The effect of labour market unionization on firms’ disclosure behaviour

(1)

THE EFFECT OF LABOUR MARKET UNIONIZATION

ON FIRMS’ DISCLOSURE BEHAVIOUR

CECILIA NEDERLOF

UNIVERSITY OF AMSTERDAM

FACULTY OF ECONOMICS AND BUSINESS

MASTER THESIS ACCOUNTANCY & CONTROL

Westermarkt 21-1 1012 ES Amsterdam

The Netherlands cecilia_nederlof@hotmail.com

Student Number: 11108940 Thesis supervisor: Dr. Réka Felleg

20-06-2016

(2)

Statement of Originality

This document is written by student Cecilia Nederlof who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract – Prior research suggests that firms have incentives to keep a certain level of information asymmetry with the market when negotiating with labour unions. The aim of my paper is to examine the effect of labour market unionization on U.S. firms’ disclosure behaviour for the period 2002-2011. Specifically, I examine the association between the level of labour unionization and managerial earnings forecasts. Drawing from a sample of 820 unionized firms with 4,789 firm-year observations, I predict that firms facing strong labour unionization are less likely to disclose information on their earnings. In other words, these firms are less likely to issue earnings forecasts on the presumption that greater company knowledge will reduce corporate bargaining power.. This is coherent with the inference that firms prefer to maintain a certain level of information asymmetry during collective wage negotiations. However, I do not find statistical support for my first hypothesis. Moreover, I predict that firms that do issue earnings forecasts are more likely to down-ward bias these forecasts, when confronted with strong labour unionization. As for my first hypothesis once again I do not find support in my statistical analyses for my second prediction. These findings therefore indicate that firms do not change their voluntary disclosure behaviour when faced with strong labour. Thus, my findings largely contradict with prior research postulating the existence of an association between labour market unionization and a firms’ voluntary disclosure behaviour. My findings should be of interest to labour unions and other market participants to better understand and interpret the earnings forecasts issued by managers.

Key words – voluntary disclosure, management earnings forecast, unionization, bargaining power, forecast bias

(4)

1 Introduction

Prior literature shows that managers’ accounting choices are systematically affected by firm contracts based explicitly on accounting numbers and circumstances which are implicitly tied to a firm’s accounting numbers (Bova, 2013; Liberty & Zimmerman, 1986). Corporate bond covenants and management compensation are examples of such contracts, which are often explicitly based on accounting numbers. These contracts and the specific terms specified in them influence managers’ accounting choices (Holthausen & Leftwich, 1983).

While there has been considerable research done on examining the explicit link of contracts and accounting numbers, there has been far fewer that examine the implicit link (Bova, Dou & Hope, 2011; Bova, 2013). An example of such an implicit contract is the collective bargaining agreement that is negotiated between unionized employees and their employers. More specifically, research shows that negotiated wages in a unionized setting are established based on a firm’s prior profitability (Christofides & Oswald, 1992). Given this implicit contract, researchers have suggested that managers may be motivated to project a negative outlook to their unionized workers in order to improve a firm’s bargaining position when negotiating wages (Bova, 2013). It is for this reason I examine how labour market unionization affects a firm’s direct disclosure behaviour. Specifically, I analyse the influence of labour market unionization on a specific type of firm’s voluntary disclosure, namely management earnings forecast.

Disclosure of financial information is an essential ingredient for the functioning of an efficient capital market (Darrough, 1993). Firms have the ability to disclose financial information through a variety of means including: regulated financial reports, management discussion and analysis, footnotes, and other regulatory filings. In addition, some firms engage in voluntary communication, such as press releases, management forecasts, and analysts’ presentations (Healy & Palepu, 2001). Firms that decide to voluntarily disclose information mitigate the adverse selection problem that arises in capital markets by reducing the information asymmetry between managers and investors, enabling greater liquidity and lowering the firm’s cost of capital (Francis, Nanda & Olsson, 2008).

According to Healy and Palepu (2001), one of the most common ways in which US managers voluntarily disclose private information is through the use of managerial earnings forecasts. Management earnings forecasts can provide firms with several benefits by representing one of the key voluntary firm disclosure mechanisms through which firms can establish or modify market earnings expectations and influence its reputation for accurate and transparent reporting

(6)

(Hirst, Koonce & Venkataraman, 2008). In addition, forecast releases can translate into higher firm values by means of giving investors the ability to better and comprehensively assess a manager’s ability to foresee economic environmental changes and to alter operations accordingly (Trueman, 1986).

Financial disclosure is often seen as an important communication platform for firm/investor relations. The ability for firms to increase their communication reach and engagement with investors through financial disclosure is seen as a significant benefit. Policymakers frequently ask U.S. organizations to increase their disclosure of forward-looking information (Breeden, 1995). However, Darrough (1993) argues that the reporting and disclosure of corporate decisions and performance can also affect a firm negatively. This can be the case when market participants make strategic use of the information that is disclosed by another firm to their advantage. This negative effect is often referred to as a ‘’proprietary cost’’, and leads to the occurrence of firms facing a trade-off between the potentially positive or negative effects of financial disclosure (Verrecchia, 1983). Hence, Cheng and Warfield (2005) postulate that proprietary costs can lead to incentives for managers to intentionally misstate actual earnings in order to meet or beat analyst forecasts.

Commonly, labour unions have no access to their employer’s financial, operating, and personnel information. However, Leap (1991) posits that labour unions need private firm information to function effectively. Nonetheless, the discretion on what information to disclose or to hide from external stakeholders remains with the managers. It is therefore important for outside stakeholders like labour unions, to create in-depth understanding of influential factors of managerial voluntarily disclosures, more specifically managerial earnings forecasts, and examine the effect on the forecasts’ accuracy.

My study is motivated by the importance placed on managerial earnings forecasts in the current financial market and existing incentives for managers to preserve information asymmetry with outside stakeholders (Hilary, 2006). A reoccurring theme discovered in prior literature is that revealing private information weakens management’s position in collective bargaining negotiations (Bova, 2013; Hilary, 2006). To the best of my knowledge, no prior study has examined the implicit link between labour unionization and direct voluntary corporate disclosure, such as management earnings forecasts.

(7)

I developed two hypotheses to test the relationship between labour market unionization and firms’ disclosure behaviour. Firstly, I examine the effect of market unionization on manager’s issuance of earnings forecast. I posit that firms facing strong labour are less likely to disclose information and therefore are less likely to issue earnings forecasts. Secondly, I perform an analysis to examine the accuracy of the earnings forecasts. I predict that firms that issue earnings forecasts are more likely to down-ward bias their forecasts, when confronted with strong labour. The data is gathered from the U.S Labour Statistics Bureau matched with data from the Compustat database. The sample period for my analyses is from 2002 to 2011. A sample of 820 unionized firms with 4,789 firm-year observations was used to test the hypotheses.

In contradiction to the aforementioned predictions, I find no empirical support for either of my hypotheses. These findings indicate that firms do not change their voluntary disclosure behaviour when faced with strong labour. Moreover, my findings largely contradict prior research arguing for an association between labour market unionization and firms’ voluntary disclosure behaviour. These results can be interpreted in several ways. Firstly, labour unions have multiple sources of firm information, which may be substitutes of each other (Tasker, 1998). I assume that the existence of these various sources of information lowers managerial incentives to withhold form issuing their earnings forecasts. Secondly, managers are less likely to bias their forecasts when it is probable that unions’ representatives are financially experienced and capable to accurately determine the firm’s financial position independently of the forecasts issued. A third probably interpretation may be the fact that analysts commonly de-bias the managerial earnings forecasts. Jennings (1987) suggests that the extent of the analyst’s forecasts revisions acts as a proxy for the credibility of a managerial earnings forecast. This would lower the incentive for managers to bias their earnings forecasts, since labour unions can rely on the forecasts produced by analysts. My study makes several contributions to prior literature. Firstly, the analyses performed in my paper are of societal value to unionized employees and other market participants. The findings of my study are useful to these outside stakeholders by creating a better understanding and interpretation of the earnings forecasts issued by managers. These stakeholders may want to give greater consideration to the degree to which managerial incentives influence the accounting choices in projecting the firm’s financial position. Secondly, I believe my paper is of value for current literature, because it offers contradictory findings to previous research and clarifies the level of current and future firm disclosure in light of specific labour market unionization. I provide a clear framework of the choices managers have once they choose to issue an earnings

(8)

forecast, which serves as an important direction for both future theory development and empirical research. Lastly, my paper informs concerned regulators about the quality, effectiveness, and frequency of management earnings forecast. My findings indicate that the extent and the accuracy of management earnings forecasts is not influenced by a greater level of unionization.

The remainder of my paper proceeds as follows. Chapter 2 includes a literature review on the foundation of my study and building upon this review hypotheses are formulated. In addition, the conceptual model is provided at the end of this chapter. Chapter 3 details a description of the sample, data collection, and the empirical research design of this paper. The results of the analyses are presented in chapter 4. To conclude, the findings and implications are provided in chapter 5.

(9)

2 Theoretical Framework

2.1 Firm’s Disclosure Behaviour

For many years, accounting and empirical researchers have shown interest in firms’ decision to disclose information to outsiders (Beuselinck, Deloof & Manigart, 2008). Darrough (1993) postulates that firm disclosure is a critical ingredient for the functioning of an efficient capital market. Firms release annual reports to numerous stakeholders to provide relevant and timely information that is useful for investment decisions, writing contracts, monitoring and rewarding performance. There are various stakeholders (creditors, shareholders, employees, government agencies, suppliers, etc.) that demand the right to access corporate financial information; which in turn, leads to the possibility that the disclosure of firm specific information has differential effects on these stakeholders (Healy & Palepu, 2001). To provide an example, detailed and direct disclosure about new products can convey information about a firm’s future prospects and strategy to its shareholders. On the other hand, it might also disclose strategic information to a firm’s market competitors, thereby negatively affecting the disclosing firm’s competitive advantage. Therefore, the disclosure of information may lead to both positive and negative effects on the prosperity of the firm’s shareholders. The negative effect that can arise from a firm’s disclosure is often referred to as a ‘’proprietary cost’’ (Verrecchia, 1983). Thus, a firm’s disclosure policy is likely to be driven by a variety of proprietary costs, such as agency costs, litigation costs and disclosure related costs (Hossain, Perera & Rahman, 1995). In the presence of proprietary costs, a firm has to trade-off both the positive and negative effects of disclosure. Agency conflicts and information asymmetry can be used to explain why variation in disclosure exists among firms. Agency theory suggests that in situations with a separation of ownership and control potential agency costs arise, because of the conflict of interests between the agents and principal. In order to prevent these costs, managers have the ability to decide to disclose private information. In fact, Dye (1985) argues that management will disclose as much information as possible as long as the advantages (e.g. limiting litigation risks, lowering cost of capital) outweigh the risks and costs (e.g. actionable in court). The assumption made here is that when price-maximizing managers withhold information, investors will be uncertain about the quality of the investment that are so great, that these investors mark down its quality to the point where managers are better off eliminating the information asymmetry.

[Type a quote from the

document or the summary of an interesting point. You can position the text box anywhere in the document. Use the Drawing Tools tab to change the formatting of the pull quote text box.]

(10)

More specifically, Milgrom and Roberts (1986), predict that firms who wish to maximize their share price often disclose all their private information on the condition of

‘’ (1) the disclosure is costless to the firm; (2) investors know that the firm has, in fact, private information; (3) all investors interpret the firm’s disclosure; (4) the firm can credibly disclose its private information; and (5) the firm cannot commit ex-ante to a certain disclosure policy ’’

These conditions lead to incentives for firms to disclose information in order to distinguish themselves from firms who do not disclose information and are thereby perceived as having undesirable information. This is the case for mostly all information disclosed, which leads to the ‘’unravelling’’ of a firm’s withheld information (Berger, 2011). However, an important aspect to consider is the fact that ‘’unravelling’’ does not necessarily suggest that all firms’ private information is disclosed in its entirety. Rather, this phenomenon explains why and when minimal disclosure is likely to occur; with the obvious reason being that one or more of the five conditions did not hold.

As a consequence of these conflicting incentives for disclosure, it remains unclear whether firms will voluntarily disclose all their relevant information. Meek, Roberts and Gray (1995) define voluntary disclosures as information revealed in excess of that which is required by law and regulations. With the arrival of mandating disclosures through regulatory agencies such as the FASB or the SEC, firms are now limited in their ability to withhold information that they would rather remain concealed. The mandates established by these bodies have a real effect on the workings of the market if it is actually able to force firms to disclose more than they would have, with potential effects on different stakeholders. In other cases, the mandating might not have any incremental effect, because the information disclosed would have been disclosed voluntarily anyway (Darrough, 1993). Therefore, Robinson, Xue and Yu (2011) argue that mandatory and voluntary disclosures are interdependent and that new mandatory disclosure regulations cannot be considered without taking the effect on voluntary disclosure into account. It is therefore important for regulatory bodies to sort out the disclosure motives of firms in order to develop more efficient disclosure policies that are in line with the goals of the bodies.

Many firms tend to disclose information voluntarily when they plan to raise capital and therefore want to influence investors’ perceptions by providing explicit information. (Healy & Palepu, 1995). Economic theory suggests that voluntary disclosures will almost always be provided with

(11)

the best intentions as managers may otherwise lose their credibility and could face possible shareholder litigation (Stocken, 2000). In addition, Holder-Webb and Cohen (2007) found that companies disclose voluntarily to avoid some distress costs, such as capital, reputation and contracting costs. Moreover, it has the possibility to decrease the company’s bankruptcy risk, default risk and uncertainty about future cash flows. In sum, economic theory suggests that firms benefit from voluntary disclosure to the capital markets (Bova et al., 2013).

Management (Earnings) Forecasts

One key voluntary disclosure mechanism by which firms establish or modify market earnings expectations and influence their reputation for transparent and accurate reporting are management earnings forecasts (MEF). Management earnings forecasts are unscheduled voluntary firm disclosures predicting earnings prior to the expected reporting date for a particular firm (Hirst, Koonce & Venkataraman, 2008). From the 1990s, management earnings forecasts have become prevalent in US capital markets (Anilowski, Feng & Skinner, 2007; Beyer, Cohen, Lys & Walther, 2010), and have been found to affect analysts’ forecasts (Baginski & Hassell, 1990), stock prices (Pownall, Wasly & Waymire, 1993), and bid-ask spreads (Coller & Yohn, 1997). Moreover, managerial incentives, firm characteristics, and biases are likely to have significant influence on the management’s decision whether or not to issue a forecast, as well as the attributes of the forecasts (Ajinkya, Bhorjraj & Sengupta, 2005; Athanasakou, Strong & Walker, 2011). Ball and Shivkumar (2008) argue that management forecasts hold higher value relevance over earnings announcements due to their diverse attributes that include more timely and forward-looking information.

Hirst, Koonce and Venkataraman (2008) argue that the incentives managers have for issuing earnings forecasts largely correspond with those of shareholders. Namely, the provision of and the demand for management forecasts is assumed to be mostly driven by stock price considerations to reduce existing information asymmetry between managers and analysts and potential investors. This suggests that earnings forecasts are highly influential in the financial market. Moreover, firm’s earnings forecasts affect the earnings forecasts issued by analysts (Baginski and Hassell, 1990). As a response to the managerial forecasts, analysts update their own forecasts, which suggests the significant effect of management earnings forecasts on intermediaries in the market (Waymire, 1986; Jennings, 1987; Cotter, 2006).

(12)

2.2 Labour Market Unionization

Within the nexus of a firm, employees are an important stakeholder group that plays a key role in the long-term success and overall performance of a corporation. Within this stakeholder group, labour unions, often highly influence employee social rights and contracts (Agrawal & Matsa, 2013). Labour unions can be defined as legally recognized representatives of employees in many different industries. Their main role is focused on collective bargaining over wages, working conditions, benefits for their members, and on representing their members in disagreements with a firm’s management over violations of contract provisions (Bova, Dou & Hope, 2011). Commonly, labour unions have little or no access to a firm’s private financial, operating, and personnel information, which can lead to unions resorting to legal action in order to obtain this valuable information (Leap, 1991). Frost (2000) notes that accessing information is one of the key factors for labour unions in being able to secure favourable positions and results.

Previous literature has found that managers’ accounting choices are systematically affected by a firm’s contracts based explicitly on accounting numbers and situations which implicitly use accounting numbers (Liberty & Zimmerman, 1986). Unlike most debt agreements and executive compensation contracts, few labour contracts are explicitly based on accounting numbers. However, Reynolds (1978, p. 390) suggests that a firm’s earnings are often important pillars used in labour contract negotiations: ‘’If profits are high and the business outlook is good a labour union can afford to make large demands. On the other hand, when a business is declining and profits are failing, the union might have to be content with holding the present wage level’’. In his line of reasoning, Reynolds (1978) assumes that labour unions desire information regarding a firm’s economic rents, which is provided by accounting earnings.

Managers have to consider economically significant stakeholders when present, and balance their actions to consider the potentially conflicting objectives (e.g. shareholders, employees). The juxtaposition from conflicting objectives is argued to affect the financial communication policy of a firm (Hilary, 2006). As Bowen et al. (1995) and D’Souza et al. (2001) show, managers facing strong labour have incentives to bias reported earnings to be less favourable. Hence, it is argued that the benefits to managers in depressing the firm’s earnings during labour contract negotiations outweigh the costs, thereby influencing both the political process and labour talks (Liberty & Zimmerman, 1986). The expected benefits to managers of manipulating earnings are the gain in greater bargaining power over employees and their lower wages. That is, wages would be higher if earnings are not reduced during contract bargaining, because labour unions would

(13)

have greater bargaining power. This inference is also confirmed by Hilary (2006), who posits and finds that firms facing labour market unionization tend to select income-decreasing accounting choices.

2.3 Hypothesis Development

2.3.1 Labour Market Unionization & Management Earnings Forecast

Kleiner and Bouillon (1988) note that information sharing is associated with elevated bargaining power among American labour unions. The authors argue that more information enables labour unions to negotiate more effectively and obtain greater results. Their findings portray that information disclosure about the entity’s financial condition, relative wages, productivity, and future investments is significantly related to considerably higher wages and remunerations for production employees. Moreover, Scott (1994) shows how Canadian firms operating in an industry with high average salaries or facing a higher likelihood of a strike curtail the amount of information provided on pension-related issues. In other words, these firms withheld from disclosing private information in order to keep a certain level of information asymmetry with their employees.

Unionized employees are said to have greater ability to extract above-market rents from their employers compared to their nonunionized peers (Hirsch, 2008). This is due to their greater collective bargaining leverage in negotiating contracts with their employers and the ability to strike if demands are not met. This leads to firms developing incentives to minimize the compensation their employees are able to extract above-competitive-market rents. Additionally, Reynolds, Masters, and Moser (1998) argue that an important feature of labour negotiations is the effort firms make to conceal or even misrepresents its true financial position.

I presume that in a setting of unionization, managers have to consider the costs (i.e., increase in a firm’s cost of capital) and benefits (i.e., decrease of union negotiation leverage, which may lead to better negotiation positions for the firm) of information asymmetry. Considering this presumption, I argue that employees can benefit from voluntary management forecasts by increasing favourable bargaining positions during negotiations with employers. This is due to the fact that earnings forecasts provide the unions with information about the future financial position of the firm and consequently decrease the level of information asymmetry (Lennox & Park, 2006). Hence, managers will have a tendency to refrain from disclosing earnings forecasts

(14)

to maintain a certain level of informational advantage. Therefore, I posit that firms facing strong labour unionization are less likely to forecast earnings.

Hypothesis 1: Labour market unionization is negatively associated with the probability of firms disclosing managerial earnings forecasts.

2.3.2 Labour Market Unionization and Management Earnings Forecast Bias

For many years there has been an interest in the accuracy of published financial reports and forecasts (Conference Board, 2003). This interest in forecast accuracy can be substantiated by several capital market issues. Keeping in mind that earnings forecasts can be effective in conveying news about firm value (Waymire, 1984; Lev & Penman, 1990), the market should be concerned with the accuracy of these published forecasts and understand to what extent these predict the future (Hartnett & Romcke, 2000). Additionally, the SEC has been concerned with the credibility of information disclosed in earnings forecasts, since it lifted its prohibition against forward-looking information such as management earnings forecasts in 1973 (Levitt, 1998, pg. 79).

Commonly, managers strive to provide accurate earnings forecasts since the accuracy of these forecasts is related to their own credibility (Hirst et al., 2008). Furthermore, Rogers and Stocken (2005) state that management earnings forecast are more accurate when investors have greater ability to detect forecast errors. The authors use earnings volatility as a proxy for investor ability to detect forecast errors and argue for four elements that affect the intentional forecasting behaviour of managers. Firstly, when the risk of litigation for biasing earnings is larger, managers tend to issue more pessimistic forecasts. Secondly, when firms are in financial distress, managers are more likely to forecast their earnings optimistically to secure their position in the market and avoid negative market valuations. Thirdly, inaccurate forecasts may be provided by managers if these lead to favourable results from insider transaction based on mispriced shares prompted by these forecasts. Lastly, managers employed by firms operating in concentrated industries might feel tempted to release pessimistic forecasts in order to discourage potential entrants from entering the market (Rogers & Stocken, 2005).

In this study I argue for another managerial incentive to issue pessimistic earnings forecasts. Liberty and Zimmerman (1986) suggest that as firm profit increases, the labour unions expect

(15)

larger wage concessions. In addition, labour unions are able to use their power to strike as a means of bargaining to increase the likelihood of realizing these concessions. In order to maintain a strong bargaining position in collective wage negotiations and deter larger wage bills, I presume managers will manipulate earnings forecasts down-ward. Thus, I predict that strong labour unionization is negatively associated with the accuracy of management earnings forecasts.

Hypothesis 2: Labour market unionization is positively associated with the probability of managers’ down-ward basing their earnings forecasts.

Figure 1depicts my paper’s conceptual model and predicts the associations between the construct variables.

(16)

3 Sample and Empirical Design

3.1 Sample

In order to test my hypotheses, I require a sample of unionized firms. To develop this sample, I retrieved data from the U.S. Labor Statistic Bureau databases, which I matched to data from the Compustat North America Fundamentals database to obtain a sample of U.S. unionized firms. The sample period runs from 2002 to 2011 and is constrained to these years, for the reason that within this period the labour union statistics were measured with the same 2000 census industry classification. The initial sample consists of 2,744 unionized firms, with 17,943 firm-year observations.

In order to obtain data on annual U.S. management earnings forecasts of earnings per common share (EPS) and analyst following, I included data from the First Call Company Issued Guidance (FCCIG) database. From this database I used the forecasts where a point forecast can be estimated, because Rogers and Stocken (2005) argue that the market has a better ability to compare these forecasts to realized earnings unambiguously. When the forecast is a range forecast with no other qualification provided by management I use the midpoint of the range, which could be the average of the two endpoints. Moreover, if there are indications provided by management concerning which end of the range would be more likely to be suitable, I use the specific end of the range as in line with prior literature (Hutton & Stocken, 2007). Additionally, open-ended forecasts are included in the sample, where the endpoint is used. In the case of open-ended forecasts structured in comparison to 0, I chose to use 0, which is in line with research by Baginski, Conrad and Hassel (1993) and Hutton and Stocken (2007). In consonance with the reasoning of Hirst et al., (2008) the first management forecast for each fiscal year-end is used, because later forecasts are shown to be more optimistic. Lastly, I exclude qualitative forecasts from my observations.

Moreover, I drop forecasts issued on or after the fiscal year-end as is done by Rogers and Stocken (2005) and Hutton and Stocken (2007) to exclude earnings pre-announcements, earnings warnings, and forecasts that might have erroneous forecast dates. I also drop observations with missing forecasts and no identifiable numerical forecasts. Financial sector observations (SIC code between 6100 and 6299) are dropped (800 observations), since firms operating in sector have different incentives. For instance, these firms (e.g. banks) are covered by

[Type a quote from the

document or the summary of an interesting point. You can position the text box anywhere in the document. Use the Drawing Tools tab to change the formatting of the pull quote text box.]

(17)

different regulatory requirements, which impacts their accounting choices (Ramesh & Revsine, 2000).

Furthermore, I match my sample with First Call actual database for the earnings announcement dates and the annual earnings numbers. I eliminate observations where more than one earnings announcement is made on the same date. In order to identify any outliers I winsorized at the 1st and 99th_{percentiles. The final sample consists of 820 unionized firms, with 4,789 firm-year} observations, of which 376 unionized firms disclosed an earnings forecast, with 1,554 firm-year observations. The sample’s data collection is summarized in Table 1.

[Insert Table 1 here]

3.2 Empirical Design

3.2.1 Design for Hypothesis 1

I examine how labour market unionization affects a firm’s voluntary disclosure behaviour. More specifically, how strong organized labour influences the decision of managers to issue earnings forecasts and the quality of these forecasts. My first hypothesis predicts that labour market unionization is negatively associated with the issuance of earnings forecasts by managers. I rely on prior research for measurement instruments in order to construct the empirical model (Choi & Ziebart, 2004; Ajinkya, Bhojraj & Sengupta, 2005; Hilary, 2006; Bova, Dou & Hope, 2011; Bova, 2013). My prediction is examined with the following empirical model:

MEF = a + β0 + β1 LMU + Control variables+ e (1)

The variables in this model are defined as follows:

MEF, my outcome variable, is an indicator variable that equals one if the firm has issued an annual management earnings forecast, and zero otherwise. In other words, I examine the probability of occurrence of management forecasts (Ajinkya, Bhojraj & Sengupta, 2005). I include management forecasts of annual earnings, because these statements are often provided enough in advance of the earnings announcement for it to be likely that these forecasts are associated with significant proprietary costs (Ali, Klasa & Yeung, 2014). In order to test my first hypothesis,

(18)

firms that issue multiple forecasts and ones that issue just one forecast in the period are treated the same.

LMU, I measure labour market unionization with an interaction of labour intensity (LINT) with a firm’s unionization rate (R). More specifically, I multiply the unionization rate per industry with the labour intensity specified for each firm (Hilary, 2006). Data for the unionization rate is obtained from the U.S. Labour Statistics Bureau. The unionization acts as a proxy for the degree of bargaining that affects wages. This is because of the fact that if the labour market would be fully competitive, wages would be determined exogenously. This rate is calculated at the two-digit NAICS code. Prior research offers an argument for why firm-specific proxies in this context are likely to underestimate the effect of unions on the information environment. Rosen (1969) argues that spill-overs are externalities initiated by the threat of market unionization. Thus, the pressure of labour unions is not restricted to their own firms, but this pressure creates a credible threat to other firms operating in the same industry. Rosen’s research shows that the direct effect is empirically dominated by the degree of the spill-over effect. This line of reasoning is confirmed by Bronars and Deere (1994), who indicate that ‘’the total negative effects of unionization on profits, after cross-firm or spill-over effects are included, are nearly three times as large as the own-firm effects’’. Concerning this statement, the authors specifically limit the firms affected by the spill-over effect as the ones operating in the same industry.

In order to measure the extent of labour intensity I divide the firm’s number of employees by the firm’s total assets (Hilary, 2006). This means that I use the labour to capital ratio to determine if unionized employees are able to significantly influence managerial decisions. Even though all firms operating in the same industry are affected by a comparable pressure from labour unions, the effect of the industry-wide unionization pressure will be firm specific through this measurement.

Control variables: Following previous research, I identified several variables which I include in my empirical model. Firstly, I include firm characteristics, which possibly influence forecasting difficulty. Prior research suggests that it is more of a challenge for management to forecast earnings if the firm is performing badly and has to report a loss (Hayn, 1995; Basu, 1997). In addition, earnings are less value relevant for firms that have to report losses (Hayn, 1995) and meeting analysts’ forecasts is less important for these loss-making firms (Degeorge, Patel & Zeckhauser, 1999). It is for these reasons I include the indicator variable loss (LOSS). Loss

(19)

equals one if the firm reported losses in the previous year, and equals zero otherwise (Choi & Ziebart, 2004).

I also control for firm size because the information environment of large firms is likely to be richer. This should have a positive effect on the accuracy of management’s forecasts (Botosan & Plumlee, 2002). Additionally, large firms tend to disclose more information, which might be motivated by a firm’s reputational needs (Hutton & Stocken, 2007; King, 1996). On the other hand, large firms tend to be more complex, which might have an effect on the accuracy of the management forecasts (Baginski & Hassell, 1997; Bamber & Cheon, 1998; Hutton, Lee & Shu, 2012). Moreover, Hribar and Yang (2011) indicate that smaller firms are more likely to issue biased forecasts. I therefore proxy for firm size (SIZE), by including the natural logarithm of a firm’s total assets (Bova, Dou & Hope, 2011).

Lastly, I control for the number of analysts following a firm, denoted as NUMEST. Hutton, Lee and Shu (2012) argue that when more analyst are covering a firm, the firm is subject to greater scrutiny, which creates stronger incentives for managers to maintain high quality corporate disclosure and a reputation of credibility. On the other hand, Hilary (2006) argues that firms facing strong unionization tend to have a lower amount of analyst coverage. This is explained by the reasoning that firms that try to remain a level of information asymmetry by disclosing less information are less likely to be covered by analysts. I therefore include analyst coverage, measured by the natural logarithm of the number of analysts following the firm on the same day as the management earnings forecast (Baginski & Hassel, 1997; Choi & Ziebart, 2004; Hribar & Yang, 2011).

I test my first hypothesis by performing a logit regression analysis. I also make sure my variables are normally distributed and that there is no multicollinearity. In order to confirm the hypothesis, I expect β1 to be negative, because I predict that there is a negative association between labour market unionization and the probability of managers issuing earnings forecasts.

3.2.2 Design for Hypothesis 2

I also examine the association between labour market unionization and the accuracy of earnings forecasts issued by managers. More specifically, I argue that strong organized labour enlarges the probability of managers issuing down-ward biased earnings forecasts. Relying on prior literature, I construct my second empirical model (Ajinkya, Bhojraj & Sengupta, 2005; Hilary, 2006; Bova,

(20)

Dou & Hope, 2011; Bova, 2013). My second hypothesis is tested with the following empirical model:

BIAS = a + β0 + β1 LMU + Control variables+ e (2)

BIAS, To measure the quality of a firm’s voluntary disclosure I use BIAS as a proxy. This instrument is measured by the difference between the management forecast of earnings per share (EPS) and the actual EPS, deflated by the stock price at the beginning of the fiscal period (Ajinkya, Bhojraj & Sengupta, 2005). Since I am interested in down-ward biased forecasts, I consider BIAS as an indicator variable that equals one if the firm issues a down-ward biased forecast, and zero otherwise.

LMU, as for my first empirical model, I measure labour market unionization by multiplying the unionization rate per industry with the labour intensity specified for each firm (Hilary, 2006). Control variables, Under the same reasoning as for my first hypothesis BIAS is controlled for firm operating loss, firm size and number of analysts following a firm. Following, Hribar and Yang (2011), I include return on assets (ROA), which I measure as the firm’s net income at the beginning of the period divided by its total assets at the beginning of the period. Hribar and Yang (2011) show that firms which perform badly have the tendency to issue biased earnings forecasts. I also test my second hypothesis by performing a logit regression analysis. I again examine whether my variables are normally distributed and that there is no multicollinearity. In order to confirm the second hypothesis, I expect β1 to be positive, because I predict that there is positive association between labour market unionization and the probability of managers issuing earnings forecasts.

(21)

4. Results

In this chapter I will present the results of my statistical analyses used to test my hypothesis. I firstly provide my sample’s descriptive statistics and demonstrate the correlations between the construct variables. Lastly, I provide the outcomes of the regression analyses in order to determine the support for the hypotheses.

4.1 Descriptive Statistics

Table 2 provides the descriptive findings for the sample used to test my hypotheses. As Panel A of Table 2 shows, my sample is fairly evenly spread across the time period, 2002-2011 with the exception for 2011. Panel B of Table 2 outlines the distribution of various industries which make up the sample selection. As we can see from Table 2, the Electric, Gas, & Sanitary Services (22.57%) and Communications (18.08%) industries are relatively stronger represented in comparison to the Mining, Quarrying, and Oil and Gas Extraction (0.67%) and Legal Services (0.19%) industries.

Table 3 shows the descriptive statistics for my dependent, independent, and control variables. Some of these results I would like to highlight in this section. Firstly, the average for MEF indicates that more than half of the unionized sample did not issue a management earnings forecast. Secondly, the average for LMU indicates that the unions’ bargaining power is relatively strong when compared to the average labour strength (0.011) which was found in Hilary’s (2006) paper. Thirdly, the average firm size of my sample is consistent with prior literature (Hilary, 2006; Bova, Dou & Hope, 2011). Lastly, I can conclude that more than half of my sample was not faced with losses at prior fiscal year-end.

Table 4 displays the amount of down-ward biased forecasts that were issued out of the total (1,554) earnings forecasts used in my sample. I can conclude that less than half of the issued forecasts were down-ward biased.

(22)

Table 5 presents the results for the Pearson correlations of my dependent, independent, and control variables. Firstly, I note that issuance of management earnings forecasts (MEF), is significantly correlated (at the one percent level) with all other construct variables included in my first model. As expected, management earnings forecast is negatively correlated with labour market unionization and, firm operating loss and positively correlated with firm size and analyst following.

Additionally, I note that down-ward biasing of forecasts (BIAS) is significantly correlated (at the one and five percent level) to all the other construct variables in my second model except the control variable ROA. Surprisingly, forecast bias is negatively correlated to labour market unionization and firm operating loss, and positively correlated with firm size and number of analysts following a firm. The found relation between BIAS and LMU contradicts my second hypothesis, because I predict that a higher level of labour market unionization is associated with a higher probability of down-ward biasing earnings forecasts.

Considering the correlations depicted in Table 5, initial evidence is provided that labour market unionization does influence a firm’s disclosure behaviour. However, I recognize the fact that the correlations only provide univariate evidence and I therefore should rely on logistic regressions before drawing conclusions.

4.2 Results of Hypothesis Tests

4.2.1 Hypothesis 1

The main results for my first logistic regression are summarized in Table 6. Based on the regression analysis, I find that labour market unionization (LMU) has no significant influence (β = 1.392, p = 0,67) on the probability of managers issuing earnings forecasts (MEF). The insignificant and positive association is not consistent with my first hypothesis, where I predict that firms facing strong labour unionizations are less likely to issue an earnings forecast. The finding is also in contradiction to prior research (Liberty & Zimmerman, 1986; Frost, 2000; Hilary, 2006). Hilary (2006) finds that firms facing organized labour have incentives to preserve a

(23)

certain level of information asymmetry with outside stakeholders. The author uses direct measures of information asymmetry by examining bid-ask spreads, probability of informed trading and analyst coverage. As disclosing earnings forecasts in collective wage negotiations lowers the information asymmetry between firms and labour unions, I expected a significant and negative result.

A probable reason for this result could be the fact that I used a raw disclosure measure (MEF) in order to measure firm’s voluntary disclosure behaviour. Labour unions and the firm’s financial markets have access to numerous alternative sources of firm information. A potential empirical challenge lies in the finding that raw disclosures channels may act as substitutes for each other (Tasker, 1998). Therefore, measuring firm’s voluntary disclosure behaviour with one of these measurement constructs may lead to spurious results on the level of firm disclosure.

Moreover, Table 6 presents that no control variables have significant influences on the issuance of earnings forecasts, when labour market unionization is included. Considering the aforementioned results, I conclude that I do not find support for my first hypothesis.

4.2.2 Hypothesis 2

Table 7 summarizes the main results for my second logistic regression. I find that labour market unionization (LMU) is not significantly and positively associated with the probability of biased earnings forecasts (BIAS) (β = -10.136, p = 0.26). This result is in contradiction with prior research, that find that unionized firms are more likely to miss mean consensus analysts’ earnings forecasts (Bova, 2013). A potential explanation for this finding is the fact that unions’ representatives have incentives to undo managers’ manipulations if these were to affect the labour bargaining. According to Reynolds (1987, p. 390), with so much at stake, the union representatives are financially experienced and capable to accurately determine the firm’s financial position independently of the forecasts issued. In other words, if the unions have the ability to undo manipulations, there remains no reason for managers to down-ward bias their earnings forecasts.

Another interpretation of this finding may be the fact that analysts commonly de-bias the managerial earnings forecasts. Jennings (1987) postulates that analyst’s reaction to a management

(24)

forecasts are reliant on the credibility of the earnings forecast. The forecast’s credibility depends on several aspects, including the ability of managers and their incentives to forecast accurately. Hence, managers’ past forecasting records are used by analysts to assess the managerial forecasts error and accuracy relative to their own forecast (Williams, 1996). Jennings (1987) draws the conclusion that the extent of the analyst’s forecasts revisions acts as a proxy for the credibility of a managerial earnings forecast. Additionally, Hassel, Jennings and Lasser (1988) argue that analyst revisions represent the analyst’s ability to identify those circumstances in which their own forecasts are more accurate than the forecasts issued by managers and are not misled by the inaccuracy of these. I argue that this line of reasoning potentially results in little incentive for managers to bias their forecasts, since labour unions can rely on the forecasts produced by analysts.

However, I do find that LOSS is significantly and negatively influencing the probability of biased forecasts (β = 1.528, p = 0.07). In other words, when firms are faced with losses at prior fiscal year-end the probability of biased earnings forecasts increases. Moreover, the table depicts that ROA significantly and positively influences BIAS (β = 5.600, p = 0.10). So, when firm’s return-on-assets increases, the probability of managers issuing biased forecasts increases. The table also presents that no other control variables are significantly associated with the probability of biased earnings forecasts. Considering the overall results of my regression, I can conclude that I do not find support for my second hypothesis.

(25)

5. Conclusion

The purpose of my study is to gain insight into the effect of labour market unionization on a firm’s voluntary disclosure behaviour. My paper is motivated by the significance of management earnings forecasts in the current financial market and existing incentives for managers to preserve information asymmetry with outside stakeholders. I examine the effect of labour strength on the issuing of earnings forecasts by managers. This chapter will start with the findings of my study, which were obtained through STATA. Moreover, I will provide the theoretical and practical implications of this study and end my paper with the limitations and indications for future research.

5.1 Findings

Managerial earnings forecasts represent one of the key voluntary firm disclosure mechanisms through which firms can establish or modify market earnings expectations and influence its reputation for accurate and transparent reporting (Hirst, Koonce & Venkataraman, 2008). Firm characteristics and managerial incentives are likely to have a substantial effect on the management’s decisions to issue a forecast (Ajinkya, Bhorjraj & Sengupta, 2005; Athanasakou, Strong & Walker, 2011). This inference is also supported by Bowen et al. (1995) and D’Souza et al. (2001) who show that when faced with strong labour unionization, managers are incentivised to down-ward bias their earnings. To the best of my knowledge, no prior study has examined the implicit link between labour unionization and voluntary corporate disclosure, such as management earnings forecasts.

The results of my regression analyses suggest that there is no negative association between the level of labour market unionization and the probability of managers disclosing earnings forecasts. This finding is not consistent with my predictions and with prior literature, which suggested that firm’s faced with strong labour prefer to preserve a certain level of information asymmetry with their outside stakeholders (Liberty & Zimmerman, 1986; Hilary, 2006; Bova, 2013). My finding can be interpreted in a way that suggests labour unions may not need earnings forecast issued by managers to retrieve private firm information. Tasker (1998) notes that unions have access to a variety of sources of corporate information, which may act as substitutes of each other. This finding makes it difficult to draw conclusions on a specific (raw) disclosure measure. However, the assumption I can make from my findings, is that managers are not influenced by unions on their issuance of earnings forecasts. I assume that the multiple sources of information lower the

(26)

managerial incentives to refrain from issuing earnings forecasts. Hence, I argue that the benefits of disclosing private earnings information outweigh the negative consequences to transparency in a unionized setting. Moreover, I do not find significant associations between my control variables included in the first empirical model and the issuance of earnings forecasts, when labour market unionization is also included.

My findings for my second empirical model suggest that there is no association between the level of labour market unionization and the probability of managers down-ward biasing their earnings forecast. The results for my second regression analysis are not consistent with my prediction and prior research (Liberty & Zimmerman, 1986; Hilary, 2006) that argue for a significant and positive effect between labour unionization and the likelihood of firms choosing to select income decreasing accounting choices. In addition, the results are not consistent with the findings from Bova’s (2013) paper, suggesting that unionized firms are more likely to miss mean consensus analysts’ earnings. A probable interpretation for this finding is that unions’ representatives are financially experienced and capable to accurately determine the firm’s financial position independently of the forecasts issued. This occurrence would leave no incentives for managers to bias their earnings forecasts. Another interpretation of this finding may be the fact that analysts commonly de-bias the managerial earnings forecasts (Jennings, 1987). Following, I draw the assumption that this lowers managerial incentives to bias their earnings forecasts, since labour unions can rely on the forecasts produced by analysts. Hence, I argue that the costs in biasing the firm’s earnings forecasts during labour talks outweigh the benefits.

My findings do suggest a significant and negative association between firm operating loss and the probability of managers biasing their earnings forecasts. The results also suggest a significant and positive association between firm return-on-assets and the probability of managers issuing down-ward biased earnings forecasts, when unionization is considered. No other control variables were found to have significant association with the down-ward biasing of earnings forecasts.

Overall, my findings suggest that there is no association between the level of labour market unionization and a firm’s disclosing behaviour regarding management earnings forecast. However, my paper is subject to several limitations, which should be considered when interpreting my results. These will be presented after I have stated the theoretical and practical implications of my paper.

(27)

5.2 Theoretical and Practical Implications

I believe my paper adds to the existing body of current literature, because it offers contradictory findings to previous research and clarifies the level of current and future firm disclosure in light of specific labour market unionization. While a great amount of research has examined the explicit link between contracts and accounting numbers, little research has been done to analyse the influence of implicit contracts. Researchers have suggested that these implicit contracts (e.g. collective bargaining agreement) may influence managerial incentives to project a negative outlook to their unionized employees to improve their own bargaining position (Bova, 2013). I provide a clear framework of choices managers have once they choose to issue an earnings forecast; which serves as an important direction for both future theory development and empirical research.

Moreover, my paper is of societal value to unionized employees and other market participants in that my findings will allow for the enhanced understanding and interpretation of earnings forecasts issued by managers. These outside stakeholders would be wise to place greater importance to which managerial incentives play a part in accounting choices used in projecting the firm’s financial position.

Furthermore, my study informs regulators who are concerned about the quality, effectiveness, and frequency of management earnings forecasts. Since 1973, when the SEC lifted its prohibition against forward-looking information such as MEF, the SEC has been concerned with the credibility of such information (Levitt, 1998, pg. 79). As my findings suggest, the extent and accuracy of managerial earnings forecasts is not influenced by a greater level of unionization.

5.4 Limitations and Future research

My paper is subject to limitations that should be considered when interpreting the results and findings. The limitations in my study pertain to the external validity of my results and findings. The fact that my study only considers U.S. unionized firms puts a limitation on the generalizability of my findings. I therefore encourage future research to study the implicit link between labour unionization and firms’ voluntary disclosure in other countries in order to replicate and extend the findings of my paper.

Moreover, I exclude managerial earnings forecasts that were qualitative in nature in my sample selection. Skinner (1994) shows in his paper that upward biased management forecasts are more

(28)

inclined to be point or range projections while down-ward biased forecasts tend to be of qualitative nature. To the extent of a difference between managers’ forecasting behaviour for forecasts of quantitative and qualitative nature my findings could potentially not be generalizable to qualitative earnings forecasts. Future research would do good to perform additional analyses to examine the influence unionization has on qualitative earnings forecasts.

Another limitation of my study is the use of managerial earnings forecasts as my sole metric for firm’s voluntary disclosure. This limitation occurs due to the fact that the accuracy of these forecasts can easily be verified by outside market participants through the actual earnings realizations (Healy & Palepu, 2001). On the other hand, with other types of voluntary disclosures, such as human capital and customer satisfaction, it is more difficult to verify the accuracy after issuance. So, even though the use of earnings forecasts will most likely lead to an increase of power of the test, the forecasts may not be generalizable to other voluntary firm disclosure mechanisms. Therefore, future research might consider using other forms of voluntary disclosure Despite the aforementioned limitations, my paper certainly contributes to the voluntary disclosure theory. I do this by broadening findings of existing research and literature concerning voluntary firm disclosures, possible managerial incentives, and by providing a contribution to a larger understanding of the effect of the direct and implicit relation between unionization and firm disclosure. Particularly this aspect is of additional value, since it has received little attention in previous research. Additionally, my paper gives new insights by focussing on management earnings forecasts as a proxy for voluntary disclosure.

(29)

6. References

Agrawal, A. K. and Matsa, D. A. (2013) Labor unemployment risk and corporate financing decisions, Journal of Financial Economics, 108(2), pp. 449-470.

Ajinkya, B., Bhojraj, S. and Sengupta, P. (2005) The association between outside directors, institutional investors and the properties of management earnings forecasts, Journal of

accounting research, 43(3), pp. 343-376.

Ali, A., Klasa, S. and Yeung, E. (2014) Industry concentration and corporate disclosure Policy, Journal of Accounting and Economics, 58(2), pp. 240-264.

Anilowski, C., Feng, M. and Skinner, D. J. (2007) Does earnings guidance affect market returns? The nature and information content of aggregate earnings guidance, Journal of Accounting

and Economics, 44(1–2), pp. 36–63.

Athanasakou, V., Strong, N. C. and Walker, M. (2011) The market reward for achieving analyst earnings expectations: does managing expectations or earnings matter, Journal of Business

Finance & Accounting, 38(1‐2), pp. 58-94.

Baginski, S. P., Conrad, E. J. and Hassell, J. M. (1993) The effects of management forecast precision on equity pricing and on the assessment of earnings uncertainty, Accounting

Review, p. 913-927.

Baginski, S. P. and Hassell, J. M. (1990) The market interpretation of management earnings forecasts as a predictor of subsequent financial analyst forecast revision, Accounting Review, pp. 175-190.

Baginski, S. P. and Hassell, J. M. (1997) Determinants of management forecast precision, Accounting Review, pp. 303-312.

Ball, R. and Shivakumar, L. (2008) How much new information is there in earnings?, Journal of

Accounting Research, 46(5), pp. 975-1016.

(30)

Antecedents and outcomes associated with forecast venue and forecast specificity choices, Journal of Accounting Research, 36, pp. 167-190.

Basu, S. (1997) The conservatism principle and the asymmetric timeliness of earnings, Journal of

accounting and economics, 24(1), pp. 3-37.

Berger, P.G. (2011) Challenges and opportunities in disclosure research - A discussion of the financial reporting environment: Review of the recent literature, Journal of Accounting and

Economics, 51, pp. 204-218.

Beyer, A., Cohen, D., Lys, T. and Walther, B. (2010) The financial reporting environment: Review of the recent literature, Journal of Accounting and Economics, 50, pp. 296-343. Beuselinck, C., Deloof, M. and Manigart, S. (2008) Private equity investments and disclosure

policy, European Accounting Review, 17(4), pp. 607-639.

Blanchflower, D., G. A. Oswald and P. Sanfey. (1996) Wages, profits, and rent-sharing, The

Quarterly Journal of Economics 111 (1), pp. 227–51.

Botosan, C. A. and Plumlee, M. A. (2002) A re‐examination of disclosure level and the expected cost of equity capital, Journal of accounting research,40(1), pp. 21-40.

Bova, F. (2013) Labor Unions and Management’s incentive to signal a negative outlook, Contemporary Accounting Research, 30 (1), pp. 14-41.

Bowen, R. M., DuCharme, L. and Shores, D. (1995) Stakeholders’ implicit claims and accounting method choice, Journal of Accounting and Economics, 20, pp. 255–295.

Breeden, R. (1995) Testimony before the Securities Subcommittee of the Senate Committee on Banking, Housing, and Urban Affairs, April 16, Government Printing Office, Washington DC

(31)

effects, Journal of Political Economy, 102(6), pp. 1281-1287.

Brown, L. D. (2001) How important is past analyst forecast accuracy?, Financial Analysts

Journal, 57(6), pp. 44-49.

Cheng, Q. and Warfield, T. D. (2005) Equity incentives and earnings management, The accounting

review, 80(2), pp. 441-476.

Choi, J. H. and Ziebart, D. A. (2004) Management earnings forecasts and the market's reaction to predicted bias in the forecast, Asia-Pacific Journal of Accounting & Economics, 11(2), pp.167- 192.

Chrisofides, L. and Oswald, A. (1992) Real wage determination and rent-sharing in collective bargaining agreements, The quarterly Journal of Economics, 107(3), pp. 985-1002.

Coller, M. and Yohn, T. L. (1997) Management forecasts and information asymmetry: An examination of bid-ask spreads, Journal of accounting research, 35(2), pp. 181-191. Conference Board (2003) Americans Slowly Regaining Trust in Companies and Markets.

Available at: http://www.irwebreport.com/trends/030701.htm.

Cotter, J. (2006) Extreme value estimation of boom and crash statistics, The European Journal of

Finance, 12(6-7), pp. 553-566.

Darrough, M. N. (1993) Disclosure policy and competition: Cournot vs. Bertrand, Accounting

Review, pp. 534-561.

Degeorge, F., Patel, J. and Zeckhauser, R. (1999) Earnings management to exceed thresholds, The

Journal of Business, 72(1), pp. 1-33.

Depoers, F. (2000) A cost benefit study of voluntary disclosure: Some empirical evidence from French listed companies, European Accounting Review, 9(2), pp. 245-263.

(32)

Finance, 46 (4), pp. 1325–1359.

D’Souza, J., Jacob, J. and Ramesh, K. (2001) The use of accounting flexibility to reduce labor renegotiation costs and manage earnings, Journal of Accounting and Economics, 30, pp. 187– 208.

Dye, R. A. (1985) Disclosure of nonproprietary information, Journal of accounting research, pp. 123- 145.

Francis, J., Nanda, D. and Olsson, P. (2008) Voluntary disclosure, earnings quality, and cost of capital, Journal of accounting research, 46(1), pp. 53-99.

Frost, A. C. (2000) Explaining variation in workplace restructuring: The role of local union capabilities, Industrial & Labor Relations Review, 53(4), pp. 559-578.

Hartnett, N. and Romcke, J. (2000) The predictability of management forecast error: A study of Australian IPO disclosures, Multinational Finance Journal, 4(1/2), pp. 101-132.

Hayn, C. (1995). The information content of losses. Journal of accounting and economics, 20(2), 125-153.

Healy, P. M. and Palepu, K. G. (1995) The challenges of investor communication The case of CUC International, Inc, Journal of Financial Economics, 38(2), pp. 111-140.

Healy, P. M. and Palepu, K. G. (2001) Information asymmetry, corporate disclosure, and the capital markets: A review of the empirical disclosure literature, Journal of accounting and

economics, 31(1), pp. 405-440.

Hilary, G. (2006) Organized labor and information asymmetry in the financial markets, Review of

Accounting Studies 11 (4), pp. 525-548.

Hirsch, B. T. (2008) Sluggish institutions in a dynamic world: Can unions and industrial competition coexist?, The Journal of Economic Perspectives,22(1), pp. 153-176.

(33)

Hirst, D. E., Koonce, L. and Venkataraman, S. (2008) Management earnings forecasts: A review and framework, Accounting Horizons, 22(3), pp. 315-338.

Holder-Webb L. and Cohen J. R. (2007) The Association between Disclosure, Distress, and Failur, Journal of Business Ethics , 75, pp. 301–314

Holthausen, R. W. and Leftwich, R. W. (1983) The economic consequences of accounting choice implications of costly contracting and monitoring, Journal of Accounting and Economics, 5, pp. 77-117.

Hossain, M., Perera, M. H. B. and Rahman, A. R. (1995) Voluntary disclosure in the annual reports of New Zealand companies, Journal of International Financial Management &

Accounting, 6(1), pp. 69-87.

Hribar, P. and Yang, H. (2011) CEO overconfidence, managerial earnings forecasts, and earnings

management, Working Paper, University of Iowa and Cornell University.

Hutton, A. P., Lee, L. F. and Shu, S. Z. (2012) Do managers always know better? The relative accuracy of management and analyst forecasts, Journal of Accounting Research, 50(5), pp. 1217-1244.

Hutton, A. P. and Stocken, P. (2007) Effect of reputation on the credibility of management forecasts, Boston College and Dartmouth College, Working Paper.

Jennings, R. (1987) Unsystematic security price movements, management earnings forecasts, and revisions in consensus analyst earnings forecasts, Journal of Accounting Research, pp. 90-110. King, R. R. (1996). Reputation formation for reliable reporting: An experimental

investigation. Accounting Review, 375-396.

Kleiner, M. M. and Bouillon, M. L. (1988) Providing business information to production workers: Correlates of compensation and profitability, Industrial & Labor Relations Review, 41(4), pp. 605-617.

The effect of labour market unionization on firms’ disclosure behaviour