|
Activity Number:
|
507
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 2, 2007 : 8:30 AM to 10:30 AM
|
|
Sponsor:
|
Business and Economics Statistics Section
|
| Abstract - #308592 |
|
Title:
|
Comparing Methods Used for Handling Negative Values in Statistical Samples of Business Records
|
|
Author(s):
|
Eric Falk*+
|
|
Companies:
|
Ernst & Young
|
|
Address:
|
1225 Connecticut Ave NW, Washington, DC, 20036,
|
|
Keywords:
|
Statistical Sampling ; Unmatched Credits ; Negatives ; Estimation ; Simulation
|
|
Abstract:
|
Negative monetary values occur in business records for several reasons and must be handled appropriately to prevent biasing estimates. Different taxing authorities require different treatment of negative values in the context of statistical sampling of tax records. Typically business expense records contain debits for purchases and expenses, and offsetting credits if purchases are returned or expenses are reversed. Many of the credits can be matched to the original debit when the sampling population is being cleaned, but usually the final frame will contain unmatched credits for various reasons. This study compares four common methods currently used or required by either the IRS or state taxing authorities for handling these remaining negative values when using statistical sampling.
|
