ARTICLE IN PRESS

Energy Policy 34 (2006) 3027–3040

www.elsevier.com/locate/enpol

Oil refining expansion criteria for Brazil

Marina Elisabete Espinho Tavares, Alexandre Salem Szklo, Giovani Vitória Machado,
Roberto Schaeffer, Jacqueline Barboza Mariano, Janaı́na Francisco Sala
Energy Planning Program, Graduate School of Engineering, Federal University of Rio de Janeiro, Centro de Tecnologia, Bloco C, Sala 211, Cidade
Universitária, Ilha do Fundão, Rio de Janeiro 21941-972, Brazil
Available online 11 July 2005

Abstract

This paper assesses different strategies for the expansion of Brazil’s oil refining segment, using criteria that range from energy
security (reducing imports and vulnerability for key products) through to maximizing the profitability of this sector (boosting the
output of higher value oil products) and adding value to Brazil’s oil production (reducing exports of heavy acid oil). The
development prospects are analyzed for conventional fuel production technology routes, sketching out three possible refining
schemes for Brazilian oil and a GTL plant for producing gasoil from natural gas. Market scenario simulations indicate that
investments will be required in Brazil’s oil refining segment over and above those allocated to planned modifications in its current
facilities, reducing the nation’s vulnerability in terms of gasoil and petrochemical naphtha imports. Although not economically
attractive, oil refining is a key activity that is crucial to oil company strategies. The decision to invest in this segment depends on
local infrastructure conditions, environmental constraints and fuel specifications, in addition to oil company strategies, steady
growth in demand and the definition of a government policy that eases institutional risks.
r 2005 Elsevier Ltd. All rights reserved.

Keywords: Oil refining in Brazil; Oil products market; Oil supplies

1. Introduction 4.9% p.a.), while considerably higher than the world

Brazil is a continent-sized country with the world’s
fifth-largest population (177 million inhabitants in 2002)
clustered in urban centers where road transportation
predominates.
The country’s gasoline and gasoil consumption has
risen by 5.5% p.a. and 4.1% p.a., respectively, since
1990, while fuel oil used by industry has remained
virtually unchanged (MME, 2002). Actually, during the
1990s, Brazil0 s oil demand growth (4.9% p.a.) was only
smaller than China and India (respectively, 7.2 and
Corresponding author. Tel.: +55 21 2562 8760;
fax: +55 21 2562 8777.
E-mail addresses: mtavares@ppe.ufrj.br (M.E.E. Tavares),
szklo@ppe.ufrj.br (A.S. Szklo), giovani@ppe.ufrj.br (G.V. Machado),
roberto@ppe.ufrj.br (R. Schaeffer), jacqueline@ppe.ufrj.br (J.B.
Mariano), jsala@ppe.ufrj.br (J.F. Sala).
average (1.2% p.a.)—EIA US-DOE (2001). In addition,
the country’s oil market growth affects the Atlantic
Basin.1
For the oil products and fuel market, Brazil has also
important specificities, related to its preeminence in the
bio-fuel use (particularly ethanol). This affects the
country’s gasoline market, but also it is an example
for countries intending to expand the use of bio-fuels
(IEA, 2004).
However, as domestic output has not been sufficient
to meet oil product demands in Brazil, particularly for
gasoil, LPG and petrochemical naphtha, imports of
1
For instance, NEPD Group (2001), a report on national energy
policy performed by a group leaded by the own US Vice-President,
Mr. Dick Cheney, emphasizes the strategic role of Brazil for the US oil
market.

0301-4215/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.enpol.2005.06.003
 ARTICLE IN PRESS
3028 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
these products have been required to supplement this
output.2 Particularly noteworthy is the increase in gasoil
imports (11% p.a.) over the past 10 years, outstripping
the average growth rates for all imports (MME, 2003).
Consequently, if Brazil maintains its economic growth,
in parallel to the steady increase in its crude output, it
will become a net low-grade crude exporter3 and a net
importer of white oil products (gasoil, naphtha and
LPG).
Therefore, this paper assesses different strategies for
the expansion of Brazil’s oil refining segment, using
criteria that range from energy security (less vulner-
ability for key products) through to maximizing the
profitability of this sector (boosting the output of higher
value oil products) and adding value to Brazil’s oil
production (reducing exports of heavy acid oil).
Oil refining processes are changing radically, ushering
in trail-blazing innovations for meeting the needs of
liquid fuel markets in terms of industrial activities, and
above all transportation, which is today the dominant
technological route for these purposes (API, 2000; EPA,
1995; Williams, 2003; Worrell and Galitsky, 2004).
Along these lines, the early decades of the XX century
were important for fine-tuning the oil refining industry,
with the development of a well-grounded technological
learning curve.
Light oil is becoming scarcer and more hotly disputed,
driving a quest for alternative energy sources through
research and development4 (Adelman, 2002; Campbell,
2002; Laherrere, 2002). Over the short term, the use of
alternative fuels and alternative technologies for produ-
cing conventional fuels—such as gas-to-liquid (GTL)
technology—is economically nebulous.
Oil products still dominate the energy market,
particularly in the transportation sector, benefiting from
an infrastructure that is totally focused on the trans-
2
There are also sponge coke imports, but here Brazil exports high-
grade (low-sulfur) coke for metallurgical uses and producing graphite
anodes, while importing coke merely to produce heat.
3
Mainly heavy and acid, Brazilian crude encounters difficulties in
finding markets, and it is sold at far lower prices than benchmark
crudes such as WTI and Brent. For further details, see Szklo et al.
(2004).
4
Alcohol, biodiesel and natural gas are examples of alternative
energy sources. The quest for fuel production solutions other than oil
refining is not as recent as it might seem. When World War II broke
out, a new option began to challenge oil refining as the dominant
technology. A wide variety of research and development programs
introduced processes for obtaining synthetic fuels from raw materials
other than crude. These processes were of the utmost importance for
Germany and Japan, which lacked their own national oil reserves and
were subject to rigorous trade blockades by the Allies during the
conflict. Other than in South Africa, the meager economic feasibility of
these processes resulted in these processes petering out during the post-
War period. However, economic and institutional factors—although
very different from the situation in World War II—are today
prompting renewed interest in synthetic fuel processes, paving the
way for a new trend in technological development.
portation and sale of liquid fuels, backed by the sheer
size of the global market and vast amounts of equipment
(particularly automotive vehicles). These factors are
reflected in economies of scale and scope that constitute
technological stumbling barriers.
In the past decade, the upsurge in demands for oil
products has varied in different parts of the world.
Wider-spread demands for medium distillates have been
driven by sharp increases in aviation and road
transportation, in parallel to an upswell in gasoil
consumption by commercial vehicles, particularly in
the developing Asian countries and Europe (Bensaid,
2005b; Brown, 2004). In more mature markets—such as
North America and Europe, whose production struc-
tures are not so energy-intensive—demands have
remained fairly stable over the past few decades (BP,
2004; Bensaid, 2005b).
Within this context, fuel demand expansion scenarios
were simulated for Brazil, as well as for oil products and
oil supplies. The following aspects were taken under
consideration:

Forecast for oil production in Brazil through to 2015,
according to the Hubbert model.5

Modifications to the currently installed refining
facilities, driven by the planned investments of
Petrobras (2004).

The estimated development of the current refining
production profile and refinery feed in Brazil.

Sizing three new basic refining schemes: two for
energy products (gasoil and gasoline) and one for
non-energy products (basic petrochemicals, particu-
larly propene).
Finally, different criteria were established for expand-
ing the refining segment in Brazil, based on independent
logic schemes for lessening energy vulnerability, step-
ping up oil output values or maximizing the profitability
of new refineries in Brazil. These criteria define the oil
products supply scenarios and lead to specific expansion
for Brazil’s refining segment, while also indicating
different investment allocations. This paper compares
these results. The next section presents the oil products
market development scenarios for Brazil, simulated in
the study. The third section presents and analyses the
various expansion criteria adopted, while the fourth
section assesses the possible refining schemes that might
be adopted by new Brazilian refineries. The fifth section
presents the findings for the expansion of oil product
supplies in Brazil through to 2015, based on the criteria
established in the study, while the sixth section closes
with the final remarks from the paper.
5
See Laherrere (2002), Moroney and Berg (1999) and Halloch et al.
(2004), for details about Hubbert model.
 ARTICLE IN PRESS
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
2. Oil products market scenarios
Two different scenarios for the Brazilian oil products
demand were elaborated: a ‘‘Business-as-Usual scenar-
io’’ (BAU scenario, hereafter), and an ‘‘Alternative
Scenario’’. Those scenarios derive from a former study
performed by some authors of this paper to the Ministry
of Mines and Energy of Brazil (Schaeffer et al., 2004a).
Both scenarios are based on the same GDP growth
rates for 2003–2015, which are derived from a study
performed for the Ministry of Planning, Budget and
Management of Brazil (MPOG, 2004): 4.26% per year
from 2005 to 2010 and 4.11% per year from 2010 to
2015. Those GDP real growth rates are in line with other
scenarios considered for Brazil such as EIA US-DOE
(2004a), IEA (2004), IMF (2005) and Wilson and
Purushothaman (2003). Also, both scenarios assume
macroeconomic stability and good business environ-
ment as fundamental preconditions for achieving those
economic growth rates.
Basically, the conceptual difference between these two
scenarios lies in the assumptions considered for sectoral
growth, intra-sectoral product mix, interfuel substitu-
tions and energy efficiency gains.
In the BAU scenario, an update of MME (2002) and
MPOG (2002), sectoral growth, intra-sectoral product
mix, interfuel substitutions and energy efficiency gains
(basically through replacing consumption equipment)
follows business-as-usual trends.
In turn, the Alternative scenario assumes a shift
toward higher value added economic activities, intra-
sectoral product mix upgrade (also higher value added),
equipment substitution programs as part of energy
efficiency policies, intermode substitution in the trans-
portation sector, and higher penetration of more
environment-friendly energy sources. The rationality of
such scenario is to make good use of Brazil’s current and
potential competitive advantages, which was elaborated
mainly according to particular studies performed for the
Brazilian Government, such as Bonelli and Gonc- alves
(1999), Coutinho et al. (2003), Embrapa (2002), and
Giambiagi (2003). Finally, the Alternative scenario
includes an uptick in Brazil’s social inequality indicators
(more specifically, the Gini Index) which leads to a
relative drop in the consumption of traditional energy
sources by low-income rural households, compared to
the BAU scenario.
Clearly, for the horizon of the single decade, the
differences between these scenarios are incremental,
although important. Despite of differences in sectoral
growth, technical changes and energy policy, drastic
modifications in the Brazil’s economic development are
not considered here.6
6
The details of technical and economic assumptions underlying the
market scenarios are described in Schaeffer et al. (2004a) and,
3029
According to those scenarios (Schaeffer et al., 2004a),
total oil products demand increases from 286,602 m3/
day in 2002 to 380,338 m3/day in 2015 in the BAU
scenario and to 352,077 m3/day in 2015 in the alternative
scenario. In order to provide a better sense of such
findings, Petrobras (2004), the major oil company in the
Brazilian market (it still holds 98% of refining industry
in Brazil), forecast a total oil products demand growth
rate of 2.3% per year in the period 2003–2010. Taking
the same period for comparison, this study’s oil product
demand projections are very close to Petrobras demand
forecast: 2.7% per year for the BAU scenario and 2.2%
per year for the Alternative scenario. This is also in line
with the most recent projections of IEA (2004). Table 1
summarizes demand growth rate by oil product for both
scenarios.
As shown in Table 2, the differences between the fuels
market findings for the two simulated scenarios indicate
possible strategies that would allow Brazil to meet its
fuels demands. Particularly through energy policies
encouraging fuel conservation and wider use of alter-
native sources such as ethanol and biodiesel and CNG/
VNG for heavy vehicles in 2015, the Alternative
Scenario indicates a total oil-products market some
7.4% smaller than the BAU Scenario. For gasoil
produced from petroleum, this difference reaches
4.4%,7 21,3% for gasoline and 1.5% for LPG.
It is worthwhile stressing that these market projec-
tions are subject to significant uncertainties in terms of
the future of bio-fuels, particularly ethyl-alcohol and
biodiesel, in view of the advent of multi-fuel auto-
mobiles and the Biodiesel Program, as well as a better-
adapted future scrapping curve for Brazil; the expansion
of natural gas in Brazil’s industrial sector, replacing
fuelwood and used fuel oils; policies fostering the
expansion of compressed natural gas for use by heavy
vehicles in Brazil (replacing gasoil); and policies urging
the certification of thermal equipment, similar to those
currently implemented for electrical equipment.
3. Expansion criteria for oil product supplies in Brazil
Faced with the uncertainty associated with refining
margins, oil prices and the future market for oil
products, some reluctance to expand nominal refining
capacities is noted in the world’s more complex refin-
ing segments, but not the treatment and conversion
facilities (which are associated with stricter fuel quality
(footnote continued)
alternatively, in Schaeffer and Machado (2005), a study performed
basically by the same team.
7
The total gasoil market includes gasoil made from crude or natural
gas (GTL plant) and biodiesel. Looking at this total gasoil market, the
difference between the two scenarios alters to 6.3%, due to blending
5% biodiesel with oil-based gasoil.
 ARTICLE IN PRESS
3030 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040

Table 1 lighter products, reducing fuel oil availability, supplying

Growth rates: oil products demand (% p.a.) products with lower sulfur and lead content levels, and
2002–2005 2005–2010 2010–2015 2002–2015
processing lower-grade oils (EIA US-DOE, 2004b).
Within this context, the scenarios for the expansion of
Business-as-usual scenario Brazilian oil supplies cannot fail to take into account the
LPG 0.44 1.84 2.11 1.62 modifications planned for the existing facilities, as
Gasoline 0.73 0.28 1.80 0.97
shown in Tables 3 and 4, underwritten by the invest-
Kerosene 5.35 3.75 4.92 4.57
Gasoil 1.51 2.97 2.27 2.36 ments scheduled by Petrobras, and the estimated oil
Residual Fuels 1.16 3.65 3.72 2.22 processed, according to the modeling for the ‘‘average’’
Fuel oil 6.09 1.54 1.15 0.42 Brazilian refining segment.
Coke 8.52 5.14 5.23 5.94 There are 13 refineries in Brazil and Petrobras owns
Naphtha 2.08 2.75 2.89 2.65
11 of them. The other two belong to private initiative.
Alternative scenario The capacity of these two refineries represent only 2% of
LPG 0.26 1.68 1.94 1.45 all country’s refining capacity and both, because of their
Gasoline 0.48 1.86 0.65 0.86
scale, sell their products to market niches. The Brazilian
Kerosene 5.99 6.25 6.27 6.20
Gasoil 1.96 2.54 1.50 2.00 oil refining is extremely concentrated in the Southeast0 s
Residual Fuels 0.55 2.38 1.98 1.59 country, near to consumer centers, because the refineries
Fuel oil 6.48 0.77 0.49 1.43 construction was optimized to maximize production
Coke 7.72 3.56 3.52 4.49 scale economies and minimize distribution scale dis-
Naphtha 2.10 2.06 1.91 2.01
economies. In this sense, the oil refining in Brazil can be
Note: residual fuels include fuel oil, coke and other oil products. simplified and represented as one refinery.
The simple applied modeling derives from a Refinery
Flows Model that rebalances mass flows for an entire
Table 2 refining segment, integrating all Brazilian refineries in
Market growth: oil products demand (m3/day) compliance with different oil standards (five types) and a
2002 2010 2015 basic scheme that integrates all the refineries in Brazil.8
In this rebalancing process, the model assigns high
Demand % Demand % Demand % priority to flows to the high conversion units such as the
Business-as-usual scenario FCC- Fluid Catalytic Cracking, HCC-Hydrocracking
LPG 33,219 11.6 35,554 10.6 37,848 10.0 and Delayed Coking, which work in their maximum
Gasoline 44,386 15.5 46,012 13.7 50,303 13.2 operational capacity, while also making good use of all
Kerosene 10,852 3.8 15,254 4.5 19,394 5.1 the operating alkylation, catalytic reform and isomer-
Gasoil 107,293 37.4 129,884 38.6 145,288 38.2
ization capacities. Nevertheless it does not consider
Fuel oil and coke 58,027 20.2 71,857 21.4 86,247 22.7
Naphtha 32,825 11.5 37,875 11.3 41,258 10.8 hydrotreating and hydrogen production units, as other
Total 286,602 100.0 336,436 100.0 380,338 100.0 auxiliary processes. Consequently, this follows a similar
logic to adding complexity to a refinery, pursuant to the
Alternative scenario
LPG 33,219 11.6 35,204 10.8 37,287 10.6 Nelson Index. Finally, this also fine-tunes the produc-
Gasoline 44,386 15.5 40,999 12.6 39,675 11.3 tion of medium and light cuts.
Kerosene 10,852 3.8 17,497 5.4 23,721 6.7 The results of the modeling in this study lead to a
Gasoil 107,293 37.4 128,905 39.5 138,863 39.4 production profile that is significantly different from
Fuel oil and coke 58,027 20.2 66,367 20.4 73,219 20.8
Brazil’s current refining segment (Table 5), with a
Naphtha 32,825 11.5 37,144 11.4 39,312 11.2
Total 286,602 100.0 326,116 100.0 352,077 100.0

specifications on these markets) (Stell, 2003). Addition-
ally, there is a trend towards convergence in oil products
production and consumption standards, with invest-
ments in new conversion and treatment plants, in
parallel to the divestment of less complex plants, within
a context of redefining strategic positions, in addition to
rising capacity in areas with higher annual growth rates
(BP, 2004; Nakamura, 2003; Williams, 2003).
Consequently, future investments in refineries in the
developing countries should include more advanced
configurations in order to meet global demands for
8
This type of integrated simulation of Brazil’s refining segment,
represented as a single refinery, is quite common in long-term models,
although simplificatory and reductionistic for fine-tuning loads. The
process units operational capacity is a model input while the oil yields
are fixed parameters, which means that each unit’s production is fixed
and does not depend on the market. In other words, the model
operates in order to produce the maximum of gasoil and gasoline,
considered the main products. Nevertheless, it can be a disadvantage
when the market does not indicate the need of maximum processing in
complex units, i.e., when the gasoline and gasoil rentabilities are not
attractive. In this case, it is assumed that the overall optimum level for
this segment is compatible with the optimum level of each refinery,
which is not always true, but nevertheless tends to occur in segments
that are controlled largely by a single enterprise (Babusiaux et al.,
1983).
 ARTICLE IN PRESS
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040 3031

Table 3
Main projects in Brazilian refineries—2004–2011

Refinery Project Capacity (m3/day) Operational start

REDUC Hydrotreating 4000 2004

FCC revamp — 2005
Coking 5000 2005
Hydrotreating (gasoil and naphta) 2000 2005
Hydrodesulfuring (gasoline) 5000 2007
Hydrocracking 5000 Between 2007 and 2011

REGAP Hydrotreating (gasoil) 3500 2004

Catalytic reform 1000 2007
Hydrodesulfuring (gasoline) 2000 2007
Hydrotreating (gasoil) 2000 Between 2007 and 2011
Hydrodesulfuring (gasoline) 2000 Between 2007 and 2011
REFAP Coking 2000 2004
Hydrotreating (gasoil) 4000 2004
Residue catalytic cracking 7000 2004
Sulfur recovery 2  50t 2004
Hydrogen Generation 550,000 Nm3 2004
Hydrodesulfuring (gasoline) 4000 Between 2007 and 2011
Hydrotreating (gasoil) 2000 Between 2007 and 2011
Hydrodesulfuring (gasoline) 2000 Between 2007 and 2011
RLAM Atmospheric destilation revamp — 2003
Atmospheric destilation revamp — 2003
Hydrodesulfuring (gasoline) 5000 2007
Hydrogen generation 500,000 Nm3 2007
Coking 5000 Between 2007 and 2011
Hydrotreating (naphta) 1500 Between 2007 and 2011
Hydrotreating (gasoil) 8000 Between 2007 and 2011
Atmospheric destilation revamp — Between 2007 and 2011
Hydrodesulfuring (gasoline) 3000 Between 2007 and 2011
Catalytic reform 2000 Between 2007 and 2011
REPAR Hydrotreating (gasoil) 5000 2004
Atmospheric destilation revamp — 2007
Coking 5000 2007
Hydrotreating (naphta) 1500 2007
Hydrotreating (gasoil) 5000 2007
Hydrodesulfuring (gasoline) 5000 2007
Catalytic reform 1000 2007
Hydrocracking 3500 Between 2007 and 2011
REVAP Coking 5000 2006
Hydrotreating (gasoil) 6000 2006
Hydrotreating (naphta) 3000 2006
Hydrodesulfuring (gasoline) 5000 2007
Hydrodesulfuring (gasoline) 3000 Between 2007 and 2011
Catalytic reform 1500 Between 2007 and 2011
REPLAN Coking 5000 2004
Hydrotreating (gasoil) 5000 2004
Atmospheric destilation revamp — 2006
Hydrodesulfuring (gasoline) 5000 2007
Hydrotreating (gasoil) 6000 Between 2007 and 2011
Hydrodesulfuring (gasoline) 5000 Between 2007 and 2011
Catalytic reform 2500 Between 2007 and 2011
RPBC Hydrodesulfuring (gasoline) 5000 2007
Catalytic reform revamp — 2007
Hydrotreating (naphta) 2500 2007

Source: Fairbanks (2003a, b).

 ARTICLE IN PRESS
3032 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040

Table 4 Expansion Criteria Scenarios

Estimated capacity of process units in Brazilian refineries (m3/day) Energy 1. Self

Vulnerability Sufficiency
2. Ceiling:20%
Process units 2010 2015
Macro Scenarios Market Scenarios Minimum 3. Ratio100%
Processing 4. Ratio 80%
Vacuum destilation 141,913 141,913
Fluid catalytic cracking 80,540 80,540 Basic Maximum 5. Margin
Basic Market Profitability
Residues fluid catalytic cracking 20,000 20,000
Macro- Scenario
Catalytic reform 6365 12,365 economic
Petrochemical 6. Petrochemicals
Gasoil hydrocracking 10,000 13,500 Integration
Scenario
Gasoil hydrotreating 47,500 65,100
Coking 30,200 35,200 Alternative Energy 7. Self
MTBE 1172 1172 Macro- Vulnerability Sufficiency
economic 8. Ceiling: 20%
Lubricants hydrotreating 5670 5670
Scenario Minimum 9. Ratio 100%
Kerosene hydrotreating 5800 5800 Processing
Alternative 10. Ratio 80%
Naphta hydrotreating 9775 10,800
Market Maximum 11. Margin
Hydrotreating 17,200 17,200 Scenario Profitability
Lubricants 3180 3180
Gasoline hydrodessulfuring 22,000 41,000 Petrochemical 12. Petrochemicals
Hydrodessulfuring (gasoil) 5000 5000 Integration

Alkilation 1100 1100

Fig. 1. Simulated oil refining segment expansion scenarios.
Source: Schaeffer et al. (2004a).

other countries (Horsnell, 1997). In this case, two

Table 5 expansion scenarios are analyzed for each oil products
Estimated production profiles for Brazil’s current refining segment, as market scenario: one for self-sufficiency in gasoil (no
shown in the modeling (%)
gasoil imports) and the other for a gasoil imports ceiling
2001a 2010 2015 set at 20% of the total market for this oil product in
2010 and 2015.
LPG 8.1 8.9 9.1
The second criterion (Minimum processing) is de-
Gasoline 19.5 21.9 23.8
Naphtha 10.5 11.3 9.6 signed to boost the profitability of domestic oil
Kerosene 2.6 5.4 5.9 production. This is worthwhile in the case of Brazil,
Gasoil 36.4 40.7 41.6 bearing in mind that the Brazilian oil is largely non-
Fuel oil and coke 22.9 11.9 10.1 conventional (acid/heavy). Other non-conventional oil
a
Modeling datum—i.e., slightly different from that actually noted producers follow a similar strategy—see, for instance,
for 2001. the case of Venezuela (EIA US-DOE, 2004b; EPA,
1995). In this case, there are once again two expansion
scenarios for each oil products market: one for total
processing of an equivalent volume of oil produced in
marked influence on the findings for the expansion of Brazil; the other for a processing/production ratio of
supplies. 80%.
However, taking into account the planned expansion The third criterion (Maximum profitability) follows
of the existing facilities, there is still a gap between oil the strategy of maximizing the profitability of a refinery
product demands and output for the horizon through to on a stand-alone basis, fine-tuning its output for
2015, as shown in the market scenario simulated in this gasoline. In this case, the feasibility of a non-integrated
study. Consequently, four criteria were adopted for refinery scenario is tested, focused on premium gasoline
adding new refineries to the current refining segment, exports to international niche markets. This criterion is
which in turn are broken down into twelve oil-products in line with current strategies undertaken by oil
supply scenarios for Brazil (see Fig. 1). refineries focused on exports—see, for instance, the
Those criteria derive from former discussions placed refineries in the Caribbean zone (EIA US-DOE, 2004a).
under the aegis of Technical Committees of the National Finally, the fourth criterion (Petrochemical integra-
Energy Policy Council of Brazil (CNPE, 2002; MME, tion) seeks to integrate the refinery with the petrochem-
2002), which were formalized by the present authors in ical industrial complex (refinery focused on petro-
Schaeffer et al. (2004a), Schaeffer et al. (2004b) and chemicals, particularly propene).9 This refining scheme
Schaeffer and Machado (2005). is focused on the direct production of intermediate
The initial criterion (Energy vulnerability) refers to the
logic of minimizing the energy vulnerability of the oil 9
It is estimated that global petrochemical product demands should
chain. This is in line with recent strategies undertaken by rise at rates higher than energy demands. By 2020, ethylene demands
 ARTICLE IN PRESS
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040 3033

Table 6 Table 7
Estimated oil production—Hubbert curve Production profiles for new refining schemes as shown in the modeling
(%)
Year 2002 2005 2010 2015 2020 2025
Gasoil refinery
Volume (million m3/day) 0.20 0.26 0.38 0.48 0.52 0.48 Capacity (39,746 m3/day) %
LPG 9.5
Gasoline 34.1
Naphtha 5.3
Kerosene 0.0
petrochemical products—i.e. skipping petrochemical Gasoil 44.6
naphtha—with concomitant fuel production. A typical Fuel oil and coke 11.6
refinery for basic petrochemicals would in principal be Gasoline refinery
focused on propylene, but could also produce medium Capacity (39,746 m3/day) %
distillates, aviation kerosene and gasoil, heavy gasoils LPG 6.0
and LPG fractions. The petrochemical integration is in Gasoline 43.0
line with recent technological innovation of the oil Naphtha 2.0
Kerosene 16.0
refining sector worldwide (Worrell and Galitsky, 2004; Gasoil 14.0
Zai-Thing et al., 2002). Fuel oil and coke 16.0
Finally, in order to draw up the scenarios for the
Petrochemical refinery
Minimum Processing criteria, modeling was required as Capacity (31,796 m3/day) %
well as an oil production projection for Brazil through LPG 2.4
to 2015, giving the figures shown in Table 6. Similar to Gasoline 22.0
the oil products market projection, this modeling is Naphtha 3.4
subject to uncertainties: initially, uncertainty regarding Propene 16.4
Gasoil 21.0
the pertinence of the Brazilian case in its model adopted Fuel oil and coke 14.0
for the production curve (Hubbert Approximate Mod-
el); and second, uncertainty regarding Brazilian reserves
to be added in the future, which would result in larger or
smaller recoverable reserves.
To a large extent, this latter uncertainty is measured about this item, particularly regarding the final fine-
through the reserves probability curve by basin in tuning of the proposed refining schemes that may vary,
Brazil, which follows a Log-normal distribution. This according to market demands and the variations in the
study considered a hypothesis based on a cumulative crude. It is also important to note that there are possible
probability of 75%. This leads to an Ultimate Recover- variations in terms of the refining schemes outlined here.
able Reserve (EUR) of 41 Gb. A 50% probability would However, these variations do not undermine the
result in the addition of 20 Gb to the production curve schemes as proposed and sized.
adopted. A 95% probability would result in the Additionally, as an alternative source of supply, a
subtraction of 12 Gb to the integral of the production GTL plant is proposed, to start up operations in 2015,
curve adopted. These are significant differences. How- focused on gasoil production and absorbing part of the
ever, the curve presented which is based on those increased natural gas output from the Santos Basin,
criteria, leads to findings similar to those obtained by resulting from recent discoveries in this region. This
Petrobras in its recent Strategic Plan (Petrobras, 2004), plant was modeled on the basis of the indirect
as well as those of IEA (2002, 2004). production route and is useful for sensitivity analyses
in terms of the expansion of the refining segment based
on refineries optimized for gasoil.10
In order to draw up economic assessments for each of
4. Future refining schemes
the criteria proposed, two oil and oil products price
scenarios were also drawn up (Table 8).11 Once again,
In order to prepare the supply scenarios in compliance
there is some uncertainty regarding not only the
with the four criteria listed above, three basic refining
international oil prices and discounts for Brazilian oil,
schemes were drawn up, two for energy products (gasoil
but also the international oil product prices. Particularly
and gasoline) and one for non-energy (basic petrochem-
noteworthy over the short term is uncertainty over the
icals, particularly propene) whose yields are presented in
impact of the ban on MTBE in the USA, in terms of its
Table 7. Once again, there are significant uncertainties
10
(footnote continued) As sized, the GTL plant has a thermodynamic yield of 63% and a
should rise by 3.5% p.a., with propylene rising at 5% p.a., benzene up capacity of 50 Kbpd (70% gasoil and 30% petrochemical naphtha).
11
4% p.a. and xylenes at 5% p.a. (Zai-Thing et al., 2002). Based on the analysis by Machado (2004).
 ARTICLE IN PRESS
3034 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040

Table 8 Table 9
International price scenarios adopted (US$/m3) Internal rate of return (IRR) for the refineries and the GTL plant
considered for expansion (% p.a.)
2001 2005 2010 2015
Price scenario 1
Scenario 1
Petroleum—WTI 165.4 249.7 167.3 176.1 2010 2015
Marlim 146.6 205.7 135.9 151.0
Gasoline 203.2 324.6 217.6 229.0 Gasoil refinery 21.8 17.0
Premium Gasoline 258.5 438.4 303.8 297.5 Gasoline refinery 28.2 15.0
Gasoil 193.7 299.4 208.8 220.2 Petrochemical refinery 18.7 10.9
Jet fuel 203.8 307.6 205.7 216.4 GTL — 17.5
Naphtha 145.9 259.8 184.9 194.4
Price scenario 2
LPG 129.6 196.2 131.5 138.4
Fuel oil-BTE 126.4 191.2 117.0 114.5 Gasoil refinery 14.5 9.6
Propene 389.4 588.7 394.4 415.1 Gasoline refinery 16.1 8.5
Scenario 2 Petrochemical refinery 12.8 4.9
Petroleum—WTI 165.4 249.7 167.3 176.1 GTL — 14.0
Marlim 146.6 205.7 135.9 151.0
Notes: The internal rate of return (IRR) calculations assume that the
Gasoline 203.2 274.9 183.7 193.7
net margin posted for the year continues throughout the useful life of
Premium Gasoline 258.5 349.1 233.4 245.9
the refinery. These are consequently approximate values that are
Gasoil 193.7 292.5 195.6 206.3
relatively conservative, offering only an indication of the magnitude of
Jet fuel 203.8 307.6 205.7 216.4
the IRR to be received by possible future investors if they were to
Naphtha 145.9 219.5 147.2 154.7
freeze the net refining margins for the year in question.
LPG 129.6 196.2 131.5 138.4
The estimated investments of the new refineries are based on the
Fuel oils-BTE 126.4 191.2 127.7 134.6
estimated Nelson Index for these facilities. The investments in these
Propene 389.4 588.7 394.4 415.1
plants reach US$ 2.92 billion for the gasoil refinery, US$ 3.01 billion
for the gasoline refinery and US$ 1.50 billion for the GTL plant (this
latter with a capacity of 7.949 m3/day). The operating costs of the
refineries were estimated on the basis of their complexity index for
effects on international gasoline prices (EIA US-DOE, 2010 and 2015.
2002). For the medium term, uncertainty over the
impacts of the rising use of gasoil in the European fleet
is particularly noteworthy, in parallel to rising demands
for medium-grade products in Southeast Asia, in terms two in 2010 and one in 2015 (processing all oil produced
of the effects on gasoil prices and availability on in Brazil).
international markets. Another key factor in the price Consequently, if the expansion criteria are focused
scenarios is that it is assumed convergence between the mainly on energy security, expansion through gasoil
prices charged by Brazilian refineries (ex-refinery prices) refineries is not necessarily the most appropriate
and the international oil products prices. solution, and the possibility of setting up a GTL plant
These price scenarios lead to estimated internal rates should be assessed. Although this plant is less versatile
of return for the new refineries, presented in Table 9. than the gasoil refinery proposed here, with ample
delayed coking and HCC capacities, in addition to a
lower refining margin than the gasoil refinery, for both
5. Expansion of oil product supplies in Brazil according price scenarios in 2015, the investments in the GTL
to selected criteria plant are some 50% of the amounts required for the new
refinery. In a scenario where capital for heavy invest-
As already mentioned, the analysis of the expansion ments is limited, or with high interest rates, the GTL
of Brazil’s refining segment simulates twelve scenarios solution may be the more robust. Consequently, these
(six for each oil products market scenario) according to findings suggest that more detailed studies are required
the four different expansion criteria mentioned in for the introduction of GTL technology into the gas and
Section 3 of this paper. The findings of these simulations oil chain in Brazil.
are presented in Tables 10 and 11. On the other hand, if the criteria are designed to add
According to these criteria, the BAU Scenario value to Brazil’s crude output through highly versatile
indicates expansion of Brazil’s refining segment by up refineries exporting gasoil in considerable volumes by
to three refineries. The simulation hovered between no 2010–2015, the ‘‘100% processing’’ may be indicated. In
additional refinery other than the expansion planned for this case, where expansion takes place through three
the current facilities (maximum gasoil imports for 20% refineries (total investments of around US$ 9 billion), an
of the total market, with the installation of a GTL plant increase in the balance of trade (oil and oil products)
in 2015); and three new refineries, optimized for gasoil, appears for 2015, varying between US$ 960 and US$
Table 10
Expansion of Brazil’s oil refining segment according to the criteria adopted—business-as-usual scenario

2015

Self-sufficiency: Test 1: Maximum gasoil Test 2: GTLa Processing/ Processing/ Maximum Petrochemicalsc:
gasoil GTLa Plant imports: 20% Plant production: Production: profitability: Integration
100% X80% gasolineb

Number of new refineries: 1 1 1 0 1 1 0 0

2015
Investments: 2015 (US$ 2.92 4.42 2.92 1.5 (GTL) 2.92 2.92 0 0
billion)
Accumulated expansion: no 2 2 1 0 3 2 1 1
new refineries: 2003–2015
Total accumulated 5.84 7.34 2.92 1.5 8.76 5.84 3.01 1.60
investment: new units (US$
billion)f
Total dependency level 23.8 21.8 23.8 21.8 23.8 23.8 23.8 23.8
without new refinery(ies)d—
(%)
Total dependency level with 4.0 2.7 13.9 — 5.9 (exports) 4.0 14.7 19.1
new refinery(ies)d—(%)
Gasoil dependency level 17.8 14.3 17.8 14.3 17.8 17.8 17.8 17.8
without new refinery(ies)d—
(%)
Gasoil dependency level with 2.2 (exports) 7.7 6.9 — 15.1 (exports) 2.2 (exports) 14.4 13.7
new refinery(ies)d—(%) (exports)
ARTICLE IN PRESS

Net balance of trade with the +0.05/0.75 +0.40/ 2.40/2.96 — +2.45/+1.42 +0.05/0.74g 2.68/3.20h —i
new refinery(ies) (US$ billion) 0.43
price scenarios 1 & 2e
a
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040

The sensitivity analysis for the GTL Plant was undertaken for expansions in 2015, focused on gasoil refining—self-sufficiency in gasoil and gasoil imports capped at 20% of the total market. In
this case, total investments include refining expansion and GTL plant.
b
The maximum profitability focused on gasoline indicates a refinery whose operations are not integrated with Brazil’s current refining segment, focused on its highest possible profitability on a
stand-alone basis when not associated with the objectives of minimizing total refining costs in Brazil.
c
In fact, the proposed refinery produces a mix of products, particularly medium and light distillates, as well as petrochemicals. Consequently, this also affects imports of key oil products,
particularly gasoil.
d
Dependency is measured by the relationship between the net imports and demand volumes for oil products (energy and naphtha).
e
The balance includes freight rates (imported oil products) or deducts the freight rates (exported oil and oil products) and assumes that the oil is exported, with no new refinery. The first figure
refers to price scenario 1, and the second to price scenario 2.
f
The accumulated total investments assume a null discount rate and do not include the effects of inflation and equipment.
g
The net operating margin of the GTL Plant is estimated at US$ 3.5/barr in price scenario 1 and US$ 2.9/barrel in price scenario 2, for 2015, with the natural gas price at US$ 1.5/MMBTU.
h
In this case, the exported oil, which would be processed by a possible future refinery, is not discounted.
i
Due to the lack of modeling accuracy for the balance of trade associated with propene (or other basic petrochemical feedstock) it was decided not to calculate the trade balances for this refinery.
3035
 3036

Table 11
Expansion of Brazil’s oil refining segment according to the criteria adopted—alternative scenario

2015

Self-sufficiency: Test 1: Maximum gasoil Test 2: GTLa Processing/ Processing/ Maximum Petrochemicalsc:
gasoil GTLa Plant imports: 20% Plant production: Production: profitability: Integration
100% X80% gasolineb

No new refineries: 2015 1 0 0 0 1 1 0 0

Investments: 2015 (US$ 2.92 1.5 (GTL) 0 1.5 (GTL) 2.92 2.92 0 0
billion)
Accumulated Expansion: No 2 1 0 0 3 2 1 1
new refineries: 2003–2015
Total accumulated 5.84 4.42 0 1.5 (GTL) 8.76 5.84 3.01 1.6
investment: new units (US$
billion)f
Total dependency level 17.7 15.6 17.7 15.6 17.7 17.7 17.7 17.7
without new refinery(ies)d—
(%)
Total dependency level With 3.7 (exports) 5.5 — — 14.4 (exports) 3.7 (exports) 7.8 12.6
new refinery(ies)d—(%)
ARTICLE IN PRESS

Gasoil dependency level 14.0 10.3 14.0 10.3 14.0 14.0 14.0 14.0
without new refinery(ies)d—
(%)
Gasoil dependency level with 8.9 (exports) 1.2 — — 20.4 (exports) 8.9 (exports) 10.4 9.7
new refinery(ies)d—(%) (exports)
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040

Net balance of trade with the +1.81/+0.91 — — — +4.21/+3.08 +1.81/+0.91 0.88/0.55 —

new refinery(ies) (US$ billion)
price scenarios 1 & 2e

Note: See notes in the previous table.

 ARTICLE IN PRESS
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
310 million. However, it is always worthwhile noting
that:

Expansion on this scale may be justified only through
refineries with high conversion capacities (high
proportion of bottom of barrel units) and high
versatility (high proportion of HCC).

There are prospects for gasoil shortages (or an
expansion in the global market for this oil product)
over the short term in Southeast Asia and Western
Europe (Bensaid, 2005b; Brown, 2004).

There are also prospects for broadening price gaps
between light and heavy acid oils over the short term,
encouraging investments in complex, versatile refi-
neries with high operating margins.
Another interesting aspect indicates that, for this
market scenario, the criteria for self-sufficiency in gasoil
and processing 80% of domestic output are indistin-
guishable: both lead to expansion with two new gasoil
refineries (one in 2010, the other in 2015). This seems to
indicate a tradeoff point between an energy security
criterion and a value-adding criterion to Brazil’s crude
output. But this trade-off is due only to the versatility of
the proposed refineries.
In turn, the petrochemical refinery not only ensures
the integration of the oil-petrochemical chain, but also
complies with the criterion of gasoil imports capped at
20%. This means that, in order to limit imports at 20%
for gasoil, a refinery focused on gasoil or a petrochem-
ical refinery may absorb investments in 2015. However,
the latter offers the advantage of presenting integration
economies, producing high-value market oil products
such as propene. This may prove a sturdy and well-
integrated alternative for expanding Brazil’s refining
segment. In this case, the margin of this refinery is less
important than the supply security that it offers for
gasoil and petrochemicals.
Finally, the gasoline refinery designed for premium
gasoline exports in the market scenario under con-
sideration here, operates with net margins that vary
considerably between 2010 and 2015, depending on the
price scenarios considered. Although the operating
mode of this refinery is not closely integrated with
Brazil’s refining segment, it nevertheless presents inter-
esting economic findings at certain points, particularly
for 2010 in Price Scenario 1, although it is very
vulnerable to the price context on the international
scenario, particularly for gasoline. This is a refinery that
must find a market for its main product. In this case, a
more interesting alternative leading to more robust
market results would be to install this refinery in its
reference market (or close to it), such as the Caribbean
or the USA, in order to produce gasoline for exports.
For the Alternative Scenario, the main results of the
previous scenario remain the same: gasoil imports
3037
capped at 20% of the total market with the installation
of a GTL plant in 2015; and processing all oil produced
in Brazil. However, the more sluggish oil products
market growth indicates that the energy security criteria
(self-sufficiency in gasoil or import levels at 20% of this
oil product) are more affected by the GTL plant coming
on-stream, simply because the gasoil market is smaller in
the Alternative Scenario. As there are uncertainties
about the future oil products market in Brazil,
particularly the gasoil market, these findings show the
advantage of drawing up robust planning for the
expansion of the refining segment in Brazil.
On the one hand, self-sufficiency in gasoil as a target
leads to the expansion of two refineries, regardless of the
GTL plant, in the higher growth scenario. Taking the
average net margins obtained in the two price scenarios
and the investments in the refinery, the internal rate of
return (IRR) of the gasoil refineries should vary between
17% and 22% p.a. for Scenario 1, and between 10% and
14% p.a. for Scenario 2. This indicates the impact of
price scenarios on the feasibility of the expansions under
consideration. In fact, given the investment volumes,
these IRRs are not as attractive as those of other
investments in the industrial sector. However, they are
compatible with those noted for infrastructure sectors,
particularly the energy sector.
On the other hand, in the lower-growth scenario, this
criterion pits the gasoil refinery against the GTL plant.
The GTL plant avoids investments in a new gasoil
refinery, buffering a lower investment option (around
one-half of the refinery investments) and a smaller scale,
although with slimmer net margins than the proposed
gasoil refinery. Once again, considering the average net
margins obtained in the two price scenarios and the
investments in the refineries, the IRR of the GTL plant
should vary between 17% p.a. in Scenario 1 and 14%
p.a. in Scenario 2, leading to very little difference in
terms of the outcome of investments in the gasoil
refinery.
It is also interesting to note that, for the self-
sufficiency in the gasoil criterion, the expansion based
on the GTL plant in 2015 and a gasoil refinery in 2010
almost offset the physical gasoil balances. This finding
contrasts with that of other criteria that lead to a greater
expansion of the refining segment (criteria based on
minimum oil output processing). According to these
criteria, in 2015 Brazil becomes a net exporter of
considerable volumes of gasoil (for instance, around
28,617 m3/day of gasoil in 2015 for the maximum
domestic oil processing criterion). In itself, this is not a
problem because initially, according to the analyses in
the study, the global gasoil market is expanding; and
second, the proposed refinery is quite versatile, based on
coking, FCC and HCC, meaning that it offers a certain
level of flexibility in terms of its output, depending on
the use of its deep conversion units.
 ARTICLE IN PRESS
3038 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
However, in order to attend the global gasoil market
and to comply with the specifications for European
gasoil, the refinery should invest quite heavily in
treatment plants. In this case, the estimated investments
for the new gasoil refinery should increase, according to
the necessary investments for HDT plants in the current
refining segment. Once again, it is always worthwhile
recalling that the IRRs of the gasoil refineries are lower
than those noted or sought by industrial sectors in more
competitive environments, although they remain com-
patible with those of the infrastructure sectors.
For the refinery specializing in premium gasoline, its
vulnerability to international market conditions is
particularly noteworthy, which is also expressed through
the variation in the IRR of this project: 8–28% p.a.,
depending on the price scenarios analyzed. In other
words, the gasoline refinery may be somewhat inade-
quate within an economic context of low premium
gasoline prices or should it prove difficult to place the
product in a specific niche market. The most reasonable
option would be to study the feasibility of this refinery
within its reference niche market. It must be recalled
that Western Europe functions as a swing exporter of
gasoline to the North American market too often, which
causes immense vulnerability for non-integrated refi-
neries that are highly exposed to the international
market.
For the petrochemical refinery, an analysis focused
only on the IRR or the net refining margin, fails to
capture the integration gains of this plant. To an
increasing extent, it is expected that the shortage of
petrochemical naphtha on the international market will
lead to greater vertical integration in the oil—petro-
chemicals chain. This is particularly important for a
refinery specializing in propene through a process of
direct oil conversion into basic petrochemicals. Another
important point is the fact that the petrochemical
refinery—which also produces gasoil—keeps Brazil’s
level of dependence at around 10% for this oil product,
below the current figure. This means that the refinery
has a dual function: integrating chains and boosting
energy security.
6. Final remarks
As it was discussed, currently Brazil is a net gasoil
importer, and these imports might grow. In addition,
some regions show a large growth in gasoil demand,
especially the Southeast Asia and Europe, which
combined with a lack of investments in refining capacity
can represent a constraint in the world gasoil availabi-
tity.
As of today Brazil exports gasoline, but this oil
product does not attend the most strict specifications in
United States, the largest gasoline market in the world.
Following the example of Venezuela or any other
country, which searches the maximum profitability for
refining its own crude, refineries can be constructed for
producing high-quality gasoline in Brazil.
Petroleum products account for 97% of the energy
consumed in road transport in the world. The purpose
of replacing these products with alternative energies is to
reduce oil dependence as well as greenhouse gas
emissions (Bensaid, 2005a). Recent developments in-
dicate that the use of bio-fuels, previously confined to a
handful of countries including Brazil12 and the United
States, is ‘‘going global’’ and a world market may
emerge. However, these prospects could eventually be
limited by constraints relative to resources and costs
(His, 2005).
In this sense, it is stressing that the simulated market
scenarios present growth rates that are quite different
for the gasoline and ethanol markets, due to the greater
penetration and wider use of multi-fuel vehicles in the
Alternative Scenario. This scenario also indicates that
the expansion of the refining sector spurs gasoline
exports, compared to the BAU Scenario. This occurs
even when the refinery is not focused on producing
gasoline.
In fact, the differences between the fuels market
figures in the two simulated scenarios indicate possible
strategies for Brazil to meet its own fuel demands. The
Alternative Scenario presents a total fuels market some
7.4% smaller than the BAU Scenario, due mainly to
energy policies designed to conserve fuels and encourage
wider use of alternative sources such as ethanol,
biodiesel and CNG/VNG for heavy vehicles in 2015.
For oil-based diesel, this difference is 4.4%; reaching
21.3% for gasoline and 1.5% for LPG.
As mentioned previously, the Alternative Scenario
does not indicate any sharp breakaway changes
compared to the BAU Scenario, and it is consequently
a possible scenario, provided that certification policies
are established for consumption appliances in domestic
sector, with broader-ranging diesel motors inspections
and replacement of heavy vehicle fleets (buses and
trucks) in parallel to support for replacing industrial
equipment and energy conservation diagnoses. These
policies trigger important synergetic effects in the
economy, ushering in technological innovations and
gains in productivity. They also lessen the need for
investments in expanding energy supplies, particularly
along the natural gas and oil chains.
A sounder and more prudent criterion for Brazil’s
energy policy would indicate that the strategy of
12
In Brazil, the Proalcohol Program, rolled out by the government in
the wake of the oil crises of the 1970s, was key to developing ethanol
production from sugarcane. Between 1973 and 1990, various measures
were implemented under this program. Nowadays the goal is to export
ethanol on a new global market in bio-fuels (His, 2005).
 ARTICLE IN PRESS
M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
investing in energy conservation endows the nation with
greater flexibility over the long term and, depending on
the expansion criteria adopted, leads to the expansion of
no more than two refineries by 2015, one optimized for
gasoil and the other optimized for gasoil or basic
petrochemicals (propene, in this case study). However,
the GTL plant may alter the situation, resulting in
expansion not taking place through refineries, depend-
ing on the criteria adopted, or expansion through only
one new refinery by 2015, optimized for gasoil or
petrochemicals. However, the second option has the
advantage of offering integration economies, producing
high-value oil products such as propene, being smaller,
adding value and making good use of natural gas
produced in Brazil. In this sense, the margin of this
refinery is less important than the supply security it
offers for gasoil and petrochemicals. Consequently,
uncertainties on the oil products market may favor the
selection of the GTL plant as the best solution.
Finally, several factors influence investment decisions
in the oil refining segment, including local infrastructure
conditions, the possibilities of partnerships among
enterprises, corporate strategies, raw material and
products transportation costs, environmental issues,
fuel specifications, the availability of oil products
worldwide, oil and oil product price scenarios and
government policies. It is the combination of all these
factors—together with the expansion criteria simulated
in this paper—that will justify future investments in
Brazil’s oil refining segment.
Acknowledgments
We thank the Brazilian Petroleum Institute (IBP),
CAPES Foundation; CNPq foundation; the National
Petroleum Agency and my colleagues with the IBP,
particularly José Fantine.
References
Adelman, M.A., 2002. World Oil Production & Prices 1947–2000. The
Quarterly Review of Economics and Finance 42 (2), 169–191.
API, 2000. Technology Roadmap for the Petroleum Industry.
American Petroleum Institute, Washington, DC, USA (February,
draft version).
Babusiaux, D., Champhon, D., Valais, M., 1983. Aggregate oil refining
models. Energy Exploration & Exploitation 2 (2), 143–153.
Bensaid, B., 2005a. Alternative motor fuels today and tomorrow.
Panorama 2005. Institut Franc- ais du Petrole. Available on line at
http://www.ifp.fr
Bensaid, B., 2005b. Road transport fuels in Europe: the explosion of
demand for diesel fuel. Panorama 2005. Institut Franc- ais du
Petrole. Available on line at http://www.ifp.fr
BP, 2004. BP Statistical Review of World Energy. Available on line at
http://www.bp.com
Brown, J., 2004. China will continue leading product demand growth
in Asia-Pacific. Oil and Gas Journal, 52–54.
3039
Bonelli, R., Gonc- alves, R., 1999. Padrões de desenvolvimento
industrial no Brasil 1980–95. IPEA, Rio de Janeiro, Brazil.
Campbell, C., 2002. Forecasting Global Oil Supply 2000–2050.
Hubbert Center. Newsletter 3, July 1–8. Available on line at
www.hubbertpeak.com/campbell
CNPE, 2002. O setor energético brasileiro: a situac- ão em 2001 e
perspectivas. Brası́lia, D.F. Brazil.
Coutinho, L., Laplane, M., Kupfer, D., Farina, E., Sabbatini, R.
[coords], 2003. Estudo de competitividade por cadeias integradas:
impactos de zonas de livre comércio. UNICAMP, UFRJ, USP,
MDIC/FINEP, Campinas, Brazil.
EMBRAPA, 2002. Cenários do Ambiente de Atuac- ão das Organizac-
ões Públicas de Pesquisa, Desenvolvimento e Inovac- ão para o
Agronegócio Brasileiro 2002–2012. Ministério da Agricultura,
Pecuária e Abastecimento. Brası́lia, D.F. Brazil.
EIA US-DOE, 2001. International Energy Outlook 2020. Energy
Information Administration, US Department of Energy, Washing-
ton, DC.
EIA US-DOE, 2002. Supply Impacts of an MTBE Ban. September.
Energy Information Administration, US Department of Energy,
Washington, DC Available on line at http://www.eia.doe.gov.
EIA US-DOE, 2004a. International Energy Outlook. Energy Informa-
tion Administration, US Department of Energy, Washington, DC
Available on line at http://www.eia.doe.gov/international/re-
ports.html#Annual.
EIA US-DOE, 2004b. Challenging Times for Making Refinery
Capacity Decisions. In: NPRA Annual Meeting. Available on line
at: http://www.eia.doe.gov
EPA, 1995. Profile of the Petroleum Refining Industry. Office of
Enforcement and Compliance Assurance. September. Available on
line at: http://www.epa.gov/compliance/resources/publications/as-
sistance/sectors/notebooks/petroleum.html
Fairbanks, M., 2003a. Petrobras—Investimentos ajustam Refino para
usar mais Óleo Nacional. Quı́mica e Derivados. n. 407. Outubro
15–19.
Fairbanks, M., 2003b. Descobertas feitas no Cinqüentenário ajudam
Companhia a alcanc- ar mais cedo a meta da auto-suficiência, além
de ampliar a oferta de óleo leve. Quı́mica e Derivados, n. 420,
Outubro 30–34.
Giambiagi, F., 2003. Um Cenário para a Economia Brasileira
com Permanência da Austeridade Fiscal e Reduc- ão da Vulner-
abilidade Externa. Working Paper, 98, BNDES, Rio de Janeiro,
Brazil.
Halloch Jr., J., Tharakan, P., Hall, C.A.S., Jefferson, M., Wu, W.,
2004. Forecasting the limits to the availability and diversity of
global conventional oil supply. Energy 29 (11), 1673–1696.
His, S., 2005. Bio-fuels Worldwide. Panorama 2005. Institut Franc- ais
du Petrole. Available on line at http://www.ifp.fr
Horsnell, P., 1997. Oil in Asia—Markets, Trading, Refining and
Deregulation. Oxford University Press, Oxford, United Kingdom.
IEA, 2002. World Energy Outlook 2002. International Energy Agency,
Organisation for Economic Cooperation and Development, Paris,
France.
IEA, 2004. World Energy Outlook. International Energy Agency,
Organisation for Economic Cooperation and Development, Paris.
IMF, 2005. World Economic Outlook. International Monetary Fund,
Washington, DC, USA.
Laherrere, J., 2002. Hubbert’s Peak: The impending world oil
shortage. International Journal of Global Energy Issues 17 (4),
360–382.
Machado, G., 2004. Geopolı́tica do Petroleo e Suas Implicac- ões sobre
as Estratégias Produtivas e Tecnológicas para o Aproveitamento
dos Recursos Petrolı́feros Nacionais. In Rio Oil & Gas Conference,
Rio de Janeiro. Brazil, October 4–7.
Moroney, J., Berg, D., 1999. An integrated model of oil production.
The Energy Journal 20 (1), 105–124.
 ARTICLE IN PRESS
3040 M.E.E. Tavares et al. / Energy Policy 34 (2006) 3027–3040
MME, 2002. Plano de Longo Prazo Projec- ão da Matriz—2022.
Sumário Executivo. Brazil. Available on line at http://www.mme.-
gov.br
MME, 2003. Balanc- o Energético Nacional de 2002. Brası́lia. Brazil.
Available on line at http://www.mme.gov.br
MPOG, 2002. Impactos econômicos do portfólio de investimentos dos
eixos nacionais de desenvolvimento: cenário tendencial 2001–2020.
Ministério do Planejamento, Orc- amento e Gestão (performed by
FIPE e Monitor Group do Brasil), Brası́lia, D.F., Brazil.
MPOG, 2004. Impactos econômicos do portfólio de investimentos dos
eixos nacionais de desenvolvimento: cenário tendencial 2001–2020.
Ministério do Planejamento, Orc- amento e Gestão (performed by
FIPE e Monitor Group do Brasil), Brası́lia, D.F., Brazil.
Nakamura, D., 2003. Refining capacity creeps higher in 2003. Oil and
Gas Journal, 64–69.
NEPD Group, 2001. National Energy Policy: Report of the National
Energy Policy Development Group. US Government, Washington,
DC.
Petrobras, 2004. Plano Estratégico da Petrobras—2015. Available on
line at http://www.petrobras.com.br
Schaeffer, R., Machado, G., 2005. Chapter 5: Energy and Economic
Development. In: IAEA (Org.). Brazil: A Country Profile on
Sustainable Energy Development. International Atomic Energy
Agency. United Nations Vienna, in press.
Schaeffer, R., Szklo, A., Machado, M., Mariano, J., Sala, J., Tavares.
M., Magrini, A., 2004a. Evoluc- ão do Mercado Brasileiro de
Derivados de Petróleo e Perspectivas de Expansão do Parque de
Refino Nacional até 2015. PPE/COPPE/UFRJ, Rio de Janeiro,
Brazil, October.
Schaeffer, R., Szklo, A., Machado, G., 2004b. Matriz Energética
Nacional 2023. Ministério de Minas e Energia, Brazil.
Stell, J., 2003. Worldwide construction update. Oil and Gas Journal.
Szklo, A.S., Machado, G., Schaeffer, R., Simôes, A.F., Mariano, J.B.,
2004. Placing Brazil‘s heavy acid oils on international markets.
Energy Policy, doi:10.1016/j.enpol.2004.07.001.
Williams, B., 2003. Refiner0 s future survival hinges on adapting to
changing feedstocks, product specs. Oil and Gas Journal, 13–18.
Wilson, D., Purushothaman, R., 2003. Dreaming with BRICs: The
Path to 2050. Global Economics Paper, No. 99, Goldman Sachs,
New York, USA.
Worrell, E., Galitsky, C., 2004. Profile of the Petroleum Refining
Industry in California. California Industries of the Future
Program. Energy Analysis Department, Environmental Energy
Technologies Department, Berkeley, USA.
Zai-Thing, L., Chao-Gang, X., Zhi-Gang, Z., Jiu-Shun, Z., 2002.
Olefin production technology with adjustable propylene/ethylene
ratio by catalytic cracking route. In: 17th World Petroleum
Congress, Rio de Janeiro, Brazil.