Aligning Cost Risk Analysis with ICE and RICS Guidance

The shift from deterministic to probabilistic risk

Infrastructure programmes have historically relied on arbitrary percentage additions for contingency, a practice often referred to as the "10% syndrome". Data from the Royal Institution of Chartered Surveyors (RICS) and the Infrastructure and Projects Authority (IPA) indicates that while a cost contingency of 10–15% of the base estimate may be appropriate for well-scoped construction projects, complex infrastructure or technology-intensive projects at early design stages frequently require 20–35% or higher. However, applying a flat percentage fails to account for the specific risk profile and uncertainties inherent in a unique programme.

The professional framework for cost estimating and risk management is heavily guided by RICS and the Institution of Civil Engineers (ICE). Both professional bodies acknowledge that deterministic, single-point estimates are insufficient for complex infrastructure. Instead, they recommend Quantitative Cost Risk Analysis (QCRA) and Quantitative Schedule Risk Analysis (QSRA) using Monte Carlo simulations. This probabilistic approach models the interaction of risks, captures correlations between related activities, and produces a distribution of possible outcomes, allowing programmes to establish precise P50 or P80 confidence levels.

ICE and RICS frameworks for cost risk

RICS sets the baseline for cost planning through its New Rules of Measurement (NRM 1 and NRM 2) and the technical standards contained within the RICS Black Book. Within its specific guidance on the management of risk, RICS explicitly defines Monte Carlo simulation as a computer-generated technique used to model outcomes and quantify probability. This formal definition embeds probabilistic modelling into the standard toolkit of the chartered surveyor.

Concurrently, ICE provides a strategic framework through RAMP (Risk Analysis and Management for Projects), developed in collaboration with the Institute and Faculty of Actuaries. The RAMP framework, currently in its third edition published in 2014, offers a structured, lifecycle approach to identifying and quantifying project uncertainty. Furthermore, ICE has published specific guidance on infrastructure cost risk via its report, "Managing cost risk and uncertainty in infrastructure projects". This report outlines nine key recommendations for managing cost risk and uncertainty throughout the project life cycle. Together, these frameworks establish a professional standard that expects risk to be quantified probabilistically rather than estimated through intuition.

The limitations of expected monetary value

A common intermediate step between arbitrary percentages and full probabilistic modelling is the use of Expected Monetary Value (EMV). EMV multiplies the probability of a risk occurring by its financial impact to generate a single-point contingency figure. While this is an improvement over a flat percentage, professional guidelines and advanced QCRA practice recognise its severe limitations for complex programmes.

The difference between Expected Monetary Value and the P80 confidence level derived from a quantitative risk analysis is not merely arithmetic. It represents the additional uncertainty that emerges when risks combine in adverse ways. Monte Carlo simulations capture these compounding effects, illustrating how multiple concurrent risks can push a programme beyond its baseline budget. Relying solely on EMV leaves programmes exposed to the tail-end risks that frequently derail major infrastructure delivery, as it fails to account for the correlation and simultaneous realisation of multiple threats.

Data reliability and cognitive bias

While professional bodies endorse Monte Carlo simulations, the methodology requires rigorous application. The primary challenge is the "garbage in, garbage out" principle: the results of a Monte Carlo simulation depend entirely on the quality of the input data. Sophisticated QSRA and QCRA software cannot correct flawed risk identification, poorly scoped schedules, or biased impact assessments.

Data reliability is a persistent issue in risk modelling. Monte Carlo simulation requires highly reliable data for effective application, yet the reliability of subjective opinions gathered during expert elicitation is often difficult to ascertain. Cognitive biases can unintentionally skew the inputs provided by project teams, leading to suboptimal risk portfolios.

Furthermore, practitioners frequently face challenges in identifying the appropriate probability distributions for individual risks. Deciding whether a risk follows a triangular, PERT, or lognormal distribution requires a deep understanding of the risk's nature and potential extremes. If the underlying data and chosen distributions are flawed, the resulting P50 or P80 figures will provide a false sense of security rather than a reliable baseline for decision-making.

What this means for your programme

For professionals running cost or schedule risk analysis, the convergence of ICE and RICS guidance provides a defensible, auditable basis for calculating contingency reserves. It mandates a shift away from arbitrary percentage additions toward probabilistically derived contingencies that satisfy the rigorous governance requirements of major clients and government bodies.

There remains a recognised tension at early project stages. RICS NRM 1 traditionally relies on a deterministic approach, equivalent to a risk-unadjusted cost estimate plus a contingency reserve at early feasibility stages. However, advanced QCRA practice bridges this gap by overlaying Monte Carlo simulations to calculate that reserve probabilistically from the outset. By aligning Monte Carlo outputs with RICS NRM cost breakdowns and the ICE RAMP strategic framework, practitioners ensure their risk models are grounded in recognised professional standards.

Establishing a defensible contingency

The ultimate goal of applying ICE and RICS guidance to quantitative risk analysis is to produce credible, justifiable figures that stand up to scrutiny. When contingency is calculated using rigorous Monte Carlo simulations based on robust data, the resulting P50 or P80 figures become highly credible to stakeholders, sponsors, and auditors.

This process ensures that contingency is treated as a probabilistically justified client-side budget rather than a hidden buffer within the contract price. By adhering to the standards set by RICS and ICE, infrastructure programmes can move beyond subjective guesswork and single-point estimates. Instead, they bring mathematical rigour and professional certainty to project cost and schedule forecasting, ensuring that the programme is adequately funded to absorb the realities of infrastructure delivery.