Population Sampling in a Damages Analysis
Written by: Chandler Simich, CPA
We were engaged by the defendant in a lawsuit to determine whether allegations involving the misappropriation of company assets had occurred. The plaintiffs were claiming that the defendant inappropriately used company monies for personal use without reimbursing the company. We reviewed and tested a client-produced population of thousands of personal and business expenses over nine years. In addition, we tested all reimbursement payments made by the defendant during this time period and calculated whether there was a net amount owed to the plaintiffs due to excessive personal use.
The first step in our damages analysis was to ensure the population of personal and business expenses was complete and accurate.
If the population is not complete and accurate, the results obtained are not a true representation that can be extrapolated across the whole population and, therefore, the population cannot be relied upon.
We started by assessing the number of transactions and types of expenses within the population to gain an understanding of the level of risk over the population. Based on these factors and our understanding that the population was prepared by our client, we selected a sample size of 5% of the entire population to test for accuracy and completeness. Any deviation in the 5% sample size would be extrapolated over the entire population to evaluate the accuracy and completeness of the listing. To substantiate these transactions, we obtained over 200 credit card statements to be used as third-party evidence. Third-party evidence is reliable when testing a data set because it is produced by an unbiased source and is unlikely to be manipulated.
We traced transactions selected from the population to the credit card statements (were the transactions properly recorded in the population?) and from the credit card statements to the transactions selected from the population (were all transactions from the credit card statements recorded in the population?). Through this process, we were able to confirm that all transactions in our sample size of 5% were properly accounted for in the population. Therefore, with no deviations in our sample, there were zero errors to extrapolate over the population; indicating that the population was complete and accurate.
The second step in our analysis was to verify that all personal expenses were properly identified.
Our client identified 800 personal expenses out of a total of 4,000 transactions within the population. We verified that the 800 personal transactions identified by the client were in fact for personal use and agreed to the credit card statements. Although there was minimal risk that the defendant purposefully misclassified expenses as personal when they actually served a business purpose, there is always a risk of error. Therefore, we believed it was important to verify the total personal expenses incurred.
In this case, the largest risk was our client not appropriately identifying all personal expenses and misclassifying them as business expenses. To test for this risk, we set a testing threshold of 5% for all business labeled expenses. For each of the expenses we sampled, we reviewed invoices, receipts, and other evidence management provided to determine the purpose of each expense. Through this process, we identified a few personal expenses incorrectly labeled as business expenses. Therefore, we took the percentage of incorrectly labeled expenses and multiplied the percentage by the total dollar value of business expenses, to extrapolate the dollar variance of these errors over the population. This resulted in an additional $4,000 owed to the company.
The third step was to verify the reimbursements that the client had indicated were made to the company.
We obtained check copies and bank statements for all 30 reimbursement payments made to the company. These were crucial to review because there was risk that personal expenses were not reimbursed or were reimbursed with company monies. Through this procedure, we were able to verify that all of the indicated reimbursements were in fact repaid to the company appropriately from the defendant’s personal accounts.
In the final step, we completed a reconciliation of our analysis to arrive at the net personal expenses owed by the defendant.
We calculated the total verified personal expenses owed to the company of $120,000, plus the extrapolated variance of $4,000 for improperly labeled business expenses, and then subtracted the verified reimbursements of $100,000. In conclusion, we were able to provide a summary of procedures performed along with the total concluded amount of $24,000 still owed to the company.
The client and attorneys appreciated our team’s experience identifying data discrepancies, synthesizing data, and our ability to clearly communicate our findings.
Overall, it is important to assess the level of risk over each data set prior to setting a sampling threshold. Then, once a sampling threshold is set, it is important to obtain sufficient evidence to properly test the sample to reach a conclusion.
Although we did not use a software program to analyze the specific dataset in the case above, Cogence Group does use a product called Valid8, a SOC2 certified, professional-grade financial verification software to extract and reconcile data utilized in many of our financial forensics projects. Using software, coupled with our Firm’s extensive auditing experience, we are able to efficiently analyze large datasets with many different types of data to best serve our clients.