Entity-wide Risk Management for Pension Funds
[this page | pdf | references | back links]
“Entity-wide Risk Management for Pension Funds”, co-authored by Malcolm Kemp
and Chinu Patel, was presented at meetings of the Institute and Faculty of
Actuaries in Edinburgh and London on 21 February and 28 February 2011. A
pre-print version of the paper is available here
and slides relating to a model example used in the presentation are available here.
Other supporting material and relevant links to the UK Actuarial Profession
website are available in the Nematrian Presentation Library.
The paper includes a reference to ‘Kemp (2011)’, i.e. this
page, in two places:
(a) In a footnote
in Section 1.8 in which this page is referred to as a source of material on
more mathematical aspects of risk management.
Please refer to relevant pages on
the Nematrian website focusing generically on risk measurement, including Introduction to Risk
(b) In Section 6.6 in
which this page is referred to as a source of models that can analyse
quantitatively the different interests that different stakeholders might have
in a pension fund
See below for a description of
the types of output such a model might produce and see here for more
material on the Nematrian website relating to such a model.
Patel (2011) express the view in Section 6.7 of their paper that a
particularly useful model to focus on in this respect is one that aims to
estimate the following:
(1) The spread (versus
risk-free) on the beneficiaries’ pension benefits implicit in the arrangement,
because payment of pension benefits will in general be contingent on the
continued health of the pension fund and ultimately therefore on the continued
health of the sponsor.
(2) The effective
(instantaneous) asset mix underlying the beneficiaries’ interest in the fund.
(3) The effective
(instantaneous) asset mix underlying the sponsor’s interest in the fund.
Their reasons for doing so include:
(a) It fits
naturally with the balance sheet characterisation described in Section 5.8 and
the Appendix of their paper;
(b) It helps
differentiate between sponsors and beneficiaries in cases where their interests
diverge, as per Section 3.7 of their paper;
(c) These outputs
potentially offer insights not directly available from traditional ALM models.
For example, they provide a more effective way of analysing sponsor covenant
risk from the perspective of beneficiaries/trustees than traditional ALM
(d) Such a model
naturally places ‘economic’ values on assets and liabilities and thus is
immediately consistent with financial economic principles. This type of model
behaviour is seen as particularly important for promoting effective ERM by Hatchett et
al. (2010); and
(e) It hopefully
simplifies the interpretation of any assumed equity risk premium (and other
similar aggregate economic assumptions). As explained above, traditional ALM
models generally include an assumed (positive) equity risk premium. Often their
primary purpose is to help quantify the trade-off between more favourable
expected outcomes and potentially more unpalatable adverse outcomes.
Beneficiaries (and sponsors) should therefore be particularly interested in
understanding how much of any apparent risk-adjusted return uplift is purely
the result of an assumed equity risk premium and how much derives from other
factors. For example, beneficiaries who are rich enough could alter the
disposition of the remainder of their assets to neutralise their implicit
equity exposure via the pension scheme. The same is true of the sponsor. The
effective (instantaneous) asset mixes in (1) and (2) indicate the magnitudes of
the adjustments that each would need to make to their other assets (and
liabilities) if they wished to hedge their investment exposures in this manner.