The information environment of the firm and the market valuation of R&D

• Date: 01 October 2018
• Published date: 01 October 2018
• Author: Lu Y. Zhang,Melissa Toffanin
• DOI: http://doi.org/10.1111/jbfa.12349

DOI: 10.1111/jbfa.12349

The information environment of the firm

and the market valuation of R&D

Lu Y. Zhang Melissa Toffanin

Schoolof Accounting and Finance, Ted Rogers

Schoolof Management, Ryerson University, 350

VictoriaStreet, Toronto, ON, M5B 2K3, Canada

Correspondence

LuY. Zhang, School of Accounting and Finance,

TedRogers School of Management, Ryerson

University,350 Victoria Street, Toronto,ON, M5B

2K3,Canada.

Email:lu.zhang@ryerson.ca

Fundinginformation

RyersonSocial Sciences and Humanities Research

Council(SSHRC) Institutional Grant

Abstract

We examine whether firms’ information environmentinfluences the

market valuation of their hard-to-value knowledge assets: R&D cap-

ital. We find that a better information environment, as measured

by greater analyst coverage, lower earnings forecast dispersion,

and greater earnings forecasting accuracy,is associated with higher

stock market valuation of firms’ R&D capital. We conduct additional

analyses to rule out alternative explanations and test for the direc-

tion of causality. Our causal evidence suggests that analysts signif-

icantly reduce the information uncertainty surrounding firms’ R&D

capital.

KEYWORDS

analyst coverage, information environment, R&D, stock market

valuation

1INTRODUCTION

The dramatic increase in the number of technologically-intensive firms overthe last few decades and the dazzling dol-

lar amounts that these firms spend on R&D raises the question of whether the R&D investments made bythese firms

and the resulting knowledge assets are correctly incorporated into their marketvaluations. Financial markets may fail

at this task because of the information asymmetry associated with R&D investments: these investments are largely

unique to their respective firms, thereby making it difficult for outside investors to evaluate the investmentsagainst

those made bypeer firms. Further complicating matters, R&D investments have no competitive market or market price

from which investors can obtain information (Aboody & Lev,2000). Some empirical studies suggest that investors are

overlyoptimistic regarding the probability of success of the investments (Daniel & Titman, 2006; Jensen, 1993; Lakon-

ishok, Shleifer, & Vishny, 1994), while others find that investors underreact to the information regarding the R&D

investments (Ali, Ciftci, & Cready,2012; Chan, Lakonishok, & Sougiannis, 2001; Cohen, Diether, & Malloy, 2013; Hall,

1993; Hirshleifer, Hsu, & Li, 2013; Porter,1992; among others). Furthermore, the accounting rules of expensing R&D

spending may dampen investors’ability to convert the reported R&D expense to future knowledge assets’ values.1

There has been limited empirical evidence about the mechanisms through which investors incorporate the value

of R&D investments into stock prices. Our paper takes on this task by investigating the role of a firm's information

environment in the market valuation of its R&D investments. Specifically, we examine whether financial analysts,

through their function as information intermediaries, improve that R&Dvaluation by the financial market.

1See,for example, Aboody and Lev (2000), Lev and Sougiannis (1996) and Wyatt (2005).

J Bus Fin Acc. 2018;45:1051–1081. wileyonlinelibrary.com/journal/jbfa c

2018 John Wiley & Sons Ltd 1051

1052 ZHANG ANDTOFFANIN

Financial analysts have been shown to independently collect and produce value-relevant information about the

firm.2In the context of valuing R&D investments, where misvaluation is more likely to occur than when valuing tan-

gible investments, analysts have greater incentivesto acquire information for potential profit opportunities. Analysts

have been found to provide greater coverage for R&D-intensive firms (Barth, Kasznik, & McNichols, 2001), conduct

more conference calls with these firms (Tasker, 1998), and produce information about firms'intangibleassets beyond

the disclosure in financial statements (Amir, Lev, & Sougiannis, 2003). Analyst coverage also helps incorporate target

firms’ R&D values into their equity valuations in mergers and acquisitions (Kimbrough, 2007).

Using analyst variables as measures of the information environment of firms, we find that analyst coverage sig-

nificantly elevates the stock market's valuation of R&D capital, as do the consensus among and accuracy of analysts’

earnings forecasts. After controlling for financial characteristics of firms, our baseline results suggest that a US$ 1

extra investment in R&D capital made by firms with high (above industry-year median) analyst following translates

into a US$ 0.55 increase in their market value, on average,while the same US$ 1 R&D investment translates into only

a US$ 0.40 increase in market value for firms with low (below industry-year median) analyst following. The valuation

difference is more pronounced in smaller firms, and among firms in R&D-intensiveindustries. These findings are robust

using an Ohlson (1995)-type linear information model as an alternative valuation model and using bid-ask spreads and

probabilities of informed trades (PIN) as alternative measures ofinformation environment.

We conduct further tests to rule out two alternativeexplanations for our findings. It may be the case that firms with

higher levels of analyst coverage, facing greater short-term performance pressure, select R&D investments that are

less risky and are potentially more easily valued by the market.It could also be the case that analysts merely self-select

into covering firms with higher R&D valuations rather than help to incorporate new information about R&D invest-

ments into firms’ stock prices. Totest for the first alternative explanation, we construct two measures of R&D risk: the

correlation between a firm's recent R&D investments and its future earnings (Cohen et al., 2013) and the correlation

between a firm's recent R&D investmentsand the standard deviation of its future earnings (Kothari, Laguerre, & Leone,

2002). We use these two measures to investigatewhether the positive association between analyst coverage and R&D

valuation is stronger among firms with lower R&D risk. We find that analysts do in fact improve the marketvaluation

of firms irrespective of the riskiness of their R&D. Our main finding cannot be explained byfirms selectively choosing

their R&D investments under pressure from analysts.

To address the second alternative explanation,we employ two instrumental variables to capture the variation in

analyst coveragethat is exogenous to the R&D valuation of the firms they cover.The two instruments are the expected

analyst coverage estimated with the change in brokerage firm size over time (He and Tian, 2014; Yu, 2008) and an

indicator variable of membership in the S&P 500 index (Aghion, Van Reenen, & Zingales, 2013; Wahal& McConnell,

2000; Yu,2008). The variation in expected analyst coverage results from the strategic business changes in the broker-

age firms, while that of coverage based on S&P 500 membership is more likelyto be the result of greater information

demands by the market for S&P 500 firms. Consequently, the variation in coveragein both cases is less likely to be

caused by the market valuation of the firms’ R&D investments. We use each instrument to estimate predicted ana-

lyst coverage and compare the market valuation of R&D among firms with high and low predicted analyst coverage.

We find consistent results of higher R&D valuations among firms with greater analyst coverage predicted by the two

instrumental variables. Our main findings are not driven by the self-selection of analyst coverage.

To further establish the direction of causality, we investigate the loss of analyst coverage from two quasi-natural

experiments: brokerage firm mergers and closures. Both eventsresult in the termination of analyst coverage caused

by the strategic changes of the brokerage firms rather than potential negative information about the coveredfirms

(Kelly & Ljungqvist, 2012; Wu & Zang, 2009). Previousstudies have used similar events to investigate the causal effect

of analysts (see, for example, Derrien & Kecskés,2013; Hong & Kacperczyk, 2010; Kelly & Ljungqvist, 2012). Employ-

ing a difference-in-difference (DiD) approach, we find that firms losing analyst coveragefrom these events experience

2There is a vast literature regarding the value-relevant information produced byanalysts. See, for example, Chung and Jo (1996), Franco and Hope (2011),

Jegadeesh,Kim, Krische, and Lee (2004), Lee, Strong, and Zhu (2014), Loh and Stulz (2011), Lui, Markov, and Tamayo(2012) and Womack (1996). For notable

reviewsof this literature, see, for example, Healy and Palepu (2001) and Ramnath, Rock, and Shane (2008).