Sensitivity analysis is a critical component of project evaluation, yet it often receives insufficient attention. When this happens, the overall assessment of a project’s financial performance can be compromised.
Under a deterministic approach, the effect of changes in key assumptions is assessed by varying one parameter at a time while holding all others constant. In this case, four cashflow parameters were examined: coal calorific value (CV), total operating costs, capital expenditure, and sales production tonnages. Each parameter was adjusted by ±10% and ±20% to measure its effect on project value, as shown in the graphs above.
The results indicate that the project is most sensitive to changes in coal CV and sales production tonnage, while operating costs and capital expenditure have a comparatively smaller effect on net present value (NPV).
In many studies, sensitivity analysis ends at this stage. However, such an approach only reveals which parameter has the greatest impact on NPV, without showing how multiple uncertainties may interact. A stochastic approach addresses this by allowing all parameters to vary simultaneously, thereby capturing a more realistic range of outcomes.
In this example, operating and capital costs were modelled using a truncated uniform distribution to avoid unrealistic values, while coal CV and sales tonnage were modelled using a normal distribution. A Monte Carlo simulation was then used to generate numerous scenarios, resulting in a distribution of possible NPVs (illustrated in the bar charts above). This distribution provides insights into the probability of unfavorable outcomes (such as a negative NPV), known as the project’s value-at-risk (VaR).
For instance, the simulation results show a 45% probability of a negative NPV under Option 1. By contrast, in Option 2, where a portion of upfront capital expenditure is shifted to a contractor miner, the probability of a negative NPV decreases to 35%. This demonstrates that while reallocating capital costs from owner to contractor has limited effect on the overall NPV distribution, it significantly reduces project risk (VaR) and improves the expected NPV (mean).
Author: Shaun Barry, Principal Consultant (Project Evaluation), SRK Consulting
