This paper examines the added-value of combining traditional valuation ratios with each other as well as with some financial statement variables in the German stock markets during the 2000–۲۰۱۵ period. The results show that combination pays off and, moreover, that the benefits of combination are greater in Germany than in most other developed stock markets. Particularly, we find strong evidence of the added-value of using Piotroski’s F-score as a supplementary selection criterion for value stocks as well as for low-accrual stocks. Our results show further that the F-score also boosts the efficacy of other valuation ratios besides the book-to-price ratio. In addition, the inclusion of F-score besides a relative value measure tends to increase the average market equity of portfolio firms. The decomposition of the full-sample-period performance into separate bull- and bear-period performance shows clearly that the better performance of F-score-boosted portfolios is mostly attributable to their outperformance during bearish periods, even though on average, they also generate higher bull-period returns than the comparable value portfolios formed without F-score. The use of F-score as a supplementary criterion also increases the proportion of stocks that earn above-market-average returns during the subsequent holding period. For the first time in the financial literature, we also document a strong relationship between high F-score stocks and momentum stocks.

Introduction

Considerable international evidence of value anomalies, which refer to the tendency of value stocks to outperform the stock market average most of the time, has been documented (e.g., Chan and Lakonishok 2004; Brown et al. 2008a; Fama and French 2006, 2012; Cakici et al. 2013; Pa¨ta¨ri et al. 2017b). This evidence has shown not only that the value anomalies in stock markets are a worldwide phenomenon but also that the relative efficacy of different valuation criteria varies across stock markets and the sample periods examined.1 In addition, some scholars have started to examine whether the performance of value stock portfolios can be further enhanced by adding other criteria besides relative value to the portfolio selection. These studies can be classified into two categories, the first of which either combines value indicators with momentum indicators (e.g., Pa¨ta¨ri et al. 2012, 2017a; Fisher et al. 2016) or double-sorts the stocks based on their value and momentum indicators (e.g., see Bird and Casavecchia 2007; Leivo 2012). In the second category, test designs attempt to pick the best-performing value stocks of the future from a larger set of value stocks based on some other firm characteristics, such as profitability, accruals or financial strength. The results of the studies of the latter type have been promising as, on average, value stocks with good profitability, a high earnings quality and/ or a sound financial condition have performed better than stocks that are just the cheapest in terms of valuation ratios.2 This study contributes to the latter category by combining these three quality dimensions with a value dimension by using German stock market data. Motivated by the evidence for the added-value of combining individual valuation ratios into composite value criteria,3 we also test whether the benefits of creating combinations within the value dimension are comparable to those achievable by combining the value and quality dimensions. For this purpose, we form some simple 2-combination value portfolios. We also examine the added-value of the inclusion of Piotroski’s (2000) F-score as the supplementary criterion in cases of forming such composite value portfolios. In general, the benefits of combining selection criteria are related to the inter-relationship between the criteria being compared. The weaker the relationship, the better the potential for the added-value of combining. For this reason, the added-value of combining individual valuation ratios is generally deemed to be limited, as they all are reasoned to represent the same (i.e., value) dimension (e.g., see Fama and French 2011).