Companies face strong incentives to meet expectations – whether their own or those of the capital markets. A wide literature shows that this can lead to deceptive behavior by firms, which can amount to anything from overly aggressive accounting to outright fraud.
Prior research (e.g., Cohen et al. (2008)) also shows that, subsequent to the 2002 Sarbanes-Oxley Act, firms are increasingly engaging in real-activities management – that is, operating decisions taken for short-term financial benefit at the expense of long-term gain – in lieu of, or prior to, accrual-based earnings management. Yet, while we now know a fair bit about firms’ decision-making processes with respect to shady-but-legal options, we don’t know much about how firms choose to engage in misconduct – primarily because of a lack of comprehensive data on non-financial misconduct.
I address this topic in a recent paper that defines financial misconduct (FM) as anything that a securities regulator would frown upon – even if it comes to light through, say, a shareholder lawsuit. Non-financial misconduct (NFM) reflects any other type of illegal activity. Common examples of NFM include misclassifying workers to avoid paying overtime, deferring required maintenance, and illegally disposing of toxic waste. An example of NFM is BP’s 2010 Deepwater Horizon disaster. Put another way, FM occurs when a firm misrepresents its underlying performance to investors, while NFM is an operational decision taken to boost performance; non-executive employees are likely to directly feel the effects of the latter, but not the former. Are firms more likely to choose one over the other, or do firms tend to view the two types of misconduct as complementary? Does this depend on the type of NFM?
To answer this question, I exploit a new database called Violation Tracker from the non-profit organization Good Jobs First. Violation Tracker consists of all fines paid by companies to U.S. federal regulators, as well as a handful of state regulators, from 2000 onward. As of this writing, there are over 350,000 observations in the database.
My primary finding is that firms that commit FM (measured using SEC enforcement actions and significant shareholder lawsuit settlements) are more likely to concurrently engage in NFM – that is, firms seem to view the two types of misconduct as complementary. This result could reflect two possibilities with very different implications for stakeholders: (i) companies strategically choose to engage in both types of misconduct when the incentive arises, or (ii) some companies are just inherently evil. If (ii) is true, observing a specific instance of NFM doesn’t necessarily tell us much about whether the company is also committing FM. However, if (i) is true, then it is also crucial to note that NFM is usually detected much faster than FM. This is because securities violations are generally based on intent (i.e., it’s not enough to blame a firm for a sudden stock price shock; you need to show that that reflects deliberate deceptive action by the firm in order for it to be FM). However, most non-securities violations are based on outcomes (it doesn’t matter whether you strategically underinvest in maintenance or just forget; you’ve still created an unsafe work environment). To that end, while the SEC conducts around 1,000 investigations a year, agencies such as OSHA might conduct almost 40,000 a year; the OSHA investigation will also take far less time than the SEC investigation. This means that if a firm commits both FM and NFM in the same year, the NFM will almost always be detected first. Thus, there is signalling value in NFM for those interested in detecting FM.
To figure out which of these is the more likely explanation, I look at the type and timing of NFM relative to FM. If the documented link between the two types of misconduct reflects inherent corporate culture, then I should also find a statistical relation between past or future NFM and FM. If, however, the documented link between the two types of misconduct reflects strategic actions, then the link between FM and NFM in a single year should vary based on relative timing. For example, if a company is slightly short of an EPS target, it might boost reported earnings and also defer required maintenance to save on immediate costs. However, it doesn’t make sense for that company to defer maintenance next year as a response to this year’s incentives; in this case, we should observe a positive relation between concurrent FM and NFM but not between FM this year and NFM next year.
I therefore classify NFM into actions that yield short-term gains (e.g., deferring required maintenance or wage theft) vs. long-term gains (e.g., Volkswagen’s diesel-emissions scandal, which came from a desire to cut long-term production costs). When doing so, I find that firms concurrently commit FM and short-term NFM, but that FM does not precede or follow short-term NFM (i.e., firms seem to engage in short-term NFM opportunistically to supplement FM). Conversely, firms only engage in long-term NFM after they have already begun FM – perhaps because they have realized that merely cooking the books isn’t sufficient for meeting their targets. This is intuitive; it suggests that NFM reflects actions rather than culture. Firms engage in NFM when we know they have the incentive to, and seem not to engage in NFM when the incentive has already expired. Furthermore, this also suggests that the literature on real vs. accrual-based decisions may not extend to outright misconduct (i.e., shady-but-legal real actions precede shady-but-legal actions; but this is not the case when the actions are outright illegal).
To better understand what drives my results, I also look at NFM based on the specific penalizing agency. I find that firms concurrently engage in FM and wage theft (underreporting workers’ hours to avoid paying them overtime; underpayment of wages; etc.). This may reflect the immediate savings from cutting wages – the firm simply keeps the money rather than transferring it to employees – whereas other types of NFM may take more time to be reflected in financial performance. Firms also seem to commit more wage theft and worker safety-related violations (defined as anything overseen by OSHA) when they are at risk of missing an earnings target. This is consistent with prior literature (e.g., Caskey and Ozel (2017); Cohn and Wardlaw (2016)) that documents a link between short-term financial constraints and OSHA violations.
I also find that FM appears to precede EPA violations. If firms attempt to cook the books and find that to be insufficient, they may then seek other ways to cut costs; my results suggest that in these cases they are more likely to resort to illegal environmental practices.
Even though Violation Tracker is a relatively new dataset, much of its contents have long been publicly available from the federal government. Why, then, is there not much prior research on how NFM and FM are related? The most likely answer appears to be based on materiality. The penalties for wage theft and most environmental violations are quite small, largely due to laws that cap penalties assessed by agencies such as the Department of Labor. This has historically made these violations unimportant to investors, who primarily care about issues that have a meaningful and direct impact on a firm’s bottom line; a $50,000 fine for environmental violations is a drop in the bucket for a firm with a market capitalization in the billions of dollars.
However, investors do have a vested interested in financial fraud and its detection. Cases of detected financial fraud are relatively rare; my paper’s findings, however, suggest that investors seeking to predict financial fraud may benefit from incorporating other, more commonly detected, types of corporate misconduct into their prediction models. An additional result in the paper further supports this point. I compare firms in “low-wage, high-violation industries” (HVIs; as described here) against firms in other industries. Because wage violations affect hourly workers but not salaried workers, these are primarily labor-intensive industries that rely on low-skill workers. I find that the link between wage theft and financial misconduct is stronger in non-high violation industries, i.e., in industries where wage theft is more likely to reflect a deliberate action by the firm rather than simply sloppiness in the course of doing business. That is, wage violations are a better predictor of FM in industries where wage theft provides a stronger signal that something may be systematically wrong within the firm.
Overall, my findings suggest that the recent trend of investors paying closer attention to environmental, social, and governance (ESG) issues has more than just feel-good effects. In light of its signalling role with respect to financial health and trustworthiness, paying attention to ESG issues may provide direct economic benefits to investors as well.
Caskey, Judson, and N. Bugra Ozel, 2017. Earnings expectations and employee safety. Journal of Accounting and Economics 63, 121-141.
Cohn, Jonathan and Malcolm Wardlaw, 2016. Financing Constraints and Workplace Safety. Journal of Finance 71(5), 2017-2058.
Cohen, Daniel, Aiyesha Dey, and Thomas Lys, 2008. Real and Accrual-Based Earnings Management in the Pre- and Post- Sarbanes-Oxley Periods. The Accounting Review 83(3), 757-787.
This post comes to us from Professor Aneesh Raghunandan at the London School of Economics and Political Science. It is based on his recent paper, “How Are Non-Financial and Financial Misconduct Related?,” available here.