When firms are economically linked by having more than 10% of their sales accounted for by one customer, insiders at the linked suppliers are able to realize abnormal returns by selling their firm’s stock on the basis of public information concerning the firm’s customers. The effect is strongest for top officials at the firms. Insiders at economically linked firms exhibit more routine and nonroutine trading activity than those at nonlinked firms. The difference between insiders’ nonroutine and routine trading activity, however, is greater for sales than for purchases. There is no evidence that linking increases the profitability of insider purchases.

The authors find that firm returns are a function of lagged supplier returns, indicating that analysts should consider the industry as a whole and pay attention to firm-specific information further down the supply chain. Additionally, the authors find that supplier trading activity is greater when customer information is released during periods when outside investors are likely to be less attentive, such as on a Friday afternoon. When firms are economically linked, supplier insiders exhibit more trading activity, although much of it is either routine or does not imply any information advantage compared with the trading activity of insiders at nonlinked firms. In years when the links are active, the proportion of nonroutine versus routine trading is higher, indicating that more profitable trading opportunities exist. Therefore, outside investors hoping to glean cues from insider trades might find it profitable to focus on nonroutine sales by top-level insiders at linked suppliers following customer announcements.

Overall, the authors’ results imply that if “outsider” investors are less attentive and lack the deep understanding of the relationship between customer and supplier possessed by insiders, then insiders are able to profit from their information processing advantage, leading to imperfections in the informational efficiency of the market.

The authors take a large sample of US firms with a common stock price above $5 and a market capitalization above $100 million for the period January 1986–December 2010. For each firm year, the sample is partitioned into hand-matched firms with an individual customer relationship worth more than 10% of annual sales (linked firms) and a control group of “never linked” firms and firms with a link that is inactive in that particular year. Nonroutine trades by insiders that are required to report transactions are examined together with subsets of trades by top officers or directors. Trades and stock returns are measured at the monthly level. The authors perform tests for differences in the abnormal returns of linked firms versus those of other firms in months following nonroutine trades.

In the second part of their analysis, the authors investigate whether the stock returns and the earnings surprises of principal customers predict the trading activity of supplier insiders. Tobit models are used to predict the number of shares traded based on the lagged abnormal returns of customers as well as dummy variables indicating whether lagged returns are positive (purchases) or negative (sales). Significant coefficients of the expected signs on these variables indicate that insiders trade when public information indicates that it will be profitable to do so.

Probit models are similarly used to predict the likelihood of insider supplier trades following announcements of an earnings surprise by the customer. Tobit models are used to predict the directional magnitude of transactions following an earnings surprise.

Extensive prior work has been undertaken to examine the information content of insider trades. But the authors make a nice contribution through their detailed decomposition of insider trades into routine and nonroutine for both sales and purchases in the context of economically linked firms. Their findings shed light on the information transmission mechanisms at play in firms with concentrated customer relationships. The results should stimulate debate about how the boundary between public and nonpublic information is defined. For example, when attentive insiders are in a better position than outsiders to interpret the implications of public customer information for their own firms, should the insights derived from their privileged position be defined as public or private information?