Three essays on conservative reporting and capital markets

• Autores: Akram Khalilov
• Directores de la Tesis: José María Marín Vigueras (dir. tes.)
• Lectura: En la Universidad Carlos III de Madrid ( España ) en 2020
• Idioma: español
• Tribunal Calificador de la Tesis: Sudipta Basu (presid.), Manuel Núñez Nickel (secret.), Miguel Duro (voc.)
• Programa de doctorado: Programa de Doctorado en Empresa y Finanzas / Business and Finance por la Universidad Carlos III de Madrid
• Materias:
  • Ciencias económicas
    • Contabilidad económica
      • Contabilidad financiera
• Enlaces
  • Tesis en acceso abierto en: e-Archivo
• Resumen

  • This thesis consists of three studies that examine the feedback effect of capital markets in general, and capital market participants in particular, on the quality of accounting information provided by firms. Specifically, I analyze the effect of conservative reporting on the dissemination of information to capital markets and whether firms benefit from such accounting practice.

    In Chapter 1, I examine the association between conservatism and insider-trading profitability from sales and purchases. We argue that firm-level accounting conservatism influences transparency and thus, the opportunities available for insiders to speculate on good and bad news. Our main focus is on conditional conservatism as it systematically affects the timeliness of good and bad news recognition. Conditional conservatism refers to the asymmetric verifiability requirements for the recognition of economic gains versus losses, which results in earnings that capture unfavourable economic events more quickly and completely than favourable events (Basu 1997), leading to asymmetric persistence of good and bad news.

    Prior literature usually regards conditional conservatism as a desirable property of accounting numbers, which results in high quality information useful to monitor management (Beekes et al. 2004; Ahmed and Duellman 2007, 2011; Louis et al. 2012; Mora and Walker 2015). We build on this literature and argue that conditional conservatism reduces insiders’ trading profitability. Two key assumptions underpin our prediction. First, that insiders can earn abnormal returns by exploiting private information (Seyhun 1986; Jagolinzer et al. 2011). Second, that conditional conservatism is positively associated with decreases in information asymmetry (LaFond and R. Watts 2008; Francis et al. 2013) and that it acts as a corporate governance mechanism that disciplines opportunistic decision-making, offsetting managerial tendency to hide bad news and accelerate good news recognition (Watts 2003).

    In particular, given that conditional conservatism leads to timely and complete bad news recognition, when conservatism is high, we expect a reduction in insiders’ opportunities to speculate on negative news. This should result in lower insiders’ profitability from sales. In contrast, the prediction on the effect over the profitability from purchases is not as straightforward. On the one hand, conditional conservatism imposes higher verification standards for the recognition of economic gains (Basu 1997). This means that gains are recognized as the associated cash flows are realized (thus, often with a lag), and could lead investors to make incorrect inferences regarding firm’s prospects. Then, conservatism would create opportunities for insiders to speculate on positive news, increasing profitability from purchases. On the other hand, prior work shows that conditional conservatism ameliorates the firm information environment, improving transparency. Conditional conservatism is associated with improvements to corporate governance, lowering the incentives for opportunistic managerial behaviour (Watts 2003; Gao 2013), and enhancing the confirmatory role of accounting, disciplining good news disclosure and increasing its credibility (Ball 2001; Garcia Osma et al. 2018). Then, conditional conservatism would act as a disciplining mechanism that leads to truthful disclosure of good news (Guay and Verrecchia 2007; LaFond and R. Watts 2008), reducing the opportunities to speculate on good news.

    We test our predictions on a large sample of U.S. firms over the period 2003 to 2014. To measure insiders’ profitability, we focus on opportunistic insiders’ sale and purchase transactions aggregated at a firm-day level, following Jagolinzer et al. (2011). We classify firms as having high (low) profitability if they earn (do not earn) abnormal returns from their transactions. To ensure the robustness of our results, we measure conservatism using two different proxies. The first one is market-based in cross-section (Basu 1997), and the second is firm-specific (Khan and Watts 2009). Both of the measures are modified following Banker et al. (2016). In addition, to address potential endogeneity issues, we run changes analyses and study the effect of an exogenous shock to conservatism: the mandatory adoption of SFAS-142. For robustness, we examine whether our results are robust to different categories of insiders (CEO and CFO, Top-5 insiders, and all other officers and directors). Finally, we study the impact of unconditional conservatism on insiders’ profitability.

    We report the following key findings. First, our results indicate timelier recognition of bad news in firms where insiders have lower profitability from sales. We find no systematic evidence of an effect over the profitability from purchases. This reduces the concerns that conditional conservatism delays the recognition of good news resulting in higher information asymmetry. In contrast, we find unconditional conservatism is associated with greater profitability from insiders’ sales. Unconditional conservatism results in an understatement of net assets that is news independent (Beaver and Ryan 2005), and may prevent the recognition of future negative news in a timely manner (Basu 2001). Our finding is consistent with the view in Ball and Shivakumar (2005) and Basu (2005) that unconditional conservatism is uninformative and largely exists to circumvent taxes and regulation. Finally, we show that the relation between conditional conservatism and insiders’ profitability is sensitive to the constraining effect of unconditional conservatism (Sunder et al. 2018). Because unconditional conservatism pre-empts the recognition of future bad news, it lowers the negative effect of conditional conservatism on insiders’ profitability. Our results are robust to the use of different measures for insider trading and conditional and unconditional conservatism, to the inclusion of additional control variables and to a battery of robustness tests.

    In Chapter 2, I corroborate evidence on price informativeness of conservative reporting and its benefits to shareholders. Importantly, prior work provides overwhelming evidence on the benefits of conditional conservatism for debt holders (e.g., Beatty et al. 2008; Zhang 2008; Li 2013; Haw et al. 2014; Franke and Muller 2019). However, there is limited evidence on the effects of conservatism for shareholders (e.g., Francis and Martin 2010; Garcia Lara et al. 2011; Biddle et al. 2016). We contribute to this work on the equity market benefits of conservatism. In particular, we study whether conditional conservatism ameliorates the information efficiency of stock prices.

    Conditional conservatism imposes more stringent verification requirements for the recognition of economic gains than losses, which results in earnings that capture unfavourable economic events more quickly and completely than favourable ones (Basu 1997). We argue that this results in two key positive outcomes that lead to more informationally efficient prices. First, conditional conservatism offsets managerial tendency to strategically reveal good news while hiding or delaying bad news disclosure (Basu 1997; Watts 2003). As a result, bad news flows into the market more quickly than unverifiable good news, reducing the risk that bad news will be hidden (LaFond and Watts 2008; Kim and Zhang 2016). Second, conditional conservatism enhances the confirmatory role of accounting. It sets a ‘hard’ benchmark against which ‘softer,’ unverifiable, disclosures can be compared ex-post. This disciplines managerial voluntary disclosure of good news, increasing its credibility (Ball 2001; Garcia Osma et al. 2018). Jointly, these effects lead to full disclosure of information (Guay and Verrecchia 2007), where no information about firm value is withheld.

    We build on this prior literature and argue that conditional conservatism improves overall market efficiency and has a positive impact on the functioning of equity markets. In particular, we study three capital market outcomes. First, we posit that conservatism helps to assess the company’s equity underlying value, reducing the probability of equity overvaluation that arises when stock price is higher than underlying value (Jensen 2005). Importantly, we expect that conditional conservatism accelerates the reversal of overvalued equity back to underlying value, limiting the duration of sustained overvaluation. Given that unsophisticated investors are more likely to value firms over-optimistically and take accounting numbers at face value, conditional conservatism, by disciplining good news disclosure and recognition, is expected to result in rational equity prices reflecting intrinsic value. Conditional conservatism is also expected to reduce equity overvaluation through better earnings quality (e.g., Chen et al. 2007; Gao 2013) and lower information asymmetry (e.g., Suijs 2008; Garcia Lara et al. 2014). Second, if conservatism decreases equity overvaluation, it should reduce short-sellers interest. In particular, through timelier recognition of losses relative to gains, conservatism should promptly signal unprofitable projects and decrease bad news hoarding (Kim and Zhang 2016), decreasing the probability that firms have hidden, unrealized losses that short-sellers could uncover and benefit from. Third, taking into consideration the benefits associated with conservative reporting, we expect that equity market participants will apply a lower discount for uncertainty when valuing more conditionally conservative firms (Guay and Verrecchia 2007), leading to higher prices and lower discounts for missing earnings targets.

    We test our predictions on a large sample of U.S. firms over the period 1991 to 2015. We report the following key findings. First, we find that conservatism reduces overall and sustained duration of equity overvaluation. High conservatism results in lower likelihood of being a firm with overvalued equity. The effect is more pronounced with the increase in the number of consecutive years of equity overvaluation. This suggests that reversals of overvaluation back to underlying value accrue faster to more conditionally conservative firms, which is supported by the survival analysis. Second, we document that conservatism reduces abnormal short interest. These results are robust to an exogenous change in conditional conservatism as a result of a regulatory change induced by SFAS 121. Finally, we provide evidence of different valuation consequences of missing earnings forecasts both in the short- and mid-run for high versus low conservative firms that miss earnings forecasts. In particular, over the 30-day (60-day) horizon equal-weighted and value-weighted BHARs of missers with high conservatism outperform missers with low conservatism by 2.31% and 1.56% (1.70% and 2.42%). Additionally, we show that in the short-run (from 1 to 30 days after earnings announcement date) and in the mid-run (1 to 60 days) the portfolio of long missers with high conservatism and short missers with low conservatism, on equal-weighted basis results in 12 (7) basis points per day.

    To corroborate our findings, we perform a number of additional tests. First, we employ two regulatory shocks affecting 1) firms’ information environment, 2) short-selling activity. We predict that firms that maintained higher level of pre-shock conservatism are expected to have ex-ante better information environment (Suijs 2008; Garcia Lara et al. 2016) and avoid bad news hoarding that is expected to mitigate the pressure from short-sellers after the SHO regulation effect on pilot firms. Our findings indicate that both regulatory shocks have a less pronounced effect on firms with higher pre-shock levels of conservatism. Second, we test whether more conservative high-tech firms had lower equity overvaluation during the Dot-Com bubble and whether they experienced lower stock price drops after the burst of the bubble. Our findings stay in line with our expectations.

    Finally, in Chapter 3, I shed the light on the benefits on keeping balance sheet numbers more conservative. Firms’ balance sheet plays a crucial role in the dissemination of business cycle shocks (e.g. Bernanke and Gertler 1989; Kiyotaki and Moore 1997) and balance-sheet asset-informativeness is beneficial for financial statement users (Chen et al. 2019). In times of crisis, investors are more likely to spot weaknesses in accounting quality compared to good times when there are more investment opportunities (Mitton 2002). This is especially relevant as market prices are affected by limited investors’ attention and processing power (e.g. Daniel et al. 2002; Hirshleifer and Teoh 2003; Hirshleifer et al. 2004). As such, investors with limited attention tend to overvalue firms with “bloated” balance sheets (Hirshleifer et al. 2004). However, preserving positive firms’ performance through biased earnings might be especially problematic when balance sheets are bloated (Barton and Simko 2002). Given this, reducing accounting bloat (i.e., having higher balance sheet conservatism) might be beneficial during periods of increased demand for reliable lower-bound estimates of net assets and increased investors’ attention. Thus, to study the importance of firms’ balance sheet conservatism (here and after, BSC), I analyze the setting of the recent financial crisis. This period is particularly interesting as it characterized by a significant credit crunch (e.g.Duchin et al. (2010) and Ivashina and Scharfstein (2010)) and overall decline in trust (Tonkiss 2009; Sapienza and Zingales 2012) that translated into increased investors’ attention (Plantin et al. 2008; Goh et al. 2015). Overall, this study aims to answer two questions. First, analyze what kind of BSC-properties helped firms to benefit during the crisis. Second, study whether firms with higher pre-crisis BSC experienced lower stock price declines and better real economic performance during the crisis.

    BSC represents a cumulative effect of past conservative accounting choices. In particular, it results from both conditional and unconditional conservatism. Unconditional (conditional) conservatism refers to persistent (i.e., news independent) understatement of net assets (timely recognition of negative information) that creates unrecognized goodwill. Jointly, this results in net assets that proxy for liquidation values (Beaver and Ryan 2005) and reduced amount of funds distributable to contracting parties. Given the aforementioned properties of BSC, I propose two main roles of BSC - informational and cushioning, that helped firms to benefit from BSC during the financial crisis.

    The informational role of BSC refers to the relative importance of hard information to market participants. Given increased vulnerability of firms’ and investors’ attention on accounting quality during the crisis, market participants are more likely to require hard information. Provided that BSC-firms transmit reliable information on lower bound esti- mates of net asset values (i.e., collateral value of the assets), BSC is expected to serve as a bonding mechanism to lenders by providing contractual information (Sunder et al. 2018). Thus, lenders would be less reluctant to provide favorable financing terms (i.e., lower cost of debt) to BSC-firms (Sunder et al. 2018). This is in line with Duffie and Lando (2001) who document that under periodic and imperfect accounting reports there is higher uncertainty of debt issuers’ net asset values that increases transparency spread and risk assessments of credit rating agencies. Therefore, maintaining more reliable lower bound estimates through BSC resolves uncertainty and reduces net assets bloat. Overall, BSC is expected to facilitate access to capital and debt renegotiation that helps firms to overcome liquidity needs and prevents bankruptcy filings (Giammarino 1989).

    On the cushioning side, BSC promotes the accumulation of “reserves” during good times that can be released in bad times (i.e., Jackson and Liu 2010). In particular, BSC creates a buffer that limits subsequent recording of losses (Roychowdhury and Watts 2007). This feature might be especially beneficial in crisis times (that is characterized by a significant fall in profits (e.g. Kim and Yi 2006), since it allows to use accumulated buffer to smooth earnings shocks, avoid covenant violations and retain higher investors’ valuation. Additionally, by providing reliable lower bound estimates of accounting numbers, BSC-firms enter unfavorable economic periods with ex-ante impaired assets and lower likelihood of equity overvaluation (Mashruwala and Mashruwala 2018). This is associated with lower probability of assets write-downs and subsequent stock price declines.

    Overall, given the informational and cushioning roles of BSC, I hypothesize that BSC- firms experienced lower stock declines during the financial crisis. The main mechanisms underpinning better performance are i) access to lower cost of debt financing, ii) lower likelihood of significant assets write-downs, iii) lower earnings volatility, and iv) lower likelihood of covenant violation.

    I test these predictions on a large sample of U.S. firms. Overall, the results suggest that BSC plays a crucial role in the transmission of negative consequences attributable to credit supply shock and a decline in overall trust during financial crises. The main model under consideration controls for firm fixed effects and captures the difference in performance before, during and after the crisis where BSC and all the controls are measured at the onset of the crisis. I estimate BSC by extracting accounting bloat from book-to-market ratio and as a robustness provide additional proxies of BSC. First, I start with the analyses of informational and cushioning roles of BSC. On the informational side, I document that firms with higher pre-crisis BSC had lower cost of debt financing during the crisis both from bank loans and from primary bond market. On the cushioning side, I document that BSC-firms had lower probability of experiencing significant assets write-downs (i.e., taking a “bath”). Additionally, I document that BSC-firms were able to reduce earnings volatility and had lower probability of covenant violation. Second, I estimate whether firms benefited from BSC during the crisis (i.e., experienced lower stock price drops and had better real economic outcomes). Regression estimates indicate that BSC firms experienced lower stock price drops. To get a better sense of the economic magnitude of pre-crisis BSC on firm performance, firms in the fourth quartile of BSC outperformed firms in the lowest quartile of BSC by 6.3 (14.1) percentage points in raw (abnormal) returns. This result is not driven by time-invariant unobservable firm characteristics. Additionally, a hedge portfolio of going long on high BSC-firms and shorting low BSC-firms earns excess returns of 112 basis points per month. Finally, I study the setting of the Great Depression and document that firms with higher balance sheet conservatism performed better within the first two years of the crisis.

    Next, I explore the causes behind the outperformance of firms with high BSC. Given that financial constraints impair firm’s ability to gain access to financing, firms are faced with a trade-off - long-term optimization (cut on investment, restructuring of employment contracts, etc.) and short-term liquidity needs. As discussed above, high BSC firms are expected to be less financially constrained and more able to raise debt during the liquidity shock of the crisis. To test this, I analyze actions taken by firms related to investment, labor practices and productivity. On the investment side, I document that low BSC-firms cut more on investment during the crisis. Further, I analyze whether high BSC-firms’ better performance is not driven by labor cost reductions. The results indicate that high BSC-firms experienced higher employment. In addition, high BSC-firms maintained higher productivity. Next, I test whether pre-crisis level of BSC translated into lower riskiness during the crisis. I report that high BSC-firms had less volatile stock returns, higher credit rating and lower distance to default.