You are on page 1of 12

STEEL

Why minimills
give the US huge
advantages in steel

44 THE McKINSEY QUARTERLY 1996 NUMBER 2


Performance, not overcapacity, has been the real cause of industry problem

Full costs for mini-sheet hot band? $275 per net ton

Look for another wave of restructuring

Louis L. Schorsch

ITHOUT DOUBT, the minimill rev-

W olution has had profound and


lasting eƒfects on the world steel
industry. Minimills – small-scale steel plants
embodying a superior technology and, more
importantly, more streamlined management
processes – emerged more or less simul-
taneously in the United States, Southern
Europe, and Japan in the 1950s. Despite a
limited product range, minimills grew to
more than 20 percent of production in all
these regions by 1990.

Such plants were still, however, precluded


for technical reasons from participating in
steel’s largest market – the sheets used in
automobiles, containers, and other major
markets. This changed with the commis-
sioning of Nucor’s Crawfordsville, Indiana
plant in 1989. The success of this facility
initiated a surge of new ones that are revi-
talizing the US steel industry. By the end of
this decade, world-class mini-sheet plants
will represent more than 25 percent of US
hot-band capacity. By 2000, almost 60 per-
cent of American sheet capacity will be
highly competitive by international stan-
dards, up from less than 20 percent in the
early 1980s – an extraordinary turnaround.

Lou Schorsch is a partner in McKinsey’s Chicago


oƒfice. Copyright © 1996 McKinsey & Company.
All rights reserved.

THE McKINSEY QUARTERLY 1996 NUMBER 2 45


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

Exhibit 1
In stark contrast, Europe and Japan have adopted this new
Installed minisheet
capability, 1995 production process at a glacial pace. All told, there are
around 10 million net tons of mini-sheet capacity operating
Percent

100% = 10 million net tons


in North America, Europe, and Japan, most of it repre-
Western
senting the “compact strip production” (CSP) technology
Europe used by Nucor and developed by the German equipment
Japan
6 supplier SMS. More than three-quarters of this capacity –
15
around 8 million tons – is in North America (Exhibit 1). Of
%
this, roughly 6 million tons have been installed within the
79 past two years. By comparison, the sole facility in Europe,
Finarvedi’s Cremona plant, began operations in 1992, as did
United States
Japan’s only facility, Tokyo Steel’s Okayama plant.

The North American lead in this technology is likely to persist for at least
the rest of the 1990s. Firm announcements – investments that are either
already under way or apparently backed by sound financing – suggest that
at least another 8 million tons of mini-sheet capacity will have been
installed in North America by the end of the decade. This compares with 2
million additional tons in Europe and none in Japan. North America will
thus be home to more than 80 percent of the developed world’s mini-sheet
capacity until at least the end of the century (Exhibit 2). Given the
competitive advantages enjoyed by such facilities, slow adoption by
European and Japanese producers is bound to create problems for those
industries in the future.
Exhibit 2

Estimated minisheet capability in 2000


Million net tons
Already Firmly Under serious Total
installed announced consideration
United States 8.0 6.0 4.0 18.0
Europe 0.6 2.0 0.6 3.2
Japan 1.5 1.8 3.3

A seminal technology
The fundamental benefits of the thin-slab process used in mini-sheet plants
over conventional integrated steel production are undeniable. It oƒfers
tremendous savings in capital, overwhelmingly the most critical challenge
facing traditional steelmaking technology. It reduces operating costs by at
least as much as the two other seminal steel innovations of the postwar era,
the basic oxygen process and continuous casting. And it creates incentives
for further innovation, thus helping to revamp steel production processes.
Saving capital and operating costs
Thin-slab casting solves the principal problem facing integrated production
techniques: investment requirements that are not supported by market

46 THE McKINSEY QUARTERLY 1996 NUMBER 2


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

prices. Underlying this problem is sustained and substantial excess capacity,


which has plagued the world’s steel industry for the past 20 years. Low
prices and profits over that period have in eƒfect signalled that capacity
should exit the industry. Traditional producers have responded by cutting
capacity to bring it more in line with consumption and thus boost prices.
Higher prices, however, encourage market entry, which in turn revives the
excess capacity problem. The static solution – closure and exit – reinforces
the dynamic problem, entry.

At its root, the global steel crisis has always been about performance rather
than capacity balancing. The source of the excess capacity is more eƒficient
entrants: entrants that have found a lower-capital-cost route for steel
production, that can earn an attractive return at typical market prices, and
that have solved the problem of creeping obsolescence that plagues
increasingly undercapitalized integrated plants. Here lies the key advantage
of the minimills.

Against this background, the importance of thin-slab casting emerges in


sharp relief. Prior to the commercialization of this technology, small-scale,
low-capital-cost options existed for each of the process steps required for
sheet production except hot rolling. This process eƒfectively acted as a
bottleneck obstructing the development of a capital-saving alternative for
sheet production. Conventional hot-strip mills have a minimum eƒficient
scale of around 4 million tons and cost around $200 to $250 per ton of
installed capacity in the United States. The CSP Exhibit 3
technology removes the bottleneck, providing a Potential thin-slab versus
minimum eƒficient scale of around 2 million tons at integrated operating costs
a cost of less than $100 per ton of capacity. Casting/rolling stage
t e d nt

$ per net ton*


an t ie
pl gra ffic
t† t

In addition, thin-slab casting oƒfers a potential


an ee

te e
pl ish

in hly
in

reduction, compared with conventional practice, of


ig
M

about 10 percent in cash operating costs for hot- Overhead


3 7
band production. For eƒficient producers, the saving Parts and
amounts to between $20 and $25 per metric ton supplies

(Exhibit 3). It reflects better performance across the


board: higher yields, lower energy requirements, 19 15
Labor
and superior productivity. Applying direct or hot
charging in a conventional operation might reduce
the gap by $3 to $4 per ton. 4 19
Energy
2 4
How do such savings compare with those derived Yield
from previous breakthrough innovations? Conven- 2 7
Total 30 52
tional slab casting had a similar impact of around
* Excluding depreciation, including plant
10 percent of total costs, perhaps more. With the overhead
† 2.6 million tons shipped from CSP facility
basic oxygen furnace (BOF), possible savings in cash

THE McKINSEY QUARTERLY 1996 NUMBER 2 47


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

operating costs were lower – sometimes less than 5 percent. Other advantages,
particularly melt quality, underlie the BOF advantage over the open hearth.

Where capital costs are concerned, thin-slab casting enjoys a definite


advantage. From a greenfield perspective, a CSP complex incurs roughly
40 percent of the capital charges per ton associated with a conventional
caster and hot strip mill. By contrast, both the BOF and conventional caster
oƒfered modest – if any – savings in capital expenditures per ton than the
technologies they replaced. From a brownfield perspective, conventional
slab casting oƒfers a higher return, but only by a small margin: a payback of
about four years, as opposed to five for thin-slab casting. Thus thin-slab
casting merits a place alongside the BOF and conventional continuous
casting as one of the critical postwar innovations in steel production.

In this light, the discrepancy in adoption rates seems even more puzzling.
With the BOF and conventional continuous casting, Europe and Japan
led the United States in adoption. Yet only the United States looks likely to
show much progress toward adopting thin-slab casting technology this
decade (Exhibit 4).
Exhibit 4

Technology adoption rates


Percent of output 10 years after commercial introduction

Basic oxygen furnace Continuous casting Mini-sheet plants*


1954–64 1962–72 1989–99
United States 15 5 32
Western Europe 16 8 3
Japan 40 16 3
* Estimated

Two arguments can be used to justify delays in adopting thin-slab casting.


The first views it as an interim technology, soon to be replaced by strip cast-
ing. This seems doubtful. Given that the output of a successful strip casting
machine tends to be relatively low, the operating
Exhibit 5

Estimate of strip casting cost advantages will probably be outweighed by


versus SMS/CSP scale diseconomies. On a full-cost basis, then, carbon
Strip casting
strip casting is unlikely to yield any cost advantage
SMS/CSP
over thin-slab casting (Exhibit 5) and in addition
Capacity 1.0 has to overcome substantial technical problems.
Million net tons per year
2.6
Cash operating cost 209
$ per net ton hot band* The second and more convincing argument con-
214
Capital cost
cerns product capabilities. CSP technology is unable
150
$ million
310 to serve the entire market. Width, gauge, surface,
Full cost at 15% return† 232 and deformation constraints limit the range of
$ per net ton
232 products it can supply to an estimated 60 percent of
* Assumes liquid steel cost of $178 per net ton market requirements (already higher than the 50
† Excluding overheads
percent that was feasible for the early commercial

48 THE McKINSEY QUARTERLY 1996 NUMBER 2


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

operation of Nucor’s Crawfordsville facility). In principle, there is no reason


why the technology, like conventional continuous casting, could not even-
tually approach 100 percent of market requirements. By contrast, a scrap-
based carbon strip casting operation is unlikely to be able to supply more
than 26 percent of the market’s needs (Exhibit 6).
Exhibit 6

Potential thin-slab and strip casting penetration post 2000


Example: United States
Percent

Total hot band production 100


Dimensional limitation 16
Thin-slab casting dimensional application range 84
Deformation limitation and scrap-based constraints 13 15
Typical thin-slab casting application range 56
Strip casting limitations (surface and flatness) 30
Strip casting application range 26

The current product limitations of thin-slab casting technology, however,


make it unsuitable as a replacement for existing casting and hot-rolling
operations. This is the chief technical reason why its adoption on a global
scale has been slower than that of other revolutionary steelmaking tech-
nologies. As its capabilities expand, this obstacle will fade.
Promoting innovation
The final reason for the importance of thin-slab casting is its contribution to
the pace of industry innovation. It is a critical element in forcing steel
producers and equipment suppliers to overcome the problem of creeping
obsolescence by developing less capital-intensive methods of steel-
making. In particular, one of the industry’s most interesting technologies,
the development and refinement of scrap substitutes, has received a
tremendous boost from thin-slab casting.

Scrap is one of the key concerns for investors in thin-slab casting tech-
nology, since such facilities must currently be linked to electric furnaces in
order to maintain their low unit capital costs. Quality requirements for
sheet products mandate low residuals – a condition that can only be met
through the substantial use of virgin inputs. The spectacular rise in thin-slab
capacity in North America has thus greatly increased the incentives for
developing scrap substitutes.

The best current example of this phenomenon is iron carbide, which is now
being commercialized – not coincidentally – by Nucor to feed its mini-sheet
plants. Nucor could have pursued a more proven technology, such as direct
reduced iron (DRI), but chose to take a risk on iron carbide, which because
of its high carbon content has a potential value-in-use advantage of at least

THE McKINSEY QUARTERLY 1996 NUMBER 2 49


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

Exhibit 7
$10 per ton (around 8 percent) over the best DRI
Estimated economic value of
scrap and scrap substitutes* (Exhibit 7). Moreover, iron carbide should be cheap-
er to produce than most currently available scrap
Example: United States
Premium (discount) over No. 1 heavy melt substitutes, both because its ore input does not
Merchant pig iron 35† require intermediate processing and because its
capital costs are relatively low (Exhibit 8).
Granulated iron 29†
Iron carbide 26†
No. 1 bundles 25
Given these advantages, the successful commer-
cialization of iron carbide will ultimately produce a
Fior HBI 15
Midrex high met DRI HBI 13 net saving for mini-sheet plants of at least $16 per
Midrex low met DRI HBI 11
ton of hot band (perhaps much more if the tech-
Shredded 5 nology works as advertised) compared with the cost
No. 1 heavy melt 0 of using more conventional DRI products. More-
over, other technologies exist that may prove supe-
rior to iron carbide, such as Fastmet. Innovation
No. 2 heavy melt –14
thus stands a good chance of solving the ferrous
input problem posed by the low residuals require-
ments of mini-sheet plants.
No. 2 bundles –27
The story does not end there. One of the limitations
of iron carbide, particularly as a feedstock for mini-
sheet plants in developing countries, is the diƒficulty
of using it for more than about 25 percent of the
total charge. If a furnace can be designed specifi-
cally for iron carbide, however, this constraint will
ease. In theory, the cost per ton of crude steel made
in such a vessel would be very low.
Turnings –65
* Quality premium estimated at $5.3 If it follows this path, iron carbide – originally
per 0.1 percent reduction in residuals
(actual midpoint of last cycle) regarded as a potential scrap substitute – will enter
† Postcombustion capability would increase
these values by about $3 per ton the race to develop new steelmaking processes on
a small scale, at low capital cost, and without the
environmentally problematic coke required by traditional blast-furnace
technology. With the exception of Corex, the various technologies now in
that race – Hismelt, DIOS, the AISI-DOE direct steelmaking research, and
so on – are all in early stages of development, leaving both their technical
feasibility and their likely economics in the realm of speculation. Never-
theless, initial estimates suggest that iron carbide could be highly compet-
itive, producing liquid steel at a cash operating cost of about $122 per net
ton – almost $25 per ton less than the next best alternative, the combination
of Hismelt and a conventional BOF (Exhibit 9).

Far in the future, the mix of steelmaking technologies may look very
diƒferent from that prevailing today. From a cost perspective, thin-slab
casting will remain the preferred method of converting liquid steel into

50 THE McKINSEY QUARTERLY 1996 NUMBER 2


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

Exhibit 8

Estimated full cost of scrap substitutes, 1993


Example: New Orleans
$ per ton of substitute
Cash cost Capital cost* Total Rated capacities
Gross tons per year

Midrex HBI 111 34 145 1,100,000


Hyl-III 113 41 154 1,100,000
Fior 106 41 147 1,100,000
Iron carbide† 87 26 113 350,000
Fastmet 80 32 112 450,000
* At 15 percent return
† Assumes one-stage process

bands, with a full conversion cost, including capital charges, of around $45
per ton. Consequently, the ideal total complex of the future would link some
new primary technology with thin-slab casting.

Should direct steelmaking from iron carbide prove successful, hot band
might be produced at a cash cost of $154 per ton (Exhibit 10) and a full cost
of less than $200 per ton – far below the next best alternative, and $75 per
ton less than the projected full cost of today’s best option, iron carbide and
Exhibit 9 Exhibit 10

Technology cash operating costs Cash operating costs of 4 technology options


$ per net ton of liquid steel $ per net ton of hot band
ce

Iro SP g a ect

st g a c t
na

nd

g d
ca kin ire
in n
e* rt
l

C in r
ur

ee

S/ ak di
ac -a

rip a d
lf

g
st

rn he
st e

SM l m ide

st l m ide
in
ee
ct id

ct

fu -t

st
re rb

ee rb

ee rb
ire

S/ d

rip d
C f

CS

ca
D e-o

SM an

st an
di ca

s t ca

s t ca
Id

F†
n

n
at

F
IS

BO

Options
EA

EA
Iro

Iro
St
A

Other Other

25 31 20 22
Labor 45 37 47 40
7 7 3 21 Labor
Energy

14 10 46 44
Energy
43 42 12 59 10 16 12 18
Materials Materials

47 77 138 82 145 146 49 48


Total 122 157 173 184 Total 214 209 154 150
* 2.7 million ton shop
† 5.0 million ton shop * Including overhead

THE McKINSEY QUARTERLY 1996 NUMBER 2 51


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

scrap plus electric furnace plus thin-slab caster. In short, new technologies
built around thin-slab casting – and encouraged by its introduction – oƒfer
potential cost reductions on a scale unlike anything we know today.

Why Europe and Japan are lagging


Solid, although probably not suƒficient, economic reasons exist for the
global imbalance in mini-sheet plants. Mini-sheet returns are ultimately
driven by the cost advantage this technology provides. Both of the key
Exhibit 11
inputs – scrap and electricity – are traditionally
Cost and price of hot band much cheaper in North America than in Europe or
$ per net ton
Japan. Other things being equal, hot band costs at a
400
380 North American mini-sheet operation will be some
1994 hot- 5 to 20 percent lower than at an identical plant in
band price 340 Europe or Japan.

Total cost 274 281 285 The competitiveness of the integrated sector also
20
Capital* 38
30 has an influence. Marginal European and Japanese
53 players are more competitive than their US counter-
Cash 53
operating 53 15 parts, raising an additional barrier to entry.
cost† Other 13
Labor 13 47
35
Energy 15 Nevertheless, investment in a mini-sheet facility
represents an attractive economic proposition for
all three regions. Year-to-year fluctuations in market
Materials 155 150 150 prices can significantly alter the relative attrac-
tiveness of the technology by region. In early 1994,
for example, the United States provided the highest
returns, while good returns are available in all three
es

regions, yet outside, its position had reversed by the


an

pa
at

Ja
St

er
d

end of the year (Exhibit 11).


G
te
ni
U

* 2 million tons CSP line, including working


Political factors are partly to blame. The steel indus-
capital and start-up costs; based on cash
returns with no reinvestment
† Including overhead tries of Europe and Japan are more vertically inte-
grated: for example, product distribution channels
are frequently tightly controlled by integrated producers. Such producers
also play a much larger role in scrap collection than they do in North
America. This creates problems for potential mini-sheet investors, who are
more dependent on high-quality scrap supplies than traditional non-sheet
minimills and, for product quality reasons, rely more heavily on distributor
customers than do their integrated sheet competitors. In sum, the greater
vertical integration of traditional steel producers in Japan and Europe
increases the risks associated with mini-sheet investment.

Finally, government intervention continues to constrain entrepreneurial


activity in the European steel industry. Though direct political involvement

52 THE McKINSEY QUARTERLY 1996 NUMBER 2


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

is less overt in Japan, the extensive links between integrated producers and
most minimills impose similar restrictions.

Progress outside the Triad


In the developing world, too, the mini-sheet revolution has been slow to take
hold, although it is now gaining momentum. Hylsa’s Monterrey (Mexico)
facility aside, no mini-sheet production yet exists in a developing country.
Until recently, announcements have been limited in relation to the overall
prospects for steel growth in these regions; and many of the projects that
have been announced were subsequently delayed or terminated.

Mini-sheet facilities should be able to earn good returns in many developing


countries, even in the face of relatively high risk premiums. For one thing,
the political constraints present in Japan and Europe are lacking. Even-
tually, mini-sheet facilities will be built in developing countries; indeed,
roughly 10 million tons of thin-slab capacity is expected to be in place by
the decade’s end (Exhibit 12).
Exhibit 12

Thin-slab casting prospects in developing countries


Estimated capacity in 1999
Thousands of net tons

Firm or highly Brazil 900


likely
Total: 10,800 India 1,300
Malaysia 2,000
South Africa 1,100
South Korea 3,300
Thailand 2,200
Under serious China 2,500
consideration
Total: 4,300 Malaysia 1,800
Speculative China 1,650
Total: 6,850
Saudi Arabia 1,900
Turkey 3,300

At least two constraints have restrained mini-sheet investments in devel-


oping countries. First, the ideal mini-sheet plant depends on high quality
ferrous inputs. Since high quality scrap is unavailable in most developing
countries, operations must rely on scrap substitutes, which traditionally
incur cost penalties. Many companies, however, are now announcing invest-
ments in scrap-substitute facilities. Second, mini-sheet operations still suƒfer
more product constraints than conventional facilities, making a larger total
market necessary to support investment. Nevertheless, the small scale of
mini-sheet plants make them ideal options for small but growing markets,
so that they are likely to make up the bulk of capacity additions in most
developing regions.

THE McKINSEY QUARTERLY 1996 NUMBER 2 53


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

Eƒfects in the United States


Ultimately, the high productivity of mini-sheet plants will force a day of
reckoning on the steel industries of Europe and Japan. For the next few
years, however, most of the impact of the mini-sheet revolution will still be
felt in the United States.

The best way to think about this impact is from a business dynamics
perspective (Exhibit 13). Mini-sheet plants enter the market because their
costs allow them to earn an attractive return at market prices. The more
they expand, however, the more they generate excess capacity and thus
depress hot-band prices, reducing the returns available to new entrants.
Additional mini-sheet plants also mean more demand, and thus higher
prices, for high-grade ferrous inputs. This again reduces the returns
earned by new entrants, while increasing returns for investors in scrap-
substitute facilities.
Exhibit 13

Business dynamics of mini-sheet entry

Impact on Impact on price


relative costs and profitability
Impact on
materials
and import
substitution

Impact on raw Mini-sheet Impact


materials prices entry on price

Returns for
investments
in scrap Impact on
substitutes profitability
Exit

How this system balances out will ultimately define both how much
mini-sheet capacity the market will support and what this capacity will
do to hot-band prices. The key variable in input prices is the cost and
availability of scrap substitutes. Iron carbide, probably produced in gas-rich
regions like Venezuela or Trinidad, may well prove the most attractive
substitute material in North America. If the technology works as promised,
iron carbide could be delivered to the Nucor facility in Hickman Arkansas
at a full cost of less than $125 per gross ton, including a 15 percent return
on capital. Even if throughput rates are only 75 percent of target, the full
cost remains below $140 per gross ton delivered.

If a material like iron carbide turns out to act as a cap on prime scrap prices
under normal market conditions, mini-sheet plants will be able to source

54 THE McKINSEY QUARTERLY 1996 NUMBER 2


WHY MINIMILLS GIVE THE US HUGE ADVANTAGES IN STEEL

prime scrap for around $140 per gross ton. Under ambitious but feasible
assumptions about operating and capital costs, the “steady state” full cost
for mini-sheet hot band is around $275 per net ton.

Mini-sheet investors will enter this market so long as they can earn their
cost of capital – that is, until trend-line prices fall to around $275 per net
ton. Plants currently operating or announced will probably be enough to
achieve this result provided imports maintain their current share. If imports
fall, there will be room for more mini-sheet plants, although the equilibrium
price will remain at $275. If iron carbide achieves a full cost of less than
$140 per gross ton delivered, there will again be room for more entrants,
albeit this time with a negative eƒfect on price.

These numbers represent a threat to traditional integrated players, since


many plants have current cash operating costs – not to mention fully loaded
costs – that exceed $275 per net ton. The pressures that reshaped the
industry in the 1980s will return once the current boom has passed.
Ultimately, the same holds true for integrated producers in Europe and
Japan. The high-productivity, low-cost formula of the minimill revolution –
now roughly 40 years old and challenging traditional producers across the
entire range of products – cannot be ignored.

THE McKINSEY QUARTERLY 1996 NUMBER 2 55

You might also like