Good Afternoon,

There was a discussion in my Compliance department regarding samples sizes when doing Compliance Reviews. Not everyone in the team was choosing a sample size consistently. One member of our team suggested following the FRB's sample size chart (Below). In my opinion this is very excessive since we frequently have population sizes over 10,000. Does your bank have a sampling methodology that you guys follow that is different than this? Thank you.

Number in Population Minimum Size of Sample Maximum
Under 99 20 or total population if smaller than 20 20
100 - 299 20 plus 10% of number above 100 40
300 - 400 40 plus 7% of number above 300 54
500 - 749 54 plus 6% of number above 500 69
750 - 999 69 plus 5% of number above 750 82
1,000 – 1,999 82 plus 4% of number above 1,000 122
2,000 – 2,999 122 plus 3% of number above 2,000 152
3,000 – 4,999 152 plus 2% of number above 3,000 192
5,000 – 9,000 192 plus 1.5% of number above 5,000 267
10,000 or more 267 plus 1% of number above10,000 267

Many auditors use statistical sampling. With a population of 10,000 and a Level of Assurance of 95% and a Rate of Occurrence of 5.0%, your random sample size would be 60. Set your standards at the levels you are comfortable. If you find any errors, then you can expand your sample.

I think more than anything you should set standards your department is comfortable with and be consistent. When you vary a sample size properly document why the change.

How does everyone select samples at their institutions?

We use a combination of statistical and judgmental sampling depending on the test being performed. In all cases it is properly documented in the audit.