Plastic Card Fraud Coverage and Its Actuarial Challenges: A Case Study

1
Plastic Card Fraud Coverage and Its Actuarial
Challenges A Case Study

• Fall 2006 MWAF
• September 29, 2006

2
Case Study Format
This morning, we will

• Describe a coverage that CMG writes.
• Present two specific actuarial challenges.
• Ask the audience to brain storm possible
  solutions or avenues of research for the two
  challenges.

3
Remember

• When we brain storm, there are no bad or dumb
  ideas.
• This session really depends on your
  participation.
• We have 50 minutes.

4
The Plastic Card Coverage

• Reimburses the insured for losses resulting from
  an Unauthorized Transaction occurring in the
  insureds Card Programs.
• Unauthorized Transactions stolen cards,
  counterfeited card, card not present internet
  purchases.
• Card Programs Credit, Debit and ATM.

5
Structure of the Plastic Card Coverage

• Account Number deductibles and limits (AKA per
  card deductibles and limits). Example 100
  deductible and a 15,000 limit
• Annual aggregate deductibles and annual aggregate
  limits. Example 25,000 AAD and a 1m AAL

6
Example of Adjusting a Plastic Card Loss

• 100 per card deductible and a 15,000 per card
  limit.
• 25,000 AAD and a 1m AAL.

7
Scale of CMGs Plastic Card Program

• CMG insures approximately 5,500 credit unions for
  Plastic Card.
• Credit union card programs range from a few
  hundred to 700,000 cards.
• 1,000 credit unions exceed their AAD each year.
  
• 90,000 card accounts will sustain losses on
  policies written in 2006.

8
We will see that the two actuarial challenges are
related to

• The annual aggregate deductible
• Frequency of Plastic Card losses

9
Looking for Analogies

• Has anyone worked with a coverage that uses an
  annual aggregate deductible?

• Possible analogies
• Excess Aggregate WC Policy
• Excess Reinsurance with AAD
• Medical Policies with Family AAD

10
The Loss Reserving Challenge

• How does the Reserving Actuary reserve for the
  Plastic Coverage, a high loss frequency coverage
  that uses an AAD ? The answer to this question
  probably depends on the answers to
• How does the Reserving Actuary count claims and
  assign loss dates?

11
A Plastic Card Loss Scenario
12
Loss Reserving Challenge Approach 1
Suppose we use the convention that we will count
and record claims at the policy level, that is,
if the AAD is exceeded, we record a single claim.

• What loss date do we use?
• Do we change the loss date if a loss at the
  cardholder level changes?
• Reserve using AY or PY histories?
• Is there is an appropriate matching of revenues
  and losses?

13
Loss Reserving Challenge Approach 2
Suppose we use the convention that we will count
and record claims at the cardholder level.

• What loss date do we use?
• What do we do with the AAD? Allocate a portion
  of it to each cardholder loss? Does the
  allocation change with additional cardholder
  losses?
• Reserve using AY or PY histories?
• Is there is an appropriate matching of revenues
  and losses?

14
Pricing Challenges

• For this complex coverage and rapidly changing
  loss environment
• What exposure base should the Pricing Actuary
  use?
• What combination of Pricing Plans (manual,
  experience and schedule rating) is best
• Structuring manual rating for both - Per Event
  Limit and Deductibles, and- Aggregate Limit and
  Deductibles

15
Pricing the Aggregate Deductible

• Is it sufficient to
• Design a table of Excess Loss Factorsby
  Deductible Amount?

(Too much variation by size of account)
16
Pricing the Aggregate Deductible

• Is it sufficient to
• Design a table of Excess Loss Factors by ratio
  to pure premium (Entry Ratio)?

(Better, but still too much variationby size of
account)
17
Sidebar Entry Ratio is not sufficient

• Are the following equivalent?
• 1 policy of 10,000 cards at AAD 1.0 E(L)
• 100 policies of 100 cards at AAD 1.0 E(L)

No Excess loss on 100 small policies will be
much higher than on a single large policy
18
Pricing the Aggregate Deductible

• Is it sufficient to
• Design a table of Excess Loss Factors by Entry
  Ratio and size grouping?

?

• How many columns do we need?
• Creates issues related to column transitions
• Model interaction with Per Event limitations?

19
Recent Plastic Card Experience

• Beginning with the 4Q 2004, Plastic Card Losses
  Have Sky-Rocketed Because of..
• Increased Credit Card Counterfeiting Related to
  File Intrusion at Merchants.
• Increased Debit/ATM Card Fraud Related to
  Phishing.
• Ground Up Losses Have More Than Tripled.