Market co-creating and shaping through investments in innovation: a comparative analysis of two public funding programmes in Brazil

ABSTRACT This paper documents the incipient emergence of an active role of the state in Brazil that has been oriented towards creating and shaping markets through investments in innovation. It draws on a comparative analysis of two case studies of funding programmes: the BNDES–FINEP Joint Programme for Supporting Industrial Technological Innovation in the Sugar-based Economy and Sugar-Chemicals Sectors (PAISS), and the Ministry of Health’s Production Development Partnerships Programme (PDPP). In the study, evidence was found within both programmes of five dimensions of effective public–private partnerships, these being that public agencies (i) seized mapped opportunities; (ii) took the lead; (iii) engaged in risk-sharing and institutional building; (iv) pursued risk diversification and competition; and (v) sought an equitable sharing of rewards. After a discussion of the main policy lessons, the paper concludes with a reflection on the specific challenges for building equitable public–private partnerships in the light of the increasingly competitive and global scope of the economy.


Introduction
Governments around the world are confronted with new challenges, in particular having to reconcile economic growth, sustainability and inclusion.In this context, and markedly since the 2008 financial crisis, they have increasingly turned to industrial and innovation policies in response.Historically these policies have been the subject of heated discussion about the positive and negative aspects of the active role of the state in the economy.Although there are different views in this debate, an increased focus on how to design and implement policies for solving societal problems of the twenty-first century can be observed (see, for example, Stiglitz 2016;Cimoli et al. 2017;Andreoni and Chang 2019;Aiginger and Rodrik 2020).
One of the practical ways being sought to steer the directions of innovation and economic growth has been through mission-oriented policies (Larrue 2021;MOISS 2019;EC 2018).In this regard, public institutionsespecially development banksare encouraged to play an entrepreneurial role and turn to the possibility of fruitful collaboration with the private sector (Mazzucato 2013).However, it is acknowledged that there can also be problems if only the risks of investments are socialized, while the rewards are privatized (Mazzucato 2013;Lazonick andMazzucato 2013, 2020).Analyses of the implementation of policy tools that enable an adequate risk-reward nexus in concrete cases are, nonetheless, still missing.
So far, studies on mission-oriented policies have mainly focused on developed countries.However, there is an emerging interest in learning from the incipient initiatives implemented in countries of the Global South, in particular in Latin America (Mazzucato andPenna 2016a, 2019;Lavarello et al. 2020;Dudrénit et al. 2021).Empirical analyses of these experiences are important to inform future innovation policies capable of addressing the specific development problems facing the region.Moreover, compared with the advanced initiatives carried out in the United States of America or Europe, the economic, political and institutional challengesas well as the conflicts over distributionappear more clearly in emerging and developing countries.These challenges, in turn, offer opportunities for developing further theoretical and practical insights into the design, implementation and assessment of public-private partnerships for investments in innovation, and on the issue of the risk-reward nexus.
This paper is an attempt to contribute to the above line of study.It aims to analyse, in real research and development (R&D) programmes, some of the key issues that decision-makers and practitioners will need to address in order to make public-private partnerships more mission-oriented and effective.To this end, a comparative analysis has been conducted of two federal R&D programmes in Brazil (2003Brazil ( -2016)), an emerging economy that has recently experienced a brief revival of active and explicit industrial policies.The two case studies are: (i) the BNDES-FINEP Joint Programme for Supporting Industrial Technological Innovation in the Sugar-based Economy and Sugar-Chemicals Sectors (PAISS); and (ii) the Ministry of Health's Production Development Partnerships Programme (PDPP).These programmes pioneered features of mission-orientation in Brazil and were considered symbolic of virtuous public-private collaboration by the local science, technology and innovation (STI) community in terms of the quality of resources mobilized and arrangements concerning reward distribution (Mazzucato and Penna 2016a).
The study draws on a literature review and interviews with multiple stakeholders.In the analysis, the following are considered: (i) the objectives of the programmes; (ii) the role of public funding agencies; (iii) the legal and institutional changes enabling public-private partnerships; (iv) the measures to increase and manage the risks taken; and (v) the measures adopted so the state can benefit from an equitable sharing of rewards in case of success.This analysis points to important lessons that can serve as guidance for other government initiatives in the Global South and elsewhere, in particular regarding the legitimation and effective mobilization of entrepreneurial public finance.The paper concludes with a reflection on both the challenges for building effective public-private partnerships in the face of a global economy, and on the implications for future research.

The market co-creating and shaping policy framework
The concerns with contemporary challenges such as slow economic growth, rising inequality and climate change, have opened a new agenda for innovation policy studies.Issues like the alignment, coordination and institutional settings that enable innovation to be steered towards the desired directions have received increased attention (Schot and Steinmueller 2018) and this is an area also relevant for industrial policy (Stiglitz 2016;Cimoli et al. 2017;Andreoni and Chang 2019;Aiginger and Rodrik 2020).Accompanying these developments, there have been critical analyses of both the theoretical position that is focused on the correction of market failures and the rationale for policy based on systems of innovation.Some critiques have led to proposals for alternative approaches (e.g.Mowery, Nelson, and Martin 2010;Kuhlmann and Rip 2018;Mazzucato 2018;Schot and Steinmueller 2018).
In this context where interest in active mission-oriented policies resurfaces, Mazzucato (2016) has argued for a new framework in which the justification for the role of the state in innovation is concerned with the creation and shaping of markets.This type of policy is based on a view of the innovation process as a systemic, rather than linear, phenomenon.As such, it recognizes that innovation crucially depends on the quality of a complex network of relationships between a variety of actors including firms, research institutions, financing bodies, regulators, users and others (Lundvall 1992;Freeman 1995).From this perspective, policymakers need to take into account the different factors that influence innovation processes in each context, and they need to ensure that relevant policies or programmes are systemic in their aspiration of stimulating the interactions that govern the creation and diffusion of knowledge within entire innovation systems.
This implies the use of a wide range of supply-and demand-side instruments, implemented in a consistent and mutually supportive way by different organizations.In mission-oriented policies, public support is aimed at mobilizing STI to address societal challenges.The complexity of the problems to be solved requires more sophisticated policies and institutional settings, including enhanced alignment and complementarity between the support instruments, as well as a higher degree of coordination between government ministries and agencies (Larrue 2021).Thus, greater emphasis is given to the importance of different types of actors in the public sector.
Accordingly, Mazzucato (2013) has put forward the notion that the 'entrepreneurial state' should become the cornerstone of an argument that governments often play a more active role as investors in innovationwhich is very different from maintaining the overall assumption that governments only fix failures.Entrepreneurial states are driven by a clear purpose and vision of economic and technological opportunities.If public finance arrives before technologies are mature enough for a business to be willing or able to invest, public agencies, especially development banks, feature as lead risk-takers, rather than mere de-riskers (Mazzucato 2013;Mazzucato and Penna 2016b).The set of implementing agencies operate through a commonly hidden, decentralized, coordinated and networked governance structure, rather than one that is centralized and top-down (Block 2008;Block and Keller 2011).History shows that this course of actionwhich goes well beyond funding basic science to reach the entire innovation chain, including the deployment and diffusion phases through supply-and-demand-side policiesenables the strategic mobilization of different types of public and private finance (Mowery 2009;Foray, Mowery, and Nelson 2012).
This perspective in which direct public funding increasingly moves downstream means that public agencies work in close collaboration with business and become more exposed to failures, particularly given the inherent uncertainties of the innovation process.The question then arises whether, in the exceptional cases of success, the nexus will be retained between those who take the risks of investing and those who will reap the rewards (Mazzucato 2013;Lazonick and Mazzucato 2013).Laplane and Mazzucato (2020) identified two sets of measures that could help public agencies act strategically upon the distribution of rewards.These are profit-sharing (involving financial gains to the state) and conditionalities (for more significant economic, social and environmental benefits).Effective implementation, however, depend on preconditions that may be present (or absent) to different degrees in each context (Laplane and Mazzucato 2020).
The vulnerabilities of the South in their productive, financial, social, institutional and political spheres have long been emphasized in the Latin American literature on development and innovation, as discussed by Lastres and Cassiolato (2017).There is growing interest in the region's incipient experience with mission-oriented initiatives being used to derive policies that could be applied to national development problems (Gadelha 2016;Mazzucato and Penna 2019;Lavarello et al. 2020;Dudrénit et al. 2021).This emerging literature highlights the importance of public sector leadershipwhich is consistent with the notion of the entrepreneurial state (Mazzucato 2013)and the need to overcome some constraints in institutional capacities.While countries of the South with weaker states may face greater challenges in accumulating public sector capabilities, these countries could benefit the most from policy learning that enables a balanced risk-reward nexus (Laplane and Mazzucato 2020).Hence, in-depth empirical research focused on these countries becomes even more relevant, providing opportunities for further refinement of the market co-creating and shaping framework.
To sum up, the aspects highlighted so far point to the theoretical possibility of building virtuous public-private collaborations based upon the risk-reward relationship.However, success will depend on a sequential path of action being taken in policy design and implementation; identifying this path provides the basic building blocks for an initial analytical framework to study concrete R&D programmes (Figure 1).
The first step along the path is the identification of clear technological opportunities related to overcoming societal challenges toward which public agencies could seek to direct resources.In the second step, public actors are required to mobilize resources, taking and sharing risks to attract private investment; public funds must target sectors, firms and projects that contribute to the development of socially desirable innovations.But since public risk-taking is not something that occurs naturally or spontaneously, it compels public agencies to engage in legal and institutional experimentation and change (the third step), and to move towards a portfolio approach to manage the risks taken (the fourth step).Moving to the final (fifth) step entails devising measures for ensuring that public agencies are able to recover a portion of the amounts invested, in the event of success, and which may allow a balanced risk-reward relationship in public-private partnerships.If the amounts recovered were used to replenish public funds, the investment process could be repeated.Naturally, this is a simplification for analytical use, as in reality, the process is often incomplete, non-linear and less clear-cut.
Following the same five-step logic described in the analytical framework, five main guiding questions for the case studies were derived: (i) What motivations and aims lie behind the programmes?(ii) Who takes the lead, the private or public sector?(iii) What legal and institutional changes enable implementation?(iv) How do federal funding agencies seek to increase and manage the risks taken?(v) How is an equitable sharing of rewards pursued?

Methodology 1
To investigate these five questions, a comparative case study of two R&D programmes in the biofuels and health sectors (covering the period between 2008 and 2016) was adopted.These programmes were unique in adopting a challenge-oriented approach to innovation policies and the strategic use of different instruments, in a period of time that followed the Brazilian government's stated goal of enhancing investments in innovation and improving the underlying legal framework (more details are given in Section 4).As such, they were well suited for studying public-private investment partnerships and how the risk-reward nexus could be achieved.By following a comparative approach, it was possible to gain insights into the commonalities, differences and limitations in how public actors engaged with business and responded to the need of risk-and reward-sharing.
A variety of data sources was used for the analysis.First, a review of the literature and documentary material on public agencies' activities, policy reports, and scientific and popular journals, provided an initial account of the goals and characteristics of the programmes, actors involved, challenges for implementation, and preliminary outcomes.Second, interviews with the key stakeholders allowed for a more nuanced understanding of their experiences with or perceptions about innovation policies, while enabling exploration of the notion of a risk-reward nexus.
Fieldwork took place between May and August 2016 and comprised in-depth semistructured interviews with 51 subjects, representing different types of actors and organizations in the public and private sectors.Given the research interest in identifying critical issues for successful policy implementation, priority was given to respondents in the public sector (41 in total).The interviews followed a standardized protocol in accordance with ethical guidelines, and this research was approved by the Sciences & Technology Cross-Schools Research Ethics Committee (C-REC) at the University of Sussex, under reference number ER/AL445/1.A deductive approach was then adopted, in which the interviews were analysed and coded (Kvale 1996) in the five steps outlined in the operational framework set out in Figure 1.Based on this framework, and seeking to answer the five guiding questions posed in section 2, the empirical analysis in section 5 has been structured around five themes.
Before proceeding to the case studies, it is important to note the political, economic and institutional developments that allowed for active innovation policies to emerge in Brazil.

The recent evolution of innovation policy in Brazil
The evolution of innovation policy in Brazil is still recent, which is a reflection of the country's late industrialization and the continuous changes in the role of the state in the economy.It was not until the late 1990s that the Ministry of Science and Technology (MCTI) 2 took incipient steps to restore and sustain STI fundingthrough the National Science and Technology Development Fund (FNDCT)and propose legislation to allow public-private partnerships (the Innovation Law).However, since those initiatives took place amid neoliberal reforms and a conservative macroeconomic administration, it has not proved possible to achieve the synergies needed, within and across the government, to integrate innovation into actual policy.
The election of President Lula da Silva, in 2002, inaugurated a gradual shift towards a more favourable political, legal and economic landscape for advancing the role of the state through policy experimentation.In May 2003, the government affirmed its commitment to long-term, environmentally sustainable, and inclusive growth, recognizing the importance of the state as a driving force (MPOG 2003).The revival of industrial policies would then place innovation at the centre of the development strategy for the first time, evolving through three successive plans (MDIC 2003(MDIC , 2008(MDIC , 2011)).
To this end, the legal and institutional frameworks had to undergo change, part of which was the passing into law of the Innovation bill in 2004 (Law 10973/04), enabling the government to offer direct financial support to the business sector.This laid the foundations for the Brazilian Innovation Agency (FINEP) to start a Subvention Programme in 2006.The legislation also permitted the use of other policy instruments, including on public procurement.Enacted in 2007, the Law of Good (Law 11196/05) similarly regulated tax incentives for R&D.There were also institutional changes at the level of public funding agencies themselves.For instance, under new directorship, the stateowned development bank (BNDES) gradually resumed its role as the chief financier of development strategy.New operational norms ensued, which prioritized support to investments in innovation with longer-term horizons, broader access, subsided credit, and the waiving of risk premium, thus signalling the bank's willingness to take risks (Lastres et al. 2017).Furthermore, in 2007, BNDES intensified its operations in venture capital, indicating the rise of public institutions acting as venture capitalists (Schapiro 2012).
Despite these changes, the initial move towards increasing investment in innovation was modest, in part because, up to that point, the political and legal environment had been extremely inflexible about the use of public funds directed to R&D.Moreover, as private finance for innovation had historically been missing (Melo and Rapini 2014), the lack of public funding meant that business remained risk-averse.The enactment of the Innovation Law and other pieces of legislation had only begun to change actors' practices and mindsets.Nevertheless, effective implementation would require learning, further adjustment and time.Another critical factor was the continuity of a misaligned macroeconomic policy approach.As the government maintained high interest rates while withholding public funds, the stifling of private investments was not surprising.
President Lula's second term (2008)(2009)(2010)(2011) opened a period of progressive reorientation towards an active role for public investment.Following the eruption of the financial crisis, countercyclical measures provided BNDES and FINEP with additional resources.From 2010 onwards, the Treasury also (temporarily) ceased the systematic retention of FNDCT funds.In that propitious context, the Government Blue Book stated the need to break new ground in terms of the legal framework, and quantity and quality of public support to innovation (MCT/CGEE 2010).
Over the next few years, federal public funding achieved new breadth; one can also observe a correlation between it and growth in business expenditure in R&D (Figures 2 and 3).Led by executives invested with autonomy and resources, and committed to encouraging risk-taking, public agencies experienced a thriving period of experimentation and learning.Meanwhile, advances in the coordination mechanisms within government, and between it and the private sector, ensued, culminating in the growing ambition of the national STI strategy (MCTI 2012).
Notwithstanding such positive developments, which contributed to consolidating a more complex innovation system in Brazil, various analysts have recognized that public support was (and still is) limited on several fronts (Bastos 2012;De Negri 2017).Limitations include the lack of articulation of innovation policy with other policies and between supporting institutions and organizations (Cassiolato, Szapiro, and Lastres 2015), as well as the fragmentation and dispersion of policies and resources (Gadelha 2016;De Negri 2017).Ultimately, a linear view of innovation still prevailed (Szapiro, Vargas, and Cassiolato 2016).In this context, again the mismatched macroeconomic policy, and the worsening of international conditions, gradually led to a major political, institutional and economic crossroads, interrupting Brazil's promising developmental path.
The two programmes selected as case studies represent clear departures from the previous status quo, whereby neither the public nor the private sectors took the high risks of investing in R&D.In addition to being relatively successful in stimulating risk-taking, the programmes were considered positive examples of public-private partnerships by Brazilian experts and decision-makers (Mazzucato and Penna 2016a).In this sense, the programmes paralleled each other sufficiently to merit study using a compare-and-contrast approach.

Two case studies on the market co-creating and shaping role of the state
In this section, two programmes are described that were initiated by the federal government to encourage R&D and innovation: the BNDES-FINEP Joint Programme for Supporting Industrial Technological Innovation in the Sugar-based Economy and Sugar-Chemicals Sectors (PAISS) and the Ministry of Health's Public-Private Production Development Partnerships (PPDP).PAISS is a supply-side stimulus aimed at tackling the key challenges that advanced technologies impose in order to maintain the country's world leadership in sugarcane-based biofuels.It offers direct financial support to business, structured across the whole innovation cycle.PPDP is a demand-side policy geared towards building capabilities and catching up mainly in biotechnologies applied to pharmaceuticals.In coordination with supply-side funding, it deploys public procurement to support partnerships for the local development, transfer and absorption of strategic health technologies.
Both programmes represent a positive turn in the historical record of policies oriented towards the sectors in question.The Brazilian pioneering production and use of vehicular biofuels based on sugarcane that dates back to the 1970s have benefited from active government support primarily through the mandatory blending of ethanol in gasoline - especially under the National Alcohol Programme created in 1975 (PROALCOOL).Conversely, the Brazilian pharmaceutical sector, which experienced a setback following the sudden liberalization of markets in the 1990s, has enjoyed the government's proactive engagement in rebuilding the industry.Key measures were the adoption of generic drugs legislation (1999) and the supply of subsidized credit under the BNDES Programme to Support the Development of the Pharmaceutical Industrial Chain (BNDES Profarma 3 ) created in 2004.Therefore, it is only after these cumulative efforts in biofuels and pharmaceuticals that PAISS and PDPP could, to varied extents, make progress towards a more systemic, purposefully driven view of innovation.
Below, the analysis of the programmes has been presented under the five categories that enabled the five guiding questions to be addressed: (i) opportunity-driven risktaking; (ii) public sector leadership; (iii) incremental legal and institutional innovations; (iv) new financing instruments, risk diversification and competition; and (v) profitsharing and conditionalities balancing the risks and rewards of investing.These categories correspond to the dimensions of effective public-private partnerships illustrated in Figure 1, which is an attempt to chart a positive path for the design and implementation of R&D programmes.

Opportunity-driven risk-taking
Both programmes aimed to stimulate high-risk private investments in R&D and innovation, following the perception of new opportunities.In the case of PAISSthe joint initiative by BNDES and FINEP launched in 2011new technologies for the processing of sugar-based biomass brought the potential for breaking new ground in productivity gains for the biofuel agroindustry.The frontier has been moving from traditional fermentation techniques (first-generation biofuels) to more efficient models based on biochemical or thermochemical processesthe second-generation (E2G) or advanced biofuels.Thus, there were advantages regarding environmental sustainability and energy efficiency, in particular because these new processes represented a promising alternative to fossil fuels.In turn, the Ministry of Health's (MS) PPDP released in 2009, came in response to the awareness that biotechnology could help to reposition the national pharmaceutical industry in the world.The imminent expiry of patents (held by global players) for drugs that represented a heavy burden on the public healthcare system's (SUS) budget opened a window of opportunity for steering private investment in directions that ensured access to medicines.Besides potentially having high spillover effects, the programme represented a systemic view in which innovation was integrated into economic and social development perspectives (Gadelha 2003(Gadelha , 2009)).Therefore, PAISS and PDPP were both oriented towards seizing the opportunities that increased risk-taking could offer, including high economic, environmental and social benefits.
Favourable supply-and demand conditions were in place that suggested that Brazil could succeed if the government acted.PAISS qualified, having natural resources, a competitive industry, and potential demand due to the prevalence of vehicles that could operate on different types of fuel ('flex-fuel vehicles') (Nyko et al. 2010).Scientific and technological capabilities in ethanol, 4 and public funding sources of sufficient scale to support internationalization strategies (BNDES), were also essential.PDPP had a mature generic drugs industry and the demand of one of the largest healthcare systems in the world.Public laboratories with some experience of biotechnology, a science base and growing STI infrastructure were also relevant.Hence, in both cases, the combination of supply-and demand-side factors, STI base and public finance constituted drivers of change in the ambitions of policymakers.
However, the challenges were enormous.In PAISS, adopting technologies under test required investors to bear inevitable failures.A viable cost of production has yet to be reached, and the high productivity levels of the first-generation biofuels set a benchmark that was difficult to overcome.There was also the risk of international players arriving first and foreclosing the window of opportunity.Similarly, PDPP faced vulnerabilities such as a poorly integrated supply chainincluding small and fragmented biotech firms.Given the high technological complexity of the products targeted for development through technology transfers, the risks were high and dependent on the accumulation of capabilities and infrastructure.Moreover, as public demand alone could not ensure a viable rate of return for supported companies, market risks remained high.Thus these two areas of industry had to deal with technological challenges that required longterm, high-risk and capital-intensive investments that the private sector could not meet on its own.PAISS and PDPP appeared in response, as possibilities for publicprivate partnerships to be governed in socially desirable directions.

Public sector leadership
The leading role of state actors in mapping the technological, economic and institutional challenges, engaging with business, and devising solutions was vital.In the case of PAISS, public officials from BNDES started investigating the reasons behind the stagnation of the Brazilian ethanol agroindustry early in 2010.Back then, confronted with a complex R&D project, BNDES and FINEP were grappling with the problem of improving coordination and the structure of funding.This situation prompted ideas on how public support could be best employed (see, Nyko et al. 2010). 5Early communications with business, signalling a willingness to share the risks of investments, would also prove strategic: We noticed that whoever managed to solve these technological bottlenecks would establish this industry … we offered financial support with a view on this future … Once we had a diagnosis, we devised a strategy and engaged with the business sector.(Interviewee 9, PublMa) In turn, PDPP stemmed from the gradual alignment of industrial, STI and health policy agendas.In 2004, the formulation of a National Policy for Science, Technology and Innovation in Health (MS 2008) was a milestone in building a shared vision among stakeholders in the public and private sectors and organized civil society. 6While public officials operating the BNDES Profarma engaged in diagnosis and prognosis (see Reis et al. 2009Reis et al. , 2010Reis et al. , 2011)), the MS was taking the first steps towards using strategic procurement.The gradually closer relationship between them (BNDES and MS) facilitated coordination.Therefore, the proactive and future-looking attitudes of public officials in decentralized agencies triggered PAISS, and both bottom-up and top-down pressures contributed to PDPP.Indeed, the public sector's leadership in PDPP went far beyond stimulating private investment, as BNDES was playing an active role in seeking collaboration between existing national players to obtain a more competitive structure.Moreover, the successful experience of engaging a public laboratory in technology transfer from a pharmaceutical company to local production in 2007, 7 offered a template upon which to structure the procurement programme.Hence, in addition to the public sector's leadership in terms of financing and coordination, one can observe a process of growing confidence to deal with business: What is it that allows us to be so bold?It's that Brazil has a large market.With … 200 million people, you can ask for certain responses.(Interviewee 33, PublMa) The mobilizing feature of the state's risk-taking is apparent.PAISS made it possible to increase the scale of public support and combine the financing instruments of BNDES and FINEP which until then had been operated independently.Before PAISS (2010), public investment amounted to R$ 415 million ($ 250 million 8 ); from 2011 to 2015, it reached R$ 5 billion ($ 2.3 billion) (Nyko et al. 2013).FINEP non-refundable grants targeted the early stages of R&D, whereas BNDES credit and equity were key for reaching the higher-risk and capital-intensive stages of industrial plants.Likewise, federal investment in health R&D gradually moved from R$ 600 million ($ 433 million) in 2004 to R$ 1.7 billion ($ 907 million) in 2013. 9The coordinated efforts within the government enabled a systemic policy approach to PDPP, having public actors who would fund R&D directly, perform it, or manufacture/buy the resulting products.

Incremental legal and institutional innovations
The move towards public-private partnerships was facilitated by incremental legal and institutional transformations.PAISS's approach to attracting private investment involved opening calls for funding applications and experimentation with priority setting.The first edition (2011-2014) targeted the development of 2EG and renewable chemical technologies, whereas the second (2014-2018) targeted agricultural bioenergy.Another novelty was the structuring of the bidding process around business plans instead of R&D projects, signalling a preference for investments with high commercial potential.In PDPP, converting an early attempt at priority setting (Portaria n. 978/2008) into a systematic practice within the MS demanded new governance structures.The Executive Coordination Group (GECIS) had representatives of ministries and public agencies, including BNDES and FINEP. 10It was responsible for publicizing the list of strategic products, which triggered procurement and guided investment decisions of both public and private actors.Another example is the National Committee for Technology Incorporation (CONITEC).With civil society participation, CONITEC defines the criteria for investment and disinvestment in technologies for the healthcare system.
Public agencies' attempts to guide the directions of innovation involved conditionalities but maintained degrees of flexibility.BNDES and FINEP included commercial application, production, and collaboration with other firms and/or research institutions as qualifying criteria for PAISS 11 while ensuring freedom for companies to choose technological paths and partners.In PDPP, the government required the involvement of public laboratories in the partnerships, quality standards and gradual price reductions of the products that, if met, would benefit from public demand over the following ten years.Public laboratories featured as mediators of the relationships between the MS and firmsde facto price regulatorsand internalisers of technological and production capabilities through transfers.This institutional design illustrates the bold attitude of the state, and reveals the legal and political-economic nature of the challenges confronted: Incremental institutional changes happened at various levels to enable implementation, and public organizations themselves had to adapt to embrace active policies.PDPP demanded extensive regulations beyond the Innovation Law's general provisions.Law 12349/10 amended the procurement legislation, establishing preferential margins for the health sector and additional margins for products developed locally.Similarly, the regulatory framework for biological products had to be built (RDCs 55 and 49).
Increased institutionalization, capacity building and adjustments were important ingredients that enabled the programmes to operate and the inevitable failures to be borne.Indeed, public officials at BNDES and FINEP described PAISS as an innovation itself, a process of 'learning by doing' (Nyko et al. 2013, 60).One could say the same about PDPP.It started with a few partnerships (around 25) and soon scaled up to over a hundred.As the programme evolved, the rules that defined procedures for, obligations on, and responsibilities for, public and private actors went through a revision (in 2014), later enabling clearance from the auditing bodies.Hence, public risk-taking went hand-in-hand with efforts to strengthen the legal and institutional frameworks.

New financing instruments, risk diversification and competition
The quest to promote investment involving higher technological risk also prompted attempts to devise new financing instruments that prompted public agencies to increase their appetite for risk.In 2016, while targeting the scaling-up phase that is critical in the case of biofuels, BNDES introduced the Hybrid Debenture for Innovation Support (THAI). 12While the aim was to encourage disruptive innovations in large companies, in practice, design and implementation proved to be challenging, not only for legal and technical reasons, but also due to internal tensions: By law, BNDES cannot operate non-refundable grants … The challenge was to design a financial instrument with features of equity but linked to the innovation project … Internally … the biggest … resistance referred to the withdrawal of collateral guarantees … (Interviewee 18, PublMa) Earlier, in the context of financial support to pharmaceuticals, BNDES had also experimented with another financing instrumentthe Future Revenue-Sharing Risk Contract (BNDES 2010).However, operational difficulties and the lack of demand for this kind of support led to it being withdrawn from use: It was difficult to assess revenues and to define clear milestones for firms.Finally, it ended up not generating returns.(Interviewee 11,PublMa) In the interviews, public officials showed they had been mindful of the importance of increasing their risk-taking, for which ensuring an equitable sharing of any potential rewards in case of success would serve as an enabler.
In order to manage the risks taken, indications are that public agencies have pursued a sensitive portfolio approach.PAISS funds were spread across many types of firms and industries such as national start-ups, medium-sized and larger firms related to the sugar and ethanol industry (GranBio, CTC S.A. and Raízen S.A.) along with larger international groups in the biotech (Amyris, Novozymes and Mascoma), chemicals (Dow and Dupont) and oil sectors (Oliveira Filho and Consoni 2016).While part of the public support covered investments in R&D such as laboratory facilities and pilot plants, others entailed demonstration and commercial facilities.Moreover, to the extent that companies were left to choose their preferred technological solutions, PAISS financing also allowed variety and exploration of technological paths.By doing so, besides mitigating the blame-game regarding inevitable losses, this financing also increased the chances of success.In contrast, PDPP took an important step towards developing a portfolio approach when, following an initial period in which the MS commissioned a sole partnership for each product, it introduced competition.While advancing risk-mitigation practices helped to alleviate accusations of 'picking winners', it would not prevent failures more than 25 partnerships have been dissolved along the way.

Profit-sharing and conditionalities balancing the risks and rewards of investing
In both programmes, concrete mechanisms were adopted to ensure that the state could benefit from the upside in case of success and compensate for the likely setbacks.In the context of PAISS, until the creation of THAI, equity was the main instrument that the public sector could employ for that purpose, and it is what has been referred to as a 'profit-sharing policy instrument' (Laplane and Mazzucato 2020).Indeed, BNDES played an active role in seeking minority shareholder participation in national players such as Granbio and the Sugar Cane Technology Centre, thus acting as a venture capitalist.
The use of this type of financing instrument has also aimed to fulfil the strategic purpose of keeping national (as opposed to foreign) control over supported companies.An interviewee in the private sector made it clear: If there weren't public support, probably the company wouldn't be Brazilian and investments would have taken longer to happen.(Interviewee 10, PrivMa) In contrast, since PDPP's anchor is public procurement, the conditionality on pricing is the primary mechanism through which the government recoups any financial rewards.Here, aside from the impact on national savings, price regulation has a concrete meaning linked to access to medicines.The starting point is the obligation for firms to incur some price reduction, not necessarily initially, but along the learning curve and throughout the product cycle.Even though price stability is expected, renegotiation is also possible in the case of supervening circumstances.Therefore, the evidence suggests that risk-and-reward sharing serves multiple and strategic purposes, but that implementation invites caution and flexibility.Public agencies' ability to negotiate and compromise is key because it is what allows the creation and development of a common path with the business sector (Laplane and Mazzucato 2020): We are flexible and offer opportunities for [companies] to defer payment [via equity] (Interviewee 19, PublMa); and The guarantee of [public procurement] will lower industry costs with marketing and commercialisation, which are up to 40% of pharmaceuticals value.Can I ask for a 40% price reduction?… What is the price?It's the price of common sense!That's why it's so complex and risky.(Interviewee 33, PublMa) Just as important as the government's predisposition to have dialogue and to negotiate with the private sector, is that of reaching agreement across public agencies.In particular, in the context of a more advanced systemic policy like PDPP, overcoming the inherent tensions between different players' rationales, interests and return expectationsfor instance, between public financiers and public purchasersis challenging but possible through coordination.Another critical issue concerns negotiating with auditing bodies, if possible ex ante, over assessment criteria beyond the standard cost-benefit analysis in order to mitigate accountability risks for public managers.
Table 1 includes a summary of the results of the comparative analysis of the two programmes.

Outcomes and key issues for policy and practice
It is beyond the scope of this study to include an impact evaluation of the programmes; however, some results are visible, particularly with regard to the mobilizing feature of public risk-taking. 13During PAISS's first phase, the announcement of R$ 1 billion of public funds prompted 39 business plans (35 selected) involving potential investments of nearly R$ 6 billion linked to 25 firms, 7 business consortia, and 10 university-industry partnerships (Nyko et al. 2013).In response, public funding increased to around R$ 3 billion.Subsequently (2015), two commercial plants were under operation, one was under construction, and one demonstration plant was operating.PDPP achieved similar outcomes in terms of attracting private investment and stimulating collaboration.In parallel with their production of generic drugs, local companies embraced the biotech endeavour.From 2009 to 2014, 105 partnerships obtained approval, involving 50 private and 19 public laboratories, covering 61 drugs, 6 vaccines, 19 health products and 5 devices (Vargas, Almeida, and Guimarães 2016).According to the same source, by the end of 2014, procurement linked to the programme amounted to R$ 2.7 billion (U$ 1.3 billion), generating R$ 1.8 billion (U$ 880 million) savings for SUS and over R$ 9.1 billion (U$ 4.4 billion) of revenue for public producers.With such savings, access to the health system is widened, contributing to an improvement in the population's quality of life.Furthermore, building work has started on two new manufacturing plants for biopharmaceuticals.
In this paper the aim has been to analyse critical issues in the implementation of public-private investment partnership programmes.To this end, an analytical framework has been proposedcalled the market co-creating and shaping frameworkwhich has been used in an empirical analysis of recent innovation policies in Brazil.The case studies offer evidence confirming that the framework is a useful analytical tool and that it is relevant to actual experiences.The studies also highlighted some complexities which need to be taken into account in the design, implementation and evaluation of this policy approach.Hence, there are lessons drawn from the Brazilian experience that can serve as guidance for other government initiatives.For the sake of brevity, attention is drawn to two issues: the policy space and the mobilization of public finance. 14

Carving out a suitable policy space
As anticipated in Section 3, a change of vision from the top by which the government commits to promoting development policies, though a necessary foundation, is not enough.Equally important is the state's convening power.It must be able to create the synergies needed, within and across ministries and agencies, to harmonize macro-and micro-economics with social policy goals.Achieving this, however, first requires a change in the rationale that sees the space for policymaking as limited to fixing market failures.A fundamental shift in mindset, capable of countering the stigma attached to state 'intervention', while increasing public agencies' stamina to act as key drivers of economic and social change is, thus, critical for the development of virtuous public-private partnerships.The Brazilian cases illustrate that such a shift is possible even after a long period of neoliberal administrations.Still, it tends to be gradual and concentrated in a few 'islands of excellence'.
The process of legitimacy building goes over and above the structural economic conditions preceding a given policy.It features as strategic throughout the policy process, even more so in young democracies like Brazil.On the one hand, the creation of a shared vision that guides large-scale initiatives must already be present when the programme is formulated.In PAISS, whose design involved mainly BNDES and FINEP, public officials perceived that engaging more actors at that stage would have helped to strengthen public support.On the other hand, a policy like PDPP was part of a more articulated strategy for the health sector, engaging a wide range of actors such as civil society, public agencies and business.This suggests that expanding participation in the formulation of the policythat is, combining both top-down and bottom-up approaches is essential for legitimation and increases the likelihood of sustaining the policy's effects in the long-run; it is also in line with the considerations that have been highlighted in studies on the need for democratizing decisions about the directions of innovation (Stirling 2008(Stirling , 2009)).
Following on from the above, the specification of the intended policy outcomes in terms of environmental, social and economic benefitson which the two programmes reliedalso plays a role in constructing a favourable political environment for targeted initiatives.However, it is not sufficient.In addition, the value of an open-ended interaction, not just across public agencies but also between them and business and civil society, became apparent.While an intergovernmental body that helped to bring coherence to public (and private) actions were missing in PAISS, it and other participatory institutions were present in PDPP, contributing to the latter's systemic approach and stronger support.In any case, the starting point for any such policy must be a predisposition of public and private actors to collaborate with one another so as to promote the socio-economic goal in question.

Making entrepreneurial public finance effective
The change in the government's strategy must translate into real financial risk-taking and sharing capacities that enable the state to become entrepreneurial in the first place.While this is an aspect that might have been overlooked in studies focusing on contexts where the public sector had been playing a leading entrepreneurial role for a long time, in the Brazilian cases, its relevance emerges clearly.Once federal agencies came under active, and forward-looking directorshipsaligned with the administration's vision, vested with autonomy and fundingproactive and innovative attitudes flourished in the bureaucracy.It then became possible to improve both the quantity and quality of public finance.
The gains in scale and scope of public support, spread across the innovation chain while reaching the supply and demand sides through a wide range of instruments, are core ingredients for stimulating high-risk private investments.In the case of PAISS, collaboration between the funding agencies was intended to achieve greater breadth of support which was then expected to enhance their performance.There was an acknowledgement that the absence of specific measures to stimulate the demand and supply sides was a shortcoming that the creation of a new market (ideally) would need to overcome (Milanez et al. 2015).A similar and yet more comprehensive mix of policy instruments was deployed in PDPP, linking up the supply and demand sides to tackle the transformation of an existing market.This is consistent with what other comparative studies on mission-oriented R&D have found (e.g.Foray, Mowery, and Nelson 2012) regarding the importance of systemic policies in positively affecting the rate and direction of private investment.
Increasing the impetus of strategic public finance further involves assigning it properties such as patience, timely investment and some flexibility.It is worth recalling that the point of departure was a change in public funding agencies' operational policies, indicating their willingness to take risks (Section 3).In the case studies, the initiative to advance entrepreneurship first came from the public sector.The interview data also supports the interpretation that companies recognized that the state took the lead, arriving before the private sector and accelerating investments that otherwise would have taken longer to happen, or which might not have happened at all.Besides, public support was well timed as it tackled challenges that benefited from windows of opportunity.
Similarly, there is evidence pointing to a nexus between the risk appetite of public and private actors which showed that public officials were aware of the importance of playing an active role; for example, there was the realization that less ambitious business behaviour mirrors cautious attitudes in the public sector (e.g. the insufficient scale of grants allocated within PAISS), and there were attempts to innovate around financial instruments (in both cases).Although the evidence presented here is limited to two case studies, which involved some constraints in terms of the scale of resources mobilized and other factors, it does provide elements relevant to the Entrepreneurial State hypothesis as put forward by Mazzucato (2013).However, there is one aspect of the evidence that contradicts that hypothesis.
Public-private partnerships can, and must, occur in explicit ways.The analysis has drawn attention to instances in which risk-sharing was considered vital, and stakeholders were outspoken about what was being agreed upon, notwithstanding the uncertainties involved.This means that some common understanding regarding risk allocation had been achieved; and that this has occurred in a context where the discourses and practices on the role of the state were somehow attuned.Thus, the rise of the new 'Developmental and Entrepreneurial State' does not depend on implementation through implicit or hidden policies, such as those observed by Block (2008) and Mazzucato (2013) in the case of the United States.If risk-sharing occurs consciously and openly, it is more likely to allow openness about the distribution of rewards.

Conclusion
The incipient emergence of a strategic role of the state in co-creating and shaping markets in Brazil has been examined in this paper.It has been possible to recognize the entrepreneurial way (i.e. one that was opportunity-driven, proactive, problemsolving, and constructive) in which public funding agencies interacted with businesses and sought to achieve an explicit link between the risks and rewards of public investments.The national development bank was instrumental in this process.
These results point to important lessons for emerging and developing (Global South) economies that have accumulated some industrial and innovation policy experience.They indicate that under certain circumstances it is possible to create spaces for advancing active and systemic innovation policies while exploring pathways for sustainability and inclusive development.Nevertheless, as shown in the Brazilian case, the challenges for institutional building and implementation must be overcome in order to develop the conditions for the state to play a more entrepreneurial role.
A critical issue is the asymmetries in states' powers and capabilities to formulate and sustain policies that confront the interests of powerful corporations.Policymakers' concern about the risk of denationalization of businesses that benefited from public support (something well documented in the literature about global value chains and productive development in Latin America) deserves more attention and offers opportunities for further theoretical and practical insights.This problemalthough particularly sensitive in the case of emerging and developing economies, where institutional vulnerabilities are usual featuresis also experienced in other economies.Not only do tech-based companies enjoy greater mobility than other types of business, but also they may be more vulnerable to takeovers by powerful global players.Future research could explore the range of institutional and policy responses to this problem so as to derive the relevant lessons.

Notes
Mazzucato and Prof. Maria Savona for their constructive comments on earlier drafts.Previous versions have also benefited from feedback by participants at the Young Scholars Initiative Pre-Julio Olivera Conference in Buenos Aires in July 2019, and by two anonymous reviewers and the Editor of the SPRU Working Paper Series.I am also indebted to the interviewees for their invaluable insights.

Disclosure statement
No potential conflict of interest was reported by the author(s).
Because[public]  labs are subject to bidding exemption [under the procurement legislation], they have been chosen to stand as 'sellers' of health goods to the MS … they are [also] in a better position to resist dumping (Interviewee 36, CivSoc); and Multinationals cannot acquire public labs, so technology knowledge can remain in Brazil.(Interviewee40, PrivMa)

Table 1 .
A comparison between PAISS and PDPP through a market co-creating and shaping lens.