SIDC Module 19 study materials
SIDC Module 19 study resources

8 Million Shares, a Chat Group and a Buy Note

SIDC Module 19 · Put your knowledge to work in context.

All case studies

8 Million Shares, a Chat Group and a Buy Note

At 09:12 a sales trader at a broking firm receives an institutional order to sell 8 million shares of a mid-cap company - roughly four days' average volume - to be worked over three days and kept strictly confidential. At 11:40 the firm's proprietary desk sells 400,000 shares of the same counter. The stock closes down 9% that day, and the prop desk closes its position at a profit of RM 310,000. Compliance's daily review shows that the sales trader and the head of the proprietary desk belong to the same private messaging group. Two days earlier the firm's research team published a "Buy" note with a 12-month target 20% above the current price. The sales trader's chat records contain the message "big one coming, don't stand in front of it" sent at 09:20. Both staff are long-serving, and the proprietary desk's mandate permits trading the counter, which is not on any restricted list.

1. Which conduct issues arise, and what should compliance do first - before interviewing anyone?

Read the worked answer

The issues, stated plainly:

  1. Misuse of confidential client order information. The prop desk traded the same stock on the same day, hours after a large and confidential sell order arrived. If the prop desk head knew of it - and the chat group plus the 09:20 message strongly suggests knowledge - the trade is front-running the firm's own client and a misuse of confidential information, not merely an unlucky coincidence.
  2. Collusion risk between desks. A shared private messaging group between a sales trader and a proprietary trader is a structural breach of the information barrier, independent of what was said in it. Private channels outside the firm's surveillance are the mechanism.
  3. Client harm and firm reputation. The client was working a large order for a firm that traded against the flow. Whether the price impact was caused by the prop trade is a causation question, but the client has an obvious complaint.
  4. The "Buy" note's timing. Possibly a separate research-integrity issue; possibly a coincidence. Not yet a finding.

What compliance does first - order matters:

  • Preserve evidence immediately: chat and messaging archives (including the group's membership and all messages), order timestamps, trade blotters for both desks, telephone and email records, access logs, and the prop desk's positions and P&L. Do this through a formal litigation-hold style instruction to IT, with a written confirmation of what was preserved and when. Notify the head of compliance and legal.
  • Put both individuals on an administrative hold at whatever level the firm's policy requires - no personal account dealing, no access to order flow, no deletion of messages.
  • Isolate the client relationship: decide who speaks to the institutional client, and say nothing substantive until the facts are assembled.
  • Do not interview first. Interviews conducted before preservation risk witnesses reconciling stories, and messages get deleted in the hours after a question is asked.

Also check whether the firm has a reporting obligation to the exchange or regulator on the front-running suspicion, and whether the client's sell order should be paused.

2. Was the "Buy" note itself a problem? Explain the distinction between research independence and misuse of information.

Read the worked answer

The note is not automatically problematic, and the two concepts are routinely conflated.

Misuse of information asks: did the person producing the recommendation possess material, non-public information and use it, or knowingly pass it on? If the analyst wrote a balanced recommendation based on public filings, channel checks and modelling, there is no misuse - even if other people in the firm happened to know about a large sell order.

Research independence asks whether the research function can form a view free of pressure from investment banking, sales and trading, or the firm's own positions. That is a governance question, not a criminal one. It is protected by structural separation, a remuneration model not tied to deal success, supervisory review of recommendations, and rules on how research is circulated to clients.

On these facts the note raises two genuine questions:

  1. Did the analyst know anything? If the research team was told "hold the target because a big seller is coming", the note is compromised and the analyst is implicated along with the trading desks. That requires checking mail, chats and meeting notes between research and sales.
  2. Was the publication consistent with the firm's own information? A firm publishing "Buy" while its sales desk is privately distributing a large client sell order is not necessarily a breach if there is a genuine barrier - but if the desk's knowledge is shown to have reached the research or supervisory chain, then the recommendation was published in misleading circumstances.

What makes the note not a problem, in the ordinary case: analysts frequently publish recommendations that conflict with what a sales desk is doing. That is the barrier working, not failing.

So compliance should treat the note as a line of inquiry - what did the analyst know and when - and not as an automatic finding. The stronger allegation on the current facts sits with the proprietary trade, not the research.

3. Recommend control changes and identify the evidence that would distinguish front-running from coincidence.

Read the worked answer

Separating front-running from coincidence - the evidence that matters:

  • Timing and sequence. The 09:12 client order, the 09:20 message, the 11:40 prop trade. If the prop desk had not traded the stock in weeks and entered within hours of the order landing, coincidence becomes hard to maintain.
  • Content of the message. "Big one coming, don't stand in front of it" is not ambiguous in a trading context. Compliance should obtain the full thread and the group membership list before drawing a conclusion, but the immediate inference is adverse for the sales trader.
  • Historic pattern. Pull 12 months of prop desk trades and compare each against the arrival of large client orders on the same counter. A single instance is suspicious; a statistical pattern is conclusive.
  • Access. Did the prop desk head have legitimate system access to client order flow, or did the information come entirely through the chat group? The difference determines whether the failure is a technical access control or a human leak.
  • Communication to clients. Did the sales trader pass anything outward - to the prop desk, to other clients, or in his own personal account? Personal account dealing records should be reviewed.
  • Direction of the prop trade. Selling ahead of a client sell order is consistent with front-running; buying would also be inconsistent with the information but for different reasons. Either way, the information was used.

Control changes:

  • Dissolve the private messaging group and prohibit unapproved communications channels for anything client-related; add messaging to the surveillance programme.
  • Enforce physical and system separation between client-flow desks and proprietary trading, including a default of no visibility of counterparty order flow to the prop desk.
  • Automated alerting on proprietary trades in counters with a large pending client order, in any direction.
  • A pre-trade restriction: if a client order is large relative to volume, the counter goes on an internal watch list for the firm's own books until the order is complete.
  • Reinforce personal account dealing pre-clearance and make breach of it a disciplinary matter, not a training matter.
  • Escalate to the exchange or regulator if the investigation confirms front-running, and record the decision either way.