Risk Analysis: The Importance of Reserve Estimation

M Armstrong 1, N Champigny2 and P H Grimley3

PREAMBLE their point of view can be catastrophic whereas the upside is

limited. In contrast to this, mining companies while sharing in
To a bank that granted his request for a loan, the late American downside losses, stand to reap more rewards from projects on the
humorist Robert Benchley responded, 'I don't trust a bank that upside. Governments stand to lose potential revenues, electoral
would lend money to such a poor risk'. support and suffer environmental degradation. Given these
TIlls quotation is unlikely to be the nonn among mining divergent objectives it is hardly surprising to find that the three
companies that secure project fmancing. However, a poor participants analyse risk in different ways.
appreciation of one of the key risks of a mining project, the
reserves, often results in <disastrous fmancial consequences. This VALUE AND RISK
conclusion is the subject of the discussion and analyses presented
Mission statements of mining companies often refer to increasing
in this paper. value for its stakeholders and customers. One representation of
value and its relationship to risk (developed by Coopers &
DIVERGENT OBJECTIVES Lybrand), is the following formula:
Before trying to estimate or characterise risk, it is important to
identify the different and often divergent aims of three important Customer & Stakeholder Value = Quality x Capacity
stakeholders in mining projects: mining company management, Cost x Time x Risk
financing agencies and government authorities. Company
management and Directors are required by shareholders to A higher product quality (eg product delivered on specs and on
represent the latter interest by ensuring survival of the company time) and a higher capacity to deliver the product, translates into
(preservation and enhancement of shareholders capital) and by increased value. Adding value is also possible through cost
producing a return on that capital (profits). In this process, they cutting. To cut cost many mining organisations focus on
must act in accordance with legal and ethical standards of reduction in the workforce, which in some cases may be
corporate behaviour. counterproductive to the team building effort needed to ensure
The risk/reward scenario for fmancing agencies is different. high quality. A 'slash and burn' approach may in fact be a
Their profit is the difference between the interest they receive on short-term measure which at best leads to a lack of emphasis on
money lent and the interest they pay to depositors. As financing longer term profitability and survival if essential activities such as
agencies rarely have equity in the project, they do not stand to exploration and R&D are curtailed. Time is also a critical
make windfall profits when things go well (for example, high variable. One only has to consider the long lead time in
conunodity prices) but can sustain serious losses if projects fail. identifying and developing projects. In addition, during
For example, if a project loan worth hundreds of millions of production, often only a small proportion of the total time from
dollars goes 'sour', it takes many successful loans to make up the the initial blast to the shipping of concentrate, can be classified as
loss. This is why bankers approach the problem of analysing risk a value-added activity. Much of the time the material is either
from such a conservative point of view. They will try to stockpiled or re-handled or transported over significant distances
minimise the downside risk while simultaneously, not foregoing (ie non-value adding activities).
sound lending opportunities. Risk has a high degree of interdependence with the other
The point of view of governments to mining ventures is variables. For example, it is tempting to increase capacity and in
different again. Taxes and mineral royalties are major sources of parallel, lower cost and time. However, this often creates an
revenue in many countries and beyond this, mining provides unacceptable risk level. A better approach may be to minimise
direct and indirect employment for many people. Even if a risk by achieving increases in quality and controlling time.
government has equity in a project (as often happens in
developing countries), it rarely invests cash. While governments RISK - THE 6 Ps
may not stand to lose in the same way as a mining company or a
financial institution, a failed project may eliminate planned for After reading share prospectuses for new mining projects one
revenues in the form of royalties and taxes as well as the possible often develops the impression that the company management has
destruction of a natural asset. In addition, there is a potential for the situation completely in hand. However, looking at the long
environmental damage which may be a drain on government list of mining failures in recent years in many mineral producing
funds and cause an electoral back-lash unless appropriate countries, it is apparent that mining continues to be a risky
environmental control measures and adequate clean-up funding venture. A representation of risk elements may be summed up
are provided by the developers. succinctly by the following 6 Ps:
To summarise the situation, three important participants in • Project risk has a direct correlation with the uncertainty in
mining projects stand to win or lose different things. Financing estimating the project's fundamental asset, ie its reserves.
agencies tend to be more conservative because the downside from Project risks also include among other risks, location
(remoteness, lack of infrastructure) and availability of trained
labour;
1. Head, Mining Group, Centre de Geostatistique, Ecole Nationale • Process risk relates to the ability to achieve projected mining
Superieure des Mines de Paris, Fontainebleau, France. and milling recoveries and to produce a saleable mineral
product;
2. Manager, Mining Practice, Coopers & Lybrand, Toronto, • People risk is the experience of the management team in
Canada. executing projects of a similar type. If management's only
strength is in identifying orebodies, project finance may be
3. Senior Consulting Associate, Mining Practice, Coopers & very difficult to obtain and may require the appointment of
Lybrand, Toronto, Canada. experienced mine operators;

4th Large Open Pit Mining Conference Perth, 5 - 9 September 1994 285
 M ARMS1RONG, N CHAMPIGNY and P H GRIMLEY

• Pollution risk is rarely the inability to meet specific Similarly in his listing of recent North American nurung
environmental standards but rather the uncertainty in failures, Canadian investment analyst and mining executive
foreseeing the impact of rapidly changing environmental Harquail (1992) attributes the disasters to 'lack of reserves' in
regulations and reacting to them. In many major mineral more than half the cases.
producing countries decommissioning is becoming more
In assessing the impact of risk, the trade-off between risk and
costly and unpredictable;
reward should be considered, and this varies throughout the life
• Price risk is one of the most significant factors by virtue of its of a project. Figures 1 and 2 depict the project risk and then the
major impact on all projects. Coopers & Lybrand's recent corporate risk at the various stages of a project from the
mining industry surveys suggest that accommodating price
exploration phase through to decommissioning. Although the
cyclicality is the prime issue for executive management. Price
risk of project failure is highest at the outset because of a number
also receives utmost attention from financing agencies; and
of unquantified variables, the corporate risk is low then because
• Politics or the political risk depends on a country's political of the magnitude of funds required. From a fmancial point of
and regulatory stability. Motivation to assess this risk results
view the risky time for a corporation is from after the start of
from the industry's experience with nationalisation and
expropriation and changes in financial parameters (royalties development until full projection levels are reached, because of.
and taxation). the magnitude of the funds required. Project failure could result
in the destruction of the corporate entity. This is particularly true
of a company whose reserves constitute its only asset and is one
PROJECT RISK reason why project fmancing by commercial banks may be
TIris paper focuses on the first of these, the project risk, and in difficult to arrange if there is no substantial collateral other than
particular on the role of reserve estimation in assessing project the reserves.
risk. Project risk is the level of uncertainty for a specific project, To further highlight the impact of reserve estimation on risk,
to become an economically viable mine. In other words, it is the an examination follows of the:
risk of project failure. Several presentations at this conference
provide more exhaustive discussions of the other risk factors. • recent deliberations of an international 'Groupe de
Reflexion'; and
Why concentrate on this rather than the other factors? Because,
in the authors' opinion, it is a major cause of many recent project • some case studies.
failures and because better practice could limit the extent of the
losses. 1bis point of view is shared by others. In a review on INTERNATIONAL 'GROUPE DE REFLEXION'
how lenders approach risk analysis in mining projects Worth and
Haystead (1991) make the following comment: In rnid-1991 M Armstrong initiated a 'Groupe de Reflexion'
(Champigny and Annstrong, 1994) on reserve estimation with the
The assessment of the ore reserves is the most belief that the 'Groupe' is a useful initiative to:
important point of any technical evaluation.
Errors in the project reserve estimate derived • emphasise financial consequences of inappropriate reserve
from optimistic characterisation of continuity or estimation using case histories;
mineability, or the incorrect application of • exchange views on the level of accuracy obtained by
calculation methodology are usually fatal. available estimation methods including geostatistical
Money and time can generally fIX errors made in methods; and
the estimation of certain operating parameters. • provide an infonnal and international forum for debating
Money cannot create ore that was not put in reserve-related issues.
place by Providence.

PROJECT
RISK

Exploration Feasibility Development Production Decommis-

sioning
TIME

FIG 1.

286 Perth, 5 - 9 September 1994 4th Large Open Pit Mining Conference
 RISK ANALYSIS: THE IMPORTANCE OF RESERVE ESTIMATION

CORPORATE
RISK

?
---_?
Exploration Feasibility Development Production Decommis-
sioning
TIME

FIG 2.

The 'Groupe' is organised in the following way: • need for a multi-disciplinary approach; the need to have
• contributions from members are anonymous; technical, management and financial professionals working
together is a major theme. This means close interaction
• the Secretary and an inner group prepare and send out an
between disciplines to define estimation parameters. To have
initial thought-provoking text for comment by a fixed
a team of geologists, mining engineers, geostatisticians,
deadline;
metallurgists and fmancial advisors responsible for and
• the Secretary summarises the comments received and signing off on the reserves would represent significant
prepares a new text; and progress in the preparation of realistic reserve estimates and
• after a certain period, a statement of problems and current the minimisation of project risk;
thinking are summarised in a final report for wide circulation. • use of economic criteria; the impact of economic criteria is
. Fr?m 1991 to the time of writing, 13 texts have been reflected in the choice of a cut-off grade and it is not
distnbuted for comments. The 'Groupe' has more than 120 surprising to see considerable debate on the
members on its mailing list from all regions of the world. The detennination/selection of cut-off grades. The maximisation
following is a discussion (and commentaries from the authors in a of a project's net present value using an ideal/correct cut-off
project risk context) of the issues that have dominated the often receives more attention than the orebody's robustness to
submissions from the 'Groupe': cut-off fluctuations both in terms of reserve figures, mineable
• sample representativity and a{:curacy; without the collection configuration and economics; and
of samples of appropriate size and geometry, the use of an
• establishment of industry ethical standards; the summation of
adequate sampling pattern and the performance of accuracy
the views is the quotation: 'The first duty of a professional is
~d precision checks, the only conclusion is 'garbage samples
that he should have due regard to the public he is reporting
m, garbage reserve figures out' implying a potentially serious
to'. Improved reserve estimation practices can only lead to a
bias and risk for the project;
better appreciation of the project risk by those who have the
• importance of geologi~al input; there is a significant danger responsibility of assessing reserves.
in carrying out reserve estimation without any prior
An international 'Groupe de R6flexion' has examined some
knowledge of the geology. Ignoring geological data is a risk
problems and issues that are part and parcel of the reserve
that no project can afford;
estimation process and which have been the underlying cause of
• applicability of geostatistical and simulation methods; the many mining project failures over the recent past. Bringing these
trends in the use of geostatistics are that it is applied either (a)
problems and issues to the attention of a broad international
with considerable geological input; Cb) as a fashionable
accessory, or (c) as a security blanket. The latter two modes audience may be the first step in minimising project risk and
of application translate into adding unnecessary risk. defIDing what are the most appropriate estimation practices.
Geostatistical simulation is a tool which creates many
alternate grade scenarios, and provides a possible outcome of CASE STUDIES
the mined grade from any given production sequence. This is
useful for risk analysis. The average of a high number of The following case studies are a representation of unsound
simulations is equal to the result provided via geostatistical reserve estimation practices. For more case studies the reader is
estimation. This raises important questions: does one referred to the pioneer work of King et al (1982) published in
simulate or estimate? How many simulations does one Australia. In North America more recent accounts are those of
need? Finally, how does one assess several alternative Clow (1991) and Harquail (1992).
orebodies?;

4th Large Open Pit Mining Conference Perth. 5 - 9 September 1994 287
M ARMSlRONG, N CHAMPIGNY and PH GRIMLEY

Gold project asks what is the objective of drilling. The only possible
procedure is to go from one category up to the next. This often
m this open pit gold project presented for project financing, the results in the overdrilling of deposits.
exploration team had identified an in situ gold resource by
Recent work by one of the authors was on an orebody drilled
surface diamond drilling. One area, interpreted as having a
e.n a 100 m by 100 m grid. m-fill drilling was about to take place
somewhat higher grade than the average, was drilled by the
using a 50 m by 50 m spacing, before the planned production
exploration team at a significantly closer spacing than the rest of
start-up 18 months later. As one can imagine, the cost of the
the deposit The in situ resource estimations were passed on to
additional drilling, sampling and analysis was considerable. .
the mine planning department and the exploration team had no
When the exploration geologists were asked what they wanted to
further involvement. Because of a relatively high strip ratio, the
fmd out from the additional drilling, no one seemed to know. Nor
mine planning department designed an open pit which would
did they know whether the decision to exploit the deposit had
commence operations in the area of least overburden even though
definitely been taken or whether this was conditional on the
this area did not include the area of highest grade. Not until after
results of new drilling campaign. They only knew that when the
the feasibility study had been completed and the project
orebody reached category B their work as the exploration team
submitted for financing was it realised that initial production was
scheduled to come from an area about which little was known, ie was over and the deposit could be sent to the mine design
one of the least drilled areas. As a result of this perceived risk, . institute. Clearly there was a problem of compartmentalisation
here.
project fmancing was not forthcoming from at least one
commercial bank and other financing institutions demanded very The second problem was that the same procedure is adopted
stiff completion guarantees. Discussions between the exploration irrespective of the sample grades. For example, suppose the
team and the mine planning group at an early stage would have deposit is a gold orebody. Generally these orebodies are only
identified this potential risk area and would have resulted in economically interesting for an underground operation if the
either the supplementing of the information in the low average grade exceeds about 5 glt. So at 5 g/t the operation is
overburden area or modification of the mining sequence. probably marginal, at an average of 20 g/t the mine is clearly
profitable, and if it runs 50 to 100 g/t on average, then it is a real
bonanza. Therefore, project risk is proportional to the spread
Base metal mine between the cut-off and the average mine grade. For a narrow
The original reserve figures from this well-known deposit can spread, the project risk is likely to be higher and will need to be
make the most sceptical base m~tal investor ecstatic: several tens minimised by the collection of additional geological and
of million tonnes grading more than 20 per cent combined sampling information. Another requirement will be a careful
lead-zinc with significant silver, and mineable by open pit at a assessment of economic and technical parameters used in the
strip ratio of less than 1:1. At project start-up, senior calculation of cut-offs. For situations where the spread is very
management predicted that the project had the potential to be the large, and project economics are very robust to changes in
company's most profitable mine. Three years after start-up the cut-offs, a higher level of reserve risk may be acceptable. If the
mine is still struggling to get above breakeven. The problems do average grade is 5 g/t ± ten per cent, reserve risk is severe. On
not directly relate to the prediction of grades or tonnages but the other hand if the average is 50 glt ± 20 per cent, reserve risk
rather to the exclusion of appropriate geological and metallurgical may be low. Unfortunately the relationship between project
data for reserve estimation. m fact, reconciliation between economics and risk levels is not recognised. This is due to the
reserve estimates and production results show good agreement. compartmentalisation that separates exploration from mine
An examination of the feasibility study is. useful in design, and both from business decisions.
understanding why benefits have not been realised from such rich At each phase in the process of getting a mining project on
grades as yet. The geologist responsible for geological stream, decision-makers need information. During exploration
interpretation and preparation of in situ resources had apparently the purpose of information may be to reduce the lease area under
little input to the preparation of the reserve estimate or the consideration to the most interesting parts as quickly as possible,
mine/mill design. Consequently, the importance of the presence then to defme additional drilling and sampling requirements.
of several geologically distinct ore types was not fully This has to be done at minimum cost but not at the expense of
recognised. Production shows that there are various types of ore, later increasing the total cost, losing valuable time and therefore
with each type having specific milling and flotation increasing risk. fustead, the right information has to be collected
characteristics, and yielding different recoveries. Considerable at the right time, and then used correctly. So the first and most
and expensive fine tuning of ore blending is now necessary to important step in the base data collection and processing
accommodate the several ore types and gradually improve sequence (for reserve estimation purposes) is to decide what
recoveries. questions the data are supposed to answer, and for whom the
Close collaboration between geologist, and mining and information is desti.'1.ed.
metallurgical specialists at the feasibility study stage, is essential
for the preparation of realistic estimates of the reserves. Full CONCLUSION
recognition of several ore types can prevent costly alterations of
metallurgical circuits during operation. m spite of exhortations from some eminent practitioners, (King et
aI, 1982), the mining industry still retains the 'functional silo'
approach of compartmentalising the geology department, the
Using the right information mining planning department, the metallurgical engineers and the
During the 1950s and 1960s many young scientists from eastern fmancial analysts instead of viewing the estimation of reserves as
Europe, Africa and China were educated in the Soviet Union, and a 'process' involving many disciplines working as a coherent
so these countries adopted reserve estimation procedures similar team. Under the 'functional silo' approach, the geological model
to the USSR. m that system, reserves are classified into arrives at the ore evaluation office already frozen hard.
categori~ going from A (Proven) down to D (Inferred). Before Similarly, when the ore estimates reach the financial analysts
exploitation can commence reserves must in be category A or B. they too are rock solid. Although multiple simulations of the
Up to this point, the procedure is quite reasonable. The orebody could be created, the time needed to carry out a planning
unfortunate part is that this leads to highly rigid rules for passing exercise on each one prevents them from being used effectively
from one category to another. Typically the drill spacing is at to characterise the inherent uncertainty in the ore calculations or
most a certain distance. As finance is not considered, no one ever to monitor the feedback effects of variations in factors like the

288 Perth, 5 - 9 September 1994 4th Large Open Pit Mining Conference
 RISK ANALYSIS: TIIE IMPORTANCE OF RESERVE ESTIMATION

metallurgical recovery as a function of the ore type, or the cut-off REFERENCES

grade that is dependent on the metal price, and so on.
Champigny, Nand Armstrong, M, 1994. An overview of reserve
The 'functional silo' approach can lead to optimisation within estimation problems by an international 'Groupe de Reflexion', elM
the 'silos' (or departments) which does not necessarily result in Bulletin, 87(977):23-25.
optimisation of the 'process' as a whole. However, even if one Clow, G G, 1990. Lessons from the front lines - recent gold mine failures
were to achieve 'optimisation' of the project, the parameters on in Canada, paper presented at the CIM 92nd Annual General
which such an optimisation is based are subject to wide Meeting, Ottawa, May 14 p.
fluctuations, they are not static. We are not living in a static Harquail, D, 1992. Survival at the top of a gold company in the '90's,
world and we have seen dramatic changes in the last 20 years speaking notes of presentation made at the 25 February meeting of
including the break-up of Russia, the convergence of the Pacific AIME, Phoenix, 8 p.
basin as an economic powerhouse and the efforts at privatisation King, H F, McMahon, D Wand Bujtor, G J, 1982. A guide to the
in many mineral producing countries. In this dynamic understanding of ore reserve estimation, Supplement to Proc Aust
environment, one can question just how appropriate a single Inst MinMet, 281:21.
'optimal' solution is, whether on the departmental or project Worth, D J and Haystead, B, 1991. How banks assess mineral properties
scale. Are local 'optima' really relevant nowadays? Going and companies, in Proceedings ofthe Mineral Economics Symposium
further and being deliberately controversial, do we need a single 'Back to Basics', 17 January, Toronto, Canada, 23 p (CIM Mineral
Economics Society).
optimal solution or one that is robust and flexible enough against
unpredictable changes? Perhaps we should rethink an approach
to project appraisal, evaluation and development and place more
emphasis on a flexible and dynamic approach rather than a static
'optimal' one.

4th Large Open Pit Mining Conference Perth, 5 - 9 September 1994 289
290 Perth, 5 - 9 September 1994 4th Large Open Pit Mining Conference