Rebuttal
Taking an opinion apart, and telling you honestly whether it comes apart
We are regularly retained to test an opinion rather than to build one. It is frequently the most efficient way to spend an expert budget: the opposing expert has already done the work of stating a number and a method, and the question is narrower than building an affirmative case from nothing.
Rebuttal work has a failure mode worth naming at the outset. An expert who is retained to criticize will usually find something to criticize, and a report that attacks every line of an opinion tells the trier of fact nothing about which criticism matters. Our objective is the opposite: to reduce the dispute to the smallest number of variables that actually move the result, and then to be right about those.
We will also tell you when the opposing model is basically sound. That answer is worth more to you than a critique we cannot sustain in deposition, and it is better delivered before you disclose a rebuttal expert than after.
Reconciling two valuations to the variables in dispute
Two competent appraisers valuing the same interest on the same date will usually agree on more than they disagree on. They tend to agree on the standard of value, the valuation date, the premise, and often the approach(es). What separates their conclusions is a handful of inputs. A rebuttal report that does not make that visible forces the trier of fact to choose between two long documents rather than to decide a short list of discrete questions.
So we start by separating what is agreed from what is contested. Then we reconcile one conclusion to the other, changing a single variable at a time, so that each step carries its own dollar effect and can be accepted or rejected on its own.
- Standard of value: fair market value
- Valuation date
- Premise: going concern
- Primary method: capitalized cash flow
- Historical financial statements as the base
- Characterization as a non-controlling interest
- Normalized owner compensation
- Personal expenses paid through the company
- Treatment of non-recurring income item
- Cost of capital
- Discount for lack of marketability
| Variable | Effect | Restated value |
|---|---|---|
| Opposing expert’s concluded value | 7,560,000 | |
| Owner compensation normalized to market survey data | (840,000) | 6,720,000 |
| Personal expenses paid through the company, which the opposing expert did not add back | 180,000 | 6,900,000 |
| Adjustment for non-recurring income item | (900,000) | 6,000,000 |
| Cost of capital, 15.0% to 18.0% | (1,000,000) | 5,000,000 |
| Marketability discount, 10% to 19% | (500,000) | 4,500,000 |
| Our concluded value | (3,060,000) | $4,500,000 |
Presented this way, the court is not asked to prefer one report to another. It is asked whether owner compensation should be normalized, how the owner’s personal expenses should be treated, whether one income item recurs, what the cost of capital is, and what the marketability discount should be. Some of those are usually conceded once the data behind them is on the table.
The recurring grounds in valuation rebuttal are narrower than the field suggests: a standard of value that does not match the purpose or the jurisdiction; guideline companies that are not truly comparable; a cost of capital that is not commensurate with the risk of achieving projected future cash flows; an unsupported marketability or minority discount; double counting between a discount and an adjustment already made in the cash flow; unrealistic projections; and personal goodwill left inside enterprise value where the jurisdiction excludes it.
Testing a damages model
A damages opinion has more places to fail than a valuation, because it has to address causation as well as amount. In testing a damages model, we commonly consider the following.
- Causation, not just correlation. Where a recession, a lost key employee, a change in regulation or the plaintiff’s own decisions contributed to the shortfall, a model that attributes the entire gap to the defendant’s conduct has not been built to answer the question the trier of fact is asked. A proper analysis accounts for other causal factors.
- The but-for construct. Every damages model rests on a description of what would have happened absent the wrongful act. We test whether that counterfactual is stated explicitly, supported by evidence, and internally consistent — including whether the plaintiff had the capacity, capital, and customer base to achieve it.
- Source and integrity of the data. We distinguish contemporaneous business records from figures assembled after litigation began, and independently verified data from data supplied by the retaining party without corroboration. We also check for selective use of favorable periods or datasets.
- Support for each assumption. Growth rates, margins, market share, contract renewals, and useful life all need a documented basis in fact or in an assumption the record supports. Assumptions that trace only to management’s expectations are identified as such.
- Methodology and its application. We confirm the method is accepted in the field and appropriate to the facts, and that it was applied correctly. Accepted methods are frequently excluded for misapplication rather than for the method itself.
- The loss period. Start and end dates drive the result. We test the basis for when losses began, when they end or would have ended, and whether any terminal or perpetuity assumption is defensible.
- Incremental economics. Lost profits means profits, net of the costs that would have been incurred to earn the lost revenue. We test whether cost behavior was analyzed on an incremental basis and whether fixed and variable costs were correctly separated.
- Mitigation and offsets. We consider replacement revenue, avoided costs, salvage, insurance recoveries, and other benefits the claimant received, along with whether mitigation efforts were reasonable under the circumstances.
- Present value, interest, and taxes. Future losses require a discount rate consistent with the risk of the cash flows being discounted. Past losses raise prejudgment interest questions. Tax treatment is addressed where it affects the measure of recovery.
- Double counting and overlapping claims. Where a case pleads multiple theories, we test whether the same economic loss appears in more than one component of the total.
- Sensitivity and reasonableness. We stress the key inputs to see how much the result moves, and we check the output against historical performance, industry benchmarks, total addressable market, and the claimant’s own documents. A model that produces implausible results under mild changes to its inputs is a model that will not hold up.
- Documentation and reproducibility. Figures should tie to a source, and a competent expert on the other side should generally be able to replicate the calculation from the workpapers.
Where the assumptions are contestable rather than wrong, the useful output is not a competing point estimate but a sensitivity analysis: the same model re-run across a defensible range for each key input, so the court can see which assumptions actually drive the number and which are noise. Frequently one or two inputs may account for the majority of the claim, which tells counsel where to spend cross-examination and where a settlement conversation should start.
Testing a tracing or investigative report
Forensic reports fail differently. The method is usually not exotic; the exposure is in the facts, the scope, and the assumptions buried in a schedule that looks arithmetic.
- Factual and mechanical errors. Transfers between two accounts the subject already owns counted as both a source and a use. Duplicate entries where a statement period overlaps. Deposits classified from a memo line rather than from the underlying item. A schedule whose totals do not tie back to the statements it was built from.
- Was the opening position established or assumed? A tracing schedule that begins in the middle of a period, with a starting balance taken as given, proves less than it appears to.
- Assumptions doing the heavy lifting. That every unexplained withdrawal was personal. That every related-party transfer lacked business purpose. That cash deposits of unknown origin were unreported income. Each may be right; none is a finding until it is tested, and each is worth isolating so its effect on the total can be seen.
- Scope that stops short of the question. This ground is frequently productive and is pressed less often than the arithmetic. Accounts known to exist but never obtained. A period that ends before the transfers at issue. An affiliated entity excluded. Credit cards examined while the line of credit funding them was not. Expanding the scope may change the conclusion more than criticizing the schedule ever would.
- Offsets and contributions ignored. In commingling matters, tracing that follows funds out of an account without accounting for what the other party put into it produces a one-sided answer.
- Source documents, or a summary of them? Reliance on a client-prepared schedule, accounting exports that were never reconciled, or a sample whose population and selection method are not stated.
Where the underlying records will not support the opposing conclusion or ours, we say so, and we identify what would have to be obtained to reach an opinion. That is a more useful answer than a schedule built on assumptions we would have to defend.
What we need, and how it starts
To scope the work we need the opposing report and its exhibits, the workpapers, the underlying data the expert relied on, and the deposition transcript if one exists. The workpapers are the item most often not requested and most often decisive: the report states conclusions, the workpapers show whether the reasoning behind them exists.
Most rebuttal engagements begin with a preliminary read — a fixed-fee written screen of the opposing report identifying whether there is a sustainable critique, which issues carry weight, and what a full engagement would require. It is a bounded decision, it takes days rather than weeks, and it sometimes ends with our advice that you not retain a rebuttal expert at all.
Inquire about a rebuttal report
Please do not send case material until we have completed a conflict check.