the role of organizational mechanisms in preventing leakage of … · 2020-01-06 · the role of...
TRANSCRIPT
12076
1
The role of organizational mechanisms in preventing leakage of unpatented knowledge
ABSTRACT
Firms need to protect their proprietary knowledge assets from imitation in order to appropriate
rents from them. A significant volume of scholarly work is devoted to legal mechanisms for
intellectual property (IP) protection such as patents. However, there is a dearth of work on how
firms can protect their knowledge that is not patented, such as valuable knowledge that is not
codified, is inchoate, or otherwise does not meet the criteria for patenting. In this paper we
explore organizational processes and mechanisms that firms can use to protect such valuable
knowledge from imitation. We explore these mechanisms using interview and survey data from
captive R&D centers of multinational firms in India, where the weak IP regime places weak
barriers to knowledge leakage.
INTRODUCTION
Non-imitability of resources is one of the cornerstones of competitive advantage (Barney,
1986; 1991; Peteraf, 1993). The protection of knowledge resources from imitation, though
important to generate rents from valuable knowledge, may also be difficult to achieve because of
Arrow’s paradox (1962). Even though significant amount of scholarship has considered how
firms innovate, we do not understand very well how firms protect such knowledge from
imitation (Liebeskind, 1996; 1997; Puranam and Reitzig, 2009), thus enabling them to
appropriate value from such knowledge. Better understanding how firms protect their knowledge
from imitation by rivals is becoming especially important in today’s business context where
firms are increasingly accessing their knowledge from a variety of dispersed sources, both within
and outside their boundaries (Chesbrough, 2003; Parmigiani and Mitchell; 2009; Helfat and
Quinn, 2006). As a consequence, such knowledge is also less tightly controlled by the firm and
12076
2
more widely available to potential imitators (Giaratana and Mariani, 2013; Alcacer, 2006;
Alcacer and Zhao, 2012). In this paper we attempt to examine the organizational mechanisms
that firms implement to minimize the leakage of their proprietary knowledge assets.
One context where the protection of knowledge from leakage is increasingly important
stems from the globalization of R&D and New Product Development (NPD) efforts. Many
multinational firms operate R&D centers across the globe, including in locations that provide
limited legal protection for intellectual property (IP). For example, a significant number of
Fortune 500 firms operate R&D centers in India and China, two locations with poor protection
for IP (Economist, 2010), and therefore are more susceptible to imitation. We employ field
interviews supplemented with a survey of R&D managers in Western based multinational firms’
wholly owned captive centers in India engaged in R&D and NPD work to understand how firms
employ organizational mechanisms to protect their proprietary knowledge from imitation.
The resource-based-view of the firm suggests that firms can derive value from their
resources not only if they are ‘valuable’, but also ‘rare, inimitable and non-substitutable’
(Barney, 1991; Peteraf, 1993). For example, knowledge of a formula or manufacturing process
for a molecule that can cure diabetes is valuable because patients are willing to pay for it, rare
when no other firm possesses this knowledge and non-substitutable when no other molecule or
manufacturing process can achieve the same results. However, this knowledge can only provide
rents to the firm when competitors cannot easily imitate this knowledge.
Rumelt (1984) introduced the notion of ‘isolating mechanisms’ to explain why rivals
might find it difficult to imitate a firm’s resources. Such isolating mechanisms include property
rights to scare resources and various quasi-rights in the form of time lags, information
asymmetries and frictions that impede imitative competition (Peteraf, 1993). Dierickx and Cool
12076
3
(1989) suggested that asset accumulation processes such as time compression diseconomies,
causal ambiguity and interconnectedness of asset stocks can function as isolating mechanisms.
Lippman and Rumelt, (1982) argued that causal ambiguity, which is uncertainty regarding how a
firm’s activities combine in order to achieve desired outcomes, is particularly important isolating
mechanism that enables firms to maintain quasi-rents (also see Porter, 1996). Building on these
foundations, Mahoney and Pandian (1992) highlight a list of 38 isolating mechanisms that make
it difficult for competitors to imitate a focal firm’s resources. They also suggest that the
relevance of these isolating mechanisms varies according of the kind of resource in question.
In case of a firm’s knowledge assets, patents are traditionally seen as vital in preventing
its competitive imitation. However, research has also documented the disadvantages of patents
(Cohen et al, 2000; Liebeskind, 1996). For example, patents require ‘disclosure’, which may be
seen as disadvantage if it leads to creation of new knowledge by rivals that may substitute for the
patented knowledge. Also, much valuable knowledge in a firm cannot be sufficiently codified or
may not meet the requirements of ‘novelty’ or ‘non-obviousness’, and consequently cannot be
patented. For such reasons, patents are an effective isolating mechanism only in the
pharmaceutical and chemical industries (Mansfield, 1985; Levin et al, 1987; Cohen et al, 2000).
Given these limitations associated with patents, firms often also have to rely on trade secrets
when they do not want to disclose their knowledge by filing a patent, or when patents are
ineffective as a deterrent to imitation.
A trade secret consists of any formula, pattern, device or compilation of information used
in a firm’s business, which gives the firm an opportunity to claim an advantage over companies
who do not know or use it (Harding, 1967). However, the knowledge protected by trade secrets
also need to be codified, i.e., they do not protect valuable tacit knowledge. In addition, unlike
12076
4
patents, trade secrets do not create a property right in knowledge; rather they are more akin to
laws on theft or illegal possession (Friedman, Landes and Posner, 1991), and therefore there are
limited legal remedies for their breach (also see discussion by Liebeskind, 1996). The leakage of
unpatented knowledge such as trade secrets can have significant impact on a firm’s competitive
position, since it could either allow imitators to directly copy products/services based on it, or
help them create ‘work-arounds’ to any patents the firm might hold related to it.
Employee mobility across firms is the primary means for such unpatented knowledge to
leak to competitors (Almeida, 1996; Almeida and Kogut, 1999; Rosenkopf and Almeida, 2003).
Employee mobility between firms overcomes other traditional barriers to knowledge transfer
such as tacitness, lack of absorptive capacity and the lack of motivation for knowledge transfer
and use (Szulanski, 1996), especially when former employees can still use their informal social
networks to access related knowledge from their former colleagues (for example, see Saxenian,
1994). Prior studies have shown that employee mobility leads to increased use of the focal firms’
knowledge by their competitors (Almeida, 1996; Zucker and Darby, 1996).1
Firms have few legal options to protect such unpatented knowledge from leaking to
competitors, especially when that happens through the mobility of employees between firms.
This is because even if the employees are bound by ‘non-disclosure’ agreements (NDAs), breach
of such NDAs is hard to detect and prove (Hyde, 2003; Freidman et al, 1991). Firms often rely
on ‘non-compete agreements’ (NCAs) that bar employees from accepting employment with
specific rivals for a certain period of time so as to protect their knowledge from leaking to those
1 Employee mobility may not help much when competitive advantage rests on knowledge that is causally
ambiguous. However, causal ambiguity typically applies to operational excellence, such as the Toyota production
system or 3M’s ability to come up with innovative new products. However, in an NPD/R&D context, once the
product, such as the sticky note has been invented, competitors may still be able to copy those products or
reengineer them fairly easily, especially when this knowledge is only poorly protected by the IP regime, thus
depriving the focal firm from realizing rents from its innovations.
12076
5
competitors (Marx, Strumsky and Fleming , 2009; Marx, 2011). But in several jurisdictions (e.g.
Silicon Valley in the US, or countries like China or India), such non-competes are either not
enforced or illegal. Therefore, firms need to find alternate isolating mechanisms to protect its
unpatented knowledge to appropriate value from it.
Mansfield (1985) argued that a firm’s knowledge will eventually leak out, but we know
little about how firms can delay leakage of their proprietary knowledge. Whereas literature has
focused a lot on investigating the effectiveness of patents in protecting proprietary knowledge
held in that form, there is relative dearth of research on understanding the set of organizational
processes and mechanisms (as opposed to legal mechanisms such as NDAs or NCAs) to prevent
the leakage (and consequently imitation) of proprietary knowledge that is not patented. The
exception is Liebeskind (1996; 1997) argued theoretically based on anecdotal data that firms can
protect their knowledge from imitation by relying on incentives that reduce employee mobility,
or design work such that it is disaggregated, and rules of conduct that monitor employees and
attempt to minimize leakage. However, since firms increasingly rely on unbounded and widely
dispersed knowledge sources that they recombine, and the difficulty in guiding and monitoring
behavior in a networked world, these elements maybe less useful now than previously. In
addition to their likely limited effectiveness now, employing these mechanisms also involve
significant costs. Therefore, we believe it is time to take a fresh look at the organizational
mechanisms that firms use to protect their unpatented knowledge from leakage, and our paper
takes a step towards addressing this research gap.
We adopt a multi-method approach to study how firms protect their valuable unpatented
knowledge. First, we conducted filed interviews with over twenty R&D managers in
multinational firms and IP lawyers to obtain a rich description of some of the organizational
12076
6
mechanisms that firms use to protect their unpatented knowledge. Next, we utilize a survey
dataset comprising of 142 offshored R&D and NPD projects carried out by foreign-based
multinational firms in their captive R&D center in India, on which we carried out three sets of
analyses. First, we use t-tests to show that individual organizational mechanisms are associated
with greater confidence on the part of captive managers in protecting their IP. Second, we use a
canonical correlation analysis to show that using these organizational mechanisms as a set is
associated with managers’ confidence in protecting IP. This analysis is very similar to the one
used by Szulanski (1996) in his pioneering work on uncovering organizational impediments to
the transfer of best practices within firms. Finally, we use a regression analysis to show that
managers’ confidence in protecting their IP mediates the relationship between using these
organizational mechanisms and the propensity of the captive center to engage in projects that
have an inherently higher risk of knowledge leakage.
Given the paucity of studies on how organizations can protect their knowledge from
leakage, our empirical analysis is indicative. Our aim in this study is to uncover preliminary
evidence that can inform future in-depth studies regarding specific mechanisms that enable firms
to protect their knowledge and their inter-relationships. We hope that our preliminary analysis
inspires future studies on how firms protect their knowledge assets.
THE STUDY CONTEXT
We explore our research question in the context of R&D offshoring to emerging economies
where multinational enterprises (MNE) headquartered in developed economies such as the US or
EU perform R&D activities in emerging economies such as India/China. Many emerging
economy R&D destinations, and especially, the two most important ones, China and India, offer
only weak legal protection for IP. Whereas the internationalization of R&D as a phenomenon has
12076
7
a fairly long history, until recently, the lion’s share of this activity was confined to multinationals
performing R&D in foreign locations that have strong IPRs, such as US companies operating
R&D centers in the EU/Japan or vice-versa. In contrast, offshoring R&D to destinations with
weak IP protection is a fairly new phenomenon, and though it is growing rapidly, it has not yet
received significant scholarly attention. We believe, that this context of R&D offshoring to
countries with weak IP regimes seems very relevant to study how firms protect their proprietary
knowledge, especially when that knowledge is unpatented.
We first provide a brief introduction to offshoring in general and R&D offshoring in
particular in order to set the context in which we explore our phenomenon of interest. Value
chains can be decoupled across two dimensions: ownership and geography. Offshoring refers to
the decoupling of value chains across geographic locations, typically between high-wage
locations such as the USA or Western Europe and low wage locations such as India or China
(Srikanth and Puranam, 2011). ‘Captive offshoring’ occurs when the value chain is decoupled by
geography but not by ownership, whereby the parent multinational establishes a wholly owned
subsidiary to execute its work in the offshore overseas location. For example, the John F. Welch
Technology Center in Bangalore, India is an offshore captive R&D unit for GE, and is GE’s
largest multi-disciplinary R&D center in the world2. This is different from ‘offshore outsourcing’
where the value chain is decoupled by both geography and ownership, so that a third party
vendor in the offshore location executes work for the multinational company on a contractual
basis. For example, Wipro Technologies and Infotech Enterprises are prominent Indian vendors
that offer contract R&D services to different MNCs.
Offshoring mainly relies on labor cost arbitrage between high-cost locations such as the
USA and low cost locations such as India/China (Farrell, 2005). However, although cost
2 http://ge.geglobalresearch.com/locations/bangalore-india/, last accessed 13 June 2013.
12076
8
reduction is one of the most cited reasons for offshoring work, Lewin et al. (2009) from their
recent survey find that accessing high quality talent in science and engineering, particularly in
India and China, is swiftly becoming one of the most important reasons for MNEs to offshore
R&D work. In the past decade there has been an explosion in offshoring of R&D and NPD work.
Booz & Company, in their 2008 survey of 1000 public corporations worldwide that invest most
in R&D, found that 83 percent of new R&D sites of MNEs were in India and China and 91
percent of new R&D staff in these firms was located in these countries. 3
This growth in captive R&D centers is accompanied by a strategic shift in offshore
operations of the MNE towards undertaking increasingly novel and complex work in offshore
locations (see Kumar and Puranam, 2012 with respect to India). While some of this offshoring is
driven by the MNE’s desire to tap into the growing markets in emerging economies, a large
amount of R&D is directed towards developing new product and process designs and services for
global (western) markets (Zhao, 2006; Mudambi and Venzin, 2011). Hegde and Hicks (2008)
argue that emerging economy R&D captive centers are now well enmeshed in the R&D value
chains of large MNEs, and several scholars have suggested that the innovations generated from
these captive centers have as high as or even higher internal impact than innovations generated in
the MNE’s R&D centers located at the headquarters country or in other advanced economies
(Alnuaimi, George and Puranam, 2012a; Alnuaimi, Singh and George, 2012b; Singh, 2008).
With captive centers now performing high value add work it is that much more important
for MNEs to protect the knowledge resident in these centers from leaking to competitors. An
obvious answer is increased patenting of knowledge produced in these centers, and indeed,
concurrently there is a large increase in patenting activity by these offshore captive centers in
3 http://www.booz.com/media/file/sb61_10408-R.pdf. Also see “Special report on innovation in emerging markets,”
The Economist, April 17, 2010.
12076
9
recent years. This includes the number of patents (co)-generated by these captive centers and
filed in the US PTO as well as patents filed in their respective host county such as China/India
(Kumar and Puranam, 2012; Hu and Jefferson, 2009; Alnuaimi, et al, 2012a; Zhao, 2006).
However, this is a rather incomplete solution to the knowledge protection challenge faced
by these centers. In several surveys, MNE executives have expressed discontent with the extent
of IP protection available in the main offshore destinations such as India and China (for example,
see Thursby and Thursby, 2006 and Kumar and Puranam, 2012), especially in the poor
implementation of IP laws – we observed the same in our fieldwork as well. Increasingly,
captive centers perform work that is relevant to the MNE’s home market such as the USA (Zhao,
2006; Mudambi, 2011). Patent regimes in India/China are irrelevant to these projects, though the
leakage of unpatented knowledge to local firms, especially to emerging market multinational
competitors can be severely damaging. Therefore, protecting patented and especially unpatented
knowledge from leakage, by other means rather than relying on the legal IP regime, is extremely
important for R&D work carried out in these captive centers.
As prior work has suggested, knowledge spillovers tend to be local, and mainly occur
through mobility of inventors between firms (Almeida, 1996; Almeida and Kogut, 1999; Marx et
al, 2009). Whereas in the US, NCAs are effective in preventing the mobility of employees across
competitors (based on the jurisdiction), such contracts are minimally enforced in many
offshoring destinations.4 For example, the Indian constitution guarantees the right to work for an
employee and an employer can place only minimal restrictions on the employment choices of an
4 Though theoretically this problem exists even in the USA in some jurisdictions such as in Silicon Valley,
appropriation may be less of a concern because firms may have evolved norms regarding the use of IP (for example,
see Liebeskind et al, 1996; Zucker et al, 1995). Our fieldwork suggests that in offshore destinations, especially with
domestic competitors, such norms may not exist. Several patent lawyers mentioned this lack of norms regarding the
use of IP from previous employers, and press reports mention cases where the primary purpose of employee
mobility is to transfer advanced knowledge from the MNE to the domestic competitor (for example, see “China and
Intellectual Property,” New York Times, December 24, 2010, p. A22, New York edition).
12076
10
employee, regardless of whether the separation is voluntary or involuntary. Coupled with the fact
that turnover among the labor pool in many prominent offshoring destinations, such as India, is
quite significant (Manning, Massini and Lewin, 2008; Wood, 2012)5 with firms often “poaching”
away talent from their rivals, protecting unpatented knowledge is a significant challenge. In this
context, many of the mechanisms Liebeskind (1996; 1997) talks about to reduce turnover such as
incentives and social sanctions have limited effectiveness. It is important to note that this threat
of knowledge leakage is significant regardless of whether the knowledge in question is mainly
relevant to the offshore market or to the global market. In such settings, firms have to rely on
organizational mechanisms to protect the IP generated in their captive centers. Thus the context
of offshoring R&D activities to emerging economies by MNEs is a particularly interesting and
appropriate setting to understand what organizational mechanisms firms use to protect their
proprietary knowledge from leaking to competitors.
THEORY AND PROPOSITIONS
To develop our proposition, we first conducted a qualitative study of several organizations
to better understand some of the practices and processes they were adopting to prevent
knowledge leakage in their offshore R&D work. In this phase, we conducted over twenty
interviews with managers of captive centers in India that perform R&D or NPD work, two
interviews with R&D managers in the headquarters location, and four interviews with IP
lawyers. Our respondents in the captive centers typically had the roles of senior project
managers, R&D heads or captive center heads in these firms. While our cases were selected
based on a convenience sample, they represented a broad spectrum of sectors where R&D or
NPD offshoring work is widely undertaken. The managers we interviewed came from larger and
5 Also see http://articles.economictimes.indiatimes.com/2013-06-07/news/39815456_1_three-employees-indian-
employees-attrition (accessed 7 May 2014).
12076
11
well-established captives across the pharmaceuticals, energy, semiconductor, IT hardware and
software industries based on contacts we were able to gather. These interviews were free
flowing, where we asked the managers to describe how they protected their knowledge from
leaking to competitors as a result of employee mobility between firms. Interviews ranged
between 45 and 100 minutes, and some of them were taped with the respondents’ permission.
The interviews were done in pairs and the interviewers took extensive notes which were
compiled the same day, along with any field observations. Managers were asked to provide
specific examples of organizational mechanisms used rather than just a general overview. These
interviews were analyzed using a repeated readings technique to understand the mechanisms
used and how they help in preventing knowledge leakage. We compared the mechanisms we
uncovered, with prior studies on R&D offshoring and knowledge leakage, to understand how
they related to prior findings. In general, MNE managers described two principal threats that
arise from leakage of their proprietary knowledge (a) domestic competitors entering the domestic
market using misappropriated IP and (b) knowledge leakage from the foreign MNEs increases
the absorptive capacity of emerging economy competitors to the extent that these firms will
become aggressive global competitors in the near future. Next we describe the mechanisms we
identified from our interviews as well as congruent findings from prior work that mitigate the
likelihood of knowledge leakage. These findings became the input to our survey study that is
described later in this paper.
Our main premise is that personnel mobility is a key factor in enabling IP leakage,
especially tacit knowledge and trade secrets, which was borne out in our interviews. Since firm
strategies for value appropriation through IP must be sensitive to this threat, in our research we
12076
12
focus mainly on uncovering mechanisms that firms can use to prevent their current or former
employees from misappropriating their IP.
The first and perhaps the most important mechanism we uncovered, in the sense that
every single manager we interviewed mentioned it, relates to the division of work; namely, how
knowledge is distributed across geographic locations. This mechanism, also referred to as ‘fine-
slicing’, has been identified in prior work on R&D offshoring as well (Zhao, 2006; Quan and
Chesborough, 2010). This mechanism entails the division of innovative labor between the HQ
location (with strong IPR) and offshore locations (with weak IPR), which protects knowledge
from leaking in two ways. First, if the innovative know-how in the different locations is
complementary, the IP generated in the offshore location may be of little value by itself without
combining it with the IP generated in the other locations. To the extent that complementary
knowledge is protected, say in the headquarters locations such as in the USA, leakage of
knowledge in the offshore location is of little concern. This strategy requires an assumption that
the complementary knowledge is patented and these patents are enforced in the relevant markets.
This assumption likely holds if the market in question is the USA, which is the sample in which
Zhao (2006) tests her theory, but is unlikely to hold if the main market for these products are the
Chinese or Indian markets.6
Second, the spillovers of required complementary knowledge in different locations
maybe less of a problem for the focal firm. This is because (a) tacit knowledge spillovers are
likely to be predominantly local (Almeida, 1996; Almeida and Kogut, 1999); and (b) potential
6 We should note that some managers did mention that there are some projects that are carried out exclusively from
their captive centre with minimal or no involvement from other locations. These managers suggested that carrying
out such projects is important for motivating and retaining high quality R&D talent. However, they also noted that
these projects are unlikely to be large or important sources of revenue for the firm. Often these involved mature
products that were shifted to the offshore centre to enable headquarters resources to concentrate on newer products,
or were products that were at least initially focused only on the domestic market.
12076
13
imitators are less likely to be successful in combining knowledge that spills over in different
geographic locations to compete against the focal firm. This is because combining knowledge
within-firms is easier than across firm-boundaries; the higher levels of common ground within
firms allow them to achieve more efficient knowledge transfer and coordination across their
geographically distributed subsidiaries than with suppliers or other partners (Srikanth and
Puranam, 2014; Almeida, Song and Grant, 2001). Therefore, by fine-slicing, the focal firm
effectively limits its spillover threat to other multinationals that are also present in the same
locations. This is likely to be a more effective mechanism in addressing knowledge leakage than
relying on legal protections for some of the knowledge in locations with strong IPR because it
protects even tacit knowledge assets and trade secrets, which are typically less effectively
protected by the patenting regime. These arguments suggest that firms that proactively configure
their R&D projects to achieve such fine slicing are less likely to be subject to knowledge leakage
even in weak IPR locations.
In practice, such fine slicing could be achieved in two ways. R&D tasks across the
different locations could be modularized such that there is minimal interaction between the R&D
teams in different locations. The second strategy is to perform work that is interdependent across
locations and rely on interactions between R&D personnel across these locations in order to
combine the complementary knowledge into innovative products and services. Prior research on
offshoring of knowledge work finds that modularization of activities across locations and
reliance on well-specified interfaces for recombination is impractical in diverse settings such as
in IT services (Srikanth and Puranam, 2014) and in offshored R&D and NPD work (Kotha and
Srikanth, 2013; Mani et al, 2014; Leonardi and Bailey, 2008). From a practical view point,
therefore, firms are more likely to implement a fine-slicing policy by implementing R&D
12076
14
projects that are interdependent across locations and witness high levels of involvement of
scientists and managers from the headquarters in the R&D work performed by the captive center.
Contrary to Liebeskind’s (1997) argument of generally limiting social interaction between
different groups within the firm to protect knowledge, we find that firms encourage such
interaction across employees in different locations.
Our interviews with R&D managers in MNEs also bear out this intuition. For example, a
manager in one of the largest R&D captives for a Fortune 50 firm located in India told us that the
firm had developed multiple R&D centers in different geographies and each of these centers
were responsible for executing work in their competence. Some of these centers had competence
in multiple technology domains, such as polymer chemistry, nano-technology, etc., whereas
others were more specialized in only one or two domains. Typically, this firm’s products were
complex systems that involved collaboration between multiple technology domains. Whereas
individual R&D centers performed cutting-edge work in their own domains relatively
independently, this tends to be more basic research, or as the manager put it ‘8-12 years to
product’. More applied work that was ‘3-5 years to product’ was systematically distributed
across multiple R&D centers in different countries. One reason for this approach was to staff the
project with the best scientific talent available. Another important reason, according to this
manager, was to distribute knowledge regarding specific products widely within the firm and not
concentrate them in one location. This suggests that fine-slicing at this firm was a deliberate
choice to minimize the leakage of the entire knowledge regarding a product from one location.
A senior manager in a global top 10 pharmaceutical company told us that as the firm’s
captive center gained in maturity, it has increasingly performed more and more valuable R&D
work that is relevant to the core offerings of the company. For instance, the captive center
12076
15
matured from performing clinical trials in that country, followed by taking over responsibility
end-to-end clinical trials management for certain drugs across geographies. This captive has now
matured into performing various critical R&D activities in drug design and formulations. The
manager said that as the R&D capability has moved closer to the firm’s core, the work in the
captive center has become more interdependent with the global discovery efforts. Whereas
earlier the firm spent a lot of time designing protocols and modularizing the development effort,
with the core R&D functions, interaction with the other R&D centers has risen significantly. The
manager suggested that the firm has deliberately created an environment of active collaboration
between inventors at its headquarters research labs and its captive center. Therefore, we suggest
the following propositions:
P1: Managers of projects in offshore captive centers that involve significant levels of
headquarters personnel in their R&D activities are less likely to be concerned about leakage of
proprietary IP.
P2: Managers of projects in offshore captive centers that involve significant levels of
interdependence between the activities conducted at the captive location and other (headquarter)
locations are less likely to be concerned about leakage of proprietary IP.
From our interviews, another aspect of project configuration that we found plays a role
in protecting IP in MNEs’ captive centers is asset specificity. Asset specificity is defined as
investments in assets that are idiosyncratic to the nature of the firm and its business such that
they generate maximum value in that specific context as compared the value that might arise in
deploying those assets elsewhere. As prior research shows this specificity can be of several kinds
(Williamson, 1985; 1991) including physical asset specificity (such as some tool or computing
architecture), human asset specificity (in term of the know-how they possess) or even procedural
12076
16
specificity (in terms of the routines and processes necessary to generate value). In our interviews
we found instances of R&D and NPD work, where the employee’s knowledge is intimately
connected to the specific physical and knowledge assets available within the firm, which may not
be available with other firms. Asset specificity in R&D helps protect IP from leakage, since a
competitor needs to make investments in similar assets in order to effectively use the
expropriated IP.
The role of asset specificity in protecting IP from leakage is different from the notion of
complementary assets (Teece, 1986) or bundling (Arora, 1996) in one important way.
Complementary assets allow for effective commercialization of a piece of IP, and may influence
investments in developing certain kinds of IP. In contrast, our discussion about asset specificity
in innovation suggests the need for investments in unique assets that enable the firm to generate
and utilize innovations by exploiting its knowledge resources.
For example, in our field work we studied a captive center for a large pharmaceutical
company that specializes in biologics and large molecule research, a technology in which patent
protection is weaker than in typical small molecule drugs. This center has made investments in
physical assets such as labs and equipment to produce small quantities ultra-pure biological
substances for its testing purposes. It also has unique procedures for validating its computational
models for designing proteins. It spends significant effort in training its employees to get them
up to speed in using these protocols. An employee who leaves this firm for a competitor is not
able to utilize much of their knowledge in drug design since the competitor will not have the
same equipment and protocols; the practically valuable knowledge is intimately tied to the lab
itself. Similarly, we witnessed a captive center in the automotive sector whose modeling
capabilities were tightly linked to the firm’s unique test facilities that operated in another
12076
17
country. Therefore, employee mobility is not a significant threat to IP protection for this
company. This suggests the following proposition:
P3: Managers of projects in offshore captive centers that involve a high level of asset specificity
are less likely to be concerned about IP leakage.
As discussed before, in case of high interdependence between the headquarters location
and the captive center, these projects are often managed by having high levels of involvement by
scientists/engineers from headquarters in the activities performed by the captive center. This
typically involves a significant level of information transfer through formalized routines of
communication and documentation between the two sites. Many R&D captive centers also
witness a greater degree of control exerted by the headquarters on their operations. Whereas in
our fieldwork we were unable to collect specific comparative data on the degree of control
exercised by headquarters on offshore captive centers in weak IPR locations when compared to
R&D labs in strong IPR regimes, anecdotal evidence in most our interviews suggests that that
may indeed be the case.
These practices are potentially effective in protecting IP in two ways. First, it provides
the headquarters adequate and timely information to spot and understand emerging nascent ideas
in the captive center – some of these could then be nurtured by providing relevant additional
knowledge from headquarters to develop them further, which in turn, will create a more valuable
idea that is also now more difficult to imitate. In centers with more headquarters control, the
headquarters could select to perform different kinds of projects that have a lower leakage threat,
for example, in areas that the focal firm has strong complementary assets. It can also direct the
course of a new project by implementing design and technical specifications that need not be
completely controlled by the captive center. Keeping partial knowledge in a different location
12076
18
reduces the value of the residual knowledge available at the captive center to a potential imitator.
Second, because the headquarters is now more informed about the projects and processes at the
captive center it can respond more swiftly to any potential leakage of these new ideas. It could do
that by either aggressively protecting promising new ideas by seeking their patent protection
before the IP leaks, or, it can pre-emptively challenge any patent infringement against departed
former employees who might want to profit from that idea after leaving the captive center.
In our field work we encountered captive centers operated by leading software
development firms adopting these practices. Some of these captives had invested in facilitating a
high level of communication between locations. These captives typically also had extensive
control exerted by the headquarters over the project selection and the design and implementation
details of the projects. For example, one firm, that is jokingly referred to as having more
managers than developers, has a number of employees whose primary task is to coordinate work
between the captive center and the parent lab in the USA. The tasks performed by this center are
important components of the overall product the firm builds and the headquarters lab is heavily
involved in relatively detailed decisions regarding the design and building of the components at
the captive center. However, such strong connection appear to be dependent on the type of work
performed by the captive center. In contrast, another software development firm’s captive center
primarily performs customization and product enhancement type work from its captive center. It
is very difficult for a local firm to imitate the core knowledge of this firm in order to produce the
same type of products. This firm also relies on its very strong complementary assets for
commercialization, and in this case, the captive center neither has a very strong connection with
the headquarters nor are its activities and projects heavily influenced by the headquarters.
However, this is distinctly the minority of cases we encountered. Thus, on average,
12076
19
P4: Managers of projects in offshore captive centers that entail a high level of control from the
headquarters are less likely to be concerned about IP leakage to competitors
P5: Managers of projects in offshore captive centers that entail a high level of information
transfer with the headquarters are less likely to be concerned about IP leakage to competitors.
In addition to the mechanisms described above, captive centers also rely on several
internal controls to protect their IP. The primary the aim of these controls is to withhold access to
confidential information by dispersing the critical knowledge within the captive – knowledge
leakage is thus prevented because bits of knowledge regarding critical aspects in the project are
now held by multiple individuals and no single person is able to access of the information
required to replicate the captive center’s innovation. This is akin to the principle of
disaggregation in job design to protect knowledge as suggested by Liebeskind (1996). In our
fieldwork, we observed several types of internal controls companies used: (a) storing information
in multiple databases each with their own access protocols; (b) storing information in servers
located outside the offshore location with controls regarding who can access such information;
(c) read, write, and download/print access restricted according to employee role in the project;
and (d) physically segregating teams working on different projects with restrictions on access to
the different work areas. In some firms, the physical layout is designed with the aim of
minimizing interactions between different project teams. Apart from these practices, many
captive centers take steps to educate employees regarding the importance of IP management and
put in place active monitoring systems. For example, confidential documents are clearly marked
as such so employees are aware that they are working with proprietary information. For example,
Kumar and Puranam (2012) provide details regarding the internal controls put in place by Intel in
its Bangalore captive center in order to minimize IP leakage. Thus,
12076
20
P6: Managers of projects in offshore captive centers that exhibit a high level of internal controls
are less likely to be concerned about IP leakage.
SURVEY OF R&D PROJECTS IN OFFSHORE CAPTIVE CENTERS
Empirical Approach
Our empirical approach is influenced by the nature of our research question, which has received
relatively little empirical attention. We are interested in how firms protect against the leakage of
unpatented knowledge that could happen with the mobility of employees across competitors, a
condition that is exacerbated in the offshore R&D context by the weak IP regimes in many
offshore destinations. Because of the novelty of this question, as described above, our approach
was to conduct field work with MNE R&D managers about their strategies for IP protection as
well as carefully study prior descriptive work on offshoring to understand whether any of the
management practices could serve the dual purpose of protecting IP from leaking as well as other
coordination and control functions. On this basis we identified the mechanisms above, and
proposed that firms that adopt these mechanisms may be more likely to minimize IP leakage.
A robust empirical testing of this idea faces several challenges. Though we interviewed
several managers we are unsure about how generalizable these mechanisms are. We do not know
whether these mechanisms are widely adopted in offshore captives or if there are sector specific
or technology specific patterns. For instance, analyzing our interviews we found that very few
firms employed the complete set of mechanisms described above. We also do not know the
relative importance of each of these mechanisms, and whether these mechanisms complement
each other or if they are substitutes. Therefore, we adopt an approach of providing ‘plausible’
empirics that demonstrates the phenomenon and provides some insight to help develop future
research. For this we surveyed R&D managers of MNE’s offshore captives located in India.
12076
21
Survey Data Description
We collected survey data from 142 R&D organizations established by MNEs in India. Survey
respondents were project managers in these centers, who were responsible for the execution and
delivery of large, strategic R&D projects in these companies. The survey was part of a larger
project that was aimed at understanding management of offshore R&D captive centers.
The sampling frame for the survey was obtained from a comprehensive census of captive
R&D centers of MNEs conducted by Zinnov Consulting in 2009. Zinnov’s census comprised an
installed base of nearly 600 captive R&D centers in India, including 452 centers established by
publicly listed MNEs. We performed a preliminary check to ensure whether these centers are
billed as R&D centers by the MNEs but are in fact engaged in routine IT or back-office
operations, and removed such firms from the sampling frame.
The unit of analysis is an R&D or New Product Development (NPD) project, and most
survey questions related to the characteristics and management of this specific project.
Specifically, managers were requested to answer project specific questions with respect to what
they actually practice in the largest and strategically the most important R&D/NPD project for
that captive center. The managers provided a short written summary of the aims of the project
that we used to validate whether the response was indeed in the context of an R&D/NPD project.
We requested and received only one response per MNE for only one of its captive centers.
The survey instrument was designed based on a review of the literature and our interviews.
The instrument was pre-tested with managers to examine content and face validity and remove
ambiguities in interpretation. The insights from this pilot were used to revise the questions. Three
hundred pre-committed surveys were mailed, with follow-up letters five weeks later. We
12076
22
received a total of 163 valid responses. 21 responses did not pertain to a R&D or NPD project,
and were therefore removed from the following analyses.
All respondents were assured that their responses would remain confidential and results
would be reported only in aggregate, thereby, addressing privacy concerns and minimizing
potential bias in self-reported data. There were no systematic differences in industry, firm or task
attributes between the respondent sample and the larger population, suggesting that concerns of
non-response bias were minimal (Armstrong and Overton 1970; Poppo and Zenger 2002). We
also performed Harman’s one factor test to check for common method bias (Podsakoff et. al.,
2003). In the analysis, a single factor did not explain a majority of the variance across all our
variables, suggesting that this may not be a significant concern in our data.
Measures
Confidence in knowledge protection: In our context, we expect managers to strongly rely
on trade secrets to protect their IP. Thus, our first dependent variable was ‘manager’s confidence
in effectively protecting their unpatented knowledge or trade secrets’ and it was measured using
a single-item measure – “Trade Secrets are an effective way of protecting our IP” on a 1 to 7
where 1= strongly disagree and 7 = strongly agree scale. We dichotomized this variable for use
in the t-tests, which is the first part of our analyses. We create an indicator variable, with
confidence in trade secrets taking a value of one if managers strongly agreed with the above
statement by scoring 6 or 7 in the above scale, and zero otherwise. In our sample, 108 managers
(76% of the data) had strong confidence in their trade secret protection, whereas 34 managers
(24% of the data) did not. We should note that in our empirical context, relying on legal
mechanisms in order to protect trade secrets is unrealistic; therefore managers who are confident
of the effectiveness of protecting their IP via trade secrets necessarily have to have put in place
12076
23
organizational mechanisms that deter IP leakage. Note that managers are answering this question
in the context of protecting IP in the specific project that they are currently working on, and not
their general belief about how IP could/should be protected.
We tested the robustness of this variable with another single item measure that is worded to
closely relate to the specific threat of leakage we identified in our context, i.e., the mobility of
R&D employees between firms. Managers rated their agreement, again on a 1-7 scale with the
following statement – “we have put in place effective safeguards to protect our knowledge if an
employee leaves the organization”. We again dichotomized this variable as described above,
with 104 managers showing strong confidence in protecting their trade secrets and 38 managers
(27% of the data) that did not.
For subsequent regression analyses we use a continuous measure of confidence in
protecting knowledge by standardizing these variables. Our results are qualitatively identical for
analyses using either of these measures or if we used an average of these two measures.
Project Riskiness with respect to Knowledge Leakage: The characteristics of the project
and the knowledge involved may influence how much risk there is of that knowledge leaking to
competitors. Thus, we created a second dependent variable to reflect this aspect. Based on our
field work, we measured it using the following four constructs. (1) Generality of the project,
which measures whether the project knowledge is generally useful to competitors. Generally
useful projects have a higher risk of leakage than knowledge that is only useful to the focal firm
and not useful to competitors. (2) Extent of use of firm’s prior knowledge, which measures the
extent to which the project uses the pre-existing knowledge from the firm. When a project uses a
firm’s pre-existing knowledge, the risk of leakage is higher since it places in jeopardy not just
the current knowledge, but also pre-existing knowledge that so far may not have been available
12076
24
to its competitors. (3) Radicalness of the project, which measures the extent to which the current
project is expected to generate a major technical advance. Radical innovations maybe more
valuable and therefore may attract more interest from competitors. (4) The codifiability of the
project, which measures the extent to which the project’s knowledge is codified. The more
codified the knowledge base, the more easily it can leak to competitors. The specific measures
for each of these constructs are also provided in Table 1. We used a 7-point scale to assess the
items; in general, the higher the levels of these four constructs, the more risky a project is from a
knowledge leakage perspective, and therefore captive center managers should be less confident
in protecting their IP.
Organizational Mechanisms to protect IP: The qualitative study identified a set of
organizational practices that are likely to render it more difficult for a competitor to access
critical knowledge from the focal firm by means of poaching its employees. These practices
include (a) involvement of personnel from headquarters in order to prevent IP leakage, (b) high
levels of task interdependence across locations, (c) performing projects that require the use of
firm-specific assets, (d) high levels of detailed information sharing across locations regarding the
project, (e) control by headquarters of critical aspects of the project, and (f) implementing
internal controls to restrict knowledge access. We measured each of these constructs using multi-
item scales. All the scales displayed a Cronbach’s alpha above 0.70, which is generally used as a
cut-off for scale reliability (Nunally and Bernstein, 1994; Srikanth and Puranam, 2011). The
items that we used to measure the extent of adoption of these mechanisms in the project along
with their respective scale reliability coefficients are shown in Table 1.
INSERT TABLE 1 HERE
12076
25
Control Variables: In our regression analyses we controlled for several objective
characteristics of the captive center such as its age and size as number of employees, R&D
intensity of the parent firm, and the sector coded from the project description filled out in the
survey. The sectors include automotive, chemicals, pharmaceuticals, semiconductor, software,
and telecom, the ‘other’ category being the left out category. We also controlled for whether the
project was aimed at the global or local markets using an indicator variable from our survey.
Analysis
Since our study is exploratory regarding how firms protect their unpatented knowledge, our
analysis is aimed at showcasing plausible evidence for our propositions. The perspective guiding
our empirical analysis is that implementing certain organizational mechanisms or project
management practices makes leakage of knowledge more difficult. Therefore, we first examine
whether the adoption of these mechanisms by projects is correlated with manager’s confidence in
protecting their proprietary unpatented knowledge. For this purpose, we first compare the mean
levels of deployment of organizational mechanisms across the different sub-samples (high-
confidence vs. low-confidence) and test whether these means are statistically different from each
other. We also show that the use of these organizational mechanisms positively correlates with
captive centers performing projects whose knowledge is at greater risk of leakage.
Next we present a canonical correlation analysis between managers’ confidence in
protecting their knowledge and adoption of the set of organizational mechanisms identified
above. Canonical correlation analysis is appropriate when attempting to show a relationship
between two ‘sets of variables’, each of which is composed of multiple correlated items. This
analysis is considered most appropriate when the constructs of interest cannot be measured or
expressed by a single variable in isolation, which may at best only be indicative of only a part of
12076
26
the overall relationship (Carmeli and Tishler, 2004). For example, our argument is that a ‘set of
organizational mechanisms’ help protect a firm’s unpatented knowledge – though each project
may adopt only a sub-set of these mechanisms. Under these conditions, regression analysis may
not show relationships between any of these single items in the LHS and RHS. Canonical
correlation analysis has been successfully employed in prior strategy research when tackling
research questions posing the same empirical characteristics as our own. For example, Szulanski
(1996) shows a canonical correlation between different measures of difficulty of transferring
knowledge within the firm and the presence of multiple factors that give rise to internal
stickiness. Carmeli and Tishler (2004) investigated the relationship between intangible
organizational elements and organizational performance using a similar analysis technique.
Finally, we use a linear regression analysis to show the association between managers’
confidence in protecting their proprietary knowledge and the implementation of multiple
organizational mechanisms after controlling for a variety of project and captive characteristics
such as captive age, captive size, R&D intensity and industry. We also use a regression analysis
to show that captive centers that are more confident of protecting their IP perform more risky
projects (from a knowledge leakage perspective). For these regressions, we created the main
dependent and independent variables by using the loadings obtained in the canonical correlations
between the constructs and their canonical covariates. In robustness checks, instead of creating
these variables using a canonical covariates, we computed them using a principal components
analysis. Our results are unchanged.
RESULTS
Our argument is that unpatented (as well as patented) knowledge is protected by employing the
organizational mechanisms we identified in the field study, which make it more difficult for an
12076
27
employee to leak critical aspects of knowledge to competitors. In other words, in projects that
involve high levels of implementation of the identified organizational mechanisms, managers
should be more confident of protecting their IP when compared to projects that do not involve
these organizational mechanisms. Table 2 shows the correlations between our variables. Table 3a
shows the mean levels of implementation of the organizational mechanisms of interest in
offshore R&D projects for the high and low confidence groups. Simple t-tests of the difference
between these groups for the organizational mechanisms suggest that our propositions are
empirically plausible. Our results are qualitatively identical regardless of which measure we use
to identify the more confident vs. less confident groups.
PLEASE INSERT Table 2, TABLE 3a and 3b
As a robustness check, we also performed regression analyses for these relationships
controlling for a set of captive and project characteristics (please refer A1 in the Appendix). We
found that with the exception of control of headquarters, all the other organizational mechanisms
we discussed have a positive and significant influence on managers’ confidence.
It is plausible that managers who are more confident about protecting their IP perform very
different projects than those who are less confident, and uncorrelated with the difference in the
deployment of the organizational mechanisms we argue for. For example, it is plausible that
managers of projects that involve innovations that are firm-specific and not valuable to
competitors are more confident about protecting their IP on average, since competitors are
simply not interested in imitating such knowledge. Similarly, managers of projects that involve
high levels of codified knowledge are likely to be less confident of protecting their IP on
average. We test for the possibility that more confident managers manage projects whose
characteristics are such that on average that knowledge is less likely to leak.
12076
28
From our field work we identified four characteristics that make a project more risky from
a knowledge leakage perspective as described in the previous section. Similar to the above
analyses, we perform simple t-tests of difference in means of the project characteristics that are
likely to be associated with greater risk of IP leakage. These results are shown in Table 2b. We
find that contrary to intuition, more confident managers are involved in projects that have an
inherently higher risk of knowledge leakage. More confident managers are likely to manage
projects that are likely to produce innovations that are generally useful to other firms rather than
being firm-specific, more radical innovations, use more prior knowledge of the MNE, and use
more codified knowledge. These results suggest that the deployment of organizational
mechanisms discussed above leads to greater confidence in protecting knowledge, which in turn
enables these managers to offshore projects that are at a greater risk of IP leakage.
As we expected, in the regression analyses that we did as robustness check (Please refer to
Table A2 in the Appendix), when all the organizational mechanisms are simultaneously entered
as independent variables, most of them have a significant relationship with project riskiness but
some do not. These results are also dependent on the measure used for performing the analyses.
As we argued earlier, this suggests that some of the mechanisms may be related to each other,
and also some may be more important than others, generally, as well as in a context specific
manner. Since we do not have a good theoretical basis for understanding these effects, we
perform an analysis that attempts to understand the joint effect of these organizational
mechanisms as a group rather than investigate them singly.
Next, we present a canonical correlation analysis between the set of organizational
mechanisms and confidence in protecting trade secrets. Specifically, we enter all the
organizational mechanisms as correlates in the RHS and the two measures of managers’
12076
29
confidence in protecting IP in the LHS. The results of this analysis are shown in Table 3, which
shows that there is a significant correlation of 0.70 between these two sets of variables.
INSERT TABLE 4 HERE
Examining the loadings, we see that both measures of confidence have a high loading on
the co-variate measuring confidence in IP protection. On the RHS, we see that some of the co-
variates have high loadings and others have much lower loadings. Specifically, headquarters
control of the project has a very low loading of 0.30 and interdependence between locations has
a medium loading of 0.56. All the other co-variates have high loadings on deployment of
organizational mechanisms that protect trade secrets. In congruence with the regression analyses
reported earlier, this suggests that headquarters control of the project may not be an important
factor in protecting knowledge when compared to the other mechanisms. Similarly, it is plausible
that interdependence between locations is not relevant to projects that are mostly performed from
a single location, even though such projects may deploy other organizational mechanisms such
as using specific assets in their execution and putting in place internal control for knowledge
access. The strength of the canonical correlation analysis is precisely where multiple
mechanisms may be used to achieve desired ends, with some of them being more important than
others, and when the mechanisms themselves may have complex interrelations as complements
and substitutes in achieving the desired results (Hair et al, 2009). In sum, this analysis suggests
that deployment of the organizational mechanisms we argued for is associated with greater
confidence in protecting knowledge. Further research is required, however, to understand the
relative importance of each of these mechanisms and how they are inter-related. In robustness
tests, we checked the canonical correlation between the organizational mechanisms as a set and
12076
30
the set of project characteristics that make project more prone or risky with respect to knowledge
leakage. The correlation and factor loadings are shown in Table A3 in the Appendix.
Finally, we perform regression analyses to understand whether the association between the
extent of usage of organizational mechanisms, managers’ confidence in knowledge protection
and knowledge leakage risk of the project. Table 4 shows the results. In model 1, the dependent
variable is the managers’ confidence in knowledge protection, and we see that usage of
organizational mechanisms is positively and significantly related to confidence. In models 2-4,
the dependent variable is the knowledge leakage risk of the project. In model 2 we see that
deployment of organizational mechanisms is positively and significantly related to project
leakage risk, and in model 3, we see that managers’ confidence is positively and significantly
related to project leakage risk. Finally, model 4 tests whether the manager’s confidence in IP
protection mediates the relationship between the deployment of organizational mechanisms and
the knowledge leakage risk. Model 4 suggests precisely such a partial mediation effect.
Comparing the coefficient of organizational mechanisms between models 2 and 4, we see that it
is significantly smaller in the expected direction (difference =0.23; std. err. =0.08; p-val <0.01).
This lends us some confidence in our conjecture that usage of these organizational mechanisms
makes it harder for competitors to expropriate IP from the offshore captive centers. Though these
analyses by themselves cannot be used to establish causality, our fieldwork suggests that these
relationships may well be causal.
INSERT TABLE 5 HERE
DISCUSSION
Our empirical analysis and results provide some indicative insights into understanding
mechanisms that organizations can use to protect leakage of unpatented knowledge to
12076
31
competitors - which, in turn, enhances their ability to better appropriate the returns associated
with such knowledge. The mechanisms we explore here are particularly relevant when the
knowledge in question is distributed and resident in geographically distant locations where
leakage of unpatented knowledge can potentially happen through the mobility of employees who
possess most of this knowledge.
Our results shows that a ‘collection or set’ of mechanisms related to how a firm
configures, manages and controls its R&D or NPD projects is positively associated with a firm’s
ability to protect its unpatented knowledge that is developed or used in that project from leakage.
These mechanisms include ‘configuring projects’ such that the activities involved are
interdependent in nature and require investments in project/firm specific assets’, ‘managing
projects’ such that they require high level of information sharing between different constituents
and have internal controls for IP capture, and ‘ controlling projects’ such that they call for greater
involvement of headquarter personnel in the work and having headquarters formally control and
coordinate all project work.’ We find that that greater usage of these mechanisms is not only
linked positively to a firm’s (or its managers) confidence in protecting its unpatented IP from
leakage to rivals, but also associated with undertaking projects where the risk of knowledge
leakage is high given the nature/attributes of project concerned. In other words, it seems firms
are indeed more willing to undertake projects that involve high risk of leakage of unpatented
knowledge if they have in place internal mechanisms and practices to potentially prevent or
minimize the leakage of the associated knowledge.
Our results also indicate that while a ‘set of organizational’ mechanisms collectively
helps in greater protection of IP/knowledge leakage, not all of the mechanisms are equally
important in that regard - some mechanisms (e.g. asset specificity or involvement of HQ
12076
32
personnel in the project) seem to be far more critical to preventing knowledge leakage than
others; and some mechanisms, such a HQ control of the project, have no significance at all. It’s
plausible that in this case, HQ control of the project ceases to be important because of the high
involvement of HQ personnel in the project anyways. Future research can delve much deeper to
examine the differential impact of various organizational mechanisms and the conditions under
which their salience becomes more critical. Our analysis also helps uncover some subtle and
counter-intuitive insights about the kind of projects firms might be willing to undertake even in
off-shoring of R&D and NPD settings. Generally one might believe that given the threat of
knowledge leakage in such cases, managers might not be very willing to take on offshore
projects that are more prone to IP leakage – but, interestingly, we find that is indeed not always
the case. High levels of deployment of some of the organizational mechanisms we have
described increases managers’ confidence in their ability to prevent or minimize leakage of
unpatented knowledge in their projects, which in turn even enables them to take on projects that
exhibit greater risk of IP leakage. This finding is critical because it not only helps understand
what firms can do to minimize the leakage of valuable but unpatented knowhow, but also then
sheds light on how firms can even potentially undertake very knowledge intensive projects in
distributed settings without the fear of easily losing the associated knowhow to rivals.
Overall, our research makes contributions to several streams of work. First, at a broad
level we contribute to the scant but increasing body of research that focuses on how firms can
better appropriate value from the valuable knowledge they create by protecting the leakage and
imitation of this knowledge to rivals. While a vast body of work has examined the important role
of patents in this regard, it is well accepted that not all knowledge is necessarily patentable and
even if it is so, the ability of patents to provide protection against leakage and imitation is linked
12076
33
to the strength and efficacy of the IPR protection regime in a given setting. Thus, it is equally
critical to examine other factors, including internal organizational factors, that can help firms
protect leakage and imitation of knowledge that by its nature is either not patentable or amenable
to sufficient protection by patents – and our paper takes an important step in adding to that body
of work. Second, our work complements work on the role of patents as mechanism for
knowledge protection in another important way. Some scholars have recently conceptualized the
notion of ‘value appropriation as an organizational capability’ (Reitzig and Puranam, 2009) – but
done so primarily by examining organizational antecedents of such a capability in the context of
obtaining protection through patents. Our work complements and enriches this concept of ‘value
appropriation as an organizational capability’ by also outlining organizational antecedents and
processes that enable protection of unpatented knowledge in firms – and thus provide a more
comprehensive perspective on what such a capability might entail. Third, we also add to the
stream of work that examines the role of knowledge creation in distributed global settings
wherein firms undertake activities and projects in multiple geographies to leverage international
differences (Zhao, 2006) – but in doing so, expose themselves to appropriation hazards due to
the differential threats to the leakage of knowledge they have developed. If firms deploy some of
the mechanisms described in our study, they can potentially minimize at least the leakage of the
unpatented knowledge that resides within their employees in dispersed geographies, including
those where the leakage of this knowledge through employee mobility is quite high.
12076
34
TABLE 1 – MEASURES
Unless indicated, all scales were measured using the following format:
Please indicate the extent to which you agree with the following statements (1-7 scale: disagree–agree):
DV: Confidence in Trade Secret Protection
Items Not Confident (1-5) Confident (6-7) Secrets are an effective way of protecting our IP 34 (24%) 108 (76%) We have put in place effective safeguards to protect our
knowledge if an employee leaves the organization 38 (27%) 104 (73%)
DV: Characteristics that increase knowledge leakage risk of the project
1. Generality of the (expected) innovation
If a competitor could access this knowledge, our competitive advantage could be lost.
2. Usage of prior firm internal knowledge
This project involves substantial level of knowledge that is proprietary to our company.
3. Radicalness of the (expected) innovation
The output innovation of this project represents a major technological advance.
4. Usage of codified knowledge (2 items; alpha: 0.77)
(1) We had extensive documentation that described all the critical parts of this project; (2) Most of the
training required to work in this project was obtained from our manuals.
IV: Organizational mechanisms used to protect knowledge from leakage
1. Involvement of headquarters personnel (2 items, alpha: 0.87)
(1) IP concerns determine the level of involvement of HQ scientists in this project; (2) IP concerns
determine the level of involvement of HQ managers in this project
2. Interdependence (2 items, alpha: 0.91)
(1) Changes to the work approach or direction in other locations led to changes in work on the project in
our location; (2) There was frequent need to talk to personnel in other locations about their work on the
project so we could adjust our direction.
3. Project Asset specificity (5 items, alpha: 0.86)
(1) This project requires investments in skills that are not easily redeployable in other projects; (2) This
project requires investments in equipment/capital that are not easily redeployable in other projects; (3)
This project requires investments in software and similar technologies that are not easily redeployable in
other projects; (4) Vendors performing this project need to be collocated with us in order to perform
effectively; (5) Vendors have to invest in routines and procedures customized to your company in order to
work effectively.
4. Headquarters Control of Project (3 items, alphs: 0.91)
Please indicate who makes decisions in your center regarding (1-5 scale: subsidiary fully autonomous –
complete control by HQ):
(1) Product design; (2) Documentation standards; (3) Frequency and format of reports for R&D results
12076
35
5. Extent of information sharing (6 items, alpha: 0.84)
(To what extent were the following project information shared between your center and other locations;
1-not shared at all to 7-shared very frequently):
(1) Quality information; (2) Schedule and delivery information; (3) Detailed cost information; (4)
Marketing information; (5) Proprietary technical information; (6) Design information
6. Internal controls for IP capture (2 items, alpha: 0.83)
(1) Implement internal controls such that only very selected few have access to all information relating to
an innovation; (2) Implement strong mechanisms for knowledge capture.
Table 2: Correlation Table
1 2 3 4 5 6 7 8
1 Confidence 1.00
2 Project Risk 0.71 1.00
3 HQ involvement 0.61 0.66 1.00
4 Interdepedence 0.39 0.35 0.46 1.00
5 Asset Specificity 0.61 0.66 0.69 0.52 1.00
6 HQ Control 0.21 0.22 0.26 0.20 0.35 1.00
7 Information Sharing 0.48 0.43 0.35 0.43 0.58 0.19 1.00
8 Internal controls 0.57 0.64 0.74 0.45 0.72 0.36 0.35 1.00
All coefficients are significant at the p<0.05 level.
12076
36
Table 3a: Difference in means on usage of suggested organizational mechanisms that reduce the
risk trade secret leakage among managers who are more confident of protecting their IP versus
managers who are less confident of protecting their IP.
High
Confidence
Low
Confidence
Difference†
Involvement of headquarters personnel 0.12 -1.02 1.14 (0.21)***
Interdependence between project locations 0.04 -0.77 0.81 (0.21)***
Projects requiring investments in specific assets 0.03 -0.80 0.83(0.13)***
Headquarters control of project 0.08 -0.32 0.40 (0.18)**
Extent of information sharing 0.09 -0.82 0.91 (0.20)***
Internal controls for IP capture 0.13 -0.79 0.92 (0.19)***
†Standard errors in parentheses; ***, ** and * are significant at p<0.01, p<0.05 and p<0.1 respectively.
Table 3b: Difference in means on performing projects whose knowledge is at a higher risk of IP
leakage among managers who are more confident of protecting their IP versus managers who are
less confident of protecting their IP.
High
Confidence
Low
Confidence
Difference†
Innovation is generally valuable to all firms (not
firm-specific)
0.05 -0.69 0.74(0.21)***
Extent of usage of firm’s prior knowledge in the
project
0.24 -1.0 1.24 (0.19)***
Radicalness of expected innovation 0.24 -0.86 1.10 (0.20)***
Usage of more codified knowledge in the project 0.12 -0.75 0.88 (0.20)***
†Standard errors in parentheses; ***, ** and * are significant at p<0.01, p<0.05 and p<0.1 respectively.
Table 4: Canonical Correlation
(loadings of individual elements on to respective canonical covariates shown in parentheses)
12076
37
Table 5: Managers’ confidence in knowledge protection partially mediates the impact of
implementing organizational mechanisms on (the propensity for the captive center to engage in)
projects that have a high risk of knowledge leakage.
DV: Confidence in
Knowledge
protection
DV: Project knowledge characteristics
(higher the DV – the more risk of knowledge leakage)
MODEL 1 MODEL 2 MODEL 3 MODEL 4
Managers’ Confidence 1.15(0.11)*** 0.86(0.12)***
Org Mechanisms 0.28(0.04)*** 0.49(0.06)*** 0.26(0.06)***
Size 0.00(0.00) -0.00(0.00) 0.00(0.00) 0.00(0.00)
Age 0.00(0.12) 0.11(0.20) 0.13(0.18) 0.11(0.17)
Global product 0.38(0.19)** -0.01(0.30) -0.45(0.28) -0.33(0.26)
Low R&D Intensity 0.09(0.32) -0.47(0.52) -0.31(0.47) -0.55(0.44)
Med R&D Intensity -0.01(0.28) -0.39(0.45) -0.12(0.41) -0.38(0.39)
Automobile -0.15(0.40) -0.49(0.65) -0.47(0.59) -0.36(0.55)
Chemicals -0.02(0.53) 0.14(0.86) -1.27(0.69) 0.15(0.73)
Pharmaceuticals 0.14(0.37) -0.86(0.60) -1.45(0.54) -0.98(0.51)
Semiconductors -0.02(0.32) -0.33(0.53) -0.25(0.48) -0.30(0.45)
Software -0.12(0.26) 0.08(0.42) 0.28(0.38) 0.19(0.36)
Telecom -0.11(0.38) -0.30(0.62) 0.03(0.56) -0.21(0.53)
Constant -0.10(0.42) 0.34(0.68) 0.06(0.61) 0.42(0.58)
N 142 139 139 139
F 9.9*** 11.43*** 16.1*** 18.4***
Adj R2 0.44 0.47 0.57 0.62
Standard errors in parentheses; ***, ** and * are significant at p<0.01, p<0.05 and p<0.1 respectively.
Difference across models 2 and 4 on effect of org mechanisms: 0.23(0.08)***
12076
38
REFERENCES
Alcacer J. 2006. Location choices across the value chain: how activity and capability influence
collocation. Management Science52(10): 1457–1471.
Alcácer J, Zhao M. 2012. Local R&D strategies and multi-location firms: the role of internal linkages.
Management Science 58(4): 734–753.
Almeida P. 1996. Knowledge sourcing by foreign multinationals: patent citation analysis in the US
semiconductor industry. Strategic Management Journal 17: 155–165.
Almeida P, Kogut B. 1999. Localization of knowledge and the mobility of engineers in regional
networks. Management Science 45(7): 905–917.
Almeida, P., Song, J., and Grant, RM. 2002. Are firms superior to alliances and markets? An empirical
test of cross-border knowledge building. Organization Science. 13(2): 147-161.
Alnuaimi T, George P, Puranam P. 2012a. Emerging economies as sources of innovation? Evidence from
Indian and Chinese R&D subsidiaries of US firms. Working paper, Imperial College, London.
Alnuaimi T, Singh J, George G. 2012b. Not with my own: long-term effects of cross-country
collaboration on subsidiary innovation in emerging economies versus advanced economies.
Journal of Economic Geography 12(5): 943–968.
Armstrong, J.S., T. Overton. 1970. Brief vs. comprehensive descriptions in measuring intentions to
purchase, Journal of Marketing Research, 8, 114-117.
Arora, A. 1996. Contracting for tacit knowledge: The provision of technical services in technology
licensing contracts. Journal of Development Economics 50(2): 233–256.
Arrow K. 1962. Economic welfare and the allocation of resources for innovation. In The rate and
direction of inventive activity. Princeton, NJ: NBER, pp609-625.
Barney, J. 1986. Strategic factor markets: Expectations, luck and business strategy. Management Science.
32(10): 1231-1241.
Barney, J. 1991. Firm resources and sustained competitive advantage. Journal of Management. 17(1): 99-
120.
Carmeli, A., Tishler, A. 2004. The relationships between intangible organizational elements and
organizational performance. Strategic Management Journal. 25(13): 1257-1278.
Chesbrough H. 2003. Open Innovation: The New Imperative for Creating and Profiting from Technology.
Harvard Business School Press: Boston, MA.
Cohen, W.M., R.R. Nelson and J.P. Walsh. 2000. Protecting their intellectual assets. Appropriability
conditions and why U.S. manufacturing firms patent (or not). NBER Working Paper No. 7552,
February 2000.
Dierickx, I., Cool, K. 1989. Asset stock accumulation and sustainability of competitive advantage.
Management Science. 35(12): 1504-1511.
Economist, 2010. The world turned upside down. Special report on innovation in emerging markets. April
2010. http://www.economist.com/node/15879369.
Farrell, D. 2005. Offshoring: value creation through economic change. Journal of Management Studies.
42(3): 675-683.
Freidman DD, Landes WM, Posner RA 1991 Some economics of trade secret law. Journal of Economic
Perspectives. 5: 61-72.
Giaratana MS Mariani M. 2013. The relationship between knowledge sourcing and fear of imitation.
Strategic Management Journal. 35(8): 1144-1163.
Hair JF Black WC Babin BJ Anderson RE 2009. Multivariate Data Analysis. Prentice Hall.
Harding, VM. 1967. Trade secrets and the mobile employee. The Business Lawyer. 22(2): 395-411.
Helfat CE, Quinn JB. 2006. Open innovation: the new imperative for creating and profiting from
technology. Academy of Management Perspectives. 20:86–88.
Hegde D, Hicks D. 2008. The maturation of global corporate R&D: evidence from the activity of U.S.
foreign subsidiaries. Research Policy 37(3): 390–406.
12076
39
Hu, AG., Jefferson, GH. 2009. A great wall of patents: What is behind China’s recent patent explosion?
Journal of Development Economics.
Hyde A. 2003. Working in Silicon Valley: Economic and Legal Analysis of a High-velocity Labor Market.
M.E. Sharpe: Armonk, NY.
Kogut B, Zander U. 1993. Knowledge of the firm and the evolutionary theory of the multinational
corporation. Journal of International Business Studies 24(4): 625–645.
Kotha, S., K. Srikanth. 2013. Managing a Global Partnership Model: Lessons from the Boeing 787
‘Dreamliner’ Program. Global Strategy Journal 3(1) 41–66.
Kuemmerle W. 1999b. The drivers of foreign direct investment into research and development: an
empirical investigation. Journal of International Business Studies 30(1): 1–24.
Kumar N, Puranam P. 2012. India Inside: The Emerging Innovation Challenge to the West. Harvard
Business Review Press: Boston, MA.
Leonardi, P.M., Bailey, D. 2008 Transformational Technologies and the Creation of New Work Practices:
Making Implicit Knowledge Explicit in Task-based Offshoring. MIS Quarterly. 32: 411-436.
Lewin, A. Y, S. Massini, and C. Peeters. 2009. Why are companies offshoring innovation? The emerging
global race for talent. Journal of International Business Studies 40(6): 901–925.
Liebeskind JP. 1996. Knowledge, strategy and the theory of the firm. Strategic Management Journal.
17(S2): 93-108.
Liebeskind JP 1997. Keeping organizational secrets: Protective institutional mechanisms and their costs.
Industrial and Corporate Change. 6(3): 623-663.
Liebeskind JP, Oliver AL, Zucker LG, Brewer MB 1996. Social networks, learning, and flexibility:
Sourcing scientific knowledge in new biotechnology firms'. Organization Science 7(4):429-442.
Lippman, SA., Rumelt, RP. 1982. Uncertain imitability: An analysis of interfirm differences in efficiency
under competition. The Bell Journal of Economics. 13(2): 418-438.
Mahoney JT and Pandian JR 1992. The resource based view within the conversation of strategic
management. Strategic Management Journal. 13(5): 363-380.
Mani D., Srikanth, K., Bharadwaj, A. 2014. Efficacy of R&D work in offshore captive centers: An
empirical study of task characteristics, coordination mechanisms and performance. Information
Systems Research.
Manning, S., Massini, S., Lewin, AY. 2008. A dynamic perspective on next-generation offshoring: The
global sourcing of science and engineering talent. The Academy of Management Perspectives,
22(3):35-54.
Mansfield E. 1985. How rapidly does new technology leak out? Journal of Industrial Economics. 34:
217-224.
Marx M. 2011. The firm strikes back: non-compete agreements and the mobility of technical
professionals. American Sociological Review 76(5): 695–712.
Marx M, Strumsky D, Fleming L. 2009. Mobility, skills, and the Michigan non-compete experiment.
Management Science 55(6): 875–889.
Mudambi R. 2011. Hierarchy, coordination, and innovation in the multinational enterprise. Global
Strategy Journal 1(3-4): 317–323.
Mudambi, R., and M. Venzin. 2010. The strategic nexus of offshoring and outsourcing decisions. Journal
of Management Studies 47(8): 1510-1533.
Nunnally, J. C., I.H. Bernstein. 1994. Psychometric Theory (3rd ed.) New York: McGrawHill.
Parmigiani A, Mitchell W. 2009. Complementarity, capabilities, and the boundaries of the firm: the
impact of within-firm and interfirm expertise on concurrent sourcing of complementary
components. Strategic Management Journal. 30(10): 1065–1091.
Peteraf, MA. 1993. The cornerstones of competitive advantage: A resource based view. Strategic
Management Journal. 14(3): 179-191.
Podsakoff, PM., MacKenzie, SB., Lee, JY., Podsakoff, NP. 2003. Common method biases in behavioral
research: a critical review of the literature and recommended remedies. Journal of Applied
Psychology, 88(5): 879-903.
12076
40
Poppo, L., T. Zenger. 2002. Do Formal Contracts and Relational Governance function as Substitutes or
Complements? Strategic Management Journal.23: 707-725.
Porter, ME. 1996. What is strategy? Harvard Business Review.
Quan, X., Chesbrough, H. 2010. Hierarchical segmentation of R&D process and intellectual property
protection: Evidence from multinational R&D laboratories in China. IEEE Transactions on
Engineering Management, 57(1), 9-21.
Reitzig, M., Puranam, P. 2009. Value appropriation as an organizational capability. The case of IP
protection through patents. Strategic Management Journal. 30(7): 765-789.
Rosenkopf L, Almeida P. 2003. Overcoming local search through alliances and mobility. Management
Science 49(6): 751–766.
Rumelt RP. 1984. Towards a strategic theory of the firm. Competitive Strategic Management, 26: 556-
570.
Saxenian, 1994 Regional advantage: Culture and competition in Silicon Valley and Route 128.
Cambridge, MA: Harvard University Press.
Singh J. 2008. Distributed R&D, cross-regional knowledge integration and quality of innovative output.
Research Policy 37(1): 77–96.
Srikanth, K., P. Puranam. 2011. Integrating Distributed Work: Comparing Task Design, Communication,
and Tacit Coordination Mechanisms, Strategic Management Journal, 32: pp. 849-875.
Srikanth, K. and P. Puranam. 2014. The firm as a coordination system: Evidence from software services
offshoring. Organization Science. (forthcoming). dx.doi.org/10.1287/orsc.2013.0886.
Szulanski, G. 1996. Exploring internal stickiness: Impediments to transfer of best practice within the firm.
Strategic Management Journal. 17(S2): 27-43.
Teece, D. J. 1986. Profiting from technological innovation: Implications for integration, collaboration,
licensing and public policy. Research Policy 15(6): 285–305.
Thursby, J., and M. Thursby. 2006. Here or There? A Survey of Factors in Multinational R and D
Locations—Report to the Government-University-Industry Research Roundtable. National
Academy Press.
Wood, J. 2012. Coming of Age: Asia’s evolving R&D landscape. The Economist Intelligence Unit.
August 20127.
Williamson OE 1985 The Economic Institutions of Capitalism. New York, NY: The Free Press.
Williamson OE 1991 Comparative economic organization: The analysis of discrete structural alternatives.
Administrative Science Quarterly. 36(2): 269-296.
Zhao, M. 2006. Conducting R&D in countries with weak intellectual property rights protection.
Management Science 52(8): 1185
Zucker, LG., Darby, MR. 1996. Star scientists and institutional transformation: Patterns of invention and
innovation in the formation of the biotechnology industry. Proceedings of the National Academy
of Sciences, 93(23): 12709-12716.
Zucker, LG, Darby, MR, Brewer, MB, Peng, Y. 1995. Collaboration structure and information dilemmas
in biotechnology: Organizational boundaries as trust production. National Bureau of Economic
Research Working paper No:5199.
7http://www.economistinsights.com/sites/default/files/downloads/Coming%20of%20age%20Asias%20evolving%20
RnD%20landscape_WEB.pdf