Peter Persson presenterar sitt examensarbete

Implementaion of a Risk Based Solvency model

Abstract

The new methodology of Solvency II is a  gigantic step to
introduce a solvency requirement that entirely depends on the risk
connected to the specific insurances. Today it is possible for
the industry to take large financial risks without any direct
effects in capital requirement. The purpose of this report is to see
how the change of calculating solvency requirement will affect
Länsförsäkringar Liv. Exactly how it is supposed to be done in
practice is today not known. The model that is implemented however is
the Swiss Solvency Test. It is a model that combines a multivariate
normal distribution from different risk aspects with some including
adverse scenarios. Every single risk aspect is assumed to be Gaussian
with an estimation of its standard deviation and expected value. In
addition, correlation between different risk aspects are estimated.
All changes among the risk aspects are also assumed to be linear with
respect to the risk-bearing capital. Therefore some non-linear risk
factors use the Delta approach. The expected shortfall on the 99%
quantile is used as a risk measure in the final multivariate normal
distribution to describe the 1-year risk capital requirement. The
aspects considered in the Swiss Solvency Test are both financial and
insurance risks. The insurance risks are due to possible events among
the policyholders and the financial risks are due to which investments
that are made with the insurance capital. Today, when calculating the
solvency requirement, it highly depends on the interest rate and
mortality among the policyholders. This will remain with Solvency II,
where especially the interest rate used to discount liabilities in the
far future is of importance. This is seen in the report where the
10-year interest rate is the most sensitive factor for the risk-bearing
capital and therefore also for the capital requirement.

Keywords: Solvency II, Swiss Solvency Test, Multivariate Normal Distribution,
Delta Approach, Makeham Distribution, Value at Risk, Expected Shortfall,
Monte Carlo Simulation.