Wake Up to LDTI

Keep Up with Innovation

Step Up Actuarial Performance
Insights into regulatory change from a technology perspective
Robert Barg, FSA, MAAA
Contents

3 4 5
Under Complex Demands, Technology Has Advanced –
Drive Efficiency with
Are Your Actuarial Shouldn’t an Actuary’s Job
Advanced Automation
Systems Up to Speed? Be Easier?

Control the Complexity

with Automated Workflow 6 Turbo-charge Data
Management 10 Control the Complexity with Automated Workflow 11
11
Why LTDI Increases Data
1. Business Process Management 6 Requirements 10 1. Distributed Databases Data Mart 12

11
How Can BPM Support Actuaries
under LDTI? 7 Write Actuarial Results
to Multiple Servers
Data Lake 12

2. Robotic Process Automation 8 2. Modern Data Storage

Architectures 12 What’s Your Data Strategy? 1312

RPA and BPM – Working Together

for Insurers 8 Data Warehouse 12

What Do You Need to Automate

9
14
and How? Conclusion
Wake Up to the Positive Power of High-performing Technology

2
Under Complex Demands, Are Your
Actuarial Systems Up to Speed?

Although designed to simplify accounting processes, long-duration targeted

improvements (LDTI) to U.S. GAAP are threatening to make life more complicated
for actuaries – and, in turn, undermine the performance of their systems.

Now, as regulatory demands continue to rise, do you have the tools to work
fast and keep up?

In the quest to improve performance, insurers have often looked to boost

processing power – typically by adding more cores to the server.

But better management of workflow and data can also bring new speed
and sophistication to the actuarial environment.

And thanks to a range of innovative technologies, the positive impact

on performance can be immense. Read on to explore the opportunities.

As regulatory demands continue to rise,

do you have the tools to work fast and keep up?

3
Technology Has Advanced –
Shouldn’t an Actuary’s Job be Easier?

Once upon a time, actuaries had to calculate reserves laboriously by hand, using a factor table,
calculator and trusty pen and paper. Today, computers can perform the necessary calculations
for hundreds of thousands of policies in minutes or seconds.

The fact is, though, that developments in computer science haven’t made actuaries’ jobs easier.
And that’s largely because the more that technology advances, the more that stakeholders expect
from actuarial calculations.

You could see this phenomenon as an example of what economists call “induced demand.”

By way of an analogy, let’s say that there’s regularly too much traffic on a highway for two lanes to
handle, causing congestion at peak times. According to the laws of induced demand, adding a third
or fourth lane won’t solve the problem – as more lanes simply encourage more vehicles.

Similarly, increased capacity for computing is driving the scope and complexity of regulatory and internal
reporting demands – and piling more and more pressure on resources, systems and processes.

However, as much as advancements in technology are increasing expectations, they can also provide
opportunities to improve actuarial performance.

The more that technology advances, the more that stakeholders expect
from actuarial calculations.

4
Drive Efficiency with Advanced Automation

Beyond extending your computing infrastructure and increasing its power, one of the
most logical ways to enhance performance is to make your operations more efficient.
Already, changes to accounting bases like LDTI are prompting insurers to make
significant, complex adjustments to their workflow. So, it makes sense to get technology
on your side and use it to support – or redesign – your actuarial processes.

How LDTI Complicates Workflow

With LDTI, U.S. insurers must update their assumptions on a regular basis, rather
than locking them in for good.
Under the previous FAS 60 regime, you could basically set assumptions once and
then forget about them. Now, you’ll need to periodically review, analyze and – if
necessary – modify your assumptions, before testing, documenting and justifying
their use in your models to auditors.
In other words, a process as seemingly simple as updating assumptions will involve
a host of new stakeholders and systems – and add multiple steps to your workflow.
At the same time, LDTI’s rollforward disclosure requirements will mean you’ll have
to perform multiple model runs every month. In this case, you’ll need to track and
systematically archive the details of your processes so you can make sense of
them downstream.

5
Control the Complexity with Automated Workflow

Two broad types of modern technology are adept at handling the complex workflows that LDTI
will generate, each in their own unique way.

1. Business Process Management

Business process management (BPM), also known as business process automation, is geared
toward improving whole end-to-end processes. Many modern insurers already use BPM tools
to automate the processing of claims.
If you’ve ever filed your own insurance claim, you might recognize the hallmarks of a BPM-driven
process. For instance, after you called your insurer and relayed your policy number, you may have
received an automated email requesting that you upload certain documentation.
Then, on receipt of your documents, another email might have informed you that your claim
needed a second-level review, based on the amount involved or the reason for the claim.
Once you uploaded the further information your insurer required, the claim was approved,
triggering a payment, and eventually you received a check in the mail.
In this classic example, think of the BPM tool as a manager whose job it is to orchestrate
– rather than carry out – the process, notifying people when they should complete a task
and routing further tasks based on their responses.

Think of the BPM tool as a manager whose job it is to orchestrate –

rather than carry out – the process.

6
How Can BPM Support Actuaries under LDTI?
Actuaries can use BPM tools to industrialize their operations. By mapping tasks
from administration systems to the general ledger and tying them all together, they
can effectively turn the end-to-end actuarial production process into an efficient,
BPM-powered assembly line.
Just as in a factory, the goal is not only to automate as many potentially manual
steps as possible, but also to streamline and make them consistent – so that day
in day out, the same tasks are managed by the same people.
Typically, the actuarial environment is made up of multiple underlying applications,
which BPM tools mainly control with APIs. When you’re implementing a BPM tool,
you’ll therefore need considerable IT support to help you access and connect the
applications and programmatically orchestrate the whole process.
For LDTI, the potential of BPM is enormous. If you’re carrying out multiple runs
to generate disclosure rollforwards, a BPM tool could help you kick off the whole
process automatically, once the right data becomes available and the assumptions
are approved.

Actuaries can use BPM tools to industrialize their operations

and turn the end-to-end production process into an efficient
assembly line.

7
2. Robotic Process Automation
Robotic process automation (RPA) is a newer technology than BPM and tends to
focus more on the automation of individual tasks. Just as your Macro Recorder
button in Excel will allow you to make the same complex set of changes to a
spreadsheet every month – delete a column, copy data from one row to another,
edit formulas and so on – an RPA tool will take a manual process that’s typically
performed by humans and complete it automatically.
To get the job done, RPA tools use bots – not actual robots sitting at a desk, but
computer programs that act like people. Unlike BPM tools, they don’t work behind
the scenes in a managerial capacity but perform the actual tasks.
Bots can interface with applications in the same way as human employees, so they
will open a web browser, drop in a user name and password, visit a website and
download a file.

RPA and BPM – Working Together for Insurers

One of the benefits of bots is that they are easy to implement and deploy. As with a
macro, you can virtually record what a human would do in a situation and teach the
bot to do the same, rather than having to write code and script to access APIs.
That makes RPA tools particularly useful for managing legacy software, which
may not come with APIs. Because bots work like human staff, they can access
applications directly from the same interfaces.
Even though they operate differently, RPA and BPM tools work very well together as
complementary technologies. In other words, BPM-automated processes can call on
RPA bots to carry out tasks, with the BPM tool acting as the manager and RPA tools
as the workforce.

8
What Do You Need to Automate – and How?
If you’re thinking of adopting BPM or RPA tools for workflow automation, an important
question to ask yourself is: Are you automating your current processes or redesigning
them and automating from scratch? For most insurers, starting over isn’t an option
– so, they need to automate what’s already in place.
Another key consideration is your existing technology and how easy it is to automate.
Some solutions – especially those with APIs – lend themselves toward BPM tools
and others more to RPAs.
Your organization’s skills and resources are another important aspect to consider,
as BPM tools are generally more difficult to implement than RPAs and require a higher
level of IT expertise.

9
Turbo-charge Data Management

Data management is a massive topic in and of itself and one that’s

of ever-growing importance for insurers. It’s also a field that’s widely
discussed but not always well understood, with a whole lexicon of
terms to baffle the actuarial profession.
More to the point, accounting changes such as LDTI are straining
actuaries’ data management architecture, calling for a more effective
approach to writing and storing results.

Why LTDI Increases Data Requirements

LDTI is giving insurers a lot more data to analyze. With changing
assumptions and a more principles-based framework, companies trying
to make sense of results have to dig back through large vectors of
cashflows from multiple runs. On top of that, companies must include
disaggregated rollforwards in their disclosures, leading to ever-more data.

LDTI is straining data management architecture,

calling for a more effective approach to writing and
storing results.

10
Meet Data Demands with the Latest Writing
and Storage Methods

Data management technology has evolved in recent years to make it easier to both write and store large volumes of data.

1. Distributed Databases
To perform thousands of calculations at a time, an advanced actuarial modeling platform can spread the load by
running different calculations on different cores. But while this distributed approach to calculation initially reduces
runtime, it may also lead to major bottlenecks in the file system.
For example, if your cores are all simulating very similar economic scenarios, with only minor variations between
them, they are likely to finish the job at around the same time. So, if you’re running 1,000 simulations on 1,000
cores, you’ll also have 1,000 sets of results to save simultaneously to the hard drive or file server. And this can
overload the hard drive or file server, resulting in a queue.
In short, your results writing needs to be as distributed as your calculations – which is where the distributed
database comes in.

Write Actuarial Results to Multiple Servers

Previously, speeding up results writing usually meant scaling vertically by upgrading your hardware,
with a more robust file server or hard drive often proving a costly way to achieve better performance.
Today, distributed database technology allows actuaries to scale horizontally and effectively save
compressed files to a network of multiple interconnected databases – housed on different servers
but managed by a dedicated database manager.
Also known as NoSQL or non-relational databases, distributed databases are much less rigid
in structure than more traditional SQL or relational databases. It’s their very flexibility that makes
them ideal for the writing of high volumes of projection results.

11
2. Modern Data Storage Architectures
Once you’ve produced and saved results for LDTI, you need somewhere to store the data that
makes it easy to access, organize and analyze. With such large quantities to sift through at any
one time, the question is how best to structure your data repositories going forward.
The names of the latest types of architecture will be familiar to you, but it’s worth pointing
out the critical differences between them and the unique roles they can play for actuaries.

Data Warehouse
The data warehouse aggregates data from various sources or databases but compiles the data
in a highly structured, standardized way.
You might use a data warehouse to store actuarial results from all your various models,
encompassing multiple platforms and lines of business. But because the warehouse is so well
organized, business users can query it easily for trending analysis – to understand, for example,
how mortality has affected results.
Nevertheless, data warehouses have their limitations – they can grow so vast in size that they
are no longer suitable for detailed analysis. That’s where the data mart comes in.

Data Mart
If, say, you just look after payout annuities, you might want to carry out the same, specific analysis
period after period. Why hunt through a giant data warehouse containing all sorts of (potentially
irrelevant) actuarial data, when you could keep only what you need in your own modest data mart?
Put another way, the data mart will store a defined subset of the data warehouse, such as data
on payout annuities, and can be structured to suit exactly the kind of analysis you want to conduct.

12
Data Lake
In contrast to the data warehouse and data mart, the data lake – as its name suggests – is basically
an unstructured repository of data. It could contain analytics on traffic to your firm’s website, financials,
images, videos and so on: a whole universe of big data, in fact.
As a result, data lakes are more suitable for and used by data scientists than business users. It’s up
to data scientists to discover correlations and relationships in the data, any structure could lead them
down the wrong path or pre-empt their findings. Thanks to its sheer lack of structure, the data lake
provides the perfect architecture for big data analytics.

What’s Your Data Strategy?

In a recent survey, 80 percent of insurers said they were in the process of modifying their data
architecture, partially as a result of accounting changes like LDTI. Their challenge is to figure out
the best data strategy for their business.
With so many results being produced, you must make sure you store the data in a way that will help you
gain the most business insight. But you also need the right people and mix of skills to build and manage
a new data architecture, while meeting increased data privacy requirements.

80 percent of insurers are in the process of modifying their data architecture,

partially as a result of accounting changes.

13
Conclusion
Wake Up to the Positive Power of High-performing Technology

No conversation about emerging technologies can ignore the potential impact of all this innovation
on the human workforce – and the fear, to put it bluntly, that automation is a job killer.

The good news is that, while computers and robots may be here to stay, they pose little threat to qualified
professionals like actuaries. Think of automation as being here to support, rather than replace, human brainpower.

By ridding actuaries of low-value, mundane manual tasks, automated technology allows them to use their time and
minds more effectively – and focus less on producing the numbers and more on analyzing the business implications.

In our age of regulatory change, the stronger the performance of your systems and processes, the better placed
your actuaries are to not only achieve compliance but also add value. What better time to invest in innovation?

LDTI compliance won’t be easy, so you need the best team and technology
on your side. Email [email protected] to learn how FIS’ award-winning
solutions can help you comply with complex regulations and enhance
actuarial performance.

14
 About FIS
FIS is a global leader in technology, solutions and services for
merchants, banks and capital markets that helps businesses
and communities thrive by advancing commerce and the
financial world. For over 50 years, FIS has continued to drive
growth for clients around the world by creating tomorrow’s
technology, solutions and services to modernize today’s
businesses and customer experiences. By connecting
merchants, banks and capital markets, we use our scale, apply
our deep expertise and data-driven insights, innovate with
purpose to solve for our clients’ future, and deliver experiences
that are more simple, seamless and secure to advance the way
the world pays, banks and invests. Headquartered in
Jacksonville, Florida, FIS employs about 55,000 people
worldwide dedicated to helping our clients solve for the future.
FIS is a Fortune 500® company and is a member of Standard &
Poor’s 500® Index. For more information about FIS, visit www.
fisglobal.com

fisglobal.com [email protected]

©2019 FIS and/or its subsidiaries. All Rights Reserved. 769983