Life insurance policyholders depend on their insurer to pay claims as promised, despite the ongoing threat of market failures or other unforeseen catastrophes. Similarly, insurers rely on key assumptions and projection techniques to set aside the amount of capital they need to meet claim demands and achieve their own performance objectives. Soon, the process and rules governing reserve setting and other projections will change from deterministic formulas that have been around for 30 years to more dynamic capital models. Milliman professionals have been tracking and participating in these changes for years. In this article, we explore the background of this issue and provide some tips for getting prepared.
click to enlarge
Historically, setting reserve requirements on life products has been a fairly straightforward, formulaic process of determining the present value of expected future benefits, less the present value of expected future premiums, based on prescribed assumptions. The key word was formulaic. This approach worked for a very long time.
That is until the paradigm for life insurance products started to shift. Companies began developing products that had significant variability due to embedded options. The result was a good deal more uncertainty as to whether appropriate reserves had been set aside to cover the benefits that would be paid out. Chief among these products were variable annuities, in which companies took on a whole new level of risk that was not captured by the prescribed reserving requirements.
click to enlarge
Meanwhile, some companies felt that the recent changes in reserve requirements on life insurance products designed to address no-lapse guarantees using a formulaic approach generated redundant reserves (no-lapse guarantees are discussed in "A Primer on Life Insurance Products"). So over the past several years, the life industry has considered changes to how it sets reserves on certain products.
Running new and unprecedented risks
New products, such as variable annuities, represent an attractive offer for customers, giving them the opportunity to invest in the equity markets while still providing protection on death, retirement, or other insurable events. On the flip side, life companies assume a significant risk of loss if the financial markets drop precipitously, unless these risks are appropriately managed and adequate reserves and capital are established. So while life companies are shifting the competitive battlefield with their new offerings, they also are venturing into areas of financial unpredictability that require a more sophisticated methodology for assessing all of the risks embedded in the new product guarantees.
Naturally, as this new paradigm began to unfold, regulators and actuaries took a closer look at what these changes meant from their respective points of view. Without the old formulas, regulators were concerned about how to set reserve requirements for products with the potential for unpredictable performance. Actuaries realized they could no longer quantify the future risks of variable products based on a single, prescribed economic environment.
Newfound sensitivity to market volatility: A coincidental driver
click to enlarge
The poor performance of the capital markets and a newfound sensitivity to this performance on the part of insurers further fueled a new approach for setting reserve requirements that would satisfy (as much as possible) the concerns of all the interested parties. Actuaries turned to stochastic economic models and assumptions based on company experience to evaluate the distribution of possible outcomes (these models are discussed in "Reserves, Required Capital, and the Role of Actuaries" and in "Nested Stochastic Modeling.") But while the use of a stochastic process and experience-based assumptions to set reserves and capital requirements made sense, it introduced a whole new dimension to the issue—judgment, which life insurance regulators view as neither easy to define nor monitor.
The industry has been discussing reserving for life and annuity products based on first principles rather than prescribed formulas for more than 20 years; hundreds of representatives from the life insurance industry, regulating bodies, and actuarial firms have weighed in on the issue. These discussions have found new traction in recent years as the industry has been unable to devise workable formulas for minimum guarantee and no-lapse guarantee products. Technology has also evolved to enable much broader, more holistic approaches. This will probably result in major changes in reserve setting on all life and annuity products.
A look at the new models
A number of key assumptions go into developing the new models, but principal among those is the economic scenario generator—the mechanism for projecting a set of possible scenarios of future equity returns and interest rates that is consistent with the current market. The American Academy of Actuaries (AAA) developed a generator from which they provided a set of 10,000 economic scenarios for companies to use in performing the C3 Phase II analysis. Companies may use a subset of the set provided or may use a different generator, as long as it can be demonstrated that the scenario set being used has the same characteristics as the set of 10,000 provided by the AAA.
The challenge for the life industry now is one of preparing for the inevitable, including understanding how the new process will work and what its implications are for key performance measures. We believe that a short list of effective preparation priorities for the near future will include the following:
- Investigate and assess company-wide hardware needs to accommodate the new modeling techniques
- Explore specially-developed software tools that support the type and number of calculations involved, not just for compliance with the new reserve and capital requirements, but to properly reflect the requirements in pricing, business planning, and risk management
- Review implications for internal planning and controls as they relate to regulatory tracking and third-party auditing of findings
- Consider additional management actions required under more dynamic modeling processes, such as risk identification and mitigation strategies, documenting and supporting judgment calls, and defining and implementing a peer review process
Let's take a closer look at how technology and management considerations underpin these priorities.
Technology needs to accommodate new modeling techniques
Two aspects of the new approach will dictate a change in both hardware and software functionality: the sheer number and size of calculations (discussed further in "Nested Stochastic Modeling") and the ability to audit, analyze, and justify results and assumptions.
Traditionally, companies had reserving systems that were "locked down," in terms of the assumptions and formulas to use. From an audit control perspective (especially post-Sarbanes-Oxley), it was easy to implement procedures around the old systems due to this controlled and prescribed methodology.
If we attempt to use locked down systems with stochastic modeling, we confront the "black box problem"—results are opaque rather than transparent. An actuary has to explain these results to management, regulators, and rating agencies—often without clear understanding of what the results mean.
So how do companies overcome the black box problem? The next generation of modeling software must allow users to drill down through and across various dimensions, to access as much detail as necessary for the actuary to understand and audit the calculations.
The cost in time and infrastructure of implementing this major change in reserve calculation presents a significant concern, particularly for smaller companies that lack the requisite human or financial resources and expertise to develop the infrastructure. It is important that software developers and consultants focus on solutions that can benefit all sizes of affected companies.
Additional management actions
As a principles-based reserves system is implemented in coming years, we expect changes to the life industry that go well beyond how the system is put to work. These changes will drive further discussions about managing and governing, particularly in assessing an organization’s appetite for risk, in establishing internal controls to monitor or contain risks, and in selecting consultants to support the company’s risk management efforts.
Turbulent financial markets and heightened regulatory scrutiny have put risk management in the spotlight, but it is shareholder interest in how well the company lives up to its intentions that will add or detract from the company's perceived value. This broader definition of risk management will influence both internal model development and the professional judgment that will be exercised in implementing the new models.
Riding the wave of the future
click to enlarge
In the end, successful handling of the new requirements will involve strength both in theory and in practice. Our firm has been involved with these changes since their inception, including the research we have undertaken to understand their implications. We have invested in developing our capabilities in both the science and the application. As a result, we’ve been able to develop a knowledge base and tools to support the implementation of the new reserve and capital requirements. Having a long history of client advisory services, we’re prepared to help our clients understand—and implement—the capital requirements of their businesses both today and into the future.
We recognize, too, that our own role as actuarial consultants is changing. Greater accountability for our "judgment calls" demands a more institutionalized approach to use of stochastic models as well as the capability to deliver a broader array of compliance skills and services. Clients will be looking for help from firms like ours who can combine knowledge of issues and needs with the tools to resolve and fulfill them. We stand ready to embrace that challenge—and that opportunity.
Pat Renzi is a principal based in Milliman’s Seattle Life Insurance Practice and is Product Manager for MG-ALFA®, the industry-leading actuarial projection and pricing software. She oversees development, planning, marketing, and client service for MG-ALFA. Pat has presented at Society of Actuaries meetings on topics directly relevant to this article, including general modeling techniques, real-time stochastic analysis, and complex liability modeling issues.