“x” companies – as “big oil”, “big pharma”, “big...
TRANSCRIPT
1
(draft: 3/30/17)
What Has Been Happening to Aggregate Concentration in the U.S. Economy in the 21st Century?
Lawrence J. White
Stern School of Business
New York University
Jasper Yang
Stern School of Business
New York University
Abstract
Fifteen years ago White (2002) provided estimates of aggregate concentration in the U.S.
economy – the percentage of aggregate economic activity that could be attributed to the largest
“X” companies – that covered primarily the last quarter of the 20th century. Those data sets
showed that despite major merger waves during that period, aggregate concentration at the end
of the 1990s was lower than it had been in the early 1980s, although there had been some upward
movement after the mid 1990s.
Since then (to our knowledge) there have been no studies that have updated/extended
those data. Major mergers are again prominent during the period since the millennium, which
has also been marked by historic business cycles as well as by frequent mentions of phrases such
as “Big Oil”, “Big Pharma”, “Big Food”, “Big Tech”, etc., in popular descriptions of the U.S.
economy.
This paper extends the earlier data series into the first two decades of the 21st century.
We find that there has indeed been a moderate but continued increase in aggregate concentration
since the mid 1990s. This increase appears in data on employment and payroll that have been
compiled by the U.S. Bureau of the Census, as well as employment and profits data that are
drawn from the annual “Fortune 500” lists. This increase does not, however, appear to have
raised aggregate concentration above the levels of the early 1980s.
This paper also computes annual Gini coefficients for 1988-2014 for employment by firm
size and payroll by firm size. We find gradual annual increases in both sets of Gini coefficients
for this time period. These increases appear to be due to increases in the sizes of larger firms
generally and not just increases by the largest firms.
Key words: aggregate concentration; value added; employment; payrolls; profits; antitrust
JEL codes: L10; L11; L19
2
I. Introduction.
The apparent aggregate concentration – the percentage of economic activity that is
accounted for by the largest X firms – seems to be growing in the U.S. economy. Media
references to “Big Oil”, “Big Pharma”, “Big Food”, “Wall Street”, etc., convey the idea that the
U.S. economy is increasingly dominated by large firms. Prominent reporting of proposed and
consummated “mega-mergers” does the same.
Surprisingly – at least to us – quantitative measures of recent levels of aggregate
concentration are not readily available. White (2002) provided estimates for the 1980s and
1990s, but since then (to our knowledge) there have been no studies that have updated/extended
those data. A partial explanation for this dearth is that the relevant data are not regularly
compiled by government agencies.
In this paper we update White’s (2002) estimates with data that cover the first and part of
the second decades of the 21st century. By combining the earlier data with our new data, we are
able to provide a longer sweep of data that – at least for some economy-wide measures of
aggregate concentration – extend back to the early 1980s (and for the manufacturing sector
extend back to 1947).
As a summary, we find that aggregate concentration in the U.S. economy – as measured
by employment, payroll, and profits – appears to have risen moderately but steadily since the
mid 1990s. This increase does not, however, appear to have raised aggregate concentration
above the levels of the early 1980s. We also compute annual Gini coefficients for 1988-2014 for
employment by firm size and payroll by firm size. We find gradual annual increases in both sets
of Gini coefficients for this time period. These increases appear to be due to increases in the
sizes of larger firms generally and not just increases by the largest firms.
3
The remainder of this paper will proceed as follows: In section II we ask why the issue
of aggregate concentration is interesting in the first place. Section III provides a literature
review. Section IV addresses measurement issues. Section V provides the employment and
payroll data from the Bureau of the Census. Section VI provides the employment and profits
data from the annual Fortune and Forbes lists of the 500 largest companies that are
headquartered in the U.S. Section VII contains a brief conclusion.
II. Why Is Aggregate Concentration Interesting?
At the beginning, it is important to separate the measurement of aggregate concentration
from the measurements of seller concentration in specific markets that are the staple of antitrust
investigations and litigation.1 The latter measurements are part of the effort to determine
whether market power – the ability to maintain a price that is significantly above competitive
levels for a sustained period – is present (or could become present) in a specific relevant market.2
The markets that are thereby delineated are often narrow: e.g., local or regional in their
geographic scope, and encompassing narrow product categories. The absolute size of a firm
usually doesn’t matter; it is the firm’s sales as a percentage of all sales in that delineated market
that does matter.
By contrast, aggregate concentration is an economy-wide measure, and the absolute size
of a firm does matter. However, aggregate concentration has little or no relevance for antitrust,
since there is no necessary connection to seller concentration in individual markets. As an easy
illustration of this point, suppose that there were only 100 giant companies that accounted for all
1 See, for example, Kwoka and White (2014). 2 Seller concentration – whether measured by the largest X firms (often X = 4) or by the Herfindahl-Hirschman
Index (HHI) – is, of course, just one component of the market-power determination.
4
of the private-sector output in the U.S. economy. This would be an extremely high level of
aggregate concentration. But if all 100 companies were present as sellers in all U.S. markets
with roughly equal market shares, the measured levels of seller concentration in those individual
markets would be far below the levels that would ordinarily raise antitrust concern.3 And, if one
believed that economies of scale and of scope could still be exploited by companies that each
represented approximately $150 billion in value added (and that the potential diseconomies from
the difficulties of managing such large enterprises were not a problem), then the static
efficiency/productivity of this hypothetical economy would be greater than for a more atomized
economy.
Nevertheless, an economy that was thoroughly dominated by only 100 – or even 1,000 –
giant companies would surely have a different “feel” from the U.S. economy in which we
actually live. To illustrate the latter: The U.S. Internal Revenue Service’s Statistics of Income
reported that for 20134 in the U.S. economy there were: 24.1 million sole proprietorships with
$1.3 trillion in sales receipts; 3.5 million partnerships with $5.9 trillion in sales receipts; and 5.9
million corporations with $30.2 trillion in sales receipts.5 An economy with only 100 giant
companies would surely feel more stifling, with no outlets for entrepreneurial spirits, far fewer
employment choices, far fewer places where innovative ideas might take root and thrive, etc.
Thus, a focus on aggregate concentration addresses questions that are largely about the
“feel” of a society, although there are serious economic (but not antitrust) issues as well.
3 Of course, one would worry about whether multi-market contact among these 100 firms – the same firms would be
“meeting” each other in multiple markets – might be an inhibition on fully competitive behavior. See, for example,
Bernheim and Whinston (1990). 4 Although 2014 data are now available for sole proprietorships and partnerships, only 2013 data are currently
available for corporations. In order to keep the data horizontally consistent, we provide the 2013 data. 5 Since the private-sector GDP in 2013 was only $14.5 trillion, there is an obvious double-counting problem that
accompanies the summation of sales receipts data across the entire private sector. We will return to this problem
below.
5
One other preliminary issue should be addressed: There is no “right” level of aggregate
concentration to which anyone can refer; but it is possible – with sufficiently good data – to
measure changes in the level of aggregate concentration over time. If one believes that aggregate
concentration matters, then the measurement of changes over time to help determine whether
things are getting better or worse (subject, of course, to the caveat about efficiency/productivity
mentioned above) is worthwhile.
It is the measurement of those changes that is the subject of this paper.
III. Previous Literature.
TBA
IV. Some Measurement Issues.
A. The measurement basis for size. What should be the measurement basis for an
estimate of aggregate concentration? Unfortunately, there is no theory to provide guidance. By
contrast, if one is trying to measure company concentration in a relevant market for antitrust
purposes, theory can provide guidance: Since the antitrust issue is usually whether firms in a
relevant market can exercise market power – either individually (as a monopoly, or as a
dominant firm that has monopoly-like market power6) or collectively (as a coordinating
oligopoly7) – the measurement basis should be the one that best represents the firms’ ability to
6 See, e.g., Stigler (1940). 7 See, e.g., Stigler (1964).
6
exercise market power. Typically, sales revenue – or possibly sales units (e.g., for automobiles)
– seems to provide a good measurement basis.8
However, in trying to measure aggregate concentration, the goal should be some indicator
of the “importance” of the largest companies in the economy. There are multiple measures of
the size of a company that are readily available – e.g., sales revenue, employment, payroll, assets
– yet there is no clear link between any one of these size measures and the “importance” of a
company. In 2015 Wal-Mart (the #1 firm by worldwide sales in the Fortune 500 list) had sales
of $298.4 billion in the U.S. (approximately 2% of U.S. private-sector GDP!) and had 1.5 million
U.S. employees;9 Facebook (#157 by worldwide sales) in 2015 had worldwide sales of only
$17.9 billion and had only 12,691 employees.10 Which company is more important? In 2015
JPMorgan Chase (#23 by worldwide sales) had North American (substantially U.S.) assets of
$1.8 trillion. Where does it fit in importance? And McDonald’s (#109 by worldwide sales) had
$8.6 billion in U.S. sales;11 but 90% of its 14,259 stores in the U.S. are franchised:12 Those
stores have stand-alone owner operators, and McDonald’s receives royalties from them (and the
royalties account for slightly more than 50% of McDonald’s U.S. total sales revenue).13 If
8 See, e.g., Stigler (1964). 9 The worldwide totals for Wal-Mart in 2015 were $482.1 billion in sales and 2.3 million employees. The
worldwide sales figure is the basis for its #1 ranking by Fortune, which annually reports the consolidated financial
information for the largest (by sales) 500 companies that are headquartered in the U.S. All U.S. data discussed in
the text are taken from the various companies’ annual reports. 10 Slightly less than half of Facebook’s sales revenues came from the U.S.; but over three-quarters of its assets were
located in the U.S. The company’s annual report for 2015 doesn’t indicate the geographic location/distribution of its
employees. Unlike a largely bricks-and-mortar retailer such as Wal-Mart, where the geography of sales is likely
mirrored by the geography of operations and employment, the more dispersed geography of Facebook’s sales
probably occur largely as exports from the company’s Menlo Park, CA, location. 11 Worldwide sales were $25.4 billion. 12 Approximately 82% of its 36,525 stores worldwide are franchised. 13 About 35% of the company’s worldwide revenue consists of royalties from franchisees.
7
instead, the franchisees’ gross sales revenues were added to the sales of the McDonald’s owned-
and-operated stores, the company’s sales would instead be $35.8 billion.14
Further, there is a clear double-counting problem in any aggregation across companies if
sales revenues are the measurement basis;15 and a similar problem arises if assets are the
measurement basis.16 The usual bases for the measurement of size for non-financial firms and
for financial firms are somewhat orthogonal: For non-financial firms, sales revenues are a
generally accepted measure of size; for financial firms, assets – which, of course, encompass
loans and investments that have been made by the firm and thus represent a central function of
the firm – is a more common measure of size; the revenues of such firms are usually not
connected with the word “sales”.
The contrasts can be striking: In 2015 Wal-Mart had worldwide sales revenues of $482.1
billion and worldwide assets of $200.0 billion; its sales/assets ratio was 2.41. JPMorgan Chase’s
2015 worldwide revenues were $101.0 billion, and its worldwide assets were $2.4 trillion; its
revenue/assets ratio was 0.04. Both firms would be considered to be large and important. But
neither sales revenues nor assets would fully convey their respective sizes. And for McDonald’s,
as was discussed above, its reported sales figures are substantially below what it would report if
14 Similarly, if the worldwide franchisees’ gross sales revenues were added to the company’s revenues from its
owned-and-operated stores, the company’s worldwide sales would instead be $82.7 billion. At this level of
aggregate worldwide “sales”, the company would have been ranked by Fortune as the 31st largest company
headquartered in the U.S., slightly ahead of IBM. 15 Suppose that a manufacturing company that previously consisted of a “components” division and an “assembly”
division (and was thus vertically integrated) decides to dis-integrate vertically into two wholly separate components
and assembly companies. As a recent example, in November 2016 the previously vertically integrated Alcoa Inc.
was split into an “upstream” aluminum smelting and refining company (Alcoa Corp.) and a “downstream”
aluminum parts manufacturing company (Arconic Inc.). On a pro forma basis the two separate companies are
together the same size as the previous vertically integrated company. But if sales revenues are the measure of size,
then the new assembly company would be considered to be the same size as the previous vertically integrated
company; and the new components company would add to the aggregated sales revenues of the economy. A vertical
merger, of course, would reverse this process. 16 This problem of double counting may be a bit less obvious: When a bank lends to a company, that loan is an asset
for the bank. But if the company, in turn, uses the borrowed funds to invest in (say) plant and equipment, the latter
become assets on the company’s balance sheet as well.
8
its franchisees’ sales revenues (instead of just the royalties that they pay to McDonald’s) were
included in the company’s aggregate sales.
In contrast to these two problematic measures, the use of value added – approximately,
the sum of labor costs plus profits – for a company avoids the double-counting problem and thus
allows for a natural aggregation across companies. Value added does provide a standardized
measure across all companies of the contribution of the company to the overall economy.
Value added, of course, does not measure consumer surplus. Accordingly, the value
added that would be recorded for a consumer-facing company such as Facebook may well
understate the measure of its importance if somehow consumer surplus could be included.
Further, for a franchisee such as McDonald’s, its value added (if reported) would be smaller than
if the value added of all of its franchisees (rather than just their royalty payments) were included
the company’s value added.17 But this latter issue is just a “fact of life” of vertical dis-
integration versus vertical integration: Presumably, McDonald’s finds it to be profit-maximizing
to operate its U.S. stores with 90% of them owned by franchisees, rather than trying to own-and-
operate all of them itself.18
Unfortunately, the drawback to the use of value added is that it is not reported
individually by companies. Furthermore, although government data do include value added for
broad industry sectors, those data are useful for our purposes only for manufacturing, where the
concentration (based on value added) of the largest X companies is regularly reported.19
17 Similar considerations would apply to other large franchising companies, such as other large fast-food companies,
large hotel chains, large rental car agencies, etc. 18 Similarly, the value added (if reported) for General Motors would be larger if it produced most of the electrical
components for its cars (as it did through its Delco subsidiary from the 1920s through the 1990s). Presumably,
General Motors found it worthwhile to spin off those operations into the free-standing Delphi Automotive company
in 1999 and instead buy those components from Delphi. 19 The Census of Mining and the Census of Construction (for 2002, 2007, and 2012) report value added by
establishment size categories. However, these data are not reported on an enterprise basis; but it is the enterprise-
level data that are relevant for our purposes.
9
However, one of the two components of value added – company profits – is reported for
publicly traded companies and can thus be aggregated across companies. And the employment
part of the other component of value added – company labor costs – is also reported for publicly
traded companies and can be aggregated across companies. We can thus achieve somewhat
imperfect approximations to the value added measures that would (arguably) be ideal.20
B. What about geography? Large companies that are headquartered in the U.S. often
have extensive non-U.S. operations and employees, and their international involvement has been
growing over time. As the data above for Wal-Mart indicate, in 2015 over a third of its sales
revenues arose outside the U.S., and over a third of its employees were located outside the U.S.;
by comparison, Wal-Mart in 1995 had less than 5% of its employees outside the U.S., and non-
U.S. sales apparently weren’t considered important enough to be reported separately.
At the same time, large companies with headquarters outside the U.S. are increasingly
producing goods and services in the U.S. Sometimes this expansion has occurred organically –
as occurred, for example, in the automobile industry, where Japanese, Korean, and European car
manufacturers have established assembly and sales operations in the U.S. This expansion has
also occurred through mergers and acquisitions by non-U.S. companies that absorb U.S.-
headquartered companies – as has additionally occurred in the auto industry, where Chrysler is
owned by Fiat (an Italian company), and has occurred in the beer industry, where Anheuser-
Busch is owned by AB InBev (a Belgian company).
How should such geographic considerations enter into the measurement of aggregate
concentration? If an aggregate concentration measure is based solely on companies that are
headquartered in the U.S., then it is likely that their worldwide operations will be included in any
20 And, as will be discussed below, the U.S. Bureau of the Census has provided special annual tabulations of labor
costs for the largest X firms, which allows us to get closer to a value added measure.
10
aggregations; and the aggregations will miss the U.S. operations of companies (such as Chrysler
and Anheuser-Busch) that are headquartered abroad.21 Aggregations that are based on the U.S.
operations of all companies that do business in the U.S. will include those foreign-headquartered
companies’ operations but will miss some of the “heft” that attaches to U.S.-headquartered
companies that have substantial operations abroad.
In sum, there are pluses and minuses to various measurement and geographic bases for
the compilation of aggregate concentration measures. And, since the data are sparse, we will
present what is available and offer commentary as we proceed.
V. Data from the U.S. Bureau of the Census.
A. Value added data. As was discussed above, value added data probably represent the
best data for measuring aggregate concentration. Nevertheless, value added is not a metric that
companies regularly report in their individual financial statements; so value added does not
appear in compilations (such as those of Fortune and Forbes) that rely on company financial
statements.
However, for the overall manufacturing sector, the twice-a-decade “Census of
Manufactures” (compiled by the Bureau of the Census) has reported the concentration of value
added by the largest X companies since 1947.22 The time series of the manufacturing sector
value added concentration data – for the largest 50, 100, 150, and 200 companies – are presented
in Table 1 and in Figure 1.23 As can be seen, there was a substantial increase in aggregate
21 This is true of the Fortune and Forbes aggregations that will be discussed below. 22 Unfortunately, the “Census of Manufactures” is the only one of the Census Bureau’s Economic Censuses for
which data of this kind are reported. 23 As can be seen in Table 1 and Figure 1, during the 1960s and 1970s there were a few additional annual
observations for non-census years, which were taken from the Census Bureau’s “Annual Survey of Manufactures”.
11
concentration between 1947 and 1954 across all of the size categories, and a more moderate
subsequent increase into the 1960s. For the next three decades, aggregate concentration in
manufacturing remained largely unchanged. Since the late 1990s, however, aggregate
concentration appears again to have risen moderately (at least through 2012).24
The data for the manufacturing sector need to be kept in context: As a percentage of
private-sector aggregates, value added and employment in manufacturing have been shrinking
since the 1950s. As of 2015, manufacturing accounted for only 13.8% of private-sector GDP
(down from 32.6% in 1953) and 10.3% of private-sector employment (down from 40.3% in
1953). Despite the substantial attention that U.S. manufacturing employment and activity
receive in political discourse and media stories, the manufacturing sector is far from a dominant
position in the U.S. economy.
B. Employment and payroll data. Since 1988 the U.S. Bureau of the Census has been
compiling and publishing employment and payroll data – the “Statistics of U.S. Businesses”
(SUSB) – annually by enterprise size categories for all companies that operate in the U.S.25
These data are currently available through 2014. They are organized into a set of firm-size
employment “buckets”, which start at 0-4 employees and extend to 10,000+ employees. These
data – for the nine firm-size buckets that have been available through all 27 years – are
reproduced in Tables 2-8 and in Figures 2-6.
As can be seen in Table 2 and in Figure 2, the numbers of firms in the largest (10,000+)
category grew between 1988 and around 2000 and has remained relatively stable since then; the
24 The next “Census of Manufactures” is scheduled to be conducted for 2017. The data for that year are unlikely to
be available before 2019 or 2020. 25 These data thus include the U.S. employment and payroll of companies that are headquartered outside the U.S.
More information about these data can be found in White (2002); see also the Census Bureau’s SUSB websites:
https://www.census.gov/programs-surveys/susb/about.html and https://www.census.gov/programs-
surveys/susb/about/glossary.html .
12
percentage figures in Table 3 and in Figure 3 tell the same story. In terms of employment, a
similar story appears in Tables 4 and 5: The aggregate employment by the largest firms grew
until around 2000 and then largely leveled out – with perhaps some growth since 2010; again,
the percentage data (also in Figure 4) show approximately the same pattern.
Table 6 makes use of the data from the previous tables to examine the average
(employee) sizes of firms: For the aggregate of all firms, the average number of employees per
firm grew between 1988 and around 2000 and then remained level. A similar pattern applied to
the largest firms (as is also seen in Figure 5).
Finally, Tables 7 and 8 provide the payroll data for the size categories. The wage data in
Table 7 are not corrected for inflation. The percentage data in Table 8 and in Figure 6 show the
same pattern as in the earlier tables: The share of aggregate payroll by the largest firms grew
between 1988 and around 2000 and has remained roughly constant since then.
An immediate qualifier to the payroll data is important: These data cover employees’
compensation26 but do not include the company’s contributions/payments toward company-
provided fringe benefits, such as healthcare, retirement programs, etc. Thus, these payroll data
will be an underestimate of the total labor costs of companies.27
The data in these tables can’t really help us address the “largest X firms” question.
However, the percentage data in the size buckets in Tables 3, 5, and 8 can be used to compute a
wider measure of concentration: Gini coefficients for firm employment and for firm payrolls.
These Gini coefficients for 1988-2014 are presented in Table 9 and in Figure 7. As can be seen,
26 See https://www.census.gov/programs-surveys/susb/about/glossary.html . 27 Also, the payroll data come from the same source as the Census Bureau’s “County Business Patterns” reports:
payroll data through March of each calendar year, as collected for Social Security/Medicare wage tax purposes.
Starting in 2013, when the Medicare tax was expanded to cover all wage/salary income, these payroll data are likely
to be comprehensive. But prior to 2013, the data would have underestimated the payroll costs for super-high earners
whose first-quarter incomes exceeded the maximum annual income that was subject to Social Security tax. In 1988
that cap was $45,000; in 2011 it was $106,800.
13
the time series for both categories show small but more-or-less steady growth over this 27-year
period.28 For the employment series, the growth from the first year to the last is almost exactly
0.001 per year (from 0.826 in 1988 to 0.854 in 2014); for the payroll series, the annual growth is
slightly larger (from 0.839 to 0.880). The consistently larger Gini coefficient for payrolls (as
compared to employment) implies that the larger firms across the economy tend to pay higher
wages and salaries to their employees than do the smaller firms. The same pattern can be seen in
Tables 5 and 8: For the larger firms, their percentage of total payrolls is larger than their
percentage of total employment.29
To put the magnitude of the Gini coefficients and their changes somewhat in perspective:
Suppose that there are only two size buckets, and that the smallest P% of firms (where P is
expressed as a decimal; i.e., 80% is expressed as 0.80) account for (1-P)% of employment (or
payroll); the remaining (1-P)% of larger firms account for the remaining P% of employment (or
payroll). Then the resulting Gini coefficient would be equal to 2P – 1.30 Thus, the initial Gini
coefficient for employment in 1988 of 0.826 would be yielded by a two-bucket employment
distribution where the largest 8.7% of all firms employed 91.3% of all employees;31 the 2014
Gini coefficient for employment of 0.854 would be yielded by a distribution where the largest
7.3% firms employed 92.7% of all employees. When seen through this lens, the increase in the
Gini coefficient does appear to be more substantial.
28 For some years the Census Bureau provided a larger number of intermediate size buckets for the SUSB data. The
time patterns for the Gini coefficients that are computed from those more finely segmented buckets are consistent
with the patterns that are shown in Table 9. 29 The size category of 500-999 employees appears to be the dividing point between the categories that pay higher
wages and the categories that pay lower wages. 30 This follows readily from the geometry of computing the Gini coefficient from two size buckets: The area under
the Lorenz curve consists of two triangles – each with an area of 0.5 x P x (1 – P) – plus a square with an area of (1
– P)2. The Gini coefficient is equal to 0.5 minus the summed area under the Lorenz curve, all divided by 0.5. 31 This 2-bucket equivalence is somewhat similar in spirit to the “equal-size firms equivalent” measure that Adelman
(1969) provided for the Herfindahl-Hirschman Index.
14
An examination of Tables 5 and 8 shows that the rising Gini coefficients for both series
are driven by rising percentages of employment and payrolls for larger firms generally (and
correspondingly falling percentages for smaller firms), with the size category of 100-499
employees showing relatively unchanged percentages. This is highlighted in Table 10 and in
Figure 8, which shows the annualized percentage changes in the shares of employment and of
payrolls by the various employment size categories.
Finally, a special data tabulation by the Bureau of the Census of these SUSB data has
yielded the “largest X” percentages for employment and for payrolls for the largest 100, 500, and
1,000 companies in the U.S. for the 1988-2014 period. These data are in Table 11 and in Figures
9 and 10. As was reported by White (2002), aggregate concentration fell from 1988 through the
mid 1990s. Since then the pattern has been mixed: Aggregate employment concentration for all
three categories has risen gradually, so that by 2014 the percentages were moderately above the
levels in 1988. For aggregate payroll concentration, however, the percentage for the largest 100
companies has hardly risen, while the percentages for the largest 500 and 1,000 companies have
risen moderately but were still (as of 2014) below the levels for 1988.
One other interesting phenomenon is found in Table 11: Until the late 1990s, the largest
100, 500, and 1,000 companies’ payroll percentages exceeded their employment percentages. As
was true of the large firm categories in Tables 5 and 8, very large firms tended to pay above-
average wages and salaries to their employees. However, since the early 2000s the payroll
percentage for the largest 100 firms has been below their employment share: The largest 100
firms pay payroll wages and salaries that (on average) are lower than is true for companies in
general. And for the largest 500 and 1,000 firms the differential between their payroll
15
percentages and their employment percentages – though consistently positive – had narrowed by
the late 1990s and have stayed at those narrower differences since then.
However, as was discussed above, the payroll data that are collected and published by the
SUSB program do not include employer-provided/funded fringe benefits. The relative decline in
the payroll percentages for these large companies may simply be an indication that they are
compensating their employees more with fringe benefits that are non-taxable to their employees
than are smaller businesses.
VI. Data from Fortune and Forbes.
The “Fortune 500” list of large companies is probably the best-known of such lists.32
Since 1955, Fortune magazine has published every spring a list of 500 large (based on annual
sales revenues) companies that are headquartered in the U.S. The data for the companies are
drawn from those companies’ financial reports for the previous calendar year and always include
the firms’ sales revenues and profits and end-of-year assets and stock market values; sometimes
total employees are included. Because the data are taken from the companies’ consolidated
financial reports, the companies’ consolidated worldwide operations are included.
Initially, the annual lists contained solely large “industrial” companies: i.e.,
manufacturing and mining companies. Services companies – e.g., large financial companies,
utilities, railroads, airlines, etc. – were not included. Starting in the 1960s, the Fortune annual
lists began to include large financial and other services companies, and the data for the 1970s
were sufficiently consistent to allow aggregate concentration estimates for that decade.33
32 For securities market investing, however, the “S&P 500” may be at least as well known. 33 See White (1981; 2002) and Golbe and White (1988).
16
However, in the 1980s and the early 1990s there was insufficient year-to-year consistency in
what categories were included or excluded. Consequently, for that period the Fortune annual
lists are not suitable for our purposes.
From 1993 onward, however, the Fortune annual lists have been sufficiently consistent in
composition to allow year-to-year comparisons. From the Fortune annual data we extract the
employment and profits of the 500 largest companies (as ranked by annual sales)34 and divide by
total private-sector employment and after-tax corporate profits, respectively. The resulting
percentages are shown in Table 12 and in Figures 11 and 12. As can be seen, the annual
employment percentages are relatively stable but show a modest increase from the mid 1990s to
the mid 2010s. The annual profit percentages are much more variable and cyclical, dipping
substantially (as would be expected) during the recession years of 2002 and 2008. Again, there
seems to be a modest increase in the annual percentages from the mid 1990s to the mid 2010s.
For a longer perspective, these Fortune-based data can be combined with annual data for
the 500 largest companies (measured somewhat differently) that were published by Forbes from
1980 through 2000.35 Again, the annual employment and profits data are divided by the relevant
national aggregates. These Forbes-based data are also presented in Table 12 and in Figures 11
and 12.36 As can be seen, the employment percentages for the two data sets for the overlapping
years (1993-2000) are quite consistent;37 they are also consistent with the SUSB-based data on
34 Again, this is based on their worldwide operations and excludes any company that is not headquartered in the U.S. 35 After 2000 the annual Forbes data became unusable for our purposes. Discussion of the 1980-2000 Forbes data
can be found in White (2002). 36 The Forbes percentage data that appear here are somewhat different than those that appeared in Table 8 in White
(2002). For the profit percentage, the difference is due to a different national corporate profits series being used as
the denominator. For the employment percentage, there seems to have been an inexplicable error in the calculations
of the percentages in the last column of Table 8 of White (2002): As compared with the percentages presented here
in Table 11, the levels there seem to be too low by about a quarter; but the year-to-year changes are quite consistent. 37 Because the Forbes annual employment data are drawn from the largest 500 firms ranked by employment,
whereas the Fortune annual employment data are drawn from the largest 500 firms ranked by sales, the Forbes
percentages would be expected to be somewhat larger than the Fortune percentages – which they generally are.
17
employment percentages for the 500 largest companies for these same years that are shown in
Table 10, as can be seen in Figure 13. The profit percentages for the Forbes-based data are
higher than those for Fortune;38 but the year-to-year changes (except for 1993-1994) are roughly
consistent. The Forbes-based data show a decline in employment percentages and profit
percentages from the early 1980s to the mid 1990s.
If we assess the two annual data sets and their overlap, the following seems to be a
reasonable conclusion: Between the early 1980s and the mid 1990s there was a decline in
aggregate concentration as shown by the annual employment percentages and profit percentages
of the 500 largest firms in the U.S. economy. Since the mid 1990s, there appear to have been
modest increases in both percentages; but those percentages do not seem to have regained the
levels of the early 1980s.
VII. Conclusion.
As was discussed at the beginning of this paper, the level of aggregate concentration in
the U.S. economy has little connection to the state of competition in relevant markets and thus to
antitrust policy in the U.S.39 However, it may tell us something about the “tone” of the society in
which we live.
Aggregate concentration in the U.S. economy has been rising moderately in the early 21st
century – but not by as much as the political and media references to large corporations might
38 This is for the same reason as applies to the employment data: The Forbes annual profit data are drawn from the
largest 500 companies ranked by profits, whereas the Fortune annual profit data are drawn from the largest 500
companies ranked by sales. 39 For statements of recent concerns about the state of competition in the U.S. economy, see U.S. Council of
Economic Advisers (2016a, b); American Antitrust Institute (2016); Jarsulic et al. (2016), and Baker (2017).
18
indicate. The data – based on the employment, payrolls, and profits of large companies – are
clear in this regard.
What is less clear (of course) is whether this trend is desirable or not. Our data cannot
provide an answer to that question. But it is surely better to have some facts for help in shaping
any discussion on this topic.
References
Adelman, Morris “Comment on the ‘H’ Concentration Measure as a Numbers-Equivalent,”
Review of Economics & Statistics, 51 (February 1969), pp. 99-101.
American Antitrust Institute, “A National Competition Policy: Unpacking the Problem of
Declining Competition and Setting Priorities Moving Forward,” Washington, DC (2016)
Baker, Jonathan B., “Market Power in the U.S. Economy Today,” Washington Center for
Equitable Growth, Washington, DC (March 2017).
Bernheim, B. Douglas and Michael D. Whinston, Multimarket Conduct and Collusive
Behavior,” Rand Journal of Economics, 21 (Spring 1990), pp. 1-25.
Golbe, Devra L. and Lawrence J. White, “A Time Series Analysis of Mergers and Acquisitions
in the U.S. Economy,” in Alan J. Auerbach, ed., Corporate Takeovers: Causes and
Consequences. Chicago, Chicago University Press, pp. 265-302.
Jarsulic, Marc, Ethan Gurwitz, Kate Bahn, and Andy Green, “Reviving Antitrust: Why Our
Economy Needs a Progressive Competition Policy,” Center for American Progress, Washington,
DC (June 2016).
Kwoka, John E., Jr., and Lawrence J. White, eds., The Antitrust Revolution: Economics,
Competition, and Policy, 6th edn. New York: Oxford University Press, 2014,
Stigler, George J. “Notes on the Theory of Duopoly.” Journal of Political Economy 34 (1940):
521-41.
Stigler, George J. “A Theory of Oligopoly,” Journal of Political Economy, 72 (February 1964), pp.
55-69.
U.S. Council of Economic Advisers, “Benefits of Competition and Indicators of Market Power,”
Washington, DC (2016a).
U.S. Council of Economic Advisers, Labor Market Monopsony: Trends, Consequences, and
Policy Response,” Washington, DC (2016b).
19
White, Lawrence J., “What Has Been Happening to Aggregate Concentration in the United
States?” Journal of Industrial Economics, 29 (March 1981), pp. 223-230.
White, Lawrence J., “Trends in Aggregate Concentration in the United States,” Journal of
Economic Perspectives, 16 (Fall 2002), pp. 137-160.
20
Table 1: Share of Total Value Added in the U.S. Manufacturing Sector, 1947-2012
Year Largest 50
Cos.
Largest 100
Cos.
Largest 150
Cos.
Largest 200
Cos.
1947 17% 23% 27% 30%
1954 23 30 34 37
1958 24 30 35 38
1962 25 32 36 40
1963 25 33 37 41
1966 25 33 38 42
1967 24 33 38 42
1970 25 33 38 43
1972 24 33 39 43
1976 24 33 39 44
1977 25 33 39 44
1982 25 33 39 43
1987 25 33 39 43
1992 24 32 38 42
1997 24 32 37 40
2002 25 34 39 42
2007 26 35 40 44
2012 26 35 40 44
Source: U.S. Bureau of the Census, “Census of Manufactures”, various years.
21
Table 2: Numbers of U.S. Firms, by Employment Size Category of Enterprise, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 4,954,645 2,979,905 923,580 540,988 430,640 66,708 6,455 5,095 631 643
1989 5,021,315 3,003,224 937,202 553,449 443,959 69,608 6,926 5,605 678 664
1990 5,073,795 3,020,935 952,030 562,610 453,732 70,465 6,948 5,666 703 706
1991 5,051,025 3,036,304 941,296 551,299 439,811 68,338 6,842 5,727 709 699
1992 5,095,356 3,075,280 945,802 551,912 439,084 69,156 6,892 5,803 723 704
1993 5,193,642 3,139,518 962,481 559,602 445,900 71,512 7,185 5,995 726 723
1994 5,276,964 3,208,235 964,985 563,097 452,383 73,267 7,415 6,093 726 763
1995 5,369,068 3,249,573 981,094 576,866 469,869 76,222 7,566 6,334 768 776
1996 5,478,047 3,327,783 996,356 585,844 476,312 76,136 7,670 6,329 796 821
1997 5,541,918 3,358,048 1,006,897 593,696 487,491 79,707 7,972 6,464 818 825
1998 5,579,177 3,376,351 1,011,849 600,167 494,357 80,075 8,055 6,568 869 886
1999 5,607,743 3,389,161 1,012,954 605,693 501,848 81,347 8,235 6,698 871 936
2000 5,652,544 3,396,732 1,021,210 617,087 515,977 84,385 8,483 6,824 903 943
2001 5,657,774 3,401,676 1,019,105 616,064 518,258 85,304 8,572 6,931 934 930
2002 5,697,759 3,465,647 1,010,804 613,880 508,249 82,334 8,326 6,722 884 913
2003 5,767,127 3,504,432 1,025,497 620,387 515,056 84,829 8,408 6,725 910 883
2004 5,885,784 3,579,714 1,043,448 632,682 526,355 86,538 8,404 6,848 905 890
2005 5,983,546 3,677,879 1,050,062 629,946 520,897 87,285 8,701 6,946 918 912
2006 6,022,127 3,670,028 1,060,787 646,816 535,865 90,560 8,974 7,213 931 953
2007 6,049,655 3,705,275 1,060,250 644,842 532,391 88,586 9,064 7,320 952 975
2008 5,930,132 3,617,764 1,044,065 633,141 526,307 90,386 9,098 7,415 975 981
2009 5,767,306 3,558,708 1,001,313 610,777 495,673 83,326 8,631 6,991 956 931
2010 5,734,538 3,575,240 968,075 617,089 475,125 81,773 8,489 6,923 906 918
2011 5,684,424 3,532,058 978,993 592,963 481,496 81,243 8,761 7,071 908 931
2012 5,726,160 3,543,991 992,716 593,641 494,170 83,423 9,061 7,249 945 964
2013 5,775,055 3,575,290 992,281 600,551 503,033 85,264 9,340 7,342 980 974
2014 5,825,458 3,598,185 998,953 608,502 513,179 87,563 9,504 7,586 992 994
Source: U.S. Bureau of the Census, “Statistics of U.S. Businesses”.
22
Table 3: Percentage Distribution of U.S. Firms, by Employment Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 100.0% 60.14% 18.64% 10.92% 8.69% 1.35% 0.130% 0.103% 0.013% 0.013%
1989 100.0% 59.81% 18.66% 11.02% 8.84% 1.39% 0.138% 0.112% 0.014% 0.013%
1990 100.0% 59.54% 18.76% 11.09% 8.94% 1.39% 0.137% 0.112% 0.014% 0.014%
1991 100.0% 60.11% 18.64% 10.91% 8.71% 1.35% 0.135% 0.113% 0.014% 0.014%
1992 100.0% 60.35% 18.56% 10.83% 8.62% 1.36% 0.135% 0.114% 0.014% 0.014%
1993 100.0% 60.45% 18.53% 10.77% 8.59% 1.38% 0.138% 0.115% 0.014% 0.014%
1994 100.0% 60.80% 18.29% 10.67% 8.57% 1.39% 0.141% 0.115% 0.014% 0.014%
1995 100.0% 60.52% 18.27% 10.74% 8.75% 1.42% 0.141% 0.118% 0.014% 0.014%
1996 100.0% 60.75% 18.19% 10.69% 8.69% 1.39% 0.140% 0.116% 0.015% 0.015%
1997 100.0% 60.59% 18.17% 10.71% 8.80% 1.44% 0.144% 0.117% 0.015% 0.015%
1998 100.0% 60.52% 18.14% 10.76% 8.86% 1.44% 0.144% 0.118% 0.016% 0.016%
1999 100.0% 60.44% 18.06% 10.80% 8.95% 1.45% 0.147% 0.119% 0.016% 0.017%
2000 100.0% 60.09% 18.07% 10.92% 9.13% 1.49% 0.150% 0.121% 0.016% 0.017%
2001 100.0% 60.12% 18.01% 10.89% 9.16% 1.51% 0.152% 0.123% 0.017% 0.016%
2002 100.0% 60.82% 17.74% 10.77% 8.92% 1.45% 0.146% 0.118% 0.016% 0.016%
2003 100.0% 60.77% 17.78% 10.76% 8.93% 1.47% 0.146% 0.117% 0.016% 0.015%
2004 100.0% 60.82% 17.73% 10.75% 8.94% 1.47% 0.143% 0.116% 0.015% 0.015%
2005 100.0% 61.47% 17.55% 10.53% 8.71% 1.46% 0.145% 0.116% 0.015% 0.015%
2006 100.0% 60.94% 17.61% 10.74% 8.90% 1.50% 0.149% 0.120% 0.015% 0.016%
2007 100.0% 61.25% 17.53% 10.66% 8.80% 1.46% 0.150% 0.121% 0.016% 0.016%
2008 100.0% 61.01% 17.61% 10.68% 8.88% 1.52% 0.153% 0.125% 0.016% 0.017%
2009 100.0% 61.70% 17.36% 10.59% 8.59% 1.44% 0.150% 0.121% 0.017% 0.016%
2010 100.0% 62.35% 16.88% 10.76% 8.29% 1.43% 0.148% 0.121% 0.016% 0.016%
2011 100.0% 62.14% 17.22% 10.43% 8.47% 1.43% 0.154% 0.124% 0.016% 0.016%
2012 100.0% 61.89% 17.34% 10.37% 8.63% 1.46% 0.158% 0.127% 0.017% 0.017%
2013 100.0% 61.91% 17.18% 10.40% 8.71% 1.48% 0.162% 0.127% 0.017% 0.017%
2014 100.0% 61.77% 17.15% 10.45% 8.81% 1.50% 0.163% 0.130% 0.017% 0.017%
Source: Table 2.
23
Table 4: Total Number of Employees in Each Employment Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 87,844,303 5,006,203 6,060,724 7,252,715 16,833,702 12,761,379 4,453,706 10,146,017 4,418,568 20,915,729
1989 91,626,094 5,054,429 6,152,151 7,420,196 17,353,444 13,373,640 4,782,882 11,141,733 4,746,232 21,605,433
1990 93,469,275 5,116,914 6,251,632 7,543,360 17,710,042 13,544,849 4,804,321 11,244,354 4,869,749 22,385,891
1991 92,307,559 5,151,143 6,174,730 7,386,939 17,146,411 13,143,390 4,716,916 11,409,214 4,929,224 22,255,352
1992 92,825,797 5,178,909 6,202,861 7,390,874 17,121,010 13,307,187 4,743,398 11,640,355 5,031,158 22,213,213
1993 94,773,913 5,258,195 6,313,651 7,498,345 17,420,634 13,825,238 4,956,676 12,083,674 5,051,450 22,366,643
1994 96,721,594 5,318,961 6,332,580 7,543,777 17,693,995 14,118,375 5,126,244 12,361,020 5,019,851 23,213,183
1995 100,314,946 5,395,432 6,440,349 7,734,080 18,422,228 14,660,421 5,226,409 12,890,471 5,366,850 24,175,902
1996 102,187,297 5,485,712 6,541,288 7,854,502 18,643,192 14,649,808 5,293,302 12,783,631 5,507,895 25,424,199
1997 105,299,123 5,546,306 6,610,374 7,962,136 19,109,691 15,316,863 5,496,614 13,088,681 5,696,683 26,482,729
1998 108,117,731 5,584,470 6,643,285 8,047,650 19,377,614 15,411,390 5,547,037 13,233,763 6,086,847 28,185,675
1999 110,705,661 5,606,302 6,652,370 8,129,615 19,703,162 15,637,643 5,662,057 13,533,807 6,064,760 29,715,945
2000 114,064,976 5,592,980 6,708,674 8,285,731 20,276,634 16,260,025 5,855,270 13,833,418 6,299,143 30,953,101
2001 115,061,184 5,630,017 6,698,077 8,274,541 20,370,447 16,410,367 5,906,266 13,957,822 6,456,068 31,357,579
2002 112,400,654 5,697,652 6,639,666 8,246,053 19,874,069 15,908,852 5,734,715 13,634,767 6,131,966 30,532,914
2003 113,398,043 5,768,407 6,732,132 8,329,813 20,186,989 16,430,229 5,787,182 13,684,578 6,389,355 30,089,358
2004 115,074,924 5,844,637 6,852,769 8,499,681 20,642,614 16,757,751 5,781,342 13,879,053 6,378,292 30,438,785
2005 116,317,003 5,936,859 6,898,483 8,453,854 20,444,349 16,911,040 6,018,347 14,091,935 6,438,639 31,123,497
2006 119,917,165 5,959,585 6,973,537 8,676,398 21,076,875 17,537,345 6,197,683 14,622,542 6,490,547 32,382,653
2007 120,604,265 6,139,463 6,974,591 8,656,182 20,922,960 17,173,728 6,257,654 14,998,669 6,628,415 32,852,603
2008 120,903,551 6,086,291 6,878,051 8,497,391 20,684,691 17,547,567 6,298,847 15,111,901 6,773,466 33,025,346
2009 114,509,626 5,966,190 6,580,830 8,191,289 19,389,940 16,153,254 5,963,102 14,235,941 6,594,104 31,434,976
2010 111,970,095 5,926,452 6,358,931 8,288,385 18,554,372 15,868,540 5,872,738 14,101,170 6,286,593 30,712,914
2011 113,425,965 5,857,662 6,431,931 7,961,281 18,880,001 15,867,437 6,060,684 14,451,424 6,365,032 31,550,513
2012 115,938,468 5,906,506 6,527,943 7,974,340 19,387,249 16,266,855 6,266,639 14,730,462 6,543,543 32,334,931
2013 118,266,253 5,926,660 6,523,516 8,058,077 19,697,707 16,617,417 6,436,233 14,944,970 6,767,037 33,294,636
2014 121,069,944 5,940,248 6,570,776 8,176,519 20,121,588 17,085,461 6,544,625 15,502,104 6,870,983 34,257,640
Source: U.S. Bureau of the Census, “Statistics of U.S. Businesses”.
24
Table 5: Percentage Distribution of Employees, by Enterprise Employment Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 100.00% 5.70% 6.90% 8.26% 19.16% 14.53% 5.07% 11.55% 5.03% 23.81%
1989 100.00% 5.52% 6.71% 8.10% 18.94% 14.60% 5.22% 12.16% 5.18% 23.58%
1990 100.00% 5.47% 6.69% 8.07% 18.95% 14.49% 5.14% 12.03% 5.21% 23.95%
1991 100.00% 5.58% 6.69% 8.00% 18.58% 14.24% 5.11% 12.36% 5.34% 24.11%
1992 100.00% 5.58% 6.68% 7.96% 18.44% 14.34% 5.11% 12.54% 5.42% 23.93%
1993 100.00% 5.55% 6.66% 7.91% 18.38% 14.59% 5.23% 12.75% 5.33% 23.60%
1994 100.00% 5.50% 6.55% 7.80% 18.29% 14.60% 5.30% 12.78% 5.19% 24.00%
1995 100.00% 5.38% 6.42% 7.71% 18.36% 14.61% 5.21% 12.85% 5.35% 24.10%
1996 100.00% 5.37% 6.40% 7.69% 18.24% 14.34% 5.18% 12.51% 5.39% 24.88%
1997 100.00% 5.27% 6.28% 7.56% 18.15% 14.55% 5.22% 12.43% 5.41% 25.15%
1998 100.00% 5.17% 6.14% 7.44% 17.92% 14.25% 5.13% 12.24% 5.63% 26.07%
1999 100.00% 5.06% 6.01% 7.34% 17.80% 14.13% 5.11% 12.23% 5.48% 26.84%
2000 100.00% 4.90% 5.88% 7.26% 17.78% 14.26% 5.13% 12.13% 5.52% 27.14%
2001 100.00% 4.89% 5.82% 7.19% 17.70% 14.26% 5.13% 12.13% 5.61% 27.25%
2002 100.00% 5.07% 5.91% 7.34% 17.68% 14.15% 5.10% 12.13% 5.46% 27.16%
2003 100.00% 5.09% 5.94% 7.35% 17.80% 14.49% 5.10% 12.07% 5.63% 26.53%
2004 100.00% 5.08% 5.96% 7.39% 17.94% 14.56% 5.02% 12.06% 5.54% 26.45%
2005 100.00% 5.10% 5.93% 7.27% 17.58% 14.54% 5.17% 12.12% 5.54% 26.76%
2006 100.00% 4.97% 5.82% 7.24% 17.58% 14.62% 5.17% 12.19% 5.41% 27.00%
2007 100.00% 5.09% 5.78% 7.18% 17.35% 14.24% 5.19% 12.44% 5.50% 27.24%
2008 100.00% 5.03% 5.69% 7.03% 17.11% 14.51% 5.21% 12.50% 5.60% 27.32%
2009 100.00% 5.21% 5.75% 7.15% 16.93% 14.11% 5.21% 12.43% 5.76% 27.45%
2010 100.00% 5.29% 5.68% 7.40% 16.57% 14.17% 5.24% 12.59% 5.61% 27.43%
2011 100.00% 5.16% 5.67% 7.02% 16.65% 13.99% 5.34% 12.74% 5.61% 27.82%
2012 100.00% 5.09% 5.63% 6.88% 16.72% 14.03% 5.41% 12.71% 5.64% 27.89%
2013 100.00% 5.01% 5.52% 6.81% 16.66% 14.05% 5.44% 12.64% 5.72% 28.15%
2014 100.00% 4.91% 5.43% 6.75% 16.62% 14.11% 5.41% 12.80% 5.68% 28.30%
Source: Table 4.
25
Table 6: Average Employee Size of Enterprise, by Enterprise Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 17.7 1.7 6.6 13.4 39.1 191.3 690.0 1,991.4 7,002.5 32,528.3
1989 18.2 1.7 6.6 13.4 39.1 192.1 690.6 1,987.8 7,000.3 32,538.3
1990 18.4 1.7 6.6 13.4 39.0 192.2 691.5 1,984.5 6,927.1 31,708.1
1991 18.3 1.7 6.6 13.4 39.0 192.3 689.4 1,992.2 6,952.4 31,838.8
1992 18.2 1.7 6.6 13.4 39.0 192.4 688.2 2,005.9 6,958.7 31,552.9
1993 18.2 1.7 6.6 13.4 39.1 193.3 689.9 2,015.6 6,957.9 30,935.9
1994 18.3 1.7 6.6 13.4 39.1 192.7 691.3 2,028.7 6,914.4 30,423.6
1995 18.7 1.7 6.6 13.4 39.2 192.3 690.8 2,035.1 6,988.1 31,154.5
1996 18.7 1.6 6.6 13.4 39.1 192.4 690.1 2,019.9 6,919.5 30,967.4
1997 19.0 1.7 6.6 13.4 39.2 192.2 689.5 2,024.9 6,964.2 32,100.3
1998 19.4 1.7 6.6 13.4 39.2 192.5 688.6 2,014.9 7,004.4 31,812.3
1999 19.7 1.7 6.6 13.4 39.3 192.2 687.6 2,020.6 6,963.0 31,747.8
2000 20.2 1.6 6.6 13.4 39.3 192.7 690.2 2,027.2 6,975.8 32,824.1
2001 20.3 1.7 6.6 13.4 39.3 192.4 689.0 2,013.8 6,912.3 33,717.8
2002 19.7 1.6 6.6 13.4 39.1 193.2 688.8 2,028.4 6,936.6 33,442.4
2003 19.7 1.6 6.6 13.4 39.2 193.7 688.3 2,034.9 7,021.3 34,076.3
2004 19.6 1.6 6.6 13.4 39.2 193.6 687.9 2,026.7 7,047.8 34,200.9
2005 19.4 1.6 6.6 13.4 39.2 193.7 691.7 2,028.8 7,013.8 34,126.6
2006 19.9 1.6 6.6 13.4 39.3 193.7 690.6 2,027.2 6,971.6 33,979.7
2007 19.9 1.7 6.6 13.4 39.3 193.9 690.4 2,049.0 6,962.6 33,695.0
2008 20.4 1.7 6.6 13.4 39.3 194.1 692.3 2,038.0 6,947.1 33,665.0
2009 19.9 1.7 6.6 13.4 39.1 193.9 690.9 2,036.3 6,897.6 33,764.7
2010 19.5 1.7 6.6 13.4 39.1 194.1 691.8 2,036.9 6,938.8 33,456.3
2011 20.0 1.7 6.6 13.4 39.2 195.3 691.8 2,043.8 7,009.9 33,888.8
2012 20.2 1.7 6.6 13.4 39.2 195.0 691.6 2,032.1 6,924.4 33,542.5
2013 20.5 1.7 6.6 13.4 39.2 194.9 689.1 2,035.5 6,905.1 34,183.4
2014 20.8 1.7 6.6 13.4 39.2 195.1 688.6 2,043.5 6,926.4 34,464.4
Source: Tables 2 and 4
26
Table 7: Aggregate Annual Payroll (in $ million) by Employment Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 $1,858,652 $108,801 $103,041 $130,326 $315,751 $244,647 $89,773 $226,384 $109,289 $530,645
1989 1,989,942 112,462 108,003 136,795 332,733 264,144 99,298 258,095 119,396 559,174
1990 2,103,971 116,857 114,006 144,451 352,391 279,452 103,095 267,415 128,763 597,528
1991 2,145,016 118,234 116,794 146,517 352,033 279,437 107,036 281,641 134,707 608,541
1992 2,272,392 124,592 122,382 152,831 368,969 298,174 114,074 308,591 146,115 636,497
1993 2,363,208 128,968 127,133 159,153 385,005 316,184 121,233 325,414 152,900 647,046
1994 2,487,960 134,649 131,667 166,476 408,053 335,574 128,876 344,831 155,000 682,945
1995 2,665,922 141,538 137,083 175,388 437,065 361,061 136,762 372,696 171,952 732,329
1996 2,848,623 150,825 144,692 185,491 465,230 384,020 146,134 388,552 186,015 797,899
1997 3,047,907 158,448 150,877 193,805 494,617 418,453 156,358 416,344 203,600 855,243
1998 3,309,406 168,433 159,689 207,063 531,231 446,353 166,642 438,498 222,289 969,207
1999 3,554,693 177,378 166,599 217,571 564,975 474,607 181,766 465,790 225,757 1,080,252
2000 3,879,430 186,176 174,384 230,564 608,446 527,545 201,266 506,746 236,381 1,207,923
2001 3,989,086 187,982 178,881 236,986 624,313 539,385 203,985 514,614 255,249 1,247,692
2002 3,943,180 193,789 182,384 241,411 623,716 535,750 200,639 518,467 242,128 1,204,895
2003 4,040,889 197,241 187,419 246,562 635,269 552,003 206,163 540,512 269,455 1,206,265
2004 4,253,996 205,948 195,519 257,803 670,418 587,676 217,589 574,527 278,397 1,266,118
2005 4,482,722 220,009 206,178 269,417 700,453 616,524 230,798 600,925 297,594 1,340,824
2006 4,792,430 229,730 214,137 282,193 741,917 660,816 247,552 652,602 302,709 1,460,774
2007 5,026,778 234,921 222,420 292,088 768,547 686,862 261,892 689,206 324,791 1,546,051
2008 5,130,509 232,063 222,505 293,534 774,589 706,477 271,128 713,055 337,598 1,579,560
2009 4,855,545 219,913 212,719 278,321 719,054 654,812 257,741 675,635 337,026 1,500,324
2010 4,940,983 226,541 212,040 283,246 719,061 665,645 265,832 697,904 337,070 1,533,644
2011 5,164,898 230,422 218,086 284,252 746,085 690,510 280,490 745,523 340,581 1,628,950
2012 5,414,256 237,897 224,438 290,991 783,572 730,638 296,705 779,674 372,593 1,697,748
2013 5,621,697 241,348 228,080 297,246 799,075 752,414 308,895 806,086 397,390 1,791,163
2014 5,940,187 251,757 235,547 309,924 838,406 803,653 324,548 860,728 418,466 1,897,158
Source: U.S. Bureau of the Census, “Statistics of U.S. Businesses”.
27
Table 8: Percentage Distribution of Aggregate Annual Payroll by Employment Size Category, 1988-2014
Year Total
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
1988 100.00% 5.85% 5.54% 7.01% 16.99% 13.16% 4.83% 12.18% 5.88% 28.55%
1989 100.00% 5.65% 5.43% 6.87% 16.72% 13.27% 4.99% 12.97% 6.00% 28.10%
1990 100.00% 5.55% 5.42% 6.87% 16.75% 13.28% 4.90% 12.71% 6.12% 28.40%
1991 100.00% 5.51% 5.44% 6.83% 16.41% 13.03% 4.99% 13.13% 6.28% 28.37%
1992 100.00% 5.48% 5.39% 6.73% 16.24% 13.12% 5.02% 13.58% 6.43% 28.01%
1993 100.00% 5.46% 5.38% 6.73% 16.29% 13.38% 5.13% 13.77% 6.47% 27.38%
1994 100.00% 5.41% 5.29% 6.69% 16.40% 13.49% 5.18% 13.86% 6.23% 27.45%
1995 100.00% 5.31% 5.14% 6.58% 16.39% 13.54% 5.13% 13.98% 6.45% 27.47%
1996 100.00% 5.29% 5.08% 6.51% 16.33% 13.48% 5.13% 13.64% 6.53% 28.01%
1997 100.00% 5.20% 4.95% 6.36% 16.23% 13.73% 5.13% 13.66% 6.68% 28.06%
1998 100.00% 5.09% 4.83% 6.26% 16.05% 13.49% 5.04% 13.25% 6.72% 29.29%
1999 100.00% 4.99% 4.69% 6.12% 15.89% 13.35% 5.11% 13.10% 6.35% 30.39%
2000 100.00% 4.80% 4.50% 5.94% 15.68% 13.60% 5.19% 13.06% 6.09% 31.14%
2001 100.00% 4.71% 4.48% 5.94% 15.65% 13.52% 5.11% 12.90% 6.40% 31.28%
2002 100.00% 4.91% 4.63% 6.12% 15.82% 13.59% 5.09% 13.15% 6.14% 30.56%
2003 100.00% 4.88% 4.64% 6.10% 15.72% 13.66% 5.10% 13.38% 6.67% 29.85%
2004 100.00% 4.84% 4.60% 6.06% 15.76% 13.81% 5.11% 13.51% 6.54% 29.76%
2005 100.00% 4.91% 4.60% 6.01% 15.63% 13.75% 5.15% 13.41% 6.64% 29.91%
2006 100.00% 4.79% 4.47% 5.89% 15.48% 13.79% 5.17% 13.62% 6.32% 30.48%
2007 100.00% 4.67% 4.42% 5.81% 15.29% 13.66% 5.21% 13.71% 6.46% 30.76%
2008 100.00% 4.52% 4.34% 5.72% 15.10% 13.77% 5.28% 13.90% 6.58% 30.79%
2009 100.00% 4.53% 4.38% 5.73% 14.81% 13.49% 5.31% 13.91% 6.94% 30.90%
2010 100.00% 4.58% 4.29% 5.73% 14.55% 13.47% 5.38% 14.12% 6.82% 31.04%
2011 100.00% 4.46% 4.22% 5.50% 14.45% 13.37% 5.43% 14.43% 6.59% 31.54%
2012 100.00% 4.39% 4.15% 5.37% 14.47% 13.49% 5.48% 14.40% 6.88% 31.36%
2013 100.00% 4.29% 4.06% 5.29% 14.21% 13.38% 5.49% 14.34% 7.07% 31.86%
2014 100.00% 4.24% 3.97% 5.22% 14.11% 13.53% 5.46% 14.49% 7.04% 31.94%
Source: Table 7.
28
Table 9: Gini Coefficients for Enterprise Employment and Payrolls, 1988-2014
Year
Gini coefficient based
on employment
Gini coefficient based
on payrolls
1988 0.826 0.839
1989 0.829 0.842
1990 0.829 0.843
1991 0.830 0.845
1992 0.831 0.847
1993 0.832 0.848
1994 0.835 0.849
1995 0.836 0.851
1996 0.837 0.852
1997 0.839 0.854
1998 0.841 0.857
1999 0.843 0.859
2000 0.845 0.862
2001 0.846 0.863
2002 0.845 0.861
2003 0.844 0.861
2004 0.844 0.862
2005 0.846 0.863
2006 0.847 0.865
2007 0.847 0.868
2008 0.848 0.870
2009 0.848 0.872
2010 0.848 0.874
2011 0.851 0.876
2012 0.851 0.877
2013 0.853 0.879
2014 0.854 0.880
Source: Tables 3, 5, and 8.
29
Table 10: Annualized Percentage Changes in Employment and Payroll by Enterprise Employment Size, 1988-2014
0-4
Empl.
5-9
Empl.
10-19
Empl.
20-99
Empl.
100-499
Empl.
500-999
Empl.
1,000-
4,999
Empl.
5,000-
9,999
Empl.
10,000+
Empl.
Employment -0.57% -0.92% -0.77% -0.55% -0.11% +0.25% +0.40% +0.47% +0.67%
Payrolls -1.23% -1.28% -1.13% -0.71% +0.11% +0.48% +0.67% +0.70% +0.43%
Source: Tables 5 and 8.
30
Table 11: Aggregate Concentration in the Private Sector of the U.S. Economy, as Measured by
Employment and by Payroll, 1988-2014
Share of private-sector employment
by largest
Share of private-sector payroll by
largest
Year
100
companies
500
companies
1,000
companies
100
companies
500
companies
1,000
companies
1988 11.3% 21.9% 27.1% 13.8% 26.3% 32.4%
1989 11.1 21.5 26.6 13.6 25.6 31.2
1990 11.0 21.4 26.6 13.2 25.3 31.5
1991 11.1 21.6 26.8 13.2 25.4 31.6
1992 10.9 21.3 26.6 12.9 24.9 31.3
1993 10.7 20.8 26.1 12.0 24.1 30.5
1994 10.6 20.7 26.1 11.8 23.6 29.9
1995 10.5 20.7 26.1 11.7 23.6 29.8
1996 10.6 20.9 26.5 11.2 23.3 30.0
1997 10.9 21.2 26.7 11.4 23.4 30.0
1998 10.8 21.5 27.1 10.9 23.4 30.5
1999 11.2 21.8 27.4 11.5 23.9 31.1
2000 11.4 22.0 27.6 11.6 25.0 31.7
2001 11.5 22.3 27.8 12.0 25.3 32.0
2002 11.6 22.3 27.9 11.8 24.8 31.4
2003 11.6 22.1 27.5 11.7 24.7 31.1
2004 11.3 22.0 27.4 11.1 24.6 30.9
2005 11.5 22.1 27.5 11.2 24.3 30.7
2006 11.5 21.9 27.4 11.3 24.2 30.9
2007 11.5 21.9 27.4 11.1 24.4 31.0
2008 11.5 21.9 27.5 11.1 24.2 31.0
2009 11.8 22.4 28.0 11.6 24.8 31.6
2010 11.8 22.4 28.1 11.6 25.1 31.9
2011 11.9 22.7 28.4 11.6 25.3 32.2
2012 11.9 22.5 28.2 11.3 24.9 31.7
2013 12.0 22.6 28.4 11.7 25.3 32.2
2014 11.9 22.6 28.3 11.5 25.2 32.0
Source: U.S. Bureau of the Census, “Statistics of U.S. Businesses”, special tabulations.
31
Table 12: Aggregate Concentration in the U.S. Economy, as Measured by Employment and by
Profits, Based on Fortune and Forbes Data, 1980-2015
Fortune 500 Forbes 500
Year
Share of private-
sector employment
Share of
corporate profits
Share of private-
sector employment
Share of
corporate profits
1980 30% 61%
1981 30% 64%
1982 29% 68%
1983 29% 72%
1984 27% 70%
1985 26% 70%
1986 25% 90%
1987 24% 78%
1988 23% 68%
1989 23% 75%
1990 23% 66%
1991 23% 53%
1992 23% 57%
1993 21% 42% 22% 61%
1994 21% 53% 21% 61%
1995 21% 52% 21% 61%
1996 20% 59% 21% 64%
1997 21% 59% 21% 65%
1998 21% 70% 22% 76%
1999 21% 81% 22% 89%
2000 22% 92% 22% 103%
2001 22% 42%
2002 22% 12%
2003 22% 61%
2004 22% 54%
2005 22% 49%
2006 22% 57%
2007 22% 50%
2008 22% 9%
2009 23% 32%
2010 23% 48%
2011 23% 58%
2012 24% 49%
2013 23% 64%
2014 23% 55%
2015 23% 53%
Source: See text.
32
15%
20%
25%
30%
35%
40%
45%
1947 1954 1958 1962 1963 1966 1967 1970 1972 1976 1977 1982 1987 1992 1997 2002 2007 2012
Figure 1: Share of Total Value Added in the U.S. Manufacturing Sector, 1947 - 2012
Largest 50 Cos. Largest 100 Cos. Largest 150 Cos. Largest 200 Cos.
33
643
994
600
650
700
750
800
850
900
950
1000
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 2: Numbers of U.S. Firms, 10,000+ Employees, 1988 - 2014
34
0.013%
0.017%
0.012%
0.013%
0.014%
0.015%
0.016%
0.017%
0.018%
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 3: Percentage Distribution of U.S. Firms, 10,000+ Employees, 1988 - 2014
35
23.81%
28.30%
23%
24%
25%
26%
27%
28%
29%
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 4: Percentage Distribution of Employees, 10,000+ Employees, 1988 - 2014
36
32,528.3
30,423.6
34,464.4
30,000
30,500
31,000
31,500
32,000
32,500
33,000
33,500
34,000
34,500
35,000
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 5: Average Employee Size of Enterprise, 10,000+ Employees, 1988 - 2014
37
28.55%
31.94%
27%
28%
29%
30%
31%
32%
33%
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 6: Percentage Distribution of Total Annual Payroll, 10,000+ Employees 1988 - 2014
38
0.826
0.854 0.839
0.880
0.820
0.830
0.840
0.850
0.860
0.870
0.880
0.890
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 7: Gini Coefficients for Enterprise Employment and Payrolls, 1988 - 2014
Gini (for Employment) Gini (for Payrolls)
39
- 0.57%
0.67%
- 1.23%
0.43%
-1.50%
-1.00%
-0.50%
0.00%
0.50%
1.00%
0-4 Empl.
5-9 Empl.
10-19 Empl.
20-99 Empl.
100-499 Empl.
500-999 Empl.
1,000-4,999 Empl.
5,000-9,999 Empl.
10,000+ Empl.
Figure 8: Annualized Percentage Changes in Employment and Payroll by Enterprise Employment Size, 1988-2014
Employment Payroll
40
11.30% 11.90%
21.90% 22.60%
27.10%
28.30%
10%
12%
14%
16%
18%
20%
22%
24%
26%
28%
30%
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 9: Aggregate Concentration in the Private Sector, as Measured by Employment, 1988-2014
Largest 100 co. Largest 500 co. Largest 1000 co.
41
13.80%
11.50%
26.30% 25.20%
32.40% 32.00%
10%
15%
20%
25%
30%
35%
1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
Figure 10: Aggregate Concentration in the Private Sector, as Measured by Payroll,
1988
- 2014
Largest 100 co. Largest 500 co. Largest 1000 co.
42
19%
21%
23%
25%
27%
29%
31%
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 11: Aggregate Concentration as Measured by Employment Share, based on Forbes and Fortune 500 Data, 1980-2015
-
Fortune 500 Share of Total Employment Forbes 500 Share of Total Employment
43
0%
20%
40%
60%
80%
100%
120%
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 12: Aggregate Concentration as Measured by Profits Share, Based on Forbes and Fortune 500 Data, 1980-2015
Fortune 500 Share of Total Corporate Profits Forbes 500 Share of Total Corporate Profits
44
19%
21%
23%
25%
27%
29%
31%
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
Figure 13: Aggregate Concentration as Measured by Employment Share, Based on Forbes, Fortune, and SUSB Data, 1980-2015
Fortune 500 Share of Total Employment Forbes 500 Share of Total Employment SUSB Largest 500 Share of Total Employment