Once upon a time I worked with a company whose numbers were, so far as I could tell, unreliable.

Not unreliable as in a rounding error. Not unreliable as in having to place asterisks in the annual report.

Unreliable as in a billion dollars a month in unaudited transactions being posted to the general ledger through improvised patch programs that gathered data from an ancient legacy system in which the “source of truth” rotated among three different databases.

Our client’s executive team assured us their financial reportage was squeaky clean. The employees we interviewed who were closer to the action, in contrast, predicted a future need for significant, embarrassing, and high-impact balance-sheet corrections.

Assuming you consider multiple billions of dollars to be significant and embarrassing, not to mention high impact, a few years later the employees were proven right.

How do these things happen? It’s more complicated than you might think. A number of factors are in play, none easy to overcome. Among them:

Confirmation bias: We all tend to accept without question information that reinforces our preferences and biases, while nit-picking to death sources that contradict them. Overcoming this — a critical step in creating a culture of honest inquiry — starts with the CEO and board of directors, and requires vigilant self-awareness. If you need an example of why leading by example matters, and how leader behavior drives the business culture, look no further.

Ponzi-ness: Ponzi schemes — where investment managers use new investor money to pay off longer-term investors instead of using it to, well, invest — often don’t start out as fraudulent enterprises launched by nefarious actors.

My informal sampling suggests something quite different: Most begin with an investment manager making an honest if overly risky bet. Then, rather than fessing up to the investors whose investments have shrunk, they find new investors, putting their funds into bets that are even more risky in the hopes of enough return to pay everyone off and get a clean start.

It’s when that attempt fails that Ponzi-ness begins.

Middle managers aren’t immunized against this sort of behavior. It’s how my former client got into trouble. A manager sponsored the effort to replace the creaky legacy system. Part of the business case was that this would replace a cumbersome, expensive, and error-prone month-end process with one more streamlined and efficient.

When the legacy replacement didn’t happen on schedule the manager was still on the hook for the business case, leading him to turn off the maintenance spigot — hence the need for improvised transaction posting programs.

Delivering pretend benefits by increasing risk is the essence of Ponzi-ness.

View altitude and failed organizational listening: Management knows how the business is supposed to work. They are, in general, several steps removed from how it actually works, depending on lower-level managers to keep them informed, who rely on front-line supervisors to keep them informed, who in turn rely on the employees who report to them to make sure (that is, provide the illusion) that they know What’s Going On Out There.

Executives enjoy the view from 100,000 feet; middle managers from 50,000. Smart ones recognize their views are at best incomplete and probably inaccurate, so they establish multiple methods of “organizational listening” to compensate.

Those who skip levels to direct the action are, rightly, called micromanagers. And yet, everyone below them in the management hierarchy has a personal incentive to keep bad news and their manager as far apart as they can. The solution is to recognize the difference between expressing interest in What’s Going On Out There and needing to direct it.

Managers should listen to everyone they can, but instruct only those who report to them directly.

Holding people accountable: As discussed in this space numerous times and detailed in Leading IT, managers who have to hold people accountable have hired the wrong people. The right people are those who take responsibility. Managers never have to hold them accountable because they handle that little chore themselves.

But those who have bought into the hold ’em accountable mantra effectively block the flow of What They Need to Know because why on earth would anyone risk telling them?

If something is amiss in an organization, someone in it knows that something is wrong, and usually knows what to do about it.

What they too-often lack is an audience that wants to know about the problem, and, as a consequence, has no interest in the solution.