IFRS Adoption Among Private Companies

Abstract

EU gave the opportunity to each Member State to oblige/allow non-listed (i.e., private) companies to use international financial reporting standards (IFRS). Considering a sample of Italian private companies that switched to IFRS in the time span from 2005 to 2008, we compare financial reporting quality between IFRS adopters and a matched sample of companies still using local (Italian) generally accepted accounting principles (GAAP). This should be of interest for the EU Commission in evaluating the impact of the current financial reporting regulation and for EU national regulators, who are left with a certain degree of flexibility in endorsing parts of the European legislation. Overall, our results show that IFRS adoption did not improve reporting quality among private companies but, on the contrary, decreased it. As companies can exploit the level of flexibility embedded in IFRS to pursue their own reporting interests, separate analyses were conducted taking into consideration firms’ incentives. In particular, assuming that entities controlled by listed companies might have switched to IFRS mainly for complying with parent company requirements and/or simplifying the financial reporting process, we run the analyses separately for this sub-sample and other firms. Findings reveal signs of earnings quality deterioration for both groups although the impact seems slightly worse for subsidiaries of listed companies.

Keywords

IFRS reporting quality private companies reporting incentives

Introduction

The aim of this study is to analyze the impact of adopting international financial reporting standards (IFRS) on financial reporting quality using a sample of Italian private (i.e., non-listed) companies that switched to IFRS.¹

The European Union (EU) introduced a common set of accounting standards in 2005, with the objective of enhancing financial reporting quality and comparability across countries. This shift was intended for all public (i.e., listed) companies in Europe preparing consolidated financial statements, but the regulation also gave each Member State the opportunity to decide whether to oblige/allow other kind of companies, for example non-listed ones, to use the same set of standards for financial reporting purposes. In most recent years, International Accounting Standards Board (IASB) and EU Commission have particularly focused their attention on the use of international accounting standards by private entities (Nobes, 2009). Indeed, the latter represent the majority of EU economy and account for more than 75% of European GDP (The European Confederation of Directors Associations [Ecoda], 2010).

Dealing with private companies only, our research should be of interest for the EU Commission in evaluating the impact of the current financial reporting regulation and for EU national regulators, who are left with a certain degree of flexibility in endorsing parts of the European legislation. Given that standard setters and regulators are asked to take into account the effects and the consequences of the standards they develop/suggest (European Financial Reporting Advisory Group, 2011), providing this kind of empirical evidence is, in our opinion, crucial. This also permits us to reply to the Maijoor (2010) call to the academia “to be more involved in policy-orientated and normative debates, of course after having done proper research based on evidence-gathering” (p. 330).

Our article focuses on an area where it is difficult to locate any prior literature: The effects of IFRS adoption on the reporting quality of private companies. As recent literature demonstrates, financial reporting quality cannot be assumed as homogeneous between public and private companies (e.g., Ball & Shivakumar, 2005; Burgstahler, Hail, & Leuz, 2006); therefore, we feel that the mere extension of the (conflicting) results obtained for listed companies is inadequate for understanding the effects on unlisted companies. The research site used in this article, Italy, can be considered a suitable setting for many reasons. First of all, the Italian environment is a typical example of stakeholder-oriented accounting system (like France, Germany, Belgium, and Spain). As emphasized in previous works (Bartov, Goldberg, & Kim, 2005; Hung & Subramanyam, 2007), IFRS are influenced by a shareholder-oriented model: Conversely, we consider a setting where the effect of the transition to the new standards is likely to be particularly marked. This is also supported by Bae, Tan, and Welker (2008) who ranked countries according to the “distance” between local generally accepted accounting principles (GAAP) and IFRS. Among a sample of 49 countries, Italy was considered to be one of those countries where local (Italian) GAAP differed most from the international ones. In particular, Italian “scored” 12 differences whereas Anglo-Saxon settings report “difference scores” below 4. For this reason, we expect that the shift to the new set of standards would lead to stronger effects in Italy compared with other countries where national standards are more similar to IFRS (like the United Kingdom and Australia). Moreover, the other most important European countries with a distance between local GAAP and IFRS similar to the Italian one (like, for example, France, Belgium, Austria, and Germany) are not allowed free choice as to preparing financial statements using either national standards or international ones² (PricewaterhouseCoopers, 2012). In addition, small- and medium-size (mainly family-owned) companies are the strongest component of the Italian economy (Economist Intelligence Unit [EIU], 2005). Finally, according to the Italian tax principle of neutrality, an equal treatment will be granted for those company adopting IFRS and those accounting according to the Italian GAAP (PricewaterhouseCoopers, 2006). Therefore, individual tax issues and, more generally, the peculiarities of the national tax system should not influence the results of our analysis.

Our purpose aims at understanding whether the level of earnings quality and, in particular, the level of abnormal accruals and timely loss recognition are different between private companies adopting IFRS and a matched sample of firms reporting under local GAAP.

We first select a sample made of all Italian private companies that decided to adopt IFRS in the time span from 2005 to 2008 with reference to their annual reports. Then, we match each IFRS adopter with a pair that, in the same time span, accounts under local (Italian) GAAP, using a propensity-score matching methodology. We then compare the level of abnormal accruals and timely loss recognition across the two samples. Overall, our findings show that IFRS adopters do not show higher earnings quality compared with local GAAP adopters. On the contrary, we find evidence indicating that companies that adopted IFRS exhibit higher levels of abnormal accruals and a decrease in timely loss recognition, according to the Ball and Shivakumar (2005) accrual-cash flow model. Evidence suggesting no improvements in the level of earnings quality among IFRS adopters is also confirmed by a set of robustness tests, where we use other proxies for earnings quality and an alternative research methodology in which earnings quality is measured before and after the IFRS adoption for the sample of IFRS adopters only, thus mitigating endogeneity problems. Taken as a whole, these findings suggest that the adoption of a set of accounting standards reputed to be of better quality than national ones (Barth, Landsman, & Lang, 2008) does not imply, per se, better financial reporting quality.

Because IFRS contain several accounting policy options (Kvaal & Nobes, 2012), there is room for discretion in IFRS application (e.g., Daske, Hail, Leuz, & Verdi, 2008; Leuz, 2010). As companies can exploit the flexibility within IFRS, separate analyses were conducted distinguishing between firms controlled by listed companies (i.e., subsidiaries the parent company of which is obliged by EU Regulator to use IFRS), which are assumed to adopt IFRS for complying with parent company requirements and/or for simplifying financial reporting process, and other firms. We find that reporting quality does not show any sign of improvement for both groups of firms, although the impact of IFRS on our earnings quality proxies seems, in some cases, worse for subsidiaries of listed entities.

The structure of the article is as follows: In “Literature Review and Hypotheses Development” section, we frame our research in the context of the extant literature; the “Method” section describes the sample and the methodology used; “Empirical Results” section provides the main empirical findings and robustness tests; “Conclusion” section concludes the article highlighting its main implications and limitations.

Literature Review and Hypotheses Development

EU regulation 1606/2002 imposed companies listed on any European country to adopt IFRS in their consolidated accounts from the 1st of January 2005 and gave the possibility to each Member State to decide whether to oblige/allow other kind of companies, that is non-listed ones, to use the same set of standards. As a consequence, in 2012, IFRS can be used by (all or some) private companies in all European countries with the exception of Austria, Belgium, France, Latvia, Romania, Spain, Sweden, and Switzerland where the adoption of international standards by this kind of firms is prohibited. On the contrary, Cyprus, Montenegro, and Serbia require all private companies to follow IFRS for the preparation of their annual report (PricewaterhouseCoopers, 2012).

Given that accounting standards directly influence reporting quality (Soderstrom & Sun, 2007), it has been observed a pressing need to understand the effects of this change on reporting quality as soon as financial statements prepared under IFRS were available. The intent of the IASB is

to develop, in the public interest, a single set of high quality, understandable and enforceable global accounting standards that require high quality, transparent, and comparable information in financial statements and other financial reporting to help participants in the world’s capital markets and other users make economic decisions (IASC Constitution).

Moreover, international accounting standards are expected to enhance the quality of companies’ annual report in most countries (Francis, Khurana, Martin, & Pereira, 2008; Leuz, 2010). As a consequence, the literature has plenty of studies that investigate the effect of the adoption of IFRS on earnings quality (for a recent review, please refer to Brown, 2011). However, this stream of literature only focuses on public companies providing for findings which are not always consistent. Although some authors document an improvement in the quality of earnings under IFRS (e.g., Paananen & Henghsiu, 2009; Tsalavoutas & Evans, 2010), most of the studies reveal very limited improvements, conflicting results across metrics or no differences in earnings quality between IFRS and local GAAP (e.g., Van Tendeloo & Vanstraelen, 2005). Finally, it is also documented that earnings quality decreases after the adoption of IFRS (e.g., Jeanjean & Stolowy, 2008).

Although the impact of IFRS adoption on earnings quality is still an open issue for public companies, at the best of our knowledge, it is even more a pending question for private firms, as no study has directly investigated this issue so far. Francis et al. (2008) investigate the determinants of voluntary adoption of IFRS by non-listed companies, analyzing whether firm-specific incentives matter in their decision to shift to a different body of accounting standards. However, they do not test whether the quality of accounting numbers changes after the adoption of international GAAP. The same is true for a more recent article by Matonti and Iuliano (2012), which focused on Italian private companies.

Earnings quality is still an important issue for private firms. Their incentives in managing financial statements may be different from those of public companies but are still present. Privately held firms have more concentrated ownership and major capital providers often have insider access to corporate information so earnings would not have to be as informative about the true economic performance. Bank is usually the major source of external funds in privately held companies, resulting in agency conflicts between bankers and owners/management (Vander Bauwhede & Willekens, 2004), which could also create earnings management incentives, exacerbated in the case of earnings-based debt covenants.³

Moreover, even in private companies, as well as in public ones, management’s bonuses might be earnings-based. Under this perspective, executive compensation might represent relevant incentives for managers to manipulate earnings, whereas shareholders have a clear interest in obtaining non-manipulated numbers. Finally, in settings where earnings reported in financial reports are the basis for determining tax obligation, management/shareholder would like to pursue tax saving objectives, whereas fiscal authorities have an interest to obtain high-quality accounting numbers. Even if it is clear that incentives to manipulate earnings are present in private companies, it is not evident yet whether these incentives are stronger or weaker compared with listed companies.

Givoly, Hayn, and Katz (2010) notice that two competing hypotheses can be used for testing the differences in incentives and, therefore, in the quality of accounting numbers produced by public and private held companies: the “demand” and “opportunistic behavior” hypothesis. The “demand” argument postulates that earnings of public firms are of higher quality due to stronger demand by shareholders and creditors for high-quality reporting. In contrast, the second (opportunistic behavior) posits that public company managers have a greater incentive to manage earnings in comparison with private ones due to the continuous pressure by investors to meet certain performance benchmarks or as a result of having stock-based compensation. This “opportunistic behavior” hypothesis is supported by the results of Beatty, Ke, and Petroni (2002) and the survey conducted by Penno and Simon (1986). In accordance with the demand hypothesis, previous research shows that European private firms engage in more earnings management than public companies (Ball & Shivakumar, 2005; Burgstahler et al., 2006), as their financial statements are not widely distributed to the public and are more likely to be influenced by tax objectives (Ball & Shivakumar, 2005).

Taking into consideration all the above, whether IFRS adoption impacts on earnings quality in private firms is a relevant but still open question. Previous literature results on the effect of IFRS adoption on earnings quality for public companies are unclear. Moreover, a simple extension of findings of previous research on public companies would not be appropriate as reporting incentives are found to be different between public and private companies. For these reasons, we do not have clear a priori expectations related to the direction with which the shift to IFRS standards affects private companies in Italy. So, our first hypothesis can be stated in the null form:

Hypothesis 1 (H1): Earnings quality among Italian private companies is not different between IFRS and local GAAP adopters.

As IFRS contain several accounting policy options (Kvaal & Nobes, 2012), it has been claimed that there is still room for discretion in IFRS application (e.g., Daske et al., 2008; Leuz, 2010). Companies can exploit the level of flexibility embedded in IFRS to pursue their own reporting interests. Earlier literature suggests that results on the consequences of IFRS adoption may depend on factors reflecting preparer incentives (Brown, 2011; Leuz, 2010; Pope & McLeay, 2011). Previous research (Cameran & Campa, 2010) that analyses the characteristics of Italian private companies that adopt IFRS on a voluntary basis in 2006 and 2007 finds that about 75% of these firms are part of a group where the parent company is a listed company thus obliged to adopt IFRS. In the sample used in the present study, 172 firms out of 270 (63.7% of our IFRS adopters) are controlled by listed entities. We assume that subsidiaries of listed companies may be forced to use international accounting standards on the basis of parent company requirements and/or for simplifying the financial reporting process. This would be consistent with what has been observed in a survey carried out by PricewaterhouseCoopers in 2011, investigating the adoption of IFRS in the United States. The survey states that “IFRS has increasingly become a fact of life for certain companies including … subsidiaries of non-US companies” (PricewaterhouseCoopers, 2011, p. 17). Indeed, according to the current EU regulation, firms listed on one of the European financial markets have to prepare their consolidated statements in accordance with the international accounting standards. Voluntary IFRS adoption by their subsidiaries would permit a simplification in their financial reporting process: They would prepare only one financial statement that would be used both for their parent company consolidated financial statement and local (national) requirements. Otherwise subsidiaries would have to prepare two different financial statements: One in accordance with local GAAP and the other with IFRS for permitting parent company to obtain its consolidated financial report. This is also in line with a PricewaterhouseCoopers (2009) executive survey that supports the conclusion that the adoption of IFRS by subsidiaries may bring “considerable benefits to the company as a whole, for example streamlining the consolidation process” (p. 2).

For other private companies, IFRS adoption may be part of private firms’ signaling strategy. According to Francis et al. (2008), “accounting is likely to play a more important role for private firms (compared with large established corporations) in addressing market imperfections in the form of agency conflicts and information asymmetry” (p. 333). A financial statement contains the only public available accounting data for some private companies’ stakeholders like customers, suppliers, employees, and government (Van Tendeloo & Vanstraelen, 2008). Also banks, that are usually the major source of finance for private companies, consider financial statement data for their decisions. The use of IFRS, a set of accounting standards reputed to be of better quality than national ones (Barth et al., 2008), could imply a “reputational effect” for the voluntary adopter and a wider possibility to contract with outside parties.

As private companies controlled by listed companies obliged to use IFRS might have different incentives to adopt IFRS in comparison with private companies not controlled by listed companies, and given that previous literature suggests that incentives matter in evaluating IFRS adoption consequences (Brown, 2011; Leuz, 2010; Pope & McLeay, 2011), we formulate the following second hypothesis:

Hypothesis 2 (H2): The effect of IFRS adoption on earnings quality of Italian private companies differs between subsidiaries of listed firms and other private companies.

Method

Sample Selection

To test our hypotheses, we select all the Italian non-listed and non-financial companies that use IFRS in their financial statements from 2005 to 2008.⁴ For each firm, we downloaded all the financial information included in their balance sheet, income statement, cash flow statement, and all the other data we need for our tests from the first year after the transition to IFRS⁵ until the most recent financial statement, which was 2009 at the time of the data collection. We excluded firms where financial information was not available. We ended up with a sample of 355 IFRS adopters. We then create a control sample of non-IFRS adopters matched with our IFRS adopters using the propensity-score matching approach developed by Rosenbaum and Rubin (1983). Employing this methodology, we estimate propensity-scores using a logistic model that relates the adoption of IFRS to firm size, leverage, profitability, and industry at the year of the transition to the new accounting standards.⁶ The dependent variable of our logistic model is a dummy variable that is equal to 1 for our sample of IFRS adopters and 0 for a large group of non-IFRS adopters extracted from AIDA database and counting 4,535 companies.⁷ We finally match, without replacement, each IFRS adopter to the non-IFRS adopter that has the closest propensity-score following the “nearest neighbor matching” procedure. As we cannot find a match for all our IFRS adopters, the final sample is composed of 270 pairs of IFRS and non-IFRS adopters and a total of 948 firm-year observations.⁸

Proxies for Earnings Quality

Earnings quality has been traditionally measured in the literature by the investigation of three dimensions: earnings management, timely loss recognition, and value relevance (Dechow, Ge, & Schrand, 2010). Because our article considers non-listed companies, the value relevance dimension is not applicable to our sample. For this reason, we only focus on earnings management and timely loss recognition. The next sections explain in detail the proxies used to estimate these aspects of earnings quality.

Earnings management

Our measure of earnings management considers discretionary or abnormal accruals. This is believed to be the part of total accruals that is more likely to be the result of managerial discretion to achieve particular goals. Prior literature that analyzes earnings management through abnormal accruals often uses Jones-type abnormal accrual measures (Jones, 1991; Kothari, Leone, & Wasley, 2005). However, when the number of observations per year/industry is limited such models might be unreliable (Wysocki, 2004). As the latter is the case of our sample, we use the DeFond and Park (2001) model to estimate abnormal working capital accruals (AWCA) as a proxy for earnings management. This proxy was already used in prior research that investigates the Italian context (e.g., Marra & Mazzola, 2014; Prencipe & Bar-Yosef, 2011). Moreover, the choice of the DeFond and Park (2001) measure of earnings management also permits to limit measurement errors. In fact, according to Kim, Chung, and Firth (2003), Jones (1991) type models have been criticized as the parameter estimates are biased and measurement errors associated therewith could potentially induce erroneous conclusions about the existence of earnings management (e.g., Bernard & Skinner, 1996). DeFond and Park (2001) methodology is independent from potential measurement errors associated with the Jones (1991) model parameters.

DeFond and Park (2001) estimate AWCA using the following formula:

AWC A_{t} = W C_{t} - (\frac{W C_{t - 1}}{S_{t - 1}}) \times S_{t},

where AWCA is the abnormal working capital accrual as defined by DeFond and Park (2001); WC is non-cash working capital accruals, which is calculated as follows: (current assets − cash and short-term investments) − (current liabilities − short-term debt); S is year’s sales.

We use absolute values of AWCA to analyze earnings management per se. This approach is employed when there is no a priori expectation about the direction of earnings manipulation (Becker, DeFond, Jiambalvo, & Subramanyam, 1998; Warfield, Wild, & Wild, 1995).⁹

We measure the relation between accounting standards and abnormal accruals using the following regression (Equation 1):

\begin{matrix} A W C A_{it} = α + β_{1} I F R S_{it} + β_{2} L E V_{it} + β_{3} CF O_{it} + β_{4} R O A_{it} + β_{5} G R O W T H_{it} + β_{6} D I S S U E_{it} \\ + β_{7} F A M I L Y_{it} + β_{8} S I Z E_{it} + β_{9} B I G 4_{it} + β_{10} Q U O T_{it} + β_{i} I N D_{t} + β_{j} Y E A R_{i} + ε_{it}, \end{matrix}

where AWCA_it = abnormal working capital accruals; IFRS_it = 1 if the company adopts IFRS and 0 otherwise; LEV_it = total debt divided by beginning total assets; CFO_it = cash flow from operations divided by beginning total assets; ROA_it = operating profit divided by beginning total assets; GROWTH_it = annual change in net sales; DISSUE_it = annual change in total liabilities; FAMILY_it = 1 if members of a family own the absolute majority of the capital in a direct or in an indirect way and 0 otherwise; SIZE_it = natural logarithm of total assets; BIG4_it = 1 for Big 4 clients and 0 otherwise; QUOT_it = 1 if the firm is controlled by a listed company and 0 otherwise; IND_t = industry dummy variables; YEAR_i = year dummy variables.

We test our first hypothesis looking at the sign and the significance of β₁. If H1 is verified, then β₁ is expected to be not significantly different from zero.

In line with the previous literature, a set of control variables is also included in the regression to control for other firm-level factors that can influence earnings management. Earlier studies have found that financial leverage is positively related with earnings management (e.g., Frankel, Johnson, & Nelson, 2002): For this reason, variable LEV is introduced in our model. Cash flow from operations and return on assets are included in the model to control for extreme performance, which may affect the level of accruals (Kothari et al., 2005; McNichols, 2000). Growth and profitability of the firm can affect the extent of earnings management (Carey & Simnett, 2006), for this reason variables GROWTH and ROA are introduced. Ball and Shivakumar (2008) indicate that there could be a significant effect between net financing changes and unexpected accruals measurement. In the same way, Shan, Taylor, and Walter (2011) show that failure to control for external financing can result in significant measurement errors and erroneous inferences. DISSUE, designed to capture the effect of debt issues, is included in the model. Previous studies (for a review, see Moores & Salvato, 2010) have shown that family firms can be differently sensitive to earnings management in comparison with non-family firms. As family-owned firms are the strongest component of the Italian economy (EIU, 2005) a dummy variable, FAMILY, is introduced to control for this effect. There are several definitions of family firms (e.g., Villalonga & Amit, 2006). They include different combinations of family ownership, management, and control. Accordingly with other studies (e.g., Bennedsen, Nielsen, & Wolfenzon, 2005), our definition is based on control. In our analyses, we classify a firm as a family firm if the members of a single family hold more than 50% of the shares either in a direct or an indirect way. We use a 50% threshold for control as opposed to the lower one used in empirical analysis of publicly held corporations because of the different ownership structures of closely held and publicly held firms. In a publicly held firm, a shareholder with a large minority stake (e.g., 10%-20%) can have effective control. However, because the number of shareholders in a close corporation is smaller, it is likely that a 50% stake is needed to achieve control (e.g., Bennedsen & Wolfenzon, 2000). SIZE, measured as the natural logarithm of total assets, is included as abnormal accruals are found to be negatively related to firms’ dimension, (Bédard, Chtourou, & Courteau, 2004; Warfield et al., 1995). The choice of a firm’s auditor is also likely to affect earnings quality and Big 4 audit firms are usually associated to less earnings management (Francis, 2004). A dummy variable controls for this effect (BIG4). Finally, consistently with our second hypothesis, a dummy variable that identifies subsidiaries of listed parent companies is introduced (QUOT).

Timely loss recognition

As stated by Ball and Shivakumar (2005), timely loss recognition is a crucial attribute of earnings quality, enhancing information usefulness for example in loan agreements. For this reason, this is the second dimension of earnings quality analyzed in this study. Our measure of timely loss recognition is based on the Ball and Shivakumar (2005) accrual-cash flow model that relates accruals (ACC) to cash from operations (CFO). We consider this methodology as it was developed to overcome other models’ shortfall (e.g., Basu, 1997)¹⁰ and was specifically implemented for a private firms’ sample. Consequently, it is better tailored for our research needs. The model is specified by Equation 2.

A C C_{it} = α + γ_{1} D C F O_{it} + γ_{2} C F O_{it} + γ_{3} D C F O_{it} \times C F O_{it} + γ_{i} I N D_{t} + γ_{j} Y E A R_{i} + ε_{it},

where ACC_it = earnings before extraordinary items minus CFO, scaled by beginning total assets; DCFO_it = 1 if CFO is negative and 0 otherwise; CFO_it = cash flow from operations divided by beginning total assets; IND_t = industry dummy variables; YEAR_i = year dummy variables.

It follows from the definition of accruals and cash flow from operations that they tend to be inversely related. For example, collecting cash from selling inventory results in higher CFO but lower ACC because the balance of inventory decreases when a sale is made. This suggests that γ₂ should be negative. However, timely loss recognition may be based on expected, not realized, cash flows and therefore attenuates this relationship. For example, if the reporting entity is experiencing a decline in demand for its products, it likely needs to recognize a loss for the possibility that inventory can only be liquidated below cost. In such a case, ACC will decrease at the same time as the entity is experiencing lower or negative cash from operations. That is, timely recognition of losses may create a positive relationship between ACC and CFO. It follows that timely recognition of unrealized losses should attenuate the negative relationship between accruals and cash from operations (Ball & Shivakumar, 2005). It is therefore expected that γ₃ is positive. We test differences in timely loss recognition by estimating model (Equation 2) separately among IFRS and local GAAP adopters investigating the difference between the coefficients γ₃. In the hypothesis that accounting quality does not change for the former group of companies, so we should not observe any significant difference between the coefficients across the two scenarios. Indeed, a significant and positive (negative) difference would suggest that the adoption of IFRS enhances (decreases) the process of timely recognition of unrealized losses.

The above presented models (Equations 1 and 2) are calculated for the entire matched sample and then separately for IFRS adopters that are subsidiaries of listed companies (with their pairs) and other firms to gather evidence about our second hypothesis. In both these models, standard errors are clustered at firm level.

Empirical Results

Descriptive Statistics and Univariate Analyses

Table 1 shows descriptive statistics for the variables used in our analyses.

Table 1.

Descriptive Statistics.

	n	M	Median	SD	First quartile	Third quartile
AWCA	948	0.118	0.048	0.169	0.015	0.143
ACC	948	−0.045	−0.030	0.050	−0.062	−0.012
IFRS	948	0.500	0.500	0.500	0.000	1.000
LEV	948	0.617	0.667	0.231	0.455	0.802
CFO	948	0.048	0.034	0.087	0.003	0.082
ROA	948	0.019	0.005	0.095	−0.005	0.043
GROWTH	948	0.143	0.049	1.049	−0.116	0.122
DISSUE	948	0.153	0.023	0.634	−0.120	0.247
FAMILY	948	0.474	0.000	0.500	0.000	1.000
SIZE	948	7.548	5.870	3.179	4.785	10.648
BIG4	948	0.371	0.000	0.483	0.000	1.000
QUOT	948	0.332	0.000	0.471	0.000	1.000

Note. AWCA is the absolute value of the abnormal working capital accrual (DeFond & Park, 2001) scaled by beginning total assets; ACC is earnings before extraordinary items minus CFO, scaled by beginning total assets; IFRS is a dummy variable that takes the value of 1 for IFRS adopters and 0 otherwise; LEV is total debt over total assets; CFO is the cash flow from operation scaled by beginning total assets; ROA is operating profit divided by beginning total assets; GROWTH is the annual change in net sales; DISSUE is the annual change in total liabilities; FAMILY is a dummy variable that takes the value of 1 if members of a family own the absolute majority of the capital in a direct or in an indirect way and 0 otherwise; SIZE is the natural logarithm of the total assets; BIG4 is a dummy variable that equals 1 if the firm’s auditor is Deloitte and Touche, Ernst & Young, KPMG or PwC; QUOT is a dummy variable that equals 1 if the firm is controlled by a listed company and 0 otherwise.

Results indicate that companies included in our sample show an average level of AWCA of 0.118 (median = 0.048). Firms run by families represent 47% of the sample while around 33% are owned by listed companies; the use of debt is relevant given that it is more than the half of total assets (61.70%). Finally, 37% of the sample is audited by a Big 4 audit firm.

Correlation Analysis

We report correlation matrix in Table 2.

Table 2.

Pearson Correlation Matrix.

	AWCA	ACC	IFRS	LEV	CFO	ROA	GROWTH	DISSUE	FAMILY	SIZE	BIG4	QUOT
AWCA	1
ACC	.154***	1
IFRS	.242***	.157***	1
LEV	.101***	.059*	−.031	1
CFO	−.017	−.478***	.041	−.195***	1
ROA	.045	−.037	−.006	−.109***	.719***	1
GROWTH	.177***	−.058*	−.067**	.083**	.121***	.155***	1
DISSUE	−.021	−.108***	−.102***	.186***	.094***	.078**	.390***	1
FAMILY	.029	−.031	−.108***	.049	.074**	.101***	.010	.023	1
SIZE	.118***	.130***	.428***	.033	.019	.035	−.037	−.028	−.061*	1
BIG4	.171***	.114***	.664***	−.042	.056*	.027	−.034	−.063*	−.069**	.472***	1
QUOT	.266***	.096***	.683***	−.033	.052	.015	−.051	−.072**	−.077**	.288***	.570***	1

, **, *** indicate that a coefficient is statistically significant at the 10%, 5%, and 1% level or better, two-tailed.

The level of AWCA is positively associated with IFRS, suggesting that the IFRS adopters exhibit larger abnormal accruals that would lead to a decrease in reporting quality. As expected, AWCA is also positively associated with ACC, and with variables LEV, GROWTH, SIZE, BIG, and QUOT. Looking at other correlations, we notice that IFRS shows a positive association with QUOT indicating that IFRS adopters are more likely to be part of a group where the parent company is listed on a financial market.

We also observe significant correlations between variables that are used as control variables in our regression. For example, BIG4 and QUOT exhibit a significant positive correlation coefficient (ρ = .570), so do CFO and ROA (ρ = .719). High correlation coefficients between control variables might affect the regression results because of potential multicollinearity problems. For this reason, we test the robustness of all our results using the Variance Inflation Factor (VIF) test and it shows that multicollinearity is not biasing our results.¹¹

Regression Analysis

Voluntary IFRS adoption and earnings management

First of all, we test whether the voluntary adoption of IFRS affected the reporting quality of non-listed companies on the basis of our discretionary accrual proxy (Table 3).

Table 3.

Voluntary IFRS Adoption and Abnormal Accruals.

DEPENDENT VARIABLE	AWCA
INTERCEPT_it	−0.005
	(0.951)
IFRS_it	0.070***
	(0.001)
LEV_it	0.091***
	(0.001)
CFO_it	−0.187*
	(0.062)
ROA_it	0.159*
	(0.078)
GROWTH_it	0.040***
	(0.000)
DISSUE_it	−0.020**
	(0.038)
FAMILY_it	0.024*
	(0.075)
SIZE_it	0.001
	(0.671)
BIG4_it	−0.023
	(0.236)
QUOT_it	0.071***
	(0.000)
Observations	948
R ²	0.162
Industry and year dummies	Yes

Note. Regression model:

$\begin{matrix} A W C A_{it} = α + β_{1} I F R S_{it} + β_{2} L E V_{it} + β_{3} C F O_{it} + β_{4} R O A_{it} + β_{5} G R O W T H_{it} + β_{6} D I S S U E_{it} + β_{7} F A M I L Y_{it} + β_{8} S I Z E_{it} + β_{9} B I G 4_{it} \\ + β_{10} Q U O T_{it} + β_{i} I N D_{t} + β_{j} Y E A R_{i} + ε_{it}, \end{matrix}$

where AWCA is the absolute value of the abnormal working capital accrual (DeFond & Park, 2001) scaled by beginning total assets; IFRS is a dummy variable that takes the value of 1 for IFRS adopters and 0 otherwise; LEV is total debt over total assets; CFO is the cash flow from operation scaled by beginning total assets; ROA is operating profit divided by beginning total assets; GROWTH is the annual change in net sales; DISSUE is the annual change in total liabilities; FAMILY is a dummy variable that takes the value of 1 if members of a family own the absolute majority of the capital in a direct or in an indirect way and 0 otherwise; SIZE is the natural logarithm of the total assets; BIG4 is a dummy variable that equals 1 if the firm’s auditor is Deloitte and Touche, Ernst & Young, KPMG or PwC; QUOT is a dummy variable that equals 1 if the firm is controlled by a listed company and 0 otherwise; IND is an industry dummy variable; YEAR is a year dummy variable. Variables of interest are highlighted in bold. IFRS = International Financial Reporting Standards.

, **, *** indicate that a coefficient is statistically significant at the 10%, 5%, and 1% level or better, two-tailed.

Results indicate that companies that switched to IFRS exhibit higher levels of AWCA (β₁ = .070; p = .001), with a corresponding decrease in reporting quality, in comparison with a control group of pairs that still use local GAAP during the same time-period investigated.

Most of our control variables show a relation with the level of AWCA generally consistent with the literature. In particular, the level of AWCA is positively associated with leverage, return on assets, firm growth and family firms while it is negatively associated with cash flow from operations and issuance of debt, given the higher level of scrutiny the company is usually subject to when asking for additional funding (Rodríguez-Péréz & Van Hemmen, 2010). Interestingly, results also highlight a positive and very significant coefficient between AWCA and QUOT (β₁ = .071; p = .000), suggesting that subsidiaries of listed companies are, on average, characterized by lower reporting quality.

Voluntary IFRS adoption and timely loss recognition

We report the analysis of the impact of IFRS adoption on timely loss recognition in Table 4.

Table 4.

Voluntary IFRS Adoption and Timely Loss Recognition.

DEPENDENT VARIABLE	(A): ACC (IFRS adopters)	(B): ACC (Non-IFRS adopters)	(A – B): (Differences)
INTERCEPT_it	−0.010	−0.046	0.036
	(0.672)	(0.190)	(0.433)
DCFO_it	0.008	0.007	0.001
	(0.231)	(0.234)	(0.962)
CFO_it	−0.216***	−0.493***	0.277***
	(0.000)	(0.000)	(0.000)
DCFO_it×CFO_it	0.313***	0.903***	−0.590***
	(0.000)	(0.000)	(0.000)
Observations	474	474
R ²	0.349	0.632
Industry and year dummies	Yes	Yes

Note. Regression model:

$A C C_{it} = α + γ_{1} D C F O_{it} + γ_{2} C F O_{it} + γ_{3} D C F O_{it} \times CF O_{it} + γ_{i} IN D_{t} + γ_{j} Y E A R_{i} + ε_{it},$

where ACC is earnings before extraordinary items minus CFO, scaled by beginning total assets; DCFO is a dummy variable taking the value 1 if CFO is negative and 0 otherwise; CFO is the cash flow from operation scaled by beginning total assets; IND is an industry dummy variable; YEAR is a year dummy variable. Variables of interest are highlighted in bold. IFRS = International Financial Reporting Standards.

***

indicate that a coefficient is statistically significant at the 1% level or better, two-tailed.

It is here matter of interest the significance of the difference between the coefficients γ₃ from model (2) estimated both among IFRS and non-IFRS adopters. The difference is negative and highly significant (γ_{3IFRS ADOPTERS}−γ_{3NON-IFRS ADOPTERS} = −0.590; p = .000), evidence that losses are less timely recognized by IFRS adopters. It suggests a decrease in timely loss recognition as a consequence of IFRS adoption with a correspondent negative impact on earnings quality, consistently with our earnings management indicator.¹²

Voluntary IFRS adoption and earnings quality between subsidiaries of listed companies and other companies

Findings, so far, did not highlight any improvement in earnings quality among private companies that voluntarily adopt IFRS in comparison with the matched group of firms that use local GAAP. On the contrary, results are consistent with lower levels of reporting quality among the former, according to both the dimensions investigated.

In this section, we test our second hypothesis analyzing whether the effect of IFRS adoption is different for firms controlled by a listed entity and other companies. Almost 63.7% (172 firms out of 270) of voluntary IFRS adopters in our sample are controlled by companies listed on a financial market that must use IFRS for their consolidated accounts on a mandatory basis.

The estimation of model in Equation 1 separately for IFRS adopters controlled by a listed company and IFRS adopters that are not controlled by a listed company, together with their corresponding pairs that do not use IFRS, is reported, respectively, in Columns (A) and (B) of Table 5.¹³

Table 5.

Voluntary IFRS Adoption and Abnormal Accruals: Firms Controlled/Not Controlled by Listed Companies.

DEPENDENT VARIABLE	(A): AWCA (IFRS adopters controlled by listed companies and their pairs)	(B): AWCA (IFRS adopters not controlled by listed companies and their pairs)	(A – B): Difference
INTERCEPT_it	0.095	−0.041	0.136
	(0.437)	(0.704)	(0.225)
IFRS_it	0.199***	0.023	0.176***
	(0.000)	(0.444)	(0.000)
LEV_it	0.065*	0.136**	−0.071
	(0.078)	(0.013)	(0.706)
CFO_it	−0.320**	0.064	−0.384*
	(0.013)	(0.714)	(0.073)
ROA_it	0.219**	0.090	0.129
	(0.046)	(0.615)	(0.509)
GROWTH_it	0.047***	0.045***	0.002
	(0.000)	(0.000)	(0.900)
DISSUE_it	−0.019	−0.036**	0.017
	(0.144)	(0.025)	(0.847)
FAMILY_it	0.019	0.006	0.013
	(0.292)	(0.805)	(0.912)
SIZE_it	0.001	−0.002	0.003
	(0.716)	(0.737)	(0.143)
BIG4_it	−0.077***	0.014	−0.091**
	(0.006)	(0.647)	(0.049)
Observations	620	328
R ²	0.235	0.102
Industry and year dummies	Yes	Yes

Note. Regression model:

$\begin{array}{l} A W C A_{i t} = α + β_{1} I F R S_{i t} + β_{2} L E V_{i t} + β_{3} C F O_{i t} + β_{4} R O A_{i t} + β_{5} G R O W T H_{i t} + β_{6} D I S S U E_{i t} \\ + β_{7} F A M I L Y_{i t} + β_{8} S I Z E_{i t} + β_{9} B I G 4_{i t} + β_{10} Q U O T_{i t} + β_{i} I N D_{t} + β_{j} Y E A R_{i} + ε_{i t}, \end{array}$

where AWCA is the absolute value of the abnormal working capital accrual (DeFond & Park, 2001) scaled by beginning total assets; IFRS is a dummy variable that takes the value of 1 for IFRS adopters and 0 otherwise; LEV is total debt over total assets; CFO is the cash flow from operation scaled by beginning total assets; ROA is operating profit divided by beginning total assets; GROWTH is the annual change in net sales; DISSUE is the annual change in total liabilities; FAMILY is a dummy variable that takes the value of 1 if members of a family own the absolute majority of the capital in a direct or in an indirect way and 0 otherwise; SIZE is the natural logarithm of the total assets; BIG4 is a dummy variable that equals 1 if the firm’s auditor is Deloitte and Touche, Ernst & Young, KPMG or PwC; IND is an industry dummy variable; YEAR is a year dummy variable. Variables of interest are highlighted in bold. IFRS = International Financial Reporting Standards.

, **, *** indicate that a coefficient is statistically significant at the 10%, 5%, and 1% level or better, two-tailed.

Results are different between these two groups of firms. The coefficient associated to the variable IFRS is positive and significant at the 1% level (β₁ = .199; p = .000) in Column (A). This suggests that IFRS adopters controlled by listed companies exhibit higher levels of abnormal accruals in comparison with a control sample of companies that still prepare annual reports in accordance with the Italian GAAP.

The same coefficient is not significant (β₁ = .023; p = .444) in Column (B) of Table 5 indicating that the use of abnormal accruals among IFRS adopters is not more pervasive than firms which use Italian GAAP if the former are not part of a group where the parent company is listed.

The difference between the two coefficients reported above is also statistically significant at the 1% level (p = .000), suggesting that IFRS adoption has a significantly worse impact on abnormal accruals among subsidiaries of listed companies.

The results reported in Table 6 refer to the timely loss recognition aspect of earnings quality.

Table 6.

Voluntary IFRS Adoption and Timely Loss Recognition: Firms Controlled/Not Controlled by Listed Companies.

Panel A: IFRS Adopters Controlled by Listed Companies and Their Pairs.
DEPENDENT VARIABLE	(A): ACC (IFRS adopters)	(B): ACC (Non-IFRS adopters)	(A – B): (Differences)
INTERCEPT_it	−0.020	0.041	−0.061
	(0.518)	(0.221)	(0.277)
DCFO_it	0.020***	0.005	0.015
	(0.009)	(0.476)	(0.315)
CFO_it	−0.178***	−0.498***	0.320***
	(0.000)	(0.000)	(0.000)
DCFO_it×CFO_it	0.325***	0.839***	−0.514***
	(0.000)	(0.000)	(0.000)
Observations	310	310
R ²	0.460	0.666
Industry and year dummies	Yes	Yes
Panel B: IFRS Adopters Not Controlled by Listed Companies and Their Pairs.
DEPENDENT VARIABLE	(A): ACC (IFRS adopters)	(B): ACC (Non-IFRS adopters)	(A – B): (Differences)
INTERCEPT_it	−0.009	−0.059	0.050
	(0.667)	(0.155)	(0.845)
DCFO_it	−0.017	0.009	−0.026
	(0.193)	(0.504)	(0.185)
CFO_it	−0.232***	−0.428***	−0.196***
	(0.000)	(0.006)	(0.001)
DCFO_it×CFO_it	−0.128	0.924***	−1.052***
	(0.617)	(0.000)	(0.000)
Observations	164	164
R ²	0.385	0.596
Industry and year dummies	Yes	Yes
Differences-in-differences			0.538 (0.893)

Note. Regression model:

$AC C_{it} = α + γ_{1} DCF O_{it} + γ_{2} CF O_{it} + γ_{3} DCF O_{it} \times CF O_{it} + β_{i} IN D_{t} + β_{j} YEA R_{i} + ε_{it},$

***

indicate that a coefficient is statistically significant at the 1% level or better, two-tailed.

Panel A of Table 6 highlights results for IFRS adopters controlled by listed companies and their non-IFRS pairs. The difference between the coefficients γ₃ between the two groups is negative and highly significant (γ_{3IFRS ADOPTERS}−γ_{3NON-IFRS ADOPTERS} = −0.514; p = .000), indicating that losses are less timely recognized among IFRS adopters. The same results are found in Panel B of Table 6 that investigates IFRS adopters that are not controlled by listed entities and their non-IFRS pairs (γ_{3IFRS ADOPTERS}−γ_{3NON-IFRS ADOPTERS} = −1.052; p = .000). The difference of the differences between the coefficients of these two sub-samples is not significant (p = .893). Differently from the evidence coming from the analysis of discretionary accruals, it suggests that earnings quality, measured as timely loss recognition, becomes worse for all IFRS adopters and the impact is similar regardless if they are part of a group where the parent company is listed or not.

Robustness Tests

To test the robustness of our results, we run a set of additional analyses.

First of all, we investigate earnings quality using an additional dimension: income smoothing. It has been found that managers tend to smooth earnings for several reasons, for example, as a signal, because current earnings can be used as a predictor of future income (e.g., Chaney & Lewis, 1995). Moreover, income smoothing could be pursued to reduce earnings volatility, usually perceived as an increasing risk factor for investors/creditors impacting on the cost of capital/debt (e.g., Barth, Landsman, & Wahlen, 1995). Finally, income smoothing could be a mean to divert political attention from too high or too low income (e.g., Watts & Zimmerman, 1986). We estimate income smoothing as the variability of annual changes in net income. However, changes in net income can be attributed to other factors such as the economic environment and characteristics of the firms. For this reason, the earnings variability metric used in this article is the variance of the residuals from the regression of change in net income on several control variables identified in prior research (e.g., Lang, Raedy, & Wilson, 2006; Tarca, 2004). In accordance with these studies, the model is run on a set of control factors, which are expressed in Equation 3, where standard errors are clustered at the firm level:

\begin{matrix} Δ N i_{it} = α + β_{1} SIZ E_{it} + β_{2} GROWT H_{it} + β_{3} LE V_{it} + β_{4} EISSU E_{it} + β_{5} DISSU E_{it} + β_{6} TUR N_{it} + β_{7} CF O_{it} + β_{8} BIG 4_{it} + β_{i} IN D_{t} + β_{j} YEA R_{i} + ε_{it}, \end{matrix}

where ΔNi_it = change in net income; SIZE_it = natural logarithm of total assets; GROWTH_it = annual change in net sales; LEV_it = total debt divided by beginning total assets; EISSUE_it = annual change in shareholder’s equity; DISSUE_it = annual change in total liabilities; TURN_it = total sales divided by beginning total assets; CFO_it = cash flow from operations divided by beginning total assets; BIG4_it = 1 for Big 4 clients and 0 otherwise; IND_t = industry dummy variables; YEAR_i = year dummy variables.

We calculate regression (Equation 3) separately for the observations under Italian GAAP and under IFRS. The residuals of the model specified above, denoted as ΔNi* and, more exactly, their standard deviation σΔNi*, is the metric used for assessing earnings smoothing. Lower values of σΔNi* are evidence of increasing earnings smoothing, and vice versa. We compare the standard deviations using Levene’s Test.¹⁴ Results are reported in Table 7.

Table 7.

Voluntary IFRS Adoption and Income Smoothing.

Panel A: All Sample.
Sample	σ_ΔNi*	Levene’s test (p)
Non-IFRS adopters	2.752	(0.000***)
IFRS adopters	0.130	(0.000***)
Panel B: IFRS Adopters Controlled by Listed Companies and Their Pairs.
Sample	σ_ΔNi*	Levene’s test (p)
Non-IFRS adopters	2.875	(0.000***)
IFRS adopters	0.148	(0.000***)
Panel C: IFRS Adopters Not Controlled by listed Companies and Their Pairs.
Sample	σ_ΔNi*	Levene’s test (p)
Non-IFRS adopters	3.054	(0.000***)
IFRS adopters	0.073	(0.000***)

Income smoothing is calculated as the standard deviation of residuals from Equation 3:

$\begin{matrix} Δ N i_{t} = α + β_{1} SIZ E_{it} + β_{2} GROWT H_{it} + β_{3} LE V_{it} + β_{4} EISSU E_{it} + β_{5} DISSU E_{it} + β_{6} TUR N_{it} + \\ β_{7} CF O_{it} + β_{8} BIG 4_{it} + β_{i} IN D_{t} + β_{j} YEA R_{i} + ε_{it}, \end{matrix}$

where ΔNi is the change in net income; SIZE is the natural logarithm of the total assets; GROWTH is the annual change in net sales; LEV is total debt over total assets; EISSUE is the annual change in shareholder’s equity; DISSUE is the annual change in total liabilities; TURN is total sales divided by beginning total assets; CFO is the cash flow from operation scaled by beginning total assets; BIG4 is a dummy variable that equals 1 if the firm’s auditor is Deloitte and Touche, Ernst & Young, KPMG or PwC; IND is an industry dummy variable; YEAR is a year dummy variable. Differences in standard deviations are calculated using Levene’s test. IFRS = International Financial Reporting Standards.

***

indicate that a coefficient is statistically significant at the 1% level or better, two-tailed.

Findings corroborate our evidence that suggests that the adoption of international accounting standards does not lead to an increase in reporting quality. Panel A of Table 7 reports the evidence for the entire sample. It highlights that the standard deviation of residuals from Equation 3 measures 0.130 among firms that voluntarily adopt IFRS while it is 2.752 among our control sample of companies that still use local GAAP. A Levene’s test indicates that the difference between the two standard deviations is strongly significant (p = .000), suggesting that IFRS increased the level of income smoothing of private companies that can be interpreted as a decrease in earnings quality, in accordance with findings derived from our discretionary accrual analysis. Panel B of Table 7 includes the results for IFRS adopters controlled by a listed company and their pairs. The standard deviation of residuals from Equation 2 is 2.875 for non-IFRS adopters while it is 0.148 for IFRS adopters, suggesting higher income smoothing and, consequently, lower earnings quality, among the latter group of companies as Levene’s test indicates that these two figures are significantly different at the 1% level. The same evidence is found in Panel C that investigates income smoothing among IFRS adopters, which are not controlled by listed companies and their pairs. The standard deviation of residuals from Equation 2 is 3.054 for non-IFRS adopters while it measures 0.073 for IFRS adopters. A test on the difference between these two values is strongly statistically significant (p = .000) and indicates that income smoothing is higher also among the IFRS adopters that are not part of a group where the company at the apex is listed. In accordance with the evidence coming from the analysis of timely loss recognition, findings in Table 7 suggest that earnings quality, measured as the level of income smoothing, becomes worse for IFRS adopters regardless of whether they are part of a group where the parent company is listed on a financial market or not.

We then re-estimate our model in Equation 1 separating income-increasing and income-decreasing abnormal accruals. Evidence (untabled) indicates that, in relation to the entire sample, IFRS adopters use abnormal accruals more extensively than their non-IFRS pairs.

With reference to income-increasing accruals, the more extensive use of abnormal accruals by IFRS adopters is documented when we consider both firms that are controlled by a listed company (β₁ = .121; p = .012) and when the investigation involves those which are not part of a group where the parent company is listed (β₁ = .146; p = .016). We do not find differences between these two sub-samples as the difference between the above highlighted coefficients is not significant (p = .398).

Findings from income-decreasing abnormal accruals highlight that their use is more pervasive among IFRS adopters that are controlled by listed entities (β₁ = −.243; p = .000),¹⁵ while a non-significant relationship between abnormal accruals and accounting standards is observed among the group of IFRS adopters that are not controlled by listed companies (β₁ = .027; p = .500). In addition, the difference between the coefficients reported above is significant at the 1% level (p = .000).

So, taking together the results of income smoothing and positive/negative accruals analyses, we can see once more that IFRS adoption does not imply any improvement from the earnings quality point of view. Again, the impact of the new set of standards on financial reporting quality seems to be slightly worse for firms controlled by listed companies.

Finally, as the economic environment went into a severe recession in 2007, we control for this event by including in all our models a dummy variable that equals 1 for the year affected by the crisis (2008-2009) and 0 for all other year (untabled). All evidence reported thus far still holds.¹⁶

As already mentioned, earnings quality might be driven by underlining incentives (Brown, 2011; Leuz, 2010; Pope & McLeay, 2011). Given that the decision to switch to IFRS standards is made on a voluntary basis for private companies, one could argue that endogeneity issues might bias our results. The matched sample methodology that we use in our main analyses should already limit this threat, as our local GAAP pairs are chosen on the basis of similar size, leverage, profitability, and industry. However, to be on the safe side, we employ an alternative methodology to control for potential endogeneity. In particular, we investigate the use of abnormal accruals and timely loss recognition using the larger group of 355 IFRS adopters and observations pre- and post-IFRS adoption.¹⁷ The comparison between local GAAP and IFRS is therefore made within the same companies. More precisely, we collect financial data under Italian GAAP for the period pre-IFRS adoption to have the same number of observations before and after IFRS introduction, symmetric around the transition. This process yields a sample of 702 firms-year observations under IFRS and 702 firm-year observations under Italian GAAP. Results are reported in Tables 8 and 9, respectively.

Table 8.

Voluntary IFRS Adoption and Abnormal Accruals (Pre- and Post-IFRS Implementation).

DEPENDENT VARIABLE	(A): AWCA (All sample)	(B): AWCA (Firms controlled by listed companies)	(C): AWCA (Firms not controlled by listed companies)	(B – C): Differences
INTERCEPT_it	0.072***	0.134***	0.067**	0.067**
	(0.010)	(0.000)	(0.045)	(0.043)
IFRS_it	0.041***	0.050***	0.025**	0.025
	(0.000)	(0.000)	(0.035)	(0.160)
LEV_it	0.075**	0.069*	0.077*	−0.008
	(0.015)	(0.066)	(0.097)	(0.893)
CFO_it	−0.066	−0.131	0.120	−0.251*
	(0.511)	(0.236)	(0.229)	(0.091)
ROA_it	0.004	0.011***	−0.002	0.013***
	(0.416)	(0.000)	(0.435)	(0.000)
GROWTH_it	0.056***	0.068***	0.035**	0.033*
	(0.000)	(0.000)	(0.035)	(0.086)
DISSUE_it	0.002	−0.001	0.015	−0.016
	(0.878)	(0.927)	(0.481)	(0.499)
FAMILY_it	0.030**	0.032**	0.021	0.011
	(0.015)	(0.040)	(0.222)	(0.645)
SIZE_it	−0.005***	−0.004**	−0.005*	0.001
	(0.003)	(0.040)	(0.077)	(0.731)
BIG4_it	−0.014	−0.030*	0.004	−0.034
	(0.278)	(0.084)	(0.803)	(0.160)
QUOT_it	0.056***
	(0.000)
Observations	1,404	948	456
R ²	0.155	0.169	0.113
Industry and year dummies	Yes	Yes	Yes

Note. Regression model:

$\begin{matrix} AWC A_{it} = α + β_{1} IFR S_{it} + β_{2} LE V_{it} + β_{3} CF O_{it} + β_{4} RO A_{it} + β_{5} GROWT H_{it} + β_{6} DISSU E_{it} + β_{7} FAMIL Y_{it} + β_{8} SIZ E_{it} + β_{9} BIG 4_{it} + \\ β_{10} QUO T_{it} + β_{i} IN D_{t} + β_{j} YEA R_{i} + ε_{it}, \end{matrix}$

where AWCA is the absolute value of the abnormal working capital accrual (DeFond & Park, 2001) scaled by beginning total assets; IFRS is a dummy variable that takes the value of 1 for the period after IFRS adoption and 0 otherwise; LEV is total debt over total assets; CFO is the cash flow from operation scaled by beginning total assets; ROA is operating profit divided by beginning total assets; GROWTH is the annual change in net sales; DISSUE is the annual change in total liabilities; FAMILY is a dummy variable that takes the value of 1 if members of a family own the absolute majority of the capital in a direct or in an indirect way and 0 otherwise; SIZE is the natural logarithm of the total assets; BIG4 is a dummy variable that equals 1 if the firm’s auditor is Deloitte and Touche, Ernst & Young, KPMG or PwC; QUOT is a dummy variable that equals 1 if the firm is controlled by a listed company and 0 otherwise; IND is an industry dummy variable; YEAR is a year dummy variable. Variables of interest are highlighted in bold. IFRS = International Financial Reporting Standards.

, **, *** indicate that a coefficient is statistically significant at the 10%, 5%, and 1% level or better, two-tailed.

Table 9.

Voluntary IFRS Adoption and Timely Loss Recognition (Pre- and Post-IFRS Implementation).

Panel A: All Firms.
DEPENDENT VARIABLE	ACC (After IFRS adoption)	ACC (Before IFRS adoption)	(Differences)
INTERCEPT_it	−0.018***	−0.011**	−0.007
	(0.000)	(0.031)	(0.185)
DCFO_it	−0.003	−0.011	0.008
	(0.622)	(0.214)	(0.401)
CFO_it	−0.272***	−0.450***	0.178***
	(0.000)	(0.000)	(0.005)
DCFO_it×CFO_it	0.366***	0.706***	−0.340***
	(0.000)	(0.000)	(0.001)
Observations	702	702
R ²	0.160	0.370
Industry and year dummies	Yes	Yes
Panel B: Firms Controlled by Listed Companies and Their Pairs.
DEPENDENT VARIABLE	ACC (After IFRS adoption)	ACC (Before IFRS adoption)	(Differences)
INTERCEPT_it	−0.020***	−0.002	−0.018***
	(0.000)	(0.733)	(0.003)
DCFO_it	0.005	−0.012	0.017
	(0.468)	(0.231)	(0.116)
CFO_it	−0.254***	−0.532***	0.278***
	(0.000)	(0.000)	(0.000)
DCFO_it×CFO_it	0.339***	0.799***	−0.460***
	(0.000)	(0.000)	(0.000)
Observations	474	474
R ²	0.146	0.453
Industry and year dummies	Yes	Yes
Panel C: Firms Not Controlled by Listed Companies and Their Pairs.
DEPENDENT VARIABLE	ACC (After IFRS adoption)	ACC (Before IFRS adoption)	(Differences)
INTERCEPT_it	−0.014**	−0.027***	0.013**
	(0.017)	(0.000)	(0.025)
DCFO_it	−0.009	−0.027	0.018
	(0.371)	(0.144)	(0.342)
CFO_it	−0.311***	−0.281***	−0.030
	(0.003)	(0.006)	(0.730)
DCFO_it×CFO_it	0.661**	0.408**	0.253
	(0.023)	(0.013)	(0.421)
Observations	228	228
R ²	0.231	0.183
Industry and year dummies	Yes	Yes
Differences-in-differences			−0.713 ** (0.031)

Note. Regression model:

$AC C_{it} = α + γ_{1} DCF O_{it} + γ_{2} CF O_{it} + γ_{3} DCF O_{it} \times CF O_{it} + γ_{i} IN D_{t} + γ_{j} YEA R_{i} + ε_{it},$

, *** indicate that a coefficient is statistically significant at the 5% and 1% level or better, two-tailed.

Table 8 investigates the abnormal accrual dimension through the estimation of model in Equation 1. Column (A) includes results for the entire sample and evidence indicates that the shift to IFRS seems to have increased the level of AWCA (β₁ = .041; p = .000) with a corresponding decrease in reporting quality. Columns (B) and (C) present results related to IFRS adopters controlled by a listed company and IFRS adopters that are not part of a group where the apex is a listed company, respectively. The coefficient associated with the variable IFRS is positive and significant for both groups of companies suggesting that the use of international accounting standards increases the level of abnormal accruals regardless the characteristics of the controlling shareholder. The coefficient β₁ is however higher and more significant among entities controlled by listed companies (β₁ = .050; p = .000) compared with other firms (β₁ = .025; p = .035). The difference between the two coefficients is statistically (weakly) significant only if considered one-tailed (p = .160), suggesting that IFRS have a slightly more negative impact on earnings quality among subsidiaries of listed companies.

Table 9 focuses on the timely loss recognition dimension. Panel A shows that, on average, losses are less timely recognized after IFRS adoption (γ_{3AFTER_IFRS}−γ_{3BEFORE_IFRS} = −0.340; p = .001). Panel B of Table 9 highlights results for subsidiaries of listed companies. The difference between the coefficients γ₃ after and before the adoption of IFRS is negative and highly significant (γ_{3AFTER_IFRS}−γ_{3BEFORE_IFRS} = −0.460; p = .000), suggesting again that losses are less timely recognized after IFRS adoption. On the contrary, Panel C reports no difference in reporting quality under the timely loss recognition dimension among firms that are not controlled by listed company after switching the type of accounting standards adopted (γ_{3AFTER_IFRS}−γ_{3BEFORE_IFRS} = 0.253; p = .421). In this case, the difference of the differences in these coefficients between the two sub-samples is significant (p = .031). It indicates that the effect of the adoption of IFRS on timely loss recognition is different between the two groups of companies.

We repeated the same analyses reported in Tables 8 and 9 on the smaller group of 270 IFRS adopters used in the main tests (i.e., the IFRS adopters sample used in match pair analysis) and the evidence (untabled) is unchanged.

Overall, findings consistently indicate that IFRS adoption does not improve earnings quality among Italian private companies, on the contrary, on average, an increase in the abnormal accruals and a deterioration of timely loss recognition is observed. The effect of IFRS on earnings quality seems to be worse when IFRS adopters are part of a group where the company at the apex is listed on a financial market, for whom we assume that IFRS adoption is driven by meeting headquarters’ requirements and/or by simplifying the financial reporting process.

Conclusion

The aim of this study is to analyze the impact of adopting IFRS on financial reporting quality of private companies. EU regulation 1606/2002 required companies listed on any European exchange to adopt IFRS from the 1st of January 2005 and gave each Member State the right to decide whether to oblige/allow other kind of companies, for example, non-listed ones, to use the same set of standards. Given the importance of understanding the effects of this regulation on accounting quality, there have been many studies exploring changes in earnings quality for publicly listed companies. However, the findings of these studies are mixed and do not reach any definitive conclusions regarding the impact of IFRS adoption on financial reporting quality. It is therefore of considerable interest to explore the effects on private firms. Private companies constitute about 75% of European GDP (Ecoda, 2010) and the effects of adopting International Accounting Standards on non-listed companies has not been addressed in earlier studies.

We compare earnings reporting quality between a group of private companies that switched to IFRS in the period 2005-2008 and a matched set of companies that, in the same period, used local GAAP. Overall, our results show that IFRS did not contribute to the improvement of financial reporting quality among private companies in Italy. On the contrary, we find evidence suggesting that the adoption of this set of standards seems to have increased earnings management (measured by absolute AWCA) and to have led to a deterioration of timely loss recognition. These findings suggest that the adoption of a set of accounting standards reputed to be of better quality than national ones (Barth et al., 2008; Leuz, 2010) does not imply, per se, better financial reporting quality.

To further investigate the last point, we repeat our analyses distinguishing private companies that are controlled by listed companies from the rest of the sample. As previous literature has pointed out that companies can exploit the level of flexibility embedded in IFRS to pursue their own reporting interests (Daske et al., 2008; Kvaal & Nobes, 2012; Leuz, 2010), separate analyses were conducted taking into consideration firms’ incentives. Overall, we find that earnings quality shows signs of deterioration in both groups of firms. However, the analysis of the differences indicates a slightly worse impact among subsidiaries of listed companies, especially in relation to the use of abnormal accruals. Robustness tests using only the set of IFRS adopters, thus controlling for endogeneity problems, confirm our main results.

We believe that our results have several policy implications. First, we document that IFRS adoption by private companies resulted in increased earnings management, which is worse among listed companies’ subsidiaries. IFRS were aimed at improving comparability and relevance of financial reporting.¹⁸ These characteristics, while being of critical importance for outside investors in making economic decisions, might be less critical for private companies, more bounded to local territory activities and not oriented to market’s sources of financing. Under this perspective, EU State Members’ decisions to enforce IFRS among private companies should be carefully considered, especially if firms might perceive that transition costs outweigh the benefits (American Institute of Certified Public Accountants, 2008).

The negative effects of IFRS on financial reporting quality seem to be particularly marked in the case of companies controlled by listed entities that are also most prone to adopt IFRS as a result of an indirect effect of the current EU regulation (Committee of European Securities Regulators, 2003). For this reason, our results should be of interest not only for the EU State Members’ regulators to allow the use of IFRS to their national private companies but also for the EU Commission in evaluating the current EU regulation on financial reporting for private companies. IASB published SME-IFRS (i.e., standards for small- and medium-sized entities) with the objective to better serve private small entities reporting needs, whereas the EU Commission issued the Directive 2012/6 of 14 March 2012 that exempts micro-firms from the preparation of the financial statement.¹⁹ Future research will have to provide empirical evidence to understand whether the introduction of such tailored standards will lead to improved reporting quality among this economically critical segment of the market.

We acknowledge that this study may suffer from a number of limitations. First of all, we assume that the decision of private entities controlled by listed companies to switch to IFRS might be influenced by one common reason: meeting headquarters’ requirements and/or simplifying the financial reporting process. Unfortunately, it is not possible for external researchers to identify the actual motivations behind the choice of adopting IFRS using publicly available data. We therefore acknowledge that, for some of those companies, other incentives might have led to IFRS adoption. Even so, our empirical evidence documents a slightly different impact of the switch to IFRS between subsidiaries of listed companies and the rest of the sample.

In this study, we limit our tests to a single country (Italy). We acknowledge that this might affect the generalizability of our results. However, this choice permits to avoid our findings being influenced by country-specific factors that may severely affect empirical results on the topic (Leuz, 2010; Soderstrom & Sun, 2007). It would still be interesting to understand whether the lack of reporting quality improvements for private companies after IFRS adoption can be detected in other countries as well, especially for those cases in which IFRS adoption has been mandatorily enforced among non-listed companies.

Footnotes

Acknowledgements

We thank the editor, Bharat Sarath, and the anonymous reviewer for the guidance and constructive comments. The research assistance of Chiara Bonfanti is gratefully acknowledged. Finally, we thank the participants of the Fourth Financial Reporting Workshop (Luiss, Rome) and of the Sixth International Workshop on Accounting & Regulation (Siena) for their comments and suggestions.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was supported by the Claudio Dematté Research Division of the SDA Bocconi School of Management (Research Projects 2010).

Notes

References

American Institute of Certified Public Accountants. (2008). International financial reporting standards (Questions and Answers). Retrieved from http://www.ifrs.com/updates/aicpa/ifrs_faq.html#q6

Bae

Tan

Welker

(2008). International GAAP differences: The impact on foreign analysts. The Accounting Review, 83, 593-628.

Ball

Shivakumar

(2005). Earnings quality in UK private firms: Comparative loss recognition timeliness. Journal of Accounting & Economics, 39, 83-128.

Ball

Shivakumar

(2008). Earnings quality at initial public offering. Journal of Accounting & Economics, 45, 324-349.

Barth

Landsman

Lang

(2008). International accounting standards and accounting quality. Journal of Accounting Research, 46(3), 1-32.

Barth

Landsman

Wahlen

(1995). Fair value accounting: Effect on bank’s volatility, regulatory capital and value of contractual cash flows. Journal of Banking & Finance, 19, 577-605.

Bartov

Goldberg

R. S.

Kim

(2005). Comparative value relevance among German, U.S., and international accounting standards: A German stock market perspective. Journal of Accounting, Auditing & Finance, 20(2), 95-119.

Basu

(1997). The conservatism principle and the asymmetric timeliness of earnings. Journal of Accounting & Economics, 24, 3-35.

Beatty

Petroni

(2002). Earnings management to avoid earnings declines across publicly and privately held banks. The Accounting Review, 77, 547-570.

10.

Becker

DeFond

Jiambalvo

Subramanyam

K. R.

(1998). The effect of audit quality on earnings management. Contemporary Accounting Research, 15, 1-24.

11.

Bédard

Chtourou

S. M.

Courteau

(2004). The effect of audit committee expertise, independence, and activity on aggressive earnings management. Auditing: A Journal of Practice & Theory, 23(2), 13-35.

12.

Bennedsen

Nielsen

Wolfenzon

(2005). The family behind the family firm: Evidence from succession in Danish firms (NYU Working Paper No. CLB-06–002). Retrieved from http://ssrn.com/abstract=1291025

13.

Bennedsen

Wolfenzon

(2000). The balance of power in closely held corporations. Journal of Financial Economics, 58, 113-139.

14.

Bernard

Skinner

D. J.

(1996). What motivates managers’ choice of discretionary accruals?Journal of Accounting & Economics, 22, 313-325.

15.

Brown

(2011). International financial reporting standards: What are the benefits? Accounting and Business Research, 41, 269-285.

16.

Burgstahler

Hail

Leuz

(2006). The importance of reporting incentives: Earnings management in European private and public firms. The Accounting Review, 81, 983-1016.

17.

Bushman

Landsman

W. R.

(2010). The pros and cons of regulating corporate reporting: A critical review of the arguments. Accounting and Business Research, 40, 259-273.

18.

Cameran

Campa

(2010). La Scelta Operata dalle Società Italiane con Riferimento ai Principi Contabili da Utilizzare per la Redazione del Bilancio d’Esercizio (IAS/IFRS versus Standard Nazionali) [Accounting standards choices from Italian companies for the preparation of the separated annual report: IAS/IFRS versus Local GAAP]. Rivista dei Dottori Commercialisti, 61, 483-505.

19.

Capkun

Cavezan-Jeny

Jeanjean

Weiss

(2011). Setting the bar: Earnings management during a change in accounting standards (Working Paper). University of Paris. Retrieved from http://ssrn.com/abstract=1870007

20.

Carey

Simnett

(2006). Audit partner tenure and audit quality. The Accounting Review, 81, 653-676.

21.

Chaney

Lewis

(1995). Earnings management and firm valuation under symmetric information. Journal of Corporate Finance, 1, 319-345.

22.

Committee of European Securities Regulators. (2003). Standard No. 1 on financial information: Enforcement of standards on financial information in Europe. Retrieved from http://www.esma.europa.eu/system/files/03_073.pdf

23.

Daske

Hail

Leuz

Verdi

(2008). Mandatory IFRS reporting around the world: Early evidence on the economic consequences. Journal of Accounting Research, 46, 1085-1142.

24.

Dechow

Schrand

(2010). Understanding earnings quality: A review of the proxies, their determinants and their consequences. Journal of Accounting & Economics, 50, 344-401.

25.

DeFond

Park

(2001). The reversal of abnormal accruals and the market valuation of earnings surprises. The Accounting Review, 76, 375-404.

26.

Directive 2012/6/EU of the European Parliament and of the Council of 14 March. (2012). Retrieved from http://ec.europa.eu/internal_market/accounting/legal_framework/annual_accounts/original_text_en.htm

27.

Economist Intelligence Unit. (2005). Italy: Country profile. Available from store.eiu.com

28.

The European Confederation of Directors Associations. (2010). Corporate governance guidance and principles for unlisted companies in Europe. Available from www.ecoda.org

29.

European Financial Reporting Advisory Group. (2011). Considering the effect of accounting standards (Discussion paper). Available from www.efrag.org

30.

Francis

J. R.

(2004). What do we know about audit quality? The British Accounting Review, 36, 345-368.

31.

Francis

J. R.

Khurana

I. K.

Martin

Pereira

(2008). The role of firm-specific incentives and country factors in explaining voluntary IAS adoptions: Evidence from private firms. European Accounting Review, 17, 331-360.

32.

Frankel

Johnson

Nelson

(2002). The relation between auditors’ fees for non-audit services and earnings management. The Accounting Review, 77(Suppl.), 71-105.

33.

Gastwirth

J. L.

Yulia

R. G.

Miao

(2009). The impact of Levene’s test of equality of variance on statistical theory and practice. Statistical Science, 24, 343-360.

34.

Givoly

Hayn

C. K.

Katz

S. P.

(2010). Does public ownership of equity improve earnings quality?Accounting Review, 85, 195-225.

35.

Hung

Subramanyam

K. R.

(2007). Financial statement effects of adopting international accounting standards: The case of Germany. Review of Accounting Studies, 12, 623-657.

36.

Jeanjean

Stolowy

(2008). Do accounting standards matter? An explanatory analysis of earnings management before and after IFRS adoption. Journal of Accounting and Public Policy, 27, 480-494.

37.

Jones

(1991). Earnings management during import relief investigations. Journal of Accounting Research, 29, 193-228.

38.

Kennedy

(2008). A guide to econometrics. Hoboken, NJ: Wiley-Blackwell.

39.

Kim

J. B.

Chung

Firth

(2003). Auditor conservatism, asymmetric monitoring, and earnings management. Contemporary Accounting Research, 20, 323-359.

40.

Kothari

S. P.

Leone

Wasley

(2005). Performance matched discretionary accrual measures. Journal of Accounting & Economics, 39, 163-197.

41.

Kvaal

Nobes

(2012). IFRS policy changes and the continuation of national patterns of IFRS practice. European Accounting Review, 21, 343-371.

42.

Lang

Raedy

Wilson

(2006). Earnings management and cross listing: Are reconciled earnings comparable to US earnings?Journal of Accounting & Economics, 42, 255-283.

43.

Lawrence

Minutti-Meza

Zhang

(2011). Can big 4 versus non-big 4 differences in audit quality proxies be attributed to client characteristics?The Accounting Review, 86, 259-286.

44.

Leuz

(2010). Different approaches to corporate reporting regulation: How jurisdictions differ and why. Accounting and Business Research, 40, 229-256.

45.

Maijoor

(2010). Discussion of “How can we measure the costs and benefits of changes in financial reporting standards?” Accounting and Business Research, 40, 329-330.

46.

Marra

Mazzola

(2014). Is corporate board more effective under IFRS? Or is “Just an illusion.” Journal of Accounting, Auditing & Finance, 29, 31-60.

47.

Matonti

Iuliano

(2012). Voluntary adoption of IFRS by Italian private firms: A study of the determinants. Eurasian Business Review, 2(2), 43-70.

48.

McNichols

(2000). Research design issues in earnings management studies. Journal of Accounting and Public Policy, 19, 313-345.

49.

Moores

Salvato

(2010). Research on accounting in family firms: Past accomplishments and future challenges. Family Business Review, 23(3), 193-215.

50.

Nobes

(2009). On researching into the use of IFRS by private entities in Europe. Accounting in Europe, 7, 212-226.

51.

Paananen

Henghsiu

(2009). The development of accounting quality of IFRS and IFRS over time: The case of Germany. Journal of International Accounting Research, 8(1), 31-55.

52.

Penno

Simon

D. T.

(1986). Accounting choices: Public versus private firms. Journal of Business Finance & Accounting, 86(13), 561-570.

53.

Pope

McLeay

(2011). The European IFRS experiment: Objectives, research challenges and some early evidence. Accounting and Business Research, 41, 233-266.

54.

Prencipe

Bar-Yosef

(2011). Corporate governance and earnings management in family-controlled companies. Journal of Accounting, Auditing & Finance, 26, 199-227.

55.

PricewaterhouseCoopers. (2006). IFRS: Tax implications for the EU financial industry—Are you ready? Available from www.pwc.com

56.

PricewaterhouseCoopers. (2009). IFRS perspective: An executive survey. Retrieved from http://www.barometersurveys.com/store/docs/ifrs-perspective-executive-survey.pdf

57.

PricewaterhouseCoopers. (2011). 2011 US IFRS outlook survey. Retrieved from http://www.pwc.com/en_US/us/issues/ifrs-reporting/assets/dt-12-0052-2011-ifrs-survey.pdf

58.

PricewaterhouseCoopers. (2012). IFRS adoption by country. Retrieved from http://www.pwc.com/us/en/issues/ifrs-reporting/publications/ifrs-status-country.jhtml

59.

Regulation (EC) No. 1606/2002 of the European Parliament and of the Council of 19 July. (2002). Retrieved from http://www.esma.europa.eu/system/files/Reg_1606_02.pdf

60.

Rodríguez-Péréz

Van Hemmen

(2010). Debt, diversification and earnings management. Journal of Accounting and Public Policy, 29, 138-159.

61.

Rosenbaum

Rubin

(1983). The central role of the propensity score in observational studies for causal effects. Biometrika, 70, 41-55.

62.

Shan

Taylor

Walter

(2011). Error in estimating unexpected accruals in the presence of large changes in net external financing (Working paper). Retrieved from http://ssrn.com/abstract=1572164

63.

Soderstrom

Sun

(2007). IFRS adoption and accounting quality: A review. European Accounting Review, 16, 675-702.

64.

Tarca

(2004). International convergence of accounting practices: Choosing between IAS and US GAAP. Journal of International Financial Management & Accounting, 15, 60-91.

65.

Tsalavoutas

Evans

(2010). Transition to IFRS in Greece: Financial statements effect and auditor size. Managerial Auditing Journal, 25, 814-842.

66.

Vander Bauwhede

Willekens

(2004). Evidence on (the lack of) audit-quality differentiation in the private client segment of the Belgian audit market. European Accounting Review, 13, 501-522.

67.

Van Tendeloo

Vanstraelen

(2005). Earnings management under German GAAP versus IFRS. European Accounting Review, 14, 155-180.

68.

Van Tendeloo

Vanstraelen

(2008). Earnings management and audit quality in Europe: Evidence from the private client segment market. European Accounting Review, 17, 447-469.

69.

Villalonga

Amit

(2006). How do family ownership, management and control affect firm value?Journal of Financial Economic, 80, 385-417.

70.

Warfield

Wild

(1995). Managerial ownership, accounting choices and informativeness of earnings. Journal of Accounting & Economics, 20, 61-91.

71.

Watts

Zimmerman

(1986). Positive accounting theory. Englewood Cliffs, NJ: Prentice Hall.

72.

Wysocki

(2004). Discussion of ultimate ownership, income management, and legal and extra-legal institutions. Journal of Accounting Research, 42, 463-474.