Why the PCAOB and External Auditors Should be Concerned about Substantive-Only Audits

Why the PCAOB and External Auditors Should be Concerned about Substantive-Only Audits

in Blog Article by Jeff Hare

This article is long overdue, but still one I have been dreading to release.  I know the audit firms could come under significant additional scrutiny from regulators such as the PCAOB.  However, there are significant risks organizations are facing because of poorly designed controls and inadequate audit procedures.  We’re now more than 20 years post-Sarbanes-Oxley, and it is time to shine the light on areas that board members and investors would be appalled by if they understood this topic.  With this said, here is my article…

In evaluating how to approach an audit, in general, and specific business processes within any audit, in particular, external auditors sometimes choose to audit ‘around the system’.  That is, they ignore the way systems are designed and test the given process or control without regards to how the system is designed.  They do so by identifying a population of transactions related to the process and testing the activity through sampling as required by their internal standards and consistent with the regulator(s) that oversee their audits (PCAOB, OCC, various government agencies, etc).

There are consistent gaps in the approach that external audit firms take that could leave the chosen populations incomplete.  Therefore, I will make the argument that a substantive-only audit is a flawed approach and one that external audit partners and regulators should be concerned about.

To illustrate this flaw, I will use arguably the most important control related to the integrity of the financial statements – the control over manual (non-standard) journal entries.  Manual journal entries inherently pose a high risk to the integrity of the financial statements because they can move any balance between any account within the trial balance.  In my experience, most external auditors don’t know how to properly identify a population of journal entries that equate to a manual journal entry.

I identify the following types of journal entries that could exist within an ERP system:

  1. Manually created in the system within the general ledger itself
  2. Uploaded from a spreadsheet – manual or non-standard type
  3. Uploaded from a spreadsheet – created by another system (i.e. manual interface)
  4. Uploaded from a spreadsheet – looking as if it came from a subledger
  5. Interfaced from another system (i.e. automated interface)
  6. Transferred from a subledger
  7. Manually created in a subledger and transferred as if it were a subledger entry


Because of the way JE controls are typically designed (i.e. to ignore subledger JE’s) organizations using Oracle’s E-Business Suite or ERP Cloud applications would need to configure each system differently in order to prohibit a user from creating a journal entry from a Source that is not subject to the manual / non-standard JE controls.  In other words, even in the case where an audit is not placing reliance on any application controls (i.e. a fully substantive audit), the improper configuration of a system COULD cause the functional auditors to leave out a portion of the JE’s that should be in their control population.

Share this post:
ERPRA Become Our Partner

Please select your preferred datasheet and download it: