Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Blowing through the range
2014 CLRS presentationSeptember 15, 2014
Christopher Andersen, FCAS, MAAARon Fowler, FCAS, MAAA
Page 2
Agenda
► Actuarial ranges and their context► Communication of the actuarial range► When things go wrong – Case studies
► Financial guarantee► Catastrophes► Asbestos
Page 3
What is in an actuarial range?
► Actuarial standards of practice No. 43► Property / Casualty Unpaid Claim Estimates
► Actuarial central estimate► An estimate that represents an expected value over the range of
reasonably possible outcomes (“range”).
► 3.3.1 …may not include all conceivable outcomes, as, for example, it would not include conceivable extreme events where the contribution of such events to an expected value is not reliably estimable.
Page 4
What disclosures are proper for an actuary to communicate?
► 3.6.7 – External conditions► Claim obligations are influenced by … potential economic
changes, regulatory actions, judicial decisions, or political or social forces. …the actuary is not required to have detailed knowledge of or consider all possible external conditions…
► 3.6.7 – Changing conditions► The actuary should consider whether there have been significant
changes in conditions …Examples include reinsurance program changes, … claims personnel [changes], …
Page 5
What disclosures are proper for an actuary to communicate? [cont.]
► 3.6.8 – Uncertainty► The actuary should consider the uncertainty … This standard does
not require or prohibit the actuary from measuring this uncertainty. … the actuary should consider the types and sources of uncertainty … [and] may include uncertainty due to model risk, parameter risk, and process risk.
► 4.2 – Additional disclosures (re: range of estimates)► In the case when the actuary specifies a range of estimates, the
actuary should disclose the basis of the range provided
Page 6
When things go wrong – Case studies
► 2008 financial crises ► Financial guarantee insurance
► Unusual catastrophes► Hurricane Katrina► Thailand flooding► New Zealand earthquake
► Asbestos
Page 7
The 2008 financial crises was more sudden than previous depressions in the financial markets
-40.0%
-30.0%
-20.0%
-10.0%
0.0%
10.0%
20.0%
30.0%DJIA – Year over year percentage change
Prepared by EY
Page 8
Industry loss ratio spiked to 200% to 300% and have still not recovered to pre-financial crisis levels
0.0%100.0%200.0%300.0%400.0%500.0%600.0%700.0%800.0%900.0%
1000.0%
Loss ratios for Industry and Top Financial Guarantee and Mortgage-Back Securities writers
P&C Industry Company 1 Company 2 Company 3 Company 4
Based on SNL data; Prepared by EY
Page 9
Unprecedented financial market performance underlined inadequate reserve amounts in financial guarantee and mortgage guarantee products
-50.0%-25.0%
0.0%25.0%50.0%75.0%
100.0%125.0%
1 year reserve development / Prior year reserves
Commercial Auto Workers CompensationMedical Professional Fidelity and suretyFinancial Guarantee and Mortgage Gaurantee
Based on SNL data; Prepared by EY
Page 10
Financial guarantee insurance
► What went wrong?► Was the cause of the reserve development due to process risk or
parameter risk?
► The development was due to a “tail event” or process risk
► How did the actuarial models uphold?► New information needed to be reflected that was not traditional to
historical analysis and development► Economic projections► TARP program
► Scenario testing is critical in developing a range of expectations
Page 11
2006 vintage curves indicated stability
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%2 8 14 20 26 32 38 44 50 56 62 68 74 80 86 92 98 104
110
116
122
128
Seve
rity
Loan Age
1998199920002001200220032004200520062007
Page 12
Post 2008 financial crisis the historical development was no longer relevant
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%2 8 14 20 26 32 38 44 50 56 62 68 74 80 86 92 98 104
110
116
122
128
Seve
rity
Loan Age
1998199920002001200220032004200520062007
Page 13
Financial guarantee insuranceCase study – Post-crisis estimates and range
► Scenario testing was employed as future state after shock was difficult to ascertain from historical data► What movement in transition probabilities is plausible?► How high can loss severities be?
Scenarios Transition probabilities Loss Severities Difference to Median
A No change Decrease by 5% -36%
B Used 1 month look back No change -28%
C Used 2004 probabilities Increased by 5% -4%
D Lag vintage 1 year Increased by 5% +4%
E Used 2005 probabilities for all older vintages Increased by 7.5% +57%
F Used 2007 probabilities for all vintages No change +118%
Page 14
Developing a range using non-traditional information and scenario testing
► Scenario testing► Search for relevant information from non-traditional sources
► Economic data and trends► Understand reform and its impact
► Discuss with client reasonability of assumptions and apply professional skepticism to avoid biases► Develop a maximum probable loss scenario► Consider industry perspective on phenomenon
Page 15
When things go wrong – Case studies
► 2008 financial crises ► Financial guarantee insurance
► Unusual catastrophes► Hurricane Katrina► Thailand flooding► New Zealand earthquake
► Asbestos
Page 16
Recent catastrophes have proven to have stipulations that make predicting ultimate losses more uncertain
► Hurricane Katrina► Wind versus water disputes► Extended recovery and rebuilding period
► Thailand flooding► Claim investigation was severely delayed to due standing water
and the inability to investigate claim sites► Largest claims were business interruption and have a time-
element
► New Zealand earthquake
Page 17
The Thailand Flooding claims have been harder to identify and investigate due to the lingering water in warehouses
Prepared by EY; based on RAA data. 1 period is equal to 3 months
0.0%10.0%20.0%30.0%40.0%50.0%60.0%70.0%80.0%90.0%
100.0%
1 3 5 7 9 11 13 15 17 19
% C
ase
incu
rred
to u
ltim
ate
Periods
Thailand FloodsKatrinaRitaWilmaNZ2 EQJapan EQWTC
► Significantly lower after 6 periods► Time-element coverage is causing
uncertainty
Page 18
Historical earthquake % of case incurred to ultimate development
Prepared by EY; based on RAA data. 1 period is equal to 3 months
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
120.0%
1 2 3 4 5 6 7 8 9 1011121314151617181920
% C
ase
incu
rred
to u
ltim
ate
Periods
NZ EQ2ChileJapanLoma PrietaNorthridgeNZ EQ1NZ EQ3
► The case incurred amounts continue to rise through 9 report periods
Page 19
New Zealand earthquake has experience high “event creep” than other recent natural catastrophes
► Event creep (New Zealand)► New Zealand Earthquakes: Christchurch, NZ
► NZ I: October 2010► NZ II: February 2011► Many smaller aftershocks
► New Zealand Earthquake Commission pays up to 100,000 NZD property, 20,000 NZD contents
► Suncorp and IAG have majority of the market share for additional insurance
► Reinsurers then provide excess cover (they get the effects of the creep)
Page 20
Reason for event creep for New Zealand earthquake
► Scope of event (largest natural CAT year in NZ)► Apportionment: losses were apportioned between the two
events, often with complicated models► Renewals: the two earthquakes are treated as two events
and renewals separate the contracts so there are new limits
► Liquefaction (some neighborhoods abandoned), this was not factored into many loss models
► Claim settlement time (government insurer handling claims, slower than private insurance to process)
► Cordoned off areas (Red Zone) causes BI losses
Liquefaction
Page 23
When things go wrong – Case studies
► 2008 financial crises ► Financial guarantee insurance► Mortgage-back security insurance
► Unusual catastrophes► Hurricane Katrina► Thailand flooding► New Zealand earthquake
► Asbestos
Page 24
A history of predictions and re-predictions
► The poor track record of estimating asbestos
0100,000200,000300,000400,000500,000600,000700,000800,000900,000
1,000,000
1988 1993 1997 2005 2013
Manville Trust – Asbestos Claims filing
Filed claims Expected Ultimate filed claims
Page 25
AM Best’s asbestos and environmental studies have a similar shape
0
10
20
30
40
50
60
70
80
90
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
P/C Industry Net Asbestos Losses
Incurred Losses AM Best Ultimate
Based on annual statement data and AM Best
Page 26
So what’s in an asbestos range estimate?
► Historical development may have been considered an (un)conceivable extreme scenario or event
► Ranges are based on the best available data at that time► Survival ratio analyses► Exposure analyses► Market share and industry estimates
► Wide range of estimates may result – what is the right range?