Barriers to the Adoption of Modular Construction in Portugal: An Interpretive Structural Modeling Approach

Abstract

Modular construction is the most advanced method of off-site construction available, generating fully-fitted modules with up to 80% of the work completed in a controlled environment prior to their transportation and installation on-site. The adoption of modular construction (AMC) has increased in several countries worldwide. However, in smaller construction markets, the adoption rates remain low, and the industry is still heavily reliant on traditional labor-intensive methods. The main research question for this study is: What are the main barriers (including the root barriers) to the adoption of modular construction in Portugal? The methodology used was a mixed methods research (MMR) approach, trying to understand if there are any specificities in the Portuguese (European Union) markets compared with the more well-documented North American and Asian markets. First, a literature review followed by a survey were used to prioritize a list of 15 critical barriers to the AMC. The results and interrelations between the barriers were analyzed through an interpretive structural modeling (ISM) approach and an impact matrix cross-reference multiplication applied to a classification (MICMAC) analysis. The combined analyses show that the main barriers are low levels of R&D, a lack of accredited organizations to certify the quality of the manufactured components and the industry’s unwillingness to innovate.

1. Introduction

Construction is one of the largest industries worldwide, holding a significant importance to a country’s economy. Construction represents around 13% of the generated gross domestic product worldwide and employs 7% of the working population [1]. However, for the past 100 years, construction methods have remained the same, with the industry displaying a reluctance to innovate and boost productivity, choosing instead to rely heavily on traditional labor-intensive methods [2,3].

Several innovative technologies can face low productivity levels in the construction industry, particularly Off-Site Construction (OSC) methods such as modular construction. The concept of modular construction is not new, but the development of digital design and management tools, the increasing housing demand and labor shortage have revitalized it as a possible solution to the current problems faced in the industry [4].

Modular construction significantly increases productivity, offering the possibility to reduce schedules while also increasing the quality of built components [5]. It can also contribute effectively to reduce waste and, therefore, reduce the significant environmental impact of construction activities [6,7] and to provide lower life cycle costs [8]. By working in an off-site facility, there is less exposure to weather, and higher degrees of standardization and reproducibility becomes achievable while also ensuring a safer and healthier environment for the labor force [9]. This leads to a better predictability over the quality and the costs of the built components [10]. Additionally, this method is more sustainable in controlling, reusing, recycling and disposing of generated waste, providing an environmentally friendlier solution [9].

This innovative method has been reasonably used in Japan, Scandinavia, Singapore, Hong Kong, the UK and the USA [11]. The list of benefits and barriers associated with modular construction depends on the type of market under analysis, particularly on its level of adoption of the method. Markets that have only recently begun to use modular construction will not be able to retrieve the same number of benefits as the markets with high adoption rates. One example is in the delivery of affordable housing, given the possibility to achieve a better control over the costs [12]. Both benefits and barriers are expected to emerge as more local research efforts and practice are put into the method [13,14]. Hence the need for studies to consider a specific context, usually at a country level, as different countries prioritize different aspects of the method, in particular regarding the barriers to its adoption. As discussed by Abdul Nabi and El-adaway [15] there is still a large gap in understanding the main decision factors regarding modular construction, particularly regarding the legal and regulatory issues, and green practices, among others.

The purpose of this paper is to identify the main barriers to the adoption of modular construction (AMC) in the European Union’s construction markets, using Portugal as a case study, and to understand how the barriers are interrelated to develop effective mitigation measures to reduce their impact and ultimately leverage the development and implementation of modular construction. The main research questions are: What are the main barriers to the adoption of AMC, and how are they interrelated?

This paper expands the existing literature in two directions: first, by providing a novel perspective over a European country—there has been an excessive concentration of studies on Asian and US markets; second, by applying an interpretive structural modeling approach and MICMAC analysis to the problem of the identification and hierarchization of the barriers.

From a practical and professional perspective, the findings of this research allow both governments and industries to identify and define a roadmap of actions that have the potential for mitigating the main barriers, thus improving the conditions for the AMC. Although the concept of modular construction can be applied to any type of physical asset, most of this research is more focused on buildings, although the authors do not expect significant differences for other types of assets.

This paper is structured as follows: after this introduction, Section 2 presents the literature review; Section 3 contains the materials and methodology; Section 4 presents the results, followed by the discussion and mitigation measures in Section 5; finally, Section 6 presents the conclusions and policy implications.

2. Literature Review

2.1. Modular Construction

Modular construction is the most advanced method of off-site construction available. This method generates fully-fitted volumetric units, known as modules, which include trim work and the installation of mechanical, electrical and plumbing systems. Up to 80% of the work can be completed in off-site facilities prior to their transportation and installation on-site, as part of the building’s structure [11].

Some researchers refer to modular construction not only as volumetric units but as the combination of any type of pre-assembled components to form a building or part of it. To avoid any confusion in the interpretation of this study, from here onwards, the term module is used to describe the largest volumetric unit (including finishes, fixtures and fittings) transportable from a facility to be later installed on-site [16]; while the term modular construction is used to describe the construction method of designing, manufacturing, transporting and installing the modules on-site to form the building.

The benefits of modular construction in improving some of the construction industry’s bottlenecks, have been thoroughly described in the existing literature. For example, the excessive geometric variability risks in modular components, that pose significant risks in terms of operational effectiveness, and need to increase skills in the construction labor, thus raising concerns by potential adopters.

The adoption rates in most countries, particularly in smaller markets, have been sluggish [17], which has led researchers to study the barriers to the AMC.

2.2. Barriers to the Adoption of Modular Construction

The adoption of modular construction is a disruption in the traditional construction methods, which is essentially centered in the on-site fabrication. By dislocating the fabrication of components and systems, it is necessary to reorganize the value chain. This disruption creates resistance. Several authors have tried to identify what the main barriers to the adoption of modular construction are, using alternatives methods in different geographies. Next, the authors present a short analysis of such studies.

Blismas et al. [18] conducted a survey to identify the barriers to the adoption of OSC methods in the UK, which resulted in the identification of four clusters. The authors analyzed the interrelationships between the clusters using a simple model and concluded that knowledge significantly influenced the remaining clusters considered, which included value, processes and supply chain.

Azhar, Lukkad and Ahmad [19] used mixed methods to study commercial modular buildings in the USA. The authors identified 12 critical decision-making factors and six critical barriers to the AMC, including clients’ reservations, lack of suppliers, decreased flexibility for design changes, the design not supporting modularization and difficulty carrying out on-site modifications.

Mao et al. [2] examined the reasons behind the slow adoption of the OSC methods in China from the perspective of developers. The survey research considered 18 of the 30 barriers as critical to the adoption of OSC, including lack of governmental regulations and incentives, high initial cost, dependence on traditional construction methods, lack of technologies and testing institutes for prefabricated components and high-cost pressure without the economies of scale effect. The identification of the critical barriers was proceeded by a factor analysis that divided the barriers into five clusters.

Hwang, Shan and Looi [11] investigated the barriers to the AMC in Singapore and devised measures to mitigate them. The questionnaire used identified the extensive coordination required, requirement for additional project planning and design efforts, increased transportation and logistics considerations, requirement for early commitment and higher initial cost, as the main barriers. The researchers proposed encouraging close collaborations between project stakeholders during the early stages of the project, using BIM to improve coordination and facilitate communication among stakeholders and offering training courses for project teams and workers to enhance their knowledge and skills, as possible measures to reduce the barriers and increase the usage of the method.

Gan et al. [20] identified the critical barriers associated with OSC in China and determined the interrelationships between the barriers. The results divided the list of barriers into five hierarchical levels and placed the barrier inadequate policies at the bottom of the hierarchical model, meaning it represents the most significant hindrance towards OSC adoption in China. Similar studies using the ISM approach were conducted focusing on the Chinese scenario. Li et al. [21] and Zhao [22] analyzed the interrelationships between the barriers associated with prefabricated buildings. However, none of these studies proposed mitigation measures to tackle the identified barriers.

Choi, Chen and Kim [23] focused on understanding the opportunities and challenges of applying modular construction in dense urban environments, using Hong Kong as a case study. A survey was conducted, followed by interviews with experts to verify the findings. The research identified site access and on-site storage, transportation logistics, distance from off-site factory to on-site location, contractor capability/experience and clients’ tendencies, as the main barriers to the use of modular construction in dense urban areas.

Wuni and Shen [24] conducted a holistic international review, identifying 120 barriers to the AMC. The barriers identified were grouped into eight clusters, and an integrated conceptual framework mapping of the interactions was proposed to identify their interrelationships. Based on the interactions identified, the authors proposed the development of combined measures to reduce the impact of the barriers.

A significant number of studies have analyzed the existing barriers towards the AMC from the perspective of different stakeholders, proving the research commitment put into understanding these barriers and developing mitigation measures to address them. However, most articles consider barriers individually or proceed to group them into clusters. To the best of the authors knowledge, only a few studies in China consider their interrelationships and how they may influence each other [20,22].

Additionally, most available information focuses on Asian, North American and Australian markets, and a single study focuses on the European (UK) market. However, that has been no information on the European Union construction markets in over 17 years. This geographical gap in the body of knowledge should be mitigated with this research, that uses Portugal as a case study.

The European Union corresponds to a market with roughly 450 million people and USD 21.5 trillion, and is the sum of several markets/countries that although they are different, share similarities in terms of culture, institutional setting and legal and regulatory context. Additionally, the European Union is a single market, with extensive interpenetration of construction companies in multiple countries.

3. Material and Methods

According to Abowitz and Toole [25], to correctly study human behavior in construction processes such as leadership, innovation and planning, it is essential to combine the quantitative and qualitative approaches in research design and data collection to improve the validity and reliability of the results. Thus, the mixed methods research (MMR) approach increases the ability to draw trustworthy and compelling conclusions from the empirical research.

In the present study, a MMR approach was adopted (Figure 1), combining the quantitative and qualitative methods [26] and comprising four stages, namely: establishment of the critical barriers to the AMC, ISM, MICMAC analyses and the proposal of the mitigation measure. A quantitative survey was used to gather the opinion of the Portuguese construction industry on the importance of the barriers to define a set of critical barriers to the AMC. These critical barriers were the input to the qualitative methods used in the present study, the ISM, and the MICMAC analysis, which were utilized to determine their hierarchy and relationships. Moreover, focus group meetings (FGMs) were held whenever construction expert opinion was needed to develop the ISM model and the mitigation measures to reduce the impact of barriers towards AMC based on the results of the ISM and MICMAC analysis.

The FGM technique was vital in the operationalizing stages II and IV of the methodology and thus, were presented before the methodology stages. An FGM is an exploratory research method that puts together the qualitative data through group interactions on an issue presented by the moderator [27]. An FGM promotes discussion among a group of experts about their perceptions, opinions, beliefs and attitudes towards a product, concept or theory [28]. This method can be used to increase the information already known on a subject or to study the subject from a different perspective, thus creating new insights.

An FGM should include between four to 12 participants [29]. Following the recommendations of Nassar-McMillan and Borders [30], for the present study, a set of eight experts was selected for the three FGMs, namely, two academics in the field of civil engineering with an average of 15 years of experience, and six practitioners, each with more than 10 years of experience in construction. The practitioners were chosen with equal distribution among the consultants, contractors and clients, two from each group, to minimize the effect of bias. All experts had knowledge of modular construction, but limited practical experience with the technique.

In the present study, three FGMs were carried out between March and April 2021, with intervals of two weeks between them. The first (FGM-1) established the relationships between the barriers, the second (FGM-2) verified the consistency of the ISM model, and the third (FGM-3) helped to define the mitigation measures for the barriers to the AMC. The FGM-1 took about two hours, the FGM-2 45 min, and the FGM-3 one and a half hours. Prior to the FGM-1, the participants received information on the main concepts of modular construction, a description of the critical barriers to the AMC and a brief explanation of the ISM and MICMAC methods for the contextualization. The FGMs were moderated by one of the authors, ensuring a clear knowledge of modular construction management and the barriers to the AMC. The moderator helped reach a wide consensus, stimulated discussion and ensured the discussion proceeded from general to specific topics to promote sincerity and to reduce bias [31]. Inevitably, sometimes there were differences in opinions between experts. In those cases, the moderator followed the principle that “the minority gives way to the majority” [32] to decide.

3.1. Stage I: Establishment of the Critical Barriers to the AMC

Stage I comprises three steps (Figure 1).

Step 1—Identification of the barriers to the AMC

A literature review was carried out to find and analyze the most common and relevant barriers to the AMC. Using a bibliometric analysis, through a Scopus search engine, an extensive keyword search under the “Title, Abstract, Keywords” field was carried out, including “modular construction”, “prefabrication”, “prefabricated”, “off-site construction”, “off-site manufacturing”, “modularization”, “modular integrated construction”, “prefabricated prefinished volumetric construction”; AND “advantages”, “benefits”, “drivers”, “barriers”, “challenges”, “factors”, “hindrances”, “constraints”. A total of 225 articles were reviewed, and based on the most relevant 18 articles, a list of 42 barriers to the AMC was assembled (

Table 1. List of barriers to the AMC.

Code	Description of the Barriers	References
B1	Clients’ misconceptions	[2,17,19,20,23,24,33,34]
B2	Uncertainty over market demand	[2,12,13,17,20,24,35]
B3	Lack of awareness of the benefits	[2,11,19,24]
B4	Uncertainty over quality and performance	[2,5,10,17,20,23,24,34]
B5	Lack of data to evaluate the benefits	[2,10,13,19,24]
B6	Risk of increasing unemployment	[2]
B7	Fragmentation of the supply chain	[2,10,17,24,34]
B8	Industry’s unwillingness to innovate	[2,10,17,19,20,23,24]
B9	Lack of market competition	[2,17,18,24]
B10	Construction industry’s risk aversion	[2,17,24,34]
B11	Low levels of R&D in the industry	[2,24]
B12	Lack of accredited organizations to certify the quality of manufactured components	[2,5,17,18,24]
B13	Lack of governmental regulations (codes and standards)	[2,10,13,20,24,34,36]
B14	Lack of governmental incentives	[2,17,20,24]
B15	Lack of legal framework	[2,24,34]
B16	Shortage of land available for large scale developments	[17,34]
B17	Higher initial costs	[2,11,13,17,24,34,36]
B18	Higher construction costs	[5,11,17,33]
B19	Difficulty achieving economies of scale	[2,5,17,24,34]
B20	Difficulty obtaining financial investment	[17,18,24,33]
B21	Inability to complete design before manufacture	[5,11,17,18,19,23,24,34]
B22	Longer lead-in times during design stage	[2,5,11,17,24]
B23	Inflexibility to apply changes to the design	[2,11,17,19,24]
B24	Need for close collaboration between stakeholders	[11,17,19,20,24]
B25	Lack of experienced designers	[2,5,11,17]
B26	Architectural design does not consider modular construction	[19,37]
B27	Difficulty defining the most suitable projects	[10]
B28	Aesthetic limitations	[2,20,24,33]
B29	Intricate design	[2,24]
B30	Lack of experienced contractors	[2,10,17,20,23,24,36,37]
B31	Lack of experienced collaboration groups	[2,5,10,24]
B32	Lack of suppliers of prefabricated components	[2,5,17,18,19,23,24,34,36]
B33	Transport restrictions	[2,10,36,11,13,17,18,19,23,24,33]
B34	Transportation costs	[10,17,24,33]
B35	Impact of mobilization on the structural integrity of components	[24,33,36]
B36	Additional protection required	[11,24]
B37	Complex logistical management	[11,20,23,24]
B38	Requirement for high-capacity cranes	[5,24]
B39	Restrictions for unloading and storing components	[2,11,18,20,24]
B40	Lack of experienced labor for on-site assembly	[2,11,18,24,36]
B41	Difficulty carrying out on-site modifications	[19,24,33]
B42	Low tolerance between manufactured components and on-site assembly	[3,17,24]

). It is important to note that some of the articles referred to the OSC methods and not specifically to modular construction. However, this is a reasonable consideration as these methods are based on the same premises, thus sharing the same benefits, drivers and barriers.

Step 2—Ranking of the barriers

The second step comprises a survey to assess the importance of the barriers to the AMC and their ranking. A survey is a non-experimental, descriptive research method extensively used to assess the attitudes and characteristics concerning various subjects [38]. An on-line questionnaire was designed and sent out to Portuguese construction industry professionals, based on the list of the barriers (Table 1). A five-point Likert scale was used to rank the importance of the barriers (1—Not at all important, 2—Slightly important, 3—Important, 4—Fairly important, and 5—Extremely important) as suggested by the literature [2,18]. The barriers were selected and adapted to the Portuguese context by the researchers, making it possible to address the concerns about some aspects being “lost in translation”, as expressed by Skitmore and Smyth [39]. Moreover, particular care was paid to avoid ambiguity and vague wording, to ensure consistent meaning for all respondents and increase the validity of the responses. Together with the questionnaire, a cover letter was also prepared to present the aim of the questionnaire and the concepts associated with the modular construction relevant to the present study.

Researchers’ work colleagues involved in the modular construction research and quantitative analysis piloted the initial version of the questionnaire and cover letter to check the questionnaire’s effectiveness in collecting data and detecting possible errors or misunderstandings. Following some minor corrections, the final version of the questionnaire was ready. The final questionnaire was distributed by email. An online questionnaire ensures anonymity, is less intrusive and can reach a broader audience in terms of time and location at a lower cost when compared to other solutions [38]. On the other hand, it is harder to verify the respondents’ eligibility, and response rates tend to be inferior, leading to a non-response bias [40]. However, efforts were made to tackle a non-response bias, for example a follow-up program was implemented to increase the response rate and also to identify the non-respondents to control whether they present characteristics that are different from the respondents, as suggested by Rungtusanatham et al. [41]. Further, all respondents were promised access to the final report generated from the research.

The final questionnaire and cover letter were disseminated to a sample of 325 potential respondents. The sample was produced by considering the type of population (consultants, contractors and clients) and the context to be studied. Different sampling methods were used in this study. Considering the contractors and consultants, 150 of each group were randomly selected from the database of the most prominent Portuguese Association of Civil Construction and Public Works Industry (AICCOPN is the Portuguese acronym), with close to 6000 members. Random sampling is advisable to ensure the sample’s representativeness, mainly when the researchers are interested in generalizing the results [38] However, as for clients, seeing as that was an arduous group to reach, snowball sampling was adopted, resulting in a sample of 25. Snowball sampling is a non-probability sampling method where currently enrolled participants help recruit future respondents for a study [20], in particular in this study in recruiting clients with some experience/knowledge in modular construction.

The dissemination of the survey yielded 90 responses (

Table 2. Survey response rate distribution by respondent group.

Respondent Group	Potential Respondents	Number of Responses	Frequency of Responses (%)	Response Rate (%)	Respondents’ Experience (Years)
					<5	5–10	10–20	>20
Contractors	150	30	33.3	20.0	1	4	7	18
Consultants	150	49	54.4	32.7	2	8	19	20
Clients	25	11	12.2	44.0	0	1	5	5
Total	325	90	100	27.7	3 (3.3%)	13 (14.4%)	31 (34.4%)	43 (47.8%)

). The survey total response rate was 27.7%, suggesting a lower degree or the inexistence of a non-response bias in the responses [42,43]. A Cronbach’s alpha of 0.944, greater than 0.7, with a 5% significance level assured the reliability of the five-point Likert scale adopted [44].

Following this, the relative importance index (RII) was used to rank the importance of each barrier. This index is frequently used in construction management research surveys to analyze respondents’ opinions based on the data collected from the Likert scales, ranking the variables according to their importance [45]. For each barrier, the RII was calculated as shown in Equation (1).

$$RII=\frac{\sum W}{A\times N}$$

(1)

where $W$ is the importance given to each barrier (1 to 5), $A$ is the highest importance given to a barrier (5), and $N$ is the total number of responses. The higher the value of the RII, the higher the importance of the barrier to the AMC. The barriers to the AMC ranked according to their importance are presented in Appendix A.

Step 3—Establishment of the critical barriers

Following the recommendation of Ma et al. [46] and Wu et al. [47], to restrain the effort demanded of the experts in the FGM-1 in the pair-wise comparisons, to determine the relationship between barriers, only the 15 highest-ranking barriers were deemed critical barriers and used in the development of the ISM model (

Table 3. Critical barriers to the AMC.

Code	Critical Barrier
CB1	Lack of awareness of the benefits
CB2	Inability to complete design before the manufacturing stage
CB3	Lack of experienced designers
CB4	Difficulty carrying out on-site modifications
CB5	Uncertainty over quality and performance
CB6	Architectural design does not consider modular construction
CB7	Lack of experienced contractors
CB8	Lack of data to evaluate the benefits
CB9	Industry’s unwillingness to innovate
CB11	Lack of accredited organizations to certify the quality of the manufactured components
CB12	Uncertainty over market demand
CB13	Difficulty defining the most suitable types of project
CB14	Lack of market competition
CB15	Low levels of R&D in the industry

). As the ranking of the importance of the barriers considers the respondents’ overall opinion, which includes consultants, contractors, and clients, it was important to examine the possible differences in the ranking between the respondents’ groups. For that, the Spearman rank correlation tests were performed considering the pairs of respondents’ groups to determine the correlations between their rankings of the barriers. With a statistical significance level of 1% [48], the correlations between the rankings perceived by the respondents’ group were either strong (Spearman’s coefficient of 0.77 between consultants and contractors) or moderate (0.47 between consultants and clients, and 0.54 between contractors and clients). These values indicate an acceptable level of agreement between groups concerning the ranking of the importance of the critical barriers to the AMC.

3.2. Stage II: ISM Model

The ISM approach was first introduced by Warfield [49]. It is a computer-assisted group learning process used to conduct systemic research on the relationships between various variables regarding a specific topic of a particular complex system [50]. It decodes vague mental models into perceptible and well-defined systems [51]. The ISM approach improves the comprehension of the system’s variables by defining their hierarchy and relationships [52]. Moreover, ISM is useful for emerging topics or with few experts [53], as in the case of modular construction.

The main advantages of ISM include incorporating experts’ opinions based on their knowledge and experience, allowing amending opinions and altering the evaluations without demanding too many operations to evaluate the systems with less than 15 variables, making it suitable to analyze real-life conditions [46,54]. Researchers frequently use this method in ordering and decomposing complex relationships between barriers to adopting innovations in construction [21,55,56,57,58,59,60].

In the present study, ISM was used to identify and evaluate the interactions between the (15) critical barriers to the AMC, allowing a graphic and hierarchical representation of the links between them and the identification of the main barriers that must be mitigated. The implementation of ISM has followed five steps well-established in the literature [52,61,62] (Figure 1).

Step 1—Contextual relationships between the critical barriers

The 15 critical barriers gave rise to 105 (15 × 14/2 = 105) different relationships. In the FGM-1, experts were asked to define the contextual relationships between each pair of barriers, according to the following symbology [46]: V—barrier i influences barrier j; A—barrier j influences barrier i; X—barriers i and j influences each other; O—barriers i and j do not influence each other. The direct relationships between the barriers to the AMC were placed in the structural self-intersection matrix (SSIM) (

Table 4. Structural self-intersection matrix.

CB (i/j) 1	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1	-	O	O	O	V	V	O	A	O	O	O	V	V	O	A
2		-	O	O	O	O	O	O	A	O	O	A	O	O	O
3			-	O	V	V	X	O	A	O	O	O	A	O	O
4				-	O	O	O	O	O	A	O	O	O	O	O
5					-	V	A	A	O	O	A	O	O	O	A
6						-	A	A	A	O	O	A	A	A	O
7							-	O	A	V	O	O	A	V	O
8								-	O	O	O	V	O	O	A
9									-	V	X	O	O	V	X
10										-	O	O	O	O	O
11											-	O	O	O	A
12												-	O	O	O
13													-	O	A
14														-	O
15															-

¹ CB (i,j)—Critical barriers in line i, or column j.

).

Step 2—Converting the SSIM into the Final Reachability Matrix

First, the SSIM was transformed into the initial reachability matrix (IRM), a binary matrix representing the direct relationships between the barriers, which was accomplished by replacing the letters with 1 s and 0 s. If the (i,j) entry in the SSIM is: $V$, then the IRM (i,j) entry became 1 and the (j,i) entry became 0; $A$, then the IRM (i,j) entry became 0 and the (j,i) entry became 1; $X$, then the IRM (i,j) and (j,i) entries became 1; $O$, then the IRM (i,j) and (j,i) entries became 0.

Second, the IRM was checked for transitivity, giving way to the FRM. If barrier i influences barrier j and barrier j influences barrier k, then barrier i indirectly influences k through barrier j, and if the entry (i,k) in the IRM is 0, then it must be changed to a 1*. Before calculating the FRM, the matrix IRM_I was obtained by adding the IRM to identity matrix I. The FRM was then obtained through the Boolean operation, which involved self-multiplication of IRM_I until it reached a stable state, as indicated in Equation (2) [54].

IRM_I ≠ IRM_I ² ≠ …≠ IRM_I ^{n − 1} ≠ IRM_I ⁿ = IRM_I ^{n + 1} = FRM

(2)

Step 3—Level partitioning of the barriers to the AMC

In an iterative process, level partitions of the FRM are performed to determine the hierarchy between barriers. For each barrier, the reachability set, the antecedent set, and the intersection set were generated to measure its influence level. The reachability set of barrier i includes all barriers that are influenced by barrier i (represented by 1 s in the row i of the FRM). The antecedent set of barrier i includes all of the barriers that influence barrier i (represented by 1 s in the column i of the FRM). The intersection set includes the barriers that are common to the reachability and antecedent sets. When the intersection and reachability sets of a particular barrier are equal, that barrier is allocated to the current iteration level. Finally, the barriers assigned to a given level are separated from the remaining sets of reachability and antecedents for the next iteration until all of the barriers are partitioned into their respective hierarchical levels.

Step 4—Development of the ISM model

The ISM model was built by drawing a diagraph based on the FRM and the hierarchical level of each barrier. First, the conical matrix of the FRM was built by clustering the barriers in the same hierarchical level across rows and columns of the matrix to assist in the set-up of the ISM model. Second, a preliminary diagraph was drawn up by placing the barriers vertically according to the level partitioning and linking the barriers using arrows according to the conical matrix. Third, the indirect links between the barriers were removed to obtain the ISM model.

Step 5—Consistency check

Finally, the FGM-2 experts were asked to check for the conceptual consistency in the hierarchical structure and the interrelationships of the barriers to AMC of the obtained ISM model, and needed corrections, if required, were made. The experts were instructed to check if there were ambiguities in the ISM model and if it correctly represented the “vague” mental model they had of the system of barriers affecting the AMC.

3.3. Stage III: MICMAC Analysis

The MICMAC analysis was developed by Duperrin and Godet (1973) based on the multiplication properties of the matrices. In the present study, it is used to classify and better understand the list of barriers to the AMC according to their driving and dependence power [46]. The driving power represents the capacity that a barrier has to influence the other barriers and the dependence power represents the capacity to which other barriers influence it [32]. Every 1, or 1* in the entry (i,j) of the FRM indicates that barrier i influences barrier j, and vice-versa, i.e., barrier j is influenced by barrier i. Therefore, the barriers’ dependence power is calculated by the sum of each column and the barriers’ driving power by the sum of each row. A driving-dependence power graph was built, and according to the barriers driving and dependence power scores, they are positioned into one of the four clusters of the MICMAC analysis, namely: independent (strong driving power and weak dependence power), linkage (strong driving power and strong dependence power), autonomous (weak driving power and weak dependence power), and dependent (weak driving power and strong dependence power). As Zaidi et al. [63] suggested, half of the maximum driving or dependence power was used to set the thresholds between weak and strong power.

Therefore, the MICMAC analysis complemented the hierarchical representation of the links between barriers provided by the ISM model, increasing the methodological rigor [64].

3.4. Stage IV: Proposal of the Mitigation Measures

The mitigation measures for the barriers to the AMC were developed with the help of experts during the FGM-3. First, the experts analyzed and interpreted the hierarchical structure and relationships between the barriers of the ISM model and the barriers’ location in the MICMAC analysis clusters. Second, the experts were demanded to establish mitigation measures for the main barriers. Third, although these measures aim to act on the main barriers, they must reach and mitigate the other barriers by accounting for the hierarchical relationships between them and their driving and dependence power. Thus, the experts were instructed to verify whether all of the barriers were mitigated by some of the measures put forward in the previous step, particularly in the case of the causes within the autonomous cluster. If not, the experts should develop additional measures to mitigate those unmitigated barriers.

Once these measures were established according to the hierarchical structure observed and the relationships between the barriers and their dependences and driving power, and were based on the experts’ professional experience, these measures are expected to be practical and effectively mitigate the barriers to the AMC.

4. Results

4.1. ISM Model

The ISM model was developed following steps 2 to 5 of the stage II of the methodology (Figure 1). In step 2, the SSIM was initially converted into the IRM (

Table 5. Initial Reachability Matrix.

CB (i/j) 1	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
1	1	0	0	0	1	1	0	0	0	0	0	1	1	0	0
2	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0
3	0	0	1	0	1	1	1	0	0	0	0	0	0	0	0
4	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0
5	0	0	0	0	1	1	0	0	0	0	0	0	0	0	0
6	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0
7	0	0	1	0	1	1	1	0	0	1	0	0	0	1	0
8	1	0	0	0	1	1	0	1	0	0	0	1	0	0	0
9	0	1	1	0	0	1	1	0	1	1	1	0	0	1	1
10	0	0	0	1	0	0	0	0	0	1	0	0	0	0	0
11	0	0	0	0	1	0	0	0	1	0	1	0	0	0	0
12	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0
13	0	0	1	0	0	1	1	0	0	0	0	0	1	0	0
14	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0
15	1	0	0	0	1	0	0	1	1	0	1	0	1	0	1

¹ CB (i,j)—Critical barriers in line i, or column j.

). Then, in step 3, the IRM was checked for transitivity, originating the FRM (

Table 6. Final Reachability Matrix.

CB (i/j) 1	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	DVP 3
1	1	1*	1*	1*	1	1	1*	0	0	1*	0	1	1	1*	0	11
2	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	1
3	0	0	1	1*	1	1	1	0	0	1*	0	0	0	1*	0	7
4	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	1
5	0	0	0	0	1	1	0	0	0	0	0	0	0	0	0	2
6	0	0	0	0	0	1	0	0	0	0	0	0	0	0	0	1
7	0	0	1	1*	1	1	1	0	0	1	0	0	0	1	0	7
8	1	1*	1*	1*	1	1	1*	1	0	1*	0	1	1*	1*	0	12
9	1*	1	1	1*	1*	1	1	1*	1	1	1	1*	1*	1	1	15
10	0	0	0	1	0	0	0	0	0	1	0	0	0	0	0	2
11	1*	1*	1*	1*	1	1*	1*	1*	1	1*	1	1*	1*	1*	1*	15
12	0	1	0	0	0	1	0	0	0	0	0	1	0	0	0	3
13	0	0	1	1*	1*	1	1	0	0	1*	0	0	1	1*	0	8
14	0	0	0	0	0	1	0	0	0	0	0	0	0	1	0	2
15	1	1*	1*	1*	1	1*	1*	1	1	1*	1	1*	1	1*	1	15
DEP2	5	7	8	10	9	12	8	4	3	9	3	6	6	9	3

¹ CB (i,j)—Critical barriers in line i, or column j; ² DEP—Dependence; ³ DVP—Driving Power; 1*—Transitive relationships.

) with the help of a Microsoft Excel VBA program. The driving and dependence power of each barrier was also determined, the former representing the sum of the respective row, and the latter the sum of the respective column in the FRM.

In the step 3, the process of partitioning the levels of barriers to the AMC resulted in seven iterations, and the corresponding seven hierarchical levels, as shown in

Table 7. Level partitioning.

Barriers	Reachability Set	Antecedent Set	Intersection Set	Iteration/Level
CB2	CB2	CB: 1, 2, 8, 9, 11, 12, 15	CB2	I
CB4	CB4	CB: 1, 3, 4, 7, 8, 9, 10, 11, 13, 15	CB4	I
CB6	CB6	CB: 1, 3, 5, 6, 7, 8, 9, 11, 12, 13, 14, 15	CB6	I
CB5	CB5	CB: 1, 3, 5, 7, 8, 9, 11, 13, 15	CB5	II
CB10	CB10	CB: 1, 3, 7, 8, 9, 10, 11, 13, 15	CB10	II
CB12	CB12	CB: 1, 8, 9, 11, 12, 15	CB12	II
CB14	CB14	CB: 1, 3, 7, 8, 9, 11, 13, 14, 15	CB14	II
CB3	CB: 3, 7	CB: 1, 3, 7, 8, 9, 11, 13, 15	CB: 3, 7	III
CB7	CB: 3, 7	CB: 1, 3, 7, 8, 9, 11, 13, 15	CB: 3, 7	III
CB13	CB13	CB: 1, 8, 9, 11, 13, 15	CB13	IV
CB1	CB1	CB: 1, 8, 9, 11, 15	CB1	V
CB8	CB8	CB: 8, 9, 11, 15	CB8	VI
CB9	CB: 9, 11, 15	CB: 9, 11, 15	CB: 9, 11, 15	VII
CB11	CB: 9, 11, 15	CB: 9, 11, 15	CB: 9, 11, 15	VII
CB15	CB: 9, 11, 15	CB: 9, 11, 15	CB: 9, 11, 15	VII

.

In step 4, the ISM model was established with the help of the conical matrix of the FRM (Appendix B) and is presented in Figure 2. Finally, in step 5, this model was discussed in the FGM-2, where experts were asked to check for possible inconsistencies. The experts agree regarding the consistency of the ISM model. Thus the model was deemed adequate, highlighting both the hierarchical structure and the interrelationships of the barriers to the AMC.

Lower-level barriers are placed at the top of the model, Level I, which includes the inability to complete the design before manufacture (CB2), the architectural design does not consider modular construction (CB6) and difficulty carrying out on-site modifications (CB4). These barriers are not expected to excerpt influence over others but are heavily influenced by other barriers.

Levels II to VI are considered intermediate levels in the model. Level II includes the uncertainty over quality and performance (CB5), lack of experienced labor for on-site assembly (CB10), uncertainty over market demand (CB12) and lack of market competition (CB14). Level III comprises lack of experienced designers (CB3) and lack of experienced contractors (CB7), and Level IV includes difficulty defining the most suitable projects (CB13). Levels V and VI include a barrier each, lack of awareness of the benefits (CB1), and lack of data to evaluate the benefits (CB8), respectively. The barriers assigned to intermediate levels influence barriers in lower hierarchal levels and are also influenced by barriers in higher hierarchical levels.

Finally, the high-level barriers are positioned at the bottom of the ISM model and considered the main barriers to the AMC in Portugal. Level VII comprises the industry’s unwillingness to innovate (CB9), lack of accredited organizations to certify the quality of the manufactured components (CB11) and low levels of R&D in the industry (CB15). These barriers exert an influence over all of the remaining barriers of the ISM model.

4.2. MICMAC Analysis

To further evaluate the main barriers to the AMC, following stage III of the methodology (Figure 1), a MICMAC analysis was developed using the barriers’ driving and dependence power (Table 6). The result is a driving-dependence power graph where the barriers are assigned to one of four clusters (Figure 3). As the maximum driving or dependence power is 15, 7.5 was used as the threshold between weak and strong power. The linkage cluster has no barrier assigned. This occurrence is not new to the literature. In Gan et al. [20], no barrier was assigned to the linkage cluster; and in Li et al. [21] no barrier was assigned to the linkage or autonomous clusters.

The autonomous cluster comprises only two barriers, the inability to complete design before manufacture (CB2) and uncertainty over market demand (CB12). These barriers present a weak driving and weak dependence power, meaning they are somewhat detached from the rest of the system.

The dependent cluster includes seven barriers, namely, lack of experienced designers (CB3), difficulty carrying out on-site modifications (CB4), uncertainty over quality and performance (CB5), architectural design does not consider modular construction (CB6), lack of experienced contractors (CB7), lack of experienced labor for on-site assembly (CB10) and lack of market competition (CB14). These barriers are characterized by a weak driving and a strong dependence power. Thus, these barriers are strongly influenced, but their influence on the other barriers is negligible.

Lastly, the independent cluster comprises la ack of awareness of the benefits (CB1), lack of data to evaluate the benefits (CB8), industry’s unwillingness to innovate (CB9), lack of accredited organizations to certify the quality of manufactured components (CB11), difficulty defining the most suitable projects (CB13) and low levels of R&D in the industry (CB15). The barriers in this cluster have a strong driving and weak dependence power; thus, they influence most of the others but are almost unaffected by any of them. According to the MICMAC analysis, these barriers are deemed the main barriers to the AMC in Portugal.

5. Discussion and Mitigation Measures

This section discusses the ISM model and the MICMAC analysis results and proposes mitigation measures for the barriers to the AMC (stage IV of the methodology). For this, the contribution of the experts in the FGM-3 is crucial to identify and design effective mitigation measures.

5.1. Discussion

Integrating the barriers’ hierarchal structure and interrelationships (ISM model) with their driving and dependence powers (MICMAC analysis) allows leveraging the discussion of the importance of the barriers and their relationships in the AMC. Thus, the main barriers from the MICMAC analysis (independent cluster) and the main barriers from the ISM model (barriers at level VII) are from now on designated as root barriers to the AMC in Portugal (respectively barriers CB1, CB8, CB9, CB11, CB15 and CB13). Comparatively, with the other critical barriers, the root barriers have a relatively low rank in terms of the perceived importance by survey respondents (Appendix A), except in the case of CB1. This result confirms the need to go further than the ranking of the barriers when the aim is to tailor mitigating measures to reduce their impact and promote modular construction. For example, CB4 is ranked as the fourth most important barrier, but it is at the lowest level of the ISM model and is positioned in the dependent cluster. This indicates that the mitigation of CB4 depends significantly on measures to mitigate other barriers, which may be perceived as less important but better positioned in the ISM model and the MICMAC analysis, as in the case of the root barriers. Measures to mitigate root barriers are thus more effective than those based on importance.

None of the barriers of Level VII, low levels of R&D (CB15), industry’s unwillingness to innovate (CB9) and lack of organizations to certify quality (CB11) are related to modular construction itself but rather to existing problems within the industry, as suggested by Ferdous et al. [36]. As insufficient R&D efforts are put into construction innovations, a lack of knowledge and understanding is generated, and the adoption rates remain low. As verified by Wuni and Shen [24], the countries where modular construction holds higher usage levels coincide with countries where more research is undertaken. The low levels of R&D allow the development of codes and standards for modular construction, affecting the number of organizations capable of certifying the quality of modular components. In fact, the creation of adequate design codes, specifications, regulations and performance management systems have been identified as critical success factors in modular construction projects [65]. The industry’s unwillingness to innovate (CB9) is related to construction companies’ risk-averse culture, the industry’s fragmentation, and the lack of government incentives to adopt innovative methods, which promote reliance on traditional labor-intensive methods. These conditions hinder the implementation of modular construction, which requires trained labor and close communication and coordination between different stakeholders throughout the duration of the project to ensure a successful application. As traditional methods are preferred, there is less interest and financial incentives for labor training, implementing digital tools to improve communication between stakeholders and generally to pursue research into modular construction or other forms of innovation [46]. The lack of quality certification (C11) leads to uncertainty over the quality and performance of modular construction, reducing the likelihood of investments from the industry. However, the opposite scenario also takes place, as the industry is not interested in innovating, fewer investments are put into developing accreditation for new construction methods and consequently, only a reduced number of organizations can certify them.

The lack of data to evaluate the benefits (CB8) and lack of awareness of the benefits (CB1), in Levels VI and V, respectively, are related to the market and the public’s perception and knowledge of modular construction and result from the small number of modular construction projects successfully implemented in Portugal. Gan et al. [20] attained similar results regarding the lack of knowledge, concluding that it influenced lack of market competition, uncertainty over market demand and uncertainty over quality. The lack of available data reduces the possibility of further investigating and extrapolating conclusions on the benefits of modular construction. Consequently, the industry and clients are less aware of the existing benefits. Considering the disruptive nature of modular construction compared to traditional methods, unless there is sufficient evidence to validate the benefits, shareholders will not be willing to invest, often considering modular construction as more expensive. This implies the need for evaluation systems explicitly developed for modular construction to emphasize the added value and ensure reasonable comparisons to traditional methods.

Even though the difficulty defining the most suitable projects (CB13), in Level IV, is considered an independent barrier, it has a greater dependence than the remaining barriers in the cluster. This barrier is directly influenced by the lack of awareness of the benefits (CB1). There have been substantial research efforts placed in defining decision-making factors to compute which type of construction projects would benefit the most from modular construction [4,19]. However, most models tend to be designed for specific purposes rather than developing a comprehensive decision-making model that considers all project requirements. Currently, decisions are based on anecdotal evidence or practitioners’ experience rather than on rigorous data because stakeholders do not have a comprehensive knowledge of the benefits. This could reduce the achievable benefits and increase the risk of poor implementation of modular construction. In addition to developing decision-making models, it is essential to promote an early collaboration between different stakeholders and decisions must be made early on to ensure the design includes all of the requirements to accommodate modular construction.

The barriers concerning the uncertainty over market demand (CB12) and inability to complete design before manufacture (CB2) in Levels II and I, respectively, are in the autonomous cluster. These barriers have little influence over others and are also not significantly influenced by others. Thus, they may require dedicated mitigation measures. However, the ISM model shows that the inability to complete the design before manufacture (CB2) is influenced by the uncertainty over market demand (CB12); and uncertainty over market demand (CB12) is influenced by the lack of awareness of the benefits (CB1), placed in Level V. The need for an early collaboration and an early design freeze highlights the fragmentation of the construction’s supply chain, which is a transversal problem and not a specific barrier preventing modular construction. However, the ranking in the survey of CB2, as the second-highest critical barrier, demonstrates the Portuguese construction industry’s conservative mindset and reluctance to change and improve current practices. Project stakeholders are reluctant to completely define the design for fear of market changes and consequent market value losses on modular buildings for not complying with demand. Nevertheless, modular designs have different requirements from traditional ones, and the scope must be defined in advance, otherwise the benefits of modular construction might be lost. Finally, as above mentioned, the lack of awareness of the benefits (CB1) influences the uncertainty over market demand (CB12). Developers hold the perception that the final client still questions the feasibility of modular construction and its performance over the building’s life cycle. Consequently, if few developers consider modular construction, the output and cost of modular construction are affected [2].

The dependent cluster includes the barriers ranging from Level I to III (CB3, CB4, CB5, CB6, CB7, CB10, CB14). The Level III barriers lack of experienced designers (CB3) and contractors (CB7) are set apart from the remainder in the dependent cluster, presenting a higher driving and lower dependence power. Therefore, these barriers can be considered as a link between the independent and dependent clusters. These barriers influence each other, seeing as if designers are not experienced with designing modular buildings and do not propose this possibility in designs, then contractors will not be driven to invest in modular construction. The opposite scenario is also a possibility, as the lack of experienced contractors inhibits designers from proposing modular designs for fear of their execution. Both are influenced by the difficulty defining the most suitable projects (CB13), which hinders both designers and contractors from investing in the method, leading to a generalized lack of experience.

The barriers concerning uncertainty over quality and performance (CB5), lack of market competition (CB14) and the lack of experienced labor for on-site assembly (CB10), in Level II, have the same driving and dependence power and are influenced by the lack of experience barriers (CB3 and CB7). Regarding the uncertainty over quality and performance (CB5), a lack of knowledge about modular construction leads designers and contractors to question the method’s efficiency. Additionally, the less successful applications of modular construction in the past are also often cited in the literature as a barrier, from the perspective of designers and contractors for fear of its inability to meet clients’ expectations, according to Pan, Lee, and Chen [66]. The lack of market competition (CB14) refers to inexperienced stakeholders, particularly contractors, which leads to a lack of internal market competition between modular manufacturers. This barrier is also related to cost concerns as the small demand for modular construction forces the existing manufacturers to increase their prices to cover the high investments and make profits. This is particularly disadvantageous from the point of view of SMEs that do not have the financial capacity to embark on large investments required for modular construction. The higher costs associated with modular construction result in higher bidding prices. In Portugal, it is common practice to choose the lowest cost option over best value [67], often disabling modular manufacturers from competing against the lower bidding prices proposed by traditional contractors. As for the lack of experienced labor for on-site assembly (CB10), because most contracting companies do not have experience handling modular construction projects, no investment is made in labor training to handle the on-site assembly stage.

The difficulty carrying out on-site modifications (CB4) and architectural design does not consider modular construction (CB6), positioned in Level I, are at the bottom of the hierarchal model and present a higher dependence power than other barriers. The difficulty carrying out on-site modifications (CB4) is directly influenced by the lack of experienced labor for on-site assembly (CB10). It has been established that an early design freeze is necessary to start the manufacturing stage, and up to 80% of the works can be completed off-site, significantly reducing the possibility of applying later changes to the designs. Nevertheless, to perform changes, the labor assigned for on-site assembly must be trained to ensure the changes do not impact the building’s structural stability.

Architectural design does not consider modular construction (CB6) is influenced by the uncertainty over quality and performance (CB5) and market demand (CB12), and by the lack of market competition (CB14). Even with the documented benefits associated with modular construction, there is still a lack of awareness and apprehension from the stakeholders over the method’s accomplishments. This apprehension extends to market demand with clients (mainly developers), showing concerns over the potential market value of modular buildings. If architects do not propose this construction method by themselves or by incentives from clients, it is improbable that the method will be implemented in subsequent stages of the project. Regarding the influence of market competition, the current number of existing manufacturers of modular components is only a small fraction of the industry, and not all modular manufacturers have the organizational structure to take on large scale projects, nor are they available at a local level, hindering architects from considering modular construction in their designs.

5.2. Mitigation Measures

Following the discussion of the results, experts were asked to propose measures to mitigate the root barriers to the AMC. Although the aim of these measures is to act upon the root barriers, they must be designed taking into consideration the remaining barriers in the hierarchical structure, foreseeing to mitigate them. Upon definition of the mitigation measures, a verification stage took place to ensure the list of measures addressed all of the barriers, otherwise further consideration would be required. A list of 12 mitigation measures were devised based on experts’ opinions, the hierarchical structure and the interrelationships between the barriers. The list includes all of the barriers which would benefit directly or indirectly from the implementation of each measure (

Table 8. List of mitigation measures for the barriers to AMC in Portugal.

Code	Description of Mitigation Measures	Root Barriers	Autonomous Barriers	Remaining Barriers Indirectly Mitigated
M1	Increasing R&D levels through governmental incentives	CB15
M2	Improving quality certification of modular components with assessment tools and guidelines for inspections (on- and off-site)	CB11		CB: 1, 5, 8, 12, 14
M3	Developing holistic, value-based evaluation methods and standard methods of data collection for modular construction	CB:9, 11		CB: 1, 5, 6, 8, 13, 14
M4	Developing codes and standards for modular construction	CB11		CB: 3, 6, 7, 14
M5	Promoting the use of BIM to improve communication and coordination among project stakeholders	CB9		CB: 2, 13
M6	Offering public training courses for project teams and workers to enhance their knowledge and skills	CB9		CB: 1, 2, 3, 5, 6, 7, 10, 13, 14
M7	Promoting the use of modular construction in public projects	CB9		CB: 3, 5, 6, 7, 8, 12, 13, 14
M8	Increasing awareness on modular construction with assessments from research institutions	CB9		CB: 1, 3, 5, 6, 7, 9, 12, 13, 14
M9	Enhancing decision-making processes between modular and traditional construction with holistic, value-based comparisons	CB9		CB: 6, 13
M10	Encouraging early on collaborations between project stakeholders			CB: 2, 3, 4, 6, 7, 13
M11	Increasing confidence in modular practitioners through quality certification		CB12	CB: 5, 6, 14
M12	Promoting early freeze and optimization of the design		CB2	CB4

).

A total of 10 mitigation measures were proposed to address the root barriers. Given their significant driving power, their mitigation should also affect the remaining dependent barriers, as suggested by Shen et al. [32]. The autonomous barriers benefit from some of the measures proposed for the independent barriers. Nevertheless, given their autonomy from the rest of the barriers in the ISM model, two additional measures were devised, which are also expected to positively impact the dependent barriers.

6. Conclusions and Policy Implications

Modular construction has a disruptive nature from traditional construction methods, and it defies the current way construction projects are planned, procured, delivered and managed. Construction stakeholders must be actively involved in the several stages of the projects to ensure its success, from design to construction, operation and maintenance stages. Modular construction can be a viable solution to increase productivity in construction as well as other problems faced in the industry.

To study the reasons behind the low levels of usage of modular construction in European Union construction markets (using Portugal as a case study), an MMR was conducted. First, a questionnaire was sent out to the Portuguese construction industry to assess the importance of the barriers to the AMC. Based on the results, 15 items were deemed as critical barriers. The ISM approach was used to study the interrelationships between barriers and the MICMAC analysis determined their power scores.

The list of critical barriers attained is representative of the barriers found in the existing literature. However, it is difficult to establish direct comparisons between the results as they depend heavily on the characteristics of the construction industry of the country under analysis. The main differences identified between the results attained and the literature is the lack of cost-related barriers in the list of critical barriers.

The differences between the quantitative and qualitative results (i.e., low levels of R&D in the industry was the last ranking critical barrier but represents one the biggest hinderances to the AMC) highlights the importance of using a MMR to define effective mitigation measures to tackle the barriers to the AMC in European construction markets. From the MMR, it is possible to conclude that the root barriers are the lack of organizations to certify the quality of built components, the low levels of R&D, the industry’s unwillingness to innovate, lack of data to evaluate the benefits, lack of awareness of the benefits, and difficulty defining suitable types of projects.

The combination of these barriers hinders its adoption. The results show that from a public policy perspective, in order to foster the development of AMC, the focus should be to increase the levels of R&D conducted in the industry, the development of value-based evaluation methods and comparison systems to traditional methods, and promote the use of digital tools (e.g., BIM) to improve communication and coordination between project’s stakeholders. This could be achieved by a set of public policies, such as training courses (mostly promoted by research institutions), improving and stimulating the certification of modular components and integrate modular thinking in the curriculum structure of engineering degrees, among others.

A common denominator to the majority of measures proposed is governmental intervention. It seems clear that there is a need, at least in Portugal, to overcome the bottlenecks in modular construction adoption, to develop a set of public policies able to leverage the private sector’s willingness to consider and apply modular construction schemes.

The main contributions of this research are two-folded: it provides a new perspective over a European Country (Portugal) expanding the literature from Asian and US based markets, and, second, by applying an interpretive structural modeling approach and MICMAC analysis to the problem of identification and hierarchization of barriers, providing specific mitigation measures. These mitigation measures can be used by governments and policy makers in order to define a roadmap of actions that can help to unlock the potential of AMC.

The findings of this research have been validated and contribute to the body of knowledge of modular construction as there is no investigation on European countries, with the large majority of studies focusing on Asia and the US. Furthermore, these findings can enlighten the industry on the benefits, drivers and current barriers to the adoption of modular construction, so that mitigation measures can be implemented to increase the usage of the method.

Nevertheless, some limitations have been identified. First, the sample population of the questionnaire was small and the results of the three respondents’ groups were generalized. Second, only the 15 critical barriers were considered in the focus group rather than the ranking of 42 barriers initially used in the questionnaire. The barriers’ ranking did not always reflect the hierarchical position of the barriers in the ISM model or their power scores in the MICMAC analysis. Therefore, considering a higher number of barriers in the qualitative analyses could lead to the identification of other lower-ranking barriers that possess strong driving powers and should be placed at the bottom of the hierarchy, representing significant barriers to the AMC in Portugal. Lastly, the MMR findings apply to the Portuguese construction industry’s context and may not reflect the existing barriers in other European markets with different characteristics regarding the construction industry.

For future research, it is necessary to develop a detailed plan to implement the proposed mitigation measures and to evaluate their efficiency in reducing the barriers to the AMC.