26
© 2016 Society of Actuaries
including variation in the asset strategy, the reinsurance strategy, the volume and pricing of new business, and the
ability to raise additional capital or restrict shareholder dividend payments.
This is not to say that organizations adopting a finite risk horizon approach do not recognize the need to address
longer-term risk issues and their potential consequences in terms of capital. They typically prefer to address such issues
through a scenario analysis of projected EC assessments over a business planning cycle (e.g., three to five years),
bringing in all the range of management actions that might reasonably be taken in each scenario. So, for example, a
prolonged period of poor equity returns might be considered, revealing a deteriorating capital and security position if
no action were to be taken. However, the management actions described above may be available in such a scenario,
and it would appear reasonable for management to rely on their utilization, instead of holding additional capital at the
outset to cover such a risk.
The stochastic liability runoff approach on the other hand does bring in all risks during the runoff of the portfolio, albeit
often only those relating to the existing portfolio (sometimes with a limited number of years’ new business also
included). A number of the actions available to management during that period may also be allowed for through
formulae included in the stochastic model. However, it is very rare for the full range of such actions to be incorporated,
as it is difficult to allow formulaically for such actions as additional capital raising and increased utilization of
hedging/reinsurance, as the capital position of the organization varies over time; also for the ability to vary new
business volume and prices in circumstances where new business is modeled.
In this context a number of pros and cons can be observed.
The finite risk horizon approach gives strong recognition to the fact that an organization’s principal ability to
control risk in the short term is through trading assets and/or liabilities, including through reinsurance and
portfolio/business transfer. However, the lack of data available to calibrate a distribution of market-consistent
prices for non-hedgeable liabilities such as mortality/morbidity may be regarded as a potentially significant
weakness.
The liability runoff approach, on the other hand, can give insufficient recognition to this ability to control risk
through asset/liability trading, unless sophisticated algorithms are built into the model to allow for it.
The finite risk horizon approach relies on deterministic adverse scenario analysis to examine longer term risks and
their management. This has a weakness in that it is reliant for its completeness on management’s scenario
selection (as opposed to using a stochastic scenario generation process), but has a strong advantage in allowing
management to make a realistic assessment and communication of all the risk management actions it might take
in such a scenario. Management can then make a conscious choice between taking such action and holding
additional capital, additional to initial EC assessment, effectively to cover their preference not to take such
management action.
The liability runoff approach aims to build longer-term management actions into the stochastic model, although in
practice this can be difficult to perform comprehensively. While this approach removes the reliance on
management scenario selection, stochastic projections of longer-term risk emergence and management thereof
can be less clear and more difficult to analyze than with a deterministic equivalent. There is a risk that EC can be
overstated through the omission of actions that might reasonably be taken, or alternatively that the reason for the
high capital requirement (a preference for, or an assumption of, less risk management action) is not clearly
understood. In addition, in an environment where management changes can occur fairly frequently, making
assumptions as to management actions over the longer term can be considered speculative.
Regardless of risk management actions that may be taken in the future, the reality of capital management and
regulatory reporting is that required capital will be calculated on (at least) an annual basis. When applied over a
one-year time period, the finite risk horizon approach acknowledges this reality and better aligns itself with the
actual management of the company. In contrast, the liability runoff approach attempts to find the amount of
capital today that will provide sufficient protection for the lifetime of the portfolio, thus ignoring the reality that
capital levels will be annually reevaluated.
A finite risk horizon approach to EC assesses the quantum of risk over a similar time period as is typically used for
shorter-term performance measurement purposes. This allows the consistent assessment of risk, capital and
performance. A liability runoff approach to EC can result in a timing mismatch with short-term performance being
compared with risk and capital assessments based on a longer-term horizon.