.jpg)
Changes in relat ive steel product ion costs among countr ies
may have a strong influence on steel trade flows. Given any
initial market situation, relative cost reductions by one
country should allow it to expand into areas formerly controlled n
.~ .
by the countries whose relative costs have increased.
This chapter examines the cost of steelma~in9 in the Uni ted
States over the past 20 years relative to the costs in Japan and
in the European Community. For reasons of data availability,
different methodologies have been used for the Japanese and the
EC compar isons. In the case of Japan, costs are compared wi th
those of the united States on a factor by factor basis for a
number of key input factors. In the case of the EC, price and
profit data are used to compare European production costs with
those of the United States for three particular steel products.
The methodology used for Japan allows a much more intensive
study of costs, including identification of the source of relative cost changes. The methodology used for the EC is not without advantages, however. It allows a more extensive coverage
of aggregate costs and allows inferences about individual prod- ~i
ucts. Both methods allow us to examine the correlation between
relative cost changes and the trade position of the United States
steel industry.
-94-
A number of studies have estimated steelmaking costs for
the United States, Japan, or the EC. Some of these studies
have done intercountry cost comparisons but, with few exceptions, they have covered only one or a few years. Although
they collectively cover many years, their methodologies are
so incompatible as to prevent conclusions from them about longterm trends in relative costs.
The only recent study discovered whi..:h. presents comparative,
cost estimates over a long per iod of time was prepared by Pifer"
Marshall, anò I~errill (P~¡M) for the American Iron anò Steel Institute (27). The PMM study includes an II-year time series of comparative costs between the United States and Japan for the same
items studied here (27, pp. 29, 33).
unfortunately, PMM did not construct independent cost
estimates for basic input items. PMM have informed the authors
that they relied on preliminary estimates made by the FTC staff
for this report. The FTC document which PMM used was in draft
form and contained errors and omissions. These shortcomings are
reflected in the figures reported by PMM and explain most of the
differences between their figures and those appearing in this
chapter.
I I. THE UNI TED STATES AND JAPAN
,Methodology
The method employed to examine steelmaking costs in this
section is quite simple in concept. Data were collected for
-95-
the Un i ted Sta tes and Japan on the quant i ty and average pr ice
of selected inputs used in the manufacture of steel dur ing each ':;!
of the years 1956 through 1976.
Weighting each quantity by the
appropr iate pr ice 9 ives the total cost for each input in each
year. Dividing each cost by the steel output in the respective 4
year yields the cost of the input per unit of output. The unit
costs for individual inputs are summed across inputs to give the
cost of all the selected inputs per uni~. of output. The movements over time of these summed costs are used to gauge changes
in relative costs between the two countries.
Changes in relative costs can be traced to the individual
inputs or groups of inputs causing them, and the extent to which
shifts were due to changes in the quantity employed or changes
in the .pr ice of the input can be determined.
The inputs selected for examination are iron ore, scrap
iron and steel, labor, and a number of energy inputs--coking
coal, other coal, fuel oil, natural gas, and electr ic power.
These inputs accounted for over 70 percent of variable steelmaking costs in the United States in recent years and over 60
percent of total costs. They are believed to have been the I;,:t ,~
inputs most important in causing relative cost changes among
countries. Among the excluded inputs are fluxes, alloys,
oxygen, water, and other purchased materials. Taxes and the
cost of transporting finished products to market were also
-96-
excluded from consideration. II Capital costs, which are not
directly comparable to var iable costs, were also excluded.
Although it would have been preferable to include all variable
inputs and cost factors, the necessary price and quantity data
were not available. Underlying the use of the unit cost
figures calculated is an implicit assumption that the relative
cost between the United States and Japan of excluded inputs has
not shifted significantly over time. ..-.. .
Even for the selected variables, price and quantity data
which are exactly comparable for the two countries do not exist
for some inputs. Attempts were made to adjust the data in some
instances to make it more comparable but the possibility of error
was not eliminated. There are a few basic comparability problems
wh ich wi 1 1 be discussed pr ior to deal ing wi th ind iv idual var iabIes.
First, comparability of input quantities suffers because
of differences among countries in the definition of the steel
industry. Particularly troublesome are differences in the
degree of vertical integration. For example, in the United
States, finishing the edges of steel plate is usually done at
the steel mill; and, for data collection purposes, the labor
and energy inputs required are considered to be employed in
the steel industry. Industry sources have said that in Japan
II Appendix 3B discusses recent transportation costs for
finished products moving between Japan and the Uniteå States.
-97-
steel plates are more often finished by the service centers or
by the users. If this is true, the labor and energy used are
not counted as steel industry inputs in Japan.
A second general problem ar ises due to the difficulty of
defining "the price" of an input. Even in specifying the
ideal, one must choose between data which reflect the marginal
cost of inputs and those which reflect the average cost. Since
many inputs are purchased under long-ter.m. c.ontrac;t in the steel,
indus~ry, there may be significant differences between the average cost and the margi~al cost; and it is not clear which would
be preferable. Marginal conditions, as reflected by spot market
prices, may give an accurate indication of opportunity costs
and an immediate signal of changed market conditions not provided
by average prices. If average prices include a large contract
component, for example they will probably understate opportunity
cost during a period of rising input prices. On the other hand,
current market pr ices may be overly responsive to transient
changes £/ and give a poor indication of fundamental conditions.
Real world data present additional problems. In few cases
do we have the choice between average and marginal indicators;
often there is only one data source available. Average,
del ivered input costs were available for many inputs for Japan;
:)
2/ This is especially true when a market is very thin due to
neavy vertical integration into the supply of the input.
i
!
-98-
but for the U.S., the available data were generally market
prices for some characteristic market or the average transfer
value of a particular commodity. Since U.S. companies are
heavily integrated into the production of iron ore and coal,
many transfers of these commodities are intracompany; and it
is unclear whether publ ished information based on these transfer values reflects average or marginal cost.
Perhaps a more ser ious obstacle to p~~cise cost comparability between the U.S. and Japan involves output data rather
than input data. Steel is not a homogeneous proóuct. There
are hunóreds of steel products which vary in many dimensions--
the type of steel used, carbon, stainless, etc.; the shape of
the product, plate, wire, pipe, etc.; the dimensions of the
product, thick, thin, light, heavy, etc.; and most difficult to
measure, the quality of the product which involves such factors
as the finish, production tolerances, and even promptness and
reI iabil ity of del ivery. The input requirements vary with the
product and unless two countr ies produce the same product mix,
cost comparisons lose their meaning. Failure to adjust for
product mix can result in misleading conclusions concerning
cost competitiveness. As a simple example, if country A produced only steel str ip and country B produced only sheet, it
would be inappropriate to conclude that A and B were cost com-
. petitive in labor simply because each had an average labor cost
of $90 per ton of output. Since strip production is very labor
intensive compared to sheet production, country
A could be
-99-
presumed to have a clear labor cost advantage. Country A
could undoubtably produce sheet for less than $90 per ton
labor cost, and it would undoubtably cost country B more than
$90 per ton for labor to produce str ip.
Compl icating the theoretical problems of adjusting for \')':
product mix is the fact that in available data sources product
classifications are limited and differ among countries, making
comparisons of output mixes difficult. ~r'example, while the -,;
U.S. industry classifies stainless and alloy steel as "specialty
steels," the Japanese incl ude under that term certain types
of carbon steel products.
The var iable input affected most by the product mix is
labor. The Bureau of Labor statistics series upon which the
labor usages in this report are based do include some corrections for differences in product type and shape (43, pp. 20-25),
but these correcti~ns are far from complete and assume no differences in product quality.
Two types of distortions in comparative costs due to data
problems can be distinguished. One is distortions which make
the two countries' relative costs at any point in time misleading; the other is distortions which make changes in the relative
costs over time mi slead ing. An example is d istor t ions wh ich
might ar ise due to an excluded input
