GROWTH of U.S. affiliates slowed significantly in 1982.
Affiliates’ employment, one of the broadest measures of their
activity, increased 1 percent, to 2,435,000, compared with a 19-percent
increase in 1981 (table 1). Total assets of affiliates, another broad
measure, increased 16 percent, to $473.0 billion, compared with a
39-percent increase in 1981 (table 2). The slowdown was mainly a result
of the U.S. recession.
The 1981 and 1982 estimates of the U.S. affiliates’
employment, total assets, and other items presented in this article were
obtained by expanding to universe totals sample data collected in
BEA’s annual survey of foreign direct investment in the United
States. The previously published estimates for 1981 have been revised
(see the technical note); the estimates for 1982 are preliminary and
will be revised next year.
Because changes in employment are not directly affected by
inflation, they tend to correspond more closely than changes in total
assets to growth in real economic activity. For this reason, the
remainder of this article will focus on changes in affiliate employment.
As a result of the U.S. recession, many U.S. affiliates laid off
employees in 1982. The all-affiliate total increased slightly because
the employment added due to acquisitions of U.S. companies by foreign
direct investors more than offset the decreases due to layoffs. A large
portion of the added employment was in retail trade.
Employment of U.S. affiliates in mining, petroleum, manufacturing,
and real estate declined. Except in real estate, the declines occurred
mainly because layoffs were widespread and the employment added due to
acquisitions was small. In real estate, the decline occurred mainly
because a Canadian company sold its minority interest in a major U.S.
real estate brokerage firm.
The largest decline in affiliate employment was in manufacturing
(61,000). Within manufacturing, employment declined in every major
subindustry. Declines were particularly large in chemicals (24,000),
mainly industrial chemicals, and in machinery (15,000), mainly
construction machinery, industrial machinery, and electronic components.
Affiliate employment increased in wholesale trade, retail trade,
finance, and “other” industries. The largest increase was in
retail trade (46,000); it mainly reflected the acquisition by foreign
investors of several sizable U.S. retailers, including national jewelry store and fast food restaurant chains and two regional department store
By country of ultimate beneficial owner (UBO), the largest
decreases in employment were by affiliates with UBO’s in the Middle
East, the United States, and Europe. Among individual countries,
affiliates with UBO’s in France and Germany had the largest
decreases (27,000 and 26,000, respectively). For French-owned
affiliates, recession-related layoffs–particularly by affiliates in
metals, glass, and transportation equipment manufacturing–caused much
of the decrease. Other factors, however, were also important. One was
the sale by French parents of their interests in several existing U.S.
affiliates, the two largest of which together had over 10,000 employees.
Another was a shift, from France to Liechtenstein, of the UBO of a major
affiliate as a result of a change in the affiliate’s ownership
structure. For German-owned affiliates, the decrease in employment
resulted partly from layoffs, particularly in chemicals and machinery
manufacturing. Also contributing was the selling or closing of a number
of unprofitable supermarkets and other facilities of a German-owned
national grocery store chain.
The largest increase in employment were by affiliates with
UBO’s in Canada, the United Kingdom, and Switzerland. UBO’s
in each of these countries acquired several large U.S. companies during
The slow growth in affiliate employment nationwide was mirrored in
most U.S. regions. Affiliate employment increased 1 percent or less in
every region except the Mideast, where it declined 1 percent, and the
Great Lakes, were it increased 4 percent (table 3).
Among States, the change in affiliate employment ranged from a
22-percent decrease in Alaska to a 21-percent increase in South Dakota.
In absolute terms, the largest decrease in employment was in
Pennsylvania (5,300). Widespread layoffs, particularly in
transportation equipment manufacturing, and sales or closures of
supermarkets by a large grocery store chain contributed significantly to
the decrease in that State. The largest increase in affiliate
employment was in Illinois (17,100). Acquisitions of an office
equipment manufacturer and of a Chicago-based department store chain
more than accounted for the increase.
The 1981 and 1982 universe estimates presented in this article were
derived from sample data reported in BEA’s annual survey of foreign
direct investment in the United States (the BE-15). In the BE-15
survey, reports were required from nonbank U.S. affiliates that had
assets, sales, or net income greater than $5 million or that owned more
than 1,000 acres of U.S. land. The universe estimates cover nonbank
U.S. affiliates that had assets, sales, or net income of $1 million or
more or that owned 200 or more acres of U.S. land; these were the size
criteria used to determine which affiliates had to file complete reports
in the 1980 benchmark survey, which is the basis for expanding the
sample data reported in the BE-15 survey to universe estimates.
Data for nonsample affiliates–those in the universe but not in the
current-year sample–were estimated. The nonsample affiliates consisted
of affiliates (1) that were below the exemption levels for reporting in
the BE-15 survey; (2) that were required to report but for some reason,
did not; or (3) that filed reports that could not be processed in time
to meet BEA’s publication schedule.
For the preliminary 1981 universe estimates published a year ago, a
simplified procedure was used to derive estimates of data for nonsample
affiliates. Since then, the estimating procedure has been refined.
This refinement, together with corrections to the data reported by the
1981 sample, resulted in revisions to 1981 universe estimates. The new
procedure, which like the old one, estimates data both for nonsample
affiliates that were in the universe in the prior year (previously
existing affiliates) and for nonsample affiliates that entered the
universe in the current year (new affiliates), is discussed below.
Previously existing affiliates
For each previously existing nonsample affiliate, each data item is
estimated for the current year. The item is calculated as the product
of two factors: (1) the prior-year data for the affiliate and (2) the
ratio of current- to prior-year data for a matched sample of affiliates
(those that reported in both the prior and current year) that were in
the same industry group as the affiliate whose data are being estimated
and that had assets, sales, or net income of less than $50 million. The
implicit assumption in this procedure is that, in a given industry
group, data for each nonsample affiliate changes at the same rate as
data for affiliates in the matched sample.
The matched sample is restricted to relatively small affiliates
because most of the nonsample affiliates are also small. Ratios are
calculated for four industry groups–manufacturing; wholesale trade;
agriculture, forestry, and real estate; and all other. These four broad
groups, rather than more disaggregated industries, are used because, for
some of the more disaggregated industries, the matched sample would have
consisted of only a few affiliates and the reliability of the resulting
ratios would have been questionable. If the calculated ratio is biased
by the data of one or two reporters, or is unrepresentative because of
low coverage, it is adjusted before being applied.
For new nonsample affiliates, estimates are separately derived for
each affiliate based on data they reported in BEA’s survey of new
foreign direct investments in the United States (BE-13). Although the
BE-15 survey covers many items not covered in the BE-13, both cover five
keys items–total assets, sales, net income, employment, and land owned.
For these items, the universe estimates include the BE-13 data, as
reported, for nonsample affiliates. For items covered by the BE-15
survey but not by he BE-13, estimates are computed as the product of two
factors: (1) the one of the five BE-13 items of the new affiliate that
is most closely related to the BE-15 survey item being estimated and (2)
the ratio of the item being estimated to the item in (1), as reported in
the BE-15 survey by affiliates that are in the same industry group as
the new affiliate and that have total assets, sales, or net income of
less than $50 million.
Because most of the new nonsample affiliates are small, the ratios
are computed only for smaller affiliates and only for two industry
groups–agriculture, forestry, and real estate, and all other. Separate
ratios are computed for the former group because a large percentage of
the new nonsample affiliates are in it and relationships among items for
affiliates in the group often differ significantly from those for
affiliates in other industries.
The procedure just described is not used where other available
information indicates that application of a ratio would not produce
meaningful estimates. In these cases, the procedure used varies
depending on the item being estimated. For example, most new nonsample
affiliates are small and do not engage in international trade. Thus,
their exports and imports are assumed to be zero and are not estimated
using the ratio procedure.
Table 4 shows, for both employment and total assets, the percentage
of the 1982 universe estimates accounted for by the 1982 sample data.
At the all-industries, all-countries level, coverage is 91.9 percent for
employment and 93.5 percent for total assets. Coverage falls
significantly below these averages only in industry and country cells
where affiliates tend to be of small average size (for example, in real
estate and “other industries,” and in Latin America and
“other Africa, Asia, and Pacific”).