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