Mack 94

Question 1

Three options for a(I-1)

Answer 1

1) Set equal to zero (if LDF = 1)
2) Extrapolate using loglinear regression
3) min [(a(I-2)^4/a(I-3)^2),min of prior two a)

Question 2

Two problems with using normal distribution as approximation of R

Answer 2

1) If data is skewed it is a poor approximation

2) Confidence interval can have negative lower limits

Question 3

Alternative to normal distribution for R

Answer 3

Lognormal

Question 4

Three variance assumptions for C

Answer 4

1) Proportional to 1
2) Proportional to C
3) Proportional to C^2

Question 5

Corresponding estimator for LDF for each variance assumption of C

Answer 5

1) Proportional to 1, C^2
2) Proportional to C, volume-weighted
3) Proportional to C^2, simple-average

Question 6

Graphical variance test to determine appropriate LDF estimator

Answer 6

Residual plots

Question 7

Major consequence of first Mack assumption

Answer 7

Factors are uncorrelated; Spearman’s rank test

Question 8

Major consequence of second Mack assumption

Answer 8

CY effects do not exist; Diagonal test

Question 9

First Mack assumption

Answer 9

Cumulative claims are linear in nature

Question 10

Second Mack assumption

Answer 10

AYs are independent

Question 11

Third Mack assumption

Answer 11

Variance of C is proportional to cumulative losses to date