Sarbanes-Oxley and Dodd-Frank have introduced checks and balances to protect shareholders from corporate misdeeds that erode shareholder value, changing the landscape of corporate governance in the process. But like many other risk mitigation regulations, these legislative actions were reactive. To avoid corporate disasters on the scale of events like Enron, Worldcom and the events leading to the 2008 financial crisis, boards bear the responsibility to ask questions and manage risks before they occur. The question, of course, is: How can they identify the right questions to ask at the right time?
C-Suite had the opportunity to speak with Andy Fastow, who served as chief financial officer at Enron from 1997 until it declared bankruptcy in 2001 at the center of one of the more visible corporate scandals in recent history. Fastow is a keynote speaker at the Equilar Executive Compensation Summit in June.
In our interview, Fastow identified how Enron went under the radar of the company’s leaders and how these types of scenarios may be avoided. In the bigger picture, he is now invested in a company called KeenCorp, a software solutions firm that uses artificial intelligence to detect employee tension. Employee tension is highly correlated to the moment in time when governance, compliance and culture problems are created. For boards, the promise of new technology means better insight into what’s going on beneath the surface at the company. By using quantifiable information to support risk management, they may be better equipped to ask hard questions before an issue becomes a crisis.
C-Suite: What should executives and boards take away from your experience at Enron?
Andy Fastow: One of the biggest challenges that directors have is that they have limited information, and it’s not their role to micromanage a company. As a result, they often do not have the opportunity to address a situation before it becomes a problem, and they are repeatedly confronted with dealing with problems after they’ve become a crisis. Some of the most frequent phrases uttered at board meetings are “Why didn’t we catch this?” or “Why didn’t we know about this?”
From a risk management perspective, what are some of the critical questions boards should be asking that they may not be?
Fastow: There’s a fundamental difference between following the rules and doing the right thing. Many directors are too deferential to experts and advisors, meaning when an auditor or an attorney says it’s legal, the discussion often stops there. Directors should ask two questions, even after the advisors and experts have approved something.
The first one I call the private company question: “Would we make the same decision if this were a private company that I owned, that had my name on the side of the building and that I wanted to leave to my grandchildren in 20 years?”
For public companies, the answer to that question will sometimes be no. That doesn’t necessarily mean they shouldn’t proceed, but it indicates that an additional discussion is warranted, one that asks why a different decision would be made because this is a public company.
The second question boards need to ask is “How do our employees really feel about this decision?” I emphasize that word really. Directors are insulated, and key employees will not necessarily tell them what they are really thinking.
What is the most interesting or surprising thing that you’ve learned based on your experience?
Fastow: The Enron disaster would have been avoided if the directors had the tools available to them today.
Most governance, compliance and ethics programs only identify problems after they have metastasized into formal complaints, lawsuits or enforcement actions. Technology today exists to identify the problems when they are being created so that they can be remediated before they cost the company money and before they do permanent damage to an organization’s culture or workforce. There are now tools that allow companies to understand when their employees are actually uncomfortable with decisions a company is making. At those moments in time, directors should ask additional questions: “Why would employees be uncomfortable? What don’t we know that we should consider?”
What are those tools?
Fastow: My objective in business life is to find or develop the tools to help companies identify these problems much earlier so that the financial cost to the company and damage to their culture can be averted.
A company called KeenCorp has a solution that is the single best tool I have found to do this. I invested in this company after I discovered their software. In my opinion, it is the best technology to identify when governance, compliance and culture problems are being created in a company. It gives directors and human capital management professionals the ability to address issues before they become lawsuits and enforcement actions.
How does the technology work?
Fastow: The software works like a dashboard warning light. When your “check engine” light goes on, you have an expert check there and check now, before your car breaks down on the highway. Most corporations do not have an early warning system, and they wait until the car is broken down.
There is an important distinction in that this is not what people would call “sentiment” software. Traditionally, sentiment software looks for key words, counts happy words vs. sad words, etc. This is based on psycholinguistics, which is more about what people do and how they react than what they say.
When a person becomes tense, certain things happen to his or her body. For example, if you witness a fight, even if you’re not involved, your tension levels will go up. Your blood pressure will rise, and your respiration and perspiration will increase if you see bad or questionable behavior. Other things happen as well, such as your voice inflection changes and your body language changes. If you went home after witnessing that fight, your spouse or partner would likely pick up on that tension, even if that tension had nothing to do with them.
There’s one other signal that shows up, which is critical to the functioning of this software: the manner in which people write changes when their tension level rises. The actual patterns and structures of their sentences change. KeenCorp machine-learning software is trained in pattern recognition of written data streams, or dysfunctional pattern recognition. Dysfunctional patterns indicate tension variations, again the best indicator of “challenged” behavior or decisions. The AI capability enables the algorithm to differentiate between positive and negative changes in employee tension.
The most important thing to note here is that people do not have to be writing about what makes them tense in order for the software to pick up the changes. The software is not evaluating content. If you see that hypothetical fight, you’re going to be tense for a while, and it will change your demeanor. Everything you write, even if you are not writing about that fight, will reflect that tension.
Here’s a concrete example. KeenCorp ran this software for a company that has 24 offices. The dashboard can be displayed like a heat map, and the company had a report with 24 boxes with various colors that represented tension level in each office. For three weeks, we ran the software, and every day it was all various shades of green, which is good. Then all of a sudden one of the boxes turned purple. That’s like the “check engine” light going on.
After investigation, it was found that the head of the office began having an affair with a young associate that week. As with many similar situations, most employees knew, but no one in the office called HR. No one complained or objected, but the situation caused the employees’ tension level to rise. The interesting thing was that the software picked up the employees’ tension level rising without anything being written in emails about that situation. No one wrote about this affair in any emails, but everyone knew about it and it made the atmosphere sufficiently tense that it was picked up as a statistically significant move in the data.
Because they picked up the signal, HR was able to remediate the situation quickly. The HR director told us afterward that remediating so quickly probably saved the company millions of dollars, and that it probably avoided damage to the culture and the loss of good employees.
There’s a famous philosopher Kahlil Gibran who said the truth of another is not found in what they reveal, but what they don’t reveal to you. In other words, if you want to understand your employees, don’t listen to what they say, listen to what they don’t say.
Directors are insulated, and key employees will not necessarily tell them what they are really thinking.
I want to jump back to a really important point. How could this software have avoided the Enron scandal specifically?
Fastow: LJM was the most infamous deal I did at Enron, and it later triggered the bankruptcy. On June 28, 1999, the board of directors approved this deal.
The blue line on the graph represents the tension level of the top 150 employees at Enron, which KeenCorp analyzed using the emails the federal government made public from those individuals after Enron went bankrupt (Graph 1). So these were the executives who knew what was going on. On the graph, up is good, while down is bad, reflecting an unspoken concern or higher tension in communications.
Over the two days after the deal was approved, the tension level of the top 150 employees spiked dramatically (a downward movement in the diagram)—statistically, this was like a Black Swan event. The KeenCorp Index was almost as low on June 30, 1999, as it was on the day of the bankruptcy in 2001. This indicates that the top 150 people in the company thought it was the stupidest decision and that it was an existential type of decision. But here’s the interesting thing: None of them raised their hands. No one challenged the deal.
If Enron’s board of directors had seen this graph on July 1, they would have stopped, called timeout, brought in top management and asked what they were thinking. If the board had seen this data and talked to management, they would have unwound this deal. But they had no signal that told them people were really concerned. So the board approved it and went about their business.
When I first saw the data displayed this way, my heart sank. The board of directors were all serious, smart people who wanted good things for the company, and if they had seen this, they would have undone this deal and history would have been different. You can see after that big dip in 1999, the index never returned to its highest level. This transaction caused permanent damage to the culture and the psyche of the organization.
Here’s the interesting thing: If the board had seen this data, and if they had undone the deal, you might have seen the opposite thing happen. The board would have sent a strong message to the employees that deals like this, which are characterized as “misleading,” would not be approved or tolerated. The culture would have been different.
Here’s an interesting postscript: A former CHRO from Enron contacted us recently when she saw this slide. At the time, her team was doing traditional surveys that were saying everything was great. So they didn’t pick up this huge dip in 1999 or the decline in 2000. She said they did start getting more complaints in 2000 and 2001 but dismissed them as opinion because all the survey data was positive.
Obviously, the corporate landscape has changed considerably since you were at Enron, particularly when it comes to things like technology, automation and even things like social media. What advice do you have for executives and boards today given these shifts?
Fastow: Be very, very scared. That’s my advice. Technology, and social media in particular, highlight board decisions and make it even more important to ask the question of whether, in addition to following the rules, we are doing the right thing.
There’s a big difference between ticking the boxes and asking the penetrating questions that make your company and its culture better.
Boards are confronted every day with a human nature problem. Management doesn’t always want to tell the board what they are really thinking. People don’t like to deliver bad news, especially if it might reflect poorly on them.
And on the other side, I understand why directors might sometimes be reluctant to ask that next question. They often have limited information, and they don’t want to be perceived as being adverse to management. But now there is more objective data available to give them reasons to feel comfortable to ask that next question.
Surveys are inherently flawed. The data is incomplete, it is biased, and it is usually dated. KeenCorp cracks the code on corporate culture. It takes something that was heretofore not measurable and makes it measureable—objectively, completely and in real time. It’s the technology that will replace employee surveys because, like the Enron CHRO example, opinion is easily dismissed, but data is not.
What are some of the things you’ve learned speaking to executives and directors now, long after the fact?
Fastow: I do not aspire to be a public figure. I hope that my talks, which are based on my failings, can lend some insight so that experts can figure out better ways of doing things.
Since I’ve been out of prison, I’ve been fortunate enough to have conversations with many directors. It seems you can generally lump them into two categories: directors who don’t really want to know what’s going on because they are afraid it might require them to do something, or that it will increase their liability if they don’t do something; and directors who really want to make their companies more valuable and do their jobs.
Every director has to take a hard look in the mirror and ask which one they want to be. If it’s the former, in today’s world, they need to step aside and make room for directors who want to proactively make their companies more successful.