People may or may not know come across Benford’s Law which is also known as the law of first digits. It identifies the first digits of numbers found in series of numerical data (refer to below illustration) which can be applied widely from stock prices to your electricity bills. According to this law, “1” is the most frequent occurring approximately 30% of the time, followed by “2” until “9” which is the least frequent occurring less than 5% of the time. The below is an illustration of the distribution.
First Digits
Distribution
The applicability of Benford’s Law depends on a few criteria:
- Data size – the larger the better, preferably 500 or more
- Data information – digit 1 to 9 in the dataset should have equal chance of being the leading digit
- Data source – computer or formula generated data will not work
- Date restrictions – data with maximum or minimum limits will not work
The output of the analysed data should approximately follow the distribution curve in the graph. Otherwise, it may cast doubts on the authenticity of the data which may lead to further investigation.
Financial institutions (FI) deal with an enormous amount of data and information in their day to day operations. Where money laundering and fraud are concerned, licensed or regulated FIs are required to monitor customers’ transactions and detect suspicious, complex, unusually large or unusual patterns. In Singapore, FIs are required to file a suspicious transaction report to the authorities within 15 days.
Compared with manually identifying anomalous transactions for amounts that stick out like a sore thumb, Benford’s Law may be useful as an initial start by simply using MS Excel to plot the distribution of transactions, saving considerable time and effort. For FIs with leaner set-up and resources, and processing vast number of transactions, this method of analysis may just be what they are looking for. Although the output is very preliminary and may not offer deep dive insights, it is still useful as a first consideration to assess if further investigation is required. If the distribution graph output does not conform to Benford’s Law and suspect an anomaly, be sure to confirm again that the dataset is valid.
Disclaimer: The views or opinions expressed are provided for general information and should not be relied upon as legal or professional advice.