“Due diligence” originated as a legal term of art that describes the duty owed by officers and directors when making decisions on behalf of a corporation. The term has a more generic meaning, however, which is to describe the investigation process related to any decision. A customer in a restaurant who asks if the fish is fresh has exercised some amount of due diligence. A customer who asks to go into the kitchen to inspect the fish has exercised even more. Before investing your hard-earned dollars in a new business venture, think about smelling the fish.
The problem with most due diligence is that its focus is procedural and not substantive. “Due diligence” checklists abound, but they merely provide a framework for collecting information. There are plenty of business deals in which the checklists were followed, but the results were still tragic. Some of the problem is a shortcoming in analyzing the information. Some of the problem is that checklists tend to seek historical information that might be of little relevance to the future. Obviously, there are some deals that go bad for reasons that could not be foreseen.
One of the most common shortcomings of most due diligence is an inability to see through statistics. England’s former Prime Minister Benjamin Disraeli once stated: “There are three levels of deceit in the world; lies, damnable lies, and statistics.” While some people may use statistics to intentionally mislead, others may do so unwittingly.
Substantive due diligence involves scrutinizing information to determine whether it matters, whether it’s accurate but potentially misleading, or whether it reveals something that was not intended. I taught an MBA course in Due Diligence, and my goal was to teach critical thinking. To do that, I took claims from a variety of sources outside the business context and tried to teach my students to scrutinize them. If you are able to accurately answer the problems that follow, you are probably capable of conducting your own substantive due diligence. If not, you may want to enlist a professional.
I started the semester by giving my students this problem:
Problem 1: You have been made the head of a charitable foundation whose mission is to promote women’s health. You are given $1 million to donate to the one charity you think will provide the most benefit. You are approached by the American Heart Association, who informs you that heart disease is the leading cause of death among women. You are also approached by the American Cancer Society, who informs you that the life expectancy for women diagnosed with breast cancer has increased more than 100% in the last ten years. Which organization will get your money?
While the cancer statistic has probably changed since the time I taught the class, procedural due diligence would involve research that would, in fact, substantially confirm both claims.
Many of my students fell for the intended misdirection and tried to decide whether the money should go to the biggest problem (heart disease) or the one where money was making the most difference (breast cancer). The real purpose of the exercise, however, was to demonstrate that both claims were either irrelevant or potentially very misleading.
The problem with the heart disease claim is that the real question is not: “What is the leading cause of death among women?” The real question is: “What is the leading cause of premature death among women?” While an internet search would confirm that heart disease is listed as the leading cause of death among women, it would require skepticism and additional research to realize that the main reason heart disease is the leading cause of death among women is because in the 1990s a decision was made that death certificates would no longer list “natural causes” or “old age” as a cause of death. Thus, when a ninety-five year old woman simply doesn’t wake up one morning, the cause of death would be that her heart stopped – and statistically that death is included with other heart disease deaths. If you have any doubt at all about this contention, think of how many women you know who have died prematurely because of heart disease and how many more you know who have died because of breast cancer.
Do not assume, however, that breast cancer is the automatic winner in the hypothetical. The claim that the life expectancy subsequent to a breast cancer diagnosis has greatly increased might mean that great strides in treatment are being made. On the other hand, it might mean that the attention to the issue has resulted in earlier diagnoses. The real question is whether treatments have increased life expectancy, and average life expectancy after diagnosis does not necessarily answer the question.
Rule No. 1 – Does the information actually answer the relevant question?
Here’s the second problem I gave my students:
Problem 2: The literacy rate in Arkansas increased 5% while Bill Clinton was the governor, while the literacy rate in Texas decreased 6% while George Bush was governor (I don’t remember the exact percentages, but Arkansas saw its rate go up and Texas was the reverse). Which one was the education president?
Don’t answer too fast, and don’t think there is a political message embedded in the hypothetical. Consider what is meant by the literacy rate and what would cause it to go up or down. Generally, literacy is defined as the ability of an adult to read and write (although some measures may look at the ability to read and write English). Also, consider the fact that most children learn to read by age 8. Now can you answer the question?
The statistics are not for high school graduation rates or for the literacy rate of high school graduates. Even if they were, the responsibility for the literacy of graduating high school students should focus on why the students did not learn to read ten years earlier. Thus, if the statistic was really meant to suggest a decline among recent high school graduates, the responsibility or blame would belong to the administrations from ten years earlier and not the men who were in office at the time of graduation.
The reality, however, is that the literacy rate is a calculation of the number of people who can read divided by the number of people total. The rate can be affected by changing either number. Thus, if you have a lot of rural poor who grew up during or before the depression and never learned to read, your literacy rate will go up as those people die. Similarly, if you have a huge influx of illiterate immigrants, your literacy rate will go down. In short, the change in literacy rate in both cases had little if anything to do with the educational system in place or the governors at the time the information was tallied.
Rule No. 2: Understand all of the variables involved in any statistical analysis.
Problem 3: This one arises out of a speech made by a Minnesota Commissioner to my Rotary Club. He asserted that the mortality rate for people who did not complete high school is significantly higher than for those who did. He further claimed that the statistics proved we could save several thousand lives a year by ensuring that all students stayed in school and graduated.
The underlying statistic is completely true. His conclusion is completely false. There is no question that there is a correlation between completing high school and premature death. For example, people who use Meth tend to be high school dropouts, so it is not surprising that there is a correlation. On the other hand, there is no apparent cause and effect. In other words, there is nothing that these casualties would have learned in school that would have made a difference.
Rule No. 3: Understand the Difference between Correlation and Cause and Effect.
Problem 4: “Since the oil boom, the violent crime rate in North Dakota has risen 40%.” The typical reaction is to view this as a grave situation. The reason for that is most people don’t understand rates and percentages. They conclude that if the violent crime rate was a mere 5% before, it must be really bad now that it’s at 45%. But that’s not the new rate. If the violent crime rate was 5%, an increase of 40% means the rate is now 7%. While any increase is unfortunate, a 40% increase is insignificant when the starting rate was relatively low. Furthermore, in this particular case, the population is probably being undercounted because there is no easy way to calculate the number of people present in the state – as opposed to the number of citizens counted at the last census. As a result, it is also likely that the numerator (number of violent crimes) is going up, but is being compared to the census population (as opposed to the number of people actually in the state).
Rule No. 4: Understand the difference between rates and actual numbers.
Problem 5: The issue of global warming is subject to a lot of debates beyond the scope of this article, but there is one misperception worth addressing. Assume the average temperature has increased by one degree in the past decade and will continue at that rate for the next five decades. Does that mean that in fifty years we will regularly experience high temperatures that are five degrees higher than what is currently typical? One of the prevailing fears is that summers would become unbearable.
In fact, even if the average temperature is rising, it is entirely possible that the high temperatures are not. Average temperature is calculated by multiplying the temperature by the period of time that it was at that temperature and then averaging it for the world (or country or other area). Thus, in simple terms, consider a situation where the temperature is 90 for twelve hours and 70 for twelve hours. The average temperature for the day is 80. If the high goes up to 91 and the low goes up to 71, the average is now 81. However, if the high stays at 90 and the low goes up to 72, the average has still gone up to 81 – even though the high has not changed. Similarly, if the high remains at 90, but it stays that temperature for a little more than thirteen hours and is 70 the rest of the day, the average has again increased to 81.
The increase in average temperature in the United States from the late nineties to the early two thousands was primarily that – warmer winters, but not hotter summers. Many people will argue that it does not make a difference because warmer winters are enough to cause problems. I am not expressing an opinion on anything other than the fact that the statistics create a conclusion that is not necessarily accurate.
Rule No. 5: Understand Averages and Weighted Averages.
None of these hypotheticals have anything to do with legal issues or typical corporate due diligence. The principles involved are, however, relevant for analyzing information. Statements and representations can be completely true, but be either meaningless or misleading. Substantive due diligence requires a literal or figurative cross-examination in order to scrutinize the relevant information.