Government Procurement Policy in Developing Countries: The Case of Petrobras

Introduction

The Public Sector is a major user of many goods and services that support the functioning of the machinery of state and help meet the basic needs of the population. Government procurement plays a leading role in the promotion of domestic industry, especially given the magnitude of this market. Some research shows that government procurement accounts for 10–16 per cent of GDP in developed countries (Audet, 2002; Georghiou et al., 2003; Hoekman and Mavroidis, 1995; Weiss and Thurbon, 2006). Hence, in light of the scope and scale of the government procurement market, national governments use their purchasing power to promote the development of domestic productive sectors.

Government procurement policy has been at the centre of recent debate on policies to support innovation in developed countries (Aschoff and Sofka, 2009). However, there is a dearth of research on the adoption of government procurement policy as an instrument to stimulate innovation in developing countries (Ribeiro, 2009). This article sets out to start filling this gap by attempting to elucidate the role that can be played by government procurement policy in fostering innovation by firms located in developing countries.

Therefore, the focus of this article is to analyse the consequences of the government procurement policy practiced by developing countries. It is important to note that the concept of innovation used here encompasses a range of activities from copying, imitation, enhancement and experimentation to more value-added activities such as the development of new products and processes. This extended view of innovation, which transcends the traditional concept, is particularly important to an assessment of technical change in developing countries (Bell and Pavitt, 1993; Dosi, 1988; Lall, 1992). In developing countries, firms are recipients rather than developers of technology, and the learning process is the main way in which they build up technological capabilities (Bell, 1984; Lall, 1980, 1982, 1992).

Taking into account the important role played by learning to companies located in developing countries, the article provides a typology to classify and evaluate learning impacts of government procurement policy in the context of these countries. This typology is applied in this work to analyse the Brazilian government procurement policy, specifically the procurement policy practiced by Petrobras.

The case of Petrobras is examined or more precisely of this state-owned enterprise’s procurement policy for its offshore projects. ¹ Despite being from a developing country, Petrobras is recognised as a company at the technological cutting edge, with respect to the exploration and production of oil and gas (O&G) in deepwaters (Furtado, 1999). Since its foundation Petrobras has pursued a procurement policy that was very important for the constitution of the Brazilian Oil Supply Industry ² (BOSI). Thus, it is an interesting case for an investigation of whether government procurement policy in developing countries is used as a stimulus to innovation.

To elucidate the impact of Petrobras’ procurement policy, this article proposes a case study approach. It is argued here that to assess the procurement policy impacts of a government entity it is necessary to select an important case and investigate it thoroughly. Only through a systematic and micro level approach it is possible to capture how the procurement policy is contributing to transform the innovative capabilities of local firms.

This article focuses on large platforms that are very large and complex technological projects, and more specifically in the case of P-51 construction. This project is not only significant by its economic size (around a billion US$) but by its technological challenges. P-51 is one of the largest semi-submersible platforms ³ ever built in the world. As for the technological challenges, it is important to understand that procurement of a platform such as P-51 involves a large number of engineering-intensive activities with custom-built items and complex interfaces. Moreover, this stationary production unit (SPU) is the first platform of its type built entirely in Brazil. In sum, the project is emblematic in several ways.

Because of its importance, the case of the P-51 platform can be considered as a showcase of Petrobras’ procurement policy ability to promote in the recent years the innovation capabilities of the BOSI. Besides, the acquisitions made by this company represent an important part of government procurement market in Brazil. Last but not least, it is argued here that this analysis contributes to the understanding of some important features of public procurement policies in developing countries.

This article presents the results of field research with the main actors involved in the design of the P-51, namely Petrobras, ⁴ its Engineering Procurement and Construction (EPC) contractors ⁵ and a sample of subcontracted firms ⁶ for the supply of items to design the platform. The three EPC contractors hired by the national carrier were studied. In relation to subcontractors firms, to build the sample for this study the following requirements were adopted: (1) find themselves in Brazil (regardless of the origin of the capital); (2) be regarded supplier of an important item in the eyes of Petrobras and EPC contractors. Furthermore, it is worth noting that the sample of subcontractors was defined to include the largest possible number of segments belonging to the oil industry.

It can be said that the interviews were semi-structured, because although previously elaborated questionnaires were used, in many cases, the interviewer introduced questions not covered in the questionnaire, for greater depth, detailing questions, etc. Each actor participating in the P-51 project had to answer a questionnaire differently prepared according to their specificities. However, it is worth noting that the key theme of all questionnaires was the same: the impact of Petrobras’ procurement policy in terms of technological learning. These interviews took place between 2007 and 2009.

The results of the field research suggest that the learning impacts of Petrobras’ procurement policy, in relation to P-51, have been limited, focusing on elementary learning. According to the information obtained in the interviews, the main factors that led to this lacklustre result, despite the potential of learning related to a participation in a project with this profile are the following: the main multinationals participating in the project do not conduct high-tech activities in Brazil; and the Brazilian firms have a low level of technological capability.

The article has six sections including this introduction. The second section analyzes government procurement policy, emphasising on innovative public procurement. The focus of the third section is the literature on technological learning and the typology created to evaluate the technological impacts of public procurement policy in developing countries. Petrobras’ procurement policy is discussed in the fourth section, and its subsection presents a characterisation of BOSI, focusing on the profile and technological dynamics of this industry, in order to contribute to the analysis about P-51. The key features and procurement of P-51 are outlined in the fifth section, and its subsection describes the model proposed in this article for evaluation of the learning experiences made possible by participation in oil production projects. The consequent subsections present the empirical findings of the study conducted to analyse P-51 procurement. The last section contains the conclusions.

Innovative Public Procurement

Goods and services are purchased by the public sector to supply the inputs required for the fulfilment of its functions, such as the provision of public goods for use by society. In addition, government procurement policy is one of the few industrial policy instruments that can still be used by national governments to promote the development of domestic industry, given the existence of protectionist safeguards in the multilateral rules that discipline this field (Audet, 2002).

Although the supranational rules applicable to this field are designed to liberalise the market, it can be argued that not all markets will be opened up to international competition, since areas such as defense spending and, to some extent, investment in research tend to be reserved for domestic suppliers. Moreover, while purchasing decisions by private players are guided by strictly market-based criteria such as price, lead time and quality, public sector agencies can use other criteria, depending on the industrial policy and science, technology and innovation (ST&I) agenda adopted by the national government (Ribeiro, 2009).

Edquist and Hommen (1998, 2000) worked with the idea that government procurement policy may be not just a modality of industrial policy, but also an instrument for stimulating innovation. They emphasise the linkages between government procurement policy and innovation theory, highlighting the public sector purchases that contribute to the development of user–supplier interaction. For Edquist and Hommen (1998), two types of government procurement can be identified: regular government procurement, and innovative public procurement. Regular government procurement occurs when a government agency buys standardised goods, such as pens and paper. Innovative public procurement occurs ‘when a government agency places an order for a product or system which does not exist at the time, but which could (probably) be developed within a reasonable period’ (Edquist and Hommen, 1998, p. 4).

Geroski (1990) propounds the view that innovative public procurement is a more efficient instrument for fostering innovation than a wide array of R&D incentives, the tool used most frequently by national governments to promote domestic business innovation. When the end-user of the procured goods is the public sector itself, ⁷ the user–supplier relationship is of a special type, insofar as the final customer has substantial purchasing power and in some cases cutting edge technological capabilities (Ribeiro, 2005). It is worth noting that innovative projects often fail to materialise owing to the risks and large amount of capital they usually involve. Thus the products, processes, services and systems generated by government orders might not be developed at all if that order did not exist.

Despite the importance of innovative public procurement to the technological development of a country’s industrial fabric, it is possible to identify cases where the end-result of such a policy has been failure, in technological and/or commercial terms. Examples include the United States nuclear power industry (Campbell, 1988; Edquist and Hommen, 1998), and the Anglo-French Concorde programme (Edquist and Hommen, 1998). Notwithstanding such failures, there are also countless success stories in the specialist literature. ⁸ Moreover, the risk of failure is a necessary element of any innovation project and it therefore cannot be said to invalidate the deployment of an innovative public procurement policy.

It should be stressed that many innovations associated with strong social needs or demands offer low rates of return to private investors. In such cases, an innovation-oriented procurement policy serves as an important driver of innovation development. Thus, considering that the market does not always offer effective mechanisms for the satisfaction of society’s needs, innovative procurement policy can help fill the gaps created by market failure, by fostering the production of innovative public goods, for example (Dalpé, 1994).

The idea that government procurement policy can act as an important stimulus to innovation is grounded in arguments generally used to defend public sector intervention in a country’s technological development. Such arguments can be summarised as follows:

Underinvestment in high-risk or long-term R&D, involving radical innovation, for example;

Despite the importance of innovation-based procurement as a government policy instrument, this means of satisfying demand has been overlooked or underestimated for many years (Edler and Georghiou, 2007). The literature on this subject is relatively recent and mostly restricted to the experience of developed countries. The main studies aim to present the achievements of innovative public procurement in the US (see for example Cohen and Noll, 1991) and Europe (Edler et al., 2005; Edquist et al., 2000). Thus, as stressed in the introduction, one of the main aims of this article is the presentation of elements that contribute to an understanding of innovative public procurement in the developing countries.

Technological Learning and Government Procurement in Developing Countries

For a long time the economic literature has neglected questions relating to the technical progress achieved in developing countries. The basic neoclassical assumption is that technology is freely available to all countries and to all firms within each country. In this context, developing countries are assumed to prefer the technologies created by developed countries, selecting the most appropriate according to the factor–price ratio, in view of their physical, capital and labour endowments (Lall, 1994).

In turn, non-conventional theory also neglected for a long time the role played by technological efforts in developing countries. Lall (1982) argues that technical change in developing-country industries differs from the Schumpeterian concept of radical innovation. The fact that technological activities by firms in these countries do not fit the Schumpeterian concept has delayed an examination of the changes that occurred during the import substitution industrialisation period, thus concealing the richness and diversity of these experiences.

The 1970s are considered a watershed in the debate on developing-country technological efforts. Here it is worth noting the importance of the groundbreaking ideas proposed by evolutionary theorists such as Nelson and Winter (1977). In their seminal work, they explain, any conscious technical change made by a firm to enable the introduction of new products and processes can be considered technical progress, even when other firms know about the resulting technology. According to this vision, therefore, any innovation requires some kind of effort by the firm in question, regardless of whether it is added to the global stock of existing technologies, reproduces technology already developed by another firm, or even simplifies or renovates existing technology.

Inspired by this contribution from Nelson and Winter (1977), studies were produced on the experience of late industrialising countries, showing that certain characteristics of externally acquired technology and of the markets in which technology is bought and sold, such as non-replicability and imperfect information, lead firms in these countries to undergo a process of technological learning with a strongly idiosyncratic and adaptive bias (Fransman, 1984; Katz, 1974, 1981, 1987; Lall, 1978). In this context, technological learning enables developing-country firms to acquire the necessary competencies to select, assimilate, adapt and improve imported technology.

According to Lall (1980), developing-country firms learn about foreign technology in three stages: elementary—operational learning and minor adaptations; intermediate—copying and enhancement of processes and products; and advanced—understanding integrated projects and innovation. The following forms of learning are part of these stages:

Elementary Stage

Learning by doing: involves the development of manufacturing skills; materialised in manufacturing skills, promoting the development of increasing skill in production so that labour costs per unit of output and the incidence of quality issues can be reduced.

Intermediate Stage

Learning by design: involves the development of skills and knowledge relating to basic design; replication of external product or process technology by the firm; and hence the possibility of opening the black box of a specific technology to capture the knowledge essential to the redesigning of complex equipment.

Advanced Stage

Learning by setting up complete production systems: involves the development of the technical expertise required to produce equipment, as well as engineering skills and adaptation of entire factories according to specific needs.

Note that these forms of learning, called stages by Lall, do not necessarily follow each other in time sequence. Furthermore, according to the author, although the categories presented above do not exhaust the possibilities and ways of technological learning, provide a plausible approach regarding the stages of technological development which a company can go through (especially those located in developing countries). The technological effort required and fruits generated from the activities described in each of given learning mechanisms make clear that there is a gradation in terms of learning, going from elementary (associated with routine activities) to advanced level (that require deliberate technological activities as R&D).

The descriptive value and timeliness of this classification proposed by Lall should be emphasised, especially if we take into account that the process of import substitution is still ongoing in many developing countries. Moreover, it is a convergent approach with the broader vision of the innovation that is adopted in this article and with the technological profile of the companies operating in developing countries.

Accordingly, taking into account the importance of technological learning for companies located in developing countries and the relevance of Lall’s classification, we propose here a typology to assess the impacts of government procurement policy in the context of these countries, based on the contributions from this author. As can be seen in Table 1, the government procurement policy is divided into three modes: (1) government procurement inductive of basic learning; (2) government procurement inductive of intermediate learning; (3) government procurement inductive of advanced learning.

OPEN IN VIEWER

Table 1

Typology of Government Procurement for Developing Countries

Characterisation of Activities Involved	Elementary Routine Experience-based	Intermediate Replicable Search-based	Advanced Inventive R&D-based
Types of Learning	Learning by doing Learning by adapting	Learning by design Learning by improved design	Learning by setting up complete production system Learning by innovation
Procurement Policy	Government procurement inductive of basic learning	Government procurement inductive of intermediate learning	Government procurement inductive of advanced learning

Source: Adapted from Lall (1980).

The typology presented in Table 1 shows the different ways of technological learning can be triggered from a government order. With this approach, we seek to escape the ‘non-innovative’/‘innovative’ dichotomy (which could lead to misinterpretation of what was framed as ‘innovative’, considering only advanced activities, based on R&D), taking into account the diverse learning opportunities engendered by participation in a government order.

The government procurement inductive of basic learning promotes knowledge about the technologies operated by the company supplying the public sector, involving a combination of factors such as skills, equipment, product specifications and production, systems and organisational methods with which it is accustomed. This accumulation of experience enables the development of skills in the manufacturing process, making it possible, for example, to reduce the incidence of quality problems (learning by doing) and the introduction of small adjustments in an industrial plant, or product (learning by adapting).

With respect to government procurement inductive of intermediate learning, it is a modality that is linked to orders that promote technological capability in the suppliers to copy and promote improvements/adaptations in specific technologies. Such learning may involve opening the black box of a particular technology, allowing to capture essential knowledge to the redesign of the original basic design of complex equipment (learning by design), or changes in products, processes or plants, adapting them to local conditions in terms of raw materials, conditions and skills (learning by improved design).

Finally, government procurement inductive of advanced learning involves ordering of items that require the performance of risky and highly complex activities by the suppliers and often involve large investments in R&D. This kind of public acquisition can promote the development of technical skills required to produce items of an equipment involving engineering skills and adaptation of entire factories or plants according to specific needs (learning by setting up complete production system), as well as the emergence of new products and processes (learning by innovation).

This typology, as noted, emphasises the importance of different learning processes in companies located in developing countries. Thus, this typology seeks to address the multiplicity of activities from which technical change is achieved, being, therefore, more adherent to developing countries’ reality. It is evident that the adoption of government procurement inductive of advanced learning appears more promising in terms of fostering suppliers’ technological capabilities. However, one should not disregard the importance for the technological reality of developing countries, of public acquisitions that promote basic and intermediate learning. Mainly, intermediate leaning is important for technical change that takes place in these countries, because it involves more deliberate effort (than the basic learning) by the public sector’s suppliers, combined with activities such as copying and/or improvement on licensed technologies or improvements in specifications given, stimulating, therefore, the capability of this firms in terms of basic engineering.

The Evolution of Petrobras’ Procurement Policy

Petrobras is a state-owned enterprise (SOE) classified under Brazilian law as a ‘mixed-capital’ corporation because a large proportion of its share capital is held by private investors but it is subordinated to the federal government of Brazil. ⁹ This company was created by the Brazilian government in 1953 to reduce restrictions on the industrialisation process of the country, in view of the limitations in terms of oil resources. At the beginning of its activities, the degree of dependence Petrobras in relation to materials, equipment and technical services imported was almost complete. Example is that in 1954, only screws, bolts and similar pieces were produced in the country for the construction of Bahia (RLAM) and Cubatao (RPBC) refineries. Completed its first year of existence, the local content index reached the insignificant level of about 10 per cent (ANP, 1999). Thus, notorious was the weakness of the BOSI in the first years of Petrobras.

Since that time the strategy was to reduce external dependence, so that Petrobras started to adopt a policy to strengthen the BOSI. In the construction projects of refineries, production infrastructure, transportation and distribution, the Brazilian SOE called the BOSI to participate (Ribeiro, 2005). Aiming to reduce its dependence on imported materials, equipment and services, Petrobras developed a pioneer and successful policy of local suppliers’ training and qualification. ¹⁰

According to Villela (1984) Petrobras may be considered the first Brazilian SOE to adopt a procurement policy directed to the domestic market. It is worth mentioning that, being a net importer, the activities performed by Petrobras at that time had a negative impact on the Brazilian trade balance. Given the low price of a barrel of oil (until 1972) in relation to the high price of its derivatives, Petrobras investments were concentrated in the area of refining, a trend observed since the company’s founding in the 1950s (Ribeiro, 2009).

From 1973, under the first oil shock and the consequent increase in the price of oil, offshore activities assumed the status of a strategic area within the Petrobras. In the context of foreign exchange constraint and undervalued exchange rate, the Brazilian government pressed Petrobras to seek self-sufficiency of the country in oil production, in order to reduce the negative impact that their activity provided the country’s trade balance, problem that has worsened with the second oil shock occurred in 1979 (Ribeiro, 2009).

As exploratory research pointed to the presence of large reservoirs of oil and natural gas in the Campos Basin, which predominate areas of great depth, Petrobras had to evaluate the possibility of using new technological concepts, which allow the economic exploitation of deepwater reserves. Thus, beyond relying on imported technologies, the company’s strategy also encompassed the development of in-house capabilities, especially for overcoming the technological challenges related to E&P in deepwater. ¹¹ Six decades after its creation, Petrobras has become one of the largest oil companies in the world, ¹² with an extensive know-how in the area of oil exploration and production in deepwaters.

The achievement of self-sufficiency and the discovery of pre-salt, a new exploratory frontier, show the company’s success. ¹³ However, the purchasing policy of Petrobras paid little attention to the BOSI’s technological capability to innovate. These companies were hired just to replicate technologies developed by CENPES ¹⁴ (Petrobras’ R&D centre), or foreign companies.

According to Erber et al. (1984), this attitude was common among Brazilian SOEs in the period of industrialisation through import substitution, requiring the use of technologies already tested abroad by its domestic suppliers. For these authors, such a requirement ‘tended to generate a vicious circle in which its domestic suppliers, for not having prior experience in basic engineering, were forced to license foreign technologies, and because of this, these firms did not develop a technological capability in the area of basic design’ (Erber et al., 1984, p. 24).

For their part, suppliers are not considered strong enough to take innovative risks. For those firms, only if the state promotes policies that diminished considerably these risks it could be advantageous to make efforts in order to develop technologies, rather than import them. Thus, the import substitution industrialisation did not promote among Brazilian suppliers the capability to develop technologies. In this sense, the BOSI had a small participation in technological efforts that have been undertaken by Petrobras to develop technologies to be adopted in the production of O&G in deepwater, area that concentrates the main technological challenges faced by the company (Furtado, 2004).

From the second half of the 1990s, the use of Brazilian SOEs’ purchasing power was removed from the country’s industrial policy, due to the emergence of the neoliberal agenda. The faith in the ‘invisible hand’ of the market and the negative effects that the ‘hypertrophy’ of the state had brought to the country led to profound changes in the role of the public sector in the Brazilian economy, especially privatisation, deregulation, trade liberalisation and the adoption of orthodox macroeconomic policies to stabilise the economy (Ribeiro, 2005).

As part of this process, Petrobras lost its monopoly. The new political, economic and regulatory framework had adverse effects on the BOSI, leading Petrobras to reduce local procurement. ¹⁵ In line with global trends in the oil and gas industry, Petrobras began outsourcing many of the functions it had previously undertaken itself in major investment projects, preferring to engage turnkey contractors. Foreign firms were better equipped to act as EPC contractors in these projects, and local procurement declined farther because the foreign firms were not interested in building ties with the BOSI (Furtado et al., 2003). However, from the beginning of the 2000s, it has been possible to perceive a redirecting of Petrobras’ procurement policy back toward the domestic market, as evidenced by the adoption of local content requirements ¹⁶ in competitive bidding procedures for offshore platform projects. The case of P-51 is emblematic because it is the first semi-submersible platform assembled entirely in Brazil.

This means that some of the activities required for the project to be executed will take place in Brazil and would probably have taken place in other countries if the local content criterion had not been included (Ribeiro, 2009). Having indicated the significance of P-51 in the context of Petrobras’ new procurement policy, we present in this article the findings of a case study focusing on P-51. Semi-structured interviews with managers of Petrobras, EPC contractors and a sample of subcontractors for P-51 were used as the main data collection tool for the field survey.

Interviews took place in 2007 and 2009. The main aim of the field survey was to find out whether participation in the P-51 project was seen by EPC contractors and subcontractors as a good opportunity for technological learning in Brazil. Therefore, this study intends to reveal whether the current profile in the Petrobras’ procurement policy repeats or eliminates the main failure of import substitution industrialisation in Brazil, that is, a reduced contribution from the point of view of stimulating technological capability of firms to innovate.

Before presenting the results of field research, the next section presents a characterisation of BOSI, focusing in the profile and technological dynamics of this sector, in order to contribute to the analysis later on the P-51.

Characterisation of the BOSI

The group of companies engaged in the supply of equipment, technologies and specialised services to the O&G companies is called oil supply industry. This industry includes four types of companies, namely: (1) companies of geological, geophysical, drilling, air transport, maintenance of refineries, etc.; (2) engineering companies specialised in the preparation of basic design of platforms, processing O&G facilities, refineries, etc.; (3) construction and assembly’s companies, responsible for the construction of ships and floating structures, platforms, refining plants, assembly of oil and gas pipelines, etc.; (4) equipment suppliers of large, medium and small size, like turbines, compressors, pressure vessels, risers, tree, air filters, etc. (Iootty, 2004).

Therefore, heterogeneity is a hallmark of the oil supply industry. Coexist in this industry knowledge-intensive and skilled labour companies, service suppliers with a high technological component, companies with large construction sites and intensive hand labour, manufacturers of custom equipment, highly customised and high added value, services companies with lower technological complexity and manufacturers of series, standardised and low value goods.

Given this heterogeneity, because it is a multifaceted industry, made up of several segments with characteristics often distinct from each other, highlights the difficulty to build a clear vision about its technological profile. By taking the European Commission study (2008) on investment in R&D of various industries, the oil supply industry (oil equipment, service and distribution is the name given to this sector in the original) is classified as a medium-low R&D intensity sector, which applies between 1 per cent and 2 per cent of its net sales in R&D.

In the Brazilian case, the diversity/heterogeneity of the oil supply industry repeats, which complicates the analysis of the technological intensity of this industry. In addition, in the National Classification of Economic Activities (NACE)—used by IBGE, the national statistics bureau of Brazil—which guides the official statistics in R&D, there is no segment called oil supply industry. However, this study uses data on manufacturing of machinery and equipment as a proxy. Clearly this is an approximation, which has certain limitations such as the fact that it does not include the segment of engineering services. However, such an approach is quite plausible, because a significant portion of the oil supply industry is actually composed of companies that supply equipment and services.

Table 2 presents data for the Brazilian manufacturing industry and some selected segments belonging to it, like as expenditures on R&D, and R&D intensity, in the period 2006–2008. Table 2 reveals that there are several industries ahead machinery and equipment industry, among which we can mention the following: manufacture of other transport equipment (2.02 per cent); manufacture of cars, vans, SUVs, trucks and buses (2.01 per cent); manufacture of other electronic and optical products (1.90 per cent); manufacture of pharmaceutical products (1.44 per cent); manufacture of electric material (1.01 per cent) and petroleum refining (0.96 per cent). In addition, machinery and equipment industry’s R&D intensity is lower than the Brazilian manufacturing industry average (which is 0.80 per cent), as well as the R&D intensity of the oil supply industry pointed by European Commission study (between 1 and 2 per cent).

OPEN IN VIEWER

Table 2

Expenditure and Intensity on R&D, According to the Selected Activities of the Brazilian Industry (2006–2008)

Selected Segments	Expenditure on R&D (1,000 R$)	R&D Intensity
Manufacturing industry	10,634.632	0.80
Manufacture of other transport equipment	650.998	2.02
Manufacture of cars, vans, SUVs, trucks and buses	2,478.631	2.01
Manufacture of other electronic and optical products	112.897	1.90
Manufacture of pharmaceutical products	430.982	1.44
Manufacture of electric material	525.089	1.01
Petroleum refining	1,698.302	0.96
Manufacturing of machinery and equipment	392.844	0.46

Source: Adapted from IBGE (2010).

The R&D intensity is an indicator internationally adopted for comparisons between industries and countries and often used as a reference of the technological level of these agents, to measure efforts to generate new knowledge that are behind the advancement in technology. Therefore, it is an indicator of input in S,T&I important and lavishly used. However, like other conventional indicators of S,T&I (as patents, percentage of qualified scientists and engineers, capital intensity, etc.), the R&D intensity presents some problems highlighted by the literature (Figueiredo, 2004; Lall, 1992; Ribeiro, 2009) such as:

R&D represents only a portion of the set of activities that firms carry out to innovate;

The importance of R&D for industries that have high net income is distorted;

The incidence of R&D laboratory in emerging economies is rare.

In order to fill the gaps resulting from the application of conventional indicators of S,T&I in the context of developing countries, this study proposes an alternative model, presented in the previous section. This model assumes the study of technological learning stage (basic, intermediate or advanced) brings important contributions to the analysis of the profile of procurement policy (particularly in the case of developing countries) and, more than that, reveals the level of technological capability of this company. Furthermore, it is noteworthy that, although this model has been applied to analyse the design of an oil platform, it can be adapted for studies related to other projects and industries.

The P-51 Platform: A Case Study

P-51 is a semi-submersible platform that came on line in January 2009. It is located in Module 2 of the Marlim Sul field in the Campos Basin offshore Rio de Janeiro State. ¹⁷ Constructed in modular fashion, P-51 comprises a hull and topside, with the latter containing all facilities and modules for utilities, processing, power, compression and accommodation. The entire rig weighs more than 40,000 tonnes, with a height of over 75 metres and a perimeter of 470 metres. It generates 100 megawatts of electricity and can operate in water depths of up to 1,800 metres. It is designed to produce 24-hours a day for 25 years. Nominal throughput is up to 180,000 barrels per day, besides compression of 6 million cubic metres of gas per day. Its accommodation module can house a crew of 200 (Reis, 2004). In Figure 1 is shown an illustration of the P-51 and some figures about Campos Basin with its different field (as Marlim Sul filed, where the P-51 was installed).

After the competitive bidding procedure was held for the construction of P-51, three contracts were signed in 2004 with three different EPC contractors: a contract for the construction and delivery of natural gas compression modules (Rolls-Royce); a contract for the construction and delivery of power generation modules (Nuovo Pignone); and a contract for hull construction and integration, topside construction, process plant construction and reception and integration of all modules (FSTP Consortium). ¹⁸

OPEN IN VIEWER

Figure 1

Source: Ribeiro (2009).

Construction of P-51 took place in 2005. With regard to P-51’s modular construction, it can be said that technological evolution in the oil and gas industry is similar to the evolution process in other industries. Division into specific modules for integration at the production site was the solution found to address the complexity of the production systems involved. This solution enabled each part of the platform to be designed, fabricated and assembled in a relatively independent manner, significantly enhancing productivity via specialisation, learning, and scale economies (Barlow, 2000).

P-51 was built under turnkey lump sum contracts. For each module, the EPC contractors followed a set of specifications known as Front-End Engineering Design (FEED), ¹⁹ comprising detail engineering, procurement of equipment, materials and services, and construction under the respective scope of work. They also assumed the risks associated with project execution cost after FEED Endorsement, ²⁰ which basically means they bore the cost of any budgeting and planning errors once they had technically endorsed the design.

Furthermore, because P-51 is a complex production system ²¹ with a series of custom-built items (Barlow, 2000) it is arguable that participation in the project could lead to a significant amount of innovative learning, especially for EPC contractors and suppliers of mission-critical items. Besides, this is an emblematic case about the recent change in Petrobras’ procurement policy, characterised by local content requirements. These facts evidence the importance of a study that establishes the linkages between the procurement policy adopted by Petrobras for the P-51 order and the technological learning opportunities provided by participation in the project for EPC contractors and local subcontractors.

Technological Learning in Platform Projects

The various types of player involved in any platform project perform functions and activities with different levels of technological complexity, which can lead to basic, intermediate, and advanced learning. The matrix shown in Table 3 is designed to describe the range of possibilities for technological learning that derives from a firm’s participation in a platform project, and to help interpret the effects of the procurement policy adopted by Petrobras.

The matrix of technological learning in platforms projects presented in Table 3 is an adaptation of the technological capabilities matrix developed by Lall (1992), later adapted by Figueiredo for empirical use in studies of steel companies (Figueiredo, 2002, 2004) and organisations that operate on the basis of information and communications technology (Figueiredo, 2006). But, unlike Lall’s matrix and the adaptation proposed by Figueiredo, our approach evaluates learning not technological capability by firms.

The matrix presented in Table 3 covers the following functions (rows): platform engineering design, fabrication of equipment and supply of services, and construction and assembly. To each of these functions, there correspond activities that may or not involve innovative learning (columns). To create this matrix was necessary the selection of key technological functions in platform design and construction, and the identification of specific activities that offer opportunities for learning in the various forms outlined in the second section.

It should be noted that classification of the activities in Table 3 into different kinds of learning took into consideration each activity’s technological complexity. In addition, it is important to stress that this model was validated during the field survey with different professionals from the Brazilian oil industry.

OPEN IN VIEWER

Table 3

Technological Learning in Platform Projects

Learning →	Level of Learning
	Basic	Intermediate	Advanced
Functions↓	RoutineExperience-based	ReplicableSearch-based	InventiveR&D-based
Platform engineering design	Detailed and routine design	FEED (learning by improved design)FEED endorsement (learning by improved design)	In-house production of basic design (learning by innovation)
Fabrication of equipment and supply of services	Replication of given specifications (learning by doing)Routine quality control to assure conformity with customer’s specifications and standards (learning by doing)Minor adaptations to given specifications (learning by adapting)	Improvements to given specifications (learning by improved design)Reverse engineering (learning by design)	Development of entirely new equipment or services (learning by innovation)
Construction and assembly	Fabrication of hull or topside modules in conformity with given specifications (learning by doing)Minor adaptations to hull or topside fabrication (learning by adapting)	Modifications to hull or topside fabrication (learning by improved design)Modifications to module integration or deck mating (learning by improved design)	Development of new platform construction and assembly methods/solutions via engineering and R&D (learning by innovation)EPC management of platform projects (learning by setting up complete production system)

Source: Adapted from Lall (1992) and Figueiredo (2002).

Technological Learning in Brazil by EPC Contractors Working on P-51

The EPC contractors who worked on P-51 are analysed in this section in light of the matrix presented above. As noted, the analysis of participation by these companies in the platform project is based on information collected from interviews with their managers. Table 4 shows the functions and activities performed in Brazil by EPC contractors on P-51 and the respective forms of technological learning involved.

The point to which attention should be paid is the fact that only one of the EPC contractors, namely the FSTP Consortium, directly performed engineering design and construction and assembly functions in Brazil. Rolls-Royce ²² and Nuovo Pignone ²³ outsourced engineering functions for power generation and gas compression modules respectively to firms and shipyards located in Brazil, fabricating the mission-critical components of their modules outside Brazil.

Rolls-Royce has a sales office and turbine maintenance and repair facility in Brazil. Nuovo Pignone only has a sales office there. Rolls-Royce produced the turbogenerators, the main components of the power generation modules on P-51, at its plant in Liverpool, England. Likewise, Nuovo Pignone produced the main components of the compression modules (reciprocating compressors) abroad, more precisely in Florence, Italy. In sum, these firms did not directly perform in Brazil any of the functions or activities in their scope of work for P-51 and hence are not included in Table 4.

The FSTP Consortium, the only EPC contractor performing some technological activities in Brazil, included two firms Keppel FELS and Technip. Singapore-based Keppel FELS is a global leader in shipbuilding and oil rig production, with a long track record as a Petrobras supplier. In Brazil it owns the Keppel FELS shipyard and manages the BrasFELS shipyard, located in Niterói and Angra dos Reis respectively. Technip, a French firm, is a world leader in engineering, technology and project management for the oil and gas industry, with an engineering office in the city of Rio de Janeiro.

With regard to platform engineering design, as shown in Table 4 FSTP partially produced detail engineering for its scope of work on P-51 in Brazil. Partially because Keppel FELS produced FEED and detail engineering for the hull in Singapore, while Technip produced FEED and detail engineering for the topside in France and Brazil. Detail engineering is based on FEED specifications and uses platform information of various kinds. The aim of this activity is to produce a set of documents with detail designs, drawings, bills of materials, specifications for facilities and materials, architectural designs and calculations, among other items. These documents then serve as key inputs for the procurement of equipment, services and materials, and for platform construction.

With regard to the production of equipment and supply of services, FSTP farmed this function out to local and foreign firms and therefore did not engage in any learning experiences associated with it. The activities performed by FSTP in Brazil were confined to hull and process module construction and assembly, as well as integration of the modules and deck mating. Hull fabrication was a new activity for the consortium in Brazil. Hull blocks were received by the BrasFELS shipyard and welded together to form two C-shaped structures. The two Cs were joined at sea in an operation without precedent in Brazilian naval engineering, giving rise to a floating ring on which four columns were anchored to serve as the foundations for the topside.

Another point stressed in the interviews with regard to the construction activities performed by the consortium is the need to use special materials to build the process plant, given the weight, viscosity and corrosiveness of the oil processed by P-51. These characteristics of the oil extracted from Module 2 of Marlim Sul required the use of special stainless steel called super duplex. Welders at the Keppel FELS shipyard were obliged to adapt the procedure used to weld regular steel for oil rigs in order to guarantee the quality of super duplex steel welds.

Also with regard to the activities performed by FSTP on P-51, fabricating the topside outside the barge used for deck mating was a significant challenge. The FS-1 barge, built by the BrasFELS shipyard for work on P-52 as well as P-51, was being used for the construction of P-52 when work began on P-51. ²⁴ The solution was to fabricate the topside deckbox on one of the slipways located just behind the shipyard’s dry dock ²⁵ and then skid the finished deckbox down the slipway to the hull. A 112-metre slipway was built to slide the deckbox down on to the barge. First the dry dock was opened up to receive the FS-1; then, once the dry dock had been drained, the deckbox was skidded to the barge using hydraulic jacks.

Finally, with regard to the deck mating operation, although FSTP had already executed a similar operation for P-52 in Brazil, it can be said that a modification was introduced for the P-51 deck mating operation: installation of the flare boom ²⁶ on the topside before this structure was mated with the lower hull. This adaptation to the platform assembly process resulted in valuable cost and time savings by dispensing with the use of barges for hoisting and installation of the flare boom after deck mating. Among the hazards avoided by this solution was the risk of topside listing or tilting, since most of the flare boom projects outside the structure. According to information collected in the field survey, the operation was a success.

Cross-referencing the information collected from FSTP with the learning matrix presented in the previous section shows that the activities performed by the consortium for the P-51 project led to some occurrences of learning. Learning by doing took place in the fabrication of the process modules in accordance with the FEED specifications. Learning by adapting consisted of the introduction of a minor adaptation to fabrication of the process module by using special steel that is difficult to weld. Learning by improved design derived from topside detail engineering and the modifications adopted for construction of the hull and topside and for the deck mating operation. Lastly, based on the information obtained in the field survey it can be said that FSTP, as the lead EPC contractor for P-51, benefited from learning by setting up a complete production system, as it managed both the P-51 and P-52

Comparing the activities of the FSTP consortium in the P-51 project (described above) with the typology of technological learning in oil platform projects, seven instances of learning have been identified, which are detailed in Table 4: two occurrences of basic learning (fabrication of process modules according to given specifications, and minor adaptations to fabrication of process modules), three occurrences of intermediate learning (detail engineering of topside partially produced in Brazil; modifications to hull fabrication and; modifications to deck mating) and one occurrence of advanced learning (EPC management of platform projects). In this sense, one can say that the balance in terms of learning for this EPC contractor can be considered positive. However, it is important to take into account that the other two EPC contractors yielded no learning in Brazil. Furthermore, it is noteworthy that there was no occurrence of advanced learning. Given these findings, it appears that the total balance of this order from Petrobras among EPC contractors cannot be considered encouraging.

OPEN IN VIEWER

Table 4

EPC Contractors: Learning Experience in P-51 Project

Learning →	Level of Learning
	Basic	Intermediate	Advanced
Functions↓	RoutineExperience-based	ReplicableSearch-based	InventiveR&D-based
Platform engineering design		Detail engineering of topside (partially produced in Brazil)
Fabrication of equipment and supply of services
Construction and assembly	Fabricationof process modules according to given specificationsMinor adaptations to fabrication of process modules	Modifications to hull fabricationModifications to topside constructionModifications to deck mating	EPC management of platform projects

Source: The authors.

Technological Learning in Brazil by Subcontractors Working on P-51

The activities of local firms subcontracted by the EPC contractors to supply equipment and services for the P-51 project were also investigated in the field survey, which included a sample of twelve such firms. These subcontractors that are presented in Table 5 belong to the BOSI. They were selected for sampling purposes according to the following criteria: (1) located in Brazil (regardless of whether the parent company was Brazilian); (2) supplier of an important item for P-51 in the eyes of EPC contractors and Petrobras.

OPEN IN VIEWER

Table 5

Sample of P-51 Subcontractors Covered by Field Survey

Group	Firm	Origin of Capital	Location	No. of Employees (2008)
Construction and assembly	Nuclep	Domestic	Itaguaí - RJ	1,200
	Usiminas	Domestic	Ipatinga - MG	8,587
Equipment	Iesa	Domestic	Araraquara - SP	2,814
	Ebse	Domestic	Rio de Janeiro - RJ	637
	Weg	Domestic	Jaraguá do Sul - SC	21,846
	Cliner	Domestic	São Gonçalo - RJ	60
	Sulzer	Foreign	Jundiaí - SP	857
	Siemens	Foreign	Jundiaí - SP	9,030
	Unicontrol	Domestic	São Paulo – SP	250
	Elco	Domestic	Curitiba - PR	320
Engineering services	Projemar	Domestic	Rio de Janeiro - RJ	350
	Aibel	Foreign	Macaé - RJ	880

Source: The authors.

Note: More information about the sample firms is shown in Appendix section.

The sample encompasses the main firms involved in hull block construction and in construction and assembly of metal structures for the power generation and compression modules, as well as some of the local firms subcontracted to supply mission-critical equipment and services for the platform (such as pressure vessels, electricity generators and the electrical system), and the firms responsible for detail engineering of four out of the platform’s eight modules.

The firms were divided into three groups according to the activities performed and the items supplied, namely: construction and assembly (Nuclep and Usiminas); equipment suppliers (Iesa, Ebse, Weg, Cliner, Sulzer, Siemens, Unicontrol and Elco); and engineering firms (Projemar and Aibel). These firms were mostly located in São Paulo State and Rio de Janeiro State, that is, the wealthiest state in Brazil and the state leader in oil and gas production.

Eight of the twelve firms (about 65 per cent) can be classed as large according to the criteria used by IBGE, the national statistics bureau. ²⁷ Thus for most of the firms size was not a factor that might have restricted their technological dynamism. Moreover, 75 per cent are Brazilian-owned, so that origin of capital cannot in principle be used as an argument for failure to perform technological activities in Brazil.

It is worth noting that Petrobras considers the equipment and services provided by firms that are included in its Corporate Register ²⁸ as essential for the full operation of its plants. Therefore, we can say that all companies in the sample, to a greater or lesser extent, provide critical items for Petrobras’ projects. The research revealed that the requirements of the Brazilian SOE force BOSI’s companies to improve their standard of quality, Security and Environment (HSE) and financial management, etc.

Based on information collected in interviews, it was found that about 85 per cent (twenty of twenty-three) of all occurrences of learning identified between subcontractors as a result of their participation in the design of the P-51 can be classified as basic-level (see Table 6).

The main form of learning of these firms was learning by doing, related to the following activities: replication of given specifications in equipment fabrication; routine quality control to assure conformity with Petrobras’ specifications and standards; minor adaptations to given specifications; minor adaptations to hull block fabrication in accordance with specifications; and construction and assembly of metal structures for generation and compression modules according to given specifications.

The occurrences of intermediate learning accounted for about 13 per cent of total cases of learning observed among the sampled firms that participated as suppliers in P-51. They are related to the platform engineering design function (two) and the equipment fabrication and service provision function (one), in activities that enabled the firms involved to internalise learning by improved design. Appendix indicates in a large measure that none of the subcontractors experienced risky, R&D-based learning that could enable them to develop innovations at the international technology frontier. Furthermore, participation in P-51 was confined to routine learning for nine of the sampled firms (75 per cent), and to learning by doing and/or learning by adapting for two firms in the construction and assembly group, and seven of the eight firms in the group of equipment and service suppliers.

OPEN IN VIEWER

Table 6

Subcontractors: Learning Experience in P-51 Project

Learning →	Level of Learning
	Basic	Intermediate	Advanced
Functions↓	RoutineExperience-based	ReplicableSearch-based	InventiveR&D-based
Platform engineering design		FEED/detail engineering for compression modules (firm Projemar)FEED/detail engineering for compression modules (firm Aibel)
Fabrication of equipment and supply of services	Replication of given specifications in equipment fabrication (Iesa, Ebse, Weg, Cliner, Sulzer, Siemens, Unicontrol and Elco)Routine quality control to maintain existing specifications and standards (firms Ebse, Weg, Cliner, Sulzer, Siemens, Unicontrol and Elco)Minor adaptations to specifications (Siemens and Unicontrol)	Improvements to specifications (Unicontrol)
Construction and assembly	Minor adaptations to hull fabrication (Nuclep)Construction and assembly of metal structures for generation/compression modules according to given specifications (Usiminas)

Source: The authors.

Thus, innovative activities was not the spotlight of the P-51 project. The almost total absence of technological efforts among the firms surveyed is one of the main factors contributing to the low technological dynamism of these firms. Besides showing weak R&D investments, the information collected in the interviews suggests that these firms have negligible relationships with universities or research institutions. The low level of technological effort in the BOSI and its weak involvement with other players in the national innovation system makes the licensing of foreign technology inevitable. The study of the P-51 project revealed that most of the subcontractors involved relied heavily on technology licensing agreements or design development by the customer (Petrobras-CENPES) or by EPC contractors.

According to information obtained in the field research, the form of licensing adopted by companies which are part of the BOSI is non-exclusive. It is a form of licensing in which there may be more than one company holding the license to exploit the patent. In addition, the contract does not provide training and/or technology transfer. Such information collected in the field research corroborates what was identified in the literature (Dantas, 1999; Furtado and Freitas, 2000; Oliveira and Rocha, 2008; Ribeiro, 2005, 2009; Ruas, 2010). The licensing strategy mitigated risk but prevented them from developing technological capabilities through more sophisticated forms of learning than learning by doing and learning by adapting.

Regarding the technological gap of BOSI, it is appropriate to present some information collected in the interview at Iesa company. This firm provided three pressure vessels for process modules of the P-51, whose internal components were purchased outside of the country, more precisely by the Norwegian Aker Kvaerner. While Iesa carried out activities related to industrial boiler, such as the manufacture of pressure vessels, Aker Kvaerner has focused on defining the internal components, critical to implementation of the physical and chemical processes within the platform. According to the interviewed from Iesa, this case serves to illustrate the technological dependence of the BOSI. This interviewee stated that,

in the contracts made with the foreign firms there is no condition requiring technology transfer. In light of this, the Brazilian company has no access to the technology in order to be able to work on its own on the next project. So, the role of local firms in Petrobras’s project was practically restricted to the supplying of industrial ‘commodities’, i.e., items of low-technology content.

According to the interviewed from Ebse,

There are some things that Cenpes outsources. In the P-55 project we are making the bottom nodes of the structure which supports the equipments, the bottom nodes of the deckbox. That platform’s basic project was developed by Cenpes but its structural calculations were done by a Dutch firm. Therefore, it can be said that Petrobras prioritizes foreign firms when establishing partnerships aiming the development of technologies.

Furtado and Freitas (2000), confirms this view, as they point out that besides the technological efforts in-house, one of the main strategies of technological catch-up of Petrobras is the partnerships with foreign companies and institutions (especially through cooperative R&D). To support this view, the authors present the cases of technological development achieved by Petrobras relating to three critical technologies for the subsea production of oil and gas: subsea multiphase flow pumping system, subsea separation system and, electrical submersible pumps in subsea wells. Furtado and Freitas (2000, p. 23) argue that these technologies were developed ‘with a majority of foreign producers or institutes.’

Companies wishing to supply equipment and services to Petrobras’ projects need to be entered in the Corporate Register of Goods and Services’ Suppliers. This register is composed of qualified and enabled firms, and that meet the provisions of Decree 2.745/98, which governs the acquisitions made by Petrobras. The Brazilian SOE evaluates companies seeking to integrate the Corporate Register, based on some criteria, namely: technical, economic, legal, Health Safety Environment (HSE), managerial and social responsibility. The aim of Petrobras with this rigorous selection is to ensure the quality of equipment and services used in its projects. Companies that seek to participate in the registration of Petrobras are audited in the areas of engineering, procurement and manufacturing. In some cases, when applicable, they must develop prototypes and have the following certifications: ISO 9001, ISO 14000 and OHSAS 18000. The supplier that passes through Petrobras’ sieve is incorporated into the Corporate Register and receives a Certificate of Registration of Registry Classification (CRCC), valid for one year.

In this regard, it should be noted that Petrobras has, besides the CRCC, a register called Brazilian Prime Vendor, composed of some Brazilian companies that can provide equipment with third party technology, since owning the domain of manufacturing process. Ebse company, for example, by integrating the Brazilian Prime Vendor, acted this way in the projects of the P-55 and P-57. According to information obtained from Ebse,

when local suppliers are contracted directly by Petrobras, the technology associated with the ordered item must be licensed by them, and this process is not accompanied by the technology transfer. Thus, the technology is delivered by the company that holds it, to the domestic manufacturer, inside a black box.

As stressed in the fourth section of this article, one of the problems identified by the literature in relation to Petrobras’ procurement policy during the import substitution industrialisation period refers to the fact that this policy has not dealt with the BOSI’s technological capacity to innovate. More recent studies indicate that this scenario has not changed lately. According to Dantas (1999), the lack of technological capability and the dependence on foreign technologies are problems that remain from the import substitution industrialisation period. Oliveira and Rocha (2008) state that even companies in this industry with higher competitive capacity are dependent on external authorisation and licensing of the technologies they use. Thus, these authors conclude that the BOSI presents a significant deficit of competitiveness linked to the fragility of national engineering and limited technological capacity to innovate. Ruas (2010, p. 41) states that despite the growth experienced by the BOSI in recent years, due to the increased volume of orders from Petrobras, ‘some deficiencies persist, limiting the potential for advances even with the advent of reserves and production in pre-salt layer’.

The results of the research presented in this article converge with these other studies, revealing that the low technological capacity of the BOSI is the main obstacle to its development. Therefore, the simple replication of externally developed technologies leads to learning of basic level. Despite its importance, it is the first stage of learning in the classification proposed by Lall (1982). The technological development of an industry requires that firms are able to move to more advanced stages in terms of learning. For such a development to take place, it requires a more active attitude by firms receiving technology coming from abroad (Ribeiro, 2009). However, only three companies in the sample (25 per cent) reported carried out R&D systematically and have a formalised department for this purpose. Therefore, it is not surprising that there was no occurrence of advanced learning between the subcontracted companies of the sample, since such learning requires a more deliberate effort on the part of companies.

According to what was found with the interviews Petrobras’ procurement policy could also help change this picture. The technological capabilities accumulated by CENPES, owns the domain of much of the technology used in the design of a platform. This domain comprises the basic engineering, responsible for defining the size and model of the platform, equipment and materials used, the operating conditions of the processing fluids extracted from wells, etc. However, the CENPES’ technological domain is not transferred to Brazilian suppliers. In accordance with the field research, what happens in relation to the basic engineering of platforms is symptomatic. For this activity, CENPES, according to some respondents, is the main competitor of engineering companies in the country. This competition prevents the learning of these companies in basic engineering of platforms, which is restricted to detailed engineering.

From this study, it was also possible to verify that the links between EPC contractors and subcontractors are purely commercial. As it was found in the field research, we can say that the EPC contractors concern is to acquire equipment or services from a company with CRCC. An indication of the fragile relationship between EPC contractors and subcontractors is the absence of clauses in the contracts between the parties providing the transfer of technologies and conducting training.

Thus, taking the case of P-51 as an illustration of the procurement policy adopted by Petrobras for its offshore projects in the recent years, it can be said that the national operator redirected its purchasing to the domestic market and thereby contributed significantly to a recovery by important segments of the BOSI, such as shipbuilding and naval engineering. Nevertheless, its procurement policy did not advance in the sense of reducing this industry’s technological dependency because it did not foster the technological learning noblest, which made it possible to perform an authentic process of catching-up. ²⁹

Conclusions

As stressed throughout the article, procurement by government entities can play a leading role in the promotion of domestic industries, above all owing to the size of this market in the national economy. Given the importance of public procurement, therefore, this article aims among other things to contribute to the debate about procurement policy by focusing on its consequences in terms of technological learning/innovativeness in developing countries.

Developing-country firms are much more recipients than developers of technology, so that the learning process is the principal means through which they can build technological capacity. The article therefore presents a methodology for analysis of public procurement policy in developing countries, constructed on the basis of the literature on technological learning. The thesis defended here is that this type of approach is a useful tool for determining the extent to which public sector procurement stimulates innovation by firms located in developing countries.

The article uses this approach to analyse the case of Petrobras, chosen mainly because of its technological vigour in deepwater oil production. Furthermore, as outlined in the fourth section of this article, the use of orders from Petrobras to promote BOSI returned to the Brazilian government’s agenda recently, particularly from Lula’s government. These factors combined justify the choice of Petrobras and P-51 for application of the framework proposed in this article. To evaluate the impact of P-51 procurement on the BOSI, the article presents the findings of a field survey comprising interviews with the key players in the project, namely Petrobras, its EPC contractors, and a sample of subcontracted suppliers.

For EPC contractors, the field survey showed that only one experienced some technological learning FSTP Consortium thanks to the P-51 project. The others contractors outsourced detail engineering, procurement of goods and services and module construction to subcontractors. This shows a lack of interest in manufacturing and conducting R&D in Brazil of some multinationals, which restrained technological learning by the BOSI.

Most of the critical, expensive and technology-intensive components ³⁰ were fabricated outside Brazil. While complying with the contractual local content requirement, the index excluded from the calculation some of the major high value imported items. One of the main conclusions to be drawn, therefore, is that notwithstanding the existence of a procurement policy with a local content, this policy was unable to promote the production of high value added, R&D-intensive equipments. The predominance of multinational EPC contractors in Petrobras’ platform projects also evidences both the almost complete absence of domestic shipbuilders and EPC contractors, and the technological and entrepreneurial fragility of the Brazilian industrial system.

Subcontracted firms supplied mainly products that involved routine activities, so that learning by doing was the most recurrent kind of learning identified. One of the obstacles to non-routine learning is the low technological capacity of the BOSI. Only three of the sampled firms (25 per cent) reported the systematic activities and the existence of a formalised department in R&D. The limited experience of the BOSI in engineering projects ends up forcing it to resort to the licensing of foreign technologies, or development of the project by the client (Petrobras–CENPES). Therefore, the main conclusion of the study is that, despite technological opportunities related to the order of a complex production equipment, with complex interfaces and standardised equipment and services, Petrobras has adopted a procurement policy mostly inductive of basic learning. It is worth noting that these results of our study corroborate what has been identified in the literature (Dantas, 1999; Furtado and Freitas, 2000; Oliveira and Rocha, 2008; Ribeiro, 2005, 2009; Ruas, 2010).

In addition, the fact that the top procurement priority for Brazil is increasing local content, rather than promoting innovation learning among Brazilian suppliers, does not contribute to a change in these firms’ technological dependency. Accordingly, Petrobras’ procurement policy nowadays, based on the study made on the P-51 order and given that this project represents an important and emblematic case concerning the acquisition strategy recently adopted by the Brazilian oil company, as mentioned in this article, by emphasising local content, reissues import substitution. Nevertheless, such a policy does not advance in order to stimulate the technological capability of local suppliers, so that the technological dependence of these companies is not eliminated. ³¹

To these authors it appears that Petrobras could contribute to an improvement in this situation, given its status company at the technological cutting edge, with respect to the exploration and production of oil and gas in deepwaters, and the massive investment it will have to make to extract oil profitably from the presalt layer far below the ocean bed, its new production frontier. However, Petrobras has so far shown reluctance to assume the role of promoting technological development among local suppliers because of the risks and costs involved in doing so.

In order to encourage Petrobras to adopt an innovation-oriented procurement policy, the Brazilian government must undertake to bear this extra cost. Thus the main thesis advocated by this article is that given the technology lag suffered by firms in developing countries; a SOE cannot bear alone the burden of fostering innovation by firms via an innovation-oriented procurement policy. Developing-country governments should make good use of their industrial and technological policy arsenal (such as, long-term financing, tax exemption and R&D subsidies) to help local firms participate in public sector procurement of items that constitute innovations and acquire higher levels of technological capability through a continuous learning process, thus supplementing their innovative public procurement efforts.

It is important to emphasise that this article aims to analyse the technological ramifications of government procurement in the context of developing nations, proposing an adequate typology for those countries’ profiles. Such typology is based on the premise that technological learning plays a decisive role in the innovative dynamics of the firms that operate within those countries, because in many cases the import substitution process remains ongoing. However, since this typology was used only to evaluate the Brazilian governments’ procurement policy concerning the oil industry, the need for new studies to evaluate the Brazilian government’s procurement policy in other areas, as well as to analyse other countries’ government procurement, based on this typology is evident. From those studies, it would be possible to test its validity, as well as to improve the typology proposal contained in this article. Regarding the typology of technological learning in the projects of oil platforms (used in the P-51 study), it is also argued here that that typology could be applied to evaluate other platform projects, with a view to making the projects more robust. Furthermore, this study deserves to be continued since it would allow shedding light on important issues that have not been addressed on this article, such as the influence on Petrobras’s procurement policy targeting the acquisition of platforms posed by the new scenario that presents itself in the Brazilian oil sector (marked by the pre-salt discoveries—with its opportunities and challenges—and by the creation of new instruments of the Science, Technology and Innovation industrial policy by the Brazilian government towards the oil sector, such as Inova Petro). Finally, that typology, originally created to identify the technological learning resulting from the participation in oil platforms projects, could be adapted for applying it in the subsea segment, in which the pre-salt’s main technological challenges present themselves.