Taking Portfolio Analytics to the Next Level

By Paul Algreen, CIO, SVP, Janus Capital Group

Paul Algreen, CIO, SVP, Janus Capital Group

Great portfolio managers are commonly great storytellers. They have a finely tuned ability to use market observations, risk management expertise, personal insights, and intuition to construct a unique story or mental model that enables them to predict upcoming plot twists in the markets and for their holdings. They leverage this story in their portfolio management process and in their communication with clients. If you look at some of the most successful and iconic portfolio managers in our industry like Warren Buffett and Bill Gross, they consistently use expert storytelling to capture the attention of their clients and develop long standing relationships. As asset management firms grapple with decisions around evolving next generation portfolio analytics platforms to support their investment process, a key focus must be on supporting the development of that storyline for both internal and external consumption.

“Today, forward-thinking technology leaders are focused on the more value-added assignment of engaging the client”

Traditionally, the CXO’s role was in developing a portfolio analytics technology strategy centered on optimizing basic dimensions like cost, features, and technology adoption. However today, forward-thinking technology leaders are focused on the more value-added assignment of engaging the client. As participants in the investment management industry struggle to remain relevant amidst plunging global yields across all asset classes, a new, pressing challenge to engage the client in new and differentiated ways has emerged. It is no longer good enough to have consistent outperforming returns; we must identify new and unique methods of capturing the attention of our clients and provide them with an experience that includes a robust accounting of their investment story.

Most legacy portfolio analytics systems were architected from the bottom-up. They were designed with the analytics library and models as the central component and architected under the burden of the technology constraints at the time. Additionally, legacy systems often arose from a single asset class solution and then later modified to handle additional asset classes and add on new capabilities. Most existing portfolio analytics systems are optimized to quickly calculate a wide array of risk measurements and capture portfolio attributes but don’t overlay broader market measures or historical trends that are a key ingredient for constructing a comprehensive portfolio story. As with most aging platforms in an asset manager’s application portfolio, the limitations of the original design create persistent headwinds for change that impedes timely progress.

Combined with aging portfolio analytics technology and design, today’s financial markets present additional hurdles. As markets test and push through previously firmly held assumptions like the “zero bound” for interest rates, legacy portfolio analytics platforms require not just model updates but significant redesigns to accommodate new paradigms. To solve today’s increasingly complex and transient market correlations and dynamics, it is more important than ever for firms to identify top-down approaches to their portfolio analytics solutions that address both the new analytics requirements as well as the need to pull the portfolio risk and performance measurements into a single, unified story.

How should firms begin to think about modernizing their portfolio analytics platform to address these challenges? First, in conjunction with the investment and risk teams, a clearly defined strategy must be agreed upon for how the portfolio analytics platform will be leveraged in the construction of the overall story. Historically, it was the sole responsibility of the portfolio manager to use their expertise and observations to build that story and using technology in that process may not always be welcomed. A mutual agreement of how portfolio analytics can, and more importantly should be used in that process will form the foundation for any forward progress. Most firms have a unique set of viewpoints, paradigms, and ultimately culture that combine to constitute their “special sauce” which must be captured and institutionalized into the portfolio analytics solution.

Second, once the role of technology and a special sauce have been mutually agreed upon, a top-down portfolio analytics strategy need be devised that meets the overarching goals. Arguably, with today’s complexity and rapidly changing technology landscape, some combination of systems reuse, application acquisition, and application development will be required to fill in all the gaps. Many firms today are using a data-centric approach to guide their systems design and overall solution architectures. This approach has merit but technology leaders would be wise to decouple the hard and present day data requirements definition from the systems design process to mitigate against an excessively narrow architecture and avoid future inhibitors to change.

Fortunately, as the challenge to construct a more capable portfolio analytics platform has arisen, so too has a cadre of innovative and game-changing technologies. Previous generations of technologists were limited to two modes of solving for their needs. With the “buy” option we were forced to constrain ourselves to the predefined vendor features available and hope to influence their roadmap and then find creative ways to integrate those monolithic systems into their technology landscape. Conversely, with the “build” option, a substantial level of effort was required to lay a tremendous amount of code to construct a system capable of handling the complexity of a portfolio analytics system. Starting with a blank canvas using .NET, Java, or C++ combined with a relational database was the standard form factor that necessitated large build teams and multiple, long-cycle iterations to achieve success.

Endowed by a host of more modular and flexible technology components, technology architects now have a third option to consider. A new “compose” mode affords us a more nimble methodology for constructing our solutions in ways that meet both the foundational needs as well as the new storyline requirements. When considering how to derive a unique story from our portfolio analytics platform, technologists are now free to build a solution using the appropriate constituent components and aggregating them using new, faster, more capable aggregation techniques. The technology landscape is riddled with candidate offerings like Hadoop, Apache Spark, Riak TS, and Airflow which might be combined to build a componentized, scalable architecture for solving complex, computationally intensive portfolio analytics. Likewise, solutioners now have the benefit of micro services using RESTful APIs, concurrency, and elasticity being built into the core vendor products that form the building blocks of a portfolio analytics solution.

Clearly now is an exciting time to solve new challenges facing our portfolio analytics teams. The mandate to construct a more comprehensive, unified, and compelling story demands creativity, partnership, and a focus on not just the needs of the portfolio managers, but also the story you want your clients to hear.