Workers’ Compensation Combined Ratio
Continues to Deteriorate
Underwriting
Expense
17. 6
18. 5
19. 8
20. 4
21. 7
23. 3
25. 4
25. 9
26. 7
28.0
26. 4
26. 1
23. 5
22. 2
22. 1
22. 3
19. 6
24. 6
24. 6
26. 2
26. 5
From clarifying underwriting and
premium auditing accuracy to detecting
precursors for unnecessary claim costs,
this multivariate analysis is improving
systems and uncovering savings in
compelling ways.
Property/casualty insurers, in general,
increasingly report benefits from its
use in underwriting, pricing, rating,
and market segmentation, according to
Towers Watson’s third annual predictive
modeling survey released in February
2012. More than 75 percent of the survey’s
60 U.S. and nine Canadian respondents
are enjoying bottom-line benefits of rate
accuracy, loss ratio improvement, and
higher profitability.
Nearly half of the respondents—
49 percent—cited a positive impact due to
expanding underwriting appetite. These
numbers are some 10 percent to 20 percent
higher than last year’s results, showing
“the enduring sustainable benefits” of
predictive modeling, the survey said.
Of the carriers that participated in the
2011 Towers Watson survey, 48 percent
of workers’ compensation insurers said
they are already using predictive modeling. An additional 36 percent plan to use
predictive modeling, while 16 percent do
not (see table below).
Where there are enough reliable data,
there’s an opportunity to apply predictive
modeling. Most of the largest workers’
comp insurers began using predictive
modeling in premium auditing, along
with underwriting, pricing, and marketing, about 10 years ago, said Peter Wu,
a director at Deloitte Consulting LLP.
Property/casualty carriers using predictive modeling for market segmentation are saving two to four
points in their loss ratio and loss adjustment expenses (LAE) after the first year of implementation, said Towers Watson director
Calendar Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
20-year
average,
1991—
2010
66.9 14. 1
10-year
average,
2001—
Source: national Council on Compensation Insurance Inc. (nCCI)
Loss
83.8
87.8
83.9
71.7
60. 5
55. 3
55. 8
55. 5
60. 1
65.9
71.2
78.0
70.8
70.7
68.9
64. 5
58. 7
60. 1
60. 3
67.6
71.0
LAE
10. 7
11. 5
13. 2
12. 4
13. 1
12. 5
13. 7
13. 8
15. 3
15. 8
15. 9
13. 8
13. 7
15.0
14. 5
14. 4
13. 6
14. 5
14. 2
14. 9
16.0
Dividends
5. 1
4. 8
4. 4
4. 7
6. 4
6.0
4. 8
5. 4
5. 3
5. 6
4. 7
3. 7
2. 8
1. 6
1. 3
1. 3
1. 3
1. 5
1. 8
1. 6
1. 5
Combined
Ratio
117.2
122.6
121.3
109.2
101.7
97.1
99.7
100.6
107.4
115.3
118.2
121.6
110.8
109.5
106.8
102.5
93.2
100.7
100.9
110.3
115.0
AVg
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.3
108.2
107.1
Workers’ Compensation Predictive Modeling
On the Rise
Year
Currently
Use Plan to Use
2009 18% 43%
2010 32% 55%
2011 48% 36%
Source: towers Watson predictive modeling survey, February 2012
No Plans
to Use
39%
13%
16%
Brian Stoll, who heads up the annual survey. By improving efficiency, the process also can reduce underwriting expenses.
The private workers’ comp insurance market in general
could use such reductions. The loss ratio has been steadily
creeping up to 71.0, according to NCCI, with the 10-year average being 67.1. At the same time, the LAE for 2010 is 16.0—the
highest it has been in at least 20 years, with the average 10-
year LAE being 14. 5.
There are other benefits to predictive modeling besides cost
savings. Zurich Financial Services, which has been using predictive modeling for workers’ comp underwriting since 2008,
is an example of just one carrier that has seen many benefits
from applying predictive modeling.
