Tales From the Trenches: How to Successfully Use Statistics—Without Being a Data Guru
In BVR’s monthly publication, Business Valuation Update (BVU), we include a new “Tales From the Trenches” series of articles on valuable lessons to be learned based on the experiences of seasoned valuation experts. While you may be able to find some of this advice in books, only real-life experiences can highlight the nuances found in the minds of the experts who have lived through many engagements. In a recent BVU article, which we’ve excerpted here, veteran expert R. James Alerding, CPA/ABV, ASA (Alerding Consulting) writes about how to successfully use statistics—without being a data guru.
You don’t have to be a statistics guru to successfully use them
Mark Twain said there are “lies, damned lies and statistics.” Some valuation experts may be uncomfortable using statistics in an engagement. But you don’t have to be a statistics guru to successfully use them when the engagement calls for it. For instance, many damages litigations include the use of regression analyses and/or statistical sampling. Both are statistical methods of determining the amount of damages in a dispute.
A regression analysis, for example, can be a method of forecasting or estimating what the future income from an interfered-with business or line of business would be “but for” the interference. It does this by measuring the “dependent variable” from let’s say an historical stream of income to predict what the income would be in the future.
A statistical sampling may also be used in a determination of damages, although sampling has been more of a recent development in litigation matters. A sample is any subset of a population, so its size can be small or large. We want a sample small enough to be manageable by our computing power, yet large enough to give us statistically significant results, so that we do not have to go through every individual item in a population in order to arrive at a conclusion.
The problem, for the financial expert, arises in the level of knowledge of the expert in the areas of regression analyses and statistical sampling. Where a statistical sample is called for in a case, I use a statistical expert to determine the sample and the results of the sample. Then I take those results and determine the damages related to that sample. Where a regression analysis is utilized, I will often perform that analysis myself (or have my staff perform the analysis). In both instances, as an expert, you should be prepared to deal with technical issues that might arise. Following are examples of cases where regression analysis or statistical sampling played a part in the outcome.
Case No. 1: How a regression analysis works
This case involved a travel agency that specialized in trips to Greece. Its primary clients were Greek nationals living in the U.S. It had a working relationship with Alitalia Airlines, which provided the agency with charter flights from Chicago to Athens. When the Olympics was announced for Athens in 2004, Alitalia saw an opportunity and discontinued its arrangement with the travel agency, causing damages to the travel agency. I was the expert for the travel agency. Alitalia did not engage an expert.
In determining the damages, we used a regression analysis. I recognized that Alitalia would come after our methodology since it did not have an expert. Since I do not have an in-depth knowledge of regression analyses, I worked with my staff person who did have such knowledge to train me to be able to answer the questions I was sure were coming on things such as the R2 and the t-test. Sure enough, those were the key questions the attorney asked on cross in an attempt to discredit my work to the jury. I was able to answer the questions well enough that the attorney did not press the issue. The jury accepted my damages.
Lesson: Always be proficient in all the analyses and methods utilized in a report. This is especially important when staff is used to assist in determining damages or valuing a business. The staff might have more expertise than you do.
Case No. 2: What a regression really does for us
The case involved a termination of an arrangement with a sewing business. The business had a contract to provide the sewing process for a business that designed women’s purses. The sewing business sued the purse business for an improper termination of its sewing services. In determining the damages, the opposing expert used a regression analysis to assist in quantifying the damages. However, the expert did not take into account other factors that might have caused a future decrease of business and thus would reduce the damages. For example, the 2008 recession occurred about the time of this action. We made the argument that the purse business would have gone down because of the recession and other factors that we identified, which would have reduced the sales of the sewing company. We also emphasized that not taking these factors into account raised issues of causality and reasonable certainty. This reduced the provable damages by a significant amount.
Lesson: The other expert also was not well versed in the application of a regression analysis and was therefore not well equipped to rebut our criticisms. Once again, the expert should be versed in the methodology used and thus able to support his or her analysis and rebut criticisms.
Case No. 3: A slam dunk using statistical sampling
The case involved an engagement where I represented the Department of Justice in a Medicare/Medicaid fraud and abuse case. The defendant was a medical testing clinic in the Northeast. The assertion was that the defendant had overtested blood panels for various diseases. The panel had the ability to perform 28 tests, but, in reality, in a majority of the tests, fewer than 28 panels were actually tested. Many thousands of tests were under scrutiny for potential fraud. The cost and time to manually go through each of these tests would have been excessive and impractical. At my suggestion, the DOJ engaged a Ph.D. statistician from a major Midwest university that I had referred to it to perform a statistical sample. We used the results of the sample to determine our opinion of the actual damages. The defendants settled immediately after deposing the statistician.
Lesson: The defendants engaged a nationally well-known statistician from Harvard to assist in their deposition of the DOJ statistician. They were unable to do any damage to him. Had I done a sample on my own using, for example, the AICPA protocols for audit sampling, I no doubt would have been ripped to shreds in deposition and the government’s case would have been in jeopardy. The lesson is to rely on other experts where that expertise is required in the engagement.
Case No. 4: Using sampling to stop unnecessary litigation
I was engaged in a case that dealt with a patent infringement. At first look, I was concerned that the potential damages might not warrant a continuation of the case. Rather than spend a lot of money and time calculating damages as though we were going to court, I suggested that we use the same statistician that we used in the DOJ case to do a sample to see what the damages might be. The sampling confirmed my concern, and the attorney and the client decided not to go forward with the lawsuit.
Lesson: It is important, as an expert in a litigation, that one have the proper training and knowledge of the methods and analyses to use in determining damages. If not, then it is imperative to either acquire that knowledge yourself or engage another professional to provide the professional opinion needed to complete the engagement.
Conclusion
Interested in more “Tales From the Trenches” articles? Be sure to check out BVR’s monthly Business Valuation Update. Other topics covered include the buildup approach, company research, and more. Do you have a lesson to share? Contact the BVU editor, Andy Dzamba, at andyd@bvresources.com.
R. James Alerding (Alerding Consulting LLC) has more than 45 years of experience working with businesses in various consulting matters and over 30 years of experience with complex valuation and litigation support matters. He lectures frequently on business valuation and forensic services and has testified on valuation and economic damages in a variety of federal and state courts. He can be reached atjim@alerdingconsulting.com.