How Do Financial Constraints Relate to Financial Reporting Quality? Evidence from Seasoned Equity Offerings

Abstract

This paper examines how constraints on firms’ financing capacity relate to managers’ discretionary accounting choices. Three hypotheses of earnings management – the opportunism hypothesis, the rational expectations hypothesis, and the signaling hypothesis – predict that constrained firms engage in greater upward earnings management than unconstrained firms when selling equity. Using a sample of seasoned equity offerings (SEOs) announced between 1983 and 2014, I find support for this prediction. The relation between financial constraints and earnings management is robust to including controls such as offer size, growth opportunities, analyst following, and chief executive officer equity holdings, as well as to using the instrumental variable approach. Investors’ reaction around and following the SEO announcement supports the rational expectations hypothesis. I find that aggressive earnings management by constrained issuers is associated with lower SEO announcement returns but is not followed by negative abnormal returns in the long run. The evidence suggests that constrained issuers’ aggressive use of income-increasing accruals is an outcome of managerial myopia caused by capital market pressure, not managerial opportunism intended to mislead investors.

Acknowledgements

This paper is based on my dissertation completed at the University of Pittsburgh. I appreciate helpful comments of Guochang Zhang (editor), an anonymous reviewer, John Chadam, Katrien Craninckx, Diane Denis, Jesse Ellis, Mei Feng, Ken Lehn (committee chair), Gershon Mandelker, Mercedes Miranda, Nandu Nagarajan, Sara Moeller, Shawn Thomas, Joseph Vu and seminar participants at the Rotterdam School of Management, Universitat Pompeu Fabra, University of Toledo, Salisbury University, the 2012 Financial Management Association meetings, the 2011 Eastern Finance Association meetings, and the 2011 Midwest Finance Association meetings.

Notes

1 My paper seeks to examine how constraints on firms’ financing capacity alter managers’ nonoperational decisions (i.e. decisions with no immediate cash flow implications). Thus, I focus on managers’ discretionary accounting choices and do not extend my analysis to real activities manipulation (e.g. cutting discretionary spending). In a related study, Campello et al. (2010) document that financially constrained firms are more likely to cut technology and marketing expenditures when facing tight economic conditions.

2 Here the term ‘utility’ refers to the usefulness of the financing transaction in terms of dollar amount of capital obtained. A manager would like to maximize total proceeds raised through the financing transaction by presenting the firm in the best possible light. Earnings management helps the manager achieve this goal.

3 Several studies support the argument that financially constrained firms do not have an established reputation in financial markets (e.g. Devos, Dhillon, Jagannathan, & Krishnamurthy, 2012; Fernando, May, & Megginson, 2012; Fortin & Pittman, 2004; Kasahara, 2008). Evidence also suggests that financially unconstrained firms act as if they wish to preserve their reputation. Law and Mills (2015), for instance, find that unconstrained firms engage in less aggressive tax planning strategies than constrained firms. Consistent with the reputation argument, constrained firms incur higher costs than unconstrained firms when accessing the capital markets (e.g. Hennessy & Whited, 2007).

4 None of the SEO firms in my sample is truly financially constrained because all of them have access to equity markets. Following Korajczyk and Levy (2003), I use the terms ‘constrained’ and ‘unconstrained’ to denote a relative relation. Nevertheless, it is worth noting that most issuers operate under tight financial conditions prior to the SEO (DeAngelo, DeAngelo, & Stulz, 2010). In fact, DeAngelo et al. (2010, p. 276) point out: ‘[M]ost issuers would have run out of cash by the year after the SEO had they not received the offer proceeds.’

5 Lamont, Polk, and Saa-Requejo (2001, p. 529) note: ‘This inability to fund investments might be due to credit constraints, or inability to borrow, inability to issue equity, dependence on bank loans, or illiquidity of assets.’ Consistent with previous research, I do not use the term ‘financial constraints’ to mean financial distress.

6 In general, the earnings management game between issuers and investors is analogous to the prisoner’s dilemma. The cooperative equilibrium would involve no earnings inflation on the part of managers and no discounting of inflated earnings by the stock market. However, this is not sustainable as a Nash equilibrium. If investors do not conjecture earnings inflation, managers will have an incentive to fool them by boosting earnings. More important, the higher the issuer’s financial constraints, the worse the problem becomes.

7 The signaling hypothesis is based on the idea that managerial discretion in earnings management improves the ability of earnings to reflect the underlying economics of the firm. Thus, the signaling hypothesis does not always imply optimism on the part of managers. However, in the context of SEOs, managers are expected to use earnings management to send a favorable signal to the market in hopes of obtaining higher proceeds from the offering.

8 To mitigate the influence of outliers, I winsorize all the variables at the top and bottom 0.5% within each quarter. I also require at least 10 observations in each industry and quarter to accurately estimate the regression coefficients.

9 I do not discard those observations in the middle tercile (according to a financial constraints criterion) and keep such opaque issuers in the sample to be used in the regression analysis. I create a dummy variable (labeled ‘Medium-FC’) that is equal to 1 if an issuer is in the middle tercile and 0 otherwise. Although I expect that these firms also manage their earnings more aggressively as compared with financially unconstrained firms, the significance of the difference in discretionary current accruals between the two groups is ultimately an empirical question.

10 I labeled the quarter of the last earnings announcement prior to the SEO announcement as quarter −1. Quarter 0 is the quarter of the first earnings announcement following the SEO announcement. Other quarters are labeled accordingly.

11 Although it is not stated in Equation (5), the natural log of the pre-offer market value is added to the model as a control variable when the payout ratio measure is used to determine the financial constraint groups. The reason that I do not control for the market value when other financial constraint measures are employed is that the market value is a proxy for the firm size, which is already captured in the other constraint measures. The correlation of the natural log of the market value with assets, payout ratio, the WW index, and the SA index is, respectively, 0.79, 0.13, −0.63, and −0.62.

12 I use a five-day window instead of a three-day or two-day window to calculate CARs because previous research highlights that the actual announcement date (identified through news searches) sometimes precedes the filing date recorded by the SDC (Carlson, Fisher, & Giammarino, 2010; Kim & Purnanadam, 2014). Using three-day or two-day CARs yields directionally consistent results.

13 Using the ‘suest’ command in Stata, I also test whether the estimated coefficients on DCA (Q₋₄ through Q₋₁) are significantly different between the constrained and unconstrained columns throughout the table. The results reveal a significant difference in estimated coefficients under the size and WW index classifications (p < 0.05) but not under the payout and SA index classifications.

14 For this particular analysis, I exclude SEOs announced between 2011 and 2015 because I examine stock return performance for the five-year post-SEO period and because stock return data are not available beyond December 2015.

15 Finding a support for this argument may also shed light on constrained issuers’ continued use of higher income-increasing accruals during the post-SEO period.

16 The rational expectations hypothesis (as laid out in the present research and related previous studies) is not focused on possible direct personal benefits managers derive from upward earnings management around SEOs (i.e. engaging in post-SEO insider selling at potentially inflated prices). Instead, the hypothesis is built on the premise that investors focus primarily on corporate-level incentives of constrained and unconstrained SEO firms when forming their expectations (i.e. boosting stock price to increase offer proceeds and thereby corporate cash holdings). Thus, it is difficult to interpret this finding in light of the rational expectations hypothesis.

17 The IV diagnostics indicate that the present analysis does not suffer from a weak instrument problem. The F-statistic of the first-stage IVs of ln(Assets), payout ratio, the WW index, and the SA index is 107.61, 10.19, 58.05, and 48.15, respectively (p < .01). In addition, the p-value of the Hansen J statistic for ln(Assets), payout ratio, the WW index, and the SA index is 0.156, 0.814, 0.290, and 0.684, respectively, suggesting that the overidentification restriction of both IVs holds under each measure of financial constraints. While I include ln(Market value) as an additional control in the OLS model under the payout classification, I exclude ln(Market value) from the 2SLS model to increase the consistency of the 2SLS estimation (because the variable is highly correlated with market share). I obtain qualitatively similar results if I keep ln(Market value) in the 2SLS model.