Data-driven quantification of public–private partnership experience levels under uncertainty with Bayesian hierarchical model
Introduction
The critical role of infrastructures, such as airports and seaports, played in economic growth and poverty alleviation in developing countries has long been internationally recognized [1]. Governments are pursuing private resources to alleviate the financial and technical deficit owing to the high cost of public debts and increasing demand in infrastructure facilities [2], [3]. Public–private partnership (PPP) is one type of private involved project, and what makes it different from other types of procurement modes are the attributes of risk-sharing between public and private sectors, together with the long duration of the contract. These make it easy to be adversely influenced by various risks, leading to the failure of projects [4], [5]. PPP contract failure, including distress or eventual termination of the contracts, is an extreme form of the poor performance of a PPP project, which has been considered as serious trouble for foreign investors [6], [7]. To avoid a contract failure, governments should be empowered with adequate capabilities in proper planning, execution, and monitoring, which are heavily reliant on the PPP experience that governments may possess [8], [9]. Knowledge about prior PPP experience would avoid a start from scratch, therefore reducing the uncertainties in the PPP project, however, the research on the role of a country’s PPP experience in lowering the contract failure rate is insufficient [10]. Therefore, there is an urgent need for assessing if and how the contract failure rate changes as a country’s PPP experience level increases.
Learning from experience to improve future infrastructure PPPs has attracted a lot of attention from policymakers, financiers, implementers, and private stakeholders [8]. For instance, Sampson et al. [11] found that institutional capabilities and performance can be strengthened by the experience. Jones et al. [12] stated that experience in collaboration over the implementation of infrastructure can contribute to a peaceful and effective solution to interest disputes. O’Shea et al. [13] pointed out a gradual change in the governance of PPPs in Ireland has been found by studying 27 PPP projects. Cohen et al. [14] stated that the lessons learned from the Milan metro line project can help in cost reduction in future projects. Hurk et al. [15] concluded that the government could adjust the contracts to the market as they developing more projects by studying four road projects in Belgium. Overall, it is sure that a country can gain experience from previous PPP projects. Unfortunately, the above-mentioned studies are commonly about the lessons learned from the previous PPP projects. The deep insight about the impact of PPP experience on the PPP performance is, however, absent.
As the experience is evolving with time, there could be some reflection points that can divide the PPP experience into various levels, whose effects on the performance of PPP projects are diverse [16]. Change points detection can be a challenging task since both the number and positions of change points are unknown but it has been used in many domains, such as finance and crime analysis [17]. Some effort has been committed to the categorization of experience levels by surveys. For example, Queiroz et al. [18] divided the PPP experience into 5 levels by using PPP Perception Index based on the results from questionnaires. Wang et al. [19] classified PPP experience as success or failure according to whether a country has concluded or failed PPP projects. It can be seen that these studies only characterized the PPP experience of a country from a qualitative standpoint. The bounds of these experience levels are, however, not properly identified. Hence, some quantitative methods were proposed by researchers in response to the above problems. For example, Marcelo et al. [8] interpreted the PPP experience as the number of developed PPP projects in the past decades based on the World Bank Private Participation in Infrastructure (PPI) Project Database. In this method, the location of the change point is determined by visual check, and the influence of experience on the contract failure rate is determined using the piecewise linear regression model. However, the method is subjective and via-based, leading to the misdetection or omission of PPP experience levels. This suggests the need for a quantitative and accurate model to determine the bounds of different experience levels.
The purpose of this paper is to quantify developing countries’ PPP experience levels by taking advantage of both the binary segmentation method and the Bayesian hierarchical model. The research questions to be addressed in this research include: (1) How to measure PPP experience in developing countries, (2) How to determine the number of PPP experience levels by the binary segmentation method, and (3) How to determine the bounds of PPP experience levels based on Bayesian hierarchical model.
The remainder of this paper is organized as follows: Section 2 introduces the related work of change point detection as well as preliminaries of the binary segmentation method and the Bayesian hierarchical model. Section 3 presents the proposed methodology with detailed step-by-step procedures. Section 4 shows the data source. Section 5 gives the developed Bayesian hierarchical model and analyzes the results. Section 6 discusses the advantages of the proposed model. Section 7 draws up the conclusions and future studies.
Section snippets
Change point detection
In the change point detection problems, two primary tasks are determining the number and locations of change points. The binary segmentation method, which can search the number and locations of change points automatically, is one of the widely cited algorithms in multiple change-point detection processes [20], [21]. It has been proved to be consistent in detecting the number of change points in a series of sectors, such as hazard rate and finance [22]. Moreover, different hybrid methods have
Methodology
From the current literature, no general model has been found with uncertainties considered for analyzing experience levels. To quantify the level of PPP experience for different developing countries, a novel probabilistic method that combines the binary segmentation method and Bayesian hierarchical model was proposed. Firstly, the number of change points, where the level of experience for a country transforms suddenly, will be searched by the binary segmentation method. Besides, the location of
Data exploration
This section provides an outline of 7096 PPP projects in developing countries, which exist in 5 PPP sectors with each sector being divided into several subsectors. Table 2 illustrates the number of PPP projects in each sector.
The distribution of PPP projects shows an unequal trend in different sectors and regions. Besides, some countries have a lot more PPP projects than other countries in every sector. Table 2 indicates that the energy sector has the most PPP projects, while the information
Analysis of results
The Bayesian hierarchical model has been developed to categorize the PPP experience levels of developing countries in different sectors. In this section, results will be presented. First, the resultant number of change points in each sector will be determined. Second, the Bayesian hierarchical model will be established based on results in the last step. Third, the convergence of the Bayesian hierarchical model was checked and the posterior distribution of the parameter was given. Finally, the
Discussions
A comparison between the proposed method and the conventional one has been conducted to further demonstrate the advantages of the proposed methodology based on the data from the transportation sector. Table 11 lists the prior distributions and results of the change point by the Bayesian hierarchical model with and without informative priors. Fig. 15 gives the posterior distribution of – for the transportation sector under noninformative priors.
As previously mentioned, the arbitrary
Conclusions and future studies
PPP experience is one of the most important advantages for public agencies to deal with the relationship with private counterparty. However, most private investors only have a qualitative understanding of the PPP experience of a country. The objective of this research is to quantify the levels of countries’ PPP experience for different PPP sectors in a data-driven manner, which is expected to provide a guideline for private investment. In this research, the definition of PPP experience was
CRediT authorship contribution statement
Yongqi Wang: Writing - original draft, Methodology, Visualization, Data curation, Validation, Formal analysis. Zengqi Xiao: Data curation, Methodology. Robert L.K. Tiong: Data curation, Methodology, Supervision. Limao Zhang: Methodology, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgment
The Ministry of Education Tier 1 Grants, Singapore (No. 04MNP000279C120; No. 04MNP002126C120) are acknowledged for their financial support of this research.
References (71)
- et al.
Performance risk assessment in public–private partnership projects based on adaptive fuzzy cognitive map
Appl. Soft Comput.
(2020) - et al.
A bi-projection model based on linguistic terms with weakened hedges and its application in risk allocation
Appl. Soft Comput.
(2020) - et al.
Supporting infrastructure development in fragile and conflict-affected states: Learning from experience
Oxf. Policy Manag.
(2012) - et al.
Change points detection in crime-related time series: An on-line fuzzy approach based on a shape space representation
Appl. Soft Comput.
(2016) - et al.
Hybrid change point detection for time series via support vector regression and CUSUM method
Appl. Soft Comput.
(2020) - et al.
A hybrid method for estimating the process change point using support vector machine and fuzzy statistical clustering
Appl. Soft Comput.
(2016) - et al.
A novel phase I fuzzy profile monitoring approach based on fuzzy change point analysis
Appl. Soft Comput.
(2018) - et al.
Application of a Bayesian hierarchical modeling for risk assessment of accidents at hydropower dams
Saf. Sci.
(2018) - et al.
Bayesian-network-based safety risk analysis in construction projects
Reliab. Eng. Syst. Saf.
(2014) - et al.
Roles of artificial intelligence in construction engineering and management: a critical review and future trends
Automat. Constr.
(2021)
An exact approach to Bayesian sequential change point detection
Comput. Statist. Data Anal.
Detecting change-point trend and seasonality in satellite time series data to track abrupt changes and nonlinear dynamics: A Bayesian ensemble algorithm
Remote Sens. Environ.
Local probabilistic model for Bayesian classification: A generalized local classification model
Appl. Soft Comput.
Borrowing strength and borrowing index for Bayesian hierarchical models
Comput. Statist. Data Anal.
A real-time fault diagnosis methodology of complex systems using object-oriented Bayesian networks
Mech. Syst. Signal Process.
A manufacturing failure mode and effect analysis based on fuzzy and probabilistic risk analysis
Appl. Soft Comput.
Modified Bayesian data fusion model for travel time estimation considering spurious data and traffic conditions
Appl. Soft Comput.
Hierarchical sparse observation models and informative prior for Bayesian inference of spatially varying parameters
J. Comput. Phys.
Estimating the number of change-points via Schwarz’criterion
Statist. Probab. Lett.
Willingness to take contractual risk in port public–private partnerships under economic volatility: The role of institutional environment in emerging economies
Transp. Policy
Using penalized contrasts for the change-point problem
Signal Process.
Approaches to risk identification in public–private partnership projects: Malaysian private partners’ overview
Dirasat Adm. Sci.
Roles of private-sector partners in transportation public–private partnership failures
J. Manage. Eng.
Public–private partnership contracts: A tale of two cities with different contractual arrangements
Public Adm.
Triggers of Contract Breach: Contract Design, Shocks, Or Institutions?
Do Countries Learn from Experience in Infrastructure PPP? PPP Practice and Contract Cancellation
Factors influencing early termination of PPP projects in China
J. Manage. Eng.
Failure mechanisms in international water PPP projects: A public sector perspective
J. Constr. Eng. Manag.
Experience effects and collaborative returns in R & D alliances
Strateg. Manag. J.
Using PPP to procure social infrastructure: Lessons from 20 years of experience in Ireland
Public Work. Manag. Policy
Governance of public–private partnerships and infrastructure delivery: Case of the Milan, Italy, metro line M4
J. Transp. Res. Board
Learning to contract in public–private partnerships for road infrastructure: Recent experiences in Belgium
Policy Sci.
Public–private partnerships in developing countries: The emerging evidence-based critique
World Bank Res. Obs.
An Overview of the Brazilian PPP Experience from a Stakeholders’ ViewpointTechnical Paper
Cited by (5)
Modelling supply chain disruption analytics under insufficient data: A decision support system based on Bayesian hierarchical approach
2022, International Journal of Information Management Data InsightsCitation Excerpt :It is mainly used when there is a need for prediction with limited number of data and the outcome is uncertain. Readers are referred to Wang et al. (2021), Bunnin and smith (2021) and Kalinina et al. (2018) to have detailed insights on BHM. This study takes the advantages of using the concepts of the BHM to make quantifiable and justifiable decisions to select suppliers subject to natural and manmade disruptions.
Closing the gap: financial closure in public private partnership markets
2023, International Journal of Critical InfrastructuresIntroducing Antifragility Analysis Algorithm for Assessing Digitalization Strategies of the Agricultural Economy in the Small Farming Section
2023, Journal of the Knowledge EconomyPublic-Private Partnership Contract Failure Prediction Using Example-Dependent Cost-Sensitive Models
2022, Journal of Management in Engineering