Even with the NCCI-mandated shift in defining hazard groups, the potential for workers' compensation insurers for better risk management is great, and granular pricing of class codes in the excess layer is a way forward.

1. Granular Pricing of Workers’ Compensation
Risk in Excess Layers
Identifying risk at a granular level and pricing it appropriately will put
carriers on a path to sound underwriting ability.
Executive Summary
The net written premiums for the workers’
compensation industry, and its share in the
overall commercial lines of business, have steadily
declined between 2007 and 2010. Net written pre-
miums have plummeted 33% from $48 billion at
year-end 2006 to $32 at year-end 2010. In 2011,
the industry-wide loss ratio for workers’ compen-
sation insurance in the U.S. increased to 118.1%.1
This is the highest level in more than a decade.
The combined ratio for workers’ compensation
has increased steadily from a value of 100% in
2006 to 119% in 2011 (see Figure 1, on next page).
These numbers are influencing insurance carriers
to look for creative pricing methodologies.
Pricing for workers’ compensation risk is heav-
ily regulated by state regulatory authorities and
offers less scope for pricing innovation in the
primary layer. However, pricing in the excess
layer is not as heavily regulated as pricing in
the primary layer and hence provides an oppor-
tunity for carriers to gain that competitive
advantage by making the pricing more effective.
The option to be creative in pricing and the
increase in the insured’s buying trend for excess
insurance makes excess lines pricing an ideal
place for insurers to concentrate. The onus is now
on carriers to develop the competency for clas-
sifying risk at a detailed level, which would enable
them to price products more efficiently.
Granular pricing would help carriers identify
specific components of risk that should be priced
in house, components that should be ceded to
reinsurers and components that should be
avoided. It would also help carriers to identify
niche areas of risk and develop appropriate under-
writing and pricing techniques to cover them.
One of the integral components in determin-
ing the premium for the workers’ compensation
line of business (LOB) is excess loss factor (ELF),
which is the ratio of expected losses in excess of
a limit to the total expected losses. Classifications
that use the same ELFs are grouped together
to form hazard groups. The National Council
on Compensation Insurance (NCCI) periodically
publishes ELFs by hazard group for select limits.
NCCI implemented a seven-hazard-group system
in 2007, replacing the previous four-hazard-group
system. This move was the result of the review
of the hazard group mapping by NCCI, which
cognizant 20-20 insights | june 2013
• Cognizant 20-20 Insights

2. cognizant 20-20 insights 2
Source: “2007 Hazard Group Mapping,” NCCI Research Publication
Figure 2
Distribution of Classes by Prior Hazard Groups (Left) and by New Hazard
Groups (Right)
While implementing the
seven hazard groups sys-
tem, NCCI offered an alter-
native collapsed new map-
ping for insurers who did
not want to switch to the
new system immediately.
In the collapsed mapping,
hazard groups A and B were
combined to form hazard
group 1; hazard groups C
and D were combined to
form hazard group 2; hazard
groups E and F were com-
bined to form hazard group
3 and hazard group G was
made hazard group 4.
The collapsed mapping provides a better platform
for comparison between the prior mapping and
the new mapping structure. If we compare the
prior mapping with the collapsed new mapping,
concluded that there was a need for more
granular classification of risk. The authors of this
white paper believe that there is scope and need
for further refinement in excess rating method-
ology by determining and using ELFs at a more
granular level.
Rate Making Using Hazard Groups
Rate making in workers’ compensation is based
on hazard groups, which are groupings of class
codes. NCCI has defined approximately 900 work-
ers’ compensation class codes and four or seven
hazard groups, for which it provides rate making
service. In the four (I-IV) hazard group mapping
system, the bulk of the exposure is concentrated
in two hazard groups. Hazard groups II and III con-
tained 97% of the total premium (see Figure 2).
As the figure indicates, after the implementation
of the seven hazard group system there is a more
homogenous distribution of premiums by hazard
groups, which improved pricing accuracy.
The industry
has benefited by
moving to the
seven hazard
groups system,
as this system
has distributed
premium values
more evenly. But
the opportunity
to improve is still
immense.
NCCI Hazard
Group
Number of
Classes
Percent of Total
Premium
NCCI Hazard
Group
Number of
Classes
Percent of To-
tal Premium
I 38 1% A 55 9%
II 428 46% B 241 17%
III 318 51% C 160 21%
IV 86 2% D 45 10%
E 224 19%
F 57 19%
G 88 5%
Source: “Best’s Aggregates & Averages - Property/Casualty,” A.M. Best, 2011.
Figure 1
85
95
105
115
125
0
20,000
40,000
60,000
2006 2007 2008 2009 2010 2011
Net Written Premium ($ Mn) Combined Ratio (%)
NWP and CR for Workers’ Compensation Line of Business

3. cognizant 20-20 insights 3
hazard group 1 (A and B) has substantial portion
of total premium value compared with hazard
group I of prior mapping (see Figure 3). Hazard
groups 2 (C and D) and 3 (E and F) have become
slightly smaller compared with their peers in the
prior mapping. Hazard group 4 (HG G) has slightly
more premium value than prior hazard group IV.
Granular Pricing Through Hazard
Groups Decomposition
The industry has benefited by moving to the
seven hazard groups system, as this system has
distributed premium values more evenly. But the
opportunity to improve is still immense. With the
seven group mapping, 69% of premium value is
still distributed between four (HG C, D, E, F) of the
seven hazard groups. Also, the ELFs as provided
by NCCI are applicable for countrywide class
codes. The ELFs should ideally be determined for
class codes at the state level for applicable limits
as one hazard group can be more hazardous in
one state as compared with other states.
If we look at the movement of classes between
the prior four mapping system to the collapsed
new mapping system, the great majority (300
classes and 37% of premium) moved down one
hazard group (see Figure 4).
However it should be noted
that the implementation of
new mapping was revenue
neutral for carriers as there
was a general increase in
the value of ELFs between
the prior mapping and the
new mapping.
Even with the seven haz-
ard group model, there are
numerous classes that have
experiencedadisproportion-
ate number of catastrophic
claims that might be inappropriately mapped to
a lower severity hazard group. If the carrier is
writing risk for these class codes that are placed
Granular pricing
in the excess layer
would ensure
better estimation
of overall expected
losses, which
would improve the
combined ratio of
the insurer in the
long run.
HG I, 1%
HG II, 46%
HG III, 51%
HG IV, 2%
HG I, 26%
HG 2, 31%
HG 3, 38%
HG 4, 5%
Source: “2007 Hazard Group Mapping,” NCCI Research Publication
Figure 3
Prior Mapping (Left) Versus Collapsed New Mapping (Right); Percent of
Premium by Hazard Group
Source: “2007 Hazard Group Mapping,” NCCI Research Publication
Figure 4
Comparison of Prior Mapping with Collapsed New Mapping with Respect to
Movement to Class Codes
63.5%
1.7%
34.5%
0.3%
0
100
200
300
400
500
600
No Movement Up 1 HG Down 1 HG Down 3 HGs
552 15 300 3
NumberofClassCodes

4. cognizant 20-20 insights 4
on the lower end of the hazard groups, it may have
significantly more excess loss exposure than
anticipated and hence the pricing might be
deficient. To further enable
homogeneous distribution
of premiums within the
excess layers, ELFs should
be referenced at the indi-
vidual class code rather than
the hazard group level. The
ELFs at the class code level
for every state can be deter-
mined by the historical infor-
mation on severity and fre-
quency of losses that each
class code had in the state.
An insurer that switched
from hazard group to class
code level pricing recently
compared the underwriting
results of deals between the
two systems. The insurer
found approximately $300
million in extra expected
losses in the excess layer that were not taken
into consideration with the earlier hazard group
pricing. Granular pricing in the excess layer would
ensure better estimation of overall expected
losses, which would improve the combined ratio
of the insurer in the long run.
Commoditization of personal insurance has been
assisted with sophisticated granular pricing tech-
niques. Robust underwriting through risk classi-
fication at a granular level in commercial lines is
key to survival for insurers.
Impact to the Carriers
• Process change: The decomposition of
hazard groups will introduce a larger set of
factors for actuaries to maintain. Underwrit-
ers/actuaries will need to adapt to the new
process of reviewing and analyzing pricing
worksheets at a deeper level for granular
pricing. The amount of time required by
actuaries will increase, with a corresponding
increase in the accuracy of the price.
• Algorithm logic/calculation mechanism:
Excess loss factors are used to calculate loss
rating limit and expected losses. The move of
assigning ELFs to each individual class code
will impact the algorithm logic of calculating
expected losses and loss rating limit.
For example, assume that a prospect has
20 class codes in its risk portfolio and that
all 20 class codes were previously mapped
to hazard group II. With the earlier pric-
ing methodology, the respective algorithm
logic for calculation of expected losses
and loss rating limit used only one ELF
assigned to hazard group II; with the granu-
lar pricing discussed in this white paper, the
carrier would use 20 ELFs, each assigned to an
individual class code.
The actuarial algorithms to calculate loss rat-
ing limit and expected losses might reside in
stored procedures or within the application
itself. A thorough impact analysis is needed to
determine if the existing processing logic will
be able to support the increased size of the
reference database. Therefore, the carrier’s
rating, pricing and policy issuance systems
should be adapted to support the new data
structure.
The sizeable increase in the support database
will require performance reengineering to
prevent the potential slowdown of the calcula-
tion mechanism.
• Actuarial models: These models are usually
embedded within the carrier’s policy adminis-
tration system and receive input from the esti-
mated losses/loss rating module and provide
output for calculation of the final premium of
the policy. Implementation of granular pricing
methodology can impact the calculation per-
formed by these models handling credit, profit
and aggregate charges. The models will need
to be modified to cope with the new reference
database created to enable granular pricing.
In addition to the modifications to the model,
an impact analysis is needed on all modules
directly interacting with the models that input
data into the model or use the output provided
by the model to ensure that they are intact and
are able to support the modifications made to
the model itself. Also, the actuarial reporting
systems will need to be adapted if they report
data at the hazard group level.
• In-flight change: By changing the pricing
methodology, underwriters will need to set
the correct expectations with brokers/clients.
For any policy that has already been quoted,
pricing it again using the new mechanism
A thorough impact
analysis is needed
to determine if the
existing processing
logic will be able
to support the
increased size
of the reference
database. Therefore,
the carrier’s rating,
pricing and policy
issuance systems
should be adapted
to support the new
data structure.

5. cognizant 20-20 insights 5
might change the premium values — which
might be difficult to explain to the broker.
Policies that have already been quoted or have
reached a particular stage of underwriting
should not be automatically impacted while
the granular pricing functionality is imple-
mented in the system. Underwriters should
be given the choice, in the system, to be able
to use the old mechanism or new mechanism
to price the in-flight policies. This dual pric-
ing can be achieved by implementing a simple
solution of factor-date logic at the instance of
calculation of premiums where the system
would provide the underwriter with the option
of using previous factors or new factors. The
database would need to store both sets of
factors for a certain period of time and, if the
database does not have time-stamps, it will
need to be implemented to support application
of factor-date logic.
Any policy that has already been quoted sho-
uld automatically follow the old mechanism
if repriced in the system to avoid any adverse
effect on the quote. Any new policy priced in
the system should automatically follow the
new mechanism.
Competitive Differentiation
With intense competition and the intent to
place risks more accurately, there is a need for
insurance carriers to optimize workers’ compen-
sation pricing. Carriers are trying to improve their
underwriting practices to achieve a distinguished
position. Identifying risk at a granular level and
pricing it appropriately will put carriers on a path
to sound underwriting ability.
Several advantages of this approach include:
• More accurate pricing: With better insight
into risk classification, carriers would be able
to determine appropriate expected losses in
excess layers for every state and hence would
be able to price in risk more accurately than
before.
• Ability to assume higher levels of risk: With
the higher accuracy of risk classification, there
will be some hazard groups, for specific states,
that will move down to a lower hazard group.
This will make the hazard less hazardous,
so to speak, and make it assumable for the
insurance carrier.
• Ability to avoid insufficiently priced risk:
Insurance carriers will be able to identify poor
risks easily and thereby avoid them, or engage
in appropriate risk transfer mechanisms.
Classes that have exhibited the propensity to
produce higher or more frequent losses can
thus be avoided.
As with the advancements in any field, we believe
that this functionality will become a hygiene
factor in the years to come. Insurance carriers
proactively adopting this approach will obtain a
competitive edge.
References
• “Revised Hazard Group Assignments,” National Council on Compensation Insurance Actuarial
Committee Agenda, Item ACT-92-57, April 20, 1993.
• John P. Robertson, “2007 Hazard Group Mapping,” National Council on Compensation Insurance
Research Paper.
Footnote
1
“Best’s Aggregates & Averages - Property/Casualty,” A.M. Best, 2011.

6. About Cognizant
Cognizant (NASDAQ: CTSH) is a leading provider of information technology, consulting, and business process
outsourcing services, dedicated to helping the world’s leading companies build stronger businesses. Headquartered
in Teaneck, New Jersey (U.S.), Cognizant combines a passion for client satisfaction, technology innovation, deep
industry and business process expertise, and a global, collaborative workforce that embodies the future of work.
With over 50 delivery centers worldwide and approximately 162,700 employees as of March 31, 2013, Cognizant is a
member of the NASDAQ-100, the S&P 500, the Forbes Global 2000, and the Fortune 500 and is ranked among the
top performing and fastest growing companies in the world.
Visit us online at www.cognizant.com for more information.
World Headquarters
500 Frank W. Burr Blvd.
Teaneck, NJ 07666 USA
Phone: +1 201 801 0233
Fax: +1 201 801 0243
Toll Free: +1 888 937 3277
Email: inquiry@cognizant.com
European Headquarters
1 Kingdom Street
Paddington Central
London W2 6BD
Phone: +44 (0) 207 297 7600
Fax: +44 (0) 207 121 0102
Email: infouk@cognizant.com
India Operations Headquarters
#5/535, Old Mahabalipuram Road
Okkiyam Pettai, Thoraipakkam
Chennai, 600 096 India
Phone: +91 (0) 44 4209 6000
Fax: +91 (0) 44 4209 6060
Email: inquiryindia@cognizant.com
­­© Copyright 2013, Cognizant. All rights reserved. No part of this document may be reproduced, stored in a retrieval system, transmitted in any form or by any
means, electronic, mechanical, photocopying, recording, or otherwise, without the express written permission from Cognizant. The information contained herein is
subject to change without notice. All other trademarks mentioned herein are the property of their respective owners.
About the Authors
Ramanujam Venkatesan is a Senior Consulting Manager within Cognizant Business Consulting’s Insur-
ance Practice. He specializes in the property and casualty space and has worked with multiple carriers
across North America, Europe and Asia-Pacific. Ramanujam has a post graduate degree in management
from the Indian Institute of Management (IIM) Indore and has a bachelor’s degree in engineering from
the University of Madras. He also holds certifications from the American Institute for Chartered Property
and Casualty Underwriters (AICPCU). He can be reached at Ramanujam.Venkatesan@cognizant.com.
Shishir Narula is a Senior Consultant within Cognizant Business Consulting’s Insurance Practice. He
specializes in commercial insurance underwriting and policy administration for workers’ compensation,
auto, general liability and excess lines of business. He holds an M.B.A. from Institute of Management
Technology (IMT) Ghaziabad and a bachelor’s degree in engineering from Indraprashtha University,
Delhi. He also holds certifications from the American Institute for Chartered Property and Casualty
Underwriters (AICPCU). He can be reached at Shishir.Narula@cognizant.com.