using oxygen as an
example, assume that both countr ies use the same amount of
oxygen but that one country's industry purchased its oxygen
I
:
-100-
and the other produced its oxygen. Using this study's measurement method, the purchaser's steelmak ing costs would appear to
be lower even if actual costs for the two countr ies were the
same. Outlays for oxygen by the purchasers would not be counted
while the producer's cost series for electricity and labor
would reflect the cost of inputs to produce oxygen. This would
give a misleading indication of relative costs at any point in
time, but the trend of relative costs ov~~t.ime might still be
accurate if oxygen costs remained a constant proportion of total
costs. If oxygen processes became an increasingly important
part of steelmaking, however, there would be a misleading indication of an increasing relative cost trend in the oxygen
producing country even if both countr ies increased their
expenditures for oxygen at the same rate.
Little imagination is needed to identify the likelihood
of numerous possible distortions in this study's relative cost
measures both at points in time and as indicators of trends
over time. The authors believe they have used the best available data, however, and must rest the validity of their results
on the assumption that most distortions are insignificant and
that even significant distortions are offset by others to yield
an insignificant net effect. They do believe that the trends
in the relative costs computed are generally correct in direction and approximate the magnitude of the actual changes.