Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

34
ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009

Transcript of Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Page 1: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

abRate Monitoring

Steven PetlickCAS Underwriting Cycle SeminarOctober 5, 2009

Page 2: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 2

abRate Monitoring

Table of Contents

Rate monitoring from the perspective of the reinsurance pricing actuary

The effect of new business

What is the reinsurance pricing actuary to do?

Effect of economic crisis on price monitor

Conclusions

Page 3: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 3

abRate monitoring from the perspective of the reinsurance pricing actuary

Standard Pricing Methodology includes:

Trend historical losses to prospective treaty period

Put historical premiums at rate level of prospective treaty period; i.e.. put them “on level”

During hard market periods, rate increases can turn dirt into gold

Opposite holds true in soft market periods

Page 4: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 4

abRate monitoring from the perspective of the reinsurance pricing actuary

Rate Monitor for XYZ Insurance Company

Casualty Excess of Loss Treaty

Rate Change2001 02002 +10%2003 +35%2004 +9%2005 +3%2006 -1%2007 -9%2008 -10%

Page 5: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 5

abRate monitoring from the perspective of the reinsurance pricing actuary

What would you do? I would immediately get

back to the client with questions:

Are the rate changes written or earned?

Are they adjusted for exposure changes?

Are they adjusted for changes in limits/ attachments/ deductibles/ SIRs?

Do they include new business or are they measured on renewals only? (most common)

Page 6: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 6

abRate monitoring from the perspective of the reinsurance pricing actuary

Do they include the effect of commission changes?

Do they reflect changes in terms and conditions?

What types of increases/ decreases are they observing on lost business?

How are the rate changes actually calculated?

Page 7: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 7

abThe Effect of New Business

Suppose you are trending and onleveling a 2007 loss ratio for a reinsurance treaty incepting at 1/1/2009. (In reality you would likely use at least 5 historical years, but, for simplicity, we will consider only 2007.) Assume the following:

2007 Ultimate Loss Ratio = 60%

Expected Annual Trend = 6%

Renewal Rate Changes = 2008 -10%, 2009 -10%

Page 8: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 8

abThe Effect of New Business

If you assume that these rate changes apply to the entire book of business (i.e. new business rate adequacy = renewal rate adequacy) then your projected 2009 loss ratio would be:

60% x 1.06 / (1-.10) X 1.06/(1-.10) = 83.2%

Page 9: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 9

abThe Effect of New Business

Now suppose that you believe that new business is less adequately priced than renewal business. How would this affect your projection?

In order to answer this, you would need a few more pieces of information:

Expected renewal retention rate

Projected premium growth

Page 10: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 10

abThe Effect of New Business

Simulation Model for New Business Effect

Initial Assumptions

2007 base portfolio of 100 policies with premium of $50mm and loss ratio of 60%

Renewal rate changes of -10% for 2008 and 2009

Renewal retention rate of 80%

Page 11: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 11

abThe Effect of New Business

Simulation Model for New Business Effect

Initial Assumptions

New business for 2008 and 2009 will consist of 20 new policies with same average premium as renewal book

“New business differential” values of 0, -10% and -20%

Page 12: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 12

abThe Effect of New Business

What is “new business differential?”

It is defined as the difference in rate adequacy between new business in the portfolio as compared to the renewal book for the same period. It is NOT the rate change on new business.

More about this later.

Page 13: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 13

abThe Effect of New Business

Simulation model operation:

Base year 2007 portfolio is simulated: premium is generated from uniform distribution on 0-$1000; loss ratio from normal, mean 60%, SD 10%

Each policy is either renewed or non-renewed for 2008 according to the renewal retention probability (80%)

If a policy is renewed, the premium reflects the renewal rate change (-10%)

Page 14: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 14

abThe Effect of New Business

Simulation model operation:

For renewed policies, the loss ratio reflects assumed loss trend of 6% and renewal rate change

20 new business policies are generated using same uniform distribution reduced for renewal rate change

Loss ratio for new business policies is simulated using renewal loss ratio adjusted for new business differential

Page 15: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 15

abThe Effect of New Business

Previous results assume that the base book of business carries a loss ratio of 60%, and after lost (i.e. non-renewed) business the renewed book is unchanged (i.e. base at 60%.) The reality is that there may be a bias in the quality of the lost business, and in the soft market we might expect “better” business to be leaving. Companies report that lost business frequently moves at rate reductions of 20-30% or more!

Page 16: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 16

abThe Effect of New Business

We next assumed that the expected retention rate for an individual policy varies according to the loss ratio of the policy. For this simulation, we assume that for each point of loss ratio variation from the mean, the probability of renewal varies by one percentage point from the expected renewal retention rate. For example, if the base loss ratio for the book of business is 60%, and the renewal retention rate is 80%, a policy with a loss ratio of 62% would have a probability of renewal of 82%.

Page 17: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 17

abThe Effect of New Business:Results of Simulation

80% Retention Rate

82.0%84.0%86.0%88.0%90.0%92.0%94.0%96.0%98.0%

0 -10% -20%

New Business Differential

Loss

Rat

io 2009 Base Case

2009 Vary Ret

Page 18: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 18

abThe Effect of New Business

We next ran the simulations assuming a renewal retention of only 60%; not unheard of for some E&S writers.

Page 19: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 19

abThe Effect of New Business:Results of Simulation

60% Retention Rate

82.0%84.0%86.0%88.0%90.0%92.0%94.0%96.0%98.0%

100.0%102.0%

0 -10% -20%

New Business Differential

Loss

Rat

io 2009 Base Case

2009 Vary Ret

Page 20: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 20

abWhat is the reinsurance pricing actuary to

How do we estimate the “new business differential? Typically, such a quantity is not included in reinsurance submissions

Ask the ceding company if they have attempted to estimate the effect of new business/ lost business on their portfolio

Some rate monitors will already include these effects: e.g. ratio of actual to benchmark

Page 21: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 21

abWhat is the reinsurance pricing actuary to

Look for ways to extract this information from the data that is provided: e.g. if you can identify new business in current bordereau of policies, then compare price per million for like limits, attachments, classes, etc.

If you cannot identify new business, then look at changes in price per million from year to year, again for like limits, attachments, classes, etc. This can at least give you an indicator of the total (new + renewal) rate change – compare to the rate change provided in the submission

Page 22: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 22

abWhat is the reinsurance pricing actuary to

Compare the rate change that the ceding company provides to that of similar portfolios and to industry benchmarks (CIAB, MarketScout.) If they are very different and you have no way of estimating the new business effect, then you might assume that the differences are at least partially attributable to new business/ lost business.

Page 23: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 23

abWhat is the reinsurance pricing actuary to

New Adjust

ed

WrittenRenew

al Rate BusinessExposu

re Rate

Year PremiumRetenti

onChang

eDifferent

ialInflatio

nChang

e

2003 40,000 80%

2004 40,000 80% 15.0% 0% 0.0% 15.0%

2005 40,000 80% 29.0% 0% 0.0% 29.0%

2006 40,000 80% 10.0% 0% 0.0% 10.0%

2007 49,470 80% 2.0% -5% 0.0% 0.3%

2008 44,409 80%

-10.0% -10% 0.0% -11.8%

2009 43,054 80% -8.0% -10% 0.0% -10.2%

Page 24: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 24

abEconomic Crisis Effect on Rate Monitors

Most rate monitors calculate rate changes based on change in premium per exposure

Exposures are typically estimated at time of policy renewal and premium is calculated based on exposure base

Rate monitors are typically created using actual premium and estimated exposures at time of pricing

For many types of policies, premium is not adjusted if exposure base turns out different than that assumed at time of pricing

In most cases, rate monitors are not adjusted

Page 25: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 25

abEconomic Crisis Effect on Rate Monitors

This can result is distortions in rate monitors if exposure estimates are materially incorrect

During current economic crisis, many exposures were overestimated, in particular for classes like Construction (sales), Manufacturing (sales)

This may have resulted in overstatement of rate decreases during 2008 and possibly 2009

Page 26: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 26

abEconomic Crisis Effect on Rate Monitors

Example: Renewal rate change

Portfolio of 20 risks with 2007 premium of $1.76m and exposure of $37.1m (could be sales)

Portfolio renews in 2008 with premium of $1.66m and exposure of $37.9m

Rate change = -7.80%

Company provides this information to reinsurer for 1/1/2009 treaty renewal

Assumed rate change for 2009 is +1.5%

Page 27: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 27

abEconomic Crisis Effect on Rate Monitors

Actual exposure for 2008 comes in on average 10.8% below projected. Since premium does not change, effective rate change becomes +3.40% (as opposed to original -7.80%.)

Does the company go back and adjust their rate monitor for the revised exposure? Probably not.

When 1/1/2010 renewal is done, the rate monitor should “self-adjust”; i.e.. the 2009 monitor should reflect the revised exposure and the effective 2009 rate change will be much higher, all other things being equal.

Page 28: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 28

abEconomic Crisis Effect on Rate Monitors

Example: Reinsurance treaty Effective 1/1/2009

Using 2008 rate change of -7.8% (2009 assumption +1.5%), indicated loss ratio = 87.6%

Using “real” 2008 rate change of +3.4% (2009 assumption +1.5%), indicated loss ratio = 78.1%

This would have a major impact on the reinsurance underwriting decision

Page 29: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 29

abEconomic Crisis Effect on Rate Monitors

Example: Renewal Effective 1/1/2010

“Real” rate change for 2009 is +1.5%; however, company does not go back and correct exposures used in 2008 rate monitor, so understated premium per exposure is used. With “real” 2009 premium per exposure (that is, reflecting reduced exposure), the 2009 rate change appears to be +13.8%. (+2% rate change is assumed for 2010.) Indicated loss ratio is 86.5%.

If 2008 exposures are corrected, and 2008 rate change of +3.4% is used, along with real rate change of +1.5% for 2009 (and same +2% for 2010), indicated loss ratio is ________

Page 30: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 30

abEconomic Crisis Effect on Rate Monitors

Example: Renewal Effective 1/1/2010

86.5%

So the process self-corrected.

However, the underwriting decision for 2008 could have be affected. In addition, if the exposures continue to drop, this process could be perpetuated over multiple years.

Page 31: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 31

abEconomic Crisis Effect on Rate Monitors

Caveats:

• Just because exposure base falls short, does that mean that actual exposure has fallen short? For example, for a Construction risk, if sales decline, was it due to less actual work or lower fees for the same work? The latter may not represent a real drop in loss exposure.

• Is exposure auditable; i.e. can premium be adjusted for audited exposure? (Example: Workers Comp.)

Page 32: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 32

abEconomic Crisis Effect on Rate Monitors

Caveats:

• Is the chosen exposure base the best measure of exposure to loss? For example, for a Trucking risk, exposure base may be number of units. In the recession, the number of units may not change, but the miles driven may be down.

Page 33: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 33

abConclusions

Summary:

Always understand how rate monitors in reinsurance submissions are computed

In addition, ask the client if new business is being included in the rate monitor

If not, request information that would enable you to estimate the new business effect

Understand how exposure base is being used in the rate monitor

Page 34: Ab Rate Monitoring Steven Petlick CAS Underwriting Cycle Seminar October 5, 2009.

Slide 34

abConclusions

Summary:

Determine if company is making corrections to rate monitor to account for revisions in exposure base

Incorporate assumptions on going forward exposure base when projecting future rate movement