Actuarial practice and technology have undeniable synergy. With the introduction of mainframe computers, actuaries gained a tool to improve the speed and accuracy of their work along with greater access to data. Personal computers unshackled actuaries, enabling them to build models that provide greater insights into actuarial risks and the financial operations of insurance companies around the globe. As computing capabilities continue to become less expensive and more powerful, the applications and analysis of actuarial models continue to expand. Regulation and financial reporting, once founded on factor-based or formulaic approaches, migrated to principle-based (read, “model-supported”) regimes—Solvency II, International Financial Reporting Standards (IFRS) 17, principle-based reserves (PBR), Actuarial Guideline 43, and Financial Accounting Standards Board (FASB) Targeted Improvements for Long Duration Contracts (LDTI) are all examples. Economic capital, asset liability management (ALM), cash-flow testing and a wide array of other actuarial analyses all have models at their foundation.
In fact, we should take the opening sentence of this article one step further: Actuarial practice and modeling, enabled through technology, have undeniable synergy. Modeling has become the foundation of actuarial practice, but something else also happened. Modeling has changed the actuarial organization. Modeling has introduced an array of activities necessary to produce model results, many of which are not truly actuarial. Modeling has created significant dependencies outside of the actuarial department, including broadening the range of technical skills necessary to build, maintain and evolve models to provide the requisite information for today’s reporting requirements. And now, as the technology that brings together massive data, computational power and algorithms moves forward, actuarial practice will also advance, with even more processes, technologies and skills being introduced.
While the pace of technology innovation is driving change and disruption, it is not the only source of pressure for actuarial organizations and their reliance upon models. Regulatory requirements continue to expand and evolve with increased expectations for accuracy, disclosure, transparency and speed. Shareholders demand greater returns and reduced expense, even while the economic and competitive environment becomes more complex. Within an insurance company, executives rely on actuaries and models to provide timely insights regarding acquisitions, capital and reinsurance strategies, investment approaches, risk management and other key decisions. Actuarial organizations are being asked to be both more strategic and more operationally efficient during a period of rapid change. This is a real challenge.
And yet as rapidly as modeling has reshaped actuarial practice, the actuarial organization has lagged in its evolution. A wide range of nonactuarial tasks—data preparation and management; model setup, execution and monitoring; maintenance of the computational grid; assembly of results; and so on—are often being done manually by actuaries. Core modeling activities around model development, input data design, validation—activities critical to having high-quality, reliable, maintainable models—are often delegated to entry-level actuarial students. Resources with modeling talent are not retained because there is no career path for a technical modeling actuary. In short, even though models are the foundation for actuarial practice, modeling roles are not well-understood and are too often undervalued.
The evolution of modeling is demanding an evolution of the actuarial operating model.1 The interplay of data, technology and methodology has expanded the end-to-end modeling process and requisite activities, and thus has diversified the skills required to complete actuarial analyses. The use of models for financial reporting and other mission-critical applications has necessitated a rigor in deployment and use of models to recognize and enforce distinct development and production stages within the modeling life cycle. Maybe most critical, actuarial organizations also are becoming cognizant of the difference between the business aspects of the modeling process and the operational aspects of modeling.
The operational side of modeling—selecting, building, maintaining and executing the end-to-end modeling process—needs to be explicitly recognized as a critical partner to the business side of modeling, which defines the assumptions, requests the projections, and then consumes, synthesizes and acts on the model results. This partnership can enable the actuarial organization to become more strategic through better use of models and the insights they provide. It also can increase operational efficiency and the effectiveness of an organization through optimized development and execution of the end-to-end modeling process. The organization will be better positioned to take advantage of rapidly emerging technologies such as cognitive models and analytics. In short, insurers need to create a role that can be a partner to the chief actuary to take responsibility for modeling operations; develop relationships with IT and other areas outside of the actuarial function; put a focus on the end-to-end modeling process across the breadth of the company; and signal to talented modelers, engineers, data scientists and other technicians both inside and outside of the organization that their skills are recognized and valued. The time has come for the chief modeling officer to enter the stage.
The Sum is Greater Than the Parts
Understanding the business versus the operational skills and activities related to modeling holds the key to the new actuarial operating model. When these differences are recognized and each function is empowered to focus on what it does best, the resulting sum of the two parallel functions is greater than the original comingled model.
What is different between the business and operational modelers? Business modelers are responsible for the “what” questions:
- What products are most capital intensive?
- What is the impact to generally accepted accounting principles (GAAP) earnings if we consistently realize better-than-expected mortality improvements?
- What is the financial impact of choosing a given historical transition date for LDTI?
The operational modelers solve the “how” questions:
- How do we make the model run optimally to reduce cost?
- How are the assumptions structured so they are easy to maintain and use for attribution analysis?
- How do we implement changes for LDTI to be consistent across all business units while still recognizing the unique requirements of each product line?
- How will we implement machine learning to improve anomaly detection?
To be clear, this is not to say an operational modeler does not need business expertise or the business modeler does not need technical skills—clearly they do. The point is the depth of expertise differs between the two, so that the roles are complementary. The skills and expertise required to be an exceptional “business” modeler are not the same skills and expertise necessary to be an exceptional “operational” modeler.
The operational modeling organization will look very different from the business organization. The operational modeling organization will have a broad range of actuarial and nonactuarial resources. The operational modeling team will have a heavy focus on technology and tools to streamline and enhance processes. The actuaries on the team will be more technical with deeper skills in modeling, analytics and other areas. Data scientists and other advanced technical resources will work alongside actuaries, and their roles and responsibilities likely will blur. Both development and production roles will exist, with development roles focused on building new capabilities and production roles focused on execution—both periodic (e.g., quarterly) and ad hoc. The operational modeling organization will be a service organization to the business users with a focus on high-quality, timely, cost-efficient production of analytical information that meets all applicable compliance and audit requirements.
A Chief Actuary’s Perspective
Andy Rallis, FSA, MAAA, EVP and global chief actuary at MetLife, has experienced firsthand the increased focus on modeling as part of actuarial practice, and he has taken bold steps to centralize the modeling function. “The modeling organization was created to address Solvency II,” says Rallis. “With 46 countries, we had to have standard approaches—if everyone was left on their own, it would have been really messy.” …
The business organization is the driver of the modeling analysis, so business actuaries are focused on insights—enabled by the efficient modeling chassis provided by their operational modeling counterparts. During development cycles, the business actuaries define the requirements for the changes and are responsible for validating the results. During the production process, the business actuaries define the analyses required, set the assumptions and focus on analyzing the results. Enhancement requests are noted for the next development and model release cycle.
Elevating the Actuarial Function
There have been many articles written on the desire for actuaries to become more strategic, typically also referencing the “mundane” tasks that consume disproportionate amounts of time and need to be automated or eliminated. More recently, cognitive technologies and robotics have been raised as tools that actuaries can leverage in order to again become more strategic.2 The unstated issue is how to actually implement and achieve these advantages. It often seems implied that all actuaries should be educated and trained to use emerging technologies and tools. While actuaries should have the opportunity to pursue continuing education and training, it makes sense to separate the technical skills from the business expertise and perspectives within an actuarial organization to make the insights actionable.
Enabling actuaries who are focused on business applications to have technical partners rather than firsthand technical expertise is a powerful source of leverage. Especially as the modeling tools and methodologies applicable to actuarial practice become more diverse, it will equally take a diversity of resources to harness these approaches. When done effectively, the insights available to the actuaries focused on the business applications will be broader, allowing for more strategic conversations within the organization.
The stature of the modeling operations team is also elevated. With the team now centralized and owning key modeling responsibilities, the productivity of the modeling operations team becomes very visible. Modeling roles are better understood, and the impact of the team responsible for the development and delivery of modeling results becomes readily apparent. With a visible, defined organization focused on modeling operations, a career path is established for actuaries with modeling talent and interest. Like-minded individuals can be paired with peers who become mentors and internal resources to further develop modeling skills. Modeling no longer is a job that entry-level actuarial students are asked to perform until they can be promoted to a “real” actuarial role—modeling is a critical actuarial function.
Why a Chief?
Driven by changes in regulation, risk management and the competitive landscape, modeling has had an increasing role and impact within the actuarial organization. The reliance on models for quantification of insurance risks and opportunities has expanded the breadth of internal stakeholders beyond the chief actuary to include other C-suite executives including the chief risk officer, the chief financial officer and the chief executive officer.
Modeling is critical to the insurance organization, serves many stakeholders, is unique in its role in the organization and is driven by highly skilled talent. It drives competitive advantage, is necessary for compliance and is an area where cost-containment and efficiency are at a premium.
Becoming efficient and effective at modeling requires a broad array of talents and specialized skills and expertise. Modeling roles have operational components that require a continual focus on process and optimization. As technology changes, best practices will need to be continually reevaluated. The modeling process is at the intersection of data, technology, risk, compliance, finance and actuarial. It is a distinct function that originated as part of the actuarial function, but it has evolved to have unique characteristics. In many respects, it most closely aligns with the information technology (IT) organization in terms of the role it plays in the organization, the processes it must follow and the skill sets on the team. However, it is distinctly different given the analytical tools and methodologies it utilizes. Further, the talent in this analytical component is what drives the value—it is more than simply procedural. “Operations” encompasses highly technical business analytics in addition to all supporting processes around the analytics.
The trend of appointing a modeling leader has begun and is picking up steam. Modeling resources are being consolidated into centers of excellence, shared service centers and corporate modeling teams. Titles such as director of modeling and controls, vice president of corporate actuarial model services, and modeling transformation and modernization lead are becoming more common. Companies recognize that consolidating all modeling responsibilities facilitates consistency in approaches, rationalization of technologies, wider use of tools, elimination of redundancies and a wide array of other operational efficiencies.
Establishing a chief modeling officer role with the responsibility for modeling operations is a natural evolution and has several tangible benefits.
- Attract and retain talent. The role recognizes the strategic importance of the modeling function, signaling this importance both internally and externally. Recognizing the function in this way is critical for attracting and retaining modeling talent, providing a career path and internal parity with their business-focused counterparts.
- Clout to deliver. Seating a chief at the executive table alongside the chief actuary, chief financial officer, chief information officer, chief technology officer and chief risk officer provides the internal clout to both partner with and serve internal clients.
- Focus. Providing the modeling organization with senior leadership not only puts focus on the modeling process, it also enables the chief actuary to focus on truly actuarial issues and allows the IT organization to relinquish some of the actuarial modeling support burden.
- Invest for return. With the operational aspects of modeling isolated, the company can make more explicit decisions on investment in modeling capabilities and evaluate return. The company also can manage costs associated with modeling operations more effectively, with the chief modeling officer having final authority.
- Capitalize on change. Technology is changing rapidly, and the impact to financial modeling will continue to evolve. A chief modeling officer will provide the leadership and focus to navigate the changing landscape.
- Identify opportunity. With a seat at the table during executive discussions, the chief modeling officer, armed with a deep understanding of modeling capabilities, can spur ideas and offer solutions to drive new products, new markets and new opportunities through modeling- enabled insights.
- Competitive advantage. Finally, with an internal champion for modeling, the organization is positioned to move beyond cost reduction and compliance to achieving competitive advantage.
While the use of models has become more strategic, the operational environment required to support the models has become more diverse and complex. New responsibilities and activities are required for actuarial analyses, many of which are not purely actuarial. It is now increasingly recognized that the complexity and breadth of models require a unique operational modeling skill set to effectively develop, maintain and execute these models. Technology, the fundamental enabler of modeling capabilities, is advancing at an exponential pace that will put even greater emphasis and pressure on modeling operations. How should an insurer respond given the potential implications on costs, compliance and competitive advantage? Understanding the business versus the operational skills and activities related to modeling holds the key to the new actuarial operating model. Given today’s environment, the trajectory of modeling technologies and applications, and the mission-critical nature of modeling within insurance organizations, the time has come for the chief modeling officer.
- 1. Wagner, Darryl, Tony Johnson, Nate Pohle, and James Dunseth. More Than Machines. The Actuary, October/November 2018, (accessed March 5, 2019). ↩
- 2. Hughes, Mike, Ian Sterling, and Raju Saxena. Robots Join the Team—Automation, Transformation, and the Future of Actuarial Work. Contingencies, Actuarial Software Now Special Section, Winter 2018, (accessed February 28, 2019). ↩
Copyright © 2019 by the Society of Actuaries, Schaumburg, Illinois.