Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Impact of Early Stage Project-Management Performance on the
Cost Performance of Capital Projects
Hong Long Chen*1
Few studies explore the relationships between early stage project-management performance
and the cost performance of capital projects at completion. This longitudinal study of 102
capital projects reveals that the relationships between early project-management performance
and project-cost performance are indeed significant. Our research findings suggest that when
Communication, Team, Scope, and Innovation variables perform well in the initiation and
planning phases of a project, the project is more likely to meet its goal of cost performance in
the closing phase. Subsequent multivariate robust regression analyses with a maximum Rsquared improvement procedure demonstrate that Team, Scope, Communication, and
Innovation performance in the project initiation and planning phases explains 41.15% of the
variation in the cost performance of capital projects in the closing phase. The results also
show that Team provides the highest explanation of the variation in the project-cost data.
Ultimately, our findings suggest that project-management performance in the initiation and
planning phases possesses a critical impact on a project's cost at completion, making the
initiation and planning phases the fountainhead of project cost performance.
JEL Codes: L74, M11, and N65
1. Introduction
A central task in the project-management study is to identify the critical determinants of
project-management performance. Of course, extensive research in the field of project
management examines and identifies a broad variety of measures to delineate projectmanagement performance and the input characteristics that affect project performance
(e.g., Chen, 2014; El-Sayegh, 2008; Hoegl and Parboteeah, 2007; Oke and Idiagbon-Oke,
2010; Scott-Young and Samson, 2008).
One recent finding, for example, is that team collaboration, a product of social
construction fostered by managerial support, is important factor in project performance
(Calamel et al., 2012). Another is that management’s perception and satisfaction, and
project characteristics significantly affect project performance (Lerch and Spieth, 2013).
Although project management is well-researched and extensively reviewed, relatively
few studies investigate how project-management performance affects the cost
performance of projects. As a result, there appears to be a lack of research using
longitudinal experiments to examine how early stage project-management performance
influences the cost performance of captial projects.
The objective of this study, therefore, is to conduct a longitudinal experiment that
develops an early project cost-performance model based on the relationships between
project-management performance in the initiation and planning phases and cost
performance in the closing phase. The rest of the paper is organized as follows. Section 2
reviews related studies, Section 3 briefly describes our research methodology, and
Section 4 depicts the sample collection, the test hypothesis and model development.
* Professor, Department of Business and Management, National University of Tainan, Tainan 700, Taiwan.
Email : along314@mail.nutn.edu.tw
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Section 5 discusses the implications of the research results. Section 6 presents the
research summary and conclusions.
2. Literature Review
Whilst project management is the application of knowledge, skills and techniques to
execute projects efficiently, successful project management that effectively achieves the
project on time, on budget, and within specifications to the project stakeholders’
stratification has been treated as key to developing a sustainable competitive edge for
organizations (Project Management Institute, 2013). Not surprisingly, qquestions with
regard to how to manage critical project issues scientifically, and thus, enhance project
performance take centre stage in the research (Chen, 2014).
Numerous researchers and practitioners (e.g., Calamel et al., 2012; Haas, 2006;
Kazanjian et al., 2000; Blankevoort, 1984; Brown et al., 1990; Scott-Young and Samson,
2008; Schwab and Anne, 2008; Shepherd et al., 2011) have performed extensive studies
to examine and identify the key determinants of project performance. For example, Keller
(1992) assesses 66 project groups in three industrial R&D organizations based on factor
analysis and regression analysis. He concludes that transformational leadership, the sum
of charismatic leadership and intellectual stimulation, considerably influences project
performance.
Keller (1994) further tests the hypothesis and concludes that a fit between the task
technology’s characteristics and information-process needs forecasts project performance
based on the data from 98 project groups in four industrial R&D organizations.
Subsequent work by Hoegl and Gemuenden (2001) uses regression analysis to examine
the impact of teamwork quality on project performance. Based on data from 145 projects
in four software development companies, they find that teamwork quality is significantly
associated with project performance posing high task innovativeness.
Using the data from 507 software project managers, Wallace et al. (2004) examine
the impact of social subsystem risk, technical subsystem risk, and project-management
risk on project performance using the structural-equations modeling technique. Their
results show that social-subsystem risk significantly affects technical-subsystem risk,
which, in turn, influences project-management risk, and ultimately, project financial
performance.
Haas (2006) examines knowledge gathering, team capabilities, and project
performance in work environments using logistic regression. Based on data from
independent quality ratings of 96 projects and survey data from 485 project-team
members collected during a multimethod field study, he concludes that slack time,
organizational experience, and decision-making autonomy moderate the relationship
between knowledge gathering and project performance.
Using data from 56 capital projects in 15 process-industry companies, Scott-Young
and Samson (2008) examine the impact of organizational context, team design, team
leadership, and team process factors on project performance using factor analysis and
regression analysis. Their results show that these factors are significant determinants of
project outcomes. Anand et al. (2010) analyze 98 projects in five companies using
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
hierarchical regression. They show that the inclusion of softer, people-oriented practices
for capturing tacit knowledge explains a significant amount of variance in project outcomes.
Calamel et al. (2012) examine two collaborative R&D projects in a large globalinnovation cluster in France using a longitudinal design based on in-depth case research.
They conclude that team collaboration—a product of social construction fostered by
managerial support—is an important factor in project performance. Recently, Chen (2014)
analyzes 121 capital projects using hierarchical robust regression analyses. He shows that
that the relationships among project innovation stimulants, innovation capacity, and
project performance are indeed significant.
Despite the panoply of studies that examine and identify the key determinants of
project performance, most studies (e.g., Haas, 2006; Ling et al., 2009; Schwab and Anne,
2008) focus on describing project performance and the input characteristics that affect the
performance. Although several studies (e.g., Cheng and Roy, 2011; Chou, 2011; Maravas
and Pantouvakis, 2012) provide reliable estimates of project final costs and fixed capital
requirements, few investigate how early project-management performance affects the cost
performance of projects at completion. Consequently, there appears to be a lack of
research using longitudinal experiments to examine how the project-management
performance in the initiating and plainning phases influences the cost performance of
captial projects in the closing phase.
3. The Research Question and Methodology
The preceding section critiques existing studies of project performance. Now the question
is: How does early project-management performance affect the cost performance of
capital projects?
The methodology to answer this research question is threefold. First, subsequent to
developing a test hypothesis that examines the relationships between projectmanagement performance in the initiation and planning phases and the cost performance
in the closing phase, we employ the Anderson-Darling (AD) test to verify normality,
followed by the respective Pearson’s correlation and Spearman’s correlation tests when
the data is normally and abnormally distributed.
Second, based on the test results of the hypothesis, we use robust regression
analysis with a maximum R-square improvement for model development. Use of robust
regression analysis not only dampens the influence of outlying observations but also
ensures that the forecasts and estimation of the model are unbiased when the normality of
the residuals is violated (Neter et al., 1996; Salama, 2005). Third, we employ the White
test (White, 1980) to examine heteroskedasticity, and hence, decide our optimal model.
4. Research Results
3.1 The Data
Prior to the data collection, a panel of experts from the National Association of General
Contractors critiqued the questionnaire for structure, readability, clarity, and completeness
of the survey instrument. Based on the feedback from these experts, the survey
instrument was then modified to strengthen its validity.
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
The final version of the survey questionnaire comprises two sections. The first section,
composed of open-ended questions, gathers detailed background information such as
project type, project contract price, project budget, contract price for project changes, and
actual project cost. Section two consists of multiple-choice questions in which respondents
indicate on a 5-point Likert scale the extent to which certain project variables likely affect
project performance.
Data collection occurred in two stages and lasted two years. In the first stage,
immediately after the end of a project's initiation and planning stages, participants respond
to the portion of the questionnaire that excludes questions regarding project actual cost,
contract price for project changes, and actual cost for project changes. In the second
stage, right after the close of the capital project, participants respond to the questions
excluded in stage one.
Of the 574 members of the National Association of General Contractors that we
randomly selected and invited to participate in this research, 102 companies participated.
Each of the 102 companies in the sample had assigned a project manager who had just
completed the initiation and planning of a capital project scheduled to finish within the next
two years. The 102 capital projects fall into two major categories: buildings (71 projects)
and industrial facilities (31 projects).
3.2 Measures and analysis
Measures of project-management performance variables, including Communication,
Team, Scope, and Innovation are based on a detailed examination of literature in the
project-management and organization-theory fields as well as consultation with several
experienced researchers and experts. The work identifies measures that are similar to the
project-performance measures most contracting organizations use internally for assessing
the performance of the overall project life cycle.
Communication (Cronbach's Alpha = 0.954) is measured according to a six-item scale
based on the representative studies, including Ling et al. (2009), and Oke and IdiagbonOke (2010). The six items include “C1: The project team identifies all the key stakeholders
of the project,” “C2: The project team meets the information needs of the stakeholders,”
“C3: The project team meets the communications needs of the stakeholders,” “C4:
Technology use in information sharing is high,” “C5: Communication with the customer is
effective,” and “C6: Communication within project team members is effective.”
Team (Cronbach's Alpha = 0.975) is measured according to a 12-item scale based on
the representative studies, including Anand et al. (2010), Ling et al. (2009), Scott-Young
and Samson (2008), and Tabassi and Bakar (2009). Sample items are “T1: Top
management support for the project team is high,” “T2: Enthusiasm about project success
is high,” “T3: Each team member's project role, responsibilities, and rights are clearly
defined,” “T4: Group participation in decision-making is high,” “T5: Interpersonal
relationships among team members is good,” “T6: Project teams' job skills and expertise
are good.” , “T7: Degree of cohesiveness of the project team is high,” “T8: Degree of
motivation of the project team is high,” and “T9: Degree of cooperation of the project team
is high”
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Scope (Cronbach's Alpha = 0.961) is measured according to a four-item scale based
on the representative studies, including Ling et al. (2009), Kwak and Ibbs (2002), and
Roman (1964). Items are “S1: Quality of contract documents including project definitions,
legal terms, specifications, design instructions, and implementation processes is good,”
“S2: Project owner defines project scope well,” “S3: Project owner has verified extent of
project scope well”, and “S4: Work breakdown structure (WBS) of the project is well
defined and manageable.”
Innovation (Cronbach's Alpha = 0.960) is measured according to a 10-item scale
based on the representative studies, including Keegan and Turner (2002), Prajogo and
Ahmed (2006), and Tranfield et al. (2003), and Urban and von Hippel (1988). Sample
items include “I1: Management support for innovation is high,” “I2: Project team applies
latest technology to the project,” “I3: Project team devotes much time and resources
toward generating innovative ideas,” “I4: Team members have diverse skills,” “I5:
Cognitive conflict among project team members is high,” “ I6: The project manager adopts
a bottom-up problem-solving style that incorporates all team members”, “I7: There are
widespread communications within/across the project,” and “I8: Extent of elaborating
information processing and coordination mechanisms within/across the project is high.”
For comparison purpose with the Communication, Team, Scope, and Innovation of
data collected using a 5-point Likert scale prior to the project execution phase, percentile
ranks categorize project cost performance in the closing phase. The data from the closing
phase measures capital-project cost performance on a 5-point scale using the computed
values of equation 1 from the 102 sample projects, where equation 1 is:
Cost = Revised Estimated Cost/Actual Cost
(1)
where the revised estimated cost includes the additional estimated cost due to changes in
project scope. Table 1 reports the percentile ranks of cost performance in the projectclosing phase from the 102 sample projects.
Table 1. Percentile ranks of project-cost performance
Percentile
Respective 5-Point
Number of
Average Cost
Likert Scale
Observations
Performance (%)
1 to <=20
1
19
82.5
21 to <=40
2
19
92.6
41 to <=60
3
23
102.7
61 to <=80
4
20
120.3
81 to <=100
5
21
185.1
Since our research is built on the proposition that early stage project-management
performance affects the cost performance of the project at completion, we test the
following hypothesis:
Hypothesis 1: Communication, Team, Scope, and Innovation performance in the projectinitiation and planning phases insignificantly affect the cost performance of capital projects
in the closing phase.
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Table 2 presents means, standard deviations, Anderson-Darling (AD) statistics, and
Spearman’s correlation matrix of the sample data of the variables. Correlation is
significant when the probability value is smaller than 0.05. The Anderson-Darling (AD)
values of the variables Cost Performance, Communication, Team, Scope, and Innovation
are 3.61, 0.81, 0.82, 1.54, and 0.76, respectively, all with associated p-values of <0.05,
indicating a significant abnormal distribution that justifies use of Spearman’s correlation
tests.
Table 2. Descriptive statistics, Anderson-Darling statistics, and Spearman’s
correlation matrix for all variables
Variables
AD
Means S.D. 1
2
3
4
5
Statistics
1. Cost Performance
3.61**
2.94
1.42 1.00
2. Communication
0.81*
3.57
0.83 0.60** 1.00
3. Team
0.82*
3.82
0.76 0.65** 0.87** 1.00
**
4. Scope
1.54
3.66
0.88 0.63** 0.84** 0.86** 1.00
5. Innovation
0.76*
3.48
0.76 0.53** 0.80** 0.80** 0.78** 1.00
*P < 0.05 and **P < 0.01.
As seen in the table, the Spearman correlation coefficients of Communication, Team,
Scope, and Innovation are 0.60, 0.65, 0.63, and 0.53 with the associated p-values of
<0.01. Therefore, we reject the null hypothesis. This suggests that Communication, Team,
Scope, and Innovation significantly influence the cost performance of projects. That is, the
better a capital project’s Communication performs in the initiation and planning phases,
the better that project’s cost performance would be in the closing phase, and likewise,
Team, Scope, and Innovation.
3.3 Model Development
The hypothesis reveals significant correlations between Cost Performance in the project
closing phase and each of Communication, Team, Scope, and Innovation in the projectinitiation and planning phases. Based on this test result, we further model the effects of
Communication, Team, Scope, and Innovation on the cost performance of capital projects.
Table 3 summarizes the robust regression results using a maximum R-square
improvement procedure for modelling the effects of Communication, Team, Scope, and
Innovation. As the table shows, the optimal Cost Performance model at step 1 (Model 1) is
the one with the Team variable, where 40.07% of the variation in the project-cost
performance data is explained.
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Table 3: Effects of Communication, Team, Scope, and Innovation on Cost
Performance and respective White tests created with robust regression analysis
using a maximum R-squared improvement
Dependent variable: Cost Performance
Variable
Model 1
Model 2
Model 3
Model 4
Coef
ChiCoef. ChiCoef. ChiCoef. Chi.
Square
Square
Square
Square
Step 1
Intercept
-1.861
-1.823 9.42** -1.805
9.71**
9.13** -1.776
8.32**
**
**
*
Team
1.256 67.04
0.896 8.23
0.835
0.841
4.90
4.81*
Step 2
Scope
0.363
1.82 0.323
1.28 0.329
Step 3
Communication
0.101
0.10 0.120
Step 4
Innovation
-0.040
R2 (%)
Change in R2 (%)
The White test
*P < 0.05 and **P < 0.01.
40.07
–
41.00
0.93
41.10
0.10
41.15
0.05
6.31*
6.57
8.51
11.08
At step 2, the optimal Cost Performance model (Model 2) is composed of Team and
Scope, capable of explaining 41.00% of the variation in the project-cost performance data,
which is 0.93% more than that of Model 1. The optimal model at steps 3 (Model 3) is
composed of Team, Scope, and Communication. Model 3 explains 41.10% of the variation
in the project-cost performance data, and it is 0.10% more than that of Model 2.
At step 4, the optimal Cost Performance model (Model 4) consists of Team, Scope,
Communication, and Innovation that explains 41.15% of the variation in the project-cost
performance data, which is 0.05% more than that of Model 3. The White test of Model 4
(in the bottom portion of Table 3) is 11.08, and the associated p-value is larger than 0.05,
suggesting the acceptance of the null hypothesis of no heteroskedasticity in the residuals
at the 0.05 level for Model 4. Consequently, we choose Model 4 as the optimal Cost
Performance model.
5. Discussions
Our findings regarding the importance of Team, Scope, Communication, and Innovation
are consistent with prior studies of the overall project life cycle (e.g., Ling et al., 2009; Oke
and Idiagbon-Oke, 2010; Scott-Young and Samson, 2008). The present research extends
the state of knowledge with regard to the effects of early stage project-management
performance on the cost performance of capital projects. This extension is a direct
contribution to the literature of the project-management field.
Specifically, our findings demonstrate that the Team, Scope, Communication, and
Innovation performance of project management in the initiation and planning phases
explains 41.15% of the variation in the cost performance of capital projects in the closing
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
phase. An important managerial implication of the findings is that project-management
performance in the initiation and planning phases possesses a critical impact on a
project's cost at completion, making the initiation and planning phases the fountainhead of
project cost performance.
6. Summary and Conclusions
The hypothesis’s test result reveals that early stage project-management performance
significantly affects the cost performance of capital projects at completion. Specifically, the
research findings suggest that when Communication, Team, Scope, and Innovation
variables perform well in the initiation and planning phases of a project, the project is more
likely to meet its goal of cost performance. Subsequent Cost Performance models using
robust regression analysis with a maximum R-squared improvement show that the
combined Communication, Team, Scope, and Innovation variables provide an optimal cost
performance model. The results also show that Team provides the highest explanation of
the variation in the project-cost data.
Future research should investigate the relationships between performance
improvements in Communication, Team, Scope, and Innovation and project costs during
the project-delivery process. This allows to predict not only the cost of a project at
completion, but the respective improved cost performance of the project. Such an
extension would be beneficial in decision-making and project cost control.
References
Anand, G Ward, PT, and Tatikonda, MV 2010, Role of explicit and tacit knowledge in Six
Sigma projects: An empirical examination of differential project success, Journal of
Operations Management, Vol. 28, No. 4, pp. 303–315.
Blankevoort, PJ 1984, Effects of communication and organization, International Journal of
Project Management, Vol. 2, No. 3, pp. 138–147.
Brown, KA, Klastorin, TD, and Valluzzi, J 1990, Project performance and the liability of
group harmony, IEEE Transactions on Engineering Management, Vol. 37, No. 2, pp.
117–125.
Calamel, L, Defélixa, C, Picqd, T, and Retour, D 2012, Inter-organisational projects in
French innovation clusters: The construction of collaboration, International Journal of
Project Management, Vol. 30, No. 1, pp. 48–59.
Chen, HL 2014, Innovation stimulants, innovation capacity, and the performance of capital
projects, Journal of Business Economics and Management, Vol. 15, No. 2, pp. 212–
231.
Cheng, MY, and Roy, AFV 2011, Evolutionary fuzzy decision model for cash flow
prediction using time-dependent support vector machines, International Journal of
Project Management, Vol. 29, No. 1, pp. 56–65.
Chou, JS 2011, Cost simulation in an item-based project involving construction
engineering and management, International Journal of Project Management, Vol. 20,
No. 6, pp. 706–717.
El-Sayegh, SM 2008, Risk assessment and allocation in the UAE construction industry,
International Journal of Project Management, Vol. 26, No. 4, pp. 431–438.
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Haas, MR 2006, Knowledge gathering, team capabilities, and project performance in
challenging work environments, Management Science, Vol. 52, No. 8, pp. 1170–
1184.
Hoegl, M, and Gemuenden, HG 2001, Teamwork quality and the success of innovative
projects: A theoretical concept and empirical evidence, Organization Science, Vol. 12,
No. 4, pp. 435–449.
Hoegl, M, and Parboteeah, KP 2007, Creativity in innovative projects: how teamwork
matters, Journal of Engineering and Technology Management, Vol. 24, No. (1-2), pp.
148–166.
Kazanjian, RK, Drazin, R, and Glynn, MA 2000, Creativity and technological learning: The
roles of organization architecture and crisis in large-scale, Journal of Engineering and
Technology Management, Vol. 17, No. (3-4), pp. 273–298.
Keller, RT 1992, Transformational leadership and the performance of research and
development project groups, Journal of Management, Vol. 18, No. 3, pp. 489–501.
Keller, RT 1994, Technology-information processing fit and the performance of R&D
project groups: A test of contingency theory,” Academy of Management Journal, Vol.
37, No. 1, pp. 167–179.
Keegan, A, and Turner, JR 2002, The Management of innovation in project-based firms,
Journal of Long Range Planning, Vol. 35, No. 4, pp. 367–388.
Kwak, YH, and Ibbs, CW 2002, Project management process maturity (PM)2 model,
Journal of Management in Engineering, Vol. 18, No. 3, pp. 150–155.
Lerch, M, and Spieth, P 2013, Innovation project portfolio management: A qualitative
analysis, IEEE Transactions on Engineering Management, Vol. 60, No. 1, pp. 18 - 29.
Ling, FYY, Low, SP, Wang, SQ, and Lim, HH 2009, Key project management practices
affecting Singaporean firms’ project performance in China, International Journal of
Project Management, Vol. 27, No. 1, pp. 59–71.
Maravas, A, and Pantouvakis, JP 2012, Project cash flow analysis in the presence of
uncertainty in activity duration and cost, International Journal of Project Management,
Vol. 30, No. 3, pp. 374-384.
Neter, J, Kutner, MH, Nachtsheim, CJ, and Wasserman, W 1996, Applied Linear
Statistical Models, McGraw-Hill, Boston.
Oke, A, and Idiagbon-Oke, M 2010, Communication channels, innovation tasks and NPD
project outcomes in innovation-driven horizontal networks, Journal of Operations
Management, Vol. 28, No. 5, pp. 442–453.
Prajogo, DI, and Ahmed, PK 2006, Relationships between innovation stimulus, innovation
capacity, and innovation performance, R&D Management, Vol. 36, No. 5, pp. 499–
515.
Project Management Institute. 2013, A guide to the project management body of
knowledge (PMBOK Guide). 5th ed., Project Management Institute, Newtown Square,
PA.
Roman, D 1964, Project management recognizes R&D performance, Academy of
Management Journal, Vol. 7, No. 1, pp. 7–20.
Salama, A 2005, A note on the impact of environmental performance on financial
performance, Structural Change and Economic Dynamics, Vol. 16, No. 3, pp. 413–
421.
Schwab, A, and Anne, SM 2008, Earning in hybrid-project systems: The effects of project
performance on repeated collaboration, Academy of Management Journal, Vol. 51,
No. 6, 1117–1149.
Proceedings of 31st International Business Research Conference
27 - 29 July 2015, Ryerson University, Toronto, Canada
ISBN: 978-1-922069-80-1
Scott-Young, C, and Samson, D 2008, Project success and project team management:
evidence from capital projects in the process industries, Journal of Operations
Management, Vol. 26, No. 6, pp. 749–766.
Shepherd, DA, Patzelt, H, and Wolfe, M 2011, Moving forward from project failure:
Negative emotions, affective commitment, and learning from the experience,
Academy of Management Journal, Vol. 54, No. 6, pp. 1229–1259.
Tabassi, AA, and Bakar, AHA 2009, Training, motivation, and performance: the case of
human resource management in construction projects in Mashhad, Iran, International
Journal of Project Management, Vol. 27, No. 5, pp 471–480.
Urban, GL, and von Hippel, E 1988, Lead user analyses for the development of new
industrial products,” Management Science, Vol. 34, No. 5, pp. 569–582.
Tranfield, D, Young, M, Partington, D, Bessant, J, and Sapsed, J 2003, Knowledge
management routines for innovation projects: Developing a hierarchical process
model,” International Journal of Innovation Management, Vol. 7, No. 1, pp. 27–49.
Wallace, L, Keil, M, and Rai, A 2004, How software project risk affects project
performance: An investigation of the dimensions of risk and an exploratory model,
Decision Sciences, Vol. 35, No. 2, pp. 289–321.
White, H 1980, A heteroskedasticity-consistent covariance matrix estimator and a direct
test for heteroskedasticity, Econometrica, Vol. 48, No. 4, pp. 817-838.