Abnormal Returns To Accounting-Based Investment Strategies by Steven De Klerck
WALTER AERTS
Universiteit Antwerpen, Prinsstraat 13, 2000 Antwerp, Belgium
University of Tilburg, Warandelaan 2, 5037 AB Tilburg, The Netherlands
walter.aerts@ua.ac.be
JAN ANNAERT
Universiteit Antwerpen, Prinsstraat 13, 2000 Antwerp, Belgium
Antwerp Management School, Sint Jacobsmarkt 9, 2000 Antwerp, Belgium
jan.annaert@ua.ac.be
MARC J.K. DE CEUSTER
Universiteit Antwerpen, Prinsstraat 13, 2000 Antwerp, Belgium
Antwerp Management School, Sint Jacobsmarkt 9, 2000 Antwerp, Belgium
marc.deceuster@ua.ac.be
STEVEN DE KLERCK
Universiteit Antwerpen, Prinsstraat 13, 2000 Antwerp, Belgium
steven.deklerck@ua.ac.be
First draft: March 9, 2015
Abstract
The literature on financial statement analysis is characterized by the working hypothesis that fine-grained and complex accounting-based investment strategies provide investors with higher abnormal returns compared to simple accounting-based investment strategies. Up to now this concerns a working hypothesis which still has not been submitted to a comprehensive empirical analysis. Thanks to the extensive number of accounting-based investment strategies that have been developed since the past decades, in this paper we provide insight into the abnormal returns of the following prominent accounting-based investment strategies: the working capital accrual strategy (Sloan, 1996), FSCORE (Piotroski, 2000), SSCORE (Penman and Zhang, 2006), PEISCORE (Wahlen and Wieland, 2010), competitive-specific SSCORE (Dickinson and Sommers, 2012), the gross profits-to-assets strategy (Novy-Marx, 2013) and the operating profitability strategy (Ball et al., 2014). Abnormal returns are documented for both the US stock market and international stock markets, i.e. Japan and Europe. In addition, we document the fundamental risk profiles of the aforementioned investment strategies. The gross profits-to-assets strategy and the operating profitability strategy emerge as the most robust accounting-based investment strategies. Future studies should focus on documenting the added value of detailed revisions and analyses of the financial statements when developing and studying accounting-based investment strategies.
Abnormal Returns To Accounting-Based Investment Strategies – Introduction
Financial statement analysis or fundamental analysis attempts to separate ex-post winners from losers on the basis of accounting information that is not correctly impounded in stock prices. Most handbooks on financial statement analysis implicitly and/or explicitly assume that fine-grained revisions and analyses of the financial statements provide fundamental investors with indispensable key insights enabling them to realize significantly higher portfolio returns compared to more simple methods. Penman (2010), for example, starts with a discussion of a basic valuation method, i.e. the use of valuation multiples such as the book-to-market ratio. It is noted that these methods ignore a lot of accounting information, information that enables financial analysts to realize more accurate estimates of future earnings and intrinsic values. This working hypothesis is also reflected in accounting papers on financial statement analysis. In many research papers (e.g. Nissim and Penman, 2001; Penman and Zhang, 2006; Wahlen and Wieland, 2010; Dickinson and Sommers, 2012), there is a strong belief that ever more detailed accounting information, often in combination with intricate methodologies, will result in better investment decisions and consequently higher investment returns. The fact that the majority of accounting-based investment strategies were developed and studied independently implies that this working hypothesis still has not been confirmed by empirical research. As a consequence, in this paper, we study in an integrated way the abnormal returns of the most fundamental accounting-based investment strategies in relation to their underlying degree of detail and complexity.
Based on a review of the literature, we are able to identify a large number of prominent accounting-based investment strategies that have been developed and/or researched over the past decades (e.g. Ou and Penman, 1989; Holthausen and Larcker, 1992; Lev and Thiagarajan, 1993; Sloan, 1996; Abarbanell and Bushee, 1998; Piotroski, 2000; Beneish et al., 2001; Penman and Zhang, 2002, 2006; Mohanram, 2005; Wahlen and Wieland, 2010; Dickinson and Sommers, 2012; Novy-Marx, 2013; Ball et al., 2014). These studies show that fundamental investors can benefit from investing on various accounting signals.
Some of these investment strategies require a vast amount of historical accounting information, extensive accounting computations and complex methodologies (e.g. Ou and penman, 1989; Holthausen and Larcker, 1992; Beneish et al., 2001; Penman and Zhang, 2006; Dickinson and Sommers, 2012), an observation first made by Piotroski (2000). Ou and Penman (1989), followed by Holthausen and Larcker (1992), implement a financial statement analysis which combines a large set of accounting variables into one summary measure, referred to as Pr. The methodology requires at least ten years of accounting data, the computation of 68 accounting variables and the entertainment of univariate and multivariate statistical analyses through logistic regression to calculate the Pr summary measure. Beneish et al. (2001) apply a two-stage statistical approach, first using twelve market-based signals to identify likely extreme performers, and then applying nine financial signals to differentiate between extreme winners and extreme losers. The objective of Penman and Zhang (2006) is similar to the one of Ou and Penman (1989), i.e. the development of a summary measure that informs about earnings sustainability. While Ou and Penman (1989) opt for the ad hoc selection of accounting variables, Penman and Zhang (2006) operate in a structured way when selecting accounting variables. Their approach requires at least six years of accounting data, the calculation of six accounting items and the use of regression analyses to calculate the forecasted summary measure of operating profitability. Dickinson and Sommers (2012) build on the framework developed by Nissim and Penman (2001) and Penman and Zhang (2006) and contribute to the literature by introducing a large number of strategically-motivated accounting variables. In addition, their strategy demands the calculation of a risk- and industry-adjusted measure of operating profitability, also substantially contributing to the complexity of their investment approach.
Besides these comprehensive accounting-based investment strategies, in the accounting and finance literature we also find less elaborate methods. In this regard, we can think of the post earnings announcement drift documented by Bernard and Thomas (1989) and the accrual accounting signal set forth by Sloan (1996) and elaborated upon in many accounting and finance papers (e.g. Richardson et al., 2005; Pincus et al., 2007; Fama and French, 2008; Papanastasopoulos, 2014). More recently, the gross profits-to-assets ratio introduced by Novy-Marx (2013) and refined by Ball et al. (2014) has received a lot of attention. These accounting-based investment strategies require the calculation of only one accounting signal and rely on a simple ranking procedure to establish stock portfolios.
The papers by Piotroski (2000), Mohanram (2005) and Wahlen and Wieland (2010), among others, strike the balance between the two extremes. The three strategies – i.e. FSCORE (Piotroski, 2000), GSCORE (Mohanram, 2005) and PEISCORE (Wahlen and Wieland, 2010) – require the computation of between six and nine accounting signals. The scoring systems range from relatively simple for FSCORE to more complex for GSCORE and PEISCORE. Piotroski (2005), for example, remarks that the implementation costs associated with GSCORE are significantly higher compared to FSCORE, again underlining the advantage of simple accounting-based investment strategies. These middle-of-the-road accounting-based investment strategies are characterized by the absence of the need to implement statistical procedures.
See full PDF below.
