Calculating scenarios: new risk-based capital requirements are changing how variable annuity writers understand risks.

Key points

* The living-benefit guarantees of variable annuities pose investment-market risks for direct writers.

* Current risk-based capital requirements do not address these risks.

* Stochastic modeling helps insurers better understand and hedge for the risks.

* Regulations on the drawing board address these techniques.

Variable annuity writers have taken much risk out of investing for their policyholders. Now regulators want to ensure that writers are reserving enough for the new risks they have assumed.

The newest optional feature, for example, guarantees that buyers can withdraw all of the money over a stipulated minimum number of years, regardless of how well their investment options perform. That limits their risks in volatile equity markets.

Other features are similar. One guarantees a minimum policy account value at a given point in time. Another guarantees that regardless of the contract's investment performance, the minimum amount available for annuitization will be the amount invested or, more often, the invested amount growing at an annual rate of 3% to 6% or the contract performance locked in annually, whichever is greater.

The trouble is that these guaranteed-minimum features--withdrawal benefits accumulation benefits and income benefits--don't correlate well with risk-based capital regulations developed in the early 1990s. As the reinsurance market dried up during the bear market of March 2000 to March 2003, the insurers writing these "living benefits" turned to stochastic scenario analysis and hedging to manage their new risks and determine appropriate reserve levels. Now the American Academy of Actuaries is proposing risk-based-capital standards using these techniques. If the National Association of Insurance Commissioners adopts a model regulation this month, new standards would be in effect nationwide by the end of next year, according to Hubert Mueller, a consultant with the Tillinghast business of Towers Perrin in the firm's Hartford, Conn., office, and leader of Tillinghast's financial management practice.

It's fair to say that everyone realized that the existing risk-based capital model in place does not pay proper attention to the risks of the more aggressive guarantees," said Mueller, a member of the Life Capital Adequacy Subcommittee set up by the academy "The current model was put in place in the late 1980s and early 1990s, when annual variable-annuity sales were in the $10 billion to $20 billion range as compared to $120 billion-plus today. The old model was appropriate for the requirements at the time, but is now outdated."

Variable annuity writers were pushed into scenario-testing of their assets and liabilities as competition drove them to develop or enhance the living-benefit guarantees of their products. "Self-insurance works when the risks are moderate and the percentage of total business is small," said Mueller. "Reinsurance works when reinsurers are in the marketplace. What happened was that reinsurers had losses because they didn't price risks properly, so they retreated from the marketplace."

Market share slips when variable annuity writers don't offer living benefits. So insurers implemented hedging strategies to stay competitive. The new, more-aggressive guarantees could jeopardize a company's long-term capital position, a concern for mutuals as well as publicly traded insurers. The academy has been instrumental in promoting the importance of stochastic models and hedging, Mueller said.

An Active Hedger

Hartford Life Insurance Co. was probably first to develop the new risk-management tools as a result of being the first company to offer a guaranteed minimum withdrawal benefit in October 2002. The company had reinsurance at the time, but sales of the new benefit "were beyond their wildest expectations," quickly exhausting the reinsurance coverage and forcing implementation of a hedging strategy, said Mueller. Hartford succeeded in convincing analysts that its strategy was sound, raising the bar for the whole industry, he said. "Nobody wanted to say they weren't doing anything, and everyone is now scrambling to put some kind of hedging into place."

Hartford Life began hedging in September 2003 due to a concern about tail risk and reserve requirements. "We hedge the equity risk," said Tom Campbell, vice president and corporate actuary. "In extreme scenarios, losses are much greater risk than we're comfortable with, and Hartford has always been a conservative manager of risk."

"The academy has taken the position that the best way to measure the risks and determine reserves and capital is to use a model-based rather than a formulaic methodology," said Campbell, also a member of the academy. "We think this method really aligns reserves and capital with risk management."

Hartford still has reinsurance for its guaranteed minimum death benefit. For its Principal First product, which offers a withdrawal benefit, new business is supported by hedging, while reinsurance covers the in-force block, Campbell said. Product design and benefit pricing are also part of Hartford's risk-management strategy, he said.

The death benefit, while still subject to equity-market risk, is easier to reserve because the policyholder must die to trigger the benefit. But scenario modeling and hedging of living benefits can be very complicated. Due to the size of its business Hartford has 50 or 60 computers running day and night to help with the process. "In order to measure the risks, you have to generate more computations," Campbell said. "You can't just look at the block as a whole. Each contract has a different level of risk, and we break those risks into manageable cells with different risk characteristics."

Hedging now helps Hartford develop products. "You move with what you can sell and people are willing to buy, but in doing the analyses, you better understand the products and risks, and that helps facilitate the generation of ideas," he said.

Without hedging, Campbell speculated that Hartford would have a smaller product lineup or would find other risk-management techniques. But he agreed with Mueller that scenario modeling and hedging represent the beginnings of "a new era in risk-based capital requirements."

Scenario modeling and hedging also may fundamentally change the relationship between direct writers and reinsurers. "We can share our techniques with reinsurers," said Campbell. "If the hedging and the analysis that supports the hedging reduces our tail risk, that could open up the reinsurance markets."

Benefits Must Outweigh Costs

Analysis of tail risk and hedging costs insurers more than the formulaic methods implemented in the past, but Mueller estimates that an effective hedging program can reduce by about half the amount a writer must hold in capital for the variable-annuity line of business. Whether these guarantees become too expensive depends on the product design and overall risk structure, Campbell said. "If you're analyzing your business properly, you have controls in place to keep that from happening," he said. "If the hedging costs are going up, you have to charge more for the benefits, and that depends on product designs and risk-management techniques."

Rating agencies such as A.M. Best Co. are interested in how companies determine their reserves. "The risks inherent in the living benefits are somewhat different than more traditional life and annuity products, where the primary risks are mortality risk and interest-rate risk," said Edward Easop, assistant vice president, Life/Health Division.

"We believe that effectively managing these products requires an integrated risk-management approach that starts with product design and understanding your target market." Easop said this constitutes "front-end" risk management. Effective management then progresses through risk-mitigation techniques such as hedging, a "back-end" technique, "and then comes back full circle to product design by applying what a company learns about how the products behave to design improved products in the future." A.M. Best published a special report in September, "Variable Annuity Risks: Secondary Guarantees," authored by Easop and Dave Tauber, financial analyst.

Overall, the industry probably welcomes the development of a more economic risk-based capital approach using stochastic modeling instead of the current factor-based approach, Easop said. A.M. Best's view of the proposed requirements is that they are "a step in the right direction to integrate more advanced tools and methods in the determination of required capital," he said. "However, with the use of these models comes increased modeling risk as companies try to develop assumptions related to market movements and policyholder behavior in an increasingly volatile economic environment. As a result, we believe companies may be exposed to low-probability, high-cost events even if the stochastic scenario testing is done thoughtfully and carefully."

Critical Actions

In a November 2003 survey of
chief financial officers of variable
annuity writers, Tillinghast found
that in the next six to 12 months
most planned to revise life and
annuity product designs and
further enhance guarantees.

Revising product design/guarantees         84
Adopting more extensive scenario testing   63
Changing product pricing                   53
Limiting sales                             32
Increasing required capital                11

Source: Tillinghast CFO Survey #6
(November 2003)

Note: Table made from bar graph.

Learn More

Hartford Life Insurance Co. A.M. Best Company # 06518 Distribution: Independent agents, brokers, banks and wirehousos

Ace Tempest Life Reinsurance Ltd. A.M. Best Company # 77030 Distribution: Mostly through reinsurance brokers

For ratings and other financial strength information about these companies, visit www.ambest.com.

RELATED ARTICLE: How living-benefits hedging works.

Proposed risk-based capital requirements for variable products with guarantees fall under the heading of RBC C-3 Phase II. The proposal covers products with equity-based risks.

The American Academy of Actuaries presented its C-3 Phase II RBC recommendation in December 2002 and submitted follow-up reports in September and December 2003, according to an April 2004 Tillinghast Update report. Life insurers offering variable annuities with guaranteed minimum living benefits would be required to use stochastic scenario analysis, while those offering only guaranteed death benefits could choose either an alternative, factor-based approach or the stochastic analysis. Tillinghast is the insurance business of Towers Perrin.

Under current factor-based RBC regulations, category C-1 applies to asset risk; C-2, pricing risk and C-4, business risk. The C-3 category applies to asset-liability mismatch, or "tracking" risk. C-3 Phase I applies to fixed annuities and single-premium life insurance. "C" stands for contingency.

The top 10 variable annuity writers shared 70% of the market in June 2004, up from 60% in 2002, according to Hubert Mueller, a Towers Perrin principal and one of the authors of the report. Almost all of them had introduced or updated their living-benefit riders in the previous 12 to 18 months. Of the top 25 variable-annuity products sold in 2003, 14 had a minimum income benefit, 12 had a minimum withdrawal benefit and five had a minimum accumulation benefit.

Examining effective hedging techniques, Tillinghast reported in its May 2004 Update that when performing stochastic modeling, most companies writing variable annuities with guarantees generate at least 1,000 different scenarios. Direct writers can expect to reduce the capital they must hold in reserve by using this modeling and a proper hedging program, the report said.

Hedging involves buying options or futures to offset the risks posed by product liabilities. If insurers do not rebalance, they are engaged in static hedging. Dynamic hedging requires frequent rebalancing.

The "Greeks" represent measures of change in options, according to the May report. Delta is a measure of how much an option price changes given a unit change in the underlying stock/futures price. Gamma is a measure of how fast Delta changes given the same unit change. Rho measures the rate of change in an option price for a 1% change in discount (interest) rates. Vega is a measure of the rate of change in an option price for a 1% increase in volatility of the underlying asset. Theta measures the rate of change in an option price with respect to the passage of time.

The simplest, most common dynamic hedging, which limits exposure to market movements, is achieved by offsetting the Delta on variable annuities with options or futures on publicly traded indices, Tillinghast said. However, the effectiveness of Delta-only hedging program is limited.

A more sophisticated and costly dynamic hedging program seeks to mitigate volatility risk by also matching the Vega of a block of business by buying options, usually exchange-traded options. This is needed because the options bought for Delta hedging have durations shorter than the duration of the underlying variable-annuity guarantee liabilities.

Tillinghast also recommends matching Rho. Since liability options tend to be very long, interest rates fluctuation could dramatically change the option's value. The hedging of Rho is typically achieved using interest-rate futures.

Tillinghast reported that in a year-end survey of 20 large variable annuity writers, only two were hedging dynamically beyond Delta. Eight were Delta hedging, one was static hedging, and nine had no hedging program in place. "Going beyond Delta hedging is expected to increase the hedge effectiveness, but will also increase the cost of the hedging program," the report said. "This is offset by the lower cost of capital resulting from the reduction in capital that would be expected to arise through an improved hedging program."

Those companies hedging beyond Delta use as many as 30 to 60 computers running overnight to "run their entire in-force business on a daily basis," but they typically buy options only on a weekly basis to reduce the trading costs of hedging," the report said.

The report mentions calculating the Greeks directly "by shocking the underlying indices using a brute force method." Mueller explained that brute force means running a lot of scenarios. (Hartford runs about 60,000 every night, he said.) "Then you shock your indices 1% up or down and then run the same scenario set to help calculate the Greeks," he said.

RELATED ARTICLE: Market demand breeds new reinsurers.

With increasing demand for reinsurance on variable annuities' guaranteed living benefits, three or four new companies are planning to enter the market, according to Hubert Mueller, a consultant with the Tillinghast business of Towers Perrin. As of late October, some were a month or two away, while others were six to 12 months from startup.

"These new entrants are not traditional reinsurers, not like the Swiss Re's or Munich Re's, but rather are companies backed by institutional investors or investment banks," he said. Two Mueller could name were Lennox Re and Catalyst Re, both of which, if formed, would be based in Bermuda.

As the availability of reinsurance for living benefits dried up in recent years, direct writers have turned to stochastic modeling and hedging to address their risks and to keep the required risk-based capital reserves from becoming onerous. But direct writers still reinsure parts of their businesses and would be eager to buy reinsurance if it becomes available at a reasonable price, said Mueller. "Product design, reinsurance and hedging of risks are the key ingredients of a state-of-the-art risk management program," he said.

Phil Bieluch, an independent consultant in Avon, Conn., is involved in discussions about setting up a Bermuda reinsurer that would use modeling and hedging to manage its risks. "For reinsurers as well as for direct writers, the new risk-based capital regulations provide a framework that will be used by rating agencies to determine capital adequacy," he said. "Given the complexity of the hedging strategies and the effect they will have on the balance sheet, reinsurance will be a preferred alternative, if available."

Direct writers are likely to prefer reinsurance over their own hedging, Bieluch said, noting "Due to GAAP accounting, hedging may cause asset losses when the economics dictate that less hedging is needed. As the market goes up, you need less hedge on variable-annuity guarantees, but you lose money on your futures portfolio."

For example, Cigna Corp. posted a pretax loss of $312 million for the second quarter of 2003 from the futures and forward contracts it held for reinsuring the guaranteed minimum death benefits of other companies, according to the company's 10-Q statement for the quarter ending June 30, 2004. Stock prices rose rapidly during that quarter. But the losses were only- $29 million for the second quarter of 2004, a time the stock market was tepid. Of course, rising stock prices reduce amounts at risk and claims.

The face amount of the futures and forward contract positions held by Cigna at June 30, 2004, was $1.7 billion. Cigna discontinued its reinsurance operations in 2000; they are now an inactive business in run-off mode, according to the 10-Q.

Bermuda-based Ace Tempest Life Reinsurance Ltd. is the only major reinsurer that still offers coverage of living-benefit guarantees. The company has a healthy portfolio, makes a lot of different transactions and reinsures a number of top writers, said Ari Lindner, senior vice president and chief life officer. "But obviously, the VA market is more than $120 billion a year, and we can't cover it all," he said.

Ace Tempest is able to offer the living-benefits reinsurance because it has a small and highly experienced team of experts in the area, Lindner said. It also has significant risk-management capability and benefits from a high-demand, low-supply environment. "We're able to get the structures and pricing to make sense," he said. Lindner added that Ace manages equity- and interest-rate exposure on a "holistic and global basis."

"Large players have built their own staffs and expertise, and they typically have six to 10 people working on hedging," Lindner said. "But it is not straightforward. There are no capital-market instruments correlated to the underlying risks." Other companies have purchased hedging plans from major consulting firms like Ernst & Young and Milliman, he said. "Will these schemes work? They might, but every plan has a cost," he said.

Bieluch said a hedging strategy "provides the Wall Street view of what the benefits are worth. "If the strategy is too expensive, it may mean the benefits are underpriced by the carriers," he said.

Bieluch said higher prices for benefits are likely. In general, the cost of hedging increases with market volatility, and volatility "is currently at relative historic lows," he said.