How do I detect fraud in my accounting system?
As you may remember, fraud creates anomalous patterns in your accounting data. To find it, we have to dive deep into that data and analyze it to find those patterns. Today we are going to cover a few patterns to look for and how they are linked to potentially fraudulent behavior.
One of the single best covers for fraud is messy records. Reconciliation is the process of matching each transaction in your accounting records for a given period with a “source of truth”, such as a bank statement. When a transaction or, most often, lots of transactions are left unreconciled, it becomes difficult to tell what is legitimate and what is not. Worse yet, you could be issuing incorrect financial statements to various stakeholders without realizing it.
That said, unreconciled transactions are usually not fraud. Most often it is a signal that the month-end close process needs to be tightened up. However, most of the fraud I’ve uncovered in my career was perpetrated by people who seemed incompetent or disorganized on the surface, and used that to their advantage. Remember, fraudsters don’t show up to work in a ski-mask or dress like the Hamburglar.
You have to look for other signs of suspicious behavior. Fraud or not, unreconciled transaction are an issue and should be identified and remediated early.
When stealing a specific amount of money, thieves fall into two categories. Either they steal a round amount or something they consider to be random. Both do so in an effort to conceal their fraud; thinking the amount they stole will be assumed to be an error of some kind rather than fraud. What they don’t know is there are ways of identifying both types of theft.
The key lies in something called Benford’s Law. If I gave you a large set of transaction data and asked you to predict how many of the transactions began with the numbers one through nine, you might reasonably expect that there would be an equal amount of each digit across all of the transactions. In fact, that would be wrong. Benford’s Law predicts the distribution of digits to be logarithmic. In other words, there are a lot more ones than twos, twos than threes, and so on. For instance, there will be over six times more ones than nines. We can use this fact to uncover the two types of thieves.
Round Amounts - if a thief is stealing round amounts of money, there will be too many transactions ending in zero. Carefully reviewing the count and type of transactions ending in zero can be a good way of quickly identifying odd patterns.
“Random” Amounts - when thieves steal “random” amounts of money, they end up not being random at all. Our brains are horrible random number generators, and eventually the amounts stolen will show themselves. This is because the frequency of the leading digit will be too high in those fraudulent transactions or the ones they use to cover their trail.
There are many different tests that are based on Benford’s Law, such as: first digit, second digit, first two digits, first three digits, last two digits, etc. A robust usage of Benford’s Law for analysis will use a combination of all of these.
If there are open invoices or bills outstanding for long periods of time in your accounting system, there is an issue. At the very least you need to tighten up collections or vendor payment processes. The longer something goes without being collected or paid, the more suspect it becomes.
For receivables, you have to question if the payment was received and not coded correctly, or if something nefarious happened in transit? Fraudsters know this, and will often misapply receivables intentionally in a scheme called “lapping”. Essentially, they steal the initial payment from the customer then use future payments to continuously offset the previous receivable. This can be avoided by regularly reviewing your collections process and sending statements to customers so they have a chance to reconcile how their payments are being applied.
For accounts payable, the same question has to be asked. Oftentimes, thieves will duplicate bills in the system under a fraudulent vendor and direct the payments for the bill to the fake vendor instead of the real one. Regularly asking your vendors for statements can avert this becoming a large issue.
To get away with their various schemes, fraudsters will often create fictitious profiles in the accounting system. The profile could be a fake employee, customer or vendor depending on the type of scheme they are running. It is important to regularly check all of these profile names, addresses, phone numbers and email addresses against each other to look for duplicates or suspiciously similar information. There are cases where there will be legitimate overlap, but they are rare.
There should never be a gap in your check numbering. Full stop. If you cut a check that has to be stopped or voided, that should be recorded in the system. Gaps in your check register numbers are dangerous because you have no idea what actually happened to the check. Was it shredded? Did someone steal it and is just waiting to use it? Should it have been voided, but was deleted by accident. Best efforts should be made on a regular basis to find any gaps and remediate them.
When a transaction is made in your accounting system, whether directly by an employee or through an integration with a third-party system, a record is created in the audit log of who created the transaction and when. When a transaction is updated, that information about who did it and when gets stored again in the audit log. If you notice that transactions are being updated after their initial creation, whether to change the payee, date, amount, or any other field, the person responsible needs to be asked their reason for performing the updates. If you notice a particular person is doing most of the updates, that person should be monitored closely as that is suspicious behavior and should not be necessary in a well-functioning team.
These are just a handful of things to look out for when perusing your books. There are many, many more. If you are feeling overwhelmed by all of the various anomalies you need to be looking for, rest assured you’re not alone. It is really difficult to find all of the anomalous data in your records. At least it was until the advent of Scrutinize. Of course, you could continue to spend hours trying to analyze your transactions manually...or you could make your life easy and use Scrutinize to find these patterns quickly and remedy them before they get out of hand.
For answers to questions you have about this or related topics, schedule your free consultation today.