Tone of 10-Ks and Liquidity Events: Baseline Results--论文代写范文精选

我们感兴趣的是如何衡量，公司面临金融摩擦会缩减他们的红利,财政拮据的公司可能有困难资助退休金。因此这篇金融paper代写范文调查的是金融约束如何预测未来股息。

Abstract 
  A number of prior studies examine the relation between financial constraints and a firm’s capital structure and payout. More constrained firms are found to have high cash holdings, keep higher leverage, and pay lower dividends. The interpretation of these results, however, is often problematic due to endogeneity concerns as financial choices and constraints are determined simultaneously. Instead, we investigate how well different measures are able to predict future developments associated with the deterioration or improvement of external financing conditions. Relating current financial constraints to future liquidity events alleviates the endogeneity issues. 

  We are interested in how our measure performs on its own and alongside other measures. We build on the insights of Cleary (1999) and Whited (2009) who argue that firms facing financial frictions would scale down their committed dividend distributions to shareholders. Farre-Mensa and Ljungqvist (2015) point out that firms facing financial constraints should be less likely to engage in equity recycling, i.e., simultaneous raising and paying out of equity. Rauh (2006) shows that financially constrained firms may have difficulty funding pension obligations to their retirees which subsequently undermines their ability to undertake investments. We therefore investigate how well measures of financial constraints predict future dividend omission and increases, equity recycling, and underfunded pension plans. 

  As a first step, how well do the KZ, SA, WW indexes and % Constraining predict liquidity events without controlling for firm characteristics? Table 4 reports summary results from 16 separate regressions. For each of the four ex post liquidity events, the KZ index, SA index, WW index, and percentage of constraining words are independent variables. In each regression, an intercept, Fama and French (1997) 48-industry dummies, and calendar year dummies are included. The standard errors in all the regressions are clustered by both year and industry. The regressions for dividend omission, dividend increases, and underfunded pensions are logits while the regression for equity recycling is OLS. 

  These regressions provide an important perspective for assessing the usefulness of the various indexes because they do not yet include standard control variables which we know will be highly correlated with the index components (e.g., the correlation between log of total assets and log of market capitalization is 0.87). In Table 4, an “X” represents a coefficient that is both significant at the 1% level and has the expected sign. Generally, traditional measures of financial constraints do a poor job of predicting the ex post liquidity events even when isolated from the inclusion of standard macro-finance variables. The SA index has a significant coefficient value for only the dividend omissions. The KZ index is significant and has the expected sign for only equity recycling and underfunded pensions. 

  The SA and WW indexes have no link with equity recycling or underfunded pensions. As noted earlier, the poor ability of the SA and WW indexes to predict equity recycling has already been documented by Farre-Mensa and Ljungqvist (2015). The two authors sharply criticize the existing literature’s measures of financial constraints. In contrast, the percentage of constraining words is significantly linked with all four liquidity events. The fact that existing measures of constraints are quite imperfect may not be surprising for all readers. One could argue that the HP and WW papers illustrate the difficulties in consistently measuring constraints across time and samples, rather than as definitive indexes of the proper way to measure constraints. At its core, the notion of financial constraints is a delicate and nuanced concept.

  Control Variables 
  Will the predictive power of the percentage of constraining words continue to be robust once additional firm level control variables are added to the regressions? Our four additional control variables are (1) natural logarithm of market capitalization (stock price times shares outstanding); (2) natural logarithm of the book-to-market ratio; (3) excess prior year buy-andhold returns; and (4) a dummy variable equal to one when the prior fiscal year income before extraordinary items is negative, zero otherwise. More detailed variable descriptions are provided in Appendix A. From Table 5 we observe that the control variables are generally statistically significant in all the regressions and represent obvious first-order factors an investor should consider when identifying financially constrained firms. As before, an intercept, Fama and French 48-industry dummies, and calendar year dummies are included in all regressions. The zstatistics (or t-statistics for the OLS regressions) are in parentheses with the standard errors clustered by both year and industry.

  Dividend Omissions 
  In the first two columns of Table 5, the dependent variable is the Dividend Omission Dummy. Since only firms with at least one dividend distribution in the prior year are included in the first two regressions, the sample size is 12,669 firm-year observations. The control variables imply that larger and better performing companies are less likely to stop paying dividends, whereas companies with negative earnings or value firms (i.e., high book-to-market value) are more likely to do so. When % Constraining words is added as an independent variable in column (2), its coefficient is positive (1.318) with a statistically significant coefficient value (z-statistic of 3.80). More constraining words in a 10-K are linked with a higher likelihood of omitting dividends in the year following the 10-K filing. The marginal effect of the coefficient is 0.023 while the standard deviation of percent constraining is 0.166. Thus, a one standard deviation increase in percentage of constraining words is associated with a 10.32% (=0.023*0.166 / dividend omission sample mean of 0.037) higher chance of a dividend omission.

  Dividend Increases 
  In columns (3) and (4) of Table 5, the dependent variable is the Dividend Increase Dummy (equal to one if the firm increased its dividend during the following year, else zero). As before, only firms with at least one dividend distribution in the prior year are included in the regressions. The column (3) regression includes only the control variables. It is worth noting that company market capitalization and past performance are related—as expected—to dividend increase: larger and better performing companies are more likely to increase dividends. Value firms and companies with negative trailing earnings are less likely to increase their dividends. In column (4), the coefficient on % Constraining (-0.733) is statistically significant at the 1% level. This implies that as the percentage of constraining words in the 10-K rises, the likelihood of the firm increasing its dividend decreases. The marginal effect of the coefficient is -0.178 while the standard deviation of percent constraining is 0.166. Thus, a one standard deviation increase in percentage of constraining words is related to a -6.46% (=-0.178*0.166 / dividend increase sample mean of 0.4572) smaller likelihood that a firm will increase its dividend in the subsequent year.aper代写)

  51Due网站原创范文除特殊说明外一切图文著作权归51Due所有；未经51Due官方授权谢绝任何用途转载或刊发于媒体。如发生侵犯著作权现象，51Due保留一切法律追诉权。
  更多paper代写范文欢迎访问我们主页 www.51due.com 当然有paper代写需求可以和我们24小时在线客服QQ:800020041 联系交流。-X(paper代写)