Organizational change and the dynamics of innovation: Formal R&D structure and intrafirm inventor networks

• Author: Brian S. Silverman,Nicholas Argyres,Luis A. Rios
• Date: 01 November 2020
• Published date: 01 November 2020
• DOI: http://doi.org/10.1002/smj.3217

RESEARCH ARTICLE

Organizational change and the dynamics of

innovation: Formal R&D structure and

intrafirm inventor networks

Nicholas Argyres

1

| Luis A. Rios

2

| Brian S. Silverman

3

1

Olin School of Business, Washington University in St. Louis, St. Louis, Missouri

2

Wharton School of Business, University of Pennsylvania, Philadelphia, Pennsylvania

3

Rotman School of Management, University of Toronto, Toronto, Ontario, Canada

Correspondence

Luis A. Rios, Wharton School of

Business, University of Pennsylvania,

3620 Locust Walk, Philadelphia, PA

19104.

Email: luisrios@upenn.edu

Funding information

Social Sciences and Humanities Research

Council of Canada; The William and

Phyllis Mack Institute for Innovation

Management

Abstract

Research Abstract: Prior research has argued and

shown that firms with more centralized R&D produce

broader innovations, but the organizational mecha-

nisms underlying this relationship are underexplored.

This gap limits our understanding of whether and how

formal R&D structure can be used as a lever to influ-

ence research outcomes. To address this question, we

study the relationship between formal R&D structure,

internal inventor networks, and innovative behavior

and outcomes. We find that centralization of R&D bud-

get authority increases the connectedness of internal

inventor networks, which in turn increases the breadth

of both innovation impact and technological search.

Surprisingly, decentralization does not have the oppo-

site effect. Our results suggest that changes in formal

structure influence innovation outcomes through

changes in inventor networks, with a lag reflecting

organizational inertia.

Managerial Abstract: Diversified corporations can

organize their R&D functions to be more or less cen-

tralized. Prior research has shown that this organiza-

tional choice is associated with different types of

Nicholas Argyres, Luis A. Rios, and Brian S. Silverman contributed equally to this study and they are listed in

alphabetical order.

Received: 28 May 2018 Revised: 19 January 2020 Accepted: 4 February 2020 Published on: 18 August 2020

DOI: 10.1002/smj.3217

Strat Mgmt J. 2020;41:2015–2049. wileyonlinelibrary.com/journal/smj © 2020 John Wiley & Sons Ltd 2015

innovative outcomes. But what happens when a corpo-

ration changes its level of R&D centralization? This

paper suggests that centralization of R&D gradually

leads to new patterns of collaboration among inventors,

which in turn will be associated with innovations that

draw on and influence a wider range of technologies.

However, future work is needed to understand why

decentralization does not appear to have the opposite

effect.

KEYWORDS

innovation, organizational structure, R&D, social networks

1|INTRODUCTION

The unquestionable importance of innovation has spurred a broad and diverse literature in stra-

tegic management. A small but influential stream of this work has argued that a firm's formal

R&D organizational structure should affect the type of innovation produced (Hounshell &

Smith, 1989; Kay, 1988; Teece, 1996). Consistent with this, empirical investigations have found

that distinct organizational forms are indeed associated with different patterns of innovation

(Argyres, 1996; Argyres & Silverman, 2004; Arora, Belenzon, & Rios, 2014). However, these lat-

ter studies have been limited to showing static, cross-sectional associations between organiza-

tion structure and innovation outcomes. We thus lack even preliminary evidence about the

mechanisms or channels underpinning such relationships, which hampers the development of

theory and of normative guidance. This study addresses these limitations by empirically investi-

gating one channel through which changes in formal R&D structure may affect innovation:

changes in the structure of the firm's internal inventor network.

While prior research has studied inventor networks, it has not investigated the ways in

which collaboration networks within a firm might be shaped by choices of formal R&D

structure—a question that is highly relevant for strategic management given its interest in the

active management of innovation. For example, prior studies have focused on relationships

between network attributes and various types of innovative activity, or on how an individual

inventor's network position impacts her and/or her colleagues' innovation (Grigoriou &

Rothaermel, 2017; Obstfeld, 2005; Reagans & Zuckerman, 2001). Indeed, relatively few studies

have documented the determinants and consequences of network change even in non-

innovation settings (Ahuja, Soda, & Zaheer, 2012).

To address this gap, we examine the relationship between changes in firms' formal R&D

organization structures (which we observe through changes in the loci of R&D budget author-

ity) and subsequent changes in firms' internal co-patenting networks. We also examine the rela-

tionship between changes in these internal networks and key firm-level innovation outcomes.

Using whole-network topology measures (Amburgey, Al-Laham, Tzabbar, & Aharonson, 2008),

we capture the structure of collaboration at the level of the entire organization, rather than

from the perspective of individual inventors as much prior work has done. While we are not

able to empirically establish causation in the relationships we study, we are able to track a large

2016 ARGYRES ET AL.

subset of all U.S. innovation-oriented public companies over a 20-year period, and to document

important patterns in the relationships between budgetary control, inventor networks, and

innovative output. To our knowledge, no prior work has systematically documented the rela-

tionships between formal organization structure and collaboration networks, despite recent

calls to do so (McEvily, Soda, & Tortoriello, 2014).

Our findings are as follows. First, increased centralization of R&D budget authority is asso-

ciated with the emergence of new collaborations among researchers who had not previously

patented together. Specifically, such centralization is associated with greater connectedness of

the firm's coinvention network; that is, a larger fraction of researchers become connected to

other researchers through coinvention of patented innovations (Amburgey et al., 2008).

1

Sec-

ond, greater network connectedness in turn is associated with an increase in the breadth of

innovative search and impact exhibited by a firm's patents. Surprisingly, we find that the

reverse may not hold-decentralization does not seem to have much of an effect on either net-

works or innovation. Finally, we provide novel facts that speak to the mechanisms through

which network connectedness mediates the relationship between changes in formal R&D struc-

ture and innovative outcomes. We find that roughly 13%–18% of the relationship between bud-

getary structure change and differences in innovation output is attributable to changes in the

informal inventor network, and that these patent changes occur in tandem with the evolution

of the coinvention network. Together, these findings suggest that a change in the firm's formal

R&D structure triggers a gradual change in its inventor network, which in turn leads to changes

in the nature of the firm's innovations.

2|THEORY AND HYPOTHESES

We bridge the literatures on R&D organization and innovation networks to suggest two under-

lying mechanisms that may interact to drive the organizational dynamics of innovation. These

mechanisms correspond to two major views of the firm that are not often reconciled: the firm

as an authority-based incentive system in which employees respond to changes in authority

and incentives (Holmstrom & Milgrom, 1994; Simon, 1944; Williamson, 1985); and, the firm as

a system for transferring knowledge via social capital (Nahapiet & Ghoshal, 1998; Zander &

Kogut, 1995). In these views, organization structure affects outcomes through the ways that it

sets incentives (Jensen & Meckling, 1992) and guides knowledge flows (Karim & Kaul, 2015;

Leiponen & Helfat, 2010).

2.1 |The organization of R&D

Prior literature has conceptualized a firm's R&D activities as centralized if lab directors report

to corporate management, and decentralized if they report to divisional management (hybrid

R&D organizations involve both types simultaneously). Researchers in centralized structures

are motivated to produce innovations that benefit the firm as a whole, while those reporting to

divisions or business units are thought to be concerned with division- or unit-specific innova-

tions (Kay, 1988). These motivations presumably reflect a desire to defer to authority in order to

1

We describe our formal measures of network connectedness in the empirical section below.

ARGYRES ET AL.2017