This paper is a summary version of the winning paper of the 2013 CFA Society Toronto & Hillsdale Canadian Investment Research Award. A full version of the paper can be found on the society’s website: www.cfatoronto.ca. The paper’s authors were presented with a cash award in the amount of $10,000 at the CFA Society Toronto Award Reception in February 2013. The judges of the award found the authors’ paper to be well researched and well written and regarded the topic and conclusions as extremely interesting and relevant to Canadian capital markets. The paper’s findings also raised important questions for Canadian brokers and securities regulators to address.

The CFA Society Toronto & Hillsdale Canadian Investment Research Award is open to global researchers, including academics and practitioners, conducting research related to Canadian capital markets. Submissions are judged on the basis of their potential contribution to areas of research into any aspect of investment management, including portfolio management, asset valuation, risk management, compliance, and performance evaluation.

Our study investigates tipping by brokers who handle insider trades in Canada. Tipping occurs when a broker leaks news of an insider’s trade to his or her other clients before the insider trade is publicly disclosed. Since public disclosure of insider trades happens with a delay, the broker’s other clients have an early opportunity to mimic insider trades.¹ Insiders are normally considered to be more informed than outsiders, and many analysts use insider trading as a signal of stock value. As such, broker tipping and trading on the basis of undisclosed insider trades may undermine fairness and equitableness of Canadian equity markets.

Broker tipping made headline news in the Martha Stewart case. It was alleged that Ms. Stewart received a tip from her broker, Peter Bacanovic, that the founder of ImClone, Sam Waksal, was selling his shares ahead of news about FDA disapproval of an experimental drug. Ms. Stewart sold her shares and avoided a loss of approximately $100,000.

According to our results, Martha Stewart’s was not a singular case. We find that soon after an insider buys (sells), there is a spike in the buying (selling) activity of other clients of the brokerage firm that handles the insider trade. As shown in Figure 1, on days when insiders sell through a particular broker, this broker’s market share of agency trades increases from 5.0% to 7.0% (compared to the prior day).² On days when insiders buy, the broker’s market share increases from 5.9% to 8.3%. In Figure 2, we isolate the spike in activity not only to the day of the insider trade but to the specific half hour when the insider trade occurs. It appears that brokers are quick to share their information with others.

We also find a significant increase in the amount of trading in the broker’s own accounts on days of insider trades. This principal trading may be motivated by potential profits from the insider information. Alternatively, if the broker is also a market maker, then the principal trading may be a result of inventory management.

We further examine the cross-sectional factors that may be related to broker tipping. We draw four conclusions:

1. Reputation affects tipping. Independent brokers experience significantly larger market share increases than brokers affiliated with banks. Brokers affiliated with banks are expected to be more concerned about their reputation, given the possibility of spillover effects.
2. Brokers condition their tipping on the quality of the insider’s information. Insider trades followed by larger price changes are accompanied by more tipping.
3. Brokers tip more about trades of insiders in small companies, consistent with greater information asymmetries in such companies and higher potential profits from mimicking insider trades.
4. Tipping is significantly smaller when insiders have been actively trading in the prior 10 days. In such circumstances, information about a new insider trade might be stale.

While previous research in the U.S. has found indirect evidence of broker tipping, this is the first study to provide a direct link between an insider, his or her brokerage firm, and the brokerage firm’s other clients. This is also the first study to find evidence of tipping in Canada. This study employs a unique dataset from the Toronto Stock Exchange that identifies insider trades and brokers on each side of the trade. Data includes all TSX stocks but is limited to the period of October 2004 to December 2006.

1. Insiders have five days to report their trades through a web-based disclosure system, SEDI. Recently, the TSX began providing an end-of-day summary of insider trades.
2. We define agency trades as all trades executed by a broker, excluding insider trades and trades for the broker’s own account.