Description
Organizations seek to accelerate decision cycles and reduce the time to insight generation.
Actuarial Modernization
and Business Intelligence:
Driving Transformation
RPM Seminar Dallas
Casualty Actuarial Society
March 2015
www.pwc.com
PwC - Copyright & All Rights Reserved
Agenda – Modernization and Business Intelligence
1. Introduction and Context
2. Data Structure and Technology
3. Tools and Outputs
4. Goal Alignment
5. User Application
6. In Conclusion
7. Q&A
2
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
3
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
March 2015 Actuarial Modernization and Business Intelligence
4
Sophistication
Speed
Slow Fast
Low
High
Future Decisions
Today’s Decisions
Source: PwC's Global Data & Analytics Survey 2014. Sophistication measures the reliance on data, the inputs used in the decision,
management preparedness scale etc. Speed measures how quickly a decision is/can be made and implemented in an organization.
Organizations seek to accelerate
decision cycles and reduce the
time to insight generation.
As innovation is enabled,
organizations increase their ability
to identify new, sophisticated
decisions which are required to
address market demands.
As cultures transform, front-line
managers are empowered to
make faster, consistent decisions.
Only by embedding data and
analytics can organizations be
prepared for the future.
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
Insurers today are pursuing ambitious,
cross functional modernization initiatives.
Business Intelligence (BI) is at the
nexus of these efforts due to its:
?Integration with IT infrastructure
?Wide stakeholder base
?Ability to improve business
decision making
5
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
Business intelligence rests on a simple concept:
Timely, accurate, relevant, digestible data leads to better
business decision making.
In actuarial and insurance processes, this occurs sub-optimally due to:
? Technology and data production issues
? Opaque data presentation
? Cultural resistance to distributing actuarial products
The time is now to unlock BI benefits by solving these challenges.
New data sources and tools address—and complicate—the task.
“What gets measured gets managed.”
-Peter Drucker
6
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Challenges
BI Technology challenges
?Transaction Process System (TPS) datasets
?Irreconcilable sources
?Access to “unsafe” fields
?Extract, Transform & Load (ETL) filters/
cleansing reduce traceability
?Unintuitive user interfaces
?Lack of prospective information
?Limited alignment with plan or business goals
?Too much!
?Low training and user ability
Warehouse
Data
Marts
Staging
Tables
BI ETL
BI User
Interface
7
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
• Collaboration across RAFT: Risk,
Actuarial, Finance, Technology
• Iterative versus “Big Bang”
Goal
Alignment
Actuarial Modernization & Business Intelligence
Overview: Challenges
Despite the myriad
symptoms, key
challenges fall into
four broad
categories:
• Non-reconciling parallel sources
• Less is more
• Prioritize and drill down
(top view requires choices)
Data
Structure
• Trust and training
• Leverage actuarial viewpoint—
or users will form their own
User
Application
• Tools ? Solutions
• Front end enables analysis—
not just production
Tools and
Outputs
8
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
9
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Context
Time spent by senior reserving
actuaries on data processing,
manipulation and reconciliation:
Is there a dedicated IT resource
supporting data extraction and
provision to the Actuarial team?
59%
29%
10%
2%
below 10% 10% to 25%
25% to 50% >75%
33%
67%
Yes No
Source: PwC Actuarial Effectiveness Survey, 2013
10
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Context
What are common sources of data issues you encounter?
0% 10% 20% 30% 40% 50% 60% 70%
Other
Inconsistency in information provided by agents/brokers
Inconsistent reporting between MGAs
Timing lag in data received from MGAs
General Ledger does not easily reconcile
IT function resource constraints
Multiple sources of actuarial data
Manual off-system adjustments
Multiple legacy systems
Source: PwC Actuarial Effectiveness Survey, 2013
11
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Sources through uses
BI systems evolved to serve many disparate users.
Complexity compels architects to make choices—less can be more.
Internal Audit Actuarial Risk/Capital Mgmt Finance
SOx
Capital
Allocation
Pricing
Loss
Reserves
Risk
Audit
Results
Solvency
Warehouse
12
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Sources through uses
BI tools often evolve incrementally
via “add on” requests.
Excess information is produced as
companies “boil the ocean.”
More information does not
necessarily lead to better decisions.
Programs lacking strategic vision
can benefit from narrower scope.
“We have multiple
committees—each with
their own metrics. We go
round the figures three
times.”
Multiple
versions
“Sifting through our packs
could take 3 weeks—leaving no
time to actually do anything.”
Too
detailed
“My guys still get pulled
into ad hoc data requests
twice a day.”
Not
intuitive
Make decisions—
not reconciliations.
13
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
The information environment journey
March 2015 Actuarial Modernization and Business Intelligence
14
EDW into
Multiple
Data Marts
• Generalized to problems
• Several routes to information
fragments and shallows BI
• Many solutions for many
individualized problems
• Reporting layers opaque
Turbulent information flows
• Focused on decisions
• Single data viewing framework
combats fragmented BI
• Single major decision making
process platform
• Transparent and credible—ties to
reporting framework
Smooth delivery momentum
Analytics Mart
and Supporting
Platform
Decision
focused
Problem Focused Decision Focused
PwC - Copyright & All Rights Reserved
Data Structure and Technology
The information environment journey
March 2015 Actuarial Modernization and Business Intelligence
15
Enterprise Data
Warehouse
Multi-Mart Production
& Reporting
Analytics-Oriented
Decision Making
• Enterprise-wide solution
with centralized storage and
access, but providing a
backwards-looking view
• Limited agility for rapidly
changing conditions
• Typically requires large-
scale IT projects, with long
lead times
• BI: “Check what’s there”
• Insight and decision based
• Aligns to business strategy and
user adoption
• Enhances information flows and
analytics tools
• Tailored to analytics models,
uses, and production process
• Speeds organizational agility
• BI: “Use our best insights”
EDW
AM
3PD
• Purpose built repositories,
designed for specific task(s) or
business unit(s)
• Enables run-the-business,
incremental efficiencies
• Customizable, but data mart
proliferation increases technology
portfolio complexity
• BI: “This is what we use”
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Development approach
Top-down strategic BI benefits:
? 80/20 Rule: Get the most
important information right
? Align disparate specialists on
key objectives
? Focus on business impact—
not data issues
? Free resource time from
reconciling differences
Left to Right Right to Left
Collect and
manage data
Analyze data
Draw insights
Make decisions
What decision
is needed?
What insight
will help?
What data
answers this
question?
Manage
targeted data
16
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Development approach
BI users must help development teams avoid the roadblocks which
typically bog down IT projects.
Less is more. Restricting BI data can improve alignment and
efficiency—of both development teams and users.
Left to Right
•Traditional “heavy” design
provides all that is possible
•Includes unneeded functionality
•Extra overhead increases
development complexity
•Complexity slows execution
•Permanent solution
Right to Left
•“Light” development provides
users only what is needed
•Needs specified by analytics
•Sandbox to experiment with
new data or processes
•Lower project complexity
•Incremental solution
17
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Tailor information to the audience
R
i
s
k
M
a
n
a
g
e
m
e
n
t
H
u
m
a
n
R
e
s
o
u
r
c
e
s
I
T
&
O
p
e
r
a
t
i
o
n
s
C
l
a
i
m
s
&
C
e
d
e
d
R
e
i
n
s
u
r
a
n
c
e
I
n
t
e
r
n
a
l
A
u
d
i
t
&
C
o
m
p
l
i
a
n
c
e
M
a
r
k
e
t
i
n
g
U
n
d
e
r
w
r
i
t
i
n
g
&
D
i
s
t
r
i
b
u
t
i
o
n
F
i
n
a
n
c
e
A
c
t
u
a
r
i
a
l
Board Pack
Functional Area Metrics
Macro
Metrics
Individual
BU Metrics
Individual
Functional
Area Metrics
Business Unit Metrics
Level 1
Board Committees
Level 2
Committees,
Entities
Level 3
Functional
Review
18
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Fast, Widespread Information
Better data structures benefit power users and the wider organization.
Business
Intelligence
Data Governance
? Fast, reliable information source
? Scrubbed data aligns with operations
? Reduces reliance on Power Users
? Promotes data based decision making
? Data issues fixed at the source, rather
than continually adjusted for by users
? Better IT understanding of business uses
? Tighter feedback loops when issues arise
S
o
u
r
c
e
:
T
a
b
l
e
a
u
19
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
20
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
Wouldn’t it be great if…
CRO
“…we had a measure of risk
adjusted return on capital.
We don’t allocate capital
effectively, and we don’t
have clearly articulated risk
appetite metrics.”
Chief Operating Officer
“…I could advise management
on the cost of a new product, a
claim or a customer phone
call.”
Director of Investor
Relations
“… I could get the
information and insights I
need to be able to explain
our results to the market.”
CEO
“…I had a clear profitability view
across:
• Markets
• Channels
• Products
and could relate that back to capital.”
Distribution Director
“…it was clear why I sell a
significant amount of product
and then get castigated for
utilising capital and delivering
low margins.”
Claims Director
“…I could understand
why the claims ratio is
tracking upwards faster
than the market trend.
What are the drivers?”
21
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
State-of-the art platforms – Focus on three areas
Data
• Regular and direct
access to data marts:
• Policy and Submissions
data for UW, leakage and
prioritization models
• Claims and case reserves
for IELR and case
management changes
• External data:
Exploratory Sandbox
versus Production Ready
Analytics
• Enabling actionable
analytics: For decision
makers to make better
decisions on better data
• Model Validation: Audit
trail easily accessible to
explore decisions
• Results: Implemented
after validation through MI
dashboards or embedded
into data marts
Visualization
• Visualizing Analytic
results: Tuned to solving
a business problem for a
decision maker - not just
visuals of data
• Clarity: What the data are
saying versus what it is not
Proactively:
Work with Procurement
teams. Support Enterprise
Data Teams to add to
Production; data and
model from Sandbox.
Proactively:
Power to connect analytics
across enterprise and
influence management and
broader portfolio strategy.
Proactively:
Visualize portfolio and
change to enable key
executives to make better
decisions using better
data.
22
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs - Reserving Survey
Context
Is an interactive dashboard currently
used to present actuarial findings?
If not, their desire to build such a
dashboard?
22%
78%
Yes No
37%
63%
Yes No
Source: PwC Actuarial Effectiveness Survey, 2013
23
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
Emerging Data Visualization Tools
• 5 years ago, Data Visualization (DV)
tools were relatively scarce
• Now, anyone with access to data can
quickly produce powerful tools and
exhibits in hours or days
• Open Source tends to have a longer
learning curve, though exceptions exist
? The right tool for your organization for
implementation speed, cost and visuals
is likely available – Not necessarily a
large investment
“Open source – Flexible, Free”
“Easy Build – Easy Use”
“Full-Service – Technical”
PwC does not express any official views on the products/tools listed
24
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Open Source R – Chain Ladder + Lattice Plots
example
R Actuarial packages include
“actuar”, “chainladder” and statistical
“lme4” (for GLMM option).
- Example of results using Mack
Method for Chain ladder estimates
- Open source options usually
require more prep and knowledge,
do not guarantee package accuracy
but may be cutting edge
- Paid options can serve a wider,
less technical group and more
likely guarantee their product
? Fits your organization
“Measuring the Variability of Chain
Ladder Reserve Estimates “
Thomas Mack,1993
25
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs - Reserving
Example: Reserve Variation by Line of Business
-
250
500
750
1,000
Example finding: “Reserve variability is most volatile in workers’ compensation primary, workers’
compensation excess, D&O and E&O. Eventual outcomes are sensitive to inflationary trends, litigation
outcome, economic/stock market conditions etc.”
Individual Bootstrapped
Estimates
Portfolio Visualize
Visualize
Data
Validate and
incorporate
to systems
England, PD and Verrall, RJ. Stochastic Claims Reserving in General
Insurance (with discussion), British Actuarial Journal 8, III. 2002
26
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Interactive Output – Visualizing & Collaborating
Reserve Development, Duration, Run-Off
Inflation
New latent
claims
Moore,
Oklahoma
Tornado
Current
U/W Year
Unfavorable
Litigation
Settlement
“Collaborative
Reporting”
“Transparent
Reporting”
Google API:
Creates
annotated
charts using
source data +
comments
Plotly API:
Allows
collaboration,
editing and
sharing through
tech capability
spectrum
BE Reserves Uncertainty Significant Run-Off
27
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs – Results and Scenario
Risk Indicators – Enterprise level
28
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
29
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
Decision enablers have to “think more about thinking”
In most insurance processes, insights are:
? Discovered within Risk, Actuarial, and Finance
? Enabled by up and down-stream Technology
Actioned by decision makers outside RAFT roles
? Most Decision Makers regularly use System 1* thinking, so base
judgments on small samples and heuristics ? cognitive biases
? e.g.: Underwriters in controlled environments with same
information calculate varying prices for same risk in experiments
Question: How can Decision Enablers better help Decision Makers?
Answer: Understand decision making process and incorporate
frameworks into analysis and technology to reduce cognitive bias:
? RAFT: Underwriting and Claims Adjusters receive better and/or
more resonant information to process new accounts (e.g. benchmarks)
? RAFT: Underwriters receive psychometric style questionnaires to check
information grasp and feedback loop impact before assigning prices
R
A
F
T
30
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
Enabling decision makers see your data driven world
Decision Makers incorporate their prior beliefs
? Biased to the individual
Decision Enablers can help by:
? Communicating the limits of data analysis
? Advising on the boundaries of decisions from the data
? Toy Example: “The data shows that the variation of WC losses
for construction companies (10% CI) is larger than for clerical
staff (2%CI) – taking all other known variables into account.”
Note: This still may not incorporate all economic, relationship and other variables pertinent to the
specific decision.
31
March 2015 Actuarial Modernization and Business Intelligence
Goal: Inform human decisions in ways that:
? Reduce unbeneficial variation when using the same information
? Increase information usage
PwC - Copyright & All Rights Reserved
Goal Alignment
Decision making enablers
Success Factor Leading Practice
People
• Management promotes full “buy in”
• Tool champions build data self sufficiency
Roll Out
• Training/ change programs prepare users
• Speak across RAFT silos
Communication
• Tailored to specific stakeholders
• Bottom up feedback guides iterations
Process
Integration
• BI mirrors business process
• BI tied to key metrics (e.g. versus plan/ industry)
Documentation
• Robust data dictionaries and metadata
• People are aware of documentation—and access it
People
Roll Out
Process Integration
Communication
Documentation
32
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
33
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
Distributing Actuarial Insights
Current State
?BI limited to transactional data
?Untrended, Undeveloped
?No credibility measures
Future State
?Granular IBNR included in BI
?Trend and Inflation Tables
?Auto-generate credibility routine
Actuarial insights can be difficult to understand. This has not
prevented incorporation in downstream business processes, such as
pricing and straight through processing.
BI should be no different.
Actuarially enabled BI helps decision makers robustly
consider information.
34
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
State; Payment Type; ZIP; Industry
State in (CT, MA); Policy_Symbol = 2
Loss_Paid; Loss_OS; Loss_Developed;
Z_Score
Group
Fact
Where
LDF, GLM, or
IBNR Allocation
Routines
Call Out
Credibility Issues
User Application
Distributing Actuarial Insights
35
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
Distributing Actuarial Insights
Restrict Distribution
?Reduce misuse risk
?No training need
?Low credibility
?No development
?Users form own view
Distribute
? Share insights
? Improve decisions
? Easy to design
? Misuse risk
? Training need
? Volatility
36
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
In Conclusion
37
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
In Conclusion
Actuarial Modernization and Business Intelligence
Modernize Business Intelligence by:
1) “Right to left” data selection
2) Dashboards facilitate understanding—pick the right tool for you
3) Communicate and train to improve decision making
4) Fill information vacuums; caveat if needed
Business Intelligence tools and data sources should:
• Tie to metrics
• Enable drill down capability
• Be internally consistent
Focusing on widely applicable metrics builds alignment.
38
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Questions?
39
March 2015 Actuarial Modernization and Business Intelligence
40
Professional background
• Tony is a Director in PwC’s predictive analytics practice with
over a decade’s experience in the financial services industry.
Tony advises clients on technical modeling, predictive analytics,
operational, technology, and data quality engagements.
• Tony has led numerous engagements quantifying, pricing, and
managing financial cash flows. He has assisted insurers, banks,
and other organizations with pricing, underwriting,
subcontractor evaluation, customer management, costing the
burden of disease, and predicting borrower default rates and
costs, and testing complex algorithms on systems integration
projects.
• Since joining PwC in 2005, Tony has split his time between the
Philadelphia and Sydney, Australia offices. Prior to joining PwC,
Tony was an actuary and underwriter with Liberty Mutual Group
in Boston where he held various insurance reserving, pricing and
underwriting responsibilities.
• Tony is a Fellow of both the Casualty Actuarial Society, and is a
Member of the American Academy of Actuaries. He is pursuing
his MBA from NYU’s Stern Business School. He graduated cum
laude with a BA in Mathematics and Economics from Boston
College, along with a minor in continental Philosophy.
Project experience highlights
• Spearheaded the actuarial testing of a major Australian bank’s
new SAP-based core banking system. To ensure that interest
and fee calculations worked “first time, every time,” Tony’s team
investigated which peculiar banking activities would stress the
system, and designed test bank accounts with these unusual
characteristics and reviewed system generated output.
Numerous high priority defects were identified and remediated.
• Led Analytics Data Mart premium requirements gathering for
the workers’ compensation business leveraging questionnaires,
interviews, working sessions, and code reviews on current
analytics uses, future state ADM use, and source data flows.
User needs were distilled for system analysis and design teams,
to ensure compatibility with existing business processes.
• A workers’ compensation insurer fundamentally changed
underwriting industry classifications. Tony led data mapping, re-
rating, and transition management efforts for the client. The
team better aligned rate with risk, and developed multiple
implementation rules to minimized extreme increases and
revenue losses.
• Evaluated Third Party Administrators claims handlers
performance over claim lifecycles using Markov Chain methods.
The model was tailored to the particular attributes of each TPA’s
settlement process. Results were used to distinguish
remuneration for good versus poor performance, and to
redistributing market share.
P&C Insurance Modernization
Tony Beirne, Director
41
Professional background
• Prashant is a manager in PwC Advisory practice using better
data and models to drive better decisions making. His focus is on
combining Data Science, Insurance Strategy and Technology to
support intelligent business decisions.
• Prashant has led analytical projects and teams focused on using
predictive analytics, machine learning and external data to
improve commercial books of business in pricing and fraud
detection.
• Prashant has also led Insurance strategy projects aimed at
identifying better Management Decision-Making through
Modern MI and Data, Improved Reserving and Reporting and
advising Large Insurance Carriers on reserve estimates.
• Prashant joins PwC’s Insurance Advisory practice in 2014,
previously he worked for large American and European carriers
in the Actuarial, Reporting and Data Science/Predictive
Analytics departments.
• Prashant holds an MBA from the University of Oxford – Said
and an Undergraduate degree in Mathematics from the
University of Texas at Austin
Project experience highlights
• Led analytics projects pricing commercial lines of business using
external data and machine learning algorithms to better predict
over incumbent models. Businesses needed better pricing to
reflect current risks and new insights around the use of external
data in pricing.
• Led project to introduce unsupervised fraud detection
methodology combining expert judgment and statistical
methods to provide robust fraud suspicion scores. Streamlined
referral process and quantified total fraud model impact using
actuarial methodology.
• Conducted a worldwide benchmarking strategy project focused
on best practice corporate decision making during key
organizational change to customer focus. Presented actionable
results to board and implemented model in foreign operating
entity.
• Led Insurance Market entry project for mid-tier US based
Insurance Broker looking to enter Commercial P&C. Advised
CEO on Product, Legal and Market dimensions.
• Led M&A Project to quantify fair value for several large
insurance company acquisitions.
P&C Insurance Modernization
Prashant De, Manager
Thank you!
This publication has been prepared for general guidance on matters of interest only, and does not constitute professional advice. You should not act upon the
information contained in this publication without obtaining specific professional advice. No representation or warranty (express or implied) is given as to the
accuracy or completeness of the information contained in this publication, and, to the extent permitted by law, PricewaterhouseCoopers LLP, its members,
employees and agents do not accept or assume any liability, responsibility or duty of care for any consequences of you or anyone else acting, or refraining to act,
in reliance on the information contained in this publication or for any decision based on it.
© 2015 PricewaterhouseCoopers LLP. All rights reserved. In this document, “PwC” refers to PricewaterhouseCoopers LLP which is a member firm of
PricewaterhouseCoopers International Limited, each member firm of which is a separate legal entity.
Tony Beirne, FCAS MAAA
Director—Actuarial
[email protected]
+1 (267) 330-1492
Prashant De
Manager—Insurance Advisory
[email protected]
+1 (646) 574-3585
doc_136207900.pdf
Organizations seek to accelerate decision cycles and reduce the time to insight generation.
Actuarial Modernization
and Business Intelligence:
Driving Transformation
RPM Seminar Dallas
Casualty Actuarial Society
March 2015
www.pwc.com
PwC - Copyright & All Rights Reserved
Agenda – Modernization and Business Intelligence
1. Introduction and Context
2. Data Structure and Technology
3. Tools and Outputs
4. Goal Alignment
5. User Application
6. In Conclusion
7. Q&A
2
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
3
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
March 2015 Actuarial Modernization and Business Intelligence
4
Sophistication
Speed
Slow Fast
Low
High
Future Decisions
Today’s Decisions
Source: PwC's Global Data & Analytics Survey 2014. Sophistication measures the reliance on data, the inputs used in the decision,
management preparedness scale etc. Speed measures how quickly a decision is/can be made and implemented in an organization.
Organizations seek to accelerate
decision cycles and reduce the
time to insight generation.
As innovation is enabled,
organizations increase their ability
to identify new, sophisticated
decisions which are required to
address market demands.
As cultures transform, front-line
managers are empowered to
make faster, consistent decisions.
Only by embedding data and
analytics can organizations be
prepared for the future.
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
Insurers today are pursuing ambitious,
cross functional modernization initiatives.
Business Intelligence (BI) is at the
nexus of these efforts due to its:
?Integration with IT infrastructure
?Wide stakeholder base
?Ability to improve business
decision making
5
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Actuarial Modernization & Business Intelligence
Business intelligence rests on a simple concept:
Timely, accurate, relevant, digestible data leads to better
business decision making.
In actuarial and insurance processes, this occurs sub-optimally due to:
? Technology and data production issues
? Opaque data presentation
? Cultural resistance to distributing actuarial products
The time is now to unlock BI benefits by solving these challenges.
New data sources and tools address—and complicate—the task.
“What gets measured gets managed.”
-Peter Drucker
6
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Introduction and Context
Challenges
BI Technology challenges
?Transaction Process System (TPS) datasets
?Irreconcilable sources
?Access to “unsafe” fields
?Extract, Transform & Load (ETL) filters/
cleansing reduce traceability
?Unintuitive user interfaces
?Lack of prospective information
?Limited alignment with plan or business goals
?Too much!
?Low training and user ability
Warehouse
Data
Marts
Staging
Tables
BI ETL
BI User
Interface
7
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
• Collaboration across RAFT: Risk,
Actuarial, Finance, Technology
• Iterative versus “Big Bang”
Goal
Alignment
Actuarial Modernization & Business Intelligence
Overview: Challenges
Despite the myriad
symptoms, key
challenges fall into
four broad
categories:
• Non-reconciling parallel sources
• Less is more
• Prioritize and drill down
(top view requires choices)
Data
Structure
• Trust and training
• Leverage actuarial viewpoint—
or users will form their own
User
Application
• Tools ? Solutions
• Front end enables analysis—
not just production
Tools and
Outputs
8
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
9
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Context
Time spent by senior reserving
actuaries on data processing,
manipulation and reconciliation:
Is there a dedicated IT resource
supporting data extraction and
provision to the Actuarial team?
59%
29%
10%
2%
below 10% 10% to 25%
25% to 50% >75%
33%
67%
Yes No
Source: PwC Actuarial Effectiveness Survey, 2013
10
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Context
What are common sources of data issues you encounter?
0% 10% 20% 30% 40% 50% 60% 70%
Other
Inconsistency in information provided by agents/brokers
Inconsistent reporting between MGAs
Timing lag in data received from MGAs
General Ledger does not easily reconcile
IT function resource constraints
Multiple sources of actuarial data
Manual off-system adjustments
Multiple legacy systems
Source: PwC Actuarial Effectiveness Survey, 2013
11
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Sources through uses
BI systems evolved to serve many disparate users.
Complexity compels architects to make choices—less can be more.
Internal Audit Actuarial Risk/Capital Mgmt Finance
SOx
Capital
Allocation
Pricing
Loss
Reserves
Risk
Audit
Results
Solvency
Warehouse
12
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Sources through uses
BI tools often evolve incrementally
via “add on” requests.
Excess information is produced as
companies “boil the ocean.”
More information does not
necessarily lead to better decisions.
Programs lacking strategic vision
can benefit from narrower scope.
“We have multiple
committees—each with
their own metrics. We go
round the figures three
times.”
Multiple
versions
“Sifting through our packs
could take 3 weeks—leaving no
time to actually do anything.”
Too
detailed
“My guys still get pulled
into ad hoc data requests
twice a day.”
Not
intuitive
Make decisions—
not reconciliations.
13
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
The information environment journey
March 2015 Actuarial Modernization and Business Intelligence
14
EDW into
Multiple
Data Marts
• Generalized to problems
• Several routes to information
fragments and shallows BI
• Many solutions for many
individualized problems
• Reporting layers opaque
Turbulent information flows
• Focused on decisions
• Single data viewing framework
combats fragmented BI
• Single major decision making
process platform
• Transparent and credible—ties to
reporting framework
Smooth delivery momentum
Analytics Mart
and Supporting
Platform
Decision
focused
Problem Focused Decision Focused
PwC - Copyright & All Rights Reserved
Data Structure and Technology
The information environment journey
March 2015 Actuarial Modernization and Business Intelligence
15
Enterprise Data
Warehouse
Multi-Mart Production
& Reporting
Analytics-Oriented
Decision Making
• Enterprise-wide solution
with centralized storage and
access, but providing a
backwards-looking view
• Limited agility for rapidly
changing conditions
• Typically requires large-
scale IT projects, with long
lead times
• BI: “Check what’s there”
• Insight and decision based
• Aligns to business strategy and
user adoption
• Enhances information flows and
analytics tools
• Tailored to analytics models,
uses, and production process
• Speeds organizational agility
• BI: “Use our best insights”
EDW
AM
3PD
• Purpose built repositories,
designed for specific task(s) or
business unit(s)
• Enables run-the-business,
incremental efficiencies
• Customizable, but data mart
proliferation increases technology
portfolio complexity
• BI: “This is what we use”
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Development approach
Top-down strategic BI benefits:
? 80/20 Rule: Get the most
important information right
? Align disparate specialists on
key objectives
? Focus on business impact—
not data issues
? Free resource time from
reconciling differences
Left to Right Right to Left
Collect and
manage data
Analyze data
Draw insights
Make decisions
What decision
is needed?
What insight
will help?
What data
answers this
question?
Manage
targeted data
16
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Development approach
BI users must help development teams avoid the roadblocks which
typically bog down IT projects.
Less is more. Restricting BI data can improve alignment and
efficiency—of both development teams and users.
Left to Right
•Traditional “heavy” design
provides all that is possible
•Includes unneeded functionality
•Extra overhead increases
development complexity
•Complexity slows execution
•Permanent solution
Right to Left
•“Light” development provides
users only what is needed
•Needs specified by analytics
•Sandbox to experiment with
new data or processes
•Lower project complexity
•Incremental solution
17
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Tailor information to the audience
R
i
s
k
M
a
n
a
g
e
m
e
n
t
H
u
m
a
n
R
e
s
o
u
r
c
e
s
I
T
&
O
p
e
r
a
t
i
o
n
s
C
l
a
i
m
s
&
C
e
d
e
d
R
e
i
n
s
u
r
a
n
c
e
I
n
t
e
r
n
a
l
A
u
d
i
t
&
C
o
m
p
l
i
a
n
c
e
M
a
r
k
e
t
i
n
g
U
n
d
e
r
w
r
i
t
i
n
g
&
D
i
s
t
r
i
b
u
t
i
o
n
F
i
n
a
n
c
e
A
c
t
u
a
r
i
a
l
Board Pack
Functional Area Metrics
Macro
Metrics
Individual
BU Metrics
Individual
Functional
Area Metrics
Business Unit Metrics
Level 1
Board Committees
Level 2
Committees,
Entities
Level 3
Functional
Review
18
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Data Structure and Technology
Fast, Widespread Information
Better data structures benefit power users and the wider organization.
Business
Intelligence
Data Governance
? Fast, reliable information source
? Scrubbed data aligns with operations
? Reduces reliance on Power Users
? Promotes data based decision making
? Data issues fixed at the source, rather
than continually adjusted for by users
? Better IT understanding of business uses
? Tighter feedback loops when issues arise
S
o
u
r
c
e
:
T
a
b
l
e
a
u
19
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
20
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
Wouldn’t it be great if…
CRO
“…we had a measure of risk
adjusted return on capital.
We don’t allocate capital
effectively, and we don’t
have clearly articulated risk
appetite metrics.”
Chief Operating Officer
“…I could advise management
on the cost of a new product, a
claim or a customer phone
call.”
Director of Investor
Relations
“… I could get the
information and insights I
need to be able to explain
our results to the market.”
CEO
“…I had a clear profitability view
across:
• Markets
• Channels
• Products
and could relate that back to capital.”
Distribution Director
“…it was clear why I sell a
significant amount of product
and then get castigated for
utilising capital and delivering
low margins.”
Claims Director
“…I could understand
why the claims ratio is
tracking upwards faster
than the market trend.
What are the drivers?”
21
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
State-of-the art platforms – Focus on three areas
Data
• Regular and direct
access to data marts:
• Policy and Submissions
data for UW, leakage and
prioritization models
• Claims and case reserves
for IELR and case
management changes
• External data:
Exploratory Sandbox
versus Production Ready
Analytics
• Enabling actionable
analytics: For decision
makers to make better
decisions on better data
• Model Validation: Audit
trail easily accessible to
explore decisions
• Results: Implemented
after validation through MI
dashboards or embedded
into data marts
Visualization
• Visualizing Analytic
results: Tuned to solving
a business problem for a
decision maker - not just
visuals of data
• Clarity: What the data are
saying versus what it is not
Proactively:
Work with Procurement
teams. Support Enterprise
Data Teams to add to
Production; data and
model from Sandbox.
Proactively:
Power to connect analytics
across enterprise and
influence management and
broader portfolio strategy.
Proactively:
Visualize portfolio and
change to enable key
executives to make better
decisions using better
data.
22
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs - Reserving Survey
Context
Is an interactive dashboard currently
used to present actuarial findings?
If not, their desire to build such a
dashboard?
22%
78%
Yes No
37%
63%
Yes No
Source: PwC Actuarial Effectiveness Survey, 2013
23
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs
Emerging Data Visualization Tools
• 5 years ago, Data Visualization (DV)
tools were relatively scarce
• Now, anyone with access to data can
quickly produce powerful tools and
exhibits in hours or days
• Open Source tends to have a longer
learning curve, though exceptions exist
? The right tool for your organization for
implementation speed, cost and visuals
is likely available – Not necessarily a
large investment
“Open source – Flexible, Free”
“Easy Build – Easy Use”
“Full-Service – Technical”
PwC does not express any official views on the products/tools listed
24
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Open Source R – Chain Ladder + Lattice Plots
example
R Actuarial packages include
“actuar”, “chainladder” and statistical
“lme4” (for GLMM option).
- Example of results using Mack
Method for Chain ladder estimates
- Open source options usually
require more prep and knowledge,
do not guarantee package accuracy
but may be cutting edge
- Paid options can serve a wider,
less technical group and more
likely guarantee their product
? Fits your organization
“Measuring the Variability of Chain
Ladder Reserve Estimates “
Thomas Mack,1993
25
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs - Reserving
Example: Reserve Variation by Line of Business
-
250
500
750
1,000
Example finding: “Reserve variability is most volatile in workers’ compensation primary, workers’
compensation excess, D&O and E&O. Eventual outcomes are sensitive to inflationary trends, litigation
outcome, economic/stock market conditions etc.”
Individual Bootstrapped
Estimates
Portfolio Visualize
Visualize
Data
Validate and
incorporate
to systems
England, PD and Verrall, RJ. Stochastic Claims Reserving in General
Insurance (with discussion), British Actuarial Journal 8, III. 2002
26
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Interactive Output – Visualizing & Collaborating
Reserve Development, Duration, Run-Off
Inflation
New latent
claims
Moore,
Oklahoma
Tornado
Current
U/W Year
Unfavorable
Litigation
Settlement
“Collaborative
Reporting”
“Transparent
Reporting”
Google API:
Creates
annotated
charts using
source data +
comments
Plotly API:
Allows
collaboration,
editing and
sharing through
tech capability
spectrum
BE Reserves Uncertainty Significant Run-Off
27
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Tools and Outputs – Results and Scenario
Risk Indicators – Enterprise level
28
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
29
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
Decision enablers have to “think more about thinking”
In most insurance processes, insights are:
? Discovered within Risk, Actuarial, and Finance
? Enabled by up and down-stream Technology
Actioned by decision makers outside RAFT roles
? Most Decision Makers regularly use System 1* thinking, so base
judgments on small samples and heuristics ? cognitive biases
? e.g.: Underwriters in controlled environments with same
information calculate varying prices for same risk in experiments
Question: How can Decision Enablers better help Decision Makers?
Answer: Understand decision making process and incorporate
frameworks into analysis and technology to reduce cognitive bias:
? RAFT: Underwriting and Claims Adjusters receive better and/or
more resonant information to process new accounts (e.g. benchmarks)
? RAFT: Underwriters receive psychometric style questionnaires to check
information grasp and feedback loop impact before assigning prices
R
A
F
T
30
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Goal Alignment
Enabling decision makers see your data driven world
Decision Makers incorporate their prior beliefs
? Biased to the individual
Decision Enablers can help by:
? Communicating the limits of data analysis
? Advising on the boundaries of decisions from the data
? Toy Example: “The data shows that the variation of WC losses
for construction companies (10% CI) is larger than for clerical
staff (2%CI) – taking all other known variables into account.”
Note: This still may not incorporate all economic, relationship and other variables pertinent to the
specific decision.
31
March 2015 Actuarial Modernization and Business Intelligence
Goal: Inform human decisions in ways that:
? Reduce unbeneficial variation when using the same information
? Increase information usage
PwC - Copyright & All Rights Reserved
Goal Alignment
Decision making enablers
Success Factor Leading Practice
People
• Management promotes full “buy in”
• Tool champions build data self sufficiency
Roll Out
• Training/ change programs prepare users
• Speak across RAFT silos
Communication
• Tailored to specific stakeholders
• Bottom up feedback guides iterations
Process
Integration
• BI mirrors business process
• BI tied to key metrics (e.g. versus plan/ industry)
Documentation
• Robust data dictionaries and metadata
• People are aware of documentation—and access it
People
Roll Out
Process Integration
Communication
Documentation
32
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
33
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
Distributing Actuarial Insights
Current State
?BI limited to transactional data
?Untrended, Undeveloped
?No credibility measures
Future State
?Granular IBNR included in BI
?Trend and Inflation Tables
?Auto-generate credibility routine
Actuarial insights can be difficult to understand. This has not
prevented incorporation in downstream business processes, such as
pricing and straight through processing.
BI should be no different.
Actuarially enabled BI helps decision makers robustly
consider information.
34
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
State; Payment Type; ZIP; Industry
State in (CT, MA); Policy_Symbol = 2
Loss_Paid; Loss_OS; Loss_Developed;
Z_Score
Group
Fact
Where
LDF, GLM, or
IBNR Allocation
Routines
Call Out
Credibility Issues
User Application
Distributing Actuarial Insights
35
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
User Application
Distributing Actuarial Insights
Restrict Distribution
?Reduce misuse risk
?No training need
?Low credibility
?No development
?Users form own view
Distribute
? Share insights
? Improve decisions
? Easy to design
? Misuse risk
? Training need
? Volatility
36
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
In Conclusion
37
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
In Conclusion
Actuarial Modernization and Business Intelligence
Modernize Business Intelligence by:
1) “Right to left” data selection
2) Dashboards facilitate understanding—pick the right tool for you
3) Communicate and train to improve decision making
4) Fill information vacuums; caveat if needed
Business Intelligence tools and data sources should:
• Tie to metrics
• Enable drill down capability
• Be internally consistent
Focusing on widely applicable metrics builds alignment.
38
March 2015 Actuarial Modernization and Business Intelligence
PwC - Copyright & All Rights Reserved
Questions?
39
March 2015 Actuarial Modernization and Business Intelligence
40
Professional background
• Tony is a Director in PwC’s predictive analytics practice with
over a decade’s experience in the financial services industry.
Tony advises clients on technical modeling, predictive analytics,
operational, technology, and data quality engagements.
• Tony has led numerous engagements quantifying, pricing, and
managing financial cash flows. He has assisted insurers, banks,
and other organizations with pricing, underwriting,
subcontractor evaluation, customer management, costing the
burden of disease, and predicting borrower default rates and
costs, and testing complex algorithms on systems integration
projects.
• Since joining PwC in 2005, Tony has split his time between the
Philadelphia and Sydney, Australia offices. Prior to joining PwC,
Tony was an actuary and underwriter with Liberty Mutual Group
in Boston where he held various insurance reserving, pricing and
underwriting responsibilities.
• Tony is a Fellow of both the Casualty Actuarial Society, and is a
Member of the American Academy of Actuaries. He is pursuing
his MBA from NYU’s Stern Business School. He graduated cum
laude with a BA in Mathematics and Economics from Boston
College, along with a minor in continental Philosophy.
Project experience highlights
• Spearheaded the actuarial testing of a major Australian bank’s
new SAP-based core banking system. To ensure that interest
and fee calculations worked “first time, every time,” Tony’s team
investigated which peculiar banking activities would stress the
system, and designed test bank accounts with these unusual
characteristics and reviewed system generated output.
Numerous high priority defects were identified and remediated.
• Led Analytics Data Mart premium requirements gathering for
the workers’ compensation business leveraging questionnaires,
interviews, working sessions, and code reviews on current
analytics uses, future state ADM use, and source data flows.
User needs were distilled for system analysis and design teams,
to ensure compatibility with existing business processes.
• A workers’ compensation insurer fundamentally changed
underwriting industry classifications. Tony led data mapping, re-
rating, and transition management efforts for the client. The
team better aligned rate with risk, and developed multiple
implementation rules to minimized extreme increases and
revenue losses.
• Evaluated Third Party Administrators claims handlers
performance over claim lifecycles using Markov Chain methods.
The model was tailored to the particular attributes of each TPA’s
settlement process. Results were used to distinguish
remuneration for good versus poor performance, and to
redistributing market share.
P&C Insurance Modernization
Tony Beirne, Director
41
Professional background
• Prashant is a manager in PwC Advisory practice using better
data and models to drive better decisions making. His focus is on
combining Data Science, Insurance Strategy and Technology to
support intelligent business decisions.
• Prashant has led analytical projects and teams focused on using
predictive analytics, machine learning and external data to
improve commercial books of business in pricing and fraud
detection.
• Prashant has also led Insurance strategy projects aimed at
identifying better Management Decision-Making through
Modern MI and Data, Improved Reserving and Reporting and
advising Large Insurance Carriers on reserve estimates.
• Prashant joins PwC’s Insurance Advisory practice in 2014,
previously he worked for large American and European carriers
in the Actuarial, Reporting and Data Science/Predictive
Analytics departments.
• Prashant holds an MBA from the University of Oxford – Said
and an Undergraduate degree in Mathematics from the
University of Texas at Austin
Project experience highlights
• Led analytics projects pricing commercial lines of business using
external data and machine learning algorithms to better predict
over incumbent models. Businesses needed better pricing to
reflect current risks and new insights around the use of external
data in pricing.
• Led project to introduce unsupervised fraud detection
methodology combining expert judgment and statistical
methods to provide robust fraud suspicion scores. Streamlined
referral process and quantified total fraud model impact using
actuarial methodology.
• Conducted a worldwide benchmarking strategy project focused
on best practice corporate decision making during key
organizational change to customer focus. Presented actionable
results to board and implemented model in foreign operating
entity.
• Led Insurance Market entry project for mid-tier US based
Insurance Broker looking to enter Commercial P&C. Advised
CEO on Product, Legal and Market dimensions.
• Led M&A Project to quantify fair value for several large
insurance company acquisitions.
P&C Insurance Modernization
Prashant De, Manager
Thank you!
This publication has been prepared for general guidance on matters of interest only, and does not constitute professional advice. You should not act upon the
information contained in this publication without obtaining specific professional advice. No representation or warranty (express or implied) is given as to the
accuracy or completeness of the information contained in this publication, and, to the extent permitted by law, PricewaterhouseCoopers LLP, its members,
employees and agents do not accept or assume any liability, responsibility or duty of care for any consequences of you or anyone else acting, or refraining to act,
in reliance on the information contained in this publication or for any decision based on it.
© 2015 PricewaterhouseCoopers LLP. All rights reserved. In this document, “PwC” refers to PricewaterhouseCoopers LLP which is a member firm of
PricewaterhouseCoopers International Limited, each member firm of which is a separate legal entity.
Tony Beirne, FCAS MAAA
Director—Actuarial
[email protected]
+1 (267) 330-1492
Prashant De
Manager—Insurance Advisory
[email protected]
+1 (646) 574-3585
doc_136207900.pdf