Chapter 6 - Data-Driven Fraud Protection

Question 1

fraud vs errors

Answer 1

errors:
unintentional
will be found throughout any data set

fraud:
is intentional
found in very few data sets
is like finding a ‘needle in a haystack’

Question 2

proactive detection vs reactive detection

Answer 2

proactive is seeking it out

reactive is waiting until you have a problem

Question 3

pros and cons of proactive fraud detection

Answer 3

pro - can find frauds that nobody knew about

con-needle in a haystack, can be time-consuming and often not fruitful

Question 4

pros and cons of reactive detection

Answer 4

you dont figure out the fraud until reasons are found, and after it’s been going on for some time

Question 5

6 steps of proactive fraud detection

Answer 5

1. understanding the business
2. identify possible frauds that could exist
3. Catalog possible fraud symptoms
4. Use technology to gather data about those symptoms
5. analyze results
6. investigate symptoms

Question 6

How to catalog fraud symptoms?

Answer 6

divide them into one of the 6 groups…

1. acct anomalies
2. internal control weakness
3. analytical anomalies
4. extravagant lifestyle
5. unusual behavior
6. tips or complaints

Question 7

if you were to investigate a kickback fraud, classify these symptoms:

• increasing prices
• larger order quantities
• increasing purchases from favored vendor
• decreasing quality
• decreasing purchases from other vendors

Answer 7

analytical symptoms

Question 8

classify these symptoms:
buyer is an outsider
buyers work habits change unexpected

Answer 8

behavioral sympt-om

Question 9

classify these symptoms:

• buy lives beyond means
• buyer buys an expensive car
• buyer builds a more expensive home

Answer 9

lifestyle symptoms

Question 10

classify these symptoms:

• all transactions are with one buyer and one vendor
• unapproved vendors being used

Answer 10

control symptoms

Question 11

classify these symptoms:

1099 from vendor to one of the buyers relatives

Answer 11

document symptom

Question 12

classify these symptoms:

• anonymous complaints about the vendor or buyer
• quality complaints about purchased products

Answer 12

tips and complaints

Question 13

Benford’s Law

Answer 13

stout of any given data set, the first and second digit of any given numbers will likely be a 1 or 2. not so true for 3rd digit on

the first digit will be 1 30% of the time, progressing to only being a 9 in 5% of instances

Question 14

what is digital analysis?

Answer 14

the art of analyzing digits that make up numbers in data sets.

Question 15

what is a strength and weakness of digital analysis?

Answer 15

strength - it’s inexpensive

weakness - correlation doesn’t necessarily mean causation and can lead to a waste of time