Local content requirements in the oil and gas industry in developing

Local content requirements in the oil and gas industry in
developing countries: the return of ISI?*
Renato Lima de Oliveira
Massachusetts Institute of Technology, [email protected]
Abstract: In the recent years, a growing number of developing countries are adopting local
content (LC) requirements as condition for exploring and developing oil and gas resources in
their home countries, with Brazil being a leading example and Mexico as the latest follower.
Similar to the Import-Substitution Industrialization (ISI) model, LC creates incentives to the
development of indigenous industries by protecting domestic markets against free imports of
goods, like drilling rigs, oil platforms and subsea equipment. Given this similarity, it is worth
questioning: Is LC a contemporary reenactment of the ISI development model? ISI has been
extensively criticized in the literature and that most developing countries – including Latin
America nations – had abandoned in the 1990s after market-driven reforms. This paper will
highlight the similarities and differences between LC in the resource sector and the historical
experience of ISI. Using a principal-agent framework and Hirschman’s concepts from “Exit,
Voice, and Loyalty”, this work remarks the monitoring and voice capabilities of oil operators in
pressuring suppliers over the quality and time to market of the goods used by the oil industry.
Drawing evidence from a recent Petrobras corruption scandal, we show how bribe rates differ
from LCP contracts in comparison to other Petrobras’ projects, using data from 88 projects
signed between 2003-2010, worth over $ 30 billion. This paper advances our theoretical
understanding of local content as part of the political effects of resource abundance, how it
differs from previous industrial policy efforts, and provides corroborating evidence from a rich
dataset of corrupt practices.
*
Paper prepared for the conference “Thirty Years of Democracy in Brazil: A Research Workshop”, Kellogg Institute,
Notre Dame University, April 20, 2015. Very early draft – please request latest version before citing. I would like to
thank the conference organizers for the invitation and participants of the Harvard's Kennedy School Latin America
National Oil Companies (NOC) Workshop, MIT’s Latin American Working in Progress Group, the Total Energy
School in Paris, and Azerbaijan Diplomatic Academy (ADA) 2014 Energy Summer School in Baku where I received
earlier comments and additional information about some of the topics developed here.
1
1. Introduction
If you are an oil company operating in Brazil, such as Shell, Statoil and Petrobras, in
addition to look for oil you most likely have to look after suppliers. Since this country opened the
oil sector for foreign operators in the mid-1990s, local content policies (LCP) imposed that oil
operators have to purchase a significant share of their capital goods and services from local
suppliers, benefitted by high tariff walls. Any delays from these suppliers can affect the rate of
production of oil operators, and thus their cash flow. This gives oil companies a strong incentive
to interact with local manufacturers up to the point of transferring technology and managerial
skills – effectively looking after their development.
Brazil is far from being alone in adopting LCP. Mexico is the latest country to adopt a
LCP in its ambitious energy reform – a move that has already be taken by countries such as
Malaysia, Nigeria, Uganda and Kazakhstan. Despite the rising importance of such policies and
its global reach, discussions of LCP has been mostly confined to the oil industry niche and has
received very little attention from a broader developmental standpoint.
Viewed from afar, the establishment of high-tariff walls to stimulate the growth of a local
industry that would replace the purchase of goods that could have been imported at cheaper
prices seems to be a reenactment of the import-substitution industrialization (ISI) model of
development. If that is the case, then it would be relatively easy to look back at the accumulated
evidence on ISI to judge the feasibility of LCP. Given the similarities, it is worth questioning: is
ISI coming back, now under the label of LCP? Or is LCP a qualitatively distinct policy, with
possibly different challenges and outcomes?
2
If LCP resembles the general characteristics of ISI, then most analyst would agree,
although not all, that the prospects would not be very optimist, as many like Bulmer-Thomas
(2003, p. 278) believes that ISI as a development model “cannot be defended.” On the other
hand, if it has characteristics that depart from shortcomings of the conventional ISI model, it may
suggests more room for optimism about the possibility of combining abundance of natural
resources with industrial upgrading that such policies claim to seek.
This paper will argue that LCP has key differences from the traditional ISI. Using a
principal-agent framework and Hirschman’s concepts from “Exit, Voice, and Loyalty”, this work
remarks the monitoring and voice capabilities of oil operators in pressuring suppliers over the
quality and time to market of the goods used by the oil industry – their incentive to look after
suppliers. In addition, LCP is less insulated from international competition and is conducted
through policies that are more transparent and less distortionary than traditionally has been the
case under ISI. This aspect is crucial to limit inefficiency. On the other hand, common challenges
to societies and governments remain on how to redirect capital to invest in upgrading
capabilities, strengthen the linkages within the supply chain and avoid rent-seeking.
In addition to the theory presented here, we draw evidence from Brazil’s adoption of LCP
and the current scandal of Petrobras. The Brazilian flagship oil company is currently under
investigation for taking part in a major kickback scheme where money from suppliers went to the
pockets of its employees and politicians. Particularly, we exploit data of bribe rates on 88 large
Petrobras projects from 2003 to 2010 and worth over R$ 70 billion (over $ 30 billion dollars).
The data, in a detailed spreadsheet was handed to the Brazilian Federal Police and public
prosecutors in a leniency agreement with Pedro Barusco, former senior manager in the
Engineering and Services directorate of Petrobras. We show that corruption was higher in the
3
downstream sector of Petrobras, responsible for building of refineries, for example, than in the
upstream division, the one responsible for fulfilling the local content mandate. We explain the
results in light of models of political economy of corruption and the challenges of LCP addressed
in this paper.
This paper is divided as follows: in addition to this introduction, section two address the
main characteristics of the ISI development model. Section three introduces LCP, traces its
genesis to the development of high-cost of production resources, and highlights the main
differences to ISI. Section four provides empirical evidence that backs the dissimilarities
highlighted in the previous section and analyzes the growth of LCP in Brazil and the current
Petrobras scandal. Finally, section five concludes.
2. The inward looking development model
Any general characterization of a development model will lose the specificities of the
application of policies within particular countries at given times. This also applies to characterize
the ISI as a development model. Many countries, including developed ones (Chang 2002), opted
at some historical moments to protect certain sectors from international competition, either to
stimulate their growth or deter their decline. However, the occasional use of an instrument does
not make one an artisan of the trade. ISI, or perhaps more appropriately the inward-looking
model, had a toolbox of incentives, implementing institutions and bureaucracy, and an ideology
behind it with centers of intellectual production and diffusion that goes much beyond the
sporadic adoption of protectionism.
4
Where these full characteristics – instruments, institutions and ideology in favor of
import replacing industrialization – could most be found was in Latin America. While the
successful East Asian countries used instruments of protection to support infant industries, ISI as
a model was not deepened nor had broad intellectual support as it had in Latin America (Haggard
1990).1 Consequently, the portrayal of the ISI model here will draw largely from the experience
of Latin American countries, particularly their major economies, Brazil and Mexico.2
Import substitution had different phases and stimuli. It can start spontaneously out of
necessity, as a reaction in times of crisis to the physical shortage of imported goods (like in
wartime), or driven by the inability to import due to hard currency reserve constraints. As a
policy tool, however, it goes much beyond the spontaneous response that occurs when market
players react to price signals using their own dispersed knowledge (Hayek 1945). Quite in
opposition to this Hayekian view, ISI is a top-down planning strategy that distorts prices on
purpose in order to induce deliberate economic outcomes.
Intellectually, this policy was influenced by the works of Prebisch (1950) and Singer
(1950) that assigned a limited role to natural resources in the process of economic development.
Commodities, in their view, have limited capacity to incorporate productivity gains, as opposed
to the industrial sector. As a result, Prebisch and Singer believed in a long-term declining terms
of trade against commodity exporters.3 From this analysis followed policy prescriptions that
1
African countries also embarked in ISI early after independence, but had much less participation in the ideological
formulation of such policies nor had the same bureaucratic apparatus of East Asian or Latin American countries.
For an analysis of the role of the African state in transferring resources from the primary sector to the industrial
and service sector – with its dire consequences – see Bates (1984) for the earlier period and Van de Walle (2001)
for the challenges of structural adjustment, especially in the 1990s.
2
In Schneider’s analysis (1999), cases of desarrollista states.
3
This hypothesis was the subject of various studies that either challenged or confirmed the original predictions,
with results depending on the periods and measures used. Overall, the Prebisch-Singer thesis has not proven to be
empirically robust over time and is based on assumptions that were sound in the 1950s – before the rise of a
5
policymakers should promote the shift of factors of production from a commodity-based
economy to industrial goods production, thus defying Latin America’s natural place in the world
division of labor. The Economic Commission for Latin America (ECLA) spearheaded this effort
and influenced policymakers and a generation of economists, such as Celso Furtado and Maria
da Conceição Tavares.
Baer (1972) considers that only after the World War II ISI became a deliberate policy
tool. In order to achieve a structural change, governments deployed a range of instruments:
protective tariffs, exchange controls, exchange auctions, multiple currencies, subsidized loans,
direct participation of the State in selected industries and in infrastructure projects, and
quantitative ban of imports that had “similar” national copies – to name a non-exhaustive list of
such instruments.
ISI was effective in industrializing a handful of Latin American countries at a high cost.
It created an inefficient industrial park that produced to closed markets using capital-intensive
technology in labor abundant countries. Incentives to protection steer away companies from
seeking foreign markets, which could have reduced unit costs by scale effects in addition to the
salutary effects of international competitive pressures to prices and quality of industrialized
goods. Instead, domestic consumers were a captive market, with few options to choose when
purchasing industrialized goods or capacity to influence import policies. Being many and
dispersed, they had less power to influence policies and product quality than the few and
organized industrial producers who faced limited competition and lower costs to lobby for
service and knowledge economy – but are now questionable. Empirically, most likely commodity prices have no
permanent trend over time (Cuddington et al. 2007).
6
government favors (Olson 1965). Commercial viability could be sustained by seeking favors,
new rounds of protection, and favorable legislation – rent-seeking, in short.
By the late 1960s and early 1970s, analysts and advocates of ISI were already concerned
about the future prospects of this development model (Baer 1972, Hirschman 1968, Tavares
1972). The signs of the limits of the import-replacing model were already clear and called for a
new source of economic dynamism, such as the development of production for mass market.
Latin America’s income inequality was an obstacle to this move, as well-paid workers were both
the labor input of protected industries and the ultimate consumers of their production.
Furthermore, social policy was tilted to support this privileged class of formal workers (Wibbels
and Ahlquist 2011), being another mechanism leading to persistent income inequality. Policies
that could reduce inefficiency of the ISI model, such as sunset clauses on subsidies and
performance requirements, faced considerable economic opposition from affected domestic
sectors, including labor unions.
By the mid-1980s up to the 1990s, adjustment was forced to Latin America as the region
had to fix problems arising from high external debt, inflationary pressures, and microeconomic
distortions of which ISI was partly to blame. Adjustment to the industry was painful, as in the
1960s and 1970s Latin American countries were “overindustrialized”, they had a much higher
share of industry as percentage of the total economy than it would be predicted by its income per
capita in a worldwide sample of countries (Moreira 2007). African countries, which adopted
many of the policies used by the ISI development model, also had to pass a process of adaptation
and market opening (Van de Walle 2001). The so-called Washington Consensus made a direct
attack to the tools used by ISI, such as multiple exchange rates, appreciated currency that favored
7
imports of capital goods to the use of the domestic industrial sector, protectionism, and state
companies (Williamson 1990, Rodrik 2002).
From the 1990s onwards, industrial policy waned and the idea of selecting specific
sectors for state support became discredited and heavily criticized (Young 1995). Instead of
having planners nudging or outright directing capital to specific sectors with the belief that this
would lead to higher economic growth, a strong support emerged for an altogether different
strategy: strengthening and building pro-market institutions and fostering good governance
(Burki and Perry 1998). This institutionalist turn in economic development led to a new round of
policy prescriptions. In addition to macroeconomic reforms that sought to “get the prices right”,
as emphasized the original Washington Consensus, institutional reforms were viewed as
necessary to achieve faster rates of growth. Domestic reforms as well as trade agreements
created new constraints to the implementation of ISI policies.
In this context critical of picking winners, contemporary advocates of industrial policy
turned to horizontal policies that promote competitiveness of the whole economy and not
particular sectors (Canêdo-Pinheiro 2007, Almeida 2013) and the facilitation of investment
coordination and the promotion of new activities (Rodrik 2007).
It is within this background of a prevailing economic policy mindset heavily critical of
sectoral policies and protectionism that LCP has paradoxically grown. It started historically in
Norway in the 1970s but gained momentum and wide adoption only by the late 1990s when
many other countries adopted variants of policies dedicated to privilege local industrial
development in the oil and gas supply chain. In Brazil, the same administration that adopted
liberalizing reforms introduced, in 1997, preferences for local content in oil and gas bidding
rounds. Nigeria, Africa’s biggest oil producer (BP 2014), passed a local content law in 2010.
8
Likewise, Mexico included LCP in its ambitious energy reform approved in 2013, part of a
broader reform package that is liberalizing economic sectors and aims to promote competition. In
the next section, I analyze the growth of LCP and why it differs from ISI in ways that have not
been previously analyzed.
3. Local content policies in the oil and gas (O&G) sector
3.1. Theories about the effects of resource abundance and the growth of LCPs
Once a paradox (Karl 1997), the view that resource abundance can actually play a
harmful effect on economic development and in political institutions is now conventional
wisdom. The perils associated with oil abundance range from being a point-source commodity
that fuels conflicts to its unusually high rents and tendency to foster enclave development
(Frankel 2012, Ross 2012).
An initial economic explanation to why resource-rich countries lagged behind in terms of
economic development focused on the Dutch Disease model and the issues associated with
currency overvaluation (Sachs and Warner 2001). More sophisticated studies, based on political
economy models, stressed the role of institutional quality and political incentives in explaining
when resources result in a blessing or a curse (Mehlum et al. 2006, Jones Luong and Weinthal
2010). Good institutions are viewed as crucial to revenue management, preventing the temptation
of spreading patronage for political gains enabled by windfall profits (Robinson and Torvik
2009) and issues associated with short-term fluctuations in oil revenues (Hausmann et al. 2014).
However, there is more to oil and gas production than just the revenue after the output is sold in
9
the market. There is an industry behind it – and the more production goes to frontier areas, such
as deep offshore or exploration of unconventional resources, the higher is the demand for goods
and services and the lower the rents available, due to higher costs of production.4
The bulk of the resource curse literature still treats indistinguishably the effects of oil
regardless of production costs or sources, what becomes more and more questionable as
unconventional production grows from deep offshore oil, shale gas/tight oil, and oil sands.5
Capital and operation costs vary considerably among different types of fields and resources (see
Table 1), therefore it is not correct to assume the same production costs across all producers.
Table 1: Variability of capital and production costs for selected projects
Country or region
Type of project
Capital cost per barrel Operating
per day of capacity
costs
($ thousand)
($/bbl)
Brazil
Canada
Deepwater pre-salt
Canadian oil sands
with upgrading
Iraq
Onshore super-giant
Kazakhstan
North Caspian
offshore
Saudi Arabia
Onshore generic
expansion
United States
Light tight oil
West Africa
Deepwater
Source: IEA, World Energy Outlook 2013
4
45-55
15-20
100-120
10-15
25-30
2
70-80
15-20
15
90-100
70-80
2-3
8
25-30
As oil rents are equal to price minus production costs.
This brings obvious measurement problems to empirical works. Ross (2012), for example, does a cross-country
study that measures oil rents as a function of total oil production times the yearly international price. Because
production costs vary according to geological conditions and type of resource, this measure will tend to
overestimate the amount of available rents in countries that produce oil in frontier areas (or even the production
of heavy oil, which is sold at a discount in the international market). This accounting method also disregards the
knowledge and skills spillovers involved in the production at technologically challenging conditions. It is fair to say
that this is a subject particularly difficult to measure. B. Smith (2012) identifies eleven different choices of
measuring oil wealth in published papers, from a simple dummy of OPEC countries to oil rents per capita. They all
have some conceptual and/or endogeneity problems. He acknowledges that endogeneity is likely to still be present
in any measure of oil wealth, what calls for recognizing this possibility and try to account for its implication.
5
10
The growth of production from unconventional sources also calls for a renewed look on
the politics of oil abundance. In countries rich in conventional oil, politics tend to be
concentrated on how oil rents are distributed – the who gets what, as the production of
conventional oil provides few job opportunities and require very little in terms of technology,
skills and industrial demand. A different dynamic exists when the demand in the upstream sector
is high – when cost of production is substantial.6 Resources that are untapped by high production
costs lead to a new distributive dynamic in the upstream sector. It becomes a political decision
where the oil platforms, drilling rigs, and supply vessels will come from and the citizenship of
their operators and management. Even countries with very modest industrial capacities, such as
Uganda, Ghana and Kazakhstan have instituted mandates that maximize local purchases of
goods and services and training of personnel (Tordo et al. 2013). Given its power over access to
subsoil resources, governments and sectors of the society can bargain with oil companies to
shape their procurement policies – as shown in the examples listed in Box 1.
As in most industries, every country can do a minimum of local supply with high
efficiency or a lot with high inefficiency. It used to be the case that the market – in this case, oil
companies – decided this mix and set their own procurement policies. LCPs are a trend where
governments are setting procurement and hiring policies for the oil companies, fixing them in
law or license contracts, sold as part of efforts to avoid the resource curse. Although this is a
policy that at first sight brings enormous risks of rent-seeking, it will be argued here that the
competitive nature of the upstream sector limits rent-seeking in comparison to simple rents
6
The oil and gas industry value chain is composed of three segments: upstream, midstream and downstream. The
upstream concentrates the activities of exploration, development and production of oil and gas fields, being the
core of the industry and the focus of this paper and most LCPs. The midstream refers to the trading and
transportation of the crude or gas. Finally, oil refining and petrochemical operations are concentrated in the
downstream (Inkpen and Moffett 2011).
11
distribution or government-supported investments in the downstream sector. In addition, this
high demand for goods and services in the upstream sector opens up the possibility of adopting
assertive industrial policies in the O&G sector, as was historically the case for the North Sea
area, the origin of LCP.7
Box 1: Examples of different local content policies in the oil industry (selected countries)
Brazil
Uganda
Kazakhstan
Nigeria
Oil operators have to
purchase goods and
services with domestic
suppliers as a requirement
for participation in bidding
rounds. For deep water,
operators have to commit
to purchase drilling rigs
and oil platforms that have
a minimum of 55% of
locally produced goods.
Operators have to
undertake the schooling
and training of Ugandan
citizens for staff positions,
including management
positions. In addition,
purchases should
maximize use of local
goods and services, where
available on a competitive
basis.
Bids for new mineral
rights have to include
commitments to
minimum levels of local
content in goods, works,
and services, and the
training of Kazakh
personnel. Ninety-five
percent of employees
have to be Kazakhstani
citizens.
FPSO (offshore) oil
platforms have to be built
with 80% of local
engineering and 50%
local content by tonnage
in the fabrication of
topsides, and 100% of
steel plates. The
government also created
a fund to support the
development of national
suppliers.
Source: ANP (http://www.anp.gov.br/brnd/round10/ingles/conteudo_local.asp), Tordo et al. 2013
3.1 Exit, Voice, and Loyalty applied to LCP
To shed light on how industrial policy done through LCP differs from ISI, it is worth
resorting to Hirschman’s framework developed in Exit, Voice and Loyalty (1970). LCP has
7
High oil prices following the first OPEC embargo made commercially viable the extraction of oil from the offshore
of the North Sea in the 1970s. The North Sea was then a frontier area and the cradle of local content policies that
is now witnessed in other countries. Norway, for example, early on adopted a policy of preferential procurement
of local companies, whenever “competitive on cost and quality,” and promotion of R&D and knowledge exchange
between IOCs and domestic suppliers. While there was no formal minimum amount set in legislation, regulations
and the direct influence of Statoil safeguarded the inclusion of supplier companies, arguably a decisive step in
creating a petroleum cluster in Norway that later went global (Ryggvik 2014).
12
dynamics and incentives that differ from ISI and even from Hirschman’s own model of fostering
forward linkages.8
Exit... is perhaps Hirschman’s most influential work and has been extensively applied to
areas beyond the author’s initial analysis of consumer responses to quality decline in firms and
organizations (Hirschman 1978, Dowding et al. 2000). Traditionally, economists tended to
concentrate on the exit option as a signal that consumers are unhappy with a firm – where the
decline of sales is an indicator that the product or service is less rewarding to consumers than
used to be. Unit sales or revenues are highly visible and easy to measure, thus quickly informing
managers and owners. However, for products with inelastic demand and some organizations
which individuals are associated (by voluntary or compulsory means, including Parent Teacher
Association, a national army or the nation state), exit may be excessively costly. Given the
limited availability of the exit option, consumers and citizens may resort to voice – express
verbally concerns in the expectation that the firm or organization will change its practices. It is
an attempt to change rather than to escape from and can be exercised from a simple petition to
the management team to collective protests.
Both options have advantages and drawbacks. Firms may ignore voice if consumers are
captive – such in a monopolistic condition, where incentives to change are low. On the other
hand, exit is a crude indicator of consumer’s preferences: it does not directly communicate where
the problem of the firm lies (if prices are high, if the quality has deteriorated or if some better
substitute good has appeared in the market). Typically, firms resort to a combination of exit and
voice indicators: the financial bottom line and surveys with consumers and market research.
8
LCP in the oil and gas sector is a specific type of backward linkage in the natural resource sector, affecting the
upstream of the O&G sector.
13
How this concept can be applied to understand the differences between ISI and LCP? The
typical ISI development model grants domestic monopolies to firms in order to stimulate
investments. Consumers are captive to domestic goods with few exit options.9 Particularly for
the consumer goods markets, voice tend to be ineffective, as consumers face high coordination
costs and the dominant position of firms in the market decrease incentives to react to consumer
complaints. A closed market also reduces the repertoire of consumers – the ability to complain
about inferior goods is connected to being exposed to goods of higher quality.
The export-led development model that characterized the Asian Tigers benefited from the
signals provided by consumers located in the developed markets. Firms had to compete on price
(foremost, as entrants to mature markets) and quality (increasingly) to please consumers from the
most sophisticated markets. A rich exchange of information between bureaucrats and
industrialists facilitated the identification and solution of bottlenecks and achievements of
developmental goals (Evans 1995). Voice flowed between bureaucrats and business actors, with
the latter having to react to market signals coming from the exit option of consumers in the
markets they exported.
LCP has an interesting combination of partial domestic exit, foreign market signals and
voice. One pivot of the policy is to limit exit – to increase the cost of contracting some goods
outside of the domestic market. However, the exit door is only semi-closed, which means that it
is semi-open. Quantitative indicators of local content mandates establish a global goal but allows
for the oil operator some room of maneuver to manage the achievement of the target. For
example, a goal of spending 65% of the total cost of an oil tanker internally can be fulfilled
9
In addition to the few options provided by firms in the domestic oligopolistic markets, smuggling is a type of exit
option, but with a potential high cost and limited scale.
14
through different combinations of inputs. In addition, legislations typically allow for goods that
cannot be found in domestic markets, or only at a prohibitive cost, to be waived, thus providing
an exit option for specific components.
Furthermore, LCP cannot insulate domestic markets from the international price
fluctuation of oil. LCP thrives where cost of production is high – where oil operators need to
invest considerable amounts in capital goods and services before oil can be produced. Such fields
are the ones most exposed to price swings. Frontier areas become economically viable by the
combination of high prices and technological gains, but risk being the first to shut down when
prices swing back. Inefficiency can be sustained under a scenario of high-prices, where
extraordinary future profits of oil production compensates the costly requirement to buy from
less competitive suppliers. However, as price goes down and a project for a given oil field
approximates the break-even point, the risk that a future investment will not pay off rises and oil
operators will refrain from undertaking new projects. More formally, let 𝑒 be the expected
earnings, 𝑦 the international price of oil, 𝑐 the exploration and production costs and 𝑧 the
premium paid to fulfill the local content requirement: 𝑒 = 𝑦 − (𝑐 + 𝑧). For 𝑒 to be positive, 𝑦
has to be higher than 𝑐 + 𝑧. As the premium paid to purchase locally increases, 𝑧, exploration
and production costs become closer to the break-even point and enters in the negative expected
earnings region, thus shutting down investments. It is clear that for the same geological
characteristics, O&G fields with higher costs due to 𝑧 will shut down before similar fields
operating in more competitive markets for goods and services. Low oil prices affect every
frontier area, but affects first the less competitive countries with local content mandates.10
10
Exploration and production costs are the combination of finding costs (expenditures on exploration,
development, and unproved property acquisition) and lifting costs (costs to get the oil out of the ground).
15
Note that 𝑧 can originate from a variety of sources, such as lower worker productivity due
to schooling gaps, infrastructural deficiencies that increase transportation costs, less developed
financial markets, bureaucracy on doing business and pure rent-seeking of domestic
entrepreneurs and politicians. Regardless of the origin, the bottom line is that the sustainability of
LCP requires reducing the difference between local and international prices of domestic goods,
otherwise new investments will be prohibitively costly when prices go down. Alternatively,
governments can abandon a LCP to keep the rate of investment in oil production and the
associated hard currency earnings. Unlike ISI, LCP has an automatic exogenous pressure to costreduction to reach similar levels of international prices for domestic goods – or abandoning the
policy altogether.11
The voice aspect is present in the relationship between oil companies and domestic
suppliers. Oil companies have a stake in creating capabilities that will lead to higher quality and
reduced costs in its supply chain. They might well oppose LCP, but if accepting them is part of
the conditionality of accessing oil reserves, they will have to engage with local suppliers – or
avoiding investing altogether. Therefore, we should expect the buildup of capabilities due to the
presence of a single principal (the oil company) with high competence to monitor its local
suppliers and labor force.
The next section analyzes the Brazilian case of LCP. We review how this policy was
adopted in Brazil, its growth and how oil companies are working with suppliers combing
strategies of voice and exit. In the sequence, we address the recent Petrobras scandal, analyzing
bribes related to LCP projects and the rest of the oil company.
11
Which would also be a form of exit.
16
4. Brazil’s oil abundance and LCP
4.1. Genesis and growth of local content
Brazil shares with many oil-rich countries history of resource nationalism, as it was the
first country in the world to have nationalized oil exploration and production even when there
was not a drip of oil under production (Smith 1976). Petrobras, Brazil’s National Oil Company
(NOC), was founded in 1953 with the mission to explore and develop Brazil’s oil resources.
Contrary to the optimistic expectations that existed, Brazil’s onshore basins proved to be of
limited potential. It was only when the company set out to sea that production boosted. The first
offshore discovery was made in 1968, in Sergipe, but Petrobras would have to wait until 1976 to
drill in the Campos basin, what showed to be the biggest oil province of the country. The
Brazilian state-owned enterprise (SOE) progressively went further offshore, registering some of
the world records in that kind of oil operation. Notwithstanding that, production was still far
below the national consumption.
In 1995, a Constitutional Ammendment allowed the federal government to license
exploratory areas to other companies, breaking Petrobras’ monopoly on the exploration and
production (E&P) sector. The opening proved to be good for the company: productivity
increased as well as production (Bridgman, Gomes and Teixeira 2011). The NOC kept their
leading role in the country, leveraging from partnerships with foreign companies to invest in
capital intensive ultra deep water projects. The strategy paid off very well when in 2007 was
announced the discovery of a new oil province located in the pre-salt of the Campos and Santos
basins, 300 km offshore and at water depths of 2,000 meters. Thanks to the pre-salt discoveries,
Brazil is expected to surpass Mexico and Venezuela in oil production by the next decade and
become an oil exporter.
17
While Petrobras had a history of developing local suppliers (Dantas and Bell 2011), a
formal local content requirement was only established in law in 1997. The requirements were
progressively increased in bidding rounds for new oil exploration and comprehend both private
operators (such as Shell, BG, Esso) and the NOC (see Annex 1 for a timeline of LCP in Brazil).
As previously addressed, developing oil under conditions of high-cost of production
subject procurement policies to politicization. In the mid-1990s, the growth of offshore oil
production in the country led Petrobras to purchase 12 new platforms, all to be built abroad. This
sparked criticism of labor unions and business associations. The Federation of Industries of the
State of Rio de Janeiro (Firjan) reacted creating, in 1998, a movement to increase the
competitiveness of national companies of the oil and gas sector so they could become suppliers
of Petrobras. The then newly established Oil Regulatory Agency (ANP) hired PUC-RJ, a
prestigious university, to study the conditions necessary to increase the participation of the
national industry in the supply chain. The study identified a number of mechanisms to achieve
this end. Finally, potential suppliers organized one specific business association (Organização
Nacional da Indústria do Petróleo, ONIP) to articulate the maximization of the national supply
industry under competitive conditions and to lobby for preferential policies. In the sequence,
ANP changed the bidding rules of future oil exploratory license auctions, increasing the
minimum local content requirements and creating a list of equipment that could be sourced
locally. The result was indeed an increase in the role of local content in awarding exploratory
blocks and in the capacity of ANP in enforcing the clause (Fernández y Fernández 2013).
Brazil now stands in the world with the most ambitious local content program, with
dozens of drilling rigs, floating production storage and offloading (FPSO) platforms, oil tankers
and platform supply vessels (PSVs) being constructed at the same time in brand new shipyards
18
that were built specifically to supply the demand of the oil sector (Tordo et al. 2013, Foster et al.
2013). The local content policy has been costly to oil operators but has strong constituencies in
the Executive, in the Congress (where a specific caucus has been established12) and business
groups and unions that are directly benefited. More than 200 thousand workers are being trained
through Prominp, a public program that was designed to meet the challenges of the growth of the
oil sector in the country.13
Petrobras has a dominant position in the Brazilian oil market – with around 90% of the
share of oil production – and is the operator of the main oil fields and future projects (ANP
2014). Therefore, the company and its partners (such as BG and Galp) are the most affected by
the LCP, although the policy is also valid for International Oil Companies (IOCs) that directly
operate oil fields in Brazil, like Shell and Chevron. By far, Petrobras is the company who has to
deal the most with local suppliers in order to abide by regulations.
In many instances, the relationship between the NOC and local suppliers are far from
harmonious. Petrobras had to face considerable delays when contracting with local shipyards
(Campos Neto 2014). While suppliers are subject to fines when projects are delayed, this may
not be sufficient to compensate the total loss of the oil operator. To speed up deliveries,
Petrobras developed a program to closely monitor shipyards, including taking measures to
increase productivity gains and improve engineering and management practices (Pires et al.
2013), examples of using voice to influence firms. For instance, in the Rio Grande Shipyard
12
The Frente Parlamentar Mista da Indústria Marítima, supported by 210 deputies (40% of the house) and 10
senators (12% of the higher chamber).
13
Prominp is a forum that identifies and propose solutions to remove bottlenecks related to industrial capacity and
human resources (Almeida, Lima-de-Oliveira and Schneider 2014).
19
(ERG), the firm even imported Chinese labor to work on the Brazilian site replacing Brazilian
workers in order to increase productivity in labor-intensive operations (Azevedo et al. 2014).
When these efforts were insufficient, the NOC decided to remove orders from local
shipyards and put them abroad – exiting from contracts. The cost of this option is the payment of
fines to the Oil Regulatory Agency (ANP), but the benefit is guaranteeing the delivery of
platforms and with that additional oil production. This happened in 2013, when Petrobras
decided to transfer to a Chinese shipyard (Cosco) part of the orders that were originally to be
made in the Inhaúma and ERG sites.14 Similar issues happened with other shipyards, such as
Atlantico Sul (EAS), which in 2012 had to bear a loss of R$ 1,4 billon because the company had
to more than double the number of employees, from 5,000 to 11,000, in order to accelerate the
delivery of oil tankers to Petrobras, after the first tanker was delayed by more than 20 months.15
Petrobras and its suppliers have been featured in the news not only about new orders and
measures to increase productivity. Far from that. In 2014, with the launch of the Operation Car
Wash (Operação Lava Jato) by the Federal Police, a major corruption scandal involving
Petrobras and all its large suppliers was revealed. A former director and a senior manager of
Petrobras entered in leniency agreements with public prosecutors, agreeing to return millions of
dollars from Swiss bank accounts and provide evidence that could be used to prosecute others
Petrobras’ employees, politicians and businessmen. The documents, which have been gradually
14
See also “Petrobrás vai à China para evitar atraso na produção. Fornecedor local reclama”, Estado de São Paulo,
02/25/2013.
15
See “Estaleiro Atlântico Sul teve prejuízo de R$ 1,4 bilhão em 2011”, Jornal do Commercio, 04/26/2012. In 2012
and 2013 the company was still having losses of over R$ 100 million, according to its financial statement available
at http://www.estaleiroatlanticosul.com.br/eas/pt/relatorio/bp2013.pdf.
20
disclosed by the courts as the investigation proceeds, reveal in detail the corruption scheme of
Petrobras, including which contracts paid higher bribes, topic that will be addressed next.
4.2. Rent-seeking and LCP: the Brazilian case
Local content policies create rents in order to stimulate companies to shift factors of
production from its current use to produce goods and services to supply the O&G sector. In the
common metaphor, rents are the carrots used by policymakers that aim to guide entrepreneurs to
sectors that are believed to show positive externalities. Sectoral industrial policies necessarily
involve the offering of carrots – which is largely carried out by the creation of rents, such as
subsidies or protection. However, effective industrial policy also depends on exercising the
power of stick to curb incentives to non-performing companies, such as sunset clauses (Rodrik
2007).
Industrial policy is easier said than done and from the outcomes sketched at the draw
board of policymakers to the actual functioning of the incentive schemes on the ground there is a
messy political process that should not be taken for granted. Rent creation by governments,
unsurprisingly, may result in rent-seeking – unless shielded by strong institutions that includes
transparency requirements and punishment of corruption.
As Rodrik (2007, p 50) puts it, “For every South Korea, there are many Zaires where
policy activism is an excuse for politicians to steal and plunder.” Starting with Luiz Inácio Lula
da Silva’s tenure, Petrobras was used as a tool to pursue industrial policy and regional politics,
such as allocating investments of refineries and petrochemicals in states governed by political
allies. This was no a secrecy – in several speeches and interviews President Lula and the then
21
minister of Mines and Energy, Dilma Rousseff, praised Petrobras for investing in Brazilian
companies.16 However, what was only recently discovered was a major kickback scheme where
a pool of contractors, organized in a cartel, paid up to 3% of the valor of large contracts to the
senior management of Petrobras and parties of the ruling coalition (PT, PMDB, PP).
In studies about corruption it is rarely the case that a researcher has access to direct
evidence and is able to compare the cost of corruption and factors that influence the level of
corrupt behavior (Olken and Pande 2012, Treisman 2007). In the case of Petrobras, given the
evidence that came to light in the recent months, it is possible to compare which projects paid the
highest amount of bribes and what where the mechanisms of payment and negotiations between
corrupt executives, politicians and contractors.
Data for this subsection comes from the recently disclosed confidential statement of
Pedro Barusco, former senior manager of the Engineering and Services (E&S) directorate of
Petrobras. Among the documents that he gave to the Federal Police and public prosecutors was a
spreadsheet with 88 large projects under his supervision, with detailed information on the project
type, date, the total amount, the bribe rate, the companies involved, how the bribe was to be split,
and the company contacts who made the payment. Barusco’s division was responsible for the
bulk of large projects of Petrobras, although some projects were directly contracted by other
divisions, such as the International, Gas and Energy and Downstream directorates. Therefore,
although this spreadsheet does not cover all the corruption found in Petrobras, it covers the
majority of it. This evidence is supplemented by Barusco’s own statement to the latest inquiry
committee in charge of investigating Petrobras (CPI da Petrobras), statements made by others
16
Lula even said that the government pressured Petrobras to invest in refineries as part of the government’s
countercyclical investment after 2008, over objections of the company’s executives (Valor 09-17-2009).
22
suspects who entered in leniency agreements, petitions made by public prosecutors and judicial
decisions.
Thirty-five companies are listed17, of which 8 are international or local subsidiaries from
a variety of countries.18 The spreadsheet has a variety of projects from 2013 to 2010, including:
parts of two new refineries and upgrades of existing ones, gas pipelines, local headquarters,
expansion of Petrobras’ R&D center, oil platforms. The last group of item is the only one
directly affected by LCP, while all others have no binding mandate to fulfil a specific target of
domestic purchase. It is possible, therefore, to compare the bribe rate of products where
Petrobras had the incentive to procure domestically due to LCP versus others which the company
freely contracted. Barusco’s accounting lists only projects where the bribe rate ranged from 1%
to 2%, while in the statement made by the former director of downstream, Paulo Roberto Costa,
bribes went from 1% to as high as 3%.19 If anything, the difference between the downstream and
upstream segment will be understated, since the projects handled by the downstream director had
a bribe rate even higher – and oil platforms and drilling rigs, LCP requirements, pertain to the
upstream segment.
Another difference between contracts from the local content requirement and other
activities of Petrobras is that platforms, drilling rigs and their components are priced in dollar.
17
Alusa, Andrade Barbosa Melo, Gutierrez, Bueno, Camargo Corrêa, Carioca, CNEC, Contreras, Construbase,
Construcap, EAS, EBE, EIT, Engeform, Engevix, Floatec, Galvão, GDK, Keppel Fels, Mendes Jr., MPE, OAS,
Odebrecht, Odebei, Promon, Queiroz Galvão, Quip, SBM, Schain, Setal, Skanska, Techint, Tomé, Toyo Setal, UTC.
18
Argentina (Contreras), Sweden (Skanska), Japan (Toyo), Singapore (Keppel Fels), United States (Quip), Nederland
(SBM), Italy (Techint), Germany (Hotchief). CNEC was originally a Brazilian company but was acquired in 2010 by
the Australian group WorleyParsons. CNEC appears in Barusco’s spreadsheet in a 2009 contract in consortium with
Camargo Correa, thus one year before it was acquired.
19
Costa was appointed to the board of Petrobras by politicians from PMDB and PP, and the additional 1
percentage point on bribes from his directorate was used to channel resources to himself and these parties. The
lion’s share of corruption in Petrobras was within the engineering directorate, which had bribes split between PT
and the employees appointed by PT leaders.
23
These are tradable goods, with international price and suppliers, while domestic construction
services hired by Petrobras are priced in Brazilian reais (BRL). Although they can have imported
components, construction services are purchased domestically and paid in BRL. On the other
hand, capital goods for the offshore oil industry follows closely international maritime and safety
standards and can be exported, as they currently are.20 These goods are mobile, standardized, and
provided by a competitive international market where price differences can easily be spotted.
Construction services, especially in complex projects, have more room for customization.
Specific terrain challenges, availability of local infrastructure and climatic events contribute
towards losses during construction that can be contingent and variable. These characteristics of
being unique goods facilitates over-invoicing and overpayment because it is more difficult to
detect bribes (Shleifer & Vishny 1993). Therefore, in political economy models of bribe
extraction, bureaucrats are assumed to choose optimal bribe rate taking into account market
forces and the probability of being caught (Olken and Pande 2012, Shleifer & Vishny 1993).
These factors would lead us to expect that bribes in the tradable goods would be lower
than in other segments of Petrobras, such as building new refineries. To test this hypothesis, we
regressed the data of bribe rate on the type of contract, which is a dummy for tradable goods.
Figure 1 presents the difference of means of the variation on bribe rate with confidence intervals:
20
Oil tankers and platforms built in Brazil are formally exported out of the country and then imported as part of a
tax break regime called Repetro that was created in 1999.
24
We also controlled for the value of the contract, as 1% of a very large contract can reward
more than 2% of a smaller contract. Alternatively, a large contract may already contain a higher
bribe percentage.21 Full regression results are displayed in Table 2 below, which uses three
different specifications:
Table 2: Corruption regressions
Intercept
Tradable/LCP
Contract value (in 100 million BRL)
Tradable/LCP*Contract value
Model 1
1.60***
(0.061)
-0.567***
(0.073)
Model 2
1.55***
(0.077)
-0.71***
(0.070)
0.11
(0.065)
Model 3
1.43***
(0.087)
-0.41***
(0.092)
0.304***
(0.070)
-0.29*
(0.07)
Robust standard errors in parentheses
21
Contracts priced in dollar were converted to BRL using the currency exchange rate of the day the contract was
signed. In the list, R$ 48 billion of projects are priced in BRL and US$ 10.7 billion in dollars.
25
As is possible to see, results are statistically and substantively significant and in line with
the theory outlined here.22 Non-tradable goods, that operate in a more competitive market even
when benefited by LCP, paid less in bribes than other contracts made by Petrobras. Model 2
shows that contract value is significant at the 10% level (t = 1.77). Once an interactive term is
added, as is done in Model 3, contract value is also significant at the 1% level (t = 4.34).
For a contract of R$ 100 million, the predicted bribe rate from Model 3 is of 1.73% for
non-tradable goods (1.43 + 0.304 = 1.734) and 1% for goods purchased under LCP (1.43 – 0.41
+ 0.304 – 0.29 = 1.034). Given the interaction effect, higher contract values shows the bribe rate
of non-tradable goods increasing faster than goods purchased under LCP. This suggest that there
are more rents to be shared with politicians and Petrobras’ executives outside of the LCP
contracts than through this policy.
Statements made by Barusco to the Federal Police and to the congressional inquiry that is
investigating the Petrobras scandal corroborates that interpretation. The data presented above
refers to contract made directly by Petrobras, but the NOC helped to set up a new company
specifically to purchase new drilling rigs to exploit the pre-salt area, Sete Brasil, formed in 2010.
Barusco was one of the first executives of Sete Brasil after retiring from Petrobras. According to
him, he had tough negotiations with national shipyards in order to bring prices to international
standards. Sete Brasil is a portfolio manager of O&G assets and has a contract to lease 28 drilling
rigs to Petrobras, with the commitment of purchasing them in Brazil, following the LCP. Sete
Brasil’s total investment was projected to be US$ 25.7 billion. According to him, bribe rates in
22
From the 88 fields, there is missing data on bribe rate for 12 projects – only one related to the upstream
segment and quoted in dollars, a contract with the Dutch SBM.
26
the Sete Brasil contracts were reduced to less than one percent in some cases in order to bring
prices closer to Asian producers and due to pressures coming from Petrobras. 23
Interpretation of this data has to be very careful. A thorough analysis of corruption in
Petrobras is worth a full paper – already under preparation. However, the evidence so far
suggests that corruption in the NOC did not start nor was higher, in percentage terms, due to
LCPs. The corrupt behavior actually matches political economy models that predict optimal
bribe rate subject to market constraints and monitoring capabilities. Local content contracts are
subject to three forces that constrain rent-seeking: 1) goods are priced in the international market,
so price monitoring is easier; 2) local suppliers are still in their learning curve trying to match
more competitive producers, so their lower efficiency means that there is less rents to be
shared24; 3) they are more exposed to international price fluctuations, putting their whole
business in jeopardy when the price of oil goes down and the LCP policy risks making domestic
oil production impracticable.
5. Conclusion
23
It is worth quoting in lengthy the following passages from his “Termo de colaboração N. 01”, made to the
Federal Police on 11/20/2014: “…Que a SETE BRASIL é uma empresa privada que nasceu de um projeto da
Petrobras para construir sondas de perfuração no Brasil; Que este projeto foi montado inicialmente dentro da
Petrobras... Que para ofertar essas 21 sondas à Petrobras, a SETE BRASIL negociou 21 contratos de construção com
vários estaleiros... Que o declarante esteve à frente desta etapa de negociações a fim de buscar preços
competitivos com as sondas ofertadas no mercado asiático, pois isso era uma exigência da Petrobras... Que a
licitação não foi direcionada para que a SETE BRASIL ganhasse... Que havia uma combinação de pagamento de 1%
de propina para os contratos firmados entre a SETE BRASIL e cada um dos estaleiros, mas esse percentual foi
reduzido em alguns casos para 0,9% por conta da competitividade do processo licitatório e a exigência da
Petrobras de que os preços estivessem em conformidade com os do mercado asiático...”
24
Another evidence of this mechanism is that the only bribe rate for oil platforms above 1% in the Barusco dataset
comes from the Keppel Fels shipyard, from Singapore, which opened a factory in Brazil, bringing their own
expertise in this business. All other shipyards are owned by Brazilian companies.
27
The growth of oil production from hard to access reserves, such as ultra-deep offshore,
creates a demand for capital goods and services in the oil industry. The more oil exploration and
production go to frontier areas, the higher the industrial demand component of oil production
vis-à-vis oil rents. The latter has been the traditional focus of the resource curse literature, while
the former is a new political and economic dynamic that deserves a closer look. Governments
and sectors of the society attempt to influence how this industrial demand will be met through
policies that require oil operators to procure domestic goods by adopting local content policies
(LCP).
At first sight, LCP resembles the Import Substitution Industrialization (ISI) by protecting
domestic sectors from international competition. This paper reviewed and contrasted the
characteristics of ISI and LCP, using Hirschman’s Exit, Voice, and Loyalty framework. We
showed that LCP combines partial or total exit with strong voice of oil operators in leading with
suppliers. It is not plain ISI – it has its own characteristics and represents a new political aspect
of resource abundance and industrial policy.
In addition to showing evidence from Brazil on how Petrobras has been dealing with
suppliers, we analyzed data from 88 large contracts that the company made with suppliers and
involved bribes. Regression results strongly suggest that LCP suppliers figure among the lowest
bribe givers of the sample. We explained these findings through the more competitive nature of
LCP goods versus non-tradable contracts and models of political economy of corruption. Results
are in line with the theoretical expectations presented here.
It is a known fact that oil-rich nations tend to rank among the least transparent and most
corrupt, with few exceptions (Treisman 2007, Ross 2012). Brazil is a newcomer to the club of
oil-rich nations and the issue of oil wealth, public policies and corruption will certainly occupy
28
many studies in the years to come. Micro-evidence from royalties and Brazilian municipalities
tend to show waste of resources, lower rate of growth and less transparency (Caselli and
Michaels 2013, Monteiro and Ferraz 2009, Oliveira 2009, Oliveira and Alonso 2014). The
effects of LCP so far has not been widely studied – in Brazil and abroad. In the Brazilian case,
LCP suppliers stand among the lowest bribe givers to Petrobras, what can be explained by the
characteristics of the goods produced. However, this position is certainly not going to help the
continuation of this industrial policy as the public impatience with corruption grows. To the
challenges previously highlighted, such as achieving international standards of productivity, LCP
as a policy will ultimately rests on the capacity of conducting clean business in Brazil and
abroad, what involves strengthening accountability institutions and preventing corruption.
29
References
Agência Nacional de Petróleo (ANP). 2014. Anuário Estatístico Brasileiro do Petróleo, Gás Natural e
Biocombustíveis: 2014. Rio de Janeiro: ANP.
Almeida, Mansueto. 2013. “Padrões de política industrial: a velha, a nova e a brasileira.” in O Futuro da
Indústria No Brasil, ed. Edmar Bacha and Monica de Bolle Baumgarten. Rio de Janeiro:
Civilização Brasileira.
Almeida, Mansueto, Renato Lima-de-Oliveira, Ben Ross Schneider. 2014. “Política industrial e empresas
estatais no Brasil: BNDES e Petrobras.” in Capacidades Estatais e Democracia: Arranjos
Institucionais de Políticas Públicas. Ed. Alexandre de Ávila Gomide e Roberto Rocha C. Pires.
Brasília: IPEA.
Azevedo, Oscar, João Guilherme Rebechi, and Brigitta Struck. 2014. “Sistema Regional de Inovação –
Rio Grande do Sul. Petróleo & Gás, Naval e Offshore.” Caxias do Sul: OAZ LTDA.
Bates, Robert H. 1981. Markets and states in tropical Africa: the political basis of agricultural policies.
Berkeley: Univ of California Press.
Baer, Werner. 1972. “Import Substitution in Latin America: Experiences and Interpretations.” Latin
American Research Review 7: 95–122.
BP. 2014. Statistical Review of World Energy. Available at
http://www.bp.com/en/global/corporate/about-bp/energy-economics/statistical-review-of-worldenergy.html.
Bulmer-Thomas, Victor. 2003. The economic history of Latin America since independence. Cambridge:
Cambridge University Press.
Burki, Shahid Javed and Guillermo E. Perry. 1998. Beyond the Washington Consensus: Institutions
Matter. Washington (DC): The World Bank.
30
Bridgman, Benjamin, Victor Gomes, and Arilton Teixeira. 2011. “Threatening to Increase Productivity:
Evidence from Brazil’s Oil Industry,” World Development, Volume 39, Issue 8. August 2011,
Pages 1372-1385, ISSN 0305-750X.
Campos Neto, Carlos Alvares da Silva. 2014. “Investimentos na indústria naval brasileira 2000-2013” In
Carlos Alvares da Silva Campos Neto and Fabiano Mezadre Pompermayer Eds. Ressurgimento
da indústria naval no Brasil : 2000-2013. Brasília: IPEA.
Canêdo-Pinheiro, Mauricio, Pedro C. Ferreira, Samuel de Abreu Pessoa, and Luiz Guilherme Schymura
de Oliveira. 2007. Por que o Brasil não precisa de política industrial. No. 644. FGV/EPGE
Escola Brasileira de Economia e Finanças, Getulio Vargas Foundation (Brazil).
Caselli, Francesco, and Guy Michaels. 2013. "Do Oil Windfalls Improve Living Standards? Evidence
from Brazil." American Economic Journal: Applied Economics, 5(1): 208-38.
Chang, Ha-Joon. Kicking away the ladder: development strategy in historical perspective. London:
Anthem Press, 2002.
Cuddington, John T., Rodney Ludema, and Shamila A. Jayasuriya. 2007. “Prebish-Singer Redux.” In
Daniel Lederman and William F. Maloney, eds., Natural Resources: Neither Curse Nor Destiny.
Stanford University Press and the World Bank.
Dantas, E., & Bell, M. 2011. “The Co-Evolution of Firm-Centered Knowledge Networks and Capabilities
in Late Industrializing Countries: The Case of Petrobras in the Offshore Oil Innovation System in
Brazil.” World Development, 39(9), 1570–1591.
Dowding, K., John, P., Mergoupis, T., & Van Vugt, M. 2000. “Exit, Voice and Loyalty: Analytic and
Empirical Developments.” European Journal of Political Research, 37(4), 469–495.
Evans, Peter. 1995. Embedded Autonomy: States and Industrial Transformation. Princeton: Princeton
University Press, 1995.
Fernández y Fernández, Eloi. 2013. “Indústria nacional de bens e serviços nos arranjos produtivos do
setor de óleo e gás no Brasil” in Fábio Giambiagi and Luiz Paulo Velloso Lucas ed., Petróleo:
Reforma e Contrareforma do Setor Petrolífero Brasileiro. Rio de Janeiro: Elsevier.
31
Foster, Maria das Graças, Paulo Alonso, Edival Dan Junior, Adília de Assis. 2013. Retomada da indústria
naval e offshore do Brasil 2003-2013-2020: visão Petrobras. Rio de Janeiro: Petróleo Brasileiro
S/A.
Haggard, Stephan. 1990. “Pathways from the periphery: The politics of growth in the newly
industrializing countries.” Ithaca: Cornell University Press.
Hausmann, Ricardo, Brad Cunningham, John Matovu, Rosie Osire and Kelly Wyet. 2014. “How should
Uganda grow?”. CID Working Paper No. 275. January 2014.
Hayek, F. 1945. “The use of knowledge in society.” The American Economic Review, 35(4), 519–530.
Hirschman, Albert O. 1968. "The political economy of import-substituting industrialization in Latin
America." The Quarterly Journal of Economics: 1-32.
Hirschman, Albert O. 1970. Exit, voice, and loyalty: Responses to decline in firms, organizations, and
states. Cambridge (MA): Harvard University Press.
Hirschman, Albert O. 1978. “Exit, Voice, and the State.” World Politics, 31, pp 90-107.
Inkpen, Andrew and Michael H. Moffett. 2011. The Global Oil & Gas Industry: Management, Strategy &
Finance. Tulsa (OK): PennWell Corporation.
Karl, Terry Lynn. 1997. The paradox of plenty: oil booms and petro-states. Berkeley: University of
California Press.
Mehlum, Halvor, Karl Moene, and Ragnar Torvik. 2006. “Institutions and the Resource Curse.”
Economic Journal 116, no. 508: 1-20.
Monteiro, Joana, and Claudio Ferraz. 2009. “Resource Booms and Voracious Politicians: The Effects of
Oil Shocks on Patronage, Rent-Seeking, and Elections,” Unpublished, PUC-Rio
Moreira, M. 2007. “Fear of China: Is There a Future for Manufacturing in Latin America?” World
Development, 35(3), 355–376.
Oliveira, Renato Lima. 2011. “Dealing with plenty: Brazil in the era of surplus oil.” Master's thesis,
University of Illinois at Urbana-Champaign.
32
Oliveira, Renato Lima and Martin Liby Alonso. 2014. “Fueling development? Assessing the Impact of
Oil and Soybean Wealth on Municipalities in Brazil,” Paper presented at the annual meeting of
REPAL, Santiago, Chile.
Olken, B. a., & Pande, R. 2012. “Corruption in Developing Countries.” Annual Review of Economics,
4(1), 479–509.
Olson, Mancur. 1965. The logic of collective action. Cambridge, MA: Harvard University Press.
Pires, Roberto Rocha, Alexandre de Ávila Gomide, and Lucas Alves Amaral. 2013. “A ver navios? A
revitalização da indústria naval no Brasil democrático.” Texto para discussão No. 1864. Brasília:
IPEA.
Postali, F. a S. 2009. “Petroleum royalties and regional development in Brazil: The economic growth of
recipient towns.” Resources Policy, 34(4), 205–213.
Prebisch, Rául. 1950. The Economic Development of Latin America and its Principal Problems. Lake
Success (NY): United Nations Department of Social Affairs.
Robinson, J. a., & Torvik, R. 2009. “A political economy theory of the soft budget constraint.” European
Economic Review, 53(7), 786–798.
Rodrik, Dani. 2002. "After neoliberalism, what?." Remarks at the BNDES seminar on New Paths of
Development, Rio de Janeiro, September. 2002.
Rodrik, Dani. 2007. “Industrial Policy for the Twenty-First Century.” In One Economics, Many Recipes.
Princeton: Princeton University Press.
Ross, Michael. 2012. The Oil Curse: How Petroleum Wealth Shapes the Development of Nations.
Princeton (NJ): Princeton University Press.
Ryggvik, Helge. 2014. Construindo uma indústria nacional de petróleo offshore: a experiência da
Noruega. Rio de Janeiro (RJ): Elsevier Editora.
Sachs, J., Warner, A., 2001. “The curse of natural resources.” European Economic Review 45, 827–838.
Schneider, Ben Ross. 1999. “The Desarrollista State in Brazil and Mexico.” In Meredith Woo-Cumings,
ed., The Developmental State. Ithaca: Cornell University Press, 1999.
33
Shleifer, Andrei and Robert W. Vishny. 1993. “Corruption.” The Quaterly Journal of Economics, 108.3
(August): 599-617.
Singer, Hans. 1950. “The Distribution of Gains Between Investing and Borrowing Countries.” The
American Economic Review, Vol. 40, No. 2.
Smith, Benjamin. 2012. “Measuring the Resource Curse: Revisiting the Politics of Oil Wealth.” Working
Paper.
Smith, Peter Seaborn. 1976. Oil and Politics in Modern Brazil. Toronto: The Macmillan Company.
Tavares, Maria da Conceição. 1972. Da substituição de importações ao capitalismo financeiro: ensaios
sobre economia brasileira. Rio de Janeiro: Zahar Editores.
Tordo, S., Warner, M., Manzano, O., & Anouti, Y. 2013. Local Content Policies in the Oil and Gas
Sector. Washington (DC): World Bank.
Treisman, D. 2007. “What Have We Learned About the Causes of Corruption from Ten Years of CrossNational Empirical Research?” Annual Review of Political Science, 10(1), 211–244.
Van de Walle, Nicolas. 2001. African Economies and the Politics of Permanent Crisis, 1979-1999. New
York: Cambridge University Press.
Wibbels, E., & Ahlquist, J. S. 2011. “Development, Trade, and Social Insurance.” International Studies
Quarterly, 55(1), 125–149.
Williamson, John. 1990. "What Washington means by policy reform." Latin American adjustment: How
much has happened 7 (1990): 7-20.
Young, A. 1995. The tyranny of numbers: confronting the statistical realities of the East Asian growth
experience. The Quarterly Journal of Economics. Vol. 110, No. 3 (Aug., 1995), pp. 641-680.
Annex 1: Oil regulations and local content initiatives in Brazil
Timeline of oil regulation and LC in Brazil (1997 to present)
Density of LC regulations
1-4 bidding rounds
Free offers of LC in bidding rounds
Count towards selecting winners
Promef 1
Transpetro / Petrobras
Program to purchase 26 oil tankers with 65% of LC
PPSA is created
E&P regulations approved
Discovery of the pre-salt
Debate over new regulations Petrobras as the sole operator
-
-
1997-01 1997-09
1998-06 1999-02
-
2002-08 2003-04 2003-12 2004-08 2005-04 2005-12 2006-08 2007-04 2007-12 2008-08 2009-04 2009-12 2010-08 2011-04 2011-12 2012-08 2013-04 2013-12 2014-08
-
-
Opening of the oil sector
Creation of ANP
Regulatory Agency
-
-
1999-11 2000-07 2001-03 2001-11
First auction of PSA
Libra
BR, Total, Shell, CNPC, CNOOC
-
7+ rounds
Minimum and maximum of LC
Required to all operators and now audited
5-6 rounds
ANP sets a minimum requirement of LC for all oil operators
Count towards selecting winners
-
Promef II
Transpetro / Petrobras
Purchase of 23 new oil tankers
-
Suezmax João Cândido
Promef
Delivery of first Suezmax