G.3. BF Method Flashcards

1
Q

BF formula for estimated ultimate claims using paid data

A

BF Ultimate = Actual Paid + Expected Unpaid
= Actual Paid + Expected Claims x % Unpaid
= Actual Paid + EP x ELR x % Unpaid

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

BF formula for estimated ultimate claims using reported data

A

BF Ultimate = Actual Reported + Expected Unreported
= Actual Reported + Expected Claims x % Unreported
= Actual Reported + EP x ELR x % Unreported

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How BF is credibility weighting of other methods

A

BF Ultimate = Development Ultimate x 1/CDF + Expected claims x (1 - 1/CDF)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Benktander formula

A

Benktander ultimate = Actual + BF Ultimate x % Unreported

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

How Benktander is credibility weighting of other methods

A

Benktander = Development Ultimate x 1/CDF + BF x (1 - 1/CDF)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Many iterations of Benktander

A

More weight to development ultimate until it approaches the development ultimate

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Assumptions of B-F method

A

IBNR or unpaid claims are better estimated using a priori estimate. Claims reported to date tell you no useful information about IBNR or unpaid claims for that exposure period

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Common uses of B-F

A

Random fluctuations or large claims at early maturities
When entering new line of business
Estimating ultimates at early maturities for long-tailed lines where early LDFs are highly leveraged

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Advantages of BF and Benktander

A

B-F: more stable estimates than chain ladder and more responsive than expected claims
Benktander: even more responsive than B-F but still more stable than chain ladder (but not more stable than BF)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

2 challenges in using BF or Banktander

A
  1. Estimating expected claims

2. Estimating expected % unreported

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

When credibility interpretation of BF and Benktander does not hold

A

Negative development, since 1/CDF > 1

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Options for dealing with downward development using BF or Benktander

A

Continue to use
Limit CDFs to minimum of 1
Rely on different technique to select ultimates with CDF

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Speedups/slowdowns and BF

A

Speedup: Paid will overestimate if speedup. Error will not be as big as chain ladder, since weight given to expected claims technique reduces amount of error

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Case reserve adequacy and BF

A

Increase: reported will overestimate. Error will not be as big as chain ladder, since weight given to expected claims technique

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Changes in claim ratios and BF

A

Do not fully react to changes in claims ratios due to Expected claims. Reported BF will be more responsive than Paid BF since more weight will be given to chain ladder when reported claims are used.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Exposure growth and BF

A

Unaffected by growth on its own. Change in accident date –> methods affected in same direction as chain ladder (Ex: growing book of business causes average accident date to be later, BF will underestimate)

17
Q

Mix of business and BF

A

Inaccurate if:
Segments of business that are changing have different development patterns (chain ladder)
Segments of business have different expected claims ratios (expected claims)