Competitive intelligence and disclosure

• Date: 01 October 2015
• Author: Susan G. Watts,Mark Bagnoli
• DOI: http://doi.org/10.1111/1756-2171.12103
• Published date: 01 October 2015

RAND Journal of Economics

Vol.46, No. 4, Winter 2015

pp. 709–729

Competitive intelligence and disclosure

Mark Bagnoli∗

and

Susan G. Watts∗

Competitive intelligence (CI) activities open new opportunities for firms to acquire and disclose

information. We show that disclosure depends on the relative usefulness of information to the

competing firms and is generally less (more) likely with Cournot (Bertrand) competition and

when firms adopt product differentiationstrategies. When CI costs are independent of information

characteristics, each firm seeks information solely useful to itself and discloses it unless it is a

Bertrand competitor with customer information. Only when the cost advantage is sufficiently

great does each firm seek information useful to itself and its rival.

1. Introduction

In recent years, firms have invested heavily in competitive intelligence (CI) activities.1

These activities are focused on developing systems to generate information and insight about a

firm’s rivals, the firm’s own customers or production costs, and, when necessary, processes for

maintaining secrecy. Much of the focus is on learning and analyzing rival business strategies and

developing appropriate responses, but significant effort is also expended studying the firm’s and

its rivals’ customers and production processes.2As a concrete example, consider the European

unit of Cisco Systems’ adoption of a CI system.3Cisco’s initial CI efforts focused on developing

systems to acquire information about their current customers, industry demand more broadly,

macroeffects on industry demand, and demand for future product characteristics. Subsequently,

Cisco expanded its CI activities to include developing systems to acquire information about

∗Purdue University; mbagnoli@purdue.edu, swatts@purdue.edu.

We thank the Editor,Kathr yn Spier,and two anonymous referees for many helpful comments and suggestions. We also

thank Jennifer Altamuro, Anil Arya, Masako Darrough, Dick Dietrich, John Fellingham, Leslie Hodder, Pat Hopkins,

Rick Johnston, Ranjani Krishnan, Brian Mittendorf, Marc Picconi, Doug Schroeder, Isabel Wang,Rick Young, workshop

participants at The Ohio State University, participants at the August 2011 American Accounting Association annual

meeting in Denver,CO, and our discussant Leonidas Enrique de la Rosa for helpful discussions. Wegratefully acknowledge

the financial support provided by the Krannert Graduate School of Management and Purdue University.

1Fuld(2007) repor ts on the first survey of CI expendituresand estimates that Fortune 1000 firms spent approximately

$1 billion in 2006 with the expectation that they would grow tenfold by2012.

2Standard references to CI activities include Prescott and Miller (2001), Carr (2003), Fuld (2006), Liebowitz

(2006), Davenport and Harris (2007), Fleisher and Bensoussan (2007), and Ayers(2008).

3Details are from an interview of Joost Drieman, Market Intelligence Director Europe (Global Intelligence Alliance,

2008).

C2015, The RAND Corporation. 709

710 / THE RAND JOURNAL OF ECONOMICS

its competitors and information about its own and its competitors’ selling channels with the

intention of acquiring information about the costs of servicing customers.4Interestingly, Cisco’s

Chief Executive Officer, John Chambers, is well known for the quantity and quality of his

voluntary disclosures about product prices, revenues, costs, product development, and general

market conditions provided in the firm’s quarterly earnings press releases, conference calls, and

8-K filings.

The objectives of this article are to extend our understanding of ex ante disclosure policy

choices to information environments that more closely capture the impact of corporate CI ac-

tivities and to study the firms’ strategic incentives to direct those activities to acquire specific

information characteristics and create the resulting equilibrium information environment. We

expect CI activities to produce information structures that differ from those examined previously

in, for example, Gal-Or (1985, 1986), Darrough (1993), and Raith (1996), among others. This

literature examines the private values case, when a firm has private information about its own

production costs, and the common values case, when a firm has private information about a

demand parameter that impacts its own and its rival’s payoffs symmetrically.5We generalize the

information structure to an intermediate values case in which the information is privately useful

to the firm and (at least) partially useful to its rival.

Our generalized information structure leads to disclosure policy choices that are quite

different from those in the prior literature. The most familiar result from this work, often referred

to as the “flipping result,” is that Cournot (Bertrand) competitors adopt a policy of disclosure

(nondisclosure) when theyhave private information about costs but adopt a policy of not disclosing

(disclosing) when they have private information about a common demand parameter (Gal-Or,

1985, 1986; Darrough, 1993). Further, in the private and common values cases, the firms have

a dominant strategy that does not depend on the degree of product heterogeneity or the firms’

underlying level of uncertainty.6

In contrast, in the intermediate values information structure that we introduce, each firm

adopts a policy of disclosure for some sets of information characteristics and a policy of nondis-

closure for others, regardless of whether firms have privateinfor mation about customers or costs

and regardless of whether they are Cournot or Bertrand competitors. Specifically,when firms are

Cournot competitors, a firm adopts an ex ante policy of disclosure if its private information is

sufficiently asymmetrically useful to itself and its rival. In contrast, when the firms are Bertrand

competitors, a firm adopts an ex ante policy of disclosure if its private information about cus-

tomers is sufficiently useful to itself and its rival or if its private information is about costs and is

sufficiently useful to the firm itself. Thus, in general, for parameters for which disclosure occurs

when firms are Bertrand competitors (usually thought of as highly competitive markets), it does

not occur when the firms are Cournot competitors (usually thought of as less competitive mar-

kets), and vice versa. That is, disclosure depends on the information characteristics and whether

the products are strategic substitutes or strategic complements.

An examination of the intermediate values information structure that we introduce also

provides insight into how the degree of product differentiation affects disclosure policy choices.

We show that for Cournot competitors disclosure of either customer or cost information is more

likely in more homogeneous product markets. However, for Bertrand competitors, customer in-

formation is less likely and cost information is more likely to be disclosed in more homogeneous

product markets. Together, these results suggest an empirical regularity that disclosure is generally

less (more) likely in less (more) competitive markets when firms adopt differentiation strategies.

4A very similar pattern in CI activities is described by Rene Loozen, Business Intelligence Manager for Royal

Vopak(Global Intelligence Alliance, 2009).

5Raith (1996) introduces imperfect signals and shows how a firm’s disclosure policy is affected by the quality of

its signal and how correlated the signals the firms receive are.

6These results depend on the assumption that each firm makes its disclosure policy choice independently. If the

firms are permitted to use “quid pro quo” disclosure policies, the “flipping result” no longer characterizes equilibrium

disclosure activity (Kirby,1988; Malueg and Tsutsui, 1998a).

C

The RAND Corporation 2015.