Benford’s Law

Benford’s Law As a forensic accountant I have many ways and techniques to spot fraud. One of the ways to detect fraud, especially when analyzing tax returns, general ledgers and other items that contain a large amount of numerical data is by applying Benford’s Law. Benford’s Law states that any random numbers will have a specific result as to which digits appear first in each data set. The law gives a prediction of frequency of leading digits using base-10 logarithms. The below chart shows what you would expect to find if the information you analyze is indeed random and not somehow altered. When you analyze data you would expect to see the first digit in the data to be a one about 30% of the time, expect a two to be the first digit in the data about 17.6% of the time. When analyzing large data sets it is easy to spot anomalies that tip us off that there is a high probability that the numbers have been manipulated. From there we perform forensic accounting techniques in order to establish fraud. This is one of those techniques that is hard to understand unless we actually go through an example. This video is being made in the time of COVID-19 and there are some concerns about the cases reported to the Centers for Disease Control and Prevention “CDC”.  I am going to perform a Benford’s Law test on the data. The first thing I am going to do is go to the CDC website: https://www.cdc.gov/coronavirus/2019-ncov/index.html From there I am going to download cases by date and state territory. This should give me the total cases per state through today’s date. I will fix the data and then I will use an Excel formula to segregate the data based on the first digit that appears. =LEFT(B2,1) =COUNTIFS ($C$2: $C$61, D2) Once I have this, I am going to then look at it from a percentage standpoint, graph it and compare it to what it should look like as noted above. This analysis is telling me that the data is not random and there is a high likelihood of fraud. You can see that the bar chart does not have an even downward slope as expected.