Auditing in a computer-based environment

Relevant to Foundation level Paper FAU and ACCA Qualification Papers F8 and P7

Specific aspects of auditing in a computer-based environment

Information technology (IT) is integral to modern accounting and management information systems. It is, therefore, imperative that auditors should be fully aware of the impact of IT on the audit of a client’s financial statements, both in the context of how it is used by a client to gather, process and report financial information in its financial statements, and how the auditor can use IT in the process of auditing the financial statements.

The purpose of this article is to provide guidance on following aspects of auditing in a computer-based accounting environment:

  • Application controls, comprising input, processing, output and master file controls established by an audit client, over its computer-based accounting system and
  • Computer-assisted audit techniques (CAATs) that may be employed by auditors to test and conclude on the integrity of a client’s computer-based accounting system.

Exam questions on each of the aspects identified above are often answered to an inadequate standard by a significant number of students – hence the reason for this article.

Dealing with application controls and CAATs in turn:

Application controls

Application controls are those controls (manual and computerised) that relate to the transaction and standing data pertaining to a computer-based accounting system. They are specific to a given application and their objectives are to ensure the completeness and accuracy of the accounting records and the validity of entries made in those records. An effective computer-based system will ensure that there are adequate controls existing at the point of input, processing and output stages of the computer processing cycle and over standing data contained in master files. Application controls need to be ascertained, recorded and evaluated by the auditor as part of the process of determining the risk of material misstatement in the audit client’s financial statements.

Input controls

Control activities designed to ensure that input is authorised, complete, accurate and timely are referred to as input controls. Dependent on the complexity of the application program in question, such controls will vary in terms of quantity and sophistication. Factors to be considered in determining these variables include cost considerations, and confidentiality requirements with regard to the data input. Input controls common to most effective application programs include on-screen prompt facilities (for example, a request for an authorised user to ‘log-in’) and a facility to produce an audit trail allowing a user to trace a transaction from its origin to disposition in the system.

Specific input validation checks may include:

Format checks
These ensure that information is input in the correct form. For example, the requirement that the date of a sales in voice be input in numeric format only – not numeric and alphanumeric.

Range checks
These ensure that information input is reasonable in line with expectations. For example, where an entity rarely, if ever, makes bulk-buy purchases with a value in excess of $50,000, a purchase invoice with an input value in excess of $50,000 is rejected for review and follow-up.

Compatibility checks
These ensure that data input from two or more fields is compatible. For example, a sales invoice value should be compatible with the amount of sales tax charged on the invoice.

Validity checks
These ensure that the data input is valid. For example, where an entity operates a job costing system – costs input to a previously completed job should be rejected as invalid.

Exception checks
These ensure that an exception report is produced highlighting unusual situations that have arisen following the input of a specific item. For example, the carry forward of a negative value for inventory held.

Sequence checks
These facilitate completeness of processing by ensuring that documents processed out of sequence are reject ed. For example, where pre-numbered goods received notes are issued to ac knowledge the receipt of goods into physical inventory, any input of notes out of sequence should be rejected.

Control totals
These also facilitate completeness of processing by ensure that pre-input, manually prepared control totals are compared to control totals input. For example, non-matching totals of a ‘batch’ of purchase invoices should result in an on-screen user prompt, or the production of an exception report for follow-up. The use of control totals in this way are also commonly referred to as output controls (see below).

Check digit verification
This process uses algorithms to ensure that data input is accurate. For example, internally generated valid supplier numerical reference codes, should be formatted in such a way that any purchase invoices input with an incorrect code will be automatically rejected.

Processing controls

Processing controls exist to ensure that all data input is processed correctly and that data files are appropriately updated accurately in a timely manner. The processing controls for a specified application program should be designed and then tested prior to ‘live’ running with real data. These may typically include the use of run-to-run controls, which ensure the integrity of cumulative totals contained in the accounting records is maintained from one data processing run to the next. For example, the balance carried forward on the bank account in a company’s general (nominal) ledger. Other processing controls should include the subsequent processing of data rejected at the point of input, for example:

  • A computer produced print-out of rejected items.
  • Formal written instructions notifying data processing personnel of the procedures to follow with regard to rejected items.
  • Appropriate investigation/follow up with regard to rejected items.
  • Evidence that rejected errors have been corrected and re-input.

Output controls

Output controls exist to en sure that all data is processed and that output is distributed only to prescribed authorised users. While the degree of output controls will vary from one organisation to another (dependent on the confidentiality of the information and size of the organisation), common controls comprise:

  • Use of batch control totals, as described above (see ‘input controls’).
  • Appropriate review and follow up of exception report information to ensure that there are no permanently outstanding exception items.
  • Careful scheduling of the processing of data to help facilitate the distribution of information to end users on a timely basis.
  • Formal written instructions notifying data processing personnel of prescribed distribution procedures.
  • Ongoing monitoring by a responsible official, of the distribution of output, to ensure it is distributed in accordance with authorised policy.

Master file controls

The purpose of master file controls is to ensure the ongoing integrity of the standing data contained in the master files. It is vitally important that stringent ‘security’ controls should be exercised over all master files.

These include:

  • appropriate use of passwords, to restrict access to master file data
  • the establishment of adequate procedures over the amendment of data, comprising appropriate segregation of duties, and authority to amend being restricted to appropriate responsible individuals
  • regular checking of master file data to authorised data, by an independent responsible official
  • processing controls over the updating of master files, including the use of record counts and control totals.

Computer Assisted Audit Techniques (CAATs)

The nature of computer-based accounting systems is such that auditors may use the audit client company’s computer, or their own, as an audit tool, to assist them in their audit procedures. The extent to which an auditor may choose between using CAATs and manual techniques on a specific audit engagement depends on the following factors:

  • the practicality of carrying out manual testing
  • the cost effectiveness of using CAATs
  • the availability of audit time
  • the availability of the audit client’s computer facility
  • the level of audit experience and expertise in using a specified CAAT
  • the level of CAATs carried out by the audit client’s internal audit function and the extent to which the extern al auditor can rely on this work.

There are three classifications of CAATs – namely:

  • Audit software
  • Test data
  • Other techniques

Dealing with each of the above in turn:

Audit software

Audit software is a generic term used to describe computer programs designed to carry out tests of control and/or substantive procedures. Such programs may be classified as:

Packaged programs
These consist of pre-prepared generalised programs used by auditors and are not ‘client specific’. They may be used to carry out numerous audit tasks, for example, to select a sample, either statistically or judgementally, during arithmetic calculations and checking for gaps in the processing of sequences.

Purpose written programs
These programs are usually ‘client specific ’ and may be used to carry out tests of control or substantive procedures. Audit software may be bought or developed, but in any event the audit firm’s audit plan should ensure that provision is made to ensure that specified programs are appropriate for a client’s system and the needs of the audit. Typically, they may be used to re-perform computerised control procedures (for example, cost of sales calculations) or perhaps to carry out an aged analysis of trade receivable (debtor) balances.

Enquiry programs
These programs are integral to the client’s accounting system; however they may be adapted for audit purposes. For example, where a system provides for the routine reporting on a ‘monthly’ basis of employee starters and leavers, this facility may be utilised by the auditor when auditing salaries and wages in the client’s financial statements. Similarly, a facility to report trade payable (creditor) long outstanding balances could be used by an auditor when verifying the reported value of creditors.

Test data

Audit test data
Audit test data is used to test the existence and effectiveness of controls built into an application program used by an audit client. As such, dummy transactions are processed through the client’s computerised system. The results of processing are then compared to the auditor’s expected results to determine whether controls are operating efficiently and systems’ objectiveness are being achieved. For example, two dummy bank payment transactions (one inside and one outside authorised parameters) may be processed with the expectation that only the transaction processed within the parameters is ‘accepted’ by the system. Clearly, if dummy transactions processed do not produce the expected results in output, the auditor will need to consider the need for increased substantive procedures in the area being reviewed.

Integrated test facilities
To avoid the risk of corrupting a client’s account system, by processing test data with the client’s other ‘live’ data, auditors may instigate special ‘test data only’ processing runs for audit test data. The major disadvantage of this is that the auditor does not have total assurance that the test data is being processed in a similar fashion to the client’s live data. To address this issue, the auditor may therefore seek permission from the client to establish an integrated test facility within the accounting system. This entails the establishment of a dummy unit, for example, a dummy supplier account against which the auditor’s test data is processed during normal processing runs.

Other techniques
This section contains useful background information to enhance your overall understanding.

Other CAATs include:

Embedded audit facilities (EAFs)
This technique requires the auditor’s own program code to be embedded (incorporated) into the client’s application software, such that verification procedures can be carried out as required on data being processed. For example, tests of control may include the reperformance of specific input validation checks (see input controls above) – selected transactions may be ‘tagged’ and followed through the system to ascertain whether stated controls and processes have been applied to those transactions by the computer system. The EAFs should ensure that the results of testing are recorded in a special secure file for subsequent review by the auditor, who should be able to conclude on the integrity of the processing controls generally, from the results of testing. A further EAF, of ten overlooked by students, is that of an analytical review program enabling concurrent performance of analytical review procedures on client data as it is being processed through the automated system.

Application program examination
When determining the extent to which they may rely on application controls, auditors need to consider the extent to which specified controls have been implemented correctly. For example, where system amendments have occurred during an accounting period, the auditor would need assurance as to the existence of necessary controls both before and after the amendment. The auditor may seek to obtain such assurance by using a software program to compare the controls in place prior to, and subsequent to, the amendment date.


The key objectives of an audit do not change irrespective of whether the audit engagement is carried out in a manual or a computer-based environment. The audit approach, planning considerations and techniques used to obtain sufficient appropriate audit evidence do of course change. Students are encouraged to read further to augment their knowledge of auditing in a computer-based environment and to practise their ability to answer exam questions on the topic by attempting questions set in previous ACCA exam papers.

Written by a member of the audit exam team