Observations on Transportation’s Risk-based Contingency vs. Industrial Risk-driven QRA

Most of my career was on industrial capital projects where risk quantification has a 60+ year history of QRA methods development including parametric in the 50s-60s, range estimating with MCS in the 60s-90s (advancing in parallel with phase-gate) and an increasing focus on a systems/complexity view, cost/schedule integration and empirically-valid methods through the 2000s. Next natural step is AI. AACE has captured much of this in their Recommended Practices. The industrial world is driven by competitiveness; reducing cost and duration. Flexible RFPs and EPCM are most common. The focus is early; the front-end and the Select phase. AACE International is the main venue for sharing by industrial sector QRA practitioners.

The last decade I have been increasingly involved in public transportation work (road/rail). What I find is that it is mainly still in the more-or-less pre-determined contingency with less QRA but in recent years transitioning to the ranging-based MCS methods of the 1990s. Phase-gate is increasingly used but about 15 years behind industrial in maturity. Some, trying to address perceived overrun tendency, are using reference class forecasting or similar non-risk driven benchmarking (its not QRA). The transport world is driven less by competitiveness and more by a (somewhat schizophrenic) mix of politics (announce a low price too early) and predictability (deliver the undeliverable). Hard-nosed Tendering and DBB is most common. The focus is late: the construction phase. Transportation and civil engineering associations are the main venues for transportation estimating practice sharing (and estimators are its primary cost risk practitioners).

The following are typical QRA / Contingency guidelines found:

For the industrial sectors, AACE International’s Professional Guidance Document (PGD-02) provides a roadmap and application guide for QRA methods. This covers all the project phases and levels of complexity: ?https://library.aacei.org/pgd02/pgd02.shtml

For transportation sectors, each country and province/state has its own venues for its estimating practitioners. For example, in the US for road transport projects there is FWHA (federal), AASTHO (state DOTs) and more generally ASCE (civil engineering). For the UK, there is IPA. And so on. One finds similar transport groups in each country. For each there are guides of various flavors for their fledging “risk-based” methods. A couple representative examples are below:

Australia: Guidance Note 3A - Probabilistic contingency estimation (2023): https://investment.infrastructure.gov.au/sites/default/files/documents/guidance-note-3A-probabilistic-cost-estimation-v2.pdf

AASTHO: Guidebook on Risk Analysis Tools and Management Practices to Control Transportation Project Costs (2010): https://nap.nationalacademies.org/catalog/14391/guidebook-on-risk-analysis-tools-and-management-practices-to-control-transportation-project-costs

UK Infrastructure and Projects Authority: Cost Estimating Guidance 2021): https://assets.publishing.service.gov.uk/media/6050c9528fa8f55d324b0c84/IPA_Cost_Estimating_Guidance.pdf

Note that in the transport sectors, the language of risk quantification is more focused on “contingency estimating” rather than integrated quantitative risk analysis revealing a lingering focus on a deterministic number (driven in turn by its tender/construction focus) despite a push for more risk-based and probabilistic methods that would support early phases.

For data, the industrial sectors have been benchmarking their project systems including cost, schedule, risks and practices since the 1980s albeit confidentially (e.g., Independent Project Analysis ) that supports risk-driven modeling while the transport sectors in the public domain have regional data records but mostly of cost at various levels and largely devoid of practice data which supports base estimating and gross RCF, but not risk-driven modeling. In both industries, AI QRA methods are hindered by limited availability and sharing of data.

My views are biased of course; I am very active in AACE and am a partner in a commercial QRA software tool ValidRisk . There is some, but not extensive overlap between AACE and the regional transport group memberships. For those with toes in contingency/QRA in one or both sectors, do these views agree with your observations regarding project cost and schedule risk quantification? Am I missing something? Can you share better examples from the sectors?

?