Understanding Artificial Intelligence and Asset Management

Abstract blue And black connected dots

John M. Foehl | May 27, 2020 |

Abstract blue And black connected dots

Like so many things disrupted by COVID-19, one of the casualties was the Strategic Asset Alliance (SAA) Insurer Investment Forum. Longtime Captive.com readers will recognize that we have talked with Alton Cogert and Dan Smereck on a number of occasions regarding investment topics of interest to captive insurers. The Investment Forum for 2020 was to be focused on how artificial intelligence (AI) will be utilized in the asset management arena.

For captive insurance leaders who haven't been exposed to this topic, we thought a brief description of some of the presentation topics and then a question-and-answer segment with Mr. Cogert and Mr. Smereck could be extremely informative. So, buckle up as we take a shallow dive into how AI is likely to power asset management in both fixed income and equities in the near future.

While we have written about AI and its implications for captive insurers, this is the first time we have looked specifically at how it affects asset management. According to an informal SAA poll, roughly 50 percent of the traditional asset manager universe for insurers has adopted some form of AI. However, this usage can encompass a fairly extensive range of areas. Artificial Intelligence: The Next Frontier for Investment Management Firms, a 2019 Deloitte global investment management group white paper, shows the breadth of how AI can be used. The Deloitte publication lists AI use cases in investment management as follows.

"Portfolio Management and Client Enablement"

  1. "Automated insight": reading or listening to earnings reports to extract pertinent information
  2. "Relationship mapping": identifying securities and market indicators relationships that are not intuitive
  3. "Alternative datasets": building and analyzing alternative data, looking for trends that can be used in asset selection or hedging strategies
  4. "Client outreach": demand-generated analytics and reporting

Operations Efficiency

  1. "Operations intelligence": using AI to automate functions
  2. Risk performance: building algorithms and machine learning to monitor compliance with risk positions and suspicious transactions and develop response protocols
  3. "Employee insights:" monitoring employee conduct and morale

So, while investment managers may report they are using AI, it might not necessarily be in the actual selection and construction of portfolios. Captives will need to specifically ask managers for examples of how they are engaging with AI.

Had the SAA conference gone forward, attendees would have participated in case studies and presentations from, among others, AllianceBernstein ("Digital Disruption Meets Fixed Income") and Allianz ("Embracing Disruption by Artificial Intelligence in Asset Management"). Note the use of the word "disruption" in both presentation titles when talking about AI. Find more discussion on AI on the websites of these managers.

Instead of trying to infer what each presenter may have talked about from their presentations, I have chosen just to provide highlights from each. If nothing else, they should assist the reader in starting to understand how AI is likely to impact asset management. Immediately following this synopsis, I have provided a "Q&A" with Mr. Cogert and Mr. Smereck from SAA on the subject matter.

Allianz provides a very neat definition of AI for asset management. It reads as follows: "AI is nothing but pattern recognition!" Asset managers are starting to use this technology to explore and manipulate new, bigger, and more unstructured data sets. The idea is to harness this information to "sustain innovation within asset management from an alpha generation perspective." In equity investing, Allianz notes, "Current AI-related technology and methodologies might be useful to generate a signal that is true 'alpha,' but there are no risk considerations that accompany this analysis, it is purely an alpha gauging exercise." This suggests that the need still exists for human interpretation and analysis of the information AI uncovers.

AllianceBernstein, in its presentation, begins by highlighting the challenges facing fixed income managers, chiefly liquidity and scarce alpha. Liquidity is stressed due to a rapidly growing bond market coupled with shrinking dealer balance sheets and highly fragmented trading. Fixed income alpha is difficult due to low prospective returns, limited dispersion, and suppressed volatility. AllianceBernstein suggests that the traditional fixed income investment process is fundamentally broken due to the fragmentation in the bond market. It suggests "successful fixed income investing is increasingly about conquering the 'big numbers' challenge in all its forms." It has built a proprietary AI solution to overcome this issue.

Alton, you note in your opening presentation that roughly 50 percent of the asset management community is currently employing AI. How quickly do you see the remainder of the universe adopting the technology?

That was the result from a supplemental questionnaire from Q4, 2019, for our Manager Select database. Since there are many different ways to use AI and investment management is a very competitive business, I suspect we will see close to 100 percent use of AI in some form within the next 2 to 3 years, at the latest.

Do you believe that firms that do not adopt AI will be operating at a competitive disadvantage? If not, why not, or if so, why?

Ultimately, the manager will be stating it uses AI, even if the exact use may be suboptimal. Not making that statement about AI will eventually be viewed as a competitive disadvantage from a marketing appearance perspective. However, whether using AI is a true competitive advantage or disadvantage will be a function of exactly for what AI is used, how it works (lots of tools available), how important it is to the investment process, and whether the manager can lucidly explain the benefits of AI and, most importantly, how the AI determined its result (attribution). By the way, that last point is a huge issue, as AI seems like a "black box" to many.

Looking specifically at the two main asset classes that captive insurers utilize—fixed income and equities, do you think AI is likely to be of greater benefit to one asset class or useful to both?

Both, but since equities produce a wider range of potential returns, the potential impact of AI is greater there.

More specifically, how is AI utilization likely to differ between various asset classes and are there specific risk asset classes that are particularly suited to the adoption and use of AI?

Great question, and I think that is more of a "frontiers of AI" question that will be best answered in the near future by the innovators and experimenters in finance. However, one thing is certain: the global economy is becoming more digitized every day. And the more digitized the world becomes, the more important using digital tools will be to solving complex problems—even those that are heavily influenced by we less than completely rational humans.

You provide a series of questions in your forum presentation to pose to a captive insurer's asset manager concerning AI. How should a captive's board interpret those responses, and what is SAA doing to assist in that regard?

Sometimes I think it is very unfair what we expect from board members, as it seems they have to collectively be an "expert" on so many subjects impacting the captive's business. In this case, I think the board member would be well served to do some background reading on AI. It doesn't have to be technical reading but should be something that highlights what AI can and cannot do today and what it might be able to do in the future. Thus, the board member would have a starting point for interpreting answers to those questions we cited in the presentations. However, adding AI to the investment process is truly in its infancy, and SAA can assist in determining the level of AI's true advantages and disadvantages in the manager's investment process. With that in mind, perhaps the presentation should have ended with a common refrain: "caveat emptor."

(The Allianz and AllianceBernstein Insurer Investment Forum XX presentations and Mr. Cogert's presentation, "Understanding Artificial Intelligence and Your Insurer," can be viewed on the Strategic Asset Alliance website.)

John M. Foehl | May 27, 2020