inventing social capital: evidence from african american...
Post on 19-Jun-2020
4 Views
Preview:
TRANSCRIPT
Inventing Social Capital: Evidence from African American Inventors, 1843-1930
Lisa D. Cook, Ph.D. Department of Economics 110 Marshall-Adams Hall
Michigan State University
East Lansing, MI 48824
lisacook@msu.edu
© May 2011 Lisa D. Cook
I am grateful to Jeff Biddle, Barry Eichengreen, Art Goldsmith, Thomas Jeitschko, Naomi Lamoreaux, Josh Lerner, Ken Sokoloff, Michèle Tertilt, and two anonymous referees for helpful comments on this and earlier versions of the paper. Conversations with Ken Arrow, Gary Becker, Bill Collins, Milton Friedman, Claudia Goldin, Avner Greif, Bob Margo, Douglass North, Patricia Carter Sluby, and seminar participants at the University of California, Berkeley, Davis, and Riverside; at Harvard University; at the University of Michigan; at NBER; and at Stanford University were helpful. I am also grateful to Priyanka Bakaya, Chaleampong Kongcharoen, Maksym Ivanya, and Christopher Tan for able research assistance and to reference librarians and staff at the Carter G. Woodson Collection at the Library of Congress and the Moorland-Spingarn Research Center at Howard University for their expert assistance. Most of this research was conducted while at Stanford University (Hoover Institution) and its generosity is acknowledged. All mistakes are my own.
Inventing – Cook May 2001
2
Abstract
Much recent work has focused on the influence of social capital on innovative outcomes. Little research has been done on disadvantaged groups who were often restricted from participation in social networks that provide information necessary for invention and innovation. Unique new data on African American inventors and patentees between 1843 and 1930 permit an empirical investigation of the relation between social capital and economic outcomes. I find that African Americans used both traditional, i.e., occupation-based, and nontraditional, i.e., civic, networks to maximize inventive output and that laws constraining social-capital formation are most negatively correlated with economically important inventive activity.
Inventing – Cook May 2001
3
Recent research using historical and contemporary data have begun to explore the social aspects of
innovation. Thomson (2004, 2009) finds that innovation during the antebellum period hinged
critically on knowledge that was socially organized, i.e., that was shared through social networks
within and across industries, as well as through civil organizations engaged in pure and applied
science. Khan and Sokoloff (2004) cite the importance of the social institution of apprenticeships as
a critical mechanism for transmitting and accumulating knowledge that accelerated the pace of
inventive and innovative activity during Second Industrial Revolution. In reference to the Second
Industrial Revolution, Neal and Davis (2007) suggest that “to realize the commercial prospects of
electricity, telephony, … a firm needs complementary inputs from a social infrastructure.”1
Lamoreaux, Levenstein, and Sokoloff (2007) show that “hub” enterprises, such as Brush Electric
Company in Cleveland, were important loci of inventor networks, including as disseminators of
technical knowledge, attractors of inventive talent, business incubators, and screeners of
opportunities for venture capital investment. The relation between ethnic groups and innovative
activity has been a particular interest among those using data from the late 20th and early 21st
centuries. Recently, much scholarship has focused on social networks derived from ethnicity and
national origin, including patenting and international knowledge flows, e.g., Trajtenberg (2001),
Agarwal, Kapur, and McHale (2007), and Kerr (2008), and entrepreneurial activities associated with
innovation, e.g., Castillo, Hwang, Granovetter, and Granovetter (2000) and Saxenian (2000). While
social capital has been studied among disadvantaged groups, e.g., Munshi (2003) and O‟Regan
(1993), it has largely related to contemporary labor-market outcomes. The contribution of this
paper is to fill two important holes in the literature. African American inventors and patentees have
not been studied as a group systematically. A new data set I have constructed allows this for the
first time. Further, since African Americans were often restricted from participation in many social
1 Neal and Davis cited in Lamoreaux and Sokoloff (2007), p. 132.
Inventing – Cook May 2001
4
networks that provided information necessary for innovation, e.g., apprenticeships and venture-
capital networks, the data allow us to examine social capital among these inventors. Did African
American inventors possess traditional social capital relevant for inventive activity historically? If so,
what kind and how was it used? If not, were there other forms of social capital that were relevant
for invention and innovation?
I find that even the most prolific African American inventors possessed relatively little traditional
social capital in the late 19th and early 20th centuries and used broader social networks to access the
information needed for invention and innovation. Further, I find that proliferation of segregation
laws, which would disrupt information flows, depressed patent outcomes, particularly among those
of economic importance.
II. Data on African American Innovation and Social Capital: Collection and Summary Statistics I have collected unique new data on African American inventors between 1843 and 1930 to test the
hypothesized relation between social capital and inventive, or economic, activity.
A. Historical Identification of African American Inventors and Patentees
Data on patents are extracted first from survey data compiled by Henry Baker, a Second Assistant
Patent Examiner, who was African American.2 Baker conducted surveys of 9,000 patent agents and
2 Throughout the paper the terms “patent” and “utility patent” will be used interchangeably. A utility patent is issued for any new and useful process, machine, manufacture, composition of matter, or any new and useful improvement thereof. From 1995, utility patents are effective for 20 years from the date of application. Utility patents constitute over 95 percent of all patents granted African Americans. While it is
Inventing – Cook May 2001
5
attorneys for the Patent Office in 1900 and in 1913. Questions focused on information about
clients who were African American. His findings were published in a pamphlet, in the Journal of
Negro History, and as a two-volume catalogue of patents.3 The pamphlet and article provide
biographical information on a small set of inventors whose patents appear in the catalogue. The
Baker volumes include images of patent records, which include first and last name of the inventor,
city of patent application, patent number, dates of application and issue, title and description of
patent, and drawings. For the period under review, the Baker data were incomplete and needed to
be extended to include patent histories of inventors in his data set that extended beyond 1917,
missing patents, and patents obtained by other inventors after 1917.4
With the names of the inventors identified by Baker, executing the first task was straightforward
using online patent-search tools, including that of the European Patent Office (EPO). Just slightly
more than half of the inventors in the current data set were originally cited in a Baker publication.
In contrast, identifying missing and additional African American patentees was considerably more
difficult, since race is missing in patent data.5 The strategy executed was to identify African
Americans among the population of inventors and likely inventors from other sources and to match
them to patent records. In addition to collecting data from the aforementioned works, this was
accomplished by collecting names from modern and historical directories of African American
scientists, engineers, and medical doctors, e.g., Drew (1950), the State Library of New York, and
standard to use patents as a proxy for innovation and inventive activity, it should be recognized that this measure has limitations as, for instance, not all inventions are patentable or patented. 3 See Baker (1913, 1917, 1921). 4 Baker missed the first known U.S. patent to an African American inventor in 1821, which suggested that others might be missing. 5 Census-based approaches, including those exploiting the recent literature related to “black names”, were attempted but were unsuccessful. These are described in the Appendix.
Inventing – Cook May 2001
6
Sammons (1990); published biographies and collections of biographies, e.g., Aptheker (1951), Haber
(1970), Polski (1967), and Work (1921); programs from or scholarship on the “Negro Building” or
“Negro Day” at fairs and exhibitions related to science and invention, e.g., from the Garrett Morgan
Papers Collection and Foner (1978); Census data; and online searches of newspapers, company
archives, and patent databases, e.g., newspaperarchive.com, Western Electric and the Great Lakes
Patent and Trademark Center. While better-known inventors may appear in directories and
biographies, newspaper and obituary searches and programs from fairs and exhibitions capture
lesser-known inventors. For the sake of comparison to existing historical data sets, inventors were
selected according to the following criteria: identified by Henry Baker (1917), the aforementioned
patent examiner, as a significant inventor; identified in published scholarly literature as an important
inventor; or obtained four or more patents.6 The last criterion is consistent with a method of
defining “prolific” in Ciarlante (1978). In the resulting data set, patents are restricted to those
obtained by an inventor with at least one utility patent by 1930. Assigned patents are total or share
of utility patents which are assigned to a person or to a firm at issue.7 Since citation data, which are
used by the literature as a quality measure, are not publicly available until 1975, patent assignment
will be the best information available on a patent‟s commercial viability.8
This data set will have several limitations. First, Baker observes that there was likely under-
reporting by patent attorneys and agents who feared revelation of their clients‟ racial identity and
6 Another reason to select an alternative population from which one might obtain a sample of African American inventors is because it has been argued in the literature that African Americans are generally underrepresented in the DAB. Bandolph (1955) offers adjustments to the DAB and to Who’s Who in America for the period 1770 to 1936. This work resulted in the addition of one inventor, James Forten, to the data set. 7 Excluded are patents assigned to inventors in the data set but on which they are not designated as inventors, e.g., as assignees. 8 This will still be a crude measure of commercializability, because assignment at issue does not account for the secondary market in patents.
Inventing – Cook May 2001
7
possible negative consequences that may have resulted.9 Second, the method employed to find
other patentees is relatively conservative. Although a broad range of sources was used, African
American inventors who are not identified as such will not be matched to patents. It is a general
problem in the literature that less information exists on lesser-known inventors. Finally, patentees
for whom unique name matches are not found are dropped from the data set. There is no evidence
that the selection process follows a systematic process, and it should not significantly affect the
results or their interpretation.
In this study, inventions are total additions to the stock of knowledge for which a patent would have
been applied in the absence of institutional constraints, e.g., limitations on types of inventions that
were legally patentable or slavery; for which the inventor sought protection of intellectual property
through non-patenting means, e.g., trademarks or trade secrecy; or for which a patent was
obtained.10 Inventions are restricted to those of inventors who were productive by 1930. Data on
non-patented inventions and their inventors were collected from Aaseng (1997), Adams (1964),
Bandolph (1955), Drew (1950), Haber (1970), Sinclair (2004), Sullivan (1998), Williams (1978), and
Yancy (1984).
Patent and invention histories are created for each inventor identified. The resulting data set –
covering the period from 1843 to 1930 – contains information on 45 innovators (all but four of
9 For example, in the Garrett Morgan Papers it is noted that orders for Garrett Morgan‟s gas masks were canceled when southern fire chiefs saw a photo of him identifying him as the inventor of the modern gas mask. 10 Design patents, which comprise the “visual ornamental characteristics embodied in, or applied to, an article of manufacture,” are included among inventions but are excluded from patents in estimation due to data constraints. Carver could not patent many of his invention until the end of his career in the early 1920s, when it became legal to patent asexually-reproduced plants.
Inventing – Cook May 2001
8
whom are patent-holders) who were responsible for 1380 inventions and 363 patents.11
The data
comprise the patent number; inventor‟s full name, full names of co-inventors, and order of
appearance of names of inventors; location of the inventor; title of the patent; date of issue;
assignment status; assignee‟s name and location; and current USPTO patent class and sub-class.
Technological categories are taken from Hall, Jaffee, and Trajtenberg (2001) and are matched to
patented and to non-patented inventions.12 In the complete data set, data on inventions, patents,
and characteristics of their inventors have been merged.
B. Invention, Innovation, and Social Networks In examining the innovation system that gave rise to the American Industrial Revolution, Thomson
(2009) demonstrates that social ties were central to the new structure of technological change.13
Embodied technological progress spread through machinist and other inventor networks, i.e., in the
occupations and industries of inventors. In this sense, Thomson (2004, 2009) shows that creation
and dissemination of new knowledge were socially organized, and social capital was formed and
used to facilitate the flow of information needed for invention and innovation among inventors.
While suitable for the present investigation, this definition would be difficult to employ using data
on African American inventors. First, there is a paucity of biographical information, e.g., in the
11 Most of the unpatented inventions are those contained in the scientific papers of George Washington
Carver. Two inventors are slaves for whom patents were sought on their behalf but were not granted. Both trade secrets and significant public disclosure protected their inventions in these two cases, and their inventions were manufactured. In general, African Americans will likely patent less than their white counterparts due to constraints related to education, etc., at this time. 12 Technological classes created by Hall, Jaffe, and Trajtenberg (2001) are designed as an alternative to the USPTO technical classification to capture broad technological categories of innovation. Patents collected are matched to broad one-digit categories and more specific two-digit sub-categories. 13 In this paper, invention will be taken to mean the act of creating additions to the stock of knowledge. Innovation will mean commercialized invention.
Inventing – Cook May 2001
9
Census data on slaves, and on blacks more generally. Second, during the period of interest, there are
existing and increasing barriers to blacks‟ participation in the social networks Thomson cites as
important for innovation. For example, Margo (1990) finds that African Americans were barred
from many apprenticeships by local custom until 1950. African Americans live largely in the South
throughout the period of interest, and many networks useful for patented invention were outside the
South. Given that the first-best definition of social capital when considering the effect of social
capital on innovation cannot be used, I seek both the data that Thomson uses, e.g., on occupations,
schooling, and mobility, and data on broader social networks. That is, in this paper, social capital
will include social ties cited in the literature as being useful or necessary for inventive and innovative
activity and social ties seemingly unrelated to the inventive process, i.e., membership in civic or
professional organizations, election or appointment to public office, engagement in employment
with significant public exposure, friendships with well-known individuals, invention-related
partnerships, and evidence of entrepreneurial activities.14
Data on inventor‟s prior occupation, current occupation, and schooling are taken largely from U.S.
Censuses from 1880, 1900, 1910, 1920, and 1930. Additional inventor characteristics related to
social capital were collected from biographies, scholarly journals, archival collections, and the
National Inventors Hall of Fame: Aaseng (1997), Adams (1964), Baker (1917), Bandolph (1955),
Burkett, et al. (1991), Carter-Ives (1987), Carter-Sluby (2004), Fouché (2003), Hermann (1981),
Jenkins (1996), Kaplan (1955), James (1989, 2004), Lowry (2003), Meade (1946), the Garrett Morgan
Papers Collection, Princeton University (2007), Schmitz (1977), Sullivan (1998), Williams (1978), and
Yancy (1984). Since data on schooling are limited, state illiteracy data are extracted from Integrated
14 The terms “social capital,” “social ties,” “social networks,” and “social interactions,” will be used interchangeably in the paper. While the social capital being considered presently relates to invention and innovation, its conception is in line with that of Manski (2000) and Montgomery (1991) in their analysis of the social dimension of economic activity.
Inventing – Cook May 2001
10
Public Use Microdata Series (2004), approximately 50,000 individuals over 10 years old, and from
the University of Virginia Library (2010), full sample, individuals over 10 years old. Data on
aggregate illiteracy rates in 1890 are taken from Collins and Margo (2003). These data are derived
from the population of 10-69 year-olds using the full count. It is possible that innovators that
migration to an area occurs because of the rich social connections that promote innovation.15
Migration data are collected from Census and patent data. A person is a migrant if an inventor
changed his or her residence among the 1870, 1900, and 1920 U.S. censuses, patented in multiple
states, or both.16
Like most inventors of their day, African American inventors largely worked alone. Co-inventors
listed on patent documents were rare but more common than among other inventors.17 Turner
Byrd, Jr. and Granville T. Woods patented with male relatives.18 A few people, in addition to the
inventors of interest, were listed as both co-inventors and assignees. Such events are considered
partnerships and evidence of social capital in estimation.
The corollary of the social-capital hypothesis is that events that diminish social capital will reduce
inventive activity. Most U.S. patent activity was in northern states during the period of interest, and
15 See Krugman (1991) or Fujita, Krugman, and Venables (1999) on economic geography. 16 One potential problem in estimation may be that inventors may have moved to destinations of enhanced educational opportunities, which sometimes coincided with hubs of inventive activity. In this case, controls for education and migration could be fairly correlated. This effect should be small, because migration for the purpose of simultaneously increasing educational and inventive opportunities is reported for only two of 45 inventors in the data set. This trend is likely general. Logan (2009), for example, shows that health was a strong predictor in migration as well, lessening the impact of education. 17 At least one patent-holder was a member of a well-known research team. Lewis Latimer worked on Alexander Graham Bell‟s patent for the telephone and was also a member of Thomas Edison‟s research team, the “Edison Pioneers,” as an engineer and later as a member of the legal team. Other inventors in the data set, such as George W. Carver and Granville T. Woods, were invited to join Thomas Edison‟s research team but declined his offer. 18 Relatively little is known in the historical literature about their brothers who were co-inventors. Such collaboration does not appear to be commonplace among comparable white inventors, e.g., in the Khan-Sokoloff (1993, 2004) or Thomson (2009) data set.
Inventing – Cook May 2001
11
African American patentees were more active in Midwestern and Mid-Atlantic states than in other
regions. From baseball to education, the introduction of legal segregation and customs in the North
would likely disrupt previously open information flows among professions. For example, unlike
economic activity in new and growing ethnic enclaves, e.g., restaurants and funeral homes, patented
inventions required interaction with patent attorneys and agents, who were white until the 1970‟s.
White-only commercial districts, in which such intermediaries would have been located, would have
increased transactions costs and reduced social-capital formation between black inventors and
critical professional networks related to invention. Data on state segregation laws, proxies for the
increasingly racially polarized social and political environment of the late 1800‟s and early 1900‟s
throughout the U.S., have also been collected from jimcrowhistory.org and matched to states in
which patents were obtained by African Americans.19 Figure 1 reports the evolution of African
American patent activity, total U.S. patent activity, and the adoption of segregation during the period
of interest. The graph shows that patent activity and passage of Jim Crow laws move in opposite
directions. This relation will be tested formally in estimation.
Given that most inventors were independent and invention could be costly, a control for access to
credit is added to the analysis, and data on black-owned banks are collected from Ammons (1996).20
For comparison, data on all historical inventors in the National Inventors Hall of Fame, who are
19 The quantitative measure of legal segregation, number of new segregation laws passed in a given year, will not fully capture the depth and scope of informal segregation, e.g., extent of discriminatory informal customs and practices; quality of legal enforcement; and laws overturned after three years. It is reasonable to assume that informal Jim Crow customs and practices for which there was significant political consensus became embedded in law. Many such practices did not rise to this level of broad agreement but remained embedded in society, e.g., racial exclusion in apprenticeships. The segregation variable can measure some informal, but not all informal, segregation. 20 Following Khan and Sokoloff (1993), to control for productivity that is due to luck or “genius,” data on birth year is also collected from Censuses and biographies. These data will be limited and only used in estimation related to total inventive activity.
Inventing – Cook May 2001
12
only patentees, have also been collected.21
C. Summary Statistics
The new data, which provide an extensive set of information on black innovators and their level of
social capital, are presented in Table 1. Inspection of Table 1 reveals that most inventors have
several years or less of schooling. One third of inventors are machinists, and another third are
artisans and merchants. Nearly half had short patenting careers, i.e., of one to five years. More than
three quarters of African American inventors are migrants. More than a third of these migrant
inventors were born in Virginia and Kentucky, and most African American inventors engaged in
inventive activity outside the South. As identified in the historical literature or in patent records,
i.e., assignment of patents to inventor‟s own firm, entrepreneurs are relatively scarce among
African American inventors in the data.
A comparison of prolific inventors from the current data set and from the Inventors Hall of Fame,
the Khan and Sokoloff (1993), and Thomson (2009) also appears in Table 1. The four groups are
most similar with respect to occupation at first invention. Engineers, machinists, or full-time
inventors constituted the largest share of inventors in each group. They are also similar in that most
patenting occurs by age 40. African American inventors were more mobile than other inventors.
Thomson (2009) reports that 27.6 percent of inventors were born outside of their region or
country of inventive activity.22 More than half of prolific inventors in most regions of the U.S.
21 A representative patent is provided by the National Inventors Hall of Fame, and patent histories are not available for these inventors. 22 Thomson (2009), p. 116, Table 4.8.
Inventing – Cook May 2001
13
are migrants in the Khan and Sokoloff (1993) study. 23 While it is not reported in this table,
prolific inventors also obtain patents related to manufacturing more than in other fields.24
D. Social Capital Differences: An Illustration and Mechanisms
How different were social networks related to innovation and invention? To illustrate, Table 2
summarizes the major social connections of four members of the National Inventors Hall of Fame
and inventor-entrepreneurs, two of whom are African American and two of whom are white. Both
black and white inventors had wide-ranging social ties, e.g., to scientific publications, trade journals,
other inventors, and religious, civic, and educational institutions. African American inventors also
participated in the social institution of apprenticeship (formal or informal), maintained these
networks, and employed these machinists or found ones through these networks. Nonetheless,
with few exceptions, these ties were relatively more localized and tenuous for black inventors than
for their white counterparts. If comparing the inventors who are active in Cleveland, Charles Brush
and Garrett Morgan, Brush‟s social ties inside and outside inventive networks were broader and
deeper. Similarly, when comparing Granville Woods, who was known as the “Black Edison,” and
Thomas Edison, Edison‟s legions of firms based on his inventions and otherwise and affiliations
with the leading inventors of the day, including by marriage, dwarfed those of Woods. African
American inventors like Woods, knew suppliers to firms, such as General Electric and
Westinghouse, rather than the founders or significant members of their inventive teams.25 Further,
African American inventor networks often had longer gestation periods. For African Americans, it
was not unusual that informal or formal apprenticeships were preceded by unrelated non-technical
positions in firms, e.g., as a sweeper in the cases of Jan Matzeliger, an inventor of the shoe lasting
23 Khan and Sokoloff (1993), p. 293, Table 2. 24 Author‟s calculation. 25 Fouché (2003), pp. 64-65.
Inventing – Cook May 2001
14
machine, and Lewis Latimer, who eventually became an Edison Pioneer. In general, more robust
social ties are apparent among white inventors and likely shortened the time from idea to patent and
patent to commercial activity, along with increasing the magnitude and intensity of each activity in
this sequence.
Mechanisms
A few examples may illustrate the mechanisms by which social capital worked.
As aforementioned, Thomson (2009) shows that machinists are important transmitters of
information relevant for invention and innovation. During the period of study, as the Charles Brush
example shows, the Midwest was an important locus of inventive and innovative activity. Table 2
reports a representative sample of inventors and their locations. As can be seen, a significant
number of African American inventors were active in the Midwest. Coupled with the large number
of machinists in the sample, it appears that both occupation and region are associated with networks
of interest to our investigation. Fouché‟s (2003) biography of Granville Woods provides details of
his relatively deep integration into the machinist networks in and around Ohio. We will test
formally the effect of occupation and region as conduits of invention-related information below.
Among the broader social ties that may be relevant for African Americans may come through
positions of high visibility. Consistent with the evidence in Table 1, a large share of prolific
African American inventors was integrated into social networks, but these were mostly unrelated to
their inventive activity. Alexander Miles, the inventor of the forerunner of the modern elevator, was
the first black president of the Chamber of Commerce of Duluth, Minnesota. After having been
chair of the Republican Party for his county, George W. Murray, an inventor of agricultural
implements, was elected Congressman from his South Carolina district in 1892. It is plausible that
Inventing – Cook May 2001
15
both may have become acquainted with various individuals involved in inventive and innovative
activity in the regular course of business.
Finally, transactions costs associated with inventive activity increased amid increasing segregation
throughout the country. A number of items in the Garrett Morgan Papers and newspaper accounts
offer evidence that Garrett Morgan, the inventor of forerunners of the modern traffic light and gas
mask, employed a variety of techniques to disguise his identity as inventor of the gas mask, including
hiring whites to impersonate him and dressing as a Native American, “Big Chief Mason,” to
impersonate himself. Shelby Davidson, a supervisor in charge of the use, testing, and maintenance
of adding machines in the Post Office Division of the Treasury Department, had one patent for
adding machines and made improvements to his and others‟ for adoption by the Post Office. Due
to President Wilson‟s order to segregate the civil service in 1913 and despite the support of Booker
T. Washington and others of influence, Davidson, like many African Americans in supervisory
positions, was fired and ultimately lost access to the networks established at the Post Office and
among its vendors.26
In the next section, we will formally test the relation between social capital and economic activity, as
measured by patents.
III. Methodology
26 See Fouché (2003) for a comprehensive analysis of Davidson‟s tenure at the Post Office and as an inventor.
Inventing – Cook May 2001
16
Did social capital affect economic activity, as measured by patent output, among African Americans
in the late 19th and early 20th centuries? Does this result vary across types of inventive activity? The
main variables through which social capital is hypothesized to work are affiliations, co-inventor
relationships, segregation laws, occupations, reputation, and migration. In order to test the
hypothesis that more extensive social-capital networks led to greater inventive activity on the part of
black inventors, I estimate negative binomial and OLS models with the newly available Cook data.
A. Panel Regressions It is possible to measure patent activity as the count of total or assigned patents or as the share
of assigned patents. A negative-binomial specification is employed, because the number of
patents granted an individual inventor over time is a non-negative count variable and because it
accounts for overdispersion and the irregular rate at which patents arrive across years and across
inventors better than OLS. Otherwise, OLS is used when the dependent variable is a continuous
variable, i.e., share of assigned patents, and cannot be represented in a count model. 27 The new
data contain both a count and a share measure. Using correlates of inventive activity as in Khan
and Sokoloff (1993) and Thomson (2009), the expected number of inventions patented by an
inventor in a given year conditional on inventor and state characteristics is
E(PATENTit) = exp[α + β1SOCAPi + β2COINVENTit + β3SEGLAWit + β4FAMEi +
β5MIGRATEi + XΨ + εit + ζit]. (1)
27
OLS is appropriate, because share of patents assigned is a continuous variable, and its mass is fairly evenly
distributed between 0 and 1.
Inventing – Cook May 2001
17
SOCAPi is the presence of social capital for inventor i; COINVENTi is average number of co-
inventors per patent; SEGLAWi is the average cumulative number of laws promoting segregation
in the state of patenting for inventor i across patent years; FAMEi is a dummy for inductee into
the National Inventors Hall of Fame incorporated to account for inventor reputation; MIGRATEi is a
dummy for migration status; and εit is the random error component of composite error term, ζ it.28
The term X is a vector of controls, including a proxy for access to credit, cumulative number of
black banks in the state of inventor i at time t; experience and experience squared in patenting,
which are calculated from year of the given patent minus year of the first patent granted to
inventor i; dummies for inventor‟s major technological class of invention – electrical and
mechanical; for prior and current occupation – artisan, farmer, machinist, professional, and
manufacturer; and for region of principal inventive activity. 29
Inventors may be active in an occupation in a given region because of the strong social connections
that have emerged to promote invention and innovation. To investigate the validity of this
extension of the Thomson hypothesis, we re-specify equation (1) to include an interaction term
between engineer, machinist, or full-time inventor and Midwest.
E(PATENTit) = exp[α + β1SOCAPi + β2COINVENTit + β3SEGLAWit + β4FAMEi +
β5MIGRATEi + β6(MIDWESTi * ENGINEERi) + XΨ + εit + ζit]. (2)
28 While the proliferation of segregation laws may have diminished previously existing social capital useful for inventive and innovative activity or constrained their development, segregation also led to the formation of an unprecedented number of black firms. Broadly speaking, social capital was likely most developed where there is a significant and active black population. The number of black banks in a state may be considered a proxy for this and is included in the regressions. 29 The dummy for patent class indicates the modal patent class for inventors with inventions in multiple classes in the pooled regression.
Inventing – Cook May 2001
18
An estimate of β6 that is positive is consistent with information flows being greater and social capital
having an even stronger impact on invention for those machinists working in the Midwest.
As a weighted average of the fixed- and between-effects estimators, a random-effects estimator is
employed, because it is likely that some omitted variables vary across time and across inventors.30
Using the time dimension of the panel data also allows us to control for experience or “learning-by-
doing” effects or enhancing one‟s reputation over time.31 In this specification, standard errors are
clustered on inventor.
The coefficients of interest are β1, β2, β3, β4, β5, and β6. Assigned patents are the inventions of
greatest economic interest.32 Descriptive statistics are reported in Table 4.
IV. Results
Results from estimating models of economic activity as measured by inventive and patent outcomes
are reported in Table 5.
Social Capital
Nontraditional forms of social capital are significantly and positively correlated with economic
outcomes as measured by patents. In the random-effects estimation, the estimated coefficients
suggest that possessing nontraditional social capital is associated with 0.68 additional assigned patent
per inventor per year and an additional share of 0.49 in assigned patents per inventor per year. In
30 In specification tests, the random-effects, rather than the fixed-effects, estimator was the estimator with minimum variance. See Wooldridge (2002) for a discussion of decisions related to fixed- and random-effects estimators. 31 See, for example, Kim, et al. (2009). 32 Nonetheless, assignment data provide lower-bound estimates of the number of patents assigned or commercialized, because patents are observed at issue and not over the lives of patents.
Inventing – Cook May 2001
19
the pooled-OLS estimation, possessing social capital is associated with five additional patents per
inventor, 0.73 additional assigned patent per inventor, and an additional share of 0.36 in assigned
patents per inventor. These findings are robust to various model specifications and definitions of
the dependent (patent) variable. We calculate that, given years of experience, inventors not endowed
with this type of social capital are predicted to receive 0.64 fewer patents per year than their
endowed counterparts.
Traditional social capital is also relevant. While the estimated association with occupation by itself is
ambiguous, the estimated coefficient on Midwestern engineer is positive and significant across
specifications, types of invention, and data sets. Although not reported, patenting in electrical and
mechanical fields is positively correlated with patenting outcomes, which is in line with information
channels likely being more open in these industries than in others. This evidence is consistent with
that of Thomson (2004, 2009), who finds that innovation hinged critically on knowledge that was
socially organized, i.e., that was shared through networks within and across industries.
Similarly, factors limiting social capital, such as segregation laws, are negatively and significantly
correlated with patent activity, especially among those of the greatest economic importance –
assigned patents. More segregation corresponds to lower assigned-patent activity, whether
considering total assigned patents or shares. multiracial social networks, among other networks,
became illegal, precarious, or infeasible. Increasing segregation, as measured by cumulative
segregation laws, made some social networks illegal and others more precarious, which would also
erode previously integrated economic relations.
Inventing – Cook May 2001
20
This result lends support to Robert Putnam‟s (2000) observation that, “Norms and networks that
serve some groups may obstruct others, particularly if the norms are discriminatory or the networks
socially segregated.” It is also consistent with Laird‟s (2006) observation that after emancipation and
during the age of segregation, it was difficult for African Americans to access social capital outside
their own community, e.g., in large corporations, and flourished in developing and leveraging their
own social capital among African Americans. To be successful, the evidence is suggestive that
African American patentees had to link the two loci of social capital.
For total patents, the relation between social capital and inventive activity holds only when using a
traditional measure of social capital, Midwest and engineer. While not reported, results from using
all inventions and patents in the data set are similar.33 This is suggestive of a relation that applies to
a specific range of inventive and economic activity. One possible explanation is that the act of
patenting forces interaction with patent intermediaries, agents and attorneys, who have access to
larger networks, professional and social.34 Interaction with patent attorneys or agents is not required
for non-patented invention, and extensive use of their networks may not be triggered simply by the
presentation of inventions with low probability of success or assignment, i.e., commercialization.
Across specifications, the estimated coefficient on COINVENT is largely negative and significant.
From historical records, the contributions of a number of co-inventors were not clear, which may
suggest a partnership or an alternative channel for social capital to affect inventive activity. For
example, Fouché (2003) reports that Granville Woods often lacked the funds to file a patent
application, and one proposed solution was to offer co-inventor or assignee status to potential
33 The regression with non-patented and patented inventions is executed as a negative binomial regression executed in pooled data, because precise dates of invention are not available for non-patented inventions. 34 According to the National Bar Association, until the 1970‟s, patent agents and attorneys were white, which would imply that their networks may have been larger than those of African American at the time.
Inventing – Cook May 2001
21
financiers.35 This finding implies that, during the age of the independent inventor, the presence of
co-inventors may have indicated a social tie but not one that would be salient for patenting or
economic outcomes. This result is not entirely surprising, since more than half of co-inventors were
relatives, e.g., Woods‟s brother Lyates, for whom the marginal contribution of social capital may
have been negligible.
IV. Conclusion and Future Research
Using the laboratory of economic history, I examine the value of social networks in economic
activity. I collect data on African American inventors, their inventions, and their personal
characteristics to execute this test. In the analysis of cross-section and panel data, I find that
social ties are important for the most economically relevant inventive activity, assigned patents,
and not important for less economically relevant activity. Further, I find that factors that disrupt
social ties, i.e., laws promoting segregation, reduce patent output. In toto, this evidence is
consistent with the existing literature that posits a positive relation between social capital and
economic outcomes.
Further research would focus on a deeper understanding of the means by which social capital was
negotiated and leveraged and of the timing of social-capital formation. It is likely that a number
of these networks were informal, and data may not exist. If they exist, the current research would be
enhanced by their collection and analysis.
35 Fouché (2003), pp. 26-81.
Inventing – Cook May 2001
22
References Aaseng, Nathan. Black Inventors. New York: Facts on File, 1997. Adams, Russell L. Great Negroes Past and Present. Chicago: Afro-American Publishing Co., 1964. Agarwal, Ajay, Devesh Kapur, and John McHale. “Birds of a Feather -- Better Together?: Exploring the Optimal Spatial Distribution of Ethnic Inventors,” NBER Working Paper No. 12823, January 2007. Ammons, Lila, “The Evolution of Black-Owned Banks in the United States Between the 1880s and
the 1990s, Journal of Black Studies, 26 (1996), 467-489. Ancestry.com, “Census Search,” 1870 to 1930, last accessed March 2008. Aptheker, Herbert. A Documentary History of the Negro People in the United States. New York: Citadel Press, 1951. Baker, Henry E. “The Colored Inventor,” The Crisis Publishing Company, 1913 (pamphlet reprinted by Arno Press, 1969). . “The Negro in the Field of Invention,” Journal of Negro History, 2 (1917), 21-36. . Negro Inventors, Vols. 1--4. 1921, Moorland-Spingarn Research Center, Howard University. ____________. Henry E. Baker Papers, Carter G. Woodson Collection, Library of Congress.
Bandolph, Richard. “The Distinguished Negro in America, 1770-1936," American Historical Review,
60 (1955), 527-547. Bertrand, Marianne and Sendhil Mullainathan. “Are Emily and Greg More Employable than Lakisha and Jamal? A Field Experiment on Labor Market Discrimination,” The American Economic Review, 94 (2004), 991 – 1013. Brush Wellman Inc., "The Life of Charles F. Brush, Sr.," Brush Wellman Inc., http://www.brushelmore.com/history-charlesbrush.asp Burkett, Randall K., Nancy Hall Burkett, Henry Louis Gates, Jr. Black Biography, 1790-1950: A Cumulative Index. 3 Volumes. Alexandria: Chadwyck-Healey, 1991.
Carter-Ives, Patricia. Creativity and Inventions: The Genius of Afro-Americans and Women in the United
States and Their Patents. Arlington, VA: Research Unlimited, 1987. Carter-Sluby, Patricia. The Inventive Spirit of African Americans. Westport, Connecticut: Praeger, 2004. Castillio, Emilio, Hokyu Hwang, Ellen Granovetter, and Mark Granovetter. “Social Networks in
Inventing – Cook May 2001
23
Silicon Valley,” in Lee, Chong-Moon, William F. Miller, Marguerite Gong Hancock, Henry S. Rowen, eds. The Silicon Valley Edge. Stanford, CA: Stanford University Press, 2000. Chernow, Ron. The House of Morgan: An American Banking Dynasty and the Rise of Modern Finance. New York: Atlantic Monthly Press, 1990. Ciarlante, Majorie Heins. A Statistical Profile of Eminent American Inventors, 1700--1860: Social Origins and Role. Ph.D. Dissertation, Northwestern University, 1978. Collins, William J. and Robert A. Margo, “Historical Perspectives on Racial Differences in Schooling in the United States,” NBER Working Paper 9770, June 2003. Cook, Lisa D., “Coase Conquers Consumer-Side Discrimination: Firm-level Evidence from the Age of Segregation,” December 2009, Michigan State University, mimeo. . “Violence and Economic Activity: Evidence from African American Patents, 1870 to 1940,” mimeo, Stanford University, July 2009. Drake, Francis S., ed. Dictionary of American Biography. Boston: Houghton, Osgood, 1879.
Drew, Charles R. “Negro Scholars in Scientific Research,” Journal of Negro History, 35 (1950), 135-
149. DuBois, W.E.B. The Negro in Business. New York: AMS Press, 1899 [1971]. Encyclopedia of Cleveland History, various entries, http://ech.cwru.edu/, last accessed December 2010. European Patent Office. “Patent Search,” www.ep.espacenet.com, last accessed January 2007. ___________. The House of Rothschild: The World's Banker: 1849-1999. New York: Penguin, 2000.
Foner, Philip S. “Black Participation in the Centennial of 1876,” Phylon, 39 (1978), 283-296. Fouché, Rayvon. Black Inventors in the Age of Segregation: Granville T. Woods, Lewis H. Latimer, and Shelby J. Davidson. Baltimore: Johns Hopkins University Press, 2003. Fryer, Roland G. and Steven D. Levitt. “The Causes and Consequences of Distinctively Black Names,” The Quarterly Journal of Economics, 119 (2004), 767 – 805. Garrett A. Morgan Papers Collection, Western Reserve Historical Society Library, June 2007. Granovetter, Mark,“The Impact of Social Structure on Economic Outcomes,” The Journal of Economic Perspectives, 19 (2005), 33-50. Great Lakes Patent and Trademark Center of the Detroit Public Library. “African American Inventors Database,” http://www.detroit.lib.mi.us/glptc/aaid/, last accessed January 2007.
Inventing – Cook May 2001
24
Haber, Louis. Black Pioneers of Science and Innovation. New York: Harcourt, Brace, and World, 1970. Hall, Bronwyn, Adam Jaffee, and Manuel Trajtenberg. “The NBER Patent Citations Data File: Lessons, Insights, and Methodological Tools,” NBER Working Paper 8498, October 2001. Hermann, Janet Sharp. The Pursuit of a Dream. Oxford: Oxford University Press, 1981. “The History of Jim Crow,” www.jimcrowhistory.org, last accessed October 2007. Integrated Public Use Microdata Series (2004) Steven Ruggles, Matthew Sobek, Trent Alexander, Catherine A. Fitch, Ronald Goeken, Patricia Kelly Hall, Miriam King, and Chad Ronnander. Integrated Public Use Microdata Series: Version 3.0 [Machine-readable database]. Minneapolis, MN: Minnesota Population Center [producer and distributor], 2004. http://usa.ipums.org/usa, last accessed January 2004. James, Portia. “Invention and Innovation, 1619-1930,” in Bruce Sinclair, ed., Technology and the Experience of African Americans. Cambridge, MA: MIT Press, 2004. ___________. The Real McCoy: African American Invention and Innovation, 1619-1930. Washington: Smithsonian Institution, 1989. Jenkins, Edward Sidney. To Fathom More: African American Scientists and Inventors, Lanham, MD: University Press of America, 1996. Johnson, Allen and Dumas Malone, eds. Dictionary of American Biography. 21 Volumes. New York, Charles Scribner‟s Sons, 1937. Kaplan, Sidney. “Jan Ernst Matzeliger and the Making of the Shoe," Journal of Negro History, 40, (1955), 8--33. Kerr, William R., “Immigrants and Spatial Adjustments in US Invention,” The Review of Economics and Statistics, 90 (2008), 518-537. Kim, Jinyoung, John Lee Sangjoon, and Gerald Marschke, “Inventor Productivity and Firm Size: Evidence from Panel Data on Inventors,” Pacific Economic Review, 14 (2009), 516-531. Laird, Pamela Walker. Pull: Networking and Success since Benjamin Franklin. Cambridge, MA: Harvard University Press, 2006.
Lamoreaux, Naomi. “Banks, Kinship, and Economic Ties,” Journal of Economic History, 46 (1986),
647-667.
Inventing – Cook May 2001
25
________, Margaret Levenstein, Kenneth L. Sokoloff, Financing Invention During the Second Industrial Revolution: Cleveland, Ohio, 1870-1920, in Lamoreaux and Sokoloff, eds., Financing Innovation in the United States, 1870 to Present. Cambridge: MIT Press, 2007. ________, and Kenneth L. Sokoloff, eds., Financing Innovation in the United States, 1870 to Present. Cambridge: MIT Press, 2007. Krugman, Paul, “Increasing Returns and Economic Geography,” Journal of Political Economy, Vol. 99, No. 31, 1999, pp. 483-499. ________, Masahisa Fujita, and Anthony J. Venables. The Spatial Economy: Cities, Regions, and International Trade. Cambridge, MA: MIT Press, 1999. Logan, Trevon D. “Health, Human Capital, and African American Migration Before 1910,” Explorations in Economic History, 46 (2009), 169-185. Lowry, Beverly. Her Dream of Dreams: The Rise and Triumph of Madam C.J. Walker. New York: Alfred K. Knopf, 2003. Manski, Charles F., “The Economics of Social Interactions,” Journal of Economics Perspectives, 14 (2000), 115-136. Margo, Robert A. “Accumulation of Property by Southern Blacks Before World War I: Comment and Further Evidence,” American Economic Review, 74 (1984), 768-776. Meade, George P. “A Negro Scientist of Slavery Days.” The Scientific Monthly, 62 (1946), 317-326. Montgomery, James D. “Social Networks and Labor-Market Outcomes: Toward an Economic Analysis,” American Economic Review, 81 (1991), 1408-1418. Munshi, Kaivan, “Networks in the Modern Economy: Mexican Migrants in the U.S. Labor Market,” Quarterly Journal of Economics, 118 (2003), 549-599.
National Inventors Hall of Fame, “Inventor Search,” www.invent.org, last accessed January 2010. Neal, Larry and Lance E. Davis, “Why Did Finance Capitalism and the Second Industrial Revolution Arise in the 1890s?” in Lamoreaux and Sokoloff, eds., Financing Innovation in the United States, 1870 to Present. Cambridge: MIT Press, 2007. Newspapercarchive.com, “Search,” last accessed January 2010. The University of the State of New York. “African-American Bibliography - Science, Medicine, and Allied Fields: Selected Sources from the Collections of the New York State Library," 1992, www.africa.upenn.edu/Bibliography/AFAM_History.html, last accessed January 2007. Ohio Historical Society Online Collection Catalog, “Newspaper Database,” http://www.ohiohistory.org/occ/menu.htm, last accessed January 2010. Polski, Harry A. The Negro Almanac. New York: Bellwether Publishing Company, 1967.
Inventing – Cook May 2001
26
Princeton University. “The Faces of Science: African Americans in the Sciences," www.princeton.edu/~mcbrown/display/faces.html last accessed January 2007. Putnam, Robert. Bowling Alone: The Collapse and Revival of American Community. New York: Simon & Schuster, 2000. Rauch, James E. and Vitor Trindade. “Ethnic Chinese Networks in International Trade,” Review of Economics and Statistics, 84 (2002), 116-130.
Sammons, Vivian O. Blacks in Science and Engineering. New York: Hemisphere Publishing, 1990. Saxenian, AnnaLee, “Networks of Immigrant Entrepreneurs,” in Chong-Moon Lee, William F. Miller, Marguerite Gong Hancock, Henry S. Rowen, eds. The Silicon Valley Edge. Stanford, CA: Stanford University Press, 2000. Schmitz, Mark D. “Economies of Scale and Farm Size in the Antebellum Sugar Sector,” Journal of Economic History, 37 (1977), pp. 959-980. Sinclair, Bruce, ed. Technology and the Experience of African Americans. Cambridge, MA: MIT Press, 2004. Sokoloff, Kenneth. “Inventive Activity in Early Industrial America: Evidence from Patent Records, 1790--1846,” Journal of Economic History, 48 (1988), 813-850. and Zorina Khan. “‟Schemes of Practical Utility‟: Entrepreneurship and Innovation Among „Great Inventors‟ in the United States, 1790-1865,” Journal of Economic History, 53 (1993), 289-307.
_____________. “Institutions and Democratic Invention in the 19th Century:
Evidence from the „Great Inventors‟ of the United States, 1790-1930,” mimeo, 2004. Sullivan, Otha Richard. Black Stars: African American Inventors. New York: John Wiley & Sons, Inc., 1998. Thomson, Ross. Structures of Change in the Mechanical Age: Technological Innovation in the United States, 1790-1865. Baltimore: The Johns Hopkins University Press, 2009. _____________, “From the Old to the New: The Social Basis of Innovation in the Antebellum United States,” BHC, 2004, http://www.thebhc.org/publications/BEHonline/2004/Thomson.pdf. Trajtenberg, Manuel. “Innovation in Israel, 1968-1997: A Comparative Analysis Using Patent Data,” Research Policy 30 (2001), 363-389. U.S. Census (1910). Inter-University Consortium for Political and Social Research. Study 00003: Historical Demographic, Economic, and Social Data: U.S., 1790-1970. Ann Arbor: ICPSR. http://fisher.lib.virginia.edu/cgi-local/censusbin/census/cen.pl, last accessed January 2007.
Inventing – Cook May 2001
27
______________. Historical Census Statistics on Population Totals By Race, 1790 to 1990, and By Hispanic Origin, 1970 to 1990, For The United States, Regions, Divisions, and States, September 2002, http://www.census.gov/population/www/documentation/twps0056/twps0056.html, last accessed January 2010. U.S. Patent and Trademark Office. “Patent Search," www.uspto.gov/patft/index.html, last accessed January 2007. University of Virginia Library, Historical Census Browser, last accessed January 2010. Washington, Booker T. The Negro in Business. Hertel, Jenkins, & Co., 1907. “Western Electric History,” www.porticus.org/bell/westernelectric_history.html, last accessed January 2010. Williams, James C. At Last Recognition in America: A Reference Handbook of Unknown Black Inventors and Their Contributions to America, Volume I. Chicago: B.C.A. Publishing Corp., 1978 Wooldridge, Jeffrey M. Econometric Analysis of Cross Section and Panel Data. Cambridge, MA: The MIT Press, 2002. Work, Monroe N, ed. Negro Year Book: An Annual Encyclopedia of the Negro, 1921-22. Tuskegee, AL: Negro Year Book Publishing Co., 1921. Yancy, Dorothy Cowser. “The Stuart Double Plow and Double Scraper: The Invention of a Slave,” Journal of Negro History, 69 (1984), 48-52.
Inventing – Cook May 2001
28
Figure 1. Total African American Inventors and Segregation Laws, 1834 to 1930
Source: Cook (2003, 2006, 2008); USPTO (2007); U.S. Census (2002).
Note: Intercensal population is estimated by linear interpolation.
05
10
Pa
tents
per
mill
ion
, A
A/N
um
be
r o
f la
ws
0
100
200
300
400
Pa
tents
per
mill
ion
, T
ota
l
1830 1850 1870 1890 1910 1930Year
Total patents African American patents
Segregation laws
Inventing – Cook May 2001
29
Table 1. Prolific Inventor Characteristics, Compared
Hall of Fame Khan-Sokoloff Thomson
African
American
Prolific Inventors Number % Number % Number % Number %
Total inventors 132 -- 160 -- 282 -- 45 --
Average patents per patentee -- -- 7.9 -- 6.8 -- 9 --
Inventors without patents 0 0.0 10 16.0 0 0.0 4 11.3
Schooling
None to several years of schooling 9 6.8 76 47.5 69 24.6 18 40.0
More than several years 17 12.9 22 13.8 53 18.8 3 6.7
Attended college 45 34.1 38 23.8 160 56.6 13 28.9
Unknown 61 46.2 24 15.0 0 0.0 11 24.4
Occupational class at first major invention
Artisan 6 4.5 24 15.0 27 9.6 8 17.8
Farmer 1 0.8 8 5.0 10 3.6 3 6.7
Engineer/machinist/full-time inventor 80 60.6 53 33.1 185 65.4 15 33.3
Merchant/professional 11 8.3 36 22.5 25 8.9 8 17.8
Manufacturer 11 8.3 37 23.1 35 12.5 1 2.2
Other/missing 23 17.4 2 1.3 0 0.0 10 22.2
Age at first major invention
<20 0 0.0 9 5.6 -- -- 1 2.2
20-29 25 18.9 41 25.6 -- -- 10 22.2
30-35 31 23.5 42 26.3 -- -- 12 26.7
36-40 19 14.4 26 16.3 -- -- 6 13.3
41-45 12 9.1 13 8.1 -- -- 4 8.9
46-55 45 34.1 21 13.1 -- -- 3 6.7
>55 0 0.0 8 5.0 -- -- 1 2.2
Unknown 0 0.0 0 0.0 -- -- 8 17.8
Duration (in years) of career in patenting
0-5 -- -- 45 28.1 -- -- 22 48.9
6-10 -- -- 11 6.9 -- -- 4 8.9
11-20 -- -- 31 19.4 -- -- 7 15.6
21-30 -- -- 37 23.1 -- -- 3 6.7
>30 -- -- 36 22.5 -- -- 9 20.0
Source: National Inventors Hall of Fame; Khan and Sokoloff (1993); Thomson (2009); Cook (2003, 2006)
Note: Hall of Fame data include inventors that have inventions between 1790 and 1929.
Khan-Sokoloff patent histories extend from 1790 through
1865.
Thomson data are for 1790 to 1865.
African American patent histories are restricted to U.S. utility patents and extend from 1834 through 1962.
The total number of U.S. patents granted between 1790 and 1930 is 1,795,701.
The total number of U.S. patents granted to African Americans between 1790 and 1930 is 710.
There are nine African Americans in the Hall of Fame data set and one reported in the Khan-Sokoloff data set.
Inventing – Cook May 2001
30
Table 2. Comparison of Social Capital of White and African American Inventors, 1843-1930
Inventor Place of Education, Main Area Pat- Firms Created Interest Significant Other Significant
Dates Inventive Training of Invention; ents Related to in Other Financial Social Connections,
Birthplace Activity Inventions Inventions Firms Connections Positions Held
Charles F. Cleveland, BS, MS, PhD Mechanics, 60 Total -- 5 Total -- 18 Euclid Ave. University of Michigan
Brush Ohio in mining electricity; Brush Electrical Engineering Linde Air National Western Reserve
1849-1929 engineering dynamo, Co. Products Co., Bank, Pres. University, Trust.
Euclid, arc lights Brush Electric Co. Found., Pres. George W. University School, Trust.
Ohio Sandusky Stockly (local Delta Kappa Epsilon
Portland telegraph Brush Foundation, Found.
Cement Co., manufacturer) Cleveland Vocal Society,
Co-found. Honorary Pres.
Cleveland Adelbert College, Trust.
Arcade Co., Cleveland School of Art,
Pres. Trust.
Euclid Avenue Lake View Cemetery,
National Trust.
Bank, Pres. Trinity Protestant Episc-
Brush- opal Church, Warden
Wellman, Inc. Cleveland Chamber of
Commerce, Pres.
Rumford Prize
Edison Medal
Franklin Medal
Garrett A. Cleveland, 6+ years of Mechanics, 4 Total -- 3 Total -- 2 Victor Sincere General Electric
Morgan Ohio schooling; electricity; G.A. Morgan Hair Refining [Cleveland] (local depart- Western Reserve University
1877-1963 Machine Traffic light, Co., Found., Pres. Call & Post, ment store Alpha Phi Alpha
Paris, repair gas mask National Safety Device Co., Found. manager) Fire chiefs
Kentucky Found., Gen. Mgr. Wakeman Local attorneys, business,
Morgan's Cut-Rate Ladies Country Club, religious, and civic leaders
Clothing Store, Found., Pres. Found., Pres. Antioch Baptist Church, Off.
NAACP, Local Off.
International Association
of Fire Engineers Medal
Continued on next page.
Inventing – Cook May 2001
31
Table 2. Comparison of Social Capital of White and African American Inventors, 1843-1930 (Cont.)
Inventor Place of Education, Main Area Pat- Firms Created Interest Significant Other Significant
Dates Inventive Training of Invention; ents Related to in Other Financial Social Connections,
Birthplace Activity Inventions Inventions Firms Connections Positions Held
Thomas A. Menlo Park, Home- Electricity; 1,093 Total -- 23 Total -- 141 J.P. Morgan Nikola Tesla
Edison West schooled, microphone, Edison General Electric Co. Kinétoscope Vanderbilts Henry Ford
1847-1931 Orange, and telegrapher, phonograph, Edison Electric Light Co. Belge Drexel, Mor- "Edison Pioneers"
Milan, Ohio Newark, machinist incandescent Edison Manufacturing Co. Newark Tele- gan, and Co. Benjamin Bredding
New Jersey light bulb Edison Portland Cement Co. graph Works (telephone inventor)
New York, Argentine Edison Light Co. Edison Wiring Lewis Miller (inventor,
New York Edison Milling Ore Co., Ltd. Co. father-in-law)
Motion Pictures Patents Co. Deutsche Edison Charles Lindberg
Drexel, Morgan, and Co. Gesellschaft Marie Curie
Edison Business Mexican Nat'l Herbert Hoover
Phonograph Co. Phonograph Co. Warren G. Harding
American Novelty Co. Toy Phono- Harvey Firestone
Port Huron and Gratiot graph Co. Naval Consulting Board
Street Railway Co. Western Union Fort Myers Civitan Club
American Telegraph Works Telegraph Co.
Thomas A. Edison, Inc. Societe du Tele-
phone Edison
Granville T. New York, Training and/ Electricity 45 Total -- 3-4 N/A James Zerbe General Electric
Woods New York or degree in Woods Electric Company Calvin Bowen Westinghouse
Columbus, Cincinnati, mechanical, Machine repair shop(s) Lucius H. Ward Leonard
1856-1910 Ohio electrical Robertson (Westinghouse supplier)
Ohio or engineering, John A. Gano
Australia machinist, Ralph Peters
railroad James E.
engineer Chandler
Source: Brush -- "The Life of Charles F. Brush, Sr.," Brush Wellman Inc., http://www.brushelmore.com/history-charlesbrush.asp;
"Charles F. Brush," Encyclopedia of Cleveland History; Lamoreaux, Levenstein, and Sokoloff (2004)
Edison -- Rutgers University, http://edison.rutgers.edu/list.htm
Morgan -- Garrett Morgan Papers; Cook (2009)
Woods -- Fouché (2003); Cook (2003, 2008)
Note: All inventors are inductees in the National Inventors Hall of Fame. While not listed, all are familiar with machinists in their fields and in
their locations. Brush and Edison are white. Morgan and Woods are African American.
Inventing – Cook May 2001
32
Table 3.
Inventions by African Americans, Selected, 1843 to 1930
Year Inventor Invention Location
1843 Norbert Rillieux Improvement in sugar-works New Orleans, LA
1850s Benjamin T. Montgomery Boat propeller Cincinnati, OH
1872 Elijah McCoy Automatic lubricator cup Ypsilanti, MI
1872 Turner Byrd, Jr. Neck yokes for wagons, holder for reins Williamsville, MI
1875 Alexander P. Ashbourne Method of preparing coconut Oakland, CA
1878 Benjamin H. Taylor Improvement in rotary engine Rosedale, MS
1881 Lewis H. Latimer Carbon filaments for electric incandescent lamp New York, NY
1883 Jan Ernst Matzeliger Automatic method for lasting shoes Lynn, MA
1887 Alexander Miles Elevator Duluth, MN
1887 Granville T. Woods Telephone system, electro-mechanical Cincinnati, OH
brake, railway telegraphy, third rail New York, NY
1890 Frank J. Ferrell Steam trap, apparatus for melting snow, valve New York, NY
1890 William B. Purvis Paper-bag machine Philadelphia, PA
1891 Albert C. Richardson Churn South Frankfort, MI
1894 George W. Murray Fertilizer distributor, planter, cotton chopper Sumter, SC
1895 Clatonia J. Dorticus Machine for embossing photographs Newton, NJ
1897 Andrew Jackson Beard "Jenny" coupler (for train operators) Eastlake, AL
1906 Madame C.J. Walker Hair and beauty products Indianapolis, IN
1910 Ned E. Barnes Indicator or bulletin Willis, TX
1912 Oscar Robert Cassell Flying machines New York, NY
1914 Garrett A. Morgan Gas mask Cleveland, OH
1920 Asa J. Taylor, Jr. Electric gear shifter Chicago, IL
1925 George W. Carver Paint and stain and producing the same Tuskegee, AL
1930 Solomon Harper Electrical hair-treating implement New York, NY
Source: Baker (1917), USPTO, EPO, Hermann (1981), Lowry (2003) Note: Year is the year of a representative patent is awarded or, for those without patents, approximate
date of the invention's introduction. Location is location of major inventive activity. Co-inventors are not listed.
Inventing – Cook May 2001
33
Table 4. Descriptive Statistics
Inventors Patentees
Total patents, Mean 8.18 9.00
per inventor S.D. (10.56) (10.74)
N 44 40
Total assigned patents, Mean 5.11 5.55
per inventor S.D. (8.91) (9.24)
N 44 40
Share of assigned patents, Mean 0.45 0.45
per inventor S.D. (0.35) (0.35)
N 40 40
Connected Mean 0.86 0.85
S.D. (0.35) (0.36)
N 44 40
Co-inventor Mean 0.10 0.10
S.D. (0.18) (0.19)
N 44 40
Migrate Mean 0.78 0.79
S.D. (0.42) (0.41)
N 37 34
Hall of fame Mean 0.18 0.20
S.D. (0.39) (0.41)
N 44 40
Segregation laws, Mean 1.83 1.83
per state S.D. (3.09) (3.09)
N 39 39
Banks, Mean 0.55 0.55
per state S.D. (1.44) (1.44)
N 39 39
Illiteracy rate, Mean 0.20 0.20
per state S.D. (0.16) (0.16)
N 39 39
Source: See text for sources.
Note: Summary statistics are reported for inventors and patentees, excluding outliers.
See text for variable definitions.
Inventing – Cook May 2001
34
Table 5. Random-Effects Estimates
Explanatory Variables
Total
Patents Assigned
Share
Assigned
Per
Inventor
Patents Per
Inventor
Patents Per
Inventor
Connected -0.196 0.674*** 0.500***
(0.940) (0.198) (0.101)
Co-inventor -0.405*** -0.299* 0.030
(0.113) (0.154) (0.067)
Segregation laws -0.004 -0.051*** -0.034**
(0.039) (0.017) (0.014)
Migrate -0.295 -0.068 0.036
(0.457) (0.209) (0.118)
Hall of Fame -0.405 -0.112 -0.109
(0.421) (0.226) (0.101)
Midwest -0.447 -0.375 -0.100
(0.277) (0.265) (0.180)
Engineer -0.392 -0.081 -0.060
(0.380) (0.167) (0.105)
Midwest x Engineer 0.827** 0.955*** 0.387**
(0.369) (0.233) (0.182)
Patent class controls Yes Yes Yes
Wald 804.403 231.853 469.956
N 195 195 195
Note: Dependent variables are total patents, assigned patents, and share assigned patents per inventor per year.
All columns are estimated as random effects models. Constants are included in estimation but not reported.
Cluster-robust standard errors on inventor are in parentheses.
Data on segregation laws, banks, and illiteracy rates correspond to the states in which patents are
obtained by inventor i. Controls for patent class are included in all models.
Co-inventor is average number of co-inventors per patent.
Coefficients marked with an asterisk (***) are significant at the 1 percent level of
significance; (**), at the 5 percent level; and (*), at the 10 percent level.
Inventing – Cook May 2001
35
Appendix
Data: Identifying African American Inventors in Patent Data
As aforementioned, it is very difficult to identify the race of a patentee, since it is not recorded in patent
records, with only one exception since 1790. The first systematic attempt to identify African American
patentees was an effort by the Patent Office, which undertook surveys in 1900 and 1913. The objective of
the surveys was to locate African American patentees whose achievements would be featured in the 1900
Paris World‟s Fair and to commemorate scientific achievements by African Americans in the 50 years
following the end of the Civil War. Directed by one of the lead examiners, Henry E. Baker, surveys were sent
to 9,000 of the approximately 12,000 patent attorneys and agents. Responses to the survey were collected
and analyzed by Baker and published in four volumes, a pamphlet, and an article in the Journal of Negro History.
A subset of the original responses were donated to Carter G. Woodson, a noted historian, and, in turn,
donated by him to the Library of Congress. The Baker data extend from 1834 to 1917. The investigation in
this paper required that the data be extended to 1940.
A first strategy to extend the data set was to include patents obtained in 1913 and beyond by inventors
already in the data set. These data were collected using the European Patent Office (EPO) search engine,
which is searchable by name from 1920. Google Patent Search, which can also conduct historical searches,
became available after 2004, when these data were originally collected. Google Patent Search misses some
historical patents, and the EPO search is more reliable.
One strategy for identifying additional black inventors would be to match patentees from USPTO data to
Census data. This method should work for inventors who are living and patenting in the same place.
However, this procedure fails, because African Americans during this period are not patenting where most
African Americans live, as Table 1 shows. Before 1940, most African American inventors obtained patents in
northern states rather than in southern states. Unlike today, specific addresses were not reported by the
Patent Office, just the city or town in which the inventor resided or from which he or she applied for a
patent. It is difficult to find a unique first- and last-name match using Census data, because of the proximity
of first and last names of African American inventors to those of other, especially British, inventors. Eight
patentees were identified as African American using this method. Only with significant additional
biographical data does this method work, and these data are available for a minority of inventors in the data
set. And if additional biographical data were introduced, the selection problem would be of greater concern,
since biographical information is available for only the most famous and prolific inventors.
Another strategy would be to match common names given to African Americans to patent data. A three-
pronged strategy in the spirit of Fryer and Levitt (2004) and Mullinathan and Bertrand (2004) was executed
but was not successful in identifying black patentees. This mechanism is described below. A second-best
method would be to match known black inventors to names in the patent data. This method was
significantly more successful in producing matches. The second method and its limitations are described in
the text.
Inventing – Cook May 2001
36
An index of black names for the period 1870 to 1940 was constructed from census data in two ways. The
first strategy answered the question: conditional on being black, which names are most likely to be observed?
Random samples of black (“Negro”), “mulatto”, and “colored” heads of households from the 1870, 1900,
and 1920 censuses were drawn for the District of Columbia and three states: Georgia, Michigan, and New
York. From these samples, frequencies were calculated for first and last names separately. There were 14 first
names and 11 last names that appeared more frequently than the median frequency and were included in the
index.
The second strategy answered the question: conditional on observing a certain name, what is the likelihood
that the person is black? First and last names of blacks (“Negroes”) and whites were extracted from the five-
percent IPUMS sample of the 1870 census. Unlike the above samples, names were not restricted to heads of
households. From these samples, frequencies were calculated for first and last names separately and by race
for names occurring at least 80 total times. Among blacks, there were 27 first names and 20 last names that
appeared more frequently than the median frequency for whites or were a larger share of the total names than
the black share of the total population and were included in the index.
A third approach was an extension of the second approach and answered the question: conditional on having
a name widely adopted by African Americans following the end of slavery, what is the probability that the
person is black? This strategy was intended to take advantage of a well-known practice among African
Americans of adopting the first and last names of presidents, e.g., George Washington, or famous people in
the black community, e.g., Booker T. Washington, as first and middle names. The entire 1900 census was
used and also was not restricted to heads of households.
These approaches yielded largely similar results from which an index of “black names” was constructed.
Results were nearly identical with respect to surnames.
Yet the composite index was not able to predict matches in the 1880 census sample of the 690 individuals
identifying their occupation as “inventor.” I was able to predict a small number of black inventors whose
names followed the post-slavery practice of adopting the names of American presidents as first and middle
names. The inventors using this convention were Andrew Jackson Beard, George Washington Carver, and
George Washington Murray. However, most names could not be matched to a patent. The index
significantly under-predicted matches to black inventors and over-predicted matches to white inventors in
New England, particularly those born in England, as was the case with the first Census-based approach.
Additional location and biographical data would have been required to obtain unique first- and last-name
matches. In general, these methods are more suitable for the current rather than historical period.
This highlights a problem associated with occupation identification and reporting among inventors. Many
identify themselves as machinists or artisans or engineers rather than inventors, irrespective of race. Thomas
Edison, among other “great inventors” who are alive and active as inventors, does not appear in the 1880
sample from Ancestry.com.
The final strategy to extend the Baker data set was to construct a broad-based data set of African American
inventors, i.e., potential patentees, and to match the resulting data to patent data. Among the historical and
contemporary sources used to create a pool of potential patentees were searches of historical newspapers,
including obituaries, e.g., from the Ohio Historical Society Newspaper online database and
newspaperarchive.com; correspondence from Carter G. Woodson, Henry E. Baker, and patent survey
participants (Library of Congress); the Garrett Morgan Papers; historical and contemporary directories of
Inventing – Cook May 2001
37
African American medical doctors, scientists, and engineers, e.g., ; academic journals, including the Journal of
Economic History and the Journal of Negro History; historical and contemporary biographies of African American
inventors and general biographies, e.g., Great Negroes Past and Present; and programs of exhibitors in the
African American sections or exhibitions of historical fairs, including the “Exhibit of American Negroes” at
the 1900 Paris World‟s Fair, the 1904 “Great Negro Fair” in Raleigh, North Carolina, and the 1933 Chicago
World‟s Fair “Negro Day”. Newspaper and obituary searches and programs of exhibitions allowed the
identification of lesser-known inventors. A complete list of sources appears in the text and in the list of
references. Not all inventors and others in the pool of potential patentees were matched to patent records
and were dropped from the data set. Others were dropped if there was not a unique first- and last-name
match, e.g., James Young in the patent data. Ultimately, while second best, this process provides a more
systematic and less ad hoc means of recovering black patentees to extend the data set.
top related