Cost Risk Factors in Construction Projects: A Contractor’s Perspective

Abstract

Cost overrun is a major challenge in the construction industry. However, there is a notable lack of data from empirical studies that exhaustively identify and analyze risk factors contributing to overruns. This study aims to address this gap by systematically identifying and analyzing these risk factors. A hybrid methodology was employed. It combined a systematic literature review, a three-round Delphi process, and fuzzy set techniques. Insights from the literature review informed the first-round Delphi questionnaire. Subsequent rounds were refined based on earlier results. In the third round, experts’ opinions on the likelihood and impact of the cost risk factors were collected using a 5-point Likert scale. Finally, a fuzzy approach was employed to assess the severity of cost risk factors based on the combined effects of their likelihood and impact. The results revealed that the primary cost risk factors include escalation and fluctuation in material prices, inflation, material shortages, the country’s political instability, the country’s economic instability, delays in payment to the contractor, and delays in material procurement and delivery. Notably, the significant cost risk factors are largely beyond the contractor’s control and are closely tied to the broader political and economic conditions of the country.

1. Introduction

The construction industry (CI) plays a crucial role in the socio-economic development of any nation, where its contribution is even higher in the economies of developing countries [1,2]. The industry’s impact is contingent upon its performance, which is assessed based on the achievement of the project goals: time, cost, and quality [3], known as the iron triangle [4]. Nevertheless, cost overrun is one of the major challenges of the industry, where its severity is more pronounced in the cases of developing countries [5]. In Ethiopia, for example, Belay et al. [6] evaluated the cost performance of building and infrastructure projects and identified a cost overrun of up to 248% and 61%, respectively. Similarly, Belachew et al. [7] reported a cost overrun of up to 83.2% among eight road projects.

Conversely, cost overruns can be managed by the implementation of systematic cost risk management [8]. Risk management is the art of dealing with risks [9], which involves the systematic application of risk identification, analysis, response, and monitoring and control [10,11]. However, the Ethiopian Construction Industry (ECI) is known for its poor project risk management practice [12,13] that have resulted in a number of unsuccessful projects in the country [14].

Risk identification involves a systematic identification and documentation of risk factors [15]. It is a fundamental step in formal risk management. For a risk to be managed, it must be identified [16]. Several studies have been conducted to identify risk factors causing cost overruns in the global CI. However, it has been claimed that the majority of the studies conducted in the case of developing countries are not context-specific, such as project type or stakeholder perspective. Therefore, identifying cost overrun risks that are specific to the project type and from a specific stakeholder’s perspective has been suggested [17]. In the case of the ECI, for example, Mitikie et al. [18] ranked risk factors affecting the project objectives (time, cost, and quality) by combining the responses of contractors and consultants. However, this approach may overlook important distinctions, since risk factors can vary depending on the context and the stakeholder, as what matters to one may not matter to another [19,20]. Therefore, identifying cost risk factors from the perspective of each stakeholder is advised. A study by Zewdu and Aregaw [21] ranked the cost risk factors from clients’, contractors’, and consultants’ perspectives independently. However, the study ranked these factors in terms of their importance only, whereas risk factors are better ranked based on the combined effects of at least their probability of occurrence and impact [20]. Furthermore, these studies have used the findings of limited research literature in developing their questionnaire, which could have a higher chance of researcher bias in the literature selection and the non-inclusion of some risk factors. In addition, these studies have solely relied on the simple relative importance index [18] or mean [21] rankings that do not entertain the uncertainties in expert judgments. Therefore, a comprehensive and stakeholder, project-type, and objective-specific cost risk identification and analysis is lacking in the ECI.

To address these gaps, this study utilized a sequential and context-sensitive framework that considers uncertainties for identifying and evaluating cost overrun factors in construction projects with empirical importance to the Ethiopian building construction projects from the contractor’s perspective. The study integrates (1) a systematic literature review (SLR) that identifies a comprehensive list of cost overrun factors by using a systematic and evidence-based screening, (2) a three-round Delphi process that is used in refining the list with contractor experts’ consensus confirming their relevance to the Ethiopian building construction projects, and (3) the fuzzy logic that is used to evaluate the probability and impact of these risk factors independently, considering the uncertainty in contractor expert judgment.

This study provides a context-sensitive ranking of cost overrun risk factors, particularly for the ECI and other developing countries, where data shortage and institutional volatility are challenges. It evaluates the probability and impact of risks independently by applying fuzzy logic. Therefore, it provides uncertainty-conscious risk evaluation techniques by improving the traditional Delphi-based average ranking. In general, though several studies have identified risk factors, the aim of this study is not to discover a new risk factor. Rather, it advances the existing body of knowledge by showing how risk severity varies when expert uncertainty, economic volatility, and specific project objectives are considered.

In addressing the aim, the following research questions will be answered.

•

RQ1. What are the main risk factors that cause cost overruns to construction projects in developing countries?

•

RQ2. What are the main risk factors causing cost overruns in the Ethiopian building construction projects from the contractors’ perspective?

2. Materials and Methods

To address the research aim, the study utilized a multi-stage Systematic Literature Review (SLR), an iteration of three-round Delphi techniques, and fuzzy logic. The aim of this multi-stage filtering process is to distinguish between risks that are frequently cited in the literature and those that remain critical when evaluated by experts under uncertain and volatile conditions. Accordingly, the SLR captured the theoretical consensus in the literature, while the Delphi process is used to contextualize the knowledge within the ECI context. Fuzzy logic accounts for uncertainty in expert judgment. Therefore, the results reflect context-sensitive cost risks rather than a generic ranking. Figure 1 summarizes the process involving these methods.

Figure 1. The research process.

2.1. The Systematic Literature Review Procedure

The SLR is a stand-alone research method that utilizes secondary data from relevant literature [22] to describe the knowledge and practices in a given subject [23]. The SLR is claimed to experience researcher bias. However, that can be addressed through the utilization of a systematic methodological process commonly known as the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) [23,24]. Accordingly, this SLR followed the PRISMA protocol to ensure transparency and reproducibility. Figure 2 shows the process of the SLR in view of the PRISMA protocol specifically utilized in this study to address the risk of bias in the document search.

Figure 2. Adopted PRISMA protocol.

Initially, a documents search was conducted on 31 December 2024 on the Scopus database to access documents published from 2010 to 2024. Scopus was selected as it has the most extensive coverage and is solely utilized in most studies [25,26]. The search protocol and database limiting criteria adopted were: TITLE-ABS-KEY ((“cost overrun” OR “cost risk” OR “budget overrun” OR “risk factors”) AND (“construction projects”)) AND PUBYEAR > 2009 AND PUBYEAR < 2025 AND (LIMIT-TO (SUBJAREA, “ENGI”)) AND (LIMIT-TO (DOCTYPE, “ar”) OR LIMIT-TO (DOCTYPE, “re”)) AND (LIMIT-TO (PUBSTAGE, “final”)) AND (LIMIT-TO (LANGUAGE, “English”)).

Then 1288 documents were exported to Excel and screened based on the predefined inclusion and exclusion criteria. Accordingly, articles (1) that are about construction projects in developing countries and (2) that have ranked the critical cost risk factors were included. Documents not meeting the above inclusion criteria simultaneously were excluded from the review. Countries were categorized based on the provisions of the World Economic Situation and Prospects report [27].

The document screening was done twice by the first writer independently, and the final result was presented to the panel of the other writers for the final decision. Accordingly, after detailed evaluation of the titles, abstracts, and full documents, fifty-one articles meeting the inclusion criteria were screened for final review (The list of the screened articles can be found from supplementary materials: Table S1). The contents of the articles were examined twice in detail by the first writer, and eighty-one cost risk factors were extracted and categorized into ten based on their nature.

Risk categorization is important in risk identification, as it facilitates the understanding of its nature for effective management [15]. Risks can be categorized based on their source [28], nature [29,30,31,32,33,34], project stage [35,36], and the originating party [34,37], where there is no consensus on which to adopt. However, classification based on nature is most commonly used. For instance, Siraj and Fayek [33] evaluated 130 construction studies and found that 50 articles (38.46%) adopted the nature-based risk categorization. A similar technique is also suggested by generic project risk management guidelines such as PMI [15].

In addition to the SLR’s primary focus on the studies of developing countries’ construction industry (CI), the majority of the articles reviewed (46 out of 51) were published within the past ten years (Figure 3). Thus, the findings of the SLR are relevant if used as an input in assessing the cost risk factors in the recent ECI.

Figure 3. Reviewed articles according to publication period.

2.2. The Delphi Technique

The Delphi technique is an alternative to focus group discussions [38]. It involves distributing a series of questions, each built upon the results of the previous rounds [39], to get experts’ consensus anonymously [40]. Delphi is becoming a robust tool commonly utilized in the research of construction engineering and management fields [41]. In order to ensure the reliability and credibility of the Delphi process, the procedures suggested by Junger et al. [42] were adopted.

Informed decision-making on the number of rounds and the quality and number of experts is the fundamental requirement of the Delphi technique. It depends upon the available resources and time, the nature of the research question, and the expert’s willingness to participate in each round [40]. Up to six-round Delphi studies are identified in the literature; however, conducting Delphi with more than three rounds is uncommon as they are costly, time-consuming, and can lead to expert fatigue and low response rates [41]. Conversely, finalizing results in one or two rounds can result in misleading conclusions [43]. A three-round Delphi is commonly utilized in most construction engineering and management studies [43,44,45]. In relation to the number of experts, it has been evidenced that a sample of 8–20 experts is common in construction engineering and management research [41].

Therefore, in this study, the consensus of 30 experts was assessed with an iteration of a three-round Delphi process. A three-round Delphi process was utilized to improve consensus and to reduce respondent fatigue while ensuring the reliability and validity of risk factors. The aim of the first round is to systematically screen risks from the literature by removing irrelevant factors. The second round is a screening process to validate the stability of the selected factors and to reduce random selection bias. The third round involved a detailed evaluation of the remaining risk factors, enabling fuzzy transformation and ranking. In general, the adopted three-round process increases robustness by progressively focusing expert attention while reducing the risk of early-stage bias.

The study adopted a predefined consensus criterion for each Delphi step. In the early exploratory components, the first two Delphi rounds adopted a majority agreement cut-off threshold (>50%). Though some literature suggested the use of a higher consensus level, such as ≥75% [42]. This study adopted a >50% cut-off threshold to avoid a premature exclusion of risk factors from the final analysis. It has been justified that there is no single and universal threshold level, and the adoption is dependent on factors such as the study aim [42,46]. Similar cut-off logic has been applied in other research [43,47,48].

In the third round, consensus was determined using the Average Fuzzy Distance (d) calculated for each risk parameter: the probability and impact independently. The overall consensus was assumed to be achieved when the calculated d ≤ 0.2 for both parameters simultaneously. Accordingly, factors meeting the consensus criteria for their probability and impact at a time will be retained, while those failing are considered to lack consensus and excluded from further analysis. Cost overrun factors with the lower values of d indicate strong consensus. The detailed step-by-step procedure and formula for calculating the value of d can be found in the literature [49]. Some fuzzy-based Delphi studies report both the average fuzzy distance and percent agreement as consensus criteria [49]. However, this study used the average fuzzy distances as the primary consensus criterion. The fuzzy distance provides an accurate and continuous measure of expert judgment proximity in fuzzy space. It is thus acceptable for capturing expert judgment differences that cannot be indicated by categorical percentage agreement. Thus, using only average fuzzy distance avoids biases in considering adjacent Likert ratings as disagreement, and it ensures a mathematically robust and theoretically consistent consensus measure.

As the study aims to identify cost overrun factors from the contractors’ perspective, the selection of experts started with the selection of the construction companies. It is assumed that experts working in companies with better project management backgrounds have a better understanding of the recent ECI. In this regard, the Ethiopian Conformity Assessment Enterprise (ECAE) was requested for the list of certified contractors, as it is responsible for certifying the management system of companies in Ethiopia. Until 2 December 2024, only seven certified contractors were found in the list, among which 4 privately owned companies were purposely selected based on their experiences and extent of their participation in different regions of the country. Though the role of public contractors in the ECI should not be undermined, private contractors are significantly participating in various regions of Ethiopian building construction projects. In addition, maintaining homogeneity among companies is advantageous in maintaining the equilibrium of respondents’ risk perception.

Initially, the heads of the companies were contacted, and a brief introduction about the aim of the research was given to them. Then, they were requested to suggest their technical people who meet the following expert criteria. For an expert to be selected, they (1) should have a minimum of a bachelor’s degree, (2) should have more than five years of experience in the ECI, and (3) should be currently participating in activities such as project risk management, cost risk management, cost estimation, or project management in general.

Accordingly, the Delphi was started with the participation of a panel of 30 experts in the first round and completed with the participation of 26 experts in the third round (86.67% response rate). Figure 4a,b show the experts’ demography, where the majority of experts have above a master’s degree (69%) and more than 10 years of experience (73%).

Figure 4. Experts’ demography.

Though Delphi is criticized for its poor response rate in subsequent rounds [40], starting the process with a face-to-face method was suggested to increase the experts’ commitment [50,51]. Accordingly, a face-to-face data collection was used in the first round, during which a detailed explanation about the procedures and number of Delphi rounds was given to the experts. In addition, experts were requested to provide their e-mail addresses to confirm their willingness to participate in the subsequent rounds. All respondents expressed their willingness and provided their e-mail addresses. In the second and third rounds, experts were contacted via their e-mail.

Initially, the questionnaire was developed based on the findings of the SLR. A pilot test of the questionnaire was conducted with two experts: one Ph.D. holder and one M.Sc. holder with more than 18 years of academic and industry experience in the risk management of the ECI. The experts were requested to review the clarity of the questionnaire and suggest additional cost risk factors common to the ECI. Additional cost risk factors were not suggested. However, some wording adjustments were made, and the final questionnaire of the first-round Delphi was developed.

In the first round, experts were requested to select the cost overrun factors they commonly face from the lists identified by the SLR. Then, frequency analysis was conducted, and factors that were selected by the majority of the experts (>50%) were forwarded to the second round. Then, in the second round, experts were again requested to select from the factors forwarded from the first round, and factors that received more than 50% approval were then forwarded to the third round for final analysis. In the third round, experts were requested to independently rate the probability and impact of risk factors exported from the second round using a five-point Likert scale—from Very Low (1) to Very High (5)—adopted from PMI [15].

Although the Delphi technique facilitates experts’ consensus, it has limitations in accounting for the uncertainties inherent in human judgments. The third-round Delphi survey used linguistic scales to evaluate the probability and impact of the risks that cannot be adequately aggregated using crisp numerical averages. Therefore, utilizing fuzzy logic is advantageous by transforming such linguistic variables into fuzzy numbers, thereby preserving uncertainty and improving the robustness of risk prioritization.

2.3. Fuzzy Technique

Fuzzy logic—introduced by Zadeh [52]—is suited to mathematically interpret subjective reasoning where a definite conclusion is required from ambiguous information in the absence of precise data [53,54]. Fuzzy logic-based approaches can make the construction decision-making process more transparent and allow experts to express themselves in linguistic terms rather than in numerical terms [53,55]. Fuzzy technique is becoming a more researched topic in construction engineering and management fields in the form of either fuzzy set alone or hybrid fuzzy techniques (such as fuzzy neural networks, fuzzy Delphi, and fuzzy analytic hierarchy process). However, the use of fuzzy hybrid techniques is effective in addressing problems that cannot be addressed by fuzzy sets alone [56,57]. Thus, this study utilized a Fuzzy–Delphi hybrid approach to analyze the experts’ subjective judgment.

For a given fuzzy set A in the universe of discourse, the membership function of the universal set X, µ_A (x), is defined as µ_A (x): X→ [0, 1], determines the degree that element x belongs to the fuzzy set A [52,54]. The membership value 0 shows that the element x does not belong to the fuzzy set A, whereas the value 1 shows the element fully belongs to the set A. Any value between 0 and 1 show the element partially belongs to the set, A [52,58]. The basic step-by-step procedures of the fuzzy technique used in this study are (1) fuzzification, (2) aggregation, (3) calculation of fuzzy risk magnitude, (4) defuzzification, and (5) cost risk factor ranking and analysis. Similar techniques have been successfully utilized in other studies [59,60]. The steps are outlined in the subsequent sections.

2.3.1. Step 1: Fuzzification

In the third round of the Delphi process, experts rated the linguistic variables, probability, and impact, using a linguistic scale from Very Low (1) to Very High (5). Linguistic variables are variables whose values are linguistic scales [61]. The linguistic scale represents a group of fuzzy numbers [60]. Thus, fuzzification is the technique of transforming the crisp values of linguistic variables into the corresponding fuzzy membership functions [62]. Several geometric mappings of membership functions, such as triangular, trapezoidal, and S-shaped, have been used [54,63]. However, this study adopted triangular membership functions as they are widely used because of their simplicity to compute and interpret [55]. A triangular membership function is basically represented by (a, b, c), where a < b < c. The parameters a and c represent the two extreme points of the base of a triangle, where b indicates the pick point of the triangle projected to its base. The triangular membership function can be symmetric or asymmetric [64]. Figure 5 and Equation (1) show the graphical and mathematical representation of the triangular membership functions, respectively [65,66]. For two triangular fuzzy membership functions A₁ = (a₁, b₁, c₁) and A₂ = (a₂, b₂, c₂) the basic fuzzy operational laws can be referred to the literature [54,67].

$${\mu }_A(x)=\left\{\begin{array}{l}{\displaystyle \frac{x-a}{b-a}},x\in [a,b]\\ {\displaystyle \frac{c-x}{c-b}},x\in [b,c]\\ 0,otherwise\end{array}\right.$$

(1)

Figure 5. Representation of triangular membership functions.

The fuzzy scale is selected based on the type of fuzzy numbers that are suitable to present the linguistic variable in a problem [60]. Table 1 and Figure 6 depict the triangular membership functions adopted in this study. Table 2 illustrates the fuzzification process of the experts’ judgments for the cost risk factor “escalation and fluctuation in material prices”.

Table 1. Adopted linguistic scales and triangular membership functions.

Linguistic Evaluation	Linguistic Scale	Triangular Fuzzy Number
Very Low (VL)	1	(0,0,0.25)
Low (L)	2	(0,0.25,0.5)
Medium (M)	3	(0.25,0.5,0.75)
High (H)	4	(0.5,0.75,1)
Very High (VH)	5	(0.75,1,1)

Figure 6. The fuzzification diagram.

Table 2. Illustrations of the fuzzification process.

Expert’s ID	Linguistic Evaluation		Likert Scale Response		Fuzzification
	Likelihood (L)	Impact (I)	Likelihood (L)	Impact (I)	Likelihood (L)	Impact (I)
EXP1	High	Very High	4	5	(0.5,0.75,1)	(0.75,1,1)
EXP2	Very High	Very high	5	5	(0.75,1,1)	(0.75,1,1)
…	…	…	…	…	…	…
EXP26	Very High	High	5	4	(0.75,1,1)	(0.5,0.75,1)
Average					(0.577,0.827,0.942)	(0.654,0.904,0.981)

2.3.2. Step 2: Aggregation

Once the experts’ judgments are fuzzified, the next step is to conduct aggregation of their judgments for the probability and impact independently. Aggregation is a technique by which several fuzzy sets are combined to produce a single fuzzy set [54]. In fuzzy aggregation, the main concern was how expert judgments are summarized into a set of consistent group judgments. In response to this, several aggregation techniques, such as t-norms, t-conforms, and simple averaging operations, are suggested in the literature [68,69]. In this study, the averaging aggregation technique is used, as it is commonly utilized in most similar studies due to its practical simplicity.

Therefore, for n number of experts participating in a study, the aggregated expert’s judgment is computed by Equation (2) provided by Klir and Yuan [54].

$${{\displaystyle A}}_{average}={\displaystyle \frac{{{\displaystyle A}}_1+{{\displaystyle A}}_2+{{\displaystyle A}}_3+\dots +{{\displaystyle A}}_n}{n}}$$

(2)

where A₁, A₂, A₃, and A_n are the fuzzy sets with fuzzy numbers (a₁, b₁, c₁), (a_n, b_n, c_n). Thus, the aggregate fuzzy numbers for the probability (FP) and impact (FI) of the ith cost risk factor and jth expert could be computed by Equations (3) and (4).

$${FP}_i={\displaystyle \frac{1}{n}}\sum _{j=1}^n{FP}_{ij}$$

(3)

$${FI}_i={\displaystyle \frac{1}{n}}\sum _{j=1}^n{FI}_{ij}$$

(4)

For example, the aggregate value for probability and impact of the cost risk factor “escalation and fluctuation in material prices” is: FP = (0.577, 0.827, 0.942) and FI = (0.654, 0.904, 0.981).

2.3.3. Step 3: Determination of Fuzzy Risk Magnitude

Basically, the magnitude of a risk is computed as the product of its probability and impact [9]. Thus, the fuzzy risk magnitude (FRF_i) of the ith cost risk factor is computed by multiplying the aggregated fuzzy values of probability (FP_i) and impact (FI_i) for each risk factor by using Equation (4).

FRF_i = FP_i × FI_i,

(5)

For instance, the fuzzy risk magnitude for the cost risk factor “escalation and fluctuation in material prices” is (0.577, 0.827, 0.942) ⊗ (0.654, 0.904, 0.981) = (0.377, 0.747, 0.924).

2.3.4. Step 4: Defuzzification

For a risk factor to be practically evaluated, the fuzzy values should be converted to a single crisp value. Defuzzification is the technique of transforming the fuzzy values (which represent uncertainty) into a single crisp value for the subsequent decision-making [58]. The literature has provided a number of defuzzification techniques with their advantages and disadvantages [70,71]. Their selection depends on factors such as their suitability to the desired fuzzification output [70]. This used the centroid method (Equation (6)) as it is commonly utilized in symmetrical fuzzy shapes such as triangular [66,72].

$${{\displaystyle x}}_{Crisp}={\displaystyle \frac{a+b+c}{3}}$$

(6)

where a, b, and c are the membership values for a given triangular fuzzy set A. For instance, the defuzzified value for the cost risk factor “escalation and fluctuation in material prices” is 0.683.

2.3.5. Step 5: Risk Factors Analysis

In this final step, cost risk factors are qualitatively analyzed based on the magnitude of their corresponding crisp value derived from defuzzification (step 4). Risks are basically analyzed by using qualitative and quantitative techniques. The qualitative technique is the critical step used to analyze the effects of individual factors, whereas the quantitative step is used to assess their joint effect on the project objective [15]. In both cases, risk probability and impact are the fundamental inputs [73].

The probability-impact (P-I) matrix is a tool for qualitative risk analysis commonly utilized in different sectors, ranging from construction to healthcare [73] due to its simplicity in application and construction [74]. The matrix categorizes risk factors using linguistic terms such as low, moderate, high, and very high, built based on the ‘if-and-then’ rules like: if the probability is ‘High’ and the impact is ‘Low’, then the risk’s severity is ‘Moderate’, shown in yellow color in Figure 7 [15].

Figure 7. Typical probability-impact matrix.

In spite of its popularity, the P-I matrix technique has limitations. For example, it does not support the decision-makers in prioritizing risk factors having the same severity category. Rather, it pushes the risk factors to the extreme corner of the linguistic terms. That means the risk severity should be low, medium, high, or very high [74]. This may mislead decision-makers during the assigning of mitigation strategies and may result in a waste of resources [73]. For instance, the matrix in Figure 7 collectively classified risk factors with different levels of severity—0.420, 0.480, and 0.603—as ‘High’. Though, they require different management attention.

In responding to this problem, researchers have proposed solutions. Acebes, Gonzalez-Varona, Lopez-Paredes and Pajares [73] suggested the use of a quantitative matrix by incorporating Monte Carlo simulation, while Levine [74] recommended the use of a logarithm-based matrix. However, these suggestions could reduce their practical utilization—simplicity—as they require structured data input and analytical skills. Therefore, incorporating a means that promotes the strengths and overcomes the weaknesses of the P-I matrix is suggested. Accordingly, this study used the advantage of fuzzy techniques described in the above steps to rank and qualitatively analyze risk factors classified under a given severity category.

Accordingly, a range of crisp values that is in agreement with the descriptions in a matrix is suggested. Thus, the risks are categorized as: low, moderate, high, or very high based on their proximity to the extreme bounds within the range [0.020–0.854]. The highest value (0.854) is where both the probability and impact of the risk are ‘Very High’, while the lowest crisp value (0.020) reflects situations where both dimensions are rated as ‘Very Low’. The crisp value ranges for each category are summarized in Table 3. Similar risk categorization techniques were used in other studies [11,59,60]. This technique is advantageous, as it can be used as a simplified risk probability and impact matrix.

Table 3. Crisp value ranges and risk categorization.

Risk Category	CRISP Value Range
Low	(0.02—0.103)
Moderate	(0.104—0.29)
High	(0.30—0.603)
Very High	(0.604—0.853)

The color in the table shows the agreement between the conventional P-I matrix risk zoning (Figure 7) and the proposed crisp value zoning.

Once the risks are linguistically classified, they are ranked based on their crisp magnitude to determine the relative importance for management. For instance, the crisp value for the cost risk factor “escalation and fluctuation in material prices” is 0.683, which is in the range of (0.604–0.853), and hence it is among the ‘Very High’ severity risk factors and ranked 1st (Table 3).

3. Results

The SLR resulted in eighty-one cost risk factors that are common in the developing countries’ CI. Table 4 shows that escalation and fluctuation in material prices, design change, poor quantity and cost estimation, variation or change order (addition/omission), and delay in payment to the contractor are the top five factors in the literature.

Table 4. Results of the SLR.

Risk Factors	References	Frequency	%	Rank
Management Risk Factors
1. Poor coordination among parties involved in the project	[75,76,77,78,79]	5	10%	31
2. Lack of project management experience and skills of the project team	[76,79,80,81,82,83]	6	12%	26
3. Inadequate project planning and scheduling	[76,77,81,84,85,86,87,88,89,90,91,92]	12	24%	9
4. Poor communication among parties involved in the project	[81,86,88,93,94,95,96,97,98]	9	18%	14
5. Poor site management and supervision by the contractor	[19,75,76,81,86,87,88,97,99,100,101,102,103]	13	25%	6
6. Poor experience of the owner in project management	[79,89,98]	3	6%	43
7. Poor project management such as cost, resource, and risk	[78,82,84,86,89,95,104,105,106]	9	18%	14
8. Delayed decision making by the client	[76,79,84,89,90,92,94,97,98,103,107]	11	22%	10
9. Poor contract management	[81,86,92,106,108]	5	10%	31
10. Unstable organizational environment	[81]	1	2%	64
Technical Risk Factor
11. Design error	[76,77,86,92,94,98,100,103,106,109]	10	20%	11
12. Design change	[75,76,77,78,82,83,84,85,86,88,89,92,93,94,95,106,107,110,111,112]	20	39%	2
13. Lack of consistency between design and specification	[78,94]	2	4%	56
14. Insufficient feasibility study before design	[19,77,95]	3	6%	43
15. Design delay	[85,97,101,113]	4	8%	38
16. Design and project scope complexity	[78,80,89]	3	6%	43
17. Consultant and design team incompetency	[76,86,94,98,101,114]	6	12%	26
18. Poor quantity and cost estimation	[6,19,76,77,81,84,86,88,89,90,93,98,102,109,110,115,116]	17	33%	3
19. Poor estimation of activity durations	[19,76,78,81,102,115,116]	7	14%	21
20. Poor project scope definition	[117,118]	2	4%	56
21. Negligence of site visit during bidding	[115]	1	2%	64
22. Lack of proper professional software	[78]	1	2%	64
23. Incomplete design and specification	[75,76,116,118]	4	8%	38
Construction Risk Factors
24. Rework due to poor workmanship and poor quality of work	[80,86,91,100,107,114]	6	12%	26
25. Rework due to construction mistake or error	[77,88,100,102,104,105,119]	7	14%	21
26. Temporary delays and interruptions causing cost overrun	[92,101,103]	3	6%	43
27. Adoption of improper or unaccepted construction methods	[77,103,120]	3	6%	43
28. Contractor’s incompetency and lack of experience in similar projects	[76,82,86,87,88,89,98,103,120]	9	18%	14
29. Pressure to deliver project on accelerated schedule	[104,115,119]	3	6%	43
30. Making tough decisions by engineers during project implementation	[117]	1	2%	64
31. Delay in execution of planned activity	[86,88,93,100,121,122]	6	12%	26
32. High level of quality requirement by client	[91]	1	2%	64
Resource Related Risk Factors
33. Shortage of skilled labor in the project area	[6,75,76,77,78,81,87,92,103,104,107,117,123]	13	25%	6
34. Shortage of materials	[6,75,82,83,94,104,123]	7	14%	21
35. Shortage of equipment	[6,75,80,82,83,94,104,119,123]	9	18%	14
36. Suppliers’ monopoly	[121]	1	2%	64
37. Poor labor productivity	[75,76,86,113]	4	8%	38
38. Sub-contractors’ failure; default of sub-contractors	[19,76,87,90,120]	5	10%	31
39. Delay in procurement and delivery of materials	[78,80,90,93,94,121,123]	7	14%	21
40. High staff turnover	[87]	1	2%	64
Site Condition Risk Factors
41. Unforeseen subsurface conditions	[80]	1	2%	64
42. Unforeseen and differing site conditions	[19,83,116]	3	6%	43
43. Poor site investigations such as geological site survey and soil test	[19,83,90,100,103]	5	10%	31
44. Delay in site possession	[80,91,111,112]	4	8%	38
45. Improper project site or location	[106]	1	2%	64
Contractual and Legal Risk Factors
46. Contradictions in documents and ambiguity in contract terms	[77,78,83]	3	6%	43
47. Deficient contracts with discrepancies and errors	[90,99]	2	4%	56
48. Project scope change	[76,99,100,104,110,113,116,120,122]	9	18%	14
49. Variation or change order (addition/omission)	[6,19,84,86,87,94,101,105,109,110,111,112,117,119,120]	15	29%	4
50. Contractual claims and disputes	[75,76,86,90,111,116]	6	12%	26
51. The lowest bid price award system	[19,83,85,87,100,104,108,115,121]	9	18%	14
52. Inappropriate procurement system and contract type	[19,89,120]	3	6%	43
53. Issuing strict instructions and regulations during crisis	[117]	1	2%	64
54. Resistance to follow regulations	[117]	1	2%	64
55. Bureaucracy in tendering method	[90]	1	2%	64
Economic and Financial Risk Factors
56. Inflation	[6,77,80,82,99,110,115,116,118,123]	10	20%	11
57. Funding problems (financial difficulty of client)	[76,82,84,89,90,92,95,97,98,103,108,114,116]	13	25%	6
58. Fluctuations in currency exchange rate	[78,93,106,108,116,121,123]	7	14%	21
59. Increase banks loan interest rates	[78,83,99]	3	6%	43
60. Escalation and fluctuation in material prices	[6,76,77,78,80,83,84,86,88,92,93,94,95,97,100,102,104,106,107,108,111,112,118,119,121]	25	49%	1
61. Delay in payments to the contractor	[76,78,84,85,86,87,90,94,95,98,99,120,122,123]	14	27%	5
62. Country’s economic instability	[79,88,90]	3	6%	43
63. High transportation costs	[77]	1	2%	64
64. High bond and insurance rates	[90,120]	2	4%	56
65. Increase in equipment and machinery cost	[77,83,90,93,115]	5	10%	31
66. Increase in labor costs or wages	[6,77,83,93,100]	5	10%	31
67. Contractor’s financial difficulty	[75,76,84,87,90,96,97,107,116,123]	10	20%	11
68. Rebates	[106,115]	2	4%	56
69. Tax liability	[78]	1	2%	64
Social and Cultural Risk Factors
70. Time and cost required for land acquisition and compensation	[80,91]	2	4%	56
71. Differences in culture, language, and religious backgrounds	[122]	1	2%	64
72. Insecurity and crime (theft, vandalism, kickbacks, and fraudulent practices)	[79,118]	2	4%	56
Governmental and Political Risk Factors
73. Changes in government laws, regulations, and policies affecting the project	[77,78,93,106]	4	8%	38
74. County’s Political Instability	[103,121,122]	3	6%	43
75. Political interferences	[110]	1	2%	64
76. Delay or refusal of the government bodies in project approval and permit	[90,107,117]	3	6%	43
77. Corruption and bribery	[78,117]	2	4%	56
Environmental and Safety Risk Factors
78. Adverse weather conditions (continuous rainfall, snow, temperature, wind)	[19,78,79,83,103,105,112,122]	8	16%	20
79. Force majeure (natural and man-made disasters which are beyond the firm’s control, e.g., fire, floods, thunder and lightning, landslide, earthquake, hurricane)	[97,98,106,110,117]	5	10%	31
80. Epidemic illness or disease	[106]	1	2%	64
81. Difficulty of applying new health and safety standards	[117]	1	2%	64

The results of the first two Delphi rounds forwarded fifty-four cost risk factors that were finally reduced to forty-six, receiving an overall consensus fuzzy distance value of d ≤ 0.2. The results of the fuzzy analysis revealed that escalation and fluctuation in material prices, inflation, shortage of materials, country’s political instability, country’s economic instability, delay in payment to the contractor, and delay in procurement and delivery of materials are the critical cost risk factors, each exerting a ‘Very High’ impact on project costs.

The results show that the majority of cost risk factors excluded by Ethiopian experts (24 out of 35 during the Delphi rounds were those with the least frequency in the literature (less than 6%)). This agreement could be because the reviewed articles are based on countries with similar economic status and, hence, construction practice to Ethiopia. However, the majority of the factors classified as ‘Very High’ magnitude in the Ethiopian context have a frequency from low to high in the SLR. This could be because the scope in the Ethiopian context was building contractors, whereas the SLR was not stakeholder and project-specific. Furthermore, the cost risk factors finally assessed have ‘High’ or ‘Very high’ severity in the Ethiopian context. This confirms the advantage of Delphi techniques, as risk factors with the least severity were excluded during the first and third rounds. Table 5 summarizes the results of the study.

Table 5. Results of the data analysis.

Cost Risk Factors	Delphi Round-1		Delphi Round-2		Delphi Round-3					Fuzzy Analysis
	Agreement%	Consensus	Agreement %	Consensus	Probability		Impact		Overall Consensus	Crisp Value	Category	Rank
					d Value	Consensus	d Value	Consensus
Management Risk Factors
1. Poor coordination among parties involved in the project	90%	Yes	85%	Yes	0.173	Yes	0.171	Yes	Yes	0.517	HI	26
2. Lack of project management experience and skills of the project team	67%	Yes	67%	Yes	0.165	Yes	0.176	Yes	Yes	0.457	HI	39
3. Inadequate project planning and scheduling	87%	Yes	78%	Yes	0.130	Yes	0.146	Yes	Yes	0.535	HI	22
4. Poor communication among parties involved in the project	70%	Yes	70%	Yes	0.149	Yes	0.154	Yes	Yes	0.505	HI	29
5. Poor site management and supervision by the contractor	70%	Yes	52%	Yes	0.170	Yes	0.136	Yes	Yes	0.484	HI	32
6. Poor experience of the owner in project management	70%	Yes	52%	Yes	0.179	Yes	0.157	Yes	Yes	0.404	HI	49
7. Poor project management such as cost, resource, and risk	93%	Yes	70%	Yes	0.098	Yes	0.122	Yes	Yes	0.592	HI	9
8. Delayed decision making by the client	93%	Yes	85%	Yes	0.194	Yes	0.150	Yes	Yes	0.582	HI	13
9. Poor contract management	80%	Yes	85%	Yes	0.142	Yes	0.141	Yes	Yes	0.58	HI	14
10. Unstable organizational environment	70%	Yes	44%	No	-	-	-	-	-	-	-	-
Technical Risk Factors
11. Design error	67%	Yes	63%	Yes	0.169	Yes	0.211	No	No	-	-	-
12. Design change	93%	Yes	93%	Yes	0.161	Yes	0.161	Yes	Yes	0.533	HI	23
13. Lack of consistency between design and specification	60%	Yes	59%	Yes	0.164	Yes	0.192	Yes	Yes	0.388	HI	52
14. Insufficient feasibility study before design	87%	Yes	85%	Yes	0.136	Yes	0.142	Yes	Yes	0.531	HI	25
15. Design delay	73%	Yes	59%	Yes	0.179	Yes	0.178	Yes	Yes	0.469	HI	37
16. Design and project scope complexity	50%	No	-	-	-	-	-	-	-	-	-	-
17. Consultant and design team incompetency	60%	Yes	56%	Yes	0.186	Yes	0.157	Yes	Yes	0.404	HI	49
18. Poor quantity and cost estimation	80%	Yes	81%	Yes	0.151	Yes	0.150	Yes	Yes	0.562	HI	17
19. Poor estimation of activity durations	73%	Yes	63%	Yes	0.166	Yes	0.170	Yes	Yes	0.509	HI	28
20. Poor project scope definition	70%	Yes	70%	Yes	0.204	No	0.172	Yes	No	-	-	-
21. Negligence of site visit during bidding	53%	Yes	52%	Yes	0.220	No	0.189	Yes	No	-	-	-
22. Lack of proper professional software	37%	No	-	-	-	-	-	-	-	-	-	-
23. Incomplete design and specification	70%	Yes	78%	Yes	0.194	Yes	0.137	Yes	Yes	0.532	HI	24
Construction Risk Factors
24. Rework due to poor workmanship and poor quality of work	70%	Yes	63%	Yes	0.177	Yes	0.183	Yes	Yes	0.374	HI	53
25. Rework due to construction mistake or error	50%	No	-	-	-	-	-	-	-	-	-	-
26. Temporary delays and interruptions causing cost overrun	73%	Yes	81%	Yes	0.174	Yes	0.211	No	No	-	-	-
27. Adoption of improper or unaccepted construction methods	40%	No	-	-	-	-	-	-	-	-	-	-
28. Contractor’s incompetency and lack of experience in similar projects	60%	Yes	56%	Yes	0.196	Yes	0.183	Yes	Yes	0.396	HI	51
29. Pressure to deliver project on accelerated schedule	70%	Yes	48%	No	-	-	-	-	-	-	-	-
30. Making tough decisions by engineers during project implementation	40%	No	-	-	-	-	-	-	-	-	-	-
31. Delay in execution of planned activity	80%	Yes	81%	Yes	0.130	Yes	0.171	Yes	Yes	0.54	HI	21
32. High level of quality requirement by client	43%	No	-	-	-	-	-	-	-	-	-	-
Resource Related Risk Factors
33. Shortage of skilled labor in the project area	67%	Yes	44%	No	-	-	-	-	-	-	-	-
34. Shortage of materials	97%	Yes	89%	Yes	0.151	Yes	0.143	Yes	Yes	0.653	VH	3
35. Shortage of equipment	90%	Yes	81%	Yes	0.160	Yes	0.152	Yes	Yes	0.585	HI	12
36. Suppliers’ monopoly	43%	No	-	-	-	-	-	-	-	-	-	-
37. Poor labor productivity	60%	Yes	52%	Yes	0.198	Yes	0.159	Yes	Yes	0.472	HI	35
38. Sub-contractors’ failure; default of sub-contractors	70%	Yes	78%	Yes	0.185	Yes	0.172	Yes	Yes	0.439	HI	43
39. Delay in procurement and delivery of materials	83%	Yes	89%	Yes	0.147	Yes	0.133	Yes	Yes	0.619	VH	7
40. High staff turnover	63%	Yes	30%	No	-	-	-	-	-	-	-	-
Site Condition Risk Factors
41. Unforeseen subsurface conditions	83%	Yes	70%	Yes	0.203	Yes	0.166	Yes	Yes	0.446	HI	42
42. Unforeseen and differing site conditions	90%	Yes	63%	Yes	0.189	Yes	0.135	Yes	Yes	0.474	HI	34
43. Poor site investigations such as geological site survey and soil test	77%	Yes	81%	Yes	0.164	Yes	0.174	Yes	Yes	0.516	HI	27
44. Delay in site possession	87%	Yes	74%	Yes	0.018	Yes	0.152	Yes	Yes	0.57	HI	15
45. Improper project site or location	50%	No	-	-	-	-	-	-	-	-	-	-
Contractual and Legal Risk Factors
46. Contradictions in documents and ambiguity in contract terms	50%	No	-	-	-	-	-	-	-	-	-	-
47. Deficient contracts with discrepancies and errors	43%	No	-	-	-	-	-	-	-	-	-	-
48. Project scope change	80%	Yes	81%	Yes	0.192	Yes	0.145	Yes	Yes	0.471	HI	36
49. Variation or change order (addition/omission)	93%	Yes	85%	Yes	0.142	Yes	0.186	Yes	Yes	0.544	HI	18
50. Contractual claims and disputes	83%	Yes	74%	Yes	0.172	Yes	0.172	Yes	Yes	0.479	HI	33
51. The lowest bid price award system	87%	Yes	67%	Yes	0.226	No	0.219	Yes	No	-	-	-
52. Inappropriate procurement system and contract type	57%	Yes	70%	Yes	0.209	No	0.210	No	No	-	-	-
53. Issuing strict instructions and regulations during crisis	40%	No	-	-	-	-	-	-	-	-	-	-
54. Resistance to follow regulations	33%	No	-	-	-	-	-	-	-	-	-	-
55. Bureaucracy in tendering method	50%	No	-	-	-	-	-	-	-	-	-	-
Economic and Financial Risk Factors
56. Inflation	97%	Yes	96%	Yes	0.142	Yes	0.142	Yes	Yes	0.679	VH	2
57. Funding problems (financial difficulty of client)	90%	Yes	74%	Yes	0.186	Yes	0.194	Yes	Yes	0.568	HI	16
58. Fluctuations in currency exchange rate	97%	Yes	85%	Yes	0.145	Yes	0.176	Yes	Yes	0.597	HI	8
59. Increase banks loan interest rates	57%	Yes	33%	No	-	-	-	-	-	-	-	-
60. Escalation and fluctuation in material prices	97%	Yes	85%	Yes	0.164	Yes	0.112	Yes	Yes	0.683	VH	1
61. Delay in payments to the contractor	97%	Yes	96%	Yes	0.143	Yes	0.156	Yes	Yes	0.62	VH	6
62. Country’s economic instability	90%	Yes	81%	Yes	0.168	Yes	0.161	Yes	Yes	0.627	VH	5
63. High transportation costs	70%	Yes	37%	No	-	-	-	-	-	-	-	-
64. High bond and insurance rates	47%	No	-	-	-	-	-	-	-	-	-	-
65. Increase in equipment and machinery cost	97%	Yes	81%	Yes	0.156	Yes	0.170	Yes	Yes	0.588	HI	10
66. Increase in labor costs or wages	63%	Yes	44%	No	-	-	-	-	-	-	-	-
67. Contractor’s financial difficulty	80%	Yes	74%	Yes	0.161	Yes	0.166	Yes	Yes	0.586	HI	11
68. Rebates	20%	No	-	-	-	-	-	-	-	-	-	-
69. Tax liability	27%	No	-	-	-	-	-	-	-	-	-	-
Social and Cultural Risk Factors
70. Time and cost required for land acquisition and compensation	87%	Yes	85%	Yes	0.189	Yes	0.164	Yes	Yes	0.543	HI	19
71. Differences in culture, language, and religious backgrounds	37%	No	-	-	-	-	-	-	-	-	-	-
72. Insecurity and crime (theft, vandalism, kickbacks, and fraudulent practices)	70%	Yes	63%	Yes	0.188	Yes	0.223	Yes	Yes	0.416	HI	47
Government and Political Risk Factors
73. Changes in government laws, regulations, and policies affecting the project	67%	Yes	63%	Yes	0.238	Yes	0.215	Yes	Yes	0.414	HI	48
74. County’s Political Instability	90%	Yes	96%	Yes	0.174	Yes	0.143	Yes	Yes	0.642	VH	4
75. Political interferences	63%	Yes	74%	Yes	0.229	No	0.169	Yes	No
76. Delay or refusal of the government bodies in project approval and permit	60%	Yes	44%	No	-	-	-	-	-	-	-	-
77. Corruption and bribery	70%	Yes	74%	Yes	0.156	Yes	0.183	Yes	Yes	0.541	HI	20
Environmental and Safety Risk Factors
78. Adverse weather conditions (continuous rainfall, snow, temperature, wind)	77%	Yes	81%	Yes	0.190	Yes	0.195	Yes	Yes	0.431	HI	45
79. Force majeure (natural and man-made disasters which are beyond the firm’s control, e.g., fire, floods, thunder and lightning, landslide, earthquake, hurricane)	63%	Yes	74%	Yes	0.216	No	0.204	No	No	-	-	-
80. Epidemic illness or disease	23%	No	-	-	-	-	-	-	-	-	-	-
81. Difficulty of applying new health and safety standards	33%	No	-	-	-	-	-	-	-	-	-	-

HI: High severity cost risk factor, VH: Very high severity cost risk factor.

4. Discussion

4.1. Cost Risk Factors in the Developing Countries

The results of the SLR revealed that escalation and fluctuation in material prices (49%), design change (39), poor quantity and cost estimation (33%), variation or change order (addition/omission) (29%), and delay in payment to the contractor (27%) are the top five risk factors causing cost overrun in the developing countries’ CI.

4.2. Cost Risk Factors in the Ethiopian Context

4.2.1. Escalation and Fluctuation in Material Prices

Escalation and fluctuation in material prices are identified as the major cause for cost overrun, aligning with the findings of Ashtari, Ansari, Hassannayebi and Jeong [80], Xie, Deng, Yin, Lv and Deng [106], and the SLR. Construction materials account for 60% of the project cost [124,125], meaning that even a small price increase can significantly impact the overall project cost. Essentially, escalation is triggered by external factors such as supply chain disruption, geopolitical instability, inflation, variation in exchange rate, and a country’s economic instability [93,126]. Thus, escalation and fluctuation in material prices remain beyond the control of the contractor and their effect is difficult to forecast [127]. However, providing a price adjustment clause in the contract document is commonly suggested in managing such factors [36,106].

4.2.2. Inflation

Inflation is the second-ranked cost risk factor, aligning with the findings of Belay, Tilahun, Yehualaw, Matos, Sousa and Workneh [6]. Inflation drives up prices for materials, labor, equipment, and hire rates, making it one of the key causes of cost overrun [128]. Though forecasting inflation is not simple, estimators should conduct extensive information and data collection to properly incorporate the effect of inflation [129]. Musarat et al. [130] suggested the use of a time series construction rate forecasting model, which incorporates the possible inflation rates. Furthermore, Ebekozien et al. [131] suggested a downward review of the monetary policy rate, control of exchange rate volatility, and addressing insecurity as a strategy to mitigate the effect of inflation.

4.2.3. Shortage of Materials

Materials shortage is the other critical cause for cost overruns, supported by Berihu, Grum, Tariku and Abebe [94]. Materials affect up to 80% of the construction activity [125]. However, the CI is challenged by material shortages mainly due to poor quantification, delays in material procurement and delivery, supply of defective materials, poor inventory control, delays in the production of special materials, and frequent material quality and specification changes [132]. Thus, the materials shortage is primarily due to the contractor’s fault, but other parties’ faults should not be ignored. Management strategies such as materials planning, logistics, handling, and stock and waste control are the suggested mitigation strategies for materials shortage [133].

4.2.4. County’s Political Instability

Political instability is among the significant cost risk factors consistent with Mahmud, Ogunlana and Hong [122]. Political instability is one of the external risk factors affecting countries’ investment activity [134]. A politically unstable environment creates conditions that escalate project costs and even lead to full project interruptions [127]. According to Bekr [135], for instance, projects under politically unstable situations significantly encountered cost overruns and delays due to the higher cost of security, corruption, and unofficial holidays.

4.2.5. Country’s Economic Instability

The country’s economic instability is the other challenge, in agreement with Shaikh [103]. It is an external risk factor that influences managerial decision-making [136]. Project cost overruns and delays are the main challenges during economic instability that can be mitigated by implementing improved planning, effective cash flow management, cost reduction measures, supply chain optimization, risk management, and fostering organizational agility [136,137]. These suggestions reinforce the importance of resilient financial and operational strategies to manage economic instability, especially in developing countries.

4.2.6. Delay in Payment to the Contractor

Delay in payment to the contractor is the other factor with a ‘Very high’ severity, aligning with [94,123]. It is commonly the client’s fault [95]. It is suggested that clients should demonstrate the funding for the project in advance [98]. In addition, selecting a financially stable contractor is also advised, as contractors with weak financial standing are more likely to suspend work during payment delays, which may result in significant cost overruns [138].

4.2.7. Delay in Procurement and Delivery of Materials

The study identified delay in procurement and delivery of materials as the seventh ‘very high’ cost risk factor, in agreement with Berihu, Grum, Tariku and Abebe [94]. In many developing countries, construction materials are imported and must be purchased using foreign currency. This makes projects highly vulnerable to exchange rate fluctuations and inflation, and hence, delays in procurement [95]. Supplier defaults also exacerbate delivery challenges [123]. In its extreme cases, extended procurement delay could result in design and material changes during construction [93]. To address this risk, contractors should adopt robust procurement planning strategies [95].

In summary, it can be concluded that cost overrun in the ECI is mainly due to macroeconomic and political factors, which is in agreement with Bedada [139], who concluded that political and economic factors significantly impact the ECI. This could be because the ECI is dependent on imported resources, rendering it vulnerable to these factors, despite significant local resource use. Political instability and lower GDP reduced investment and raised inflation in Ethiopia [140]. To withstand those challenges, organizations should acquire strong stability and experience. However, the organizational capability of Ethiopian contractors remains weak. According to Donka and Mengistu [141], most of the contractors in the ECI exit the industry within a decade of their startup, and the industry is dominated by young companies with small capital and inadequate organizational capability.

Although the level of severity varies according to the specific context of the country, similar reports have been recorded in other developing countries. In Tanzania, for example, macroeconomic variables such as inflation rate, interest rate, labor, and technology are significant reasons for project failure [142]. According to Coleman et al. [143], political factors are among the critical reasons for unsuccessful projects in Ghana. According to a study by Draleti, Sengonzi and Kakitahi [118], the cost of materials, inflation, fraudulent practices, and kickbacks are among the critical causes of cost overrun in Uganda. Similarly, financial difficulty by the client, delays in payments of completed work, and material price fluctuations are among the critical causes of cost overruns in Ghana [95].

In contrast, construction project cost overruns in developed countries are mainly due to internal risk factors. For example, a study by Eliasson [144] evaluated the causes of cost overruns in Swedish transport infrastructure projects constructed from 2004 to 2022 and concluded that the decision-making process is the main cause of cost overruns. According to Steininger et al. [145], scope changes, geological conditions, high risk-taking propensity, and extended implementation are among the critical causes of cost and time overruns in Germany.

Thus, from the above comparisons, it can be concluded that though cost overrun is a global challenge, its extent and nature vary across nations depending on factors such as economic, political, organizational, and technical factors. A study by Ismail, Ramly and Hamid [17] conducted a systematic review of cost overruns in the CI of developed and developing countries and concluded that the effect of cost overruns in the cases of developing countries is more significant due to economic hurdles that often cause financial tightness in the construction projects. CI in developing countries is being challenged by issues such as political instability, a shortage of human resources, and the effect of inflation.

4.3. Implication and Limitation

The suggestion of the mitigation strategies provides practical guidance for tackling the effects of the risk factors with very high severity. The literature classified the possible mitigation strategies to be adopted as risk avoidance, risk reduction, risk sharing, risk transfer, and risk acceptance [15]. Contractors can use one or a combination of these strategies depending on their organizational capabilities and the nature of the risks to be mitigated. For instance, economic risk factors such as inflation and escalation can be mitigated through contractual procedures like the price escalation provision in the contract document (sharing the risks with the client) [36,146]. While material delivery issues can be addressed through proactive risk reduction strategies such as material hedging, early procurement, use of alternative suppliers, and stocking critical materials [147]. Furthermore, the mitigation of macroeconomic and political risks requires government intervention.

This study is significant to both the researchers and practitioners in the CI. The findings of the study highlight that cost overruns in the cases of the ECI are mainly due to factors that are beyond the control of the contractor. The study also suggests possible mitigation strategies in tackling those risks. For researchers, the study significantly contributes to the existing body of knowledge by identifying the critical causes of cost overrun in the ECI. Furthermore, its methodological approach, particularly the sequential SLR, Delphi, and fuzzy methods, offers a clear reference for other researchers conducting similar studies.

Though the study provides valuable insights into understanding the nature of cost overruns in the recent ECI, it is not without limitations. As the study is based on expert judgments, there could be a problem of subjectivity reflecting the experience of the selected experts. Although the Delphi sample is defensible, the Delphi panel may not fully represent contractors, and there could be a problem of generalizability. Furthermore, the study indicated critical risks in the ECI during data collection and does not imply that lower-ranked risks are not significant. Rather, the severity of a risk may vary depending on factors such as project type and economic situation, and hence, a risk that is critical in one situation may not be critical in another situation. Therefore, the findings of the study should not be interpreted as a universal ranking; it is rather a context-specific ranking. The other limitation is that, as the experts are from the ECI, the findings are context-specific and may not be directly generalizable to other countries. The other limitation of this study is it does not explicitly model the interaction among risks. However, these limitations present opportunities for future research, including cross-country comparative studies to enhance the external validity of the results, validation using longitudinal project cost data, and integration of probabilistic modeling techniques such as Monte Carlo simulation for cost overrun forecasting and modeling the interaction among the risks.

5. Conclusions and Recommendations

Cost overrun is among the critical challenges of the CI, where its impact is more significant in developing countries like Ethiopia. However, only limited attempts have been made to critically identify and analyze the risk factors contributing to cost overruns in the ECI. This study utilized a systematic literature review (SLR), a three-round Delphi, and a fuzzy set to fill the gap by assessing the developing countries and the Ethiopian context. The SLR resulted in eighty-one cost risk factors common to the developing countries. The Delphi resulted in fifty-four cost risk factors common to ECI, ranked by using fuzzy logic.

The findings of the SLR indicated that escalation and fluctuation in material prices, design change, poor quantity and cost estimation, variation or change order (addition/omission), and delay in payment to the contractor are the top five cost risk factors in the literature. From the context of the ECI, escalation and fluctuation in material prices, inflation, material shortages, the country’s political instability, the country’s economic instability, delay in payment to the contractor, and delay in procurement and delivery of materials are cost risk factors with a ‘Very High’ severity in the ECI, indicting cost overrun in the ECI is mainly due to the macroeconomic and political factors.

As these risks are predominantly external risks, stakeholders in the ECI should adopt proper mitigation strategies. Clients and contractors should incorporate contractual clauses that support price escalation and inflation-indexed contracts. Furthermore, those risks cannot be managed by the traditional contingency allocation, and hence, they should adopt advanced contingency estimating techniques, such as simulation, if risk acceptance is desired. As a policymaker, the government’s involvement is critical. Some of the activities to be performed by the government are facilitating the availability of foreign exchange for construction material suppliers and importers and promoting stable procurement policies.