Investigation into the performance of oil and gas projects
Zhenhua Rui a, *
, Fei Peng b
, Kegang Ling c
, Hanwen Chang d
, Gang Chen a
, Xiyu Zhou a
a
University of Alaska Fairbanks, USA
b
China University of Petroleum (Beijing), China
c
University of North Dakoda, USA
d
Nexen, USA
a r t i c l e i n f o
Article history:
Received 16 September 2016
Received in revised form
27 October 2016
Accepted 24 November 2016
Available online 27 November 2016
Keywords:
Oil and gas project
Project performance
Cost performance
Project management
Project economics
a b s t r a c t
Many oil and gas projects experienced significant cost overruns, which is a major concern for the industry.
The objective of this study is to investigate the cost performance of oil and gas projects by
analyzing the data of approximately 200 public oil and gas projects. The average cost overrun of the
projects is 18% with a standard deviation of 29%. The results also indicate that the error of underestimation
is more frequent and greater than that of the overestimation. The projects' cost performance is
also examined in terms of project size, type, region, joint venture information, and Final Investment
Decision (FID) year. All effects of each factor are tested with statistical methods, and various drivers for
cost performances are suggested to explain the differences in cost performance. The findings of this
research will provide some guidance and references for future improvement in project performance.
© 2016 Elsevier B.V. All rights reserved.
1. Introduction and background
Average annual global oil & gas (O&G) project capital project
investment is about US$1 trillion per year between 2011 and 2035
(IEA, 2011). However, the performances of the O&G projects did not
get any attention until the recent oil price crashes. There are many
metrics to evaluate project performance (e.g., cost, schedule,
operability, and production). Cost performance is the key metric for
determining whether a project is successful as cost is the aggregated
outcome of all project scopes. Therefore, investigating the
cost performance is a great approach for understanding the projects'
performance. Few reports or papers have mentioned the
performance of O&G megaprojects (greater than US$1 billion in
cost). About 78% of the megaprojects have serious cost overruns
and schedule slip (Merrow, 2012). EY (2014) reported similar
findings with 64% of O&G megaprojects experiencing a cost overrun.
More scholars have conducted research to investigate project
performance in other industries. Pohl and Dubravko (1992) investigated
more than 1000 projects from the 1940se1980s and they
had an average overrun of 21.5%. Flyvbjerg et al. (2004) found out
that more than 250 public projects had an average overrun of more
than 25%. Bertisen and Davis (2008) examined mining projects and
found that about 13.5% of them went over budget. Bordat et al.
(2004) reported more than half of Indiana public transportation
projects had a 5% overrun. Kolltveit and Grønhaug (2004)
concluded that Norwegian projects have experienced overruns
ranging from 6% to 160%. Rui et al. (2011a, 2011b, and 2012a)
investigated more than 200 compressor station projects and they
had an average 11% cost overrun. Rui et al. (2012a, 2012b, and 2013)
and Rui and Wang (2013) also found that the average cost overrun
of 400 pipeline projects is 7%. Shehu et al. (2014) found a 2% average
cost overrun for 359 Malaysian construction projects. Jacboy (2001)
found that 74 projects had an average 25% cost overrun. Jahren and
Ashe (1990) concluded that the naval facility project cost overruns
were between À10% and 100%. Given the reported research, it can
therefore be concluded that cost overruns are a global phenomenon
and cost estimating error exists across many different types of
projects. In addition to reporting the cost overruns, many researchers
tried to identify their causes. Some authors reported that
schedule delays are a key driver (Bordat et al., 2004). Others proposed
that project teams were incentivized to develop low cost
estimates in some cases (Flyvbjerg, 2014; Bertisen & Davis, 2008).
Bordat et al. (2004) and Jahren and Ashe (1990) proposed that large
projects have high complexity and there is a balance between
competitive bidding and economies-of-scale for large projects.
Merrow (2012) indicates that the quality of project Front-End-* Corresponding author.
E-mail address: zhenhuarui@gmail.com (Z. Rui).
Contents lists available at ScienceDirect
Journal of Natural Gas Science and Engineering
journal homepage: www.elsevier.com/locate/jngse
http://dx.doi.org/10.1016/j.jngse.2016.11.049
1875-5100/© 2016 Elsevier B.V. All rights reserved.
Journal of Natural Gas Science and Engineering 38 (2017) 12e20
Loading drives project outcomes. Rui et al. (2012a, 2012b, and 2013)
indicates that location, project size, and weather are factors
affecting project performance. EY (2014) pointed out that internal
factors (commercial context, project development, and project
delivery) and external factors (regulatory challenges and geopolitical
challenges) affect project performance. The performance of
public-private-partnership (PPP) projects is more competitive than
that for projects with traditional procurement; other studies indicated
that PPP projects have worse performance than infrastructure
projects using the typical procurement (Gleave, 2009). Flyvbjerg
(2004) mentioned the effect of ownership on project performance
and also explained the cost overrun issues using technical, psychological,
and political-economic factors. Le-Hoai and Lee (2008)
proposed that design issues, market issues, financial difficulty,
and governmental regulations are drivers of cost overruns. Nawaz
et al. (2013) found the drivers to be political interests, poor site
management, corruptions, and change orders. Takim (2005) indicated
that large contractors have worse performance, leading to
cost overruns. Frimpong et al. (2003) determined the major factors
affecting the performance of water projects, which included difficulties
of monthly payments, procurement management,
contractor issues, technical performance, and cost escalation.
Sambasivan and Soon (2007) analyzed 150 surveys and identified
major causes of schedule delays: poor contractor performance,
finance and payment, labor supply, resource availability, lack of
communication, dispute and litigation, and total abandonment.
Ahsan and Gunawan (2010) investigated 100 projects and found
that the major causes for the cost overruns to be the devaluation of
local currency, bidding price, and large contingency budget. Kaliba
et al. (2009) identified weather, scope changes, environmental
protection, strikes, technical challenges, inflation, local governments,
staffing, equipment unavailability, and poor management as
major reasons for the cost escalation and schedule delay for Zambian
projects.
The previous research on O&G project performance is very
limited. This paper will try to explore O&G project performance
with an in-depth analysis and provide some guidelines or references
for project teams. The mission of an O&G project is to extract
underground O&G for commercial use. Major process or steps for
oil and gas projects include: locate the O&G, produce the O&G from
underground, process the crude O&G, and transport it to a refinery
or customer. O&G upstream and midstream projects are quite
unique and specialized compared to other process projects or
public transportation and building projects. O&G projects are
relatively more capital intensive and natural resource based. Due to
various process steps, an O&G project involves a high number of
scopes: reservoir mapping, well drilling and completion, production
facility building, and transportation system developing. An
O&G project can be developed onshore, offshore, or via a combination
of both. Therefore, the complexity of O&G projects tends to
be high because of the numerous characteristics involved. Reducing
project costs is now the first priority for all O&G companies. Understanding
O&G project cost performance is critical for successfully
implementing cost reductions or controlling current and
future projects. Therefore, the objective of this paper is to investigate
the performance of O&G projects and identify different drivers
of the cost overruns.
2. Database and methodology
Project data are the foundation for quantifying the project performance
and understanding the project outcomes and root causes.
Scholars employed different ways to collect project performance
data. Some researchers designed and distributed a survey to owner
companies, contractors, and consultants for data collecting. The
survey contained information regarding the project characteristics.
The information includes project size, location, type of project,
project sector, project nature, procurement types, tending methods,
causes for cost overruns and delays, etc. (Le-Hoai and Lee, 2008;
Shehu et al., 2014). Some scholars worked with large international
organizations or government agencies to collect projects funded or
supported by these organizations and governments. For example,
Ahsan and Gunawan (2010) received data from 100 projects from
Asian development banks and Pohl and Dubravko (1992) collected
data from more than 1000 projects from the World Bank. Other
groups of researchers collected project data by using public sources,
such as publications, public databased, etc. For example, Rui, et al.
(2011a, 2013) collected more than 200 U.S. pipeline projects and
400 compressor projects from O&G journals. The data for this study
were collected from different public sources. Based predominantly
on exclusive reviews of all literature and public sources regarding
O&G project performance, data from 206 oil & gas projects were
collected from the following major data sources: Business monitor
international, the Global LNG Info website, the offshore technology
website, the Subseaiq website, the quest offshore website, Oil & Gas
journal, analyst reports from Thomson one, company websites and
annual reports, and the Society of Petroleum Engineers. The project
information was prepared on the basis of best-effort. The project
data include the project cost, FID year, project types, location,
company, joint venture (JV) information, etc.
The average project cost is US$765 million, with the individual
project costs spread over a wide range from less than US$10 million
to about US$9 billion. The average FID year is 2008, ranging from
2002 to 2013. The O&G projects in this sample are also well
distributed around the world and all O&G production regions are
included: 27% of the projects are from Europe, 23% are from Asia,
20% are from North America,11% from Africa,10% from Oceania, and
9% from South America. In addition, the project sample also includes
all concepts in Industry, including fixed platforms, floating
platforms, subsea systems, offshore wells, onshore wells, and
onshore production facilities. In general, the Industry we observed
was well represented in this project sample. The estimated costs
are the approved project budget or cost at the FID gate for project
development; the actual or final costs are the accumulated costs of
the completed project. A cost overrun rate was used to evaluate the
project cost performance. A positive rate indicates a cost overrun or
underestimate and a negative rate indicates an underrun or
overestimate.
To better understand the data pattern and differences among
the groups, various statistical techniques are employed in this study
for data analysis. A test was used with two sets of data to determine
if they are significant different from each other. Each test will
produce a p-value, which is used to determine whether the null
hypothesis is true, and generally, a p-value less than 5% is considered
to be a significant difference between two groups. The MannWhitney
test (MW) determines if it is equally likely that a randomly
selected value from one sample will be lower than or greater than a
randomly selected value from a second sample. The MW also produces
a p-value.
3. Findings
3.1. Overall cost performance
Cost performance provides great insight in understanding a
project's performance and identifying the drivers. At the FID gate,
the estimated project cost is one of the key factors affecting senior
management's FID decision and the unexpected consequence of the
overestimation or underestimation significantly influences the
project outcomes. In some cases, poor estimations even lead to
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e20 13
company financial difficulties due to large capital cost overruns.
Therefore, cost performance is an important metric for evaluating
O&G projects and companies.
As shown in Fig. 1, a relatively small number of cost overrun
rates are narrowly spread around 0 and a large number of cost
overrun rates are distributed far above 0. If the frequency of the
overestimated cost is the same as the underestimated cost, the
histogram would be symmetrical. Fig. 1 shows that cost overrun
rates are widely spread and have non-symmetric distribution. More
than 150 (74.0% of the total) projects were underestimated and 50
projects (25.0% of the total) were overestimated. The average cost
overrun rate was 18.2%, which indicates that the average final cost
of the collected projects was 18% higher than the estimated cost.
The overrun cost was significant due to the project's large size. The
overrun rates ranged from À32% to 169% with a large standard
deviation (SD) of 29% and only 36% of the projects were completed
within the À10% to 10% cost overrun range. The large range and SD
imply that O&G project performance is very unstable and there are
serious cost predictability issues as the performance is much worse
than the expected accuracy range of À20% to 30% (AACE, 2016). The
high average cost overrun and large deviation indicate that
reducing cost overruns and improving O&G project cost predictability
are badly needed.
To understand the biased level and accuracy level of cost estimation
for O&G projects, a binomial statistical test is used, which
confirms whether the frequency of overestimated projects is equal
to that of underestimated projects. The test results (Pr < 0.01)
indicate that underestimating is more frequent than overestimating
for O&G projects. Hence, the O&G project cost estimations
are biased. The MW test results (Pr < 0.01) show that there is a
higher magnitude of underestimating errors than overestimated
errors. Therefore, O&G projects are more likely to have underestimating
errors than i overestimating errors. This conclusion
aligns with the findings from projects in other industries (Flyvbjerg
et al., 2004). Flyvbjery et al. (2014) and Bertisen and Davis, (2008)
suspected that project teams tend to underestimate project costs.
EY (2014) also mentioned that project teams of megaprojects
underestimated the risks associated the projects in the planning
phase. Based on our observations and O&G project cost performance,
O&G project estimators or project teams are also more
likely to have aggressive target setting problems and optimistic bias
issues. We also observed that many potential opportunities and
projects in the portfolio channels are waiting for management
approval, especially for large O&G companies. A similar observation
from Pergler and Rasmussen (2014) indicates that a large number of
projects or opportunities have different priorities in the portfolio,
but the company has only enough capital to execute some of them.
To compete with other projects or opportunities, many project
teams or project managers are very likely to underestimate the cost
by ignoring risk or hiding risk associated projects. This irresponsible
action later on causes serious issues for project execution and
on the outcomes. On the other hand, many project teams fail to
recognize the project risks and complexity due to a lack of capacity
and experience, which is very common for new technology projects
or megaprojects. The data show that a high percentage of O&G
projects cannot meet their planned goals because of the high cost
overrun rate issue. Thus, fully understanding the drivers for project
cost overestimating or underestimating will provide some useful
guidance for improving project performance. Therefore, the more
detailed analysis of project performance regarding the project
characteristics is conducted in the following sections.
3.2. Cost performance in terms of project sizes
Project cost is used as a measurement of project size in this
study. Many scholars have also mentioned that the size of projects
affects cost overrun. Flyvbjerg et al. (2004) suggested that larger
bridge and tunnel projects tend to have large cost overruns, but this
is not found in road projects. Cantarelli et al. (2012) found that the
cost overruns of road and fixed links decrease with project size, but
the cost overruns of rail projects increase with project size. Rui et al.
(2011a), Rui (2012c and 2012d) concluded that cost overruns of
pipeline project total costs are affected by project size and large
pipeline projects have large cost overruns, but not for individual
component costs. Rui et al. (2013) also indicated that project size
does not influence the cost performance of compressor station
projects. Shehu et al. (2014) found that smaller Malaysia construction
projects have larger cost overruns. Previous research
literature indicated that project size has different degrees of effect
on cost overruns in terms of the different types of projects. This
section is to quantify the effect of project size on O&G project cost
performance.
The O&G project size is much larger than that reported for
projects in other industries (PWC, 2014). According to the
0
5
10
15
20
25
30
35
40
45
-30% -20% -10% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 105% +
NumberofProjects
Cost Overrun Rate
Fig. 1. Overall cost performance of the O&G projects collected.
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e2014
distribution of estimated project costs, project is classified into
three size groups. Small-sized projects are defined as those that
cost less than US$200 million; medium-sized projects are defined
as those between US$200 million and US$1 billion; and projects
costing more than US$1 billion are defined as large size projects or
megaprojects. In this study, each size project group has a decent
number of representative projects: 80 projects in the small size
project group with an average cost of US$104 million, 76 projects in
the medium size project group with average cost of US$461 million,
and 49 projects in the large project group with an average cost of
US$2. 3 billion.
The cost performance for each project size group is shown in
Table 1. The overrun rate is incrementally higher with the
increasing project size, but the relationship between the overrun
rate and project size is not straight linear. The cost overrun rates
between the small and medium group are not significantly
different (Pr > 0.5) with cost overrun rates of 15.5% and 14.6%,
respectively. However, it clearly shows that megaprojects have the
largest cost overrun rate of the three at 25.5% and the highest SD
(Pr < 0.01). In summary, project size affects project cost performance
and the performance for large projects is much worse than
small and medium groups.
We try to explore some potential causes for large O&G project
cost overruns in this section. With project size increasing, there is a
higher possibility for a larger number of JV partners. The practice of
a JV with multiple partners has become more common. However, as
the number of partners grows, it becomes more and more difficult
to align the objectives and secure funds from multiple partners. EY
(2015) indicates that a JV takes about 18 months to establish and
the failure rate of JVs is as high as 70%. In addition to the high
number of partners, a large project often is a high profile project
under pressure and being influenced by various parties around the
project, such as non-governmental organizations, local residents,
local governments, the federal government, etc. Typically, the
period of planning and negotiation with different stakeholders will
take 5 years or longer for large O&G projects. The effort and time to
reach an agreement with local residents or local natives and receive
permits from multi-government bodies or different countries are
significantly higher for larger projects than smaller projects. One
prime example is the purposed keystone XL pipeline project from
the Canada to the United States, which took more than 10 years to
be ultimately rejected by the U.S. government in November 2015.
The difficulty and risk are even higher when dealing with unstable
governments. Therefore, the associated cost of dealing with
different stakeholders is hard to accurately estimate and forecast
due to its high uncertainty. As is known, the mission of an O&G
project is to extract O&G from subsurface. Since the first commercial
well for production was drilled on August 28, 1859, a large
amount of the easy oil from underground has been produced or is
in production to meet global energy demand and the remaining
undeveloped O&G reserves are becoming more difficult to be
accessed and produced. To meet the growing world energy demand,
O&G companies started to explore risker and more technically
challenging regions with undeveloped natural resources, such
as the Arctic regions, ultra-deep offshore fields, etc. They also
developed a new technology for producing the O&G that previously
was impossible to be commercially produced. The shale O&G
development in North America using the new horizontal fracture
technology and advanced stimulation methods is an example (Zhao
et al., 2015a; Zhao et al., 2015b; He et al., 2016; Ren et al., 2016; Sun
et al., 2016; Zhao et al., 2016). To justify the economics of producing
difficult oil, more complicated large size O&G projects are being
planned and developed and these projects tend to employ more
new technologies and have a large number of sub-scopes. The new
technology may increase operation efficiency in the future, but
implementing the new technology application during initial
development is not an easy task and will incur a higher cost and
longer duration than a proven technology. Further, the supply of
new technology is limited. For example, there are only three or four
O&G service companies in the world that can provide subsea
equipment and installation services. The lack of experienced teams
and historical information on new technologies are common.
Merrow et al. (1979) reported that new technology can add an extra
10%e20% of the total cost. Combinations of these factors make
predicting the associated cost of implementing a new technology
extremely difficult. Meanwhile, large projects often have a high
number of sub-scopes with a high complexity due to their large
capacity, complicated reservoirs, extreme weather conditions, etc.
The high number of sub-scope increases the number of interfaces
exponentially, and each additional interface increases the difficulty
of integration, communication, project management, etc. Further,
most large O&G fields are located in remote areas or developing
countries (e.g., Africa). Unavailable or the lack of established
infrastructure is a huge potential risk that can cause projects to not
be completed within the proposed budget and time. This is because
the companies have to spend extra time and cost to build the
infrastructure for accessing the resources or transporting the produced
O&G to clients. Megaprojects were also subject to labor issues
and geopolitical challenges that are not common in relative
small projects because megaprojects have substantial effects on
society, the environment, and the government's fiscal revenue.
Therefore, it is not surprising to see that large O&G projects
experienced significantly higher cost overruns.
3.3. Cost performance in terms of region
Various O&G projects are developed in different regions to
extract natural resources. The characteristics of the O&G projects in
these different regions vary greatly, so the geographical region may
be affecting project performance. Some researchers have already
looked at the effect of the region on project performance. Flyvbjerg
(2014) indicates that Dutch projects perform better with lower cost
overruns compared to other geographical locations and suggested
the geography factor should be taken into consideration. Olaniran
et al. (2015) reported different megaproject cost overruns for projects
in different regions, finding that different local or state governments
may affect separated contract items. Cost performance is
related to location because the technical requirements and Best
Practices for engineering and installing offshore platforms vary in
different regions (Hall and Delille, 2011). They indicated that
pipeline project performances are significantly different for the
different U.S. regions and proposed that weather, the surface
Table 1
Cost performance in terms of different project size groups.
Project size groups Num. of projects Average cost overrun SD Min Max
Small Size Project 80 14.6% 28.8% À28.0% 157.0%
Medium Size Project 76 15.4% 26.0% À25.0% 109.0%
Large Size Project 49 25.5% 30.5% À14.0% 169.0%
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e20 15
conditions, remoteness, and availability of supplies are variables for
determining the regional difference (Rui et al., 2013; Zhao, 2000;
Zhao and Rui, 2014). Very limited information is available on
O&G project cost overruns across different regions from public
sources. This section will demonstrate the cost performance for
different regions and identify which regions perform better. In
addition, the regional variation will be examined and possible
causes for the overrun will be provided.
As seen in Table 2, each region has a good number of projects in
the dataset. Europe has the largest share of the project sample at
about 27%, followed by Asia with 23%, North America with 20%,
Africa with 11%, Oceania with 10%, and South America with 9%. All
O&G producing regions are included. It is obvious that the average
cost overrun rate varies among the regions. The African projects
average cost overruns of 35.3%, followed by South America at 21.0%,
Oceania at 19.7%, Asia at 15.7%, North America at 15.6%, and Europe
at 13.6%. The test results (Pr < 0.01) confirm that the average cost
overruns vary across the different regions, and Europe and Africa
are significantly different from the other groups in terms of the cost
overrun rate. The African project cost performance has the largest
SD, which is significantly higher than the other groups (Pr < 0.01).
The high average cost overrun and SD imply that African projects
have serious cost overrun issues and the poorest cost predictability.
European projects have the lowest cost overruns and relatively
small SD, which indicates that European projects are much better
controlled than those in other regionals. Therefore, more investigation
on projects from Europe and Africa will provide more
valuable information to better understand the performance drivers.
Projects in Europe have the lowest cost overrun rate, and the
cost of living in Europe is relatively high. The cost performance of
European projects indicate that the cost of living-related factors
may not significantly influence the high cost overrun rate, and that
other factors may be more influential. The data limitations and lack
of experience for certain regions may result in a high cost overrun.
Some scholars (Flyvbjerg et al., 2004; Rui et al., 2011a; Cantarelli
et al., 2012) proposed that the standard of professionalism and
decision-marking system or culture of governance may also play
role in project performance in different regions. Among all regions,
Europe has the highest number of projects and most of them are
located in the North Sea region as offshore projects. The concentration
of European projects is much higher than that for other
regions. Therefore, project teams in European regions gathered
more practical experience, had a good understanding of the subsurface
and surface situation, and had better access to the historical
information. In addition, Europe generally has a high standard of
professionalism and has produced a large amount of good engineers.
These advantages make project performance in Europe very
competitive.
The statistic test results also show that African projects have
significantly higher cost overruns with most inconsistent performance
compared to the other regions. The cost overrun rate is 10%
higher than the global average overrun and more than two times
the average cost overrun for European region projects. Industry has
recognized the difficulty of implementing projects in Africa, and
many factors have been found to drive the high cost overrun rate.
PWC (2014) indicates that inadequate pre-engineering, weak
project management, internal procurement, funding issues,
governmental complications, and supply chains significantly affect
African project performance. However, many drivers of the difficult
development of O&G projects in Africa have not been very well
identified and documented. In addition to the regular causes of cost
overruns, there are still a few unique reasons or serious issues
associated with African O&G project development (e.g., regulatory
uncertainty and government stability). A World Bank report indicates
that all African O&G production countries are the most
difficult countries for doing business (World Bank Group, 2014). A
good example is the local content policy. Local content, or the requirements
for the O&G industry to generate more welfare or
benefit to the local economy in addition to directly adding value, is
a common policy in the O&G sector in Africa or other oil-producing
regions (Tordo et al., 2013). The local content policy requirement is
implemented to improve local employment, skill development, and
local industry or company participation. O&G-producing countries
have established various levels of local content requirements and
most of them require a high percentage of local participation (Tordo
et al., 2013). Africa is natural resource rich continent, and there are a
huge amount of underground resources waiting to be developed.
Most projects involve a large reserve and high complexity and are
relative large. Therefore, almost all large O&G projects are developed
by foreign O&G companies with a JV. According to the local
content requirement, over 90% of the labor and technical professionals
should be local residents and most equipment and
fabrication should be supplied or done by local contractors (Tordo
et al., 2013). Due to the high natural complexity and uncertainty
of the projects, a significant amount of the goods and services are
highly technological and specialized in nature. In reality, at present,
African O&G-producing countries are incapable of providing
adequate qualified local staff and contractors as required. Therefore,
there is always a conflict between the local content requirement
and local capacity in O&G project development in Africa. This
conflict leads to unexpected schedule delays and cost overruns
because more resources and time are needed to build the local
capacity to complete the project. This situation worsens due to the
frequent changes to these local requirements. It is a wellrecognized
fact that most Africa O&G-producing countries have
serious government instability and diplomatic and security issues.
These issues are critical for effectively implementing O&G projects,
and the risks involved are very likely to add extra cost. All uncertain
factors are summed up to significantly increase the difficulty of
developing projects in Africa. Therefore, a set of Best Practices is
greatly needed for improving O&G project performance in Africa.
For example, more investigation work regarding the local region or
government should be conducted to understand the true uncertainty
and risk and prepare a contingency plan or assign a good
amount of contingency in the cost estimate.
Therefore, it can be concluded that cost performance differs by
region, and the different regions have common and unique factors
that drive the cost performance, and these regional factors should
Table 2
Cost performance in terms of region.
Region Num. of projects Average cost overrun SD Min Max
North America 41 15.6% 22.3% À25.0% 94.0%
Europe 56 13.6% 25.4% À32.0% 108.0%
Africa 23 35.3% 42.5% À26.0% 152.0%
South America 19 21.0% 28.3% À17.0% 169.0%
Oceania 21 19.7% 31.9% À14.0% 132.0%
Asia 47 15.7% 31.3% À28.0% 157.0%
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e2016
be considered in the planning and estimation phase for O&G
projects.
3.4. Cost performance in terms of the joint venture
With the increasing project size and complexity, the JV approach
is widely used in the O&G industry to share risk and increase the
capability between organizations or companies. The JV partners
may include companies, government bodies, and/or financial institutions.
A JV at a corporate level has been widely discussed and
researched in the financial and management sectors. However,
there is only a handful of literature on the JV project performance.
Each company or organization has different objectives or reasons to
pursue JV projects. EY (2015) listed eight major positive reasons for
entering a JV partnership: capital intensity, risk mitigation, access
to technology, access to resource, supply chain optimization, market
position and scale, regulatory requirement, and political
sensitivity. Meanwhile, the challenge incurred by the JV is also not
easy to overcome and cost overruns of JV projects may be higher
than in non-JV projects. The reports (EY, 2015) also mentioned
seven major challenges regarding JV projects: the JV structure,
alignment of goals and strategy, trusted relationship, integration,
project leadership, timing of exit, and dispute resolution. All advantages
and disadvantages aggregate to affect JV project performance
as a whole. An investigation into project performance
associated with JVs will help to identify drivers and establish Best
Practices for ensuring JV projects are successful.
Just 19.5% of the projects are fully owned by one company. More
than 80% were developed though a JV, meaning a JV is a common
industry practice in the O&G industry for project development. This
section tries to answer three questions regarding JVs. First, is the
number of JV partners driven by project size? Because large projects
tend to have s high capital requirement and complexity, large
size projects are expected to have a high number of partners.
Table 3 shows the number of JV partners ranges from 1 to 7 with the
average number being 2.9. The average number of JV partners is
almost the same between the small and medium size groups.
However, large projects have an average of 3.5 JV partners, which is
significantly higher than the other two groups (Pr < 0.01). This
result confirms that as project size grows, so does the number of JV
partners on O&G projects. The second question is whether the JV
partnership affects the cost performance difference between JV
projects and non-JV projects. As seen in Table 4, the average cost
overrun rate for non-JV projects is 17.4%; for JV projects, it is 18.3%.
In addition, the test (Pr > 0.5) shows that the cost differences between
JV and non-JV projects is insignificant in terms of the average
cost overrun and variance. Thus, it can be concluded that the JV
type does not affect the cost overrun for an O&G project as a whole.
The final question is whether the number of JV partners influences
the cost overrun performance for JV projects. Table 5 shows that JVs
with two or three partners are the most common JV format, accounting
for 48 percent of the total. However, there are still a good
amount of projects in the other groups. The cost overrun rate varies
in terms of the different number of JVs, and the differences are
significant (Pr < 0.05). The average cost overruns among the five
different groups range from 13.5% to 25.7%. Projects with two JV
partner have an average cost overrun of 25.7% with an SD of 38%,
which is significantly higher than the other groups (Pr < 0.01).
In summary, the number of JV partners increases with project
size. On average, the differences in the cost performance for JV and
non-JV projects are insignificant, but the number of JV partners
significantly affects the O&G project performance of JV projects and
projects with two JV partners have significantly higher cost overruns.
The mix of advantages and disadvantages in JV projects
accumulate together to affect O&G project performance. Different
factors associated with JV projects may affect the project's cost
performance at different degrees or in different directions (e.g.,
partner type, equity share, etc.). The findings from the study are
informative and valuable, but further research regarding JV projects
will focus on quantifying the risk of the factors with more detailed
information.
3.5. Cost performance in terms of company type
The petroleum industry is comprised of a wide diversity of
companies of a different size and type. Most large O&G reserves or
producing countries have national oil companies (NOC), which are
either fully owned or partially owned with majority shares held by
the federal or local governments. An example is Saudi Aramco, the
largest NOC in the world. About 75% of the world's oil production
and 90% of the proved reserves are dominated by NOCs (Tordo,
2011). The general impression around NOCs is that they are characterized
by operational inefficiencies and a lack of competition.
Non-NOC O&G companies include international oil companies
(IOC) and/or independent companies. IOCs are integrated upstream
and downstream companies with worldwide operations. The top
five major IOCs are ExxonMobil, Royal Dutch Shell, BP, Total, and
Chevron. Independent companies normally are medium or small
sized private or public O&G companies (e.g., Anadarko). The major
difference between IOCs and independent oil companies are
whether the companies have downstream sectors. In this study, all
projects in this sample are upstream and midstream projects, so
comparisons are made between NOC projects and non-NOC projects.
Most NOC-developed projects are executed in the native
country, but we have seen dramatic changes in recent years to NOC
operations. Some NOCs are increasing their project investment
outside of their native countries. Most current O&G projects have
multiple JV partners and the partners can be different combinations
of the different company types. We, therefore, defined a project as
an NOC project if more than 50% of it was owned by the NOC;
otherwise, it is considered a non-NOC project. Research also found
some findings regarding the performance of public (government)
owner and private owner projects. It is widely accepted that publicly
owned projects are slow, costly, and of poor quality compared
to privately owned projects. Hwang et al. (2010) conducted a study
on the cost performance of publicly owned infrastructure projects
compared to privately owned projects by analyzing the performance
of 341 projects, and found that privately owned projects
have a strong focus on cost control and profit. Flyvbjerg (2004)
reported that there is a significant cost overrun difference between
private and state-owned projects, but that the finding was not
conclusive due to the small sample size. He also suspected that
publicly owned projects lack the pressure and competition for
Table 3
Number of JV partners in terms of project size.
Cost group Average num. of partners SD Min Max
Small Size Project 2.7 1.6 1 6
Medium Size Project 2.8 1.6 1 6
Large Size Project 3.5 2.0 1 7
Overall 2.9 1.7 1 7
Table 4
Cost performance in terms of JV and non-JV.
JV forms Num. of projects Average cost overrun SD Min Max
JV 170 18.3% 30.8% À32% 169%
Non-JV 36 17.4% 25.5% À28% 132%
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e20 17
improving performance and accountability. Regrettably, there is
very limited information on project performance in terms of
company type, aside from some general information regarding the
corporate performance of NOCs and non-NOCs. NOC O&G projects
differ from the above mentioned government projects because the
major objective for NOC projects is the same as that for privately
owned projects, maximizing profit and minimizing cost, whereas
most government or publicly owned projects involve public interest
and welfare, such as highways, hospitals, public facilities, etc.
As shown in Table 6, non-NOC projects, which account for 59% of
the data sample, have an average cost overrun of 17.6%. NOC projects
have a slightly higher overrun rate of 18.9%. However, the test
results (Pr > 0.5) show that the effect of company type on the
average cost overrun between non-NOC and NOC groups is insignificant.
However, the SD test (Pr < 0.05) shows that the NOC SD of
0.33 is significantly larger than that for the non-NOC group of 0.28,
meaning the performance of NOC projects is worse than that for
non-NOC projects. Although the NOCs had widespread technology,
project management, and other technical capability issues, they
may also have some advantages: access to energy sources, local
experience, social and economic development, etc. (Wainberg and
Foss, 2007). NOCs in the OPEC groups perform better than their
private sector competitors in terms of capital and labor efficiency
and profitability (Wolf, 2008); NOCs outside the OPEC group are
less competitive than their counterparts from the private sectors
(Wolf, 2008). Our research shows that the cost performance difference
is insignificant for different types of ownership, but the
performance of NOC projects is less stable than that for non-NOC
projects.
3.6. Cost overrun in terms of project type
Numerous O&G projects for different functions and sizes are
developed around the world each year. All projects can be categorized
into two groups: greenfield and brownfield. The greenfield
projects are normally built on a new and undeveloped site; the
brownfield projects refer to expansion or revamp projects on
existing facilities. Both project types are common in the petroleum
industry, and selecting greenfield or brownfield is determined by
various factors. Greenfield and brownfield projects each have their
own advantages and disadvantages. In general, the advantages of
developing greenfield projects include no demolition, a flexible
design, new regulation requirements, operation efficiencies, etc.
The major disadvantages of greenfield projects include access to the
site, lack of historical subsurface and surface data, lack of prior local
experience, planning permit approval, etc. Many of the advantages
and disadvantages of brownfield projects are that opposite of those
related to greenfield projects. The advantages of brownfield projects
include historical data, existing environmental or regulation
approvals, existing infrastructure and facilities, prior local experience,
etc. The disadvantages of brownfield projects include the
limited design options, operation efficiency, difficulty of site operations,
removing an unused existing facility, etc. Skitmore et al.
(1990) found that refurbishment projects had higher cost overruns
than new projects. The same was found for Malaysian construction
projects (Shehu et al., 2014). However, there is no public
information that shows that performance difference between
greenfield and brownfield O&G projects.
In the data sample (see Table 7), 43% of the projects are
brownfield projects and the rest are greenfield projects. Seen from
Table 7 the average brownfield project cost overrun is 16.7% with a
range from À32% to 152%. The greenfield projects have an average
cost overrun of 18.1% with a range from À28% to 169%. The statistical
tests (Pr > 0.5) indicate that the cost performance between the
greenfield and brownfield groups is insignificant in terms of the
average cost overrun and SD. Therefore, project type appears to
have no significant effect on O&G project cost performance and
neither greenfield nor brownfield projects are more prone to cost
overruns. Minor quantified cost differences between greenfield and
brownfield projects eliminate some concerns for companies to
make decisions on using a greenfield or brownfield project.
3.7. Cost performance over the years
Some literature indicated that cost overruns have been common
for years for different industries or different project types. Through
an analysis of large transport projects in The Netherlands, Cantarelli
et al. (2012) concluded that there is no relationship between year
and cost overrun, and cost estimates do not improve over time. The
cost overruns are stable for the 88-year period from 1910 to 1998
for more than 200 transportation projects in 14 countries worldwide
(Flyvbjerg, 2014). In this study, O&G project performance is
examined over time. With the help of new techniques and/or tools
for estimation and advanced knowledge of projects, improved cost
performance over time is expected. The year of FID is used as the
year for measurement instead of the year of completion because
the estimated cost is one of the major factors influencing the final
decision on whether to proceed with the project (Flyvbjerg, 2014).
Fig. 2 shows the cost overrun fluctuation over the years, but there is
no clear trend over the years. The test results (Pr > 0.1) show that
there is no significant difference between 2002 and 2013, which
implies that new methods or new techniques for cost estimation
are not leading factors that affect cost performance, which is in line
with the observations from other scholars (Flyvbjerg, 2014;
Table 5
Cost performance in terms of number of JV partners.
JV partners group Num. of projects Average cost overrun SD Min Max
2 Partners 55 25.7% 37.7% À23.0% 169.0%
3 Partners 44 17.1% 26.9% À26.0% 109.0%
4 Partners 27 13.5% 24.0% À14.0% 118.0%
5 Partners 22 14.7% 21.1% À21.0% 93.0%
6 or more Partners 22 15.0% 20.3% À18.0% 86.0%
Table 6
Cost performance in terms of company type.
Company type Num. of projects Average cost overrun SD Min Max
NOC 84 18.9% 28.1% À32.0% 169.0%
Non-NOC 122 17.6% 32.0% À28.0% 157.0%
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e2018
Cantarelli et al., 2012). New estimation or technical forecast tools
were developed and applied over time. However, the O&G project's
complexity and the associated risk also climb over the years due to
O&G reserve's nature. Karev (2014) indicates that the increased
reserve risk leads to a higher average risk profile for O&G megaprojects
over the years because of increased capital spending on
new cutting edge projects. In summary, cost overruns show some
variations over the years, but the O&G project cost performance has
not systemically improved over time.
4. Conclusion and suggestion
This study investigates O&G project cost performance in terms
of project size, company type, JV, region, project type, and year.
There are major findings as below: 1. The overall average cost
overrun of O&G projects is 18% with an SD of 29%. The overruns for
O&G projects are more frequent than underruns and the magnitude
of overrun is also significantly higher than the magnitude of underrun.
All findings indicate that the overall cost performance of
O&G projects is worse than expected, implying that project teams
fail to incorporate the risks in the planning phase. 2. The cost
overruns for the large project group are significantly higher than
that for the small size project group because large projects have
more uncertainties. 3. The cost overrun rate varies across regions.
European projects have the lowest cost overruns, but African projects
have significantly higher cost overruns. Projects in each region
have their own unique factors that affect cost performance.
Learning experience, professional standards, the local content
policy, the local capacity, and government stability were factors
that potentially could be driving the regional difference. 4. The
analysis of JV projects indicates that the number of JV partners
increases with project size, but there is no statistical difference
between the performance of JV projects and non-JV projects.
However, the number of JV partners affects the cost performance
for JV projects. 4. The difference between NOC and non-NOC project
performance is insignificant in terms of the average cost overrun,
but NOC project performance is worse because of larger variations
in NOC project performance. 5. The performance differences between
greenfield and brownfield projects are very minor. The cost
performance varies over the years, but there appears to be no
improvement over time.
The differences in cost overruns across the many different
groups have been analyzed and tested, and potential causes or
drivers for the underruns have been investigated or suggested;
these findings will provide some insights into the cost estimators or
project teams during planning. The project team should conduct a
significant amount of preparation work to understand the project
better and quantify the different risks related to the unique characteristics
of each project. Then, different project risk settings or
contingencies should be assigned based on project-specific situations
to avoid cost overruns. Future work will include collecting
more data and characteristics from projects that will allow us to use
multi-factor methods to analyze the data to provide more meaningful
insights around oil and gas project performance.
Table 7
Cost performance in terms of project type.
Project type Num. of projects Average cost overrun SD Min Max
Brownfield Projects 88 16.7% 30.5% À32.0% 152.0%
Greenfield Projects 118 18.1% 28.8% À28.0% 169.0%
-30
-20
-10
0
10
20
30
40
50
60
70
80
2002 2004 2006 2008 2010 2012
Average Cost Overrun (%) Mean -SD (%) Mean + SD (%)
Fig. 2. Cost performance over years.
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e20 19
References
AACE, 2016. Cost Estimate Classification System-As Applied in Engineering, Procurement,
and Construction for the Process Industries. AACE, WV.
Ahsan, K., Gunawan, I., 2010. Analysis of cost and schedule performance of international
development projects. Int. J. Proj. Manag. 28 (1), 68e78.
Bertisen, J., Davis, G., 2008. Bias and error in mine project capital cost estimation.
Eng. Econ. 53 (2), 118e139.
Bordat, C., McCullouch, B., Labi, S., Sinha, K., 2004. An Analysis of Cost and Time
Overrun of INDOT Projects. Purdue University, IN.
Cantarelli, C., Molin, E., Wee, B., Flyvbjerg, B., 2012. Characteristics of cost overruns
for Dutch transport infrastructure projects and the importance of the decision
to build and project phases. Transp. Policy 22, 49e56.
EY, 2014. Spotlight on Oil and Gas Megaprojects. Ernst & Yong, UK.
EY, 2015. Joint Ventures for Oil and Gas Megaprojects. Ernst & Young, UK.
Flyvbjerg, B., 2014. What you should know about megaprojects, and why: an
overview. Proj. Manag. J. 45 (2), 6e19.
Flyvbjerg, B., Skamris, M., Buhl, S., 2004. What causes cost overrun in transport
infrastructure projects? Transp. Rev. 24 (1), 5e18.
Frimpong, Y., Oluwoye, J., Crawford, L., 2003. Causes of delays and cost overruns in
construction of groundwater projects in developing countries; Ghana as a case
study. Int. J. Proj. Manag. 21 (5), 321e326.
Gleave, S., 2009. Ex-Post Evaluation of Cohesion Policy Programmes 2000-2006.
European Commission Directorate General for Regional Policy, Loudon.
Group, W. B., 2014. Doing Business 2015: Going beyond Efficiency. World Bank
Group, Washington, D.C.
Hall, N., Delille, Sarah, 2011. Cost Estimation Challenges and Uncertainties Confronting
Oil and Gas Companies, 55th Annual Meeting of the AACE International,
WV.
He, J., Rui, Z., Ling, K., 2016. A new method to determine Biot's coefficients of Bakken
samples. J. Nat. Gas Sci. Eng. 35, 259e264.
Hwang, B., Thomas, S., Caldas, C., 2010. Performance metric development for
pharmaceutical construction projects. Int. J. Proj. Manag. 28 (3), 265e274.
IEA, 2011. IEA World Energy Outlook. IEA, Paris.
Jacboy, C., 2001. Report on Supplemental Agreement Reasons: AASHTO-FHWA
Project Cost Overrun Study. Washington, D.C.. Federal Highway Administration,
US Department of Transportation, Washington, D.C.
Jahren, C., Ashe, A., 1990. Predictors of cost-overrun rates. J. Constr. Eng. Manag. 116
(3), 548e551.
Kaliba, C., Muya, Mundia, Mumba, K., 2009. Cost escalation and schedule delays in
road construction projects in Zambia. Int. J. Proj. Manag. 27 (5), 522e531.
Karev, A., 2014. Oil and Gas Reality Check 2014. Deloitte Global Services Limited,
London.
Kolltveit, B.J., Grønhaug, K., 2004. The importance of the early phase: the case of
construction and building projects. Int. J. Proj. Manag. 22 (7), 545e551.
Le-Hoai, L., Lee, Y., 2008. Delay and cost overruns in Vietnam large construction
projects: a comparison with other selected countries. KSCE J. Civ. Eng. 12 (6),
367e377.
Merrow, E., 2012. Oil and gas industry megaprojects: our recent track record. Oil
Gas Facil. 1 (2), 38e42.
Merrow, E., Chapel, S., Worthing, C., 1979. A Review of Cost Estimation in New
Technologies Implications for Energy Process Plans. The Rand Corporation, CA.
Nawaz, T., Shareef, N., Ikram, A., 2013. Cost performance in construction industry of
Pakistan. Ind. Eng. Lett. 3 (2), 19e33.
Olaniran, O.J., Love, P., Edward, D., Olatunji, A., Mattews, J., 2015. Cost overruns in
hydrocarbon megaprojects: a critical review and implications for research. Proj.
Manag. J. 46 (6), 126e138.
Pergler, M., Rasmussen, A., 2014. Making Better Decisions about the Risks of Capital
Projects. Mckinsey & Company, UK.
Pohl, G., Dubravko, M., 1992. Project evaluation and uncertainty in practice a statistical
analysis of rate of return divergences of 1015 world bank projects. World
Bank Econ. Rev. 6, 255e257.
PWC, 2014. Capital Projects and Infrastructure in Africa. PricewaterhouseCopper,
South Africa.
Ren, Z., Wu, X., Liu, D., Rui, R., Guo, Wei, Chen, Z., 2016. Semi-analytical model of the
transient pressure behavior of complex fracture networks in tight oil reservoirs.
J. Nat. Gas Sci. Eng. 35 (a), 497e508.
Rui, Z., 2012c. Study confirms feasibility of in-state Alaska gas pipeline. Oil Gas J. 110
(10), 112e120.
Rui, Z., 2012d. Study finds in-state gas pipeline feasibility. Oil Gas J. 110 (11),
124e130.
Rui, Z., Wang, X., 2013. A comprehensive analysis of natural gas distribution pipeline
incidents. Int. J. Oil, Gas Coal Technol. 6 (5), 528e548.
Rui, Z., Metz, P.A., Reynolds, D., Chen, G., Zhou, X., 2011a. Historical pipeline construction
cost analysis. Int. J. Oil, Gas Coal Technol. 4 (3), 244e263.
Rui, Z., Metz, P.A., Reynolds, D., Chen, G., Zhou, X., 2011b. Regression models estimate
pipeline construction costs. Oil Gas J. 109 (14), 120e127.
Rui, Z., Metz, P.A., Chen, G., 2012a. An analysis of inaccuracy in pipeline construction
cost estimation. Int. J. Oil, Gas Coal Technol. 5 (1), 29e46.
Rui, Z., Metz, P.A., Chen, G., Zhou, X., Wang, X., 2012b. Regressions allow development
of compressor cost estimation models. Oil Gas J. 110 (1a), 110e115.
Rui, Z., Metz, P.A., Chen, G., Zhou, X., Wang, X., 2013. Inaccuracy in pipelinecompressor-station
construction cost estimation. Oil Gas Facil. 2 (5), 71e79.
Sambasivan, M., Soon, Y., 2007. Causes and effects of delays in Malaysian construction
industry. Int. J. Proj. Manag. 25 (5), 517e526.
Shehu, Z., Endut, I., Akintoye, A., Holt, G., 2014. Cost overrun in the Malaysian
construction industry projects: a deeper insight. Int. J. Proj. Manag. 32 (8),
1471e1480.
Skitmore, M., Stardling, S., Tuohy, A., Harry, M., 1990. Accuracy of Construction Price
Forecasts. University of Salford, UK.
Sun, J., Gamboa, E., Schechter, D., Rui, Z., 2016. An intergrated workflow for characterization
and simulation of complex fracture networks utilizing microseismic
and horizontal core data. J. Nat. Gas Sci. Eng. 34, 1347e1360.
Takim, R., 2005. A Framework for Successful Construction Project Performance.
Ph.D thesis. Glasgow Caledonian University, UK.
Tordo, S., 2011. National Oil Companies and Value Creation. World Bank Group,
Washington, D.C.
Tordo, S., Warner, M., Manzano, O., Anouti, Y., 2013. Local Content Policies in the Oil
and Gas Sector. World Bank Group, Washington, D.C.
Wainberg, M., Foss, M., 2007. Commercial Framework for National Oil Companies.
University of Texas at Austin, Texas.
Wolf, C., 2008. Does Ownership Matter? the Performance and Efficiency of State Oil
Vs. Private Oil. University of Cambridge, UK.
Zhao, J., 2000. Diffusion, Cost and Learning in the Development of International Gas
Transmission Lines. International Institute for Applied Systems Analysis,
Austria.
Zhao, Y., Rui, Z., 2014. Pipeline compressor station construction cost analysis. International
Journal of Oil. Gas Coal Technol. 8 (1), 41e61.
Zhao, X., Rui, Z., Liao, X., Zhang, R., 2015a. The qualitative and quantitative fracture
evaluation methodology in shale gas reservoir. J. Nat. Gas Sci. Eng. 27, 486e495.
Zhao, X., Rui, Z., Liao, Xinwei, Zhang, R., 2015b. A simulation method for modified
isochronal well testing to determine shale gas well productivity. J. Nat. Gas Sci.
Eng. 27, 479e485.
Zhao, X., Rui, Z., Liao, X., 2016. Case studies on the CO2 storage and EOR in heterogeneous,
highly water-saturated, and extra-low permeability Chinese reservoirs.
J. Nat. Gas Sci. Eng. 29, 275e283.
Z. Rui et al. / Journal of Natural Gas Science and Engineering 38 (2017) 12e2020