In a 2016 case before the U.S. Court of Appeals for the D.C. Circuit, a former investment adviser lost a petition to review and vacate the decision of an SEC administrative law judge relating to the improper use of backtested information. The case was mainly followed by market participants who are following the challenges to the SEC's use of administrative law judges to handle disciplinary cases and other alleged violations of the federal securities laws. However, the case is also an illustration of how the improper presentation of backtested information can lead to trouble under the SEC's rules and regulations.
The Court of Appeals' decision may be found at: Raymond J. Lucia Cos. v. SEC, 832 F.3d 277 (DC Cir 1986).
The SEC's September 2015 opinion discussed by the appellate court may be found at the following link: https://www.sec.gov/litigation/opinions/2015/34-75837.pdf
At issue in the case were presentations of "backtested" historical performance for an investment strategy entitled "Buckets of Money," which, among other things, involved the shifting of assets between different investment types under various market conditions.
The SEC determined that investors who received the presentations were misled by statements that the investment strategy was "backtested." The presentations were based on a combination of historical data, together with assumptions about the rate of inflation and the rate of return on REITS, which were one type of strategy on which the strategy was based.
The presentations indicated that the strategy would have been effective during periods of market volatility, but did not indicate that the presentations were in fact showing "abstract hypotheticals", without indicating that significant assumptions were made. The presentations were also misleading in that they did not implement a key part of the strategy: moving assets from the riskiest buckets of assets to safer buckets of assets.
According to the SEC, the presentations were misleading in that they represented that the use of the investment strategy would have shown the value of investments increasing, even though they were based on flawed assumptions: underestimating the effect of inflation and the expected REIT returns, thereby dramatically departing from historical reality. The failure to move assets from one class to another resulted in the presentation figures being based on an artificially high percentage of assets in stocks during the time when stock market values were increasing. In fact, had the presentation used more realistic estimates, and shifted the assets as the investment strategy actually contemplated, the strategy would have run out of assets, as opposed to increasing in value.
The SEC did not agree with the advisors claims that the term "backtest" did not necessarily relate to a presentation that was based solely on historical data. While certain presentation slides did include cautionary language, the slides did not provide any indication to investors as to the extent to which the results that were presented in fact differed substantially from how the strategy was supposed to perform in practice and what its actual results would have been had that strategy been followed.
In this particular case, the SEC found a fairly reckless pattern of behavior in presenting the relevant information. In creating information of this kind, particularly for use in securities offering documents, most of today's market participants exercise a fairly high degree of caution and a careful degree of review. In presenting this type of information, it is critical to understand and accurately present the manner in which any departures from actual historic asset performance might be occurring or if any departures are being made from the relevant strategy or index rules. And of course, if it is actually impossible or impracticable to set forth how a strategy or index would have actually performed using actual historical values and the actual rules, it may be appropriate to consider whether that information should be used at all in selling an investment product.