Changing the Program and Re-Using the Old Track Record
A CTA has traded its 'Financials Program' for three years across 14 interest-rate and currency futures markets at roughly 2x leverage. Its disclosure document shows 11.0% annualized and a 9% maximum drawdown, computed as a composite of its six client accounts. In April the CTA backtests a narrower version - 8 markets, 3.5x leverage - showing 15% annualized and a 12% drawdown, and decides to switch. Two of the six client accounts are funded at $75,000 and never held the full 14-market list; one $4,000,000 client has been traded at 3x leverage since 2023 with the CTA's written consent. The CTA wants to mail a one-page performance table to prospects in May showing the three-year record next to the backtest.
1. Is the switch a material change, and what must happen before the CTA solicits prospects for the new version?
Read the worked answer
Yes - materially. Two of the three defining characteristics of the program change at once:
- Universe: 14 markets → 8 markets (the risk concentration rises even as the count falls)
- Leverage: 2x → 3.5x (a 75% increase in exposure per dollar of capital)
The historical record is the performance of a different program, and the backtest is the performance of no program at all.
Required before solicitation
- Amend the disclosure document and file the amendment with NFA. The description of the trading program, the markets, the leverage, and the risk factors are all material and must be rewritten - not stickered.
- Deliver the amended document to every prospect before any solicitation, and to existing clients before the change is implemented in their accounts.
- Present the three-year record only as the predecessor program, labelled as such, with the differences in market list and leverage disclosed side by side. It cannot be the headline of the new program's table.
- Label the backtest hypothetical, carry the Rule 4.41(b) cautionary statement, disclose the assumptions and that leverage of 3.5x was selected after observing the outcome, and present it entirely separately from actual results.
- Do not place the actual record and the backtest in one contiguous row with a single 'annualized' column - that construction implies continuity that does not exist.
Judgement: the CTA is not prohibited from changing its program. It is prohibited from implying that a 2x, 14-market record is a 3.5x, 8-market record.
2. Which accounts belong in a program performance table, and how must the two $75,000 accounts and the 3x-leverage account be handled?
Read the worked answer
Governing principle: performance must be presented for the accounts the CTA actually directed pursuant to the program. Selection must be by rule, applied consistently, not by how the results look.
The two $75,000 accounts
They could not hold the full 14-market list - that is a capacity constraint, not a discretion issue, and it makes them non-representative of the program as sold. The defensible treatments are:
- Exclude them and disclose it: state the number of accounts excluded and why (insufficient capital to implement the program), and disclose the aggregate performance of the excluded accounts. Silence is cherry-picking; disclosure is not.
- Or define the program so that it explicitly contemplates a reduced market list for smaller accounts, and include them as a distinct composite.
Do not simply drop them without a word.
The 3x account
Leverage of 3x against a 2x program is a materially different risk profile; pooling it into a 2x composite overstates the returns the composite investor would have received and understates the drawdowns. Either:
- present it in a separate composite for accounts traded at 3x, and disclose the leverage and that the consent was client-specific; or
- include it, but present returns and drawdowns normalized to program leverage using notional funding adjustment, with the method disclosed.
Never blend a leveraged account into an unleveraged composite and call the result 'program performance'.
Also required: disclose that the composite is asset-weighted or equal-weighted, how accounts enter and leave, and whether fees are gross or net of the CTA's fee. A composite whose weighting method is undisclosed cannot be verified and should not be published.
3. The CTA proposes to email the one-page table to 900 prospects on its list before the amended disclosure document is ready, 'so the pipeline doesn't go cold'. Evaluate.
Read the worked answer
This is the one clear-cut violation in the fact pattern.
Solicitation of a CTA program requires delivery of a current disclosure document before the prospect is asked to invest. Marketing the new program before the document is amended and filed means every one of the 900 emails is an offer of an undisclosed program - and the document on file describes a 2x, 14-market program that no longer exists.
Specific consequences
- Solicitation using a materially stale disclosure document.
- Performance for a program that is not the one being offered.
- Hypothetical results distributed without the required cautionary statement if the backtest column is included.
- Every recipient becomes evidence of the violation: 900 dated emails are not defensible as an oversight.
What to do instead
- Hold the campaign until the amended document is filed and ready to send with it. This is days or weeks, not months.
- If the pipeline genuinely cannot wait, send a generic, non-performance communication that makes no reference to the new program, no return figures, and no invitation to invest - and send the compliant package when it is ready.
- Notify the existing clients now, in writing, of the change in markets and leverage and obtain their acknowledgement; they are exposed to 3.5x whether or not the marketing is ready.
Judgement: the pressure to keep the pipeline warm is exactly the pressure that produces most advertising violations. The correct answer is always to move the document faster, not to move the marketing earlier.