Forecasting the Nickel Market
Put yourself in the position of senior management of one of the world’s
largest nickel mining companies in 1990. Currently your company has about
a 30 percent share of world nickel sales, and your largest mine, in Canada,
produces nickel more cheaply than anyone else. The company correctly
predicted the high growth in demand for nickel in the 1980s and profited
handsomely from it. Management has just completed a long-range forecast
(for the next decade) of the nickel market. The forecast predicts that:
1. World nickel sales (in tons) will increase at an average rate of 6
percent per year during the coming five years and 4 percent for the
remainder of the decade.
2. Nickel prices will remain strong, growing at an average of 7 percent
per year, considerably faster than the overall rate of inflation.
3. Continuing the trend of the past five years, new competitors will
expand into the industry. The likely effect is an erosion of the firm’s
market share from 30 to 25 percent.
4. Together, input prices (labor, plant and equipment, energy), extraction
costs, and transport charges will raise nickel’s cost per ton by an
average of 4.5 percent per year over the decade.
In light of these forecasts, the company has decided to undertake a $1
billion investment to expand production facilities in Guatemala and
Indonesia. With this expansion in capacity, the company should profit
from growth in the nickel market and minimize the potential loss of
market share. From its past record of forecasting accuracy, management
is confident in its long-range plan. But a final evaluation of its strategy
can be made only seven to ten years from now.
Discussion
The large mining company was “bullish” on the prospects of world nickel
sales during the 1990s after seeing its own shipments steadily increase
during the 1980s. In fact, based on a linear time trend fitted to the past data,
it estimated its own growth in nickel shipments roughly at an additional 12
thousand tons per year (see Exhibit 1.a). At this rate, demand would outstrip
capacity in fewer than five years. Accordingly, the firm began expansion of
its Indonesian and Guatemalan facilities. So long as actual nickel sales
John Wiley & Sons
4-?
tracked the company’s projections, it would earn healthy (perhaps even
“obscene”) profits and remain the world’s leading producer of nickel.
With 20-20 hindsight (Exhibit 1.b), one observes that the demand for the
company’s nickel did not quite unfold as forecasted. Shipments in 1990 were
242 thousand tons and only 259 thousand tons per year a decade later.
Shipments declined steeply in the early 1990s, recovered in the middle of the
decade, then took a smaller dip before the decade closed. Used in many
metal products (especially stainless steel), electronic products, and industrial
processes, nickel demand closely tracked the business cycle falling during
the 1990-1991 recession in the United States and collapsing further with the
prolonged 1991 1994 European recession. With the worldwide recovery,
nickel demand rebounded, only to fall again during the East Asian financial
crisis and recession.
Not only did the company’s nickel shipments stagnate; the price of
nickel during the decade dropped as well. Nickel prices dropped from
$9,000 per ton in 1990 to as low as $6,000 a ton in 1993. Although prices
recovered, they dipped again during 1998 and finally regained the $9,000
level by 2000. The price drops were due to cyclical declines in nickel
demand, but also to over-capacity in the industry. The large nickel company
was not alone in expanding capacity in the early 1990s. The third and fourth
largest companies also expanded their production facilities, driving down
prices. All three firms failed to make any profits in the early 1990s. The
number one firm returned to profitability in the middle years of the decade,
beginning to pay back its investment in expanded capacity. However, higher
energy prices at present threaten profits. (Processing the kind of ore found in
the tropics takes one and one-half times as much energy as Canadian ores
require.)
John Wiley & Sons
4-?
Exhibit 1a
1982 1984 1986 1988
1990
26
0
24
0
22
0
20
0
18
0
16
0
14
0
12
0
The large company’s myriad problems resulted from a combination of
bad forecasting and bad luck. Could the company have planned better for the
coming turbulence in world markets? The “bigger” picture of the nickel
market depicted in Exhibit 1.b suggests that the answer is yes. The
company’s 1980s time series of shipments (magnified in part a) appears in
the small box in part b. The bigger picture suggests the following
conclusions:
1. The company’s shipment history during the 1982 to 1990 period was
unusually steady. However, during 1981 and 1982 (a time of worldwide
recession), its shipments had fallen significantly. In fact, part of the
company’s high growth probably represented recovery from previous
depressed levels. Moreover, a look at the nickel industry as a whole would
have revealed much greater volatility in demand year to year during the
during the decade. Though the firm’s nickel sales grew steadily in the
1982-90 period, how long could the company remain shielded from the
reality of global fluctuations?
John Wiley & Sons
4-?
26
0
24
0
22
0
20
0
18
0
16
0
14
0
12
0
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000
300
250
200
150
100
50
Exhibit 1b
2. A linear time trend probably overstates the company’s long-term growth
prospects. (A quadratic specification, allowing for slackening growth,
probably would have fit the data better.) Whatever the specification, the
company should be aware of the standard error surrounding its forecast.
Had anyone asked, the standard error for the linear time trend was large
plus or minus 8 thousand tons around the annual forecast. This error will
increase as forecasts become increasingly long range.
3. The time series did not take into account a key determinant of total nickel
demand; the worldwide level of economic activity (especially steel
production). Nor did it include an important factor influencing the
company’s share of demand: the supply and prices of its leading
competitors. These omissions increase dramatically the margin of error in
the company’s forecast.
4. Japan’s decade-long recession, the East Asian financial crisis, and recent
oil price shocks were all nearly impossible to predict. The company’s
exposure to these adverse developments could well be attributed to bad
luck.
The moral of this story? Often, the worst mistake a company can
make is to ignore uncertainty. Even if its forecasts had been correct on
John Wiley & Sons
4-?
average, the company faced the risk that nickel demand could fall well
below its expected projection. Given the uncertainty surrounding nickel
demand, such a scenario was a significant possibility. Thus, the firm’s
production planning should have taken uncertainty into account.
V.Quips and Quotes
It is a capital mistake to theorize before one has data. Insensibly, one begins
to twist facts to suit theories, instead of theories to suit facts.
(A Conan Doyle)
The only relevant test of the validity of a hypothesis is a comparison of its
predictions with experience. (Milton Friedman)
One accurate measurement is worth 100 expert opinions.
A straight line prediction is the shortest distance between two errors.
In economics, everything depends on everything else — and in more than one
way. (A simultaneity problem?)
He uses statistics like a drunken man uses a lamp post: for support rather
than illumination. (Attributed to Benjamin Disraeli and Andrew Lang)
The future is not what it used to be.
The great tragedy of science: the slaying of a beautiful hypothesis by an ugly
fact. (Thomas Huxley)
Beggar: I haven’t eaten in three days.
Statistician: How does that compare to last year?
A 17-year old dropped out of school and rode his motorcycle to […Fill in
your choice of state], thereby raising the average IQ in both states.
John Wiley & Sons
4-?
The best test of an expert in a given area is the ability to win money in a
series of bets on future events in that area.
The job of forecasters: to make the unknown, the unknowable, plausible to
those who insist on knowing.
Economists are generally right in their predictions but wrong in their dates.
Forecasters: Often wrong, never in doubt.
Even a stopped clock is right twice a day.
Make 3 correct guesses consecutively and you will establish a reputation as
an expert.
Two things one should never see in the making: sausages and economic
forecasts.
He tortured the data until it spoke. (But was it telling the truth?)
John Wiley & Sons
4-?