Oversight Recommendations on AI and Quantitative Investment Strategies

W.With the growth of Artificial Intelligence (AI) and quantitative strategies like Smart Beta, fund directors and asset managers are increasingly being asked to oversee funds that use AI and quantitative methods to manage investments. Fund directors, in particular, need to understand the basics of these technologies in order to properly manage the funds they oversee.

ON research paper developed by Brian Bruce, CEO & CIO of Hillcrest Asset Managementprovides a framework for directors and asset managers to understand how these funds are monitored. The paper discusses the concepts of artificial intelligence and quantitative investing, with particular focus on key areas that directors should understand. The research raises important questions and offers important perspectives for these increasingly used investment approaches and instruments that can be helpful for asset managers, advisors and investors.

Hortz: How did you define and characterize quantitative investing and artificial intelligence in investment management for your research paper?

Bruce: Quantitative investment management is simply the use of computers and data in portfolio management. AI is a series of investment strategies where the machine not only does the data acquisition and processing, but also the judgment behind the decisions.

Hortz: What do you think are the most important governance priorities when monitoring these strategies?

Bruce: The main priorities for these strategies are:

Does the fund advisor have the expertise, knowledge and resources to implement the proposed strategy? Before each launch of the Fund, the Directors should determine whether the advisor has the expertise, knowledge and resources to implement the proposed strategy of the new Fund. However, it can be more difficult to gauge the advisor’s skills in relation to an AI strategy because the application is new and there are no significant track records to refer to.

How is quality assurance monitored? When adding AI funds, the consultant may need to change the approval process. The procedures are solid for current strategies, but AI is much more of a black box than most of the existing strategies. This means directors will have to learn additional information and ask various questions to ensure that the AI ​​is properly tested and implemented.

Hortz: Can you discuss with us some of your most important findings in your research that you think we need to take into account in order to monitor these strategies?

Bruce: Some important considerations would be:

  1. Quantitative funds have been around for many years. You invest based on a set of rules drawn up by the investment team. It differs from artificial intelligence funds in that there are no predetermined rules: AI looks at large amounts of data and creates its own rules.
  1. Quantitative and AI strategies involve some sort of backtesting to confirm that the strategy will work in the future. In our research report, we ask questions of the investment team to make sure the testing was done properly, as well as SEC rules on backtesting disclosure.
  1. Directors should establish a framework for assessing the AI ​​process and its structure. You should also set criteria for expected outcomes to approve AI funds. Finally, directors should determine for management what to expect from future AI efforts and how this will be communicated. It is vital that the directors have a process in place that will result in a consistent metric that will enable them to better manage these new funds.

Hortz: What are the most common problems the SEC should have with hypothetical backtesting?

Bruce: There are quite a number of problems the SEC has with hypothetical backtesting. We’ve also put together an appendix to our report with specific questions you need to ask in order to evaluate a backtest in response to these SEC concerns:

  1. Failure to Disclose Limitations. A common claim is that organizations do not fully disclose the limitations of hypothetical backtested performance (HBP).
  2. Insufficient backup data. The SEC will attempt to verify that you have maintained sufficient backup data to support your HBP claims.
  3. Time periods for the cherry harvest. Many companies are breaking SEC marketing rules by choosing a period of time during which their HBP will look better.
  4. Misleading information. Hidden or confusing HBP disclosures will create enforcement concerns for the SEC.
  5. Retrospective model changes. Organizations cannot continue tinkering with their models to improve HBP results.
  6. Use of incorrect historical market inputs. The SEC can review market data from past periods. So make sure you use the correct numbers.
  7. Apply different models. The SEC has set red flags when HBP differs significantly from audited or live performance information using the same models.
  8. Use the wrong model rules. Organizations have gone astray by applying different model rules to the back-tested data they use to manage real-world accounts.
  9. There was no investment. The SEC will be calling HBP which contains investments that were not available at the time.
  10. Bad algorithm. Incorrect programming can lead to excessive performance figures.

Hortz: What are the main risks for fund directors for you?

Bruce: I think AI funds can be managed in the same way as quant funds. AI requires understanding another set of complexities that can cause the computer to find the wrong answer. A famous example is an AI built to distinguish a dog from a wolf. A series of training images was provided and was then able to identify the difference very precisely. It wasn’t until various photos were tested that it failed. The researchers found that the AI ​​chose “wolf” when there was snow in the background because all training photos of wolves were in the snow. The boards have to be sure that the investing AI won’t find snow.

Hortz: What are your main conclusions?

Bruce: Monitoring AI requires due care and understanding of its limitations. Boards require at least one member who has experience or knowledge of AI in order to properly manage AI funds.

Hortz: What other research or additional information would you recommend in order to better understand these investment strategies or instruments?

Bruce: I particularly like the following piece from the CFA Institute because it is comprehensive. Our study is the only one that we know addresses these issues for boards.

Artificial intelligence in asset management until CFA institute:: https://www.cfainstitute.org/en/research/foundation/2020/rflr-artificial-intelligence-in-asset-management

The link to my study that we talked about was posted on the website Mays Business School’s innovation research center:: https://mays.tamu.edu/innovation-research-center/wp-content/uploads/sites/101/2020/09/Fund-Board-Oversight-of-Artificial-Intelligence-and-Quantitative-Investments.pdf

The Institute for Innovation Development is an education and business development catalyst for growth-minded financial advisors and financial services companies determined to conduct their business in an operating environment where business and cultural change is accelerating. We position our members with the necessary ongoing innovation resources and best practices to drive and nurture their growth, differentiation, and unique next-generation customer / community engagement strategies. The institute was launched with the support and foresight of our founding sponsors – NASDAQ, Ultimus Fund Solutions, Pershing, Fidelity, Voya Financial and Charter Financial Publishing (publishers of Financial Advisor and Private Wealth magazines).

The views and opinions expressed are those of the author and do not necessarily reflect those of Nasdaq, Inc.

Source link


Read Previous

Scientists Search For The Anthropocene : NPR

Read Next

REBEL Bourbon Caramel Hard Latte, Hard Irish Cream | 2021-03-17

Leave a Reply

Your email address will not be published. Required fields are marked *