- Email: [email protected]
Competitive cost analysis in the mineral industries
Competitive cost analysis in the mineral industries The example of nickel
Thomas
The theoretical basis of competitive cost analysis is described and linked with the practical aspects of constructing and analysing industry supply curves. Concepts of cost categories such as fixed and variable, avoidable, reactivating, net operating, cash, and total are discussed, as well as costs under capacity and less than capacity operating rates. The production unit is defined and cost determination by activity is discussed. Nickel production costs are used as an example to demonstrate the use of net and cash break even costs as proxies for variable costs in constructing industry supply curves and predicting behaviour of individual producers. The author is Associate Professor, Mineral Resource Economics, West Virginia University, 214 White Hall, Morgantown, WV 26506, USA. ‘For one of the earliest discussions of competitive costs in modern times see J. Viner, ‘Cost curves and supply curves’, Zeitschriff fur Nationalokonomie, Vol I II, 1931, pp 23-46. reprinted in The American Econ&c Associ&ion, Reading in Price Theory, Richard D. Irwin, Chicago, IL, 1952, pp 198-232. ‘R.L. Davidoff, ‘Supply analysis model (SAM): a minerals availability system methodology’, USBM IC 8820, US Bureau of Mines, 1980. ‘There have possibly been more cost studies of copper than any other nonferrous metal. See eo P. Folev and J. Clark, ‘US copper supply - an &onomic/ engineering analysis of cost-supply relationships’, Resources Policy, September 1981, pp 171-187; L. Kovisars, Copper continued on page 194
0301-4207/88/030193-12$03.00
0
F. Torries
managers and economists have long been interested in industry supply curves and the competitive position of the individual firms comprising the industry.’ This interest was heightened during the 1970s when inflation and structural changes in world economies made forecasting prices in the process of evaluating new mineral projects a dubious exercise. Since 1980 a number of studies have been undertaken, attempting to derive costs of individual mining firms and the industry cost or supply curves. The best known of these is the supply analysis model (SAM), a minerals availability system methodology of the US Bureau of Mines.2 The USBM has completed analyses of a number of mineral industries including copper, manganese, nickel, phosphate, molybdenum, lead and zinc, in which production costs of individual mineral producers in the world were estimated. These analyses resulted in the estimation of long-run industry supply curves for each mineral commodity. Other similar studies have been completed for copper,3 molybdenum,4 nickel,” uranium6 and coal.’ Determining costs of individual mineral producers and constructing an industry supply curve from the aggregate costs of individual producers is fraught with practical as well as theoretical difficulties. On the other hand, this process produces good results when properly applied and can be used to identify and solve many practical problems involving availability and prices of mineral commodities. In spite of the increasing literature on case studies of competitive cost analyses, little has been presented on the underlying theory or practical aspects of competitive cost analysis.8 It is the purpose of this paper to outline the procedures used to determine cost of individual producers in the mineral industries and the resultant industry supply curve and to identify the problems, constraints, and advantages of this process. Nickel is used as a mineral commodity example. Production
The competitive cost curve Economic
1988 Butterworth
theory
indicates
& Co (Publishers)
Ltd
that
an
industry
supply
curve
can
be 193
Conzpetilivr cos/ analysis irl rhr rninrral irlduslries
AC, MC
ACC, MCC
t
Figure 1. Short-term marginal
cost
curves
average of
and
individual
A, 6, and C, with constant marginal costs up to maximum capacity and infinite thereafter.
firms
continued from page 193 7982-2000. MET Research. Dallas. TX. 1982; Commodities Research Unit, Cop: per Studies, Commodities Research Unit, Auoust 1984; RD. Rosenkranz et al, ‘Coppe; availability - domestic: a minerals availabilitv svstem aooraisal’. USBM IC 8809, US Bureau of ‘Mines, 1979; P.R. Thomas et al, ‘The depletion allowance and domestic minerals availability: a case study in copper’, USBM IC 8874, US Bureau of Mines, 1982; A. Markowski and M. Radetzki, ‘State ownership and the price sensitivity of supply: the case of the copping mining industry’, Resources Policy, Vol 13, No 1, 1987, pp 1934. “L. Kovisars, Molybdenum 1982-2000, MET Research, Dallas, TX, 1982. ?See T.F. Torries and I. Martens, Nickdata 1985: A Comprehensive Nickel Industry Cost Data Base and Costing Program, Nickdata Inc., Toronto, Ontario, 1985 and D.I. Bleiwas, ‘Nickel availabilty - market economy countries: a minerals- availability oroaram aooraisal’. USBM IC 8995, US B&au of iines, 1984. 6L. Kovisars, Uranium 7981-2000, MET Research, Dallas, TX, 1982. ‘A number of coal costing models exist for the United States. For a summary comparison see J.P. Price, ‘Coal supply models: the state of the art’, in B. Lev eta/, eds, continued on page 195
1
qA
I
qB
I
-
qc
Q
constructed from the summation of the marginal cost curves of individual firms as long as the actions of one firm do not affect the costs or actions of any other firm in the industry.” For graphical simplicity, assume that the average and marginal cost curves of individual firms are flat up to the point of plant capacity and are vertical thereafter.” In a competitive industry, it is then possible to construct an industry stepwise supply curve, as is shown in Figure 1, by summing the marginal cost curves of the individual firms across the output range (horizontal axis). Production costs of the firm can be classified in many ways but for now all costs will be classified as being either fixed or variable. Under this simplistic classification, production decisions of the firm will be based on average and marginal variable costs. Average variable and total costs for each firm for which costs can be calculated are plotted on Figure 2. In the minerals industries it is common for some portions of production to be contributed by producers unresponsive to changes in commodity price. Such producers include those who produce the mineral commodity as a byproduct (such as the South African platinum producers who also produce byproduct nickel), and producers in planned economies. No meaningful costs can be calculated for the production of nickel as a single commodity in the case of joint production. The quantity of nickel produced as a joint product and by planned economies is represented on the industry supply curve by Oq” in Figure 2. Production from firms responsive to prices and for which costs can be calculated are added to the right of Oq’. d
P, TC, AC, MC t
I
I TC AVC, MVC
Figure 2. Industry cumulative cost curves based on total costs and average and marginal variable costs. Production Oq” is that for which no meaningful costs can be calculated.
194
d’ I
* Q
9’
RESOURCES
POLICY
September
1988
Competitive
continued from page 194 Analytic Techniques for Energy Planning, Elsevier Science Publishers, Amsterdam, 1984, pp 3-18 and M.L. Holloway, ‘On coal resource uncertainty and its importante in modeling for public policy analysis’, in J. Quirk et al, eds, Coal Models and their Use in Government Planning, Praeper, New York, 1982, pp 5-21. _ M. Porter, Competitive Strategy Techniques for Analyzing Industries and Competitors. The Free Press. New York. 1980: ‘R. Adams, ‘The long-term issues affecting the US mineral industry’, Minerals and Materials: A Bimonthly Survey, US Bureau of Mines, June/July 1986, pp 41-45; and L. Kovisars, ‘Competitive analysis’, in R. Tinsley, et al, eds, Finance for the Mineral Industry, AIME, New York, 1985. ‘C.E. Ferguson and J.P. Gould, Microeconomic Theory, Richard D. Irwing, Homewood, IL, 1975 and F.M. Scherer, fndustrial Market Structure and Economic Performance, Houghton Mifflin, Boston, 1980. “‘The assumption of flat average and marginal cost curves for individual production units up to capacity is not unreasonable. For capital-intensive industries, such as mining, marginal costs are often relatively constant over most of the feasible range of production and rise significantly onlv when capacitv is approached. See M. Radetzki, ‘Long -run ‘price aspects for aluminum and copper’, Natural Resources Forum, January 1983, and Commodities Research Unit, ‘A study to determine current costs of producing primary copper and future trends’, London, 1974.
RESOURCES
POLICY
September
cost analysis
in the mineral
industries
An industry supply curve, as shown in Figure 2, represents those costs in effect at a specific time period. Coincident with this supply curve there is also a demand function for the commodity. Short-run demand curves for mineral commodities are typically highly inelastic, and, as in the case shown in Figure 2, can be represented by a nearly vertical line. From the intersection of the demand curve and the cumulative marginal variable cost curve (MVC), an indication of commodity price can be determined, with the assumption that the industry is competitive. In the case shown, Oq’ would be produced at a price of 0~‘. Of course, other factors such as stocks and speculation influence prices, but the price indicated by the industry supply curve does indicate the stable price about which spot prices must vary. Several other observations need to be made about Figure 2. Price of the commodity, Op’, is determined by aggregate supply, demand, and the highest cost producer required to fill demand. Note that producers with variable costs greater than Op’ are shut down while those firms operating with total costs greater than Op’ are not recovering their fixed costs. If return on invested capital is considered to be a portion of fixed cost, these operating firms are making less than expected profits. On the other hand, those firms operating with total costs less than Op’ are making more than expected profits and are collecting rents. Economic theory suggests that whenever rents are available, new firms will enter the industry until such rents no longer exist. Such a situation characterized much of the mineral industry from 1980 to 1987. The shape of the industry supply curve indicates the availability of the product at various prices. Since 1975 the shape of the world supply curve for nickel has changed from one with a very wide range of costs between the high and low cost nickel producers to one in which most producers have similar production costs. This was caused by the entry of new nickel operations which eventually attained sufficiently low production costs to be competitive with the existing low cost producers. The higher cost producers went out of business. In reality, two supply schedules exist instead of one. Plants that close because of low prices do not immediately reopen as soon as prices increase to the point of equalling average variable costs. Instead, managers want assurance that prices will remain high for a sufficiently long period of time before the plant is reopened. Therefore a premium must be offered to the plant manager before the plant is reopened. In addition, costs (such as the cost of rehiring and training personnel), are often incurred in reopening. From this must be subtracted certain costs of care and maintenance and ongoing costs while the plant is inactive. In equation form, this can be represented by p’
>
AC + PR
where PI AC PR E’ C’
m
= = = =
price necessary for reactivation average variable cost f’(E’,-c’) - m price premium determined by managerial expectations future higher prices I = reactivation costs = maintenance and other ongoing costs when shut down
In the case that m < f’(E’, c’), the price necessary be higher than average variable cost.
1988
for reactivation
about
will
195
Competitive cost analysis in the mineral industries
Likewise, a manager will not close a plant until he is assured that prices will remain low for some sufficiently long period of time, considering closing and maintenance costs. In equation form this can be represented by p
PC
where P AC PC E c m
= = = =
price necessary for shutting down average variable cost f(E, c) + m price discount determined by managerial expectations future lower prices = closing costs = maintenance and other ongoing costs when shut down
about
When m < f(E, c) the price necessary to induce closing down is less than average variable costs. Other ongoing costs associated with closing a plant include contract cancellation penalties, loss of market share, loss of foreign exchange earnings and costs of unemployment. The resulting two average variable cost or industry supply curves are shown in Figure 3. Curve S’ is for a contracting industry with falling prices and curve S* is for an expanding industry with increasing prices. The existence of two supply curves for the US copper industry has been demonstrated by the Commodities Research Unit (CRU).” The effective supply curve at any given point would then be S1 for prices less than p and S* for prices greater than p’. This means that when demand is close to q and prices are between p and p’, changes in prices have little effect on output.
Structure of costs Costs can be defined and classified in a number of ways depending on what is needed to be shown. Bases of cost classifications often used include the following:‘* “Op cif, Ref 3, Commodities Research Unit. “H. Bierman and T. Dyckman, Managerial Co.9 Accounting, Macmillan, New York, 1976.
1. reaction to changes in activity (fixed or variable); 2. responsibility (plant, department or cost centre); 3 costs directly involved in production (direct or indirect); 4: natural characteristics (labour, energy, materials etc); 5. functional (manufacturing, administrative, selling); Pi
d I I
.s2
p,_____$----___
I
I I
Figure 3. Short-run industry supply curves under conditions of declining prices and plant closures (S’) and increasing prices and plant reactivations (9).
196
I I I I I 0
I
I I d’
c
q”
Q
4
RESOURCES
POLICY
September
1988
Competitive
6. reference to a particular decision able); and 7. time frame (short, long run).
Table 1. Cost categories. Labour Energy Materlals and supplies Local taxes and insurance Transportation Royalties
in the mineral
avoidable,
industries
unavoid-
Winning usable materials from ores usually involves several definable stages which may include mining, milling, smelting, refining, transporting and selling. Costs can be determined for each of these activities and added together to determine total cost of production. In practice, a single mine and mill may provide feed to more than one smelter, and a single smelter or refinery may obtain feed from more than one source. These cost centres may or may not have the same location or owner. It is therefore necessary to determine costs on the basis of the production unit, which is defined as that combination of activity centres (mine/mill, smelter, refinery) required to produce a quantity of metal. Costs for each item in a production unit are based on the grade and recovery of the material being processed and the total throughput of all material at each activity centre (mine, smelter and refinery). Major cost items for the mine, smelter and refinery are shown in Table 1. Direct operating costs include the costs of factors used directly in mining, smelting, refining, and transporting ore, intermediate and finished products. This cost gives a measure of the efficiency of an operation. From this, byproduct credits are subtracted to derive net operating costs. This method of treating byproducts is not without fault but does allow consistent treatment of costs among the producing units. The net operating cost is a figure familiar to most production supervisors and is a measure of the competitive nature of the production unit considering ore grade as well as production efficiencies without regard for externalities of overheads, capital and financial charges. Overhead charges are required if the production unit is to function, but the level of these charges is to some degree arbitrary. Similarly, interest charges are functions of debt and interest rates and may have little to do with the efficiency of the production unit. These two charges are real expenses for all operations and represent cash that must be paid out of operations. However, these charges can be controlled to some degree by management, or in the case of interest, postponed or not paid at all (such as when debt is exchanged for equity). Other cash costs include avoidable costs, such as those associated with shutting down or reactivating an operation, and cash items, such as ongoing capital costs. Net production costs less overhead and interest charges equals cash break even costs. In the long run an operation must recover all cash expenses to stay in business. Although short-term decisions are made on the basis of variable costs plus appropriate adjustments according to the situation, cash break even cost is an important longer-term decision variable for managers. The last item to be considered is recoupment of invested capital and return on investment. For ongoing operations initial investment represents a sunk cost and is of interest only to determine the investment profitability of the operating unit. For a new operation, there must be strong indications that total costs, which include capital, will be recovered in order for the project to be undertaken.
Direct operating costs Byproduct credits Net production costs Sales and administrative Interest Other cash costs
cost analysis
(opportunity,
costs
Cash break even costs Initial capital recoupment Return on investment Total costs
Variable, avoidable and cash break even costs Economic
RESOURCES
POLICY
September
1988
theory
suggests
production
decisions
made
by the firm are
197
Competitive cost analysis in the mineral industries
13M. Friedman, Price Theory, Aldine, New York, 1976. 140p cif, Ref 3, Commodities Research Unit and Kovisars.
based on variable costs and the industry supply curve. Under appropriate conditions, the supply curve can be obtained by summing the marginal variable costs of the individual firms. In practice, it is often difficult to identify what is variable and what is fixed. The category for any particular cost may change over time and be different among firms or operations. This problem is sometimes recognized by referring to unavoidable (fixed) and avoidable (variable) costs,13 but this does little to resolve the underlying identification problems. Labour costs are often difficult to classify into fixed and variable costs for a particular mining operation. In some countries, such as those with strong labour unions, labour is sometimes considered a fixed cost in the short run but variable in the long run. In others, labour may well be a fixed cost in the long run as well as the short run because of the lack of employment opportunities and the commitment of government to support unemployed citizens. The distinction between fixed and variable costs also depends to some degree on the range of options open to the firm. For example, some costs may be avoided as long as the operation continues but become unavoidable when operations cease (such as political costs of unemployment). The terms avoidable and unavoidable therefore refer to fixed costs associated with shutting down an operation either on a short- or long-term basis. Attempts have been made to identify avoidable and unavoidable costs for each production unit of an industry in order to define the industry supply curve. I4 This is usually accomplished by estimating variable operating cost in the short run and adding a component to represent the hidden cost of shutting down or reactivating. Costs associated with shutting down or reactivating a production unit are summarized in Table 2. It is obviously difficult to make estimates of certain costs, such as the political cost of unemployment. It may also be difficult to correctly assess the effect of less controversial costs such as contractual penalties, loss of market share, and loss of trained personnel. The difficulty in determining variable costs suggests there may be other more easily determined costs that might be used as proxies for variable costs. Net and cash break even costs can be calculated in a consistent manner for all producing and potentially producing production units in the longer run. The cash break even costs represent the minimum price a production unit must receive for its product in the long Table 2.
Costs and considerations
associated with shutting down or reactivating
production
units. Shutting down
Reactivating
Continuing labour costs Loss of skilled workers
Costs of rehiring and retraining labour
Contractual commitments have to be honoured.
for input factors may
Renegotiating contracts for input factors may result in different factor costs.
Losing market share
Regaining discounts
market
share,
usually
by
Continuing care and maintenance
Expending rejuvenation
Assurance that prices will remain low
Assurance that prices wrll remain high
price
capital
Social costs Loss of foreign exchange earnings Political cost of unemployment Secondary economic effects
198
RESOURCES
POLICY September
1988
Competitive
cost analysis
in the mineral
industries
run in order to survive. Consequently, a cumulative industry cash break even cost curve reveals much about relative profitabilities of individual production units and gives an indication of the longer-term stable price necessary to keep the highest cost marginal producer from going bankrupt. However, if the amount of interest to be paid or level of overhead charges are overestimated, the net production cost plus avoidable costs gives a better indication of possible actions of individual production units.
Calculation of costs Methods of determining production costs for individual production units include analysis of published historical cost data, engineering cost analysis and corporate interviews. Historical cost data such as are presented in annual reports are readily obtainable but may not be indicative of expected future costs; may not present cost data on individual operations; and may not be collected and presented in a consistent manner that allows meaningful comparisons between companies. More often, a combination of the three methods is required to obtain usable results. Engineering cost estimating procedures have great advantage in that, for any given process, input factors can be identified and quantified. Once these are known the effects of changes in factor prices can be easily determined. Changes in factor prices can be caused by changes in exchange rates or in factor scarcity. This approach allows all costs to be treated in the same manner for all production units but the process is time consuming, difficult to accomplish and expensive.
Level of costing detail Costs of mineral production can be calculated in great detail by the level of activity and by cost category. The degree of detail required depends on the number of production units in the industry being studied and the purposes of the study. In the case of nickel there are fewer than forty production units for which costs can be calculated. While calculating detailed costs for this many production units is a large task, it is not insurmountable. Since there are relatively few production units in the nickel industry the costs for each unit must be calculated carefully, for errors made in the costs of a single large production unit will seriously affect the results of overall industry analysis. In the case of copper there are nearly 400 production units for which costs could be calculated. In this case, obtaining detailed costs for all production units is a much larger proposition. If the purpose of the study were to determine future copper prices rather than investigate the economics of individual production units, costs of individual units would not have to be determined in as great a degree of detail or accuracy. Compensating errors among the large number of production units would result in the satisfactory definition of the cumulative cost curve. If, however, the purpose of study were to investigate the relative economics of specific production units, attention to detailed production unit costs would be required. Finished metal products, such as cathode nickel or copper or ferronickel, are relatively uniform in character. This means that the costs of using these finished metal products are similar among all metal
RESOURCES
POLICY
September
1988
199
Competitive
cosl analysis in the mineral
indusfries
users and no adjustments need to be made in the supply curve to account for differential user costs. This is not the case of coal, where the characteristics of the finished product are heterogeneous and are very much related to the costs of the final user. In addition, mining, processing, transporting and combustion costs are related. For commodities in which the characteristics of the final product differ significantly or in which processing methods and costs are related to the costs of using the final product, user costs must be included in the competitive cost calculations in order to predict producer behaviour and to define a cost curve suitable for price determination.
Operating rate, capacity and costs
“This is clearly illustrated by the practice of the Canadian nickel producers scheduling production shut downs in addition to vacation shut downs to rationalize production with market demand. This illustrates that it is more economical to operate at full capacity rates and experience complete production shut downs rather than operate continuously at lower production rates. By extension, it is even more economical to operate at full capacity rates with no production shut down, assuming a constant product price. See INCO, Annual Reports, 1982-l 986.
200
Capacity of an operation is defined as the effective maximum production capability given the particular operating constraints in effect at a particular time. Capacity has three dimensions: quantity, economics and time. It is therefore necessary to specify the limiting conditions when specifying capacity in order for the term capacity to be meaningful. For a mineral operation, limiting conditions include time and economics, ore grade and recovery, the state of repair of the plant, the throughput capacity of individual sectors of the operation (such as mine, mill, smelter or refinery), and the level of capital expenditures required to remove bottlenecks. A ‘soft’ definition often found useful when defining capacity of a mineral operation is that output that can be obtained in a short period of time without ‘excessive’ capital expenditures. This definition takes into account current economics and time, and usually results in the identification of a narrow range of possible capacity levels. Other definitions of capacity, such as rated or nameplate capacity, are often misleading in that an operation may not be able to reach rated capacity without extensive capital expenditure or using higher grade ores than are actually available. Effective capacity must be stated in proper units for each activity. For example, mining and milling capacity must be stated in terms of tonnes of ore mined, whereas smelter and refinery capacity must be stated in terms of tonnes of material processed. Changes in grade of ore or feed material affect production capacity. All costs can be calculated for two production levels: at actual operating rates, which may be less than full capacity, and at full capacity operating rates. Costs at actual production rates are of interest in that the current financial conditions of the production unit can be identified. However, costs at full capacity production rates are more instructive. Firms attempt to operate at the rate that yields the greatest profit, which is usually near full capacity. Therefore, managers plan to operate plants at full capacity’” even though this is often not achieved. The fact that full capacity production is the goal of each plant manager means that the relevant cost in determining the behaviour of the production units and the industry is at full capacity. Plants do operate at less than full capacity and sometimes have lower operating costs in the short run than if they had operated at capacity. Certain costs, such as maintenance, research and exploration can be postponed in the short run, resulting in lower costs whether the plant operates at capacity or not. More often, higher than average grade ore is mined to lower costs. In such cases it may not be physically possible for the mine to produce sufficient high grade ore to run a plant at capacity
RESOURCES
POLICY
September
1988
Competitive cost unalysis in the mineral industries
rates. When higher than average ore grade is mined it is difficult to tell if this is a temporary feature or a permanent situation resulting in more rapid depletion of reserves. If raising the grade of ore mined at the expense of shortening the life of the mine is the only way to reduce costs sufficiently to prevent bankruptcy, then high grading of reserves becomes a permanent feature as long as prices do not increase.
Competitive cost analysis of the nickel industry
160p
cit.
Ref 5, Torries
and
Martens,
annual issues, 1981-1985.
RESOURCES
POLICY
September
Production data were collected and costs calculated for individual nickel production units on a worldwide basis from 1980 to 1985.16 The 1985 costs are calculated for 33 production units, which includes all major and most minor nickel producers in the world. Costs are accumulated on an itemized basis for each activity (mine, smelter or refinery) and are based on specific prices for input factors such as oil, electricity, coal, labour, material and supplies and exchange rates in which the various trades occur. Changes in any of the input factor prices or exchange rates change the results of the cost calculations. Estimates made of 1985 world nickel industry production costs and resulting industry cost curves at 1985 actual and full capacity operating rates are given in Figures 4 and 5 respectively. Total, break even, and net operating costs for each production unit are shown at actual and full capacity operating rates. Total demand for primary nickel in 1985 was about 1.3 million lb. At this rate of production, the cash break even cost curve at full capacity production suggests that a stable nickel price would be in the neighbourhood of $2.30 per lb. During the first half of 1985 the nickel price rose to $2.62 per lb because of strong nickel demand and declining nickel stocks. During July 1985, demand slackened and deliveries increased, causing nickel prices to retreat to the $2.20 to $2.30 per lb range. As shown in Figure 4, the higher cost producers did not recover their cash break even costs at this price level, and some did not even recover their net production costs. Most production units did not recover their total costs, which meant that few production units achieved a positive return on capital during 1985. A rational response of those units indicated in Figures 4 and 5 as not recovering cash break even costs would have been for them to shut down. Some did, but some did not. There are a number of possible reasons to explain why all high cost producers did not shut down. It is likely that the costs shown in Figures 4 and 5 are too high for those producers because shut down costs (m) and price expectations (PC) were not adequately treated in the analysis. During 1985 nickel producers anticipated an increase in nickel demand (which largely did not happen), a decline in nickel stocks, and an increase in nickel prices. These incentives may well have contributed to the continued operation of the unprofitable production units. In addition, all production units were at that time embarking on or completing extensive cost reduction programmes and anticipated improved profits through lowered costs. Even though prices started to increase in 1985, partially because of speculative reasons, there was no evidence that prices higher than $2.30 could be sustained on the basis of production costs. When demand failed to materialize as expected, prices immediately fell back to the $2.30 per lb range, or that price range indicated by the cash break even cost curve.
1988
201
Competitive
cost analysis
in the mineral
industries
4’oo 9 3 3.50 .u c G & 3.00 Total 2.50
-
2.00
-
cost
1.50 -
1.00 -
0.50
Figure 4. industry cumulative cost curve for nickel at actual 1985 production rates.
0.00
Break
even cost
5 f >
I-
O
200
I
I
I
400
600
I
I
1000
600 Million
I
1200
1400
lb nickel
Later in 1985 two other events reduced nickel production costs; oil prices declined and exchange rates adjusted relative to each other and the US dollar. The probable results of the combination of these two events were tested in the nickel model. The results of this analysis indicated that the cash break even costs for the marginal producer to supply the anticipated 1.4 million lb of nickel required in 1986 were around $1.70 to $1.80 per lb. This was based on oil prices of $15 per barrel. By the beginning of 1986 nickel prices were in the $1.70 per lb range. These results again indicate that the model and the use of the cash break even cost curve do result in reasonable predictions.
2.50
-
2.00
-
1.50
-
1.00
-g
et operating
? x Break
f 0.50
Figure 5. Industry cumulative curve for nickel at full capacity production
202
rates.
cost 1985
0.00
-:
’
0
even cost
E E ’
1’
200
I
400
cost
I
I
600
800 Million
I
I 000
I
1200
l
1400
I
1600
1
1800
lb nickel
RESOURCES
POLICY
September
1988
Competitive cost analysis in the mineral industries
The assumption of flat marginal and average cost curves can be tested by comparing the costs of the individual producers at full and less than full capacity operations. Costs at less than full capacity are generally higher than at full capacity if differences in ore grade, interest and overhead charges are not changed, indicating that average costs are not flat but do decline up to some capacity figure. There are sufficient exceptions to this to prevent generalizations. Since 1982 a significant feature of the world nickel industry has been the continued and dramatic decrease in the cost of producing nickel. For example, break even production costs of Falconbridge’s Canadian operation have declined a full $1 .OO per lb since 1982. I7 The decrease in costs for all producers has caused the industry supply curve to shift downward, as is shown in Figure 6. In the case of INCO’s Sudbury operations, costs have been cut in half. The process by which INCO reduced costs at Sudbury illustrates some of the pitfalls that may be present in calculating competitive production costs for individual production units. INCO operated Sudbury at less than half of full capacity in 1982 and 1983. During this period INCO laid off half the workers at Sudbury. When INCO did increase production starting in 1984, it did so without rehiring any of the workers previously laid off. This was accomplished by increasing labour efficiency in mining and production, by using less labour-intensive mining and processing methods, and by changing the mix of products produced to one requiring less labour.”
Data gathering and cost of information Actual cost data on individual production units in an industry are of use to many economic researchers. However, gathering these data is time consuming and expensive even though computers make managing and updating the data relatively easy. In addition, factor prices and usages, demand for the final product and exchange rates constantly change, quickly making estimated costs obsolete. The nickel study for which the results are given in this paper was first
’ 71bid. “Op tit, Ref 15.
1982
/ 3.50
-
3.00
-
s c 2.50
-
$ ;
2 % 2.00 2 II r, 1.50
-
s
Figure 6. Decline in cash break even 1982 to of nickel production 1985 at full capacity production rates.
POLICY
September
-
0.50
-
0.00 L 0
costs
RESOURCES
1.00
I 500
I 1000
I 1500
2000
Million lb nickel
1988
203
Competitive cost analysis in the mineral industries
started in 1980. Since then, the computer program and the database have been revised and improved yearly. Duplicating a similar effort would require considerable time and would involve visiting most of the nickel operations in the world, conducting research in process metallurgy and production costs, and developing a computerized database and costing program. Because of the perishable nature of production data, it is necessary to update the data and costs on at least a yearly basis. In addition, it is necessary to visit each production unit at least every other year after the initial visit to keep abreast with changes in factor costs and usage. The significant cost of yearly updates must be considered when an industry cost study is being contemplated.
Conclusions Competitive cost analysis is a powerful tool that can be used to evaluate mineral projects, predict the behaviour of individual producers within an industry and, under appropriate conditions, give indications of future prices. Cumulative cost curves can be determined for an industry and linked theoretically to true industry supply curves. The key to competitve cost analysis is to determine costs in a consistent and rigorously defined manner for all firms within an industry. Accomplishing this is sometimes difficult because classifications of costs vary among firms and change over time. Marginal variable costs are particularly difficult to determine although these costs, under conditions of perfect competition, do form the theoretical basis of the industry supply curve. It is, however, possible to determine net, cash break even and total costs for each firm in an industry, and use these costs in meaningful ways to construct cumulative cost curves that approximate industry supply curves. Each of these curves can be generated for actual and full capacity production rates. The curve showing the most predictive ability, both in theoretical and applied terms, is the cash break even curve at full capacity. The existence of two supply schedules, one for increasing prices and an expanding industry and one for decreasing prices and a shrinking industry, has not been widely recognized. Entry and exit costs plus managerial expectations create the two supply schedules. The two curves create the situation in which changes in prices over certain ranges do not result in changes in quantities produced by the industry. Different methods must be used in determining costs for producers selling homogeneous products, such as cathode nickel, and for producers with heterogeneous products, such as coal. In the case of commodities such as coal, user costs must be included in the analysis. Competitive costs of the nickel industry have been determined during a five-year period. While exit and entry costs were not adequately treated, the resulting industry cash break even curve did indicate which producers were likely to stay in business and which were likely to shut down. In addition, the cash break even curve, in conjunction with an estimate of primary nickel demand, did give indications of future nickel prices.
204
RESOURCES
POLICY
September
1988
Thomas
The theoretical basis of competitive cost analysis is described and linked with the practical aspects of constructing and analysing industry supply curves. Concepts of cost categories such as fixed and variable, avoidable, reactivating, net operating, cash, and total are discussed, as well as costs under capacity and less than capacity operating rates. The production unit is defined and cost determination by activity is discussed. Nickel production costs are used as an example to demonstrate the use of net and cash break even costs as proxies for variable costs in constructing industry supply curves and predicting behaviour of individual producers. The author is Associate Professor, Mineral Resource Economics, West Virginia University, 214 White Hall, Morgantown, WV 26506, USA. ‘For one of the earliest discussions of competitive costs in modern times see J. Viner, ‘Cost curves and supply curves’, Zeitschriff fur Nationalokonomie, Vol I II, 1931, pp 23-46. reprinted in The American Econ&c Associ&ion, Reading in Price Theory, Richard D. Irwin, Chicago, IL, 1952, pp 198-232. ‘R.L. Davidoff, ‘Supply analysis model (SAM): a minerals availability system methodology’, USBM IC 8820, US Bureau of Mines, 1980. ‘There have possibly been more cost studies of copper than any other nonferrous metal. See eo P. Folev and J. Clark, ‘US copper supply - an &onomic/ engineering analysis of cost-supply relationships’, Resources Policy, September 1981, pp 171-187; L. Kovisars, Copper continued on page 194
0301-4207/88/030193-12$03.00
0
F. Torries
managers and economists have long been interested in industry supply curves and the competitive position of the individual firms comprising the industry.’ This interest was heightened during the 1970s when inflation and structural changes in world economies made forecasting prices in the process of evaluating new mineral projects a dubious exercise. Since 1980 a number of studies have been undertaken, attempting to derive costs of individual mining firms and the industry cost or supply curves. The best known of these is the supply analysis model (SAM), a minerals availability system methodology of the US Bureau of Mines.2 The USBM has completed analyses of a number of mineral industries including copper, manganese, nickel, phosphate, molybdenum, lead and zinc, in which production costs of individual mineral producers in the world were estimated. These analyses resulted in the estimation of long-run industry supply curves for each mineral commodity. Other similar studies have been completed for copper,3 molybdenum,4 nickel,” uranium6 and coal.’ Determining costs of individual mineral producers and constructing an industry supply curve from the aggregate costs of individual producers is fraught with practical as well as theoretical difficulties. On the other hand, this process produces good results when properly applied and can be used to identify and solve many practical problems involving availability and prices of mineral commodities. In spite of the increasing literature on case studies of competitive cost analyses, little has been presented on the underlying theory or practical aspects of competitive cost analysis.8 It is the purpose of this paper to outline the procedures used to determine cost of individual producers in the mineral industries and the resultant industry supply curve and to identify the problems, constraints, and advantages of this process. Nickel is used as a mineral commodity example. Production
The competitive cost curve Economic
1988 Butterworth
theory
indicates
& Co (Publishers)
Ltd
that
an
industry
supply
curve
can
be 193
Conzpetilivr cos/ analysis irl rhr rninrral irlduslries
AC, MC
ACC, MCC
t
Figure 1. Short-term marginal
cost
curves
average of
and
individual
A, 6, and C, with constant marginal costs up to maximum capacity and infinite thereafter.
firms
continued from page 193 7982-2000. MET Research. Dallas. TX. 1982; Commodities Research Unit, Cop: per Studies, Commodities Research Unit, Auoust 1984; RD. Rosenkranz et al, ‘Coppe; availability - domestic: a minerals availabilitv svstem aooraisal’. USBM IC 8809, US Bureau of ‘Mines, 1979; P.R. Thomas et al, ‘The depletion allowance and domestic minerals availability: a case study in copper’, USBM IC 8874, US Bureau of Mines, 1982; A. Markowski and M. Radetzki, ‘State ownership and the price sensitivity of supply: the case of the copping mining industry’, Resources Policy, Vol 13, No 1, 1987, pp 1934. “L. Kovisars, Molybdenum 1982-2000, MET Research, Dallas, TX, 1982. ?See T.F. Torries and I. Martens, Nickdata 1985: A Comprehensive Nickel Industry Cost Data Base and Costing Program, Nickdata Inc., Toronto, Ontario, 1985 and D.I. Bleiwas, ‘Nickel availabilty - market economy countries: a minerals- availability oroaram aooraisal’. USBM IC 8995, US B&au of iines, 1984. 6L. Kovisars, Uranium 7981-2000, MET Research, Dallas, TX, 1982. ‘A number of coal costing models exist for the United States. For a summary comparison see J.P. Price, ‘Coal supply models: the state of the art’, in B. Lev eta/, eds, continued on page 195
1
qA
I
qB
I
-
qc
Q
constructed from the summation of the marginal cost curves of individual firms as long as the actions of one firm do not affect the costs or actions of any other firm in the industry.” For graphical simplicity, assume that the average and marginal cost curves of individual firms are flat up to the point of plant capacity and are vertical thereafter.” In a competitive industry, it is then possible to construct an industry stepwise supply curve, as is shown in Figure 1, by summing the marginal cost curves of the individual firms across the output range (horizontal axis). Production costs of the firm can be classified in many ways but for now all costs will be classified as being either fixed or variable. Under this simplistic classification, production decisions of the firm will be based on average and marginal variable costs. Average variable and total costs for each firm for which costs can be calculated are plotted on Figure 2. In the minerals industries it is common for some portions of production to be contributed by producers unresponsive to changes in commodity price. Such producers include those who produce the mineral commodity as a byproduct (such as the South African platinum producers who also produce byproduct nickel), and producers in planned economies. No meaningful costs can be calculated for the production of nickel as a single commodity in the case of joint production. The quantity of nickel produced as a joint product and by planned economies is represented on the industry supply curve by Oq” in Figure 2. Production from firms responsive to prices and for which costs can be calculated are added to the right of Oq’. d
P, TC, AC, MC t
I
I TC AVC, MVC
Figure 2. Industry cumulative cost curves based on total costs and average and marginal variable costs. Production Oq” is that for which no meaningful costs can be calculated.
194
d’ I
* Q
9’
RESOURCES
POLICY
September
1988
Competitive
continued from page 194 Analytic Techniques for Energy Planning, Elsevier Science Publishers, Amsterdam, 1984, pp 3-18 and M.L. Holloway, ‘On coal resource uncertainty and its importante in modeling for public policy analysis’, in J. Quirk et al, eds, Coal Models and their Use in Government Planning, Praeper, New York, 1982, pp 5-21. _ M. Porter, Competitive Strategy Techniques for Analyzing Industries and Competitors. The Free Press. New York. 1980: ‘R. Adams, ‘The long-term issues affecting the US mineral industry’, Minerals and Materials: A Bimonthly Survey, US Bureau of Mines, June/July 1986, pp 41-45; and L. Kovisars, ‘Competitive analysis’, in R. Tinsley, et al, eds, Finance for the Mineral Industry, AIME, New York, 1985. ‘C.E. Ferguson and J.P. Gould, Microeconomic Theory, Richard D. Irwing, Homewood, IL, 1975 and F.M. Scherer, fndustrial Market Structure and Economic Performance, Houghton Mifflin, Boston, 1980. “‘The assumption of flat average and marginal cost curves for individual production units up to capacity is not unreasonable. For capital-intensive industries, such as mining, marginal costs are often relatively constant over most of the feasible range of production and rise significantly onlv when capacitv is approached. See M. Radetzki, ‘Long -run ‘price aspects for aluminum and copper’, Natural Resources Forum, January 1983, and Commodities Research Unit, ‘A study to determine current costs of producing primary copper and future trends’, London, 1974.
RESOURCES
POLICY
September
cost analysis
in the mineral
industries
An industry supply curve, as shown in Figure 2, represents those costs in effect at a specific time period. Coincident with this supply curve there is also a demand function for the commodity. Short-run demand curves for mineral commodities are typically highly inelastic, and, as in the case shown in Figure 2, can be represented by a nearly vertical line. From the intersection of the demand curve and the cumulative marginal variable cost curve (MVC), an indication of commodity price can be determined, with the assumption that the industry is competitive. In the case shown, Oq’ would be produced at a price of 0~‘. Of course, other factors such as stocks and speculation influence prices, but the price indicated by the industry supply curve does indicate the stable price about which spot prices must vary. Several other observations need to be made about Figure 2. Price of the commodity, Op’, is determined by aggregate supply, demand, and the highest cost producer required to fill demand. Note that producers with variable costs greater than Op’ are shut down while those firms operating with total costs greater than Op’ are not recovering their fixed costs. If return on invested capital is considered to be a portion of fixed cost, these operating firms are making less than expected profits. On the other hand, those firms operating with total costs less than Op’ are making more than expected profits and are collecting rents. Economic theory suggests that whenever rents are available, new firms will enter the industry until such rents no longer exist. Such a situation characterized much of the mineral industry from 1980 to 1987. The shape of the industry supply curve indicates the availability of the product at various prices. Since 1975 the shape of the world supply curve for nickel has changed from one with a very wide range of costs between the high and low cost nickel producers to one in which most producers have similar production costs. This was caused by the entry of new nickel operations which eventually attained sufficiently low production costs to be competitive with the existing low cost producers. The higher cost producers went out of business. In reality, two supply schedules exist instead of one. Plants that close because of low prices do not immediately reopen as soon as prices increase to the point of equalling average variable costs. Instead, managers want assurance that prices will remain high for a sufficiently long period of time before the plant is reopened. Therefore a premium must be offered to the plant manager before the plant is reopened. In addition, costs (such as the cost of rehiring and training personnel), are often incurred in reopening. From this must be subtracted certain costs of care and maintenance and ongoing costs while the plant is inactive. In equation form, this can be represented by p’
>
AC + PR
where PI AC PR E’ C’
m
= = = =
price necessary for reactivation average variable cost f’(E’,-c’) - m price premium determined by managerial expectations future higher prices I = reactivation costs = maintenance and other ongoing costs when shut down
In the case that m < f’(E’, c’), the price necessary be higher than average variable cost.
1988
for reactivation
about
will
195
Competitive cost analysis in the mineral industries
Likewise, a manager will not close a plant until he is assured that prices will remain low for some sufficiently long period of time, considering closing and maintenance costs. In equation form this can be represented by p
PC
where P AC PC E c m
= = = =
price necessary for shutting down average variable cost f(E, c) + m price discount determined by managerial expectations future lower prices = closing costs = maintenance and other ongoing costs when shut down
about
When m < f(E, c) the price necessary to induce closing down is less than average variable costs. Other ongoing costs associated with closing a plant include contract cancellation penalties, loss of market share, loss of foreign exchange earnings and costs of unemployment. The resulting two average variable cost or industry supply curves are shown in Figure 3. Curve S’ is for a contracting industry with falling prices and curve S* is for an expanding industry with increasing prices. The existence of two supply curves for the US copper industry has been demonstrated by the Commodities Research Unit (CRU).” The effective supply curve at any given point would then be S1 for prices less than p and S* for prices greater than p’. This means that when demand is close to q and prices are between p and p’, changes in prices have little effect on output.
Structure of costs Costs can be defined and classified in a number of ways depending on what is needed to be shown. Bases of cost classifications often used include the following:‘* “Op cif, Ref 3, Commodities Research Unit. “H. Bierman and T. Dyckman, Managerial Co.9 Accounting, Macmillan, New York, 1976.
1. reaction to changes in activity (fixed or variable); 2. responsibility (plant, department or cost centre); 3 costs directly involved in production (direct or indirect); 4: natural characteristics (labour, energy, materials etc); 5. functional (manufacturing, administrative, selling); Pi
d I I
.s2
p,_____$----___
I
I I
Figure 3. Short-run industry supply curves under conditions of declining prices and plant closures (S’) and increasing prices and plant reactivations (9).
196
I I I I I 0
I
I I d’
c
q”
Q
4
RESOURCES
POLICY
September
1988
Competitive
6. reference to a particular decision able); and 7. time frame (short, long run).
Table 1. Cost categories. Labour Energy Materlals and supplies Local taxes and insurance Transportation Royalties
in the mineral
avoidable,
industries
unavoid-
Winning usable materials from ores usually involves several definable stages which may include mining, milling, smelting, refining, transporting and selling. Costs can be determined for each of these activities and added together to determine total cost of production. In practice, a single mine and mill may provide feed to more than one smelter, and a single smelter or refinery may obtain feed from more than one source. These cost centres may or may not have the same location or owner. It is therefore necessary to determine costs on the basis of the production unit, which is defined as that combination of activity centres (mine/mill, smelter, refinery) required to produce a quantity of metal. Costs for each item in a production unit are based on the grade and recovery of the material being processed and the total throughput of all material at each activity centre (mine, smelter and refinery). Major cost items for the mine, smelter and refinery are shown in Table 1. Direct operating costs include the costs of factors used directly in mining, smelting, refining, and transporting ore, intermediate and finished products. This cost gives a measure of the efficiency of an operation. From this, byproduct credits are subtracted to derive net operating costs. This method of treating byproducts is not without fault but does allow consistent treatment of costs among the producing units. The net operating cost is a figure familiar to most production supervisors and is a measure of the competitive nature of the production unit considering ore grade as well as production efficiencies without regard for externalities of overheads, capital and financial charges. Overhead charges are required if the production unit is to function, but the level of these charges is to some degree arbitrary. Similarly, interest charges are functions of debt and interest rates and may have little to do with the efficiency of the production unit. These two charges are real expenses for all operations and represent cash that must be paid out of operations. However, these charges can be controlled to some degree by management, or in the case of interest, postponed or not paid at all (such as when debt is exchanged for equity). Other cash costs include avoidable costs, such as those associated with shutting down or reactivating an operation, and cash items, such as ongoing capital costs. Net production costs less overhead and interest charges equals cash break even costs. In the long run an operation must recover all cash expenses to stay in business. Although short-term decisions are made on the basis of variable costs plus appropriate adjustments according to the situation, cash break even cost is an important longer-term decision variable for managers. The last item to be considered is recoupment of invested capital and return on investment. For ongoing operations initial investment represents a sunk cost and is of interest only to determine the investment profitability of the operating unit. For a new operation, there must be strong indications that total costs, which include capital, will be recovered in order for the project to be undertaken.
Direct operating costs Byproduct credits Net production costs Sales and administrative Interest Other cash costs
cost analysis
(opportunity,
costs
Cash break even costs Initial capital recoupment Return on investment Total costs
Variable, avoidable and cash break even costs Economic
RESOURCES
POLICY
September
1988
theory
suggests
production
decisions
made
by the firm are
197
Competitive cost analysis in the mineral industries
13M. Friedman, Price Theory, Aldine, New York, 1976. 140p cif, Ref 3, Commodities Research Unit and Kovisars.
based on variable costs and the industry supply curve. Under appropriate conditions, the supply curve can be obtained by summing the marginal variable costs of the individual firms. In practice, it is often difficult to identify what is variable and what is fixed. The category for any particular cost may change over time and be different among firms or operations. This problem is sometimes recognized by referring to unavoidable (fixed) and avoidable (variable) costs,13 but this does little to resolve the underlying identification problems. Labour costs are often difficult to classify into fixed and variable costs for a particular mining operation. In some countries, such as those with strong labour unions, labour is sometimes considered a fixed cost in the short run but variable in the long run. In others, labour may well be a fixed cost in the long run as well as the short run because of the lack of employment opportunities and the commitment of government to support unemployed citizens. The distinction between fixed and variable costs also depends to some degree on the range of options open to the firm. For example, some costs may be avoided as long as the operation continues but become unavoidable when operations cease (such as political costs of unemployment). The terms avoidable and unavoidable therefore refer to fixed costs associated with shutting down an operation either on a short- or long-term basis. Attempts have been made to identify avoidable and unavoidable costs for each production unit of an industry in order to define the industry supply curve. I4 This is usually accomplished by estimating variable operating cost in the short run and adding a component to represent the hidden cost of shutting down or reactivating. Costs associated with shutting down or reactivating a production unit are summarized in Table 2. It is obviously difficult to make estimates of certain costs, such as the political cost of unemployment. It may also be difficult to correctly assess the effect of less controversial costs such as contractual penalties, loss of market share, and loss of trained personnel. The difficulty in determining variable costs suggests there may be other more easily determined costs that might be used as proxies for variable costs. Net and cash break even costs can be calculated in a consistent manner for all producing and potentially producing production units in the longer run. The cash break even costs represent the minimum price a production unit must receive for its product in the long Table 2.
Costs and considerations
associated with shutting down or reactivating
production
units. Shutting down
Reactivating
Continuing labour costs Loss of skilled workers
Costs of rehiring and retraining labour
Contractual commitments have to be honoured.
for input factors may
Renegotiating contracts for input factors may result in different factor costs.
Losing market share
Regaining discounts
market
share,
usually
by
Continuing care and maintenance
Expending rejuvenation
Assurance that prices will remain low
Assurance that prices wrll remain high
price
capital
Social costs Loss of foreign exchange earnings Political cost of unemployment Secondary economic effects
198
RESOURCES
POLICY September
1988
Competitive
cost analysis
in the mineral
industries
run in order to survive. Consequently, a cumulative industry cash break even cost curve reveals much about relative profitabilities of individual production units and gives an indication of the longer-term stable price necessary to keep the highest cost marginal producer from going bankrupt. However, if the amount of interest to be paid or level of overhead charges are overestimated, the net production cost plus avoidable costs gives a better indication of possible actions of individual production units.
Calculation of costs Methods of determining production costs for individual production units include analysis of published historical cost data, engineering cost analysis and corporate interviews. Historical cost data such as are presented in annual reports are readily obtainable but may not be indicative of expected future costs; may not present cost data on individual operations; and may not be collected and presented in a consistent manner that allows meaningful comparisons between companies. More often, a combination of the three methods is required to obtain usable results. Engineering cost estimating procedures have great advantage in that, for any given process, input factors can be identified and quantified. Once these are known the effects of changes in factor prices can be easily determined. Changes in factor prices can be caused by changes in exchange rates or in factor scarcity. This approach allows all costs to be treated in the same manner for all production units but the process is time consuming, difficult to accomplish and expensive.
Level of costing detail Costs of mineral production can be calculated in great detail by the level of activity and by cost category. The degree of detail required depends on the number of production units in the industry being studied and the purposes of the study. In the case of nickel there are fewer than forty production units for which costs can be calculated. While calculating detailed costs for this many production units is a large task, it is not insurmountable. Since there are relatively few production units in the nickel industry the costs for each unit must be calculated carefully, for errors made in the costs of a single large production unit will seriously affect the results of overall industry analysis. In the case of copper there are nearly 400 production units for which costs could be calculated. In this case, obtaining detailed costs for all production units is a much larger proposition. If the purpose of the study were to determine future copper prices rather than investigate the economics of individual production units, costs of individual units would not have to be determined in as great a degree of detail or accuracy. Compensating errors among the large number of production units would result in the satisfactory definition of the cumulative cost curve. If, however, the purpose of study were to investigate the relative economics of specific production units, attention to detailed production unit costs would be required. Finished metal products, such as cathode nickel or copper or ferronickel, are relatively uniform in character. This means that the costs of using these finished metal products are similar among all metal
RESOURCES
POLICY
September
1988
199
Competitive
cosl analysis in the mineral
indusfries
users and no adjustments need to be made in the supply curve to account for differential user costs. This is not the case of coal, where the characteristics of the finished product are heterogeneous and are very much related to the costs of the final user. In addition, mining, processing, transporting and combustion costs are related. For commodities in which the characteristics of the final product differ significantly or in which processing methods and costs are related to the costs of using the final product, user costs must be included in the competitive cost calculations in order to predict producer behaviour and to define a cost curve suitable for price determination.
Operating rate, capacity and costs
“This is clearly illustrated by the practice of the Canadian nickel producers scheduling production shut downs in addition to vacation shut downs to rationalize production with market demand. This illustrates that it is more economical to operate at full capacity rates and experience complete production shut downs rather than operate continuously at lower production rates. By extension, it is even more economical to operate at full capacity rates with no production shut down, assuming a constant product price. See INCO, Annual Reports, 1982-l 986.
200
Capacity of an operation is defined as the effective maximum production capability given the particular operating constraints in effect at a particular time. Capacity has three dimensions: quantity, economics and time. It is therefore necessary to specify the limiting conditions when specifying capacity in order for the term capacity to be meaningful. For a mineral operation, limiting conditions include time and economics, ore grade and recovery, the state of repair of the plant, the throughput capacity of individual sectors of the operation (such as mine, mill, smelter or refinery), and the level of capital expenditures required to remove bottlenecks. A ‘soft’ definition often found useful when defining capacity of a mineral operation is that output that can be obtained in a short period of time without ‘excessive’ capital expenditures. This definition takes into account current economics and time, and usually results in the identification of a narrow range of possible capacity levels. Other definitions of capacity, such as rated or nameplate capacity, are often misleading in that an operation may not be able to reach rated capacity without extensive capital expenditure or using higher grade ores than are actually available. Effective capacity must be stated in proper units for each activity. For example, mining and milling capacity must be stated in terms of tonnes of ore mined, whereas smelter and refinery capacity must be stated in terms of tonnes of material processed. Changes in grade of ore or feed material affect production capacity. All costs can be calculated for two production levels: at actual operating rates, which may be less than full capacity, and at full capacity operating rates. Costs at actual production rates are of interest in that the current financial conditions of the production unit can be identified. However, costs at full capacity production rates are more instructive. Firms attempt to operate at the rate that yields the greatest profit, which is usually near full capacity. Therefore, managers plan to operate plants at full capacity’” even though this is often not achieved. The fact that full capacity production is the goal of each plant manager means that the relevant cost in determining the behaviour of the production units and the industry is at full capacity. Plants do operate at less than full capacity and sometimes have lower operating costs in the short run than if they had operated at capacity. Certain costs, such as maintenance, research and exploration can be postponed in the short run, resulting in lower costs whether the plant operates at capacity or not. More often, higher than average grade ore is mined to lower costs. In such cases it may not be physically possible for the mine to produce sufficient high grade ore to run a plant at capacity
RESOURCES
POLICY
September
1988
Competitive cost unalysis in the mineral industries
rates. When higher than average ore grade is mined it is difficult to tell if this is a temporary feature or a permanent situation resulting in more rapid depletion of reserves. If raising the grade of ore mined at the expense of shortening the life of the mine is the only way to reduce costs sufficiently to prevent bankruptcy, then high grading of reserves becomes a permanent feature as long as prices do not increase.
Competitive cost analysis of the nickel industry
160p
cit.
Ref 5, Torries
and
Martens,
annual issues, 1981-1985.
RESOURCES
POLICY
September
Production data were collected and costs calculated for individual nickel production units on a worldwide basis from 1980 to 1985.16 The 1985 costs are calculated for 33 production units, which includes all major and most minor nickel producers in the world. Costs are accumulated on an itemized basis for each activity (mine, smelter or refinery) and are based on specific prices for input factors such as oil, electricity, coal, labour, material and supplies and exchange rates in which the various trades occur. Changes in any of the input factor prices or exchange rates change the results of the cost calculations. Estimates made of 1985 world nickel industry production costs and resulting industry cost curves at 1985 actual and full capacity operating rates are given in Figures 4 and 5 respectively. Total, break even, and net operating costs for each production unit are shown at actual and full capacity operating rates. Total demand for primary nickel in 1985 was about 1.3 million lb. At this rate of production, the cash break even cost curve at full capacity production suggests that a stable nickel price would be in the neighbourhood of $2.30 per lb. During the first half of 1985 the nickel price rose to $2.62 per lb because of strong nickel demand and declining nickel stocks. During July 1985, demand slackened and deliveries increased, causing nickel prices to retreat to the $2.20 to $2.30 per lb range. As shown in Figure 4, the higher cost producers did not recover their cash break even costs at this price level, and some did not even recover their net production costs. Most production units did not recover their total costs, which meant that few production units achieved a positive return on capital during 1985. A rational response of those units indicated in Figures 4 and 5 as not recovering cash break even costs would have been for them to shut down. Some did, but some did not. There are a number of possible reasons to explain why all high cost producers did not shut down. It is likely that the costs shown in Figures 4 and 5 are too high for those producers because shut down costs (m) and price expectations (PC) were not adequately treated in the analysis. During 1985 nickel producers anticipated an increase in nickel demand (which largely did not happen), a decline in nickel stocks, and an increase in nickel prices. These incentives may well have contributed to the continued operation of the unprofitable production units. In addition, all production units were at that time embarking on or completing extensive cost reduction programmes and anticipated improved profits through lowered costs. Even though prices started to increase in 1985, partially because of speculative reasons, there was no evidence that prices higher than $2.30 could be sustained on the basis of production costs. When demand failed to materialize as expected, prices immediately fell back to the $2.30 per lb range, or that price range indicated by the cash break even cost curve.
1988
201
Competitive
cost analysis
in the mineral
industries
4’oo 9 3 3.50 .u c G & 3.00 Total 2.50
-
2.00
-
cost
1.50 -
1.00 -
0.50
Figure 4. industry cumulative cost curve for nickel at actual 1985 production rates.
0.00
Break
even cost
5 f >
I-
O
200
I
I
I
400
600
I
I
1000
600 Million
I
1200
1400
lb nickel
Later in 1985 two other events reduced nickel production costs; oil prices declined and exchange rates adjusted relative to each other and the US dollar. The probable results of the combination of these two events were tested in the nickel model. The results of this analysis indicated that the cash break even costs for the marginal producer to supply the anticipated 1.4 million lb of nickel required in 1986 were around $1.70 to $1.80 per lb. This was based on oil prices of $15 per barrel. By the beginning of 1986 nickel prices were in the $1.70 per lb range. These results again indicate that the model and the use of the cash break even cost curve do result in reasonable predictions.
2.50
-
2.00
-
1.50
-
1.00
-g
et operating
? x Break
f 0.50
Figure 5. Industry cumulative curve for nickel at full capacity production
202
rates.
cost 1985
0.00
-:
’
0
even cost
E E ’
1’
200
I
400
cost
I
I
600
800 Million
I
I 000
I
1200
l
1400
I
1600
1
1800
lb nickel
RESOURCES
POLICY
September
1988
Competitive cost analysis in the mineral industries
The assumption of flat marginal and average cost curves can be tested by comparing the costs of the individual producers at full and less than full capacity operations. Costs at less than full capacity are generally higher than at full capacity if differences in ore grade, interest and overhead charges are not changed, indicating that average costs are not flat but do decline up to some capacity figure. There are sufficient exceptions to this to prevent generalizations. Since 1982 a significant feature of the world nickel industry has been the continued and dramatic decrease in the cost of producing nickel. For example, break even production costs of Falconbridge’s Canadian operation have declined a full $1 .OO per lb since 1982. I7 The decrease in costs for all producers has caused the industry supply curve to shift downward, as is shown in Figure 6. In the case of INCO’s Sudbury operations, costs have been cut in half. The process by which INCO reduced costs at Sudbury illustrates some of the pitfalls that may be present in calculating competitive production costs for individual production units. INCO operated Sudbury at less than half of full capacity in 1982 and 1983. During this period INCO laid off half the workers at Sudbury. When INCO did increase production starting in 1984, it did so without rehiring any of the workers previously laid off. This was accomplished by increasing labour efficiency in mining and production, by using less labour-intensive mining and processing methods, and by changing the mix of products produced to one requiring less labour.”
Data gathering and cost of information Actual cost data on individual production units in an industry are of use to many economic researchers. However, gathering these data is time consuming and expensive even though computers make managing and updating the data relatively easy. In addition, factor prices and usages, demand for the final product and exchange rates constantly change, quickly making estimated costs obsolete. The nickel study for which the results are given in this paper was first
’ 71bid. “Op tit, Ref 15.
1982
/ 3.50
-
3.00
-
s c 2.50
-
$ ;
2 % 2.00 2 II r, 1.50
-
s
Figure 6. Decline in cash break even 1982 to of nickel production 1985 at full capacity production rates.
POLICY
September
-
0.50
-
0.00 L 0
costs
RESOURCES
1.00
I 500
I 1000
I 1500
2000
Million lb nickel
1988
203
Competitive cost analysis in the mineral industries
started in 1980. Since then, the computer program and the database have been revised and improved yearly. Duplicating a similar effort would require considerable time and would involve visiting most of the nickel operations in the world, conducting research in process metallurgy and production costs, and developing a computerized database and costing program. Because of the perishable nature of production data, it is necessary to update the data and costs on at least a yearly basis. In addition, it is necessary to visit each production unit at least every other year after the initial visit to keep abreast with changes in factor costs and usage. The significant cost of yearly updates must be considered when an industry cost study is being contemplated.
Conclusions Competitive cost analysis is a powerful tool that can be used to evaluate mineral projects, predict the behaviour of individual producers within an industry and, under appropriate conditions, give indications of future prices. Cumulative cost curves can be determined for an industry and linked theoretically to true industry supply curves. The key to competitve cost analysis is to determine costs in a consistent and rigorously defined manner for all firms within an industry. Accomplishing this is sometimes difficult because classifications of costs vary among firms and change over time. Marginal variable costs are particularly difficult to determine although these costs, under conditions of perfect competition, do form the theoretical basis of the industry supply curve. It is, however, possible to determine net, cash break even and total costs for each firm in an industry, and use these costs in meaningful ways to construct cumulative cost curves that approximate industry supply curves. Each of these curves can be generated for actual and full capacity production rates. The curve showing the most predictive ability, both in theoretical and applied terms, is the cash break even curve at full capacity. The existence of two supply schedules, one for increasing prices and an expanding industry and one for decreasing prices and a shrinking industry, has not been widely recognized. Entry and exit costs plus managerial expectations create the two supply schedules. The two curves create the situation in which changes in prices over certain ranges do not result in changes in quantities produced by the industry. Different methods must be used in determining costs for producers selling homogeneous products, such as cathode nickel, and for producers with heterogeneous products, such as coal. In the case of commodities such as coal, user costs must be included in the analysis. Competitive costs of the nickel industry have been determined during a five-year period. While exit and entry costs were not adequately treated, the resulting industry cash break even curve did indicate which producers were likely to stay in business and which were likely to shut down. In addition, the cash break even curve, in conjunction with an estimate of primary nickel demand, did give indications of future nickel prices.
204
RESOURCES
POLICY
September
1988