sustainability

Article
Loss Assessment for Sustainable Industrial
Infrastructure: Focusing on Bridge Construction and
Financial Losses
Ji-Myong Kim, Taehui Kim and Sungjin Ahn *
Department of Architectural Engineering, Mokpo National University, Mokpo 58554, Korea;
jimy@mokpo.ac.kr (J.-M.K.); thkim@mokpo.ac.kr (T.K.)
* Correspondence: sunahn@mokpo.ac.kr; Tel.: +82-61-450-2456




Received: 22 April 2020; Accepted: 28 June 2020; Published: 1 July 2020


Abstract: Bridges are important infrastructures for urban growth and the economic development of a
country, because bridges allow a large volume of logistics and transportation by connecting rivers,
canyons, islands and lands. As such, massive resources including financial, material and human
resources are invested for bridge construction and management. However, although the latest bridge
construction is undergoing rapid development of new technologies and designs, the management
and prevention of risks still tend to rely on qualitative practices, which, as a result, calls for more
quantified and systematic measurement and, thus, more sustainable management of potential risks.
As part of efforts in managing risks to achieve quantitative risk management, this study aimed to
predict losses of financial resources by identifying statistically significant risk factors based on the
past record of insurance claim payouts (compensation for a loss that occurred as a result of a material
damage in bridge construction projects) from a major insurance company in Korea, and conducted a
multiple regression analysis to identify the loss indicators and to develop a loss estimation model.
The statistical analysis confirmed that superstructure types, superstructure construction methods,
and construction duration are the three significant risk factors that affects financial losses of bridge
construction projects among the seven variables adopted as independent variables, which included
the superstructure type, maximum span length, superstructure construction method, foundation
type, floods, typhoons, and construction duration. Such findings, and the consequentially developed
risk prediction model of this study, will contribute to sustainable construction management through
cost reduction by predicting and preventing the future financial loss factors of bridge construction.

Keywords: quantitative loss assessment; bridge construction; sustainable risk management; insurance

1. Introduction
As the sizes of recent construction projects grow and the methods of construction become highly
diversified, a variety of uncertainties are on an increasing trend, accompanied by many new risk
factors. Thus, the need for a more systematic and reliable risk management is continuously being
emphasized in the entire process of construction projects. However, such construction risk management
techniques seem rather insufficient in the field, as they have not yet been established in actual practice
in trend [1]. For example, the risk assessment used in most construction projects tends to rely on
subjective and qualitative information based on the experiences and opinions of individual contractors
and construction managers [2,3], rather than on technical analysis and verification. Especially in bridge
construction, various accidents occur, which can result in significant human injuries and material
damages, because bridge construction is exposed to various types of risks that tend to be more extreme
and severe than those of other types of construction sites, due to environmental conditions such as

Sustainability 2020, 12, 5316; doi:10.3390/su12135316 www.mdpi.com/journal/sustainability

Sustainability 2020, 12, 5316 2 of 16

streams, rivers, and canyons, as well as work conditions such as high place work and heavy equipment
operations. In addition, many recently built bridges tend to be long-spanned, complicated structures,
as a result of a high degree of the development of new technologies and design strategies, which can
also cause extensive losses if accidents happen with the increased risk factors associated with the
complexity of the work processes. Yet, as briefly mentioned above, decisions relating to risk assessment
and management are often made based on insufficient and/or subjective information [2–4]. Therefore,
more research-based and empirical risk factors are required to be verified and used in analyzing and
predicting to prepare for the possible damages and accidents. As such, risk management measures
should identify various risk factors throughout the whole process of the construction in advance,
in order to estimate and analyze the extent of the potential damages and losses according to their causes,
based on as minimal information as reasonably possible [5–8]. With this background, the purpose
of this study is to provide the risk prediction model that can contribute to minimizing risks in more
systematical and evidence-based way. In other words, this study aims to first identify the statistically
significant risk factors in bridge construction projects inherent in the process from onset to completion
and to present a risk prediction model, while reflecting the actual record of damages which occurred in
bridge construction projects, both of which ultimately aim to achieve sustainable risk management.

2. Review of Literature

2.1. Need for Quantifying the Risks and Loss of Construction Projects
The objective of the risk analysis is to establish effective and sustainable strategies to proactively
prevent and manage possible risks, ultimately to better allocate and commit financial resources in
construction projects. The concept of risk is variously addressed in a wide range of fields and areas in
an effort to identify, measure and control the risks. Table 1 below lists the various Interpretations of
risk from different studies.

Table 1. Interpretations of risk.

Author Interpretations of Risk

Morgan and Henrion (1990) [9] Degree of exposure to damage or loss
UN DHA (1992) [10] Loss expected due to a disaster in a particular period and region
Stenchion (1997) [11] Probability of the occurrence of an undesirable event
Probability of losses consisted of three elements: disaster,
Crichton (1999) [12]
exposure, and vulnerability
A combination of probability of a potential event causing
Jones and Boer (2003) [13]
damage and the consequence

Analyzing risk provides a means to cope with many possible risks, as it helps identifying
appropriate and effective strategies to confront, avoid and reduce the risks. Nonetheless, risk analysis
has not been in the center of attention for long in the literature, compared to such topics as cost and
schedule management in construction projects. Furthermore, it has traditionally been qualitative
methodologies used for risk analysis in construction projects, rather than quantitative methodologies,
even though quantitative methodologies enable not only more factual and accurate risk analyses,
but also provide a more reliable and generalizable measurement for examining and analyzing the
risks. With quantitative risk analysis methodologies, it is possible to model and analyze nearly all
construction projects by quantifying the factors of risks, their actual occurrence and impact. For this
reason, recent studies in risk analysis have started to adapt qualitative methodologies, although much
of the previous research has been conducted qualitatively.
The construction industry inclines to complexity and specificity of the projects and is considered a
high-risk area, as it involves a large number of project participants, long period of project time and the
large cost scale. Such technical and strategic complexities, in turn, create a great deal of risks based on
Sustainability 2020, 12, 5316 3 of 16

uncertainty [14]. Yet, although many studies have been conducted on risk management, the topic of
risk assessment remains a controversial issue [15], in terms of its data and methodologies of analysis.
For example, conventional risk management data and methodologies remained in the qualitative
realm, which can be inadequate in precisely measuring, quantifying and enumerating the risks. On the
contrary, quantitative data and methodologies can provide more sophisticated, accurate and reliable
risk assessments and prediction, because it adopts more mathematical and statistical way to perceive
and cope with the risks.
Although the construction industry is considered a high-risk industry compared to other industries,
e.g., insurance and finance, the attitude toward and application of risk assessment are reported in
many studies to be rather inadequate [16]. For example, Baker et al. [17] inquired construction
companies to examine the effective risk assessment techniques the companies used. They found
that individual owners’ or experts’ personal experiences and engineering judgment were the most
frequently used assessment strategies. Wood and Ellis [2] also reported that personal experiences and
underlying judgement are the main risk assessment factors and that risk management is generally
conducted in rather simple and primitive ways, such as checklists and surveys. Similarly, noting that
qualitative methods were more frequently used than quantitative methods in studies on construction
risk assessment, Dikmen et al. [3] addressed the need for a quantified approach to organize subjective
assessment factors such as individual experts’ experiences, knowledge and intuitive decisions regarding
the identification of possible risk factors.
As can be seen, the literature regarding risk assessment has pointed out that the field has lacked
in quantitative approaches in assessing and analysis risks, although quantitative measures can provide
factual and numerical values and, in turn, a more systematic understanding for decision making [18].
Upon the findings that the sensitivity analysis was frequently used for risk analysis in construction
projects, Warszawski and Sacks [19] concluded that, because construction projects tended to be one-off
enterprises and the input information required is not sufficient, more advance and refined analysis
methods approaches were not generally adopted. Many studies [20–23] have frequently adopted risk
cost as impact measurement scale for risk analysis. Although many researchers have done risk analysis
on different project goals, they have used little of the actual risk cost due to limits of data acquisition.
For this reason, using risk cost for assessment scale seem to make much improvement in increasing the
validity and accuracy of risk analysis. Keshk et al. [24] examined various types of risks to consider in
quantitative and qualitative analysis for construction project management and indicated that, although
it is impossible to locate the one strategy to deal with all risks, quantification of types and sizes of risks
can greatly affect the strategies for responding to the risks.
As discussed, the literature has focused on various types of approaches to examine the factors for
risk management. Many studies pointed out that it is almost impossible to exclude qualitative data such
as opinions of experts and their personal experiences and called for the ways to quantitatively interpret
such qualitative data. On that account, in risk assessment, quantitative assessment is necessary for
advancing and improving the validity of risk factor analysis, while being accompanied by quantifying
the available and reasonable data.

2.2. Risk Management for Sustainable Construction

Traditional or typical enterprise risk management has tended to overlook the unexpected risks
and other risks caused from unpredictable events; in turn, such risks have failed to be encompassed
in assessing Enterprise Risk Management (ERM) strategies. For this reason, many business and
companies are now realizing that considering emerging or unforeseeable risks is essential for their
survival in the industry for the long run [25]. One way for more effective, long-term risk management
strategies is to adopt the concept of sustainability. The concept of sustainable construction is to provide
the stakeholders with long-term economic feasibility, high quality and efficiency, while reducing
negative environmental impacts and enhancing economic sustainability [26]. In addition, sustainable
risk management requires appropriate management of information and knowledge of risks and
Sustainability 2020, 12, 5316 4 of 16

uncertainties [27]. In particular, sustainable risk management in the field of construction needs to meet
the environmental challenges, to respond to social needs and to deliver economic improvement. For this
reason, sustainable risk management in construction focuses on the development of infrastructure to
achieve development that meets the economic, social and environmental needs of future generations,
as infrastructure is not just a means of generating corporate profits, but rather a type of public good
that plays an important role in influencing economic development and its impact on social needs [28].
These infrastructure development projects are mostly large, complex and capital intensive, therefore
innovative financial strategies are needed [29]. For instance, the Hong Kong government adopted
three weighted criteria (65% finance, 20% engineering, and 15% planning of operation and transport)
to evaluate contractors for the infrastructure project. The weights assigned to the financial criteria in
this assessment reflect the importance of sound financial planning for the success of infrastructure
projects [30].
Analysis and evaluation are essential elements of risk management, and the analysis consists of the
extent of the results and the best evaluation means that can be used to characterize the probability [31].
One of the many aspects of sustainable risk management that is distinct from traditional or typical
risk management is that, as mentioned earlier, it encompasses the management of risk factors and
potential uncertainties that are difficult to predict. This study addresses and contributes to the concept
of sustainable risk management in that, based on existing loss records, potential risk factors and
environmental disaster factors have been identified and were used to develop the loss estimation
model in bridge construction, one of infrastructure development. Specifically, it provides essential
and key data for plan and management that minimizes losses in advance to prepare for emerging and
unforeseeable risks and to optimize the overall life of bridges.

2.3. Risks in Bridge Construction

Various studies on risk factors specifically in bridge construction have also been conducted in
many regions and countries. Among them, the following Table 2 shows the key elements and analysis
methods for the bridge construction risk analysis in a few major relevant studies.

Table 2. Previous research on bridge construction projects.

Key elements in Bridge

Author Risk Data Collection Analysis Method
Construction Risk Analysis
(1) Safety, (2) Functionality, (3) Team of bridge Fuzzy Group Decision
Wang and Elhag (2007) [32]
Substantiality, (4) Environment experts’ rating Making (FGDM) approach
Databases: National Bridge
(1) Traffic loads, (2) Environmental Inventory (NBI)), U.S. Probabilistic evaluation of
Decò and. Frangopol (2011) [33]
attacks, (3) Scour, (4) Earthquakes Geological Survey (USGS), time-dependent
water data (USGS2010)
(1) Delayed payment on contract
and extras, (2) Shortage of labor,
Questions and interviews
Hashemi et al. (2011) [34] material, and equipment (3) Bootstrap Technique
with bridge experts
Construction permitting issues (4)
Poor relationship among parties
(1) Economic risk, (2) Contract and
law risk, (3) Building technology
Survey questionnaires from
Li et al. (2011) [35] risk, (4) Design risk, (5) Factor Analysis
project stakeholders
Environment risk, (6) Staff risk, (7)
Material and equipment risk
(1) Financial risks, (2) Contractual
risks, (3) Design risks, (4) Health
Interview and interviews
Choudhry et al. (2014) [36] and safety risks, (5) Management Monte Carlo simulation
with expert panels
risks, (6) Construction risks, (7)
External risks

As can be seen, various studies used different analysis methods to derive the factors of risk
management in bridge construction projects. Most of them aimed to interpret qualitative data in
quantitative ways. However, it would not be unreasonable to exclude the limitations of subjective and
Sustainability 2020, 12, 5316 5 of 16

intuitive aspects of the qualitative data itself. Therefore, it is necessary to statistically verify the risk
factors and develop a prediction model using a quantified and specific database, such as risk costs and
loss amount.

3. Risk Factors for Bridge Construction

3.1. Construction Insurance

Construction All Risks (CAR) insurance is insurance protection under which the contractor,
builder, and others involved in construction projects are collectively covered for damages to the
construction object, materials used, and equipment, injuries and deaths, or damages to the property of
a third party incurred by unexpected accidents during the construction projects. The coverage of the
CAR insurance policy is classified into a bridge, port, railway, road, building, tunnel, plant and water
resources-related construction, and covers the risks during the entire construction period from the
start to the transfer to the contractor after completion. As mentioned, this study adopted the record of
insurance claim payout. However, the scope of this study is limited to focus on the damages to the
materials incurred in the phase of construction only, excluding the maintenance phase.
Bridge Material loss means any damages for construction materials within the construction sites
occurred during construction and civil engineering works, such as temporary construction work,
main construction, and damage to the construction materials. Various accidents caused material loss
are as follows

 Damage caused by fire, lightning, explosion, and fire extinguishing operations;

 Damage caused by floods, floods, rainfall, snowfall, tidal waves, etc.;
 Damage caused by storms, earthquakes, ground subsidence, soil, rock collapse, etc.;
 Theft;
 Various electrical accidents such as short circuits, short circuits, etc.;
 Damage caused by a mechanical defect, worker’s carelessness, or construction defect.

For risk analysis, this study collected data from construction insurance coverage of the bridge
construction project from 1999 to 2017, among which 367 material damage insurance claim payout data
were selected and used. Among many possible elements for measuring risks in bridge construction
found in the data, this study perceived and articulated three types of information including accident
information (accident dates, accident location address and details), technical information (superstructure
types, foundation types, superstructure construction methods, and the maximum span length) and
project scale (project duration).

3.2. Risk Factors

3.2.1. Superstructure
When deciding the superstructure type in a bridge construction project, a comprehensive
consideration of economic constructability and safety, along with topographical and environmental
conditions, should be given at the planning phase. Previous studies have suggested that the
characteristics of the superstructure type of a bridge are distinguished by the conditions, and estimated
construction cost, and that the optimal superstructure can thus be selected according to the
conditions [37,38]. The Korean Public Procurement Service classified the grade of the bridge into Class
A (suspension bridge, cable stayed bridge, bridge 200 m or more in span), Grade B (bridge connecting
islands and land, bridge connecting island and island, arch bridge, truss bridge, bridge 100 m or
more but less than 200 m in span), Grade C (bridge 50 m or more in span, general bridge over
1000 m), and Class D (general bridge over 500 m) according to the difficulty of construction in the
contractor Pre-Qualification (PQ) criteria for bridge construction [39]. This distinction by size and
difficulty of constructing can be used in consideration for the main measure of risk analysis for bridge
Sustainability 2020, 12, 5316 6 of 16

construction. In this study as well, the superstructure types were classified in order of nominal scale of
arch bridges, pre-stressed concrete I-type (PSCI) and cable-stayed bridges by the distribution of the
average compensation.

3.2.2. Maximum Span Length

As with the superstructure types, the length of the bridge is an important criterion for the difficulty
of rating of the PQ review for constrictors, and for long-spanned bridges, the risk can be increased due
to the relative increase in construction period, cost and wind speed [40–42]. In this study, nominal
scales were classified on the basis of below 50, 100, and 500 m.

3.2.3. Foundation Types

In terms of deciding bridge foundations, the quantity and rate of river flow play a significant role.
If the quantity and rate of river flow increase rapidly in a short period of time due to torrential rain or
flooding, the sudden scour of the bridge foundation can happen and it may cause unexpected bridge
collapse accidents. As such, scour can be an important point in maintaining the life or stability of a
bridge [43]. The foundation of a bridge is a crucial structural element in bridge construction and holds
significant risk factors and uncertainties, which necessitate the hazard management [44]. Therefore,
the information on the bridge foundation types from the collected data was also utilized as a risk factor
in the analysis and, based on the comprehensive consideration of the construction difficulty and loss
compensation, the nominal scale classification for the analysis was established as follows: precast
concrete pile, cast in place, open caisson.

3.2.4. Superstructure Construction Method

Previous studies have evaluated the optimal structural safety design method by the bridge
construction method, focusing on the influence of the load according to the construction method
and the need for economic design methods [45–48]. In other words, the risk analysis based on the
characteristics of the construction method and its classification are necessary. Kim and Cho classified
the typical construction methods of the bridge superstructure as ILM (Incremental Launching Method),
FCM (Free Cantilever Method), FSM (Full Staging Method), and MSS (Movable Scaffolding System),
applying them to the rough construction cost estimation, which allowed identifying differences in
characteristics of superstructure construction method [49]. In this study, the criteria for classification
of superstructure construction methods was established on the basis of the nominal scale of ILM,
FCM, FSM, and MSS, based on the average amount of compensation sorted by the superstructure
construction method.

3.2.5. Total Duration

The duration of the construction projects was often used as a measure of risk analysis in previous
studies [50,51], and it could also be a significant measure of loss. For construction projects, it was
confirmed in studies that the longer the construction period, the less the amount of the loss. This study
as well adopted the factor of construction duration, i.e., number of months spent for the projects,
in the analysis.

3.2.6. Typhoon and Flood

The risks of natural disasters are calculated by the indicators of floods and typhoons, using the
reinsurer’s Natural Disaster Assessment Network map. It is designed to accurately evaluate the
characteristics of a particular place, with regard to natural disasters such as floods, storms and
earthquakes. This study selected and used the risks as to the intensity of typhoons and occurrence
of floods, because floods and typhoons are the typical natural disasters in Korea that affect bridge
construction. NATHAN (Natural Disaster Assessment Network) provided the differentiation in the
 Sustainability
Sustainability 2020,
2020, 12,12, x FOR
x FOR PEER
PEER REVIEW
REVIEW 7 of7 16
of 16
Sustainability 2020, 12, 5316 7 of 16
floods,
floods, because
because floods
floods andand typhoons
typhoons arearethethe typical
typical natural
natural disasters
disasters in in Korea
Korea that
that affect
affect bridge
bridge
construction. NATHAN (Natural Disaster Assessment Network) provided the differentiation in in
construction. NATHAN (Natural Disaster Assessment Network) provided the differentiation thethe
regional
regional
regional zones
zones
zones which
which
which then
then
then represented
represented
represented the
thetheintensity of
intensity
intensity typhoons
of of and
typhoons
typhoons andoccurrence
and of
occurrence
occurrence floods
of of using
floods
floods location
using
using
information
location
location (address)
information
information from insurance
(address)
(address) from claim payout
frominsurance
insurance claim
claim data. In other
payout
payout words,
data.
data. the words,
In Inother
otherflood occurrence
words, thetheflood and
flood
typhoonand
occurrence
occurrence intensity
and
typhoonwere
typhoon chosen
intensity
intensity aswere
independent
were chosen
chosen variables.
as as Figure
independent
independent 1 shows
variables.
variables. the nominal
Figure
Figure scale
1 shows
1 shows thethebynominal
typhoon
nominal
scaleintensity
scale
byby and
typhoon
typhoon flood occurrence.
intensity
intensity andand flood
flood occurrence.
occurrence.

Natural
Natural Hazards
Hazards

Tropical
Tropical Cyclones
Cyclones Flood
Flood

Max. speed (km/h) Frequency

Frequency of flood
of flood (per(per year)
year)
Max. speed (km/h)
Zone 0: 76~141 Zone
Zone 1: 1 1: 1
Zone 0: 76~141
Zone 1: 142~184 Zone
Zone 2: 22: 2
Zone 1: 142~184
Zone 2: 185~212 Zone 3: 33: 3
Zone
Zone 2: 185~212
Zone
Zone 4: 44: 4
Zone 3: 213~251
Zone 3: 213~251
Zone 4: 252~290
Zone 4: 252~290 Zone
Zone 5: 55: 5

Zone
Zone 5: 300~
5: 300~

Figure
Figure
Figure1.1.Natural
1. Natural hazards
hazards
Natural factors.
factors.
hazards factors.

TheThe
The Figure
Figure
Figure 2 2indicates
2 indicates thethe
indicates theaverage
average compensation
compensation
average distribution
distribution
compensation byby
distribution byeach
each variable,
variable,
each which
which
variable, show
show
which thethe
show the
criteria
criteria for selecting
for selecting
criteria the
the the
for selecting ordinal
ordinal scale.
scale.
ordinal scale.

₩100
₩100 ₩250
₩250
₩80
₩80 ₩200
₩200
₩60
₩60 ₩150
₩150
₩40
₩40 ₩100
₩100
₩20
₩20 ₩50
₩50

₩0₩0 ₩0₩0
Arch PSCI Cable under under
under under under
under over
over
Arch PSCI Cable
-stayed 50m50m 100m
100m 500m
500m 500m
500m
-stayed

(a)(a) Superstructure
Superstructure (b)(b) Max.
Max. Span
Span Length
Length
₩70
₩70
₩1,400
₩1,400 ₩60
₩60
₩1,200
₩1,200 ₩50
₩50
₩1,000
₩1,000 ₩40
₩40
₩800
₩800 ₩30
₩30
₩600
₩600 ₩20
₩20
₩400
₩400 ₩10
₩10
₩200
₩200 ₩0₩0
₩0₩0 Precast Cast
Precast Cast
in in Open
Open
concrete Place
concrete Place caisson
caisson
ILMILM FCM FSM
FCM FSM MSS
MSS
pilepile

(c)(c) Construction
Construction Method
Method (d)(d) Foundation
Foundation

Figure 2. Cont.
Sustainability 2020, 12, 5316 8 of 16
Sustainability 2020, 12, x FOR PEER REVIEW 8 of 16
Sustainability 2020, 12, x FOR PEER REVIEW 8 of 16

₩200 ₩150
₩200 ₩150
₩150
₩100
₩150
₩100 ₩100
₩100 ₩50
₩50
₩50
₩50
₩- ₩-
₩- 1 2 3 4 ₩- 1 2 3 4
1 2 3 4 1 2 3 4
(e) Flood (f) Typhoon
(e) Flood (f) Typhoon
₩160
₩140
₩160
₩120
₩140
₩100
₩120
₩80
₩100
₩60
₩80
₩40
₩60
₩20
₩40
₩0
₩20
₩0 1 2 3 4 5 6 7 8 9
1 2 3 4 5 6 7 8 (yrs)
9
(yrs)
(g) Total Duration
(g) Total Duration
Figure2.2.The
Figure Theaverage
averagecompensation
compensation distribution.
distribution.
Figure 2. The average compensation distribution.
4. Data
4. DataAnalysis
Analysis andandResult
Result
4. Data Analysis and Result
4.1.4.1.
DataDataCollection
Collectionand and
Method Method
4.1. Data Collection and Method
TheThe risk management
risk managementmeasures measuresininconstruction
construction are classified
classified into intofive
fivecategories,
categories,such suchasas risk
risk
The
retention,
retention, risk management
avoidance,
avoidance,reduction, measures
reduction, transfer in construction
transfer and andsharing are
sharing[52]. classified
[52].In the into
In general five
the general categories,
construction
construction such as risk
industry,
industry, risk
retention,
risk management
management avoidance, reduction,
focuses
focuses transfer
on reducing
on reducing and
risksharing
risk factors along [52].
factors along
with Inthethe general
with the construction
transferring transferring
risks through industry,
risks risk
through
insurance
management
[53]. In this
insurance focuses
[53].study, theon
In this reducing
risk
study, factors risk factors
in bridge
the risk factors along
constructionwithare
in bridge theanalyzed,
transferring
construction arerisks
based on through
the data
analyzed, insurance
regarding
based on the
[53].
367 In
casesthis
of study,
material the risk
damages factors
in in
actualbridge construction
projects and ensuingare analyzed,
insurance
data regarding 367 cases of material damages in actual projects and ensuing insurance claim payouts based
claim on the
payouts data regarding
(compensation
367
madecases
(compensationforoflosses
material
made thatdamages
for losses in
occurred actual
that projects
as occurred
a result asand
of ensuing
material
a result insurance
damages
of material claim payouts
indamages
bridge (compensation
construction
in bridge projects),
construction
made
projects), for
collected losses
from a from
collected that occurred
majorainsurance as a result
company
major insurance of material
in Korea,
company inin damages
orderin
Korea, in
toorder
develop bridge construction
a damage
to develop prediction
a damage projects),
model
prediction
collected
that fromthe
estimate a major
possibleinsurance
financialcompany
losses ininbridge
Korea,construction.
in order to developMore a damage prediction
specifically, in this model
study, the
model that estimate the possible financial losses in bridge construction. More specifically, in this study,
that estimatesignificant
statistically the possiblerisk financial
factors losses
were indrawn
bridge through
construction.the More payout
claim specifically,
data inbythis
a study, the
regression
the statistically significant risk factors were drawn through the claim payout data by a regression
statistically
analysis; in significant
process,risk factors were drawnand through the claim payout data by a regression
analysis; in thetheprocess, thethe derived
derived risk risk factors
factors and the loss
the ratio
loss ratio were
were selected
selectedasasindependent
independent and
and
analysis; in the process, the derived risk factors and the loss ratio were
dependent variables, respectively, to identify the correlation. Figure 3 outlines the flow of analysis.selected as independent and
dependent variables, respectively, to identify the correlation. Figure 3 outlines the flow of analysis.
dependent variables, respectively, to identify the correlation. Figure 3 outlines the flow of analysis.
Bridge Loss Event
Bridge Loss Event
↓
↓
Raw Data
Raw Data
Insurance Payouts Records
Insurance Payouts Records

Engineering information, Address

Engineering information, Address
Project Information, Accident Details
Project Information, Accident Details
↓
↓
Risk Indicators
Risk Indicators

Figure 3. Cont.
Sustainability 2020, 12, 5316 9 of 16
Sustainability 2020, 12, x FOR PEER REVIEW 9 of 16

Dependent Variable
Financial loss: Loss Ratio

Independent Variables

Engineering Natural Hazard Project

Superstructure Flood

Max. span Length Duration

Construction Method Typhoon

Foundation

Regression Analysis

Loss Prediction Model

Figure 3.
Figure 3. Analysis
Analysis procedure.
procedure.

The
The risk
risk prediction
prediction model
model that
that estimates
estimates thethe possible
possible financial
financial lossloss
in in bride
bride construction,
construction, which
which was
developed in this study is intended to utilize the basic or minimal information in the context of
was developed in this study is intended to utilize the basic or minimal information in the context of a
a lack
lack of existing
of existing accumulated
accumulated data
data regardingdamages
regarding damages and
and losses,
losses, in
inorder
ordertotoprovide
providea plan
a planforfor
thethe
development of quantitative risk models and a basis for sustainable risk management.
development of quantitative risk models and a basis for sustainable risk management.

4.2.4.2. RegressionAnalysis
Regression Analysis

4.2.1.
4.2.1. DependentVariable
Dependent Variable

In In order
order totoobjectively
objectivelyandandquantitatively
quantitatively measure
measure the
theloss,
loss,thethedependent
dependentvariable
variable is is
expressed
expressed
as “loss ratio,” which is the amount of the insurance payout divided by the Total Sum Insured (TSI),
as “loss ratio,” which is the amount of the insurance payout divided by the Total Sum Insured (TSI),
corresponding to each accident case of insurance claim. This concept and term of “loss ratio” was
corresponding to each accident case of insurance claim. This concept and term of “loss ratio” was
developed in this study to reflect the possible differences and particularities depending on the size
developed in this study to reflect the possible differences and particularities depending on the size
and scale of each bridge construction project. In other words, it seems reasonable to consider the fact
and scale of each bridge construction project. In other words, it seems reasonable to consider the fact
that, even when the loss amount is relatively low in a project, if the size of the project is rather small
that, even when the loss amount is relatively low in a project, if the size of the project is rather small
the extent of the loss can be more detrimental and severe; Although seemingly paradoxical, it is also
thepossible
extent of theeven
that, loss when
can bethemorelossdetrimental and severe;
amount is relatively Although
high, if the sizeseemingly paradoxical,
of the project is rather it is also
large,
possible that,ofeven
the extent when
the loss canthebe loss amountsomewhat
considered is relatively high,
benign andif slight.
the size of loss
The the project is rather large,
ratio is expressed in
thethe
extent of the loss can
following Equation (1): be considered somewhat benign and slight. The loss ratio is expressed in the
following Equation (1):
Loss Ratio = IP/TSI (1)
Loss Ratio = IP/TSI (1)
Here, IP: Insurance Payout and TSI: Total Sum Insured.
Here, IP: Insurance Payout and TSI: Total Sum Insured.
In each case, the material’s losses were relatively very small compared to the total sum insured.
In each case, the material’s losses were relatively very small compared to the total sum insured.
The distribution of loss ratio is a positive-skew distribution, which means it has a long right tail when
The distribution
expressed of loss ratio
by Equation (1). is
Fora positive-skew
this reason, eachdistribution,
dependentwhich
variable,means
loss itratio,
has awas
longtransferred
right tail when
by
expressed
natural log in order to fit the normal distribution. The value of the transferred loss ratio,transferred
by Equation (1). For this reason, each dependent variable, loss ratio, was dependent by
natural log used
variable, in order toregression
in the fit the normal distribution.
analysis is shown in The value of
Equation (2):the transferred loss ratio, dependent
variable, used in the regression analysis is shown in Equation (2):
Transformed Loss Ratio = Ln(LR) (2)
In the model presented by this Transformed
study, theLoss Ratio =value
predicated Ln(LR)of y(ŷ) is the logarithm of the loss(2)
ratio. The value of the predicted loss ratio represents the value of the loss ratio within the range of 0–
In the
1. No lossmodel presented
whatsoever by this study,
is represented the predicated
by ‘0′ and ‘1′ representsvalue of y(loss.
an entire ŷ) is The
the Ln(Predicted
logarithm of Ratio)
the loss
ratio. The
value value as
derived of athe predicted
result lossregression
of a valid ratio represents
analysis the
is value of the
expressed as loss ratio within
a decimal number the
byrange of 0–1.
Equation
No(4),
loss whatsoever
which is represented
allows the by ‘0’ andof‘1’quantified
relative comparison representsscale
an entire loss. Thethe
to determine Ln(Predicted
level of the Ratio) value
loss. This
derived
ratio in fact was also used in several other studies as a measure of quantitative loss and risk [50,53,54].(4),
as a result of a valid regression analysis is expressed as a decimal number by Equation
Sustainability 2020, 12, 5316 10 of 16

which allows the relative comparison of quantified scale to determine the level of the loss. This ratio in
fact was also used in several other studies as a measure of quantitative loss and risk [50,53,54].

Ln(Predicted Ratio) = β0 + β1 ·X1 + β2 ·X2 + β3 ·X3 + . . . (3)

where, β0 = Y-intercept; β1 , β2 , β3 = Coefficient; X1 , X2 , X3 = Independent variables.

Predicted Ratio = e β0 + β1·X1+ β2·X2 + β3·X3+ . . . (4)

4.2.2. Independent Variables

In order to set up the group of independent variables, the abovementioned data and information
(Section 3.2) were utilized and the following risk indicator categories were developed: (1) bridge
engineering factors, (2) natural disaster factors, and (3) project information of bridge construction.
The first category, which is bridge engineering factors, consists of the superstructure types, foundation
types, superstructure construction methods, and maximum span length. The second category, natural
disaster factors, includes the intensity of flooding and typhoons. The third group, project information,
includes the construction duration and the total construction cost. Table 3 organizes the categories and
included variables.

Table 3. Categories of risk indicator.

Category Variables Unit Description

1: Arch bridge
Superstructure Number (Ordinal Scale) 2: PSCI beam bridge
3: Cable-stayed bridge
Max span Meter Maximum span length
1: ILM (Incremental Launching Method)
Engineering factor
2: FCM (Free Cantilever Method)
Construction Method Number (Ordinal Scale)
3: FSM (Full Staging Method)
4: MSS (Movable Scaffolding System)
1: Precast concrete pile
Foundation Number (Ordinal Scale) 2: Cast in place
3: Open caisson
Flood Number (Zone) Frequency of flood in local area
Natural hazard factor
Typhoon Number (Zone) Max. speed of Typhoon in local area
Project factor Duration Month Total duration of the project

Here, variables such as superstructure types, foundation types, and construction methods were
not used to evaluate the levels of safety, but to identify the change in the loss rate and thus influence on
the financial losses. This was done by sorting the loss levels of each type into nominal scales, based on
the actual loss data of the particular type. This process was to provide reasonable evidence for the
analysis by determining statistical significance.

4.2.3. Validation of the Model

To provide statistical evidence of the significant risk factors in bridge construction, the normality
of the data has been tested using Shapiro–Wilk tests. When each value was put into the equation,
the Loss Ratio is inclined toward zero with a high skewness value, because the majority of the project
cost tend to be extremely large, compared to the loss. Yet, the distribution of the Loss Ratio satisfied
the normality through natural log transformation. Shapiro–Wilk normality tests also provides the clear
evidence of the normal distribution (Table 4).
Sustainability 2020, 12, x FOR PEER REVIEW 11 of 16

Table 4. Normality test of dependent value.

Shapiro–Wilk Test
Sustainability 2020, 12, 5316 11 of 16
Statistic df sig.
LR 0.377 367 0.000
Ln(LR) 0.948 Table 4. Normality367
test of dependent value. 0.087
Shapiro–Wilk Test
Additionally, the normal P-P plot of regression standardized residual and residual scatterplots
Statistic df sig.
are used to test whether the basic assumptions of regression analysis, normality of standard residuals
LR 0.377 367 0.000
and homoskedasticity are satisfied. The normal probability plot in Figure 4a shows the scatters of the
Ln(LR) 0.948 367 0.087
residuals’ approximately linear pattern without any extreme outliers. This means that the error terms
are indeed normally distributed.
By plotting thethe
Additionally, standardized
normal P-P residuals against the
plot of regression standardized
standardized predicted
residual values asscatterplots
and residual shown in
Figure 4b,tothe
are used testassumption
whether theofbasic
homoscedasticity
assumptions offor regression
regression analysis
analysis, was not of
normality violated.
standardIn residuals
general,
the
andscatter plot shouldare
homoskedasticity be satisfied.
distributed
Thebetween −2 and +2 at
normal probability theinregression-standardized-predicted
plot Figure 4a shows the scatters of the
value (the approximately
residuals’ horizontal axis), andpattern
linear the regression-standardized residuals
without any extreme outliers. This(the vertical
means axis)
that the areterms
error also
considered to be gooddistributed.
are indeed normally for homogeneity of variance when they are distributed between −2 and +2.

(a) Normal P-P plot (b) Standardized residual scatter plot

Figure
Figure4.
4. Validation
Validation of
of regression
regression model.
model.

By plotting
Table the standardized
5 displays the results of residuals against
the regression the standardized
model. As mentioned predicted
earlier, forvalues as shown in
the independent
Figure 4b,seven
variables, the assumption
variables in of homoscedasticity
total were adapted forincluding
regressionsuperstructure
analysis was not violated.
types, maximumIn general,
span
the scatter plot should be distributed between −2 and +2 at the regression-standardized-predicted
length, foundation types, superstructure construction methods, typhoon intensity, flood intensity, value
(the total
and horizontal axis), and
construction the regression-standardized
duration. The analysis found that, residuals
among(the vertical
these sevenaxis) are also considered
independent variables,
to be good for homogeneity of variance when they are distributed between
three of them including superstructure types, superstructure construction methods and the total −2 and +2.
Table 5 displays
construction durationthe areresults of the regression
the significant model.that
risk indicators As mentioned
affected theearlier, for the
loss ratio, independent
in other words,
variables,
the amount seven
of thevariables
loss ininbridge
total were adapted including
construction. In addition,superstructure
the maximum types,
spanmaximum
length, span length,
foundation
types, typhoon intensity, and flood occurrence were not statistically significant, which indicatestotal
foundation types, superstructure construction methods, typhoon intensity, flood intensity, and no
construction
correlation duration.
with the lossThe analysis
ratio. In the found that,analysis
regression among theseof theseven independent
loss ratio of the bridgevariables, three of
construction,
them
as canincluding
be seen insuperstructure
Table 5, the revisedtypes,R2superstructure
value is 0.315, construction
which indicates methods and of
the 31.5% the total construction
explanation power
duration are the significant risk indicators that affected the loss ratio, in
of the regression model. The R value can be increased when additional significant variables other
2 other words, the amount of
the loss
than in presented
those bridge construction.
are reflected. In Furthermore,
addition, the maximum
the range ofspan VIF length,
values isfoundation
between 1.033 types,andtyphoon
1.050,
intensity,
which and that
means floodthere
occurrence were not statistically
is no multicollinearity betweensignificant, which A
the variables. indicates no correlation
standardized with the
beta coefficient
loss ratio. the
compares In the regression
strength of theanalysis
effect ofofeachthe loss ratio ofindependent
individual the bridge construction, as can
variable to the be seen in
dependent Table 5,
variable.
theother
In revised R2 value
words, is 0.315, which
a standardized indicates coefficient
regression the 31.5% of explanation
indicates the power
importanceof theofregression model.
the regression
The R 2 value can be increased when additional significant variables other than those presented are
coefficient; and the higher the value of the beta coefficient of the variable, the greater the effect on the
reflected. Furthermore,
dependent variable. When the range of VIF values
the absolute value of is between 1.033 and 1.050,
the beta coefficient of eachwhich meansisthat
indicator there is no
expressed in
multicollinearity between the variables. A standardized beta coefficient compares the strength of the
effect of each individual independent variable to the dependent variable. In other words, a standardized
regression coefficient indicates the importance of the regression coefficient; and the higher the value
Sustainability 2020, 12, 5316 12 of 16

of the beta coefficient of the variable, the greater the effect on the dependent variable. When the
absolute value of the beta coefficient of each indicator is expressed in descending order, it would
be organized as follows: (1) bridge superstructure type (beta coefficient = 0.284) (2) superstructure
construction method (beta coefficient = 0.262) (3) construction duration (beta coefficient = −0.247).
This can be interpreted that the longer the construction duration, the lower the loss ratio, and the
loss ratio increases according to the increases in the nominal scale of the superstructure type and the
superstructure construction method.

Table 5. Summary of the regression model.

Variables Coef. Beta Coef. p > |z| VIF

Engineering factor
Superstructure 0.683 0.284 0.024 1.033
Construction method 0.272 0.262 0.031 1.050
Project factor
Construction duration −0.814 −0.247 0.039 1.044
Number of Observations = 367
F value = 8.924
Adj-R2 = 0.315

5. Discussion and Conclusions

The risks in construction projects refer to uncertain conditions and events, and, if they occur, there
are negative results in terms of construction costs, time, quality, and so on. Therefore, potential risk
factors may be the important aspects that can determine the success or failure of the entire project.
Therefore, effective and sustainable risk management can reduce and minimize losses from various
potential risk factors. For this reason, understanding and managing possible risk factors should be an
essential consideration in construction projects.
It is expected that the recent increase of bridge construction projects and the risk factors resulting
from the introduction of the new construction technologies will continue to be on rise. Risk factors
in bridge construction projects due to the use of large construction equipment under complex and
unstable environmental conditions, and simultaneous ground and high-altitude work have yet to be
systematically identified and managed. However, there are limitations in preventing construction
disasters only with safety facility equipment, safety guidelines and training, as done so far. Therefore,
it has been strongly called for to predict and prepare for possible losses in bridge construction accidents
in more reliable and sustainable ways. In this study, the risk factors in bridge construction were
analyzed based on the actual data on insurance claim payout from a Korean insurance company,
which included 367 material damage claims. The risk prediction model was also developed to predict
the future bridge construction losses. Various studies in the past have suggested the analysis of
risk factors in general construction as well as bridge construction and pointed out the need for the
quantitative data and methodologies in risk analysis. However, due to the lack of detailed loss records,
the tendency of companies that hesitate to disclose the actual loss records, the qualitative analyses
based on expert surveys and interviews were performed in most previous studies. These previous
studies include possible limitations of subjective and intuitive judgment that rely on personal opinions
and experiences.
Compared to such qualitative data, on the contrary, it can be said that insurance claim payout
data used in this study is by far the most specific and accurate. Further, these quantitative data can
be used as objective data for assessing the risks of potential projects and predicting them, using the
characteristics particular to each project. In this study, the extent of potential risks through actual losses
was able to be identified and the impact on losses was determined through statistical verification with
actual case data. As a result, this study found that the three statistically significant risk factors that can
Sustainability 2020, 12, 5316 13 of 16

be used for bridge risk analysis, which are superstructure types, superstructure construction methods
and the construction duration.
The findings of this study also demonstrate consistency with the previous studies. That is,
the superstructure types indicate that the order of the nominal scale criteria of arch, PSCI and
suspension bridges are significant and that it is similar to the classification by the PQ review criteria
grade. Additionally, this study found that the level of the risk in the superstructure construction
methods tended to be similar to the order of the approximate cost of the construction methods.
More specifically, the order of the levels of risk identified in this study was I.L.M, F.C.M, F.S.M and
M.S.S, from bottom to top, and the order of the approximate cost of construction identified in a previous
study [36] was I.L.M, F.S.M, F.C.M, and M.S.S. In addition, the loss ratio of damages tended to be
relatively lower, as the construction duration was extended. This was also similar to the previous
studies on general construction and plant construction projects [37,38]. However, it is important to note
that, although the derived indicators in this study have a correlation with the amount of the insurance
payout, or the losses, it does not account for the causality and, thus, further studies to develop the
detailed risk prevention measures are required.
As for the implications of this study, the central and local governments and government-invested
institutions, which are the main stakeholders of the bridge construction, should develop new
technologies in order to improve construction productivity, but also to make an overall improvement
of risk management techniques, and to make predictions about future losses based on basic bridge
construction planning information, thus enabling preparation for the financial risks. The utilization
of the findings of this study is expected to contribute to the establishment of effective measures of
risk management reflecting the characteristics of bridge construction and the prevention as well as to
minimize losses. In addition, as mentioned earlier, the collected data in this study are quantified and
enumerated in numerical values, and thus more objective and reliable. Yet, in terms of the amount of
detailed and specific information, the data provided no more than basic and minimal information about
the bridge construction projects or damages occurred in the projects. For this reason, for construction
companies and insurance companies, the framework of deriving risk factors and developing the risk
prediction model presented in this study can also contribute to estimating the financial losses with the
available minimum information.
Insurers and reinsurers, particularly, can use the risk factors from this study to reconstruct their
loss/risk prediction model. Insurers can reconstruct their own model of measuring potential risks for
a particular bridge construction by estimating the possible loss, and use the results as key data for
calculating the rates of premiums. Construction companies can assess the maximum possible losses in
bridge construction and advance bridge construction planning that minimizes the material losses by
optimizing bridge superstructure type, construction method, and construction duration, which were
found to be the three significant risk factors in this study. Ultimately, the framework of developing the
loss prediction model suggested in this study and the risk factors derived from the framework can be
used to efficiently distribute and allocate various resources.
We believe that the value of this study is not a low number to describe the natural phenomena
statistically. However, the adjusted R2 value is 0.315, and the remaining variability of the damage
is described by an indeterminate predictor. Therefore, future research is needed to determine other
potential predictors and add them to the model. Regarding what constitutes as relevant R2 value,
we have referred to varying opinions of scholars as following:

(1) Falk and Miller [55] suggested that R2 values should be equal to or greater than 0.10 in order for
the variance explained of a particular endogenous construct to be deemed adequate.
(2) Cohen [56] recommended R2 values for endogenous latent variables are assessed as follows:
0.26 (substantial), 0.13 (moderate), 0.02 (weak).
(3) Chin [57] suggested R2 values for endogenous latent variables based on 0.67 (substantial),
0.33 (moderate), 0.19 (weak).
Sustainability 2020, 12, 5316 14 of 16

In order to identify more diverse risk factors and for advancing the accuracy of quantitative
risk prediction models in the future, government agencies and construction participators responsible
for design and engineering will need to make continuous efforts to establish a broader and more
sophisticated loss-recording database system and to establish the technical and institutional foundation
to accumulate data in the event of damages and losses.

Author Contributions: Conceptualization, S.A.; Data curation, S.A., J.-M.K.; Funding acquisition, T.K.;
Investigation, S.A., J.-M.K.; Methodology, S.A.; Project administration, T.K.; Software, J.-M.K.; Writing—original
draft, J.-M.K.; Writing—review and editing, S.A. All authors have read and agree to the published version of
the manuscript.
Funding: This research was supported by a grant (NRF-2019R1A2C1009398) from the National Research
Foundation of Korea by Ministry of Science, ICT and Future Planning.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Kim, J.M.; Kim, T.; Son, K.; Bae, J.; Son, S. A quantitative risk assessment development using risk indicators
for predicting economic damages in construction sites of South Korea. J. Asian Archit. Build. Eng. 2019, 18,
472–478. [CrossRef]
2. Wood, G.; Ellis, R.C.T. Risk management practices of leading UK cost consultants. Eng. Constr. Archit. Manag.
2003, 10, 254–262. [CrossRef]
3. Dikmen, I.; Birgonul, M.T.; Arikan, A.E. A critical review of risk management support tools. In Proceedings
of the 20th Annual Conference of Association of Researchers in Construction Management, Association of
Researchers in Construction Management, Edinburgh, UK, 1–3 September 2004; Heriot-Watt University:
Edinburgh, UK, 2004; pp. 1145–1154.
4. Taroun, A.; Yang, J. Dempster-Shafer theory of evidence: Potential usage for decision making and risk
analysis in construction project management. Built Hum. Environ. Rev. 2011, 4, 155–166.
5. Cooper, D.F.; Grey, S.; Raymond, G.; Walker, P. Project Risk Management Guidelines: Managing Risk in Large
Projects and Complex Procurements; Wiley: Chichester, UK, 2005.
6. California Department of Transportation (Caltrans). Project Risk Management Handbook; Office of Project
Management Process Improvement: Sacramento, CA, USA, 2007.
7. Ebrahimnejad, S.; Mousavi, S.M.; Mojtahedi, S.M.H. A fuzzy decision-making model for risk ranking with
application to the onshore gas refinery. Int. J. Bus. Contin. Risk Manag. 2009, 1, 38–66. [CrossRef]
8. Tavakkoli-Moghaddam, R.; Mojtahedi, S.M.H.; Mousavi, S.M.; Aminian, A. A jackknife technique to estimate
the standard deviation in a project risk severity data analysis. In Proceedings of the 39th International
Conference on Computers and Industrial Engineering, Troyes, France, 6–8 July 2009.
9. Morgan, M.G.; Henrion, M. Uncertainty; Cambridge University Press: New York, NY, USA, 1990.
10. United Nations Department of Humanitarian Affairs (UNDHA). Internationally Agreed Glossary of Basic Terms
Related to Disaster Management; UNDHA: Geneva, Switzerland, 1992.
11. Stenchion, P. Development and Disaster Management. Aust. J. Emerg. Manag. 1997, 12, 40–44.
12. Crichton, D. The risk triangle. In Natural Disaster Management; Ingleton, J., Ed.; Tudor Rose: London, UK,
1999; pp. 102–103.
13. Jones, R.N.; Boer, R. Assessing current climate risks. In Adaptation Policy Frameworks for Climate Change:
Developing Strategies, Policies and Measures; Cambridge University Press: Cambridge, UK, 2004; pp. 91–117.
14. Zou, P.X.; Zhang, G.; Wang, J. Understanding the key risks in construction projects in China. Int. J. Proj.
Manag. 2007, 25, 601–614. [CrossRef]
15. Baloi, D.; Price, A.D. Modelling global risk factors affecting construction cost performance. Int. J. Proj. Manag.
2003, 21, 261–269. [CrossRef]
16. Laryea, S. Risk pricing practices in finance, insurance and construction. In Proceedings of the Construction and
Building Research Conference of the Royal Institution of Chartered Surveyors, Dublin, Ireland, 4–5 September
2008; University of Reading: London, UK, 2008; pp. 1–16.
17. Baker, S.; Ponniah, D.; Smith, S. Techniques for the analysis of risks in major projects. J. Oper. Res. Soc. 1998,
49, 567–572. [CrossRef]
Sustainability 2020, 12, 5316 15 of 16

18. Khan, F.; Rathnayaka, S.; Ahmed, S. Methods and models in process safety and risk management: Past,
present and future. Process Saf. Environ. 2015, 98, 116–147. [CrossRef]
19. Warszawski, A.; Sacks, R. Practical multifactor approach to evaluating risk of investment in engineering
projects. J. Constr. Eng. Manag. 2004, 130, 357–367. [CrossRef]
20. Ben-David, I.; Raz, T. An integrated approach for risk response development in project planning. J. Oper. Res.
Soc. 2001, 52, 14–25. [CrossRef]
21. Fan, C.F.; Yu, Y.C. BBN-based software project risk management. J. Stat. Softw. 2004, 73, 193–203. [CrossRef]
22. Molenaar, K.R. Programmatic cost risk analysis for highway megaprojects. J. Constr. Eng. Manag. 2005, 131,
343–353. [CrossRef]
23. Cagno, E.; Caron, F.; Mancini, M. A multi-dimensional analysis of major risks in complex projects. Risk Manag.
2007, 9, 1–18. [CrossRef]
24. Keshk, A.M.; Maarouf, I.; Annany, Y. Special studies in management of construction project risks, risk concept,
plan building, risk quantitative and qualitative analysis, risk response strategies. Alex. Eng. J. 2018, 57,
3179–3187. [CrossRef]
25. Aziz, N.A.A.; Manab, N.A.; Othman, S.N. Sustainability Risk Management (SRM): An extension of Enterprise
Risk Management (ERM) concept. Int. J. Manag. Sustain. 2016, 5, 1–10. [CrossRef]
26. Majdalani, Z.; Ajam, M.; Mezher, T. Sustainability in the construction industry: A Lebanese case study.
Constr. Innov. 2006, 6, 33–46. [CrossRef]
27. Bal, M.; Bryde, D.; Fearon, D.; Ochieng, E. Stakeholder Engagement: Achieving Sustainability in the
Construction Sector. Sustainability 2013, 5, 695–710. [CrossRef]
28. Shen, L.; Wu, Y.; Zhang, X. Key assessment indicators for the sustainability of infrastructure projects. J. Constr.
Eng. Manag. 2011, 137, 441–451. [CrossRef]
29. Adetola, A.; Goulding, J.S.; Liyanage, C.L. Collaborative engagement approaches for delivering sustainable
infrastructure projects in the AEC sector: A review. Int. J. Constr. Supply Chain Manag. 2011, 1, 1–24.
[CrossRef]
30. Zhang, X.Q.; Kumaraswamy, M.M. Procurement protocols for public-private partnered projects. J. Constr.
Eng. Manag. 2011, 127, 351–358. [CrossRef]
31. Wilderer, P.A.; Renn, O.; Grambow, M.; Molls, M.; Mainzer, K. (Eds.) Sustainable Risk Management; Springer
International Publishing: Cham, Switzerland, 2018; p. 274.
32. Wang, Y.M.; Elhag, T.M.S. A fuzzy group decision making approach for bridge risk assessment. Comput. Ind.
Eng. 2007, 53, 137–148. [CrossRef]
33. Decò, A.; Frangopol, D.M. Risk assessment of highway bridges under multiple hazards. J. Risk Res. 2011, 14,
1057–1089. [CrossRef]
34. Hashemi, H.; Mousavi, S.M.; Mojtahedi, S.M. Bootstrap technique for risk analysis with interval numbers in
bridge construction projects. J. Constr. Eng. Manag. 2011, 137, 600–608. [CrossRef]
35. Li, Q.F.; Li, Z.X.; Niu, J. Application of factor analysis to risk evaluation of bridge construction. Adv. Mater.
Res. 2011, 243, 1848–1853. [CrossRef]
36. Choudhry, R.M.; Aslam, M.A.; Hinze, J.W.; Arain, F.M. Cost and schedule risk analysis of bridge construction
in Pakistan: Establishing risk guidelines. J. Constr. Eng. Manag. 2014, 140, 04014020-1-9. [CrossRef]
37. Yun, S.Y.; Kim, C.H.; Kang, L.S. Development of model for selecting superstructure type of small size bridge
using dual classification method. J. Korean Soc. Civ. Eng. 2015, 35, 1413–1420. [CrossRef]
38. Malekly, H.; Mousavi, S.M.; Hashemi, H. A fuzzy integrated methodology for evaluating conceptual bridge
design. Expert Syst. Appl. 2010, 37, 4910–4920. [CrossRef]
39. Public Procurement Agency. Available online: http://www.law.go.kr/ᄒ ᆼᄌ
ᅢ ᆼᄀ
ᅥ ᅲᄎ ᆨ/ᄌ
ᅵ ᅩᄃ ᆯᄎ
ᅡ ᆼᄋ
ᅥ ᆸᄎ
ᅵ ᅡ
ᆯᄎ ᆷᄀ
ᅡ ᅡᄌ ᅡᄀᆨᄉ
ᅧ ᅡᄌᆫᄉ
ᅥ ᆷ
ᅵ
ᅡᄀ
ᄉ ᅵᄌᆫ/(9746,20191209) (accessed on 3 March 2020).
ᅮ
40. Koh, H.M.; Kim, H.J.; Lim, J.H.; Kang, S.C.; Choo, J.F. Lifetime design of cable-supported super-long-span
bridges. In Proceedings of the 5th International Conference on Bridge Maintenance, Safety and Management,
Philadelphia, PA, USA, 11–15 July 2010; CRC Press: London, UK, 2010; pp. 26–42.
41. Jo, B.W.; Park, J.C.; Kim, C.H. Wind characteristics of existing long span bridge based on measured data.
J. Civ. Eng. 2005, 9, 219–224. [CrossRef]
42. Yang, Y.; Gang, Y.; Wei, F.; Qin, W. Buffeting performance of long-span suspension bridge based on measured
wind data in a mountainous region. J. Vibroeng. 2018, 20, 621–635. [CrossRef]
Sustainability 2020, 12, 5316 16 of 16

43. Elsaid, A.; Seracino, R. Rapid assessment of foundation scour using the dynamic features of bridge
superstructure. Constr. Build. Mater. 2014, 50, 42–49. [CrossRef]
44. Giroux, R.P. Relevance of Roebling. J. Perform. Constr. Facil. 2009, 23, 2–4. [CrossRef]
45. Won, J.H.; Kim, G.Y. Structural Safety Analysis of a Long Span Cable-stayed Bridge with a Partially Earth
Anchored Cable System on Dynamic Loads during Construction. J. Korean Soc. Saf. 2016, 31, 104–110. [CrossRef]
46. Sampaio, A.Z.; Martins, O.P. The application of virtual reality technology in the construction of bridge:
The cantilever and incremental launching methods. Autom. Constr. 2014, 37, 58–67. [CrossRef]
47. Kwak, H.G.; Son, J.K. Determination of design moments in bridges constructed with a movable scaffolding
system (MSS). Comput. Struct. 2006, 84, 2141–2150. [CrossRef]
48. Pan, N.F. Fuzzy AHP approach for selecting the suitable bridge construction method. Autom. Constr. 2008,
17, 958–965. [CrossRef]
49. Kim, S.B.; Cho, J.H. Development of the approximate cost estimating model for PSC box girder bridge based
on the breakdown of standard work. J. Korean Soc. Civ. Eng. 2013, 33, 791–800. [CrossRef]
50. Kim, J.M.; Kim, T.; Bae, J.; Son, K.; Ahn, S. Analysis of plant construction accidents and loss estimation using
insurance loss records. J. Asian Archit. Build. Eng. 2019, 18, 507–516. [CrossRef]
51. Ryu, H.; Son, K.; Kim, J.M. Loss prediction model for building construction projects using insurance claim
payout. J. Asian Archit. Build. Eng. 2016, 15, 441–446. [CrossRef]
52. Vaughan, E.; Vaughan, T. Essential of Insurance: A Risk Management Perspective; John Wiley & Sons: New York,
NY, USA, 1995; p. 656.
53. Kim, J.M.; Kim, T.; Son, K.; Yum, S.G.; Ahn, S. Measuring Vulnerability of Typhoon in Residential Facilities:
Focusing on Typhoon Maemi in South Korea. Sustainability 2019, 11, 2768. [CrossRef]
54. Kim, J.M.; Woods, P.K.; Park, Y.J.; Kim, T.; Son, K. Predicting hurricane wind damage by claim payout based
on Hurricane Ike in Texas. Geomat. Nat. Hazards Risk. 2016, 7, 1513–1525. [CrossRef]
55. Falk, R.F.; Miller, N.B. A Primer for Soft Modeling; University of Akron Press: Akron, OH, USA, 1992.
56. Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Lawrence Erlbaum Associates, Inc.:
Mahwah, NJ, USA, 1988.
57. Chin, W.W. The Partial Least Squares Approach to Structural Equation Modeling. In Modern Methods
for Business Research; Marcoulides, G.A., Ed.; Lawrence Erlbaum Associates: Mahwah, NJ, USA, 1998;
pp. 295–358.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).
 © 2020. This work is licensed under
http://creativecommons.org/licenses/by/3.0/ (the “License”). Notwithstanding
the ProQuest Terms and Conditions, you may use this content in accordance
with the terms of the License.