Does voluntary balance sheet disclosure mitigate post-earnings-announcement drift?

Abstract

Theory suggests that balance sheet information such as total assets, total equity, or total liabilities complements earnings information in helping investors assess a firm’s profitability and estimate earnings growth. The voluntary disclosure of balance sheet information at earnings announcement could help investors gather and process this information at a lower cost. We therefore predict that voluntary balance sheet disclosure at the time of an earnings announcement helps investors promptly understand the implication of current earnings news for future earnings and subsequently reduces post-earnings-announcement drift (PEAD). Consistent with these predictions, our results show that when firms provide voluntary balance sheet disclosures, the earnings response coefficient in the event window is significantly higher and the corresponding PEAD is significantly lower. We further find that the impact of voluntary balance sheet disclosure on PEAD is more pronounced when the magnitude of balance sheet value surprise is larger, when balance sheet value is more informative about future earnings, when earnings uncertainty is higher, or when information cost is higher, consistent with our conjectures that helping investors to better understand future earnings performance and lowering information costs are key mechanisms underlying the effect of voluntary balance sheet disclosure on PEAD.

Introduction

The accounting literature has indicated the important role of balance sheet information such as total assets, total equity, or total debts in predicting firm performance and valuing equity. For instance, Ou and Penman (1989) show that book value of equity helps predict earnings persistence. Ohlson (1995) finds that information on a firm’s total assets can be used to predict the normal component of its future earnings. Billings (1999) indicates that information on a firm’s debt to equity ratio affects equity valuation because debt to equity ratio is associated with expected earnings growth. More recently, Chen et al. (2019) indicate that balance sheet information such as the amount of assets provides information about firms’ capital stocks. The level of a firm’s capital stock directly affects its future cash flows and the information on capital stock can help financial statement users assess firms’ future cash flows, estimate intrinsic values and make investment decisions. The findings in these studies suggest that balance sheet information complements earnings information in helping investors understand the implication of current earnings for future earnings and thus price the current earnings information more efficiently.¹

However, despite the suggested importance of balance sheet information for pricing earnings information, there is surprisingly little evidence on whether or not the balance sheet information affects the pricing efficiency of earnings news. We aim to fill this gap in the literature by examining whether the concurrent disclosure of balance sheet at earnings announcement improves the pricing efficiency of current earnings news and subsequently reduces the magnitude of post-earnings-announcement drift (PEAD).

It is well-documented in the literature that stock returns drift in the direction of earnings surprises for several weeks after an earnings announcement.² Informational inefficiency is one of the most important factors leading to this most protracted financial anomalies documented in the literature (e.g., Bernard and Thomas, 1989, Bhushan, 1994). The informational inefficiency may result from a lack of information at the time of earnings releases that could help investors better understand a firm’s profitability and future earnings or a high cost of collecting and processing information. Consistent with the above explanations, existing studies find that the concurrent issuance of earnings forecasts by analysts (Zhang, 2008) or managers (Zhang, 2012) mitigates PEAD. By examining the impact of concurrent balance sheet disclosure on PEAD, we extend these studies by focusing on the role of non-earnings information in mitigating PEAD. As discussed above, balance sheet disclosure provides important information such as book values of total assets, total equity and total liabilities that could help investors better assess future earnings and thus the persistence of current earnings news. Voluntary disclosure at earnings announcement also helps investors to get the information with lower costs. We therefore predict that the voluntary balance sheet disclosure at earnings releases helps investors promptly understand the implication of current earnings news for future earnings and subsequently mitigates PEAD.

However, whether investors use the above mentioned balance sheet information at earnings announcement is an empirical question. Prior studies suggest that the quality of supplementary information matters in price formation. For example, Zhang (2012) investigates the impact of management forecast at earnings announcement on PEAD and finds that only high quality supplementary disclosure (i.e., accurate management forecast) helps mitigate PEAD. If investors do not find supplementary balance sheet information useful in predicting future earnings or if investors can easily calculate the balance sheet information on their own using historical data (last year’s balance sheet and this year’s earnings) with the assumption of actual clear surplus, disclosing balance sheet information at earnings announcement may add little value to the investor decisions.

We begin our analyses by examining whether the voluntary disclosure of balance sheet at earnings announcements affects market reactions to earnings surprises. This analysis shows that the earnings response coefficient in the event window is significantly higher for firm-quarters with concurrent voluntary balance sheet disclosure. It also shows that the corresponding PEAD for these firms is significantly lower. Specifically, balance sheet disclosure reduces the abnormal return in the drift window by 31 percent.

Our results are potentially subject to two types of endogeneity: self-selection and omitted variables. Self-selection implies that firms that disclose balance sheet information might be different from firms without such a disclosure, in observable or unobservable aspects. To address the potential self-selection issue, we use a Heckman (1979) two-stage approach in our main specification. We also use an instrumental variable (IV) approach as an alternative way to mitigate the self-selection concern. Furthermore, we use an entropy balancing (EB) approach to mitigate the concern that our results may be affected by an imbalance between balance sheet disclosure subsample and non-disclosure subsample. The omitted variable concern is that our results in respect to greater pricing efficiency of earnings can be explained by an omitted variable (e.g., other disclosure) correlated with the balance sheet disclosure. To address the omitted variable concern, we control for a number of other disclosures or concurrent information that we can identify around an earnings announcement, including conference calls, management forecasts, cash flow disclosure, dividend declaration, Street earnings, and analyst forecasts. To control for potential unobservable omitted variables related to firm fixed characteristics, we further include firm fixed effects in our regression model. Our results regarding the effect of balance sheet disclosure on PEAD are robust to these endogeneity tests.

To further strengthen our hypothesis regarding the effect of balance sheet disclosure on PEAD, we conduct several cross-sectional analyses. First, we examine whether the magnitude of balance sheet value surprise will affect the usefulness of balance sheet disclosure in mitigating PEAD. We conjecture that the balance sheet disclosure will be more useful when balance sheet value conveys more new information. We use the magnitude of changes in total assets, total liabilities and total equity to measure the extent to which balance sheet disclosure conveys new information (surprises). Our results show that the effect of balance sheet disclosure on PEAD is more pronounced when the balance sheet value surprises are larger.

Second, we examine the strength of the effect of balance sheet disclosure on PEAD conditional on the informativeness of balance sheet value for future earnings. The intuition is that when a firm’s balance sheet value is more informative about future earnings, investors will find balance sheet disclosure more useful in valuing the firm. Our results from this analysis show that the effect of balance sheet disclosure on PEAD is more pronounced when the balance sheet value of assets, equity or liabilities is more informative, consistent with our conjecture.

Third, we examine the strength of the effect of balance sheet disclosure conditional on earnings predictability. Our intuition is that when earnings are harder to predict, investors likely find balance sheet information more useful to help them evaluate the firm’s current and future performance. Thus, we expect that the effect of balance sheet disclosure on PEAD to be more pronounced when earnings are harder to predict. Consistent with our prediction, the findings from this analysis show that the effect of voluntary balance sheet disclosure on PEAD is more pronounced where earnings volatility or cash flow volatility is high.

Last but not the least, we examine whether the effect of balance sheet disclosure is more pronounced when the information cost is high. Even though balance sheet information can be calculated using historical data with the assumption of clear surplus relation, investors may have to bear the additional cost of data collection and analysis. Directly providing balance sheet information can reduce this cost of information collecting and processing. Thus we would expect the effect of balance sheet disclosure on PEAD is larger when the information cost is high. Investors are more likely to benefit more from the direct disclosure of balance sheet information when they are less capable of processing information or the firm’s information environment is relatively poor. The results from this analysis confirm our conjecture and show that the effect of balance sheet disclosure is more pronounced when the firm’s institutional ownership is low or when the analyst coverage is low.

In addition to examining the role of voluntary balance sheet disclosure in reducing PEAD, we examine whether mandatory disclosures in the form of 10Q/10K filings play a similar role. Specifically, we investigate whether investors correct the underpricing of earnings news in the subsequent 10Q/10K filings a few days to several weeks after the earnings announcement. 10Q/10K filings tend to be lengthier, more complex, and less predictable in their release timing compared to earnings press releases (Li, 2008, Loughran and McDonald, 2015). As a result, investors likely have much weaker incentives to gather and process information from 10Q/10K filings. Our analysis provides no evidence that the market corrects its under-reaction to earnings news around 10Q/10K filing dates for firms that did not disclose balance sheet information at the time of their earnings announcements. For these firms, we find that the stocks returns continue to exhibit a greater drift in the direction of earnings surprises several weeks after the 10Q/10K filing dates. These results are consistent with prior research’s finding that the market pays strong attention to the information in earnings announcements but little or no attention to the information in 10Q/10K filings (Stice, 1991, Easton and Zmijewski, 1993, Hollie et al., 2005, Louis et al., 2008, Li and Ramesh, 2009).³

Our study contributes to the accounting literature that examines the usefulness of balance sheet information to investors. While prior literature identifies the importance of balance sheet information such as total assets, total equity, or total liabilities for equity valuation (Ou and Penman, 1989, Penman, 1992, Penman, 1996, Ohlson, 1995, Billings, 1999, Zhang, 2000, Penman and Reggiani, 2013; among others), little is known about whether or not balance sheet information complements earnings information in equity valuation, especially in a short window. Our evidence confirms that voluntary balance sheet disclosure at the time of an earnings announcement improves the pricing efficiency of earnings news by promoting information assimilation. We also explore the mechanisms through which balance sheet disclosure mitigates PEAD. Our findings indicate that the impact of balance sheet disclosure on PEAD increases with (1) changes in key balance sheet items (assets, equity, and liabilities), (2) informativeness of balance sheet value for future earnings, (3) earnings uncertainty, and (4) information cost.

We also contribute to the PEAD literature by providing new evidence in support of the idea that PEAD is due to the market correction of its under-reaction to earnings news (e.g., Bernard and Thomas, 1989, Bernard and Thomas, 1990). Although some studies find that additional information that helps predict future earnings tends to reduce PEAD, these studies focus on the concurrent issuance of earnings forecasts by analysts (Zhang, 2008) or managers (Zhang, 2012). In contrast, we focus on the impact of concurrent non-earnings accounting information such as balance sheet information on PEAD. Our results suggest that non-earnings accounting information disclosure can also promote information assimilation and increase the information efficiency in the capital markets.⁴

Our findings also have policy implications for companies and regulators. The results in our study suggest that bundling balance sheet information with earnings information in earnings announcements increases the usefulness of those announcements to investors. Our findings lend support to FASB’s proposition that income statement data should be interpreted in conjunction with the balance sheet (FASB, 2008). Our findings thus imply that companies can improve the usefulness of their earnings announcements by adopting a policy of bundling earnings information with balance sheet information in their earnings announcements. Since disclosures in earnings announcements are voluntary in nature, regulators (e.g., the SEC) may encourage companies to make balance sheet information available to investors in their earnings announcements to enhance the informativeness of those announcements to investors.