recent National Knowledge Platforms Program. In recent years, the country has formulated industrial policies in addition to the incentive for high technological sectors such as information and communication technologies, industrial complex for health, oil and gas, defence, aerospace, nuclear, biotechnology, nanotechnology, green economy and S&T for social development (IPEA, 2015).

In terms of R&D expenditures in relation to GDP, Brazil presented an average rate of 1.15 per cent while Russia presented an average of 1.09 per cent, demonstrating that Brazil invested proportionally more than Russia in the period 2008–2014. India for the period 2008–2011 and South Africa for the period 2008–2013 allocated less (an average of 0.84 and 0.78 respectively) of their GDP. China presented an expressive amount of financial resources and reached an average of 1.49 per cent of its GDP, confirming its aggressive policies of ST&I. In Brazil, the major part of the funds comes from public sources from companies such as Petrobrás, Embraer and Embrapa, or from public research institutes such as INPE, FIOCRUZ, INCA, CETEX, IpqM, DCTA, and Embrapii, a special agency created in 2013, all belonging to the federal government (IPEA, 2015). The country presents a bottleneck because private companies do not invest much in R&D activities, which requires mechanisms to internalise investments in the sector, through fiscal incentives, strategic public policies and the creation of a more stable economic environment for more foreign direct investment/FDI. In general, the analysis of the indicators led to the conclusion that Brazil has a lot to do to reach out its partners, especially China, in terms of the National ST&I system. China has made a great financial and human resources effort, based on the policies devoted to the sector of knowledge and innovations, even considering that a great part of the resources invested comes from public sources. Therefore, in addition to domestic public policies, it is imperative to collaborate internationally with partners who, in general, can contribute positively to the creation of an environment conducive to the generation of new technologies. Not only Brazil, but the other BRICS members will rivalise in a more competitive level in the global arena with nations more evolved economically and technologically which demands more improvements in their National Systems of Innovation.

Conclusions

What motivated this study was the claim that scientific production is central to the goals of a society and country. Given this, we found it necessary to further our understanding of this phenomenon with respect to the stock of knowledge that forms the basis of human development and which is presented as a source of new discoveries. Based on this, we sought to identify the main variables that influence Brazilian scientific production through an econometric study and by designing a model that explains the power of this form of production. We conducted a quantitative regression analysis through which we estimated parameters of variables that shape scientific production. As a way of representing this form of production, we used the variable number of published articles to measure national scientific production for 1994–2014.

Our regression results, which were subjected to a series of tests and statistical analysis, show that the model is statistically significantly and explains Brazilian scientific production. The most significant variables are the number of postgraduate programmes, the number of masters (MsC) and doctors (PhD), the Brazilian population denoted by the number of inhabitants and the expenditures on R&D. The number of postgraduate programmes is the best variable for explaining scientific production for the study period. Therefore, we suggest that these programmes can serve as targeted initiatives aimed at strengthening scientific production through the actions of universities and external factors, such as policy makers of federal, state and municipal agencies. It is also important to encourage the private education sector to participate more actively in educational activities. Policies that stimulate the private sector to conduct research activities to increase national scientific production require the presence of a favorable economic environment, as such institutions have a market vision for their activities.

We also found that the graduation of masters, and especially doctors, enables the formation of a potential quota for research and boosts scientific production. We show that higher education institutions feed Brazil’s postgraduate system by contributing to the formation of a potential postgraduate sector quota for the scientific production system, in turn translating into further research potential. The concentration of scientific production among a few public institutions highlights the importance of policies that strengthen all other institutions in the public sector. Furthermore, educational policies must focus on the private education sector to stimulate postgraduate investment and consequently, the potential for scientific production. In the future, it is expected that an increasing number of new MsCs and especially PhDs will become engaged in scientific research and will contribute to scientific production in Brazil in the presence of stronger postgraduate programmes.

The only exogenous variable that showed statistical significant power in explaining scientific production was the Brazilian population as measured as the number of inhabitants. From this analysis, we draw conclusions on the supply of new scholars who, after completing postgraduate training, can carry out research and contribute to scientific production in Brazil. As this is not dependent of population size or on the rate of population growth, the importance of the population lies in its existence rather than in its quantity. To optimise scientific production based on the Brazilian population, it is necessary to change the Brazilian educational system, and a long-term perspective must be employed to achieve desired levels of quality education in the country.

We also highlight the importance of R&D spending. Despite its endogenous characteristics, because it was included in the S&T system in this study, R&D spending was considered as an economic variable. The literature points to its fundamental role to knowledge generation process.

These results are interesting due to Brazil’s importance at a global level and this importance is strengthened when considering that the country is a member of BRICS, a multilateral block that aggregates five important developing economies and therefore requires national policies to be convergent to a process of cooperation that transforms knowledge into wealth and place these countries in a more competitive condition when facing developed economies. This challenge is not easy because there are asymmetries due to remarkable differences among these countries in the respective National Systems of Innovation. China is an example of BRICS which confirms that aggressive policies linked to a high level of investments in ST&I is the key factor to reaching a new pattern of development and international recognition, demonstrating that strategic decisions towards an agenda which aims at the local potentiality and foreign experiences should be implemented. Therefore, Brazil needs to improve its NSI mainly with the participation of the private sector in R&D activities, which requires policies that encourage the private investment in a more favorable environment to new inversions in the long run. As the major part of the investment in R&D comes from the public sector, it is strategic to convert private resources into knowledge to consolidate the potential for innovation.

The Brazilian structure of knowledge is well developed but there is a gap when analysing it for innovation. The Brazilian NSI must operate to improve this important part of the system and, therefore, conquer a new level in the global context of nations with a higher level of development and a new position in the frontier of knowledge and technology. Finally, we expect that this study will offer a stronger understanding of conditions of scientific production in Brazil while also highlighting ways of promoting improvements to knowledge generation system and the conversion of that into innovations to benefit Brazilian society.

Footnotes

Note

References

Akcaldi

B. Y.

& Sismanoglu

(2015). Innovation and the effect of research and development (R&D): Expenditure on growth is some developing and developed countries. Elsevier. Procedia Social and Behavorial Sciences, 195, 768–775. doi:10.1016/j.sbspro.2015.06.474

Almeida

J. A. S.

Pais

A. A. C. C.

& Formosinho

S.J.

(2009). Science indicators and science patterns in Europe. Journal of Informetrics, 3(2), 134–142. doi:10.1016/j.joi.2009.01.001

Arora

Ashish

David

Paul A.

& Gambardella

(1988). Reputation and competence in publicly funded science: Estimating the effects on research group productivity. Annales d’économie et de statistique, 49/50, 163–198.

Babones

S. J.

(2002). Population and sample selection effects in measuring international income inequality. Journal of World-Systems Research, VIII(I), 8–28.

Balbachevsky

Elizabeth

(2009). A pós-graduação no Brasil: novos desafios para uma política bem sucedida. Retrieved 22 November 2016, from https://www.researchgate.net/publication/237073967

Bernardino

M. C. R.

& Alentejo

E. S.

(2014). Ranking da produção científica dos programas de pós-graduação em ciência da informação no Brasil. Brazilian Journal of Information Science-Research trends, 8(1).

BRICS-STI (2014, February). First BRICS Science, Tecnhnology and Innovation. Ministerial meeting: Cape Town declaration. Cape Town.

Cassiolato

J. E.

& Soares

M. C. C.

(2012). The state and the national system of innovation: A comparative analysis of the BRICS economies. New Delhi: Routledge.

Cassiolato

J. E.

& Soares

M. C. C.

(Eds). (2013). Introduction: BRICS National Systems of Innovation. In The Role of the State. London: Routledge.

10.

Castellani

Montresor

Schubert

& Vezzani

(2015). Multinationality, R&D and productivity: Evidence from the top R&D investors worldwide. International Business Review, 26(3), 405–416.

11.

CGEE/MCT/ABC (2002). Parcerias estratégicas. Edição Especial, v.1, n.14 (junho de 2002). Centro de Gestão e Estudos Estratégicos/Ministério de Ciência e Tecnologia/Academia Brasileira de Ciências/CGEE/MCT/ABC: Brasilia.

12.

Foley

J. A.

DeFries

Asner

G. P.

Barford

Bonan

Carpenters

S. R.

… Snyder

P. K.

(2005). Global consequences of land use. Science, 309, 570–574. doi:10.1126/science.1111772.

13.

Geuna

& Crespi

G. A.

(2008). An empirical study of sicentific production: A cross country analysis, 1981–2002. Elsevier, Research Policy, 37, 565–579.

14.

Gonzales-Brambila

C. N.

Reyes-Gonzales

Veloso

& Perez-Angón

M. A.

(2016). The scientific impact of developing nations. PLoS One, 11(3), e0151328. doi:10.1371/journal.pone.0151328.

15.

Gujarati

Damodar N.

(2000). Basic econometrics. São Paulo: Pearson Education do Brasil.

16.

Hasan

S. A.

& Luthra

(2014). Comparative performance of India with other BRICS countries in publishing science and engineering research papers. Currente Science, 106(12), 1654–1657.

17.

IPEA (2015). Towards a long-term strategy for BRICS: A proposal by the BRICS (Think Tanks Council/BRICS). Institute of Economic and Applied Research, Brasília. 184p.

18.

IPEA (2016). Inovação no Brasil: crescimento marginal no período recente (Nota Técnica n. 34). Institute of Economic and Applied Research, Brasília.

19.

Jessop

(2000). The state and the contradictions of the knowledge-driven economy (Knowledge, Space, Economy, pp. 63–78). London, UK: Routledge.

20.

Jones

C. I.

(1998). Introduction to economic growth. New York/ USA: W.W. Norton & Company, Inc.

21.

MCTIC (2012). Estratégia Nacional de Ciência, Tecnologia e Inovação: 2012–2015 (Balanço das atividades estruturantes 2011). Ministry of Science, Technology, Innovations and Communications/MCTIC, Brasilia.

22.

MCTIC (2016). Estratégia Nacional de Ciência, Tecnologia e Inovação: 2016–2019. Ministry of Science, Technology, Innovations and Communications/MCTIC, Brasilia.

23.

Meadows

H. M.

Meadows

D. L.

Randers

& Behrens

W. W.

III (1972). The limists to growth. New York/ USA: Potomac Associates Book.

24.

Motta e Albuquerque

(2004). Ideias Fundadoras: Christopher Freeman - The ‘National System of Innovation’ in Historical Perspective. Revista Brasileira de Inovação, 3(1), 9–13.

25.

Nature Index (2015). Retrieved 25 March 2016, from https://www.natureindex.com/annual-tables/2015/country/all

26.

Parker

A. L.

(2011). Os periódicos brasileiros e a comunicação da pesquisa nacional (Revista USP, no. 89, pp. 26–61). São Paulo.

27.

Pilinkiené

(2015). R&D investment and competitiveness in the Baltic states. Procedia-Social and Behavioral Sciences, 213, 154–160.

28.

Pintec (2014). Intituto Brasileiro de Geografia e Estatística/IBGE (2016). Pesquisa de inovação: 2014. Coordenação de Indústria. Rio de Janeiro.

29.

Reddit (2014). Top 40 countries by the number of scientific papers published. Retrieved 22 August 2017, from https://www.reddit.com

30.

Rosenbloon

J. L.

Ginther

D. K.

Juhl

& Heppert

J. A.

(2015). The effects of research and development funding on scientific productivity: Academic chemistry, 1990–2009. PloS One, 10(9), e0138176. doi:10.1371/journal.pone.0138176.

31.

RSA (2013). National survey of research and experimental development: 2010/11—Fiscal year: Main results. Pretoria: Department of Science and Technology.

32.

Soares

M. S. A.

(2002). A educação superior no Brasil. IESALC/UNESCO. Porto Alegre, Brazil.

33.

Solow

R.M.

(1956). A contribuition to the theory of economic growth. Quartely Journal of Economics, 70, 65–94.

34.

Tian

Wen

& Hong

(2008). Global scientific production on GIS research by bibliometric analysis from 1997 to 2006. Journal of Informetrics, 2, 65–74. doi:10.1016/j.joi.2007.10.001

35.

United Nations (UN) (2015). World population prospects. The 2015 revision. New York. Retrieved 6 October 2016, from https://esa.un.org/unpd/wpp/publications/files/key_findings_wpp_2015.pdf

36.

Wanzenbock

& Piribauer

(2016). R&D networks and regional knowledge production in Europe: Evidence from a space-time model (Vienna University of Economics and Business. Department of Economics Working Paper No. 207). doi:10.1111/pirs.12236

37.

Yan

(2015). Disciplinary knowledge production and diffusion in science. Journal of the Association for Information Science and Technology, 67(9), 1–29.