One of the requirements of NI 43-101 for Item 22 Economic Analysis is “sensitivity or other analysis using variants in commodity price, grade, capital and operating costs, or other significant parameters, as appropriate, and discuss the impact of the results.”
 The result of this 43-101 requirement is typically the graph seen below, which is easily generated from a cashflow model.  Simply change a few numbers and then you get the new economics.  The usual main conclusions derived from this chart are that metal price has the greatest impact on project economics followed by the operating cost.   Those are probably accurate conclusions, but is the chart itself telling the true story?
 DCF Sensitivity GraphI myself have created the same chart in several economic studies so I understand the limitation with it.   The main assumption is that sensitivity economics are generated on the exact same reserve and production schedule as for the base case.  That assumption may be applicable when applying a variable capital cost but may not be applicable when applying varying metal prices and operating costs.   Does anyone think that in the example show, the NPV is still $120M with a 20% decrease in metal price or 20% increase in operating cost? Potentially a project could really be uneconomic with such a significant decrease in metal price but that is not shown by the sensitivity analysis.
Increasing the operating cost changes the cutoff grade, which changes the waste-to-ore ratio within the same pit.  So assuming the same the life-of-mine production tonnage is not entirely correct in this scenario.
Reducing the metal price would also result in a change to the cutoff grade.  If one were to go all the way back, these changes in economic parameters would impact on the original pit optimization used to define the pit upon which everything is based.  A smaller pit size results in a different pit tonnage, which may require a smaller processing plant, which would then have new (higher) operating and lower capital costs than assumed.  A smaller reserve would produce a different production schedule and shorter mine life.  It can all get quite complex.
So due to all the changes these sensitivities generate, it does require a lot of work to properly examine them. However generally the project proponent does not want to incur the costs necessary to run multiple pit designs and multiple life of mine plans simply to examine sensitivities.  Hence the shortcut is to simply change inputs to the cashflow model and generate outputs that are questionable but meet the 43-101 requirements.

One thought on “26. Cashflow Sensitivity Analyses – Be Careful

  1. hardrockminer

    To be frank, this is one section of a 43-101 where I don’t bother wasting my time reading. Even metal price sensitivity can be distorted if the mine produces more than one product, which is more often than not the case. What I DO look at is the metal price assumptions, the capital cost estimate, the operating cost estimate, the resource estimate (for how tonnes and grade are calculated) and the estimated recoveries. If these factors are reasonable then the cash flows will likely be reasonable.

    Most financial houses can supply metal price estimates, and one can develop a more or less “normal” distribution if one has 20 or more of them. It’s a simple matter then to decide how much risk to take on metal price. I generally drop back by one sigma to ensure my assumed price will be less than the consensus at least 83% of the time.

    Grade and recovery have killed a lot of mines, so I pay a lot of attention to how the resource grade was determined and how much metallurgical work was done. Cash flow is very sensitive to these two factors. Issues such as the composite length, block size, geological constraints, etc need to be correctly determined for the orebody in question. Any sensitivity should include these factors.

    Lastly, one needs to decide what discount rate is appropriate. Most operators would accept 8% in today’s world for an operating mine, but not for a project. Projects still carry significant risk, and therefore should be evaluated at a higher discount rate. If the project is at FS level a discount rate of 15% might be acceptable. At any earlier stage the discount rate should be even higher.

    over at Jack’s site there was a very interesting discussion on decoupled NPV. (DNPV) by David Esponoza and Jeremy Morris. It’s worth reviewing their work to learn how they de-risk NPV.

Leave a reply