Real Earnings Management and Accrual-based Earnings Management in Family Firms

Abstract

We examine the effects of family firms on real earnings management (REM) and accrual-based earnings management (ABEM). Using socioemotional wealth as a theoretical framework and considering the different implications of REM and ABEM on family firms' transgenerational sustainability, we hypothesise and find for a sample of 402 German listed family firms during 1998–2008 that family firms engage less in REM and exhibit more earnings-decreasing ABEM policies as compared to a sample of 436 non-family firms. We further provide evidence that family firms as compared to non-family firms treat REM and ABEM as substitute rather than complementary tools for earnings management. Overall, our findings suggest that family firms use earnings management activities strategically, avoiding those that inhibit the firm's long-term value (i.e. REM) and engaging in those that help families retain transgenerational control (i.e. ABEM).

View correction statement:

Correction to European Accounting Review 23(3) (2014), Special Issue on ‘Accounting and Reporting in Family Firms’

Acknowledgements

We thank Joseph Astrachan, Sasson Bar-Yosef, Annalisa Prencipe, Pureum Kim, Suresh Radhakrishnan, Bharat Sarath, Marco Trombetta, Stanley Baiman, and workshop participants at the 2010 IFERA meeting at Lancaster University and at the 2012 European Accounting Review Conference on Accounting and Reporting in Family Firms at Bocconi University. We are also grateful to three anonymous referees for their helpful comments and suggestions received during the review process. All errors remain our own.

Notes

¹Studies show, however, that Germany is only one of many countries where firms report high family ownership concentration (see Prencipe, Bar-Yosef, & Dekker, 2014).

²It should be noted that some researchers have recently called for an investigation into REM and ABEM in family firms. For example, Salvato and Moores (2010, p. 198) write: … . Reflecting a tendency in the broader accounting literature, the totality of works on earnings management in family firms focuses exclusively on cosmetic manipulations. … These results show a diverse portrait of the impact of family ownership on corporate disclosure, hence calling for additional research efforts placing the ‘controlling family’ entity at center stage. For instance, do different types of family firms favor different earnings management strategies? If so, why? Is the extent of cosmetic versus real earnings management different in family firms relative to their nonfamily counterparts?

³Note, however, that although abnormal accruals are traditionally viewed as capturing management's financial reporting decisions (or discretion), Wang's and Tong's research questions focus more on earnings quality in family firms because they do not impose a directional sign on management's expected financial reporting decisions. Wang (2006) uses earnings management and earnings quality interchangeably. Specifically, in the description of his empirical ABEM metric, he asserts: ‘A higher value means a greater level of earnings management or lower earnings quality’ (p. 630). In contrast, our research question is strictly related to earnings management and does impose a directional prediction (as we explain below, our focus is on the signed value of abnormal accruals).

⁴A number of current studies and surveys also evaluate the economic importance of German family firms relative to other EU family firms. Faccio and Lang (2002), for example, document that German founding family firms account for the largest share of their sample of 2332 EU listed firms (20%). German family firms are also larger in terms of total turnover. Franks et al. (2009) find that median sales in German family firms is 1.16 billion Euros in 2006, which compares to 0.54 billion Euros, 0.77 billion Euros, and 0.20 billion Euros in France, UK, and Italy, respectively.

⁵In La Porta et al. (2006), the median values of disclosure requirements, liability standards, and public enforcement are 0.5, 0.5, and 0.37, respectively, while for Germany these values are all below the medians (0.42, 0, and 0.22, respectively), indicating a low level of investor protection. In Leuz et al. (2003), the median value for the aggregate accrual earning management score is 16.5, while for Germany this value is above the median (21.5), thus indicating a high level of accrual earning management.

⁶As noted by Roychowdury (2006), by including a scaled intercept, α₁(1/A_t−1) we avoid a spurious correlation between scaled CFO and scaled sales due to variation in the scaling variable, total assets. The unscaled intercept, α₀ ensures that the mean abnormal CFO for every industry-year is zero. Including these intercepts allows the average CFO_t/A_t−1 for a particular industry-year to be non-zero even when the primary explanatory variables in the model, sales and change-in-sales, are zero.

⁷Discretionary expenses correspond to the sum of R&D, advertising, selling, general and administrative (SG&A) expenses. As long as SG&A expenses are available, R&D are set to zero if information on this item is missing in Worldscope.

⁸The average explanatory power of this piecewise nonlinear abnormal accruals estimation for our pooled sample is 43.34% and is similar to that reported in Wang (2006); however, it is larger than that reported in traditional linear accrual estimations models, such as in Dechow and Dichev (2002) (37.36%).

⁹We acknowledge that the construction of our FF_DUMMY variable is somewhat different from that used in Anderson and Reeb (2003) and Wang (2006). Specifically, Anderson and Reeb (2003) and Wang (2006) set this dummy variable equal to one if family members are on the management board of the firm, regardless of the percentage of voting rights of the founding family. In contrast, we use the 25% ownership threshold because, under German corporate law, this is the ownership level that allows shareholders to block important decisions of the board. With this research design we ensure thus the ability of the founding family to affect board decisions, in cases when no member of the founding family is represented in none of the boards.

¹⁰We focus on the presence of members of the founding family in the management board or in the supervisory board because, under the German corporate governance regulation, these two organs are directly involved in the determination of (real and financial) actions affecting the earnings decision process.

¹¹The CDAX is a German stock market index calculated by the German Stock Exchange (Deutsche Börse AG). It includes all stocks traded on the Frankfurt Stock Exchange that are listed in the General Standard or Prime Standard market segments. The CDAX Index represents the German equity market in its entirety, i.e. all companies listed on the Frankfurt Stock Exchange. We obtain data on the yearly composition of the CDAX from the Weighting Files of the German Stock Exchange.

¹²For example, the Gesetz zur Kontrolle und Transparenz im Unternehmensbereich (KontraG) is intended to strengthen the role of the supervisory board in monitoring management and to increase the accountability of board members and auditors. Furthermore, it introduces rules on risk management systems, requires firms to report on risks and risk structures and to provide cash flow statements and segment reporting in their consolidated financial statements. The Kapitalaufnahmeerleichterungsgesetz (KapAEG) allows listed firms in Germany to prepare consolidated financial statements under IFRS or US GAAP without being required to disclose additional consolidated financial reports under German GAAP.

¹³We verify data on firm ownership and board structure using the following databases: Bureau van Dijk's Amadeus, database, Commerzbank's Wer gehört zu wem, Bundesanstalt für Finanzdienstleistungsaufsicht (BaFin).

¹⁴The field used to identify the type of accounting standards adopted in Worldscope is ‘Accounting Standards Followed’. We code firm–year observations as German GAAP if one of the following cases applies: 01 (local standards), 08 (local standards with EU and International Accounting Standards Committee (IASC) guidelines), 10 (local standards with some EU guidelines), 17 (local standards with some Organization for Economic Cooperation and Development (OECD) guidelines), 18 (local standards with some IASC guidelines), and 19 (local standards with some IASC and OECD guidelines). We code firm–year observations as IFRS in the following case: 23 (IFRS). Prior studies, however, indicate that Worldscope contains data errors in coding accounting standards (Daske, Hail, Leuz, & Verdi, 2008). We cross-check the information on accounting standards followed contained in Worldscope with data obtained from firms' own financial statements.

¹⁵This pattern is similar to that outlined by Wang (2006) for a sample of 207 Standard & Poor's (S&P) 500 firms during 1994–2002. Specifically, Wang summarises his descriptive statistics showing that, on average, family members own 10.35% of common stock. The average (median) family ownership (measured by percentage of common equity) is 11.54% (6.04%) in 1994, and it decreases to 7.40% (4.05%) in 2002. Regardless of this similarity between our sample of family firms and Wang's, we note that the magnitude of family ownership in our sample (35.80%) is critically larger than that reported by Wang's (10.35%) and by Anderson and Reeb (2003) (17.9%). Such a difference increases our confidence that any documented effect of family firms on earnings management strategies suffers no more than other studies from power concerns.

¹⁶Descriptive information about the variable HERF indicates that non-family firms are significantly more concentrated than family firms. In additional tests we find that this trend is attributable to the larger percentage of outside blockholders (e.g. strategic investors, banks, private equity, and other institutional shareholders) in non-family firms than in family firms.