soa post-level term experience resultssoa sponsored post-level term reports 2007 – report on the...
TRANSCRIPT
Derek Kueker, FSA
SOA Post-Level Term
Experience Results
Actuaries’ Club of the Southwest
June 19, 2014
Introduction to Post-Level Term
Background
Assumption Survey Results
Experience Study Results
Predictive Model
Acknowledgements
Table of Contents
2
Introduction to Post-Level Term
3
Introduction to Post-Level Term
10-Year Term: Premium Structure with Jump to ART
-
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Level Period Post-Level Period
An
nu
al P
rem
ium
Duration
10-Year Term45-Year Old, $500,000 Death Benefit
Super-Preferred Non-Smoker
$7,395
Duration 11
Premium
20 times
Premium Jump
$375 Level
Premium
4
Background
5
SOA Sponsored Post-Level Term Reports
2007 – Report on the Survey of Post-Level Premium Period Lapse and Mortality
Assumptions and Experience For Level Premium Term Plans
2009 – Report on the Survey of Post-Level Premium Period Lapse and Mortality
Assumptions for Level Premium Term Plans
2010 – Report on the Lapse and Mortality Experience of Post-Level Premium
Period Term Plans
2013 – Report on the Survey of Post-Level Premium Period Lapse and Mortality
Assumptions for Level Premium Term Plans
2014 – Report on the Lapse and Mortality Experience of Post-Level Premium
Period Term Plans
Background
6
Assumption Survey Results
7
Sent to top 100 term providers based on 2012 term insurance sales
Responses from 41 companies
Responses represented 62% of 2012 term sales
Assumption Survey Results
Overview
8
Assumption Survey Results
Product Distribution
x=5% x=15% x=30%
5 Year Term 1.20% 3 1 0
10 Year Term 22.70% 42 36 8
15 Year Term 8.60% 32 6 1
20 Year Term 52.00% 44 44 42
25-30 Year Term 15.10% 33 23 9
Other 0.40% 5 2 1
Level Premium Term Product Mix by Level Period
Product Level PeriodAggregate Distribution
for Respondents
Number of companies where product
represents at least x% of individual
company's term sales
9
Assumption Survey Results
Post-Level Premium Structure
Dominant Structure is Jump to an ART Scale
Product Structure Responses
Premium jump to ART 40
Premium grade to ART 4
Jump to new level period 3
Face amount decrease 1
Product terminates 2
Flexible Premiums (Term UL) 1
Post-Level Product Design
10
Assumption Survey Results
Expected Changes to Post-Level Premium Structure
No change expected for most term new business
Minimal changes reported to inforce term business
No change 23
Grade premiums into an ART scale over 'x' years 3
Other 3
Changes to Post-Level Premium Structure for Term New Business
DescriptionImplemented in
last 5 yearsConsidering
Lower post-level premiums 1 5
Grade into an ART scale 3 3
Other 0 2
Changes to Post-Level Premium Structure for Term In Force
11
Assumption Survey Results
Structure of Current & Guaranteed Ultimate Rates
Trend towards increasing level of guaranteed ultimate rates
Relationship between current and guaranteed continues to vary
Description 2013 2009
% of 1980 CSO
Between 150-300% 2 8
% of 2001 CSO
Less than 200% 5 1
Exactly 200% 7 12
Between 200-300% 11 7
Exactly 300% 11 12
More than 300% 7 1
Other 3 5
Structure of Guaranteed Ultimate Rates
Description 2013 2009
Product has Guaranteed Rates only 14 15
Current Rates = Guaranteed Rates 15 12
Current Rates < Guaranteed Rates 14 16
Current Rates grade to Guaranteed Rates 1 2
Relationship between Current Ultimate and Guaranteed Ultimate Premiums
12
Assumption Survey Results
Post-Level Lapse Rate Structure
Most companies assume only one shock lapse, generally grading
down thereafter
10-Year Term 20-Year Term
Duration L Shock, then level 8 7
Duration L Shock, then grade down 12 10
2 Shocks, Duration L>=L+1 6 5
2 Shocks, Duration L+1>L 3 2
Responses
Lapse Rate Trend By Duration
Description
13
Assumption Survey Results
Post-Level Lapse Rate Assumption by Premium Jump
Broad range of assumptions by premium jump
Trend is somewhat inconsistent with experience
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.0 10.0 20.0 30.0 40.0
Sh
oc
k L
ap
se
Premium Jump
Premium Jump vs. Shock Lapse10 Year term
14
Assumption Survey Results
Post-Level Lapse rate Assumptions
Today’s skewness assumptions do not always follow past experience
Response Respondents
Lapses are uniformly distributed 18
Lapses occur on premium payment modes 10
Lapses occur at the end of the year 7
Other 4
No response 5
Monthly Lapse Skewness During Level Premium Period
Response Respondents
Lapses are uniformly distributed 5
Lapses occur on premium payment modes 3
Lapses occur at the end of the year 17
Lapses graded toward end of the year with shock in month 12 12
No response 7
Monthly Lapse Skewness During Year of Shock Lapse
Response Respondents
Lapses are uniformly distributed 6
Lapses occur on premium payment modes 7
Lapses occur at the end of the year 9
Lapses skewed to the beginning of L+1, Uniform thereafter 8
No response 14
Monthly Lapse Skewness During Post-Level Period
15
Assumption Survey Results
Post-Level Mortality Methodology
Dukes-MacDonald (or derivatives) and Flat Multiple are the primary
assumptions used in developing mortality after the shock lapse
Method Responses
N/A - 100% shock lapse 9
Dukes-MacDonald or derivatives of Dukes-MacDonald 14
Canadian Institute of Actuaries Valuation Technique Paper #2 4
Flat Multiple 13
Other: Set by reinsurer/external consultant 3
Other: Internally developed method 3
Method of Developing Mortality Assumption
16
Assumption Survey Results
Post-Level Mortality Assumptions vs. Cumulative Lapse Rate
Mortality deterioration increases gradually with increasing lapse rates
0%
100%
200%
300%
400%
500%
600%
700%
800%
30% 40% 50% 60% 70% 80% 90% 100%
Mo
rtality
Dete
iora
tio
n D
ura
tio
n 1
2
Cumulative Lapse Rate Duration 10 through 11
Shock Lapse vs. Mortality Deterioration10 Year term
DM
CIA
Flat Multiple
Other
17
Assumption Survey Results
Post-Level Mortality Assumptions vs. Premium Jump
Mortality deterioration assumptions do not differ dramatically by
premium jump, inconsistent with experience
0%
100%
200%
300%
400%
500%
600%
700%
800%
900%
1000%
0.0 10.0 20.0 30.0 40.0
Mo
rtality
Dete
rio
rati
on
Du
rati
on
11
Premium Jump
Premium Jump vs. Mortality Deterioration10 Year term
18
Assumption Survey Results
Post-Level Mortality Assumptions vs. Prem Jump and Lapse Rate
When split by lapse rate, it does appear assumptions vary by
premium jump
0%
100%
200%
300%
400%
500%
600%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Mo
rtality
Dete
iora
tio
n D
ura
tio
n 1
2
Cumulative Lapse Rate Duration 10 through 11
Shock Lapse vs. Mortality Deterioration by Premium Jump10 Year term
1-10X Jump
11-20X Jump
20X Jump or Greater
19
Assumption Survey Results
Post-Level Mortality Deterioration Structure
Assumptions vary broadly by structure of mortality deterioration
10-Year Term 20-Year Term
Level all durations 7 6
Decreasing or grade to ultimate level 13 11
Increasing then decreasing 6 1
Generally increasing 1 3
Description
Mortality Deterioration Assumption Trend By Duration
Responses
20
Experience Study Results
21
2000-2012 Policy-Year Study
Lapse Study Experience Results
Overview
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
2014 Study 2010 Study Change
Number of Companies w/ Post Level Experience 36 25 144%
Post-Level Lapses with Premiums 436,307 87,544 498%
Post-Level Lapses without Premiums 258,030 170,171 152%
Number of Companies w/ Post Level Experience 15 7 214%
Post-Level Lapses with Premiums 37,673 12,191 309%Post-Level Lapses without Premiums 1,364 1,359 100%
10-Year Term
15-Year Term
22
Lapse Study Experience Results – Jump to ART
Lapse Rates by Duration
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
1
10
100
1,000
10,000
100,000
1,000,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
6 7 8 9 10 11 12 13 14 15 16+
Level Period Post-Level Period
Nu
mb
er
of
La
pse
s
Lap
se
Rate
T10 Jump to ART Lapse Rates by Duration
Lapse Rate Median Lapse Rate Number of Lapses (right axis)
23
Lapse Study Experience Results – Jump to ART
Lapse Rates by Premium Jump
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
1
10
100
1,000
10,000
100,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Nu
mb
er
of
La
pse
s
Lap
se
Rate
Duration 11/10 Premium Jump Ratio
T10 Jump to ARTDuration 10 Lapse Rate by Duration 11/10 Premium Jump Ratio
Duration 10 Lapse Rate Count Duration 10 Lapses
24
Lapse Study Experience Results – Jump to ART
Lapse Rates by Duration & Premium Jump
Duration 11/10
Premium Jump
Ratio Band
Dur 11 /
Dur 10
Dur 12 /
Dur 11
1.01x - 2x 79.5% 59.5%
2.01x - 3x 77.0% 46.7%
3.01x - 4x 73.2% 44.9%
4.01x - 5x 73.1% 42.7%
5.01x - 6x 70.5% 45.4%
6.01x - 7x 73.5% 45.1%
7.01x - 8x 73.2% 46.3%
8.01x - 9x 74.5% 45.9%
9.01x - 10x 78.7% 54.7%
10.01x-12x 79.1% 49.7%
12.01x-14x 79.7% 49.4%
14.01x-16x 82.9% 56.8%
16.01x + 86.0% 59.8%
Subtotal Prem Data Available 59.4% 42.9%
No Prem Data Available 55.0% 43.1%
Grand Total 57.4% 42.9%
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Duration 11/10 Premium Jump Ratio
T10 Jump to ART Lapse Rate
by Duration 11/10 Premium Jump Ratio
Duration 10 Lapse Rate Duration 11 Lapse Rate Duration 12 Lapse Rate
Dur 11 / Dur 10 Dur 12 / Dur 11
25
Lapse Study Experience Results
Lapse Rates by Duration and Issue Age
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0-19 20-29 30-39 40-49 50-59 60-69 70+
Lap
se
Rate
Issue Age
T10 Lapse Rates by Issue Age
Duration 6-9 Duration 10 Duration 11 Duration 12+
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
26
Lapse Study Experience Results
Lapse Rates by Duration & Gender
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Male Female
Lap
se
Rate
Gender
T10 Lapse Rates by Gender
Duration 6-9 Duration 10 Duration 11 Duration 12+
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
27
Lapse Study Experience Results
Lapse Rates by Duration & Risk Class
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Super-Preferred
Preferred Non-Preferred Undiff Preferred Non-Preferred Undiff
Non-Smoker Smoker Unk/Agg
La
ps
e R
ate
Risk Class
T10 Lapse Rates by Risk Class
Duration 6-9 Duration 10 Duration 11 Duration 12+
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
28
Lapse Study Experience Results
Lapse Rates by Duration & Face Amount Band
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
< $50,000 $50,000 -$99,999
$100,000 -$249,999
$250,000 -$999,999
$1,000,000 -$4,999,999
$5,000,000 +
La
ps
e R
ate
Face Amount
T10 Lapse Rates by Face Amount Band
Duration 6-9 Duration 10 Duration 11 Duration 12+
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
29
Lapse Study Experience Results
Lapse Rates by Duration & Premium Payment Mode
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Annual Semi-Annual Quarterly Monthly
La
ps
e R
ate
Premium Payment Mode
T10 Lapse Rates by Premium Payment Mode
Duration 6-9 Duration 10 Duration 11 Duration 12+
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
30
Lapse Study Experience Results – Jump to ART
Lapse Rates (duration 10) Premium Jump and Issue age
1
10
100
1,000
10,000
100,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
<40 40-49 50-59 60+ <40 40-49 50-59 60+ <40 40-49 50-59 60+ <40 40-49 50-59 60+ <40 40-49 50-59 60+
1.01x - 3x 3.01x - 5x 5.01x - 7x 7.01x - 9x 9.01x +
Nu
mb
er
of
La
ps
es
Lap
se
Rate
T10 Jump to ARTDuration 10 Lapse Rate
by Issue Age and Duration 11/10 Premium Jump Ratio
Duration 10 Lapse Rate Duration 10 Lapses
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
31
Lapse Study Experience Results – Jump to ART
Lapse Rates (duration 10) Premium Jump and Face Amount Band
1
10
100
1,000
10,000
100,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
<100k 100k-249k
250k-999k
1M+ <100k 100k-249k
250k-999k
1M+ <100k 100k-249k
250k-999k
1M+ <100k 100k-249k
250k-999k
1M+ <100k 100k-249k
250k-999k
1M+
1.01x - 3x 3.01x - 5x 5.01x - 7x 7.01x - 9x 9.01x +
Nu
mb
er
of
La
ps
es
La
ps
e R
ate
T10 Jump to ARTDuration 10 Lapse Rate
by Face Amount and Duration 11/10 Premium Jump Ratio
Duration 10 Lapse Rate Duration 10 Lapses
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
32
Lapse Study Experience Results
Skewness of Lapses
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
6-9 10 11 12+
Pro
po
rtio
n o
f L
ap
se
s
Duration
T10: Lapse Skewness by MonthCompanies without Grace Period Adjustments
12
11
10
9
8
7
6
5
4
3
2
1
33
Lapse Study Experience Results
Comparison of RGA to SOA
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
1.0
1 -
2.0
0
2.0
1 -
3.0
0
3.0
1 -
4.0
0
4.0
1 -
5.0
0
5.0
1 -
6.0
0
6.0
1 -
7.0
0
7.0
1 -
8.0
0
8.0
1 -
9.0
0
9.0
1 -
10
.00
10
.01
- 1
1.0
0
11
.01
- 1
2.0
0
12
.01
- 1
3.0
0
13
.01
- 1
4.0
0
14
.01
- 1
5.0
0
15
.01
- 1
6.0
0
16
.01
- 1
7.0
0
17
.01
- 1
8.0
0
18
.01
- 1
9.0
0
19
.01
- 2
0.0
0
20
.01
- 2
1.0
0
21
.01
- 2
2.0
0
22
.01
- 2
3.0
0
23
.01
- 2
4.0
0
24
.01
AN
D U
P
Lap
e C
ou
nt
Lap
se R
ate
Premium Jump
10-YR Term Duration 10 - Jump to ART
SOA LapseCount
RGA LapseCount
SOA LapseRate
RGA LapseRate
34
Mortality Study Experience Results
Overview
2000-2012 Calendar Year Study
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
2014 Study 2010 Study Change
Number of Companies w/ Post Level Experience 36 24 150%
Post-Level Claims with Premiums 2,651 382 694%
Post-Level Claims without Premiums 729 381 191%
Number of Companies w/ Post Level Experience 6 5 120%
Post-Level Claims with Premiums 432 145 298%
Post-Level Claims without Premiums 12 13 92%
10-Year Term
15-Year Term
35
Mortality Study Experience Results – Jump to ART
Mortality by Duration
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
1
10
100
1,000
10,000
0%
50%
100%
150%
200%
250%
300%
350%
400%
450%
6 7 8 9 10 11 12 13 14 15 16+
Level Period Post-Level Period
Nu
mb
er
of
De
ath
s
Mo
rtali
ty R
ela
tive
to
Du
rati
on
s 6
-10
T10 Jump to ARTMortality by Duration
A/E Mortality Median A/E Mortality Number of Deaths (right axis)
36
Mortality Study Experience Results – Jump to ART
Mortality by Duration & Premium Jump
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
1
10
100
1,000
10,000
0%
200%
400%
600%
800%
1000%
1200%
1400%
1600%
1800%
2000%
1.01x -2x
2.01x -3x
3.01x -4x
4.01x -5x
5.01x -6x
6.01x -7x
7.01x -8x
8.01x -9x
9.01x -12x
12.01x-15x
15.01x-18x
18.01x +
Nu
mb
er
of
De
ath
s
Mo
rtali
ty R
ela
tive
to
Du
rati
on
s 6
-10
Duration 11/10 Premium Jump Ratio
T10 Jump to ARTDuration 11 Mortality Relative to Level Period
by Duration 11/10 Premium Jump Ratio
Relative Mortality Ratio Number of Deaths (right axis)
37
Mortality Study Experience Results – Jump to ART
Mortality by Duration & Issue Age
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
50%
100%
150%
200%
250%
300%
350%
400%
0-29 30-39 40-49 50-59 60+
Issue Age
T10 Jump to ART2008 VBT Mortality Ratios
by Issue Age
Duration 6-10 Duration 11+ Post-Level Relative to Dur 6-10
38
Mortality Study Experience Results – Jump to ART
Mortality by Duration & Gender
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
50%
100%
150%
200%
250%
300%
350%
Male FemaleGender
T10 Jump to ART2008 VBT Mortality Ratios
by Gender
Duration 6-10 Duration 11+ Post-Level Relative to Dur 6-10
39
Mortality Study Experience Results – Jump to ART
Mortality by Duration & Risk Class
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
50%
100%
150%
200%
250%
300%
350%
400%
450%
500%
550%
Super-Preferred
Preferred Non-Preferred Undiff Preferred Non-Preferred Undiff
Non-Smoker Smoker Unk/Agg
Risk Class
T10 Jump to ART2008 VBT Mortality Ratios
by Risk Class
Duration 6-10 Duration 11+ Post-Level Relative to Dur 6-10
40
Mortality Study Experience Results – Jump to ART
Mortality by Duration & Face Amount Band
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
0%
50%
100%
150%
200%
250%
300%
350%
400%
450%
500%
< $50,000 $50,000 -$99,999
$100,000 -$249,999
$250,000 -$999,999
$1,000,000 +
Face Amount Band
T10 Jump to ART2008 VBT Mortality Ratios
by Face Amount Band
Duration 6-10 Duration 11+ Post-Level Relative to Dur 6-10
41
Mortality Study Experience Results
Mortality by Monthly Duration
0
25
50
75
100
125
150
175
0%
50%
100%
150%
200%
250%
300%
350%
1 2 3 4 5 6 7 8 9 10 11 12
Nu
mb
er
of
De
ath
s
20
08
VB
T M
ort
ali
ty R
ati
os
Months since start of Duration 11
T10 Duration 112008 VBT Mortality Ratios by Month
Companies without Grace Period Adjustments
2008 VBT Mortality Ratios Number of Deaths (right axis)
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
42
Mortality Study Experience Results – Jump to ART
Mortality Deterioration by Shock Lapse
0%
150%
300%
450%
600%
750%
900%
1050%
1200%
1350%
1500%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Du
rati
on
11 M
ort
ality
Re
lati
ve
to
Du
rait
on
s 6
-10
Duration 10 Shock Lapse
T10 Jump to ARTDur 10 Shock Lapse vs. Dur 11 Mortality Deterioration
by Company
>=35 claims
<35 claims
Expon. ()
Source: http://www.soa.org/Research/Experience-Study/Ind-Life/Persistency/research-2014-post-level-shock.aspx
43
Mortality Study Experience Results – Jump to ART
Mortality Deterioration by Shock Lapse
0
20
40
60
80
100
120
140
160
0%
200%
400%
600%
800%
1000%
1200%
1.01 -3.00
3.01 -5.00
5.01 -7.00
7.01 -9.00
9.01 -11.00
11.01 -13.00
13.01+
Cla
im C
ou
nt
A/E
08
VB
T
Premium Jump
10-YR Term Duration 11 - Jump to ART
RGA ClaimCount
SOA ClaimCount
RGA A/E
SOA A/E
44
Predictive Model
45
Multivariate Lapse Rate Model – T10 Duration 10 Shock Lapse
Model:
Generalized Linear Model (GLM)
Target variable follows distribution in the exponential familly
Response variable = observed lapse count
Follows a Poisson distribution
Benefits:
Elimination of possible bias from a uni-variate approach
Systematic way of controlling lapse assumption complexity
Transparent insight into true drivers of lapse rates
Distribution of target variables
Predictive Model
46
Predictive Model – Model 1 Model Parameter
Intercept - -2.676 <2.0E-16
Issue Age Numerical 0.0552 <2.0E-16
(Issue Age)^2 Numerical -0.000316 <2.0E-16
Risk Class Categorical
Super-Pref NS 0 1.00 0 11.3% 82.6% 82.6% 100.1%
NS -0.06736 0.93 <2e-16 76.9% 69.1% 68.8% 100.4%
SM -0.01296 0.99 0.0339 11.8% 63.5% 63.5% 100.0%
Face Amount Categorical
<50K 0 1.00 0 1.2% 14.9% 17.9% 83.2%
50-100K 0.600 1.82 <2.0E-16 3.9% 61.8% 62.7% 98.5%
100K-250K 1.262 3.53 <2.0E-16 51.9% 67.5% 67.4% 100.1%
250K-1M 1.559 4.75 <2.0E-16 36.9% 74.4% 73.8% 100.8%
>1M 1.585 4.88 <2.0E-16 6.1% 80.7% 80.4% 100.4%
Premium Mode Categorical
Annual 0 1.00 0 22.9% 85.8% 85.3% 100.5%
Semi/Quarter -0.1365 0.87 <2.0E-16 24.5% 74.9% 75.1% 99.7%
Monthly/BiWeekly -0.3506 0.70 <2.0E-16 45.4% 56.9% 56.6% 100.5%
Other/Unknown -0.08481 0.92 <2.0E-16 7.2% 85.3% 84.8% 100.7%
Cross Term Mixed
Issue Age:Face Amt <50 0 0
Issue Age:Face Amt 50-100K -0.001288 0.3162
Issue Age:Face Amt 100-250K -0.01074 <2e-16
Issue Age:Face Amt 250K-1M -0.01653 <2e-16
Issue Age:Face Amt >1M -0.0172 <2e-16
Validation of Results
Data
Proportion
Actual
Lapse
Rate
Predicted
Lapse
Rate
Variable Type P-ValueCoefficientActual /
PredictedFactor
47
Predictive Model – Model 1
0
10,000
20,000
30,000
40,000
50,000
60,000
70,000
80,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0-19 20-29 30-39 40-49 50-59 60-69 70+
Exp
osu
re
Lap
se R
ate
Issue Age
Model Predicted vs. Actual Lapse Rate
Exposure
Predicted
Actual
48
Predictive Model – Model 2
0
2,000
4,000
6,000
8,000
10,000
12,000
14,000
16,000
18,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Exp
osu
re
Lap
se R
ate
Premium Jump
Model Predicted vs. Actual Lapse Rate
Exposure
Predicted
Actual
0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0-19 20-29 30-39 40-49 50-59 60-69 70+
Exp
osu
re
Lap
se R
ate
Issue Age
Model Predicted vs. Actual Lapse Rate
Exposure
Predicted
Actual
49
Acknowledgements
50
Contributing Companies
SOA and SOA Staff
Project Oversight Group Jeff Beckley
Brian Carteaux
David Wylde
RGA Contributors Derek Kueker
Tim Rozar
Michael Cusumano
Susan Willeat
Tatiana Berezin
Tony Phipps
Richard Xu
Scott Rushing
Minyu Cao
Kathleen Broom
Acknowledgements
Special Thanks
51