Three essays on corporate social responsibility and financial reporting
- Autores: Irina Gazizova
- Directores de la Tesis: Beatriz García Osma (dir. tes.), Juan Manuel García Lara (codir. tes.)
- Lectura: En la Universidad Carlos III de Madrid ( España ) en 2020
- Idioma: español
- Tribunal Calificador de la Tesis: Edward Lee (presid.), Josep Antoni Tribó Giné (secret.), Hans B. Christensen (voc.)
- Programa de doctorado: Programa de Doctorado en Empresa y Finanzas / Business and Finance por la Universidad Carlos III de Madrid
- Materias:
- Enlaces
- Tesis en acceso abierto en: e-Archivo
- Resumen
This Doctoral Thesis consists of three empirical papers that focus on the intersection between financial reporting quality (FRQ) and corporate social responsibility (CSR). In the light of the existing demands that firms take steps to ensure that CSR is inherent in their businesses, the questions of how to implement CSR practices and how to measure the impact of CSR on firms’ other attributes (e.g. accounting quality), remain largely unanswered. I study these issues in this Thesis.
The first paper is titled “Corporate Social Responsibility and Financial Reporting Quality: Evidence from CSR-Disasters” (co-authored with Juan M. Garcia Lara and Beatriz Garcia Osma). In this paper, we introduce the concept of CSR-disasters and argue that stakeholders become more critical of CSR performance in industries affected by the disasters. CSR-disasters are large technological disasters that can be attributed to particular firms and that trigger attention for CSR performance. CSR-disasters enable us to isolate plausibly exogenous increases in CSR performance for firms that operate in affected industries. These settings allow us to study whether socially enforced increases in CSR performance create sufficient improvements in firms’ corporate culture for having a positive effect on financial reporting quality.
We argue that CSR-disasters plausibly provide a randomly assigned increase in CSR performance and thereby help to eliminate the problem of simultaneity. In other words, this setting allows us to attribute any change in financial reporting quality to the influence of CSR. Further, observing ex-post the sign of the change (or no change) in financial reporting quality helps to distinguish between `true' and `window dressing' CSR, where, following prior literature, we define CSR as voluntary stakeholder-oriented actions that improve social conditions and that are not required by the law and extend beyond firm’s profit maximization (McWilliams and Siegel 2000; Godfrey et al. 2009; Benabou and Tirole 2010; Liang and Renneboog 2017).
In this study, we define CSR-disasters as technological disasters that (1) are sufficiently large to affect the whole disaster-affected industry; and (2) that, plausibly, could have been prevented or mitigated if a firm had gone further than just meeting the minimum formal obligations imposed by law. As a result of these CSR-disasters, same-industry firms (treated firms) (1) are exposed to a negative stakeholder reaction (negative shock); and (2) may undertake efforts to mitigate it (Blacconiere and Patten 1994; Desai 2011; Diestre and Rajagopalan 2014; Pek et al. 2018).
In our tests, we exploit a set of major technological disasters, as reported by The International Disaster Database (EM-DAT) that occurred between 2004 and 2012 in the US. We measure CSR performance using MSCI (formely KLD) data. Using a research design similar to Flammer (2015), we apply a differences-in-difference approach to estimate the effect of the disasters on CSR. More specifically, if a firm operates in an industry that is exposed to a technological disaster, we compute the difference in CSR before and after the catastrophe. Then we compare this difference with the corresponding difference in industries that are not affected by the catastrophe.
We argue that a CSR-disaster is a negative shock to the relationship between a firm and its stakeholders. Given that stakeholders' positive attitude in the form of firm’s social capital has a valuable effect on firm’s financial performance (Lev et al. 2010; Cheng et al. 2014; Shiu and Yang 2017), managers have to undertake actions to restore this relationship. Thus, we expect that firms improve CSR performance in the post-disaster period. We propose two possible mechanisms through which strengthening CSR performance may lead to strengthening the relation between a firm and its stakeholders. First, because high CSR performance can help to differentiate treated firms from the guilty firm by signaling the low operational risk and high preparation for the possible regulatory changes associated with the disaster (Hein and Wallace 2017). Second, an increase in CSR performance can signal firms’ social awareness and high environmental and social standards that do not necessarily mitigate firms’ risk. Prior literature suggests that CSR helps to build social capital and to form stakeholders’ positive attitudes, which would mitigate the negative consequences of a possible disaster (Godfrey 2005; Godfrey et al. 2009). We find strong evidence that treated firms improve CSR performance in the post-disasters periods.
Next, we address whether firms’ response to a CSR-disaster is sensitive to the pre-disaster level of CSR. Prior literature suggests that firms with previously accumulated social capital in the form of high CSR performance can mitigate negative market reactions because (1) market participants expect that these firms have lower costs associated with the disaster (Godfrey et al. 2009); and (2) because these firms have social trust and stakeholders’ loyalty (Godfrey 2005; Shiu and Yang 2017). Further, incremental increases in CSR performance may not be equally useful for firms with different pre-disaster CSR performance. For instance, Clarkson et al. (2004) show that in the pulp and paper industry only low-polluting firms (e.g. firms with high environmental CSR performance) extract incremental economic benefit from environmental expenditures. We show that only firms with low pre-disaster CSR improve CSR in the post-treatment periods.
Finally, we study whether socially enforced CSR performance leads to better FRQ. Following Kim et al. (2012) we argue that improvement in true CSR performance leads to enhancement in all corporate ethical standards. Firms that exert efforts and spend resources to achieve high social, ethical, and environmental standards may apply these standards to all their business decisions, including financial reporting. For instance, Atkins (2006) argue that being socially responsible means being transparent in firms' financial reporting. Conversely, if CSR-disasters generate entirely window dressing improvements in CSR performance, then there will be a negative or no relationship between CSR and FRQ in the post-disaster period.
Our results show that for firms with low pre-disaster level of CSR, there is a negative relationship between CSR and FRQ in the post-disaster period. These firms increase earnings management through abnormal discretionary accruals and have lower readability of their disclosure. Furthermore, the increase in CSR performance varies depending on pre-disaster levels of CSR. In line with the prior research, we argue that CSR builds goodwill in crises time, when stakeholders’ scrutiny inevitably increases.
I turn to the question of how financial analysts view CSR firms in my second paper, entitled “Do Markets Reward CSR Firms? Evidence from Target Beating Behavior” (co-authored with Antonio B. Vazquez). We study the capital markets consequences of missing analyst forecasts for CSR and non-CSR firms. Despite this increasing role of CSR, and the amply held views that CSR is associated with long-term value creation, little is known about how CSR shapes firms' incentives to meet and beat analyst consensus earnings forecast, which may be a signal of managerial short-termism and myopia (Graham et al. 2005; Bhojraj et al. 2009). The objective of this paper is to address this question.
Following prior literature, we define CSR as voluntary, stakeholder-oriented actions that (1) aim to improve social and environmental conditions, (2) are not required by law, and (3) extend beyond firm’s profit maximization (Godfrey et al. 2009; Benabou and Tirole 2010; Liang and Renneboog 2017). Building on this definition, we split firms into CSR and non-CSR, and predict that, compared with non-CSR firms, CSR firms have a longer-term horizon, engage less in target beating behavior, and as a consequence, suffer lower penalties when missing their earnings targets. Two key arguments underpin our predictions. First, CSR firms may have different investors that are less critical of short-term performance (Eccles et al. 2014; Serafeim 2015). Second, high reputation of CSR firms and their strong connection with stakeholders may help to prevent negative market expectations regarding future prospects (Godfrey et al. 2009; Lins et al. 2017; Shiu and Yang 2017) and, thus, mitigate negative price revisions if they miss an earnings target.
In line with our primary argument that CSR firms have longer horizons and attract long-term investors that are less critical of short-term goals, Eccles et al. (2014) show that CSR firms are more long-term oriented and consistently engage with stakeholders over longer windows. Serafeim (2015) argues that firms that provide integrated reporting are more likely to attract long-term investors, as these investors value information about long-term firms' prospective.
Regarding reputational motives, several empirical papers emphasize high ethical standards in CSR firms that help to improve trust with stakeholders. Prior literature suggests that corporate culture has an impact on real economic decisions including incidents of earnings management (Kim et al. 2012), and tax avoidance (Hoi et al. 2013). In their survey study, Graham et al. (2005) show that managers believe that target missing is associated with uncertainty about firms' future prospects and is a signal of previously unknown problems and, thus, leads to market penalties. If high reputation and stronger connections with stakeholders help CSR firms reduce this threat, CSR firms may receive a lower penalty when they miss their earnings targets. This assumption is consistent with the studies on the insurance role of CSR against idiosyncratic (Godfrey et al. 2009; Shiu and Yang 2017), and systematic risk (Lins et al. 2017; Albuquerque et al. 2018). This `insurance' perspective implies that high social capital maps into a stronger relationship with stakeholders, which has positive real effects in the case of bad news or crisis.
We follow the design in Bhojraj et al. (2009) and split firms that miss their analysts forecasts into firms with high and low CSR performance. Second, we identify firms that meet their earnings targets, to test whether CSR firms receive differential market rewards for target beating. CSR data comes from KLD, which has been used extensively in prior research to operationalize the CSR construct (Kim et al. 2012; Di Giuli and Kostovetsky 2014; Flammer 2015; Lins et al. 2017).
We report the following key findings. First, we find that CSR firms have lower negative price revisions proxied by 3-day returns surrounding the release of an earnings announcement. The positive effect of CSR is stronger for firms with a prevailing proportion of long-run institutional investors, measured by Bushee's classification. This result is consistent with our conjecture that CSR firm investors are less critical of short-run financial performance. To test our 'reputation' hypothesis, we examine whether CSR firms have better accounting quality, as a proxy for firms' commitment to high ethical standards in communication with their stakeholders. Consistent with Kim et al. (2012), we find that CSR firms have lower earnings management both when they miss and when they beat their targets. This suggests that firm commitment to CSR may help establish stronger, more trusting relationships with stakeholders which, in turn, pays off when CSR firms miss their target. This lower penalty could indicate lower stakeholders’ concerns regarding future prospects. However, according to Bartov et al. (2002), higher accruals quality does not guarantee higher reward for target beating to CSR firms. We turn to the target beating case next.
Regarding target beating, we do not find evidence that CSR firms receive an extra reward for exceeding analysts' earnings expectations. This is consistent with the evidence in Lins et al. (2017) of asymmetric CSR effects, where CSR pays back in the bad, but not in the good times. This may indicate that the benefits to CSR firms are constrained to limiting penalties for target missing, which is consistent with Graham et al. (2005) conjuncture that market rewards to target beating may be homogeneous across firms. Further, it draws attention to the `insurance' role of CSR (Godfrey 2005; Godfrey et al. 2009; Lins et al. 2017).
CSR reflects a firm choice, which may correlate with unobservable firm characteristics that also affect the penalty for `missers' or target beating incentives. For instance, top executives' compensation structure may affect both managers' commitment to CSR and their incentives. We address this empirical challenge by following Flammer and Kacperczyk (2016b) methodology and applying the enactment of constituency statutes as a plausibly exogenous shock for CSR. Using 2SLS approach with predicted values of CSR as an instrument, we confirm our previous findings.
Finally, the third paper “Different Similarities and Similar Differences. New Evidence on Corporate Social Responsibility” investigates heterogeneous CSR strategies in firms that have the same CSR performance and its consequences. In particular, I aim to address the following questions: Do heterogeneous CSR strategies that underpin the same CSR performance matter? What can we learn about CSR beyond CSR performance? Following prior literature, I define CSR as voluntary stakeholder-oriented actions that (1) aim to improve social conditions, (2) go beyond compliance or legal requirements, and (3) extend above solely profit maximization (McWilliams and Siegel 2000; Godfrey et al. 2009; Benabou and Tirole 2010; Liang and Renneboog 2017). For instance, firms' practices aimed at reducing carbon footprint, improving employee policies, and diversity in the workplace can be considered as CSR dimensions, which together contribute to firms' CSR performance. Building on this definition, I argue, that within the same CSR performance firms may have different CSR strategies depending on the combinations of CSR dimensions.
Prior research has already studied CSR performance and proposed to study independently positive (strengths) and negative (concerns) aspects of corporate social actions as they may reflect different underlying mechanisms (e.g., Kacperczyk (2009), Kim et al. (2014), Ioannou and Serafeim (2015), and Fernando et al. (2017)) or focus on a specific CSR dimension (e.g., Clarkson et al. (2004) and Fernando et al. (2017)). I continue this line of research by arguing that firms with the same CSR performance may have different number of strengths and concerns, which reflect various CSR strategies.
Anecdotal evidence points toward this conjecture. To illustrate, I use the example of Nike, a public US company that sells their athletic footwear and apparel virtually in all countries around the world (Nike 2019b).. Nike has long been suspected of child labor, sweatshops, and block of basic labor rights particularly in Nike's Asian factories. According to The Guardian, between 25% and 50% of the Nike's Asian factories restrict access to toilets and drinking water during the work day and deny workers at least one day off in seven (Teather 2005). Nike's culture of sexual harassment and gender bias has led to a number of gender discrimination lawsuits from their former employees (Hsu 2018). At the same time, Nike is constantly recognized as one of the most environment-friendly companies in the world. According to the Nike Impact Report from 2019, they continuously minimize their environmental footprint by using recycled materials and consuming renewable energy (Nike 2019a). According to this example, Nike has a slightly negative CSR performance, as high positive social actions are outweighed with negative ones. First, I argue that a firm's positive and negative activities as well as actions along different social dimensions are interconnected. Second, I posit that another firm with the same CSR performance but with different number of positive and negative social actions (e.g., a firm that has only one concern) has different CSR strategy.
I test my predictions using a large sample US publicly-listed firms over the period 2003 to 2013. My CSR data is obtained from MSCI (formerly known as KLD), which rates firms' CSR performance on a variety of strengths and concerns along different social dimensions. I calculate the score for each dimension by taking the difference between corresponding strengths and concerns. I measure average firms' CSR performance (CSR_Score) by adding the scores of the individual dimensions. Following the prior research, I consider firms with the same CSR_Score as firms that have similar average CSR performance. I proxy potential heterogeneity among CSR strategies by calculating summation of the strengths and concerns for firms within the same CSR_Score (net_CSR_Score). I argue that regardless of the average CSR performance, firms that are in different net_CSR_Score quantiles may have different CSR strategies.
I document the following findings. First, I find that firms with the same average CSR performance but with different number of strengths and concerns vary among different firms' characteristics. Specifically, firms that have higher total numbers of strengths and concerns (hereafter H_CSR) are older, bigger, more profitable, have less cash and more leverage, and are less research intense. Second, I show that comparing with the firms that have the same CSR performance, H_CSR firms have higher future CSR performance. Finally, I show that H_CSR firms have a negative association with future accounting performance and stock returns even controlling for CSR performance. These results are consistent with the prediction that net_CSR_Score provides additional information about firms' CSR strategies, which is complementary to CSR performance. Overall, these results provide additional insights on CSR that have not been previously captured by CSR performance. I contribute to the literature that studies how firms differ along different CSR dimensions. I show that there is a heterogeneity between firms with the same average CSR performance and this heterogeneity has an association with future CSR and financial performance.
