Transcript Slide 1

Town Hall Session/ RCM-2
CAS 2006 Ratemaking Seminar,
March 13, 2006
Louise Francis, FCAS, MAAA
Francis Analytics and Actuarial Data Mining
[email protected]
www.data-mines.com
How Many Risk Load Actuaries
Can Dance on the Head of a
Shamrock?
Remarks
• Risk adjusted discount rates/internal rate of return
approaches still relevant
• Alternative beta calculations
• Alternative denominator for cost of equity
• Some older approaches such as risk loads based on
percentiles still make sense in some contexts
• The market seems to allow for superior strategy
returns
• Should we think of Risk Load as a topic within
ERM?
• Should we charge a “bad data” risk load?
Financial & Actuarial Perspective
Converge?
• Risk Load= Systematic
+
Frictional Cost
Simpler Methods are Sometimes
Better
• To some managements, our methods of
computing risk loads seems very complex
• Loads based on percentiles still used often for
self-insurance pools and captives
Augmentations of Beta
• Base on correlation with insurance
liabilities
• SUMBETA – (Ibbotson, Risk Premium
Project)
• Take lag effects into account
• For insurance companies generally results
in larger betas
Cost of Capital Denominator
• What is an appropriate denominator
• Usually GAAP surplus
• Smith says this is the wrong denominator
• Economic Value of Company = Market value
of assets - Discounted value of liabilities +
franchise value – insolvency put
• Measure change in this value
Return for Having an Edge:
Geometric Maximization
Edge
f 
Odds
f  Amount of capital to allocate to bet
“Bad Data” Risk Load
• Moral Hazard problem
• Customers supplying bad data may be
poorer risks due to poorer management
• Information has a cost
• More uncertainty is associated with
pricing based on sparse or inadequate
data