Analytical Aspects of Audit
Stephen Allen, ACBA
Aspects Covered
• Sample selection from paper lists or multiple sources
• Simulating applications in MS Excel
• Presentational power of the ‘If…’ function in MS Excel
Sample Selection
• Methods - advantages & disadvantages
• ACBA function for random selection– preparation for and use of the function– presentation of results and audit trail
• Access to ACBA Random Sampling
Selection Methods
Method Advantages Disadvantages
Judgement Quick Poor coverage, biasedcoverage
Each nth transaction Avoids bias Requires preparation toensure full coverage
Random selection Full coverage, avoidsbias – can be used as abasis for understandinga system
Requires detailedpreparation and a needto remember randomnumber methodologyand how to use it. Oftentakes too long.
Preparation - complete the form
Preparation Explained• Size of sample & test name
– decide how many transactions you want to select– how do you want to identify this test?
• For each source– give it a unique name - one that you can recognise– how many pages are there?– count how many transactions there are
• on the first page
• on the inner pages
• on the last page
• Click ‘Enter Details’– you will be asked whether you want to include another source
Select the sample
• Click ‘Yes’ to the above message box
• unless you have variable numbers of transactions on the inner pages of your list
The Results• You will be presented with
– a list of selected transactions– in ascending order - for ease of locating them– without duplicates– showing
• the random number which determines the page selected
• the page number within the source
• the name of the source
• the nth transaction selected within that page
It will look like
ACBA Random Sampling function - availability
• Available from ACBA web site
• no charge for single copies
• anonymous download
• web site address http://www.acba.co.uk
Simulating Applications
• Deciding what to simulate• Design on a spreadsheet• Gathering input
– input controls
• Processing – creating formulae
• The Results– what do they really mean?
• Summarising the results
What to Simulate• Standard system
– input & associated control
– processing & associated control
– output
• Primary interest is in ‘processing’– all other elements including the system’s controls
can usually be discarded
– otherwise you are, in effect, rebuilding the system
Design on a spreadsheet
• Separate permanent data from transaction data
• Use columns and rows as if you were writing a working paper– left to right then downwards
• Use different colours or other presentational devices to show the data entered and the associated results
• Hide columns or rows of formulae in preference to using spreadsheet protection– protection hinders access to useful functionality
• If possible, keep to a single screen without scrolling
Design on a spreadsheet - in practice
Permanent data separate
Green - to enter data, Red - for result
Rows 14 - 18 hidden for protection
Spreadsheet - impractical to fit on one screen without scrolling
Data input
• You will need – all the transactional and permanent data
necessary to carry out the calculation, and– the results from the original system, so that you
can make a comparison
• Do not allow comparative results to show unless all the necessary data is there
Data input - practice 1
• Wherever practicable explain the type of input required
• Validate the type of input if there is serious risk of error
Data Input - practice 2
• monitor entries as they are made• make sure you gather all the data you need• Cell ‘J6’ states - if Cell ‘I6’ is a number enter fig 1
otherwise leave blank• Cell ‘J9’ states - if all the required entries contain valid
data enter fig 3 otherwise leave blank
Processing
• Mimicking the system’s calculations– break the calculations into easy to understand chunks
– ensure you process them in the correct and logical order
– compare results to system’s output
• Note, when handling cumulative calculations, the first formula is always different to all subsequent ones
Data input and processing- practice 3
The Results
• Literally, that the two approaches do or do not agree– always more likely that the simulation has an error if
agreement not achieved
– validate simulation independently before assuming system error.
• Correction of errors in simulation generates a much improved understanding of the main system
Summarising the Results
• Automated summaries should be adopted only where the simulated system has been independently validated
• Often a good idea to provide for auditor interpretation of results first, even with validated systems
The ‘IF’ Statement
• The classic method of branching in computing
• Structure – set evaluation criteria– action if criteria are met– action if criteria are not met
• Syntax in MS Excel – ‘=IF(Criteria, True, False)’
‘IF’ - a worked example• =IF(D6 <> “”, D6, “”)
• this translates as, IF– the cell at D6 is not blank <> “”– places the contents of D6 in this cell– otherwise leave this cell blank
• very useful for avoiding situations where you do not want a calculation to take place before the appropriate data has been put in