Market Data

The Crucial Role of Accurate Market Data in Asset Liability Management

Life insurance companies protect their policyholders not only from longevity and mortality risks, but also from market (investment) risks. In the complex and ‘asset-heavy’ world of life and annuity insurance, managing the delicate balance between asset and liability risk (duration/cashflows, etc.) is of paramount importance. One critical input in this process high-quality derived market data, such as interest rate curves (MIBOR, MIOIS, MIFOR, MMFOR and Government Bond curves) and other market-consistent estimates.

Parametric vs Non-Parametric

Parametric models can be loosely understood as “formulae and parameter based”. Such formulae and parameters (aka constants) are fitted to the observed market data.   Non-parametric models, on the other hand, do not rely on pre-defined formulas or parameters, and let the actual data govern the functional relationship between independent and dependent variables.

The global shift toward market-consistent risk models

The ongoing global shift towards the use of ‘market-consistent’ valuation and risk models in the life insurance industry undoubtedly favors the use of non-parametric (data-driven) models. IFRS17 and GAAP LDTI both require the use of market-consistent valuation and reserving methods for their liabilities (annuities and life insurance policies). To produce analyses that are truly  ‘market-consistent’, one needs a representation of the market (such as the interest rate yield curve) that is truly consistent with the market conditions ‘as they persisted’. In other words, a ‘market-consistent’ yield curve is truly representative of the market conditions on a particular date or event. 

Risk-Neutral and Market-Consistent mean the same thing

Yes, indeed.

IFRS 17 and market-consistent discount rates

IFRS 17 requires life insurers to use market-consistent approaches to pricing and reserving their liabilities. IFRS 17 allows for two different approaches to constructing market-consistent discount curves viz, the “top-down” and the “bottom-up” approach. Here, at Annuity Risk, we focus almost exclusively on the bottom-up approach. Bottom-up is also the most common method implemented in the United States (both GAAP and SOA) in similar requirements. The discount curve produced by the bottom-up approach will also (theoretically) lead to more consistency in valuation between companies and across time periods within the same company. In the bottom-up approach, we build a liquid risk-free yield curve using government bonds. We then overlay a precisely calibrated liquidity term structure. Then we overlay multiple ratings-dependent and market-calibrated credit risk term structures.

Minor kinds in the yield curve are ok

In a world where interest rates and currencies are managed by central banks, we have come to accept minor kinks in the yield curve as normal, even though they might present transient arbitrage opportunities. We see large and persistent dislocations in the government bond and REPO markets in the US, which strongly suggest that kinks (pointy corners) in the yield curves should be acceptable.

Economic Scenario Generator (ESG)

Risk-neutral (aka market consistent) stochastic scenarios require accurate and well-calibrated discount curves and commonly accepted parametric models. These stochastic interest rate scenarios can be tuned to risk-free yields, or to varying degrees of credit risk. These scenarios are used to price asset and liability cash flows. This method is well-suited for pricing discontinuous cash flows and other non-standard liabilities and assets. This method produces a true, model-free market-consistent valuation.

Summary

The importance of using high-quality yield curve data cannot be understated. Insurers must begin implementing ‘market-consistent’ valuation and reserving methods as soon as possible, in order to avoid discovering large-scale fluctuations or mismatches in their balance sheets later. Annuity Risk’s proprietary derived data technology produces accurate and market-consistent yield curves, and other derived datasets that become the primary valuation inputs in a wide range of valuation methodologies.