Figure 3 shows the annual publication statistics for research in the domain of RAS in PPPs. There is a clear increase in research interest from the year 2010 however, the number of annual publications does not sustain a regular trend. This agrees only partially with the trends observed in general PPP research as determined by Zhang et al. (2016). In the first 5 years of the review period (2000–2005), research studies only came out of China, Hong Kong, Singapore, India and the United Kingdom with the highest number of studies published in the construction management and economics (CME) journal. In the next 5 years (2006–2010), most of the studies were published in the international journal of project management (IJPM) with new countries that started showing interest in this research domain including Indonesia, Spain, Australia, and Greece, whereas China, the United Kingdom, and Singapore continued to work in this period. Other countries that jumped in during the next 5 years (2011–2015) period include Iran, the United States, Thailand, Italy, Malaysia, Ireland, and Taiwan. IJPM continued to publish the largest number of studies in this period. Ghana, Vietnam, Turkey, Portugal, the UAE, Colombia, and Egypt appeared as new contenders in the next 5 years period (2016–2020). The journal of construction engineering and management (JCEM) and the sustainability (Switzerland) journal published the largest number of studies in this period. Nepal was the only new country that saw a relevant paper coming out in the year 2023. Engineering, Construction, and Architectural Management (ECAM) journal published the largest number of studies in this period. It is understandable that as time goes on, governments across the world are finding interest in this mode of procurement and therefore the increase in interest to study the critical issue of RAS is also understandable.

The reviewed papers were extracted from 20 journals in total. As evident from the information in Table 1, the six journals that published 62% of all the papers reviewed include the JCEM, IJPM, Construction Management and Economics (CME), ECAM, Sustainability (Switzerland), and the International Journal of Strategic Property Management. This result is also partially supported by the findings of previous more general reviews on PPPs (Tang et al., 2010; Cui et al., 2018).

The United Kingdom and Australia are considered to be the leaders when it comes to promotion of the PPP mode of infrastructure procurement (Table 2). However, China contributed the most by accounting for 26% of total reviewed studies in this research, followed by Australia and the United Kingdom. It should also be noted that the statistics shown in Table 2 were developed by considering the country of institutional affiliation of the first author only. This assumption is based on the logic that the position of the author in a publication is a direct indicator of the level of research contribution, thus justifying the country of association of the lead author to be the country of origin of research. Co-authors may contribute in various ways in any research and could be multinational thus their countries of association were not considered in the research origin analysis. On the contrary, cases were observed where the lead author in a paper is affiliated with one country whereas the paper reports data on PPP RAS by covering the case of another country. Examples include the research by Mazher et al. (2019) and Almarri et al. (2019), where the lead authors show their institutional affiliations with Hong Kong and the UAE while presenting data and analysis on PPP projects in Pakistan and Saudi Arabia, respectively. Table 2 also shows how so many other countries where PPP projects are implemented remain under-represented in research in PPP RAS.

It is common in research on RAS to focus on a particular infrastructure sector context (either economic or social infrastructure). This is essentially true because considerations for RAS vary significantly from one infrastructure sector to another (e.g., transport and power sectors), from one subsector to another (e.g., power plants and transmission grid) and from one technology to another (e.g., hydropower plant and solar power plant) (Mazher et al., 2019). In the review sample, 29 research papers did not specify any specific infrastructure sector whereas all the rest focused on one particular sector (Table 3). The transport sector seems to be the focus of most research, followed by the water and power sectors. This trend is clearly visible from the research coming out of China and Australia where more papers deal with transport infrastructure than water infrastructure projects. Other infrastructure sectors saw a low interest in RAS research in PPPs. According to the private participation in infrastructure (PPI) database by the World Bank, investment statistics up to the year 2018 show that the power and transport infrastructure sectors account for the largest share of global private investment in public infrastructure in low- and middle-income countries (Mazher et al., 2018). A troubling finding here is that the social infrastructure sector which includes prisons, hospitals, student hostels, etc. are largely ignored by the extant research. Moreover, even in the domain of economic infrastructure, which includes highways, water supply systems, power plants etc., many sectors and subsectors are under-represented by the existing research. Energy infrastructure for natural gas systems, waste collection and management systems for municipal solid waste, solar power systems, are a few examples of the infrastructure sectors and subsectors that need attention in RAS research.

Most of the research reviewed adopted mixed methods for research whereby several data collection and analysis techniques were employed in a single research study. More research methods are better to cross validate the research findings, which makes the study more rigorous and convincing (Zhang et al., 2016). Through content analysis method, the research methods of each paper are identified and counted, as shown in Table 4. Case study method appeared to be the most popular method of all the methods adopted. Out of the 80 reviewed papers, 42 papers adopted this method to either discuss RAS schemes on projects or to use the case study data for testing and validation of a proposed model. Case study method is the best method to explore the intricacies of proportioning risks between the pubic and the private sector. Interviews with project experts and review of project documents are some of the data collection methods used in case studies that lead to discovery of specific details on a project. Literature reviews (28), questionnaire survey (24), and interviews (12) are other popular data collection methods that were adopted by researchers. Literature reviews are mostly adopted to position research in the existing knowledge landscape. Questionnaire surveys are more popular than interviews as they serve as a versatile medium to obtain large amounts of data in a short amount of time and with limited resources. Monte Carlo Simulation (MCS) simulation and NPV modeling techniques have also been popularly adopted to understand the influence of various risk allocation and sharing scenarios and the effect of risk mitigation strategies on project revenues and VfM. Various MCDM techniques have been used in research to propose RAS models that provide a structured decision-making methodology which avoids weaknesses of human qualitative judgement.

It was also found from the review exercise that most of the research (37.5%) used primary data for modeling and analysis whereas only 7.5% of the reviewed articles adopted secondary data (Table 5). For example, Kukah et al. (2024) used Delphi surveys to propose a fuzzy quantitative risk allocation model for Ghanian PPP power projects (a typical case where primary data was used) whereas Xu et al. (2014) collected 10 PPP project cases from journal papers, doctoral/master thesis, and news reports from official media, in identifying how government guarantees influence the success of PPP projects (a typical case where secondary data was employed). 15% of the papers reviewed used both primary and secondary data whereas the remaining 40% of the papers did not mention the type of data collected.

Table 6 shows the sub-themes of risk allocation and sharing that were established to conduct a detailed systematic review of research in RAS in PPPs. To better understand the basis of this proposed categorical framework for this review, a general RAS decision framework is proposed in Figure 4. It can be clearly seen that risks are to be proportioned based on the capabilities of the public and the private sector to shoulder the risk individually or together in a shared mode, while always ensuring that the arrangement leads to a commercially feasible and a bankable project that also ensures VfM for the public. This exercise of determining and assessing the capability of a party (public and private sectors) is complex as it is based on stakeholder characteristics, project characteristics, and a project’s environmental characteristics, all of which could be uncertain and hard to estimate. Therefore, a lot of research efforts have been exerted to determine how risks should be ideally proportioned. This led researchers to study RAS preferences of PPP stakeholders. In other cases, the researchers adopted a different approach whereby they leaned on various modeling approaches such as artificial intelligence (AI), MCDM, Game theory, Real options, MCS, NPV, etc., to quantify the capability of the stakeholders and to analyze project characteristics to properly apportion risks. While transferring certain risks to the private sector, and in an effort to attract private investment, the governments provide various types of support including guarantees which are essentially a risk allocation and sharing exercise. In some instances, during a PPP project’s operational lifecycle phase, uncertainties may lead to situations where the project stakeholders are forced to renegotiate the project’s original risk allocation and structuring leading to undetermined effects on VfM. Therefore, the sub-themes proposed in Table 6 condenses all similar research and summarizes developments in five proposed sub-themes.

A large number of researchers with interest in the subject of RAS in PPPs/PFIs have sought to explore perceptions of experts on fair and equitable allocation of risks using surveys. It is imperative to understand that risks and their allocation are heavily influenced by the context in which a particular project is executed. Preferences for allocation and sharing of risks vary by geographical location of a project, the infrastructure sector, technology involved in the project, and local economic, financial, and political factors, among others (Mazher et al., 2019). The most common research methods adopted to capture PPP experts’ judgements include questionnaire surveys and interviews. These approaches are well-recognized and widely used in social science, general management and project management research (Thomas et al., 2003; Wibowo and Mohamed, 2010; Mazher et al., 2018). The main research question is to seek a consensus-based opinion on whether a particular risk should be allocated to the public sector or to the private sector or shared. A risk shall be transferred to the private sector if they have a superior ability or comparative advantage to manage it. Alternatively, it should be retained by the public sector. Where none of the parties show adequate capability in managing a risk, then it should be shared (Asian Development Bank, 2000; Irwin, 2007). General risk allocation preferences have been explored in previous research in the United Kingdom (Bing et al., 2005), Nigeria (Ibrahim et al., 2006), China (Ke et al., 2010b) Greece (Roumboutsos and Anagnostopoulos, 2008), Singapore (Hwang et al., 2013), Saudi Arabia (Almarri et al., 2019), in a comparative study of China and HK (Ke et al., 2010a), and in a cross-country comparison between Taiwan, Singapore, China, the United Kingdom, and Indonesia (Chou and Pramudawardhani, 2015). Some researchers have focused on expert’s risks allocation preferences in particular infrastructure sectors. Thomas et al. (2003) conducted one of the first studies that explored experts’ perceptions on risk allocation for BOT road projects in India. Kate and Patil (2020) explored risk allocation preferences for power transmission line projects in India. Shrestha et al. (2018) indicated risk allocation inefficiencies in Chinese PPP water projects whereas, Nur et al. (2023) investigated risk allocation perceptions in Indonesian geothermal projects. Other infrastructure sectors surveyed from the perspective of RAS include: Egypt’s domestic wastewater treatment and seawater desalination PPP projects (Rafaat et al., 2020), Indonesian water supply projects (Wibowo and Mohamed, 2010), and energy from waste technologies in Indian PPP waste treatment projects (Dolla and Laishram, 2021).

A particular stream of research effort has been directed at developing models to assist in risk allocation decision making. Researchers have adopted MCDM, game theory, genetic algorithms, and AI techniques to model and optimize the risk allocation decision making process. The risk allocation decision process can be likened to a MCDM problem where a utility function can be employed in evaluating multiple risks (Mazher et al., 2019). Wang et al. (2023) adopted the Best-worst multi–criteria decision-making method and comprehensive fuzzy evaluation method to develop a practical and effective risk sharing model for PPP construction waste recycling projects. Based on a set of established risk-allocation criteria, Ameyaw and Chan (2015) and Ameyaw and Chan (2016) developed fuzzy synthetic evaluation-based models to assist in risk allocation in water supply infrastructure projects in Ghana. Valipour et al. (2016), Mazher et al. (2019) and Kukah et al. (2024) also developed different fuzzy MCDM quantitative risk allocation models to guide decision-making on risk allocation in PPP projects. Valipour et al. (2019) presented a hybrid SWARA-COPRAS approach to examine risk allocation, particularly for PPP water supply and sewerage projects in the context of Malaysia. Li et al. (2017) contended that the risk allocation can be analyzed as a game to more fully reflect the risk allocation bargaining process among the parties i.e., the public and private sectors involved in PPP projects. Lai (2018) explored optimal risk-sharing scheme between the public sector and the private sector for six main kinds of risks in the development of PPP based sustainable characteristic town through use of a game model. A different approach was adopted by Alireza et al. (2014) to develop a Genetic Algorithm based optimization model to enhance risk allocation process in PPP projects. Jin and Zhang (2011) and Jin (2011) proposed and validated methodologies for modelling risk allocation decision-making process in PPP projects based on artificial neural network and neuro-fuzzy techniques, respectively.

For complex social science and real-life events, the case study method enables systematic observation of a phenomenon in its natural context. This method can provide a deeper understanding of the phenomenon under consideration. In case study research, both qualitative and quantitative data can be collected using other methods (Eisenhardt, 1989). Data are usually collected using interviews, primary and secondary documentary analysis which may include contract documentation, government and private sector reports, journal articles, conference presentations, and newspaper articles, among other sources (Ng and Loosemore, 2007; Heravi and Hajihosseini, 2012). Ng and Loosemore (2007) discussed difficulties in distributing risk appropriately while presenting a case study of the troubled New Southern Railway project in Sydney, Australia. Nguyen et al. (2018) conducted very interesting research by evaluating 21 PPP project contracts in the U.S. highway market and baselined risk allocation and provisioning among contemporary contracts. Similarly, other notable research reviewed that adopted case studies as the primary research method include a case study of the Tehran–Chalus Toll Road project in Iran (Heravi and Hajihosseini, 2012), risk allocation in Chilean highway PPPs (Castelblanco et al., 2020), risk allocation in water sector projects in China (Xu et al., 2011), and risk allocation in the context of a river management Design-Build-Finance-Operate-Transfer project (Liu and Xue, 2019). Such research provides in-depth insights on a project’s contextual aspects and why certain risks are apportioned in a particular way with some authors providing suggestions to improve risk allocation practices for future projects.

A small number of papers in the reviewed sample adopted a unique approach to their investigations pertaining to risk allocation by focusing on particular risks including revenue risk allocation for transport infrastructure PPP projects (Roumboutsos and Pantelias, 2015), demand risk in Irish PPP toll road projects (Burke and Demirag, 2015) and forex risk-allocation (Kuikel et al., 2023). Roumboutsos et al. (2020), in one of the first efforts, introduced two composite indicators as tools to guide contracting parties in efficient demand risk allocation.

PPP projects usually exhibit a significant exposure to high impact risks which necessitate support provided by the governments to attract private sector interest. Risks related to political, legal, and economic stability mainly influence the confidence of investors. Other risks of concern for private investors include those that can potentially negatively influence cash flows due to uncertainty in demand, variation in toll/tariff, different project revenue and debt denominations, and force majeure, among other risks. Government support often required includes revenue guarantee, foreign exchange risk guarantee, debt repayment security, price adjustment, restrictive competition guarantee, and the compensation for force majeure events (Ye and Tiong, 2000; Wang and Liu, 2015). Such government intervention protects the project stakeholders by shielding them from the full financial costs of risks (Iyer and Sagheer, 2011). In PPP projects, provision of guarantees by the government to the investors is an important risk-sharing scheme. Another concern is that information asymmetry between the government and the private investors leads to speculation since the government does not have the same information as the private investors (Wang et al., 2019). Provision of guarantees by the government may potentially open doors for opportunistic behavior by the private sector where either this provision may weaken the capacity of the private sector to reduce risks or such a provision may be abused by the private sector and they may qualifying for government guarantees by purposefully limiting revenues and keeping them lower than the expected threshold thereby reducing cost commitments (Medda, 2007; Wang et al., 2018; Wang et al., 2019).

The extant literature recognizes the difficulty in assessing the real impact of these supports on project’s performance and the stakeholders. Firstly, these supports cause the government to assume all contingent liabilities thus potentially leading to a situation where the government has to face significant future budgetary impacts and it may also become penalizing for the public sector (Pellegrino, 2021). Secondly, the arrangements under these supports must ensure an even risk distribution between the public and the private sector, while also balancing the distribution of benefits that result from the application of such supports and avoiding windfall gains to one party at the expense of the other (Carbonara et al., 2014b; Wang and Liu, 2015; Zhang et al., 2021). Guarantees are expensive and too many guarantees can induce government credit risk whereas inadequate guarantees provided may make the project uneconomical (Sobhiyah et al., 2009). Xu et al. (2014) argued that many PPP projects have been heavily criticized in China because the government provided the private sector with too many guarantees and suggested that both the public and the private sectors need to enhance their understanding of risk sharing to obtain a reasonable guarantee structure. A similar observation was shared by Bel et al. (2017) where the authors summarized three PPP failure case studies from Spain. The authors showed that the local taxpayers and users absorbed losses as the state’s financial liability activated and resulted in the government assuming heavy financial obligations. Alonso-Conde et al. (2007) presented the case study of the Melbourne CityLink Project and used the real options approach to model the financial incentives/guarantee provided by the government. They analyzed the consequential partial risk transfer back from the private to the public sector and how these options affect the incentive to invest and further estimated the value the public sector may be transferring to the private sector through government guarantees. Liu et al. (2017) and Pellegrino (2021) proposed quantitative models to guide governments’ decisions when choosing among the support mechanisms when the financial viability of a project is uncertain. The authors contended that there is a real risk of surplus fiscal support ex ante or unexpected liabilities ex post in case of a government’s inability to evaluate various alternatives.

A plethora of research exists that has modeled and explored fair traffic and revenue risk allocation and the influence of traffic guarantees, and government minimum revenue guarantee and traffic revenue cap (Medda, 2007; Vassallo, 2010; Iyer and Sagheer, 2011; Ashuri et al., 2012; Kokkaew and Chiara, 2013; Carbonara et al., 2014b; Wang and Liu, 2015; Buyukyoran and Gundes, 2018; Wang et al., 2018; Wang et al., 2019; Du et al., 2019; Zhang et al., 2021; Wu et al., 2022; Jin et al., 2023). Liu et al. (2019) adopted the real option theory and explored non-competitive guarantee in PPP rental retirement villages and developed a model to evaluate government compensation and guarantee costs in consideration of benefit redistribution if the governments are unable to keep the promise on guarantee provision.

In BOT projects, concession period represents the span of time within which the private sector must construct the project and then operate it to recover the investment and to generate a return on investment (Carbonara et al., 2014a). The risk burden shifts to the government if the concession period is too long and it shifts to the private sector if it is too short. The stakeholders need a reasonable length of concession period to ensure that the private sector makes a favorable profit, and the public sector gets VfM (Shen et al., 2002; Guo et al., 2021). The design of concession period involves the design of period structure, the determination of period length and incentive schemes. Each concession has its duration, which may be fixed or variable where the choice is influenced by various risk factors such as completion time, product prices and market demand (Ye and Tiong, 2003b). Shen et al. (2002) developed an NPV based approach for determining the concession period that incorporates both the investor’s and the government’s interests. Ye and Tiong (2003b) systematically explored the types of concession structures for BOT projects and evaluated the effectiveness of different concession period structures on financial return and completion risk management through mathematical analyses and computer simulations. Carbonara et al. (2014a) developed a model for calculating the concession period as the best instant of time that creates a “win–win” solution for both the concessionaire and the government and allows for a fair risk sharing between the two parties, while taking into account the uncertainty that affects the PPP projects using the MCS. Ma et al. (2018), Lomoro et al. (2020), and Guo et al. (2021) proposed more advanced and robust approaches for the determination of the optimal concession period.

A related issue is with regard to concession pricing or tariff determination and adjustment. A similar concern exists here in the determination of the concession price where the private sector wishes to ensure a reasonable return on investment and the public sector wishes to achieve VfM (Zhang, 2009). It is one of the most important issues during the PPP contract negotiations and it embodies risk sharing between the government and the private sector (Xu et al., 2012). Xu et al. (2012) developed a general concession pricing model from the perspective of the private sector. Ye and Tiong (2003a) utilized simulation analysis to systematically explore tariff adjustment frameworks for PFI projects, given project risk exposure and incentives, and evaluated the effect of adjustment frameworks on the tariff charged to customers and the rate of return earned by the project company.

In PPP, the greatest VfM driver is risk transfer. It means that higher value can be achieved in terms of cheap and high-quality services when compared to conventional procurement methods, if risks are appropriately transferred to the private sector based on its capability (Clayton, 2006). However, in practice, the greatest risk management capability is difficult to be clearly determined (Lam et al., 2007). Xu et al. (2010) identified the critical criteria for equitable risk allocation and established a quantitative model for equitable RAS. Jin and Doloi (2008) contended that risk allocation mechanism lacked theoretical support and that the claim of allocating risks based on capability is often “violated” by current industrial practice. The authors proposed and validated a theoretical framework for understanding risk allocation practice in PPP projects. Jin and Zuo (2011) studied the risk management environmental uncertainty factors to achieve efficient risk allocation and minimize risk management-related costs in a long-term view. Shrestha et al. (2017) studied three critical parameters in the principal agent relationship in PPP wastewater treatment projects, i.e., (i) competition, (ii) monitoring and (iii) incentives, and demonstrated how risks can be efficiently transferred to the private sector using these parameters. Shrestha et al. (2019) proposed a theoretical framework composed of a thirteen-step process for the allocation of risk in PPP projects using the principal agent theory and risk allocation capability of stakeholders.

Optimal allocation of risk between the public and private sector is instrumental to the enhancement of VFM for PPP infrastructure projects (Wibowo and Sundermeier, 2020). If risks are not appropriately transferred to the private sector, enhanced risk premiums and inefficient management of risks may make the project more expensive than what could be procured using traditional procurement methods. Dikmen et al. (2011) described the mechanism of PFI used by the United Kingdom Government and evaluated three case studies in achieving the essential characteristics of adequate risk transfer and VfM to the taxpayer. Wibowo and Sundermeier (2020) offered a novel approach to evaluate VFM of a PPP infrastructure project that considers the risk mitigation capability of the public and private sector stakeholders and enables early project-based decision-making for the public sector before the tendering begins.

Abednego and Ogunlana (2006) discovered the perception of proper risk allocation of each party involved and utilized the findings as the foundation to develop the concept of good project governance. One of the downsides of adopting long-term PPP contracts is the possibility of renegotiations. In long-lasting cooperation, contractual elements of risks and benefits distribution and project implementation between government and private sector are subject to change (Chen et al., 2023). Variation in identified risks and emergence of unidentified risks are the main reasons for the renegotiation phenomenon during the concession period of PPP projects, where the purpose of these renegotiations is risk re-sharing on new terms to “save the deal” in a bilateral, asymmetric, and noncompetitive environment (Fernandes et al., 2019; Wang and Yu, 2020). Factors contributing to these renegotiations could include poor estimations and forecast, weak project documentation, and inadequate technical and economic evaluations, and unilateral changes by the government that involves a change in risk-sharing structure of the contract, among a range of other factors (Fernandes et al., 2019). Guasch et al. (2014) reported that more than 50% Latin American transport concession contracts and contracts in the water sector were renegotiated. Similarly, more than half of Portuguese concessions studied in different sectors were renegotiated (Cruz and Marques, 2013).