Content uploaded by Julian Poniewierski
Author content
All content in this area was uploaded by Julian Poniewierski on Aug 18, 2016
Content may be subject to copyright.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016 1
INTRODUCTION
It is not the purpose of this paper to be an introduction to cut-off grade theory or to be a treatise on
cut-off grades – for that the author refers readers to Hall (2014) or Lane (1997). It is the purpose of
this paper to elaborate on a few interesting aspects and to hopefully trigger the reader’s interest into
a journey of discovery in the fi eld of cut-off grade knowledge and application.
The author notes that despite Ken Lane’s theories on how to optimise net present value (NPV) by
judicious choice of cut-off grades having been in existence for more than 50 years (Lane, 1964; Lane,
1997), the mining industry remains wedded to the use of break-even cut-off grades for stating ore
reserves. It has become evident to the author that the mining industry is unlikely to ever be weaned
off the use of break-even cut-off grades. It is therefore appropriate that some of the perils and pitfalls
associated with the use of break-even grades, resulting in what the author calls ‘negatively geared
ore reserves’, are outlined so that they can be avoided.
Negatively Geared Ore Reserves –
A Major Peril of the Break-even Cut-o Grade
J Poniewierski1
ABSTRACT
Recognising that the mining industry is wedded to the break-even cut-off grade rather than
a cut-off grade that optimises value, the author examines some perils and pitfalls in the use
of break-even grades that lead to what the author calls ‘negatively geared ore reserves’.
Negatively geared ore reserves occur when a portion of the ore reserves actually has a
negative value and is effectively being subsidised by the more profi table portion of the ore
reserves. This situation is generally diffi cult to detect with the level of information available
to the public investor. The ore reserves can still be quite profi table in total, but now possess
a level of risk that may not be appreciated.
The risk profi le (fi nancial risk) of a negatively geared ore reserve is far greater than for a
non-negatively geared ore reserve. Like a negatively geared investment property in a falling
property market (for readers familiar with the Australian tax minimisation investment
strategy), when the price of a commodity falls, a company that possesses a negatively
geared ore reserve loses value at a magnifi ed rate relative to a company that has a neutral
or positively geared ore reserve. In the extreme case of a major commodity price reduction,
as happened with the gold price in mid-2013, the negatively geared ore reserve can easily
become negative value in total, not just in part.
The negatively geared ore reserve is sometimes intentional – for instance in order to
increase published ore reserve tonnages and meet the CEO’s key performance indicators –
but is also often unintentional, resulting from various errors in the break-even calculation.
Common errors in calculating a break-even grade include using fi xed metallurgical
recoveries, not including allowances for ideal laboratory versus real plant results, ignoring
available operational data and using modal rather than mean parameter values in cost and
performance calculations.
This paper examines a common value distribution profi le in many ore reserves and the
effect of common break-even grade errors on this profi le.
1. FAusIMM(CP), Principal Mining Engineer, AMC Consultants Pty Ltd. Email: jponiewierski@amcconsultants.com
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
2
Negatively geared ore reserves occur when a portion of the ore reserves actually has a negative
value and is effectively being subsidised by the more profi table portion of the ore reserves.
The negatively geared ore reserve is often a result of various errors in the break-even calculation.
A number of the more common errors will be examined in this paper.
By defi nition, negatively geared ore reserves imply that there is value lost that need not occur if the
negative value portion of the reserve is eliminated. Of course, this has to be taken within the context
of the whole mine plan.
The risk profi le (fi nancial risk) of a negatively geared ore reserve is far greater than for a non-
negative geared ore reserve. Like a negatively geared investment property in a falling property
market (for readers familiar with the Australian tax minimisation investment strategy), when the
price of a commodity drops, a company that possesses a negatively geared ore reserve loses value
at a magnifi ed rate relative to a company that has a neutral or positively geared ore reserve. In the
extreme case of a major commodity price reduction, as happened with the gold price in mid-2013,
the negatively geared ore reserve can easily become negative value in total, not just in part.
Executives should recognise that the inclusion of subsidised and marginal material in their mine
plans increases the risk that modelling of the project cash fl ows will demonstrate a poor response to
downside price, recovery and cost assumptions and will make securing of project fi nance both more
diffi cult and more expensive.
DISTRIBUTION OF VALUE
Although it is not often appreciated, a key feature of many stated ore reserves estimates is that a
signifi cant portion of the ore reserves tonnage is marginal in value.
While this is well known and understood by mining engineers heavily involved in the preparation
of ore reserve estimates and strategic long-term planning, it has become apparent to the author that
the vast majority of mining executives do not understand how marginal much of their company’s
ore reserves actually are, especially those executives from a fi nancial or legal background.
Figure 1 presents an example of the distribution of an ore reserve tonnage by value for an open pit
copper–gold mine (Mine A). The distribution shown is typical of many stockwork-style orebodies
when a break-even cut-off value (grade) is used to estimate the ore reserves. A signifi cant portion of
the ore reserve tonnage is of low or marginal value.
The distribution of total value for this ore reserves estimate is shown in Figure 2.
In this particular distribution, an interesting symmetry was present. The bottom 27 per cent of the
ore reserve held 6 per cent of the value. The equivalent tonnage top 27 per cent of the ore reserve
held 60 per cent of the value, which equated to ten times the value for the same tonnage. In effect, a
lot of the ore reserve is being mined and processed for little more than ‘practice’.
FIG 1 – Distribution of tonnage by value in an ore reserve for a real and typical copper–gold open pit mine.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
NEGATIVELY GEARED ORE RESERVES A MAJOR PERIL OF THE BREAKEVEN CUTOFF GRADE
3
THE IMPLICATION OF AN ERROR IN BREAKEVEN CUTOFF GRADE
Using the cut-off value–tonnage distribution for Mine A, it can be seen that a marginal break-even
cut-off grade for the ore reserves estimate can have a signifi cant negative effect if there is an error in
calculating the break-even cut-off grade. Such an error would result in a signifi cant portion of the
stated ore reserves becoming negative in value.
If the error is equivalent to $5/t in value (be it a cost or revenue error, or a combination of both),
the distribution shown in Figure 1 would be modifi ed to a value distribution as shown in Figure 3.
In total, 27 per cent of the ore reserve tonnage is now in the -$5/t to $0/t value bin and a similar
tonnage is in the $0/t to +$5/t value bin. A total of 55 per cent of the ore reserve now effectively has
zero value.
In addition, the cost of capital works (eg waste stripping) justifi ed by this portion of the ore reserve
is destroying value such that a potentially signifi cant loss of value is occurring.
THE ERROR LEVERAGE EFFECT WITH A PRICE FALL
To highlight the risk of using a break-even grade (or break-even value) with an error in it, let us
examine the leverage effect of an error in the cut-off grade by assuming an ore reserve of 100 Mt
using the distribution of Figure 3 with an error in the cut-off value of -$5/t. The key values of
tonnage, total value (undiscounted) and the average value per tonne of such an ore reserve are
shown in the second column of Table 1.
FIG 2 – Distribution of value in an ore reserve for the same copper–gold open pit mine.
FIG 3 – Distribution of tonnage by ‘true’ value in the example ore reserve with a $5 value error.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
4
In addition, two possible alternative ore reserves with different cut-off values are given in Table 1.
Note that the total value for the case of a +$5/t cut-off value is effectively the same as the total value
with the -$5/t cut-off value, although the ore reserve tonnage is only 45 per cent that of the erroneous
negatively geared ore reserve. Additionally a potential NPV has been calculated for each case using
a basic cash fl ow model with assumed but realistic movement schedules and costs. Note that the
lower tonnage, higher cut-off case has the greatest NPV. (For simplicity, the author has ignored the
likely effects of mine redesign with the changed cut-off values in this example.)
Table 2 shows what will happen should there be a commodity price fall equivalent to $15/t value.
It is important to note that from effectively the same starting value of $735 M, the negatively geared
ore reserve has dropped in value by $1500 M, which is 2.2 times the $674 M value drop of the +$5/t
cut-off value ore reserve.
It is also interesting to note that after the $15/t equivalent price fall, the +$5/t ore reserve is now
also a negatively geared ore reserve, but is still positive in value.
WHAT DOES A $5 VALUE ERROR LOOK LIKE?
It is useful to understand what a $5/t error in a break-even cut-off grade calculation might look
like. Consider the case for a gold mine.
At the time of writing, the gold price was trading between US$1200/oz and US$1250/oz. This
range is similar to that used in reporting ore reserves as at the end of December 2012. For example,
Newcrest used US$1250/oz for estimating its 2012 ore reserves and Barrick used US$1200/oz for
its Cortez ore reserves. In 2013, Newmont used US$1400/oz and subsequently announced in early
2014 that it would use US$1300/oz for its long-term pricing assumption for ore reserve reporting
purposes in 2014.
At US$1200/oz, a US$5/t error in revenue can be caused by a recovered gold difference of 0.13 g/t
(at the Newmont price of US$1400/oz, a difference of 0.11 g/t results in a US$5/t error). Similarly, a
cut-off grade calculation error of 0.13 g/t will result in a US$5/t difference in expected revenue. The
loss in profi ts associated with this loss in revenue will be greater than US$5/t as a result of the costs
incurred to generate the expected US$5/t of revenue.
For copper at a price of US$2.40/lb, a US$5/t error can be caused by an error in grade of 0.10 per cent.
At the 2011 peak price of US$4.5/lb, a US$5/t error will result from an error in grade of 0.05 per cent,
which is effectively equivalent to the rounding error in most ore reserve calculations.
In subsequent sections of this paper, it will be seen how commonly observed errors are inducing
errors of such value or more.
Cut-o value -$5/t $0/t +$5/t
Tonnage (Mt) 100 73 45
Total value (undiscounted) ($M) 735 803 734
Average value per tonne ($/t) 7.35 11.05 16.34
Potential NPVa (M$) 77 146 152
a. pre-tax, 8 per cent discount rate, 5 Mt/a processing cost, 250 Mt total movement, 2 $/t mining cost.
TABLE 1
The example ore reserve at three di erent cut-o values.
Cut-o value -$5/t 0$/t +$5/t
Tonnage (Mt) 100 73 45
Total value (undiscounted) ($M) -765 -287 60
Average value per tonne ($/t) -7.65 -3.95 1.34
Change in total value ($M) -1500 -1090 -674
TABLE 2
The example ore reserve at three di erent cut-o values after a $15/t equivalent price fall.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
NEGATIVELY GEARED ORE RESERVES A MAJOR PERIL OF THE BREAKEVEN CUTOFF GRADE
5
THE FIXED RECOVERY ERROR
The author has witnessed the erroneous assumption of fi xed recovery for all grades down to the
break-even cut-off grade on numerous occasions, but will discuss one specifi c actual case of an open
cut gold mine that we will refer to as Mine B.
A break-even cut-off grade is usually defi ned as the grade of material that pays for itself. This
also means that it should pay for itself when it is by itself. So the recovery of this material by itself
is the recovery that should be used in determining the break-even grade – not the recovery of the
total material when this cut-off grade is used (which results in the higher-grade material effectively
subsidising the lower-grade material).
In the case of Mine B, the feasibility study indicated a 94 per cent recovery for the planned ore
reserve gold head grade of 2.2 g/t. The mine operator used an elevated cut-off grade of 0.9 g/t Au
as a result of limited tailings storage capacity.
During construction, the joint venture (JV) partner (a 50:50 partnership) indicated that it wished
to take over long-term planning and ore reserves work. In doing so, the JV partner decided to
effectively ignore these topographic-related physical constraints and declare the ore reserves based
on a calculated break-even cut-off grade of 0.4 g/t using the assumption of 94 per cent processing
recovery at this break-even grade. Note that this implies a break-even cut-off recovered grade of
0.38 g/t and a tails grade of 0.02 g/t. This enabled the JV partner to increase the declared ore reserves
tonnage by over 50 per cent.
For most deposits and minerals, it is generally accepted that mill recovery is not constant over a
range of varying grades and commonly decreases towards lower grades.
It is common for a fi xed tails grade to be used for recovery calculations, with the processing
recovery being determined by the difference between head grade and the fi xed tails grade. This is
not strictly true as the tails grade is likely to decrease at lower grades, but it is a good conservative
rule of thumb, remembering that it is diffi cult to go broke when making money.
In the Mine B example, the feasibility study processing recovery of 94 per cent at a head grade of
2.2 g/t implies a tails grade of 0.13 g/t. For a fi xed tail of 0.13 g/t, the processing recovery at the
0.4 g/t break-even cut-off grade used by the JV partner would be 68 per cent.
At the calculated break-even recovered grade of 0.38 g/t, the fi xed tail of 0.13 g/t implies a break-
even cut-off grade of 0.5 g/t. The resulting break-even cut-off grade error of 0.1 g/t is equivalent to
a US$5/t error at a US$1400/oz gold price (and a US$4.3/t error at a US$1200/oz price).
If the ore reserves estimate for Mine B had a value distribution similar to that shown in Figure 1,
approximately 20 to 30 per cent of the ore reserves estimate would have negative value due to the
fi xed processing recovery error in the calculation.
It should be noted that one of the major key performance indicators (KPIs) of the CEO of the
JV partner, and therefore also the KPI for the executives and senior management, was growth in
ore reserves. Signifi cant bonuses were paid to this executive team that year and for a number of
subsequent years.
Before the reader starts objecting to the use of fi xed tails as being overly conservative, the author
notes that while a fi xed tail is not always true, fi xed tail cases do commonly exist. Some examples of
these from the open literature include:
•in an update on the Kanmantoo deposit, Hillgrove Resources (2013) stated that the ‘process
recovery model was updated with a fi xed tail model utilised for primary ore recovery [as this
gave] a better refl ection of operational performance measured since the commencement of
operations’
•Kinross’ Round Mountain deposit uses a fi xed tail of around 0.2 g/t up to a head grade of 1.4 g/t
after which it reverts to a fi xed processing recovery of 85 per cent, ‘developed from the mine’s
operating history [and] consistent with life of mine operating experience’ (Kinross, 2006)
•recent test work for the Mankarga 5 gold deposit ‘indicated that the recovery of gold [was]
governed by a fi xed tail’, varying by weathered state from 0.10 g/t for strongly oxidised material
to 0.70 g/t for fresh ore (West African Resources, 2015).
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
6
Of course, in some instances, the use of a fi xed tail can also be an overly conservative position, but
again the author emphasises that it is an assumption that has a lower fi nancial risk. An example of a
situation where a fi xed tail is overly conservative is shown in Figure 4, which is based on laboratory
test data for Pascua-Lama (Silver Wheaton, 2009). Figure 4 also shows the error of assuming a fi xed
processing recovery. A percentage tail model is often required to best refl ect a deposit’s recovery
characteristics.
It is therefore highly recommended that during test work programs, the processing recovery should
be measured over a range of grades at lower than expected average grades and particularly down
to expected break-even grades (and at the grade of potential low-grade stockpiles if a stockpiling
strategy is proposed).
LABORATORY VERSUS REAL MILL PERFORMANCE
Continuing with the Mine B example, as well as the error associated with the assumption of a fi xed
recovery, it has become apparent after almost six years of operation (at the time of writing) that
using an unadjusted tail grade as determined through the laboratory-based test work has resulted
in considerable error in assessing the actual mill performance. As shown in Figure 5 (taken from
publicly available quarterly reports), the actual fi xed tail grade of Mine B is 0.25 g/t, which is almost
double the 0.13 g/t implied in the feasibility study.
FIG 4 – Example of the error of a xed recovery assumption and the conservatism of a xed tail recovery assumption..
FIG 5 – Historical tails gold grade for Mine B as per quarterly reports versus the feasibility study implied tails grade.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
NEGATIVELY GEARED ORE RESERVES A MAJOR PERIL OF THE BREAKEVEN CUTOFF GRADE
7
This increased tails grade implies a change in the break-even cut-off grade to 0.6 g/t, rather than
the 0.4 g/t used by the JV partner. For a value distribution as shown in Figure 3, this implies that
potentially half of the stated ore reserve was actually negative in value due solely to the processing
recovery assumptions.
As a fi nal note on the recovery error issue, nothing beats reality to confi rm a recovery model.
IGNORING DATA
Puzzling though it may be, the author has encountered numerous cases where the ore reserves
estimates for operating mines have been found to still use parameter values fi rst expounded in the
feasibility study, despite the availability of several years of operating data.
As an example, consider Mine C, which was using a fi xed tails gold grade of 0.11 g/t, but there
was no discussion of the provenance of this value. Upon further investigation, it transpired that this
value was the one used in the feasibility study some seven years earlier and no-one had bothered to
reconcile this with the past fi ve years of operating data.
The author compiled shift-based tails grade data for the previous 18 months (acquired with an
email to the mill manager) and produced the distribution shown in Figure 6 (all in half an hour).
The tails grade of 0.11 g/t used to estimate the ore reserve had occurred less than four per cent of
the time in the previous 18 months. The median tails grade was a somewhat larger value of 0.26 g/t.
Referring back to the earlier discussion on value distribution in a break-even ore reserve, this
difference alone implied that potentially 20 to 30 per cent of the ore reserve was likely to be negative
in value.
As such, the author is usually suspicious of stated ore reserve parameters at operating mines if there
is no historical comparison or analysis presented to put a parameter into context. It is recommended
that such data is presented in ore reserve estimation documentation to provide context on the
parameter chosen, as shown in Figure 7.
MEANS, MEDIANS AND MODES
A number of years ago, the author had an epiphany on some of the underlying problems in feasibility
studies and ore reserves estimates. It was observed that median and modal values were being used
as input parameters rather than mean values. This is an issue that is well known in the area of project
management literature (Emhjellen, Emhjellen and Osmundsen, 2001).
The statistical distribution of many parameters used in ore reserves estimates (and in concomitant
feasibility studies) are heavily skewed (similar to the tails grade values in Figure 6; skewed to the
FIG 6 – Distribution of 18 months historical shift tails gold grade for Mine C.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
8
right or ‘positively’ skewed), and study managers have a bias for using modal or median values
rather than mean values as they don’t believe that the larger mean values are true.
In a brainstorming session for parameters and risks, it is invariably modal or median values that
are generated. Yet empirical studies show that the distribution of actual/estimate cost data nearly
always has a very long, bimodal tail on the high side (Hollmann, 2014).
To illustrate this effect, consider the drill and move times for a raise bore rig drilling stope slots at
Kidd mine, as shown in Figure 8.
In this example, the median value for the drill and move time for a stope slot raise-bore hole is
11.5 days (ie it will take less time 50 per cent of the time and more time 50 per cent of the time).
FIG 7 – Example of an ore reserve recovery parameter against historical data (MMG, 2013).
FIG 8 – Distribution of raise bore slot drill and move time at Kidd mine (Resoort, 2010).
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
NEGATIVELY GEARED ORE RESERVES A MAJOR PERIL OF THE BREAKEVEN CUTOFF GRADE
9
However, the distribution is skewed to the right as a result of extraneous events (for example, the
vent bag needing repair or pipes being blown down), and the mean time is 13.6 days.
The mean result is 18 per cent greater than the median time, which is the value that the author
generally observes being used in planning and scheduling (although a far more aggressive value is
often mandated by management). As a result, the fi xed labour costs will be 18 per cent greater than
planned and, on average, each scheduled stope start time in the long-term mine plan will slip by
2.1 days, signifi cantly impacting on the annual planned production.
An interesting aspect of this median versus mean issue is the question of, ‘what value should be used
in scheduling?’ The author suggests that the median time should be used in short-term scheduling as
it will be the value most likely to occur, with subsequent short-term plans adjusted for the events that
cause the mean to be greater than the median; however, for long-term scheduling, the mean value
should be used.
Ultimately, ore reserves (and projects in general) are made up of numerous items that are positively
skewed. The use of median or modal values for even a few of these inputs will result in an incorrect
estimate. Indeed, even the sum of means from distributions that are positively skewed will not equal
the mean of the sum. The central limit theorem dictum that we absorbed as students – that the mean
of the total will equal the sum of the means – is only true if, and only if, every single distribution of
those means is normal, which is clearly not the case.
Therefore, it is obvious that there will be a problem with most of our estimates. By default, the
estimated mean of a system of individual components will be underestimated. This underestimate
ignores the further effects of project selection bias:
•typically, the projects with the most optimistic internal cost estimates will be pursued by the
investing fi rm
•during tender selection, competition sees to it that tenders with pessimistic and realistic cost
estimates are ruled out (Emhjellen, Emhjellen and Osmundsen, 2001).
Is it any wonder that ten per cent of large projects overrun their budgets by 70 per cent or more
(Hollmann, 2014)? Or that ore reserves estimates are often not truly representative of reality?
These same project biases also affect ore reserve estimates, resulting in a negatively geared ore
reserve.
SIMPLE COST ERRORS AND OMISSIONS
While in the previous section it was shown that total costs are likely to be underestimated due
to the positively skewed statistical distribution of cost estimates, the author has also come across
numerous examples of single individual costs just being simply wrong or being omitted through
lack of oversight. These have ranged from simple errors in data use, such as trucks being scheduled
with a full load of ‘dry tonnes’ rather than ‘wet tonnes’ in a tropical mine environment, to errors of
cost basis, such as exclusion of equipment ownership (sustaining capital recovery) costs in operating
cost estimates.
A specifi c example of cost errors and their potential effect on the ore reserves will be examined for
a gold mine, Mine D.
As shown in Table 3, Mine D estimated its total costs in year X-1, the year before construction
started. In year X, when the mine was under construction and key management and supervisory
roles had been fi lled, the cost estimate was redone by the new but highly experienced staff.
Item Year X-1 Year X Increase (%)
General and administration ($M/a) 5.0 12.0
General and administration ($/t) 1.28 3.56 178
Mining ($/t) 2.24 2.28 2
Milling ($/t) 8.92 11.90 33
Total ($/t) 12.44 17.74 43
TABLE 3
Example of a cost base change in the space of one year at Mine D.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
10
It can be seen that the cost estimates increased by slightly over $5/t. Again, referring back to the
break-even cut-off grade-based ore reserve value distribution shown in Figure 3, it is possible that
some 20 to 30 per cent of Mine D’s previous year’s stated ore reserve is now negative in value.
The major causes of Mine D’s cost increases were as simple as:
•omitting freight and customs costs
•severely underestimating the maintenance cost of 30+ km of unsealed access road in a wet tropical
environment
•omitting the maintenance costs of other infrastructure access roads
•misinterpreting the test data for estimations of carbon-in-leach cyanide consumption
•omitting the operational costs of changes to the mill design equipment selection
•power cost increases.
SUSTAINING CAPITAL COSTS
Sustaining capital costs is an area that is often omitted from cut-off grade calculations.
As stated by Hall (2014), the ‘…fi nancial accounting distinction between capital and operating
costs is largely irrelevant for cut-off derivation purposes. These classifi cations have been developed
for public accounting reports…’, not for engineering costing purposes.
There is a misconception that these costs need not be considered. However, sustaining capital costs
are an artefact of an accountant’s defi nition, and in reality they are ‘lumpy’ or irregular operating
costs used to maintain operating capacity and just happen to take more than one year to consume.
It is ‘by virtue of its spending pattern rather than its behaviour’ (Hall, 2014) that sustaining capital
costs are termed ‘capital’ by accountants. Operating expenses are recognised in a company’s accounts
in the year that they occur, whereas sustaining capital expenses are accumulated and depreciated
over the ‘life’ of the asset to which the capital expense relates.
One area of sustaining capital costs that is often omitted is the cost of the tailings storage facility
(TSF). The TSF incurs a cost to build and to maintain. Filling the TSF with tailings derived from
marginal break-even ‘ore’ that has not included an allowance for the cost of wall-lift builds will
result in a loss from the most marginal material.
Figure 9 shows an analysis of a multilift TSF with respect to lift construction volume versus
cumulative storage capacity achieved by each lift. It indicates that an average of 0.24 m3 of lift
embankment volume has to be constructed for each tonne of tailings placed in the TSF. For this
example, using this average TSF embankment volume per tonne of tailings stored and the average
lift construction costs per embankment volume placed, it was determined that the TSF-related cost
for inclusion in the break-even grade calculation was US$0.77/t of ore.
FIG 9 – Example of tailings storage facility lift embankment volumes versus tailings storage capacity.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
NEGATIVELY GEARED ORE RESERVES A MAJOR PERIL OF THE BREAKEVEN CUTOFF GRADE
11
A good starting point when deciding what costs to include in an incremental cut-off policy is to
include all costs that are incurred as a consequence of the decision to process that particular piece of
rock. For example, mills wear out by processing tonnes, so their sustaining capital is relevant. TSFs
are only built higher if there are tonnes to place, so these costs are relevant. Where a mine is mill-
constrained, a decision to process a tonne extends the mine life incrementally so fi xed general and
administration costs are relevant. As a mine is expanded, the closure and rehabilitation costs will
increase, suggesting that there should also be a component of these costs included.
CLUES TO NEGATIVELY GEARED ORE RESERVES
As evidence that many ore reserves estimates are tending to be negatively geared, a study by Creech
(2014) reported that the average quantity of reserves reported prior to closure by coalmines in NSW
and Queensland was equivalent to 20 to 30 years of production of the average mine studied, yet
75 per cent of the mines gave ‘uneconomic operations’ as reasons for closure.
Given that the defi nition of an ore reserve is that it is the ‘economically mineable part of a resource’,
this tends to suggest the prevalence of negatively geared ore reserves and the likelihood that the
reserves left at closure were substantially the subsidised portion of that negatively geared ore reserve.
The presence of any of the errors discussed in this paper indicates the potential existence of
negatively geared ore reserves. For an operating site, a lack of comparative data for justifi cation of
parameters should give cause for suspicion.
However, perhaps one of the biggest causes of negatively geared ore reserves is executive KPIs
that provide generous fi nancial rewards that relate to the size of ore reserves – be they tonnes-based
or metal (ounces)-based. These executives are able to enrich themselves by meeting a KPI while
increasing the risk profi le of the company. This risk is not typically appreciated by board members
setting the ‘increase ore reserves’ KPI. A secondary driver is the analysts that ‘reward’ companies
with share price target upgrades for reserve growth without due consideration of the value of those
ore reserves. A close examination of some of the major losers of market capital during the 2013 gold
price correction will uncover a number of such companies.
Sometimes, all that is needed to suspect negatively geared ore reserves is an examination of cost
performance against grade. Figure 10 shows the total cost versus the head grade for a current
operating gold mine (Mine E) based on publicly available data.
From Figure 10, it can be seen that there is a general relationship, though not statistically strong,
between the total cost and the mine’s head grade. The most recently published ore reserves for
Mine E (as at the end of December 2014) had a grade of 1.7 g/t at a gold price of A$1470/oz. The
relationship in Figure 10 suggests that the most likely total cost for an ore reserve at such a grade
is closer to A$1900/oz, although the range could vary from A$1300/oz to A$2250/oz. It can be
FIG 10 – Total cost of production versus gold head grade for Mine E.
PROJECT EVALUATION 2016 / ADELAIDE, SA, 8–9 MARCH 2016
J PONIEWIERSKI
12
confi dently said that Mine E has a signifi cant portion of negative value ore reserves (if in fact it isn’t
close to 100 per cent).
CONCLUSIONS
Due to the preponderance of marginal-value material in many ore reserves estimates that rely on the
use of a break-even cut-off grade, only small errors in the inputs are required to cause a large portion
of an ore reserves estimate to have negative value. In some cases, rounding errors alone can cause a
signifi cant percentage of an ore reserve to have negative value.
An ore reserves estimate that contains negative-value tonnages can still quite legitimately be called
an ore reserve under the defi nition of the JORC Code (2012) if it is still economic as a whole. However,
such negatively geared ore reserves will put a company at greater fi nancial risk should commodity
prices fall, as has been increasingly witnessed over the last year of tumbling commodity prices. Such
experiences are analogous to the value of a negatively geared house in a falling property market.
The most common errors seen by the author that lead to negatively geared ore reserves include:
•the assumption of a fi xed processing recovery down to an inappropriate low grade
•ignoring real data, be it laboratory derived or operation derived
•using median or modal values rather than mean values for parameters informing the analysis of
the cut-off grade
•cost errors, often from inexperience
•cost omissions, such as ignoring sustaining capital
•deliberate aggressiveness by companies to increase the ore reserve size without new discoveries.
Ideally, the ore reserves estimate should be based on a cut-off grade that maximises the value
objective of an organisation (such as NPV). However, if a break-even cut-off grade must be used, it
is imperative that the cut-off grade calculation is undertaken using:
•the best available data, preferably based on actual operation or suitably modifi ed laboratory data
•a full set of cost data that includes all contributory costs.
The break-even cut-off grade calculation should be prepared by a Competent Person with
signifi cant experience and knowledge of the key parameters and their impact on the outcome for
the ore reserves estimate and a mature ability to reason across the risks involved.
REFERENCES
Creech, M, 2014. Reserves, reserves and not a tonne to mine – a study of reserves reported prior to mine closure, in Mineral
Resource and Ore Reserve Estimation, second edition, pp 627–634 (The Australasian Institute of Mining and Metallurgy:
Melbourne).
Emhjellen, K, Emhjellen, M and Osmundsen, P, 2001. Cost estimates and investment decisions, presentation to the 24th
International Conference of International Association for Energy Economics, Houston, Texas, April 25–27, 12 p.
Hall, B, 2014. Cut-off Grades and Optimising the Strategic Mine Plan, 311 p (The Australasian Institute of Mining and
Metallurgy: Melbourne).
Hillgrove Resources, 2013. Kanmantoo copper mine ore reserve – 25% increase in contained copper, ASX release,
1 November 2013, 17 p.
Hollmann, J K, 2014. Risk analysis at the edge of chaos, 2014 AACE international technical paper, 19 p.
JORC, 2012. Australasian Code for Reporting of Exploration Results, Mineral Resources and Ore Reserves (The JORC
Code) [online]. Available from: <http://www.jorc.org> (The Joint Ore Reserves Committee of The Australasian
Institute of Mining and Metallurgy, Australian Institute of Geoscientists and Minerals Council of Australia).
Kinross, 2006. Round Mountain mine, technical report, 30 March, 85 p.
Lane, K F, 1964. Choosing the optimum cut-off grade, Colorado School of Mines Quarterly, 59(4):811–829.
Lane, K F, 1997. The Economic Defi nition of Ore, second edition (Mining Journal Books: London).
MMG, 2013. Mineral Resources and Ore Reserves statement 2013, technical appendix, p 156.
Resoort, P, 2010. The root causes of stope slippage at Kidd Mine, Canada, MSc thesis (unpublished), Delft University of
Technology.
Silver Wheaton, 2009. NI 43–101: technical report: Pascua-Lama Project, Region III, Chile/San Juan Province, Argentina,
SRK Project Number 2CS019.003, 9 September, 224 p.
West African Resources, 2015. Technical report 43–101 pre-feasibility study, Mankarga 5 gold deposit, Tanlouka gold
project, Burkina Faso, 14 April, 204 p.
