Doner and Schneider (2019 , 2016 ) recently set up the debate on upgrading coalitions by bringing business policies to the heart of the literature on the middle-income trap. Although the literature on economic upgrading has never neglected the role of politics in changing the economic structure, the focus has most often been on the State.

Thanks to the seminal work of Gerschenkron (1962) , State initiatives are widely perceived as one of the most important mechanisms for economic upgrading. In fact, the importance of public incentives became even more widely appreciated after the publication of Mariana Mazzuccato’s books on the subject (2019, 2013), which outlined the crucial role played by the U.S. Government in the development of some of today’s most important technologies. However, the fact that States can build institutions does not mean that they have the capacities or the political will to do so ( EVANS, 1980 ). As Gourevitch (1986) famously put it, ‘policies require politics’.

Upgrading policies carry an aura of national cohesion with strong societal appeal; after all, who would oppose high quality education, technological autonomy and economic catch-up? Yet the politics of upgrading are far from smooth. From an evolutionary point of view, upgrading is not only redistributive but also destructive (SCHUMPETER, 1943). According to Nelson (2008) :

In this sense, lagging behind may not only be profitable, but also a survival strategy for those who hold economic power. This is the essence of the ‘trap’. BBs might deliver better innovation results than other sized firms, which may serve as a justification for government investment in and incentives in favor of their activities. However, governments cannot count on them to support institution-intensive solutions that could enable a country to exit the trap.

Due to their advantages of scale and scope, BBs are expected to be leading agents of innovation because they tend to be more efficient in developing new technologies than smaller firms ( COHEN, 2010 ; COHEN and KLEPPER, 1996 ; NELSON and WINTER, 1982 ). Consequently, BBs have become the primary targets of public incentives aimed at fostering innovation in less-developed countries. Both the Japanese and the South Korean experiences are used as successful empirical cases to criticize the ‘small is beautiful’ idea ( SCHUMACHER, 1973 ) and to defend the concentration and centralization of the capital of nationally-owned businesses. According to such a point of view, this is the only way for national businesses to compete globally, as they enjoy economies of scale and scope ( CHANDLER JR., 1990 ) that enable them to fight for disputable market share and can draw on resources sufficient to invest in R&D ( SCHUMPETER, 2003 ).

Notwithstanding the foregoing, the patterns of interaction between states and BBs are often fuzzy. The experience of East Asian big businesses contrasts with that in Latin America, where BBs are mostly concentrated in commodities, regulated sectors (especially banking and utilities), natural oligopolies (such as cement and beer) and low-technology manufacturing ( DONER and SCHNEIDER, 2016 ). The contrasting roles of BBs in East Asia and Latin America cannot be understood without looking at the development paths followed by the two regions.

While East Asia has chosen a nationalist path and sought to lay the foundations for indigenous technological development, Latin America has relied heavily on foreign capital and technology and remained dependent on the developed world ( AMSDEN, 1989 ; CHANG, 2003 ; EVANS, 1980 ; KOHLI, 2009 ). Latin American countries have given excessive importance to foreign capital, which has been assigned the role of increasing the technological capability of the region by supporting product and process innovations produced by transnational companies ( KATZ, 2000 ). The entrance of foreign direct investment (FDI) into the region has created profound economic segmentation by enhancing structural heterogeneity ( FURTADO, 1959 ) and feeding a ‘modernization-marginalization’ duality (FURTADO apud ALBUQUERQUE, 2005 ) that is characterized by high levels of inequality and informality ( OECD, 2018 ). In addition, time has shown that the emphasis on technology imports to the detriment of local development (with the expectation that the growth of transnational participation would result in technology transfers) has proven this policy to be a failure in overcoming underdevelopment ( ERBER, 2000 ).

The existence of domestic companies and powerful MNCs in Latin America lies at the core of the political obstacles to building the institutions necessary to implement upgrading policies in the region ³ ( ALBUQUERQUE, 2005 ; FURTADO, 1987 ). Without sustained and cohesive support from a broad segment of BBs, it is especially difficult to build high-cost, long-term upgrading institutions, especially in democratic settings where short-term political thinking weakens incentives for long-term investment ( DONER and SCHNEIDER, 2019 , 2016 ).

Moreover, there is a correlation between the heterogeneity of the structures in place and the proliferation of obstacles to the building of coalitions around common institutions, especially in respect of two fields that are essential for economic upgrading: education (human capital) and Research and Development (R&D). The former is a particularly interesting case and has been extensively explored in the literature ( DONER and SCHNEIDER, 2019 , 2016 ).

There is a universal consensus on the need for more and better education in order to foster economic development. However, most LDCs have not made a major effort to advance their education policies. The absence of effective measures to improve education can be explained by the absence of a unified voice or any political force that could support massive investments in education ( DONER and SCHNEIDER, 2019 , 2016 ). In effect, although almost all BBs complain about a lack of skilled workers, their real need is too feeble to support any governmental imposition of the heavy costs that would be needed to improve the quality of education.

As for domestic BBs, most of them tend to be most competitive in the areas of commodities (petroleum, agro-industry, minerals and metals, etc), low-technology manufacturing (tfood, beverages and clothing) and low-complexity services (utilities). Such sectors generally require small numbers of highly-skilled workers and large numbers of less-skilled workers. When specific knowledge is needed, firms can pay for in-house training.

By contrast, MNCs are more competitive in high-technology sectors and demand more skilled labor. However, they have the advantage of moving relatively more complex activities to places where more highly qualified and skilled human capital is abundant. When MNCs do need to hire skilled labor (especially when workers are less mobile due to the locations of natural resources or large domestic markets), they have the advantage of offering higher wages than local businesses or attracting employees directly from prestigious universities ( HEYES and RAINBIRD, 2011 ). Indeed, empirical evidence demonstrates that MNCs routinely provide higher wages in LDCs than their local counterparts do (BROWN, DEARDORFF, and STERN, 2003).

A similar argument can be made regarding the demand-side for innovation. Increasing investment in R&D is widely recognized as key to overcoming underdevelopment. As in education, R&D investments cannot be achieved overnight: they require a range of institutions that promote cooperation among universities and business, property rights protection, innovation financing, technical training and so on.

Generally, countries with the highest levels of gross expenditure on R&D (GERD) are the ones with the highest private sector participation in R&D ( Table 01 ) funding. In South Korea and Japan, private expenditure represented more than 76% of total GERD (in 2017), and both countries’ R&D investments represent more than 03% of GDP. In Latin America, by contrast, the innovation shortfall is accompanied by a lack of private R&D investment. In Argentina and Mexico, the private sector accounts lass then 20% of total GERD (in 2017), followed by Chile (31.43%) and Brazil (47.46%). The GERD as a percentage of the GDP in those countries has settled at less than 01% (except in Brazil, which has maintained its GERD/GDP at around 01%). Incredibly, China increased its GERD from 0.89 in 2000 to 2.15 in 2017. This increase, while coordinated by the central government, was pushed by the private sector and resulted in a jump from 57.59% to 76.48% of the country’s total GERD.

Table 01

	GERD, % of GDP					Percentage of GERD funded by the private sector
	2000	2005	2010	2015	2017	2000	2005	2010	2015	2017
European Union**	1.75	1.78	1.97	2.12	2.13	55.36	53.71	53.15	54.64	57.60
OECD	2.27	2.20	2.36	2.44	2.56	64.08	62.24	58.50	61.41	62.33
UK	1.63	1.56	1.66	1.67	1.66	48.31	42.06	44.05	48.96	51.77*
USA	2.62	2.51	2.74	2.72	2.79	69.03	63.31	56.93	62.54	62.48
France	2.09	2.05	2.18	2.27	2.19	52.52	51.93	53.50	54.04	56.08
Germany	2.39	2.42	2.71	2.91	3.02	65.85	67.51	65.52	65.58	66.18
Middle Income	0.64	0.84	1.11	1.36	1.52	-	-	-	-	-
Latin America	0.57	0.59	0.75	0.78	0.74*	-	-	-	-	-
Argentina	0.44	0.42	0.56	0.62	0.54	23.28	31.02	-	17.22	16.52
Brazil	1.00	1.00	1.16	1.34	1.26	44.16	50.44	47.04	45.55	47.46
Chile	-	-	0.33	0.38	0.36*	-	-	25.44	32.77	31.43
Mexico	0.31	0.40	0.53	0.52	0.49*	29.52	41.51	33.02	17.38	19.05
China	0.89	1.31	1.71	2.07	2.15	57.59	67.04	71.69	74.73	76.48
Japan	2.91	3.18	3.14	3.28	3.21	72.42	76.12	75.93	77.97	78.27
South Korea	2.18	2.63	3.47	4.22	4.55	72.38	74.96	71.80	74.55	76.23

Source: The authors’ own. Data sourced from the World Bank regarding GERD/GDP and from the OECD regarding GERD funded by the private sector. Dada regarding GERD funded by private sector in Argentina, Brazil and China was sourced by the Brazilian Ministry of Science, Technology, Innovations and Communications.

Note: (*) data for 2016; (**) 28 countries.

Low levels of private investment in R&D are one of several signs of a weak NSI. This weakness is particularly problematic in light of the high vulnerability of public revenues in middle-income countries. Latin American firms cannot rely on public support for long-term innovation projects, except those that are ring-fenced against budget cuts, which are frequent occurrences and de rigueur upon a change of government. Firms cannot invest in additional R&D if the duration of government support is uncertain ( GUELLEC and POTTERIE, 2003 ).

Here again the split between big domestic and MNCs de-incentivizes private investment in R&D. Although MNCs might be willing to support and even benefit from the local NSI, they always have the option of relying on the ST&I actors available at home or simply obtaining what they need from elsewhere, especially from somewhere whose NSI is mature. Despite increases in R&D performed abroad, many studies have shown that most such activities remain concentrated in MNCs’ home countries ( ZACHAREWICZ, SANZ-MENÉNDEZ, and JONKERS, 2017 ) and that MNCs have few incentives to base their R&D activities abroad, given that no more than 10% of their technological effort takes place overseas ( ARCHIBUGI and PIETROBELLI, 2003) .

To illustrate the previous discussion, it is interesting to look at some figures relating to the U.S. case, as U.S. companies are those that invest the most in domestic activities to develop new technologies through R&D, being far ahead of Japanese, German, French and British companies in this respect ( CHIARINI, 2017 ). U.S. private R&D abroad is continually increasing ( CHIARINI et al., 2020 ; LAURENS et al., 2015 ; NSB, 2016) . Investments in R&D in the host countries of majority-owned foreign affiliates of U.S. parent companies account for only about 9.0% of the total R&D performed in the U.S. (average for the 1997–2017 period). Figure 01 shows the increase trend of U.S. GERD and U.S. R&D invested abroad.

Figure 01 Source: The authors’ own. Data sourced from OECD and from the Bureau of Economic Analysis (https://www.bea.gov/).Note: Current USD values were deflated using the GDP deflator (year base 2015), data available from the World Bank.

In Latin America, FDI is characterized by an inverse statistical relationship with R&D investment (CRESPI, TACSIR, and VARGAS, 2016), which corroborates the finding of other studies conducted in the region that MNCs do not develop disruptive technologies locally ( CRESPI and ZUÑIGA, 2012 ; PAGÉS, 2010 ; RAFFO, LHUILLERY, and MIOTTI, 2008 ). Still in respect of the U.S. case, one can see that while Latin America receives a considerable amount of U.S. FDI investment (16.76% in 2017), it receives only a small amount for R&D activities (3.95% in 2017). On the other hand, the European Union receives the lion's share of both total U.S. FDI abroad (59.09%) and U.S. R&D captive offshoring (48.39% in 2017) ( Table 02 ).

Table 02

		U.S. direct investment position abroad on historical-cost basis (all industries)						R&D performed abroad by majority-owned foreign affiliates of U.S. parent companies (all industries)
		2000		2005	2010	2015	2017	2000	2005	2010	2015	2017
European Union		52,22		54,01	54,37	58,15	59,09	60.97	64.25	55.27	49.32	48.39
UK		17.53		15.68	13.40	11.96	12.43	20.10	19.55	14.51	11.01	11.33
France		3.24		2.70	2.09	1.54	1.42	7.16	8.13	5.07	3.98	3.55
Germany		4.22		4.48	2.76	2.28	2.26	15.23	16.67	16.84	15.63	14.45
Latin America		20.25		16.93	20.12	17.07	16.76	3.24	3.04	6.40	4.25	3.95
Argentina		1.33		0.45	0.31	0.28	0.25	0.19	0.08	0.29	0.27	0.23
Brazil		2.79		1.38	1.79	1.07	1.14	1.24	1.46	3.48	1.58	1.57
Chile		0.76		0.50	0.82	0.53	0.43	0.05	0.03	0.05	0.06	0.09
Mexico		2.99		3.29	2.29	1.92	1.82	1.48	0.88*	0.84	1.20	0.89
China		0.85		0.85	1.58	1.74	1.79	2.47	2.42	3.85	6.10	6.45
Japan		4.34		3.62	3.03	2.02	2.15	7.97	6.21	4.54	4.46	5.08
South Korea		0.68		0.88	0.70	0.73	0.69	0.70	1.65	2.07	1.76	1.79
Total**		1,766,550.28		2,687,005.00	4,079,778.52	5,289,071.00	5,837,334,21	27,455.58	33,146.81	43,488.52	56,096.00	54,941.47

Source: The authors’ own. Data sourced from the Bureau of Economic Analysis, Survey of U.S. Direct Investment Abroad (annual series).

Note: (*) value for 2007; (**) Absolute value is in millions of constant U.S. dollars. Current U.S. dollars were deflated by the GDP deflator (year base 2015) available from the World Bank database.

Domestic businesses in Latin America have evolved in a productive structure that provides few incentives to innovate. Many of them were able to grow and survive in sectors free of pressure to build innovative capabilities. The kind of R&D carried out by domestic firms is commonly oriented towards innovative processes and products for the domestic market rather than cutting-edge technologies. Their efforts to innovate tend to be even more tepid than those of MNCs ( ZUCOLOTO and CASSIOLATO, 2013 ).

Low demand for innovation reinforces the mismatch between the production of knowledge and technology in the region. Consequently, the increase in scientific output observed in the last decade (measured by the participation of Latin American countries in the worldwide publication of scientific papers) has not been accompanied by an increase in patent applications, which suggests that private sector demand has not evolved at the same pace as scientific supply ( DUTRÉNIT and ARZA, 2015 ). To illustrate this, we have depicted in Figure 02 the values for scientific and technical journal articles published, and patent applications registered, per million inhabitants for selected countries. It is interesting to note that Latin American countries are clustered at the bottom left of the graph, which shows that Latin American scientific papers and patent applications are scarce in comparison with those from more developed countries.

Figure 02 Source: The authors’ own. Data sourced from the World Bank and from USPTO.

Weak demand for higher-quality education ( DONER and SCHNEIDER, 2019 ), combined with scarce private expenditure on R&D call into question the capacity of business incumbents to invest in NSIs in Latin America. It is true that the direction of causality remains controversial. Is the lack of business investment in R&D attributable to the absence of institutional framing or is it the case that failure to invest in institution building can be put down to weak demand for R&D? In this article, we shall take a detour off the conventional path followed by NSI scholars – who tend towards the former explanation – and propose that the latter be explored.

Chang (2007) notes that in spite of a wide consensus on the idea of co-evolution between the ‘mode of production’ of a society and institutions, historical observation endorses the argument that it is very difficult to create new institutions through political means alone. The introduction of new economic activities that increase demand for different kinds of institutions has proved more efficient in driving institutional change. However, in periods of radical technological change (CHANG, 2007), it is natural to see institutional inertia serving as a roadblock as it protects the status quo.

The picture here resembles the argument put forward by Acemoglu, Johnson, and Robinson (2005), who point out that producers are the main ‘demanders’ of institutions to foster economic activity. Yet, at the same time that they demand institutions to protect their property rights, they erect barriers against the entry of new entrepreneurs, thereby blocking economic development ( ACEMOGLU, 2008 ). Once established, incumbent businesses prefer to maintain the barriers to entry, thereby inhibiting the access on the part of newcomers to skilled labor and technology so as to defend their incomes ( PRZEWORSKI, 2007 ).

What makes this line of argument tricky for Latin America is that incumbent businesses drive innovation in the region and should therefore demand the strengthening of the NSI. Moreover, considering their privileged bargaining position, it would not be difficult for BBs to unify their demands and coordinate lobbying in favor of a strong NSI ( FAIRFIELD, 2013 ; SANTOS, 2014 ). The argument to be corroborated is whether BBs in Latin America are active supporters or active blockers of strengthening the NSI.

Note: ‘Scientific and technical journal articles’ denotes scientific and engineering articles published in the following fields: physics, biology, chemistry, mathematics, clinical medicine, biomedical research, engineering and technology, as well as earth and space sciences.

In the following sections, we expand the framework proposed by Doner and Schneider (2019 , 2016 ) to shed light on the participation of BBs in upgrading coalitions in Brazil. Focusing on both public and private initiatives, we analyze the political obstacles that impede Brazil in its climb up the technology ladder.

Brazil industrialized rapidly in the period between 1930 and 1980. Briefly, we can say that Brazil has been able to build a relatively diversified and integrated productive structure based on the welcoming of MNCs and on the establishment of big public companies ( DAHLMAN and FRISCHTAK, 1990 ) such as in the oil and gas, metal-mechanical and chemical sectors. Nonetheless, there is a consensus that Brazil formulated no comprehensive policy strategy for the absorption of productive knowledge by domestic private businesses, and that emphasis was instead placed on the role of public enterprises and public research laboratories. The domestic private sector was not considered important to the innovation process and their efforts were limited to the adaptation of products and processes contingent on the importation of machinery and equipment, or the spill-over effects of state-owned or multinational enterprises ( KATZ, 2000 ).

Brazil has also achieved relatively vigorous scientific production vis-à-vis counterpart countries and has succeeded in establishing a considerable number of ST&I institutions with the enforcement of a reasonable intellectual property rights protection system ( MAZZUCATO and PENNA, 2016 ). However, the two actors – for-profit companies and universities – were always separate and had weak relationships. Instead, Brazil had so-called ‘islands of innovation’, especially in respect of relationships between publicly-owned laboratories and companies ( SUZIGAN and ALBUQUERQUE, 2011) .

In the last two decades, governments with divergent ideologies ⁴ have taken important steps to foster ST&I activities and spur innovation. During this time, a myriad of institutional arrangement such as National Acts, Decrees, Programs, Action Plans and Strategic Plans ( Table 03 ) have been put in place, while a constitutional amendment was promulgated that altered the text related to ST&I activities in order to highlight their role in innovation processes (85A). This resulted in the then ‘S&T Ministry’ (MCT) becoming the ‘ST&I Ministry’ (MCTI).

Table 03

Governments	Institutional Instruments	Publication year	Brief description
President Cardoso (1995-2002)	Law 9,478	1997	Creation of Oil and Natural Gas Sector Fund (CT-Petro)
	Law 9,991	2000	Creation of Energy Sector Fund (CT-Energ)
	Law 9,992	2000	Creation of Transportation Sector Fund (CT-Transp)
	Law 9,993	2000	Creation of Hydro Resources Sector Fund (CT-Hidro) Creation of Mineral Resources Sector Fund (CT-Mineral)
	Law 9,994	2000	Creation of Aerospace Sector Fund (CT-Espacial)
	Law 10,052	2000	Creation of Telecommunications Sector Fund (FUNTTEL)
	Law 10,168	2000	Creation of University–Industry Collaboration Fund (Fundo Verde Amarelo)
	Law 10,197	2001	Creation of R&D Infrastructure Fund (CT-Infra)
	Law 10,176	2001	Creation of Information Technology Sector Fund (CT-Info) Creation of Amazon Sector Fund (CT-Amazônia)
	Law 10,260	2001	Creation of Higher Education Student Fund (FIES)
	Law 10,332	2001	Creation of Biotechnology Sector Fund (CT-Bio) Creation of Agribusiness Sector Fund (CT-Agro) Creation of Health Sector Fund (CT-Saúde) Creation of Aeronautics Sector Fund (CT-Aeronáutico)
	Decree 4,078	2002	Creation of the Center for Management and Strategic Studies (CGEE)
President Lula (2003-2010)	Law 10,973	2004	Innovation Act (Lei da Inovação)
	Law 11,077	2004	New Computing Act (Nova Lei da Informática)
	Law 11,080	2004	Creation of the Brazilian Agency for Industrial Development (ABDI) Creation of the National Industrial Development Council (CNDI)
	Law 11,096	2005	Creation of University for All Program (PROUNI)
	Law 11,105	2005	Biosecurity Act (Lei da Biossegurança)
	Law 11,196	2005	Good Act (Lei do Bem)
	Law 11,437	2006	Creation of Audiovisual Sector Fund (FSA)
	Law 11,487	2007	New incentive for technological innovation and new rules to speed up amortization of R&D investments
	Action Plan	2007	Action Plan for ST&I (PACTI) 2007-2010
	Decree 6,041	2007	Creation of the Biotechnology Development Program
	Decree 6,096	2007	Creation of Federal University Restructuring & Expansion Program (PROUNI)
	Decree 6,259	2007	Creation of the Brazilian Technology System (SIBRATEC)
	Law 11,892	2008	Creation of Federal Institutes of Education, Science and Technology (IFETs)
President Rousseff (2011-2016)	541PM	2011	'S&T Ministry' became 'ST&I Ministry'
	Decree 7,642	2011	Creation of Science Without Borders Program (Ciência sem Fronteira)
	Action Plan	2011	Creation of Innovate Company (Inova Empresa)
	Decree 7,819	2012	Creation of Inovar-Auto Program to support innovation for automotive sector
	Strategic Plan	2012	National Strategy for ST&I (ENCTI) 2012-2015
	Decree 02.09	2013	Creation of the Brazilian Enterprise for Research and Industrial Innovation (Embrapii)
	Decree 8,269	2014	Creation of the National Knowledge Platform Program
	Law 13,123	2015	Biodiversity Act (Lei da Biodiversidade)
	85A	2015	Constitutional amendment which updates the treatment of ST&I activities
	694PM	2015	Reduction of fiscal incentives proposed by Good Act
	Strategic Plan	2015	National Strategy for ST&I (ENCTI) 2015-2019
	Decree 8,637	2016	Increases the production rate of Brazilian suppliers of oil and gas sector through technological innovation
	Law 13,243	2016	Legal Framework for ST&I Act (Marco Legal da Inovação)
President Temer (2016-2018)	95A	2016	Constitutional amendment capping public spending for the next two decades
	Decree 8.898	2016	Regulates the ST&I National Council that advises the President on the formulation and implementation of ST&I policies
President Bolsonaro (2019-?)	-	2019	Entrepreneurial and Innovative Institutes and Universities Program (Future-se)

Source: Authors’ own.

Probably one of the main contributions of the last decades in respect of ST&I policies was the design of a new funding mechanism based on sectoral funds. These funds were conceived to answer industry needs for innovation-oriented research ( PACHECO, 2019 ; PELAEZ et al., 2017 ; SÁ, 2005) and promote financial resource stability in the Brazilian Innovation System ( PEREIRA, 2005 ; SZAPIRO, VARGAS, and CASSIOLATO, 2016 ). In total, sixteen funds were formed: there were thirteen thematic funds – oil and natural gas; energy; transportation; hydro resources; mineral resources; aerospace; telecommunications; information technology; biotechnology; agribusiness; health; audiovisual; and aeronautics – and three horizontal funds – the university-industry collaboration fund; the Amazon fund; and the infrastructure fund.

Other institutional advancements include the creation of a handful of legal instruments to deal with different aspects of the Brazilian Innovation System: to facilitate university-industry interactions; to reduce bureaucratic procedures that hamper public-private cooperation focused on R&D activities; to provide legal support and set incentives for the commercialization of the results of scientific and technological research developed in public universities and research institutes; and to allow tax exemptions on corporate investments in R&D – the so called ‘Innovation Act’, the ‘Lei do Bem’ and the ‘New Computing Act’.

Moreover, there was a great effort to boost the performance of such public financing institutions as the National Bank of Economic and Social Development ⁵ , the Finance Agency of Studies and Projects ⁶ , ( BASTOS, 2012 ) national ST&I funding agencies, and the Brazilian Enterprise for Research and Industrial Innovation ⁷ . Moreover, other agencies were created to provide better coordination between pro-innovation initiatives and provide interaction with other institutions ( STEIN and HERRLEIN-JR., 2016 ; TONI, 2007) , such as the establishment of the Brazilian Industrial Development Agency ⁸ whose aim was to promote the implementation of industrial policies, especially those contributing to job creation, in line with the foreign trade and ST&I policies and the National Industrial Development Council ⁹ with a view to at advising the president on specific policies and other measures to promote industrial development in Brazil (Law 11,080 of December 30, 2004; Table 03 ).

It is also interesting to note that national ST&I strategies were also put in place during this period. The 2015 strategy publication (ENCTI 2015–2019), acknowledged that a determining factor for the success of nations in fostering innovation was the creation of a strong and articulated NSI capable of mobilizing actors and instruments in a consistent and result-oriented manner that would generate socio-economic benefits.

Despite there being no doubt Brazil has built a diversified set of ST&I policies and programs, it faces two paramount challenges that are of concern to us. First, there is a lack of articulation and coordination between actors and instruments within the Brazilian government. Studies have shown that Brazil lacks a clear coordination between such instruments and other government initiatives within government itself ( NEGRI and RAUEN, 2018 ; SUZIGAN, GARCIA, and FEITOSA, 2020 ; ZUÑIGA et al., 2016) .

These coordination failures across innovation policies were identified in a report by the Federal Court of Accounts ¹⁰ , which was one of the first such publications on the role of Brazilian public agencies and entities in the context of the NSI and whose subject was not a specific program, but the innovation agenda ( TCU, 2018 ). According to the report, Brazil lacks an active government coordination structure for pro-innovation policies. Pro-innovation programs and action plans are spread throughout diverse ministries due to their transversal character and there is no clear definition of how they should be coordinated with national policies. The coordination responsibilities of federal public policies for innovation are not clearly defined between the ST&I Ministry and other ministries in a whole-of-government approach ( TCU, 2018 ).

Second, the Brazilian ST&I agenda is unstable. To highlight the instability of the ST&I agenda, some scholars ( PACHECO, 2019 ; PELAEZ et al., 2017) have proposed looking at the terms of office of ST&I Ministers and Industry and Commerce Ministers, who are responsible for proposing policies in this field ( Figure 03 ). They show that the irregularity of ministerial terms of office goes a long way to explain recurrent discontinuities in industrial and ST&I policies.

Figure 03 Source: Authors’ own, based on Pelaez et al. (2017) .

The churn rates for ST&I Ministers and Industry and Ministers indicate that innovation policies were also subject to the pressures of Brazilian 'coalitional presidentialism', which reinforces fragmentation of the ST&I agenda ( LIMOEIRO and SCHNEIDER, 2017 ). It is possible to see from Figure 03 that ST&I Ministry churn reached its peak during the Rousseff administration when a political crisis dominated in Brazil and the political command system crumbled. As correctly stated by Pelaez, this means that "long-term political agendas are rarely set, while it also holds back the formation of skills needed to balance all the different interests at stake"( PELAEZ et al., 2017, p , p. 801). As a consequence the ST&I Ministry is a weak ministry with poor coordination capacity ( PACHECO, 2019) .

Still looking at the ST&I agenda instability, we now turn our gaze to the Annual Budget Law (Lei Orçamentária Anual ( (LOA) ¹¹ ) for key ST&I activities: 01. the 'Higher Education Program' of the Office for the Improvement of Higher Level Personnel ¹² (Capes); 02. the 'ST&I Program' of the National Council for Scientific and Technological Development ¹³ (CNPq); and 03. the Sectoral Funds ¹⁴ (FNDCT).

According to Pelaez el al. (2017), the Brazilian dependency on macroeconomic stabilization goals has perpetuated the government’s habit of cutting back Sectoral Funds (FNDCT) and prioritizing austerity over long-term industrial policy. Brazil suffers from a lack of coordination between long-term policy and its macroeconomic regime ( NASSIF, BRESSER-PEREIRA, and FEIJO, 2018 ). In fact, it is not only the Sectoral Funds that suffer as a result of this habit, but also other ST&I program budgets ( Figure 04 ).

Figure 04

Note: Current BRL values were first deflated using the GDP deflator (year base 2018) and then converted into USD using the official exchange rate (year average); both sets of data were available from the World Bank database.

It is interesting to note that historically most innovation efforts in Brazil have not been performed by private big businesses. Indeed, by breaking down gross domestic expenditure on R&D (GERD ¹⁵ ) by fund source, we can see that the share financed by the public sector accounted for more than 50% in the last decades and reached its peak in 2013, when it accounted for 57.7% of total GERD.

The diversified set of ST&I policies, programs and action plans summarized in Table 03 in the previous section had a certain impact on the increase of the total GERD, however, this was mostly a result of the government’s own expenditure on R&D activities. Private businesses maintained their historically low level of R&D intensity, which was not sufficient to increase their R&D participation in total GERD. Looking at Figure 05 , we can see that, while total GERD increased in absolute terms from 2000 to 2017 – especially during the Lula and Rousseff administrations – the GERD share financed by private businesses accounted for 47.16% of the total in 2000, 47.46% in 2017

Figure 05

Note: The current USD figure was deflated by the GDP deflator (year base 2017) available from the World Bank database. ‘Others’ refer to higher education sector and other non-profit and foreign institutions.

We have addressed some institutional and political dynamics involving Brazilian BBs and the State. Our point of departure was the following question: how are the innovative practices and political behavior adopted by BBs able to shape the Brazilian NSI? The assumption behind our question was that the way in which businesses are organized in each country is necessary to understand the political obstacles to its economic upgrading.

Brazil offers an intriguing scenario in which to examine the interactions between BBs and the State in the consolidation of the NSI. On the one hand, the country has developed an active private sector, with domestic BBs and MNCs operating in the most important economic sectors. On the other, the Government has increased its participation in innovation efforts, especially during the 2000s. However, the private share of the GDP in R&D has not increased, and the rate of innovation of BBs has decreased.

Our claim is that there are mismatches within the public sector (exacerbated by the lack of coordination between actors and instruments within the public sector itself and by the instability of the ST&I agenda) and there are also mismatches within private sector. We believe that both the Brazilian government and businesses are fragmented and failed to form a strong coalition with a strong interest in fostering innovation as a matter of State with a long-term perspective. As a result, Brazil faces a low-innovation trap, by which we mean a situation in which the interactions between the most important actors responsible for carry out innovation do not enhance the institutions necessary to foster innovation in the rest of the economy. The reasoning behind this argument lies in the way in which BBs are organized in Brazil, which seems to stymie the formation of political coalitions that could push for policies to strengthen the NSI.

In line with our argument, the ongoing lack of coordination of the main institutions of the Brazilian NSI may be a sign that the big players in the system do not care about building institutions that could provide public goods – otherwise, they would be the first ones to raise their voices against the retreat of the State from the front. This suggests that the efforts of the State to stimulate innovation have not created the right incentives for enticing BBs into supporting the NSI. Those that do not depend on public instruments to innovate, such as MNCs and some domestic BBs, benefit from the existing resources, such as public funding, human capital and cooperation networks, without contributing to their provision. In this sense, by providing goods unconditionally, the Brazilian State becomes at the same time responsible for and a victim of the low-innovation trap. This situation is reinforced in a negative cycle especially when there is an economic downturn like the one Brazil has been suffering since 2013, which has caused companies to reduce their innovative efforts.

Rather than suggesting actions to escape from the trap, our intention is a distinctly modest one: to call attention to the fact that the strategies and practices adopted by BBs in Brazil are creating a trap for the NSI. Much of the debate on policy prescriptions and institution building in the NSI will be fruitless if the structural and political foundations that block upgrading changes are not fully understood.

Moving this agenda forwards implies a change of focus from institutional arrangements to institutional origins, i.e., recognition that BBs have their own interests that may not be congruent with the upgrading of institutions. The exit from the trap may be found in developing new patterns of interaction among domestic businesses, MNCs and the State. This essay was our first attempt in that direction.