casualty actuarial society ratemaking seminar tampa, 2002 int-1 introductory data management 101 al...

Post on 04-Jan-2016

215 Views

Category:

Documents

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Casualty Actuarial SocietyRatemaking Seminar

Tampa, 2002

INT-1

Introductory Data Management 101

Al Hapke

ASPECT, Inc.

Your Role in Data Management

• Demanding User

• Lack of defined needs

• Lack of knowledge about information technology

• Lack of business knowledge in the IT staff

Therefore, you must communicate your goals effectively and clearly.

Objective of Data Management:

To store and organize data in a way that allows the analyst to answer questions about the business.

These questions should help direct and guide the management of the business.

Processing Systems are not adequate to satisfy the analytical needs of the company.

They’re designed to do work, not answer questions.

Steps That Help Communication:

• Formulate many specific questions– Brainstorm yourself– Talk to your customers/clients/boss– Read actuarial papers– Review competitive rate filings

• Write them down

• Design your spreadsheet or model to answer the question

• Determine what you need to populate the spreadsheet

Example of Questions:

What do we expect to pay for claims in this class vs. other classes?

1. Age/experience of driver

2. WC class code

3. Property construction

4. State/territory/location

5. Other characteristics such as credit rating,

new/renewal, etc.

Issues with this Question:

• Volume in each class or characteristic

• How much history?

• Can premium be appropriately matched with losses?

• Can earned exposures be captured?

• Can class definitions be multidimensional?

Other Questions:

• What is our exposure to maximum loss?– Answer can be found in limit/location studies

• How much development can we expect on reported losses?– Considerations:

• Report date

• Historical claim distributions

Other Questions (continued)

• Has our underlying book of business changed?– What types of losses are we seeing?

• This is only meaningful if we have been profiling our mix of claims so that we can see what’s different.

• e.g– Size of loss at consistent points in time

– Minor coverage detail

– Cause of loss

Other Questions (continued)

• Who, how much, and what is your producer selling?

What are the characteristics of the business being brought in the front door?

Or…. Leaving through the back door?

Issues to Consider in Your Management Information System

• Ease of Access - level of independence

from programmers

• Flexibility - new classes, new

characteristics, new products

• Quality of data

top related