THE VALUE IN FUNDAMENTAL ACCOUNTING INFORMATION

• DOI: http://doi.org/10.1111/jfir.12119
• Published date: 01 March 2017
• Date: 01 March 2017

THE VALUE IN FUNDAMENTAL ACCOUNTING INFORMATION

H. J. Turtle

Colorado State University

Kainan Wang

University of Toledo

Abstract

We examine the role of fundamental accounting information in shaping portfolio

performance. Using a conditional performance approach, we address the concern that

the positive relation between Piotroski’sF Score and ex post returns is due to risk

compensation. Our results show that portfolios of firms with strong fundamental

underpinnings generate significant positive and time-varying performance. One

potential source of these performance gains is an underreaction to public information

(such as momentum and F Score) when information uncertainty (proxied by size,

illiquidity, and idiosyncratic volatility) is high. In addition, conditional performance

benefits seem prevalent in periods of high investor sentiment.

JEL Classification: G11, G12, G14, M41

I. Introduction

There is extensive and long-standing support for the use of fundamental valuation tools

in equity choice from the early seminal work of Graham and Dodd (1934) to the current

investment philosophy of their modern adherent Warren Buffett.

1

In addition, numerous

authors document the potential economic benefits of fundamental analysis through

the judicious use of available financial accounting information (cf. Ou and Penman 1989;

Lev and Thiagarajan 1993; Haugen and Baker 1996; Asness 1997; Frankel and Lee 1998;

Griffin and Lemmon 2002; Piotroski 2000; Mohanram 2005; Piotroski and So

2012; Novy-Marx 2013).

In recent years, the F Score, a comprehensive measure of firm fundamentals, has

drawn attention from academics and practitioners. Introduced by Piotroski (2000), the

F Score is constructed from nine accounting signals capturing three aspects of a firm’s

financial strength: profitability, liquidity, and operating efficiency. Piotroski finds that

The authors are grateful for the helpful comments and suggestions of the editors, associate editor, referee, and

seminar participants at the 2010 Northern Finance Association meetings and the 2012 Financial Management

Association meetings.

1

For example, in his 2008 Letter to Shareholders, Buffett quips, “Price is what you pay. Value is what you

get.”In the Foreword to the First Edition in The Warren Buffett Way (Hagstrom 1994), Peter Lynch describes

Buffett’s thought process regarding company valuation as “the critical investment factor is determining the

intrinsic value of a business and paying a fair or bargain price”(p. xxi).

The Journal of Financial Research Vol. XL, No. 1 Pages 113–140 Spring 2017

113

© 2017 The Southern Finance Association and the Southwestern Finance Association

RAWLS COLLEGE OF BUSINESS, TEXAS TECH UNIVERSITY

PUBLISHED FOR THE SOUTHERN AND SOUTHWESTERN

FINANCE ASSOCIATIONS BY WILEY-BLACKWELL PUBLISHING

the F Score has strong predictive power for ex post returns. Piotroski and So (2012)

further show that the predictability of the F Score for ex post returns is stronger when

there is substantial disagreement, or incongruence, between a firm’s fundamental value

(proxied by the F Score) and market-perceived value (measured by book-to-market).

Walksh€

ausl (2016) further examines the returns to these incongruent portfolios with

strong F Scores and large book-to-market ratios. He finds strong unconditional portfolio

performance is also present in international markets, and there is some further indirect

evidence that the observed predictability is at least partially related to a financing-based

mispricing factor.

Although the literature finds a strong relation between F Score and ex post

returns, the economic rationale for this relation has been debated by multiple authors. In

particular, although portfolios of high F Score stocks often produce large ex post returns,

it is not clear if these returns simply represent appropriate compensation for risk

premiums. Fama and French (2006) present a valuation framework to jointly test the

relations among expected profitability, expected investment, book-to-market ratio, and

expected returns.

2

They conclude that the sources of predictability in expected returns

due to F Score could be both rational (i.e., high F Score stocks are more risky) and

irrational (i.e., high F Score stocks are more prone to mispricing). Similarly, in a recent

survey on accounting anomalies, Richardson, Tuna, and Wysocki (2010) point out that

researchers examining the relation between accounting attributes and ex post returns

must ensure that plausible risk-based explanations are precluded as alternative

explanations.

The extant literature on risk-based pricing is primarily cast in the context of an

unconditional asset pricing model in which required asset returns do not vary with

changes in the underlying information set. Recent advances in conditional performance

evaluation have the ability to assess the impact of information on marginal performance

to directly address the risk premium counterargument. For example, Ferson and Schadt

(1996) and Christopherson, Ferson, and Glassman (1998) propose conditional alpha

measures based on an underlying time-varying beta model. Jha, Korkie, and Turtle (2009)

extend this work to provide a conditional alpha that is consistent with an underlying

conditional mean–variance decision framework. The resultant conditional alpha has

desirable properties and may be readily obtained from a simple unconditional regression.

We adopt this framework to examine the importance of fundamental accounting

information in conditional performance.

The conditional alpha measures the risk-adjusted return contribution from a

conditional model of equilibrium returns. For a given equilibrium model (with

2

Through an extension of the basic valuation equation for future expected dividends, and with the addition of

a clean surplus accounting condition from Ohlson (1995), they obtain,

Pt¼X1

t¼1

E EPStþtdBtþt

ðÞ

1þrðÞ

t,

where P

t

is the stock price per share at time t,B

t

is the book value per share at time t,EPStþtis the earnings per share

at time tþt,dBtþt¼BtþtBtþt1is the similarly indexed change in book value per share, EðÞisthe expectation

operator, and ris the internal rate of return on expected dividends (cf. Fama and French 2006, p. 492). Dividing both

sides of the equation by the book value of equity produces the results discussed.

114 The Journal of Financial Research