Financial Reporting & Market Integrity — Earnings Management & Reserves · Public Companies / Finance & Accounting
The Quarter Is Two Cents Short of Consensus, So the CFO Asks the Accounting Team to “Revisit” the Warranty Reserve. Every Number Stays Within a Defensible Range. The Gap Closes. Eight Quarters Later, a Director Notices the Reserve Is Released Every Time Earnings Would Otherwise Miss. Is It Still Earnings Management?
A three-perspective earnings management compliance scenario — the CFO who asked the team to revisit an estimate, the senior accountant who owns the reserve model and was asked to move an assumption, and the audit committee chair who sees the smoothing pattern only when she looks across eight quarters at once.
Quick Answer
Can a company adjust its reserves to hit an earnings target if every estimate stays “within a defensible range”?
No. Accounting estimates — warranty reserves, bad-debt allowances, returns reserves — must reflect management’s best estimate based on the facts, not be tuned to produce a desired earnings number. Building extra reserves in strong quarters (“stocking the cookie jar”) and releasing them in weak quarters to smooth earnings is improper earnings management, even when every individual estimate falls inside a range an auditor would accept. The violation is using the estimate as a lever to hit a target.
A CFO has the authority to challenge assumptions and demand rigor. A CFO does not have the authority to direct an estimate to a number that closes an earnings gap. The correct behavior at every seat is the same: set the estimate from the facts, document the basis, and report the result the facts produce — even when that result is a miss.
Four Pressure Types Active in This Scenario
Pressure 1 — The Number
“We’re two cents short of consensus. Two cents. The estimate has a little room in it. Nobody is asking anyone to invent revenue.”
Pressure 2 — Authority
“The CFO asked me to revisit the assumption. He owns the financials and signs the certifications. Re-running the model is doing my job for the person above me.”
Pressure 3 — Prudence / Normalization
“Estimates always have a range. Being conservative in good quarters and easing off in tight ones is just smoothing the curve — every company manages the estimate a little.”
Pressure 4 — The Lane Rationalization
“Reserves get reviewed by the controller, the CFO, and the external auditors. If my assumption were really wrong, one of them would catch it. It is not my call to make.”
Three Moments. No Fake Document Anywhere.
An ask that sounded like rigor. An assumption moved a few points within its range. And a pattern that only became visible when someone finally looked across eight quarters at once.
The Leader’s Moment — The Ask
Glenn Saunders is CFO of Quorvane Instruments, a NASDAQ-listed maker of medical and industrial instruments with about $500 million in annual revenue. It is the final week of the quarter, and the preliminary numbers land two cents below consensus EPS. A miss this small, against a long streak of meeting the number, will be read by the market as the streak breaking.
Glenn does not ask anyone to invent revenue or hide a cost. He asks the technical accounting team to “revisit” the warranty reserve. The reserve, he points out, has been conservative for years — the product has been reliable, and the assumed failure rate looks high relative to recent experience. “Make sure the assumption reflects reality,” he says. “If the number’s been too cautious, fix it.”
Lowering the assumed failure rate a few points releases warranty reserve back into income — and closes the two-cent gap almost exactly.
Glenn believes he is doing his job: challenging a stale assumption, demanding the numbers reflect reality. He has the authority to question any estimate in the building. What he does not register is the sequence — that he reached for the reserve only after he saw the gap, that the “reality” he wants the assumption to reflect is the EPS target, and that he is using an accounting estimate as a dial to produce a number rather than letting the facts produce it.
The Accountant’s Moment — The Assumption
Rita Almeida is the senior technical accountant who owns Quorvane’s warranty reserve model. The request comes to her: re-run the model with a lower assumed failure rate. The new rate is not absurd — it sits inside the band of recent product data, and she could write a memo supporting it that an auditor would likely accept.
But Rita knows what she actually saw. The failure data did not change this week. The model did not get better information. What changed is that the quarter came in two cents light, and now the assumption needs to move in the one direction that releases reserve. She has been here long enough to recognize the rhythm — the estimate gets “revisited” in tight quarters and quietly rebuilt in strong ones.
She knows three things. The CFO asked, and he signs the financials. The close is tonight. And she is not certain whether picking a defensible point inside a defensible range — at someone else’s direction — is her problem or his.
She is deciding now. The model is open on her screen, the new assumption already typed into the input cell, the gap already closing in the output.
The Director’s Moment — The Pattern, Eight Quarters Later
Marjorie Vance chairs Quorvane’s audit committee. A former public-company CFO and CPA, she is the board’s designated financial expert. Preparing for a routine reserve review, she asks for something the committee rarely sees laid out together: the warranty and bad-debt reserve roll-forwards for the last eight quarters, lined up against reported EPS versus consensus.
The picture is quiet but unmistakable. In every quarter the company would otherwise have missed, a reserve was released just enough to meet the number. In strong quarters, reserves were rebuilt. No single quarter’s estimate is indefensible. Each memo is reasonable on its own. But the reserves move with the EPS target, not with the underlying experience — the signature of earnings smoothing.
Marjorie knows what she is looking at, and she knows what it implies: that the estimates may have been used to manage earnings, that “within the range” is not the standard, that prior financials the CEO and CFO certified may be misstated, and that a correction of any kind would pull in materiality analysis, the auditors, internal-control conclusions, and the clawback rule.
She also knows the hardest part: nobody forged anything. Each individual number had a memo attached to it. She has to decide what to do about a pattern, not a document.
Three Sets of Choices.
For Glenn, before he reached for the reserve. For Rita, before she changed the assumption. And for Marjorie, deciding what to do about a pattern she can see but no single memo confirms.
For Glenn — What Should He Have Done Instead?
Choice A. Do what he did — ask the team to “revisit” the warranty reserve once he saw the gap, knowing a lower assumption closes it, and frame it as making the estimate reflect reality.
Choice B. Separate the two questions completely. If he genuinely believes the warranty assumption is stale, raise it with the controller and auditors on its own timeline, supported by the failure data — never in close week, never sized to a gap. And if the quarter is two cents short, report it two cents short. A small, honest miss is survivable; a managed estimate is not.
Choice C. Ask the team to “take a fresh look at all the estimates” and find the two cents wherever it is least visible — spreading the adjustment across several reserves so no single change looks large enough to question.
For Rita — What Should She Do?
Choice A. Change the assumption. It’s within the range; she can document it. The CFO directed it, and the close is tonight. The estimate is ultimately his to own.
Choice B. Hold the change and put the basis in writing. Set the estimate to what the current failure data actually supports, document why, and flag — to the controller and, if needed, the audit committee or auditors — that a change was requested in the coming week with no new information behind it. Her job is the integrity of the estimate, not the EPS result.
Choice C. Make the change, but write a thorough memo justifying the lower rate — building the strongest defensible case for the new assumption so the file looks clean if anyone ever asks.
For Marjorie — What Should She Do With the Pattern?
Choice A. Let it go. Every quarter had a memo, every estimate was inside its range, and the external auditors signed off on it each year. Raising “a pattern” with no single wrong number risks looking like she’s chasing ghosts and second-guessing management.
Choice B. Treat the pattern as what it is. Direct an independent review through the audit committee — with independent counsel and forensic accountants reporting to the committee, not to management — preserve the records, and scope the review to determine whether the estimates were used to manage earnings. Run the SAB 99 materiality analysis on the affected periods, assess whether the conduct indicates a material weakness in internal control, determine restatement versus revision, weigh disclosure and any SEC self-report, and run the mandatory clawback analysis. Decide on Glenn after the facts are scoped, with counsel.
Choice C. Handle it quietly inside management. Tell Glenn privately to stop using the reserve to smooth or tighten the estimate process going forward, and to avoid the disruption and exposure of a formal look-back at prior periods.
The Right Calls
For Glenn: Choice B — separate the estimate from the target, and report the miss.
Choice A is the category error: reaching for an estimate after seeing the gap, and sizing the change to the gap, turns a best-estimate judgment into a dial. The tell is the sequence — the number drove the assumption, not the facts. Choice C is worse because spreading the adjustment to avoid detection is evidence of intent. A genuine view that the assumption is stale is legitimate — but it gets raised on its own timeline, with data, away from close week. Two cents reported honestly is a footnote; two cents manufactured from a reserve is a restatement.
For Rita: Choice B — set the estimate to the facts and document the request.
Choice A is the most dangerous because “within the range” and “the CFO told me to” feel like cover and are not: an estimate moved to hit a target is improper, no matter who asked or how defensible the endpoint is. Choice C is the subtle trap — a polished justification memo for a change with no new information doesn’t protect her, it documents that the change was reverse-engineered. Rita’s role is to ensure the integrity of the estimate. The facts didn’t move this week; the assumption shouldn’t either, and the request to move it is exactly what the controller and audit committee need to know.
For Marjorie: Choice B — investigate the pattern independently, then decide.
Choice A mistakes “no single wrong number” for “nothing wrong” — but earnings management lives in the pattern, and a financial-expert director who sees it and looks away has a governance problem of her own. Choice C is the trap dressed as pragmatism: a quiet management fix leaves the audit committee investigating conduct through the very people whose conduct is in question, forfeits independence, and skips the analysis the law requires once a possible error surfaces. Choice B is the only path that treats this as an accounting determination, an internal-control conclusion, a disclosure decision, and a clawback question — handled by the committee, not by management.
Why This Is Harder Than It Looks
There is no fake document — the fraud is in the judgment.
Every other financial-statement scheme leaves an artifact: a backdated invoice, a hidden side letter, a stuffed channel. Cookie jar reserves leave a memo that looks correct. The misconduct is not in the number itself; it is in why the number was chosen. That is what makes it the hardest form of earnings management to see, to prove, and to train people to refuse.
“Within the range” is not the test — intent is.
A defensible range is where the honest estimate lives, not a menu to choose from based on the EPS you need. The same endpoint can be perfectly proper if the facts lead there and improper if the target led there. The standard is the best estimate from the evidence, and an estimate selected to hit a number fails that standard even when an auditor would accept the endpoint in isolation.
The pattern is the proof — and no one sees it from inside a single quarter.
From inside any one close, each adjustment is small and explainable. The earnings management only becomes visible in the reserve roll-forward across many quarters, where the releases line up with the misses and the build-ups line up with the cushion. This is why the SEC’s long-standing scrutiny of “cookie jar” reserves — dating to Arthur Levitt’s 1998 “Numbers Game” speech — focuses on trends over time, and why the person most likely to catch it is a director looking across quarters, not a preparer looking at one.
The cushion feels like prudence, which is exactly how the jar gets stocked.
Over-accruing in a strong quarter feels responsible, even virtuous — who objects to being conservative? But deliberately building reserves beyond the best estimate to create a cushion for later is the same violation as releasing them, just running the other direction. And once a pattern of managing earnings through estimates exists, it can signal a material weakness in internal control over financial reporting, implicate the CEO and CFO’s SOX certifications, and — if a restatement results — trigger a mandatory, no-fault clawback of incentive pay under SEC Rule 10D-1. (As with any live matter, confirm the specific accounting, internal-control, and disclosure treatment with your auditors and counsel.)
Frequently Asked Questions
What is a cookie jar reserve?
A “cookie jar” reserve is an accrual — a warranty reserve, bad-debt allowance, returns reserve, or similar estimate — that is deliberately set higher than the best estimate in strong periods to create a cushion, then released into income in weak periods to boost reported earnings. Using reserves this way to smooth earnings is a form of improper earnings management. It became a focus of SEC enforcement after the agency’s late-1990s campaign against earnings management.
Isn’t it just prudent to be conservative with reserves?
No. Accounting estimates are supposed to reflect management’s best estimate based on available evidence — not the most conservative number or the most favorable one. Deliberately over-accruing to build a cushion is a misstatement in the period the reserve is over-built, just as releasing an unsupported reserve is a misstatement when it is released. “Being conservative” is not a defense when the conservatism is a tool for managing future earnings rather than a faithful estimate of the obligation.
If an estimate is “within a defensible range,” how can changing it be wrong?
Because the violation is in the intent, not the endpoint. The proper estimate is the one supported by the evidence; a range exists because reasonable people weighing the same facts can land in slightly different places. Selecting a point in that range because it produces a desired earnings number — rather than because the facts lead there — is earnings management even if the chosen number, viewed alone, looks acceptable. The tell is the sequence: when the target is identified first, and the assumption is moved to meet it, the estimate has become a lever.
How would anyone ever detect earnings smoothing through reserves?
Rarely from a single period — each quarter’s estimate looks reasonable in isolation. It surfaces in the reserve roll-forward over time: releases that consistently coincide with quarters that would otherwise have missed earnings targets, and build-ups that coincide with quarters that had cushion to spare. When reserve movements track the earnings target rather than the underlying experience, that correlation is the signature that auditors, audit committees, and the SEC look for.
What is the difference between a legitimate change in an estimate and earnings management?
A legitimate change is driven by new information or a genuine change in facts — updated failure data, a shift in customer payment behavior, a new contract term — and it gets made when that information arrives, in whatever direction the evidence points. Earnings management is driven by the desired result: the estimate is revisited because a number needs to move, the timing coincides with reporting pressure, and the direction is always the one that helps the target. Same input cell, opposite purpose.
Does earnings management through reserves trigger SOX and clawback consequences?
It can. A pattern of using estimates to manage earnings can indicate a material weakness in internal control over financial reporting, and the CEO and CFO certify both the financial statements and the effectiveness of those controls under SOX Sections 302 and 906. If correcting the affected periods requires a restatement, SEC Rule 10D-1 requires a recovery analysis of executive incentive compensation, and that requirement applies to both Big R restatements and little r revisions, on a no-fault basis. Organizations should confirm the specifics with their own auditors and counsel.
How to Use This Scenario in Training
Recommended for finance and accounting teams, controllers, technical accounting and reserve owners, FP&A, internal audit, and audit committees at public and pre-IPO companies. Because the misconduct lives in judgment rather than in a forged document, it is especially useful for experienced accounting professionals who would never falsify a record but may not recognize a directed estimate for what it is. It works well as a mixed session — preparers in the room with the controller, the CFO’s organization, and a director or internal audit — so the people who set estimates and the people who review patterns hear each other.
This scenario demonstrates the authority-boundary and normalization patterns from the Decision Readiness Engine™. The debrief question for finance audiences: “How would we tell the difference, in our own close, between updating an estimate because the facts changed and moving it because the number needed to?”
It pairs naturally with the Side Letter and Channel Stuffing scenarios — three different roads to the same restatement — and with Compliance Conversations Episode 13, which covers the SAB 99 and Rule 10D-1 machinery in depth.
More Financial Reporting & Market Integrity Scenarios
Sibling Scenario
The Side Letter →
A VP closes the quarter with verbal promises that never reach the order form — revenue recognition pressure on the deal side.
Sibling Scenario
Channel Stuffing →
A VP loads distributors with product they haven’t sold to make the number, and the new CFO inherits the air pocket.
Browse the Cluster
Financial Reporting & Market Integrity Scenarios →
Revenue recognition, channel stuffing, earnings management, audit interference, SOX certification, and the recognition behaviors that keep small compromises from becoming restatements.
Want Financial Reporting Scenarios for Your Program?
Xcelus builds scenario-based training for finance and accounting teams, controllers, and audit committees — covering earnings management, revenue recognition, and recognition behaviors that prevent misstatements before they’re booked.
© 2005–2026 Xcelus LLC. All rights reserved. Scenario content is original work protected by copyright. You may link freely — reproduction or adaptation without written permission is prohibited.
Compliance Reinforcement Kit — Leadership Facilitation Guide
Running This Scenario as a 20-Minute Leadership Discussion
This guide is for the leader or facilitator running the discussion. No deep accounting expertise required — the format follows five steps, and the answers are above.
Step 1 — Read the Situation (3 min)
Read all three moments aloud — Glenn’s ask, Rita’s assumption, Marjorie’s pattern. Ask the room to sit with it for sixty seconds before anyone speaks.
Step 2 — Name the Pressure (3 min)
Ask: “Which of the four pressures would be hardest for you to push back against?” Finance rooms usually name the prudence/normalization pressure — “it’s within the range, everyone does it a little” — which is the one worth dwelling on, because it is the rationalization that makes cookie jars feel responsible rather than wrong.
Step 3 — Show of Hands (2 min)
Ask for a show of hands on Rita’s choices — A (change it), B (hold and document), or C (change it but write a strong memo). Choice C often draws the most support from experienced accountants, because a well-supported file feels like diligence. Use that to make the key point: a justification memo for a reverse-engineered change documents the intent rather than curing it.
Step 4 — The Right Answer and the Key Concept (5 min)
The answer is B at all three seats. The concept to land: the standard is best estimate from the facts, not “anywhere inside a defensible range” — and the tell for earnings management is the sequence, the number driving the assumption instead of the facts driving it. If the room pushes back with “but it was within range,” the response is: “A range is where the honest estimate might land. It is not a menu you pick from based on the EPS you need.”
Step 5 — The Key Question (7 min)
Ask: “In our last few closes, has an estimate ever been ‘revisited’ late, in the direction that helped the number? Would we even be able to see the pattern if it were happening?”
This moves the scenario from hypothetical to a real read on your close process. The follow-up if the room goes quiet: “Who in this company actually looks at the reserve roll-forward across quarters — and do they look at it next to EPS versus consensus?”
© 2005–2026 Xcelus LLC. All rights reserved. This content is for training and discussion only and is not legal advice; consult qualified counsel about your organization’s specific obligations.