Bias Financial Expert Witness
Introduction: Your Only Real Product is Credibility
In financial expert witness work, there are two opinions that matter: yours and the judge’s opinion of you.
Attorneys have clients. Experts have credibility. The moment a jury or a judge thinks you were hired to say something rather than hired to find something, your damage model, your business valuation, or your forensic accounting analysis doesn’t matter anymore.
I have seen great analyses get thrown out — not because the math was wrong — but because the bias was obvious.
Bias in expert work is rarely intentional. It’s subtle, it creeps in, and you have to actively guard against it. Here are the 6 biases that trip up financial experts the most, and how to stay clean.
1. Advocacy Bias: The Client Who Pays You Is Not Your Client
This is the big one. The retaining attorney is paying your invoice, talking to you weekly, and being very persuasive about their theory of the case.
Your duty is not to the attorney. The expert’s duty is to the trier of fact.
Advocacy bias sounds like this in your report: “As Plaintiff correctly notes…” or ignoring inconvenient documents because counsel told you they were “not relevant.”
The fix: Write your report as if it will be read by the opposing expert, opposing counsel, and the judge — because it will be. If you wouldn’t say it on cross-examination, don’t put it in your report. The ABA Model Rules and AICPA/ASA standards all say the same thing: you are an independent advisor to the court.
2. Confirmation Bias: Looking for What You Want to Find
We all do it. You form a hypothesis early — “this company was definitely losing money before the breach” — and then you subconsciously give more weight to the data that confirms it.
In damages work, this shows up as cherry-picking comparable companies, selecting only the high-growth forecast from management’s files, or accepting one data source without testing it.
The fix:
-
- Actively search for disconfirming evidence. Ask: “What document would prove me wrong?”
-
- Document what you requested and what you received. If you asked for 5 years of financials and only got 2 good years, say so.
-
- Run a sensitivity. Show the court what happens if your key assumption changes by 10%.
Judges love an expert who acknowledges weaknesses. It makes the rest of your opinion more believable.
3. Anchoring Bias: The First Number Isn’t Always the Right Number
The mediation demand is $10M. Plaintiff’s lost profits calculation is $8M. The defendant’s counter is $0.
If you let those numbers anchor you, you are no longer independent. I see this constantly in lost profits and business valuation cases where an expert starts with a party’s number and just tweaks it.
The fix: Build your model from zero, using source documents only. Don’t look at the other expert’s report until your draft is done. Then compare and reconcile. That process is also highly defensible on the stand.
4. Hindsight Bias: “It Was Obvious All Along”
This kills experts in fraud, breach of fiduciary duty, and shareholder disputes.
Looking back from 2026, it’s easy to say management should have known a product line would fail or a loan would default. But what did they know at the time, with the information reasonably available then?
The fix: Put a date fence around your analysis. Every opinion should start with “As of…” Use only information that was known or knowable as of that date. If you use subsequent information, you must explicitly state why it is relevant to confirm or refute what was knowable.
5. Selection Bias and Information Bias: Garbage In, Gospel Out?
We are dependent on what we are given, but we are responsible for what we rely on.
Did you rely solely on QuickBooks without checking the tax returns? Did you use management’s unaudited projections without testing them against historical performance? Did counsel only give you the “good” emails?
Under Federal Rule of Evidence 702 and the Daubert standard, you must show that your data is sufficient and your principles were reliably applied.
The fix:
-
- List all documents considered, not just relied upon.
-
- Apply basic forensic checks: Does the P&L tie to the tax return? To the bank statements? To the inventory?
-
- If data is limited, qualify your opinion. A limited, well-supported opinion is better than a broad, unsupportable one.
6. Affiliation Bias: We Like People Who Are Like Us
You worked at a Big 4 for 20 years. The business owner reminds you of your dad. The opposing counsel is rude. These things shouldn’t matter, but our brains are wired to favor the in-group.
The fix: Standardize your process. Use the same checklist, same valuation date memo, same document request list, and same reliability tests for every case, whether it’s a $200k wage case or a $20M shareholder dispute.
A Practical Checklist Before You Submit Any Report
Before I sign any expert report, I run through this:
-
- Would I give the same opinion if I was retained by the other side? If not, why not?
-
- Have all assumptions been disclosed and why I chose them?
-
- Did I address the contrary evidence? One paragraph explaining why you rejected the alternative makes you 10x stronger on cross.
-
- Can a non-CPA understand my methodology? If the jury can’t follow it, they won’t trust it.
-
- Did I do the work, or did counsel do the work for me? Your workpapers should tell the full story.
Final Thoughts
Attorneys want to win. That’s their job.
Your job is to be independent, objective, and ruthlessly honest about the numbers — even when it hurts the party that hired you. In the long run, that’s how you get hired again. Attorneys remember the expert who told them the bad news early in private, not the one who got embarrassed in public.
Bias doesn’t make you a bad expert. Unchecked bias does.
