The Bureau of Labor Statistics and Stigler-Kindahl Data.
Our examination of the Bureau of Labor Statistics (BLS) price
ser ies for steel mill products, finished steel products, and
semi-finished steel products reveals an almost steadily nondecreasing trend. As the BLS obtains its data from list rather
than from transactions pr ices, it is natural to ask whether the
true prices would manifest a cyclical trend even if the BLS
series does not¡ i.e., might nonreported discounts off list be
more frequent dur ing the trough of the business cycle.
Stigler and Kindahl (33) published what they regard as
transactions pr ice indexes in major industr ies on a monthly
basis for the years 1~57-66.
These indexes were computed
i
¡
-171-
from confidential data supplied by purchasers of the products
reported.
with respect to steel, they compared their aggregate index
of the nine steel products in their study with the BLS index
of finished steel products and found:
The BLS and NB prices of steel products move
together so closely that a description of one
is a description of the other. The upward
trends in price are essentially th~.same: .05
percent monthly (BLS) vs. .03 percent monthly (NB).
Neither index displays a noticeable cyclical movement in either expansion or contraction. Nor are
the short-run fluctuations of appreciable size.
Figure I is a graph of the da ta discussed in the above
quotation. There were two recessions in the time period covered
by the Stigler-Kindahl data:
July 1957 to April 1958 and May
1960 to February 1961. The absence of noticeable cyclical
movement in the steel industry surpr ised St ig ler and K indah 1.
It stood in contrast to the other industries they studied; for
each of the other industries, their index of transactions prices
showed more cyclical behavior in prices than were exhibited
by the BLS indexes.
Generally, BLS indexes of steel pr ices move in the manner
predicted by the administered pricing hypothesis. Critics of
the hypothesis allege that actual or transactions prices differ
markedly from the published or list prices that the BLS reports.
Thus, if the Stigler-Kindahl data, which purport to measure
actual prices and move closely and noncyclically with BLS prices,
are relevant to current steel pricing, then administered pricing
-172-
may be a fair characterization of industry pricing policy.
There is considerable evidence, however, that this is
incorrect.
The Stigler-Kindahl data cover the decade beginning in
January 1957. In section I of this chapter, it was suggested
that a change in pricing policies occurred in the steel
industry around 1960. This implies that conclusions based on
that part of the Stigler-Kindahl data wh~ch. is pre-1960 might
not apply to current pricing practices
. There were two contractions in the per ioà covered by the Stigler-Kindahl data.
The data in table 4.5 reflect a slight decline in prices in
the post-1960 contr act ion wh i Ie the coun ter-cycl ical pr icing
was for the pre-1960 contraction.
The Steel Trade Press Data. An examination of the steel
trade literature of the past 10 years reveals considerable
evidence that contradicts the impression of "administered
pr icing" obtained from the BLS and Stigler-Kindahl data.
In early 1968, steel buyers began accumulating inventories
in anticipation of the expiration of the contract between the
Un i ted Steelwor ker s and the maj or steel compan ies. Despi te
the hedge buying, certain areas, such as Florida and the Great
Lakes region, were experiencing heavy import competition.
United States Steel secretly offered to meet the low pr ices
of foreign mills to certain customers on some important products including gas and water pipe, galvanized sheet