Issuers that link structured products to a new index with limited, or no, performance history typically present hypothetical historical performance, or backtested, data to show how the index would have performed prior to the inception of the index.

In this article, we present guidelines for backtested data disclosure and also discuss any guidance from U.S. regulators on the subject.

Some common sense disclosure principles:

  • Backtested data must be objective and capable of reproduction;
  • Backtested data must cover a time period long enough to include a variety of market conditions, including stressed periods;
  • Appropriate disclosure about the backtesting methodology and appropriate risk factors highlighting the difference from historical data must be included.

Objective and replicable data. Backtested data must be based on objective formulae and models applied to past historical data. The methodology used to create the backtested data should be the same methodology used to calculate the index, with appropriate disclosure as to any differences in the application. An independent third party, or a skilled investor, should be able to independently reproduce the backtested data. Necessarily, this excludes presenting backtested data for indices that are "managed"; i.e., where discretion is used in calculating the level of the index. That discretion could not be reproduced by an independent third party. Understandably, some discretion might be required to create the backtested data. The required discretion should be minimal and fully disclosed to investors.

Representative time periods. Backtested data should be presented for a time period long enough to encompass different market conditions. Including stressed periods, such as the recent financial crisis, is particularly important.

Currently, presenting a five-year period would show how an index would have performed during and after the financial crisis. However, as time goes on, a five-year period will not include the period from 2008 - 2010. Issuers should consider using longer periods, when necessary, to include periods representing different market conditions in order to show investors how the index might perform in similar future periods. Issuers should avoid "cherry picking," or presenting backtested data that shows only periods when the index would have performed optimally.

Full disclosure. Backtested data should be accompanied by additional disclosure that explains how the data was derived (including any assumptions and the impact of fees), that the data is hypothetical and not actual, was created with the benefit of hindsight and that the backtested data cannot accurately predict future results. Issuers should also include risk factor disclosures that explain that backtested data is not necessarily accurate and is not predictive of future performance. Issuers should disclose that there may be conditions (e.g., very high or low interest rate environments) for which the backtested data may not be effective. If the index has actual historical performance results, the backtested data may be presented alongside the historical data, with an explanation of the two time periods and any differences in the data.

Regulatory Guidance

As many readers know, there are no specific Securities and Exchange Commission or Financial Industry Regulatory Authority, Inc. regulations that address the use of backtested data. Accordingly, market participants look to general disclosure principles, and other statements and materials from the regulators, in order to guide their practices.

In the SEC’s April 2012 sweep letter sent to structured products issuers (question 12), the SEC asked issuers whether they use backtested data and, if so, whether the information was provided to investors and how it was presented.3 After receiving detailed responses from a variety of major issuers, the SEC did not object to the inclusion of such data. Absent specific guidance from the SEC, the key test remains whether the data is presented in a manner that is not misleading.

FINRA recently provided written guidance in the context of broker-prepared materials for exchange-listed products provided to institutional investors.4 FINRA’s approval of the use of backtested data in these materials is limited to a narrow set of circumstances. In its guidance, FINRA reiterated its historic position that the presentation of backtested data to retail investors does not comply with its disclosure standards. FINRA also warned that the backtested data should not be given excess weight in a recommendation to an investor.

It is possible that backtested data aimed solely at educating investors about product risks, together with providing a balanced perspective without over-emphasizing a product’s positive features, can be differentiated from the use of backtested data to which FINRA objects.

In private discussions, FINRA has recently reiterated its position against the use of backtested data in "retail communications," as that term is defined in FINRA Rule 2210(a)(5), relating to structured products (not just exchange-listed products). In "institutional communications," as that term is defined in FINRA Rule 2210(a)(3), relating to structured products, FINRA indicated that its application of the content standards of Rule 2210(d) to backtested data would not be applicable in the same manner as it is to retail communications. Consequently, the use of backtested data may be appropriate in institutional communications relating to structured products, in the discretion of the broker-dealer.

Conclusion

Backtested performance data may be used in issuer-prepared offering documents for structured products. The data itself should be objective and reproducible, and cover a sufficient time period so as not to be misleading. The presentation must contain appropriate disclosure as to the methodology used to create the backtested data and appropriate risk factors. Broker-dealers preparing materials relating to structured products containing backtested data should consider FINRA’s guidance, and not include such data in any retail communications.