Figure 2 displays the benefit rate of normalized citation impact, scientific excellence, and technological impact. Differences in percentages of all and leading papers determine the benefit earned by each country when the given country does not lead collaboration. As noted earlier, the lower the benefit, the more autonomous the country; the higher the benefit, the more dependent the country is on foreign collaborations for citation gains. For instance, we find that China and India derive very little benefit in terms of citation or excellence and only moderate advantages for technological impact. However, China’s citation rates exceed those of India. This displays the complicated notion of autonomy—autonomy does not necessarily equate to higher or lower research performance. The inverse is also true. For example, Chile and Cuba both attain relatively high technological impact: for Chile, there is high dependency, for Cuba nearly total autonomy.
Figure 3 compares the benefits of collaboration to technological impact and normalized citation. Russia, Chile, and Mexico are relatively autonomous. Cuba leads publications cited in patents (just two papers in the period), showing a higher capacity for knowledge transfer than Colombia or Venezuela. Cuba is a particular case showing a negative benefit in technological impact, which means that the country would not need international collaboration to be autonomous. However, this finding could give rise to misinterpretations, as there are few documents cited in patents and Cuba leads international collaboration. The growth of collaboration most likely helps the country to leverage its competitiveness, taking advantage of its autonomy. If so, Cuba might reinforce this aspect in order to attract international participation and increase its potential.
Figure 4 illustrates the trends behind the influence of different types of collaboration (x axis) in the normalized citation of each country (y axis). For each country, we show the evolution over time of the normalized citation by types of collaboration. Each line shows the tendency by indicating the initial year with a circle and the final one with an arrow. The dotted line located in each graph refers to the worldwide average of normalized citation.
Brazil presents a clearly growing trend in its total and leading international collaborations, yet the percentage of papers published by one single institution and in collaboration with several domestic institutions decreases. This pattern positively affects the normalized citation of the country, especially in domestic collaborative papers. Total and leading international collaboration is seen to have a positive effect on impact, yet international collaboration is the most determinant for high citation. Furthermore, output in collaboration allows Brazil to eventually obtain normalized citation indexes above the world average.
In Mexico, the panorama is quite different. There is a rising trend for the impact of output involving international collaboration, as well as national and international collaboration, though a reverse trend appears for the levels of impact of output coming from leading collaboration. Non-collaboration and its impact decrease over the period of study, while national collaboration increases, but is not accompanied by markedly higher citations.
The impact of Argentina’s internationally collaborative output has increased, even though the percentages of publications has decreased considerably. There is a citation gain for leadership roles in mixed international and national collaboration, while a decrease is seen for leadership on purely international projects. Hence, the association of domestic and foreign institutions produces the most visible outcomes. Impact is on the rise across all collaboration types, with the exception of non-collaborative output, which remains below the world average throughout the period studied.
Chile's output with all and leading international collaboration increases steadily, in both cases reaching impact values above the world mean by the end of the period analyzed. The greatest gains in impact are seen in leadership of international and national collaborations. Note that while papers and impact in national collaboration decrease, the association of national and international collaboration led by the country put Chile on the map at the international level. This pattern is relevant to orient national strategies intended to foster the visibility of Chilean research.
For Colombia, nearly all output is below the world mean at the end of the period, showing virtually no growth. Thus, this country obtained more benefits from international collaboration with leadership but did not increase its share of production. It shows no real decline in non-collaborative output, though its impact is substantially reduced. The greatest change is in the rise of impact of its leading international output, though the volume of that output has not considerably increased.
Cuba increases only its output in international collaboration, and much more in output with leadership than without leadership. This means it also considerably increases the levels of impact, which at the end of the period reach an index > 2. Non-collaboration shows a considerable decline in both output and impact. The national and international collaboration without leadership grows slightly in volume and impact, while collaboration with leadership grows in volume but ends up with somewhat less impact.
The case of the Russian Federation is noteworthy considering that the expenditures of business enterprise, government and the higher education sector in this field are far greater than in countries such as Japan and Korea (OECD 2013). Yet all its NST output—except that entailing international collaboration—has impact levels below the world mean at the end of the period, showing virtually no percentage growth apart from leading international collaboration. Thus, this country increases considerably the number of papers published by one single institution, with a significant drop in citations respect to the world average. It shows no real decline in domestic collaborative output, though its impact is substantially reduced. The greatest change is the decrease in national and international output, with or without leadership, although in either case, the impact achieved lies below the worldwide rate.
South Africa increases only its output in non-collaborative papers with an increase in its impact that does not reach the global standard. South African output decreases in all the other types of collaboration, with a dramatic drop in citations, especially in international collaboration: at one point, it surpassed the world average, reaching 2.3, but at the end of the period it lies at the same level of visibility as domestic collaboration.
A different pattern is observed in India, where non-collaborative papers are the most common output and surpass the world citation average at the end of the period. At the same time, national collaboration achieves high citations despite a decrease in the number of publications. The greatest decrease in the publications is found in international collaboration, yet this output is the most visible over the period. In contrast, there is a drop in visibility with respect to national and international collaboration, with or without leadership.
China is the only country where any kind of output is above the world average, despite the fact that non-collaborative and domestic papers accumulate more than half of its entire production. The trends in these two models follow different strategies—while non-collaborative papers decrease, domestic papers increase. The international collaboration also suffers a drop but is accompanied by an increase in citations. At the end of the period, national and international collaboration are the best associations in terms of citations.