ING Direct 2008 Annual Report Download - page 230

2.1 Consolidated annual accounts
Risk management (continued)
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 Economic
Capital. Second, they assist Business Units in hedging strategies and management of Economic Capital. Third, the process of replicating
portfolio calculations increases the understanding of the complex nature of insurance liabilities in a market consistent environment.
Replicating portfolios are currently determined from a single factor RN ESG interest rate model. This limits the ability of the replicating
portfolios to pick up sensitivity to non-parallel shifts of the term structure of interest rates. Hence RW ESG interest rate scenarios for the
Value at Risk calculations are generated using a single factor model as well. However both RN ESG and RW ESG models are consistent
with respectively, the RN ESG and RW ESG volatility structure of interest rates.
Economic Capital Calculation
ECAPS uses Monte-Carlo simulation to determine diversification benefits for the complete ‘portfolio hierarchy’, from business unit level
up to an ING Group level. All diversification calculations are done within ECAPS and are driven by the Gaussian copula of all risk drivers
using the underlying distributions applicable for each risk type.
For the calculation of Economic Capital, ING uses a one-year time horizon. In practice, the model calculates instantaneous quarterly
shocks and then annualises the resulting VaR statistic to determine an annualised EC. The quarterly shock is used to stabilise the results,
to ensure the shocks are within a range that can be more credibly valued for assets and liabilities, to better capture the impact of dynamic
hedge strategies, to more reasonably use weekly correlations of risk factors, and to get closer to actual risk practices and reporting cycles.
Using Monte-Carlo simulation, ING’s Economic Capital model generates 20,000 possible ‘states-of-the-world’, by randomly simulating all
risk drivers – simultaneously. For each state-of-the-world, the market value of assets and liabilities are recalculated and the change in
value of the Market Value Surplus (MVS) is stored. All these changes in MVS are then sorted, and the 99.95% worst-case change in MVS
is identified, to provide the Economic Capital level for the given level of aggregation.
ECONOMIC CAPITAL GROUP
ING’s Group Economic Capital and Bank-Insurance diversification benefit is determined by applying one common aggregation approach to
the banking and insurance businesses. The starting point is the actual reported Economic Capital figures for ING Bank and ING Insurance,
excluding inter-risk diversification. In addition an aligned set of best-estimate correlation assumptions is constructed by applying the
weighted average of the Bank and Insurance specific inter-risk correlation assumptions for each of the five major risk types i.e. credit,
market, insurance, business, and operational (See also Economic Capital model sections of Bank and Insurance).
The group diversification benefit is calculated by applying a ‘Gaussian-copula’ simulation approach. Due to the inherent uncertainties
around correlation assumptions and changes in risk exposures the results are put to extensive sensitivity tests.
ING Group Annual Report 2008
228