Duplicate Payments

To help detect duplicate payments, you should follow these steps:

  • Identify all the payment systems and data files relevant to performing the analysis
  • Determine the data elements available for testing, and what they should contain, along with their meaning
  • Perform an initial analysis to identify the key attributes of the population to be tested, e.g. how many transactions, histogram of distributions, major vendors, etc.
  • Identify the tests to be performed
  • Perform the detailed analysis

Any definition of a duplicate payment will generally fall under the category of “Same, Same, Same”, e.g. same vendor, same invoice number, same payment date.

There are numerous variations and combinations on this theme and the number of variables can range from as few as two to as many as five or more. All of this needs to be carefully considered upfront. Just a few examples of potential duplicate payments include:

  • Same invoice number, amount, invoice date, and vendor.
  • Same invoice number, amount, and vendor. 
  • Same invoice number and vendor. 
  • Same amount and vendor. 
  • Same purchase order, vendor, and amount.
  • Same numbers in the invoice number, amount, invoice date, and vendor fields (removes many unneeded characters in invoice number such as punctuation). 
  • First four digits of invoice number, amount, invoice date, and vendor (e.g., identifies duplicates where a user enters the invoice number and then a duplicate payment with an invoice number with an “A” at the end).
  • Same invoice number, amount, invoice date, and different vendor (identifies same invoice payments to vendors with a different name in the accounts payable system such as “IBM” and “IBM Corp”). 
  • Same invoice number, amount, vendor, and different check (identifies same invoice payments made on different days to vendors that may normally go undetected). 
  • Report of all vendors that are comprised solely of duplicate payments. 
  • Report of all vendors where only two similar payments were made (e.g., only two $395 payments in the period reviewed).

There are alternative procedures available to make these identifications:

Proc SSS ("Same, Same, Same") can identify potential duplicate payments. In the case of transposed invoice numbers, split payments, etc. Proc CR is better suited.

Example screen shot of tests for duplicate payments:

Analysts interested in Proc CR may also be interested in Proc Means and Proc Univariate.
 
 

Web Page last updated on 10-26-2006
© EZ-R Stats, LLC 2005-2007

© EZ_R Stats

Visit EZ_R Stats on the web at:

www.ezrstats.com