Business Innovation Homepage > Business Agility

If one were to attempt to identify a core of conservatism in the traditionally conservative business of life insurance, one probably would point to the assumptions built into the formulaic statutory accounting by which insurers have demonstrated their ability to cover the risks they assume.

The conservatism of those assumptions imparted a certain financial solidity to the industry, and its tendency to cause insurers to err on the side of overpricing was easily reconciled within the industry's long-standing mutual company system: When end-of-year accounting revealed that premium had been overpriced, carriers could simply send their policyholders higher dividends.

Statutory reserving -- and its one-rule-fits-all-companies application of prescribed risk assumptions -- still prevails today, but it is not long for this world of demutualization and financial services consolidation. Life insurers increasingly sell variable annuities and other products that compete with financial services institutions -- and that require different risk management approaches to establish proper reserves and ensure carriers' solvency.

These issues, along with other market pressures, are leading regulators and rating agencies to demand more-sophisticated reporting, and many insurers are taking advantage of the competitive edge provided by the more advanced stochastic modeling technologies (which generate multiple complex, random scenarios) needed to calculate the risks implied by equity market-sensitive products. But it is not only newer products that are driving a shift to a more current and company-specific approach to risk and financial management. It simply makes more sense in today's world to get a more realistic picture of an insurer's assets and liabilities by including company-specific assumptions into risk calculations.

Recognition of that fact is driving the National Association of Insurance Commissioners (NAIC; Kansas City, Mo.) to issue different requirements for demonstrating capital adequacy, and it eventually will drive the industry to principle-based reserving (PBR). An economic rather than a statutory approach already prevails in Europe and most of the English-speaking world, including Canada. And as the United States moves in the same direction, actuaries are working in closer collaboration with their companies' IT organizations and adopting more-sophisticated technology tools. >>

Sophisticated risk modeling technology is not new to the insurance industry, however. While actuaries' proclivity for the use of Excel spreadsheets is legendary, they have been using stochastic modeling for quite some time, according to Ed Robbins, the Chicago-based director of life actuarial services, SMART Business Advisory and Consulting (Devon, Pa.), and current president of the Society of Actuaries. "Actuaries have had to 'model-up' their companies' assets and liabilities since the early 1990s," he says. "What has happened is that the processes themselves have become much more sophisticated."

Since those days, carriers have used increasingly sophisticated tools predominantly developed by actuarial firms or specialist vendors, such as Towers Perrin's (Stamford, Conn.) TAS-MoSes, Milliman's (Seattle) ALPHA, GGY's (Toronto) AXIS and SS&C's (Windsor, Conn.) PTS. Newer, more robust systems are also being produced by the usual suspects, for example in the case of Towers Perrin's RiskAgility, but technology vendors such as SunGard (Wayne, Pa.) are also supplying products in response to new demands, and others, such as SAS (Cary, N.C.), are beginning to enter the financial modeling space. The increasing sophistication these solutions imply is also shaping the requirements of the regulators and rating agencies.

"Investments have become so much more sophisticated than they were five or 10 years ago. And the technology to be able to capture and analyze this information continues to improve, and we're continuing to improve our tools," comments Matthew Mosher, group VP for rating agency A.M.Best (Oldwick, N.J.). "Insurers have more-sophisticated asset/liability management tools, and we're asking for more information in that regard."

Mosher affirms that the industry is moving in the direction of economic- or principles-oriented rather than statutory-based risk management and that the raters are expecting carriers to keep up with the trend. "We do expect companies to be able to move with it," he says. "It's not as if the sophistication of the portfolios is going to decrease. [Carriers] are going to need additional data requirements for internal management, and we expect them to provide us with similar information."

Sea Change
Insurers have long been using stochastic modeling on a companywide basis for risk management purposes, but have not applied it to separate product lines -- which will be necessary if and when PBR is adopted, SMART's Robbins points out. "Getting away from prescribed assumptions for individual reserves and product line reserves will constitute a tremendous expansion of what's going to be required from a regulatory perspective," he says. "That's where the sea change is going to happen, and that's where technology will play a larger part."

The transition from universally applied statutory assumptions to entity-specific assumptions will require ramping-up the capability of insurers' staff to execute experience analyses -- that is, assessments of what ought to be assumed based on a carrier's particular experience of risk factors. "For example, you're going to have to do good expense analyses to figure your future expense assumptions; you're going to have to do lapse-rate assumptions based on your experience of mortality -- all of those assumptions will be based on your experience and the current markets," Robbins explains. That will require "getting people familiar with how to crunch these numbers to come up with these assumptions quickly."

The generation of proper dependencies between those assumptions will be a major technology issue because of the complexity of the variables they include, Robbins predicts. "You have to have a good sense not only of your experience assumptions, but also where they're headed and how they link to each other," he says. "For example, with interest-sensitive life contracts, you have a chain of the yield curve [the graphic profile of investment returns], your crediting rate and your [policy] lapse rate."

Multiplying the complexity factor is that insurers will be required to do such number-crunching repeatedly. Currently carriers do their reserves quarterly. To do so under a PBR regime would be a far greater challenge.

The bright side is that principle-based reserving is unlikely to be required before 2011 or so. Several issues are yet to be resolved, including addressing the difficulties small companies face in bringing to bear the needed resources to do PBR, and also what proportion of PBR will be tax deductible. Further, remarks Robbins, "There is still a lot of wrestling with some of the concepts right in the model regulations, and it might be premature for a company to prepare to hit what is really a moving target."

Ahead of the Curve
Tara Hansen, senior actuarial adviser in Ernst & Young's (New York) insurance and actuarial advisory services (IAAS), counters that, even without new economic capital adequacy requirements or PBR, sophisticated modeling can be used to better understand a company's business from a risk analysis standpoint. "If the timeline gets delayed, there's no harm in getting ahead of the curve, and you can get ahead of your competitors because you understand your business better," she asserts.

In addition to Robbins' caveats, Hansen sees PBR being delayed by state regulators' reluctance to relinquish their statutory reins and give insurers a free hand to determine their own assumptions. "They are not convinced that we can put an adequate review structure around that," she comments.

That hardly matters, Hansen opines, since the capital reserve requirements being put forth by the NAIC on a national basis will come sooner and may present a greater challenge. "Only business issued as of the date of adoption will be valued on the PBR methods, whereas the capital will be required on all in-force business," she explains. "So in many respects, in terms of technology, the capital is a much bigger job -- you're going to have all your business up and running on this stochastic platform, and you're going to have to look at period-to-period results based on complex stochastic runs that will require data management systems."

Data investments also will be required for the input of original data traditionally captured within isolated spreadsheets, according to Hansen's colleague, Steve Goren, senior manager, IAAS, Ernst & Young. "Many of our clients say that they are moving as far from spreadsheets as possible, to fully automated, integrated data repositories," he says.

Uncontrolled Environment
"Some companies are putting in place completely locked-down modeling environments -- in the past you would have actuaries going in and coding the models and then running them themselves," Goren relates. "It was a pretty uncontrolled environment."

On the reporting end, companies also are moving away from capturing output from models in spreadsheets, according to Goren. "Instead of spending a lot of time and integrating and aggregating the data [in the spreadsheets], they are capturing the data into structured repositories and using business intelligence tools such as Cognos [Ottawa, Ontario] and Business Objects [San Jose, Calif.] and putting some pretty robust dashboards in place."

But the most significant technology development, in Goren's view, is the move toward more stochastic modeling, along with forays into the high-performance computing initiatives that are needed to support it.

The computing requirements for advanced risk-based stochastic modeling are "changing the game," according to Chris Brown, assistant vice president, life advanced technologies, Hartford Life, the Simsbury, Conn.-based division of The Hartford Financial Services Group (Hartford; 2006 revenues of $26.5 billion). "It has basically gone beyond the abilities of spreadsheets, and because of that it is moving into an arena where there is a need to leverage enterprise grade technology -- and that means there is a need for the actuarial and technology areas to partner up," he says.

The issue today is one of scale, Brown insists. Not only are their modeling tools insufficient for new computing challenges, there is a shortage of actuaries with the right stuff. "I don't think they are finding enough highly technical actuaries to even fill the need as it is today, which is another reason to look to the technology organization to help fill those gaps," he says.

Some adjustment also will be needed on the part of actuaries in both product development and risk management who already are accustomed to working independently to execute high-speed iterations of their models, Brown asserts. But the technology side must adapt, too. "It needs to be a group that can facilitate that iterative analysis and really support the way the actuaries need to work, while at the same time bringing them into a different realm of capability that includes novel processes and systems," he says.

Grid Facilitates Modeling
Hartford Life was driven by the nature of its products and a tight reinsurance market to implement a grid computing solution based on the University of Wisconsin's Condor open-source software in 2004. "It had to do with hedging our variable annuity guarantees, and it has to do with the fact that we need to model equity market behavior instead of strictly applying a formulaic analysis," Brown says. "We needed to come up with a statistical distribution that demonstrated what could happen in the markets."

At the same time, Hartford Life was well aware of the industry's movement toward the use of stochastic modeling for valuation, Brown notes. "We realized that what we built positioned us to begin to run those sorts of analyses as well," he says. "We in fact built upon what we started with hedging and took it forward into valuation, particularly capital valuation. It took us to a better place in terms of understanding the way our different products' risks offset one another and how the entire set of risks moves as external conditions move."

Hartford Life's grid computing platform today consists of about 800 servers, each with two processors, which in turn have two cores, according to Brown. "These are used in a large computer grid shared by different product lines and different functional groups within our company, and basically exposed to be a sort of utility computing concept," he says. "As regulatory changes come down the pike and as we continue to explore how we can constantly increase the sophistication of our analysis, we're continuing to leverage the platform."

That includes use for the proliferation of life insurance products with built-in guarantees similar to those of annuities, Brown notes. "These features are attractive to consumers, and so, as a consequence, we're seeing the need to adapt to that as we develop products and do our risk management," he concludes.

Similar product-related market demands are shaping Minnesota Life's use of technology for risk management, according to Bob Reynolds, director, life product management and illustration actuary, Minnesota Life, an affiliate of St. Paul-based Securian Financial Group ($2.4 billion in 2006 revenue). "On the product development front, we continue to see a convergence between the capital and insurance markets," he says. "Risk-adjusted pricing with derivatives to support embedded options in life and annuity products certainly calls for more-sophisticated models than we've used in the past."

Reynolds reports that three drivers have pushed Securian's actuaries to utilize technology solutions where they had never been used before. "Enterprise risk management [ERM] and PBR both employ stochastic economic valuation versus the formulaic approach we currently use, and though it may seem odd, Sarbanes-Oxley has also driven us to look for technology opportunities," Reynolds says. "As we review our financial controls, we've woven in a desire to become more efficient with our reporting processes as well by reducing the stacks of spreadsheets typically used and looking for more production [software] usage."

Finding a Prophet
To support those goals, Securian searched for technology solutions beyond the traditional actuarial company software packages. "This system has more of a technical bent, and that's what we felt we needed," Reynolds relates. "We were looking for something that will help us better project our financials through the use of stochastic modeling, as well as meeting what will be demanded by PBR."

The carrier issued RFPs in the middle of last year and by fall decided to purchase SunGard's iWORKS Prophet enterprisewide actuarial modeling system. "We didn't buy it just for individual life, but for Securian as a whole," says Reynolds.

Implementation of Prophet has been ongoing since the purchase, according to Reynolds. "We're looking to have it completely in place this year, running the system in parallel with our existing other systems generating our financials," Reynolds says. "We want to start generating some economic valuation numbers on Prophet as well for each strategic business unit, and then the big deliverable this year will be to use the system in our goals and planning for 2008 and beyond, developing our cash flows on the statutory side."

In addition to those goals, Reynolds says, "We are planning to come up with what we call a 'common currency' for measuring risk, meaning that we will be able to say that a dollar of risk is the same in individual life as group life as retirement services."

All these applications of the software demonstrate Securian's commitment to ERM, according to Reynolds. "We're taking ERM very seriously, and we felt we had to purchase this system as a technology enabler to get to milestones we would have been challenged to reach otherwise."

Securian won't comment on the cost of Prophet, but Reynolds says the company is expecting a return on investment within three years. "The efficiency of everyone using the system consistently, the reuse of the modeling because we've centralized it more than ever before, and then being able to use it more deeply than any other system, precluding the manual steps we had -- all those benefits will help us recover our investment sooner," he argues.

The value the system affords Securian is, above all, the ability to master capital management, Reynolds emphasizes. "We're going to get a much clearer picture of whether we're living up to our risk appetite with the capital that we're holding and the products we're offering," he says. "That's a picture we couldn't see as clearly in the past."

That kind of picture suggests the emergence of a more-transparent world, where the true and current economic condition of a company is more easily discovered, leading to a different attitude toward questions of financial controls, in Reynolds' view. "As far as regulation and compliance go, we're more excited about what's coming at us than we have been in the past, and it's thanks in large part to technology," he remarks. "Suddenly you can move from a strictly compliance-oriented mind-set to something you can use to show value for your company."

New Thinking
Technology, when combined with the implications of economic valuation, will also shape the actuarial role within the insurance enterprise, Reynolds believes. Applying statutory formulas by rote, as it were, is a far cry from the experience analysis and assumption projection required within an economic valuation paradigm, forcing actuaries to apply more common sense and business-like thinking. "It's not as if you can simply fall back and use a formula you've learned; you actually have to come up with your own set of assumptions and project that forward to see if things make sense," he explains. "So I think technology is transforming the way actuaries think."

Technology also is changing the way actuaries work with IT, in Reynolds' opinion -- at least at Securian. "Without a marriage between actuarial and technology, I don't see our being able to get to the risk management state we have in our vision," he remarks. "In today's environment, risk management calls for a great deal of calculations of scenarios generated through stochastic modeling -- you're not going to get there unless you have technology behind you."

Click here for more Business Agility articles