Measuring The Impact Of Corporate Governance On Innovation

Updated on

Measuring The Impact Of Corporate Governance On Innovation

Michael J. Cooper

University of Utah – David Eccles School of Business

Anne Marie Knott

Washington University in Saint Louis – John M. Olin Business School

Wenhao Yang

University of Utah

Abstract:

Innovation is vital to firm value. Accordingly a number of recent papers in the finance literature have begun examining the impact of governance and other firm measures on innovation. To date these papers have used patent-based measures to capture innovation. However such measures ignore the 50% of firms who do R&D but don’t patent their innovations. To address that concern, we examine the feasibility of using a more universal measure of innovation from the management literature — the firm’s research quotient (RQ), defined as the firm-specific output elasticity of R&D. We show that RQ is more strongly associated with firm value than previous patent-based measures. Given RQ’s universality advantages and stronger relation with firm value relative to patents, we propose using RQ in future studies of innovation.

Measuring The Impact Of Corporate Governance On Innovation – Introduction

Firm innovation and related investment are vital to firm growth. While the economics and management literatures have examined innovation since the watershed Solow paper (1957), the study of firm innovation is a relatively new topic in the finance literature. The finance fields’ interest stems from the realization that innovation is linked to firm value and that governance and other firm characteristics play substantial roles in a firm’s ability to innovate. Accordingly, research examines the impact on innovation of governance measures such as takeover provisions (Becker-Blease 2011), analyst coverage (He and Tian 2013), institutional ownership (Becker-Blease 2011, He and Tian 2013, Aghion, Van Reenan and Zingales 2013), illiquidity (He and Tian 2013, Gormley and Milbourn 2013, Fang, Tian and Tice 2014), and leverage (Becker-Blease 2011, He and Tian 2013, Gormley and Milbourn 2013).1 An important concern with many of these papers is that they (like much of the historical work in economics and management) tend to rely on patent-based measures. However, as we discuss below, these measures have weaknesses that potentially lead to flawed or incomplete inferences regarding the factors affecting innovation.

In this paper, we propose a new measure of innovation, research quotient (RQ), as a better measure of firm innovation. The measure is not based on patent or citation data, so it avoids the potential flaws and limitations of such measures. As we will show, our RQ measure is better at capturing the value creation to a firm of innovation.

One of the primary concerns with patent-based measures is they are not universal. Fewer than 50% of the firms in COMPUSTAT who conduct R&D actually file patents for their innovations. Thus half the firms conducting R&D will appear to have no innovation. This limits the sample size and decreases the power in any test based on these measures for publically traded firms.2

Beyond the practical problem of limited sample size, is the identification issue that the decision to patent is endogenous to firm innovation policy. Firms choose which (if any) innovations to patent. Because patents are costly both financially (the cost to file and defend) as well as competitively (they require disclosure of the fundamental knowledge underpinning the innovation), firms file patents only under certain circumstances. They do so primarily to prevent copying when their innovations are easy to invent around (Cohen, Nelson and Walsh 2000), though also for strategic reasons, such as to block other firms’ patents (82% of surveyed firms), to prevent lawsuits (59%), to use in negotiations (47%), or to enhance their reputation (48%).

In a paper which explicitly tackles concerns with patent-based measures, Abrams, Akcigit and Popadak (2013) develop a model that jointly considers productive and strategic rationale for patenting. They expect and find empirically that the relationship between citations and patent value is non-monotonic—suggesting that firms are less likely to patent their most valuable innovations! Thus firms who patent and the innovations they patent may be structurally different from those who don’t. If so, then policy recommendations emerging from these studies may be misguided.

See full PDF below.

Leave a Comment