عرضه سهام فصلی: کیفیت اطلاعات حسابداری و هزینه های شناور مورد انتظار

Flotation costs represent a significant loss of capital to firms and are positively related to information asymmetry between managers and outside investors. We measure a firm's information asymmetry by its accounting information quality based on two extensions of the Dechow and Dichev [2002. The quality of accruals and earnings: the role of accrual estimation errors. Accounting Review 77, 35–59] earnings accruals model, which is a more direct approach to assessing the information available to outside investors than the more commonly used proxies. Our main hypothesis is that poor accounting information quality raises uncertainty about a firm's financial condition for outside investors, though not necessarily for insiders. This accounting effect lowers demand for a firm's new equity, thereby raising underwriting costs and risk. Using a large sample of seasoned equity offerings (SEOs), we show that poor accounting information quality is associated with higher flotation costs in terms of larger underwriting fees, larger negative SEO announcement effects, and a higher probability of SEO withdrawals. These results are robust to joint determination of offer size and flotation cost components and to adjustments for sample selection bias

Flotation costs in seasoned equity offers (SEOs) represent an economically important portion of gross proceeds. Many studies show that underwriting fees range between 3% and 8% of SEO gross proceeds and that SEO announcement effects range between −2% and −3%. The extant literature has generally concluded that a substantial portion of SEO flotation costs are caused by asymmetric information between issuers and outside investors. (See the discussion in Eckbo, Masulis, and Norli, 2007, survey of the security offering literature.) However, information asymmetry is not directly observable, and no generally agreed upon measure exists for it. As a result, many SEO flotation cost studies employ a wide range of distinctly different measures of information asymmetry. This makes it difficult to assess the importance of information asymmetry or pinpoint the other key determinants of flotation costs. Common measures of asymmetric information used in the finance literature include stock return volatility, analysts’ earnings forecast dispersion, proportion of intangible assets, debt rating, and stock bid–ask spread (or a component). Many SEO studies employ proxies for information asymmetry and price uncertainty. Drucker and Puri (1999), Altinkilic and Hansen (2000), and Corwin (2003) use stock return volatility; Marquardt and Wiedman (1998) use analysts’ earnings forecast dispersion; Liu and Malatesta (2006) use debt ratings; and Corwin (2003) uses bid–ask spreads. While heavily used in empirical analysis, none of these variables has a strong theoretical claim to being a clear or complete measure of information asymmetry between issuers and outside investors. Moreover, these measures are likely to capture other economic effects beyond asymmetric information. For example, stock return volatility is also a widespread measure of uncertainty and is influenced by industry- and economy-wide shocks, for which firm managers are unlikely to have a significant information advantage relative to other investors. Dispersion in analyst forecasts can be affected by the number and quality of analysts following a stock, analyst herding, and whether the analysts are affiliated with investment banks, to name just a few of the problems that researchers have highlighted. Debt ratings have been criticized for being slow to incorporate new information and to be more focused on the solvency of a firm, which is strongly related to its leverage. The proportion of intangible assets is also a proxy for a proportion of a firm's value represented by growth opportunities, which could be modest for many firms with sizable information asymmetries. Finally, bid–ask spread is strongly affected by the stock's market microstructure environment, such as exchange rules, trading activity, execution costs, and dealer borrowing costs needed to support inventory positions. It is also often used as a liquidity measure. This liquidity measure is also found to be related to SEO flotation costs as shown in Butler, Grullon, and Weston (2005). In short, none of these commonly used proxies represents a clean measure of asymmetric information between insiders and outside investors regarding a firm's expected future financial performance. In this study, we examine the relation of expected flotation costs to an alternative measure of information asymmetry that is directly related to the information available to outside investors about firm performance. We argue that the quality of a firm's accounting information, which is taken from the current accounting literature, is a reasonable proxy for asymmetric information between managers and outside investors. Our view is that, because accounting statements are the primary source of information about firm performance available to outside investors, its quality should be directly related to investor uncertainty about a firm's financial health and past performance. Because managers have better internal sources of information, financial accounting statement quality is unlikely to cause a similar rise in manager uncertainty, implying that this rise in uncertainty represents an asymmetric information effect. The accounting literature measures accounting information quality by a number of alternative, but related, approaches. The early accounting literature focuses on manager manipulation and earnings management as the primary cause for reduced information quality. In contrast, the more recent literature places more emphasis on uncertainly about operating fundamentals, which Dechow and Dichev (2002, hereafter DD) measure by firm size, length of the operating cycle, sales and cash flow volatility, frequency of negative earnings, and size of accruals, as a major cause of reduced information quality, but also continue to include the effects of manager discretion over accounting decisions. Dechow and Dichev (2002, pp. 46–49) examine the relation of these firm fundamentals to accruals quality and find they have significant explanatory power. Thus, the primary measures of poor accounting information quality we study reflect operating fundamentals and managerial discretion, both of which make firm valuation and earnings forecasting inherently difficult. We decompose accruals quality into its operating fundamentals and discretionary components to assess whether expected flotation costs of SEOs are associated with both causes of impaired accruals quality. As accounting quality deteriorates, investor uncertainty about a firm should rise and demand for its equity should fall. In addition, as issuer accounting quality falls, investment bankers, who write fixed-price underwriting contracts that have a put option structure, are likely to price their underwriting services more dearly. (Smith, 1977, was first to emphasize the put option characteristic of fixed price underwriting contracts.) So we expect greater investor uncertainty to lead to increased equity underwriting and distribution costs. However, we are unaware of any existing studies that directly examine the relation of accounting information quality to equity flotation costs.1 We address this current gap in the literature by investigating the relation of accounting information quality to SEO offer size and expected flotation costs. For this purpose, we focus on three major components of expected flotation costs, namely, underwriting fees, offering announcement effects, and the probability of issue withdrawal. In a typical SEO underwriting contract, a syndicate of investment banks guarantees to purchase an issuer's entire equity offering at a fixed price, bearing the entire price risk associated with reselling the shares to the public once the contract is signed. By signing this contract, the underwriters accept the risk of an unexpected reduction in investor demand for the SEO. When firms with poor accounting information quality announce SEOs, the decision can increase investor uncertainty about the value of issuers’ common stock and, thus, lower investor demand for these equity issues. At the end of the registration period, underwriters have several choices if faced with weak investor demand. They can increase their underwriting fees or they can decline the underwriting assignment. In either case, this represents an increase in the issuer's expected flotation costs.2 If investment banks decide to underwrite an SEO despite the higher risk associated with poor accounting information, they must raise their underwriting fees, which represent a significant expense to SEO issuers. Issuers with poor accounting information face greater investor uncertainty about their stocks’ market value and as a result lower issue demand. An investment bank must offset the higher expected underwriting losses and distribution expenses associated with these less attractive issues by charging higher underwriting fees. We test for a negative relation between underwriter gross spreads and issuer accounting information quality, using several recently developed measures of accounting information quality. If an investment banking syndicate declines to underwrite an offering, it generally forces SEO cancellation, which represents a significant expected cost to an issuer. The inability to raise external capital is one of the greatest costs that a company can face if it delays valuable investment opportunities or forces the company to turn to more costly sources of external capital. An SEO issuer also loses registration fees, accounting expenses, and management time devoted to the offering process when an issue is withdrawn from Securities and Exchange Commission (SEC) registration. Thus, we view the probability of issue cancellation as an important component of expected flotation costs. Because poor issuer accounting quality raises investor uncertainty about an issue's value and concern about adverse selection, it can reduce issue demand and increase the likelihood of offer cancellation. Thus, accounting information quality is expected to be negatively related to our second component of expected flotation costs, the frequency of SEO withdrawals. The expected stock market reaction to an equity offering announcement is yet another important component of expected flotation cost. Because an SEO can occur only after it is publicly announced and SEOs on average have significantly negative announcement effects ranging between −2% and −3% for US industrial firms, a rational manager must expect to sell new stock at a discount below its current stock price. (See Masulis and Korwar, 1986; Asquith and Mullins, 1986; Mikkelson and Partch, 1986; Bhagat and Hess,1986; Eckbo and Masulis, 1992; Eckbo, Masulis, and Norli, 2007.) For this reason, the quality of issuers’ accounting information has a further implication. Because poor quality accounting information can obscure a firm's financial health and its performance, it increases the information asymmetry between issuers and outside investors. On initially hearing news of an equity offering, investors are likely to more heavily discount their valuation of a firm with poor quality accounting information to take into account the greater agency problems and adverse selection risk that investing in such a firm entails. So we expect to observe issuers with worse accounting information quality having more negative announcement returns relative to issuers with better accounting information quality. The accounting literature often measures the quality of a firm's accounting information by its accruals quality. Until recently, accruals quality was primarily measured in terms of discretionary accruals using a variant of the Jones (1991) model, such as the modified Jones model (Dechow, Sloan, and Sweeney, 1995), or performance-matched discretionary accruals (Kothari, Leone, and Wasley, 2005). The rationale for studying discretionary accruals is that managers can exploit their discretion over accounting decisions to enhance reported earnings. However, even in the absence of intentional earnings management, accounting information is affected by volatility in a firm's fundamental operating environment, as well as its industry- and firm specific-characteristics. Poor accruals quality creates more uncertainty for outside investors about a firm's true performance, regardless of whether it is created through earnings management or not. Thus, we follow the more recent financial accounting literature by using an accruals measure that does not distinguish between intentional earnings management and unintentional estimation errors from models of earnings quality, because both imply poor accounting information quality. See Francis et al., 2004 and Francis et al., 2005 for a further discussion of these measures. Following DD, our primary measure of accounting information quality is based on the standard errors of residuals from a model mapping yearly current accruals into operating cash flows in the prior, current, and subsequent years estimated in each of the past five years, where larger standard errors imply poorer quality accounting information. This model was modified by McNichols (2002) to control for changes in sales revenue and property, plant, and equipment (PPE) and is called the modified DD model (hereafter MDD), which we use as our first proxy of accounting information quality. However, measuring the quality of accounting information is complicated by the fact that applicable accounting standards and transparency of accounting information vary considerably across companies and industries. This is a major reason that we propose a new measure of accruals quality, which extends the MDD model to incorporate a firm fixed effect (hereafter FDD). In other words, this measure adjusts the MDD model for unobserved firm characteristics that are time invariant, such as internal accounting policies and cash flow characteristics. This adjustment mitigates possible omitted variable problems associated with the MDD measure. In addition, this new accruals quality measure directly adjusts for a major source of heteroskedasticity in the MDD model. It follows that standard errors from the FDD model should be lower and their cross-sectional variability should better reflect differences in firm accounting information quality. Poor accruals quality increases the information asymmetry between managers and outside investors, which is expected to induce increased investor risk aversion toward investing in these firms’ SEOs. As a result, underwriting SEOs with poor accruals quality should be more risky and costly. Therefore, we predict that issuers of otherwise identical SEOs, except for poor accruals quality, should be associated with larger expected flotation costs. We also require that these accounting-based measures of information asymmetry are not simply capturing stock price uncertainty by also including stock return volatility as a competing explanatory variable in our analysis. We test this flotation cost hypothesis with the well-known MDD model and our refinement of it, the FDD model. To assess the robustness of our findings and their potential causes, we examine a variety of alternative accruals quality measures and then decompose the MDD model into its operating fundamentals and discretionary components. To preview our results, we find empirical evidence strongly supporting the hypothesis that poor accruals quality is associated with larger expected flotation costs. Using a large sample of completed and withdrawn SEOs by US firms between 1990 and 2002, we find that issuers with lower quality accounting information tend to raise more equity capital. However, the trade-off is that poor accruals quality is associated with larger underwriting fees, a more negative market reaction to equity offer announcements, and a higher probability of issue withdrawal. These results are robust to taking into account the endogenous offer size choice, various alternative accruals quality measures, and controlling for potential sample selection bias. Both operating fundamentals and management discretion appear to be partial explanations for our findings. The rest of this paper is organized as follows. In Section 2, we introduce and discuss our accruals quality measures. Data sources and sample characteristics are discussed in Section 3. In Section 4, we examine how accruals quality is related to a firm's choice of equity offering size. In 5, 6 and 7, we investigate the relation between accruals quality and three major components of equity flotation costs: underwriting fees, announcement returns, and issue withdrawal probability, respectively. This is followed by sensitivity analysis of the results and a decomposition of our accruals quality measures into their operating fundamentals and discretionary components in Section 8. Finally, Section 9 presents the conclusions and highlights the contributions of this study.

The fundamental question this study investigates is whether increases in asymmetric information between issuers and outside investors are associated with larger flotation costs in SEOs. This is an important question because flotation costs can consume a large portion of the capital raised in an equity offering. While many prior studies report a positive relation between flotation costs and asymmetric information, there is no generally agreed-upon measure of asymmetric information. Examples of alternative measures of asymmetric information frequently used in the finance literature include stock return volatility, dispersion in analysts’ earnings forecasts, debt rating, and a component of bid–ask spread. However, each of these measures has its weaknesses, so we pursue an alternative approach to measuring asymmetric information based on the issuer's financial accounting information. Given that accounting earnings is one of the most commonly used measures of firm performance, we measure asymmetric information between managers and outside investors based on the quality of this information. More specifically, we examine the reliability of the accruals component of accounting earnings. This is in part motivated by evidence of opportunistic use of accounting discretion by managers for the purposes of window-dressing and manipulation of investor expectations about stock values. Because these actions are more likely to occur around major corporate events such as equity offers, accruals quality around major corporate events has received considerable attention in the accounting literature. While many earlier financial accounting studies use variations on the discretionary accruals model as a proxy for accounting information quality or earnings management, we measure accounting information quality using a relatively new current accruals model developed by Dechow and Dichev (2002) and McNichols (2002). Unlike the earlier approaches, this measure captures not only managerial actions aimed at manipulating accounting information, but also any estimation errors caused by uncertainty about firm operations and its industry conditions. The modified DD measure is based on a track record of an issuer's accruals surprises. Financial accounting researchers generally consider it a better approach to capturing the quality of accounting information than other existing earnings management or discretionary accruals models. We also improve on this model by augmenting it with firm fixed effects to take into account unobservable firm characteristics that are time invariant such as its unobservable internal accounting policies. We argue that this augmented DD model is more appropriate for evaluating the effects of information asymmetry on the flotation costs of SEOs. With our two accounting based measures of asymmetric information, we examine the effects on three major components of expected flotation costs in SEOs; underwriting fees, expected announcement effects, and the likelihood of offering withdrawal. From a large sample of recent SEOs, we find a significant relation between poor accounting information quality and larger underwriting fees, more negative market reactions to offering announcements, and a higher probability of issue withdrawal. These results are consistent across our two measures of accounting information quality. Our conclusions are also robust to controlling for the endogeneity of the offer size decision, potential sample selection bias, and a number of alternative accruals quality measures. We find that measures of accounting information quality have significant explanatory power for equity flotation costs when competing measures of asymmetric information are included as regressors. Furthermore, among these competing measures of asymmetric information, only stock return volatility has a significant marginal effect on equity flotation costs when measures of accounting information quality are included as regressors. Moreover, a strong argument can be made that return volatility is a better measure of stock price uncertainty than asymmetric information. We show that the effects on accounting information quality are due to both the effects of firm operating fundamentals and managerial discretion over accounting decisions. In summary, the results reported here present persuasive evidence that major components of equity flotation costs are positively related to measures of information asymmetry between issuers and uninformed outside investors and that our measures of accounting information quality are credible proxies for this information asymmetry.