Meyers Flashcards

1
Q

Reasons models may not accurately predict distributions of data (3)

A
  1. insurance process is too dynamic to be captured by a single model
  2. other models could better fit data
  3. data used to calibrate model was missing crucial information (ex: changes in claims handling, underlying business changes, etc.)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Tests for uniformity and descriptions (2)

A
  1. histogram - bars of equal height
  2. p-p plot and Kolmogorov-Smirnov (K-S) test - plots predicted percentiles against expected percentiles and looks for a diagonal line along y=x
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

K-S test

A

compare expected vs. model predicted percentiles and set D = max ( abs. value ( difference ) )

reject the null hypothesis of uniformity if D > critical value

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

Possible test outcomes and interpretations of the histogram and p-p plots (4)

A
  1. uniform - relatively flat histogram and p-p plot tightly distributed around diagonal line
  2. light tailed - histogram is higher at endpoints and p-p plot has an “S shape”
  3. heavy-tailed - histogram is highest in the center and p-p plot has a “sideways S shape”
  4. biased high - histogram has the highest frequency at the lowest percentiles, and decreases as expected loss increases; and p-p plot is the curved below the diagonal (right side of a U)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Mack model results on cumulative incurred data

A

symmetric with light tails

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

Significance of light tails on model results

A

understates the variability of the predictive distribution

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

Bootstrap ODP model results on incremental paid data

A

non-symmetric and biased high

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

Significance of being biased high on model results

A

expected losses will be overstated (b/c actual < expected the majority of the time)

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

Potential explanations for why neither the Mack incurred or bootstrap paid ODP model validated (2)

A
  1. changes in the insurance environment which are not yet observable
  2. other models exist which can validate
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Methods to increase the variability of the predictive distribution (2)

A
  1. treat the level of the AY as random
  2. allow for correlation between AYs (vs. independence)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Mean of the leveled CL model (LCL) - aka mu(w,d)

A

level of each AY = mu(w,d) = a_w + b_d

a_w = level of log losses for an AY
exp(b_d) = % of losses paid to date

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

Mean of the correlated CL model (CCL) - aka mu(w,d)

A

a_w + b_d + p*[ln(C_w-1,d) - mu(w-1,d)]

a_w = log losses in each AY
exp(b_d) = % losses reported
p = correlation coefficient

parenthesis: previous AY cumulative claims - previous AY mean

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

Description of the leveled CL (LCL) and correlated CL (CCL) methods

A

LCL: includes a parameter for the level of each AY

CCL: LCL + parameter for correlation between AYs

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

Relationship between sigma and development age (d) for cumulative vs. incremental losses along with rationale for each

A

cumulative: as d increases sigma decreases - fewer open claims in later development periods that are subject to random fluctuations

incremental: as d increases sigma increases - smaller, less volatile claims will be settled first

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

Leveled CL results on cumulative incurred data

A

improvement over Mack model but still produces light tails

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

Correlated CL results on cumulative incurred data

A

model validates despite thin tails

17
Q

Mack, ODP bootstrap, and CCL method results on incremental paid losses

A

all biased high

18
Q

Consequences of using a CY trend (2)

A

Trend applies to incremental losses so:

  1. data tends to be right skewed
  2. occasionally has negative values
19
Q

Description of correlated incremental trend (CIT) and leveled incremental trend (LIT) methods

A

CIT - allows correlation between AYs and includes a CY trend parameter (tau)

LIT - CIT without correlation of AYs

20
Q

Description of the changing settlement rate (CSR) method and loss data used

A

includes a claim settlement rate parameter

uses cumulative loss data

21
Q

Mean of the CIT model - aka mu(w,d)

A

level of AY = mu(w,d) = alpha(w) + beta(d) + tau * (w + d - 1)

correlation is applied by using a mixed lognormal-normal distribution

22
Q

Mean of the CSR method - aka mu(w,d)

A

level of AY = mu(w,d) = a_w + b_d * (1 - gamma) ^ (w-1)

23
Q

CIT and LIT model results

A

still biased high, no noticeable improvement over Mack or ODP bootstrap methods

24
Q

CSR model results and explanation

A

model validates and corrects bias in other models

> > recognizes speedup in claim settlement over time because claims are reported and settled faster due to advancements in technology

25
Meyers' findings on relative size of parameter risk vs. process risk
parameter risk is very close to total risk i.e. there is minimal process risk