1、 1 外文翻译 原文 Voluntary Disclosure, Earnings Quality, and Costs of Capital Material Source: Internet .Author: Jennifer Francis, Dhananjay Nanda, Per Olsson. We investigate the relation between voluntary disclosure and earnings quality, and the pricing effects of voluntary disclosure unconditionally and
2、 conditional on earnings quality. Our proxy for voluntary disclosure is a coded index of financial information from firms annual reports and 10-K filings. By earnings quality, we mean the precision of the earnings signal emanating from the firms financial reporting system. Such imprecision affects t
3、he capital markets demand for, as well as a firms motive to supply, disclosures that are useful to current shareholders and prospective investors in assessing firm value. Our main proxy for a firms earnings quality is the common factor identified by factor analysis performed on three measures of ear
4、nings quality commonly used in the literature: accruals quality, earnings variability, and the absolute value of abnormal accruals. Analytical research provides conflicting predictions about how earnings quality influences firms disclosure decisions. One strand argues that the information asymmetry
5、between firm insiders and shareholders creates a demand for disclosure and provides an incentive for firms to disclose because the value of additional information is greater in these settings (Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983). An implication is that firms with poor (good) ea
6、rnings quality will issue more (less) expansive disclosures because information asymmetry between the firm and investors is higher (lower) in such firms; we interpret this as a substitutive relation. Another strand shows that as information quality increases (such that the quality of the managers in
7、formation increases) managers have incentives to disclose more (Verrecchia, 1990). Under this view, firms with poor (good) earnings quality will issue less (more) expansive disclosures, because investors will treat such disclosures as less (more) credible; we interpret this as a complementary relati
8、on. As discussed in section 2.1, the differing implications from this literature stem from the modeling of earnings quality. 2 A separate but related issue is whether voluntary disclosures result in a lower, or higher, overall cost of capital to the firm. Most prior studies suggest that greater volu
9、ntary disclosure should lower informationasymmetry and, therefore, reduce the cost of capital (Diamond and Verrecchia, 1991). Other research argues, howe ver, that increasing cost of capital effects may occur if the disclosures themselves lead to a more asymmetric information environment than would
10、exist in their absence (Kim and Verrecchia, 1994; McNichols and Trueman, 1994; Zhang, 2001). This body of work motivates our second hypothesis concerning the relation between disclosure and cost of capital. How disclosure relates to earnings quality affects the link between disclosure and the cost o
11、f capital. If voluntary financial disclosure is fundamentally driven by earnings quality (either in a complementary way or in a substitutive way), then earnings quality will have a first order effect on the cost of capital and disclosure will have a second order effect. An implication is that tests
12、which do not control for earnings quality may find significant relations between costs of capital and disclosures, but they may be driven by the omission of the underlying (and correlated) primitive construct earnings quality. For example, if voluntary disclosures complement earnings quality, then f
13、irms with better earnings quality will disclose more. In this setting, tests which do not control for earnings quality will show that more disclosure reduces the cost of capital, but this effect is driven by high disclosure firms also having good earnings quality (the true driver of the lower cost o
14、f capital). Our third hypothesis addresses the importance of conditioning on the first-order effect of earnings quality in investigating the second-order effect of disclosure. Results of tests conducted on 677 firms disclosures for 2001 show a significant complementary association between earnings q
15、uality and voluntary disclosure, implying that firms with good earnings quality select higher levels of disclosure than do firms with poor earnings quality. We find that greater voluntary disclosure is associated with a lower cost of equity, unconditional on other factors. This relation disappears,
16、however, when we control for earnings quality, indicating that voluntary disclosure has no distinct pricing effect. These results are robust to industry-adjustments, estimation procedures (parametric and non-parametric), and alternative proxies for both earnings quality and costs of equity. We furth
17、er find that voluntary disclosure is associated with innate earnings quality, or earnings quality that is related to a firms fundamental business model, and is unrelated to 3 discretionary earnings quality, or the component that is, at least in part, due to managerial choices. These findings are con
18、sistent with earnings quality being a determinant of voluntary disclosure rather than disclosure choices determining a firms earnings quality. Our main findings are, however, sensitive to other proxies for voluntary disclosure, notably measures based on management forecasts, conference calls, and pr
19、ess releases. In general, press releases and conference call activity exhibit opposite relations to those found for our self-constructed score (i.e., press releases and conference call activity exhibit a substitutive, not a complementary, relation with earnings quality firms, and management forecast
20、s are unrelated to earnings quality). Further, while press releases show no meaningful associations with cost of equity, management forecasts and conference call activity are associated with higher, not lower, costs of equity. Our paper makes at least two contributions to the voluntary disclosure li
21、terature. First, we demonstrate the importance of viewing voluntary financial disclosure as a response to earnings quality when analyzing the cost of capital consequences of such disclosures. Second, by examining alternative proxies for all constructs voluntary disclosure, earnings quality, and cost
22、s of equity for a constant sample of firms, our study addresses the source of (some) conflicting results found in this literature. As noted above, we find that our results are robust to various operationalizations of the constructs for earnings quality and cost of equity, but not for the construct o
23、f voluntary disclosure. We conclude that how one operationalizes the voluntary disclosure construct is crucial for drawing inferences about how voluntary disclosure relates to both its determinants (such as earnings quality) and its outcomes (such as cost of equity). Hypotheses on the Relations amon
24、g Earnings Quality, Voluntary Disclosure, and Costs of Capital The discretion exercised by a privately informed manager or firm in disclosing information has been examined by a large literature. The primary insight from theoretical work is that managers disclose their private information because rat
25、ional buyers would otherwise interpret non-disclosure as unfavorable new sand consequently discount the value of the firms assets (see Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983, 2001). The managers disclosure mitigates the adverse selection problem in capital markets by reducing infor
26、mation asymmetry between the firm and investors, enabling greater liquidity anlowering the firms cost 4 of capital (Glosten and Milgrom, 1985; Diamond and Verrecchia, 1991). In this section, we describe the role played by the quality of information arising from the firms underlying financial reporti
27、ng system in enabling disclosure. We begin by summarizing research on the relation between a firms disclosure choice and its earnings quality (section 2.1) and its cost of capital(section 2.2). In section, we link these literatures by emphasizing the need to examine voluntary disclosure conditional
28、on earnings quality. 2.1. Earnings quality and disclosure Early work studying voluntary disclosure treats the quality of the managers private information as exogenous (see Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983). In such settings, one obtainsthe result that disclosure mitigates the
29、 information asymmetry in the market, so that firms with greater asymmetry increase disclosures to improve shareholders information environment. If a measure of the firms earnings quality is used to proxy for information asymmetry (under the belief that earnings quality is causally related to the in
30、formation asymmetry),the implication is that the level of a firms disclosure is inversely correlated with earnings quality. That is, poor (good) earnings quality firms disclose more (less), or a substitutive relation. This intuition ignores, however, the fact that in such a setting the firms disclos
31、ures would also be based on poor quality information and, hence, a rational expectations market will place less credence on such disclosures. This argument demonstrates the need to endogenize the disclosure decision and recognize that disclosures made by the manager will originate from an underlying
32、 information system that may be of poor (or of high) quality. Several theoretical studies model the endogenous relation between disclosure choice and information quality. Dye 1985 and Jung and Kwon 1988 model information quality as the probability that a manager is privately informed, and investigat
33、e his decision to disclose or withhold his information. As the probability of being informed (i.e., information quality) increases, the probability of disclosure also increases since the market is more likely to interpret non-disclosure as bad news and consequently discount the firms value. This set
34、ting, therefore, predicts a complementary association between information quality and voluntary disclosure. An alternative characterization of information quality (which also yields a complementary association) is provided by Verrecchia 1990, who models it as the precision of the information signal
35、observed by the firms manager. He shows that the equilibrium disclosure threshold decreases and the probability of disclosure 5 increases as the precision of a managers private information increases. The intuition here is that because market participants know the precision of the managers private in
36、formation, if a firm with high quality information withholds information from the market, a rational expectations marketwill discount the value of the firms assets. This force causes the firms disclosure threshold to decrease andthe probability of disclosure to increase, resulting in the prediction
37、of more (less) disclosures for firms with good (poor) information quality, or a complementary relation. Finally, in Pennos 1997 model disclosure choice is based on two countervailing forces: as the probability of being informed increases, disclosure is more likely (as in Dye, 1985), but the concomit
38、ant reduction in information quality makes disclosure less likely (as in Verrecchia, 1990). If higher information quality significantly lowers the probability of being informed, Penno shows that disclosure may substitute for information quality. 译文 自愿性信息披露，收益质量与资本成本 资料来源 :因特网 作者： Jennifer Francis, D
39、hananjay Nanda, Per Olsson. 我们调查自愿性披露与收益质量之间的关系和无条件自愿披露的定价与有条件收益质量之间的 关系。我们代理制作企业的年度报告和 10 - K 申报财务信息的自愿信息披露的编码索引。收益质量，即企业的财务报告制度产生的报告信息所带来收益的精密程度。 收益质量的不精确影响资本市场的需求，就如企业的理念之于供应，对现有股东和公司价值评估潜在的投资者来说，信息披露是了解企业的捷径。我们对一个企业的盈余质量主要代理是三个度量标准中执行的因子分析确定的共同因素，这三个因素常被用于文献中：应计质量，盈利变化，以及异常应计项目绝对值。 分析研究提供了有关如何做出
40、公司收益质量的披露决策和冲突的预测。 一种观点是 公司内部人和股东之间的信息不对称制造了一个公开的需求，为企业提供了一个信息披露的理由，因为信息披露的价值就是更好的解决了信息不对称这一问题（ Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983）。另一种观点是，收益质量的好差取决于公司发布信息披露的内容的多少，因为在这样的企业，公司内部人和股东之间的信息不对称程度也会随之变化，我们解释为替代的关系。还有一种说法表明，随着信息质量提高（例如，该经6 理所提供的信息质量变好）经理们便会给出更多的激励信息披露 机制（ Verrecchia，
41、 1990）。根据这种观点，企业盈利质量的好差影响发布膨胀的披露的多少，因为投资者将看到披露的可信程度，我们解释为互补关系的。正如下面讨论的，这些文献的不同影响，源于收益质量建模。 一个独立但相关联的的问题，自愿披露给总成本对公司带来的结果究竟是更低还是更高。以前的研究表明，想让自愿信息披露做的更好应降低信息不对称，因此，减少资金（ Diamond and Verrecchia， 1991）的成本。其他的研究认为，资本效应中增长的成本的影响可能会引起他们自己的披露反而会导致比他们缺席 的情况下更不对称的信息环境（ Kim and Verrecchia， 1994; McNichols and
42、Trueman, 1994; Zhang, 2001）。这个工作核心指示了我们的第二个假设有关信息披露和资本之间成本的关系。 如何披露有关披露与资本成本之间收益质量的影响。如果自愿披露财务资料的根本驱动收益质量（无论是在互补的方式或替代的方式），那么收益质量对资本和披露成本的影响将有一阶二阶效应。一种是，测试一下哪个不受收益质量控制就可能找到资本成本和信息披露的显着关系，但它们可能是由深层次（及相关） 的遗漏而运作的原始构造 -收益质量。例如，如果自愿披露收益质量的补充，然后用更好的收益质量来让公司披露更多。在这种设定下，试一下哪个不受收益质量控制将表明，更多的披露降低资金成本，但这种影响是由
43、高披露也有良好的收益质量（对资本成本较低真正的动力）公司运作为前提条件。我们的第三个假说地址调查披露二阶效应的调节对于收益质量的一阶效应的重要性。 测试了 677 家在 2001 年进行披露的公司，他们的收益质量和信息披露之间有着显著的互补，这意味着收益质量良好的公司会选择比收益质量不佳的公司更高的披露水平。我们发现，更大的自 愿性披露会使股权成本降低，不受其他相关因素影响。去掉这种关系，发现当我们控制收益质量，自愿披露没有明显的价格的影响。这些结果对行业调整、预算步骤（参数和非参数）还有收益质量与权益的其他代理带来很大冲击。我们进一步发现，自愿披露与原先的收益质量或者一个企业的基本经营模式相
44、关的收益质量有关，与任意的收益质量或其组件无关，至少部分是由于管理的选择。这些研究指出了收益质量是一个自愿披露信息披露的选择，而不是决定因素决定一个企业的收益质量。 我们的主要发现的是自愿披露其他代理的灵敏性，特别是在电话会议和新闻稿的基础 上对管理措施进行预测。一般来说，新闻发布和电话会议展览活动于我们的自设标准相反（即新闻稿和会议呼叫活动表现为取缔，而不是互补的，公司收益质量的关系，发现这些预测和收益质量管理无关）。此外，尽管新闻表示按照公平，管理预测和电话会议协会活动的成本没有真正意义上的公平与更7 高的质量成本相关。 我们的论文至少为自愿披露作出了两个贡献。首先，我们证明了观看作为收益
45、质量在分析这种披露资本成本的反应后果自愿财政披露的视点重要性。其次，通过检查所有结构替换的代理 - 自愿披露、盈利质量和资产成本 - 对于一个作为恒量样品 的企业，我们的研究解决了在这些文献发现某些冲突的结果的来源。如上所述，我们发现，我们的成果是稳固了各种 付诸实施 的收益质量和股本成本结构，但并非为构建自愿披露。我们的结论是如何付诸实施的一个自愿性信息披露建设是至关重要的描述自愿披露有关如何既涉及到它的决定因素（例如盈利质量）和其成果（如股权成本）的推论。 对盈余质量、自愿披露、资本成本之间关系的假设： 由经理或信息披露公司行使的酌情权已大量审查通过。从理论工作的主要观点来看，管理人员公开
46、其私人信息，是为了在理性买家们面前解释那些不利的信息披露，由此贴现公司 的资产值（ see Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983, 2001）。这位经理的披露减轻和减少了公司和投资者之间的信息不对称，使更大的流动性降低了公司的资本成本（ Glosten and Milgrom, 1985; Diamond and Verrecchia, 1991）。 在本节中，我们描述了从企业的基础在使所披露的财务报告信息的质量体系发挥的规则。我们开始总结一家公司的信息披露的选择与其收益质量和它的资本的成本关系的研究。 我们强调有必要
47、审查自愿性披露的盈余品质有条件这些文献。 收益质量与信息披露： 早期的工作中学习自愿披露与待管理人的私人资料外源性（ see Grossman and Hart, 1980; Milgrom, 1981; Verrecchia, 1983）。在这样的环境下， 一得到披露结果 就 减轻了市场信息不对称 ，使更大的不对称性增加，必须提高公司披露股东信息环境。如果该公司的收益质量的衡量标准是代理信息的不对称程度（抛去信仰的说法普通认为收益质量与信息不对称是因果关系），其含义是，一个公司的信息披露水平与收益质 量成反比密切相关。也就是说，公司披露程度差带来更多的收益质量，或者是一种替代的关系。这种直观
48、上的忽略的事实是在这种环境下的公司的信息披露建立于贫乏的质量信息，因此，市场理性预期将会给予披露较少信任。这种说法表明，内源性披露的决定是有必要的，并承认由经理作出披露将源于一个基本的信息系统，质量不分好差。 研究的若干理论模型之间的选择和信息披露质量存在内在关系。 Dye 1985 和 Jung 和 Kwon 1988模型做了一个经理披露概率信息的质量和调查，他可以决定公开或隐瞒他的信息。由于信息质量被告知概率 的增加，披露的概率也会增加，因为市场更可能不披露坏消息。因为此设置，预测自愿披露和信息的质8 量之间的互补的关联。一个信息的质量（也产生了一个互补的关联）替代特性是由把它作为信号的精
49、度的信息由企业的经理观察模型的 Verrecchia 1990。他表明，平衡的披露门槛降低，披露精度提高就好比一个经理人的私人信息披露提高。比较直观，因为市场参与者知道该经理的私人信息的精度，如果一个高质量的信息资料被隐瞒，市场理性则会预测贴现该公司的资产价值。这种力量使公司的披露界线以降低信息披露进行变化的可能性增加，在高信息披 露良好信息的质量，或公司的预测结果的互补关系。最后， Pennos 1997模型披露的选择是基于两个反补力量：作为被告知者的可能性增加，信息披露的程度可能更大（ as in Dye, 1985），但在信息的质量也相应降低，使信息披露的可能性减小（ as in Verrecchia, 1990）。如果更高的信息质量明显降低了被告知的可能性， Penno 表示，信息披露或许会取代信息的质量。
