ING Direct 2009 Annual Report Download - page 250

Model Disclosures
Further details on Economic Capital model
Market Data and Scenario Generation
ING Insurance uses ING Bank’s Global Market Database (GMDB) as a provider of market price and risk data for financial risk drivers. All
market data is obtained from reputable data providers such as Reuters and Bloomberg. The GMDB operational team then validates the
market data and calculates relevant risk parameters. This validated data is then automatically delivered to the ECAPS system.
Since ING Insurance operates in many developing financial markets, extrapolation algorithms are in place for extending beyond observable
market data when this is needed for the calculation of the Market Value Liabilities and the Economic Capital. These algorithms are based
on comparable data in mature markets.
Based on the market data from GMDB, ING calibrates two economic scenario generators:
Risk Neutral Economic Scenario Generator (RN ESG): capable of generating multiple equity indices and exchange rates, consistent with a •
multi-currency dynamic term structure model. Scenarios are used in the cash flow projection to determine replicating portfolios. RN ESG
scenarios are consistent with observed market prices of equity, FX and interest options;
Real World Economic Scenario Generator (RW ESG): capable of jointly simulating all risk types, i.e. all market risks, credit risk, business •
risk, operational risk, life risk, morbidity risk and P&C risk. Diversification between risks is taken into account through a Gaussian copula,
allowing for different marginal probability distributions at the risk driver level. RW ESG scenarios are consistent with historical time series
of the market risk drivers using 5 years of weekly data observations. The volatilities are scaled from weekly to quarterly and the weekly
correlations are used directly as estimates of quarterly correlations.
Stochastic Cash Flows and Non-Market Risk Capital
The market risks in assets and liabilities are captured in and represented by stochastic cash flows in 500 scenarios. Business units are
responsible for generating these cash flows, the modelling of embedded options and guarantees and a proper mapping of risk drivers in
the scenario set to cash flow determinants such as policyholder behaviour and management actions restricted to dynamic hedge programs
and setting of crediting rates/profit sharing. To better capture the behaviour in the tails of the distribution, the set of scenarios consist of
300 Risk Neutral scenarios and 200 ‘Risk Volatile’ scenarios with double volatilities. The average of the 300 Risk Neutral scenarios provides
a check on the market value of the replicating portfolio. It should be noted that this serves only as a check, and that the actual market
value of liabilities is derived directly from the replicating portfolio. The 200 Risk Volatile scenarios ensure that the replicating portfolio is
calibrated against enough extreme scenarios such that it can be used safely in Economic Capital calculations.
Non-market risk Economic Capital is calculated by business units, Corporate Credit Risk Management and Corporate Operational,
Information and Security Risk Management and inputted into ECAPS at the sub risk level. ECAPS than aggregates 21 sub-risk types (e.g.
mortality and trend risk) to 9 non-market risk types using a bottom-up Economic Capital diversification approach based on a matrix of tail
correlations. The information inputs relate to 9 sub risk types:
Credit risk;•
Business risk;•
Operational risk;•
Life risk catastrophe;•
Life risk non-catastrophe;•
Morbidity risk catastrophe;•
Morbidity risk non-catastrophe;•
P&C risk catastrophe;•
P&C risk non-catastrophe.•
The inputs are used to calibrate marginal distributions for these risk types. These distributions, in combination with the Gaussian copula,
are then used in the Economic Capital Calculation to measure diversification between market and non-market risks.
Replicating Portfolios
To handle the full complexity of calculating diversification by Monte Carlo simulation, ING maps its assets and liabilities to a set of standard
financial instruments. The set of standard instruments consists of zero coupon bonds, market indices, equity forwards, swaptions, callable
bonds, FX options and equity options. Assets and the financial components of the liabilities are represented by a portfolio of this standard
set of instruments. A user interface allows the selection of different types of replicating instruments for different cash flow types. Then an
optimal replicating portfolio is created that matches the risk profile of the stochastically generated cash flows as good as possible. The
resulting replicating portfolio is used in the calculation of Economic Capital.
Through the inclusion of equity options, FX options and swaptions in the set of replicating instruments, ING is able to incorporate implied
volatility risk in the considered risk types. The same holds for the credit spread risk through the inclusion of credit risk bearing zero coupon
bonds in the set of replicating instruments.
The quality of the replicating portfolio is monitored by several statistical criteria including R-squared and benchmarked against market
value sensitivities such as duration, convexity, and changes in value for larger interest rate and equity shocks. High quality replicating
portfolios are important in several ways. First, they ensure a good reflection of the actual risk profile and an accurate calculation of
Risk management (continued)
2.1 Consolidated annual accounts
ING Group Annual Report 2009
248