The authors study the effect of customer-imposed reputational sanctions on firms that commit fraud. They find that fraudulent firms experience a decrease in revenues and an increase in selling costs. The authors also attempt to quantify the effect of the sanctions on the fraudulent firm’s wealth. They find that a 1% reduction in the firm’s percentage of sales to a particular customer translates to a median loss of $1.23 million in net income. This loss in wealth as a result of reputational sanctions is much larger than the direct fines incurred as a result of US SEC and US Department of Justice (DOJ) investigations.

The authors’ results have two interesting implications for practitioners. First, the authors show that costs imposed by customers on fraudulent firms are real and that they are not trivial. They show that the reputational sanctions imposed by customers lower revenues and increase selling costs. Moreover, they argue that these indirect costs are much greater than fines imposed by the SEC or DOJ. For example, the authors state that the direct costs associated with government fines have a mean value of $1.89 million and the mean costs of class action settlements are $11.72 million, whereas the mean wealth effect in lost sales is $104 million and increased selling, general, and administrative expenses (SG&A) are $138 million. This difference is because of the persistent effect of lost sales and increased SG&A expense on the value of the firm that committed fraud.

Second, the authors find that a fraudulent firm’s large customers also experience a decrease in firm value as measured by a negative and significant stock return on class action filing dates and on dates that trigger the lawsuit. This result implies that the effects of the fraud are not limited only to the fraudulent firm but also affect the fraudulent firm’s large customers.

The authors begin with a sample of 2,645 class actions from 1996 to 2009 obtained from the Stanford Law School Securities Class Action Clearinghouse. They match each firm with Compustat segment-level data to identify whether a firm has a large customer. They then retrieve detailed financial information for each firm and their respective large customers.

Next, the authors identify a control group that includes firms within the same industry with similar sizes and profitability metrics. Specifically, they seek out firms with the same two-digit SIC (standard industrial classification) code as the fraudulent firms, identify whether the comparable firms also have large customers, and determine whether the comparable firms have a similar asset size. The authors then select from among the comparable firms the firms that have a return on assets closest to that of the fraudulent firms during the year prior to the class action filing. After applying these filters, they end up with a sample of 168 fraudulent firms and 168 comparable firms.

Finally, the authors test three direct measures of the intensity of customer-imposed reputational sanctions on firms that commit fraud. The first measure is the likelihood of a break-up in the business relationship between the fraudulent firm and its large customer after the detection of the fraud. The second measure is the decrease in the fraudulent firm’s sales dependency on its large customer. The third measure is the decrease in the large customer’s cost of goods sold attributed to the fraudulent firm. To strengthen their results, the authors perform several additional tests and robustness checks.

The results of the authors’ work are interesting in that they show the costs to firms and their large customers of committing fraud. Unfortunately, the incentive to commit fraud will continue to exist if the benefits of committing fraud outweigh the potential losses incurred when the fraud is exposed. It would be interesting to see a study that makes that comparison for both the firms that commit fraud and their large customers.