Full text of DOI:10.2747/0272-3638.23.2.154

This article was downloaded by: [Selcuk Universitesi]
On: 20 December 2014, At: 17:14
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Urban Geography
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/rurb20
Interlocking Directorates: An Example of Tacit
Knowledge Transfer
Sean O'Hagan a & Milford B. Green b
a Nipissing University
b University of Western Ontario
Published online: 16 May 2013.
To cite this article: Sean O'Hagan & Milford B. Green (2002) Interlocking Directorates: An Example of Tacit Knowledge
Transfer, Urban Geography, 23:2, 154-179, DOI: 10.2747/0272-3638.23.2.154
To link to this article: http://dx.doi.org/10.2747/0272-3638.23.2.154
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

INTERLOCKING DIRECTORATES:
AN EXAMPLE OF TACIT KNOWLEDGE TRANSFER
Sean O’Hagan11
Department of Geography
Nipissing University
North Bay, Ontario
Canada
Milford B. Green
Department of Geography
Social Science Centre
University of Western Ontario
London, Ontario
Canada
Abstract: This paper illustrates tacit knowledge flows between American and Canadian met-
ropolitan areas. Using the spatial distribution of interlocking in Canada and the United States a
spatial component is added to the resource dependency paradigm. Using a poisson regression
model, components of cities that initiate and attract interlocking, and thus knowledge transfer,
can be identified. From the results, the authors propose the concept of a “knowledge threshold”
essential for the transfer of tacit knowledge. In Canada the “knowledge threshold” encompasses
the entire country while it is more regionally based in the United States. [Key words: interlocking
directorates, tacit knowledge transfer.]
Over the past decade, theories of the firm have suggested that the sources of firm pros-
perity have changed. Today knowledge, instead of land, labor, and capital, is the most
viable path for competitive advantage. In this view, knowledge allows the firm to develop
superior technological know-how and create innovative products better than their rivals.
The ultimate challenge to management for sustained growth is to create knowledge that is
not easily imitated by competitors and to use this knowledge for innovative purposes.
Business theorists have, for the sake of convenience, categorized knowledge into two
forms: tacit and explicit. Explicit knowledge can be communicated as words and num-
bers. This makes the composition of explicit knowledge clear and distinct and therefore
easily transmitted. Its codified nature allows explicit knowledge to be expressed as hard
data, blueprints, computer code, and scientific formulas. These are freely processed by
computer and transmitted electronically. The transmission of explicit knowledge is often
carried out by impersonal means, such as through a computer or a technical manual.
1Correspondence concerning this article should be addressed to Sean O’Hagan, Department of Geography,
Nipissing University, North Bay, Ontario, Canada P1B 8L7; telephone: 705-474-3461; fax: 705-474-1947;
e-mail: seano@nipissingu.ca
154
Urban Geography, 2002, 23, 2, pp. 154–179.
Copyright © 2002 by V. H. Winston & Son, Inc. All rights reserved.

TACIT KNOWLEDGE TRANSFER
155
Tacit knowledge on the other hand is highly subjective, less teachable and less observ-
able. Its ambiguous nature makes it difficult to process and communicate. With few
exceptions (video conferencing and selected telephone conversations), tacit knowledge
can only be learned through face-to-face communication. Nonaka and Takeuchi (1995)
pointed out that subjective insights, intuitions, and hunches are examples of tacit knowl-
edge. While technological developments warrant the effortless movement of explicit
knowledge, distance constrains tacit knowledge transfer, resulting in two important ram-
ifications. First, the codified nature of explicit knowledge allows it to be imitated rela-
tively easily by competitors. The firm must therefore use tacit knowledge if sustainable
competitive advantage is to be created. Second, geography is important for tacit knowl-
edge transfer. Since face-to-face contact is necessary for tacit knowledge transfer, local-
ized relationships are essential. Therefore, when attempting to explain the competitive
advantage of firms, not only should the transfer of tacit knowledge be studied, but the
geographical implications for tacit knowledge should be studied as well.
For long-term prosperity the firm must create new knowledge. In order to do this,
external tacit knowledge must be introduced to the firm. One of the most important ways
for the firm to access external tacit knowledge is through its board of directors (Useem,
1984). Considered informed and accomplished professionals, directors provide indepen-
dent knowledge on the external economic climate. Being removed from everyday activi-
ties of the firm, directors offer fresh ideas for firm strategy. Perhaps the most important
way a board member can obtain knowledge is to serve as a board member of another firm.
Termed an interlocking directorate, sitting on two or more boards allows an individual to
pass knowledge of one firm to another.
Past research on interlocking has been dominated by sociologists who have attempted
to explain “why” these connections take place. From this literature, resource dependency
has emerged as the dominant explanation for interlocking directorates (Pfeffer and Salan-
cik, 1978). This theory suggests that firms can internally produce only a portion of their
inputs. Consequently, they must enter into relationships to obtain all the other required
resources. Perhaps the most important resource today is knowledge. By connecting to
external sources, knowledge of the environment is gained and uncertainty decreased. One
way the firm can link to the environment is through its board of directors.
But why does resource dependency research limit research to other firms? Why not
extend the model to include additional sources where important knowledge can be
obtained? The aim of this paper is to follow the principles of resource dependency but
develop a spatial component. It is important to build upon the original resource depen-
dency model because a number of factors influencing knowledge transfer do not exist at
the firm level. Using metropolitan-level data, this study denotes these factors by analyz-
ing the spatial distribution of interlocking in Canada and the United States. Using a pois-
son regression model, components of a metropolitan area that initiate and attract
interlocking, and thus knowledge transfer, can be identified.
KNOWLEDGE, THE FIRM, AND THE BOARD OF DIRECTORS
Today, knowledge is considered the most strategically important resource. In fact,
Drucker (1993) argued that knowledge is the only meaningful resource. Continuous gen-
eration of new and superior knowledge enhances a company’s competitive position. A

156
O’HAGAN AND GREEN
firm maintaining superior knowledge understands how to capitalize on their traditional
resources to create new products. This constant creation of new knowledge and products
increases the resistance to imitation from rival firms.
Knowledge sources may lie within or outside the firm. Internal knowledge is strategi-
cally valuable because of its uniqueness to that company. Being firm specific makes it
more difficult for competitors to imitate. To continually innovate though, new knowledge
must be acquired. External sources are ideal for procuring such new knowledge. When
the external and internal knowledges are combined, new ideas and concepts translate into
new processes and innovative products.
There are several ways this can be accomplished. For example, Japanese firms are
quite successful in acquiring this external knowledge. The Japanese keiretsu, known as
societies of business, facilitate cooperation among a set of firms for knowledge transfer to
occur. Regulated by Japanese law, keiretsu are structured in a way that collaboration is
essentially required. A typical network consists of a few dozen members including a bank
or large financial institution, large manufacturing firms and general trading companies.
Within each group, members hold one another’s shares, engage in intragroup financing,
have interlocking directorates, maintain trade, and joint research and development ven-
tures. The end result is a highly innovative network (Branstetter, 2000).
In Canada and the United States, the business network operates differently. Alliances,
foreign direct investment, acquisitions, and mergers are official arrangements that sanc-
tion knowledge transfer between North American firms. Links to suppliers and buyers,
interlocking directorates, start-ups by individuals belonging to another local firm, staff
movement, and rumors are a few examples of less stringent but very important knowledge
networks. The aim of this paper will be to use interlocking directorates to measure knowl-
edge transfer. Since most employees deal with internal questions, many firms appoint a
board of directors which has both internal and external expertise. One of the fundamental
responsibilities of these directors is to determine what external knowledge should be inte-
grated with existing resources to create innovative processes and products. As pointed out
by Charon (1998), the board is the best platform for creative thinking about firm strategy.
As stated in the introduction, a distinction is often made between different types of
knowledge, explicit and tacit. Nonaka and Takeuchi (1995) argued that the two types of
knowledge are not mutually exclusive activities. In fact, explicit and tacit knowledge
compliment each other. Social interaction between tacit and explicit knowledge actually
converts old knowledge into new knowledge.
To illustrate this relationship Nonaka and Takeuchi (1995) presented a dynamic model
of knowledge creation. They offer four modes of knowledge conversion that essentially
make up the knowledge creation process: (1) from tacit to tacit, called socialization; (2)
from tacit to explicit, externalization; (3) from explicit to explicit, combination; and (4)
from explicit to tacit, called internalization. This paper is primarily concerned with the
first two stages of the model of knowledge conversion. In other words, this paper deals
with the avenue whereby external tacit knowledge can be introduced to the firm for the
purpose of internalizing, thus creating new knowledge.
Nonaka and Takeuchi (1995) termed the interaction between the four modes of knowl-
edge as a knowledge spiral. As suggested by Teece (1998), how a firm configures this
knowledge spiral determines firm competitiveness. Today, the competitive nature of the

TACIT KNOWLEDGE TRANSFER
157
firm stems not from owning tangible assets, but from the ability to create knowledge
through this knowledge spiral (O’Hagan and Green, 2001).
Unfortunately it is difficult to visually display the socialization and externalization
phases of the model of knowledge conversion. The difficulty of researching the transfer
of tacit knowledge lies in its intangibility. How does one measure the quality and quantity
of tacit knowledge transferred from one firm to another? The characteristics of tacit
knowledge are precisely what make it difficult to measure. Clearly it has not been easy to
empirically measure tacit knowledge transfer. Glazer (1998, p. 177) verified this limita-
tion by suggesting that if our understanding of knowledge is to increase, “it is crucial to
develop valid and reliable measures of knowledge. For, despite the flurry of activity, it is
becoming clear that no real progress can be made in our efforts to treat knowledge either
as a variable to be researched or as an asset to be managed unless we come to terms with
the issue of measurement/valuation.”
Here we face a theoretical dilemma. It is important that research increasingly recog-
nize the importance of firm knowledge for competitive advantage. However, it is prob-
lematic for ideas of knowledge transfer to disregard empirical validation. The solution to
measure the flow of knowledge is to take actors that possess a great deal of tacit knowl-
edge and a medium that allows for this tacit knowledge to be transferred. We argue that
directors are important actors of the firm that possess a great deal of quality tacit knowl-
edge. They understand company goals and company strategy to achieve these goals. Inter-
locking directorates are a means by which these directors can pass tacit knowledge on
from one firm to another. In other words, the knowledge transferred from one company to
another company via an interlock is in the form of tacit knowledge.
It is important to study the concept of tacit knowledge transfer because what the com-
pany does with this new knowledge can help lead to its prosperity. As recognized by
Kuznets (1966) overall performance of a company depends on its ability to turn its knowl-
edge resources into new knowledge. This new knowledge creates value and allows the
firm to continually innovate. We propose that directors play an important role in this pro-
cess because they introduce external knowledge to the firm.
Returning to Nonaka and Takeuchi’s (1995) concept of the knowledge spiral, it is pos-
sible to display the significance of the knowledge that directors procure through knowl-
edge links. Serving on the board of a second firm through an interlocking directorate
allows a director to gain tacit knowledge from the other directors sitting on that board
(socialization). The knowledge gained from his/her experiences with the second firm can
then be brought back to the first firm where eventually it will be displayed for employees
in an explicit manner (externalization). This explicit knowledge can be combined with old
explicit knowledge already possessed by the firm to create new explicit knowledge. This
new knowledge can then go through the process of internalization where the workers of a
company turn the explicit knowledge back into a new form of tacit knowledge for further
knowledge creation.
GEOGRAPHICAL RESEARCH ON KNOWLEDGE
Most economic geography research over the past ten years discusses knowledge and
information as it relates to a locale. This research dates back to Marshall (1920) who
suggested firms in spatial agglomerations benefit from technological spillovers. Ever

158
O’HAGAN AND GREEN
since, academics have acknowledged the region as an important source of intangible
assets necessary for competitive advantage.
Malmberg (1996), Maskell and Malmberg (1999), and Morgan (1997) claimed that
physically proximate firms interact to create innovative processes and products. They
termed these areas dynamic regions, and suggested they are an important source of new
knowledge. Areas of specialization with their high levels of direct face-to-face relations
support a commitment toward collective learning and shared prosperity. As they relate to
one another, trust develops between individuals that encourages the diffusion of technical
knowledge and skills. In particular, Morgan (1997) suggested that feedback loops are an
important source of knowledge. A lack of feedback indicates activities like research and
development have little chance to learn from past experiences.
There is no doubt that regional interaction is important for the transfer of tacit knowl-
edge. But to consider local interdependencies as always superior is inadequate. For exam-
ple, even in the entertainment business, not all relevant firms and holders of knowledge
are located in Hollywood. Oinas and Malecki (1999) pointed out that relationships with
firms in other regions contribute unique experiences and provide a basis to contrast local
ideas. They called the combination of sustained local and nonlocal relationships a spatial
innovation system.
The relationships beyond the immediate area are less apparent as sources for knowl-
edge (Amin and Cohendet, 1999). What matters most are not the number of relationships,
but the quality of these relationships. For example, individuals adjacent to each other
protecting their knowledge may actually prevent knowledge transfer and innovation. In
contrast, the likelihood of knowledge transfer increases in these spatial innovation sys-
tems where individuals trust one another, no matter what the spatial separation.
Interlocking directorates provide a unique setting for studying the sharing of knowl-
edge. Over time, relationships are built up not only with other directors in the same urban
area, but also with directors in distant locations. While local relationships are important,
it needs to be increasingly emphasized that nonlocal connections also play a significant
role in sustaining competitiveness. These provide additional sources of knowledge about
producers, consumers, and regulators.
As suggested by Holtshouse (1998) the key to the creation of new knowledge involves
“tapping into and unlocking” individual tacit knowledge. This requires communication
between individuals in a face-to-face work environment. The face-to-face contact is vital
because the transfer of tacit knowledge is procured, for example, through firsthand obser-
vation and body language. Interlocking provides a forum through which tacit knowledge
can be transferred from one company to another no matter what the distance between
firms. Indeed, it seems clear that interlocking to local or distant regions is a means of
achieving competitive advantage through knowledge transfer.
PREVIOUS RESEARCH ON INTERLOCKING DIRECTORATES
A substantial amount of literature exists on the study of interlocking directorates.
From this literature, resource dependency has emerged as the dominant theory to explain
why interlocking directorates occur. This theory states that firms are unable to generate
all the necessary inputs for production. Firms respond with relationships to external ele-
ments of the environment to ensure a supply of the required resources. Organizations are

TACIT KNOWLEDGE TRANSFER
159
Fig. 1. Relationships among dimensions of organizational environments. Adapted from Pfeffer and Salan-
cik, 1978.
viewed as actively manipulating and controlling their environment to secure a steady sup-
ply of inputs. For the most part, firms increase control of their environment through rela-
tionships with other firms. One way to establish a relationship with another firm is
through the board of directors.
Resource dependency asserts that the board is an integral component of the effective
firm. In this model, the board is used to gain access to scarce resources and knowledge
(Boyd, 1990). If a firm is connected to boards of several other firms, it is in the center of
a network and has access to a valuable resource: knowledge. Obtaining knowledge is
difficult because of its embedded nature in the firm and in the people of the firm
(Badaracco, 1991). But a director sitting on the board of two firms allows knowledge
about corporate policy to be transferred. Pennings (1980) and Solvell and Zander (1995)
argued that an interlocking directorate provides a link by which knowledge can be com-
municated between two firms. Lorsch and MacIver (1989, p. 27) supported this argument
with firsthand evidence. Citing a chief executive officer the authors suggested,“You can
learn so much about situations that you may become faced with.”
But why is knowledge conveyed through interlocks so significant? Haunschild and
Beckman (1998) suggested that reasons include the inexpensive, trustworthy, credible
nature of directors. All publicly traded firms in the U.S. are required to have a board of
directors of at least three people. In Canada these same firms are required to have a board
of directors with at least one person. Therefore, the knowledge that is gained from a direc-
tor that sits on the board of another company is an inexpensive by-product of this manda-
tory regulation.
Limitations of the Resource Dependency Model
The following section is divided into five parts. First, the resource dependency model
proposed by Pfeffer and Salancik (1978) is further analyzed. Second, shortcomings of the
resource dependency model are raised. Third, modifications to the resource dependency
model are proposed. Fourth, research that explores the geographical implications of inter-

160
O’HAGAN AND GREEN
locking is summarized. Finally, improvements to both the resource dependency model
and geographical research on interlocking are presented.
A three-tiered model of resource dependency proposed by Pfeffer and Salancik (1978)
is provided in Figure 1. From this model, they display how structural characteristics of
environments are associated with relationships among social actors. The combination of
these two dimensions results in uncertainty. Interestingly, they suggested the dimensions
of a firm’s environment are not mutually exclusive. Moreover, they argue a causal rela-
tionship exists between these three dimensions.
If we assume the relationship between these three dimensions is causal, then a number
of factors influence the level of uncertainty possessed by the firm. According to resource
dependency, the level of uncertainty possessed by the firm impacts the performance of a
firm. Using interlocking directorates as a means of decreasing uncertainty, a great deal of
literature documents the relationship between interlocking and firm prosperity.
Unfortunately, these studies do not offer a distinct definition of a firm’s environment.
Herein lies a problem with the resource dependency approach. For example, Pfeffer
(1972) observed that the number of attorneys on a board increased as the level of regula-
tion increased. In other words, the level of uncertainty for a firm’s environment centered
on regulation. In another study, Pennings (1980) found that profitability increased with
ties to financial institutions. Thus, the environment for firms in Penning’s (1980) study
surrounded financial uncertainty.
Defining the firm’s environment in different ways has led to conflicting results. Some
have found positive effects of interlocking on firm profits, albeit weak relationships (Pen-
nings, 1980; Burt, 1983). As interlocking increases, a firm becomes more aware of its
external environment. By increasing information, it decreases uncertainty and is more
likely to make sound business decisions. On the other hand, others have found negative
effects (Fligstein and Brantley, 1992).
In the past, these results have been justified in two ways. First, firm prosperity is a
complex process. Interlocks and the resources they secure are only a portion of what
makes a firm successful. Second, Mizruchi (1996) suggested that interlocking may be
both a cause and effect of profitability. For instance, unprofitable firms may be more
likely to interlock. Richardson (1987) found that bankers are more likely to join a board
of directors when that company is in financial difficulty. Thus, it is precisely when profits
are lowest, that interlocking occurs.
In response, we as geographers suggest that asking the question, does interlocking
increase the profitability of a firm is to answer only one question. An alternative question
is where do interlocking directorates occur. From this one can suggest reasons why inter-
locking directorates provide a basis for knowledge transfer. This type of study does not
examine the position of the firm at the time of the interlock or how management of the
company uses this knowledge. It can only be assumed that increased knowledge intro-
duced by the director enables superior intelligence that decreases firm uncertainty and
leads to better business decisions.
While resource dependency research argues that knowledge varies across firms, it does
not suggest that knowledge varies across space. This lack of a spatial component is a
deficiency of the resource dependency and interlocking literature. Lundvall and Johnson
(1994) indicated how learning is not a placeless process. Sassen (1991) added to this
argument by suggesting that an important attribute of knowledge is its irregularity over

TACIT KNOWLEDGE TRANSFER
161
space. It exists in certain places but not in others. Consequently, geographical differences
should emerge as an important rationale when attempting to explain the acquisition of
knowledge in resource dependency and interlocking directorates.
This raises the question, how does one define the firm’s environment? Put another way,
how does one define the three dimensions of Pfeffer and Salancik’s model. As argued by
Boyd (1990, p. 421), “a comprehensive definition of environmental characteristics is
essential if we are to conduct an accurate test of resource dependence.” Since knowledge
is an increasingly important component of a firm’s environment, and knowledge varies
over space, any meaningful explanation of resource dependence must encompass loca-
tional factors that affect the flow of knowledge.
Perhaps a better explanation of the dimensions of an organization’s environment
would follow Pfeffer and Salancik’s model but include the individual, firm, and metropol-
itan area. At the base, uncertainty is heightened and can only be reduced through knowl-
edge transfer from one individual to another. A greater chance of knowledge transfer
occurs when the individual is employed by a firm that is interdependent with a number of
other firms or where conflict between the firm he/she works for and additional firms is
diminished. The quality and quantity of knowledge transfer to these firms should be
dependent upon the extent to which knowledge is distributed over the environment (con-
centration), the relative level of knowledge available in the environment (munificence),
and the degree of interconnection among elements in the environment (interconnection).
While Pfeffer and Salancik claimed that the firm should be the main unit of analysis of
the third tier, the authors of this paper contend that the metropolitan area is also important.
Dess and Beard (1984, p. 52) suggested that “the resources required for organizational
survival are the most relevant focus in defining organizational environments.” From this,
the importance of other firms in an organization’s environment is obvious. Perhaps less
obvious though, are other sources of knowledge when defining a firm’s environment.
Firms cannot maximize their knowledge by only linking to other firms. Additional
sources are crucial if the firm is to obtain knowledge. Surely, the quantity and quality of
tacit knowledge in one metropolitan area will be somewhat different from the knowledge
that exists in another metropolitan area. Social, economic, and political institutions will
vary from metropolitan area to metropolitan area. Media, political figures, and research
institutions are just a few examples of knowledge sources that exist at the metropolitan
area level but not at the firm level.
Even Pfeffer and Salancik (1978, p. 12) suggested, “the environment includes every
event in the world which has any effect on the activities or outcomes of the organization.”
Therefore, a locational explanation is just as meaningful for the third level of the hierar-
chy. This paper addresses the limitations of Pfeffer and Salancik’s (1978) definition of a
firm’s environment by considering the concentration and munificence of knowledge at
the metropolitan-area level. In addition, how connected a metropolitan area is to another
should influence the flow of knowledge between these centers. This helps us determine
whether interlocking directorates provide the basis for knowledge transfer.
The Geography of Interlocking Directorates
There are only three studies that conceptualize the spatiality of knowledge transfer
through interlocking directorates. Green (1980) and Green and Semple (1981) were the

162
O’HAGAN AND GREEN
first researchers to examine the geographic conceptualization of interlocking directorates.
By examining the interlocking of 183 American firms spatially, the authors observed that
firms in cities of the U.S. manufacturing belt preferred to interlock with firms in cities of
the same region. From their findings the authors proposed that this regional internaliza-
tion denied firms of access to useful information that existed in other regions of the
United States. This study was a start to the spatiality of knowledge as it relates to inter-
locking. Unfortunately the theoretical foundation is outdated and techniques at the time
did not allow for the analysis of a large number of firms.
Geographers are not the only researchers to acknowledge the importance of space for
the study of interlocking directorates. Kono et al. (1998) argued that sociological research
on interlocking directorates tends to ignore the power of the city, focusing instead on
individual leaders and firms. The ability of large urban areas to provide a diversity of
functions in the same area should be an important contributing factor in the study of inter-
locking directorates.
The administrative headquarters of large corporations are drawn to large metropolitan
centers to enjoy the advantages of proximity to financial establishments on which they
depend, the specialized business services that such cities offer and the information pro-
cessing industries that congregate in the city. Top managers can acquire strategic informa-
tion informally from elites of other firms headquartered in its city by living in the same
neighborhoods, sitting on the same corporate boards, and belonging to the same social
clubs. Thus cities that have many business service firms, financial institutions, and corpo-
rate headquarters should be the best sites for the acquisition of strategic knowledge.
Using a “city as a growth machine” model, Kono et al. (1998) applied sociological
research on interlocking directorates to a geographical scale. For the top 500 American
firms in 1964, the authors suggest that some types of interlocks are better suited for local
links while others are better suited for nonlocal links. Kono et al. (1998, p. 895) argued
that if they did not separate the interlocks geographically, they would have “erroneously
concluded that hypothesized determinants had no influence on interlocking at all.” For
example, when local and nonlocal interlocks were aggregated into the same sample, the
independent variable “presence of exclusive upper-class social clubs in a corporation’s
headquarters city” did not influence corporate interlocking.
However, when the sample was broken down, it was found that the independent vari-
able “presence of exclusive upper-class social clubs in a corporation’s headquarters city”
did influence local and nonlocal interlocks. When corporations were headquartered in
cities with exclusive upper class social clubs, they were more likely to maintain local
interlocks. On the other hand, corporations with headquarters in cities without exclusive
upper class social clubs were more likely to maintain nonlocal interlocks.
By examining the problem spatially, the authors were able to prove that these finding
supported the class hegemony model. However, their theoretical contribution also appears
to support the resource dependency model.Yet they failed to investigate this. In addition,
Kono et al. (1998) examined sociological variables as either intracity or nonlocal. In other
words, a nonlocal interlock of 40 km was treated the same as an interlock of 4000 km.
While resource dependency advocates (Pfeffer and Salancik, 1978) sufficiently
examined interlocks as a method of knowledge transfer between firms, they failed to
consider the spatiality of knowledge. Conversely, geographers (Green, 1980; Green and
Semple, 1981; Green, 1983; Rice and Semple, 1993) have recognized the power of the

TACIT KNOWLEDGE TRANSFER
163
city as it relates to interlocks. But geographers have failed to point out the spatiality of
tacit knowledge transfer as it relates to interlocks. The present study acknowledges the
strengths and weaknesses of both of these divergent ideals to measure interlocking as a
means of knowledge transfer between cities. Using interlocking data on directors of
firms, the firms where these directors work, and the cities where these firms are located,
this paper unveils where tacit knowledge flows across Canada and the United States and
why it flows there.
DATA
Data used in this study include all direct interlocks of the 1996 Fortune 500. A direct
interlock occurs when one company’s director also sits on the board of a different firm.
The board of directors was obtained from Standard and Poor’s Register of Corporations,
Directors and Executives. For Canadian interlocking, the top 200 industrial firms and top
50 financial institutions were selected from The Financial Post 500 (1997). The director-
ships of each of these 250 companies were then retrieved from the publication Directory
of Directors.
While there are thousands of firms with boards of directors in Canada and the United
States, only the largest 750 firms were selected for study. It is argued that interlocking is
most important to these firms for two reasons. It seems reasonable that small firms will
not expand their market reach to the extent of large firms. Therefore, it is not as important
for smaller firms to access knowledge that exists in distant cities. Second, large firms have
a greater number of employees and the necessary resources to obtain superior knowledge
on the local environment.
Using the above sources three samples were constructed. Sample USA-CAN, consist-
ing of both Canadian and American interlocking, produces a 750 × 750 adjacency
matrix. Sample USA includes only American interlocking to other American firms, gen-
erating a 500 × 500 adjacency matrix. Finally, Sample CANADA is composed of domes-
tic Canadian interlocks to create an adjacency matrix that is 250 × 250.
DISTRIBUTION OF INTERLOCKS IN CANADA AND THE UNITED STATES
Of the top 500 American firms 90% had at least one direct link. Of the top 250 Cana-
dian firms 82% also had at least one direct link. In total, the 8,909 directors produced a
web of 7,512 interlocks, an average of 10 interlocks per company. It was hypothesized
that firm size should be strongly related to number of interlocks. But a statistical test did
not verify this. Using revenue as a proxy for firm size produced an r2 = .20 (assets as a
proxy for firm size produced an r2 = .17). This may be due to the fact that we are using a
censored sample of the largest companies in Canada and the United States. A list of dom-
inant firms in the network of interlocks is supplied in Table 1. Together these 25 compa-
nies generate 21% of all the corporate interlocks. Most connected were Royal Bank of
Canada (62 interlocks), Canadian Imperial Bank of Commerce, (55 interlocks) and Chase
Manhattan (45 interlocks). From this table, the importance of finance and insurance com-
panies as well as New York City (six of the top 25 companies are within commuting
distance to New York City) and Toronto (nine of the top 25 companies) is impressive.

164
O’HAGAN AND GREEN
TABLE 1.—TOP AMERICAN AND CANADIAN INTERLOCKING COMPANIES, 1996
Company
Headquarters location
Direct interlocks
Royal Bank of Canada
Canadian Imperial Bank of Commerce
Chase Manhattan
Toronto
62
55
45
44
44
43
43
39
39
38
38
37
36
36
35
35
34
34
33
33
33
33
33
33
31
Toronto
New York City
Montreal
Calgary
Canadian Pacific
NOVA
Novagas Clearinghouse
Noranda Mines Ltd.
Morgan Stanley
Calgary
Toronto
New York City
New York City
Chicago
Dean Witter Group
Sara Lee
Brascan Ltd.
Toronto
Avenor Inc.
Montreal
Toronto
Toronto-Dominion Bank
Manulife Financial
CT Financial Services
Power Corporation
Bank of Nova Scotia
Union Carbide
Toronto
Toronto
Montreal
Toronto
Danbury
Chicago
Ameritech
The Great-West Life Assurance
Sun Life Assurance
Bell Canada Enterprises
Prudential
Winnipeg
Toronto
Montreal
Newark, NJ
Calgary
Norcen Energy Resources
Metropolitan Life Insurance
New York City
It is also hypothesized that certain economic sectors possess a greater than average
number of interlocks. In an industry where continual innovation of products and pro-
cesses is important, interlocks should result in greater rewards. For example, the commu-
nications industry is more likely to benefit from the knowledge transfer through
interlocks than wholesale trade.
In order to test this hypothesis, the Standard Industrial Classification Codes (SICs) for
all 750 firms were obtained. Canadian firm SICs were acquired from Canadian Key Busi-
ness Directory (1997) and American firm SICs were obtained from D&B Million Dollar
Directory: America’s Leading Public and Private Companies (1997). Table 2 is an amal-
gamation of firms into selected industries. Here the importance of finance, communica-
tions, as well as mining is unmistakable. Previously, theorists would suggest that this
disproportionate representation of banks could be explained by the need for large

TACIT KNOWLEDGE TRANSFER
165
TABLE 2.—Interlocking of American and Canadian Industries, 1996
Industry
Linkages
Linkages per company
Agriculture products and services
Mining
13
476
30
6.50
14.42
4.29
Construction
Food and kindred products
Chemicals and allied products
Petroleum refining
421
468
235
337
349
457
790
283
317
421
308
481
1072
879
175
11.38
10.88
13.82
9.11
Industry and commercial machinery
Transportation equipment
Paper, printing, and publishing
Other manufacturing
Transportation
12.03
12.03
10.26
11.32
15.85
8.77
Communications
Electric, gas, and sanitation services
Wholesale trade
5.60
Retail trade
6.59
Finance
15.31
11.88
2.78
Insurance and real estate
Services
amounts of short-term capital. The concern for quick access to capital is diminished with
a connection to a large bank. But the authors of this paper suggest that interlocks with
banks can be explained by the fact that industrial corporations now require knowledge
capital. In other words, the prosperity of a firm is no longer asset based, but knowledge
based. Banks are an excellent source of knowledge on the general economic climate.
Communications and mining are extremely technical, resulting in a disproportionate
amount of interlocking in these industries. The communications industry is highly com-
petitive with increasingly shorter product lifecycles. On the other hand, mining firms are
always searching for superior techniques for extracting resources. Knowledge through
interlocks allows firms to stay in tune with essential developments in both of these indus-
tries. This provides companies with new knowledge necessary for competitive advantage.
Figure 2, a simplified coded table, explores this further through interindustry linkages.
The four categories are separated into hinges. Hinges are similar to separating data into
quartiles except that the median is used instead of the mean. Hinges are a synopsis of
values in the middle of each half of a data set (Velleman and Hoaglin, 1981). It was deter-
mined that the median should be used because the linkages are not normally distributed
across industries.
Figure 2 reveals how firms from almost all industries are extensively linked to finance
as well as insurance and real estate. This is because banks, insurance companies, and real
estate firms possess extensive knowledge on the general economic climate. On the other
hand, we know from Table 2 that mining companies possess abundant linkages. But these

166
O’HAGAN AND GREEN
Fig. 2. Simplified coded table of American and Canadian industry interlocking, 1996.
linkages are concentrated within the mining sector. Mining companies would largely ben-
efit from knowledge linkages to other firms in their industry. This will allow them to
procure superior techniques for extracting resources.
As evident from Figures 3A and 3B, Eastern Canada and the Northeastern United
States are the dominant locations for interlocking. Toronto is the most important with
1,090 total interlocks (13.63 per firm). NewYork City produces the next largest number
of interlocks with a total of 815. However, New York City firms exhibit more interlocks
per firm with 15.67 per firm. With a few exceptions, a hierarchical pattern emerges in
Figure 3A. The top 20 cities account for over 71% of interlocking in Canada and the
United States.
This map supports the notion that interlocks are initiated by and attracted to cities that
are the most important for conducting business at the regional, national and global levels.
As pointed out by Sassen (1991), the most important cities now function in four new
ways. First, cities act as highly concentrated control points. Second, the most important
cities are locations for finance and for specialized service firms. These industries have
displaced manufacturing as the leading economic sectors. Third, these cities are sites of
innovation. Finally, the most important cities are markets for the products and innovations
produced.

TACIT KNOWLEDGE TRANSFER
167

168
O’HAGAN AND GREEN

TACIT KNOWLEDGE TRANSFER
169
The top eight top cities in Figure 3A generally adhere to these four principles with
respect to their domestic economies. In the United States, New York, Chicago, and Los
Angeles are generally the three most important cities with respect to Sassen’s four princi-
ples. The same holds true for the Canadian cities of Toronto, Montreal, and Vancouver.
The only exception in the United States, Atlanta, may be explained by its dominant
position in southeastern United States. Figure 3B visually displays this importance. The
data further verify the importance of Atlanta, which possesses 162 interlocks. The next
most important metropolitan area in the U.S. Southeast is Richmond with 76 interlocks.
In Canada, Calgary is an exception because so many companies want to tap into the
knowledge in the oil industry that exists in that city. Of the 419 linkages with Calgary,
76% are associated with firms in the oil and gas industry. Following Malmberg (1996)
and Morgan (1997), firms want to link to this metropolitan area because it is a crucial for
new knowledge in this industry.
Figure 3B reveals regional hubs in the knowledge network. The importance of Atlanta,
Houston, Los Angeles, San Francisco, Pittsburgh, Chicago, New York, Toronto, and
Montreal is evident. This suggests regional metropolitan areas are fundamental to the
transfer of tacit knowledge in the Canadian and American network.
SPATIAL INTERACTION MODEL
The aim of this paper is to explain how concentration and munificence of knowledge
in cities and the interconnectedness of these cities influence interlocking directorates. By
using a spatial interaction model, the paper seeks to answer two questions. First, what
influence does distance have on interlocking between two cities? Second, what character-
istics of cities promote the flow of knowledge between them?
The independent variables are tested on the three models previously mentioned where
the dependent variable is the number of interlocking directorates between one metropoli-
tan area and another. With interlocks focused among a small number of cities, a standard
linear regression is not appropriate. The bulk of cities in the sample have a small number
of interlocking directorates. In fact of the 7,512 interlocks that take place, one-third of the
cities have less than 10 interlocks. Consequently, a combination of cities with numerous
interlocks (such as the 165 interlocks between Toronto and Montreal) is rare. As a result,
the dependent variables for Sample USA-CAN, Sample USA, and Sample CANADA
possess a distribution that is highly skewed. This makes a poisson regression approach
more applicable than a least squares approach (see Liao, 1994, pp. 70-79).
INDEPENDENT VARIABLES
Population
Large cities possess more social, economic, and political resources than smaller cities.
This translates into a greater amount of existing knowledge. The snowballing of informa-
tion and knowledge in larger cities procures more knowledge and thus results in more
knowledge links. Firms want to associate with other firms in such centers where knowl-
edge already exists. Following Pred’s (1977) principle of circular cumulative causation, it
seems logical that population and knowledge do not have a direct linear relationship. Fig-

170
O’HAGAN AND GREEN
Fig. 4. Patents versus metropolitan population for the United States.Sources: U.S. Patent Office (1997) and
U.S. Bureau of the Census (1996).
ure 4 reveals the relationship between patents (used here as a proxy for knowledge) and
the square root of metropolitan area population for the United States in 1996. Testing
produced a Pearson’s correlation coefficient of r2 = .52. From this, it is hypothesized that
cities with large populations should be the focus of more interlocking activity than cities
with small populations and this interlocking should expand at an increasing rate.
Headquarters
Similar to population, cities that contain a greater number of headquarters should be
the best sites for the acquisition of strategic knowledge. When attempting to explain the
increase in Japanese competitiveness during the 1970s and 1980s, Ohmae (1995) sug-
gested that only a few firms guided by strong leaders were responsible for this incredible
achievement. These individuals were an important component of a management team
located at the firm’s headquarters. Since firms want to tap into the knowledge that these
people possess, cities with a greater number of headquarters should experience more
interlocking. Of course, headquarters are highly correlated to population. But the two
variables do display somewhat different patterns and therefore produce different results.
The variable headquarters were obtained from Corporate Affiliations Plus (1997). In this
data set, 717 firms were headquartered in Canadian cities and 12,641 were headquartered
in American cities.
Metropolitan Area Growth
An economically expanding metropolitan area maintains the necessary climate for
growth. Within this milieu firms and individuals either possess or learn the knowledge

TACIT KNOWLEDGE TRANSFER
171
essential to remain competitive today. External firms need to access this knowledge to
determine what makes internal firms successful in this economic environment. To maxi-
mize profit (which is assumed to be the desire of all companies), firms must continually
acquire the most recent techniques and processes. These techniques most likely exist in
cities experiencing the highest rates of development. Metropolitan area growth was cal-
culated as the population change from 1991 to 1996.
Industry Concentration
Michael Porter (1990) suggested that industries should try to obtain the knowledge
that spills over between firms. He argued that local firms in the same industries should
expand at a faster rate because geographically concentrated companies can learn from
each other better than isolated ones. For example, it is certainly beneficial for a software
development company to link to the knowledge that exists in Silicon Valley.
But today more than ever, firms use input from and produce output for diverse indus-
tries and locations. A metropolitan area where employment is distributed evenly over a
number of sectors should contain knowledge in a wide array of industries. Jacobs (1969)
and Glaeser et al. (1992) wrote that the most important knowledge transfers come from
outside the core industry. As a result, neighboring firms producing a variety of products
encourages innovation and growth. For example, if General Motors were to exclusively
link to sources of knowledge in Detroit, the firm could fail to recognize shifting trends
within the industry as it relates to suppliers and buyers.
The model needs to ascertain whether interlocks adhere to the theory proposed by Por-
ter (1990) or the theory proposed by Jacobs (1969) and Glaeser (1992). This can be
accomplished by using the coefficient of variation, which determines the relative variabil-
ity between data distributions. Using industrial sector employment from U.S. Bureau of
the Census and the Statistical Profile on Canadian Cities, the coefficient of variation was
used to calculate the concentration of a metropolitan area’s economy in particular indus-
tries. A greater coefficient of variation suggests greater concentration of employment in
one industry. It is hypothesized that most firms in today’s global economy cannot limit
acquisition of tacit knowledge to certain areas possessing specific knowledge. It is impor-
tant for firms to access a wide range of knowledge to remain competitive. Therefore,
interlocks should be inversely related to industry concentration.
Research Institution Presence
Universities play an important role in the accumulation of knowledge at the metropol-
itan area level. They potentially contribute knowledge in two ways. First, they are a criti-
cal source of knowledge and learning activities. Second, they provide essential
knowledge infrastructure. Since possessing both the physical and intellectual capital is
tremendously beneficial it is hypothesized that interlocking will have a positive relation-
ship with universities.
Based upon a number of components, U.S. News and World Report (1997) ranks
American universities from Tier 1 to Tier 4. Tier 1 universities yielded the highest weight
while Tier 4 universities yielded the least weight. Canadian universities are ranked by
Maclean’s magazine (1997). They are separated into three categories, Doctoral, Compre-

172
O’HAGAN AND GREEN
hensive, and Primarily Undergraduate. Universities with doctoral programs were given a
rank similar to Tier 1 universities in the United States, Comprehensive a rank similar to
Tier 2 firms, and Primarily Undergraduate a rank similar to Tier 3 universities. All univer-
sities and their weights were then amalgamated for a metropolitan area total.
Spatial Separation
From previous research it appears that interlocks have spatial attributes and spatial
determinants. Saxenian (1994, p. 161) suggested, “geographic proximity promotes the
repeated interaction and mutual trust needed to sustain collaboration.” As a relationship
develops, individuals build a basis for sustained trust that is conducive to the diffusion of
knowledge. The hypothesis of this study suggests that the degree of knowledge transfer
will decrease as distance increases. Spatial separation was calculated as the straight-line
distance from metropolitan-area headquarters of one firm to the metropolitan-area head-
quarters of the second firm involved in the interlocking.
Finance, Insurance, and Real Estate (FIRE) Employment
An alternative explanation to resource dependency model is financial control. Similar
to resource dependency, financial control suggests that firms need to maximize control
over resources necessary to conduct business. But financial control essentially supports
the notion that capital is the key input for large corporations to remain competitive in
today’s business environment. All firms require capital. Theorists suggest that the need
for large amounts of short-term capital make industrial corporations dependent on finan-
cial institutions. The concern for quick access to capital is diminished with a connection
to a large bank. By electing a banker to the board of directors, a company may expect to
have more ready access to bank funds.
Previously access to large amounts of short-term capital made industrial corporations
dependent on financial institutions. But today, large corporations can generate the neces-
sary capital internally more easily. This poses the question, why would firms still place an
emphasis on interlocking with financial institutions?
It is hypothesized by the authors that financial control model has undergone dramatic
change in the “knowledge economy.” In other words, interlocking with financial institu-
tions has moved beyond the bank as a source of funding. This can be explained by the
necessity for corporations to procure knowledge capital. As stated previously, banks by
their nature are an excellent source of knowledge on the economic environment. Since
borrowing and lending are contingent upon the overall economic climate, banks must be
knowledgeable of this to turn a profit. This knowledge is what provides banks with a
competitive advantage and more interlocking.
Two variables were created to measure the relationship between interlocking and
financial services. First, percentage of employment in FIRE examines the concentration
of a metropolitan area in finance, insurance, and real estate. It was calculated by dividing
total employment in finance, insurance, and real estate by total employment for all indus-
tries. The second variable associated with finance and insurance is the number of char-
tered banks in a metropolitan area. The authors hypothesize that firms want to link to

TACIT KNOWLEDGE TRANSFER
173
cities that provide banking and investment services. Therefore, both variables should be
positively associated with interlocking.
Region
Storper (1997) suggested that the region is now the fundamental basis of obtaining
knowledge. Providing the example of the Hollywood entertainment industry, Storper
argues that firm interdependence within a region can result in (among other things)
greater knowledge and information transfer.
The variable region is defined as an interlock to the largest metropolitan area of a
region. A dummy variable was created where those interlocks to the largest metropolitan
area were coded 1, and 0 otherwise. By interlocking to the largest metropolitan area, firms
are able to secure tacit knowledge on the regional market. Knowledge in the largest met-
ropolitan area should be more plentiful than any other metropolitan area within the
region. Interlocking to the largest cities in the eight U.S. census regions was used as the
level of measurement for the United States. For Canada, a director was considered inter-
locked if he was connected to the largest metropolitan area of the province.
EMPIRICAL RESULTS
For comparison purposes, two models from each of the three samples are presented in
Table 3. Results for Sample USA-CAN produce r2’s of .50 and .54. All estimated coeffi-
cients have the expected sign with the exception of industry concentration. The results
suggest that interlocks are likely to increase as metropolitan-area population, metropoli-
tan-area growth, research institution presence, and finance and insurance presence
increases. The largest metropolitan area in the region is likely to have a positive relation-
ship with interlocking as well. Although not indicated in Table 3, statistical testing veri-
fied that headquarters and banks also maintained a positive relationship with interlocking.
The only variable with a negative relationship on interlocking is distance. This is of
particular interest from a geographer’s perspective. The negative relationship shows that
firms interlock with other firms in the immediate area. Interlocking with regional firms
provides important knowledge on the local economy where a majority of their suppliers
and buyers are located. This consideration is verified by the fact that firms interlock with
other firms in the largest metropolitan area of the region.
Unfortunately the spatial interaction models were unable to determine with any degree
of certainty whether industry concentration was negatively or positively associated with
interlocking. The results suggest that some firms adhere to the principles put forth by
Porter (1990) and link to metropolitan areas that concentrate in similar production to their
own. Alternatively, other firms adhere to the principles put forth by Jacobs (1969) and
Glaeser et al. (1992) by linking to metropolitan areas that contain knowledge in a wide
array of industries.
To resolve the conflicting results, the residuals produced by industry concentration
were examined. All models (not only Sample USA-CAN but also Sample USA and Sam-
ple CAN) in Table 3 drastically underestimate metropolitan areas with large linkages.
Since these linkages generally occur in large metropolitan areas, we argue that the impor-

174
O’HAGAN AND GREEN
TABLE 3.—POISSON REGRESSION RESULTS,
CANADIAN AND AMERICAN INTERLOCKING, 1996
Sample USA-CAN
Sample USA
Model 1 Model 2
Sample CANADA
Variablea
Model 1
Model 2
Model 1
Model 2
Population 1
1.211*
0.796*
1.139*
0.681*
2.386
Population 2
2.047*
Headquarters 1
1.892*
2.209*
2.169*
0.476
Headquarters 2
FIRE as % of Employment 1
FIRE as % of Employment 2
Industry concentration 1
Industry concentration 2
Region
1.642*
0.873*
5.133*
–2.049*
0.807*
0.045*
0.015
1.987*
0.524*
5.139*
–2.075*
0.661*
0.039*
0.011*
1.228*
1.232*
–0.628*
–6.118*
0.992*
2.247*
–1.101**
-0.625
1.843*
–4.648*
–0.416*
0.438*
0.447*
0.012*
0.003*
1.093*
1.071*
-0.617
1.165*
–0.299*
0.004
Population change 1
Population change 2
University presence 1
University presence 2
Distance
0.008*
–0.653*
-0.62300
–8.564*
–0.434
0.544
–0.424
–23.068*
–697.859
4416.460
0.000
Constant
–16.467*
Log likelihood
–4190.807 –3930.808 –1910.729 –1913.170
–285.927
5240.330
0.000
LR chi2(10)
8560.000
0.000
9079.990
0.000
4078.380
0.000
4073.490
0.000
Prob>chi2
Pseudo R2
0.500
0.540
0.520
0.520
0.900
0.760
aUnder variables: 1 indicates origin city; 2 indicates destination city.
*p > .01. **p > .05. ***p > .10.
tance of dispersion is overwhelmed by the sheer quantity of knowledge that exists in these
large urban centers.
Sample USA results (American interlocking) are broadly similar to those estimated for
Sample USA-CAN. Table 3 reports two r2’s of .52. These models suggest that domestic
interlocking in America will increase as metropolitan-area population, metropolitan-area
growth, research institution presence, and finance and insurance presence increases. The
largest metropolitan area in the region is likely to have a positive relationship with inter-
locking as well. Once again, the number of headquarters located in a metropolitan area,
and the number of banks maintained a positive relationship with interlocking but were not
revealed in the best models.
Distance and industry concentration were inversely related to interlocking. The inverse
relationship with industry concentration would suggest that American firms are more

TACIT KNOWLEDGE TRANSFER
175
likely to interlock to cities with a dispersed economy. Care should be taken when making
this conclusion though because the relationship was only significant at the 0.10 level.
Poisson regression models for the Sample CANADA (Canadian interlocking) are also
displayed in Table 3. They yield r2’s of .90 and .76. Interlocks increase as metropolitan-
area population, the number of headquarters located in a metropolitan area, population
change, and finance and insurance presence increases. On the other hand, distance and
region are inversely related to interlocking.
Once again it is difficult to determine if a definite relationship between interlocking
and industry concentration exists. While one of the industry concentration variables
retains a positive relationship, the other maintains a negative relationship. Unfortunately,
it is difficult to construct a number of models for the Canadian sample because the vari-
ables headquarters, population, total FIRE employment, university presence, and banks
all maintained an intercorrelation greater than 0.8. In testing though, all five variables
maintained a positive relationship with interlocking.
The negative relationship for region is surprising. This can be explained if we explore
the strength of the independent variables between samples. In Sample USA-CAN and
Sample USA the negative influence of the variable distance was stronger than for Sample
CANADA. In other words, distance was not as much of a deterrent for knowledge transfer
in Canada. In addition, interlocking in Sample CANADA possessed a much stronger rela-
tionship with metropolitan area population than it did in Sample USA and Sample USA-
CAN. Thus, tacit knowledge transfer in Canada was greatly influenced by city size.
These results suggest a hierarchical interlocking structure in Canada. Distance is still
an important determinant of interlocking in Canada because of intra-city interlocking. In
fact, in Canadian cities with more than five interlocks, the intracity link was most preva-
lent in seven of the nine Canadian cities. This suggests that obtaining tacit knowledge on
the local market is important to Canadian firms. But the results also imply that regional
structures of interlocking are nonexistent in Canada.We propose that this is attributable
to a knowledge threshold essential for the transfer of tacit knowledge.
A knowledge threshold is an area that must be encompassed to ensure adequate knowl-
edge. In this instance, the knowledge threshold is the minimum distance required to
obtain knowledge necessary to make qualified business decisions. In Canada, the knowl-
edge threshold encompasses the entire country. Here, distance is overwhelmed by two
things. First, the dispersed Canadian economy makes it necessary to obtain knowledge at
an increased geographic scale. Second, a smaller base population of reputable and accom-
plished professionals is only available over a greater geographic area. Therefore, firms
must extend interlocking further distances to obtain the tacit knowledge possessed by
these individuals. From Figure 3A and Figure 3B, it is possible to argue that the only
concentrated interlocking in Canada occurs in southern Ontario. After the local metropol-
itan area, tacit knowledge transfer will occur in the largest metropolitan areas with dis-
tance being immaterial.
On the other hand, the network of interlocking has developed more extensively in the
United States. Figures 3A and 3B suggest the flow of tacit knowledge is more regionally
based and widely distributed in the United States when compared to Canada. A greater
number of headquarters, research institutions, and banks allow American firms to link
shorter distances to obtain the knowledge necessary for sound business decisions. This is
verified by the inverse relationship of interlocking with distance and the positive relation-

176
O’HAGAN AND GREEN
ship with the variable region in Sample USA. In addition, large regional populations do
not force firms to procure knowledge over greater distances. A large pool of capable
directors is available at shorter distances. Hence, the knowledge threshold in America is
much shorter.
CONCLUSIONS
The main goal of this research was to develop a spatial component of the resource
dependency model. This was accomplished by using variables that measure knowledge
existing at the metropolitan-area level. Results support the resource dependency theory.
All poisson regression models verified that interlocks increase as metropolitan-area pop-
ulation, number of headquarters located in a metropolitan area, metropolitan-area growth,
research institution presence, and finance and insurance presence increases. These results
suggest that firms want to access knowledge-rich metropolitan areas. As stated earlier,
firms are unable to generate all the necessary inputs for production. They respond with
relationships to external elements of the environment to ensure a supply of the required
resources and decrease uncertainty. Perhaps the most important source is knowledge. By
linking to knowledge-rich metropolitan areas, firms ensure a supply of the most important
resource.
The data were then divided into Canadian (Sample CAN) and American (Sample
USA) interlocking. We hypothesized that different spatial patterns and economy ages
influence the level of interlocking. In other words, the spatial patterns and metropolitan
area characteristics influencing knowledge transfer in Canada and the United States are
different. Poisson regression models verified our expectations. Examining the sign of
coefficients suggests that the only difference between Canada and the United States is
region. When combined with results of the variables distance and population, a signifi-
cant distinction emerges between Canada and the United States. It appears that distance
is less likely to discourage tacit knowledge transfer in Canada than in the United States.
We argue this is due to the fact that the knowledge threshold in the United States is geo-
graphically condensed when compared to Canada.
Figures 3A and 3B further this argument. One can see that Canadian interlocking was
highly concentrated in four cities (supported by the r2 of .90): Toronto, Montreal, Cal-
gary, and Vancouver. Of total interlocking that took place in Canada, 85% was inclined to
take place in these four cities. On the other hand, while New York City was the most
important origin and destination for interlocking in the United States, the top 20 cities
account for less than 70%. This suggests that tacit knowledge transfer, and thus knowl-
edge creation, takes place primarily in a few geographically dispersed cities across Can-
ada. American interlocking is more diversified and scattered.
This paper is a foundation for future research on the spatial dynamics of interlocking,
and thus tacit knowledge transfer. Since this research suggests a geographical threshold,
the next step is to verify if one does exist. From Figures 3A and 3B it appears that knowl-
edge transfer is a spatially complex phenomenon. But it also reveals that knowledge
transfer is associated with the region. A network analysis would allow a more appropriate
identification of prime cities existing within distinctive regions of the United States and
Canada. It also allows for the recognition of any existing hierarchical patterns. Further, a
longitudinal study on interlocking allows for the examination of the concept of a knowl-

TACIT KNOWLEDGE TRANSFER
177
edge threshold over time. Has the knowledge threshold expanded or contracted over
time? How do changes to the Canadian knowledge threshold compare to changes to the
American knowledge threshold? The theoretical concept of the knowledge threshold is
interesting because it introduces additional questions to be answered.
REFERENCES
Amin, A. and Cohendet, P., 1999, Learning and adaptation in decentralized business net-
works. Environment and Planning D, Vol. 17, 87–104.
Badaracco, J., 1991, The Knowledge Link. Boston, MA: Harvard Business School Press.
Boyd, B., 1990, Corporate linkages and organizational environment: A test of the
resource dependence model. Strategic Management Journal, Vol. 11, 419–430.
Branstetter, L., 2000, Vertical keiretsu and knowledge spillovers in Japanese manufactur-
ing: An empirical assessment. Journal of the Japanese and International Economies,
Vol. 14, 73–104.
Burt, R., 1983, Corporate Profits and Co-optation. New York, NY: Alliance.
Canadian Key Business Directory, 1997. Toronto, Canada: Dun and Bradstreet of
Canada.
Charon, R., 1998, How Boards Work: How Corporate Boards Create Competitive Advan-
tage. San Francisco, CA: Jossey-Bass.
Corporate Affiliations Plus [CD-ROM], 1997. New Providence, NJ: Bowker-Reed Refer-
ence Electronic Publishing.
D&B Million Dollar Directory: America’s Leading Public and Private Companies, 1997.
Bethlehem, PA: Dun and Bradstreet.
Dess, G. and Beard, D., 1984, Dimensions of organizational task environments. Adminis-
trative Science Quarterly, Vol. 29, 52–73.
Directory of Directors, 1997. Toronto, Canada: Financial Post.
Drucker, P., 1993, Post-Capitalist Society. Oxford, UK: Butterworth Heinemann.
The Financial Post 500, 1997. Toronto, Canada: Financial Post.
Fligstein, N. and Brantley, P., 1992, Bank control, owner control, or organization dynam-
ics. American Journal of Sociology, Vol. 98, 280–307.
The Fortune 500, 1997. Fortune, Vol. 135, F1–F20.
Glaeser, E., Kallal, H., Scheinkman, J., and Shleifer, A., 1992, Growth in cities.The Jour-
nal of Political Economy, Vol. 100, 1126–1152.
Glazer, R., 1998, Measuring the knower: Towards a theory of knowledge equity. Califor-
nia Management Review, Vol. 40, 175–194.
Green, M., 1980, A Geographical Examination of Interlocking Directorates for Large
American Corporations. Unpublished doctoral dissertation, Department of Geogra-
phy, Ohio State University.
Green, M., 1983, The interurban corporate interlocking directorate network of Canada
and the United States: A spatial perspective. Urban Geography, Vol. 4, 338–354.
Green, M. and Semple, R., 1981, The corporate interlocking directorate as an urban spa-
tial information network. Urban Geography, Vol. 2, 148–160.
Haunschild, P. and Beckman, C., 1998, When do interlocks matter?: Alternate sources of
information and interlock influence. Administrative Science Quarterly, Vol. 43, 815–
845.

178
O’HAGAN AND GREEN
Holtshouse, D., 1998, Knowledge research issues. California Management Review, Vol.
40, 277–280.
Jacobs, J., 1969, The Economy of Cities. New York, NY: Vintage.
Kono, C., Palmer, D., Friedland, R., and Zafonte, M., 1998, Lost in space: The geography
of corporate interlocking directorates. American Journal of Sociology, Vol. 103, 863–
911.
Kuznets, S., 1966, Modern Economic Growth: Rate, Structure, Spread. New Haven, CT:
Yale University Press.
Liao, T., 1994, Interpreting Probability Models: Logit, Probit, and Other Generalized
Linear Models. Sage University Paper Series on Quantitative Applications in the
Social Sciences, No. 101. Newbury Park, CA: Sage.
Lorsch, J. and MacIver, E., 1989, Pawns or Potentates: The Reality of America’s Corpo-
rate Boards. Boston, MA: Harvard Business School Press.
Lundvall, B. and Johnson, B., 1994, The learning economy. Journal of Industry Studies,
Vol. 1, 23–42.
Maclean’s, Universities 1997: The 7th annual rankings, 1997. Vol. 110, November, entire
issue.
Malmberg, A., 1996, Industrial geography: Agglomeration and local milieu. Progress in
Human Geography, Vol. 20, 392–403.
Marshall, 1920, Principles of Economics. New York, NY: Macmillan.
Maskell, P. and Malmberg, A., 1999, Localized learning and industrial competitiveness.
Cambridge Journal of Economics, Vol. 23, 167–185.
Mizruchi, M., 1996, What do interlocks do? An analysis, critique, and assessment of
research on interlocking directorates. Annual Review of Sociology, Vol. 22, 271–298.
Morgan, K., 1997, The learning region: Institutions, innovation and regional renewal.
Regional Studies, Vol. 31, 491–503.
Nonaka, I. and Takeuchi, H., 1995, The Knowledge Creating Company. New York, NY:
Oxford University Press.
O’Hagan, S. and Green, M., in press. Region, knowledge, and competitiveness. In R.
McNaughton and M. Green, editors, Global Competition and Local Networks. Alder-
shot, UK: Ashgate.
Ohmae, K., 1995, The End of the Nation State: The Rise of Regional Economies. New
York, NY: The Free Press.
Oinas, P. and Malecki, E., 1999, Spatial innovation systems. In E. Malecki and P. Oinas,
editors, Making Connections: Technological Learning and Regional Economic
Change. Aldershot, UK: Ashgate, pp. 7–34.
Pennings, J., 1980, Interlocking Directorates. San Francisco, CA: Jossey-Bass.
Pfeffer, J., 1972, Size and composition of corporate boards of directors: The organization
and its environment. Administrative Science Quarterly, Vol. 17, 218–228.
Pfeffer, J. and Salancik, G., 1978, The External Control of Organizations: A Resource
Dependence Perspective. New York, NY: Harper and Row.
Porter, M., 1990, The Competitive Advantage of Nations. New York, NY: Free Press.
Pred, A., 1977, City-Systems in Advanced Economies: Past Growth, Present Processes
and Future Development Options. London, UK: Hutchinson.
Rice, M. and Semple, K., 1993, Spatial interlocking directorates in the Canadian urban
system. Urban Geography, Vol. 14, 375–396.

TACIT KNOWLEDGE TRANSFER
179
Richardson, R., 1987, Directorship interlocks and corporate profitability. Administrative
Science Quarterly, Vol. 32, 367–386.
Sassen, S., 1991, The Global City: New York, London, Tokyo. Princeton, NJ: University
Press.
Saxenian, A., 1994, Regional Advantage: Culture and Competition in Silicon Valley and
Route 128. Boston, MA: Harvard University Press.
Solvell, O. and Zander, I., 1995, Organization of the multinational enterprise: The home-
based and the hierarchical MNE. International Studies of Management and Organiza-
tion, Vol. 25, 17–38.
Standard and Poor’s Register of Corporations, Directors, and Executives, 1997. New
York, NY: Standard and Poor’s.
Storper, M., 1997, The Regional World: Territorial Development in a Global Economy.
New York, NY: Guilford.
Teece, D., 1998, Capturing value from knowledge assets: The new economy, markets for
know-how, and intangible assets. California Management Review, Vol. 40, 55–79.
Useem, M., 1984, The Inner Circle. New York, NY: Oxford University Press.
U.S. News and World Report, America’s best colleges, 1997. Vol. 121, No. 11, entire
issue.
Velleman, P. and Hoaglin, D., 1981, Applications, Basics, and Computing of Exploratory
Data Analysis. Boston, MA: Duxbury.