In the wake of the recent global credit crisis and resulting market volatility, the insurance industry is certainly no exception in having to deal with its own set of industry challenges. Numerix’s Mark Hadley discusses the trends in pricing and risk management of variable & fixed Annuities along with the actuarial challenges posed by the current market environment.
Q.It is safe to say that insurance companies are becoming bigger and bigger consumers of derivatives. Where do you see this growth and usage coming from?
This increasing derivatives activity is coming from many angles and serves many purposes within the insurance enterprise. Specific to actuarial / product / liability groups, insurers are changing the way they produce and manage retirement savings products (e.g. Variable Annuity guarantees, traditional fixed annuities, Long-Term Care, et.al.) and adopting similar infrastructures and best practices as utilized by an investment bank in managing its derivatives book.
Some specific derivative ‘management’ changes include:
– calculating arbitrage-free prices for new product ideas
– determine fair premiums for newly issued policies
– valuing the liabilities on their (quarterly and annual) financial statements using market- consistent valuations for the derivatives embedded in their insurance products, as outlined in FAS133 and FAS157.
Q. As Insurers evolve their internal capabilities to manage these instruments, what are some of the key issues companies are thinking about?
Firstly, within the Asset Liability Management (ALM) groups, or Enterprise Risk Management (ERM) groups, they are looking at;
• Economic scenario generation (ESG) in a cross-asset, cross-currency setting; in either the “risk-neutral” or “real-world” measure and how these scenarios can be generated from a single central model and distributed to all working groups within the enterprise
• Hedging the liabilities with derivative trades in the capital markets using risk-neutral valuation models. Finding new dimensions of risk to hedge, including various cross Greeks like rate/gamma which has huge impact on GMXB values.
• Projecting different hedging strategies into the future. Defining a hedging strategy includes the determination of hedging instruments and rebalancing thresholds. Hedging strategies might be:
-fully dynamic: as in Greek-based hedging, like a delta/vega/rho strategy
-static: as in a “risk minimizing” or “replicating” portfolio approach which is based on optimization or “best fit” algorithms
-semi-static: which combines elements of both dynamic and static approaches
Additionally within the asset/investment management groups, they are reviewing:
• Replicating Synthetic Asset Transactions (RSATs) are derivative transactions executed by insurance companies, and approved by their regulators, for the purposes of income generation. They are becoming more and more common among insurance companies. The most common RSAT is the combination of a cash Treasury bond and selling CDS protection. This is equivalent to buying a cash Corporate bond. Exploiting arbitrage profits is the main motivation for insurers to engage in this activity. For example, “synthetic mortgages” try to mimic the prepayment option in MBS with Bermudan call features in exotic interest rate swaps.
• Explaining the day-to-day performance of their derivatives positions through performance attributions analysis:
– Value at risk calculations for better collateral management
– Managing counterparty credit risk
– Managing structured credit portfolios (either synthetic or cash CDOs)
Q. What are the challenges confronting most issuer’s today, and what is required of financial models for fixed / variable products that take into consideration current market factors such as high market volatility and low interest rates?
One area we continue to do a lot of work with insurance companies is helping VA writers as they take into consideration the joint impact that these market factors are having on the product. This is quite often referred to as “cross-Greeks” where the correlation between Rates, Equities and Volatility is specified in the model. All of the above mentioned activities are hugely computationally intensive, especially those which require huge “stochastic on stochastic” or “nested stochastic” calculations. Distributed computing technology is a must.
On the regulatory front, setting aside economic capital reserves and meeting other regulatory requirements for capital adequacy will become increasingly important. VACARVM is the latest of these types of calculations, which goes live at year end 2009 and will be required by regulators going forward.
Q. What are some of the variable and fixed product guarantees that are new to market?
There are several new emerging product variations that have been recently introduced. For example, policyholders are very concerned about high inflation rates in the future, which might erode the value of their nest egg. The response has been an introduction of inflation-indexed products.
Moreover, insurers have also been scaling back on the generosity of their product features to lower their exposure to volatility. One common practice is reducing the “roll up” rate on their VA products. As a result actuaries are requiring new scenario type solutions where various features such as roll ups can be manipulated and show potential impact on a policy.
Finally, basis risk, or the risk that the hedging instruments do not match the underlying funds, posed serious problems at many insurers last year, when many actively managed funds severely under-performed their benchmarks. Insurers are responding by altering the menu of mutual funds in their VA products, and are showing a clear preference for passively managed index funds. Also, funds with controlled or target volatility are appearing in VA products.
Q. What new trends with respect to how insurers are assessing the risk of their fixed and variable annuity products are you observing?
Many insurers are starting to look more at replicating portfolios, otherwise referred to as “risk minimizing portfolios”. These are asset portfolios which match the risks associated with the liability block of business. They can be tracked on a daily basis and subjected to many market scenario shocks.
More and more insurers are increasing their breadth of hedge instruments. Credit default swaps, total return swaps, and hybrid derivatives like levered puts are all hedges that might help insurers better manage their capital markets risk.
Lastly, policyholder behavior in insurance products has always been a difficult to model. One answer to this is to estimate values assuming “optimal” policyholder behavior, where the policyholder behaves completely rationally and based on the expected future contract value versus the value of immediate lapse. The fixed income derivatives market has many contract features that are very similar to policyholder lapsation, such as Bermudan options, early exercise and callability features. These features can be modeled and valued the same, using Numerix’s Generic Tree, which is based on the pioneering work of Longstaff-Schwarrz.
Q. How do Numerix pricing and analytics solutions go beyond the traditional actuarial software?
Many of the major actuarial software products on the market today are able to capture the insurance / underwriting risks of various liability products adequately. However, the challenge is keeping pace with advancements in capital market modeling that influence the values and hedging strategies of many insurance products. Clients can leverage Numerix within their actuarial software packages to generate scenarios, path by path, that can be integrated into existing actuarial cash flow models and systems.
* Mark Hadley is an Actuarial Financial Engineer at Numerix
