This category is the core and the traditional way for performing CBA, serving as the prerequisite ground for the existence of the other categories. Consequently, several sources in the literature (18% of the studied papers) have utilized this method of calculation; papers detailed attributes can be seen. Indicatively, DeCorla-Souza et al. [26], Greer and Ksaibati [27] and Uddin and Mizunoya [28] developed Mi-D-S tools for evaluating conventional parameters of CBA, such as reduction of accident, travel time and vehicle operating costs.

While all the studies reported above highlighted a specific CBA framework, Hanssen et al. [30] investigated how CBAs outcomes differ against the applied model by comparing three different national CBA models from the Nordic countries, concluding that the choice of national model is crucial since different results can derive depending on the model used.

Furthermore, several national authorities and organizations developed practice guidelines for assessing policies via CBA in a Mi-D-S environment. Four out of the five national guidelines, Minnesota DOT [31], Transport Infrastructure Ireland [32], UK Department for Transport [33] and World Bank [34], offered technical instructions or theoretical principles concerning consumer surplus, producer surplus and externalities. In contrast, California DOT [35] developed the so-called Cal-B/C software, a Microsoft Excel spreadsheet-based tool that provides economic CBA for a range of TIPs. The basic version of Cal-B/C does not contain risk analysis and it is focused on specific project-based applications, meaning that a series of input variables, namely annual average daily traffic (AADT), international roughness index (IRI), vehicles speed, etc., should be provided by the user for software’s running.

However, crucial considerations might be lost in conventional CBAs leading to suboptimal investment strategies because, in reality, markets are distorted direct and indirect impacts may differ [20]. Hence, this category seeks to narrow down the gap between actual and captured impacts by incorporating additional impacts ignored in the conventional CBAs for estimating over and above traditionally measured project’s user benefits. The WEIs studied in the literature ranged from economic to environmental to social aspects.

Two extant studies, Calthrop et al. [16] and Kidokoro [36], developed general equilibrium (GE) models to explicitly incorporate all effects of transport investment on all economic markets, advocating that capturing only the direct costs and benefits of an investment may yield misleading unrealistic CBA results. The former [16] considered distortions on all markets and distributional effects for adapting the traditional cost–benefit rules to correct the unrealistic assumptions considered in the conventional CBAs. The latter [36] investigated a basic agglomeration economy model to incorporate the agglomeration effect into conventional CBAs and produce results comparable with conventional CBAs and directly applicable to individual TIPs, such as roads. Accordingly, Laird and Venables [37] recognized the importance of an appraisal framework that ensures all relevant impacts are captured in CBAs beyond conventional benefits. Specifically, they analysed three additional WEIs types, namely the productivity effects, the investment and land-use changes and labour market effects. These WEIs types had been further examined by ITF/OECD [38], which emphasized mechanisms through which transport may create wider benefits and presented well-established methodologies for their comprising into CBAs, as adopted for instance by UK Department for Transport. Moreover, from a UK perspective, OECD/ITF [39] theoretically reviewed the state of WEIs of agglomeration economies, imperfect competition benefits and labour supply effects.

Moreover, Pienaar [40] conducted a regional economic income analysis displaying the significant contribution of the RIP to the primary macroeconomic goal of local wealth. An and Casper [41] conducted CBA combined with regional travel demand analysis using the commercial TREDIS software for evaluating regional transportation projects by examining their economic impacts. Gühnemann et al. [42] developed an innovative procedure for modifying CBA results to facilitate CBA and multi-criteria analysis (MCA) combination as a means of providing a closer alignment between transport policy and the tools used to support projects’ effective prioritization.

As regards environmental terms, Manzo and Salling [43] integrated the UNITE-DSS model with a life-cycle assessment (LCA) module for evaluating how the indirect environmental impacts affect the final project evaluation. From the comparison made between CBA’s results of the two alternative approaches, namely with and without the LCA module, it was concluded that the LCA module highly affects the CBA’s socioeconomic indicators.

From a social point of view, Turró and Penyalver [44] introduced the Intergenerational Redistributive Effects Model (IREM) that incorporate intergenerational fairness into present decisions for detecting investments that could reduce the wellbeing of affected future generations. Contrary to conventional CBAs that assumes that projects are generationally neutral, IREM provides indicators on the intergenerational redistributive effects arising from major TIPs. Thus, IREM is useful to establish to what extend the project’s impact is positive for society from a broader perspective than the traditional CBA.

Until the mid-2000s, the road infrastructure field lacked a generally approved comprehensive way of combing CBA with quantitative risk analysis (QRA). The first attempt to provide risk-based CBA of RIPs was conducted by Salling and Leleur [24,52]. They introduced an Excel-based software, the so-called CBA-DK model, for assessing TIPs combing deterministic CBA with MCS via @Risk software. In this way, decision makers could have more profound and informed knowledge since CBAs’ outputs were presented within confidence intervals rather than single-point estimates. A set of suitable PDFs, defined by a PhD thesis [76] for critical input variables into the CBA-DK framework, was used. In general, for running QRA into the CBA-DK, users should choose the PDF type among those available and define their specific parameters, thus limiting its use to those who have detailed knowledge for producing a high level of knowledge (LoK) PFD for each input parameter in order to avoid bias issues derived from low LoK PDFs.

Hence, a special issue recognized by Salling and Banister [59] was the pursuit of the most representative PDFs for capturing the inherent uncertainty of input parameters into the CBA-DK framework. The authors proposed the Reference Class Forecasting (RCF) method for shifting the LoK concerning processing uncertain input variables from a low LoK to a high one by expressing the variables with a statistical distribution formulated by similar projects values. Thereafter, Salling and Salling and Leleur [54,55,56] presented the UNITE-DSS, an Excel-based decision support model, which contains an integrated approach to socio-economic analysis, risk-based simulation and the so-called UP database containing almost 200 specific European TIPs (e.g.,: roads, fixed-links, rails) between 2009–2013. Again, among various CBA’s inputs, construction costs and demand forecasts were proven to be affected by a substantial degree of uncertainty and their PDFs were further examined for obtaining reliable estimates.

Furthermore, various “case-study” papers performed traditional CBA combined with MCS each of them examining a specific objective within roads’ economic evaluation. All these papers used commercially available risk assessment tools integrated with Microsoft Excel that can be applicable for RIP evaluation, such as @Risk and Crystal-Ball software. Particularly, Korytárová and Papežíková [46] estimated the total inaccuracy of economic efficiency ratios calculation in ex-ante project appraisals using QRA and noted that the benefits inaccuracies between ex-ante and ex-post approaches presented very inconsistent results, with the travel time savings occupying the largest share of inaccuracy among studied benefits. Del Giudice et al. [45], Prakash [48], Vagdatli and Petroutsatou [23] and Varbuchta et al. [47] supported that probabilistic CBAs of RIPs render the evaluation process more transparent and responsible since they provide additional information to the decision makers compared to the deterministic ones. All these papers used PDFs with low LoK, while the last two acknowledged that a database of similar projects would be beneficial for extracting more appropriate PDFs for input variables, leading to more robust probabilistic NPV results. Except for the well-established MCS used for copying with uncertainty of RIPs, some authors considered the uncertainty in the CBA models using alternative risk-based procedures. Maravas et al. and Maravas and Pantouvakis [50,51] presented an alternative mathematical approach based on fuzzy set theory for modelling the inherent uncertainty of TIP into CBA. They concluded that fuzzy-CBA is much easier to computerize than MCS obtaining useful results very quickly. Likewise, Bağdatlı et al. [58] investigated the utility of a fuzzy cognitive map approach for minimizing the effects of uncertainty in highway CBAs. Nguyen et al. [62] presented an enhanced functional CBA framework providing six functions and four main processes regarding the holistic picture of project evaluation.

Additionally, there is a broad grey literature covering the issue from a Mi-P-S perspective. Seven government and organizational reports incorporate risk analysis into the CBA framework and provide either theoretical or technical guidelines for accomplishing it. Four of these reports, British Columbia [70], OECD/ITF [65], Queensland Treasury [71] and Treasury Board of Canada [64], were conceptually compiled, referring solely to the conventional impacts and risk analysis being addressed in the structure of CBA within Canadian, Mexican, Australian and British Columbia framework, respectively. All these guides reported sensitivity analysis as a method for considering uncertainty, while British Columbia [70] added the scenario analysis (SA) and the Treasury Board of Canada and Queensland Treasury [64,71] added the MCS for risk assessment. From a quantitative assessment point of view, the Asian Development Bank and European Commission [68,69] offered methods for sensitivity analysis and MCS for identifying projects’ critical variables, allocating appropriate PDFs to them and performing QRA. Finally, the State of Queensland [66] proposed a PC-based tool, the so-called CBA6 software, for evaluating rural and urban RIPs. As in the case of Cal-B/C software, CBA6 performs project-based CBA, implying that a series of input variables—namely AADT, vehicles speed, road length, life-cycle costs, etc., should be provided by the user for software’s running. Regarding the risk analysis, CBA6 conducts sensitivity analysis for some specific input variables, such as vehicle operating cost and travel-time savings.

Articles in this category differ from those in the previous one only in terms of the examined markets, namely they consider a macroeconomic approach considering the interaction between the transport sector and the overall economy. The two approaches that complement the analysis remain probabilistic and static.

Two studies, Salling et al. and Shiau [53,60] combined CBA and MCA for assessing a macroeconomic set of distributional and other impacts under uncertainty. The former [53] presented a special hybrid version of the CBA-DK, the Excel-based CLG-DSS model, for decision makers to be facilitated to assess various uncertainties of TIPs. This model consisted of two modules, the COSIMA-module (CBA and MCA combination) and the computable general equilibrium (CGE) model. Their coupling was regarded as well-suited to address both the direct and indirect effects of TIPs. They examined the WEIs of network and mobility, employment and logistics and goods effects. Ten different scenarios regarding the regime of the market mechanism were considered in the CLG-DSS model, with MCS via @RISK software to be used for handling uncertainty. The latter [60] introduced a hybrid approach using the Dempster–Shafer theory for handling uncertainty due to missing information and synthesized monetary and non-monetary criteria into a utility unit. Moreover, Parker and Rommelaere [57] synthesized a reliability ratio for integrating the travel time reliability into the CBA as an additional benefit of TIPs and created the AutoCASE model, a commercial version of the proposed models applicable in the USA and Canada regions. All model’s results were presented in probabilistic terms using MCS.

Six national guidelines addressed the CBA in Ma-P-S terms. In five out of six national guidelines (European Investment Bank [67], United Kingdom HM Treasury [72], Commonwealth of Australia [73], Transport of New South Wales [74] and AIReF [75]) the most common techniques for risk were sensitivity and full risk analysis, with the latter making use of MCS for providing a comprehensive picture of the potential variability of a project. Additionally, decision trees [72] and SA [75] completed the reported set of risk analysis techniques. OECD [63] analysed except sensitivity analysis and various other techniques for risk assessment, such as comparative risk assessment, risk–benefit analysis, and risk–risk analysis. Furthermore, equity [63] and distributional effects [73] were among the most highly reported WEIs.

Two articles considered the Ma-D-D perspective for performing CBA (4% of the total cases). Nguyen et al. [61] introduced a hybrid conceptual framework for capturing the dynamic relationships between CBA’s costs and benefits over time, using a combination of agent-based modelling (ABM) and System Dynamics (SD) with multi-criteria method (MCM). The first two approaches captured the behaviour of the heterogeneous agents in the transport, supply chain system and real estate market leading to a macroeconomic consideration of CBA and examined the socioeconomic factors and their relationships. Rothengatter [18] presented various approaches for WEIs measurement and analysed the usefulness of connecting them with conventional CBA. Among the analysed approaches was the SD, which allows the investigation of dynamic relationships between the elements in order to understand complex systems and their development in space and time.

A crucial feature of most of the extant literature is that “risk” and “uncertainty” terms were interchangeably used when conducting probabilistic CBA. Only Li and Madanu [49] separated the two concepts and proposed a new methodology for highway project life-cycle CBA, which except for certainty analysis, introduced risk analysis for input variables with known probabilities and uncertainty analysis for inputs with unknown probabilities. Under a Mi-P-D environment, they approached the dynamic relationship among the examined annual maintenance and user costs using a geometric growth rate between two successive treatment intervals based on the first year’s values for each interval.

3. Gaps of CBA Approach and Future Research Directions

3.1. Model’s Probabilistic Inputs Updating

3.2. Dynamic Interdependent Framework

3.3. Homogenous Data Collection and Componentization of Road Assets

3.4. Proposed Conceptual Model for Future Research