Are regulatory interventions in financial markets delayed reactions to market failures, or can regulators pre-empt corporate misbehavior? Given the high economic and social costs associated with corporate scandals, and the substantial resources countries dedicate to preventing such misconduct, the answer to this question is of utmost importance.
Anecdotal evidence suggests that regulatory activity has a strong reactive component. History offers several prominent examples: The British Joint Stock Companies Act of 1844 followed widespread business failures and bankruptcies, the U.S. Securities Act of 1933 and the Securities Exchange Act of 1934 are often seen as a reaction to the Great Depression and the Kreuger Crash, and the China Securities Regulatory Commission (CSRC) was created in the wake of riots by disgruntled investors during the infamous Shenzhen 8.10 incident in 1992. More recently, the Sarbanes-Oxley Act of 2002 was passed amid corporate scandals like Enron, WorldCom, and Dynegy. According to the narrative of such a crisis theory of regulation, regulators are always less flexible and informed than private corporations, and react rather than prevent corporate misbehavior.
To test this argument empirically, we construct a time line of corporate scandals and regulation, spanning 26 countries over a period of more than 200 years. This longitudinal dataset allows us to examine regulatory effectiveness by analyzing the lead-lag relations between corporate scandals and regulation. We publish the findings in our study “Corporate Scandals and Regulation.” If regulators were effective, we would expect them to reduce the incidence of corporate scandals by ex-ante preventing such misbehavior or quickly remediating it ex-post.
We conduct the analysis in two steps. First, we collect the yearly number of instances that the terms “scandal” and “regulator” (or the local language equivalents) are mentioned in the leading newspapers in each country. For instance, in the United Kingdom, we search the archives of the Financial Times beginning in 1888, in India we use the Times of India beginning in 1838, and in France we start with Le Figaro in 1826.
The analysis of media mentions produces several insights. Scandals and regulation were discussed in the press over the entire period. Yet, they have attracted intense media attention only in recent decades. The number of media mentions starts to rise dramatically in the 1950s and the increase accelerates after 2000. The newspaper coverage of these two topics accurately traces national and global events such as the 1910 rubber stock crisis in China, the 1930 Great Depression, the 1973 worldwide oil crisis, and the bursting of the dot-com bubble in the early 2000s. Newspaper reporting on scandals and regulation is highly correlated with each other and closely mirrors the economic development in a country. These patterns suggest that the evolution of financial markets plays a key role in shaping corporate (mis)behavior and regulatory (re)action.
In our second step, we go beyond a descriptive analysis of the media mentions and conduct a detailed content analysis of the newspaper coverage. Our goal is to identify episodes of corporate scandals and regulation as they were reported in the media as far back as 1800. We build on the assumption that the media – at least in part – plays a watchdog and public awareness role, and that intense newspaper coverage creates pressure on regulators and politicians to act.
We focus on accounting regulation and scandals in our search, but do so in a broad sense. For regulation, we also consider self-regulation and other voluntary agreements. On the scandal side, we include incidents of financial fraud, embezzlement, investment schemes, tax evasion, etc. We identify more than 2,000 episodes of corporate scandals in our global sample, with scandals occurring most frequently in the United States, the United Kingdom, and Japan and least frequently in Israel and Finland. From 1800 to 1969, there are 4.2 (2.4) episodes of scandals (regulation) in any given year, but this number jumps to 32.5 (14.6) over the 1970 to 2015 period. Thus, frequent scandals and extensive regulation are a relatively recent phenomenon.
When we conduct a Granger causality test that examines whether one time series is useful in forecasting the other, we find that both scandals and regulation are highly persistent. Corporate misconduct and regulatory action are not isolated events, but come and go in waves. More to the point, we find that corporate scandals act as an antecedent to future regulatory intervention. This finding indicates that regulators are less flexible and informed than private entities and take a reactive approach to regulation. At the same time, we find no evidence that regulations can curb corporate misconduct. Rather, today’s regulations are a strong predictor of future fraudulent behavior, because firms are quick to adapt and move their activities to unregulated areas, or because regulators rely on narrowly defined rules to identify and prosecute corporate wrongdoing. We also find differences in the lead-lag relations of scandals and regulation over time and across countries.
Overall, our analysis of 200 years of corporate scandals and regulation provides evidence of strong time-series patterns. Not only do scandals lead regulatory action but the relation also goes the other direction. Corporate scandals follow past attempts at regulatory reform. This finding prompts the question of how to break this cycle of scandals followed by bouts of regulatory activism followed by new scandals while still keeping the long-term effects for the economy in mind. Even though we cannot answer this question, our results cast doubt on the effectiveness of regulatory action from a public interest point of view.
This post comes to us from Professor Luzi Hail at the University of Pennsylvania’s Wharton School, Professor Ahmed Tahoun at the London Business School, and Professor Clare Wang at Northwestern University. It is based on their recent paper, “Corporate Scandals and Regulation,” available here.