Management Control: A Letter from Prison

Unit 3, Professor Eugene Soltes's Control course, A Letter from Prison.

The contracts were real, the products were delivered, and the cash came in. So one manager came to believe that booking revenue a few days early was fundamentally different from the more sensational financial scandals. This case makes you see again that time itself is part of the facts.

Here goes — enjoy:

Case Summary

Computer Associates was an enterprise software company. Stephen Richards joined the company's sales force in 1988 and rose all the way to head of worldwide sales. High growth, high targets and high pay accompanied his climb, and the company lived under constant pressure from quarterly results.

Under the accounting treatment of the time, as the case describes it, once a software licence contract met conditions such as signature, delivery and reasonable assurance of collection, the licence fee could be recognised all at once in that quarter. A customer might buy a term running several years, yet a signature a few days late would change which quarter the entire licence revenue landed in. Customers, too, gradually learned to wait for better terms at quarter-end.

Some managers stretched the quarter operationally, booking contracts signed after the quarter closed as if they belonged to it. The deals could be real; the periods were not. In Q2 of FY2000, correctly recorded revenue was about US$1.047 billion, and another roughly US$558 million was recognised early; reported earnings per share were US$0.60, just above the market expectation of US$0.59 — stripped of the improper recognition, they would have been US$0.05.

After media questions and regulatory investigations surfaced, the company's initial internal investigation did not reveal the full picture. A further independent investigation found the evidence in documents and electronic records that had been left out. The case also describes executives misleading the investigation, and the head of internal legal later admitting to directing this behaviour.

Richards pleaded guilty in 2006 and was sentenced to seven years in prison. In 2008, he wrote from prison to Eugene Soltes, then a student, explaining the targets, incentives, recognition for performance and managerial pressure he had faced. He also wrote that managers often find themselves in grey areas.

The professor's questions therefore have two layers: how serious was the conduct, really; and if the same numbers had been reached using the flexibility the accounting standards allow, would the judgement change? Finally, we have to answer what we would do if we stood in that position, and what consequences we would be willing to bear.

My Thinking

A few days' difference can change the whole picture of the business

Which quarter revenue belongs to affects how others understand a company. In this case, the misattribution made a result that fell well short of expectations look as though it had just beaten them. What it changed was the performance investors saw.

My own judgement is that the fact that the deals were ultimately real cannot offset the fact that the reporting at the time was not. Nor should the total recognised early be called stolen cash or investors' total losses. Criticising a behaviour clearly also requires staying accurate about the numbers.

Understanding the pressure doesn't require cancelling the responsibility

The mechanics of pressure are worth understanding, and the choices individuals made also need to be seen. Richards describes how analysts' expectations found their way into internal targets, then were passed down to teams through the organisational structure, incentives and management style. That is an important explanation.

Another point in the letter struck me: beyond money, recognition for performance and self-identity can also exert enormous force. When "always delivering" becomes part of who you are, admitting to missing a target even once can become extraordinarily hard. A control system that only watches bonuses may miss this quieter driver.

The same numbers call for asking how they were formed

Reasonable accounting judgement and deliberately changing the facts should be evaluated separately. If an estimate is supported by evidence, applied consistently and properly disclosed, it can't be deemed wrong simply because it helped hit a target.

But legality alone can't answer every management question either. Postponing valuable R&D might raise current-period profit while weakening the future; a barely defensible accounting treatment can also make it hard for readers to see how the business is changing. I would look at the facts, the basis for the judgement, the transparency and the long-term value together, rather than only at the profit on the bottom line.

Controls have to hold when the target is about to be missed

The real test happens on the last day, not on the day training is completed. If everyone knows the date can't be changed but no one can bear the consequences of missing the target, the rules will keep getting reinterpreted.

I would record signing, revenue recognition and collection separately, preserving the real dates and the original evidence. For requests that fall outside the normal process, I would provide a channel that allows independent review. That the internal investigation failed in this case is also a reminder: handing a problem to someone with a formal title does not automatically mean it has been handled independently.

The right choice should also spell out its cost

Refusing to extend the quarter could mean smaller bonuses, pressure on the share price, even losing your job. If the classroom answer is just "I would stick to my principles," it skips the genuinely hard part.

My own choice would still be to keep the real dates, report the actual shortfall, and push for independent review. At the same time, I would fix forecasting and incentives so the company can face the business issues of late-signed contracts and customers' quarter-end bargaining. Accepting the short-term cost doesn't guarantee an immediate reward; it simply lets subsequent decisions be rebuilt on credible information.

Takeaways for Arcus42

Conclusions should be traceable to the information available at the time. Original sources, dates and the reasoning process are the foundation for later review. You can't use answers learned afterwards to rewrite the uncertainty of the moment.

Allow people to report honestly that something isn't done. If a team can only report good news, it may gradually shift its energy to maintaining the appearance of completion. Only real progress can support planning the next step.

Dissent needs a usable path. When the person who sets the targets also decides all the evaluations, doubts are hard to voice in full. Someone needs to be able to hear the problem independently.

Metrics should be examined at the level of behaviour. Every seemingly clear number encourages certain choices. Besides asking whether the target was met, we should also ask what people gave up in order to meet it.

Before Class

My pre-class judgement is that this was a serious distortion of reporting. Genuinely reasonable accounting choices deserve a different assessment; practices that mislead about the facts can't be lightly glossed over just because the deals eventually happened.

Reliable records preserve, for the future, a starting point that doesn't need to be explained over and over.

---

This piece is a personal study analysis based on the course case and the letter in it. The facts and accounting context are limited to the period of the case, and nothing here constitutes current legal, accounting or investment advice. It does not cite the preface to Paper to Practice, which was assigned separately and has yet to be obtained.